Phd Thesis Aims at Developing Self-Reporting Systems Based on Chemiluminescence for Tracking of Bond Formation and Cleavage

QUEENSLAND UNIVERSITY OF TECHNOLOGY FACULTY OF SCIENCE SCHOOL OF CHEMISTRY AND PHYSICS

SUBMITTED IN FULFILMENT OF THE REQUIREMENTS FOR THE DEGREE OF

Doctor of Philosophy

Chemiluminescent Self-Reporting Macromolecular Transformation

Fabian R. Bloesser MSc Chemistry

2021

“We scholars like to think science has all the answers, but in the end it’s just a bunch of unprovable nonsense.”

Sorcerio, in Matt Groening’s Disenchantment, S1E3

Statement of Original Authorship The work contained in this thesis has not been previously submitted to meet requirements for an award at this or any other higher education institution. To the best of my knowledge and belief, the thesis contains no material previously published or written by another person except where due reference is made.

21 June 2021

QUT Verified Signature ...... (Fabian R. Bloesser)

Abstract

Self-reporting systems play a major role in the detection and localisation of damages and mechanical stress in materials, the formation and reversion of networks, the detection of drug release as well as the presence of toxins in cells. While a change in colour or fluorescence have been the detection mode of choice in the past, chemiluminescence (CL) systems have attracted increasing interest in the past years, as CL provides a high sensitivity, allows for real-time monitoring and quantification, and does usually not require sophisticated equipment. Therefore, the current study focuses on the development of self-reporting CL systems for the quantification of bond formation and as a tool for on-line kinetic analysis where conventional analyses fail at providing valuable information. The para-fluoro – thiol reaction (PFTR) has recently been reported as an efficient tool for the formation of precision networks.1 While the PFTR forms relatively stable C-S bonds, it also releases a fluoride ion for every bond formed, which can be detected via CL using silyl-protected phenolic dioxetanes.2 Chapter 3 presents a preliminary study of the PFTR investigating its self-propagating properties. A pentafluorobenzyl (PFB) linker was reacted with a set of three structurally different thiols in three solvents of different polarity. It was found that both acidity of the thiol as well as solvent polarity do not only play an important role in the kinetic and efficiency of the reaction, but also that - depending on the exact thiol-solvent combination - the reaction proceeds quantitatively within only a few moments when understoichiometric amounts of base were used. In fact, the fluoride being released during the PFTR is sufficiently basic to deprotonate another thiol, thus propagating the PFTR. The fluoride itself, however, thus formed HF and was removed from the reaction system. These findings pathed the way for an in-depth study of the quantification of network bonds during PFTR via CL in Chapter 4. With the results of the first chapter at hand, the second part of the PhD program addresses the quantification of PFTR events via CL. Therefore, in a first step, a suitable PFB-trilinker was reacted with three different thiols at various concentrations. The fluoride that was released during the PFTR proved capable of cleaving the silyl-ether of a CL probe, thus triggering the emission of light. The integrated emission of all three thiols correlated linearly with the concentration of thiol used, i.e. with the expected conversion, as well as with the conversion as determined by 1H nuclear magnetic resonance (NMR) spectroscopy and liquid chromatography (LC). Moreover, the emission was coherent with the emission of pure TBAF solutions at identical concentrations. Next, a 2000 g·mol-1 polyethylene glycol thiol (PEG-SH) was employed for formation of a three-arm star-polymer. Again, the total CL emission was recorded and, this time, the conversion

vi was back-calculated based on the emission. Here too, the fluoride concentration as determined from CL was in good agreement with the conversion according to 1H NMR and LC. In a last step, PFTR of two bis-thiols was employed for the formation of two distinct networks. The supernatant solutions after network formation were subsequently analysed via CL read-out to obtain the absolute number of PFTR events and therefore network junctions. Critically, the presented method exhibits high sensitivity of down to 5·10-9 mol of fluoride and circumvents the necessity of degrading the networks. The last project of the present thesis employs the CL of phenyl oxalates (POs) to track the unfolding of single-chain nanoparticles (SCNPs). A set of linear polymer chains bearing photo-active ortho-methylbenzaldehyde (o-MBA) moieties as well as fluorophore moieties was synthesised. The o-MBA units were subsequently crosslinked intramolecularly using a previously reported bis-maleimide-PO linker3 and characterised via 1H NMR, diffusion-ordered NMR spectroscopy (DOSY) and size exclusion chromatogra-

phy (SEC) to prove successful folding into SCNPs. Addition of hydrogen peroxide (H2O2) allowed for the targeted degradation of the PO-linker and the subsequent emission of light, which was recorded as a function of time. The SCNPs were then analysed via the methods mention above to conﬁrm complete unfolding. The time-dependant emission data was subsequently employed for a parameter estimation using the PREDICI® software package. Hence, the presented method allows for on-line analysis of the SCNP unfolding and provides a qualitative assessment of the mechanism of SCNP unfolding. In conclusion, the present doctoral thesis established advanced optical read-out and characterisation methods for the in-depth analysis of chemical reaction system, such as quantiﬁcation of reaction events or kinetic analysis, via state-of-the-art chemiluminescence systems.

vii Contents

Contents

Abstract ...... vi Contents ...... viii Publications Included in the Present PhD Research Program ...... x Additional Publications During the Candidature ...... x Acknowledgements ...... xii List of Abbreviations ...... xiii

1 Motivation1

2 Introduction2 2.1 Networks ...... 4 2.1.1 Network Characterisation ...... 7 2.1.2 para-Fluoro – Thiol Reaction ...... 10 2.2 Single-Chain Nanoparticles ...... 14 2.2.1 Photo-Induced Chemistry as an Efﬁcient Ligation Method . . . . . 16 2.2.2 Analysis of Single-Chain Nanoparticles ...... 20 2.3 Self-Reporting Systems ...... 24 2.4 Chemiluminescence ...... 26 2.4.1 1,2-Dioxetanes ...... 28 2.4.2 Peroxyoxalates ...... 32 2.4.3 Luminol ...... 39

3 Self-Propagated para-Fluoro – Thiol Reaction 40 3.1 Abstract ...... 40 3.2 Introduction ...... 41 3.3 Results and Discussion ...... 42 3.4 Conclusion ...... 49

4 Chemiluminescent Read-Out of para-Fluoro – Thiol Reaction Events 50 4.1 Abstract ...... 50 4.2 Introduction ...... 51 4.3 Results and Discussion ...... 52 4.4 Conclusion ...... 58

5 Chemiluminescent Unfolding of Single-Chain Nanoparticles 60 5.1 Abstract ...... 60

viii CONTENTS

5.2 Introduction ...... 61 5.3 Results and Discussion ...... 62 5.4 Conclusion ...... 68

6 General Discussion 69 6.1 Summary and Key Outcomes ...... 69 6.2 Conclusion and Future Perspective ...... 72

7 References 82

8 Appendix 104 8.1 Statements of Contribution ...... 106 8.2 Supporting Information for Chapter 3 ...... 110 8.3 Supporting Information for Chapter 4 ...... 121 8.4 Supporting Information for Chapter 5 ...... 145 List of Figures ...... 174 List of Schemes ...... 181 List of Tables ...... 184

ix Publications Included in the Present PhD Research Program

Publications Included in the Present PhD Research Program

Section 2.4.2:

All Eyes on Visible Light Peroxyoxalate Chemiluminescence Read-Out Systems L. Delafresnaye, F. R. Bloesser, K. B. Kockler, C. W. Schmitt, I. M. Irshadeen, C. Barner-Kowollik, Chem. Eur. J. 2020,26, 114-127.

Chapter 3:

Self-Propagated para-Fluoro – Thiol Reaction F. R. Bloesser, F. Cavalli, C. Barner-Kowollik, L. Barner, Chem. Eur. J. 2019, 25, 10049-10053.

Chapter 4:

Chemiluminescent Read-Out of para-Fluoro – Thiol Reaction Events F. R. Bloesser, F. Cavalli, Sarah L. Walden, L. Barner, C. Barner-Kowollik, Chem. Commun. 2020, 56, 14996-14999.

Chapter 5:

Chemiluminescent Self-Reported Unfolding of Single-Chain Nanoparticles F. R. Bloesser, Sarah L. Walden, Ishrath M. Irshadeen, Lewis C. Chambers, C. Barner- Kowollik, 2020, Chem. Commun. 2021, 57, 5203-5206.

The Statements of Contribution signed by all co-authors are attached to the Appendix.

Additional Publications During the Candidature

[1] Light-fueled dynamic covalent crosslinking of single polymer chains in non-equilibrium states Kodura, D.; Houck, H. A.; Bloesser, F. R.; Goldmann, A. S.; Du Prez, F. E.; Frisch, H.; Barner-Kowollik, C. Chem. Sci. 2021, Advance Article.

[2] 2D Fabrication of Responsive Soft Matter Materials from a Single Photoresist Michalek, L.; Bialas, S.; Walden, S.; Bloesser, F.R.; Frisch, H.; Barner-Kowollik, C. Adv. Funct. Mater. 2020, 30, 2005328.

x CONTENTS

[3] Wavelength Selective Folding of Single Polymer Chains with Different Colors of Visible Light Frisch, H.; Kodura, D.; Bloesser, F. R.; Michalek, L.; Barner-Kowollik, C. Macromol. Rapid Commun. 2020, 41, 1900414. (Special Issue dedicated to 100 Years of Macromolecular Hypothesis, journal cover)

[4] Mapping the Compaction of Discrete Polymer Chains by Size-Exclusion Chro- matography Coupled to High Resolution Mass Spectrometry Nitsche, T.; Steinkoenig, J.; De Bruycker, K.; Bloesser, F. R.; Blanksby, S. J.; Blinco, J. P.; Barner-Kowollik, C. Macromolecules 2019, 52, 2597-2606.

[5] Controlling Chain Coupling and Single Chain Ligation by Two Colours of Visible Light Frisch, H.; Bloesser, F. R.; Barner-Kowollik, C. Angew. Chem. Int. Ed. 2019, 58, 3604-3609.

[6] Fatty Acid-Derived Aliphatic Long Chain Polyethers by a Combination of Catalytic Ester Reduction and ADMET or Thiol-Ene Polymerization Dannecker, P-D.; Biermann, U.; Sink, A.; Bloesser, F.R.; Metzger, J.O.; Meier, M.A.R. Macromol. Chem. Phys. 2018, 1800440.

[7] A Combined Photochemical and Multi-Component Reaction Approach to Precision Oligomers Konrad, W.; Bloesser, F. R.; Wetzel, K. S.; Boukis, A. C.; Meier, M. A. R.; Barner- Kowollik, C. Chem. Eur. J. 2018, 24, 3413-3419.

[8] Photochemistry in Conﬁned Environments for Single Chain Nanoparticle Design Frisch, H:, Menzel, J. P.; Bloesser, F. R.; Marschner, D.; Mundsinger, K.; Barner- Kowollik, C. J. Am. Chem Soc. 2018, 140, 9551-9557. (journal cover)

[9] Polyselenoureas via Multicomponent Polymerizations using Elemental Selenium as Monomer Tuten, B. T.; Bloesser, F. R.; Marshall, D. L.; Michalek, L.; Schmitt, C. W.; Blanksby, S. J.; Barner-Kowollik, C. ACS Macro Lett. 2018, 7, 898-903.

xi Acknowledgements

Acknowledgements

First of all, I am greatly indebted to Prof. Christopher Barner-Kowollik for his excellent guidance and supervision along the way of this exciting journey. Thanks for all the input and the support over the past 3.5 years in good times as in bad times, thanks for the freedom which let me steer my projects and the entire thesis in the direction I felt was right. I would also like to thank Dr. James Blinco for being a great co-supervisor, for all his input and feedback and for being a great head of the soft matter lab. I furthermore acknowledge QUT for key support in the form of a Supervisor Scholarship and HDR Tuition Fee Sponsorship. A special thanks goes to Dr. Laura Delafresnaye (Broqua) for being a great section leader, for fruitful discussions and feedback, for proof-reading this thesis and for listening to all my worries and complaints. I am greatly indebted to her and to Kaz Hosokawa for keeping the lab alive and running. – Merci Beaucoup! – どうもありがとう! Grazie mille to Dr. Federica Cavalli for an excellent collaboration. The project had its ups and downs, but in the end we managed to get what we wanted. In this context also thanks to Prof. Leonie Barner and Dr. Sarah Walden for their support, input and discussions. Another thanks goes to Dr. Hendrik Frisch, an exceptional Post-Doc and section leader, who always had an answer to scientic problems. A special thanks to Dr. Aaron Micallef for his outstanding support regarding any NMR questions and problems and for the trust he put in me operating the spectrometers and experimenting with them on my own. Thanks also to everyone I worked with over the past 3.5 years, everyone who let me be part of their project and of the entire Soft Matter Lab.

However, the past 3.5 years did not only consist of work and science. Thus, I would also like to thank Matthias Van De Walle for the great time we had and for much more to come. Hartelijk bedankt! We’ll see each other again soon. Tausend Dank! – Thanks so much to our covfefe group for (almost) daily coffee breaks and chats. I miss you and I’m sure Gerbino’s is going broke without us. Furthermore, I’d like to acknowledge the SEF HDR student society as well as Prof. Andrew Bradley and Prof. Esa Jaatinen for their commitment to make students’ lives easier.

Last, but not least I want to thank my family and my friends in Germany, Australia, and anywhere else in the world who accompanied me over the past years, who had my back and supported me in my journey!

xii CONTENTS

List of Abbreviations

β-CD β-cyclodextrin a.t. ambient temperature ACN acetonitrile AF4 asymmetrical ﬁeld ﬂow fractionation AFM atomic force microscopy AIBN azobisisobutyronitrile AT adamantyl thiol ATRP atom-transfer radical-polymerisation BDT 1,4-benzendimethane thiol BT benzyl thiol BTA benzene-1,3,5-tricarboxamide CH cyclohexane CIEEL chemically initiated electron exchange luminescence CL chemiluminescence CT charge transfer CTIL charge-transfer induced luminescence CuAAC Copper-catalysed Azide-Alkyne Cycloaddition DA Diels-Alder DBU 1,8-Diazabicyclo[5.4.0]undec-7-ene DCM dichloromethane DDT dodecanethiol

dH hydrodynamic diameter Ð dispersity DL limit of detection DLS dynamic light scattering DMF N,N-dimethylformamide DMSO dimethylsulphoxide DNA deoxyribonucleic acid DNPO bis-(2,4-dinitrophenyl)oxalate DODT triethylene glycol dithiol DOSY diffusion-ordered NMR spectroscopy DPA 9,10-diphenyl anthracene DPM diphenylmethane EA ethyl acetate

xiii List of Abbreviations

EBT electron back-transfer EP-1 3-(1,4-dihydro-1,4-epidioxy-4-methyl-1-naphthyl)propionic acid ESI-MS electrospray ionisation mass spectrometry ET electron transfer EXSY exchange NMR spectroscopy FRP free radical polymerisation FTIR Fourier-transform infrared spectroscopy

H2O2 hydrogen peroxide HCl hydrochloric acid HDA hetero-Diels-Alder HEI high energy intermediate IC internal conversion IR infrared ISC intersystem crossing LC liquid chromatography LUMO lowest unoccupied molecular orbital MALLS multiangle laser light scattering MBM 4-methoxybenzyl mercaptan MDCPO bis(2,6-dichloro-4-N-maleimido) phenyl oxalate MMA methyl methacrylate Mn number-average molecular weight MRI magnetic resonance imaging MS mass spectrometry mTrEGT methoxy triethylene glycol thiol NDS network disassembly spectroscopy NITEC nitrile imine mediated tetrazole-ene cycloaddition NMR nuclear magnetic resonance NOESY nuclear Overhauser effect NMR spectroscopy o-MBA ortho-methylbenzaldehyde PEG polyethylene glycol PEG-SH polyethylene glycol thiol PFB pentafluorobenzyl PFP pentafluorophenyl PFPMA pentafluorophenyl methacrylate PFS pentafluorostyrene PFTR para-fluoro – thiol reaction

xiv CONTENTS

PhOx phenyloxalate PMA poly(methyl acrylate) PMMA poly(methyl methacrylate) ppm parts per million PO phenyl oxalate PO-CL peroxyoxalate chemiluminescence PS polystyrene Qm swelling ratio RAFT reversible addition-fragmentation chain transfer RDRP reversible-deactivation radical polymerisation rg radius of gyration rH hydrodynamic radius dRI differential refractive index ROMP ring-opening metathesis polymerisation ROP ring-opening polymerisation ROS reactive oxygen species

S0 electronic ground state S1 excited singlet state SANS small-angle neutron scattering SAXS small-angle X-ray scattering SCNP single-chain nanoparticle SEC size exclusion chromatography

SNAr aromatic nucleophilic substitution StyPy styryl pyrene

T1 excited triplet state T 2 spin-spin relaxation time TAD triazolinedione TBA tetrabutylammonium TBABr tetrabutylammonium bromide TBAF tetrabutylammonium ﬂuoride TBAOH tetrabutylammonium hydroxide TBD 1,5,7-triaza-bicyclo-[4.4.0]dec-5-ene TCPO bis-(2,4,6-trichlorophenyl) oxalate TEM transmission electron microscopy T g glass-transition temperature THF tetrahydrofuran

xv List of Abbreviations

TMS tetramethylsilane ToF-SIMS time-of-ﬂight secondary ion mass spectrometry TP thiophenol TPE tetraphenylethylene UPy ureido-pyrimidinone uSCNP unfolded SCNP UV/Vis ultraviolet/visible XIC extracted ion chromatogram XPS X-ray photoelectron spectroscopy

xvi Motivation

1 Motivation

In-depth analysis of polymers is usually exceptionally complex and highly time-consuming as polymerisations almost exclusively yield molecular weight distributions rather than specific molar masses. Whenever more than one monomer is employed for a polymerisation, the monomers can be incorporated in an alternating fashion, randomly along the polymer backbone or as distinct blocks, causing different properties of the polymer. Moreover, depending on the starting materials and the procedure, various macromolecular architectures can be achieved, such as macrocycles, star- or brush- copolymers. Commonly, such architectures are characterised via a combination of e.g. nuclear magnetic resonance (NMR) techniques, size exclusion chromatography (SEC), high-resolution mass spectrometry (MS) or electron microscopy, since each of these techniques provides a piece of information.4,5 However, each of the aforementioned analytical techniques requires costly instrumentation and extensive expertise. In addition to the distinct conformation of a polymer (i.e. stretched, spherical coil, rod-like, etc.) or its specific architecture, the transition from one conformational state (or architecture) to another is of high interest. While conventional analyses can provide excellent information on the distinct conformation of a polymer, they only allow for limited on-line tracking of conformational changes due to long measurement times or insufficient signal-to-noise ratio. The present thesis aims at expanding the toolbox of analytical techniques by establishing self-reporting chemiluminescence (CL) systems that are straightforward to execute and easy to interpret. Such CL-based self-reporting read-out systems are therefore expected to be of significant importance for the characterisation of insoluble macromolecular architectures like networks, and to provide excellent conditions for the in-depth and on-line investigations of macromolecular transformations like folding or unfolding of polymers. Passing on to the real world, many processed materials are either insoluble or lose their characteristics upon dissolution thus limiting the utilisation of solution based analysis techniques like NMR, SEC or MS for the characterisation of such materials. Furthermore, materials like adhesives, coatings or composite materials employed in the automotive industry are often exposed to changing and occasionally harsh conditions such as severe

1 temperature changes, extreme dryness or humidity, or UV irradiation due to sunlight. While the topic of smart materials with self-healing properties has been prospering over the past years, it still exhibits a major drawback. Many self-healing processes proceed autonomously and therefore unnoticed, and even more self-healing processes occur irreversibly, i.e. once a certain area of a material had been damaged and subsequently healed, the self-healing properties are depleted within the area of damage and further damage cannot be healed.6–10 Reporting such damages is therefore key to prevent material failures. Colour changes and fluorescence were proven to be suitable tracers for mechanical, thermal, light or pH stimuli in materials on a laboratory scale.10 However, colour changes are only qualitative, not quantitative and fluorescence usually is barely visible to the naked human eye, especially in solid matter. Here, CL is a promising candidate for the tracking of such stimuli in real world materials, as it is usually visible for the naked human eye and can be quantified in terms of emitted photons. Besides the field of macromolecular architectures and processed materials, chemical transformations of in vitro cell cultures are challenging to characterise on molecular levels. In this field, however, self-reporting fluorescent probes are widely employed, highlighting the importance of self-reporting systems as a means of characterisation.11–13 Here, Shabat and co-workers pioneered the use of various CL probes and proved the critical importance of CL for self-reporting systems in vitro and even in vivo.14–21 Given the significant success of CL in academic applications so far, the present thesis submits that CL will be of critical importance for the monitoring of real world applications in both biological systems as well as materials science.

2 Introduction

Over the past one hundred years, the synthesis of precise macromolecular architectures has attracted increasing attention from researchers in various ﬁelds. In particular, micelles are commonly employed for medical applications and single-chain nanoparticles (SCNPs) represent promising materials for catalysis or medical applications.22–29 Polymeric networks ﬁnd application as adhesives, vulcanised rubbers, gels or sor- bents.30 However, analysis of macromolecular structures can be tedious and time-consuming,

2 Introduction

as polymerisations practically solely yield disperse molecular weights and a number of different polymerisation products. The in situ formation of SCNPs as well as their unfolding is often characterised by a multitude of techniques. The probably most commonly employed characterisation method for covalently collapsed SCNPs is SEC. An increase in retention time is observed for SCNPs compared to the according linear polymer, 31,32 indicating a decrease in hydrodynamic radius (r H). Light scattering experiments, such as dynamic light scattering (DLS) and multiangle laser light scattering (MALLS) yield similar results. However, Tuten et al. pioneered the combination of SEC with in-line MALLS measurement in order obtain an improved method for SCNP characterisation.33 1H NMR spectroscopy is used in order to show a decrease of functional cross-linkable groups. Moreover, an increase of the diffusion coefficient after SCNP collapse can be observed by diffusion-ordered NMR spectroscopy (DOSY).34 Recently, the group of Barner-Kowollik reported the successful characterisation of methyl methacrylate (MMA)- based SCNPs via electrospray ionisation mass spectrometry (ESI-MS).35,36 Yet, neither of the aforementioned methods allows for the in situ tracking of SCNP formation or unfolding or the non-destructive analysis of a processed material. Little progress has been made in order to provide methodology for in situ tracking. Still, many self-reporting systems for folding employing fluorescence have been reported.37–40 Networks, however, lack the ability to dissolve and therefore require different characterisation methods. X-ray photoelectron spectroscopy (XPS) is employed in order to determine the elemental composition and binding energies of the corresponding elements. An inside in the chemical structure of networks is provided by time-of-flight secondary ion mass spectrometry (ToF-SIMS) by sputtering the surface of the sample and detecting ionised fragments of the network. ToF-SIMS, however, is an at least partially destructive characterisation method. Electron microscopy allows for insight in the topological structure of networks. To the best of the candidate’s knowledge, emission of photons via chemiluminescence has not yet been reported as a self-reporting system for macromolecular transformations. Macromolecular architectures, such as SCNPs or polymeric networks, cover a wide range of possible applications and diverse characterisation methods have been developed (vide supra). Yet, especially for in vivo applications and processed materials, on-line tracking of macromolecular transformations is highly challenging and only little progress has been made. While substantial efforts have been made in the synthesis, characterisation of materials is often only performed on a bulk or material level and little to no information is gained on a molecular level. Chemiluminescent self-reporting systems would allow for facile on-line tracking of folding and unfolding events by emission of

3 Networks

photons. Thus, the present PhD thesis aims at developing self-reporting systems based on chemiluminescence for tracking of bond formation and cleavage. Therefore, a mechanistic connection between formation and cleavage of covalent bonds, respectively, and a trigger for chemiluminescence needs to be established. After establishing sufﬁciently effective self-reporting systems, the novel method can be employed for mechanistic studies, such as SCNP folding and unfolding or network formation. Thus, a readout of the number of ligation points or bond cleavage as a function of time can be obtained.

2.1 Networks

Networks are macromolecular architectures that exhibit a high degree of crosslinking via covalent or non-covalent interactions, “in which essentially each constitutional unit is connected to each other constitutional unit and to the macroscopic phase boundary by many paths through the structure”.41 These distinct features give rise to properties that are specifically characteristic for networks, e.g. deformation elasticity and the ability to swell in a solvent rather than dissolve. Once a network is swollen, it is referred to as hydrogel (if the swelling agent is water) or organogel (if the swelling agent is an organic solvent). Networks and gels find widespread applications in vulcanised rubbers,42–44 adhesives,45–47 superabsorbents,48 drug delivery,48,49 scaffolds for tissue engineering,50–52 dental fillings48 and many more.47–49,51,53 The exact properties of a network are dependant on (i) the chemical structure, i.e. the choice of monomer and (ii) the topology of the network. Figure 2.1A depicts a section

of an ideal network, which is characterised by the number of elastic chains νel, the number of junctions (crosslinks) µel and the functionality of the junction f. These three parameters correlate via: 2 µ = ν (2.1) el f el The dimensionless cycle rank ξ describes the number of independent cycles in the

network and can be determined from νel and µel according to equation 2.2:

ξ = νel – µel + 1 (2.2)

-1 Moreover, the (average) molecular weight of elastic chains Mc in g·mol is calculated from: V0NA Mc = ρ (2.3) νel -1 with the Avogadro number NA in mol , the volume of the network V 0 in mL and the

4 Introduction

density ρ of the network in g·mL-1. With all the above parameters deﬁned, one can combine equations 2.1 to 2.3 into a single equation, correlating all the parameters:

2 ρV N 2 ξ = ν · 1 – = 0 A · 1 – (2.4) f Mc f

Figure 2.1: A) Representation of an ideal network showing no inhomogeneities. νel and textitµel represent the number of elastic chains and number of junctions (crosslinks), respectively. In the present example, νel is equal to twelve, µel is equal to nine and the functionality f of µel is four. The (average) molecular weight of the elastic chains is described by Mc. The cycle rank ξ is four. B) The network exhibits various defects, such as a primary loop (1), a dangling chain end (2) and two entangled chains (3).

To understand the macroscopic elasticity of networks, two major models have been proposed,54 the affine model by Kuhn55 as well as Wall and Flory,56 and the phantom model by James and Guth.57 The affine model assumes that the crosslinks are tightly locked within the network, while the single chains within the sample move independently and affinely to the entire network. The phantom model, on the other hand, suggests movement of the junctions while the network chains rest in place. The two models are considered the limiting cases and real networks are situated in-between these limitations. Moreover, real networks usually are far from ideal and exhibit defects, causing inhomogeneities and changing the overall network properties. State-of-the-art models, such as the constrained-junction model, tube model or slip-tube model, therefore attempt to bridge the gap between the two case-limiting models.58 A selection of basic network defects is depicted in Figure 2.1B. When both ends of a network chain are linked to the same junction, a primary loop is formed (1). Such loops, as well as loose chain ends (2) cause less dense networks and therefore larger mesh sizes. Conversely, entanglements of two or more chains (3) can be considered junctions, limiting the elasticity of chain segments.

5 Networks

Non-covalent networks usually consist of polymer chains that are crosslinked via intermolecular interactions such as hydrogen bonding, ionic interaction, metal ion complexation, π-π stacking and host-guest interactions.59,60 Networks formed via physical interactions such as the above commonly respond to external stimuli such as temperature, pH, pressure and others. The synthesis of covalent or chemical networks involves the polymerisation of monomers via chain- or step-growth protocols with multifunctional crosslinkers or the post-polymerisation crosslinking of polymer precursors. During network formation via free radical polymerisation (FRP), chain-growth leads to the formation of high molecular weight polymers at early stages and a broad size distribution due to chain termination and transfer at an early stage. Furthermore, the high dilution of active radicals in a solution of monomer favours cyclisation and nanogel formation over intermolecular crosslinking. Only at later stages, when more monomer has been consumed, do these nanogels interconnect, causing gelation on a macroscopic level. This rather uncontrolled process leads to an inhomogeneous network exhibiting domains with highly variable crosslinking densities. Nevertheless, FRP network formation is often applied due to its straightforward ease of use.49,61 As reversible-deactivation radical polymerisation (RDRP) techniques provide more control over molecular weights and dispersities in the synthesis of linear polymers, RDRP protocols have been widely employed for the preparation of chain- growth networks. While more control over homogeneity of network topologies via RDRP has been suggested, no hard evidence for this claim was demonstrated yet.62–65

In a step-growth network formation, an A-B type monomer is reacted with an Af crosslinker (f > 2) in the presence of a B-B type monomer. Such step-growth leads to a network topology with a high degree of heterogeneity and crosslinking in a too early or too late stage can be even more disadvantageous to the structure of the network. Still, network formation via step-growth polymerisation is commonly employed for the synthesis of epoxy-resins, phenol formaldehyde resins, alkyd resins, polyurethanes, polysiloxanes, and others.66,67 Moreover, linear polymer chains can be crosslinked via pendant groups positioned randomly along the polymer backbone, also referred to as ’curing’. The most common examples for curing are the vulcanisation of rubber with elemental sulfur and copolymeri- sation of pendant double bonds with styrene or MMA. However, more recent approaches employ RDRP protocols for the synthesis of well-deﬁned linear precursor polymers bearing crosslinkable moieties along the backbone. The use of narrow-dispersity linear precursors allows for the preparation of networks with less inhomogeneities.68,69 Additionally, photochemical crosslinking introduces the possibility of spatio-temporal con-

6 Introduction

trol.68,70–73 When reacting α,ω-homotelechetic polymers with crosslinkers (f > 2) via efficient ligation methods such as Copper-catalysed Azide-Alkyne Cycloaddition (CuAAC), Diels-Alder (DA) reactions, photo-ligation techniques, and others, networks with an even more regular structure can be obtained.45,74–81 Cavalli et al. recently reported the synthesis of precision networks via the hihgly efficient para-fluoro – thiol reaction (PFTR).1 In the current thesis, similar networks are

prepared by reacting a B-B type thiol with a A3-type pentaﬂuorobenzyl (PFB)-linker and the network junctions are subsequently quantiﬁed. An overview of the PFTR is provided in Section 2.1.2.

2.1.1 Network Characterisation

As mentioned above, contrary to linear and branched polymers, networks do not dissolve and therefore conventional polymer analytics such as 1H NMR, SEC or MS are of little to no use when it comes to the characterisation of networks. Routine analysis of networks usually involves the characterisation of bulk properties such as the swelling ratio or rubber elasticity. The swelling ratio (Qm) is determined by immersing the network into an appropriate solvent and letting solvent molecules diffuse into the network and thereby expand it, until an equilibrium is reached. The dimensionless Qm is subsequently calculated from the mass of the dry network mdry and the mass of the swollen network mswollen: mswollen Qm = (2.5) mdry

The swelling ratio is inversely proportional to the concentration of crosslinks νc according to the Flory-Rehner equation:

2 Vmol,s · ρn –[ln(1 – vn) + vn + χs,n · vn] νc = = 1 (2.6) Mc 3 vn vn – 2 f where 1 vn = (2.7) [(Q – 1) 1 + 1 ] · ρ m ρs ρn n -1 and V mol,s is the molar volume of the solvent in mL·mol , ρn and ρs is the density -1 of the network or the solvent in g·mL , respectively, vn is the volume fraction of the network, and χs,n is the dimensionless Flory-Huggins interaction parameter of solvent and network.82 Unfortunately, the equation only applies to ideal networks without any defects.

7 Networks

In rheology, the network’s response to externally applied force (stress and strain) is measured. In viscoelastic materials, the storage modulus G’(ω) describes the energy that is stored (elastic fraction) and the loss modulus G”(ω) describes the dissipated energy (viscous fraction). During network formation, early reaction stages are dominated by the viscous portion (G”(ω) > G’(ω)), whereas after the start of gelation, the elastic portion dominates (G’(ω) > G”(ω)). Moreover, the exact start of gelation can be determined via G’(ω) = G”(ω). The expected storage modulus can be calculated for both the affine and the phantom model via: 2h G 0 = k T ν 1 – (2.8) affine B c f where kB is the Boltzmann constant, T the temperature and h is a constant that depends on the employed model. For the case limiting models, i.e. the affine model or the phantom model, h is equal to zero or one, respectively.83 However, also the elastic modulus is affected by network defects, leading to deviations between theoretical and experimental values.84,85 While NMR spectroscopy is frequently employed for the characterisation of small molecules and polymers, and a multitude of correlation spectroscopic methods allow for the determination of the exact structure of a molecule, the technique is highly dependant on the analyte being soluble, as dipole-dipole interactions between nuclei as well as quadrupolar interactions are averaged to zero in solution, and therefore simplify the spectra. In solid-state NMR, however, due to these interactions, the resonances experience line-broadening to such extend that resonances are indistinguishable or (1) disappear completely. First-order quadrupolar coupling ωQ can be described by the following equation: (1) 3πCQ 1 ω (Θ) = × (3cos2θ – 1) (2.9) Q I (2I – 1) 2 Q where Θis the molecular orientation, CQ is the quadrupolar coupling constant, I is the spin of the nucleus, and θis the angle between the axis of the external magnetic field and the axis around which the sample is rotated. For θ= 54°44’ - the so-called magic angle - the entire term becomes zero, thus at least eliminating quadrupolar interactions from the spectra. Nonetheless, the limited chemical shift range of 1H NMR and the multiplicity of resonances due to coupling, the identification in 1H NMR spectra is still eminently limited. 13C NMR, on the other hand, does not show homo-coupling, is usually proton-decoupled and exhibits a chemical shift range about 20 times that of 1H NMR, thus facilitating the interpretation of spectra and the assignment of resonances.86,87 Solid-state 13C NMR can therefore be an excellent tool for the analysis of insoluble

8 Introduction

networks and was exploited as such on a regular basis.88–91 Moreover, the transverse

relaxation decay or spin-spin relaxation time (T 2) is greatly affected by molecular motion and thus, T 2 decreases with decreasing molecular motion. Yet, in order to observe any molecular motion, the relaxation experiments have to be conducted above the glass-transition temperature (T g) of the polymer. The molecular motion of network strands depend on i) the length or molecular weight of the strands Mc, ii) the crosslinking density µc, and iii) the nature and number of defects present in the network. While longer network strands, less junctions and defects like primary loops and lose ends

increase the chain mobility, thus increasing T 2, shorter network strands, more junctions and defects like entanglements decrease the mobility of chains, thus decreasing T 2. Furthermore, chain segments that are closer to network junctions exhibit less mobility than chain segments further away from junctions.92–95 The group of Wilhelm performed

T 2 relaxation experiments on swollen networks to obtain information on the topology as well as crosslinking density of networks synthesised via reversible addition-fragmentation chain transfer (RAFT) compared to networks synthesised via FRP.64,96 The same group also investigated the relaxation behaviour and the mobility of water molecules and alkali

ions in swollen polyacrylate hydrogels by T 2 relaxation experiments and DOSY NMR, respectively. They found increasing diffusion coefficients and decreasing relaxation rates times with an increasing swelling degree, thus directly linking the structure and dynamics of the network to interactions between ions, water and the network itself.97,98 Differential scanning calorimetry and thermogravimetric analysis provide information about the thermal properties and potentially presence of certain structural motifs of the network.62,70,99,100 Cavalli et al. employed XPS and ToF-SIMS to confirm the chemical composition of their network and the presence of structural motifs,1 while Clark et al. monitored the conversion of crosslinkers during network formation via on-line Fourier-transform infrared spectroscopy (FTIR).101 The group of Johnson pioneered an experimental method for the quantification of primary loops and dangling ends within a network.102 In network disassembly spectroscopy (NDS), the network is disassembled at selected sites to form distinct products that can readily be identified by common analytical techniques. In their approach,

Johnson and co-workers prepared a norbonene-functionalised A2-type monomer and a trifunctional tetrazine linker B3 that effortlessly form a network via DA chemistry. The A2- type monomer furthermore exhibits an ester-functionality that, upon cleavage, produces two distinct fragments, a short one (S) and a long one (L) (refer to Figure 2.2). At hypothetical 100% conversion of all functional groups, there are four feasible combinations in which B3 can have reacted: B3 can have reacted with three long chain ends (LLL) of A2,

9 Networks

one short and two long chain ends (SLL), two short and one long chain end (SSL) or

three short chain ends (SSS). Importantly, as A2 consists of an S and an L segment, loop formation is only possible for SLL and SSL, not for SSS or LLL. After network formation, the DA product was oxidised to form a highly fluorescent moiety that facilitates detection and thus quantification. Subsequently, the network was degraded to form the four star-like fragments LLL, SLL, SSL, and LLL as well as species with unreacted functionalities (refer to Figure 2.2). The thus formed network fragments were analysed via liquid chromatography (LC)-MS and quantified. Based on the probabilities to form either of the species, as well as the ratio of SLL to LLL (or the ratio of SSL to SSS), the fraction of primary loops can be determined. They furthermore performed network formation at varying monomer concentrations to investigate the effect of monomer concentration on network and primary loop formation. Interestingly, even the highest investigated starting concentration of 80 mM exhibited 9% primary loops, and starting concentrations below 50 mM only produced soluble fractions.102 While the work of Johnson and co-workers constitutes an elegant example for the quantification of network defects, it requires high synthetic efforts, complex analysis and most importantly degradation of the network. The development of a more facile approach for the quantification of network junctions and defects without the necessity to degrade the material is thus imperative.

2.1.2 para-Fluoro – Thiol Reaction

Given its high abundance on Earth, it appears extraordinary that fluorine is hardly incorporated into natural organic molecules. However, considering the high bonding energy of fluorine, it is not surprising that fluorine-chemistry still is a substantial challenge for most chemists. While aliphatic fluorocarbons are almost unreactive due to the high electronegativity of the fluorine atom and the resulting high binding energy between the fluorine and the carbon, perfluoroarenes display an enhanced reactivity especially towards aromatic nucleophilic substitutions (SNAr’s). Due to persubstitution with fluorine atoms, the electron density of aromatic system is drastically decreased, making it an 103 excellent target in SNAr reactions. The earliest examples of SNAr’s on perfluorinated aryl compounds date back to the 1950s, and soon hydroxides, alkoxides, amines and even alkyl lithium compounds were reported as suitable nucleophiles. However, these nucleophiles usually required exceedingly harsh reaction conditions, such as strong bases and high temperatures over long times.104–112 In contrast, thiols are easily deprotonated by mild organic bases and the corresponding thiolates readily react at ambient temperature (a.t.) providing high yields.113,114

10 Introduction λ Product # varies with n 3 λ yields a network in 3 and B 2 , for a trifunctional network; λ , respectively. The number of primary loops is captured in loop and p ideal fraction of loop junctions, n , which is assumed to be constant.) The probabilities for formation of each vs. 3 , 3 λ monomer. Placement of a degradable group (orange star) at a noncentral position 2 . (D) Plot of the loop ratio, 3 λ . (A) Schematic depiction of an A backbone leads to an S chain (blue) and an L chain (black) after cleavage. (B) End-linking of A NDS 2 Figure 2.2: along the A LLL junctions. (C) Disassembly ofsize, the or network relative yields mass, productstrifunctional whose does product masses not at depend include ideal on thethe or the ratio mass junction loop [LLL]:[SLL] source junctions of or of are B thataccording listed product. to as Eq. ( p 1. Fully looped “dumbbell” polymers have loop-linear-loop architectures. Reprinted from [102]. Copyright 2012 National Academy of Sciences of the USA. which each network junction is unique in terms of the orientation of S and L chains. Primary loops (asterisks) cannot reside at SSS or

11 Networks

F F SH F F base + F F F + F δ S δ−F

SH H H base +

H δ− H H δ+

χH = 2.2; χC = 2.6; χF = 4.0

Scheme 2.1: General reaction scheme of an SNAr between a thiol and a PFB moiety.

When pentafluorophenyl (PFP) or PFB moieties are employed for SNAr reactions, regioselectivity towards the para-position can be observed, hence the term para-fluoro – thiol reaction (PFTR). An explanation for this regioselectivity is provided by Kvícala˘ et al. (refer to Scheme 2.2). When a nucleophile attacks the carbon of a PFB moiety in para-position relative to substituent Y, a mesomeric structure of the transition state can be postulated bearing the negative charge in α-position to the substituent Y. If the substituent Y can thus stabilise the negative charge at the α-carbon, the para-position is preferred over the meta-position. This applies accordingly to nucleophilic attack in the ortho-position. If the substituent Y is not able to stabilise the charge in α-position, substitution in meta-position to Y is preferred.115 The PFTR has attracted increasing interest especially in polymer chemistry, as it provides a highly efficient and versatile tool for polymer ligation as highlighted in several review articles.113,114 Becer et al. copolymerised pentafluorostyrene (PFS) in a RDRP and subsequently functionalised the obtained polymer via quantitative substitution of the para-fluorine atom with a protected thio-glucose derivative. The reaction was performed employing the mild base trimethylamine and a relatively short reaction time of only four hours.116 Kang and co-workers prepared a poly(ε-caprolactone)-block-PFS copolymer with an alkyne end-group by subsequent ring-opening and atom-transfer radical-polymerisation (ATRP). After transforming the bromine end-group of the ATRP initiator into an azide, a macrocycle was formed via CuAAC. Finally, the PFS moiety was reacted with polyethylene bearing a thiol end-group in order to provide jellyfish- shaped block-copolymers.117 Varadharajan et al. investigated the selective functionalisation of two pentafluorophenyl moieties in a block-copolymer backbone, namely pentafluorophenyl methacrylate (PFPMA) and PFS. They found that PFPMA readily

12 Introduction

Nu F Nu F Nu F Nu F F F F F F F F F

F F F F F F F F Y Y Y Y Nu F Nu F F

F F Y

F F F F F F F F F F F F Nu Nu Nu Nu

F F F F F F F F Y Y Y Y

Scheme 2.2: Depending on the nature of the substituent Y, SNAr on a PFB moiety takes place in either para- (or ortho) or meta-position. The scheme is adapted from [115].

reacts with amines in order to form amides, contrary to PFS. Subsequently, the remain- ing PFS moieties selectively reacted with thiols in a PFTR. Unfortunately however, the reverse pathway of first performing PFTR and subsequent formation of amides was not viable, as PFPMA would compete with PFS in the PFTR.118 In a similar approach, Turgut et al. examined the orthogonality of PFTR and thiol-ene reactions.119 Therefore, they synthesised a block-copolymer exhibiting PFS and allyl methacrylate units. By carefully choosing the reaction conditions, either the allyl moiety or the PFS moiety was chemoselectively addressed for addition of a thiol. When employing basic conditions, a nucleophilic substitution on the PFS moiety occurs, whereas after addition of a radical starter, the thiol-ene reaction takes place. Moreover, it was shown that the order of addition could also be reversed with the exact same results and the reaction be performed as one-pot synthesis. Furthermore, Turgut et al. pioneered the PFTR in aqueous conditions on poly(N,N-dimethylacrylamide)-co-PFS. PFTR of the prepared polymer and mercaptoethanol readily occurred at pH 11 and above as could be evidenced by19F NMR. Next, the polymer was grafted to Petri dishes and subsequently functionalised with a cysteine-containing peptide, thus allowing for the grafting of peptides via PFTR.120 Recently, Cavalli et al. introduced a novel route for the synthesis of precision networks based on PFTR. By employing a linker molecule exhibiting three pentafluorobenzyl moieties along with a bifunctional thiol derivative, a highly defined network could be synthesised as was shown by XPS and ToF-SIMS.1 In a separate study, Cavalli et al. performed a kinetic study on the PFTR of various benzylic and aliphatic as well as

13 Single-Chain Nanoparticles polymeric thiols, supported by Monte Carlo modelling.121 Due to its ease of use and high efficiency, Engelke et al. employed the PFTR as a crosslinking reaction in their in-depth study on the compaction of SCNPs as described in Section 2.2.2.122,123 The starting materials for a PFTR are oftentimes commercially available or easy to synthesise. PFTR shows great versatility, rapid reaction times ranging from seconds to minutes (sometimes hours), high yields and orthogonality to many other organic chemistry reactions. PFTR has therefore attracted significant attention from chemists over the past years, however, little to no attention has been paid to the side product of a PFTR, the fluoride ion. The current thesis will focus on performing preliminary kinetic studies on the PFTR, and subsequently, pioneer a CL characterisation approach that is not based on the PFTR product, but rather the fluoride that is released during the reaction.

2.2 Single-Chain Nanoparticles

For the past one hundred years, chemists have developed a multitude of techniques to gain structural control. Today, a vast toolbox of synthetic methods allows for the manipulation of primary structures and thus enabling chemists to build macromolecular architectures such as brush and comb-like polymers, macro-cycles, star polymers, den- drimers and many more.24 In fact, carefully designed synthesis nowadays even allows for the preparation of monodisperse macromolecules with absolute control over sequence of repeating units thus advancing the mimicking of natural monodisperse polymers such as deoxyribonucleic acid (DNA).124,125 The ﬁrst mention of intramolecularly crosslinked polymers dates back to 1962 describing the difference between inter- and intramolecular crosslinking at different concentrations.126 Hence, the ﬁeld of single-chain nanoparticles (SCNPs) emerged, attempting to emulate the secondary and tertiary structure of proteins, thus enabling researchers to prepare highly selective, tailor-made nanoreactors and catalytic sites. While a plethora of ligation methods for the compaction of polymers to SCNPs has been reported, ranging from coordinative and host-guest interactions to hydrogen bonds to covalent bonds and including both organic and inorganic chemistry, only little progress has been made to gain control over the secondary and tertiary structure of synthetic polymers and the precision of their natural paragons is still unrivalled. Potential applications of SCNPs range from catalysis and nanoreactors to nanomedicine to sensors.25,32,127–129 An early example of covalent bond crosslinking in SCNPs is the Friedel-Crafts alkylation of narrow-dispersity polystyrene (PS) by Antonietti et al.130 By using para-

14 Introduction

bis(chloromethyl)benzene as an external crosslinker at high dilution, these authors were able to achieve of up to 10%. In a similar manner, Davankov et al. employed chlorodimethyl ether for the partial chloromethylation of PS and subsequently crosslinked the PS intramolecularly via Friedel-Crafts alkylation without the necessity of an external crosslinker. The thus obtained SCNPs exhibited a high crosslinking density as well as porosity as demonstrated by Argon adsorption-desorption measurements, hence the term ’nanosponges’.131,132 The groups of Loinaz and Pomposo utilised the CuAAC ’click’ reaction for the formation of SCNPs on various polymer backbones.27,34,133,134 In their approach, the linear polymers exhibited either both the alkyne and the azide functionality and were crosslinked without an external crosslinker, or only exhibited the azide functionality and were crosslinked via an external bifunctional alkyne crosslinking agent. By incorporating Gd(III) ions into the crosslinker, a rather rigid environment of the Gd(III) ion is created upon folding of the SCNPs, which provides enhanced relaxitivity in magnetic resonance imaging (MRI). The thus formed SCNP represents a promising candidate as an MRI contrast agent.27 Hydrogen-bonding as a method for the collapse of linear polymers to form SCNPs has been extensively studied on a multitude of structural motifs and shows good resemblance of the stabilising interaction within the α-helices of natural proteins. Prime examples of SCNP formation via hydrogen-bonding are the association of the Hamilton’s wedge and cyanuric acid derivatives,135,136 the ureido-pyrimidinone (UPy) dimerisation137 and the helix-like self-aggregation of benzene-1,3,5-tricarboxamide (BTA).26,138,139 Hosono et al. synthesised an ABA triblock copolymer, bearing UPy moieties in the A blocks and BTA moieties in the B block. After thermal treatment and photoirradiation, the block copolymer folded in a two-step fashion. Via variable NMR measurements, the group was able to elegantly show that the two moieties fold orthogonally, although both moieties fold via hydrogen-bonding.140 By installing acetoacetoxy groups in the polymer backbone, Sanchez-Sanchez et al. were able to collapse SCNPs via coordination around Cu(II) species.29 Mavila et al. prepared poly(cycloocta-1,5-diene) via ring-opening metathesis polymerisation (ROMP). When chlorobis(ethylene)rhodium(I) dimer was added, the Rh(I) species coordinated to the residual double bonds in the polymer backbone, thus collapsing it to form SCNPs.141,142 In a joint project, the groups of Barner-Kowollik and Roesky investigated the formation of SCNPs via complexation of transition metals and lanthanides with phenanthroline, triphenylphosphine or benzoic acid bearing polymers.143–145 Moreover, complexation of a phosphine bearing PS with Pt(II) ions led to the formation of SCNPs

15 Single-Chain Nanoparticles

that were easily used as recyclable homogeneous catalysts for the amination of allyl alcohol.28 One of the most elegant folding procedures was reported by Chao et al. who synthesised a poly(oxanorbornene anhydride-co-cyclooctadiene) via ROMP and folded their linear precursor in three subsequent steps. In a ﬁrst step, the anhydride moieties along the backbone were partially functionalised with aniline tetramer units, introducing electroactive properties to the polymer and allowing for a ﬁrst compaction step via π- πstacking of the aniline tetramer units. Next, residual anhydride segments were reacted with p-aminoaniline causing in further compaction of the polymer. Last, cyclooctadiene moieties were reacted with a dithiol in a thiol-ene reaction, resulting in a total compaction of 70% compared to the linear parent polymer.146 In addition to the various techniques mentioned above, light-gated folding of SCNPs has attracted increasing interest of chemists over the past decade and will therefore be discussed in the following section.

2.2.1 Photo-Induced Chemistry as an Efﬁcient Ligation Method

Nowaday, light-gated chemical reactions play an important role in macromolecular as well as biomolecular chemistry as they allow for catalyst-free ligation under mild conditions. Moreover, they proved to be highly efﬁcient, provide spatio-temporal control, and - in some cases - are photoreversible. Commercially available light sources and even sunlight provide a pathway for environmentally benign and sustainable reaction pathways. Upon absorption of a photon, a molecule is usually transferred from its electronic

ground state (S0) to a vibrational state of its excited singlet state (S1) followed by non- radiative decay to the lowest vibrational level of S1. The remainder of the previously absorbed energy is then emitted as ﬂuorescence. The mechanism of ﬂuorescence is

depicted in Figure 2.3. In some cases where there is a excited triplet state (T1) with a geometry similar to the one of the S1 state, the molecule can undergo intersystem crossing (ISC) and enter a T1 state. Radiative emission from a T1 state is referred to as phosphorescence and is depicted in Figure 2.3. However, relaxation is not always radiative. Internal conversion (IC) is the non-radiative transition from one vibrational

level to a vibrational level of another state of the same multiplicity (e.g. S1 to S0). Most importantly for the present work, electronically excited molecules can undergo photochemical relaxation. Such photochemical relaxation can occur via isomerisation of molecules, bond formation via photochemical reactions and even bond dissociation. Over the past decades, a great number of photo-induced reaction systems has been

16 Introduction

S 1 S 0 y y g g r r T 1 e e n n e e

l l a a i i t t n n

e n o n - r a d i a t i v e d e c a y e

t t I n t e r s y s t e m c r o s s i n g o o p p

r r a a l l

u S u S c 0 c 0 e e l l o o m m

P h o s p h o r e s c e n c e F l u o r e s c e n c e

A b s o r p t i o n A b s o r p t i o n

i n t e r n u c l e a r s e p a r a t i o n i n t e r n u c l e a r s e p a r a t i o n

Figure 2.3: Overview about the possible electronic transitions after a molecule was excited by ultraviolet/visible (UV/Vis) irradiation. Left: Excitation pathway leading to ﬂuorescence. Right: Excitation pathway leading to phosphorescence. The ﬁgure is adapted from [147].

reported for macromolecular ligation and especially for the compaction of SCNPs by our group and others.25,148 An early example of light-induced SCNP folding is the dimerisation of cinnamates via [2+2] cycloaddition by the group of Liu. They prepared a PS-block-poly(2-cinnamoylethyl metharylate) and irradiated it at high dilution with broadband UV light to form tadpole-like SCNPs.149,150 Hecht and Khan synthesised a poly(m-phenylene ethynylene) bearing a cinnamate motif on each repeating unit. The polymer self-assembled in a helical structure and was subsequently crosslinked via irradiation at 316 nm.151 Similarly, Frank et al. reported SCNP formation via photochemical [4+4] cycloaddition of anthracenes upon irradiation with 350 nm UV light. UV/Vis spectroscopy showed that irradiation with UV light centred at 254 nm led to partial reversibility of the [4+4] cycloaddition, however, no increase in size of the particle was detectable via SEC.152 Willenbacher et al. prepared a PS bearing both proﬂuorescent phenyl-tetrazole and furan-protected maleimide moieties. Upon irradiation UV light (centred at 320 nm), the tetrazole moieties released molecular nitrogen, forming

17 Single-Chain Nanoparticles

highly reactive nitrile-imine species. The nitrile-imine subsequently reacted with the furan-protected maleimide in a cycloaddition to form fluorescent SCNPs. By employing an excess of tetrazole compared to maleimide, residual tetrazole was subsequently reacted with maleimide functionalised microspheres.153 By substituting the phenyl group pendant to the tetrazole with a pyrene moiety, the group was able to shift the reactivity of the tetrazole into the visible range at 410 nm.154,155 A stepwise light-induced SCNP formation was reported by Claus et al. These authors co-polymerised MMA with two photo-reactive monomers bearing an ortho-methylbenzaldehyde (o-MBA) unit and a phenacylsulfide moiety in a two-block fashion. The two blocks were subsequently folded at two different wavelenghts in the UV region. Under 355 nm irradiation, the phenacylsulfide would undergo Norrish Type II elimination and readily react with a thiol-bislinker. The o-MBA moiety was triggered at 320 nm to iso- merise to a diene and undergo a DA reaction. However, although the two photoreactive groups were addressed at distinct wavelength, they proved to be not fully orthogonal and the o-MBA moitey needed protection prior to activation of the phenacylsulfide moiety and subsequent deprotection.156 SCNP formation via [2+2] photo-dimerisation of styryl pyrene (StyPy) units at 445 nm was reported by Frisch et al. They elegantly showed that the intramolecular crosslinker- free dimerisation benefits greatly by the confined environment of the polymer. Not only was the quantum yield of the polymeric StyPy an order of magnitude higher than that of free solution StyPy, but the confined environment also allowed for intramolecular crosslinking at unprecedented high concentrations.157 In a subsequent study, Frisch et al. combined the [4+4] cycloaddition of anthracenes as reported by Frank et al.152 with their own StyPy [2+2] photo-dimerisation in a single polymer chain. In a firs step, the StyPy moieties were reacted selectively under irradiation with blue light (λmax = 470 nm) enabling a first compaction of the SCNP, while the anthracene units remained entirely unreacted. Next, the irradiation wavelength was switched to violet light at 415 nm and the anthracene dimerisation was induced, compacting the SCNP even further. When irradiating the linear precursor with violet light directly, however, both photo-reactive moieties were triggered simultaneously, resulting in a slightly less compacted SCNP compared to stepwise folding.158 Dynamic light-fueled SCNPs have recently been reported by Kodura et al. A poly(methyl methacrylate) (PMMA) based polymer bearing naphthalene moieties on the backbone was reacted with a triazolinedione (TAD)-bislinker under green light irradiation (525 nm). The thus formed SCNPs were in a non-equilibrium state, i.e. they were stable under the irradiation of green light, however, started to degrade in the dark, constituting

18 Introduction

the ﬁrst example reversible light-induced SCNP formation.159 The photo-ligation technique chosen in the present thesis is the so-called ’photo-enol’ ligation, ﬁrst reported by Yang and Rivas in 1970.160 Here, upon irradiation, an o-MBA

is excited from its S0 state to its S1 state (refer to Scheme 2.3). Via ISC, the singlet state is transformed into a T1 state. Alternatively, the singlet state can relax without emission of radiation to its ground state by IC. Instead of relaxation from the triplet state to the ground state via phosphorescence, fragmentation or hydrogen abstraction from the γ-position can take place yielding two isomeric o-quinodimethanes. In contrast to the (E)-isomer, the (Z)-isomer exhibits a half-life of only 4 s, undergoing [1,5]-sigmatropic

rearrangement yielding the o-MBA in its S0 ground state. The (E)-isomer, however, due to its missing ability to undergo sigmatropic rearrangement, shows half-lives of up to 250 s, depending on substitution and solvent. Therefore, it can be readily trapped in DA cycloaddition with electron-poor dienophiles, such as maleimides or acrylates. Interestingly, an additional driving force in the [4+2] cycloaddition of o-quinodimethanes is the restoration of aromaticity. As a result, reversibility of the DA reaction is hampered even at elevated temperatures, since reversion implies annulling of the aromaticity.

O O O OH hν ISC H-abstraction R R R R IC

E-isomer

EWG H-abstraction

R [1,5]-sigmatropic HO R R = H, Ph EWG rearrangement OH

Z-isomer

161 Scheme 2.3: Mechanistic details of photo-induced enolisation of o-MBA. The S0 state is excited to the S1 state upon irradiation. The T1 state can undergo hydrogen abstraction forming the o-quinodimethane. The enol species shows different half-lives depending on the isomer that was formed. The Z-isomer undergoes [1,5]-sigmatropic rearrangement, whereas the E-isomer can be trapped in a DA reaction.161

While the photo-responsive properties of o-MBA and its derivatives have been utilised for various macromolecular modiﬁcations, such as polymer conjugation,162 surface functionalisation163 or polymerisations,164 the group of Barner-Kowollik reported its ﬁrst example of SCNP formation.165

19 Single-Chain Nanoparticles

2.2.2 Analysis of Single-Chain Nanoparticles

SCNPs are most commonly characterised via SEC with regard to their apparent molecular weight. In SEC, the analyte is dissolved in an suitable solvent and subsequently passed over a highly porous stationary phase, commonly crosslinked PS. Shorter polymer chains are able to diffuse deeper into the pores of the stationary phase and elute slowly, whereas longer polymer chains diffuse less and elute faster. However, standalone SEC is a relative method and only provides data relative to a standard. Thus, using standalone SEC the analysis of SCNPs is limited to the comparison of appar-

ent molecular weight, or better the r H, before and after folding. The group of Hawker elegantly demonstrated that upon SCNP formation a reduction in apparent molecular weight of up to almost 75% is feasible according to SEC without changing the actual molecular weight. They synthesised a PS-co-poly(vinylbenzocyclobutane), which, upon thermal treatment, folds to SCNPs by forming cyclobutadiene moieties. Importantly, this crosslinking event occurs without any changes in actual molecular weight. Their results were furthermore supported by light scattering experiments, highlighting the importance of additional characterisation techniques.166 Light scattering has frequently been employed in macromolecular chemistry and physics to obtain radii of gyration (rg’s) and absolute molecular weights of solutions and suspensions. Light scattering thus proves an invaluable method to obtain quantitative information on polymers in solution. However, low optical contrast due to decreasing size of SCNPs (especially below 10 nm), increasing dispersity and isotropic scattering, light scattering techniques become more challenging. Viscometry correlates the intrinsic viscosity of a polymer solution to the molecular weight of the polymer via the Mark-Houwink-Sakurada equation 2.10,

[η] = K · M α (2.10) where [η] is the intrinsic viscosity, M the molecular weight, and K and α describe the

polymer-solvent interaction. The molecular weight can in turn be correlated to the r H via the Einstein-Simha equation 2.11,

s 3 3 M [η] rH = · (2.11) 4π 2.5NA with NA being the Avogadro number. Again, the group of Hawker investigated the use of viscometry for SCNP characterisation. An isocyanate-functional polymer was compacted in a crosslinker-mediated fashion using a bis-amine. While linear control

20 Introduction

polymers exhibited increasing intrinsic viscosities with increasing molecular weight, the SCNPs were virtually identical although featuring a difference in molecular weight of about 50%.167 Tuten et al. pioneered the hyphenation of conventional SEC with a differential refractive index (dRI) detector to both an external MALLS detector as well as a differential viscometer. This so-called Triple-SEC allowed not only for the detection of absolute

molecular weights, but also intrinsic viscosity and therefore r H values in a single run. Moreover, MALLS unveilled the presence of multi-chain aggregates invisible in the dRI signal of SEC. Importantly, standalone light scattering experiments were unlikely

to detect the multi-chain aggregates and probably would have overestimated the r H of the SCNPs.33 Engelke et al. synthesised a library of tert-butylacrylate polymers bearing PFB moieties varying the target molecular weight and the feed ratios. The linear polymers were intramolecularly crosslinked with a bis-thiol linker in a PFTR. Subsequently, quadruple detection SEC (conventional SEC coupled to a MALLS, DLS, viscosity and dRI detector) was employed for the analysis of the precursor polymers as well as the SCNPs and the results were compared to asymmetrical field flow fractionation (AF4). The AF4, in turn, was coupled to the same detectors as the SEC plus an additional UV/Vis detector. AF4 is a solution based separation technique without a stationary phase. A field flow with a parabolic profile is created inside a channel along which the analyte travels and a separation flow is created perpendicular to the field flow. The separation of AF4 is based on the diffusion of the analyte in the channel. Smaller molecules diffuse more than larger molecules and thus, on average, spend more time in the faster flow lines of the channel and elute faster, whereas larger molecules are more likely to be found in slower flow lines, thus eluting more slowly. Consequently, the elution profile of AF4 is inverted compared to that of SEC. Results of both techniques were well reproducible and Engelke et al. were able to perform an in-depth characterisation of the formed SCNPs, obtaining not only absolute molecular weights, but also radii of the polymers and SCNPs as well as intrinsic viscosity. Moreover, they produced on-line Mark-Houwink-Sakurada plots, providing further inside in the internal structure of the SCNPs.123 Steinkoenig et al. pioneered the use of SEC coupled to ESI-MS. They prepared a PMMA-co-polyglycidyl methacrylate which was subsequently folded via cationic ring- opening polymerisation (ROP). While the presented folding approach offers two distinct pathways for crosslinking, i) bimolecular coupling or ii) propagation, SEC–ESI-MS enabled the team to clearly distinguish between the two avenues, which could not be identified by any other means.35 In a subsequent study by Steinkoenig et al., a ter-

21 Single-Chain Nanoparticles

and quater-polymer prepared by Passerini multi-component reaction was collapsed via free-radical crosslinking. The employed polymer exhibits a series of key features to probe to the limitations and strengths of SCNP characterisation via SEC–ESI-MS. First, the polymer backbone possesses three or four different repeat units that need to be identiﬁed in the MS spectrum. Second, being a step-growth polymer, it features a high-dispersity, and last, free-radical induced crosslinking allows for i) termination via recombination of two chain-end radicals or a chain-end radical and an initiator radical or ii) disproportionation to a terminal hydrogen and a terminal double bond. Despite the immense variety in the polymer backbone, Steinkoenig et al. were able to elucidate the structure of the highly complex SCNP, highlighting the capabilities of their analytical approach.36 Nitsche et al. further advanced SEC–ESI-MS as a means of SCNP analysis by investigating the folding of linear PS via the nitrile imine mediated tetrazole-ene cycload-

dition (NITEC) reaction. In a NITEC reaction, exactly one molecule of N2 is released. Therefore, each crosslinking reaction can be precisely traced in the the mass spectra. Interestingly, the hyphenation of SEC to ESI-MS allows for the connection of speciﬁc molecular masses to individual retention time proﬁles, so-called extracted ion chro- matograms (XICs), similar to an SEC chromatogram. Examining the XIC of a PS bearing exactly two UV reactive moieties before and after irradiation, the group was able to show-

case the shift in retention time after the release of one or two N2 molecules, referring to one or two crosslinks per SCNP.168 By expanding their approach to the formation of SCNPs by intramolecular Friedel-Crafts alkylation of PS-co-poly(chloromethylstyrene), similar to reports by Davankov et al. (vide supra),132 and the loss of one chlorine atom per crosslink associated with it, the group was even able to identify up to ﬁve crosslinks in a single SCNP and illustrate the corresponding shift in retention time.169 Despite these excellent examples of SCNP analysis by means of ESI-MS, the technique is limited to crosslinking events that involve change in molar mass. Another frequently employed method for the characterisation is NMR spectroscopy. Although most commonly employed as one-dimensional 1H NMR to trace the appearance and disappearance of resonances upon folding, more elaborate techniques have been

developed. He et al. conducted spin-spin relaxation time (T 2) experiments in order to observe changes in molecular motion before and after SCNP formation via photo-induced

dimerisation of coumarin moieties. The authors were able to observe two T 2 relating to fast intrachain interaction and slow polymer-solvent interaction. While the intrachain

T 2 decreased with increasing crosslinking degree, indicating a loss of chain ﬂexibility, the polymer-solvent T 2 remained constant. This is unsurprising, as the polymer-solvent

22 Introduction

interaction should not be affected by SCNP collapse. Moreover, they recognised that

the fraction of fast interaction T 2 increased with increasing crosslinking density, as chain segments on the inside of an SCNP are less likely to be exposed to interaction with solvent molecules.170 Ormategui et al. were the ﬁrst ones to employ DOSY NMR for the characterisation of SCNPs.34 As the name suggests, DOSY determines the diffusion coefﬁcient D of molecules in solution. Via the Stokes-Einstein equation 2.12,

the diffusion coefﬁcient can be correlated to the r H:

k T r = B (2.12) H 6πηD

-1 with the Boltzmann constant kB in J·K , the temperature T in K and the viscosity of the solution η in Pa·s. As the crosslinking of a linear polymer chain proceeds, forming

SCNPs, the r H decreases and the diffusion coefﬁcient increases. The technique has since been frequently used by our group and others.26,34,36,157,168,171 Small-angle scattering techniques like small-angle neutron scattering (SANS) or small-angle X-ray scattering (SAXS) overcome the limitation in regard to particle size associated with light scattering techniques while still providing the r g and information on the morphology. However, both techniques demand a relatively high amount of sample compared to light scattering as well as access to special facilities. Still, both SAXS and SANS were employed by various research groups in order to characterise SCNPs.25,31,32 Last, but not least, microscopy techniques, such as atomic force microscopy (AFM) and transmission electron microscopy (TEM) were used to characterise the morphology of SCNPs. The group of Berda employed TEM for determination of SCNP size, conﬁrm- ing their results obtained by SEC.33,152 Moreover, TEM was capable of providing insight on the shape and morphology of SCNPs.152,172 Pomposo and colleagues were even able to quantify the size and number of local domains within SCNPs via TEM imaging.173 While similar information can also be obtained from AFM measurements,152,172 the group of Meijer employed their AFM instrument for the mechanical unfolding of SCNPs, thus providing insight on the structure and kinetics of the folding. They prepared a set of linear, telechelic polymer precursors bearing varying amounts of either UPy or BTA moieties, protected by photo-cleavable o-nitrobenzyl groups. Furthermore, each of the polymers was carrying two dithiolane endgroups which allowed attachment of the polymer to a gold surface and a gold AFM cantilever. Upon irradiation, the o-nitrobenzyl groups were cleaved and the polymers self-assembled into SCNPs via UPy dimerisation or BTA stacking. The collapsed SCNPs were subsequently deposited onto gold surfaces

23 Self-Reporting Systems and captured with a gold-coated AFM cantilever. The force applied to the cantilever is recorded as a function of chain extension and displays multiple rupture events. Their results also showed that SCNPs with a higher amount of crosslinker also featured more rupture events, as expected. Meijer and colleagues thus provide an in-depth analysis technique for the internal structure of SCNPs.174 Despite the plethora of synthetical tools for SCNP synthesis reported in the last two decades, the characterisation of SCNPs still is mostly limited to comparison of radii before and after SCNP collapse and only few analytical techniques allow for an in-depth investigation of SCNP structure. The development of new analytical approaches for the characterisation of SCNP is therefore imperative.

2.3 Self-Reporting Systems

Self-reporting systems have been used for various applications and different probes have been developed in the past decades. An overview about recently developed systems will be provided within this section. Among the probes employed for self-reporting systems are UV/Vis absorption (i.e. change in colour), fluorescence and chemiluminescence. Schenzel et al. have reported hetero-Diels-Alder (HDA) reaction of cyclopentadiene and a phosphoryl dithioester for reversible adhesion and dynamic covalent polycarbonate networks.45,75 Upon reaction with the cyclopentadiene and thus formation of the network, the deep purple phosphoryl dithioester loses its colour and thus reports the degree of conversion. When heating the network, the cycloreversion takes place and the purple colour is regained. Mechanoresponsive materials have been reported by the group of Sottos.175 By incorporating spiropyrans in poly(methyl acrylate) (PMA) chains, they created predetermined breaking points in the polymer. When applying tensile force to the polymer, the colourless spiropyran is converted into a merocyanine exhibiting a red colour. The reaction is reversible by irradiation with visible light. However, more commonly employed are fluorescence probes for self-reporting systems. Noel and co-workers synthesised polycarbonates of ferulic acid and tyrosine that showed inherent fluorescence.176 The polycarbonates were post-functionalised with polyethylene glycol (PEG) by CuAAC and were subsequently able to self-assemble in aqueous media. The thus formed micelles were found to degrade in basic media into the corresponding monomers. The degradation is eventually self-reported by a loss of intensity of fluorescence of up to 90%. Liposomes for self-reporting drug release were reported by the group of Liu.12 Liposomes designed by the group of Liu exhibited a fluorescent probe in spatial proximity to a thiol-responsive disulfide. Presence of thiols in

24 Introduction

cancer cells lead to cleavage of the disulfide bond and subsequent breakdown of the liposome and drug release. Additionally, the fluorophore incorporated in the liposome experiences a shift of its emission maximum from 459 nm to 515 nm.12 Robb et al. have developed a strategy for self-reporting detection of material damage in a micrometer scale. Core-shell microcapsules containing a solution of tetraphenylethylene (TPE) in hexyl acetate were embedded in a polymeric material. Upon mechanical damage to the composite and thus the capsules, the TPE solution is released and deposited in the damaged region. Subsequently, fluorescence of the TPE deposit readily identifies the damaged area as dissolved TPE does not fluoresce.177 Self-reporting damage to glass- and carbon-fibre-reinforced polymer composites was reported by Bruns and co-workers.178,179 A fluorescent protein was immobilised on the carbon or glass fibres and retained its fluorescence after embedding of the fibre into an epoxy resin. However, mechanical damage to the material was self-reported by a loss of fluorescence as the protein was mechanically unfolded. The group of Barner-Kowollik has pioneered a self-reporting step-growth polymerisation employing NITEC. A bifunctional tetrazole also exhibiting a RAFT agent and a bismaleimide monomer were polymerised in an equimolar ratio. Neither of the two monomers showed any fluorescence prior to polymerisation. The NITEC adduct, however, displayed significant fluorescence. Observation of the polymerisation progress with both, NMR and fluorescence spectroscopy, the group was able to show that the intensity of fluorescence is directly correlated with the degree of polymerisation.180 Also employing NITEC, the group provided fluorescent self-reporting SCNPs.171,181 A polymer release system with self-reporting properties was also described by the group of Barner-Kowollik.182 A profluorescent polymer was synthesised containing both nitroxide moieties with unpaired spins and pyrene moieties. The pyrene moiety was incorporated into the polymer via a HDA reaction. As unpaired spins of nitroxides are able to quench fluorescence, the polymer only shows weak fluorescence. When the pyrene moiety was replaced by ethyl sorbate by a thermally triggered reversed DA reaction, the increasing distance between pyrene and nitroxide prevented suppression of fluorescence. Thus, by increasing fluorescence, the release of pyrene was self-reported. Recently, also self-reporting systems using chemiluminescent probes have attracted interest. The group of Shabat has reported two systems employing modified Schaap’s dioxetane. An aqueous solution of a dioxetane precursor was treated with 3-(1,4-dihydro- 1,4-epidioxy-4-methyl-1-naphthyl)propionic acid (EP-1). Upon heating, EP-1 releases singlet oxygen that reacts with the dioxetane precursor and oxidises it to a dioxetane. The dioxetane subsequently decays under emission of light. By functionalising a peptide with

25 Chemiluminescence

the dioxetane precursor, the molecule was also able to penetrate cells and was employed as an intracellular singlet oxygen probe.11 Furthermore, by combining quinone-methide elimination with the mechanism of Schaap’s dioxetane, the group was able to provide a self-immolative polymer with chemiluminescencing with each degradation step.183 After cleaving the polymer from a substrate, the polymer degrades in a domino-like fashion following the quinone-methide elimination mechanism. Subsequently, the monomeric unit, quinone-methide dioxetane, decomposes via the dioxetane chemiluminescence pathway. Moreover, the group of Sijbesma has introduced predetermined cleavage points of dioxetanes in various polymers. The dioxetanes were stabilised by bulky adamantyl moieties in order to prevent degradation. Upon application of mechanical stress to the polymers, the dioxetanes would still break and emit bright blue light, indicating damage to the material. However, since luminescence is not triggered by a chemical reaction but rather mechanical action, one should speak of mechanoluminescence rather than chemiluminescence.184,185 Geiselhart et al. recently reported a chemiluminescent self-reporting system for dynamic supramolecular complex formation based on luminol CL.186 They synthesised MMA-methacrylate or MMA-methacrylate-styrene-based copolymers and post- functionlised them with luminol and either guanidine or 1,5,7-triaza-bicyclo-[4.4.0]dec- 5-ene (TBD). Subsequently, β-cyclodextrin (β-CD) was added to the polymers and the bases would readily form host-guest complexed with the β-CD as witnessed by nuclear Overhauser effect NMR spectroscopy (NOESY) and DLS. Upon addition of hydrogen

peroxide (H2O2), however, the host-guest complex was dismantled and the guanidine base was furthermore oxidised to urea. Simultaneously, catalysed by the now uncom- plexated base, the luminol was oxidised to 3-aminophthalic acid and emitted a striking blue light, thus reporting the decomplexation. While Geiselhart et al. established a prime example of a self-reporting system that allows for the facile detection of reactive oxygen

species (ROS) such as H2O2, the CL of luminol in their system is not directly linked to the disassembly of the host-guest complex, but is merely prompted simultaneously by the same trigger.186 Yet, in a truly self-reporting system the supramolecular transformation would trigger a reaction cascade ending with the emission of light.

2.4 Chemiluminescence

Bio- and chemiluminescence is the emission of light triggered by a (bio-)chemical reaction. In contrast to ﬂuorescence and phosphorescence, also referred to as photolu- minescence, bio- and chemiluminescence are therefore not triggered by irradiation of

26 Introduction

molecules. First mentions of bioluminescence date back to 1000-1500 BC in Chinese literature outlining the luminescence of glow-worms and fireflies.187 However, it was not until ca. 350 BC that Aristotle first reported the luminescence of several marine species. He also observed that – in contrast to the flame of a candle – the luminescence of these marine species did not occur in combination with heat, and therefore coined the term cold light.188 Pliny the Elder subsequently described a variety of luminescent aniMALLS in his Naturalis Historia, including a species of molluscs that were regarded as a delicacy by the Romans.189 While detailed studies of nature in general were rare during the middle ages due to oppression by the church, luminescence has drawn the interest of many well-known Renaissance scientists, philosophists and explorers, such as Sir Francis Bacon, Renee Descartes, Galileo Galilei, Robert Boyle, Charles Darwin or Christopher Columbus.190 In 1888, French chemist Raphaël Dubois pioneered the extraction of luciferin and luciferase, the two compounds that create the bioluminescence.190 It was also only 11 years prior that Bronisław Radziszewski reported the first example of pure CL. By dissolving lophine in alkaline alcohol, an intense white light was emitted from the solution. Most importantly, when the reaction was conducted under the exclusion of oxygen no chemiluminescence was observed.191 A common structural motif of both lophin and luciferin as well as many other bio- and chemiluminescent compounds is the 1,2-dioxetane (refer to Scheme 2.4). The CL of 1,2-dioxetanes and its derivatives will be explained hereinafter.

KOH O N 2 N O N N O H H

lophine

HO S S Luciferase HO S S O2 O N N OH N N O O O firefly luciferin

Scheme 2.4: Reaction of lophine with oxygen under basic conditions (top) and enzymatically catalysed reaction of luciferin with oxygen (bottom). The reaction products are thermodynamically unstable and decompose under formation of a electronically excited state. Notably, both oxygenated species exhibit a 1,2-dioxetane moiety (red).

27 Chemiluminescence

2.4.1 1,2-Dioxetanes

In 1968, Kopecky and Mumford were the ﬁrst to describe the luminescence in the thermal decomposition of 3,3,4-trimethyl-1,2-dioxetane.192 However, different mechanism for the CL mechanism have been suggested. A concerted mechanism resembling a pericyclic rearrangement were proposed by McCapra and Turro.193,194 According to McCapra, formation of the carbonyl groups will result in an excited state although this thermal route is actually symmetry forbidden.193 O’Neal and Richardson postulated a biradical mechanism, where the O-O bond is homolytically cleaved ﬁrst and two carbonyl groups are formed subsequently. However, molecular orbital calculations by C. and J. Tanaka suggest that neither mechanism is correct and an intermediate of the two mechanisms is more accurate.195 Both the concerted and the biradical mechanism are depicted in Scheme 2.5.

biradical concerted

hν hν O O O O O O OO O O + + H H H H H

Scheme 2.5: Two proposed mechanisms of 1,2-dioxetane chemiluminescence, a biradical on the left and a concerted on the right.

By providing bulkier substituents, the stability of the dioxetane can be improved. Adamantylidenadamantane -1,2-dioxetane, for example, only starts to decompose at 150 °C.196 While the chemiluminescence of ’naked’ 1,2-dioxetanes usually proceeds via a triplet state, it has been shown that addition of a suitable fluorophore such as anthracene or pyrene significantly increases CL quantum yields and emission then occurred from a singlet state.192 In such a case, the 1,2-dioxetane forms a charge- transfer complex with the fluorophore allowing for an energy transfer from the dioxetane to the fluorophore und subsequent emission from the fluorophore’s excited singlet state. While synthetic dioxetanes reported until then showed triplet state excitation and low CL quantum yields, firefly luciferine showed bioluminescence quantum yield of close to 88%.197 Therefore, White et al. further investigated the effect of substitution at the aromatic ring on the quantum yield of luciferin. When they substituted the hydroxy functionality with an amino functionality, the molecule was still able to produce light. However, when forming the amide of aminoluficerin, no emission could be observed.198 Moreover, when changing the position of the arylhydroxy functionality to any other position, the luciferine isomer also lost its luminescent properties.199 Inspired by White’s findings,

28 Introduction

O O O O O O kET kCCC O O

O O O O

kEBT

hν S1 O O O O

O O

Scheme 2.6: The CL mechanism of Schaap’s dioxetane. After formation of the phenoxide via deprotonation or deprotection, an electron transfer from the phenoxide to the dioxetane occurs, followed by cleaved of the O-O bond and immediate cleavage of the C-C bond. The electron back-transfer follows an intermolecular pathway from the adamantyl species to the phenoxide (upper pathway) to form the excited singlet state species. Relaxation occurs via emission of light.

Schaap et al. synthesised a phenol-substituted 1,2-dioxetane in order to investigate its chemiluminescent properties. While the phenol-substituted 1,2-dioxetane barely shows any emission behaviour by itself, deprotonation of the phenol prompted a blue flash luminescence. Indeed, the half-life of the protonated species in toluene at -35 °C has been determined to be about 17 years, whereas the half-life of the deprotonated form was only a few minutes.200 Similarly, they prepared a naphthyl acetate-substituted 1,2- dioxetane bearing an adamantyl moiety as second substituent providing the dioxetane with substantial thermal stability. Upon addition of aryl-esterase the acetate was cleaved off the naphthyl moiety and triggered a bright blue light constituting the first example of enzymatically triggered CL.201 Probably one of the best-known examples of dioxetane CL is their adamantyl-siloxyaryl substituted 1,2-dioxetane, which is commonly referred to as ’Schaap’s dioxetane’. Schaap’s dioxetane exhibits a half-life of almost four years at ambient temperature but decays rapidly under emission of a bright blue light upon addition of fluoride ions. It also shows excellent CL quantum yields of up to 25% and chemiexcitation efficiencies of up to 57%.2 The secret of its high CL quantum yields lies within its degradation mechanism. Rather than degrading via a thermal route and being excited to a triplet state with low quantum yields, phenol-substituted dioxetanes degrade via an intramolecular electron-transfer mechanism as proposed by Koo and Schuster (refer to Scheme 2.6).202 After forming the phenoxide – either via deprotonation or

29 Chemiluminescence by deprotection – an electron is transferred to the dioxetane, the O-O bond is cleaved followed by an immediate homolytical cleavage of the C-C bond. Intermolecular electron back-transfer (EBT) from the adamantyl species to the phenolate renders the latter in an excited singlet state emitting blue light upon relaxation. While an intramolecular EBT has been part of mechanistic discussions, it could be proven via viscosity experiments that this pathway is unlikely to occur (vide infra).203–205 Subsequently, research groups have developed a large range of dioxetanes bearing not only hydroxy functionalities as triggers for the CL decay, but also sulphur206–208 and nitrogen analoga209 and even CL decay via formation of a carbanion has been reported.210 Since Schaap first described his dioxetane, a vast range of leaving group modifications have been investigated in order to detect various analytes (refer to Figure 2.4, C). Possible analytes range from enzymes such as alkaline phosphatase and galactosidase to bio- or environmental toxins, such as hydrogen peroxide, hydrazin or hydrogensulfide .17,21,183,211–214 While some analytes simply cleave off the leaving group as it is the case with alkaline phosphatase and galactosidase, analytes like formaldehyde or peroxynitrite require more sophisticated leaving groups. Some leaving groups, for example, react with the corresponding analyte to para-benzylic ethers that undergo quinone-methide elimination and produce the phenoxide-dioxetane.183,211,213 Notably, the leaving group reported by Bruemmer et al. forms an imine with formaldehyde, undergoes an aza-Cope rearrangement, subsequently hydrolyses and finally undergoes beta-scission to generate the chemiluminescent phenoxide-dioxetane species.17

30 Introduction

A O O O O O O O O O O O O O O O HO O Si O O OH O O Si

base triggered fluoride triggered

B O O O O O O O O O O NO2 O O S O S NH CH2 S NC O2N

sulfide triggered base triggered

HOOH C O O OH B Br OH O O P O O O HO O O OH

triggered by: H2O2 N2H4 Alkalinephosphatase Galactosidase

N3 H NH2 N O O O O O O O triggered by: O H S O O 2 N O- H H

Figure 2.4: A. An overview of early dioxetanes synthesised by Schaap and colleagues, with the Schaap’s dioxetane on the far right.2,200,201 B. Exchange of the hydroxy-functionality with sulfur, nitrogen or even carbon species still allows for the CL decay of the dioxetane with an appropriate trigger.206–210 C. An incomplete selection of protecting group modiﬁcations that allow for the detection of numerous analytes via CL.17,21,183,211–214

31 Chemiluminescence

2.4.2 Peroxyoxalates

This section has been extracted and adapted from the candidates ﬁrst author publication ’All Eyes on Visible Light Peroxyoxalate Chemiluminescence Read-Out Systems’ in Chem. Eur. J. 2020, 26, 114-127. In 1963, the ﬁrst reports of what is today known as peroxyoxalate chemiluminescence (PO-CL) were published by Chandross regarding a simple reaction of oxalyl

chloride with H2O2 in the presence of 9,10-diphenyl anthracene (DPA), creating a ’bluish- white light’.215 The vapours of the reaction were described to induce fluorescence of anthracene-impregnated filter paper and a ’metastable excited electronic state or some other highly energetic species’215 present during the reaction was suggested. His hypothesis was largely supported by the work of Rauhut et al., who provided a thorough study of varying substituted oxalate esters, finding oxalate esters bearing electronega- tively substituted phenyl groups to be the most effective ones.216–218 Their conclusions allowed interesting mechanistic insights into the PO-CL and revealed a correlation between the substituent of the phenol and the quantum yield as well as emission lifetime of the corresponding chemiluminescent reaction. Early experiments of Chandross already indicated one of the key features of PO-CL compared to other luminescent reactions such as luminol or the catalysed CL of phenolate-substituted dioxetanes: the possibility to tune the wavelength of the emitted light by adding different fluorophores to the reaction mixture rather than changing the substitution pattern on the CL molecule.215 In contrast to the intramolecular luminescence of phenolate-substituted dioxetanes, where the wavelength is dependent on the exact substitution of the dioxetane itself, the PO-CL system produces intermolecular luminescence. Thus, the need of such a fluorophore also makes the system more versatile since the emitted wavelength can be tuned on demand by employing a wide variety of different fluorophores, spanning over the complete visible spectrum and even into the infrared area. A detailed mechanistic study of the PO-CL, including the energy transfer to the fluorophore. Nowadays, the most widely employed oxalate esters are bis(2,4,5-trichlorophenyl-6-carbopentoxyphenyl)oxalate, bis-(2,4,6-trichlorophenyl) oxalate (TCPO), and bis-(2,4-dinitrophenyl)oxalate (DNPO). Quantum efficiencies for certain phenyl oxalate systems were reported of 30% and more,190,219 being competitive with their natural counterpart, the luciferin/luciferase system and opening a variety of possible applications. Due to the high sensitivity towards specific activating species, high photon output, and long luminescence lifetimes, PO-CL reactions find widespread application ranging from analytical purposes like the 220 221 detection of H2O2 in water, viable microorganisms in food, or blood traces in

32 Introduction

forensic science,222 to common everyday objects such as glow sticks or emergency lighting devices.223,224 In the previous years, CL has gained interest in the ﬁelds of nanomaterials, biological imaging, molecular sensors, and in polymer science.

O O X O O O X H O O HO - HX X 2 2 X OH OO OO O - HX O + HX

Xa = Cl b = aryl c = imidazole

Scheme 2.7: Formation of 1,2-dioxetanedione. Oxalyl chloride, aryl oxalates or oxalyldiimi- dazolide react with H2O2 in a nucleophilic substitution to form the oxalic peracid derivative. Ring-closure leads to 1,2-dioxetanone derivative, subsequent elimination of HX results in the formation of 1,2-dioxetanedione.

The quest for a high-energy intermediate

The ﬁndings of Chandross and Rauhut sparked investigations into the mechanistic details of CL reactions and the search for the high-energy intermediate involved in the production of light. As previously stated, in their comprehensive study of substituted aryl oxalates Rauhut et al. established important relationships between the substitution of the phenols

and the quantum yield ΦCL on the one hand, as well as a correlation between substitution and emission lifetime on the other hand. With an increasing electronegativity of the substitution, the quantum yields of the PO-CL increased to a total of up to 15%, while the emission lifetimes decreased from several hours to only a few minutes. In agreement with Chandross, Rauhut et al. also found that the CL emission spectra were identical to the ﬂuorescence spectra of the employed ﬂuorophores, except for a minor bathochromic

shift that they attributed to small changes in the CL environment caused by H2O2. Interestingly, infrared (IR) spectroscopy measurements revealed that more than 98% of

ester groups have disappeared only four minutes after the addition of H2O2. However, about 75% of the chemiluminescence was observed after these initial four minutes.

Subsequently, Rauhut et al. prepared mixtures of DNPO and H2O2 in dimethyl phthalate and added the fluorescent dye with a delay of 27 and 70 minutes, respectively. Although they observed reduced total quantum yields compared to when the fluorescer was initially added, it was evident that a significant amount of CL manifested after no CL was visible anymore in a standard experiment.217 In order to prove the existence of a high energy intermediate (HEI) as originally suggested by Chandross, they prepared two solutions,

33 Chemiluminescence

one of them containing DNPO and H2O2, the other one containing DPA or rubrene. When gas streams were passed through the DNPO-H2O2 solution into the ﬂuorescer solution, bright CL with a short lifetime was observed. Based on these results, they proposed 1,2-dioxetane-3,4-dione as the metastable intermediate.217 Although, direct observation of 1,2-dioxetane-3,4-dione via IR or MS by Rauhut et al. was unsuccessful, Cordes et al. pioneered the MS investigation of the PO-CL (refer to Scheme 2.7). By

preparing a solution of TCPO and H2O2 in a high boiling point phthalate and its injection into a mass spectrometer, the group was able to detect high intensity signals at 28, 32 and 44 amu, corresponding to carbon monoxide, oxygen and carbon dioxide, respectively. Moreover, they found peaks at 88 and 60 amu, corresponding to a carbon dioxide dimer 225 (possibly 1,2-dioxetanedione) and CO3. Ab initio NMR calculations in combination with 13C and exchange NMR spectroscopy (EXSY) were performed by the group of Barnett.226,227 In a ﬁrst set of experiments, 13C-labelled oxalyl chloride was reacted 13 with H2O2 in d8-tetrahydrofuran (THF) in the presence of DPA at -70 °C. On-line C NMR revealed a single carbon species at a chemical shift of 154.5 ppm which was in good agreement with ab initio calculations.226 In subsequent 13C EXSY experiments, detecting chemical exchange and therefore allowing for the identiﬁcation of reaction pathways, Barnett and co-workers were able to provide evidence for a nucleophilic

substitution reaction of H2O2 and oxalyl chloride to 2-chloro-2-oxoethaneperoxoic acid and subsequent elimination of hydrochloric acid (HCl) and ring-closure to form 1,2- dioxetanedione (refer to Scheme 2.7).227 However, as the NMR data has not been directly linked to the CL properties of the system, the presented investigations do not provide clear evidence that dioxetanedione is the previously suggested HEI. In a very elegant experiment, Stevani and Baader attempted the trapping of a cyclic peroxide HEI.228 It has been shown previously that triphenylantimony readily inserts into the O-O bond of tetramethyl-1,2-dioxetane.229 Stevani and Baader thus synthesised the corresponding triphenylantimony – dioxetanedione adduct via an alternative route and characterised it in order to provide a reference compound. Subsequently, they performed the peroxyoxalate reaction in the absence of a ﬂuorophore and passed a gas stream

of N2 over the reaction mixture that led into a solution of triphenylantimony. Upon contact of the volatile intermediate carried in the gas stream with triphenylantimony, insertion of the antimony species into the O-O bond was anticipated. However, no such reaction was observed. Nevertheless, although no insertion was detected, the property of 1,2-dioxetanedione as HEI has not been disproven, as catalytic degradation of the dioxetane can possibly compete with the insertion reaction.228,229 In addition to dioxetanedione, a substituted 1,2-dioxetanone has been discussed as a potential

34 Introduction

HEI.230–234 Aoyama and co-workers investigated the CL properties of different 2,4,6- trichlorophenyl N-aryl-N-tosyloxamates revealing a Hammett relationship between some of the investigated structures. They found reaction constants ρof +1.75 for the release of the tosylanilides and +2.66 for the emission of light. These results suggest that formation of a HEI and its decay with an activator is faster than the release of tosylanilides and N-aryl-N-tosyl substituted 1,2-dioxetanones are the reactive intermediate.231 Similarly, Murayama et al. provided ρ-values for the PO-CL reaction of different aryl oxalates with a set of distyrylbenzenes and compared these with the lowest unoccupied molecular orbital (LUMO) energies of corresponding 1,2-dioxetanones and 1,2-dioxetanedione based on ab initio calculations. They described a linear dependency of the ρ-values and the pKa-values of the corresponding phenols correlating to the observed CL decay rates. Solely bis(2,6-dichlorophenyl) oxalate and bis(2,4,6-tricholorophenyl) oxalate exhibited deviations from the general trend reported, which can be attributed to the para- substitution and its steric hindrance.235 Thus, in agreement with Aoyama,231 Murayama et al. proposed aryloxy-substituted 1,2-dioxetanones as the HEI. Lindh and co-workers have conducted theoretical studies on the binding energies of 1,2-dioxetanes and 1,2- dioxetanones with ﬂuorophores employing naphthalene as a model compound.236 Their calculations showed that unsubstituted 1,2-dioxetanes exhibit a lower binding energy than 1,2-dioxetanones, and that binding energies decrease even further upon methylation of the peroxides. Moreover, after formation of the complex, the ground state energy of the cyclic peroxides was found to increase with O-O bond elongation, whereas the energies of the n,σ* and the π,σ* states decrease. This decrease in energy is more pronounced for the unsubstituted 1,2-dioxetanone as the carbonyl group provides additional planarity of the molecule.236 Although the presented theory has only been applied to 1,2-dioxetanes and 1,2-dioxetanones, one can speculate about an even more pronounced effect for 1,2-dioxetanedione, which will be part of their future work.236 Despite various attempts, to the best of our knowledge, no unequivocal evidence for or against the proposed HEI structures has been provided yet (vide supra).

Catalysed degradation of the HEI

Early findings that (i) CL emission spectra were identical to the fluorescence spectra,215–218,237 and (ii) there was no significant CL observed without any fluorescer215–218 led to the assumption that the HEI must somehow interact with the fluorescer in order to produce CL. In a set of experiments, Rauhut et al. reacted oxalyl chloride and

H2O2 at constant amounts with varying amounts of DPA. They observed an increase

35 Chemiluminescence

O O

hν

kCT

O O

CO2 kET kEBT

O O O O

O O kCCC

CO2

Scheme 2.8: Schuster’s CIEEL mechanism on the example of 1,2-dioxetanedione. Adapted from [237] .

36 Introduction

in quantum yield with increasing DPA concentration. Moreover, a linear correlation was found when plotting the quantum yields against DPA concentration in a double reciprocal fashion.218 This correlation has been supported by the work of various research groups.202,230,237–240 Furthermore, Schuster and co-workers have correlated the bimolecular rate constant of the CL reaction of various organic peroxides with different activators to the oxidation potential of the corresponding activators. The linear dependency of the natural logarithm of the rate constant on the oxidation potential suggests an electron transfer (ET) or at least charge transfer from the fluorophore to the peroxide prior to CL.202,237–240 Based on this data, Schuster postulated the so-called chemically initiated electron exchange luminescence (CIEEL) mechanism, involving an electron transfer and back-transfer EBT between organic peroxide and an activator (refer to Scheme 2.8).202,237–240 The group of Baader confirmed that these findings also apply to the HEI of the PO-CL.239–241 The CIEEL mechanism involves five key steps: (i) formation of a charge transfer (CT) complex of an activator with the cyclic organic peroxide within a solvent cage facilitating O-O bond elongation, (ii) electron or at least partial charge-transfer from the activator to the peroxide and almost simultaneous O-O bond cleavage, (iii) concerted cleavage of the carbon-carbon bond (CCC) and release of one equivalent of carbon dioxide, (iv) electron (or charge) back-transfer EBT to and singlet excitation of the activator together with the release of a second equivalent of carbon dioxide, and (v) decay of the singlet excited state via emission of light.190,202,237–240,242,243 Furthermore, Baader and co-workers correlated the singlet quantum yields of the PO-CL to the change in free energy in the EBT leading to an excited state of the activator.239–241 Nonetheless, it was unclear whether the two radical ions would remain within a solvent cavity until the EBT step or if they would diffuse away from each other preventing EBT and resulting in low quantum yields of the CL reaction.237,244 Adam et al. have studied the influence of solvent viscosity on the CL of a phenyl-adamantyl-substituted dioxetane which follows the CIEEL mechanism in an intramolecular fashion.203–205 However, two pathways for the EBT seem feasible (refer to Scheme 2.8). After electron transfer from the phenol moiety to the dioxetane and O-O bond cleavage, pathway one suggests C-C bond cleavage and formation of a phenol radical and an adamantyl anionic radical. While both species are still present within a solvent cage, EBT from the adamantyl species to the phenol occurs intermolecularly. The second route follows C-C bond cleavage and formation of an anionic biradical and adamantanone followed by intramolecular EBT. In case of an intramolecular EBT, the system and therefore the quantum yields should be unaffected by changes in solvent viscosity. In a set of fluoride-triggered CL reactions, Adam et al. have elegantly con-

37 Chemiluminescence

firmed the viscosity dependence of the system by employing a range of solvent mixtures of benzene and diphenylmethane (DPM). These experiments therefore support an intermolecular EBT for the CL of phenyl-adamantyl-substituted dioxetanes.203–205 Baader and co-workers have subsequently investigated the solvent viscosity dependency of the PO system and compared it to the dioxetane system. When gradually changing the solvent from pure toluene to 90% DPM, and thus increasing the viscosity by a factor of 4.6, the CL quantum yields of two different dioxetane systems increased by a factor of 2.2 and 2.6, respectively, whereas the quantum yield of the PO-CL increased by a factor of 9.4.245 In more polar solvents, however, the effect of solvent viscosity on the CL was less pronounced as a 34-fold increase in viscosity from pure ethyl acetate to pure DPM only resulted in an increase in quantum yields of a factor of five for the PO-CL.219 The presented viscosity dependency of both CL systems provides evidence for a rather stable complex within a solvent cage that facilitates EBT before the species can diffuse away. Still, the question has been raised whether the interaction of the organic peroxide with an activator molecule and subsequent decomposition of the peroxide involves an actual electron transfer or if a charge transfer would be more conclusive. In accordance with the Franck-Condon principle, electron transfer and back-transfer are about 100 times faster than that of nuclear motions. Thus, bond breaking and relocating of nuclei would have to happen close-to simultaneously with electron-transfer.244 As this requirement seems highly unlikely, the charge-transfer induced luminescence (CTIL) mechanism was proposed as an alternative to the CIEEL mechanism. In the CTIL mechanism, a charge at the O-O bond is gradually developed which results in a concerted O-O and C-C bond cleavage. This partial charge-transfer instead of a complete electron-transfer also attributes to the generally high quantum yields of PO-CL.246,247 Various mechanisms for the PO-CL have been proposed in the past 50 years since its discovery and numerous experiments have been conducted in order to confirm suggested mechanisms and their intermediates. Our group has recently demonstrated the effect of confined environments, which polymers provide, on photochemical SCNP-folding.157 Similarly, a copolymer consisting of two blocks bearing different fluorophores with distinct emission spectra could provide such a confined environment. If one of the two blocks were to be selectively cross-linked via phenyloxalate moieties and subsequently treated with

H2O2, a resulting 1,2-dioxetanedione HEI would be free to diffuse and interact with and produce CL of both fluorophores adjacent to the polymer backbone. On the other hand, if phenyl-substituted 1,2-dioxetanone is the HEI of PO-CL, it would still be bound to the polymer backbone, unable to diffuse freely and only yield the CL of one fluorophore. Thus, we envisage that the confined environments present a key opportunity for further

38 Introduction

insights into the mechanics of PO-CL.

2.4.3 Luminol

One of the oldest known chemiluminescent reactions is the one of luminol with basic hydrogen peroxide. Although luminol has not been used in the present thesis, it should briefly be mentioned here, as it represents another prominent example of chemiluminescence with a real world application. It was first described by Albrecht in 1928.190,222 The most famous application of luminol nowadays is probably the detection of presump- tive bloodstains on crime scenes.222 Despite its use in forensic sciences, luminol is also commonly employed for the detection of polyphenols, which represent persistent environmental pollutants.248 Luminol represents an aminophthalyl hydrazide. By addition of a sufficiently strong base, usually sodium hydroxide, the hydrazide is deprotonated and a tautomeric equilibrium is established. When hydrogen peroxide is subsequently added to the solution, the aminophthalate in a singlet excited state is formed and nitrogen is released from the molecule. Finally, relaxation of phthalate occurs as bright blue luminescence. The mechanism of the luminol chemiluminescence is depicted in Scheme 2.9.190,249

* NH2 O NH2 O NH2 O NH2 O 2 OH- NH N N H2O2 O

NH N N - N2 O O O O O

Scheme 2.9: Mechanism of the luminol chemiluminescence.

39 3 Self-Propagated para-Fluoro – Thiol Reaction

3.1 Abstract

We introduce a protecting-group-free concept for the powerful PFTR employing a source of ﬂuoride ions as base. The reaction is shown to be self-propagating, with understoichiometric amounts of base, thus effectively foregoing the need for high base concentrations. Careful tuning of the solvent-thiol combination allows for quantitative conversion, in some cases within a short timeframe, when only minimal amounts of base are used, allowing the PFTR reaction to essentially proceed base-free.

40 Self-Propagated para-Fluoro – Thiol Reaction

3.2 Introduction

SNAr on perfluorinated aryl compounds have been known as early as the 1950s when hexafluorobenzene was reacted with an alkali methoxide or hydroxide in order to form pentafluoroanisol or pentafluorophenol, respectively.104–106 Since then, a wide range of nucleophiles have been employed for the nucleophilic substitution of aryl fluorides, including alcohols, amines and thiols.108–112,115,250 While early studies have shown that for most pentafluoro-aromatic groups the nucleophilic substitution occurs in the para- position, more recent calculations by Kvícala˘ et al. provided mechanistic explanations for the observed regioselectivity.115 In a PFTR, a thiol is commonly deprotonated by a base in order to provide the corresponding thiolate. Subsequently, the thiolate acts as a nucleophile and substitutes the para-fluorine of the aromatic ring. During nucleophilic attack at the para-position, pentafluorophenyl moieties exhibit an increased polarisability and therefore enhanced stability of the de-aromatised transition state compared to the same nucleophilic attack in ortho- or meta-position.110,112,115 Notably, using thiols requires less harsh reaction conditions due to the increased acidity and nucleophilicity of thiols compared to alcohols and amines.116 Furthermore, it was found that aromatic and glycosidic thiols exhibit the highest reactivities, followed by primary, secondary and lastly tertiary thiols.251 In the last ten years, the PFTR has attracted increasing interest especially in polymer chemistry, as it provides a highly efficient and versatile tool for polymer ligation, as highlighted in recent review articles.113,114 Aside kinetic studies, the orthogonality of PFTR towards amidation of activated esters as well as thiol-ene reactions has been investigated.118,119 Thus far, the PFTR has been employed for post-polymerisation modification,117–120,252,253 protein conjugation120,254 as well as surface functionalisation.120,255 Moreover, PFTR was used by our group as a powerful tool in the synthesis of precision networks.1 Recently, Hedrick and co-workers reported an organocatal- ysed thiol-fluorine substitution.232 By employing perfluorinated aryl monomers and trimethylsilyl-protected dithiols in the presence of base or fluoride ions, they were able to synthesise fluorinated poly(aryl thioethers). In their work, the catalytic amounts of base or fluoride added at the beginning of the reaction deprotect the silyl-thiol generating the nucleophilic thiolate, able to start PFTR. After the first step, for each PFTR event a fluoride ion is released, known for its ability to deprotect silyl groups. This allows for a continuous generation of thiolate until full conversion is reached as summarized in Scheme 3.1. Although providing an excellent example of catalysed PFTR, their approach necessitates the use of protecting groups, which did not only include additional synthetic

41 Results and Discussion

steps for generating the starting material, but also required the use of inert conditions due to the instability of the protecting groups.

X X = H, SiMe X 3 S F R

S F R

F F

F F F F F S R F F F F S F F R F

Scheme 3.1: Mechanism of a self-propagated PFTR. When X is a silyltrimethyl moiety, the fluoride triggers the deprotection of the thiol as reported by Hedrick and co-workers.232 When X is a proton, the fluoride deprotonates the thiol and forms hydrogen fluoride. The thiol subsequently reacts with a pentafluorobenzyl moiety in an addition-elimination mechanism yielding the PFTR product and another free fluoride that can re-initiate the PFTR.

3.3 Results and Discussion

Herein, we introduce a self-propagating pathway for the PFTR that allows for the understoichiometric use of base and avoids protecting groups. A model compound, 3PFB (refer to Scheme 3.2), consisting of three pentaﬂuorobenzyl moieties, was synthesised and reacted in a PFTR with dodecanethiol (DDT), 4-methoxybenzyl mercaptan (MBM) and thiophenol (TP), respectively. Tetrabutylammonium hydroxide (TBAOH), tetrabutylammonium ﬂuoride (TBAF) and tetrabutylammonium bromide (TBABr) as well as 1,8-Diazabicyclo[5.4.0]undec-7-ene (DBU) were investigated as bases. The conversion of the PFTR was followed by 19F NMR spectroscopy over the course of the reaction (refer to Figure 7.4, Supporting Information for Chapter 3 for more details).

42 Self-Propagated para-Fluoro – Thiol Reaction

F F F F F S 10

F F base O O F O O F + HS F O O F 10 F O O F THF F F F F O O a.t. O O

F F F F 10 S F F S 10 F F F F 3PFB

Scheme 3.2: Base initiated PFTR of the 3PFB linker with DDT.

As each single thiol requires one equivalent of base for deprotonation, the fluoride being released from the PFTR needs to be able to deprotonate another thiol species in order to achieve a self-catalysed reaction. In order to investigate this hypothesis, TBAF was selected as fluoride source as the tetrabutylammonium (TBA) counter-ion provides the best means of solubilising and stabilising the fluoride ion in an organic solvent. Further, this concept is expanded to other bases such as TBAOH, which shows structural similarity to TBAF and will form one equivalent of TBAF for every consumed hydroxide and to commonly employed bases in PFTR, for example, DBU. However, the exact nature of the resulting DBU-HF adduct after PFTR is unknown and it was not clear if the species would allow for any further deprotonation of thiols. In order to exclude any possible effects of the TBA cation itself, TBABr was employed as a "blank" base, as the bromide anion does not show any basicity in water or organic solvents. In a first investigation, the reaction of DDT with 3PFB employing 0.10 and 0.50 equivalents of TBABr as base in THF was performed as depicted in Scheme 3.2. As expected, neither 0.10 equivalents of TBABr nor 0.50 equivalents showed any conversion of PFTR in 19F NMR spectroscopy. Reactions with TBAF, TBAOH and DBU, on the other hand, showed 10 or 50% conversion, respectively, after a few minutes or one hour maximum (refer to Figure 3.1 and Figure 3.2). Moreover, in all six cases, the conversion of PFTR increased further than based on the initial amount of base employed. Indeed, when starting with 0.10 equivalents of TBAOH, TBAF, and DBU as base, approximately 30, 25 and 20% conversion, respectively, was achieved after two hours. However, after 24 h, the conversion had reached 35, 30 and 27% conversion, respectively, and did not increase further. Similarly, employing 0.50 equivalents of base led to maximum conversions of 90, 75 and 70% respectively after 24 h. Interestingly, in both cases, the same trend in the reactivity of the bases was observed. This difference in reactivity can be explained by + 256–258 the different pKa values of the bases (H2O 31.4, HF 15, DBU-H 13.9). Although literature values were only found for pKa in dimethylsulphoxide (DMSO), it is assumed

43 Results and Discussion

that the tendencies are the same for THF. The higher pKa value of TBAOH allows for an enhanced reactivity in early stages of the PFTR and a higher pH of the reaction mixture throughout the reaction. Additionally, hydrogen bonding between the protonated DBUH+ and the fluoride are a possible explanation for the lower efficiency of the DBU initiated, under-stoichiometric PFTR. Despite minor differences, in each case, conversions higher than the employed amount of base can be explained by the basicity of the fluoride anion released after PFTR. Exhibiting a pKa of approximately 3.2 in aqueous solution, fluoride shows drastically increased basicity in organic solvents.257 Accordingly, the released fluoride can act as a base itself, thus sustaining the PFTR.

1.0

TBAOH

0.9

TBAF

TBABr

0.8

DBU

0.7

0.6

0.5

0.4 / % PFB group reacted

0.3

0.2 conversion

0.1

0.0

0 10 20 30 40 50

time / hours

Figure 3.1: PFTR conversion of DDT and 3PFB linker vs. reaction time employing 0.10 equivalents of TBAOH (green), TBAF (red), TBABr (orange) or DBU (blue), respectively ([SH]0=75 mM, [D8]THF). The red dashed line indicates the theoretical conversion. Error estimations have been performed and are provided in Supporting Information for Chapter 3, Table 7.1 and Table 7.2. Error bars have been omitted here for better visualisation.

A proposed mechanism for the self-propagated PFTR is depicted in Scheme 3.1. Although we have shown that conversions higher than the employed amount of base can be achieved in PFTR, no full conversion was reached for any of the bases. In order to exclude side reactions of the ﬂuoride released during PFTR, for example, interaction with the glass wall of the reaction vessel, the experiments were repeated in plastic

44 Self-Propagated para-Fluoro – Thiol Reaction

1.0

TBAOH

0.9

TBAF

TBABr

0.8

DBU

0.7

0.6

0.5

0.4 / % PFB group reacted

0.3

0.2 conversion

0.1

0.0

0 10 20 30 40 50

time / hours

Figure 3.2: B) PFTR conversion of dodecanethiol and 3PFB linker vs. reaction time employing 0.50 equivalents of TBAOH, TBAF, TBABr or DBU, respectively, in regards to the thiol ([SH]0=75 mm, [D8]THF.). The red dashed line indicates the theoretical conversion. Error estimations have been performed and are provided in Supporting Information for Chapter 3, Table 7.1 and Table 7.2. Error bars have been omitted here for better visualisation. vials. Moreover, the influence of temperature and concentration has been investigated. However, no significant changes have been observed when the reaction was carried out in plastic vials, at elevated temperatures of up to 50 °C, or at concentrations in the range of 15 to 150 mM in regard to the thiol (refer to Figure 7.5, Supporting Information for Chapter 3). Interestingly, the relative conversion was higher when less base was initially added. When 0.10 equivalents of TBAF were added the conversion reached a maximum of 25% which correlates to a 2.5-fold increase in conversion with respect to the amount of base employed. However, when 0.50 equivalents of base were added, 75% of p-fluorines reacted, corresponding to only a 1.5-fold increase in conversion with respect to the base. Consequently, the effect of the addition rates was evaluated. Here, we investigated the difference in reactivity between adding the fluoride ion at once or in a stepwise addition. Therefore, the 3PFB linker was reacted with DDT in the presence of 0.10 equivalents of TBAF. After 24 hours, the conversion reached 28% (2.8-fold conversion increase

45 Results and Discussion

1.0

single addition of TBAF

multiple addition of TBAF

0.8

0.6 / % PFB groups

0.4 conversion 0.2

0.0

0.1 0.2 0.3

total equivalents of TBAF

Figure 3.3: PFTR conversion of dodecanethiol and 3PFB linker with 0.10, 0.20 and 0.30 equivalents of TBAF,respectively added in a single step (blue) or as a multiple addition of 0.1 equivalent each time (green). The red dashed line indicates the theoretical conversion. Error bars are based on a systematic error in NMR of 10%. For error propagation calculations refer to Table 7.3, Supporting Information for Chapter 3. with respect to base) and did not increase any further. A subsequent addition of another 0.10 equivalents of TBAF led to an increase in conversion up to 57%, which still corresponds to a 2.8-fold conversion increase with respect to base. On the other hand, when PFTR is performed with an initial amount of 0.20 equivalents of TBAF a conversion of 46%, corresponding to only a 2.3-fold conversion increase with respect to base, is achieved Figure 3.3. Finally, a third portion of 0.10 equivalents of TBAF was added to the reaction mixture and an overall conversion of 93% (3.1-fold conversion increase with respect to base) was achieved, whereas a single addition of 0.30 equivalents of TBAF resulted in only 65% conversion (2.2-fold conversion increase with respect to base, refer to Figure 3.3). The red dashed lines in Figure 3.3 refer to the theoretical conversion in case no self-propagated behaviour was present. The experiment evidently proved that a step-wise addition of base improves the conversion of the reaction compared to the same amount of base added at time zero. One hypothesis for the incomplete conversion is possibly associated with changes in the pH level of the reaction mixture with advancing

46 Self-Propagated para-Fluoro – Thiol Reaction

conversion. Although the idea of a pH value is predominantly a concept for aqueous solutions, it can be transferred - to a certain extent - to organic solutions. As each single substitution above stoichiometric conversion generates an equivalent of HF, the pH of the solution is expected to decrease with increasing conversion, thus inhibiting the further deprotonation of the thiol. If this hypothesis is correct, a more acidic thiol would still be deprotonated in the same environment. Thus, in order to investigate the inﬂuence of the thiol or its pKa value on the reaction pathway, the 3PFB linker was reacted with the more acidic MBM and TP thiol derivatives. When performing PFTR with the benzylic thiol MBM rather than the aliphatic thiol DDT in the presence of 0.10 equivalents of TBAF, the maximum conversion increased from 35 to 50% (refer to Figure 3.4) and when employing the aromatic thiol, the conversion reached values of almost 60%, providing evidence for a high pH dependency of the PFTR and the thiol when working in sub-stoichiometric conditions.

1.0

T P

0.9

MBM

DDT

0.8

0.7

0.6

0.5

0.4 / % group/ PFB reacted

0.3

0.2 conversion

0.1

0.0

0 20 40 60 80 100 120

time / hours

Figure 3.4: PFTR conversion vs. time of DDT (purple), MBM (orange) or TP (cyan) with the 3PFB linker in THF. 0.10 equivalents of TBAF were employed as base. The red dashed line indicates the theoretical conversion based on the amount of base added. Dotted lines are intended as guides for the eye. Error estimations have been performed and are provided in Table 7.4, Supporting Information for Chapter 3. Error bars have been omitted here for better visualisation.

47 Results and Discussion

Furthermore, as nucleophilic substitution reactions are enhanced in more polar solvents,259 the influence of the solvent on PFTR was investigated for each thiol. Specif- ically, polar protic solvents were excluded as the solvent would most likely form hydrogen bonds with the nucleophilic thiol, and thus retard the PFTR. On the other hand, solvents with lower polarity would yield even lower conversions than THF or completely prevent the reaction. Thus, the PFTR was performed in N,N-dimethylformamide (DMF) employing 0.1 equivalents of TBAF with respect to the thiol. DMF exhibits a higher polarity than THF (0.386 relative to water, compared to 0.207 for THF) and its beneficial effect towards the conversion when performing PFTR is well known in literature.114,251 However, no reports on the self-propagated approach are available. Similar to the cases mentioned above, the conversion increased from 35 to about 70% for DDT in DMF compared to 35% in THF with 0.10 equivalents of TBAF employed (refer to Figure 3.5, violet purple line). When performing the same reaction with MBM, 84% conversion of para-fluorines was achieved (compared to 50% in THF, orange line in Figure 3.5). Quantitative conversion of the PFTR was accomplished for the reaction of TP with 3PFB in DMF after 4 hours. Given these results, we decided to test an even more polar solvent than THF or DMF. Particularly, we selected DMSO, which exhibits a polarity of 0.444 relative to water. Unfortunately, DDT proved to be immiscible with DMSO, consequently, no PFTR of DDT in DMSO was viable. Nevertheless, the nucleophilic substitution of the 3PFB linker with MBM was quantitative after 48 hours. Notably, when employing the more acidic TP for PFTR in DMSO with the same 1:0.10 ratio (SH:TBAF), the reaction led to over 90% yield within 5 min and subsequently quantitative conversion within half an hour (refer to Figure 3.5, cyan line).

48 Self-Propagated para-Fluoro – Thiol Reaction

TP/THF TP in DMF TP/DMSO

MBM/THF MBM in DMF MBM/DMSO

DDT/THF DDT in DMF DDT/DMSO

1.0

0.9

0.8

0.7

0.6

0.5 / % PFB group reacted % group / PFB

0.4

0.3

0.2 conversion conversion

0.1

0.0

0 20 40 60 80 100 120

time / hours

Figure 3.5: PFTR conversion vs. time of dodecanethiol (purple), 4-methoxybenzyl mercaptan (orange) or benzenethiol (cyan), with 3PFB linker in THF (circle), DMF (triangle) or DMSO (square), respectively. In each case 0.10 equivalents of TBAF in respect to the functional groups were employed. DDT proved to be insoluble in DMSO, thus the reaction data was excluded. The red dashed line indicates the theoretical conversion. Dotted lines are intended as guides for the eye. Error estimations have been performed and are provided in Table 7.4 to Table 7.6, Supporting Information for Chapter 3. Error bars have been omitted here for better visualisation.

3.4 Conclusion

In summary, we introduce a protecting group-free self-propagating para-fluoro – thiol reaction in polar solvents. In our hands, a source of fluoride is not only able to act as a sufficiently strong base in PFTR, but also the fluoride being released during the reaction can sustain the reaction, even when a nitrogen-containing base such as DBU is used initially. Moreover, step-wise addition of base is shown to increase the overall conversion of the reaction by a factor of 1.4 compared to the addition of the exact same amount of base in a single portion. Interestingly, the conversion as well as the kinetics of the sub-stoichiometric PFTR were strongly enhanced by careful tuning of both thiol and solvent. Featuring a sub-stoichiometric use of base, the present approach provides a foundation for even more sophisticated PFTR systems and highlights its importance for organic synthesis as well as polymer modification in presence of base-labile groups.

49 4 Chemiluminescent Read-Out of para-Fluoro – Thiol Reaction Events

4.1 Abstract

We exploit the ﬂuoride that is released via the para-ﬂuoro – thiol reaction (PFTR) to cleave silyl ethers, turning the PFTR reaction into an effective self-reporting CL probe. The cleavage induces chemiluminescence and hence provides an optical read-out for the conversion of the PFTR. The PFTR chemiluminescence read-out is established on small molecule thiols, and subsequently expanded to polymers and networks.

50 Chemiluminescent Read-Out of para-Fluoro – Thiol Reaction Events

4.2 Introduction

In a PFTR, a thiol is deprotonated by a base and subsequently reacts in a nucleophilic substitution reaction with the para-carbon of a pentafluorobenzyl (PFB) moiety releasing a fluoride ion. Due to the increased polarisability of the PFB moiety during the nucleophilic attack in the para-position, the de-aromatised system features an enhanced stability compared to substitution in ortho- or meta-position.115 The versatility of the PFTR and its use in network formation has recently been reported.1,113,114,121,260–262 While the conversion of solution based PFTR has been studied via 19F NMR spectroscopy, quantitative information regarding the fluoride ions eliminated in a PFTR as well as solid state samples, such as networks, is lacking. Despite the synthetic advances in the preparation of polymeric networks, their characterisation on a molecular level remains challenging. Most of the common analytical techniques employed for network characterisation, such as rheology, thermal analyses, gravimetry and swelling ratios, rely on macroscopic quantities and only describe the average properties of the material. To gain information on a molecular level, techniques such as XPS,1,71 ToF-SIMS1 or FTIR are often employed. While these analyses provide more insight into the chemical composition of the material, they require meticulous sample preparation, are only semi- quantitative at best, and provide little to no information on a molecular level such as the number of actual cross-linking points or the presence of structural defects, e.g. loops or entanglements. Johnson and colleagues pioneered approaches accessing the number of loops upon targeted degradation of networks and thorough analysis of the obtained fraction via NMR or SEC.102,263–265 Recently, our group introduced a fluorescence- based approach for the quantification of network cross-links. However, degradation of the networks prior to fluorescence-based read-out was required.40Developing a quantitative method that provides a chemiluminescent read-out for the quantification of the released fluoride ion – and thus the number of PFTR reaction events in any reaction system – is of critical importance. 1,2-Dioxetanes and their phenolic derivatives have been studied by Schaap et al., finding that their deprotonation increased their decomposition and CL output drastically and introduced fluoride-cleavable silyl ethers that also produce deprotonated phenols upon deprotection.2,200,201 One of the best known 1,2-dioxetanes is perhaps the sterically hindered adamantyl-substituted phenolic 1,2-dioxetane,2 commonly referred to as Schaap’s dioxetane. Numerous derivatives of Schaap’s dioxetane have been designed as CL probes for various analytes, such as fluoride ions, hydrogen peroxide, hydrogen sulfide or peroxynitrite, and many more, by varying the protecting or leaving

51 Results and Discussion

group.17,183,211,213,214,266,267 A clear advantage is that, unlike fluorescence, CL does not require an external light source and thus provides a superior signal-to-noise ratio. Herein, we introduce the quantification of reaction events occurring during PFTR using a CL read-out. While detection of free fluoride via Schaap’s dioxetane has been reported,266–268 to the best of our knowledge this is the first report of it being utilised for reaction monitoring and quantification. In our concept, the fluoride ion eliminated in the PFTR acts as a trigger for CL of phenolic 1,2-dioxetane and thus as a quantitative read-out method e.g. during network cross-link formation, without the necessity of degrading the network. A fluoride ion is released for each arm of the crosslinker reacting with a thiol and subsequently is capable of cleaving a silyl ether off Schaap’s dioxetane triggering CL (Scheme 4.1).

4.3 Results and Discussion

To demonstrate our read-out technique, a small molecule study was carried out employing a trifunctional pentafluorobenzyl linker 3PFB along with a series of monofunctional thiols. Tetrabutylammonium hydroxide (TBAOH) was chosen as the base for the deprotonation of thiols, as it subsequently provides tetrabutylammonium fluoride (TBAF), which is a suitable source of fluoride for deprotection of the phenol moiety of the CL probe. To avoid unwanted triggering of the CL by TBAOH, equimolar quantities of thiol and base were used, ensuring the base is completely reacted. To establish a guideline, the CL output of TBAF in acetonitrile (ACN) was recorded for ten solutions with fluoride concentrations ranging from 1 to 10 mM. Each of the fluoride samples was mixed with a 12 mM solution of the CL probe, thus ensuring that every fluoride is able to react, and that varying amounts of the CL probe do not affect the total emission. The CL was subsequently recorded as a function of time over the complete emission range of the probe (refer to Supporting Information for Chapter 4, Figure 7.11). The CL emission of each solution was integrated over the complete measurement time to obtain the total emission. Subsequently, the total emission was plotted vs the concentration and a linear fit was applied. The obtained slope of 6.13×109 cts·L mmol-1 for the linear fit was subsequently used for comparison with the CL of PFTR read-outs (Figure 4.1, black squares and line). Additionally, 10 mM solutions of TBA bromide, chloride, cyanate, hydrogensulfate, iodide and perchlorate were investigated as CL triggers. However, only TBA cyanate showed minor CL (2.3% compared to equimolar amounts TBAF). None of the other alternative anions showed significant CL (refer to Supporting Information for Chapter 4, Figure 7.19). Next, the PFTR of

52 Chemiluminescent Read-Out of para-Fluoro – Thiol Reaction Events + tBDMS-F adamantanone probe, thus triggering its + as a base (left side of the CL O O O TBAOH O O CL probe O O Cl Cl Si 'Schaap's dioxetane' TBA O O O linker and a thiol employing PFB R TBAF + HS TBAOH between the 3 , which subsequently deprotects the phenol moiety of the PFTR TBAF R F S R = F F F F F F F O F F F O O F F S O O R F 3PFB F F F O F F F F F General reaction scheme of the F F F S R F F Scheme 4.1: scheme). A ﬂuoride isdecay released and in subsequent the emission form (right of side of the scheme).

53 Results and Discussion

adamantyl thiol (AT) was investigated. 3PFB was reacted with different amounts of AT, and equivalent base, ranging from 0.1 to 0.9 equivalents relative to PFB moieties. Aliquots of each sample were submitted for analysis via 19F NMR spectroscopy and LC to determine the conversion of para-fluorines (refer to Supporting Information for Chapter 4, Equations (7.4) to (7.9)). In the 19F NMR spectra, four new resonances at -131.4, -144.1, -146.4 and - 159.5 ppm emerged. While the resonances at -131.4 and -144.1 ppm could be readily assigned to the meta- and ortho-fluorine of the reacted 3PFB linker, the appearance of the resonances at -146.4 and -159.5 ppm was unforeseen. However, the unexpected resonances could be assigned to the hydroxy substituted 3PFB linker. Due to the relatively high pKa value of AT, it seems likely that the thiol is not completely deprotonated, and the residual hydroxide will undergo nucleophilic aromatic substitution on 3PFB. Despite this side reaction, 19F NMR and LC of the PFTR indicate that the employed thiol has reacted quantitatively. The complete conversion of thiol can be explained by earlier findings of our group, which indicated that in polar aprotic solvents, such as the employed ACN, the fluoride eliminated from PFTR is sufficiently basic to deprotonate another thiol, enabling the use of an under-stoichiometric amount of base.260 LC traces had to be deconvoluted before integration of the signals of 3PFB as well as the mono-, di-, and trisubstituted 3PFB, allowing for the determination of the conversion and therefore the total concentration of eliminated fluoride. In addition, an aliquot of each sample was mixed with the CL probe in ACN and the emission profile was recorded as a function of time (refer to Supporting Information for Chapter 4, Figure 7.12 to Figure 7.18). An excess of the CL probe compared to the maximum number of para-fluorine atoms was used to ensure full read-out of fluoride. This is important as non-chemiluminescent, degradation of the CL probe upon reaction with thiols has been reported recently.269 The integrated emission was plotted against the amount of fluoride as calculated from 19F NMR and LC, and against the theoretical yield calculated from the quantity of employed thiol and base, and finally, linear fits were performed (refer to Figure 4.1, red data). While the slope is slightly higher (6.40×109 cts·L mmol-1, +4.4%) than that of the TBAF samples used as a reference (black line), a linear correlation between the emission vs conversion (blue line) is evident, thus demonstrating the feasibility of the CL probing system for quantification of the PFTR. Moreover, the total integrated CL emission of the reaction solutions after addition of the CL probe correlated linearly with the conversion of thiol only, not with the total amount of fluoride released. In other words, the fluoride released from the hydroxide substitution does not affect the CL read-out as it deprotonates the residual thiols producing HF, which no longer contributes to CL. An

54 Chemiluminescent Read-Out of para-Fluoro – Thiol Reaction Events

overview of the described reaction cascade is provided in the Supporting Information for Chapter 4, Scheme 7.1. Consequently, the number of PFT reactions can still be determined accurately, despite the undesirable side-reaction.

8x10

TBAF

10 AT, average 7x10

mTrEGT, average

10 BT, average

6x10 / cts /

5x10

4x10

3x10 totalemission

2x10

1x10

0 1 2 3 4 5 6 7 8 9 10 11

conc [F ] / mM

Figure 4.1: Total CL emission vs concentration of ﬂuoride (averaged over NMR, LC and expected conversion) compared to the TBAF guideline (black squares) of adamantyl thiol (AT) (red), methoxy triethylene glycol thiol (mTrEGT) (blue), and benzyl thiol (BT) (green). Solid lines represent linear ﬁts of the obtained data.

To confirm these findings, two more thiols, methoxy triethylene glycol thiol mTrEGT and benzyl thiol BT were employed for a series of PFTRs in a similar fashion to AT and analysed via 19F NMR and LC. While the hydroxy substitution is evident for mTrEGT, it was not observed for BT. This can be attributed to the lower pKa value of BT, compared to AT and mTrEGT, which facilitates complete deprotonation of the thiol. Nevertheless, BT showed more scattering in the yields calculated from 19F NMR and LC, and the linear fit of the yields exhibited a more significant deviation from the TBAF reference (5.43×109 cts·L mmol-1, -10.4%, green). mTrEGT, on the other hand, provided a linear fit with a slope of 5.90×109 cts·L mmol-1 ( 3.7%, blue) which is in excellent agreement with the TBAF reference and AT. In contrast to commonly employed techniques for network characterisation, the

55 Results and Discussion

current CL read-out provides a non-destructive reaction event counter. To demonstrate this fact, a 2000 g·mol-1 poly(ethylene glycol) bearing a thiol end-group PEG-SH was selected as a test polymer. In a similar manner as above, 3PFB was reacted with different amounts of PEG-SH and TBAOH (0.1, 0.3, 0.5, 0.7, and 0.9 equivalents with respect to PFB moieties) in 10 mL of ACN. Subsequently, a small aliquot was taken from each sample, the CL probe was added, and the total emission was recorded as a function of time (refer to Supporting Information for Chapter 4 Figure 7.26). As the CL read-out proved to be accurate for small molecules, the fluoride concentration in each polymer sample was back calculated based on the total emission of each sample and the linear fit of the TBAF solutions. The results were verified by 19F NMR and SEC. Here too, 19F NMR spectroscopy showed the occurrence of the hydroxy substitution as a side reaction, but it did not interfere with the CL read-out. Deviations in the emission vs PFTR conversion plot for both 19F NMR and SEC from the TBAF standard (7.61×109 cts·L mmol-1, +24.1%) can be attributed to an increased viscosity due to the presence of polymers in solution. Baader and co-workers previously showed that the viscosity of solvents has a significant effect on the quantum yield of CL reactions.245,270 Therefore, the concentration calculated from the emission shows increasing deviations from the concentrations obtained from 19F NMR and SEC with increasing polymer concentration and size, and thus increasing viscosity, as depicted in Figure 4.2. The actual number

of PFTR (nPFTR) events can be readily calculated from the concentration of fluoride (cCL) determined from CL via nPFTR = cCL·V. Thus obtaining 2.81, 29.6, 54.4, 85.6 and 120 µmol PFTR events for experiments with 10, 30, 50 60 and 90 µmol respective expected events, ignoring any changes in viscosity. To further expand the technique from polymers to networks, 3PFB was employed in a PFTR with five different amounts of 1,4-benzendimethane thiol (BDT) and triethylene glycol dithiol (DODT), respectively, in as little solvent as possible. After reaction comple- tion, the networks were swollen in 10 mL of ACN and sonicated to remove the released fluoride from the network and transfer it into solution. An aliquot of each of the supernatant solutions was employed for CL read-out to determine the number of covalent bonds in each network. Furthermore, the solutions were analysed via 19F NMR and SEC to determine if any soluble fraction was present. The low conversion samples show a small soluble fraction, indicated by the golden and cyan regions in Figure 4.3, whereas at higher conversions no significant amount of soluble fraction was detected (refer to Supporting Information for Chapter 4, Figure 7.27 to Figure 7.30). Thus, the CL read-out of samples at lower conversions also contains a small portion of emission from the soluble fraction and the actual number of PFTR events in the polymer network will be

56 Chemiluminescent Read-Out of para-Fluoro – Thiol Reaction Events

8x10

TBAF

10 PEG-SH, average

7x10

PEG, back calculation

6x10 / cts /

5x10

4x10

3x10 totalemission

2x10

1x10

0 1 2 3 4 5 6 7 8 9 10 11

conc [F ] / mM

Figure 4.2: Total CL emission vs concentration of ﬂuoride (averaged over NMR, LC and expected conversion) compared to the TBAF guideline (black squares) and the back calculated conversion (stars) of the PFTR of polyethylene glycol thiol (PEG-SH). Solid lines represent linear ﬁts of the obtained data.

lower than that calculated from CL. The effect of concentration on the network formation is well known, although the exact threshold value for the formation of soluble networks is unique for each integrated CL emission and the linear fit of the TBAF solutions. For both thiols, the fluoride concentration obtained from CL read-out in each of the five samples aligns well with the expected values. By immersing the networks produced from each reaction in 10 mL of solvent, the actual number of PFTR events can be readily calculated

from the concentration of ﬂuoride as determined from CL via nPFTR = cCL·V. Thus, 0.19, 0.57, 1.18, 1.83 and 2.42 mmol PFTR events per gram of network were determined for the ﬁve BDT networks, respectively (golden stars). For the DODT networks, 0.13, 0.55, 1.04, 1.65 and 2.34 mmol PFTR events were obtained per gram of network, respectively (cyan stars).

57 Conclusion

8x10

7x10

6x10 / cts /

5x10

4x10

3x10 totalemission

TBAF

2x10

BDT, theo

BDT, back calculation

1x10

DODT, theo

DODT, back calculation

0 1 2 3 4 5 6 7 8 9 10 11

conc [F ] / mM

Figure 4.3: Total CL emission vs theoretical ﬂuoride concentration (dots) and the concentration as back calculated from the total emission (stars) for the read-out of PFTR networks employing 1,4-benzendimethane thiol (BDT) (gold) and triethylene glycol dithiol (DODT) (cyan) as bis-thiols. Solid lines represent linear ﬁts of the obtained data. The rectangles indicate areas of ’soluble network’ formation for the according thiols.

4.4 Conclusion

We introduce a chemiluminescent read-out system that allows for the quantitative determination of PFTR reaction events. Upon formation of each PFTR-link, a fluoride is released that can remove the silyl-ether of a CL probe triggering CL. A three-arm PFB linker was employed for PFTR with varying amounts of three thiols, i.e. adamantyl thiol, methoxy triethylene glycol thiol and benzyl thiol. The concentration of fluoride in the solutions was determined from the conversion according to 19F NMR and LC. After the addition of the CL probe, the emission was recorded and plotted vs the concentration of fluoride. We demonstrate that the emission of the PFTR solutions is in good agreement with the emission of pure TBAF solution used for comparison. The feasibility of the CL read-out was confirmed by formation of a three-arm star-polymer, where the total CL emission was recorded and the concentration of fluoride back calculated. The deter-

58 Chemiluminescent Read-Out of para-Fluoro – Thiol Reaction Events

mined fluoride concentration was shown to be coherent with the conversions obtained from 19F NMR and SEC. Finally, the CL read-out was assessed on more complex systems such as network formation via PFTR. The resulting concentration of free fluoride and the absolute number of PFTR events can be determined via a chemiluminescent read-out. While solution-based NMR and LC require sample volumes in the range of 10-7 to 10-6 mol, the presented CL based approach allows for the detection of down to 5·10-9 mol of fluoride while providing similar accuracy. Moreover, the sensitivity of the presented approach can readily be increased by increasing the integration time of the instrument. Finally, our CL read-out allows for the quantification of reaction events without the necessity of dissolving or degrading networks. The approach thus provides an advanced optical characterisation method for the counting of reaction events in PFTR systems.

59 5 Chemiluminescent Unfolding of Single-Chain Nanoparticles

5.1 Abstract

We demonstrate the light-induced, crosslinker mediated collapse of linear polymer chains into SCNPs capable of self-reporting their unfolding. The crosslinker entails a phenyl oxalate (PO) motif allowing for the targeted degradation of the SCNPs via addition of

H2O2 that triggers CL. The time-dependant CL emission can serve as a guide to follow the time dependent unfolding of the SCNPs, allowing for a qualitative assessment of the underlying mechanism.

60 Chemiluminescent Unfolding of Single-Chain Nanoparticles

5.2 Introduction

Over the past two decades, the manipulation of single polymer chains into macromolecular architectures, aiming to mimic proteins and their unique behaviour, has been one major focus in polymer chemistry.128,271,272 While substantial progress was made mimicking the precision of primary structures, the secondary and tertiary structures found in nature are still unrivalled.273 A vast synthetic toolbox allows for the collapse of single polymer chains into SCNPs via covalent or non-covalent bond formation.25,274 Applica- tions of SCNPs span from catalysis28,145 to nanoreactors,26,170 and nanomedicine27,275 to sensors.139 However, thus far, the mechanism of folding and unfolding of SCNPs has only scarcely been investigated. Previously, Diesendruck and Pomposo and co-workers investigated the scission of SCNPs due to shear force upon sonication, obtaining insight on the inﬂuence crosslinking has on the mechanical stability of SCNPs.276–280 As SCNP folding requires highly dilute conditions, most analytical methods are not applicable for on-line monitoring of the intricate folding process. Quantitative investigation of SCNP 281,282 unfolding has mostly been achieved by comparing hydrodynamic diameters (dHs). To the best of our knowledge, only the group of Meijer conducted further in-depth and on-line investigations of SCNP unfolding by mechanical unfolding employing single- molecule force spectroscopy obtaining insight on the folded structure, as well as the course of unfolding.174 Our own group pioneered SCNPs that allowed for a self-reporting ﬂuorescence read-out after folding and unfolding. However, no on-line tracking of the (un)folding process was possible.283 Herein, we introduce a method via which unfolding of SCNPs can be monitored using CL. CL is the generation of light as a result of a chemical reaction. Gnaim and Shabat employed self-immolative chemiluminescent polycarbonates allowing for the on-line tracking of depolymerisation.183 One of the most commonly employed CL reactions is the PO-CL reaction. Besides its common use in commercial glowsticks, PO-CL has been employed for various analytical methods such as food analysis,284,285 medicinal and biological sensors,286–288 and immuno-essays.289 PO-CL is considered

a three component CL reaction, as it requires a phenyloxalate (PhOx), H2O2 and a suitable fluorophore. In the first step, H2O2 reacts with the PhOx to form a meta-stable 1,2-dioxetanedione, commonly referred to as HEI. Although the HEI will readily degrade, it cannot emit light by itself. In the presence of a suitable fluorophore, however, it can

transfer its energy to the fluorophore and degrade to CO2 (refer to Scheme 5.1). The fluorophore in turn is transferred to a singlet excited state and finally relaxes back to its ground state via the emission of light.237 The quantum yield of such PO-CL depends

61 Results and Discussion

on various parameters, such as the exact structure of the PhOx, the nature of the fluorophore (i.e. its redox potential), the proximity of the PhOx to the fluorophore, as well as the viscosity of the reaction mixture.290 Unlike fluorescence, CL does not require a photoexcitation source, and therefore exhibits significantly higher signal-to-noise ratios.

R O OH R kHEI O O R O O + H2O2 + OO I O PhOx HEI PhOH

S1 O O kcat + CO2 + OO II

HEI Pyr Pyr*

S1 kfluo III

Pyr* Pyr

Scheme 5.1: Overview of the reaction steps of the PO-CL as well as key steps in the chemiluminescent unfolding of SCNPs.

5.3 Results and Discussion

While our group has developed a solid state chemiluminescent read-out system based on PO-CL earlier,3 the current study introduces a solution-based PO-CL read-out system that allows for tracing of SCNP unfolding via PO-CL. Thus, a set of three polymers, bearing o-MBA moieties as the photoinduced crosslinking units and pyrene moieties as the ﬂuorophores, were synthesised via RAFT. The obtained polymers were crosslinked to generate SCNP1-3 and were subsequently degraded resulting in a measurable CL event. While the total amount of o-MBA and pyrene moieties was varied for different polymers, the ratio between o-MBA-functionalities and ﬂuorophore was, where possible, kept constant in order to minimise the number of parameters affecting the chemiluminescence quantum yield. The synthesised parent polymers P1-3 were characterised via SEC and 1H NMR to determine molecular weights and comonomer ratios. DOSY was employed

to obtain dHs. Table 5.1 provides an overview of molecular weights and monomer ratios

62 Chemiluminescent Unfolding of Single-Chain Nanoparticles

of the parent polymers. In order to crosslink the o-MBA moieties, a diphenyl oxalate, bearing maleimide- functionalities in the para-position, was synthesised according to literature procedures.3 Chlorine atoms in the ortho-position of the phenol decrease the electron density of the phenol, increasing both the kinetics and the quantum efficiency of the PO-CL. Previous studies showed that the herein employed bis(2,6-dichloro-4-N-maleimido) phenyl oxalate (MDCPO) provided only slightly lower light intensities compared to TCPO, which is commonly used in glow sticks and analytical methods.3,286,289,291,292 SCNPs were prepared by dissolving polymers P1-3 and MDCPO in dry, deoxygenated dichloromethane (DCM) (0.1 mg·mL-1 polymer concentration) and irradiating the solution with UV-A lamps for one hour. The resulting solutions containing SCNP1-3 were evaporated to dryness and purified via preparative SEC (Sephadex® LH 20, THF) to remove any unreacted crosslinker. SCNP1-3 were subsequently analysed via SEC, 1H NMR, and DOSY to confirm the successful folding (refer to Figure 5.1A, as well as Figure 7.32, Figure 7.33 and Figure 7.35 to Figure 7.40, Supporting Information for Chapter 5). The disappearance of the 1H NMR resonances at 10.58 and 2.50 ppm, associated with the aldehyde and methyl groups respectively, along with the appearance of the resonance at 7.59 ppm, confirm the successful and complete conversion of the o-MBA moiety. Moreover, the absence of a resonance at around 6.9 ppm confirms that no unreacted maleimide remained in solution. The compaction degree ranged from 21% to 33% according to SEC and from 7.4% to 16% according to DOSY, depending on the chain-length and the relative number of crosslinking moieties. It should be noted that, in the presented case, DOSY is the more reliable method as the functional groups on the polymer interact with the column material, potentially influencing the SEC results. Of the two similar weight parent polymers P1 and P2, P1 shows the higher compaction, which is expected as it contains more crosslinking units. Despite the parent polymers P2 and P3 exhibiting the same relative amount of crosslinkable units, the higher molecular weight polymer P2 produces stronger compaction. It should be noted that the employed crosslinker is relatively large and only exhibits a few freely rotatable bonds, which likely inhibits compaction beyond a certain polymer size, or above a certain crosslinking content, due to steric hindrance. Subsequently, SCNP1-3 were dissolved in dry DMF at concentrations of exactly -1 1 mg·mL . 6 µL of a 30% H2O2 solution was added to 100 µL aliquots of the SCNP solutions, and the CL intensity in the range of 360 to 700 nm was recorded as a function of time. Under these conditions, double the amount of crosslinker per SCNP does not necessarily mean double the amount of crosslinker per sample. After 60 min, a

63 Results and Discussion * m/n/o / O [c] H HN d nm O O 9 OH O O O O N -1[a] O / Cl Cl mol · n HO M kg 2 2 / O 2 DMF H - CO [c] H d nm m/n/o O -1[a] HN / mol · n O O 9 kg M SCNP uSCNP OH O / O O O N [c] O H d nm Cl Cl O O [b] o 1 h DCM UV-A Cl O [b] O Cl O O N n N O O O Cl O Cl [b] m/n/o Characterisation of parent polymers P1-3, SCNP1-3 and unfolded SCNPs (uSCNPs)1-3. m O HN Table 5.1: -1[a] O O 9 / mol · O n O M kg O Molecular weights were obtained viaMonomer chloroform ratios SEC. were obtained via 1H NMR. Hydrodynamic diameters were obtained via DOSY NMR. O P1 P2 33.7P3 33.1 182 (81%)28.3 30 (13%) 248 (91%) 13 (6%) 15 206 (6%) (91%) 11.0 14 7 (6%) (3%) 22.9 (-32%) 7 10.3 (3%) 9.3 (-16%) 22.1 12.6 (-33%) (-45%) 9.5 9.1 (-12%) 22.5 (-21%) 9.8 (+5.4%) 17.6 (-20%) 8.8 (-7.4%) 14.7 10.0 (-35%) (+9.9%) 9.3 (+5.7%) [a] [b] [c] Entry

64 Chemiluminescent Unfolding of Single-Chain Nanoparticles

second portion of H2O2 was added to ensure complete degradation of the PO-linker. The CL emission of each SCNP was measured ﬁve times and then averaged for analysis. The mean time-dependent emission data and the standard error of each SCNP are depicted in Figure 5.1B. For the single emission proﬁles refer to Supporting Information for Chapter 5, Figure 7.44 to Figure 7.49. An initial spike in CL intensity was observed after approximately one minute followed by a rapid decrease, which was attributed to

the addition of H2O2 and a resulting local high concentration prior to complete diffusion throughout the sample solution. The main CL emission peak – and therefore also the

maximum rate of unfolding of SCNP1-3 – is observed ten to twenty minutes after H2O2 injection. Interestingly, the most tightly compacted SCNP, SCNP, shows two distinct CL maxima after about eight and 18 min. While we do not fully understand the underlying causes of the additional peak, we propose it is due to the high number of crosslinking points in this polymer leading to a high degree of entanglement, and hence a stepwise unfolding process is observed as the degree of freedom increases. After approximately

one hour, the emission decreased to baseline level and a second portion of H2O2 was added, causing a third increase in CL intensity. The area under this third peak, however, only constituted 2-7% of the overall CL emission for any of the SCNP1-3. To obtain deeper insights into the unfolding kinetics, the emission data was processed with the PREDICI® software and rate coefﬁcients were estimated using a ﬁtting procedure. The chemiluminescent unfolding of SCNP1-3 consists of three consecutive

steps, as depicted in Scheme 5.1. First, the PhOx linker is cleaved by H2O2, and the HEI 1,2-dioxetanedione is formed. In the presence of pyrene, the HEI subsequently

degrades to CO2, exciting the pyrene moiety. Finally, the excited pyrene emits a photon and relaxes to its ground state. As luminescence is relatively fast compared to chemical 12 6 reactions (kfluo ≈ 10 to 10 s), the final step can be regarded as instantaneous, and the pyrene a mere catalyst in this system. Moreover, the ratio of crosslinker to pyrene is approximately one. Thus, the chemiluminescent unfolding of SCNP1-3 can be approxi- mated by a two-step process; the first step producing the HEI follows 2nd order kinetics with a rate constant kHEI, and the second step describing the HEI degradation follows st 1 order kinetics with the rate constant kcat and produces photons as the final reaction ® product. In order to obtain a qualitative estimate of kHEI and kcat via the PREDICI software, the emission time profile of SCNP2 was selected as a model system and integrated to obtain the cumulative number of photons. Based on the experimentally

obtained cumulative emission, a parameter ﬁt of kHEI and kcat was performed and the cumulative emission was simulated for each SCNP (refer to Figure 5.1C and to Supporting Information for Chapter 5, Figure 7.50 to Figure 7.55 for SCNP1&3). The HEI

65 Results and Discussion 2 (solid golden line) SCNP1 SCNP2 SCNP3 SCNP 1-3 (dotted lines) in the diffusion / min

/ cts / intensitry cumulative uSCNP 1.6G 1.4G 1.2G 1.0G 0.8G 0.6G 0.4G 0.2G 0.0G emission upon unfolding of 10 20 30 40 50 60 70 80 90 100 110 120 sim. [PhOx] sim. [Diox] sim. [hv] [hv] exp. CL 0 1-3 (solid lines) and

1.5E+06 1.0E+06 5.0E+05 0.0E+00

/ cts s cts / Intensity -1 SCNP / s time 12 P1 P2 P3 SCNP1 SCNP2 SCNP3 uSCNP1 uSCNP2 uSCNP3 11 10 20 30 40 50 60 70 80 90 100 110 120 1-3. C) Simulated cumulative of P1-3 (dashed lines), 0 / nm H

1.0 0.8 0.6 0.4 0.2 0.0 1-conversion r 10 SCNP DOSY NMR 9 8

1.0 0.8 0.6 0.4 0.2 0.0

/ a.u. / intesity number normalised A) Projection of the Figure 5.1: dimension conﬁrming successful compaction andof subsequent the expansion chemiluminescent of unfolding thecompared of polymer to chain. the B) experimental Meancrosslinker cumulative time-dependant and emission emission the (dashed proﬁles HEI, golden respectively. line). The green and purple line depict the population of the the

66 Chemiluminescent Unfolding of Single-Chain Nanoparticles and catalytic rate coefficients for SCNP2 were found to be 160±30 s-1 and 470±4 s-1, respectively. It is important to note that the determined rate coefficients provide a qualitative description of the emission behaviour only and do not constitute a full kinetic analysis. Despite this, the simulated cumulative emission of SCNP2 was in good agreement with the measured data (refer to Figure 5.1C and to Supporting Information for Chapter 5, Figure 7.50 to Figure 7.55 for SCNP1&3). The estimated rate coefficient shows a rapid formation of the HEI dioxetanedione, indicating a fast unfolding of SCNPs1-3. The rate coefficient reported here for the hydrolysis of the PhOx linker by H2O2 is several orders of magnitude larger than literature values for the hydrolysis of similar small molecule POs with water.230,293 This can likely be attributed to the limited rotational freedom of the employed MDCPO linker. Due to the relatively high stiffness of the crosslinker, the unfolding of such SCNPs is therefore energetically favoured and the hydrolysis highly accelerated. According to literature, reaction rates for the catalysed decomposition of the HEI are highly system specific, with reported values varying by over four orders of magnitude.241 The simulated cumulative emission was differentiated with respect to time to obtain a simulated time-dependant emission profile, which was comparable to the experimental emission profile (refer to Supporting Information for Chapter 5, Fig- ure 7.50 to Figure 7.55). As expected, the simulated emission profile deviates from the experimental data at the start of the trace due to the initial H2O2 injection peak. Finally, the unfolded SCNPs uSCNP1-3 were analysed via SEC, DOSY and 1H NMR measurements. A shift of the 1H NMR resonance at 7.59 ppm, associated with the aromatic protons in meta-position of the PhOx, to 7.28 ppm confirmed full crosslinker degradation and therefore complete unfolding of SCNP1-3 (refer to Figure 7.34, Support- ing Information for Chapter 5). Moreover, DOSY confirmed an increase in hydrodynamic diameters of 5.4, 9.9, and 5.7% for uSCNP1-3 respectively. As expected, for the two SCNPs with the same relative number of crosslinking units, SCNP2 which was formed from a larger parent polymer, had the larger increase in hydrodynamic radius, compared to SCNP3. The same expected trend was not observed for SCNP1 and SCNP2, where the SCNP with the lower percentage of crosslinking points produced the largest increase in hydrodynamic radius. This unexpected trend is not expected to have an influence on the CL output however, since DOSY confirmed full unfolding was still achieved. It does however demonstrate the limitation of the presented CL method as a means of predicting hydrodynamic diameters. In contrast to the results of DOSY, SEC indicated a further decrease in apparent molecular weight for all SCNPs, as a result of the unfolding, despite an expected increase in actual molecular weight due to the residual maleimide moieties from the oxalate crosslinker. The observed decrease in apparent molecular weight can

67 Conclusion

be attributed to the formation of additional hydroxy functionalities formed during folding and unfolding of the SCNPs. The additional hydroxy functionalities are likely to interact with the SEC column material causing the uSCNP1-3 to elute slower than the respective SCNP, suggesting a decrease in apparent molecular weight. This is evidenced by the signiﬁcant tailing observed in all uSCNP traces, which is more pronounced in polymers with more hydroxyl groups (Figure 7.35 to Figure 7.37, Supporting Information for Chap- ter 5). For these reasons, and those outlined above, we consider DOSY to be a more reliable way of characterising the unfolding of these particular SCNPs.

5.4 Conclusion

In conclusion, the light induced folding of an o-MBA bearing polymer with a degradable maleimide-crosslinker was pioneered on a set of well-deﬁned polymers. The choice of crosslinker allowed for light induced intramolecular crosslinking, followed by the emission

of light via PO-CL upon unfolding of the SCNPs with H2O2. The time-dependent emission proﬁle of the unfolding SCNP2 was employed as a model system to probe the kinetics of SCNP unfolding. We thus submit that chemiluminescence is an ideal tool for the on-line monitoring and for the further kinetic investigations of SCNP unfolding processes, including for polymer backbones with different stiffness.

68 General Discussion

6 General Discussion

A multitude of self-reporting systems based on colour changes, fluorescence or chemiluminescence (CL) has been reported over the past years. However, while self-reporting systems that use a change of colour as an indicator provide a facile read-out for changes in pH or thermal treatment and straightforward method for the localisation of mechanical stress, they are only qualitative, i.e. they signal e.g. mechanical or thermal stress, but barely give any indication on the quantity of stress, heat or pH change. In contrast, phosphorescence as a read-out in self-reporting systems allows for localisation and quantification of stimuli, but requires more sophisticated set-ups including a light source and possibly a detector, if the emission wavelength is outside the visible wavelength regime. CL, however, enables the localisation and quantification of stimuli, does not require an external light source and is visible to the naked eye. Still, CL has mostly been used in the context of detecting and localising analytes, toxins or reactive oxygen species and little effort has been made in the quantification of stimuli or the use of CL as a means of on-line reaction monitoring and kinetic analysis.10,290,294,295 The present thesis thus aimed at providing a straightforward tool for the quantification of reaction events on the one hand – especially in systems where conventional analytical methods fail – and, on the other hand, probe for the use of CL as a tool for the evaluation of reaction rates and kinetics.

6.1 Summary and Key Outcomes

Prior to the in-depth investigation of para-fluoro – thiol reaction (PFTR) event quantification via CL, the mechanistics of the PFTR as such demand closer examination, as previous studies suggested that the PFTR proceeded base-catalysed without providing definite proof or disproof.114 If the PFTR was indeed base-catalysed, the exact nature of this catalysis could have an influence on the fluoride that is released during a PFTR and thus on the CL read-out. Chapter 3 addressed this issue by first investigating the PFTR of dodecanethiol (DDT) and a pentafluorobenzyl (PFB) moiety employing 0.5 or 0.1 equivalents of three different bases in tetrahydrofuran (THF), namely 1,8-

69 Summary and Key Outcomes

Diazabicyclo[5.4.0]undec-7-ene (DBU), tetrabutylammonium hydroxide (TBAOH) and tetrabutylammonium fluoride (TBAF). While the PFTR conversion according to 19F nuclear magnetic resonance (NMR) was significantly higher than expected in all cases, it still plateaued without reaching full conversion. Importantly, when tetrabutylammonium bromide (TBABr) was employed as a “blank” base, no conversion was observed at all. We proposed a mechanism, suggesting that the fluoride released from the reaction was indeed sufficiently basic to deprotonate another thiol, thus propagating the PFTR to higher conversions than initially suggested. As HF must be formed simultaneously, changes in the pH value during the reaction can be assumed, leading to a higher but not complete conversion of PFB moieties. In addition, it was found that adding the base in several smaller portions in 24 hour intervals led to higher conversions than adding the same amount of base as a single portion. Considering a pH dependency of the fluoride propagated PFTR, in a next step, the PFTR of two more acidic thiols, 4-methoxybenzyl mercaptan (MBM) and thiophenol (TP), was investigated using only TBAF as a base. Here again, the conversion of PFB moieties according to 19F NMR was higher than the amount of base added and, as expected, the higher the acidity of the thiol, the higher the conversion. In a last set of experiments, the influence of the solvent on the conversion of the self-propagated PFTR was investigated by employing the three thiols for PFTR with PFB moieties in solvents of varying polarity, using only 0.1 equivalents of TBAF as base. Not only did the conversion follow the same trend as before in each of the solvents, i.e. the higher the acidity of the thiol, the higher the conversion, the experiments also clearly demonstrated an increase of conversions in more polar solvents. In other words, while the least acidic DDT only provided 35% yield in THF when 0.1 equivalents of TBAF were used initially, employing TP with the highest acidity of all investigated thiols for PFTR in highly polar dimethylsulphoxide (DMSO), the conversion was quantitative within 30 min. The study clearly showed that the fluoride released from the reaction is sufficiently basic to propagate the PFTR when understoichiometric amounts of base were used. As a result, in order to provide a read-out for the PFTR via CL, stoichiometric amounts of base were necessary, as self-propagation of the PFTR would remove the fluoride from the reaction in the form of HF. With the findings of the preliminary mechanistic study at hand, the development of a CL read-out for the PFTR was investigated in Chapter 4. The CL probe that was chosen for the detection of PFTR events was a Schaap’s dioxetane, bearing a silyl ether that can be cleaved by fluoride ions, thus triggering CL. To establish a proof-of- principle, three small molecule thiols were reacted with a three-arm PFB linker 3PFB. The reactions were carried out at various concentrations ranging from 1 to 10 mM

70 General Discussion

of thiol and equimolar amounts of base to prevent self-propagation of the PFTR and thus false CL read-outs. The conversion of each PFTR was determined via 19F NMR and liquid chromatography (LC). After addition of the CL probe, the emission was recorded as a function of time and integrated. When the integrated emission was subsequently plotted against the conversion, a linear correlation of the two became apparent for all three thiols and was in good agreement with pure TBAF samples of identical concentrations. To rule out any interference from other ions, several other tetrabutylammonium (TBA) salts were investigated for their ability to trigger CL, however, none of the investigated salts showed significant amounts of CL. Moreover, it was found that a side reaction - substitution of the PFB moiety with hydroxide ions - occurred, however, the substitution of PFB moieties with thiols remained unaffected. This could be attributed to the findings of the preliminary study of the PFTR, as the fluoride released from the hydroxy substitution deprotonates a thiol, thus propagating the PFTR and being removed from the system as HF. Therefore, only the thiol substitution accounts for the CL read-out, but not the hydroxy substitution. In a next step, the principle was applied to a more sophisticated macromolecular system - namely a three-arm star polymer. For this purpose, 3PFB was reacted with a 2,000 g·mol-1 polyethylene glycol thiol (PEG-SH) at varying concentrations, conversions were determined via 19F NMR and size exclusion chromatography (SEC) and the emission was recorded and integrated. Again, a correlation between integrated emission and conversion of PFTR was evident, however, deviated marginally from the emission of pure TBAF samples, which was attributed to an increase in viscosity of the solution due to star polymer formation. Last, two bis-thiols were employed for network formation with 3PFB via PFTR. This time, determination of conversion via 19F NMR, LC or SEC was futile due to insolubility of the formed networks. Thus, the supernatant solutions of the networks were diluted and investigated for their chemiluminescent properties. After recording and integrating the CL emission, the fluoride concentration and thus conversion of PFTR in each sample was back-calculated based on the emission, resulting in conversions that were coherent with the expected conversion based on the amount of thiols employed. The demonstrated method thus presents a high sensitivity, state-of-the art optical read-out for the formation of networks via the PFTR without the necessity of degrading the networks. In a third project, presented in Chapter 5, peroxyoxalate chemiluminescence (PO-CL) was employed to visualise the unfolding process of single-chain nanoparticles (SCNPs). A group of linear polymer precursors was synthesised, bearing photo-reactive ortho- methylbenzaldehyde (o-MBA) moieties for crosslinker mediated folding, as well as pyrene moieties that act as fluorophores in the PO-CL. A previously reported bis-

71 Conclusion and Future Perspective

maleimide phenyl oxalate (PO) linker with structural similarity to commercial bis-(2,4,6- trichlorophenyl) oxalate (TCPO) was employed for the intramolecular crosslinking of the linear polymer precursors to provide degradable SCNPs. Successful folding was subsequently conﬁrmed via 1H NMR, diffusion-ordered NMR spectroscopy (DOSY), and SEC showing compaction between 7.4 and 16% depending on size and o-MBA

content. Addition of hydrogen peroxide (H2O2) triggered a cascade reaction, starting with the cleavage of the PO linker, formation of a high energy intermediate (HEI), degradation of the HEI catalysed by the fluorophore and finally, emission of light from the fluorophore, which was recorded as a function of time. The time-dependant CL intensity was integrated to obtain the cumulative emission intensity as a function of time. Using the PREDICI® software package, the rate coefficients of the HEI formation and the catalytic degradation were estimated along with the population of the intact linker, the HEI and the total emission over time. The obtained rate coefficients were discussed and the estimated emission data was coherent with the recorded data confirming the validity of the estimation. The unfolded SCNPs (uSCNPs) were finally analysed via 1H NMR, DOSY, and SEC to prove complete degradation of the linker and thus unfolding of the SCNPs. PO-CL thus provides an excellent tool for kinetic and mechanistic investigations of polymeric unfolding processes.

6.2 Conclusion and Future Perspective

The present thesis constitutes a pioneering study investigating the self-propagation of the PFTR and developed a readily available optical read-out for the PFTR via CL. The CL read-out allows the facile quantification of PFTR events on a small molecule as well as polymer scale in a concentration regime 1000-times lower than conventional techniques like NMR or LC. Critically, the method presents a major breakthrough in the quantification of PFTR events within insoluble samples such as networks without the necessity to degrade the sample. Sauer et al. recently reported the use of a Schaap’s dioxetane as a read-out method for the formation of disulfides from thiols.269 However, in their approach, the dioxetane is employed as an oxidising agent and the read-out is performed based on the CL quenching due to decomposition of the dioxetane, i.e. the emission output is inversely proportional to the conversion. Thus, at this point, it is uncertain whether an on-line read-out of the PFTR is feasible. When employing all four reagents, i.e. the thiol and the base, the PFB moiety, and the CL probe, in one batch, the thiol can either undergo PFTR with the PFB moiety or oxidation by the dioxetane, as depicted in Scheme 6.1.

72 General Discussion

Thus, a future study has to investigate which of the two pathways in Scheme 6.1 is preferred, PFTR or oxidation. In case the PFTR pathway is significantly more favoured over the oxidation and the oxidation in turn is negligible, the on-line monitoring of PFTR via CL can be investigated. However, if the amount of oxidation, and thus disulfide formation and degradation of the CL probe, appears to be significant, the reaction conditions of the PFTR or the CL read-out itself need to be optimised further to enable the on-line detection of fluoride release. Thus far, the PFTR read-out via CL has only been investigated in acetonitrile (ACN) as the solvent, employing TBAOH as base, while disulfide formation by dioxetane reduction was mostly performed in methanol using potassium carbonate as a base. As methanol is disadvantageous for both the PFTR and the CL read-out, the feasibility of the disulfide formation in other solvents such as ACN should be investigated first.

O O O F F O O HO F F O F F F HBase TBSO S R + S R S R F F S TBSO R

Scheme 6.1: In the presence of a PFB moiety and a dioxetane-based CL probe, a deprotonated thiol can either undergo a PFTR (left reaction pathway) or oxidative disulﬁde formation (right reaction pathway).

While both reactions proceed within minutes to a few hours in their respective reaction environment, finding such a reaction environment that specifically addresses the PFTR, and not the oxidation, is key to the on-line monitoring of the PFTR. Folding of SCNPs for example, is commonly characterised by a reduction of hydrodynamic radius before and after the compaction. Observation of the folding process itself, however, is highly challenging to achieve, as conventional analytical techniques such as NMR, SEC or light scattering either require sample concentrations well above the concentration at which the folding takes place or instead operate on a time-scale much larger than that of the folding process. CL on the other hand, is highly sensitive and can detect concentrations as low as pmol. With an optimised CL read-out that allows on-line reaction monitoring of the PFTR, the folding of SCNPs could thus directly be observed and the folding behaviour evaluated. Yet, the reported read-out only allows for the quantification of bonds formed via

73 Conclusion and Future Perspective

PFTR and potentially other aromatic nucleophilic substitutions (SNAr’s) that release a ﬂuoride ion. However, in order to develop further CL read-out methods, the employed leaving group of the phenyl dioxetane needs to be adjusted to suit different crosslinking reactions. The obvious choice for a crosslinking reaction seems to be a nucleophilic substitution reaction, where a bond is formed whilst simultaneously a trigger molecule is released. However, while the employed nucleophiles usually show at least decent reactivity, common leaving groups like chloride, bromide, iodide, tosylate or triﬂate ions are mostly unreactive, thus unable to trigger any subsequent reactions (refer to Scheme 6.2).

CF 3 Cl O SO O O O O ? O Br + O O O O S LG O I X X

Scheme 6.2: Common leaving groups in nucleophilic substitution reactions are chloride, bromide, iodide, tosylate and triﬂate among others. However, these anions exhibit relatively low reactivity, thus complicating the design of a CL probe triggered by them.

Alternatively, the CL probe itself can be modified to function as a leaving group. Thioester-protected phenyl dioxetanes have been reported by Matsumoto et al.208,296 Thioester are generally known for their reactivity towards aminolysis. The reaction of a thioester phenyl dioxetane with an amine thus results in an amide and a deprotonated thiophenyl dioxetane undergoing a chemiluminescent degradation (refer to Figure 6.1A). A highly electron deficient phenyl dioxetane, such as the one depicted in Figure 6.1B, could also act as a leaving group, however, perfluorination of the phenyldioxetane seems impractical. A more facile approach is probably the use of a perfluorinated 4-(hydroxymethyl)phenol spacer (Figure 6.1C). The spacer exhibits sufficient electron- deficiency to act as a leaving group and, importantly, once the linker is cleaved, it undergoes quinone methide elimination, thus cleaving also the dioxetane and enabling

its chemiluminescent degradation. Moreover, careful design of R1 and R2 will allow the formation of sophisticated macromolecular architectures such as star polymers, SCNPs or networks. Interestingly, not only substitution reactions that release a trigger molecule allow for a potential optical CL read-out. A system designed by Bruemmer et al. targeted

74 General Discussion

A O R O R O 2 O 1 O H2N O R2 + S R1 N S O H O

B O O O O O O X O + R2 + F F R2 F F O R1 X

R1 O F O F X = O, NH, ... F F

C O O O O O O O X F + R2 + F R2 F R1 X F O O O

R1 O F O F F F

F O F O + O O F X = O, NH, ... F O

Figure 6.1: A. Thioester-protected phenyl dioxetane readily undergo aminolysis under emission of light while forming stable amide bonds.208,296 B. Perfluorinated phenyl dioxetanes as activated ester would allow for a direct CL emission upon cleavage. However, perfluorination of the phenyldioxetane seems impractical. C. A perfluorinated 4-(hydroxymethyl)phenolic ester functions as a spacer and a leaving group. Upon cleavage of the activated ester, the spacer undergoes quinone methide elimination, thus triggering the CL of Schaap’s dioxetane.

75 Conclusion and Future Perspective

the CL detection of formaldehyde via a cascade reaction starting with imine formation and ending with the emission of a visible light signal.17 In their approach, depicted in Scheme 6.3, the Schaap’s dioxetane is functionalised with a 4-aminohept-6-en-1-ol. In a ﬁrst step, the amino functionality forms an imine with the formaldehyde. After an aza-Cope rearrangement, the imine is hydrolysed and the 4-hydroxybutanone undergoes β-scission, deprotecting the Schaap’s dioxetane. While the reported system employs only formaldehyde as a trigger for the cascade reaction, a more sophisticated design, where R1 and R2 are part of crosslinker molecules or polymers, would allow for the chemiluminescent formation of networks or SCNPs.

76 General Discussion oe ewrso CP a efre,rsetvl,adsbeunl analysed subsequently and respectively, formed, be can SCNPs or networks bone, R and group methyl Sc ee6.3: heme R 1 Bruemmer et al .: R acd ecinfrteceiuiecn eeto ffradhd codn orfrne[7.I hi ok R work, their In [17]. reference to according formaldehyde of detection chemiluminescent the for reaction Cascade NH R O 1 2 2 O sahdoe.We R When hydrogen. a is O O O NH R O 2 R 1 2 = H = -CH3 2 O O O H O R 1 2 n R and β -scission R 2 1 r hsna i n r-ucinllneso spnatgop naplmrback polymer a on groups pendant as or linkers tri-functional and bi- as chosen are N O R 2 O R O 1 O O NH O R 3 2 O O O aza-Cope via hi msinbehaviour. emission their R 1 R N 1 O O R 2 O O NH O O R 3 2 O O O 1 sa is

77 Conclusion and Future Perspective

In addition to the quantification of PFTR events via the CL of Schaap’s dioxetane, the mechanism of SCNP unfolding was investigated by means of PO-CL. The method allowed for the on-line tracing of the unfolding and enabled the estimation of rate constants as well as the population of intermediate species. However, due to the slightly more complex reaction mechanism of PO-CL as well as the confined environment of SCNPs, a direct quantification of crosslinking moieties was not achieved. This can mostly be attributed to i) changes in viscosity due to the polymeric environment, ii) the unknown exact nature of the HEI and iii) the molecular motion of the HEI to the polymer based fluorophore. As a solution of a polymer shows increased viscosity compared to the pure solvent depending on the molecular weight of the polymer, and because the CL quantum yield depends on the viscosity, each of the employed chemiluminescent SCNPs will exhibit an individual quantum yield. Still, the viscosity of solution is relatively easy to determine, and the CL quantum yield of the PO linker can readily be determined as a function of viscosity for known concentrations. Moreover, the exact structure of the HEI is still not known (refer to Section 2.4.2). Although dioxetanedione is the commonly accepted HEI in PO-CL, some researchers still argue in favour of a phenoxy-dioxetanone. While this debate is reasonably circumstantial for small molecule PO-CL, polymeric PO-CL can be drastically affected by one structure

or the other, as depicted in Scheme 6.4. After addition of H2O2 and degradation of the linker, a dioxetanedione HEI can diffuse freely in the solution to find and react with the fluorophore, causing the emission of light. A phenoxy-dioxetanone HEI, however, would still be bound to the polymer backbone, thus being unable to diffuse freely like the dioxetanedione, instead having to rely on the molecular motion of the backbone to get into proximity of the fluorophore. Consequently, a dioxetanedione HEI would not only exhibit faster kinetics compared to a phenoxy-dioxetanone HEI, but also provide a higher CL quantum yield as the intermediate can potentially degrade before it has the chance to react with the fluorophore. Furthermore, it is unknown to which degree the diffusion of a dioxetanedione HEI is obstructed by the polymeric scaffold, i.e. after degradation of the crosslinks the polymer might still adopt a coiled stature, shielding the fluorophore from the dioxetanedione and vice versa. Early investigations of the HEI by Rauhut et al. included an experiment where they passed a gas stream through a solution of bis-(2,4-dinitrophenyl)oxalate (DNPO) and

H2O2 and subsequently conducted the gas stream into solutions of different fluophores. Finding that fluorophore solutions displayed bright CL where they were contacted by the gas led the group to the assumption that a HEI must exist in the first place. Yet, if such an experiment was repeated on a polymeric basis, dioxetanedione species would again

78 General Discussion

O O O O

H2O2 H2O2

OH OH HO O O O O O O O O

Scheme 6.4: Schematic depiction of polymer-based PO-CL and the competition between dioxetanedione and phenoxy-dioxetanone as the HEI. While a dioxetanedione HEI can diffuse freely to the ﬂuorophore, a phenoxy-dioxetanone HEI is still connected to the polymer backbone and has to rely on the rotation of the backbone to get into proximity of the ﬂuorophore.

be suited to induce CL in the fluorophore solutions. However, if the HEI was indeed a phenoxy-dioxetanone, no CL should be visible in the fluophore solutions, as a high molecular weight polymer-based phenoxy-dioxetanone will not go over to the gaseous phase. The experiment will thus shed light on the nature of the HEI. Once these limitations are overcome, more in-depth studies of the chemiluminescent unfolding and the quantification of crosslinks will become feasible. However, contrary to Schaap’s dioxetane, no leaving group modification is possible on POs and the only

trigger capable of triggering PO-CL is a reactive oxygen species (ROS) such as H2O2, limiting the applications of PO-CL with respect to the trigger. Therefore, PO-CL is more suited for the degradation of macromolecular architectures by ROS, rather than the formation. Yet, the fluorophore is easily adapted as it is not chemically bound to the oxalate, allowing for the detection of different fluorophores rather than the detection of ROS. PO-CL can furthermore find potential application in the degradation of micelles by ROS. As phenols bearing electronwithdrawing groups do not only present excellent

79 Conclusion and Future Perspective

leaving groups in PO-CL (refer to Section 2.4.2), they also possess a relatively low pKa-value. Thus, after cleavage of the corresponding PO, the phenol can be readily deprotonated. This can e.g. be exploited in pH sensitive micelles (refer to Scheme 6.5). A hypothetical co-polymer backbone is presumed that allows for the formation of micelles that are stable under neutral to basic conditions and degrade under acidic conditions. Given that the outer shell was furthermore functionalised with a PO and the inner core was functionalised with a fluorophore, then addition of H2O2 would degrade the PO, forming acidic phenols and dioxetanedione. The phenols will subsequently deprotonate, decreasing the pH of the solution and causing the degradation of the micelles. Upon degradation, the fluorophore is furthermore released from the core of the micelle and can subsequently finally react with the dioxetane to emit CL. In summary, an optical read-out for the quantification of PFTR events was established employing the CL of Schaap’s dioxetane on one hand, and PO-CL was employed for the qualitative assessment of SCNP unfolding on the other hand. The advantages and limitations have been discussed in the current chapter. While future research on the CL of Schaap’s dioxetane should mostly focus on investigating new combinations of crosslinking reactions and CL triggers, PO-CL will mainly require further mechanistic studies before it can be used for detailed investigations of polymeric systems. Never- theless, both CL systems present promising tools for the in-depth characterisation of macromolecular architectures and their transformations.

80 General Discussion

Cl- N O N+

Cl Cl Cl O O O O O O OO O O Cl Cl Cl Cl Cl O O Cl Cl

ROS H

O O

Cl- N O N+

Scheme 6.5: Schematic depiction of the degradation of pH-sensitive micelles upon cleavage of electronpoor phenyloxalates (green). The released phenols are sufﬁciently acidic to decrease the pH of the solution causing the micelles to deteriorate and release the dye (red), which can subsequently react with the dioxetanedione, causing CL.

81 7 References

[1] F. Cavalli, H. Mutlu, S. O. Steinmueller, L. Barner, Polym. Chem. 2017, 8(25), 3778–3782, DOI 10.1039/c7py00812k.

[2] A. P. Schaap, T.-S. Chen, R. S. Handley, R. DeSilva, B. P. Giri, Tetrahedron Lett. 1987, 28(11), 1155–1158, DOI 10.1016/S0040-4039(00)95313-9.

[3] L. Delafresnaye, C. W. Schmitt, L. Barner, C. Barner-Kowollik, Chem. Eur. J. 2019, 25(54), 12538–12544, DOI 10.1002/chem.201901858.

[4] S. Sinnwell, A. J. Inglis, M. H. Stenzel, C. Barner-Kowollik, in Click Chemistry for Biotechnology and Materials Science, (Edited by J. Lahann), John Wiley & Sons, Ltd, Chichester, UK, 2009, 89–117, DOI 10.1002/9780470748862.ch6.

[5] F. Szillat, B. V. Schmidt, A. Hubert, C. Barner-Kowollik, H. Ritter, Macromolecular Rapid Communications 2014, 35(14), 1293–1300, DOI 10.1002/marc.201400122.

[6] X. K. Hillewaere, F. E. Du Prez, Prog. Polym. Sci. 2015, 49-50, 121–153, DOI 10.1016/j.progpolymsci.2015.04.004.

[7] F. Zhang, P. Ju, M. Pan, D. Zhang, Y. Huang, G. Li, X. Li, Corros. Sci. 2018, 144(December 2017), 74–88, DOI 10.1016/j.corsci.2018.08.005.

[8] Q. Zhang, S. Niu, L. Wang, J. Lopez, S. Chen, Y. Cai, R. Du, Y. Liu, J. C. Lai, L. Liu, C. H. Li, X. Yan, C. Liu, J. B. Tok, X. Jia, Z. Bao, Adv. Mater. 2018, 30(33), DOI 10.1002/adma.201801435.

[9] H. Wang, P. Wang, Y. Feng, J. Liu, J. Wang, M. Hu, J. Wei, Y. Huang, ChemElec- troChem 2019, 6(6), 1605–1622, DOI 10.1002/celc.201801612.

[10] C. M. Geiselhart, H. Mutlu, C. Barner-Kowollik, Angew. Chem. Int. Ed. 2020, DOI 10.1002/anie.202012592.

[11] D. Shabat, N. Hananya, O. Green, R. Blau, R. Satchi-Fainaro, Angew. Chem. Int. Ed. 2017, 56(39), 11793–11796, DOI 10.1002/anie.201705803.

82 References

[12] W. Chen, Q. Huang, W. Ou, Y. Hao, L. Wang, K. Zeng, H. Guo, J. Li, Y.-N. Liu, small 2014, 10(7), 1261–1265, DOI 10.1002/smll.201302698.

[13] M. P. Robin, A. B. Mabire, J. C. Damborsky, E. S. Thom, U. H. Winzer-Serhan, J. E. Raymond, R. K. O’Reilly, J. Am. Chem. Soc. 2013, 135(25), 9518–9524, DOI 10.1021/ja403587c.

[14] O. Green, T. Eilon, N. Hananya, S. Gutkin, C. R. Bauer, D. Shabat, ACS Cent. Sci. 2017, 3(4), 349–358, DOI 10.1021/acscentsci.7b00058.

[15] S. Gnaim, A. Scomparin, S. Das, R. Blau, R. Satchi-Fainaro, D. Shabat, Angew. Chem. Int. Ed. 2018, 57(29), 9033–9037, DOI 10.1002/anie.201804816.

[16] T. Eilon-Shaffer, M. Roth-Konforti, A. Eldar-Boock, R. Satchi-Fainaro, D. Shabat, Org. Biomol. Chem. 2018, 16(10), 1708–1712, DOI 10.1039/c8ob00087e.

[17] K. J. Bruemmer, O. Green, T. A. Su, D. Shabat, C. J. Chang, Angew. Chem. Int. Ed. 2018, 57(25), 7508–7512, DOI 10.1002/anie.201802143.

[18] N. Hananya, O. Green, R. Blau, R. Satchi-Fainaro, D. Shabat, Angew. Chem. Int. Ed. 2017, 56(39), 11793–11796, DOI 10.1002/anie.201705803.

[19] M. E. Roth-Konforti, C. R. Bauer, D. Shabat, Angew. Chem. Int. Ed. 2017, 56(49), 15633–15638, DOI 10.1002/anie.201709347.

[20] S. Das, J. Ihssen, L. Wick, U. Spitz, D. Shabat, Chem. Eur. J. 2020, 26(16), 3647–3652, DOI 10.1002/chem.202000217.

[21] S. Gnaim, A. Scomparin, A. Eldar-Boock, C. R. Bauer, R. Satchi-Fainaro, D. Sha- bat, Chem. Sci. 2019, 10(10), 2945–2955, DOI 10.1039/c8sc05174g.

[22] M. Karayianni, S. Pispas, in Fluorescence Studies of Polymer Containing Systems, (Edited by K. Prochakzka), 1 Ed., Springer International Publishing Switzerland, chap. 1, 2016, 27–63, DOI 10.1007/978-3-319-26788-3.

[23] D. Schmaljohann, Adv. Drug Deliver. Rev. 2006, 58(15), 1655–1670, DOI 10.1016/j.addr.2006.09.020.

[24] N. Hadjichristidis, A. Hirao, Y. Tezuka, F. D. Prez, Complex Macromolecular Architectures, 1st Ed., John Wiley & Sons, Inc., Singapore, 2011.

[25] C. K. Lyon, A. Prasher, A. M. Hanlon, B. T. Tuten, C. A. Tooley, P. G. Frank, E. B. Berda, Polym. Chem. 2015, 6(2), 181–197, DOI 10.1039/C4PY01217H.

83 [26] T. Terashima, T. Mes, T. F. A. De Greef, M. A. J. Gillissen, P. Besenius, A. R. A. Palmans, E. W. Meijer, J. Am. Chem. Soc. 2011, 133(13), 4742–4745, DOI 10.1021/ja2004494.

[27] I. Perez-Baena, I. Loinaz, D. Padro, I. García, H. J. Grande, I. Odriozola, J. Mater. Chem. 2010, 20(33), 6916, DOI 10.1039/c0jm01025a.

[28] N. D. Knöfel, H. Rothfuss, J. Willenbacher, C. Barner-Kowollik, P. W. Roesky, Angew. Chem. Int. Ed. 2017, 56(18), 4950–4954, DOI 10.1002/anie.201700718.

[29] A. Sanchez-Sanchez, A. Arbe, J. Colmenero, J. A. Pomposo, ACS Macro. Lett. 2014, 3(5), 439–443, DOI 10.1021/mz5001477.

[30] K. Dušek, M. Dušková-Smrcková,ˇ Prog. Polym. Sci. 2000, 25(9), 1215–1260, DOI 10.1016/S0079-6700(00)00028-9.

[31] J. Engelke, J. Brandt, C. Barner-Kowollik, A. Lederer, Polym. Chem. 2019, 10(25), 3410–3425, DOI 10.1039/c9py00336c.

[32] H. Frisch, B. T. Tuten, C. Barner-Kowollik, Isr. J. Chem. 2020, 60(1-2), 86–99, DOI 10.1002/ijch.201900145.

[33] B. T. Tuten, D. Chao, K. Lyon, E. B. Berda, Polym. Chem. 2012, 3, 3068–3071, DOI 10.1039/c2py20308.

[34] N. Ormategui, I. García, D. Padro, G. Cabañero, H. J. Grande, I. Loinaz, Soft Matter 2012, 8(3), 734–740, DOI 10.1039/C1SM06310C.

[35] J. Steinkoenig, H. Rothfuss, A. Lauer, B. T. Tuten, C. Barner-Kowollik, J. Am. Chem. Soc. 2017, 139(1), 51–54, DOI 10.1021/jacs.6b10952.

[36] J. Steinkoenig, T. Nitsche, B. T. Tuten, C. Barner-Kowollik, Macromolecules 2018, 51, 3967–3974, DOI 10.1021/acs.macromol.8b00577.

[37] Z.-Y. Qiao, C.-Y. Hou, W.-J. Zhao, D. Zhang, P.-P. Yang, L. Wang, H. Wang, Chem. Commun. 2015, 51(63), 12609–12612, DOI 10.1039/C5CC03752B.

[38] A. Kulkarni, P. Rao, S. Natarajan, A. Goldman, V. S. Sabbisetti, Y. Khater, N. Ko- rimerla, V. Chandrasekar, R. A. Mashelkar, S. Sengupta, P. Natl. Acad. Sci. USA 2016, 113(15), E2104–E2113, DOI 10.1073/pnas.1603455113.

[39] Y. Luo, L. Huang, Y. Yang, X. Zhuang, S. Hu, H. Ju, B. Y. Yu, J. Tian, Theranostics 2017, 7(5), 1245–1256, DOI 10.7150/thno.18187.

84 References

[40] D. Estupiñán, C. Barner-Kowollik, L. Barner, Angew. Chem. Int. Ed. 2018, 57(20), 5925–5929, DOI 10.1002/anie.201713388.

[41] J. Alemán, A. V. Chadwick, J. He, M. Hess, K. Horie, R. G. Jones, P. Kratochvíl, I. Meisel, I. Mita, G. Moad, S. Penczek, R. F. Stepto, Pure Appl. Chem. 2007, 79(10), 1801–1829, DOI 10.1351/pac200779101801.

[42] C. Goodyear, Polytechnisches Journal 1838, 69, 382–384.

[43] C. Guise-Richardson, Technol. Cult. 2010, 51(2), 357–387.

[44] M. Akiba, A. S. Hashim, Prog. Polym. Sci. 1997, 22, 475–521, DOI 10.1016/S0079-6700(96)00015-9.

[45] A. M. Schenzel, C. O. Klein, K. Rist, N. Moszner, C. Barner-Kowollik, Adv. Sci. 2016, 3(3), 1–5, DOI 10.1002/advs.201500361.

[46] J. Li, A. D. Celiz, J. Yang, Q. Yang, I. Wamala, W. Whyte, B. R. Seo, N. V. Vasilyev, J. J. Vlassak, Z. Suo, D. J. Mooney, Science 2017, 357(6349), 378–381, DOI 10.1126/science.aah6362.

[47] C. W. Peak, J. J. Wilker, G. Schmidt, Colloid Polym. Sci. 2013, 291(9), 2031–2047, DOI 10.1007/s00396-013-3021-y.

[48] Y. Osada, J. P. Gong, Y. Tanaka, J. Macromol. Sci. Pol. R. 2004, 44(1), 87–112, DOI 10.1081/MC-120027935.

[49] J. Kopecek,ˇ J. Polym. Sci. A1 2009, 47, 5929–5946, DOI 10.1002/pola.23028.

[50] Z. Gu, K. Huang, Y. Luo, L. Zhang, T. Kuang, Z. Chen, G. Liao, Nanomed. Nanobiotechnol. 2018, 10, 1–15, DOI 10.1002/wnan.1520.

[51] J. Boekhoven, S. I. Stupp, Adv. Mater. 2014, 26(11), 1642–1659, DOI 10.1002/adma.201304606.

[52] S. S. Liow, Q. Dou, D. Kai, A. A. Karim, K. Zhang, F. Xu, X. J. Loh, ACS Biomateri. Sci. Eng. 2016, 2(3), 295–316, DOI 10.1021/acsbiomaterials.5b00515.

[53] A. S. Hoffman, Adv. Drug Deliver. Rev. 2012, 64(SUPPL.), 18–23, DOI 10.1016/j.addr.2012.09.010.

[54] Y. Akagi, J. P. Gong, U. I. Chung, T. Sakai, Macromolecules 2013, 46(3), 1035– 1040, DOI 10.1021/ma302270a.

85 [55] W. Kuhn, J. Polym. Sci. 1946, 1, 380–388, DOI 10.1002/pol.1946.120010505.

[56] F. T. Wall, P. J. Flory, J. Chem. Phys. 1951, 19(12), 1435–1439, DOI 10.1063/1.1748098.

[57] H. M. James, E. Guth, J. Chem. Phys. 1943, 11(10), 455–481, DOI 10.1063/1.1723785.

[58] M. Rubinstein, S. Panyukov, Macromolecules 2002, 35(17), 6670–6686, DOI 10.1021/ma0203849.

[59] L. Voorhaar, R. Hoogenboom, Chem.Soc. Rev. 2016, 45(14), 4013–4031, DOI 10.1039/c6cs00130k.

[60] C. Heinzmann, C. Weder, L. M. De Espinosa, Chem.Soc. Rev. 2016, 45(2), 342– 358, DOI 10.1039/c5cs00477b.

[61] O. Okay, Prog. Polym. Sci. 2000, 25(6), 711–779, DOI 10.1016/S0079-6700(00) 00015-0.

[62] A. J. Scott, A. Nabifar, A. Penlidis, Macromolecular Reaction Engineering 2014, 8(9), 639–657, DOI 10.1002/mren.201400004.

[63] E. Mehravar, A. Agirre, N. Ballard, S. Van Es, A. Arbe, J. R. Leiza, J. M. Asua, Macromolecules 2018, 51(23), 9740–9748, DOI 10.1021/acs.macromol.8b01648.

[64] F. Cavalli, C. Pfeifer, L. Arens, L. Barner, M. Wilhelm, Macromol. Chem. Phys. 2020, 221(1), 1900387, DOI 10.1002/macp.201900387.

[65] A. J. Scott, A. Penlidis, Can. J. Chem. Eng. 2020, 1–26, DOI 10.1002/cjce.23855.

[66] G. Odian, Principles of Polymerization, 4th Ed., John Wiley & Sons, Inc., Hoboken, New Jersey, 2004.

[67] M. L. Hallensleben, in Handbook of Polymer Synthesis, (Edited by H. R. Kricheldorf, O. Nuyken, G. Swift), 2nd Ed., Marcel Dekker, New York, chap. 15, 2005, 846– 885.

[68] A. S. Quick, H. Rothfuss, A. Welle, B. Richter, J. Fischer, M. Wegen- er, C. Barner-Kowollik, Adv. Funct. Mater. 2014, 24(23), 3571–3580, DOI 10.1002/adfm.201304030.

86 References

[69] V. T. Tan, D. H. Nguyen, R. H. Utama, M. Kahram, F. Ercole, J. F. Quinn, M. R. Whittaker, T. P. Davis, J. Justin Gooding, Polym. Chem. 2017, 8(39), 6123–6133, DOI 10.1039/c7py01038a.

[70] T. Krappitz, F. Feist, I. Lamparth, N. Moszner, H. John, J. P. Blinco, T. R. Dargaville, C. Barner-Kowollik, Mater. Horiz. 2019, 6(1), 81–89, DOI 10.1039/c8mh00951a.

[71] S. Bialas, L. Michalek, D. E. Marschner, T. Krappitz, M. Wegener, J. Blinco, E. Blasco, H. Frisch, C. Barner-Kowollik, Adv. Mater. 2019, 31(8), 1–5, DOI 10.1002/adma.201807288.

[72] L. Michalek, S. Bialas, S. L. Walden, F. R. Bloesser, H. Frisch, C. Barner-Kowollik, Adv. Funct. Mater. 2020, 30(48), 1–7, DOI 10.1002/adfm.202005328.

[73] M. Gernhardt, H. Frisch, A. Welle, R. Jones, M. Wegener, E. Blasco, C. Barner-Kowollik, J. Mater. Chem. C. 2020, 8(32), 10993–11000, DOI 10.1039/d0tc02751k.

[74] M. Zhong, R. Wang, K. Kawamoto, B. D. Olsen, J. A. Johnson, Science 2016, 353(6305), 1264–1268, DOI 10.1126/science.aag0184.

[75] A. M. Schenzel, N. Moszner, C. Barner-Kowollik, Polym. Chem. 2017, 8(2), 414– 420, DOI 10.1039/C6PY01840H.

[76] H. W. Ooi, K. S. Jack, H. Peng, A. K. Whittaker, Polym. Chem. 2013, 4(17), 4788–4800, DOI 10.1039/c3py00653k.

[77] M. Kaupp, K. Hiltebrandt, V. Trouillet, P. Mueller, A. S. Quick, M. Wegen- er, C. Barner-Kowollik, Chem. Commun. 2016, 52(9), 1975–1978, DOI 10.1039/c5cc09444e.

[78] K. K. Oehlenschlaeger, J. O. Mueller, J. Brandt, S. Hilf, A. Lederer, M. Wilhelm, R. Graf, M. L. Coote, F. G. Schmidt, C. Barner-Kowollik, Adv. Mater. 2014, 26(21), 3561–3566, DOI 10.1002/adma.201306258.

[79] B. J. Adzima, H. A. Aguirre, C. J. Kloxin, T. F.Scott, C. N. Bowman, Macromolecules 2008, 41(23), 9112–9117, DOI 10.1021/ma801863d.

[80] Y. Fan, C. Deng, R. Cheng, F. Meng, Z. Zhong, Biomacromolecules 2013, 14(8), 2814–2821, DOI 10.1021/bm400637s.

87 [81] V. X. Truong, F. Li, F. Ercole, J. S. Forsythe, ACS Macro. Lett. 2018, 7(4), 464–469, DOI 10.1021/acsmacrolett.8b00099.

[82] A. Metters, J. Hubbell, Biomacromolecules 2005, 6(1), 290–301, DOI 10.1021/bm049607o.

[83] M. Rubinstein, R. H. Colby, in Polymer Physics, (Edited by M. Rubinstein, R. H. Colby), 1st Ed., Oxford University Press, New York, chap. 7, 2003, 253–307.

[84] P. J. Flory, Chem. Rev. 1944, 35(1), 51–75, DOI 10.1021/cr60110a002.

[85] J. M. Charlesworth, Polym. Eng. Sci. 1988, 28(4), 230–236, DOI 10.1002/pen.760280406.

[86] T. N. Mitchell, B. Costisella, in NMR - From Spectra to Structures An Experimental Approach, 2nd Ed., Springer, Heidelberg, Berlin, chap. 6, 2007, 73–83.

[87] M. H. Levitt, in Spin Dynamics Basics of Nuclear Magnetic Resonance, 2nd Ed., John Wiley & Sons, Inc., Chichester, chap. 9, 2008, 195–228.

[88] D. R. Morgan, S. Kalachandra, H. K. Shobha, N. Gunduz, E. O. Stejskal, Biomate- rials 2000, 21(18), 1897–1903, DOI 10.1016/S0142-9612(00)00067-3.

[89] M. Bertmer, A. Buda, I. Blomenkamp-Höfges, S. Kelch, A. Lendlein, Macro- molecules 2005, 38(9), 3793–3799, DOI 10.1021/ma0501489.

[90] J. Brus, M. Urbanová, A. Strachota, Macromolecules 2008, 41(2), 372–386, DOI 10.1021/ma702140g.

[91] M. El Hariri El Nokab, P. C. van der Wel, Carbohyd. Polym. 2020, 240(April), DOI 10.1016/j.carbpol.2020.116276.

[92] D. J. P. Harrison, W. R. Yates, J. F. Johnson, J. Macromol. Sci. R. M. C. 1985, C25(4), 481–549, DOI 10.1080/07366578508081963.

[93] Y. E. Shapiro, Prog. Polym. Sci. 2011, 36(9), 1184–1253, DOI 10.1016/j.progpolymsci.2011.04.002.

[94] K. Saalwächter, D. Reichert, in Encyclopedia of Spectroscopy and Spectrometry, (Edited by D. Lindon, John C; Tranter, George E; Koppenaal), 2nd Ed., Elsevier, London, 2010, 2221–2236, DOI 10.1016/B978-0-12-803224-4.00032-7.

88 References

[95] M. H. Levitt, in Spin Dynamics Basics of Nuclear Magnetic Resonance, 2nd Ed., John Wiley & Sons, Inc., Chichester, chap. 19 and 20, 2008, 507–595.

[96] J. Höpfner, G. Guthausen, K. Saalwächter, M. Wilhelm, Macromolecules 2014, 47(13), 4251–4265, DOI 10.1021/ma500558v.

[97] X. Guo, S. Theissen, J. Claussen, V. Hildebrand, J. Kamphus, M. Wilhelm, B. Luy, G. Guthausen, Macromol. Chem. Phys. 2019, 220(2), 1–11, DOI 10.1002/macp.201800350.

[98] X. Guo, C. Pfeifer, M. Wilhelm, B. Luy, G. Guthausen, Macromol. Chem. Phys. 2019, 220(10), 1–9, DOI 10.1002/macp.201800525.

[99] K. J. Jovic, T. Richter, C. Lang, J. P. Blinco, C. Barner-Kowollik, Macromolecules 2018, 51(21), 8712–8720, DOI 10.1021/acs.macromol.8b01347.

[100] K. Jovic, K. De Bruycker, T. Richter, C. Barner-Kowollik, J. P. Blinco, Macromol. Rapid Commun. 2020, 41(18), 1–5, DOI 10.1002/marc.202000183.

[101] T. Clark, L. Kwisnek, C. E. Hoyle, S. Nazarenko, J. Polym. Sci. A1 2009, 47(April), 14–24, DOI 10.1002/pola.23089.

[102] H. Zhou, J. Woo, A. M. Cok, M. Wang, B. D. Olsen, J. A. Johnson, P. Natl. Acad. Sci. USA 2012, 109(47), 19119–19124, DOI 10.1073/pnas.1213169109.

[103] S. Chowdhury, E. P. Grimsrudk, T. Heinis, P. Kebarle, J. Am. Chem. Soc. 1986, 108(13), 3630–3635, DOI 10.1021/ja00273a014.

[104] J. A. Godsell, M. Stacey, J. C. Tatlow, Nature 1956, 178(4526), 199–200, DOI 10.1038/178199a0.

[105] E. J. Forbes, R. D. Richardson, M. Stacey, J. C. Tatlow, J. Chem. Soc. 1959, 2019–2021, DOI 10.1039/JR9590002019.

[106] J. M. Birchall, R. N. Haszeldine, J. Chem. Soc. 1959, 13–17, DOI 10.1039/JR9590000013.

[107] A. K. Barbour, M. W. Buxton, P. L. Coe, R. Stephens, J. C. Tatlow, J. Chem. Soc. 1961, 808–817, DOI 10.1039/JR9610000808.

[108] D. G. Holland, G. J. Moore, C. Tamborski, J. Org. Chem. 1964, 29(6), 1562–1565, DOI 10.1021/jo01029a069.

89 [109] C. Tamborski, E. J. Soloski, J. Org. Chem. 1966, 31(3), 746–749, DOI 10.1021/jo01341a023.

[110] R. Filler, N. R. Ayyangar, W. Gustowski, H. H. Kang, J. Org. Chem. 1969, 34(3), 534–538, DOI 10.1021/jo01255a011.

[111] R. Filler, R. C. Rickert, J. Flu. Chem. 1981, 18(4), 483–495, DOI 10.1016/S0022-1139(00)82665-2.

[112] T.-W. Lin, D. G. Naae, Tetrahedron Lett. 1978, 19, 1653–1656, DOI 10.1016/S0040-4039(01)94631-3.

[113] S. Agar, E. Baysak, G. Hizal, U. Tunca, H. Durmaz, J. Polym. Sci. A1 2018, 56(12), 1181–1198, DOI 10.1002/pola.29004.

[114] G. Delaittre, L. Barner, Polym. Chem. 2018, 9(20), 2679–2684, DOI 10.1039/c8py00287h.

[115] J. Kvícala,ˇ M. Beneš, O. Paleta, V. Král, J. Flu. Chem. 2010, 131(12), 1327–1337, DOI 10.1016/j.jﬂuchem.2010.09.003.

[116] C. R. Becer, K. Babiuch, M. Gottschaldt, D. Pilz, S. Hornig, T. Heinze, U. S. Schubert, Macromolecules 2009, 42(7), 2387–2394, DOI 10.1021/ma9000176.

[117] T. Cai, W. J. Yang, K. G. Neoh, E. T. Kang, Polym. Chem. 2012, 3(4), 1061–1068, DOI 10.1039/c2py00609j.

[118] D. Varadharajan, G. Delaittre, Polym. Chem. 2016, 7(48), 7488–7499, DOI 10.1039/c6py01711h.

[119] H. Turgut, G. Delaittre, Chem. Eur. J. 2016, 22(4), 1511–1521, DOI 10.1002/chem.201503844.

[120] H. Turgut, A. C. Schmidt, P. Wadhwani, A. Welle, R. Müller, G. Delaittre, Polym. Chem. 2017, 8(8), 1288–1293, DOI 10.1039/c6py02108e.

[121] F. Cavalli, L. De Keer, B. Huber, P. H. Van Steenberge, D. R. D’Hooge, L. Barner, Polym. Chem. 2019, 10(22), 2781–2791, DOI 10.1039/c9py00435a.

[122] J. Engelke, B. T. Tuten, R. Schweins, H. Komber, L. Barner, L. Plüschke, C. Barner-Kowollik, A. Lederer, Polym. Chem. 2020, 11(41), 6559–6578, DOI 10.1039/d0py01045f.

90 References

[123] J. Engelke, S. Boye, B. T. Tuten, L. Barner, C. Barner-Kowollik, A. Lederer, ACS Macro. Lett. 2020, 1569–1575, DOI 10.1021/acsmacrolett.0c00519.

[124] M. A. Meier, C. Barner-Kowollik, Adv. Mater. 2019, 31(26), DOI 10.1002/adma.201806027.

[125] S. C. Solleder, R. V. Schneider, K. S. Wetzel, A. C. Boukis, M. A. Meier, Macromol. Rapid Commun. 2017, 38(9), 1–45, DOI 10.1002/marc.201600711.

[126] W. Kuhn, G. Balmer, J. Polym. Sci. 1962, 57(165), 311–319, DOI 10.1002/pol.1962.1205716524.

[127] O. Altintas, C. Barner-Kowollik, Macromol. Rapid Commun. 2012, 33(11), 958– 971, DOI 10.1002/marc.201200049.

[128] S. Mavila, O. Eivgi, I. Berkovich, N. G. Lemcoff, Chem. Rev. 2016, 116(3), 878– 961, DOI 10.1021/acs.chemrev.5b00290.

[129] A. M. Hanlon, C. K. Lyon, E. B. Berda, Macromolecules 2016, 49(1), 2–14, DOI 10.1021/acs.macromol.5b01456.

[130] M. Antonietti, H. Sillescu, M. Schmidt, H. Schuch, Macromolecules 1988, 21(3), 736–742, DOI 10.1021/ma00181a031.

[131] M. P. Tsyurupa, T. A. Mrachkovskaya, L. A. Maslova, G. I. Timofeeva, L. V. Dubrov- ina, E. F. Titova, V. A. Davankov, V. M. Menshov, React. Polym. 1993, 19(1-2), 55–66, DOI 10.1016/0923-1137(93)90010-D.

[132] V. A. Davankov, M. M. Ilyin, M. P. Tsyurupa, G. I. Timofeeva, L. V. Dubrovina, Macromolecules 1996, 29(26), 8398–8403, DOI 10.1021/ma951673i.

[133] A. R. De Luzuriaga, N. Ormategui, H. J. Grande, I. Odriozola, J. A. Pom- poso, I. Loinaz, Macromol. Rapid Commun. 2008, 29(12-13), 1156–1160, DOI 10.1002/marc.200700877.

[134] A. R. De Luzuriaga, I. Perez-Baena, S. Montes, I. Loinaz, I. Odriozola, I. García, J. A. Pomposo, Macromol. Symp. 2010, 296(1), 303–310, DOI 10.1002/masy.201051042.

[135] O. Altintas, P. Gerstel, N. Dingenouts, C. Barner-Kowollik, Chem. Commun. 2010, 46(34), 6291–6293, DOI 10.1039/c0cc00702a.

91 [136] O. Altintas, E. Lejeune, P. Gerstel, C. Barner-Kowollik, Polym. Chem. 2012, 3(3), 640–651, DOI 10.1039/c1py00392e.

[137] E. J. Foster, E. B. Berda, E. W. Meijer, J. Am. Chem. Soc. 2009, 131, 6964–6966, DOI 10.1021/ja901687d.

[138] T. Mes, R. Van Der Weegen, A. R. Palmans, E. W. Meijer, Angew. Chem. Int. Ed. 2011, 50(22), 5085–5089, DOI 10.1002/anie.201100104.

[139] M. A. Gillissen, I. K. Voets, E. W. Meijer, A. R. Palmans, Polym. Chem. 2012, 3(11), 3166–3174, DOI 10.1039/c2py20350b.

[140] N. Hosono, M. A. Gillissen, Y. Li, S. S. Sheiko, A. R. Palmans, E. W. Meijer, J. Am. Chem. Soc. 2013, 135(1), 501–510, DOI 10.1021/ja310422w.

[141] S. Mavila, C. E. Diesendruck, S. Linde, L. Amir, R. Shikler, N. G. Lemcoff, Angew. Chem. Int. Ed. 2013, 52(22), 5767–5770, DOI 10.1002/anie.201300362.

[142] S. Mavila, I. Rozenberg, N. G. Lemcoff, Chem. Sci. 2014, 5(11), 4196–4203, DOI 10.1039/c4sc01231c.

[143] H. Rothfuss, N. D. Knöfel, P. Tzvetkova, N. C. Michenfelder, S. Baraban, A. N. Unterreiner, P. W. Roesky, C. Barner-Kowollik, Chem. Eur. J. 2018, 24(66), 17475– 17486, DOI 10.1002/chem.201803692.

[144] N. D. Knöfel, H. Rothfuss, C. Barner-Kowollik, P. W. Roesky, Polym. Chem. 2019, 10(1), 86–93, DOI 10.1039/c8py01486h.

[145] N. D. Knöfel, H. Rothfuss, P. Tzvetkova, B. Kulendran, C. Barner-Kowollik, P. W. Roesky, Chem. Sci. 2020, 11(38), 10331–10336, DOI 10.1039/d0sc03579c.

[146] D. Chao, X. Jia, B. Tuten, C. Wang, E. B. Berda, Chem. Commun. 2013, 49(39), 4178–4180, DOI 10.1039/c2cc37157j.

[147] P. Atkin, J. de Paula, Physical chemistry, 8th Ed., Oxford University Press, New York, 2006, 1072.

[148] S. Mavila, O. Eivgi, I. Berkovich, N. G. Lemcoff, Chem. Rev. 2016, 116(3), 878– 961, DOI 10.1021/acs.chemrev.5b00290.

[149] J. Tao, G. Liu, Macromolecules 1997, 30(8), 2408–2411, DOI 10.1021/ma961422p.

92 References

[150] G. Njikang, G. Liu, S. A. Curda, Macromolecules 2008, 41(15), 5697–5702, DOI 10.1021/ma800642r.

[151] S. Hecht, A. Khan, Angew. Chem. Int. Ed. 2003, 42(48), 6021–6024, DOI 10.1002/anie.200352983.

[152] P. G. Frank, B. T. Tuten, A. Prasher, D. Chao, E. B. Berda, Macromol. Rapid Commun. 2014, 35(2), 249–253, DOI 10.1002/marc.201300677.

[153] J. Willenbacher, K. N. Wuest, J. O. Mueller, M. Kaupp, H. A. Wa- genknecht, C. Barner-Kowollik, ACS Macro. Lett. 2014, 3(6), 574–579, DOI 10.1021/mz500292e.

[154] P. Lederhose, Z. Chen, R. Mueller, J. P. Blinco, S. Wu, C. Barner-Kowollik, Angew. Chem. Int. Ed. 2016, 55(40), 12195–12199, DOI 10.1002/anie.201606425.

[155] C. Heiler, S. Bastian, P. Lederhose, J. P. Blinco, E. Blasco, C. Barner-Kowollik, Chem. Commun. 2018, 54(28), 3476–3479, DOI 10.1039/c8cc01054d.

[156] T. K. Claus, J. Zhang, L. Martin, M. Hartlieb, H. Mutlu, S. Perrier, G. Delait- tre, C. Barner-Kowollik, Macromol. Rapid Commun. 2017, 38(16), 1–7, DOI 10.1002/marc.201700264.

[157] H. Frisch, J. P. Menzel, F. R. Bloesser, D. E. Marschner, K. Mundsinger, C. Barner-Kowollik, J. Am. Chem. Soc. 2018, 140(30), 9551–9557, DOI 10.1021/jacs.8b04531.

[158] H. Frisch, D. Kodura, F. R. Bloesser, L. Michalek, C. Barner-Kowollik, Macromol. Rapid Commun. 2020, 41(1), 1–6, DOI 10.1002/marc.201900414.

[159] D. Kodura, H. A. Houck, F. R. Bloesser, A. S. Goldmann, F. E. D. Prez, H. Frisch, C. Barner-Kowollik, Chem. Sci. 2020, in press, DOI 10.1203/pdr.0b013e31821ec47f.

[160] N. C. Yang, C. Rivas, J. Am. Chem. Soc. 1961, 83, 2213, DOI 10.1021/ja01470a053.

[161] G. Porter, M. F. Tchir, Chem. Commun. 1970, 6(20), 1372–1373, DOI 10.1039/C29700001372.

[162] T. Gruendling, K. K. Oehlenschlaeger, E. Frick, M. Glassner, C. Schmid, C. Barner-Kowollik, Macromol. Rapid Commun. 2011, 32, 807–812, DOI 10.1002/marc.201100159.

93 [163] T. Pauloehrl, G. Delaittre, V. Winkler, A. Welle, M. Bruns, H. G. Börner, A. M. Greiner, M. Bastmeyer, C. Barner-Kowollik, Angew. Chem. Int. Ed. 2012, 51, 1071–1074, DOI 10.1002/anie.201107095.

[164] M. A. B. Meador, M. A. Meador, L. L. Williams, D. A. Scheiman, Macromolecules 1996, 29(27), 8983, DOI 10.1021/ma960881n.

[165] O. Altintas, J. Willenbacher, K. N. Wuest, K. K. Oehlenschlaeger, P. Krolla- Sidenstein, H. Gliemann, C. Barner-Kowollik, Macromolecules 2013, 46(20), 8092–8101, DOI 10.1021/ma4015033.

[166] E. Harth, B. Van Horn, V. Y. Lee, D. S. Germack, C. P. Gonzales, R. D. Miller, C. J. Hawker, J. Am. Chem. Soc. 2002, 124(29), 8653–8660, DOI 10.1021/ja026208x.

[167] J. B. Beck, K. L. Killops, T. Kang, K. Sivanandan, A. Bayles, M. E. Mackay, K. L. Wooley, C. J. Hawker, Macromolecules 2009, 42(15), 5629–5635, DOI 10.1021/ma900899v.

[168] T. Nitsche, J. Steinkoenig, K. De Bruycker, F. R. Bloesser, S. J. Blanksby, J. P. Blinco, C. Barner-Kowollik, Macromolecules 2019, 52(6), 2597–2606, DOI 10.1021/acs.macromol.9b00203.

[169] T. Nitsche, S. J. Blanksby, J. P. Blinco, C. Barner-Kowollik, Polym. Chem. 2020, 11(10), 1696–1701, DOI 10.1039/c9py01910c.

[170] J. He, L. Tremblay, S. Lacelle, Y. Zhao, Soft Matter 2011, 7(6), 2380–2386, DOI 10.1039/c0sm01383h.

[171] K. N. R. Wuest, H. Lu, D. S. Thomas, A. S. Goldmann, M. H. Stenzel, C. Barner-Kowollik, ACS Macro. Lett. 2017, 1168–1174, DOI 10.1021/acsmacrolett.7b00659.

[172] C. F. Hansell, A. Lu, J. P. Patterson, R. K. O’Reilly, Nanoscale 2014, 6(8), 4102– 4107, DOI 10.1039/c3nr06706h.

[173] J. A. Pomposo, A. J. Moreno, A. Arbe, J. Colmenero, ACS Omega 2018, 3(8), 8648–8654, DOI 10.1021/acsomega.8b01331.

[174] N. Hosono, A. M. Kushner, J. Chung, A. R. Palmans, Z. Guan, E. W. Meijer, J. Am. Chem. Soc. 2015, 137(21), 6880–6888, DOI 10.1021/jacs.5b02967.

94 References

[175] D. A. Davis, A. Hamilton, J. Yang, L. D. Cremar, D. Van Gough, S. L. Potisek, M. T. Ong, P. V. Braun, T. J. Martínez, S. R. White, J. S. Moore, N. R. Sottos, Nature 2009, 459(7243), 68–72, DOI 10.1038/nature07970.

[176] A. Noel, Y. P. Borguet, K. L. Wooley, ACS Macro. Lett. 2015, 4(6), 645–650, DOI 10.1021/acsmacrolett.5b00227.

[177] M. J. Robb, W. Li, R. C. Gergely, C. C. Matthews, S. R. White, N. R. Sottos, J. S. Moore, ACS Cent. Sci. 2016, 2(9), 598–603, DOI 10.1021/acscentsci.6b00198.

[178] K. Makyła, C. Müller, S. Lörcher, T. Winkler, M. G. Nussbaumer, M. Eder, N. Bruns, Adv. Mater. 2013, 25(19), 2701–2706, DOI 10.1002/adma.201205226.

[179] S. Lörcher, T. Winkler, K. Makyła, C. Ouellet-Plamondon, I. Burgert, N. Bruns, J. Mater. Chem. A. 2014, 2(17), 6231, DOI 10.1039/c3ta14803c.

[180] D. Estupiñán, T. Gegenhuber, J. P. Blinco, C. Barner-Kowollik, L. Barner, ACS Macro. Lett. 2017, 6(3), 229–234, DOI 10.1021/acsmacrolett.7b00024.

[181] C. Heiler, J. T. Offenloch, E. Blasco, C. Barner-Kowollik, ACS Macro. Lett. 2017, 6(1), 56–61, DOI 10.1021/acsmacrolett.6b00858.

[182] H. Mutlu, C. W. Schmitt, N. Jasinski, H. Woehlk, K. Fairfull-Smith, J. P. Blinco, C. Barner-Kowollik, Polym. Chem. 2017, 8, 6199–6203, DOI 10.1039/C7PY01437F.

[183] S. Gnaim, D. Shabat, J. Am. Chem. Soc. 2017, 139(29), 10002–10008, DOI 10.1021/jacs.7b04804.

[184] Y. Chen, R. P. Sijbesma, Macromolecules 2014, 47(12), 3797–3805, DOI 10.1021/ma500598t.

[185] Y. Chen, A. J. H. Spiering, S. Karthikeyan, G. W. M. Peters, E. W. Meijer, R. P. Sijbesma, Nat. Chem. 2012, 4(7), 559–562, DOI 10.1038/nchem.1358.

[186] C. M. Geiselhart, H. Mutlu, P. Tzvetkova, C. Barner-Kowollik, Polym. Chem. 2020, 11(26), 4213–4220, DOI 10.1039/d0py00332h.

[187] P. R. Goldin, in The Columbia History of Chinese Literature, Columbia University Press, New York, 2001.

[188] Aristotle, De Anima, 350BC.

95 [189] P. the Elder, in Naturalis Historia, 77AD.

[190] A. Roda, Chemiluminescence and bioluminescence, Royal Society of Chemistry, Cambridge, 2011.

[191] B. R. Radziszewski, Ber. Dtsch. Chem. Ges. 1877, 10(1), 70–75, DOI 10.1002/cber.18770100122.

[192] K. R. Kopecky, C. Mumford, Can. J. Chem. 1969, 47(4), 709–711, DOI 10.1139/v69-114.

[193] F. McCapra, Chem. Commun. 1968, 3, 155–156, DOI 10.1039/c19680000155.

[194] N. J. Turro, A. Devaquet, J. Am. Chem. Soc. 1975, 97(13), 3859–3862, DOI 10.1021/ja00846a075.

[195] C. Tanaka, J. Tanaka, J. Phys. Chem. A 2000, 104(10), 2078–2090, DOI 10.1021/jp9931004.

[196] J. H. Wieringa, J. Strating, H. Wynberg, W. Adam, Tetrahedron Lett. 1972, 13(2), 169–172, DOI 10.1016/S0040-4039(01)84269-6.

[197] H. H. Seliger, W. D. McElroy, Arch. Biochem. Biophys. 1960, 88(1), 136–141, DOI 10.1016/0003-9861(60)90208-3.

[198] E. H. White, H. Worther, H. H. Seliger, W. D. McElroy, J. Am. Chem. Soc. 1966, 88(9), 2015–2019, DOI 10.1021/ja00961a030.

[199] E. H. White, H. Wörther, J. Org. Chem. 1966, 31(5), 1484–1488, DOI 10.1021/jo01343a039.

[200] A. P. Schaap, S. D. Gagnon, J. Am. Chem. Soc. 1982, 104(12), 3504–3506, DOI 10.1021/ja00376a044.

[201] A. P. Schaap, R. S. Handley, B. P. Giri, Tetrahedron Lett. 1987, 28(9), 935–938, DOI 10.1016/S0040-4039(00)95878-7.

[202] J. young Koo, G. B. Schuster, J. Am. Chem. Soc. 1978, 100(14), 4496–4503, DOI 10.1021/ja00482a030.

[203] W. Adam, I. Bronstein, A. V. Troﬁmov, R. F. Vasil’ev, J. Am. Chem. Soc. 1999, 121(5), 958–961, DOI 10.1021/ja982999k.

96 References

[204] W. Adam, M. Matsumoto, A. V. Troﬁmov, J. Am. Chem. Soc. 2000, 122(36), 8631–8634, DOI 10.1021/ja000408w.

[205] W. Adam, A. V. Troﬁmov, J. Org. Chem. 2000, 65(20), 6474–6478, DOI 10.1021/jo000495a.

[206] S. Sabelle, P. Y. Renard, K. Pecorella, S. De Suzzoni-Dézard, C. Créminon, J. Grassi, C. Mioskowski, J. Am. Chem. Soc. 2002, 124(17), 4874–4880, DOI 10.1021/ja0171299.

[207] S. Sabelle, J. Hydrio, E. Leclerc, C. Mioskowski, P. Y. Renard, Tetrahedron Lett. 2002, 43(20), 3645–3648, DOI 10.1016/S0040-4039(02)00617-2.

[208] M. Matsumoto, K. Ueda, T. Mizuno, N. Watanabe, in Proceedings of the 81st Annual Meeting of the Chemical Society of Japan, Tokyo, (Abstract 3PC066).

[209] M. Matsumoto, H. Murakami, N. Watanabe, in Proceedings of the 72nd Annual Meeting of the Chemical Society of Japan, Tokyo, (Abstract 2J242).

[210] M. Matsumoto, T. Mizuno, N. Watanabe, Chem. Commun. 2003, 3(4), 482–483, DOI 10.1039/b210857g.

[211] J. Cao, R. Lopez, J. M. Thacker, J. Y. Moon, C. Jiang, S. N. Morris, J. H. Bauer, P. Tao, R. P. Mason, A. R. Lippert, Chem. Sci. 2015, 6(3), 1979–1985, DOI 10.1039/c4sc03516j.

[212] N. Hananya, A. Eldar Boock, C. R. Bauer, R. Satchi-Fainaro, D. Shabat, J. Am. Chem. Soc. 2016, 138(40), 13438–13446, DOI 10.1021/jacs.6b09173.

[213] J. Cao, W. An, A. G. Reeves, A. R. Lippert, Chem. Sci. 2018, 9(9), 2552–2558, DOI 10.1039/c7sc05087a.

[214] J. Liu, J. Jiang, Y. Dou, F. Zhang, X. Liu, J. Qu, Q. Zhu, Org. Biomol. Chem. 2019, 17(29), 6975–6979, DOI 10.1039/c9ob01407a.

[215] E. A. Chandross, M. Hill, Tetrahedron Lett. 1963, 12, 761–765, DOI 10.1016/S0040-4039(01)90712-9.

[216] M. M. Rauhut, B. G. Roberts, A. M. Semsel, J. Am. Chem. Soc. 1966, 88(15), 3604–3617, DOI 10.1021/ja00967a025.

97 [217] M. M. Rauhut, L. J. Bollyky, B. G. Roberts, M. Loy, R. H. Whitman, A. V. Iannotta, A. M. Semsel, R. A. Clarke, J. Am. Chem. Soc. 1967, 89(25), 6515–6522, DOI 10.1021/ja01001a025.

[218] M. M. Rauhut, Accounts Chem. Res. 1969, 2(3), 80–87, DOI 10.1021/ar50015a003.

[219] S. P. Souza, M. Khalid, F. A. Augusto, W. J. Baader, J. Photoch. Photobio. A 2016, 321, 143–150, DOI 10.1016/j.jphotochem.2016.02.001.

[220] A. Tahirovic,´ A. Copra,ˇ E. Omanovic-Mikli´ canin,ˇ K. Kalcher, Talanta 2007, 72(4), 1378–1385, DOI 10.1016/j.talanta.2007.01.072.

[221] S. Yamashoji, A. Asakawa, S. Kawasaki, S. Kawamoto, Anal. Biochem. 2004, 333(2), 303–308, DOI 10.1016/j.ab.2004.05.043.

[222] F. Barni, S. W. Lewis, A. Berti, G. M. Miskelly, G. Lago, Talanta 2007, 72, 896–913, DOI 10.1016/j.talanta.2006.12.045.

[223] D. A. Beisswanger, B. H. Baretz, E. T. Rockwell Jr, E. T. Rockwell Sr, Emergency Lighting Device, 1989.

[224] B. Dubrow, E. Torrance, D. Guth, Packaged Chemiluminescent Material, 1973.

[225] H. F. Cordes, H. P. Richter, C. A. Heller, J. Am. Chem. Soc. 1969, 91(25), 7209, DOI 10.1021/ja01053a065.

[226] R. Bos, N. W. Barnett, G. A. Dyson, K. F. Lim, R. A. Russell, S. P. Watson, Anal. Chim. Acta 2004, 502(2), 141–147, DOI 10.1016/j.aca.2003.10.014.

[227] S. A. Tonkin, R. Bos, G. A. Dyson, K. F. Lim, R. A. Russell, S. P. Watson, C. M. Hindson, N. W. Barnett, Anal. Chim. Acta 2008, 614(2), 173–181, DOI 10.1016/j.aca.2008.03.009.

[228] C. V. Stevani, W. J. Baader, J. Chem. Research 2009, 2002(9), 430–432, DOI 10.3184/030823402103172752.

[229] A. L. Baumstark, M. E. Landis, P. J. Brooks, J. Org. Chem. 1979, 44(24), 4251– 4253, DOI 10.1021/jo01338a008.

[230] C. L. R. Catherall, T. F. Palmer, R. B. Cundall, J. Chem. Soc. Farad. T. 2 1984, 80, 837–849, DOI 10.1039/F29848000837.

98 References

[231] R. Koike, Y. Kato, J. Motoyoshiya, Y. Nishii, H. Aoyama, Luminescence 2006, 21(3), 164–173, DOI 10.1002/bio.901.

[232] N. H. Park, G. D. P. Gomes, M. Fevre, G. O. Jones, I. V. Alabugin, J. L. Hedrick, Nat. Commun. 2017, 8(1), 1553, DOI 10.1038/s41467-017-01129-8.

[233] C. V. Stevani, D. F. Lima, V. G. Toscano, W. J. Baader, J. Chem. Soc. Perk. T. 2 1996, 5, 989, DOI 10.1039/p29960000989.

[234] C. V. Stevani, I. P. D. A. Campos, W. J. Baader, J. Chem. Soc. Perk. T. 2 1996, 8, 1645–1648.

[235] T. Maruyama, S. Narita, J. Motoyoshiya, J. Photoch. Photobio. A 2013, 252, 222–231, DOI 10.1016/j.jphotochem.2012.12.006.

[236] F. A. Augusto, A. Francés-Monerris, I. Fdez Galván, D. Roca-Sanjuán, E. L. Bastos, W. J. Baader, R. Lindh, Phys. Chem. Chem. Phys. 2017, 19(5), 3955–3962, DOI 10.1039/c6cp08154a.

[237] G. B. Schuster, Accounts Chem. Res. 1979, 12(1977), 366–373, DOI 10.1021/ar50142a003.

[238] S. P. Schmidt, G. B. Schuster, J. Am. Chem. Soc. 1980, 102(1), 306–314, DOI 10.1021/ja00521a049.

[239] S. M. Silva, K. Wagner, D. Weiss, R. Beckert, C. V. Stevani, W. J. Baader, Lumi- nescence 2002, 17(6), 362–369, DOI 10.1002/bio.694.

[240] C. V. Stevani, S. M. Silva, W. J. Baader, Eur. J. Org. Chem. 2000, 4037–4046, DOI 10.1016/j.aca.2008.03.009.

[241] L. F. Ciscato, F. H. Bartoloni, E. L. Bastos, W. J. Baader, J. Org. Chem. 2009, 74(23), 8974–8979, DOI 10.1021/jo901402k.

[242] L. F. M. Ciscato, F. A. Augusto, D. Weiss, F. H. Bartoloni, S. Albrecht, H. Bran- dl, T. Zimmermann, W. J. Baader, Arkivoc 2012, 2012(3), 391–430, DOI 10.3998/ark.5550190.0013.328.

[243] M. Vacher, I. Fdez Galván, B. W. Ding, S. Schramm, R. Berraud-Pache, P.Naumov, N. Ferré, Y. J. Liu, I. Navizet, D. Roca-Sanjuán, W. J. Baader, R. Lindh, Chem. Rev. 2018, 118(15), 6927–6974, DOI 10.1021/acs.chemrev.7b00649.

99 [244] T. Wilson, Photochem. Photobiol. 1995, 62(4), 601–606, DOI 10.1051/jphyscol: 1980419.

[245] E. L. Bastos, S. M. Da Silva, W. J. Baader, J. Org. Chem. 2013, 78(9), 4432–4439, DOI 10.1021/jo400426y.

[246] H. Isobe, Y. Takano, M. Okumura, S. Kuramitsu, K. Yamaguchi, J. Am. Chem. Soc. 2005, 127(24), 8667–8679, DOI 10.1021/ja043295f.

[247] Y. Takano, T. Tsunesada, H. Isobe, Y. Yoshioka, K. Yamaguchi, I. Saito, Bull. Chem. Soc. Jpn. 1999, 72(2), 213–225, DOI 10.1246/bcsj.72.213.

[248] S. Li, X. Li, J. Xu, X. Wei, Talanta 2008, 75, 32–37, DOI 10.1016/j.talanta.2007.10.001.

[249] G. Merenyi, J. Lind, T. E. Eriksen, J. Biolum. Chemilum. 1990, 5, 53–56, DOI 10.1002/bio.1170050111.

[250] S. J. Shaw, K. J. Elgie, C. Edwards, R. W. Boyle, Tetrahedron Lett. 1999, 40(8), 1595–1596, DOI 10.1016/S0040-4039(98)02653-7.

[251] J. M. Noy, M. Koldevitz, P. J. Roth, Polym. Chem. 2015, 6(3), 436–447, DOI 10.1039/c4py01238k.

[252] P. Boufﬂet, A. Casey, Y. Xia, P. N. Stavrinou, M. Heeney, Chem. Sci. 2017, 8(3), 2215–2225, DOI 10.1039/c6sc04427a.

[253] Y. Pei, J. M. Noy, P. J. Roth, A. B. Lowe, Polym. Chem. 2015, 6(11), 1928–1931, DOI 10.1039/c4py01558d.

[254] E. A. Qian, A. L. Rheingold, E. H. Moully, D. Jung, D. Mosallaei, Y. Han, M. S. Messina, A. I. Wixtrom, P. Rehak, H. D. Maynard, J. C. Axtell, J. Y. Wang, P. Král, A. Saebi, A. M. Spokoyny, S. Chow, A. T. Royappa, Nat. Chem. 2016, 9(4), 333–340, DOI 10.1038/nchem.2686.

[255] L. Dumas, E. Fleury, D. Portinha, Polymer 2014, 55(11), 2628–2634, DOI 10.1016/j.polymer.2014.04.002.

[256] W. N. Olmstead, Z. Margolin, F. G. Bordwell, J. Org. Chem. 1980, 45(16), 3295– 3299, DOI 10.1021/jo01304a032.

[257] F. G. Bordwell, Accounts Chem. Res. 1988, 21(12), 456–463, DOI 10.1021/ar00156a004.

100 References

[258] A. Kütt, S. Selberg, I. Kaljurand, S. Tshepelevitsh, A. Heering, A. Darnell, K. Kaup- mees, M. Piirsalu, I. Leito, Tetrahedron Lett. 2018, 59(42), 3738–3748, DOI 10.1016/j.tetlet.2018.08.054.

[259] J. Chen, L. Dumas, J. Duchet-Rumeau, E. Fleury, A. Charlot, D. Portinha, J. Polym. Sci. A1 2012, 50(16), 3452–3460, DOI 10.1002/pola.26136.

[260] F. Cavalli, F. R. Bloesser, C. Barner-Kowollik, L. Barner, Chem. Eur. J. 2019, 25(43), 10049–10053, DOI 10.1002/chem.201901290.

[261] N. Busatto, J. L. Keddie, P. J. Roth, Polym. Chem. 2020, 11(3), 704–711, DOI 10.1039/c9py01585j.

[262] H. Turgut, N. Dingenouts, V. Trouillet, P. Krolla-Sidenstein, H. Gliemann, G. Delait- tre, Polym. Chem. 2019, 10(11), 1344–1356, DOI 10.1039/c8py01777h.

[263] H. Zhou, J. A. Johnson, Angew. Chem. Int. Ed. 2013, 52(8), 2235–2238, DOI 10.1002/anie.201207966.

[264] Y. Gu, K. Kawamoto, M. Zhong, M. Chen, M. J. Hore, A. M. Jordan, L. S. T. Korley, B. D. Olsen, J. A. Johnson, P. Natl. Acad. Sci. USA 2017, 114(19), 4875–4880, DOI 10.1073/pnas.1620985114.

[265] J. Wang, T. S. Lin, Y. Gu, R. Wang, B. D. Olsen, J. A. Johnson, ACS Macro. Lett. 2018, 7(2), 244–249, DOI 10.1021/acsmacrolett.8b00008.

[266] S. Hisamatsu, S. Suzuki, S. Kohmoto, K. Kishikawa, Y. Yamamoto, R. Motokawa, T. Yaita, Tetrahedron 2017, 73, 3993–3998, DOI 10.1016/j.tet.2017.05.084.

[267] S. Gnaim, D. Shabat, Chem. Commun. 2018, 54(21), 2655–2658, DOI 10.1039/c8cc00521d.

[268] B. Gu, C. Dong, R. Shen, J. Qiang, T. Wei, F. Wang, S. Lu, X. Chen, Sensor. Actuat. B-Chemical 2019, 301(September), 127111, DOI 10.1016/j.snb.2019.127111.

[269] C. S. Sauer, J. Köckenberger, M. R. Heinrich, J. Org. Chem. 2020, 85, 9331–9338, DOI 10.1021/acs.joc.0c00569.

[270] M. Khalid, S. P. Souza, L. F. Ciscato, F. H. Bartoloni, W. J. Baader, Photochem. Photobiol. Sci. 2015, 14(7), 1296–1305, DOI 10.1039/c5pp00152h.

[271] K. A. Dill, J. L. Maccallum, Science 2012, 338(6110), 1042–1046, DOI 10.1126/science.1219021.

101 [272] J. P. Cole, A. M. Hanlon, K. J. Rodriguez, E. B. Berda, J. Polym. Sci. A1 2017, 55(2), 191–206, DOI 10.1002/pola.28378.

[273] J.-F. Lutz, M. Ouchi, D. R. Liu, M. Sawamoto, Science 2013, 341, 628–636, DOI 10.1126/science.1238149.

[274] M. Ouchi, N. Badi, J.-F. Lutz, M. Sawamoto, Nat. Chem. 2011, 3(12), 917–924, DOI 10.1038/NCHEM.1175.

[275] A. Sanchez-Sanchez, S. Akbari, A. Etxeberria, A. Arbe, U. Gasser, A. J. Moreno, J. Colmenero, J. A. Pomposo, ACS Macro. Lett. 2013, 2(6), 491–495, DOI 10.1021/mz400173c.

[276] M. González-Burgos, E. González, J. A. Pomposo, Macromolecular Rapid Com- munications 2018, 39(6), 1–5, DOI 10.1002/marc.201700675.

[277] V. Kobernik, I. Berkovich, A. Levy, N. G. Lemcoff, C. E. Diesendruck, Chemistry - A European Journal 2020, 26(68), 15835–15838, DOI 10.1002/chem.202003330.

[278] A. Levy, F. Wang, A. Lang, O. Galant, C. E. Diesendruck, Angewandte Chemie - International Edition 2017, 56(23), 6431–6434, DOI 10.1002/anie.201612242.

[279] A. Levy, R. Feinstein, C. E. Diesendruck, J. Am. Chem. Soc. 2019, 141(18), 7256–7260, DOI 10.1021/jacs.9b01960.

[280] A. Levy, H. Goldstein, D. Brenman, C. E. Diesendruck, Journal of Polymer Science 2020, 58(5), 692–703, DOI 10.1002/pol.20190217.

[281] T. S. Fischer, D. Schulze-Sünninghausen, B. Luy, O. Altintas, C. Barner- Kowollik, Angew. Chem. Int. Ed. 2016, 55(37), 11276–11280, DOI 10.1002/anie.201602894.

[282] J. Zhang, J. Tanaka, P. Gurnani, P. Wilson, M. Hartlieb, S. Perrier, Polym. Chem. 2017, 8(28), 4079–4087, DOI 10.1039/c7py00828g.

[283] T. S. Fischer, S. Spann, Q. An, B. Luy, M. Tsotsalas, J. P. Blinco, H. Mut- lu, C. Barner-Kowollik, Chemical Science 2018, 9(20), 4696–4702, DOI 10.1039/c8sc01009a.

[284] W. J. Cheong, S. H. Yang, F. Ali, J. Sep. Sci. 2013, 36(3), 609–628, DOI 10.1002/jssc.201200784.

102 References

[285] P. S. Sharma, Z. Iskierko, A. Pietrzyk-Le, F. D’Souza, W. Kutner, Electrochemi. Commun. 2015, 50, 81–87, DOI 10.1016/j.elecom.2014.11.019.

[286] J. Shu, Z. Qiu, Q. Zhou, Y. Lin, M. Lu, D. Tang, Anal. Chem. 2016, 88(5), 2958– 2966, DOI 10.1021/acs.analchem.6b00262.

[287] X. Zhen, C. Zhang, C. Xie, Q. Miao, K. L. Lim, K. Pu, ACS Nano 2016, 10(6), 6400–6409, DOI 10.1021/acsnano.6b02908.

[288] R. Weissleder, V. Ntziachristos, Nat. Med. 2003, 9(1), 123–128, DOI 10.1038/nm0103-123.

[289] C. Dodeigne, L. Thunus, R. Lejeune, Talanta 2000, 51(3), 415–439, DOI 10.1016/S0039-9140(99)00294-5.

[290] L. Delafresnaye, F. R. Bloesser, K. B. Kockler, C. W. Schmitt, I. M. Ir- shadeen, C. Barner-Kowollik, Chem. Eur. J. 2020, 26(1), 114–127, DOI 10.1002/chem.201904054.

[291] Y. Jiang, J. Chen, C. Deng, E. J. Suuronen, Z. Zhong, Biomaterials 2014, 35(18), 4969–4985, DOI 10.1016/j.biomaterials.2014.03.001.

[292] T. He, G. N. Wang, J. X. Liu, W. L. Zhao, J. J. Huang, M. X. Xu, J. P. Wang, J. Liu, Food Chem. 2019, 288(March), 347–353, DOI 10.1016/j.foodchem.2019.03.031.

[293] H. Neuvonen, J. Chem. Soc. Perk. T. 2 1995, 945–949, DOI 10.1039/p29950000945.

[294] S. Gnaim, O. Green, D. Shabat, Chem. Commun. 2018, 54(17), 2073–2085, DOI 10.1039/c8cc00428e.

[295] N. Hananya, D. Shabat, ACS Cent. Sci. 2019, 5(6), 949–959, DOI 10.1021/acscentsci.9b00372.

[296] M. Matsumoto, J. Photoch. Photobio. C 2004, 5(1), 27–53, DOI 10.1016/j.jphotochemrev.2004.02.001.

[297] D. R. Lide, Viscosity of Liquids, CRC Press, Boca Raton, FL, 2005, 6–186 –6–190.

[298] M. Van De Walle, K. De Bruycker, J. P. Blinco, C. Barner-Kowollik, Angew. Chem. Int. Ed. 2020, 59(33), 14143–14147, DOI 10.1002/anie.202003130.

[299] C. L. Catherall, T. F. Palmer, R. B. Cundall, J. Chem. Soc. Farad. T. 2 1984, 80(7), 823–834, DOI 10.1039/F29848000823.

103 8 Appendix

104 Appendix

105 Statements of Contribution

8.1 Statements of Contribution

Statement of Contribution of Co-Authors for Thesis by Published Paper

The authors listed below have certified that:

1. they meet the criteria for authorship and that they have participated in the conception, execution, or interpretation, of at least that part of the publication in their field of expertise; 2. they take public responsibility for their part of the publication, except for the responsible author who accepts overall responsibility for the publication; 3. there are no other authors of the publication according to these criteria; 4. potential conflicts of interest have been disclosed to (a) granting bodies, (b) the editor or publisher of journals or other publications, and (c) the head of the responsible academic unit, and 5. they agree to the use of the publication in the student’s thesis and its publication on the QUT’s ePrints site consistent with any limitations set by publisher requirements.

In the case of this chapter:

Please state the publication title and date of publication or status:

All Eyes on Visible Light Peroxyoxalate Chemiluminescence Read-Out Systems, Chem. Eur. J. 2020, 26, 114-127.

Contributor Statement of contribution* • Literature review Fabian R. Bloesser • Preparation and editing of manuscript • Literature review Laura Delafresnaye • Preparation and editing of manuscript • Literature review Katrin B. Kockler • Preparation and editing of manuscript • Literature review Christian W. Schmitt • Preparation and editing of manuscript • Literature review Ishrath M. Irshadeen • Preparation and editing of manuscript • Overarching research concept Christopher Barner-Kowollik • Supervision and editing of manuscript

Principal Supervisor Confirmation

I have sighted email or other correspondence from all Co-authors confirming their certifying authorship. (If the Co-authors are not able to sign the form please forward their email or other correspondence confirming the certifying authorship to the GRC).

Christopher Barner-Kowollik QUT Verified Signature 18 January 2021 Name Signature Date

106 Appendix

Statement of Contribution of Co-Authors for Thesis by Published Paper

The authors listed below have certified that:

In the case of this chapter:

Please state the publication title and date of publication or status:

Self-Propagated para-Fluoro-Thiol Reaction, Chem. Eur. J. 2019, 25, 10049-10053.

Contributor Statement of contribution* • Research design • Synthesis of linker Fabian R. Bloesser • PFTR experiments • Analysis • Preparation and editing of manuscript • Research design • Synthesis of linker Federica Cavalli • PFTR experiments • Analysis • Preparation and editing of manuscript • Overarching research concept Christopher Barner-Kowollik • Supervision and editing of manuscript • Overarching research concept Leonie Barner • Supervision and editing of manuscript

Principal Supervisor Confirmation

Christopher Barner-Kowollik QUT Verified Signature 18 January 2021 Name Signature Date

107 Statements of Contribution

Statement of Contribution of Co-Authors for Thesis by Published Paper

The authors listed below have certified that:

In the case of this chapter:

Please state the publication title and date of publication or status:

Chemiluminescent Read-Out of para-Fluoro – Thiol Reaction, Chem. Commun. 2020, 56, 14996-14999.

Contributor Statement of contribution* • Research design • PFTR experiments Fabian R. Bloesser • Analysis • Chemiluminescence Read-Out • Preparation and editing of manuscript • Research design • PFTR experiments Federica Cavalli • Analysis • Chemiluminescence Read-Out • Preparation and editing of manuscript • Assistance CL read-out and analysis Sarah L. Walden • Editing of manuscript • Overarching research concept Leonie Barner • Supervision and editing of manuscript • Overarching research concept Christopher Barner-Kowollik • Supervision and editing of manuscript

Principal Supervisor Confirmation

Christopher Barner-Kowollik QUT Verified Signature 18 January 2021 Name Signature Date

108 Appendix

Statement of Contribution of Co-Authors for Thesis by Published Paper

The authors listed below have certified that:

In the case of this chapter:

Please state the publication title and date of publication or status:

Chemiluminescent Self-Reported Unfolding of Single-Chain Nanoparticles, submitted.

Contributor Statement of contribution* • Research design • Synthesis of monomers, polymers, linkers, SCNPs Fabian R. Bloesser • Analysis • Chemiluminescence Read-Out • Preparation and editing of manuscript • Assistance with interpretation chemiluminescence data and analysis Sarah L. Walden • Assistance with research design • Editing of manuscript • DOSY measurements and analysis Ishrath M. Irshadeen • Editing of manuscript • SEC instrument preparation, measurements and analysis Lewis C. Chambers • Editing of manuscript • Overarching research concept Christopher Barner-Kowollik • Supervision and editing of manuscript

Principal Supervisor Confirmation

Christopher Barner-Kowollik QUT Verified Signature 18 January 2021 Name Signature Date

109 Supporting Information for Chapter 3

8.2 Supporting Information for Chapter 3

Supporting Information

Self-Propagated para-Fluoro – Thiol Reaction

Federica Cavalli,+[a] Fabian R. Bloesser,+[b], Chrisopher Barner-Kowollik,*[b,c,d] and Leonie Barner*[a,b,d] chem_201901290_sm_miscelaneous_information.pdf

110 Appendix

I. Materials

1,8-diazabicyclo[5.4.0]undec-7-ene (DBU, ≥98%, Merck), thiophenol (TP, ≥99%, Sigma Aldrich), 4-methoxy-α-toluenethiol (MBM, 90%, Sigma Aldrich), dodecanethiol (DDT, ≥98%, Sigma Aldrich), tricarballylic acid (98%, Alfa Aesar), 2,3,4,5,6-pentaﬂuorobenzyl

bromide (97%, Acros Organic), caesium carbonate (Cs2CO3, 99%, Sigma Aldrich), magnesium sulphate (Mg2SO4, ≥99.8%, Roth) were used as received. N,N-dimethylformamide (DMF, AnalR), tetrahydrofuran (THF, AnalR), dichloromethane (DCM, AnalR), cyclohexane (CH, AnalR) and ethylacetate (EA, AnalR), dry DMF (>99%, Acros Organic) were used as solvents. Deuterated solvent such as chloroform-d (CDCl3, >99.8%), N,N-dimethylformamide-d7 (99.5%), dimethylsulfoxide-d6 (99.8%), tetrahydrofuran-d8 (99.5%) were purchased from Eurisotop and used as received.

II. Characterisation Methods

Nuclear Magnetic Resonance Spectroscopy

NMR spectra were recorded on a Bruker AM400 spectrometer at 298.1 K (400.3 MHz for 1H, 376.6 MHz for 19F and 100.7 MHz for 13C, respectively). Chemical shifts are expressed in parts per million (ppm) relative to tetramethylsilane (TMS) and referenced 1 on characteristic H solvent resonances as internal standards [CDCl3: 7.26 ppm; DMF- 19 13 d7: 8.03 ppm; DMSO-d6: 2.50 ppm; THF-d8: 1.72 ppm]. F and C spectra were referenced via the according Ξ values (19F: Ξ = 94.094; 13C: Ξ = 25.145) based on the corresponding 1H NMR spectrum. 1H NMR are reported as follows: chemical shift (δ in ppm), multiplicity (s for singlet, d for doublet, t for triplet, q for quartet, p for pentet, m for multiplet), coupling constant(s) (Hz), number of protons (concluded from the integrals), specific assignment. 19F NMR spectra were subjected to baseline correction via a multipoint fit function. 13C-{1H} and 19F NMR spectra are reported in terms of chemical shift and specific assignment.

Mass Spectrometry

electrospray ionisation mass spectrometry (ESI-MS) spectra were recorded on a Q Exactive (Orbitrap) mass spectrometer (ThermoFisher Scientific, San Jose, CA, USA) equipped with a HESI II probe. The instrument was calibrated in the m/z range of 74 - 1822 by using a premixed calibration solutions from Thermo Scientific. A constant spray voltage of 3.6 kV and a dimensionless sheath gas flow rate of 5 was applied. The capillary temperature and the S-lens RF level were set to 320 °C and 62.0, respectively.

111 Supporting Information for Chapter 3

The samples were dissolved in a mixture of THF and MeOH (3:2) containing 100 µmol sodium triﬂuoracetate (NaTFA) and were injected with a ﬂow rate of 5 to 10 µL min-1.

III. Syntheses

Tris(2,3,4,5,6-pentaﬂuorobenzyl)(1,2,3-Propanetricarboxylate) (3PFB)

In a pre-dried Schlenk ﬂask, Cs2CO3 (17.5 mmol, 6.174 g, 5.00 equiv.) was dispersed in dry DMF (15.00 mL) under argon atmosphere before adding tricarballylic acid (3.50 mmol, 0.616 g, 1.00 equiv.). After stirring for 30 min, 2,3,4,5,6-pentaﬂuorobenzyl bromide (10.9 mmol, 1.64 mL, 3.10 equiv.) was added and the reaction mixture was stirred

overnight (16 h) at ambient temperature. The following day, Cs2CO3 was filtered off and the filtrate diluted with water (1x 20 mL) and extracted with DCM (2x 20 mL). The collected organic phases were washed with brine (1x 40 mL), dried over magnesium sulphate and concentrated under reduced pressure. The crude product was subjected to purification on a silica gel column (7:3 CH:EA) to afford a white crystalline compound (yield 90%). 1 H NMR (CDCl3): δ = 5.20 (m, 6H, CH2-O), 3.27 (p, 6.5 Hz, 1H, CH), 2.78 (dd, 17.1 Hz, 6.5 Hz, 2H, CH2-COO), 2.64 (dd, 17.1 Hz, 6.5 Hz, 2H, CH2-COO) ppm. 19 F NMR (CDCl3): δ = -142.01 (m, 2F, ortho), -152.14 (m, 1F, para), -161.51 (m, 2F, meta) ppm. 13 C NMR (CDCl3): δ = 171.93 (COO), 170.41 (COO), 145.81 (CFmeta), 142.89 (CFpara), 137.66 (CFortho), 109.01 (Carom,q), 54.30 (CH2-O), 53.81 (CH2-O), 37.09 (CH), 34.70 (CH2-COO) ppm. + HR-ESI-MS: m/z = 739.0192 (M+Na , calculated: 739.0208, ∆abs = 0.0016, ∆rel = 2.27 ppm).

General Procedure for PFTR

The thiol (3.00 equiv.) and the 3PFB (1.00 equiv.) were dissolved in either THF, DMF -1 or DMSO at a.t. to afford a solution where [thiol]0 = 75 mmol L . Next, an understoichiometric amount of the base (DBU, TBAOH, TBAF or TBABr) was added to the reaction mixture to start the reaction, for the equivalents refer to the corresponding ﬁgure caption. At desired intervals, an aliquot is withdrawn and immediately submitted for 19F NMR in order to follow the kinetics. N.B.: All reactions have been carried out in single-use glass vials, acting as a HF quencher.

112 Appendix

19 F NMR (CDCl3): δ = -135.48 (m, 2F, ortho), -143.57 (m, 2F, meta) ppm.

(3PFB)(Phenyl) thioether

3PFB (53.7 mg, 75.0 µmol, 1.00 equiv.) and thiophenol (23.1 µL, 24.8 mg, 225 µmol, 3.00 equiv.) were dissolved in DMSO or DMF (3 mL) and TBAF (1 M in THF, 22.5 µL, 22.5 µmol, 0.30 equiv.) was added. The reaction was allowed to proceed at ambient temperature for 30 min (DMSO) or 4 hours (DMF). Subsequently, the mixture was diluted with DCM (10 mL), passed over basic alumina in order to remove any unreacted HF, and ﬁnally washed with water (10 mL). The phases were separated, the organic phase was

dried over MgSO4 and the solvent was removed under reduced pressure to yield 51.3 mg (69.3%, DMSO) or 55.2 mg (74.6%, DMF) of an off-white oil. 1 H NMR (CDCl3): δ = 7.37 (m, 3H, CHarom), 7.28 (m, 12H, CHarom), 5.22 (s, 2H, CH2-O), 5.19 (s, 4H, CH2-O), 3.29 (p, 6.5 Hz, 1H, CH), 2.80 (dd, 17.1 Hz, 6.5 Hz, 2H, CH2-COO), 2.66 (dd, 17.1 Hz, 6.5 Hz, 2H, CH2-COO) ppm. 19 F NMR (CDCl3): δ = -132.67 (m, 2F, meta), -141.63 (m, 2F, ortho) ppm. 13 C NMR (CDCl3): δ = 171.98 (COO), 170.44 (COO), 146.84 (CF), 145.47 (CF), 132.70 (Carom,q), 131.05 (CHarom), 129.51 (CHarom), 128.19 (CHarom), 115.25 (Carom,q), 54.65 (CH2-O), 53.15 (CH2-O), 37.12 (CH), 34.71 (CH2-COO) ppm. + HR-ESI-MS: m/z = 1136.1834 (M+NH4 , calculated: 1136.1825, ∆abs = 0.0009, ∆rel = 0.79 ppm).

(3PFB)(4-methoxybenzyl) thioether

3PFB (53.7 mg, 75.0 µmol, 1.00 equiv.) and 4-methoxy-α-toluenethiol (31.3 µL, 34.7 mg, 225 µmol, 3.00 equiv.) were dissolved in DMSO (3 mL) and TBAF (1 M in THF, 22.5 µL, 22.5 µmol, 0.30 equiv.) was added. The reaction was allowed to proceed at ambient temperature for 48 hours. Subsequently, the mixture was diluted with DCM (10 mL), passed over basic alumina in order to remove any unreacted HF, and ﬁnally washed with water (10 mL). The phases were separated, the organic phase was dried over MgSO4 and the solvent was removed under reduced pressure to yield 65.2 mg an off-white oil (77.7%). 1 H NMR (CDCl3): δ = 2.65 (dd, 17.1 Hz, 6.5 Hz, 2H, CH2-COO), 2.78 (dd, 17.1 Hz, 6.5 Hz, 2H, CH2-COO), 3.27 (p, 6.5 Hz, 1H, CH), 3.76 (s, 9H, CH3-O), 4.11 (s, 6H, CH2-S), 5.17 (s, 4H, CH2-O), 5.19 (s, 2H, CH2-O), 6.78 (d, 8.6 Hz, 6H, CHarom), 7.18 (d, 8.6 Hz, 6H, CHarom) ppm. 19 F NMR (CDCl3): δ = -133.59 (m, 2F, meta), -142.55 (m, 2F, ortho) ppm.

113 Supporting Information for Chapter 3

13 C NMR (CDCl3): δ = 171.99 (COO), 171.45 (COO), 159.28 (COarom), 146.81 (CF), 145.19 (CF), 130.17 (CHarom), 128.26 (Carom,q, 115.77 (Carom,q,), 114.16 (CHarom), 113.74 (Carom,q,), 55.37 (CH3-O), 54.67 (CH2-O), 54.17 (CH2-O), 41.04 (CH2-S), 37.10 (CH), 34.70 (CH2-COO) ppm. + HR-ESI-MS: m/z = 1004.1058 (M+NH4 , calculated: 1004.1038, ∆abs = 0.0020, ∆rel = 1.99 ppm).

1 Figure 8.1: H NMR spectrum (CDCl3, 400 MHz) of 3PFB.

114 Appendix

19 Figure 8.2: F NMR spectrum (CDCl3, 377 MHz) of 3PFB.

13 Figure 8.3: C NMR spectrum (CDCl3, 101 MHz) of 3PFB.

115 Supporting Information for Chapter 3

IV. Calculation of Conversion

Conversion of PFTR was calculated from 19F NMR by comparing the ratios of parent ortho- (o), meta- (m) as well as para-ﬂuoro (p) resonances with the ortho’ (o’)- and meta’-ﬂuoro (m’) resonances of the PFTR product, as shown in Figure 7.4, according to equation 7.1.

19 Figure 8.4: F NMR spectrum (CDCl3, 377 MHz) of 3PFB after approx. 30% conversion in a PFTR.

0.5 ∗ m0 0.5 ∗ 1.71 conversion(c) = = ≈ 0.461 ≡ 46.1% (8.1) 0.5 ∗ m0 + p 0.5 ∗ 1.71 ∗ 1.00

V. Error Propagation

As NMR commonly exhibits errors of up to 10%, error propagation was performed according to equation 7.2, where ∂/∂c is the derivation of the conversion (refer to 7.1) and ∆ is the error of the according integral:

116 Appendix

0 ∂m0 ∂p0 0 0 Error = ∗ ∆m + ∗ ∆p ∂c ∂c

2p –m0

= 0 2 ∗ 1.71 ∗ 0.1 + 0 2 ∗ 1.00 ∗ 0.1 (m + 2p) 2(p + 0.5 ∗ m ) (8.2)

2 ∗ 1.00 –1.71 = ∗ 0.171 + ∗ 0.1 (1.71 + 2 ∗ 1.00)2 2(1.00 + 0.5 ∗ 1.71)2 ≈ 0.0497 ≡ 4.97%

Table 8.1: Conversion and error propagation calculation for the reaction of DDT with 3PFB in THF employing 0.10 equivalents of base.

TBAOH TBAF DBU time / h rel. conv. ∆conv. time / h rel. conv. ∆conv. time / h rel. conv. ∆conv. 0.08333 0.26199 0.03867 0.01667 0.23372 0.03582 0.06667 0.0566 0.01068 0.16667 0.26199 0.03867 0.16667 0.23954 0.03643 0.08333 0.07407 0.01372 0.21667 0.27273 0.03967 0.33333 0.24812 0.03731 0.16667 0.0991 0.01786 0.33333 0.27273 0.03967 0.41667 0.24812 0.03731 0.25 0.12281 0.02155 0.5 0.28571 0.04082 0.83333 0.25926 0.03841 0.33333 0.1453 0.02484 0.83333 0.29329 0.04145 1.5 0.2674 0.03918 0.41667 0.15254 0.02585 2 0.29825 0.04186 2.5 0.27273 0.03967 0.5 0.16318 0.02731 2.5 0.30556 0.04244 4 0.28058 0.04037 0.66667 0.17012 0.02824 3 0.31034 0.04281 7 0.28571 0.04082 0.83333 0.18699 0.03041 6 0.32203 0.04367 10 0.29078 0.04125 1 0.19355 0.03122 11 0.3311 0.04429 24 0.30314 0.04225 2 0.21875 0.03418 15 0.33333 0.04444 48 0.32203 0.04367 4 0.23664 0.03613 24 0.34211 0.04501 6 0.25651 0.03814 34 0.35275 0.04566 16 0.2674 0.03918 48 0.35897 0.04602 24 0.27273 0.03967 32 0.28571 0.04082 40 0.29078 0.04125

117 Supporting Information for Chapter 3

Table 8.2: Conversion and error propagation calculation for the reaction of DDT with 3PFB in THF employing 0.50 equivalents of base.

TBAOH TBAF DBU time / h rel. conv. ∆conv. time / h rel. conv. ∆conv. time / h rel. conv. ∆conv. 0.06667 0.72376 0.03999 0.05 0.6124 0.04747 0.06667 0.13793 0.02378 0.08333 0.74359 0.03813 0.08333 0.63636 0.04628 0.25 0.33333 0.04444 0.13333 0.75 0.0375 0.13333 0.64912 0.04555 0.28333 0.34426 0.04515 0.16667 0.76744 0.03569 0.25 0 0.35 0.375 0.04688 0.25 0.78022 0.0343 0.33333 0.67213 0.04407 0.63333 0.44444 0.04938 0.33333 0.78836 0.03337 0.5 0.67742 0.0437 1 0.5 0.05 0.41667 0.79675 0.03239 1 0.69697 0.04224 1.38333 0.5122 0.04997 0.5 0.80392 0.03153 3 0.71831 0.04047 2.33333 0.57895 0.04875 0.66667 0.80952 0.03084 9 0.729 0.03951 3 0.59016 0.04837 0.83333 0.81395 0.03029 24 0.75 0.0375 4 0.61977 0.04713 1 0.82456 0.02893 6 0.64912 0.04555 1.5 0.83471 0.02759 9 0.67213 0.04407 2 0.84375 0.02637 13 0.68254 0.04334 3 0.84375 0.02637 24 0.65517 0.04518 5 0.84962 0.02555 9 0.86207 0.02378 16 0.87097 0.02248 24 0.87879 0.0213

Table 8.3: Conversion and error propagation calculation for the reaction of DDT with 3PFB in THF when adding TBAF in a single portion or in a step-wise manner.

Single addition Step-wise addition equiv. of TBAF rel. conv. ∆conv. equiv. of TBAF rel. conv. ∆conv. 0.1 0.28 0.053 0.1 0.28 0.04014 0.2 0.46 0.04967 0.2 0.58 0.04889 0.3 0.65 0.04555 0.3 0.93 0.01327

118 Appendix

Table 8.4: Conversion and error propagation calculation for the reaction of TP, MBM, and DDT, respectively, with 3PFB in THF employing 0.10 equivalents of TBAF.

thiophenol 4-Methoxybenzyl mercaptan Dodecanethiol time / h rel. conv. ∆conv. time / h rel. conv. ∆conv. time / h rel. conv. ∆conv. 2 0.34426 0.04515 2 0.3234 0.04444 1 0.25926 0.03841 4 0.38838 0.04751 4 0.34426 0.04515 3 0.27273 0.03967 24 0.48718 0.04997 30 0.39394 0.04775 7 0.28571 0.04082 48 0.53488 0.04976 48 0.43503 0.04916 24 0.30314 0.04225 90 0.58333 0.04861 72 0.46368 0.04986 48 0.31034 0.04281 120 0.50488 0.04976 72 0.33 0.04444 96 0.35 0.04579

Table 8.5: Conversion and error propagation calculation for the reaction of TP, MBM, and DDT, respectively, with 3PFB in DMF employing 0.10 equivalents of TBAF.

thiophenol 4-Methoxybenzyl mercaptan Dodecanethiol time / h rel. conv. ∆conv. time / h rel. conv. ∆conv. time / h rel. conv. ∆conv. 0.11667 0.75 0.0375 2 0.5 0.05 1 0.375 0.04688 0.13333 0.76744 0.03569 5 0.62617 0.04682 4 0.45205 0.04954 0.16667 0.78723 0.0335 30 0.76744 0.03569 24 0.54545 0.04959 0.2 0.79592 0.03249 48 0.81308 0.0304 55 0.63784 0.04767 0.21667 0.81025 0.03075 72 0.82759 0.0254 80 0.69231 0.0426 0.25 0.81818 0.02975 96 0.85075 0.02854 120 0.74684 0.03781 0.26667 0.82456 0.02893 120 0.86667 0.02311 0.38333 0.85507 0.02478 0.41667 0.86207 0.02378 0.55 0.88372 0.02055 0.58333 0.8913 0.01938 0.71667 0.90244 0.01761 0.75 0.90476 0.01723 0.86667 0.91561 0.01545 0.88333 0.91736 0.01516 1.11667 0.93127 0.0128 1.3 0.94792 0.00987 1.63333 0.95122 0.00928 1.78333 0.95918 0.00783 2.11667 0.9661 0.00655 3.11667 0.97403 0.00506 4 0.985 0.00351

119 Supporting Information for Chapter 3

Table 8.6: Conversion and error propagation calculation for the reaction of TP and MBM, respectively, with 3PFB in DMSO employing 0.10 equivalents of TBAF.

thiophenol 4-Methoxybenzyl mercaptan time / h rel. conv. ∆conv. time / h rel. conv. ∆conv. 0.08333 0.91011 0.01636 3 0.83333 0.02778 0.1 0.93485 0.01218 6 0.90148 0.01776 0.13333 0.9537 0.00883 24 0.95833 0.00799 0.15 0.96383 0.00697 48 1 0 0.18333 0.97203 0.00544 0.21667 0.97826 0.00425 0.23333 0.985 0.00296 0.3 1 0

Additional Experiments

1.0

75 mM, 25 °C, glass vial

15 mM, 25 °C, glass vial

150 mM, 25 °C, glass vial

0.8

75 mM, 50 °C, glass vial

75 mM, 25 °C, plastic vial

0.6

0.4

0.2 rel. functional group conversion

0.0

Figure 8.5: Inﬂuence of lower (blue) as well as higher concentration (green), temperature (magenta), and reaction vial (orange) on the conversion. Samples were taken after 40 h of reaction time. Neither of the three parameters showed signiﬁcant change in the conversion and minor changes are well within error margins.

120 Appendix

8.3 Supporting Information for Chapter 4

Supporting Information

Chemiluminescent Read-Out of para-Fluoro – Thiol Reaction Events

Fabian R. Bloesser,‡[a], Federica Cavalli,‡[b], Leonie Barner*[a] and Chrisopher Barner- Kowollik,*[a]

121 Supporting Information for Chapter 4

Materials

Trihydroxybenzene (Sigma-Aldrich, 99.0%), 2,3,4,5,6-Pentafluorobenzyl bromide (Acros Organics, 95%), caesium carbonate (Acros Organics, 95%), tetrabutylammonium hydroxide (1.0 M in methanol, Sigma-Aldrich), dodecanethiol(Sigma-Aldrich, 98.0%), benzyl mercaptan (Sigma-Aldrich, 99.0%), Poly(ethylene glycol) methyl ether thiol (Mn 2,000, Sigma-Aldrich), 1,4-Benzenedimethanethiol (Sigma-Aldrich, 98%), tetrabutylammonium fluoride (1.0 M in THF, Sigma-Aldrich), 2-Chloro-3-hydroxybenzaldehyde (Sigma-Aldrich, 97%), p-Toluenesulfonic acid monohydrate (Merck, 99%), trimethyl or- thoformate (Combi-Blocks, 98%), triethylamine (Ajax, 95%), imidazole (Sigma-Aldrich, 99.5%), tert-butyldimethylsilyl chloride (Combi-Blocks, 98%), trimethyl phosphite (Alfa Aesar, 97%), titan(IV) chloride (Sigma-Aldrich, 99.9%), sodium bicarbonate (Sigma- Aldrich, 99.5%), magnesium sulphate (Merck, 98%), lithium diisopropylamide solution (2.0 M in THF/heptane/ethylbenzene, Sigma-Aldrich) 2-adamantanone (Sigma-Aldrich, 99%), methylene blue (Alfa Aesar, 95%), compressed oxygen (Supagas, 99.5%) were used as received. N,N-dimethylformamide (DMF) (Ajax), ACN (RCI labscan), THF (Ajax), dichloromethane (DCM) (Fisher Scientific), methanol (Ajax), diethyl ether (Ajax), cyclohexane (CH) (Ajax)

and ethyl acetate (EA) (Ajax), were used as solvents. Deuterated solvent such as CDCl3 (99.8%), and ACN-d3 (99.8%) were purchased from Novachem and used as received.

II. Characterisation Methods

Nuclear Magnetic Resonance Spectroscopy

NMR spectra were recorded on a Bruker System 600 Ascend LH, equipped with a BBO- Probe (5 mm) with z-gradient (1H: 600.13 MHz, 13C: 150.90 MHz, 19F: 564.63 MHz, respectively). Chemical shifts are expressed in parts per million (ppm) relative to tetramethylsilane (TMS) and referenced to characteristic residual 1H solvent resonances 19 as internal standards [CDCl3: 7.26 ppm; ACN-d3: 1.94 ppm; THF-d8: 1.72 ppm]. F spectra were referenced via the according Ξvalues (19F: Ξ= 94.094) based on the corresponding 1H NMR spectrum. 1H and 19F NMR spectra are reported as follows: chemical shift (δin ppm), multiplicity (s for singlet, d for doublet, t for triplet, q for quartet, p for pentet, m for multiplet,), coupling constant(s) (Hz), number of protons (concluded from the integrals), specific assignment. 19F NMR spectra were subjected to baseline correction via a multipoint fit function. 13C-{1H} NMR spectra are reported in terms of chemical shift and specific assignment.

122 Appendix

Size Exclusion Chromatography

The SEC measurements were conducted on a PSS SECurity2 system consisting of a PSS SECurity Degasser, PSS SECurity TCC6000 Column Oven (35 °C), PSS SDV Column Set (8× 150 mm 5 µm Precolumn, 8× 300 mm 5 µm Analytical Columns, 100000 Å, 1000 Åand 100 Å) and an Agilent 1260 Infinity Isocratic Pump, Agilent 1260 Infinity Standard Autosampler, Agilent 1260 Infinity Diode Array and Multiple Wavelength Detector (A: 254 nm, B: 360 nm), Agilent 1260 Infinity Refractive Index Detector (35 °C). HPLC grade THF, stabilized with BHT, is used as eluent at a flow rate of 1 mL· min-1. Narrow disperse linear poly(styrene) (number-average molecular weight (Mn): 266 g· -1 6 -1 -1 mol to 2.52× 10 g· mol ) and poly(methyl methacrylate) (Mn: 202 g· mol to 2.2× 106 g· mol-1) standards (PSS ReadyCal) were used as calibrants. All samples were passed over 0.22 µm PTFE membrane filters. Molecular weight and dispersity analysis was performed in PSS WinGPC UniChrom software (version 8.2).

Liquid Chromatrography – Mass Spectrometry

LC-MS measurements were performed on an UltiMate 3000 UHPLC System (Dionex, Sunnyvale, CA, USA) consisting of a pump (LPG 3400SZ), autosampler (WPS 3000TSL) and a temperature-controlled column compartment (TCC 3000). Separation was performed on a C18 HPLC column (Phenomenex Luna 5µm, 100 Å, 250×2.0 mm) operating at 40 °C. Water (containing 5 mmol L-1 ammonium acetate) and acetonitrile were used as eluents. A gradient of acetonitrile:H2O 5:95 to 100:0 (v/v) in 7 min at a flow rate of 0.40 mL· min-1 was applied. The flow was split in a 9:1 ratio, where 90% of the eluent was directed through a DAD UV-detector (VWD 3400, Dionex) and 10% was infused into the electrospray source. Spectra were recorded on an LTQ Orbitrap Elite mass spectrometer (Thermo Fisher Scientific, San Jose, CA, USA) equipped with a HESI II probe. The instrument was calibrated in the m/z range 74-1822 using premixed calibration solutions (Thermo Scientific). A constant spray voltage of 3.5 kV, a dimensionless sheath gas and a dimensionless auxiliary gas flow rate of 5 and 2 were applied, respectively. The capillary temperature and was set to 300 °C, the S-lens RF level was set to 68, and the aux gas heater temperature was set to 100 °C.

Chemiluminescence Kinetics

Emission intensities of chemiluminescence was investigated using a Tecan Spark mul- timode microplate reader. CL measurements were performed using an OptiPlate-96

123 Supporting Information for Chapter 4

Black Opaque microplate (Polystyrene, PerkinElmer). The investigated fluoride concentration of the solutions ranged from 1 to 10 mM. The photon count was measured in luminescence mode in 30 s intervals for 15 h in the range of 400 to 650 nm using an integration time of 10 ms (software Tecan SparkControl). At the beginning of each interval, the reader plate was shaken mechanically in a double-orbital to ensure sufficient mixing of the solution. By adding 50 µL of fluoride solution into each well of the reading plate, down to 5 nmol of fluoride could be detected per well, when using the settings as described above. The procedure is as follows: 50 µL of a fluoride containing sample (1-10 mM, either TBAF or PFTR crude mixtures) were added to a well of the reader plate. Subsequently, 50 µL of a 12 mM dioxetane solution in ACN was added to each well, the well plate was covered with an appropriate cover slide and the measurement was started immediately. The limit of detection (DL) under the present instrument settings was determined from the below formula as per user’s manual:

n ∗ 3 ∗ SD DL = fluoride blank ≈ 2 ∗ 10–5mol ≡ 20µmol (8.3) meanfluoride – meanblank

With nfluoride being the moles of fluoride per well, nfluoride the average CL intensity in cts for samples of the same nfluoride, meanblank the average intensity of blank wells in cts and SDblank the standard deviation of the blanks in cts. It should be noted though that the DL can be further decreased by increasing the integration time from 10 ms to up to 1000 ms. Thus, concentrations of down to 5· 10-9 mol (5 nmol) can be detected and distinguished from blanks.

124 Appendix

Photoreactor

The samples were irradiated in a Luzchem LZC-4V photoreactor using LZC-VIS lamps, emitting in the complete visible range (see spectrum below). Ten lamps were installed for side and top irradiation. Homogeneous irradiation was achieved by stirring the sample solutions during irradiation. The internal chamber was ventilated to maintain ambient temperature during the entire experiment.

1.0

0.8 / a.u.

0.6

0.4 normalised intensity

0.2

0.0

200 300 400 500 600 700 800

/ nm

Figure 8.6: Emission spectrum of the LZC-Vis lamps.

III. Syntheses

1,3,5-Tris(2,3,4,5,6-pentaﬂuorobenzyl) benzene (3PFB)

In a pre-dried Schlenk ﬂask, Cs2CO3 (12.9 g, 39.7 mmol, 5.00 equiv.) was dispersed in dry-DMF (50.0 mL) under argon atmosphere before adding trihydroxybenzene (1.00 g, 7.93 mmol, 1.00 equiv.). After stirring for 30 min, 2,3,4,5,6-pentaﬂuorobenzyl bromide (3.59 mL, 6.21 g, 23.8 mmol, 3.00 equiv.) was added and the reaction mixture was stirred

for two days at 40 °C. Subsequently, Cs2CO3 was ﬁltered off and the ﬁltrate diluted with water (1× 20 mL) and extracted with DCM (2× 20 mL). The collected organic phases were washed with brine (1× 40 mL), dried over magnesium sulphate and concentrated under reduced pressure. The crude product was recrystallised from acetonitrile to give

125 Supporting Information for Chapter 4

the desired product (1.81 g, 34%). 1 H NMR (CDCl3): δ = 6.26 (s, 1H, CHAr), 5.08 (s, 6H, CH2-O) ppm. 19 F NMR (CDCl3): δ = -142.26 (dd, 22.2 Hz, 8.6 Hz, 6F, ortho), -152.14 (t, 20.6 Hz, 3F, para), -161.51 (dt, 21.8 Hz, 8.3 Hz, 6F, meta) ppm. 13 C NMR (CDCl3): δ = 160.06 (CqO), 145.90 (CF), 142.01 (CF), 137.73 (CF), 109.92 (Carom,q), 95.60 (CHAr), 57.64 (CH2-O) ppm. + HR-ESI-MS: m/z = 666.0377 (M+Na , calculated: 666.0312, ∆abs = 0.0065, ∆rel = 9.76 ppm).

1 Figure 8.7: H NMR spectrum (CDCl3, 600 MHz) of 3PFB.

126 Appendix

13 Figure 8.8: C NMR spectrum (CDCl3, 150 MHz) of 3PFB.

19 Figure 8.9: F NMR spectrum (CDCl3, 564 MHz) of 3PFB.

127 Supporting Information for Chapter 4

PFTR

The PFTR linker (111 mg, 167 µmol, 1.00 equiv.) was dissolved in ACN (25 mL), purged with Argon for 10 min and split into ﬁve solutions of 5 mL each. Separately, solutions of thiol (10.0 µmol, 0.30 equiv. to 100 µmol, 3.00 equiv.) and TBAOH (10.0 µL, 10.0 µmol, 0.30 equiv. to, 100 µL, 100 µmol, 3.00 equiv.) in ACN (5 mL) were prepared. The thiolate solutions were subsequently added to the linker solutions and the mixtures were stirred at 50 °C for 12 h. The reactions were carried out in ACN as it is an excellent solvent for conducting PFTR (i.e. polar and aprotic), and the CL reaction of the CL probe in ACN with TBAF is a well-established system with reported quantum yields. 19 F NMR (CDCl3): δ = -132.23 (m, 6F, ortho), -144.95 (m, 6F, meta) ppm.

Chlorinated Schaap’s dioxetane

Schaap’s dioxetane bearing a chlorine substitution in ortho-position has been synthesised according to literature.212

1 Figure 8.10: H NMR spectrum (CDCl3, 600 MHz) of the CL probe ’Schaap’s dioxetane’.

128 Appendix

1.0

0.8 / a.u.

0.6

0.4 normalised intensity

0.2

0.0

300 400 500 600 700 800

/ nm

Figure 8.11: CL mission spectrum of the CL probe in acetonitrile. The yellow box highlights the wavelength regime that was recorded and integrated.

129 Supporting Information for Chapter 4

IV. Time-dependant Emission

8.0M

1 mM

2 mM

3 mM

6.0M 4 mM

5 mM -1

6 mM

7 mM / cts s

8 mM

4.0M

9 mM

10 mM Intensity

2.0M

0.0

0 120 240 360 480 600 720 840

t / min

Figure 8.12: Time-dependant CL intensity of TBAF at various concentrations over the course of 15 h.

130 Appendix

10.0M

1 mM

3 mM

5 mM

8.0M

7 mM

9 mM -1

6.0M / cts s

4.0M Intensity

2.0M

0.0

0 120 240 360 480 600 720 840

t / min

Figure 8.13: Time-dependant CL intensity after PFTR of adamantyl thiol (AT) at various concentrations over the course of 15 h.

1 mM

8.0M

3 mM

5 mM

7 mM

9 mM -1

6.0M / cts s

4.0M Intensity

2.0M

0.0

0 120 240 360 480 600 720 840

t / min

Figure 8.14: Time-dependant CL intensity after PFTR of methoxy triethylene glycol thiol (mTrEGT) at various concentrations over the course of 15 h.

131 Supporting Information for Chapter 4

8.0M

1 mM

2 mM

3 mM

6.0M 4 mM

5 mM -1

7 mM

8 mM / cts s

9 mM

4.0M

10 mM Intensity

2.0M

0.0

0 120 240 360 480 600 720 840

t / min

Figure 8.15: Time-dependant CL intensity after PFTR of benzyl thiol (BT) at various concentrations over the course of 15 h.

10.0M

1 mM

3 mM

5 mM

8.0M

7 mM

9 mM -1

6.0M / cts s

4.0M Intensity

2.0M

0.0

0 120 240 360 480 600 720 840

t / min

Figure 8.16: Time-dependant CL intensity after PFTR of PEG-SH at various concentrations over the course of 15 h.

132 Appendix

8.0M

1 mM

3 mM

5 mM

6.0M 7 mM

9 mM -1 / cts s

4.0M Intensity

2.0M

0.0

0 120 240 360 480 600 720 840

t / min

Figure 8.17: Time-dependant CL intensity after PFTR of 1,4-benzendimethane thiol (BDT) at various concentrations over the course of 15 h.

10.0M

1 mM

3 mM

5 mM

8.0M

7 mM

9 mM -1

6.0M / cts s

4.0M Intensity

2.0M

0.0

0 120 240 360 480 600 720 840

t / min

Figure 8.18: Time-dependant CL intensity after PFTR of triethylene glycol dithiol (DODT) at various concentrations over the course of 15 h.

133 Supporting Information for Chapter 4

V. Anion Testing

13G

12.0G

10.0G / cts

8.0G

12G

6.0G

4.0G total emission

2.0G

600.0M / cts /

0.0

iodide

fluoride

chloridecyanate bromide 400.0M

acetonitrile

perchlorate

TBAF blank

1 mM fluoride

dioxetane blank hydrogensulfate

200.0M total emission total

iodide

fluoride

chloride cyanate bromide

acetonitrile

perchlorate

TBAF blank

1 mM fluoride

dioxetane blank hydrogensulfate

Figure 8.19: Total integrated CL emission of 10 mM solutions TBA bromide, chloride, cyanate, hydrogensulfate, iodide, perchlorate and fluoride as well as a 1 mM solution of TBAF. After 2 h, cyanate shows only 2.3% of the integrated CL intensity compared to equimolar amounts of fluoride and only 57% of integrated CL compared to 0.1 equivalents of fluoride. Integrated CL intensity for the other investigated anions lies between 16 and 70 ppm.

The reason for the good selectivity of fluoride compared to other anions is the high affinity of fluoride towards silicon (binding energy of Si-F is 595 kJ mol-1, compared to Si-Br 329, Si-Cl 398, Si-I 234 and Si-O 444 kJ mol-1) as well as its smaller radius of 133 pm compared to cyanate 159 pm, chloride 181 pm, bromide 196 pm, iodide 220 pm, perchlorate 240 pm and hydrogensulfate 258 pm.

134 Appendix

VI. Calculation of Conversion

Conversion of PFTR was calculated from 19F NMR by comparing the ratios of parent ortho-(o), meta-(m) as well as para-ﬂuoro (p) resonances with the ortho’(o’)- and meta’-ﬂuoro (m’) resonances of the PFTR product, as shown in Figure 7.20, according to eq. (7.4).

19 Figure 8.20: F NMR spectrum (ACN-d3, 564 MHz) of 3PFB after approx. 24% conversion in a PFTR.

m0 0.48 conversion(c) = = ≈ 0.242 ≡ 24.2% (8.4) m0 + m 0.48 + 1.50 As NMR commonly exhibits errors of up to 5%, error propagation was performed according to eq. (7.5), where ∂/∂c is the derivation of the conversion (refer to eq. (7.4)) and ∆is the error of the according integral:

∂m0 ∂m 0 Error = ∗ ∆m + ∗ ∆m (8.5) ∂c ∂c

135 Supporting Information for Chapter 4

19 Figure 8.21: F NMR spectrum (ACN-d3, 564 MHz) of 3PFB after approx. 24% conversion in a PFTR and 22% hydroxy substitution.

m0 0.90 conversion(c) = 0 0 = ≈ 0.226 ≡ 22.6% (8.6) m + m + mhydroxy 0.90 + 2.22 + 0.86

0 0 ∂m0 ∂m ∂m 0 hydroxy 0 Error = ∗ ∆m + ∗ ∆m + ∗ ∆mhydroxy (8.7) ∂c ∂c ∂c Conversion of PFTR was furthermore calculated from LC by comparing the weighted ratios of unsubstituted as well as mono-, bis- and trisubstituted linkers, as shown in Figure 7.22, according to eq. (7.8). The conversion of PEG-SH was calculated from SEC in the same manner.

136 Appendix

150.0k

100.0k

50.0k / a.u.

mono

bis Area=19086.23

Area=20577.15

0.0 intensity

3PFB tri

-50.0k

Area=8473.973 Area=8026.002

-100.0k

8.4 8.6 8.8 9.0 9.2

retention time / min

Figure 8.22: LC chromatogram of 3PFB after approx. 50% conversion in a PFTR.

A + 2 ∗ A + 3 ∗ A 84318.5 conversion(c) = mono bis tri = ≈ 0.500 ≡ 50.0% (8.8) 3 ∗ (A3PFB + Amono + Abis + Atri) 168490

Error calculations were performed based on the difference between the deconvoluted LC trace and the original LC trace (∆A), according to:

∂A 0 ∂A ∂A ∂A 3PFB mono bis tri Error = ∗ ∆A3PFB + ∗ ∆Amono + ∗ ∆Abis + ∗ ∆Atri ∂c ∂c ∂c ∂c (8.9)

137 Supporting Information for Chapter 4

VII. Linear Fits of Emission vs Conversion

n / µmol

0 10 20 30 40 50 60 70 80 90 100 110

80G

TBAF

AT, theo

70G

AT, NMR

AT, LC

60G / cts /

50G

40G

30G total emission emission total

20G

10G

0 1 2 3 4 5 6 7 8 9 10 11

conc [F ] / mM

Figure 8.23: CL emission vs ﬂuoride concentration for AT according to expected yields (red dots), 19F NMR conversion (blue diamonds), and LC (purple triangles) compared to the TBAF reference (black squares).

138 Appendix

n / µmol

0 10 20 30 40 50 60 70 80 90 100 110

80G

TBAF

70G

mTrEGT, theo

mTrEGT, NMR

60G

mTrEGT, LC / cts /

50G

40G

30G total emission emission total

20G

10G

0 1 2 3 4 5 6 7 8 9 10 11

conc [F ] / mM

Figure 8.24: CL emission vs ﬂuoride concentration for mTrEGT according to expected yields (red dots), 19F NMR conversion (blue diamonds), and LC (purple triangles) compared to the TBAF reference (black squares).

n / µmol

0 10 20 30 40 50 60 70 80 90 100 110

80G

TBAF

70G

BT, theo

BT, NMR

60G BT, LC

50G / cts /

40G

30G total emission emission total

20G

10G

0 1 2 3 4 5 6 7 8 9 10 11

conc [F ] / mM

Figure 8.25: CL emission vs ﬂuoride concentration for BT according to expected yields (red dots), 19F NMR conversion (blue diamonds), and LC (purple triangles) compared to the TBAF reference (black squares).

139 Supporting Information for Chapter 4

n / µmol

0 10 20 30 40 50 60 70 80 90 100 110 120 130

80G

TBAF

PEG, theo

70G

PEG, NMR

PEG, SEC

60G

PEG, back calculation / cts /

50G

40G

30G total emission emission total

20G

10G

0 1 2 3 4 5 6 7 8 9 10 11 12 13

conc [F ] / mM

Figure 8.26: CL emission vs ﬂuoride concentration for PEG-SH according to expected yields (red dots), 19F NMR conversion (blue diamonds), and LC (purple triangles) compared to the TBAF reference (black squares). The golden stars depict the concentration of ﬂuoride as calculated from the CL emission.

140 Appendix

VIII. Supplementary Network Data

The PFTR networks described in the manuscript were subjected to 19F NMR and SEC analysis in order to determine if any soluble fraction was present.

19 Figure 8.27: F NMR spectrum (ACN-d3, 564 MHz) of the supernatant solutions after network formation employing BDT. Traces 3 to 5 do not show any signiﬁcant amount of PFB resonances.

141 Supporting Information for Chapter 4

1.0

0.8 / a.u.

0.6

0.4

0.2 normalised intensity

0.0

sample number

0 5 10 15 20 25 30 35 40

retention volume / mL

Figure 8.28: SEC spectra of the supernatant solutions after network formation employing BDT. Toluene was added to the SEC samples (1 mg mL-1, at35.35 mL retention volume) in order to provide a reference. Traces 3 to 5 do not show any signiﬁcant amount of soluble fractions.

19 Figure 8.29: F NMR spectrum (ACN-d3, 564 MHz) of the supernatant solutions after network formation employing DODT. Only trace 5 does not show any signiﬁcant amount of PFB resonances.

142 Appendix

1.0

0.8 / a.u.

0.6

0.4

0.2 normalised intensity

0.0

sample number

0 5 10 15 20 25 30 35 40

retention volume / mL

Figure 8.30: SEC spectra of the supernatant solutions after network formation employing DODT. Toluene was added to the SEC samples (1 mg mL-1, at35.35 mL retention volume) in order to provide a reference. Only trace 5 does not show any signiﬁcant amount of soluble fractions.

143 Supporting Information for Chapter 4

IX. Hydroxy-Substitution of p-Fluorine

In some cases, the employed base TBAOH partially undergoes nucleophilic substitution with the 3PFB linker rather than deprotonate the thiol. In these cases, additional resonances at around -146.3 and -159.5 ppm appear in the 19F NMR spectrum (refer Figure 7.21). While the hydroxy-substitution also releases one equivalent of fluoride per reaction, the fluorine will subsequently deprotonate a thiol and form HF. Due to similar pKa values of the thiol and HF, the thiol and the fluoride are in an equilibrium with HF and the thiolate. However, as the thiolate is subsequently reacted with the 3PFB linker, the equilibrium is eventually completely shifted to HF and the thiolate. On the one hand, the deprotonation of thiols via fluoride ensures that full conversion of thiols can still be achieved, on the other hand the formed HF does not trigger significant amounts of CL und thus, does not falsify the PFTR read-out. Refer to Scheme S1 for more information.

F F F F

F F F F F OH

OH F HF

R SH R S F

F F F F

F F F F F S R

Scheme 8.1: A hydroxide reacts with the 3PFB linker and releases a fluoride. The pKa of HF-fluoride acid-base pair is low enough to allow deprotonation of a thiol, which can subsequenly undergo PFTR and release another fluoride ion. Contrary to a free fluoride, the HF formed upon deprotonation of the thiol will not trigger significant chemiluminescent. Thus, the CL read-out obtained only refers to thiol substitution, but not hydroxy substitution.

144 Appendix

8.4 Supporting Information for Chapter 5

Supporting Information

Chemiluminescent Read-Out of para-Fluoro – Thiol Reaction Events

Fabian R. Bloesser,[a], Sarah L. Walden[a], Ishrath M. Irshadeen[a], Lewis C. Chambers[a] and Chrisopher Barner-Kowollik,*[a]

145 Supporting Information for Chapter 5

Materials

3,5-Dimethylphenol (Sigma-Aldrich, 98%), caesium carbonate (Acros Organics, 95%), 11-bromo-1-undecanol (Combi-Blocks, 98%), methacryloyl chloride (Sigma-Aldrich, 97%), 1-aminopyrene (Sigma-Aldrich, 97%), 4-Amino-2,6-dichlorophenol (Combi-Blocks, 98%), maleic anhydride (Sigma-Aldrich, 99%), triethylamine (Ajax, 95%), N,N’-dicyclo- hexylcarbodiimide (Sigma-Aldrich, 99%), oxalyl chloride (Sigma-Aldrich, 98%), methyl methacrylate (Sigma-Aldrich, 99%), 2-Cyano-2-propyl benzodithioate (Sigma-Aldrich, 97%), hydrogen peroxide (Ajax, 30w% in water), sodium bicarbonate (Sigma-Aldrich, 99.5%), magnesium sulphate (Merck, 98%), and magnesium chloride (anhydrous, Sigma- Aldrich, 98%) were used as received. azobisisobutyronitrile (AIBN) (Sigma-Aldrich, 12w% in acetone) was recrystallised from methanol. DMF (Ajax), acetone (Ajax), ACN (RCI labscan), THF (Ajax), DCM (Fisher Scientiﬁc), toluene (Fisher Scientiﬁc), diethyl ether (Ajax), methanol (Ajax), CH (Ajax) and EA (Ajax), were used as solvents.

Deuterated solvent such as chloroform-d (CDCl3 , 99.8%), and DCM-d2 (CD2Cl2, 99.8%), were purchased from Novachem and used as received.

II. Characterisation Methods

Nuclear Magnetic Resonance Spectroscopy

Diffusion Ordered NMR Spectroscopy

1 DOSY experiments based on H NMR were performed in DCM-d2 at 296 K on a Bruker 400 UltraShield spectrometer equipped with a Quattro Nucleus Probe (QNP) with an

146 Appendix

operating frequency of 400 MHz (1H). A sequence with longitudinal eddy current delay (LED) using bipolar gradients was employed in order to compensate eddy currents. A bipolar gradient δ = 5 ms and a diffusion delay ∆ = 100 ms were used. Gradient strength was linearly incremented from 2% at 0.96 G to 95% at 45.7 G in 64 steps. The obtained data was processed with TopSpin 4.0.6. After Fourier transform of the 1D spectra, the signal decay along the gradients G was ﬁtted to

–D·G2·γ2·δ2·∆– δ 4 f (G) = I0 · e 3 · 10 (8.10) with the gyromagnetic ratio γ and the full signal intensity I0. hydrodynamic radii (rH’s) were calculated from the Stokes-Einstein equation:

k T r = B (8.11) H 6πηD

Where kB is the Boltzmann constant, T the temperature and η the solvent viscosity (DCM: 0.413 mPa·s).297

Size Exclusion Chromatography

The SEC measurements were conducted on a Waters Breeze QS HPLC system consisting of a Waters 1515 Isocratic HPLC Pump, a Waters 1500 Series Column Heater (35 °C), PSS SDV Column Set (8×150 mm 5 µm Precolumn, 8×300 mm 5 µm Analytical Columns, 100000 Å, 1000 Åand 100 Å), a Waters 2707 Autosampler, a Waters 2414 Re- fractive Index (RI) Detector (35 °C), and a Waters 2489 UV/Visible Detector (Wavelength A: 254 nm, Wavelength B: 360 nm). Analytical grade chloroform, stabilized with amylene, -1 is used as eluent at a ﬂow rate of 1 mL·min . Narrow disperse linear poly(styrene) (Mn: -1 6 -1 -1 266 g·mol to 2.52×10 g·mol ) and poly(methyl methacrylate) (Mn: 202 g·mol to 2.2×106 g·mol-1) standards (PSS ReadyCal) were used as calibrants. All samples were passed over 0.22 µm PTFE membrane ﬁlters. Molecular weight and dispersity analysis was performed in the Waters Breeze 2 software.

THF-SEC ‘triple detection’

The SEC measurements were conducted on a PSS SECurity2 system consisting of a PSS SECurity Degasser, PSS SECurity TCC6000 Column Oven (35 °C), PSS SDV Col- umn Set (8×150 mm 5 µm Precolumn, 8×300 mm 5 µm Analytical Columns, 100000 Å, 1000 Åand 100 Å) and an Agilent 1260 Inﬁnity Isocratic Pump, Agilent 1260 Inﬁnity

147 Supporting Information for Chapter 5

Standard Autosampler, Agilent 1260 Infinity Diode Array and Multiple Wavelength De- tector (A: 254 nm, B: 360 nm), Agilent 1260 Infinity Refractive Index Detector (35 °C), PSS SLD7100 Multiangle Laser Light Scattering (MALLS) Detector and PSS DVD1260 four capillary viscometer. HPLC grade THF, stabilized with BHT, is used as eluent at a flow rate of 1 mL·min-1. All samples were passed over 0.22 µm PTFE membrane filters. Molecular weight and dispersity analysis was performed in PSS WinGPC UniChrom software (version 8.2). Molecular weights were calculated via an iterative algorithm using light scattering and viscosity signals.

Chemiluminescence Kinetics

Emission intensities of chemiluminescence was investigated using a Tecan Spark mul- timode microplate reader. CL measurements were performed using an OptiPlate-96 Black Opaque microplate (Polystyrene, PerkinElmer). The investigated SCNP solutions exhibited a concentration of 1 mg·mL-1 polymer in DMF. The photon count was measured in luminescence mode in 12 s intervals for 3 h in the range of 360 to 700 nm using an integration time of 1000 ms (software Tecan SparkControl). At the beginning of each interval, the reader plate was shaken mechanically for three seconds at 270 rpm in a double-orbital to ensure sufﬁcient mixing of the solution. By adding 6 µL of a 30%

H2O2 solution into each well of the reading plate, down to 5 nmol of ﬂuorophore could be detected per well, when using the settings as described above. The procedure is as follows: 100 textmu L of 1 mg·mL-1 SCNP in DMF added to a well of the reader

plate. Subsequently, 6µL of a 30% H2O2 solution was added to each well, the well plate was covered with an appropriate cover slide and the measurement was started immediately.

Photoreactor

The samples were irradiated in a Luzchem LZC-4V photoreactor using LZC-UVA lamps, centred at ≈350 nm (see spectrum below). six lamps were installed for side irradiation. Homogeneous irradiation was achieved by stirring the sample solutions during irradiation. The internal chamber was ventilated to maintain ambient temperature during the entire experiment.

148 Appendix

1.0

0.8 / a.u.

0.6

0.4 normalised intensity

0.2

0.0

200 300 400 500 600 700 800

/ nm

Figure 8.31: Emission spectrum of the LZC-UVA lamps.

Interchim XS420 + SofTA Model 400 ELSD

Flash chromatography was performed on an Interchim XS420+ flash chromatography system consisting of a SP-in-line filter 20 µm, a UV-VIS detector (200-800 nm) and a SofTA Model 400 ELSD (55 °C drift tube temperature, 25 °C spray chamber temperature, filter 5, EDR gain mode) connected via a flow splitter (Interchim Split ELSD F04590). The separations were performed using an Interchim dry load column and an Interchim Puriflash Silica HP 30 µm column.

III. Syntheses

a. Synthesis of the photo-enol-monomer (1)

NEt3 Cl MgCl2 Cs2CO3 OH OH O HO O n Br 9 OH NEt3 O O O O ACN DMF DCM O 9 O 90 °C, 4.5 h a.t., 16 h O 9 OH a.t., 1 h 2 3 4 1

149 Supporting Information for Chapter 5

2-hydroxy-4,6-dimethylbenzaldehyde (3)

3 was synthesised according to a literature procedure.298 1 H NMR (CDCl3): δ = 11.93 (s, 1H, OH), 10.21 (s, 1H, CHO), 6.60 (s, 1H, CHAr), 6.51 (s, 1H, CHAr), 2.53 (s, 3H, CH3), 2.29 (s, 3H, CH3) ppm. 13 C NMR (CDCl3): δ = 195.33, 163.15, 142.15, 141.87, 137.38, 118.52, 116.04, 18.31, 18.03 ppm.

2-((11-hydroxyundecyl)oxy)-4,6-dimethylbenzaldehyde (4)

3 (3.00 g, 20.0 mmol, 1.00 equiv.) and caesium carbonate (7.71 g, 40.0 mmol, 2.00 equiv.) were dissolved in DMF (200 mL). 1-Bromoundecanol (5.27 g, 21.0 mmol, 1.05 equiv.) was added and the mixture was allowed to stir at ambient temperature overnight. Subsequently, caesium salts were removed by filtration, the filtrate was diluted with DCM and washed with water. Finally, the solvent was removed under reduced pressure and the crude product was subjected to flash chromatography on silica using n-hexane as the solvent. After removal of the solvent, the pure product 3 was obtained as a colourless oil (6.17 g, 96%). 1 H NMR (CDCl3): δ = 10.60 (s, 1H, CHO), 6.62 (s, 1H, CHAr), 6.59 (s, 1H, CHAr), 4.01 (t, 6.3 Hz, 2H, CH2-O), 3.63 (t, 6.6 Hz, 2H, CH2-OH), 2.53 (s, 3H, CH3), 2.32 (s, 3H, CH3), 1.80 (m, 2H, CH2), 1.56 (m, 2H, CH2), 1.45 (m, 2H, CH2), 1.30 (m, 12H, CH2) ppm. 13 C NMR (CDCl3): δ = 192.04, 163.13, 145.74, 141.89, 124.81, 120.99, 110.64, 68.60, 62.89, 32.82, 29.57, 29.51, 29.47, 29.44, 29.30, 29.18, 26.11, 25.78, 22.16, 21.59 ppm. + HR-ESI-MS: m/z = 321.2430 (M+H , calculated: 321.2424, ∆abs = 0.0006, ∆rel = 1.87 ppm).

11-(2-formyl-3,5-dimethylphenoxy)undecyl methacrylate (1)

1 (6.17 g, 19.3 mmol, 1.00 equiv.) and triethylamine (4.03 mL, 2.92 g, 28.9 mmol, 1.50 equiv.) were dissolved in diethyl ether (100 mL). Methacryloyl chloride (2.26 mL, 2.42 g, 23.1 mmol, 1.50 equiv.) in diethyl ether (25 mL) was added and the mixture was stirred at ambient temperature for one hour. Subsequently, the mixture was washed with bicarb solution and dried over MgSO4. The solvent was removed under reduced pressure and the crude product was subjected to ﬂash chromatography on silica using n-hexane as the solvent. After removal of the solvent, the pure product 1 was obtained as a colourless oil (6.79 g, 91%). 1 H NMR (CDCl3): δ = 10.61 (s, 1H, CHO), 6.62 (s, 1H, CHAr), 6.60 (s, 1H, CHAr), 6.09

150 Appendix

(m, 1H, CHcis), 5.54 (m, 1H, CHtrans), 4.13 (t, 6.7 Hz, 2H, CH2-OOC), 4.02 (t, 6.4 Hz, 2H, CH2-O), 2.54 (s, 3H, CH3), 2.33 (s, 3H, CH3). 1.94 (s, 3H, CH3), 1.81 (m, 2H, CH2), 1.66 (m, 2H, CH2), 1.47 (m, 2H, CH2), 1.32 (m, 12H, CH2) ppm. 13 C NMR (CDCl3): δ = 192.04, 163.13, 145.74, 141.89, 124.81, 120.99, 110.64, 68.60, 62.89, 32.82, 29.57, 29.51, 29.47, 29.44, 29.30, 29.18, 26.11, 25.78, 22.16, 21.59 ppm. + HR-ESI-MS: m/z = 411.2501 (M+Na , calculated: 411.2506, ∆abs = 0.0005, ∆rel = 1.22 ppm).

b. Synthesis of 1-pyrenyl methacrylamide (5)

O Et3N H + H2N N Cl DCM a.t. overnight O 5

Aminopyrene (1.00 g, 4.60 mmol, 1.00 equiv.) and triethyl amine (770 µL, 559 mg, 5.52 mmol, 1.20 equiv.) were dissolved in DCM (50 mL) and cooled to 0 °C. Methacryloyl chloride (540 µL, 577 mg, 5.52 mmol, 1.20 equiv.) was added dropwise and the mixture was allowed to stir overnight. The crude mixture was washed with sodium bicarb solution

and dried over MgSO4. The solvent was removed under reduced pressure and the crude product was subjected to ﬂash chromatography on silica using a gradient from 100% n-hexane to 100% ethyl acetate as the eluent. After removal of the solvent, the pure product 5 was obtained as a green solid (1.24 g, 94%). 1 H NMR (CDCl3): δ = 8.47 (d, 8.2 Hz, 1H, CHO), 8.19 (m, 4H, CHAr), 8.11 (d, 1H, CHAr), 8.03 (m, 4H, CHAr), 6.05 (m, 1H, CHcis), 5.61 (m, 1H, CHtrans), 2.23 (s, 3H, CH3) ppm. 13 C NMR (CDCl3): δ = 192.04, 163.13, 145.74, 141.89131.49, 130.92, 130.37, 129.34, 128.21, 127.50, 127.05, 126.33, 125.70, 125.41, 125.33, 125.25, 124.93, 123.68, 122.32, 120.57, 120.15, 19.15 ppm. + HR-ESI-MS: m/z = 308.1057 (M+Na , calculated: 308.1046, ∆abs = 0.0011, ∆rel = 3.57 ppm).

c. Synthesis of the crosslinker (6)

bis(2,6-dichloro-4-N-maleimido) phenyl oxalate (MDCPO) 6 was synthesised according to a literature procedure.3 1 H NMR (CDCl3): δ = 7.63 (s, 4H, CHAr), 6.92 (s, 4H, CH) ppm. 13 C NMR (CDCl3): δ = 169.04, 152.25, 140.84, 135.08, 132.00, 127.37, 127.02 ppm.

151 Supporting Information for Chapter 5

O Cl 1.) acetone Et N Cl N Cl OH 3 Cl O O O OH a.t., 1 h O oxalyl chloride + O O O O O 2.) DCC, DCM N Cl ethyl acetate O Cl H2N Cl reflux, o.n. 0 °C, 1 h N Cl O O 6

+ HR-ESI-MS: m/z = 308.1057 (M+Na , calculated: 308.1046, ∆abs = 0.0011, ∆rel = 3.57 ppm).

152 Appendix

d. Polymerisation

RAFT agent O AIBN O O + + O O toluene O 9 O 80 °C, 16 h N H

9 O O O O HN O S

S N m/n/o

Monomers (45.2 mmol, 1000 equiv. different ratios), AIBN (1.48 mg, 9.04 µmol, 0.20 equiv.) and 2-cyano-2-propyl dithiobenzoate (10.0 mg, 45.2 µmol, 1.00 equiv.) were dissolved in dry toluene (5 mL) and degassed via three consecutive freeze-pump- thaw cycles. The mixture was subsequently immersed into a preheated oil bath at 80 °C and the reaction was allowed to proceed for 16 h. Subsequently, the crude polymers were precipitated ﬁrst from ice-cold pentane, then from ice-cold methanol and ﬁnally dried in vacuo to obtain polymers P1-3. 1 H NMR (CDCl3): δ = 10.63 (s, br, CHO), 8.61-7.98 (m, CHPyr), 6.64 (s, br, CHAr), 6.61 (s, br, CHAr), 4.03 (s, br, CH2-O), 3.95 (s, br, CH2-O), 3.61 (s, br, O-CH3), 2.55 (s, br, CH3), 2.34 (s, br, CH3), 2.11-0.76 (m, aliphatic H) ppm. SEC (PMMA cal.): P1 = 33.7 kg·mol-1, Ð 1.90; P2 = 33.1 kg·mol-1, Ð 1.46; P3 = 28.3 kg·mol-1, Ð 1.59; Triple-detection-SEC (PMMA cal.): P1 = 41.9 kg·mol-1, Ð 1.62; P2 = 39.2 kg·mol-1, Ð 1.63; P1 = 41.3 kg·mol-1, Ð 1.20;

153 Supporting Information for Chapter 5

1 Figure 8.32: H NMR spectrum (CDCl3, 600 MHz) of P1.

154 Appendix

d. SCNP folding

O N Cl O O O + Cl O 9 O Cl O O O O O HN O S O

S Cl N N m/n/o O

Cl O

O N

O Cl O OH O UV-A

DCM 9 a.t., 1 h O O O O HN O S

S N m/n/o

P1-3 (50 mg) and bis-maleimide linker 6 (5.00 equiv. per o-MBA moiety in the backbone) were dissolved in dry DCM (500 mL). After purging with argon for 15 min, the mixture was irradiated with UV-A light (centred at 350 nm) for 60 min at ambient temperature. Subsequently, the solvent was removed under reduced pressure and the polymer separated from the linker via a sephadex® LH20 column using THF as the eluent. 1 H NMR (CDCl3): δ = 8.35-7.89 (m, CHPyr), 7.59 (s, br, CHAr,linker), 6.72-6.54 (m, CHAr), 5.89 (s, br, OH), 4.01-3.84 (m, CH2-O), 3.56 (s, br, O-CH3), ), 3.42-3.03 (m, CHindol), 2.32 (s, br, CH3), 2.05-0.74 (m, aliphatic H) ppm. SEC (PMMA cal.): P1 = 22.9 kg·mol-1, Ð 1.65; P2 = 22.1 kg·mol-1, Ð 1.1.64; P1 = 22.5 kg·mol-1, Ð 1.53;

155 Supporting Information for Chapter 5

1 Figure 8.33: H NMR spectrum (CD2Cl2, 600 MHz) of SCNP1.

156 Appendix

d. SCNP unfolding

SCNP1-3 (10 mg, 201-355 nmol crosslinker) were dissolved in dry DMF (1 mL). An

aliquot of the solution (100 µL) was added to the well reader plate and 30% H2O2 (6 µL, 58.8 µmol) was added. The reaction was allowed to proceed at ambient temperature for

1 h, then a second portion of 30% H2O2 (6 µL, 58.8 µmol) was added and allowed to react to ensure complete unfolding. 1 H NMR (CDCl3): δ = 8.43-7.95 (m, CHPyr), 7.28 (s, br, CHAr,linker), 6.67-6.63 (m, CHAr), 5.88 (s, br, OH), 3.92 (s, br, CH2-O), 3.57 (s, br, O-CH3), 3.34-3.03 (m, CHindol), 2.31 (s, br, CH3), 2.05-0.74 (m, aliphatic H) ppm. SEC (PMMA cal.): P1 = 12.6 kg·mol-1, Ð 2.72; P2 = 17.6 kg·mol-1, Ð 1.2.09; P1 = 14.7 kg·mol-1, Ð 2.26;

1 Figure 8.34: H NMR spectrum (CD2Cl2, 600 MHz) of uSCNP1.

157 Supporting Information for Chapter 5

IV. SEC Data

1.0 -1

M = 33.7 kg mol , Ð = 1.90

n / a.u.

SCNP1

-1

M = 22.9 kg mol , Ð = 1.65

0.8

uSCNP1

-1

M = 12.6 kg mol , Ð = 2.72

0.6

0.4

0.2 normalised number intensity

0.0

19 20 21 22 23 24 25 26

retention volume / mL

Figure 8.35: SEC chromatogram of P1 (solid), SCNP1 (dashed) and uSCNP1 (dotted).

158 Appendix

1.0

-1

M = 33.1 kg mol , Ð = 1.46

n / a.u.

SCNP2

-1

M = 22.1 kg mol , Ð = 1.64

0.8

uSCNP2

-1

M = 17.6 kg mol , Ð = 2.09

0.6

0.4

0.2 normalised number intensity

0.0

19 20 21 22 23 24 25 26

retention volume / mL

Figure 8.36: SEC chromatogram of P2 (solid), SCNP2 (dashed) and uSCNP2 (dotted).

1.0

-1

M = 28.3 kg mol , Ð = 1.59

n / a.u.

SCNP3

-1

M = 22.5 kg mol , Ð = 1.53

n 0.8

uSCNP3

-1

M = 14.7 kg mol , Ð = 2.26

0.6

0.4

0.2 normalised number intensity

0.0

19 20 21 22 23 24 25 26

retention volume / mL

Figure 8.37: SEC chromatogram of P3 (solid), SCNP3 (dashed) and uSCNP3 (dotted).

159 Supporting Information for Chapter 5

V. DOSY Data

1.0 P1

SCNP1

uSCNP1 / a.u.

0.8

0.6

0.4

0.2 normalised number intesity

0.0

9.0 9.5 10.0 10.5 11.0 11.5

r / nm

Figure 8.38: 1H DOSY NMR projection in the diffusion dimension of P1 (solid), SCNP1 (dashed) and uSCNP1 (dotted).

160 Appendix

1.0 P2

SCNP2

uSCNP2 / a.u.

0.8

0.6

0.4

0.2 normalised number intesity

0.0

8.5 9.0 9.5 10.0 10.5 11.0

r / nm

Figure 8.39: 1H DOSY NMR projection in the diffusion dimension of P2 (solid), SCNP2 (dashed) and uSCNP2 (dotted).

1.0 P3

SCNP3

uSCNP3 / a.u.

0.8

0.6

0.4

0.2 normalised number intesity

0.0

8.5 9.0 9.5 10.0

r / nm

Figure 8.40: 1H DOSY NMR projection in the diffusion dimension of P3 (solid), SCNP3 (dashed) and uSCNP3 (dotted).

161 Supporting Information for Chapter 5

VI. THF-SEC Data ’Triple-Detection’

Figure 8.41: SEC-multiangle laser light scattering (MALLS) chromatogram of P1.

162 Appendix

Figure 8.42: SEC-MALLS chromatogram of P2.

Figure 8.43: SEC-MALLS chromatogram of P3.

163 Supporting Information for Chapter 5

VII. Time-dependant Emission

2 addition of H O

2 2

1.5M -1

1.0M / cts s Intensity

500.0k

0.0

0 30 60 90 120 150 180

t / min

Figure 8.44: Averaged time-dependant CL intensity of P1 over the course of 180 min

164 Appendix

well1

well2

well3 1.5M

well4

well5 -1

1.0M / cts s Intensity

500.0k

0.0

0 30 60 90 120 150 180

t / min

Figure 8.45: Time-dependant CL intensity of ﬁve identical samples of P1 over the course of 180 min.

1.5M -1

1.0M / cts s Intensity

500.0k

0.0

0 30 60 90 120 150 180

t / min

Figure 8.46: Averaged time-dependant CL intensity of P21 over the course of 180 min

165 Supporting Information for Chapter 5

well1

well2

1.5M

well3

well4

well5 -1

1.0M / cts s Intensity

500.0k

0.0

0 30 60 90 120 150 180

t / min

Figure 8.47: Time-dependant CL intensity of ﬁve identical samples of P2 over the course of 180 min.

1.0M -1 / cts s

500.0k Intensity

0.0

0 30 60 90 120 150 180

t / min

Figure 8.48: Averaged time-dependant CL intensity of P3 over the course of 180 min

166 Appendix

well1

well2

well3

well4

well5 -1 / cts s

500.0k Intensity

0.0

0 30 60 90 120 150 180

t / min

Figure 8.49: Time-dependant CL intensity of ﬁve identical samples of P3 over the course of 180 min.

167 Supporting Information for Chapter 5

VIII. Parameter Estimation

The unknown rate coefficients kHEI and kcat were estimated via the simulation of three separate experimental CL emission profiles. Within a first approximation step, the kinetic scheme given in Scheme 7.2 was fitted to the experimental data given in Figure 7.50, Figure 7.52 and Figure 7.54 with the parameter estimation function of the PREDICI® simulation package (version 11). This feature allows for the simultaneous fit of individually obtained data sets to a common kinetic scheme. This estimation procedure suggests

possible best fit solutions for the rate coefficients that are unknown (in this case kHEI and kcat). It is very important for the estimation process to pre-select appropriate starting values for the rate coefficients that are estimated. The starting values used in this work were obtained by careful consideration of what may be expected for the individual rate coefficients on the basis of the chemistry involved. The starting value of 4.5×10-3 s-1 for

kHEI was selected based on the hydrolysis of phenyloxalates as reported by Catheral et 293,299 4 -1 al. and Neuvonen. The starting value of 4.2×10 s for kcat was selected based on values reported by Ciscato et al.241

R O OH R kHEI O O R O O + H2O2 + OO I O PhOx HEI PhOH

S1 O O kcat + CO2 + OO II

HEI Pyr Pyr*

S1 kfluo III

Pyr* Pyr

Scheme 8.2: Overview of the reaction steps involved in the chemiluminescent unfolding of SCNP2 and simulated in PREDICI®.

168 Appendix

Table 8.7: Rate coefﬁcients of HEI formation (kHEI) and catalytic pyrene emission (kcat) leading to qualitative description of CL emission behaviour.

-1 -1 kHEI / s kcat / s

SCNP1 11±5 260±4 SCNP2 160±24 470±4 SCNP3 90±5 470±4

IX. Simulated Emission Data

2.5G

1.0

2.0G sim. [PhOx]

0.8

sim. [Diox]

sim. [hv]

exp. [hv] 1.5G

0.6

1.0G

0.4 1-conversion

0.2 500.0mG

0.0

0 10 20 30 40 50 60 70 80 90 100 110 120

time / s

Figure 8.50: Simulated concentration of the PO crosslinker (black), of the HEI (red) and the photons (black) compared to the experimental concentration of photons (cyan) for the unfolding of SCNP1

169 Supporting Information for Chapter 5

1.5M -1

1.0M / cts s Intensity

500.0k

0.0

0 30 60 90 120 150 180

t / min

Figure 8.51: Emission of the chemiluminescent unfolding of SCNP1 (black) and the simulated emission (grey).

1.6G

1.0

1.4G

sim. [PhOx]

1.2G

0.8 / cts sim. [Diox]

sim. [hv]

1.0G

exp. [hv]

0.6

0.8G

0.6G 0.4 1-conversion

0.4G cumulative intensitry

0.2

0.2G

0.0 0.0G

0 10 20 30 40 50 60 70 80 90 100 110 120

time / s

Figure 8.52: Simulated concentration of the PO crosslinker (black), of the HEI (red) and the photons (black) compared to the experimental concentration of photons (cyan) for the unfolding of SCNP2

170 Appendix

1.5M -1

1.0M / cts s Intensity

500.0k

0.0

0 30 60 90 120 150 180

t / min

Figure 8.53: Emission of the chemiluminescent unfolding of SCNP2 (red) and the simulated emission (light red).

1.0

0.8G

sim. [PhOx]

0.8 / cts sim. [Diox]

sim. [hv]

0.6G

exp. SCNP3

0.6

0.4G

0.4 1-conversion cumulative intensitry

0.2G

0.2

0.0

0 10 20 30 40 50 60 70 80 90 100 110 120

time / s

Figure 8.54: Simulated concentration of the PO crosslinker (black), of the HEI (red) and the photons (black) compared to the experimental concentration of photons (cyan) for the unfolding of SCNP3

171 Supporting Information for Chapter 5

1.0M -1 / cts s

500.0k Intensity

0.0

0 30 60 90 120 150 180

t / min

Figure 8.55: Emission of the chemiluminescent unfolding of SCNP3 (blue) and the simulated emission (light blue).

172 Appendix

X. Control Experiment

O Cl O HO O 9 + Cl N O O O O HN O S O S N m/n/o

Cl O

HO N

Cl O OH O UV-A

DCM 9 a.t., 1 h O O O O HN O S

S N m/n/o

Scheme 8.3: Light-induced crosslinker-mediated folding procedure to obtain SCNP1-3.

To conﬁrm that the decrease in apparent molecular according to chloroform-SEC as well as the incomplete return to initial hydrodynamic diameter (dH) values can be attributed to polymer-column and polymer-solvent interactions, respectively, a control experiment was performed. Here, parent polymer P1 was functionalised with 2,6- dichloro-4-maleimidophenol (intermediate in the synthesis of 6, cf. section III ) to yield side-chain functionalised polymers with an identical structure as the completely unfolded uSCNP1. While chloroform-SEC was unavailable due to instrument failure and it was impossible to get replacement parts in the current COVID environment, results from THF- SEC were inconclusive. The hydrodynamic radius of the post-modiﬁed P1 obtained from

DOSY, however, was not only lower than the dH of the parent polymer but even lower than the dH of uSCNP1, thus proving the successful unfolding of the SCNPs despite an apparent decrease in molecular weight according to SEC and despite the incomplete

return to initial dH values, and the presence of polymer-column and polymer-solvent interactions.

173 List of Figures

List of Figures

2.1 A) Representation of an ideal network showing no inhomogeneities. νel and textitµel represent the number of elastic chains and number of junctions (crosslinks), respectively. In the present example, νel is equal to twelve, µel is equal to nine and the functionality f of µel is four. The (average) molecular weight of the elastic chains is described by Mc. The cycle rank ξ is four. B) The network exhibits various defects, such as a primary loop (1), a dangling chain end (2) and two entangled chains (3). .5 2.2 network disassembly spectroscopy (NDS). (A) Schematic depiction of

an A2 monomer. Placement of a degradable group (orange star) at a noncentral position along the A2 backbone leads to an S chain (blue) and an L chain (black) after cleavage. (B) End-linking of A2 and B3 yields a network in which each network junction is unique in terms of the orientation of S and L chains. Primary loops (asterisks) cannot reside at SSS or LLL junctions. (C) Disassembly of the network yields products whose masses depend on the junction source of that product. (#Product size, or relative

mass, does not include the mass of B3, which is assumed to be constant.) The probabilities for formation of each trifunctional product at ideal or

loop junctions are listed as pideal and ploop, respectively. The number of primary loops is captured in the ratio [LLL]:[SLL] or λ3. (D) Plot of the loop ratio, λ3, vs. fraction of loop junctions, nλ, for a trifunctional network; λ3 varies with nλ according to Eq. 1. Fully looped “dumbbell” polymers have loop-linear-loop architectures. Reprinted from [102]. Copyright 2012 National Academy of Sciences of the USA...... 11 2.3 Overview about the possible electronic transitions after a molecule was excited by ultraviolet/visible (UV/Vis) irradiation. Left: Excitation pathway leading to ﬂuorescence. Right: Excitation pathway leading to phosphorescence. The ﬁgure is adapted from [147]...... 17

174 List of Figures

2.4 A. An overview of early dioxetanes synthesised by Schaap and colleagues, with the Schaap’s dioxetane on the far right.2,200,201 B. Exchange of the hydroxy-functionality with sulfur, nitrogen or even carbon species still allows for the CL decay of the dioxetane with an appropriate trigger.206–210 C. An incomplete selection of protecting group modiﬁcations that allow for the detection of numerous analytes via CL.17,21,183,211–214 ...... 31

3.1 PFTR conversion of DDT and 3PFB linker vs. reaction time employing 0.10 equivalents of TBAOH (green), TBAF (red), TBABr (orange) or

DBU (blue), respectively ([SH]0=75 mM, [D8]THF). The red dashed line indicates the theoretical conversion. Error estimations have been performed and are provided in Supporting Information for Chapter 3, Table 7.1 and Table 7.2. Error bars have been omitted here for better visualisation. 44

3.2 B) PFTR conversion of dodecanethiol and 3PFB linker vs. reaction time employing 0.50 equivalents of TBAOH, TBAF, TBABr or DBU, respec-

tively, in regards to the thiol ([SH]0=75 mm, [D8]THF.). The red dashed line indicates the theoretical conversion. Error estimations have been performed and are provided in Supporting Information for Chapter 3, Table 7.1 and Table 7.2. Error bars have been omitted here for better visualisation...... 45

3.3 PFTR conversion of dodecanethiol and 3PFB linker with 0.10, 0.20 and 0.30 equivalents of TBAF,respectively added in a single step (blue) or as a multiple addition of 0.1 equivalent each time (green). The red dashed line indicates the theoretical conversion. Error bars are based on a systematic error in NMR of 10%. For error propagation calculations refer to Table 7.3, Supporting Information for Chapter 3...... 46

3.4 PFTR conversion vs. time of DDT (purple), MBM (orange) or TP (cyan) with the 3PFB linker in THF. 0.10 equivalents of TBAF were employed as base. The red dashed line indicates the theoretical conversion based on the amount of base added. Dotted lines are intended as guides for the eye. Error estimations have been performed and are provided in Table 7.4, Supporting Information for Chapter 3. Error bars have been omitted here for better visualisation...... 47

175 List of Figures

3.5 PFTR conversion vs. time of dodecanethiol (purple), 4-methoxybenzyl mercaptan (orange) or benzenethiol (cyan), with 3PFB linker in THF (circle), DMF (triangle) or DMSO (square), respectively. In each case 0.10 equivalents of TBAF in respect to the functional groups were employed. DDT proved to be insoluble in DMSO, thus the reaction data was excluded. The red dashed line indicates the theoretical conversion. Dotted lines are intended as guides for the eye. Error estimations have been performed and are provided in Table 7.4 to Table 7.6, Supporting Information for Chapter 3. Error bars have been omitted here for better visualisation...... 49

4.1 Total CL emission vs concentration of ﬂuoride (averaged over NMR, LC and expected conversion) compared to the TBAF guideline (black squares) of adamantyl thiol (AT) (red), methoxy triethylene glycol thiol (mTrEGT) (blue), and benzyl thiol (BT) (green). Solid lines represent linear ﬁts of the obtained data...... 55

4.2 Total CL emission vs concentration of ﬂuoride (averaged over NMR, LC and expected conversion) compared to the TBAF guideline (black squares) and the back calculated conversion (stars) of the PFTR of PEG-SH. Solid lines represent linear ﬁts of the obtained data...... 57

4.3 Total CL emission vs theoretical ﬂuoride concentration (dots) and the concentration as back calculated from the total emission (stars) for the read-out of PFTR networks employing 1,4-benzendimethane thiol (BDT) (gold) and triethylene glycol dithiol (DODT) (cyan) as bis-thiols. Solid lines represent linear ﬁts of the obtained data. The rectangles indicate areas of ’soluble network’ formation for the according thiols...... 58

5.1 A) Projection of the DOSY NMR of P1-3 (dashed lines), SCNP1-3 (solid lines) and uSCNP1-3 (dotted lines) in the diffusion dimension conﬁrming successful compaction and subsequent expansion of the polymer chain. B) Mean time-dependant emission proﬁles of the chemiluminescent unfolding of SCNP1-3. C) Simulated cumulative CL emission upon unfolding of SCNP2 (solid golden line) compared to the experimental cumulative emission (dashed golden line). The green and purple line depict the population of the the crosslinker and the HEI, respectively...... 66

176 List of Figures

6.1 A. Thioester-protected phenyl dioxetane readily undergo aminolysis under emission of light while forming stable amide bonds.208,296 B. Perfluori- nated phenyl dioxetanes as activated ester would allow for a direct CL emission upon cleavage. However, perfluorination of the phenyldioxetane seems impractical. C. A perfluorinated 4-(hydroxymethyl)phenolic ester functions as a spacer and a leaving group. Upon cleavage of the activated ester, the spacer undergoes quinone methide elimination, thus triggering the CL of Schaap’s dioxetane...... 75

1 8.1 H NMR spectrum (CDCl3, 400 MHz) of 3PFB...... 114 19 8.2 F NMR spectrum (CDCl3, 377 MHz) of 3PFB...... 115 13 8.3 C NMR spectrum (CDCl3, 101 MHz) of 3PFB...... 115 19 8.4 F NMR spectrum (CDCl3, 377 MHz) of 3PFB after approx. 30% conversion in a PFTR...... 116 8.5 Inﬂuence of lower (blue) as well as higher concentration (green), temperature (magenta), and reaction vial (orange) on the conversion. Samples were taken after 40 h of reaction time. Neither of the three parameters showed signiﬁcant change in the conversion and minor changes are well within error margins...... 120 8.6 Emission spectrum of the LZC-Vis lamps...... 125 1 8.7 H NMR spectrum (CDCl3, 600 MHz) of 3PFB...... 126 13 8.8 C NMR spectrum (CDCl3, 150 MHz) of 3PFB...... 127 19 8.9 F NMR spectrum (CDCl3, 564 MHz) of 3PFB...... 127 1 8.10 H NMR spectrum (CDCl3, 600 MHz) of the CL probe ’Schaap’s dioxetane’.128 8.11 CL mission spectrum of the CL probe in acetonitrile. The yellow box highlights the wavelength regime that was recorded and integrated. . . . . 129 8.12 Time-dependant CL intensity of TBAF at various concentrations over the course of 15 h...... 130 8.13 Time-dependant CL intensity after PFTR of AT at various concentrations over the course of 15 h...... 131 8.14 Time-dependant CL intensity after PFTR of mTrEGT at various concentrations over the course of 15 h...... 131 8.15 Time-dependant CL intensity after PFTR of BT at various concentrations over the course of 15 h...... 132 8.16 Time-dependant CL intensity after PFTR of PEG-SH at various concentrations over the course of 15 h...... 132

177 List of Figures

8.17 Time-dependant CL intensity after PFTR of BDT at various concentrations over the course of 15 h...... 133 8.18 Time-dependant CL intensity after PFTR of DODT at various concentrations over the course of 15 h...... 133 8.19 Total integrated CL emission of 10 mM solutions TBA bromide, chloride, cyanate, hydrogensulfate, iodide, perchlorate and fluoride as well as a 1 mM solution of TBAF. After 2 h, cyanate shows only 2.3% of the integrated CL intensity compared to equimolar amounts of fluoride and only 57% of integrated CL compared to 0.1 equivalents of fluoride. Integrated CL intensity for the other investigated anions lies between 16 and 70 ppm. 134 19 8.20 F NMR spectrum (ACN-d3, 564 MHz) of 3PFB after approx. 24% conversion in a PFTR...... 135 19 8.21 F NMR spectrum (ACN-d3, 564 MHz) of 3PFB after approx. 24% conversion in a PFTR and 22% hydroxy substitution...... 136 8.22 LC chromatogram of 3PFB after approx. 50% conversion in a PFTR. . . . 137 8.23 CL emission vs fluoride concentration for AT according to expected yields (red dots), 19F NMR conversion (blue diamonds), and LC (purple triangles) compared to the TBAF reference (black squares)...... 138 8.24 CL emission vs fluoride concentration for mTrEGT according to expected yields (red dots), 19F NMR conversion (blue diamonds), and LC (purple triangles) compared to the TBAF reference (black squares)...... 139 8.25 CL emission vs fluoride concentration for BT according to expected yields (red dots), 19F NMR conversion (blue diamonds), and LC (purple triangles) compared to the TBAF reference (black squares)...... 139 8.26 CL emission vs fluoride concentration for PEG-SH according to expected yields (red dots), 19F NMR conversion (blue diamonds), and LC (purple triangles) compared to the TBAF reference (black squares). The golden stars depict the concentration of fluoride as calculated from the CL emission.140 19 8.27 F NMR spectrum (ACN-d3, 564 MHz) of the supernatant solutions after network formation employing BDT. Traces 3 to 5 do not show any significant amount of PFB resonances...... 141 8.28 SEC spectra of the supernatant solutions after network formation employing BDT. Toluene was added to the SEC samples (1 mg mL-1, at35.35 mL retention volume) in order to provide a reference. Traces 3 to 5 do not show any significant amount of soluble fractions...... 142

178 List of Figures

19 8.29 F NMR spectrum (ACN-d3, 564 MHz) of the supernatant solutions after network formation employing DODT. Only trace 5 does not show any significant amount of PFB resonances...... 142 8.30 SEC spectra of the supernatant solutions after network formation employing DODT. Toluene was added to the SEC samples (1 mg mL-1, at35.35 mL retention volume) in order to provide a reference. Only trace 5 does not show any significant amount of soluble fractions...... 143 8.31 Emission spectrum of the LZC-UVA lamps...... 149 1 8.32 H NMR spectrum (CDCl3, 600 MHz) of P1...... 154 1 8.33 H NMR spectrum (CD2Cl2, 600 MHz) of SCNP1...... 156 1 8.34 H NMR spectrum (CD2Cl2, 600 MHz) of uSCNP1...... 157 8.35 SEC chromatogram of P1 (solid), SCNP1 (dashed) and uSCNP1 (dotted). 158 8.36 SEC chromatogram of P2 (solid), SCNP2 (dashed) and uSCNP2 (dotted). 159 8.37 SEC chromatogram of P3 (solid), SCNP3 (dashed) and uSCNP3 (dotted). 159 8.38 1H DOSY NMR projection in the diffusion dimension of P1 (solid), SCNP1 (dashed) and uSCNP1 (dotted)...... 160 8.39 1H DOSY NMR projection in the diffusion dimension of P2 (solid), SCNP2 (dashed) and uSCNP2 (dotted)...... 161 8.40 1H DOSY NMR projection in the diffusion dimension of P3 (solid), SCNP3 (dashed) and uSCNP3 (dotted)...... 161 8.41 SEC-MALLS chromatogram of P1...... 162 8.42 SEC-MALLS chromatogram of P2...... 163 8.43 SEC-MALLS chromatogram of P3...... 163 8.44 Averaged time-dependant CL intensity of P1 over the course of 180 min . 164 8.45 Time-dependant CL intensity of five identical samples of P1 over the course of 180 min...... 165 8.46 Averaged time-dependant CL intensity of P21 over the course of 180 min 165 8.47 Time-dependant CL intensity of five identical samples of P2 over the course of 180 min...... 166 8.48 Averaged time-dependant CL intensity of P3 over the course of 180 min . 166 8.49 Time-dependant CL intensity of five identical samples of P3 over the course of 180 min...... 167 8.50 Simulated concentration of the PO crosslinker (black), of the HEI (red) and the photons (black) compared to the experimental concentration of photons (cyan) for the unfolding of SCNP1 ...... 169

179 List of Figures

8.51 Emission of the chemiluminescent unfolding of SCNP1 (black) and the simulated emission (grey)...... 170 8.52 Simulated concentration of the PO crosslinker (black), of the HEI (red) and the photons (black) compared to the experimental concentration of photons (cyan) for the unfolding of SCNP2 ...... 170 8.53 Emission of the chemiluminescent unfolding of SCNP2 (red) and the simulated emission (light red)...... 171 8.54 Simulated concentration of the PO crosslinker (black), of the HEI (red) and the photons (black) compared to the experimental concentration of photons (cyan) for the unfolding of SCNP3 ...... 171 8.55 Emission of the chemiluminescent unfolding of SCNP3 (blue) and the simulated emission (light blue)...... 172

180 List of Schemes

List of Schemes

2.1 General reaction scheme of an SNAr between a thiol and a PFB moiety. . 12 2.2 Depending on the nature of the substituent Y, SNAr on a PFB moiety takes place in either para- (or ortho) or meta-position. The scheme is adapted from [115]...... 13 2.3 Mechanistic details of photo-induced enolisation of o-MBA.161 The electronic

ground state (S0) state is excited to the excited singlet state (S1) state upon irradiation. The excited triplet state (T1) state can undergo hydrogen abstraction forming the o-quinodimethane. The enol species shows different half-lives depending on the isomer that was formed. The Z -isomer undergoes [1,5]-sigmatropic rearrangement, whereas the E-isomer can be trapped in a Diels-Alder (DA) reaction.161 ...... 19 2.4 Reaction of lophine with oxygen under basic conditions (top) and enzymatically catalysed reaction of luciferin with oxygen (bottom). The reaction products are thermodynamically unstable and decompose under formation of a electronically excited state. Notably, both oxygenated species exhibit a 1,2-dioxetane moiety (red)...... 27 2.5 Two proposed mechanisms of 1,2-dioxetane chemiluminescence, a biradical on the left and a concerted on the right...... 28 2.6 The CL mechanism of Schaap’s dioxetane. After formation of the phenoxide via deprotonation or deprotection, an electron transfer from the phenoxide to the dioxetane occurs, followed by cleaved of the O-O bond and immediate cleavage of the C-C bond. The electron back-transfer follows an intermolecular pathway from the adamantyl species to the phenoxide (upper pathway) to form the excited singlet state species. Re- laxation occurs via emission of light...... 29 2.7 Formation of 1,2-dioxetanedione. Oxalyl chloride, aryl oxalates or oxa-

lyldiimidazolide react with H2O2 in a nucleophilic substitution to form the oxalic peracid derivative. Ring-closure leads to 1,2-dioxetanone derivative, subsequent elimination of HX results in the formation of 1,2-dioxetanedione. 33

181 List of Schemes

2.8 Schuster’s CIEEL mechanism on the example of 1,2-dioxetanedione. Adapted from [237] ...... 36

2.9 Mechanism of the luminol chemiluminescence...... 39

3.1 Mechanism of a self-propagated PFTR. When X is a silyltrimethyl moiety, the fluoride triggers the deprotection of the thiol as reported by Hedrick and co-workers.232 When X is a proton, the fluoride deprotonates the thiol and forms hydrogen fluoride. The thiol subsequently reacts with a pentafluorobenzyl moiety in an addition-elimination mechanism yielding the PFTR product and another free fluoride that can re-initiate the PFTR. 42

3.2 Base initiated PFTR of the 3PFB linker with DDT...... 43

4.1 General reaction scheme of the PFTR between the 3PFB linker and a thiol employing TBAOH as a base (left side of the scheme). A ﬂuoride is released in the form of TBAF, which subsequently deprotects the phenol moiety of the CL probe, thus triggering its decay and subsequent emission (right side of the scheme)...... 53

5.1 Overview of the reaction steps of the PO-CL as well as key steps in the chemiluminescent unfolding of SCNPs...... 62

6.1 In the presence of a PFB moiety and a dioxetane-based CL probe, a deprotonated thiol can either undergo a PFTR (left reaction pathway) or oxidative disulﬁde formation (right reaction pathway)...... 73

6.2 Common leaving groups in nucleophilic substitution reactions are chloride, bromide, iodide, tosylate and triﬂate among others. However, these anions exhibit relatively low reactivity, thus complicating the design of a CL probe triggered by them...... 74

6.3 Cascade reaction for the chemiluminescent detection of formaldehyde

according to reference [17]. In their work, R1 is a methyl group and R2 is a hydrogen. When R1 and R2 are chosen as bi- and tri-functional linkers or as pendant groups on a polymer back bone, networks or SCNPs can be formed, respectively, and subsequently analysed via their emission behaviour...... 77

182 List of Schemes

6.4 Schematic depiction of polymer-based PO-CL and the competition between dioxetanedione and phenoxy-dioxetanone as the HEI. While a dioxetanedione HEI can diffuse freely to the fluorophore, a phenoxy- dioxetanone HEI is still connected to the polymer backbone and has to rely on the rotation of the backbone to get into proximity of the fluorophore. 79 6.5 Schematic depiction of the degradation of pH-sensitive micelles upon cleavage of electronpoor phenyloxalates (green). The released phenols are sufficiently acidic to decrease the pH of the solution causing the micelles to deteriorate and release the dye (red), which can subsequently react with the dioxetanedione, causing CL...... 81

8.1 A hydroxide reacts with the 3PFB linker and releases a fluoride. The pKa of HF-fluoride acid-base pair is low enough to allow deprotonation of a thiol, which can subsequenly undergo PFTR and release another fluoride ion. Contrary to a free fluoride, the HF formed upon deprotonation of the thiol will not trigger significant chemiluminescent. Thus, the CL read-out obtained only refers to thiol substitution, but not hydroxy substitution. . . . 144 8.2 Overview of the reaction steps involved in the chemiluminescent unfolding of SCNP2 and simulated in PREDICI®...... 168 8.3 Light-induced crosslinker-mediated folding procedure to obtain SCNP1-3. 173

183 List of Tables

List of Tables

5.1 Characterisation of parent polymers P1-3, SCNP1-3 and uSCNPs1-3. . . 64

8.1 Conversion and error propagation calculation for the reaction of DDT with 3PFB in THF employing 0.10 equivalents of base...... 117 8.2 Conversion and error propagation calculation for the reaction of DDT with 3PFB in THF employing 0.50 equivalents of base...... 118 8.3 Conversion and error propagation calculation for the reaction of DDT with 3PFB in THF when adding TBAF in a single portion or in a step-wise manner...... 118 8.4 Conversion and error propagation calculation for the reaction of TP, MBM, and DDT, respectively, with 3PFB in THF employing 0.10 equivalents of TBAF...... 119 8.5 Conversion and error propagation calculation for the reaction of TP, MBM, and DDT, respectively, with 3PFB in DMF employing 0.10 equivalents of TBAF...... 119 8.6 Conversion and error propagation calculation for the reaction of TP and MBM, respectively, with 3PFB in DMSO employing 0.10 equivalents of TBAF...... 120

8.7 Rate coefﬁcients of HEI formation (kHEI) and catalytic pyrene emission (kcat) leading to qualitative description of CL emission behaviour...... 169

184