Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Spectroscopic studies of halogen bonding in model systems : from one end of the electromagnetic spectrum… Mustoe, Chantal 2018

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Notice for Google Chrome users:
If you are having trouble viewing or searching the PDF with Google Chrome, please download it here instead.

Item Metadata

Download

Media
24-ubc_2019_may_mustoe_chantal.pdf [ 10.03MB ]
Metadata
JSON: 24-1.0372344.json
JSON-LD: 24-1.0372344-ld.json
RDF/XML (Pretty): 24-1.0372344-rdf.xml
RDF/JSON: 24-1.0372344-rdf.json
Turtle: 24-1.0372344-turtle.txt
N-Triples: 24-1.0372344-rdf-ntriples.txt
Original Record: 24-1.0372344-source.json
Full Text
24-1.0372344-fulltext.txt
Citation
24-1.0372344.ris

Full Text

Spectroscopic Studies of HalogenBonding in Model SystemsFrom one end of the electromagnetic spectrum to theotherbyChantal MustoeB.Sc., California Institute of Technology, 2011A THESIS SUBMITTED IN PARTIAL FULFILLMENT OFTHE REQUIREMENTS FOR THE DEGREE OFDOCTOR OF PHILOSOPHYinThe Faculty of Graduate and Postdoctoral Studies(Chemistry)THE UNIVERSITY OF BRITISH COLUMBIA(Vancouver)September 2018c© Chantal Mustoe 2018The following individuals certify that they have read, and recommend tothe Faculty of Graduate and Postdoctoral Studies for acceptance, the thesisentitled:Spectroscopic Studies of Halogen Bonding in Model Systems:From one end of the electromagnetic spectrum to the othersubmitted by Chantal Mustoe in partial fulfillment of the requirementsfor the degree of Doctor of Philosophy in Chemistry.Examining Committee:Pierre Kennepohl, ChemistryCo-supervisorSuzana Straus, ChemistryCo-supervisorLawrence McIntoshSupervisory Committee MemberGlenn SammisUniversity ExaminerJanis McKennaUniversity ExaminerDavid BryceExternal ExaminerAdditional Supervisory Committee Members:Hongbin Li, ChemistrySupervisory Committee MemberDavid Chen, ChemistrySupervisory Committee MemberiiAbstractAt its simplest, chemical bonding involves a combination of two dominantcontributions: direct electrostatics (ionic) and electron sharing (covalent).The relative importance of these contributors has been the subject of signif-icant study in primary (intramolecular) chemical interactions. For example,the relevance and importance of covalent contributions has been a primaryfocus of transition metal chemistry for decades. For weaker secondary chem-ical interactions such as hydrogen bonding (HB) and halogen bonding (XB),the prevailing view in the literature is that electrostatic interactions are sodominant that covalent contributions are negligible. A notable exception isthat of so-called symmetric hydrogen bonds, which exhibit large covalentcontributions.With X-ray Absorption Spectroscopy (XAS), we have provided the firstdirect experimental evidence of covalency in XB. From such studies, we ob-serve that XB exhibit a significantly higher degree of covalency comparedwith HB counterparts of similar bond strength. Notably, the degree of co-valency in certain XBs is equivalent to that observed in transition metalhalides. Our studies provide information of the electronic changes that oc-cur in both the charge donor and charge acceptor in model systems, affordingus a unique experimental view of these weak interactions. We also demon-strate the importance of covalent contributions in XBs by showing the effectof covalency in the electron transfer properties in XB-modified dye sensi-tised solar cells. These results lead us to conclude that XBs should moregenerally be classified as coordinate bonds (and thus identified using an ar-row) to distinguish them from significantly less covalent HBs and other weakinteractions.iiiLay SummaryUnderstanding weak transient attractions between two different chemicalscan play a major role in explaining the behaviours of these chemicals. Inthis thesis, the weak chemical interaction known as a halogen bond hasbeen studied in depth to resolve an argument within the scientific literatureabout the nature of this interaction. It has been shown that this bond canplay a key role in some systems like solar cells, which use chemical dyesto generate electricity, while in other environments, specifically experimentsdesigned to model Alzheimers Disease, other weak chemical interactions aremore important. Redefining the nature of this interaction has helped enableus to explain its importance in these environments.ivPrefaceChapter 2The work in section 2.2 is based on a publication in the Journal of the Amer-ican Chemical Society: Sean W. Robinson, Chantal L. Mustoe, Nicholas G.White, Asha Brown, Amber L. Thompson, Pierre Kennepohl, and PaulD. Beer. Evidence for halogen bond covalency in acyclic and interlockedhalogen- bonding receptor anion recognition. 137(1):499-507, 2015. Thepublished manuscript was written in collaboration with S. Robinson, P.Kennepohl and P. D. Beer. The complexes analysed in sections 2.2 and 2.4were prepared by Paul Beer’s lab, University of Oxford. The Beer lab veri-fied the structure of these complexes by NMR (not included in this thesis)and measured the binding energies. The XAS data collection and analysisfor these complexes was done by the author. The data in section 2.3 waspreviously published in Faraday Discussions: Chantal L. Mustoe, Mathu-san Gunabalasingam, Darren Yu, Brian O. Patrick, and Pierre Kennepohl.Probing covalency in halogen bonds through donor K-edge X-ray absorptionspectroscopy: polyhalides as coordination complexes. 203:79-91, 2017. Thepublished manuscript was written in collaboration with P. Kennpohl. Thesamples analysed in section 2.3 were prepared by undergraduate mentee, D.Yu. The XAS data collection and analysis for these complexes were done bythe author. The work in section 2.4 is further analysis of the complexes insection 2.2. The XAS data collection and analysis was done by the author.This work has not been published previously.Chapter 3The work in sections 3.1 and 3.2 is based on a publication in Nature Com-munications: Fraser G.L. Parlane, Chantal L. Mustoe, Cameron W. Kellett,Sarah J.C. Simon, Wesley B. Swords, Gerald J. Meyer, Pierre Kennepohl,and Curtis P. Berlinguette. Spectroscopic detection of halogen bonding re-solves dye regeneration in the dye-sensitized solar cell. 8(1):1761, 2017. ThevChapter 4published manuscript was written in collaboration with F. Parlane, C. Kel-lett, P. Kennepohl and C. Berlinguette. The work in these sections was acollaborative effort with Curtis Berlinguette’s lab and completed as follows:XAS data was collected by the author and Fraser Parlane. XAS data anal-ysis was done by the author. Photosensitive dyes were synthesised by C.Kellett. DSSC were built by Fraser Parlane. Special thanks to F. Parlanefor his preparation of the graphics for the published work. The work insection 3.3 is further analysis of the dyes synthesised by C. Kellett in sec-tion 3.2. The XAS data collection and analysis was done by the author.This work has not been published previously.Chapter 4The sample preparation, data collection, data analysis, and python scriptwriting for Chapter 4 was done by the author. This work has not been pub-lished previously. Jake Lever helped visualise the output from the author’sNMR analysis scripts for figure 4.19. While none of her data is presented inthis thesis, undergraduate mentee Melanie Backer should also be acknowl-edged for her many attempts to determine why the amyloid beta peptideinitially failed to aggregate.viTable of ContentsAbstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iiiLay Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ivPreface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vTable of Contents . . . . . . . . . . . . . . . . . . . . . . . . . . . . viiList of Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ixList of Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xList of Abbreviations . . . . . . . . . . . . . . . . . . . . . . . . . xiiiAcknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . xivDedication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xvi1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.1 Halogen Bonding . . . . . . . . . . . . . . . . . . . . . . . . . 31.2 X-ray Absorption Spectroscopy . . . . . . . . . . . . . . . . . 81.3 Thesis Aims . . . . . . . . . . . . . . . . . . . . . . . . . . . 152 Evidence of Covalency in Halogen Bonds . . . . . . . . . . . 182.1 Calibration: XAS of Halide Salts . . . . . . . . . . . . . . . . 182.2 Using XAS to probe for XB: XAS of a halide electron donor 202.2.1 Qualitative XAS Analysis . . . . . . . . . . . . . . . . 212.2.2 Quantitative XAS Analysis . . . . . . . . . . . . . . . 252.3 XB Stoichiometry by XAS for interactions between KBr andI2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 322.4 Using XAS to probe for XB: XAS of the XB electron acceptor 332.4.1 I L3-edge XAS: Comparing different systems with thesame counterion . . . . . . . . . . . . . . . . . . . . . 35viiTable of Contents2.4.2 I L3-edge XAS: Comparing analogous systems withdifferent counterions . . . . . . . . . . . . . . . . . . . 402.5 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . 422.6 Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 422.6.1 XAS Sample Preparation . . . . . . . . . . . . . . . . 422.6.2 XAS Data Collection . . . . . . . . . . . . . . . . . . 423 Using XAS to Probe XB in Dye-Sensitised Solar Cells . . 453.1 Dye-Sensitised Solar Cells - Background . . . . . . . . . . . . 453.2 XAS of DSSC . . . . . . . . . . . . . . . . . . . . . . . . . . 473.3 Solution XAS of Dye-Br . . . . . . . . . . . . . . . . . . . . . 553.3.1 Solution XAS of Dye-Br: Br K-edge . . . . . . . . . . 563.3.2 Solution XAS of Dye-Br: Cl K-edge . . . . . . . . . . 593.4 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . 623.5 Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 624 Using XAS, NMR and CD to Probe Amyloid β Aggregation 664.1 Background: Aβ40 aggregation modulation by fluorescein deriva-tives . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 674.2 XAS Studies of Aβ40 Aggregation . . . . . . . . . . . . . . . 704.3 Nuclear Magnetic Resonance Spectroscopic Studies of Aβ Ag-gregation . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 824.4 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . 924.5 Methods . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 935 Conclusions and Future Directions . . . . . . . . . . . . . . . 96Bibliography . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 101AppendicesA BlueprintXAS Parameters . . . . . . . . . . . . . . . . . . . . 110B Python Programs . . . . . . . . . . . . . . . . . . . . . . . . . . 117B.1 TOCSY Program - finding spin systems . . . . . . . . . . . . 117B.2 NOESY Program - finding neighbouring spin systems . . . . 121B.3 Generate TOCSY Spectrum Program . . . . . . . . . . . . . 124C Generated Spin Systems . . . . . . . . . . . . . . . . . . . . . 132viiiList of Tables1.1 Orbitals from which an electron is excited and resulting ex-cited states for relevant XAS edges. . . . . . . . . . . . . . . . 122.1 Cl K-edge XAS E0 values for Chloride Salts . . . . . . . . . . 192.2 Association constants (Ka) of 2-Cl , 3-Cl, 2-Br , and 3-Bras measured by Beer and coworkers . . . . . . . . . . . . . . . 232.3 Filtering Fits . . . . . . . . . . . . . . . . . . . . . . . . . . . 282.4 The average (μ), standard deviation (σ), and the coefficientof variation (CV = σ/μ) are reported for the pre-edge featureintensity. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 292.5 Total Cl3p charge donation as calculated by BlueprintXASAnalysis. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 312.6 The number of intermolecular bonds per system, per iodinesubstituent or per chloride ion for 1-X, 2-X, 3-X, and 4-X. . 362.7 The first maximum of the first derivative of the I L3-edgeXAS for all compounds in figures 2.9 and 2.8. . . . . . . . . . 373.1 Experimental contribution of Cl3p in σ∗XB and relevant infor-mation as determined by analysis of pre-edge feature in XASdata for the Dye-X series. . . . . . . . . . . . . . . . . . . . . 544.1 NMR sample preparation by amount of each component. . . . 94A.1 XAS fit parameters for 1-Cl . . . . . . . . . . . . . . . . . . . 111A.2 XAS fit parameters for 2-Cl . . . . . . . . . . . . . . . . . . . 112A.3 XAS fit parameters for 3-Cl . . . . . . . . . . . . . . . . . . . 113A.4 XAS fit parameters for 4-Cl . . . . . . . . . . . . . . . . . . . 114A.5 XAS fit parameters for NOBF4 + Cl– . . . . . . . . . . . . . 115A.6 XAS fit parameters for Dye-F+. ← Cl– . . . . . . . . . . . . . 115A.7 XAS fit parameters for Dye-Br+. ← Cl– . . . . . . . . . . . . 116A.8 XAS fit parameters for Dye-I+. ← Cl– . . . . . . . . . . . . . 116ixList of Figures1.1 Bond strength variability for common primary and secondarychemical bonds . . . . . . . . . . . . . . . . . . . . . . . . . . 21.2 Schematic of a HB and a XB . . . . . . . . . . . . . . . . . . 51.3 Electrostatic potential map of bromomethane showing an areaof relative positive charge on the covalently-bonded bromineknown as the σ-hole. . . . . . . . . . . . . . . . . . . . . . . . 61.4 Electrostatic potential maps of halomethanes showing the in-crease in size of the σ-hole as the atomic radius, and thuspolarisability, of the halogen increases. . . . . . . . . . . . . . 71.5 Visualisations of electrostatic (a) and covalent (b) interpreta-tions of XB. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 81.6 Mass extinction coefficients for Si, Cl and Fe. . . . . . . . . . 101.7 Br K-edge XAS spectrum of NaCl. . . . . . . . . . . . . . . . 111.8 (Left) Cl K-edge XAS of CuCl2–4 and ZnCl2–4 complexes. (Right)A molecular orbital diagram depicts the transition correspond-ing to the pre-edge feature observed in the XAS spectrum ofthe CuCl2–4 complex. . . . . . . . . . . . . . . . . . . . . . . . 131.9 (Left) Cl K-edge XAS pre-edge features of chlorometal bondsin different MCl2–4 complexes (Right) Calculated Cl3p char-acter for one M – Cl coordinate bond for each complex. . . . . 142.1 Cl K-edge XAS data for a variety of chloride salts. . . . . . . 202.2 Chemical structures of the halide trapping molecules in whichXB covalency was identified by XAS . . . . . . . . . . . . . . 222.3 (a) Cl K-edge and (b) Br K-edge XAS data for halide trappingmolecules. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 242.4 The average of the calculated fits for: (a) 1-Cl, (b) 2-Cl,(c)3-Cl and (d) 4-Cl after fits had been filtered. . . . . . . . . . 262.5 The average of the calculated fits for: (a) 1-Br, (b) 2-Br,(c)3-Br and (d) 4-Br after fits had been filtered. . . . . . . . . 272.6 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32xList of Figures2.7 The molecular orbital diagram of a covalently-bound iodineinvolved in an XB. . . . . . . . . . . . . . . . . . . . . . . . . 352.8 I L3-edge XAS: Comparing different systems with the samecounterion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 392.9 I L3-edge XAS: Comparing analogous systems with differentcounterions . . . . . . . . . . . . . . . . . . . . . . . . . . . . 413.1 Schematic of a DSSC synthesised and built by the Berlinguettelab. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 463.2 Regeneration rate constants (kreg) for the reaction of Dye-X+. on TiO2 substrates with 0.5-10mM of I– (orange) or 10-150mM of Co(bpy)2+3 (blue). . . . . . . . . . . . . . . . . . . 473.3 Increasing σ-hole on Dye-X series. DFT models of the singly-oxidised dyes, Dye-X+. (where X is F, Cl, Br, and I), revealan increasingly electropositive σ-hole on the terminus of thehalogen substituents as the size of the halogen increases. . . . 483.4 Relevant electron transitions induced by Cl K-edge XAS ifXB present between Dye-X and Cl–. . . . . . . . . . . . . . . 493.5 Cl K-edge XAS data of Dye-I in the ground (Dye-I) and oxi-dised (Dye-I+.) states in the presence of Cl–. . . . . . . . . . 513.6 The average calculated fits for the Cl K-edge XAS data of (a)NOBF4+Cl–, (b)Dye-F+. ← Cl–, (c)Dye-Br+. ← Cl– and(d)Dye-I+. ← Cl–. . . . . . . . . . . . . . . . . . . . . . . . . 523.7 The molecular orbital diagram of the covalently-bound brominein Dye-Br+. involved in an XB with Cl–. . . . . . . . . . . . 553.8 Br K-edge XAS data of (a) Dye-Br with and without NBu4Cl,and oxidised Dye-Br (Dye-Br+.)(b) oxidised Dye-Br (Dye-Br+.) with and without (x2 repeats) NBu4Cl . . . . . . . . . 573.9 The molecular orbital diagram of the covalently-bound brominein Dye-Br+. involved in an XB exhibiting pi-type characterwith Cl–. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 583.10 Cl K-edge XAS data of oxidised Dye-Br (Dye-Br+.) returningto the ground state dye in the presence of NBu4Cl as thesample melts (10 minute scans). . . . . . . . . . . . . . . . . . 603.11 Cl K-edge XAS data of oxidised Dye-Br (Dye-Br+.) returningto the ground state dye in the presence of NBu4Cl as thesample melts (3 minute scans). . . . . . . . . . . . . . . . . . 614.1 Chemical structures of the six dyes used in Aβ40 aggregationmodulation studies . . . . . . . . . . . . . . . . . . . . . . . . 68xiList of Figures4.2 CD spectra of Aβ monomer and Aβ aggregates modulated byfluorescein derivatives . . . . . . . . . . . . . . . . . . . . . . 694.3 CD spectra for 30μM Aβ40 incubated with and without EOY,EOB, PHB . . . . . . . . . . . . . . . . . . . . . . . . . . . . 734.4 CD spectra for 30μM Aβ40 incubated with and without FLN,EOY, PHB . . . . . . . . . . . . . . . . . . . . . . . . . . . . 744.5 Br K-edge XAS spectra of Aβ40 incubated with EOY overthe course of 5 days. . . . . . . . . . . . . . . . . . . . . . . . 754.6 Br K-edge XAS spectra of Aβ40 incubated with EOB overthe course of 5 days. . . . . . . . . . . . . . . . . . . . . . . . 764.7 Br K-edge XAS spectra of Aβ40 incubated with PHB overthe course of 5 days. . . . . . . . . . . . . . . . . . . . . . . . 774.8 Br K-edge XAS spectra of 1.2 equivalents of Aβ40 incubatedwith EOY. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 784.9 Br K-edge XAS spectra of 1.2 equivalents of Aβ40 incubatedwith EOB. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 794.10 CD spectra for 30 μM Aβ40 incubated with and without EOY,EOB, PHB . . . . . . . . . . . . . . . . . . . . . . . . . . . . 804.11 Br K-edge XAS spectra of EOY and PHB in presence andabsence of Cl–. . . . . . . . . . . . . . . . . . . . . . . . . . . 814.12 A comparison of various 1H-NOESY spectra illustrating theappearance of crosspeaks upon mixing EOY with Aβ40. . . . 834.13 A comparison of various 1H-NOESY spectra illustrating theappearance and then disappearance of crosspeaks upon mix-ing EOY with Aβ40. . . . . . . . . . . . . . . . . . . . . . . . 844.14 A comparison of various 1H-NOESY spectra illustrating theappearance of additional crosspeaks upon mixing EOY withAβ40. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 854.15 A comparison of various 1H-NOESY spectra illustrating thedisappearance of apo-dye and apo-peptide crosspeaks uponmixing EOY with Aβ40. . . . . . . . . . . . . . . . . . . . . . 864.16 The algorithm logic for the python program developed forTOCSY analysis . . . . . . . . . . . . . . . . . . . . . . . . . 874.17 The algorithm logic for the python program developed forNOESY analysis . . . . . . . . . . . . . . . . . . . . . . . . . 884.18 A screenshot of the TOCSY generator program . . . . . . . . 894.19 Network graphs of spin system distance restraints . . . . . . . 915.1 XB network involving Br– and Br2 studied by Mosquera andcollaborators. . . . . . . . . . . . . . . . . . . . . . . . . . . . 98xiiList of Abbreviations#-X see figure 2.2← charge transfer through either electron transitions or covalencyAβ amyloid beta peptidebpy bypyridineCD Circular DichroismCV coefficient of variationDFT Density Functional TheoryDSSC Dye Sensitised Solar Cell(s)Dye-X see figure 4.1EOB EOsin BEOY EOsin YEPR Electron Paramagnetic ResonanceERB ERythrosine BFLN FLuorosceiNHB hydrogen bond(s)(ing)HOMO Highest Occupied Molecular OrbitalLUMO Lowest Unoccupied Molecular OrbitalNMR Nuclear Magnetic ResonanceNOESY Nuclear Overhauser Effect SpectroscopyPHB PHloxine BROB ROse BengalSSE Sum of Squares ErrorSSRL Stanford Synchrotron Radiation LightsourceTDDFT Time Dependent Density Functional TheoryTFE TetraFluoroEthanolTOCSY TOtal Correlation SpectroscopYUV-Vis Ultra Violet - VisibleX halogenXAS X-ray Absorption SpectroscopyXB halogen bond(s)(ing)Z atomic numberxiiiAcknowledgementsFirstly, I would like to thank my PhD supervisors, Pierre Kennepohl andSuzana Straus for allowing me to be a part of their research groups. I amvery grateful to both of them for their flexibility, support, and encourage-ment. I would like to thank Suzana for giving me the opportunity to delveinto the details of a field which I’ve always loved, Nuclear Magnetic Reso-nance (NMR), and I would like to thank Pierre to opening my eyes to thefascinating power of X-ray Absorption Spectroscopy (XAS).I would like to acknowledge all past and present members of Kennepohland Straus Groups for their friendly support. In particular, I would liketo extend many thanks to Mario Jaime-Delgado for his time and patiencein teaching me to use his software, BlueprintXAS, and for his never-endingwillingness to help. I would also like to recognise my undergraduate menteesfor their patience as I learned how to be a mentor. In particular I’m gratefulto Connie Tang for her friendship and wisdom, and Melanie Backer for herpatience and solidarity in working on a particularly frustrating project.I would like to thank Lawrence McIntosh, the McIntosh lab, and LindsayEltis for my first foray into multidimensional NMR. In particular, I’d like tothank Soumya De for showing me command line, Sparky, and Pymol. I amalso grateful to Mark Okon for his patience in showing me the particularsof NMR data acquisition and processing.I also thoroughly enjoyed collaborating with the Berlinguette group atUBC. I would particulary like to thank Fraser Berlinguette for the laughsand limericks at the synchrotron, and Cameron Kellett for our excellentscientific discussions. I’m very grateful that Sarah Simon approached me todiscuss halogen bonds, thus beginning the conversation which kickstartedthis collaboration.I would also like to thank the other lectures and staff in the chemistrydepartment of the University of British Columbia for making my time atUBC enjoyable. In particular, I am grateful to Anne Thomas for giving methe opportunity to work in course development and to Erin Lindenberg forher continuous support of my outreach endeavours.I am very lucky to have worked with beamline scientists Matthew Lat-xivAcknowledgementstimer and Erik Madsen and safety officer, Cindy Patty, at the StanfordSynchrotron Radiation Lightsource (SSRL). It is because of their sarcasmand positivity that I always look forward to the long hours and frustra-tions of the synchrotron trips. Despite a multitude of demands on theirtime, Matthew and Erik always take time to explain the quirks and tricksof the beamlines, and Cindy’s organisation make the SSRL labs an absolutepleasure to work in.Finally, I would like to thank my family and my partner Jake for theirendless support.xvDedicationThis thesis is dedicated in equal parts to my dad, my mum, my sister, andmy partner, Jake:• To my dad for teaching me that forces cannot be seen, and inspiring mewith “discrete elephants”, engineering demos, and coffee-shop calculus.• To my mum for inspiring me with her love of numbers, puzzles, andpoetry.• To my sister, Elisha, for inspiring me with her humour, love, andcreativity.• And to Jake for the laughs, the love, and sharing and encouraging mypassion for science.xviChapter 1IntroductionMany different types of chemical bonds exist from the very strong to thevery weak, and from exclusively electrostatic to predominantly covalent.While many factors can affect the properties of a chemical bond includingeffective nuclear charge, competing interactions, and orbital sizes, there areonly two phenomena which cause the distance between two atoms to beless than the sum of their van der Waals radii: electrostatics and orbitaloverlap. Another school of thought combines these two effects when under-standing bonding properties. This competing theory represents atoms aspoint charges to explain electron distribution within a bond, thus avoidingthe concept of electron ownership and covalency [1, 2]. While this simpli-fication is attractive, it fails to explain electron transition allowedness andprobabilities. Therefore, this thesis will focus on discussing chemical bondproperties using both electrostatics and covalency.While both electrostatics and covalency can influence chemical bonds,some bonds are exclusively electrostatic. A chemical bond which results onlyfrom electrostatic attractions involves atoms whose electronic structures areindependent from one another. Some examples of these include ionic bondsand van der Waals attractions, notably the strongest and weakest chemicalbonds known. Bonds in which the electronic structures of the atoms areinterdependent have some degree of covalency (also referred to as electronsharing, charge donation and orbital overlap). Unlike exclusively electro-static bonds, bonds which exhibit at least partial covalency have bondingand antibonding orbitals with wavefunctions that include contributions fromboth bonding atoms (equation 1.1).Ψ = α |ψA〉 –√1 – α2 |ψB〉 (1.1)Where:• Ψ is the wavefunction for the bonding orbital• ψA and ψB are the wavefunctions for the orbitals involved in bonding• α is the degree of charge donation from atom A to atom B1Chapter 1. IntroductionExamples of these bonds include traditional covalent bonds and coordinatebonds [3–5]. As defined by the International Union of Pure and AppliedChemistry (IUPAC), a coordinate or dative bond is a bond involving twoatoms, one atom which donates an electron pair to the bond and anotheratom which acts as an acceptor. In other words, unlike a covalent bond whichpreferentially results in homolytic cleavage of the electron pair upon bonddissolution, breaking a coordinate bond will result in heterolytic cleavage ofthe electron pair [6].Figure 1.1: Bond strength variability for common primary and secondarychemical bonds [7]. Reprinted with permission from author.Figure 1.1 includes both primary bonds and secondary bonds, i.e. bondswhich are often referred to as intermolecular forces. This thesis uses the term‘bond’ to apply to any force holding to atoms together to emphasise thatthese interactions exist on a continuum. In bonds, electrostatics and cova-lency determine the bond strength and the role of a bond in a system. Thedependency of bond strength on electrostatics as well as covalency meansthat while bond strength varies drastically for different types of bonds (fig-ure 1.1), it is not directly dependent on the relative distribution of the atoms’21.1. Halogen Bondingelectrons. For example, covalent bonds have the highest degree of covalencyfollowed by coordinate bonds and finally ionic bonds, a trend which is notmirrored in bond strength.This difference in covalency causes the two strongest types of bonds, ionicand covalent, to exhibit very different chemical properties in polar solvents.The exclusively electrostatic ionic bonds are easily dissolved in a polar sol-vent by competing electrostatic interactions. On the other hand, covalentinteractions cannot be broken by using the competing electrostatic inter-actions arising from dissolving the compound in a solvent. Here, althoughoverall bond strength of a covalent bond is similar and often weaker thanan ionic bond, the overlap of electron wavefunctions in the covalent bondensures that this bond remains intact while an ionic bond is often broken.Although secondary bonds are considerably weaker than primary bonds(figure 1.1), these bonds can also play an important role in the behaviourof a system. While figure 1.1 illustrates the most well known types of sec-ondary chemical bonds, there are many other types of weak chemical bonds.Other secondary bonds not represented in figure 1.1 include tetrel, pnicto-gen, chalcogen, and halogen bonds which involve elements in groups 14-17on the periodic table, respectively. These four types of bonds are generallysimilar to hydrogen bonds in that they involve covalently-bonded electron-deficient atoms (specifically a covalently-bonded tetral, pnictogen, chalcogenor halogen atom) interacting with an electron-rich species [8]. The involve-ment of a tetral, pnictogen, chalcogen or halogen atom in a covalent bondresults in an area of electron depletion opposite the bond (this area is re-ferred to as a sigma hole and will be discussed further in the next section).These interactions form when an electron-rich species interacts with thesigma hole [8]. Like hydrogen bonds, these interactions are commonly re-ferred to as non-covalent interactions despite an ongoing debate over therelative contributions of electrostatics and covalency[1, 9–12].This thesis will focus on identifying and characterising halogen bonds indifferent systems, specifically in terms of electrostatics and covalency. Theimportance of these bonds in each system will also be discussed.1.1 Halogen BondingHalogen bonds (XB) are interactions between an electron donor and acovalently-bonded halogen acting as an electron acceptor [13, 14]. The appli-cations of XB include but are not limited to ion recognition [15–17], crystalengineering [18–20], catalysis [21, 22], and molecular self-assembly [23].31.1. Halogen BondingIn understanding the nature of a bond, it is useful to compare to otherlike interactions. Theoretically, the most straightforward way to comparetwo bonds would be to determine the importance of electrostatic and cova-lent contributions in each bond. However, as different contributions to bondstrength are not experimental observables, it is very difficult to determinethe relative importance of electrostatics and covalency in bond strength. Infact, even computationally, we cannot separate electrostatic and covalentcontributions as different computational models yield different results. Twoexperimental observables that are useful in comparing bonds are (1) percentcovalency in terms of maximum possible orbital overlap and (2) overall bondenergy or bond strength. These properties can be used to determine the de-gree of importance of electrostatics and covalency in bonding properties bycomparing bonds which have similar values for (1) but different values for(2) or vice versa.While there are many different types of bonds with which XB can becompared, a comparison of XB with HB and coordinate bonds will provemost useful in this thesis. The comparison with the HB is interesting notonly because this is a common comparison in the literature [16, 23–26], butalso because these two bonds have analogous structures and are comparablein strength [27].To the author’s knowledge, the comparison between XB and coordinatebonds is not present in the literature outside of the work presented in thisthesis. This comparison is nonetheless useful because while these two bondsare not comparable in strength, they exhibit similar degrees of charge do-nation (as will be shown in chapters 2 and 3).While tetrel, pnictogen, and chalcogen bonds are also structurally similarto XB, this comparison is not as illuminating simply due to the fact that,especially in comparison to HB and coordinate bonds, the properties ofthese bonds are not as well understood. Indeed, most research on thesebonds is crystallographic or computational [12, 28]. Other more delocalisedsecondary bonds such as dipole-dipole and van der Waals interactions arestructurally dissimilar to XB and a comparison with these bonds is lessinformative as a result.Two conventions regarding the depiction and discussion of XB in theliterature are worth discussing. First, both XB and HB are often discussedin terms of a bond donor, referring to the covalently-bonded halogen or hy-drogen respectively, and a bond acceptor, the electron-rich species involvedin the bond. As this thesis is mainly concerned with the distribution of elec-trons throughout these bonds, the terms donor and acceptor will be usedto refer to the electron donor and electron acceptor (halogen or hydrogen41.1. Halogen BondingFigure 1.2: Schematic of a HB and a XB. D has been used to represent theelectron donor in each bond. As this thesis is concerned with the distribu-tion of electrons within these bonds, the terms electron donor and electronacceptor will be used rather than bond donor and bond acceptor, as is morecommon in the literature.involved in the strong covalent bond). This unusual representation and ter-minology has been used not to flout tradition but rather to facilitate thediscussion of the electronic properties of XB. Secondly, the use of a dottedline in the literature to illustrate XB and HB serves to emphasise both rela-tively weak bond energy and the limited charge transfer from the halogen orhydrogen, respectively to electron-poor species upon formation of the bond.The convention of using a dotted line to illustrate XB will be challenged inchapter 2.As mentioned previously, XB are analogous to hydrogen bonds (HB),in that they are also weak interactions involving an electron-rich and anelectron poor species (figure 1.2). However, a notable difference betweenXB and HB is the range of possible bond angles. Unlike HB which canhave highly variable bond angles [29], XB usually exhibit bond angles of170 – 180o[13, 14]. This discrepancy in geometries can be explained by51.1. Halogen Bondingthe presence of a region of electron deficiency present in covalently-bondedhalogens referred to as the σ-hole [13, 14] (figure 1.3). Unlike the positiveregion on the H in CH3Br (seen in figure 1.3, blue) which is surrounded byan area of slowly decreasing positive charge, the σ-hole on the covalently-bonded halogen is surrounded by a donut shaped ring of negative charge(red). While the surrounding donut of negative charge is expected as halo-gens are more electronegative than carbon, the σ-hole can also be explainedas it reflects the location of both the filled σC–Br orbital (electron densitybetween C and Br) and the unoccupied σ∗C–Br orbital. The location of thiselectron deficient σ-hole surrounded by an area of negative corresponds witha preferred bond angle of 180o.Figure 1.3: Electrostatic potential map of bromomethane showing an areaof relative positive charge on the covalently-bonded bromine known asthe σ-hole. Electrostatic potentials were generated with Gaussian04 us-ing the B3LYP functional and Basic(3-21G) basis set and then visualisedwith WebMO[30]. Two different views included for clarity. Note: whilebright blue and bright red do represent -0.035V and 0.040V, respectively,the change in electrostatic potential depicted by change in hue as depictedby the scale is approximate for all electrostatic potential maps visualisedwith WebMO [30] in this thesis.Due to the size and polarisability of the different halogens, the size ofσ-hole of a covalently-bonded halogen increases with the atomic radius (fig-ure 1.4). As covalently-bonded fluorine lacks a discernable σ-hole, it is rarelyable to act as the electron acceptor in an XB. On the other hand, covalently-bonded iodine with the largest and most positive σ-hole is capable of formingthe strongest XB. [14].The above explanation of XB focuses on the electrostatic nature of theXB, and until recently such interactions were generally considered to be61.1. Halogen BondingFigure 1.4: Electrostatic potential maps of halomethanes showing the in-crease in size of the σ-hole as the atomic radius, and thus polarisability, ofthe halogen increases. Images generated with WebMO[30] molecular orbitalcalculations of fluoro-, chloro-, bromo- and iodomethane.predominantly electrostatic [13, 20, 25]. Indeed, the current widespreadview in the literature is that the bonds are an electrostatic interaction inwhich an electron-rich species is attracted to the positively charged σ-holeformed in a covalently-bonded halogen [13, 14, 31]. However, a competingand less widespread view is that these bonds are partially covalent due toan overlap between the σ∗ orbital of the covalently-bonded halogen andan orbital of the same symmetry on the electron donor. In figure 1.5b,the electron donor is a halide and thus the σ∗ orbital is overlapping withthe X- LUMO pz orbital. This partially covalent model of the XB couldfurther explain the discrepancy between the geometries of HB and XB asthe stability gained by orbital overlap between σ∗R–X and pz (for example)orbitals is lost as the bond angle decreases. It is important to note that thisexplanation assumes that XB are more covalent in nature than HB; thisassumption will be explored further in chapter 2. Indeed, this thesis aimsto address the validity of using the partially covalent model to describe XB.Despite the appearance of XB in the literature several decades ago, theability to identify and characterise these bonds is mostly, though not entirely,limited to X-ray crystallography [12, 32]. X-ray crystallography can be usedto identify when the distance between a halogen and its neighbour is lessthan the sum of the atomic radii, and if this is true for two or more atomsnear a particular halogen, the halogen is involved in one or more XB. The useof NMR to indirectly detect XB has also been investigated [32]. This thesisaims to establish a new method to detect XB in non-crystalline systems,71.2. X-ray Absorption SpectroscopyFigure 1.5: Visualisations of electrostatic (a) and covalent (b) interpre-tations of XB. In the partially covalent model (b), orbitals σR–X of thecovalently-bonded halogen and pz of the electron donor are depicted to il-lustrate the potential overlap between these orbitals. In XB, X can representany covalently-bonded Cl, Br, or I, and D can be any electron donor. Elec-trostatic potentials were generated with Gaussian04 using the B3LYP theoryand Basic(3-21G) basis set and then visualised with WebMO [30].discussed in the following section [27]. With the increasing applicationsof these interactions in the literature [15–23], it is useful to both identifythe presence of these bonds in non-crystalline systems and understand theeffects the distribution of charge within an XB can have on such systems.1.2 X-ray Absorption SpectroscopyX-ray Absorption Spectroscopy(XAS) is an element-specific technique whichenables us to probe the nature of low-lying unoccupied (or partially occu-pied) orbitals using high energy X-ray radiation from a synchrotron to excitecore electrons. In XAS, specific wavelengths of X-ray radiation produced bythe acceleration of electrons are selected using a double-crystal monochro-mator. The resulting beam of X-ray radiation passes through the sample to(an) ion chamber detector(s). The relative angles of the two crystals in themonochromator can be changed in order to scan over a range of energies.The use of the synchrotron as a lightsource yields a high flux of photons ata wide range of energies. For example, XAS experiments in this thesis wereconducted at various energies ranges within 2 keV to 15 keV.This ability to select energies of high energy photons enables the exci-tation of core electrons. XAS experiments are conducted at the ionisation81.2. X-ray Absorption Spectroscopyenergy of a particular core electron of a particular element. For example,to probe the transitions available to a Br1s electron, the researcher startsthe experiment approximately 100 eV below the ionisation energy of Br1selectron and slowly increases the energy of the incident photon beam untilat least 100 eV above the ionisation energy. The detection of one of fourof the following (1) transmission of incident beam, (2) total electron yieldfrom sample, (3) partial X-ray fluorescence yield, or (4) total X-ray fluores-cence yield results in a spectrum similar to that seen in figure 1.7. In anXAS spectrum, a sudden increase in signal intensity reflects the ionisationof electrons from an orbital of interest (figure 1.6). The inflection point ofthis feature is referred to as the edge (figure 1.7).It is important to note that in XAS peak intensity refers to the area underthe peak rather than the height of the peak by convention. This terminol-ogy is used as peak area rather than peak height is directly proportional tothe probability of the associated transition. When peak width and shaperemain constant, peak height is also proportional to associated transitionsbut this is not the case in XAS. In XAS, peak shapes can be Gaussian dis-tributions (arises from experimental noise), Lawrencian distributions (arisesfrom Heisenberg’s uncertainty principle) or Voight functions, a combinationof the two. As peak shape changes based on the different contributions ofexperimental noise, line-broadening from Heisenberg’s uncertainty principle,and physical phenomena contributing to the peak (photon/electron emissionand/or scattering), it is convention in XAS to use peak intensity to refer tothe ‘area under the peak’ rather than the ‘height of the peak’.91.2. X-ray Absorption SpectroscopyFigure 1.6: Mass extinction coefficients for Si, Cl, and Fe. The almoststepwise increase in intensities correspond to the ionisation of the followingelectrons: Si2s at 150 eV, Si1s at 1839 eV, Cl2s at 270 eV, Cl1s at 2822 eV,Fe3s at 91 eV, Fe2s at 845 eV and Fe3s at 7122 eV. Figure reprinted withpermission from the director of the Centre for X-ray Optics [33].101.2. X-ray Absorption SpectroscopyFigure 1.7: Br K-edge XAS spectrum of NaCl. Edge, pre-edge region andBr5p←1s peak have been labelled for clarity. Data was collected at beamline7-3 of the Stanford Synchrotron Radiation Lightsource (SSRL).The use of the term edge feature can be somewhat ambiguous as itcan refer to either the sudden increase in signal intensity or to the firstpeak at or after the energy of ionisation. For clarity, in this thesis theterm edge will refer to the sudden increase in signal intensity, and the termedge peak will refer to the peak immediately following this rise. For thespectrum in figure 1.7, the energy of the edge is 13474 eV, and the edge peakpredominantly corresponds to the peak identified as the Br5p←1s transition.All peaks to the left of the edge peak in an XAS spectrum are referredto as pre-edge features or pre-edge peaks. Other physical phenomena canalso contribute to the edge feature including other high energy transitions,here Br6p←1s for example, and scattering. Thus, this feature can be verybroad (approximately 10 eV) and is not used to quantify the probability ofassociated transitions.In XAS, the nomenclature K-edge, L-edge, and M-edge are used to referto the excitation of electrons from the n = 1, n = 2 and n = 3 shells,respectively. For edges with more than one possible angular momentum, anumerical subscript is used to clarify the resulting excited state. Table 1.1includes all edges discussed in this thesis.The two main criteria which dictate the intensity of features correspond-111.2. X-ray Absorption SpectroscopyEdge Orbital Resulting Excited StateK-edge 1s 2S1/2L1-edge 2s2S1/2L2-edge 2p2P1/2L3-edge 2p2P3/2Table 1.1: Orbitals from which an electron is excited and resulting excitedstates for relevant XAS edges. Atomic state configuration given as 2S+1LJwhere S denotes spin, L denotes angular momentum and J = L – S, L – S +1, ...L + Sing to these transitions are (1) spatial overlap between the initial orbital andthe final acceptor orbital and (2) the angular momentum of the acceptor or-bital. The first criteria tells us that the acceptor orbital must be at leastpartially located on the same atom as the initial orbital. The second cri-teria arises from the fact that the change in the angular momentum of theelectron must equal that of the incident photon (Δl = ±1 where l is thequantum number for orbital angular momentum). For example, the inten-sity of an electron transition from an s orbital to an acceptor orbital willdirectly correlate to the %p character of the acceptor orbital.One of the strengths of XAS is that due to the large separation of theedges of core electrons in different elements and different edges within thesame element, XAS is an element-specific technique. For the spectrum infigure 1.7, all signal intensity is a direct result of excitation of only Br1selectrons. The next closest ionisation edges are the Se K-edge (12658 eV)and the Kr K-edge (14326 eV). Thus, any contamination in the samplewould be clearly visible as a second edge jump >700 eV before/after theBr K-edge edge jump. The element specificity of XAS enables experimentswhich probe the properties of a single element within a sample, albeit thesignal is an average of all atoms of that particular element within a sample.This particular strength is what has enabled the use of XAS to probe XBin this thesis.Although XAS has not been used to probe the nature of XB prior tothe work done in this thesis, XAS is a well established method to probe thecovalency of ligand-metal donor-acceptor interactions [4]. The first use ofXAS to probe the covalency of donor-acceptor systems was Solomon andco-workers’ use of Cl K-edge XAS to evaluate the degree of delocalization inchloro-metal bonds [3, 4, 34]. In such cases, a pre-edge feature is observedin the spectrum that formally corresponds to excitation of a Cl1s electron121.2. X-ray Absorption SpectroscopyFigure 1.8: (Left) Cl K-edge XAS of CuCl2–4 and ZnCl2–4 complexes. (Right)A molecular orbital diagram depicts the transition corresponding to the pre-edge feature observed in the XAS spectrum of the CuCl2–4 complex. Figureadapted with permission from [3] c© 2000 American Chemical Society.to empty valence d-orbitals on the transition metal acceptor. The pre-edge feature in the CuCl2–4 spectrum corresponds to the σ∗Cu←Cl ← Cl1stransition arising from orbital mixing between the filled Cl3p orbital andthe partially unoccupied Cu3dx2–y2orbital. By contrast, overlap of the filledCl3p and the filled Zn3dx2–y2orbital in d10 ZnCl2–4 complex does not giverise to a similar Laporte allowed (Δl = ±1) transition [3].There are various other experimental methods which can be used toevaluate covalency of coordinate bonds including analysis of the g valuesand coupling in electron paramagnetic resonance spectroscopy (EPR) andligand to metal charge transfer transitions in UV-Vis spectroscopy, and XAS[3, 35]. Each of these techniques has associated limitations. EPR can onlybe used for a sample with at least 1 unpaired electron. In UV-Vis, theexistence of ligand and metal character in the orbital of excitation requiresthat the probability of the transition depends on both the ligand and metalcharacter.Solomon and co-workers also pioneered the use of XAS to quantify thedegree of covalency in coordinate bonds. Evaluation of coordinate bondcovalency in XAS is possible as the intensity of the pre-edge feature inthe ligand K-edge XAS is directly proportional to the amount of np ligand131.2. X-ray Absorption SpectroscopyFigure 1.9: (Left) Cl K-edge XAS pre-edge features of chlorometal bonds indifferent MCl2–4 complexes (Right) Calculated Cl3p character for one M – Clcoordinate bond for each complex. Figure adapted with permission from [3]c© 2000 American Chemical Society.character in the acceptor orbital where np denotes the frontier p orbitalligand. We can show this relationship using the following equations.The intensity of the pre-edge feature corresponding to the Ψ∗ ← Cl1stransition is defined as follows:IΨ∗←Cl1s = k| 〈Ψ∗|~r|Cl1s〉 |2 (1.2)where Ψ∗ is the antibonding orbital of a metal-coordinate bond, I is theintensity of the peak corresponding to this transition, k is a constant, and ~ris the electric dipole operator.We can also rewrite equation 1.1 for the wavefunction of metal-chloridecoordinate antibonding orbital:Ψ∗ =√1 – α2 |Mnd〉 – α |Cl3p〉 (1.3)where nd are the valence d orbitals on the metal (M).As Ψ∗ ← Cl1s transition only occurs due to spatial overlap between theCl3p and Cl1s, orbitals, we can rewrite the intensity of this transition asfollows:IΨ∗←Cl1s = α2k| 〈Cl3p|~r|Cl1s〉 |2 = α2(ICl3p←Cl1s) (1.4)We can see in equation 1.4 that the intensity of the pre-edge feature isdirectly proportionate to the amount of charge transfer(α2) from the Cl3p141.3. Thesis Aimsorbital, as the donor in this example, to the electron acceptor (in this casea metal d orbital). Solomon and coworkers took advantage of the fact thatthe ICl3p←Cl1s term is not affected by different electron acceptors. Thus, theratio of metal-chloride charge transfer to the intensity of the XAS pre-edgefeature corresponding to the Ψ∗ ← Cl1s transition will be the same for allmetal-chloride bonds, and indeed for any bond in which a chloride acts aselectron donor.Providing a relevant calibration standard is available, comparing the in-tensity of the pre-edge feature of interest to the pre-edge feature in the cali-bration standard is a convenient method of obtaining a quantitative measureof the degree of ligand and metal character in the acceptor orbital [3–5]. Us-ing EPR, XAS and density functional theory (DFT) calculations, Solomonand coworkers established calibration standards for evaluating covalency inmetal-chloride and metal-sulfide bonds (figure 1.9 for chlorometal bond co-valencies). As the intensity of the pre-edge feature is directly proportionalto the percent Cl3p in the acceptor orbital of the MCl2–4 complex, thesecomplexes can be used as calibration standards for systems in which a com-parison to EPR is not possible. Specifically, with its 1 electron hole andwell-defined pre-edge feature (no overlap with edge), CuCl2–4 will be used asa reference for this thesis.1.3 Thesis AimsXAS will be used to probe for XB in three different applications in this thesis:ion trapping complexes, dye-sensitized solar cells (DSSC) and Alzheimers-associated amyloid β peptide aggregation (chapters 2,3, and 4 respectively).Despite the dissimilitude between these three systems, they are all unitedby the importance of secondary chemical bonds. Both halogen bonds andhydrogen bonds can be used to bind ions in ion recognition complexes [27].DSSCs require an interaction between the light sensitive dyes and a redoxactive electrolyte species to regenerate the excited dyes [36]. The amyloid βpeptide also aggregates via hydrogen bond formation, amongst other inter-actions, and this aggregation can be modulated by halogenated fluoresceinderivatives [37]. XAS will be used to elucidate the role, if any, of XB in these3 different systems. Other model systems with less apparent applicationswill also be employed to characterise these bonds.The aim of thesis is three fold:151.3. Thesis Aims1. To establish the use of XAS to probe the properties of halogenbonds in different systems: In chapter 2, XAS will be employed tostudy XB in systems through XAS of both the XB electron donor andthe XB electron acceptor. To the author’s knowledge, XAS has notbeen previously used to study XB. The benefits and drawbacks of bothapproaches will be discussed. Some of this work has been previouslypublished by the author [27].2. To provide evidence that halogen bonds are partially covalentin nature using XAS: In chapter 2, an XAS method analogous tothat used to probe coordinate bond covalency will be used to bothprovide evidence for and directly quantify the degree of charge transferin XB. This work has been previously published by the author [27, 38].3. To probe systems in which halogen bonds may play an im-portant role: In chapters 3 and 4, the methodology established inchapter 2 will be applied to systems in which XB have not previouslybeen identified. It is known that halogens play a key role in bothDSSC [39] and in the modulation of amyloid β peptide aggregation byhalogenated fluorescein derivatives [37]. In DSSC, it has been shownthat both the presence and identity of halogen substituents can affectDSSC efficiency [39]. The halogenation of fluorescein derivatives hasbeen shown to modulate amyloid β aggregation [37]. The work donein chapters 3 and 4 of this thesis will establish whether XB play a rolein the importance of halogen substituents in these two systems. Someof this work has been previously published by the author [36].161.3. Thesis AimsWhen looking at chemical bonds,We care if they’re short or they’re long.Bond strength is keyIn this inquiryBut so are the shared electrons.17Chapter 2Evidence of Covalency inHalogen Bonds ∗As presented in chapter 1, the nature of the XB is debated within the liter-ature with the purely electrostatic representation being the predominantlyaccepted theory. In the current chapter, direct evidence will show that XBin model compounds synthesised by the Beer group at Oxford have partialcovalency (figure 2.2). This is the first direct evidence of XB covalency inthe literature.XAS will be used to assess the degree of covalency of both XB and HB inthese systems. To establish the nature of the interactions between the halide(Cl– or Br–) donor and the electron acceptor, we have employed Cl and BrK-edge XAS. As mentioned previously, Solomon and co-workers establishedCl K-edge XAS as a technique for probing the covalency of metal-chloridecoordinate bonds via the Cl– ligands [3, 4]. However, using XAS to probe thecovalency of bonds between two non-metals has not been done previously.This method will be used to assess the covalency of both XB and HB presentin Beer and co-workers’ model systems, thus enabling a comparison of theproperties of these two weak interactions. As XB do not necessarily havehalide electron donors, I L1- and L3-edge XAS will also be used to probethe XB via the XB electron acceptor. If successful, this method could begeneralisable to all XB.2.1 Calibration: XAS of Halide SaltsXAS requires the use of calibration standards for calibration of the energyof electron excitation. As XAS is an element-specific technique, it is usuallynecessary to calibrate using a standard containing the same element. Agenerally accepted calibration standard for metal XAS is the metal itself in∗Section 2.2 is based on the author’s contributions in previously published work [27].The data in section 2.3 was also previously published by the author [38]. Unless noted,all work was done by the author. See Preface for more details.182.1. Calibration: XAS of Halide SaltsCompound E0 (eV)LiCl 2825.75NaCl 2825.95KCl 2825.35RbCl 2825.15CsCl 2825.45CaCl2 2825.65MgCl2 2825.25NBu4CuCl4 2824.70Table 2.1: Cl K-edge XAS E0 values for Chloride Saltselemental form. As this is not possible for XAS at halide edges, differenthalide salts have been used for calibration.Although CuCl2–4 salts are the generally used for Cl K-edge XAS due totheir distinct pre-edge feature at 2820.2 eV (figure 2.1(b)) [3, 4], other Cl–containing compounds can be used. A variety of Cl– salts have been analysedand referenced to CuCl2–4 . The first inflection point in the XAS spectrumhas been reported for all salts analysed (table 2.1). As both KCl and NaClwere used as calibration standards in this thesis, these experiments wereneeded to ensure calibration consistency. Furthermore, as it is reasonableto use any of these salts as calibration standards, publishing a consolidatedtable of these values was deemed useful for future researchers.As expected, XAS of halide salts shows that there is no pre-edge featurein Cl K-edge XAS of Cl– (figures 2.1(a) and 2.1(b)). The only salt whichexhibits a pre-edge feature is tetrabutyl copper chloride. In CuCl2–4 , chlorideis involved in a partially covalent coordinate bond with copper. The cova-lency of this bond results in the new allowed transition (σ∗Cu←Cl ← Cl1s),and this transition corresponds to the pre-edge feature visible at 2820.2 eV.XAS experiments were also performed at the Br K-edge and the I L3-edge. Similar studies for different iodide and bromide salts were not donedue to time constraints at the synchrotron. KBr was used as a standard forBr K-edge XAS (figure 1.7) and KI was used a calibration standard for I L3-edge. The energy values used to calibrate the spectra of these compoundsare the ionisation energies of the Br1s (E0 = 13474 eV) and I2p (E0 =4557 eV) electrons as reported in the X-ray Data Booklet [33]. The trueionisation energies for the reference compounds KBr and KI are unlikely tobe exactly 13474.0 eV and 4557.0 eV as ionisation energies for two electronsin the same element at the same energy level in different compounds are192.2. Using XAS to probe for XB: XAS of a halide electron donor(a) (b)Figure 2.1: Cl K-edge XAS data for a variety of chloride salts including(a) the alkali metals and (b) alkaline earth metals and tetrabutyl copperchloride.not the same (table 2.1). However, as all spectra are calibrated using thesame method (unless otherwise noted), any energy discrepancy between theionisation energies reported in the X-ray Data Booklet for the Br1s and I2pelectrons and the true ionisation energies for KBr and KI will not affectconclusions drawn from comparing multiple XAS spectra collected at thesame edge.2.2 Using XAS to probe for XB: XAS of a halideelectron donorThe systems described in this chapter (figure 2.2) were synthesised by theBeer Group at Oxford to investigate the use of XB and HB to trap halides.The Beer group synthesised these model compounds to demonstrate the ap-plications of XB in controlling the specificity of halide trapping systems.Despite the similarity of the strengths of XB and HB [24], the limited exam-ples of using these bonds to coordinate anions in competing solvents showdifferent behaviour between the halogen and HB systems [15, 26, 40]. Thesemodel systems are designed to compare the anion selectivity of these weak202.2. Using XAS to probe for XB: XAS of a halide electron donorinteractions. Specifically, these systems allow for three different compar-isons: (1) 1 XB (compound denoted as 1-X in figure 2.2) vs. 2 XB (3-X)(2) 2 HB (2-X) vs. 2 XB (3-X) and (3) 2 XB and 2 HB in direct competition(4-X) vs. 2 XB and 2 HB in isolation (2-X and 3-X, respectively).Iodine was chosen as the electron acceptor in the XB because, as thelargest and most polarisable of the halogens, iodine substituents are moreelectron deficient than Br or Cl substituents in analogous systems. Thusiodine is capable of forming the strongest XB. Beer and coworkers analysedthe anion coordinating properties of these systems for Cl–, Br–, I–, C2H3O–2,H2PO–4, NO–3, and SO2–4 . However, only the systems using Cl– and Br– wereanalysed with XAS as XAS of elements with Z <17 is particularly difficult toanalyse via XAS due to atmospheric signal attenuation and self absorption[41].Furthermore, as XAS is an element-specific technique, the XAS spectrumresults from signal from all atoms of a particular element. This means thatthe use of I– as both the electron donor and the electron acceptor in thehalogen would result in a spectrum with competing information from boththe electron donor and electron acceptor of the XB. The use of differentelements for the electron donor and electron acceptor reduces the complexityof each spectrum and allows for the separate analysis of the electron donorand acceptor.Prior to the XAS studies, Beer and co-workers characterised these sys-tems by NMR spectroscopy and X-ray crystallography [27]. NMR was usedfor product confirmation and to determine binding constants of halides tothe different systems. X-ray crystallography confirmed both the structureof the systems and the presence of strong XB as the bond distances betweenhalogen substituents and free halides were shown to be 84-86% of the sumof the van der Waals radii of the atoms of interest. My contribution focuseson the use of Cl K-edge and Br K-edge XAS to investigate the nature ofthese bonds.2.2.1 Qualitative XAS AnalysisIf XB were purely electrostatic interactions, one would expect a halide ion(Cl– or Br–) to exhibit K-edge XAS spectra with no pre-edge features.Electric-dipole allowed Cl3p ← Cl1s (or Br4p ← Br1s) transitions would notbe observed since these valence p-states are filled (ns2np6). Covalent delocal-ization of the filled Cl3p orbital with an empty acceptor orbital (e.g. via HBor XB) results in the possibility of a new allowed transition corresponding tocharge transfer from the chloride to its bonding partner (σRX←Cl∗ ← Cl1s).212.2. Using XAS to probe for XB: XAS of a halide electron donorFigure 2.2: Chemical structures of the halide trapping molecules in whichXB covalency was identified by XAS (figures 2.3(a) and 2.3(b)). The com-pounds analysed in these studies contained either Cl or Br as X as indicatedby the naming scheme (e.g. 1-Cl vs. 1-Br). It should be noted 3-X isthe HB analogue of 2-X. The compounds have been drawn in colours thatcorrespond to the XAS data presented throughout the rest of the chapter.Given that the intensity (i.e. area under the peak, section 1.2) of such tran-sitions is directly proportional to the amount of Cl3p in the final state’swave function, σRX←Cl∗, the intensity of any observable pre-edge featureprovides us a direct measure of the charge transfer in a XB.For the complexes containing XB, 1-Cl, 2-Cl and 4-Cl, the near-edgeregions of the XAS spectra exhibit an intense pre-edge feature that is notpresent in ionic chloride salts (figure 2.3(a)). The presence of this intensefeature can only result from charge transfer between the chloride donor and222.2. Using XAS to probe for XB: XAS of a halide electron donorits partners [3, 4]. These results clearly demonstrate charge transfer fromthe donor via a degree of covalency in the XB.For comparison, data was also collected on the HB analogue of 2-Cl (3-Cl). The more intense pre-edge feature of 3-Cl shows that, with a chloridedonor, XB interactions are significantly more covalent in character thancomparable HB interactions. The comparison of these two bonds is relevantas Beer and coworkers showed that the analogous XB and HB of complexes2-Cl and 3-Cl are similar in bond strength despite their different covalencies(see 2.2). These findings are consistent with previous reports that chargetransfer is an important factor in XB bonds [31].The presence of a low-energy shoulder is also evident in the Br K-edgeXAS data, indicating a pre-edge feature similar to that observed in the ClK-edge data. The pre-edge shoulder also appears larger for the X-bondedsystems than that of the H-bonded system (3-Br).Compound # of XB # of HB Ka(m–1) ΔGo of Binding (kJ/mol)2-Cl 2 0 387 ± 20 -15 ± 33-Cl 0 2 206 ± 11 -13 ± 22-Br 2 0 238 ± 12 -14 ± 23-Br 0 2 106 ± 3 -11.6 ± 0.6Table 2.2: Association constants (Ka) of 2-Cl , 3-Cl, 2-Br , and 3-Br asmeasured by Beer and coworkers by WinEQNMR2 [42]. The solvents usedwere CD3Cl and d6DMSO for 2-X and 3-X, respectively. Binding energieswere calculated from association constants measured by Beer and coworkers.Binding energy uncertainties (εΔG) were calculated according to the rules oferror propagation (note: although uncertainties would normally be denotedby σ, ε is used to avoid confusion with σ orbitals) [43].232.2.UsingXAStoprobeforXB:XASofahalideelectrondonor(a) (b)Figure 2.3: (a) Cl K-edge and (b) Br K-edge XAS data for Cl– and Br– trapping molecules, respectively. Inplot (a), the overlay of the four spectra makes the pre-edge feature of 4-Cl difficult to see so the peak shouldercorresponding to the σ∗XB ← Cl1s is indicated by an arrow for clarity.242.2. Using XAS to probe for XB: XAS of a halide electron donor2.2.2 Quantitative XAS AnalysisSixPack [44] and BlueprintXAS [45, 46] were used in order to obtain valuesfor the pre-edge feature intensities. SixPack is used to calibrate and averagespectra [44]. Spectra with high levels of noise and abnormalities due tofaulty channels in the 30-channel germanium detector were removed fromthe dataset. NaCl (E0 = 2820.2 eV) and KBr (E0 = 13474 eV) spectra wereacquired at the same time and used as calibration standards for Cl K-edgeand Br K-edge spectra, respectively. BlueprintXAS version 2.7 was used forbackground subtraction and normalisation. Although SixPack is standardsoftware used to workup XAS data, BlueprintXAS is not as frequently usedand does not have an associated standard protocol. Thus the methodologyused for BlueprintXAS analysis is described below.To minimise the user bias introduced during data work-up, Blueprint-XAS fits the spline, peaks and background concurrently (figures 2.4 and2.5) [45, 46]. While the parameters for each variable are user defined (seebelow), the fits were run in AUTO mode as in this mode, a Monte Carlomethodology is used to choose the starting point of each fit to further removeuser bias. Each fit contained the following components: “piecewise spline +edge”; “peak” for the edge peak; “normalised peak” for the pre-edge peak.These components can be visualised in figures 2.4 and 2.5. In these fig-ures, data has yet to be normalised. Additionally, the “piecewise spline +edge” component has been spilt into the “edge peak” and “background +spline” to aid visualisaton. The final values for each parameter are given inappendices (appendix A).Background subtraction and normalisation are standard methods usedin XAS data to allow for comparison of spectra [44]. Background subtrac-tion involves fitting linear functions to the data before and after the edgejump. A sum of these linear functions is considered the “background” andis subtracted from the data. The spectra in figures 2.4 and 2.5 have al-ready been background subtracted to aid visualisation. Data is considerednormalised when the height difference between the background before andafter the edge jump equals 1. In figure 2.4, the edge jump is the heightdifference between the data below 2820 eV in energy and above 2830 eV(energies are approximate). In the Br K-edge spectra (figure 2.5), the edgejump is the height difference between the data below 13465 eV in energyand above 13510 eV in energy. As the height of the edge jump will dependon the intensity of the incident beam and number of atoms of the elementof interest in the sample, normalisation allows for the comparison of thespectral features of different samples.252.2.UsingXAStoprobeforXB:XASofahalideelectrondonor(a) (b)(c) (d)Figure 2.4: The average of the calculated fits for: (a) 1-Cl, (b) 2-Cl,(c) 3-Cl and (d) 4-Cl after fits had beenfiltered. The ‘acquired data’ plotted is neither normalised nor background subtracted as this is done post-fittingby subtracting the ‘background + spline’ from the ‘acquired data’ and dividing the result by the intensity of theedge jump. Consequently, the intensity of the edge jump for ‘acquired data’ is proportional to the number ofatoms excited and is thus different for each dataset. Finally, the energy of the edge jump for these fits does notcorrespond to the energy of the edge jump in figure 2.3(a) as the calibration of the chloride datasets was doneafter the fitting.262.2.UsingXAStoprobeforXB:XASofahalideelectrondonor(a) (b)(c) (d)Figure 2.5: The average of the calculated fits for: (a) 1-Br, (b) 2-Br,(c) 3-Br and (d) 4-Br after fits had beenfiltered. 50 additional fits were run for the bromide data sets due to a larger number of failed fits (i.e. fit reachedmaximum number of iterations without determining a valid solution). The ‘acquired data’ plotted is neithernormalised nor background subtracted as this is done post-fitting by subtracting the ‘background + spline’ fromthe ‘acquired data’ and dividing the result by the intensity of the edge jump. Consequently, the intensity of theedge jump for ‘acquired data’ is proportional to the number of atoms excited and is thus different for each dataset.272.2. Using XAS to probe for XB: XAS of a halide electron donorFits were filtered, i.e. removed from final set if (a) any error occurredduring fit calculation, (b) the fit significantly deviated from spectra, i.e. sumof square errors (SSE) was too large, and/or (c) if fit is a significant outlierfor peak height, peak energy or E0. The “time average” option, a feature ofBlueprintXAS (used for all fits except 1-Cl and 3-Cl) calculates the averagetime taken to calculate each fit and filters any fits above this value. Thissetting was used to remove failed fits as the fits with the longest calculationtimes usually corresponded with the fits that failed. Failed fits for 1-Cl and3-Cl were removed from the dataset manually (table 2.3).Dataset # Fits Run # Fits Remaining after Filtering1-Cl 100 802-Cl 100 843-Cl 100 464-Cl 100 561-Br 150 1042-Br 150 383-Br 150 704-Br 150 51Table 2.3: Filtering Fits: A total of a 100 and 150 fits were run for the chlo-ride and bromide datasets, respectively. Due to the large number of failedfits for 2-Br and 4-Br, the total number of fits run was increased to obtainedon the order of 50 fits after filtering (recommended by the BlueprintXAS’sdeveloper). To ensure consistency of workup for all Br K-edge data, thetotal number of fits run for all Br compounds was increased. Although fitswere calculated for the Br spectra, the resulting fits were not considered vi-able as explained in the main text. In fact, the significant number of failedfits was the first indication of issues with these fits.As the purpose of fitting the data is to quantify the pre-edge featureintensity for each spectrum, the validity of the fits was assessed by lookingat the coefficient of variation for the pre-edge intensity. As the ratio of thestandard deviation to the mean, the coefficient of variation is a measurementof the precision. The coefficient of variation is lowest for the datasets for 1-Cl, 3-Cl, 4-Cl, and 4-Br (table 2.4), indicating a larger degree of confidencein the calculated pre-edge intensity values for these fits.The fitted pre-edge intensities of 3-Cl and 3-Br exhibit the two highestcoefficients of variation and thus require further inspection into the validityof these fits. A qualitative inspection of the pre-edge feature for 3-Cl reveals282.2. Using XAS to probe for XB: XAS of a halide electron donorDataset μI1 σI1 CVI11-Cl 0.90 0.05 0.062-Cl 1.59 0.04 0.033-Cl 0.13 0.03 0.234-Cl 1.00 0.06 0.061-Br 0.15 0.02 0.132-Br 0.62 0.09 0.153-Br 0.66 0.23 0.354-Br 0.19 0.01 0.05Table 2.4: The average (μ), standard deviation (σ), and the coefficient ofvariation (CV = σ/μ) are reported for the pre-edge feature intensity. Unitsare in normalised intensity for all values except the coefficient of variationwhich is unitless.that it is significantly smaller than the pre-edge features 1-Cl and 2-Clwhich corresponds to the trend observed in the fitted values. Moreover, theseparation of the pre-edge and edge peaks is larger than all other spectraexcept the peak separation in 1-Cl and 2-Cl, which are comparable. Thus,the high coefficient of variation for 3-Cl is likely due to the fact that asthis pre-edge feature has the smallest intensity of pre-edge features in all 8datasets, it will be most affected by experimental noise.The high coefficient of variation is signficantly more concerning as aqualitative inspection of the XAS spectrum of 3-Br reveals that, due to thelarge overlap between the pre-edge and edge peaks, there are very few datapoints which define the shape of the pre-edge feature. This means that asmall change in shape of the edge feature could result in large changes inthe calculated intensity of the unresolved pre-edge intensity feature. Similardegrees of large overlap and lack of peak resolution are a cause for concernin all Br fits. Indeed, 1-Br and 2-Br had the next highest coefficients ofvariation.Due to a combination of the large degree of peak overlap, lack of peakresolution, higher coefficient of variations and large number of failed fitsfor 1-Br, 2-Br, 3-Br, these fits were deemed inconclusive. While the fitfor 4-Br had a low coefficient of variation, indicative of a viable fit, thevalues obtained for this fit are not useful as there is no data to compareit to. Consequently, a quantitive comparison of the pre-edge intensities forthe four bromide data sets was not possible. This result is not unexpectedas lifetime broadening at the higher energy Br K-edge, as well as smaller292.2. Using XAS to probe for XB: XAS of a halide electron donorenergy separation between features, leads to poorer resolution of the pre-edge features of interest. However, it is interesting to note that if the pre-edge intensity value for 3-Br is discarded (the fitted pre-edge intensity valuewith the highest coefficient of variation), we observe a similar trend in pre-edge intensities for both the Cl and Br compounds (2-X >> 4-X > 1-X, i.e.% X electron donation is largest when the halide is involved 2 XB, followedby 2 XB in competition with 2HB, followed by 1 XB, as is expected - thistrend is discussed in more detail later).The chloride fits were deemed sufficiently successful to continue withfurther quantitative analysis. Having obtained intensity values for the pre-edge features from the fits, the degree of covalency in the XB bond canbe quantified, and the degree of charge transfer in different systems canbe compared. This direct comparison of charge transfer in XB and HB inanalogous systems is unprecedented in the literature. Percent charge transferof a metal-chloride coordination bond was determined by Solomon and co-workers [3] and a similar method is employed for this dataset. As discussedin Chapter 1, equations 1.2-1.4, the ratio of the charge-transfer of bond inwhich a chloride acts as as the electron donor metal-chloride bond to thenormalised intensity of pre-edge feature in Cl K-edge XAS correspondingto the D ← Cl– transition. To calculate the percent covalency (α2) for oursystems, we will use the ratio calculated by Solomon and coworkers withCuCl–24 [3].α2CuCl–24ICu←Cl–= 12.5 =α2RI←Cl–IRI←Cl–(2.1)where α2 is the charge transfer (percent covalency) and I is the intensity ofthe Cl K-edge XAS pre-edge feature. The data presented in this thesis isthe first calculation of bond covalency in non-metal systems involving halidesalts.When quantifying the charge transfer of each interaction, the total num-ber of XB and HB interactions in each complex must be considered. In1-Cl, where only a single XB interaction is possible, we find that the degreeof donation is consistent with 6% charge transfer to the iodinated triazoleacceptor. In complex 2-Cl where two XB are present, the total charge do-nation from the halide ion almost doubles to 11%, which implies that thetwo XB in the complex are mostly independent and additive. By contrast,replacement of the iodine acceptors for protons in the (3-Cl) leads to analmost complete loss of intensity in the Cl K-edge XAS pre-edge feature,reflecting very poor charge donation through HB in this system. Data forthe catenane 4-Cl, where both HB and XB are present, indicate that charge302.2. Using XAS to probe for XB: XAS of a halide electron donordonation from the chloride anion decreases substantially (as compared to 2-Cl). This presumably reflects weakened XB due to competition with amideH-bond interactions.The degree of charge transfer observed for these systems by XAS isalso well supported and substantiated by TDDFT simulations for the samesystems. Moreover, the TDDFT simulations indicate that the degree ofcovalency is essentially the same for both Cl– and Br– donors [27].Compound(XAS data)Pre-EdgeEnergy(eV)NormalisedPre-EdgeIntensity% Cl3pchargedonation*InteractionsXB HB1-Cl 1 - 2816.9 ± 0.1 0.90 ± 0.05 6.4 ± 2.12-Cl 2 - 2817.3 ± 0.1 1.59 ± 0.04 11.3 ± 2.03-Cl - 2 2816.2 ± 0.2 0.13 ± 0.03 0.9 ± 1.94-Cl 2 2 2818.4 ± 0.2 1.00 ± 0.06 7.1 ± 2.3Table 2.5: Total Cl3p charge donation as calculated by Blueprint-XAS Analysis. Calculated errors are obtained from statisti-cal distribution of fit results for >50 fits for each data set.*Total donation is calculated in comparison with Cl K-edge XASdata on a CuCl2–4 reference; errors include an estimate of the error in chargetransfer of this reference compound.Notably, the magnitude of charge donation/covalent character in the XBin 1-Cl, 2-Cl, and 4-Cl is similar to that which is commonly observed intransition metal complexes where covalent contributions are considered tobe of great importance in defining chemical properties. For example, insimple divalent metal chlorides ([MCl2–4 ], where M = Cu, Ni, Co, Fe), bondcovalencies have been determined to range from 6% to 9% [3]. To put thesevalues into perspective, it is important to note that in a fully symmetriccovalent bond, the charge donation would be an equal 50% from each atom.Thus both coordinate bonds and XB can be described as having up to onefifth of the maximum possible covalency.To emphasise the comparable degree of covalency between XB and co-ordinate bonds, we propose that XB be depicted with an arrow reflectingthe direction of charge transfer rather than a dotted line. Although XB areusually represented by a dotted line, the choice of the arrow is based on animproved understanding of the electronics of these bonds. The use of arrowsto depict XB is used in this thesis from this point forwards.312.3. XB Stoichiometry by XAS for interactions between KBr and I22.3 XB Stoichiometry by XAS for interactionsbetween KBr and I2To test this new methodology of using XAS to probe XB, Br K-edge XASexperiments were conducted on model systems involving KBr and I2. Thepurpose of these experiments is to determine when XB interactions are ata maximum for the electron donor (in this case Br–). To do this the ratioof KBr:I2 was increased in each successive experiment until the pre-edgefeature corresponding to the σ∗XB ← Br1s no longer increased.Figure 2.6: Br K-edge XAS data of KBr and I2 in various ratios. Sampleswere prepared by undergraduate mentee, Darren Yu.There are two possible phenomena which explain the lack of increase inpre-edge feature intensity from the KBr:I2 1:1.3 spectrum to the KBr:I2 1:3spectrum. This data could indicate that for these systems, each Br– can beinvolved in XB with a maximum of 1.3 I2 molecules on average. In otherwords, when the Br– is involved in two XB with two different I2 molecules,322.4. Using XAS to probe for XB: XAS of the XB electron acceptoreach XB exhibits less charge transfer than a Br– involved in a single XB withan I2. Alternatively, the Br– could be involved in more than two significantlyweaker XB. Either way, we know that the charge transfer from Br– to I2 ismaximised at a ratio of approximately KBr:I2 1:1.3. This experiment bothprovides a satisfying proof of concept for our methodology, and it illustratesthat this methodology can provide stoichiometry information about XB too.We can also see that the XB between Br– and I2 exhibits more covalencyand a greater orbital overlap between the Br4p and the σ∗I–I than the XBinvolving Br– in the previous sections. This can be seen in the larger pre-edge feature intensity.An interesting though more difficult experiment would be to run a similarexperiment at an iodine edge, slowly increasing the amount of KBr until nomore changes are observed in the pre-edge feature of the I L3-edge data.The feasibility of studying XB using XAS of the XB electron acceptor willbe explored in the remaining sections of this chapter.2.4 Using XAS to probe for XB: XAS of the XBelectron acceptorIn section 2.2, XAS was established as a method to probe the degree ofcovalency of an XB, specifically for XB involving a Cl– or Br– as the elec-tron donor. While the purpose of these XAS experiments was to probe theelectronic properties of XB, XAS can also be able to probe for the presenceXB in any systems involving a halide electron donor (this will be exploredin chapters 3 and 4).The ability to use XAS in addition to X-ray crystallography to probe forXB allows us to study a wide range of systems for which obtaining singlecrystals would be difficult if not impossible. However, the main drawback ofthis XAS diagnostic approach is that it can only be used to study XB whichhave a halide ion acting as the XB electron donor. In XB involving otherelectron donors such as covalently-bound oxygen, nitrogen and sulphur, XASof the electron donor is not possible. This is because XAS of low-Z elements(Z<17, i.e. elements with atomic number less than 17) is much more difficultdue to atmospheric signal attenuation and self absorption [41].In this section, we explore the use of XAS to probe XB via the covalently-bound halogen that acts as the XB electron acceptor. If successful, thistechnique could be generalisable to all XB systems as, by definition, all XBinvolve a covalently-bound halogen.To explore this approach, the halide trapping complexes discussed in332.4. Using XAS to probe for XB: XAS of the XB electron acceptorsection 2.2 (1-X, 2-X, and 4-X) are analysed by XAS. As all compoundsinvolve an iodine acting as the XB electron acceptor only XAS experimentsat I edges were necessary. Both I L3-edges and I L1-edges were studied(excitation of the I2s and I2p electrons, respectively), but only the I L3-edgesare reported here. The non-linear nature of the post-edge region of our I L1-edge data rendered normalisation of the spectra difficult and unambiguousintercomparison of the spectra impossible. Consequently, all I L1-edge datawas deemed inconclusive.XAS data of the electron acceptor for 3-X is not included in this section.As 3-X involves HB rather than XB, it is not possible to conduct XASexperiments on the H electron acceptor.I L3-edge XAS experiments for the KBr:I2 systems discussed in the pre-vious section is also not reported. This is due to the fact that experimentsshowed that I L3-edge XAS data for I2 is significantly affected by slightchanges in sample preparation method (i.e. more/less grinding with a mor-tar and pestle). This is most likely due to the many different possible polyio-dides which can form [47]. The changes resulting from a slight change insample prep resulted in far more significant changes in the edge region thanwould likely arise from XB.However, XAS data of the synthetic precursors of 1-X, 2-X, and 4-X, (1-PF6, 2-PF6, and 4-PF6, respectively) in which PF6 acts as thecounterion is included in this section. These systems do have a covalently-bound iodine but do not involve XB and therefore allow us to compare theXAS of iodine substituents in analogous compounds in the presence andabsence of XB. The lack of XB in these compounds can be assumed as thesteric bulk of PF–6 suggests that this counterion is unable to fit into therelatively small binding pockets of the various complexes. All compoundswere synthesised by the Beer group [27].To facilitate data interpretation: all XAS data of 1-X (1 XB present)and 1-PF6 is blue, all XAS data of 2-X (2 XB present) and 2-PF6 is green,and all XAS data of 4-X (2 XB and 2 HB present) and 1-PF6 is red. Thiscolour scheme is the same as figure 2.3 which shows the Br K-edge and ClK-edge XAS data for these same systems.Probing XB through the XB electron acceptor instead of a Cl– or Br–acting as a XB electron donor is significantly more challenging. In the XASspectrum of a halide involved in an XB, we observe a new peak correspondingto the σ∗XB ← Cl– or σ∗XB ← Cl– to indicate XB formation. In the XASspectrum of a covalently-bound halogen, there are no new transitions allowedupon XB formation. Therefore, we expect changes due to XB in the XASspectrum of the electron acceptor to manifest as changes in peak intensity342.4. Using XAS to probe for XB: XAS of the XB electron acceptorand shifts in peak energy. The electron transition which is expected to bemost affected in the systems studied is the σ∗C–I ← I2p (figure 2.7).Figure 2.7: The molecular orbital diagram of a covalently-bound iodineinvolved in an XB. The transitions of interest arising from an I L3-edgeXAS experiment for the complex in the absence and presence of XB withCl– are shown in purple and green, respectively.Upon partially covalent XB formation the antibonding orbital of the C-Ibond (σ∗C–I) overlaps with the Cl3p (in the case of Cl– acting as the XBelectron donor) to form two new orbitals: σXB and σ∗XB (figure 2.7). Theelectron transition which is lowest in energy is therefore the σ∗C–I ← I2p inthe absence of XB and σ∗XB ← I2p in the presence of XB. The energy of theσ∗XB ← I2p is expected to be higher than σ∗C–I ← I2p, due to the stabilisationof the σXB orbital, and the probability of the σ∗XB ← I2p would expected tobe lower than σ∗C–I ← I2p due to the decreased iodine character in the σ∗XBorbital. This will be discussed further in sections 2.4.1 and 2.4.2.2.4.1 I L3-edge XAS: Comparing different systems with thesame counterionFigure 2.8 plots the I L3 edge spectra for the 9 compounds of interest.These 9 spectra are grouped in order to visualise the changes between the352.4. Using XAS to probe for XB: XAS of the XB electron acceptorspectra of three different iodine XB electron acceptor compounds when thesecompounds are involved in XB with the same halide.An inspection of the XAS spectra suggests that the feature at approx-imately 4560 eV corresponds to the σ∗C–I ← I2p or σ∗XB ← I2p transitions(σ∗C–I ← I2p for systems with no XB and σ∗XB ← I2p for systems with XB).The relatively small size of the pre-edge feature when compared with theedge is not unexpected as the relative electronegativities of iodine and car-bon suggests that σ∗C–I orbital will be localised on the C atom. Furthermore,this transition is only dipole allowed due to the presence of I5s character inthe σ∗C–I orbital.Although all of these systems involve iodine substituents, the direct com-parison of systems which such widely different scaffolding may seem am-bitious. However, as the pre-edge and edge regions of XAS are affectedpredominantly by the electronics of the atom studied and its immediatebonding partners, a comparison between these systems which all contain aniodine bound to an sp2 hybridised carbon is reasonable, particularly if theidentity of the halogen bond electron acceptor remains the same. Indeed,the pre-edge and edge features are nearly identical for the three differentsystems (1-PF6, 2-PF6, and 4-PF6) (figure 2.8(c)). However, when theiodine substituents of these three systems are involved in different degreesof XB (figures 2.8(a) and 2.8(b)), the pre-edge features change. Specifi-cally, the intensity of the pre-edge feature is arguably largest for the 4-Xcompounds, very closely followed by the 2-X compounds with the pre-edgefeatures of the 1-X compounds having the smallest intensity. The energyof the pre-edge features of the 1-X compounds are also noticeably higher(1-2 eV) than the energies of the corresponding features in the 2-X and 4-Xspectra.To understand these changes, it is useful to think about the degree ofcovalency for each system, and thus we return to the results from section 2.2.There are three ways to count the number of XB in each system: the numberof XB in which Cl– participates, the number of XB in which each iodine sub-stituent participates and the total number of XB in each system (table 2.6).1-X 2-X 3-X 4-XTotal Bonds 1 XB 2 XB 2HB 2 XB & 2 HBBonds per I 1 XB 1 XB – 1 XBBonds per X– 1 XB 2 XB 2HB 2 XB & 2HBTable 2.6: The number of intermolecular bonds per system/per iodine sub-stituent or per chloride ion for 1-X, 2-X, 3-X, and 4-X.362.4. Using XAS to probe for XB: XAS of the XB electron acceptorIf we take the total %Cl3p contribution in each σ∗XB (table 2.5) and divideby the number of XB in each system in which Cl– participates (table 2.6),we calculate the degree of %Cl3p contribution per XB as follows: 6.4% for1-Cl, 5.7% for 2-Cl , and 3.6% for 4-Cl . These values tell us that XBin 1-Cl are the most covalent while the XB in the 4-Cl are least covalent.As a single Cl– ion serves as the sole electron donor for each system, thecompeting bonds in 2-Cl and 4-Cl decreases the %Cl3p contribution perXB.1-Cl 4560.1 eV 1-Br 4560.1 eV 1-PF6 4559.4 eV2-Cl 4559.5 eV 2-Br 4559.7 eV 2-PF6 4559.5 eV4-Cl 4559.5 eV 4-Br 4559.3 eV 4-PF6 4559.2 eVTable 2.7: The first maximum of the first derivative of the I L3-edge XAS forall compounds in figures 2.8 and 2.9. These values can be used to approx-imate the location of the centre of the pre-edge feature for each spectrum.The pre-edge features of 1-Cl and 1-Br are noticeably higher in eV thanall other pre-edge features. This is expected as these two systems shouldexhibit the most XB orbital overlap per iodine as the halide acting as theXB electron donor is only involved in one XB (table 2.6). The resolution ofthese experiments is 0.2 eV.With this information in mind, we return to understanding XB fromthe perspective of the iodine electron acceptor. As the XB in these systemsform via the overlap of the Cl3p and σ∗C–I orbitals, we expect the trend in% contribution of the σ∗C–I to the σ∗XB to be inverse of %Cl3p contribution.Thus, 4-Cl is expected to have the highest % contribution of the σ∗C–I,followed by 2-Cl, and then 1-Cl. As the probability of the transition σ∗XB ←I2p transition is directly proportional to the amount of iodine character ofthe σ∗XB orbital (which is determined by the % contribution of the σ∗C–I ), weexpect the trend in probability of the σ∗XB ← I2p transition to be as follows:4-Cl>2-Cl>1-Cl. Finally as the intensity of the feature corresponding tothis transition is proportional to the probability, we expect the same trendin the intensities of the pre-edge features in the XAS spectra. Indeed, avisual approximation of the intensities of pre-edge features for figure 2.8(a)involving the Cl– donor and figure 2.8(b) involving the Br– donor seems toagree with this expected trend. Note: while the pre-edge feature of 2-Clappears to be to the left of 1-Cl, the actual location of the pre-edge featureas determined by where the second derivative is zero is approximately thesame (table 2.7).Furthermore, a similar logic can explain the observed shift in the pre-372.4. Using XAS to probe for XB: XAS of the XB electron acceptoredge features for both figure 2.8(a) and figure 2.8(b). As orbital overlapimproves in the XB, i.e. higher %Cl3p contribution in the σ∗XB and higher% contribution of the σ∗C–I to the σXB, the energy of the σ∗XB is expected toincrease as the energy of the σXB decreases. Thus as % contribution of theσ∗C–I to the σ∗XB (seen as the decrease in intensity of the σ∗XB ← I2p feature),the energy of this feature is expected to increase (shift right). Again thisapproximately corresponds with what we observe for both the XAS datain figure 2.8(a) involving the Cl– donor and figure 2.8(b) involving the Br–donor.In conclusion, I L3-edge XAS of the XB electron acceptor can be used toobserve changes in XB bond covalency, i.e. the overlap between donor andacceptor orbital, for as overlap increases, the pre-edge feature correspond-ing to the σ∗XB ← I2p will decrease in intensity and increase in energy aspredicted by the molecular orbital diagram.382.4.UsingXAStoprobeforXB:XASoftheXBelectronacceptor(a) (b)(c)Figure 2.8: I L3-edge XAS: Comparing different systems with the same counterion as follows (a) Cl– (b)Br– (c)PF–6392.4. Using XAS to probe for XB: XAS of the XB electron acceptor2.4.2 I L3-edge XAS: Comparing analogous systems withdifferent counterionsIn this section, we will compare the same 9 compounds as the previoussection. Figure 2.9 includes the same 9 XAS spectra as figure 2.8. However,these 9 spectra are grouped differently in order to visualise the changesbetween the spectra of each iodine XB electron acceptor compounds as thecounterion is changed.In comparing these spectra, we observe similar trends to those discussedin section 2.4.1. The systems with PF–6 counterions exhibit the highestpre-edge feature intensity indicative of the least (or in this case absenceof) XB. The difference in intensities of the pre-edge features for the Br–and Cl– analogues is difficult to analyse qualitatively as all features are verysmall. However, we can say that for at least 1-X, the Br– and Cl– analoguesappear to exhibit nearly identical pre-edge feature intensities. Interestingly,this agrees with the TDDFT calculations reported for these compounds [27],as the calculations suggest that a similar degree of XB covalency is expectedfor the Cl– and Br– analogues for all systems.Finally, there is a noticeable trend in the intensity of the edge peak forall three acceptor systems (specifically edge peak intensity is largest for thePF–6 analogues > Br– > Cl–). This change is worth mentioning as it isconsistent for the 3 different series (1-X, 2-X, 4-X). Due to their proximityin energy, both I6s ← I2p and I5d ← I2p transitions could contribute to theintensity of this peak. Thus, the decrease in intensity of this feature suggeststhe involvement of either or both of these orbitals in XB as orbital overlapwith an electron donor would result in a decrease in % iodine character,and thus a decrease in the probability of these transitions and decrease inintensity of the associated feature in the XAS spectrum. It also suggeststhat the involvement of these orbitals would be most significant for the Cl–systems. As other physical phenomena can contribute to the intensity ofthe edge feature including scattering off nearby atoms, this is merely one ofmany possible explanations.402.4.UsingXAStoprobeforXB:XASoftheXBelectronacceptor(a) (b)(c)Figure 2.9: I L3-edge XAS: Comparing analogous systems with different counterions (a) 1-Cl, 1-Br, and 1PF6;(b) 2-Cl, 2-Br, and 2PF6; (c) 4-Cl, 4-Br, and 4PF6412.5. Conclusions2.5 ConclusionsBeer and coworkers demonstrated the importance of XB in halide recog-nition systems as 2-X display a marked enhancement in anion recognitionover 3-X, their HB analogues. Cl and Br K-edge XAS revealed the pres-ence of intense pre-edge features characteristic of charge transfer betweenthe halide donor and the XB acceptor. Quantitative fitting of the pre-edgefeatures in the Cl K-edge XAS data provided a direct measure of the degreeof covalency in the XB interaction, which is comparable to that observedin coordinate bonds transition metal complexes. Furthermore, XAS datashowed that perpendicular XB interactions in the Cl– donor systems areindependent and additive, but that the degree of XB covalency can be mit-igated through the presence of HB donors. I L3-edge XAS studies of theiodine substituents showed that as XB covalency increases, the peak corre-sponding to the σ∗XB ← I2p transition decreases in intensity and increasesin energy, as predicted by molecular orbital theory.Most importantly, these results offer (1) the first experimental evidenceof covalency in XB, revealing that although XB are comparable to HB interms of bond energy, they are comparable to coordinate bonds in terms ofcovalency and (2) a methodology to study XB in a wide variety of systemswithout the requirement for single crystals.2.6 Methods2.6.1 XAS Sample PreparationSamples were received from Beer and coworkers at Oxford and used as re-ceived. For Cl K-edge and Br K-edge XAS analysis, samples were mountedas a finely ground powder diluted (1:1) with boron nitride (BN) dusted onsulfur-free Kapton tape across the window of an Al plate. The resultanthomogeneous, finely dispersed powders were pressed into a 0.5 mm thickaluminium spacer, sealed on both sides with Kapton tape.2.6.2 XAS Data CollectionBr K-edge XAS data was collected at SSRL on beamline 7-3 under ringconditions 80-100 mA at 3.0 GeV. This beamline has a 20-pole, 2 T wigglers,0.8 mrad beam, and a Si (220) double crystal monochromator that wasdetuned by 50% intensity to attain harmonic rejection. The incident X-ray intensity (I0), sample absorption (I1), and Br reference absorption (I2)422.6. Methodswere measured as transmittance using argon-filled ionization chambers. Sixto eight sweeps were taken for each sample, and all data were measuredat 19 ± 7 K within an Oxford Instruments CF1208 continuous-flow liquidhelium cryostat. KBr calibration scans were performed concurrent with eachsample.Cl K-edge and I L3-edge XAS data was collected at beamline 4-3 SSRLusing a modified low Z setup allowing for low temperature data acquisitionunder ring conditions of 3 GeV and 60-100 mA. The setup is a 54-pole wigglerbeamline operating in high field (10 kG) mode with a Ni coated harmonicrejection mirror and a fully tuned Si (111) double crystal monochroma-tor. Signal was detected with a N2 fluorescence (Lytle) detector at ambienttemperature and pressure (1 atm, 298 K). Calibration scans for the Cl K-edge data were performed before and after every data set to ensure stablemonochromator readings.The I L3-edge XAS data in section 2.4 was not calibrated due to failedcalibration scans. However, this does not pose a problem for the intercom-parison of this data as the overall stability of the beam during the beamtimein question was very good, i.e. the calibration energy did not drift duringthe Cl K-edge experiments performed on the same beamline before and afterthe I L3-edge experiments. Thus, while the overall energies of the I L3-edgeXAS data is only approximate, this data can still be accurately compared toother data collected during the same run. As it is standard to only compareXAS data collected at the same beamline during the same beamtime, theoverall stability of the beam means the inability to calibrate each experimentindividually has no effect on conclusions drawn from comparing this data.432.6. MethodsXB nature is often debated,But all of us can now be sated,For halogen bonds,Have shared electronsAs XAS data has stated.44Chapter 3Using XAS to Probe XB inDye-Sensitised Solar Cells∗In this chapter, the role of XB in dye-sensitised solar cells (DSSC) is ex-plored. The DSSC studied involve a photosensitive dye with halogen sub-stituents (Dye-X series) (figure 3.1). As we have showed in chapter 2 thatXAS can be used to characterise XB, particularly XB involving halide elec-tron donors, XAS will be used to probe for the existence of XB in DSSCbetween the dyes and halide electrolytes. Previous work provided computa-tional and indirect experimental evidence that XB may play an importantrole in these systems [39, 48]. The work for this chapter was a collaborativeeffort and completed as follows: XAS data was collected by the author andFraser Parlane. XAS data analysis was done by the author. Photosensi-tive dyes were synthesised by Cameron Kellett. DSSC were built by FraserParlane.3.1 Dye-Sensitised Solar Cells - BackgroundIn avoiding the use of silicon and instead using dyes to enhance the photo-sensitivity of the semiconductor, DSSC are emerging as a cheaper alternativeto silicon based solar cells and a more efficient alternative to other thin filmsolar cells [49]. These systems involve photosensitive dye molecules whichare adsorbed onto a titanium dioxide (TiO2) layer (figure 3.1). Excitationof electrons in the dyes creates a driving force for electron transfer from thedye to the TiO2 layer, and from there to the cathode via connected trans-parent conducting oxide surfaces. Regeneration of the oxidised dye occursvia a redox process involving an electrolyte [49]. One method of improvingthe efficiency of DSSC is to improve the dye regeneration rate.∗The work in sections 3.1 and 3.2 is based on the author’s contributions to previouslypublished work [36]. The breakdown of work for these sections is described in the text.Section 3.3 has not been published previously and comprises of additional analysis of thedyes in section 3.2. All data collection and analysis for this section was done by the author.See Preface for more details.453.1. Dye-Sensitised Solar Cells - Backgroundregenerationrecombinationback electrontransferphotoexcitationinjectionOO CNSNTiO2XI-= F, Cl, Br, IhνXσ-hole12355434I3-e-e--0.35 V0.79 V0.29 VE E(I3– / I2•–, I–)E(Dye-X•+ / Dye-X*)E(Dye-X•+ / Dye-X)TiO2 Dye ElectrolyteE(I2•– / 2I–)Dye-FDye-ClDye-IDye-Br1.22V1.27V1.28Vkreg2 kinjkCRkBET E(I3– / I–)VOCB.A.EFFigure 3.1: Schematic of a DSSC synthesised and built by the Berlinguettelab. Figure used with coauthors’ permissions ( c©Creative Commons) [36].Berlinguette and coworkers have explored the role of halogenated dyeswith an I3/I– electrolyte in DSSC. These dyes have been designed so thatthere is a physical separation between the HOMO and LUMO of the dye.This physical separation allows for better charge separation, thus minimisingcharge recombination upon excitation. The LUMO is located on the sectionof the dye closest to surface while the HOMO is located farther from thesurface. Thus electron excitation moves the electron physically closer to thetitanium dioxide surface and away from the electrolyte.Previous work from Berlinguette and coworkers noted that the regen-eration rate of the dye changed as the identity of the halogen substituenton the dye changed. Specifically, the dye regeneration rate increased as thepolarisability of the halogen increased (from F to Cl to Br to I) (figure 3.2,orange line). Differences in the thermodynamic properties of these dyes wasunable to explain this trend. Indeed, when the electrolyte was changed fromI–/I–3 to a cobalt complex (figure 3.2, blue line), the increase in regenerationrate tracked with ΔGrxn instead [36, 39, 48].As all other structural and electronic components of the synthesised dyeswere the same, the presence of XB between dye and electrolyte was proposedto explain the difference in the regeneration rates of the Dye-X series [39].As XB strength between the electrolyte and the halogenated dye series isexpected to increase as the size and polarity of the halogen substituent in-creases, XB between dye and electrolyte could explain the dependency of463.2. XAS of DSSCFigure 3.2: Regeneration rate constants (kreg) for the reaction of Dye-X+. on TiO2 substrates with 0.5-10mM of I– (orange) or 10-150mM of[Co(bpy)3]2+ (blue). Figure used with coauthors’ permissions ( c©CreativeCommons) [36, 39, 48].regeneration rate on halogen substituent size, i.e. as the size of the σ-holeon the halogen substituent increases (figure 3.3), the strength of the XB in-teraction increases which results in faster regeneration rates. This rationalewas supported with DFT calculations showing that the interatomic distancesbetween the I– electrolyte and Dye-X were smaller than the sum of the vander Waals radii for Dye-Br and Dye-I, an indication of XB. Furthermore,DFT calculations predicted that oxidation of the Dye-X series increases thecharge donation of the I– to the σ∗C–X of the Dye-X, another indication ofXB [48].3.2 XAS of DSSCTo probe for the existence of XB between dye and electrolyte, we set outto use XAS as a tool to directly measure the electronic coupling (or orbitaloverlap) between the dye and an electrolyte species. Although the DSSCinvolve suspected XB formation between an I– and the halogen substituentsof the dye, experiments in this chapter focus on XB formation between Cl–and the halogen substituents of the dye. Cl– is an effective surrogate for the473.2. XAS of DSSCFigure 3.3: Increasing σ-hole on Dye-X series. DFT models of the singly-oxidised dyes, Dye-X+. (where X is F, Cl, Br, and I), reveal an increasinglyelectropositive σ-hole on the terminus of the halogen substituents as the sizeof the halogen increases. The electrostatic potential is plotted over a spherecorresponding to the van der Waals radius of the respective halogen sub-stituent. Figure used with coauthors’ permissions ( c©Creative Commons)[36].I– species used in the DSSC because the nucleophilicity of Cl– is comparablewith I– [50]. Also, as the reduction of the photosensitive dyes by Cl– isless favourable than by I– (equations 3.1-3.3) [50, 51], longer timescales areavailable for measuring XB interactions between the oxidised dyes and Cl–than the analogous experiment with I–.2Cl– → Cl2 + 2e– Eo = –1.36 V (3.1)2I– → I2(s) + 2e– Eo = –0.54 V (3.2)3I– → I–3 + 2e– Eo = –0.53 V (3.3)Equations 3.1-3.3: Oxidation potentials of Cl– and I–. The lower Eovalue for the Cl– oxidation reaction illustrates that this oxidation is lessfavourable than the oxidation of I– [52].As shown in chapter 2, Cl K-edge XAS can be used to quantify the degreeof covalency in XB donor-acceptor systems, specifically for XB involving a483.2. XAS of DSSCFigure 3.4: Relevant electron transitions induced by Cl K-edge XAS if XBpresent between Dye-X and Cl–. Figure used with coauthors’ permissions( c©Creative Commons) [36].Cl– electron donor. Charge transfer from the Cl– to a XB electron acceptor isdetected by a lower energy, pre-edge feature in the Cl K-edge XAS spectrumthat arises from charge depletion of the filled Cl3p orbital (chapter 2.2). Theintegrated intensity of this pre-edge feature is directly proportional to thedegree of mixing between the Cl3p orbital and the electron-accepting orbital.Cl K-edge XAS is used to probe for the presence of XB in these systemsusing the method established in chapter 2. Specifically, a pre-edge featurewith a distinctively lower energy than the Cl4p ← Cl1s transition most likelyresults from an electron transition from the Cl1s orbital to the antibondingorbital of a partially covalent XB (σ∗RX←Cl ← Cl1s), as seen in figure 3.4.Thus, if present, a pre-edge feature in a Cl K-edge spectrum of Cl– with thedyes likely indicates the presence of XB.The XAS spectra recorded for Dye-I on TiO2 does not show a pre-edgefeature (figure 3.5; black line) due to either poor orbital overlap or the ab-sence of XB. Both possibilities lead to the same interpretation that minimalXB occurs between Dye-I and the Cl–. A poor interaction between Cl3p andthe σ∗RX may be due to the σ∗RX being significantly higher in energy thanthe Cl3p electron donor orbital, or the σ∗RX being localised more towards the493.2. XAS of DSSCcarbon rather than the halogen substituent. Previous XB experiments haveshown the importance of placing strong electron-withdrawing substituentsaround the XB donor to realize the σ-hole, particularly when in solution[53]. It is therefore not surprising that Dye-I would not participate in a XBsignificantly enough to result in a detectable XAS signal due to their lackof electron-withdrawing units in proximity to the halogen.To determine if XB formation occurs with the photoexcited dyes, theXAS experiment was repeated using dyes which had been oxidised usingNOBF4 to obtain a radical species (Dye-X+.) that mimic the effects of pho-toexcitation. This system is worth investigation as the increased electrondeficiency in the photoexcited/oxidised dyes could facilitate XB with theelectrolyte. Unlike the Dye-I + Cl– system, The Dye-X+. + Cl– series doesexhibit this feature at approximately 2824 eV (figure 3.5; red line). Uponoxidation of the dye, the σ∗RX is lowered in energy and the increase in po-larization of the carbon-halogen bond causes the σ∗RX to be more localizedtowards the halogen. This scenario is more compatible with XB formation,resulting in a new dipole-allowed σ∗RX←Cl ← Cl1s transition. Note that ClK-edge XAS data is not presented for Dye-Cl as this data would yield acombination of features from both the electrolyte and the dye that rendersanalysis more difficult.503.2. XAS of DSSCFigure 3.5: Cl K-edge XAS data of Dye-I in the ground (Dye-I) and oxidised(Dye-I+.) states in the presence of Cl–. The presence of the pre-edge featureat 2825 eV in the Dye-I+. system is indicative of XB formation. Likewise,the absence of a pre-edge feature in the Dye-I system is indicative of no XBformation.513.2.XASofDSSC(a) (b)(c) (d)Figure 3.6: The average calculated fits for the Cl K-edge XAS data of (a) NOBF4+Cl–, (b)Dye-F+. ← Cl–,(c)Dye-Br+. ← Cl– and (d)Dye-I+. ← Cl–.523.2. XAS of DSSCThe pre-edge features in the XAS spectra of the oxidised dyes were quan-tified using the same method described in chapter 2 (figure 3.6 for fittingfunctions and table 3.2 for values). Two distinct mechanisms are responsiblefor the pre-edge features that arise upon oxidation of the surface-adsorbeddyes. Firstly, Cl– may form XB with the oxidised dye (Dye-X+. ← Cl–).Secondly, the oxidation of the TiO2 surface (with NOBF4) generates defectsites that result in some covalent binding of halide ions to the surface asobserved in the control experiment. Notably, there also exists a pre-edgefeature in the XAS spectrum of NOBF4 + Cl–. We also observe two pre-edge features in the Cl K-edge spectrum for (Dye-Br+. ← Cl–), a smallerfeature at higher energy most probably corresponding to Cl– interactionswith TiO2 defect sites and NO+, and a larger feature at lower energy thatonly occurs in the presence of the oxidised dye.We can thus estimate the minimum pre-edge intensity arising from XBinteractions, and thus minimum %Cl3p contribution in the XB, by sub-tracting contributions from Cl– interactions not involving the dyes from theintensity of the observed pre-edge features (table 3.2). The pre-edge inten-sity observed in the absence of dyes (NO+ and TiO2 with Cl–) allows us toestimate charge transfer arising from Cl– interactions not directly involvingdyes. The use of this spectrum as a baseline is justified by the fact that,upon addition of dyes to the system, interactions of Cl– with either thedefect sites or NOBF4 will likely decrease due to the competing XB interac-tions. Using this method of estimation, we note that the pre-edge intensityfrom XB interactions is essentially zero in Dye-F+., and quite significant inboth Dye-Br+. and Dye-I+.. This experimental trend matches the behaviourobserved in our computational modelling.The XAS data presented in this study is evidence of an interfacial XBbetween Cl– and Dye-Br+. and Dye-I+.. At first glance, these results seemto contradict the trend observed in figure 3.2 as the XAS data indicateshigher charge transfer between Cl– and Dye-Br+. compared with that be-tween Cl– and Dye-I+. (table 3.2). As Cl– is significantly smaller than I–,it is reasonable to expect greater orbital overlap between the bromine sub-stituents on Dye-Br and Cl– than Dye-I and Cl–. In the DSSC where I– isused, greater I– orbital overlap should exist with Dye-I+. rather than withDye-Br+. resulting in the observed trend in kreg values (figure 3.2).In conclusion, XAS was used to directly confirm an interfacial XB inter-action between a soluble halide species and immobilized dyes bearing dif-ferent halogen substituents. This work confirms that these intermolecularinteractions only exist when the dyes bear polarizable halogen substituentslike bromine and iodine, as opposed to non-polarizable substituents like fluo-533.2. XAS of DSSCXAS Normalised XAS pre-edge PercentExperiment XAS pre-edge feature contribution(compound) feature energy ( eV) of Cl3p in σ∗XBNO++Cl– 0.19 ± 0.04% 2818.6 Not applicableDye-I+Cl– No contribution No contribution No contributionDye-F+.+Cl– 0.27 ± 0.07% 2818.6 1.0-3.5%Dye-Br+.+Cl– 0.59 ± 0.03% 2818.3 5.2-7.7%Dye-I+.+Cl– 0.43 ± 0.02% 2818.5 3.2-5.6%Table 3.1: Experimental contribution of Cl3p in σ∗XB and relevant informa-tion as determined by analysis of pre-edge feature in XAS data for the Dye-Xseries. Ranges for percent contribution of the Cl3p orbital to the σ∗XB wasdetermined as follows: the maximum value was calculated assuming 100% ofthe pre-edge feature intensity arose from the XB interactions and the mini-mum value was calculated assuming the same contribution from backgroundinteractions as observed for NOBF4 + Cl– sample. For example: minimum%Cl3p contribution for Dye-F+.+Cl– is equal to pre-edge feature intensity(0.27) - NOBF4 pre-edge feature intensity (0.19) multiplied by %Cl3p contri-bution in CuCl2–4 (9.8) then divided by pre-edge feature intensity for CuCl2–4(0.75) [3, 4].rine, as expected from XB theory [23, 54]. This work also demonstrates thatthese XB interactions occur after electron injection into the TiO2, and thuscan be an important factor only during the short lifetime of the oxidised dye.Furthermore, computational methods show that these XB interactions arenot significantly affected by the identity of the nucleophilic halide species[36], and therefore we believe that the results presented here with Cl– canbe extended to the dye-iodide interactions that occur within DSSCs. Assuch, the previously observed increases in dye regeneration rate by I– [55]can be attributed to the increased XB donating ability of Dye-Br+. andDye-I+.. It is particularly striking that such a weak intermolecular interac-tion involving a distinctively minority species can have such a huge effecton the measured photovoltage, and it suggests that encouraging other in-termolecular interactions between oxidised dye and electrolyte species mayhave similar beneficial effects on dye performance.543.3. Solution XAS of Dye-Br3.3 Solution XAS of Dye-BrDye-Br in solution with and without Cl– was further studied by XAS todetermine if the presence of XB corresponded with changes in the XASspectrum of the XB electron acceptor (Br K-edge XAS). These XAS exper-iments are analogous to the I L3-edge experiments in section 2.4. Similarto the I L3-edge experiments, the changes observed in this XAS data areexpected to be small as this data reflects changes in an electron transition(figure 3.7) rather than the appearance of a new allowed transition (as is thecase for XAS analysis of XB through the electron donor). As can be seenin the molecular orbital diagram, the σ∗XB ← Br1s is expected to be higherin energy with less Br character. This change corresponds to a decrease inpre-edge peak intensity and a shift to higher energy for this feature.Figure 3.7: The molecular orbital diagram of a covalently-bound bromine inDye-Br+. involved in an XB with Cl–. The transitions of interest in Br K-edge XAS are shown in orange and green for ground state Dye-Br (absenceof XB) and oxidised DyeBr (presence of XB), respectively.The combination of Dye-Br and Cl– was chosen for these experiments asthis pairing exhibited the highest degree of overlap as shown by the sampleyielding the most intense pre-edge feature (figure 3.6(c)). Furthermore, thedifferent XAS instrumentation used for each experiment yields data with less553.3. Solution XAS of Dye-Brnoise for the Br K-edge. This is due to a larger number of detector channels(30) available for Br K-edge experiments. By comparison, the instrumenta-tion used for the I L3-edge only has 1-4 detector channels (1 channel for theLytle fluorescence yield detector and 4 channels for the Vortex fluorescenceyield detector).As the aim of these experiments is to elucidate the effects of XB presentin solution, the adsorbance of the dye to TiO2 was deemed unnecessary.Thus, due to instrument constraints for the Br K-edge XAS, the sampleswere not prepared on a mesoporous TiO2 film deposited in glass (as insection 3.2). Instead, all samples analysed are acetonitrile solutions. As thesample prep was changed from the experiments in section 3.2, XAS datawas also collected for the XB electron donor (Cl K-edge XAS) to confirmthat XB are present in these systems.The effects on XB formation from chemical oxidation of Dye-Br byNOBF4 is also explored. This experiment is challenging as the oxidised dyeis only briefly present in solution (the dark green colour change observedupon addition of saturated NOBF4 in acetonitrile lasted for approximately2 seconds before the solution returned to the initial orange colour). Thesolutions were flash frozen immediately following the addition of NOBF4.However, green and orange areas were both visible in the frozen solutionsindicating that both ground state and oxidised were present. Therefore, forall solution XAS data reported in this section for oxidised Dye-Br (Dye-Br+.), it is important to note that Dye-Br is also present in these samplesin unknown ratios.3.3.1 Solution XAS of Dye-Br: Br K-edgeBr K-edge XAS data of Dye-Br with and without Cl– shows no changefor the ground state dye (figure 3.8(a)). This is expected as the resultsdiscussed earlier suggested that there was no XB formation between Cl– andthe ground state dyes. Dye-Br and the oxidised dye, Dye-Br+., show littledifference indicating no significant charge transfer between the oxidised dyeand any unused oxidant NOBF4 or side products. As there is little differencebetween the edge energies for Dye-Br and Dye-Br+., it can also be assumedthat there is not significant charge delocalisation onto the Br substituent asto change the oxidation state of the Br.By contrast, the addition of NBu4Cl to the oxidised dye results in adecrease in intensity of the peak at approximately 13475 eV and the ap-pearance of a small shoulder peak at approximately 13460 eV (figure 3.8(b)).The peak at approximately 13475 eV corresponds to the σ∗C–Br of the Dye-Br563.3. Solution XAS of Dye-Br(a) (b)Figure 3.8: Br K-edge XAS data of (a) Dye-Br with and without NBu4Cl,and oxidised Dye-Br (Dye-Br+.)(b) oxidised Dye-Br (Dye-Br+.) with andwithout (x2 repeats) NBu4Cl. The data for oxidised Dye-Br (Dye-Br+. -blue line) is included in both graphs for the sake of comparison.(for ground state dyes). A decrease in the intensity of this peak is expectedupon XB formation as orbital overlap between the covalently-bound Br inDye-Br and the Cl– would result in a decrease of Br character in the resultingorbital and thus a decrease in the probability of the transition. This orbitaloverlap should also result in a change in the energy of the transition (a peakshift). A slight peak shift of +0.2 eV upon addition of NBu4Cl is observedbut as this shift in energy is close to the resolution of this technique, it isnot conclusive.The appearance of the small shoulder feature at approximately 13460eV is also noteworthy. Our preferred explanation for this feature is thepresence of pi-type character in XB. Upon one electron oxidation of Dye-Br,a lower energy transition (piXB ← Br1s) is now possible (figure 3.9). Theexistence of XB pi-type character is further supported by data in section 3.3.2which suggests that the Cl1s electron has an additional allowed transition. Itshould be noted that this small shoulder feature could also result from Dye-Br acting as both a donor and acceptor in an XB or cation-pi interactionsbetween the NBu+4 and the oxidised Dye-Br. Both of these interactionsmay result in a new low energy feature for the Br K-edge XAS, but neither573.3. Solution XAS of Dye-Braccount for the changes observed in the Cl K-edge XAS in the followingsection.Figure 3.9: The molecular orbital diagram of a covalently-bound brominein Dye-Br+. involved in an XB exhibiting pi-type character with Cl–. Thetransitions of interest in Br K-edge XAS are shown in orange and green forground state Dye-Br (absence of XB) and oxidised DyeBr (presence of XB),respectively.Figure 3.8(b) shows two different repeats of the oxidation of Dye–Br+. inthe presence of NBu4Cl. Although the difference between the two spectra isinitially concerning, it can be explained by the limitations of the experiment.As mentioned earlier, upon oxidation at room temperature in the presence oftrace amounts of water, the dark green Dye-Br+. returns to the ground state(orange Dye-Br) in less than two seconds. Although the samples analysedwere flash frozen in liquid nitrogen less than one second after oxidation,an appearance of both green and orange indicate the presence of both theoxidised and ground state dyes. Thus, the different intensities observedbetween the two repeats of the same sample can be explained by the differentratio of ground state and excited state dyes present in solution.583.3. Solution XAS of Dye-Br3.3.2 Solution XAS of Dye-Br: Cl K-edgeTo confirm the presence of XB in acetonitrile solutions of Dye-Br+. and Cl–,XAS data was also collected for the XB electron donor (Cl K-edge XAS).As the instrument set up (liquid helium flow onto the sample into aplastic bag under vacuum) was insufficient to keep the samples frozen forthe duration of the data acquisition (3-10 min/scan for these experiments),the data acquired gives insight into changes in the system as the oxidiseddye returned to the ground state. While samples were not visible duringanalysis, no green solution (oxidised dye) was visible when the sample wasremoved from the sample chamber.In figure 3.10, the first XAS data collection scan shows the diagnosticpre-edge feature (approximately 2821 eV) indicative of partially covalentXB. By this point in the second scan (10 minutes later), this feature hasdisappeared suggesting that there is no longer significant XB formation.Interestingly, there is a second less intense pre-edge feature at a lower energy(approximately 2818 eV) than the first pre-edge feature. This feature can beexplained by several phenomena. Similar to the pre-edge feature observedin the Cl K-edge spectrum of NOBF4 with Cl– in the previous section, thispre-edge feature could merely be a result of interactions between the Cl– andother non-dye species. However, the pre-edge feature in figures 3.10 and 3.11is at 2818 eV while the ‘background’ feature corresponding to interactionsbetween NOBF4 and Cl– in figure 3.6 is at 2824 eV.The second possible explanation for the feature at 2818 eV is the presenceof pi character in the XB. Any pi-type interactions contributing the XB wouldbe expected to be lower in energy and less intense than the pre-edge featurecorresponding to the σ character of the XB.The experiment shown in figure 3.10 was repeated with shorter scans,approximately 3 min/scan (figure 3.11). The first four scans of this exper-iment also show a pre-edge feature at approximately 2821 eV indicative ofpartially covalent XB. Additionally, the first two scans also show a featureat approximately 2818 eV. However, in both of these scans, this feature isdefined by a single data point and cannot be regarded as conclusive.To verify if these two pre-edge features are indicative of pi-type interac-tions in halogen bonding, there are two follow up experiments: (1) repeatexperiments (figures 3.10 and 3.11) in a cryostat to prevent the sample melt-ing and (2) perform Cl K-edge XAS of NOBF4 and Cl– in acetonitrile todetermine if interactions between NOBF4 and Cl– pre-edge give rise to oneof the pre-edge features present in figures 3.10 and 3.11. These experimentshave not yet been conducted due to the standard time constraints associated593.3. Solution XAS of Dye-BrFigure 3.10: Cl K-edge XAS data of oxidised Dye-Br (Dye-Br+.) returningto the ground state dye in the presence of NBu4Cl as the sample melts. Asthe sample is only visible before the first scan and after the final scan, it isnot possible to determine when the sample was completely melted. However,as there is little change after the second scan, it is reasonable to assume thatthe sample was fully melted by the end of the first scan (10 minutes). Eachscan lasts approximately 10 minutes (283 data points/scan).603.3. Solution XAS of Dye-BrFigure 3.11: Cl K-edge XAS data of oxidised Dye-Br (Dye-Br+.) returningto the ground state dye in the presence of NBu4Cl as the sample melts. Sim-ilar to the data collected in figure 3.10, the sample is only visible before thefirst scan and after the final scan. Thus, it is not possible to determine whenthe sample was completely melted. However, as there is little change afterthe fourth scan, it is reasonable to assume that the sample was fully meltedby the end of the third scan (9 minutes). Each scan lasts approximately 3minutes (81 data points/scan).613.4. Conclusionswith allocated beamtime at synchrotrons. Although this is only preliminarydata, this Cl K-edge XAS data could be the first direct experimental evi-dence that an XB can have σ and pi-type contributions.3.4 ConclusionsIn this chapter, XAS is used to show that only the oxidised dyes are ableto form partially covalent coordinate bonds with the Cl– electrolyte. Ourdata suggests that increasing the covalency of these interfacial XB is di-rectly responsible for the increased dye regeneration rate as the radius ofthe dye’s halogen substituents increase. These results are particularly sug-gestive due to the short lifetime of the oxidised dye. The strength of usingXAS to probe XB is particularly evident in these systems as the XB occurat a solid-liquid interface which would be difficult to characterise by otherspectroscopic techniques.XAS of the XB acceptor was also conducted on the oxidised Dye-Br insolution. As in section 2.4, a decrease in the pre-edge feature of the theseXAS spectra corresponded to the presence of XB. The limitations of thetechnique prevent us from determining if this decrease in intensity is alsoaccompanied by the expected increase in the energy of the peak (as predictedby molecular orbital theory). An additional feature in the Br K-edge spectrais potentially indicative of pi-type character in these XB. This observation isfurther supported by preliminary Cl K-edge XAS of the XB donor. Whilethese experiments should be repeated for validation, this data could be thefirst direct experimental evidence of pi-type character in XB.3.5 MethodsSample Preparation Samples analysed on beamline 14-3 were preparedon a sensitized, mesoporous TiO2 film deposited on glass. Each film wasstained with Dye-F, Dye-Br, or Dye-I, then sealed with X-ray-transparentpolypropylene before being filled with an acetonitrile solution of 100mMNBu4Cl. The oxidised forms of the dyes on TiO2 were obtained via chem-ical oxidation by washing the films with a saturated solution of NOBF4 inacetonitrile prior to filling with electrolyte. To account for possible chloride-nitrosonium interactions, an undyed film was also washed with NOBF4 be-fore filling with electrolyte to serve as a control.Samples analysed on beamlines 4-3 and 7-3 were prepared as acetonitrilesolutions with 1mM dye and 2mM NBu4Cl. For samples involving the ox-623.5. Methodsidised dye (Dye-Br+.), a saturated NOBF4 acetonitrile solution was addedto a solution of 2mM NBu4Cl and 1mM of the Dye-Br using a syringe.Oxidation of the relevant dye was instantaneous (dark green colour) and toprevent the dye from returning to its ground state (orange colour), solutionswere frozen in liquid nitrogen less than 1 second after was NOBF4 solutionadded.XAS data collection Cl K-edge XAS data was collected at beamlines 14-3 (TiO2 films), 7-3 (acetonitrile solutions) and 4-3 (acetonitrile solutions)at the SSRL. For beamline 14-3, the samples were analyzed at ambienttemperatures and pressures in a helium environment. The beam is unfocusedover a size of 1 mm x 6 mm with an energy resolution of approximately1x104 ΔE E–1 with ring conditions of 3 GeV and 500 mA to allow for highenergy resolution measurements on homogeneous samples. Data points weretaken at several points across the surface of the cell to ensure homogeneity.Cl K-edge data collection at beamline 4-3 used a modified low Z setupallowing for low temperature data acquisition under ring conditions of 3GeV and 500 mA. The setup is a 54-pole wiggler beamline operating inhigh field (10 kG) mode with a Ni coated harmonic rejection mirror and afully tuned Si (111) double crystal monochromator. Calibration scans wereperformed before and after every data set to ensure stable monochromatorreadings. Signal was detected with a N2 fluorescence (Lytle) detector atambient temperature and pressure (1 atm, 298 K). Br K-edge XAS data werecollected at SSRL on beamline 7-3 under ring conditions 500 mA at 3.0 GeV.This beamline has a 20-pole, 2 T wigglers, 0.8 mrad beam, and a Si (220)double crystal monochromator that was detuned by 50% intensity to attainharmonic rejection. The incident X-ray intensity (I0), sample absorption(I1), and Br reference absorption (I2) were measured as transmittance usingargon-filled ionization chambers. Signal was detected with a 30-channel Gedetector. All data was measured at 20 ± 15 K within an Oxford InstrumentsCF1208 continuous-flow liquid helium cryostat.XAS data analysis SixPack was used to calibrate and average XAS spec-tra [44]. NaCl (E0 = 2825.95 eV) and KBr (E0 = 13474) spectra acquiredat the same time as sample data were used as a calibrant for Cl K-edge andBr K-edge spectra, respectively. For data collected on the beamline 14-3(TiO2 films), BlueprintXAS version 2.7 was then used for background sub-traction and normalization. To minimize the user bias introduced duringdata work-up, BlueprintXAS fits the spline, peaks, and background con-633.5. Methodscurrently [45, 56], while the parameters for each variable are user defined(tables A.5-A.8 for final values for each parameter) The fits were run inAUTO mode as in this mode, a Monte Carlo methodology is used to choosethe starting point of each fit to further remove user bias. Each fit containedthe following components: edge peak fit; spline + edge peak fit; pre-edgepeak fit.643.5. MethodsAgain we probe for XB,Twixt solid and liquid they be,And now we can prove,That they help to moveCharge through the DSSC.65Chapter 4Using XAS, NMR and CDto Probe Amyloid βAggregationThe importance of XB is increasing across many fields including biochem-istry. While halogenated molecules are less commonly associated with bi-ological systems, halogen substituents are surprisingly common in both bi-ological molecules and drug candidates. There are over 3500 metabolitescontaining halogens, and both nucleic acids and proteins can undergo halo-genation in response to oxidative stress [57–63]. It is likely that some of thesesystems involve XB. Indeed, upon discovering that XB were responsible foran unexpected structure of brominated DNA, Auffinger and collaboratorsembarked upon an analysis of Protein Data Bank structures which identifiedXB in 72 additional structures: 66 protein and 6 nucleic acid systems whereintermolecular halogen-oxygen bond distances less than that of the sum ofthe Van der Waals radii [57]. Other researchers have demonstrated the im-portance of XB in biological systems by replacing hydrogens with halogensubstituents to improve drug binding specificity [64–66] and using XB tostabilize cancer-associated p53 mutants [67].In this chapter, we will probe one biological system which may benefitfrom the presence of XB: the inhibition of amyloid β peptide aggregationby fluorescein derivatives. The aggregation of various forms of this peptideinto β-sheet oligomers and fibrils is associated with the onset and diagnosisof Alzheimer’s Disease [68, 69]. The two most common forms of the ag-gregated peptide are Aβ40 and Aβ42, respectively the first 40 or 42 aminoacid residues in the Amyloid Precursor Protein. In the search for moleculeswhich prevent Aβ40 aggregation, recent work reveals that small halogenateddyes can modulate Aβ40 aggregation [37]. Furthermore, these studies showthat an analogous molecule with two nitro electron withdrawing groups inplace of the two halogen substituents is unable to modulate Aβ40 aggrega-tion (discussed further in section 4.1). This result suggests that the electron664.1. Background: Aβ40 aggregation modulation by fluorescein derivativeswithdrawing nature of the halogen substituents by itself is insufficient tomodulate Aβ aggregation. To investigate whether the importance of halo-gen substituents is a result of XB in the systems, we will use the XASmethodology discussed in Chapters 2 and 3.Using XAS to probe for XB in peptide aggregation is particularly advan-tageous as it removes the need for single crystals, XAS sample requirementsbeing either a solid powder or frozen solution containing the element ofinterest. Aβ40’s secondary structure is very sensitive to different samplepreparation; indeed Aβ40 can interconvert between random coil, α-helix, orβ-sheet secondary structures with small changes in pH, temperature, andsolvent [70]. This complex behaviour of Aβ40 makes probing for XB in thissystem an ideal experiment for study by XAS.In this section, XAS will be coupled with 2D Nuclear Magnetic Reso-nance (NMR) spectroscopy. As probing Aβ40 aggregation with XAS onlyprovides information about the electronic environment of the halogen sub-stituents, NMR will be used to investigate the binding site of these dyes.4.1 Background: Aβ40 aggregation modulationby fluorescein derivatitvesAs mentioned previously, the importance of halogen substituents in the mod-ulation of Aβ40 aggregation was initially identified by Wong and collabora-tors [37]. This study shows that addition of halogen substituents to fluores-cein derivatives (figure 4.1) changes their effectiveness in modulating Aβ40aggregation when incubated with the peptide. Circular dichroism (CD) ex-periments reveal that after incubation with fluorescein (FLN) and eosin b(EOB), the secondary structure of Aβ40 is predominantly β-sheet, indica-tive of aggregation. By comparison, after incubation with eosin y (EOY),phloxine B (PHB), erythrosine B (ERB), and rose bengal (ROB), the sec-ondary structure of the peptide is predominantly random coil, indicative offree monomer (figure 4.2).This CD data is further supported by transmission electron microscopyand cytoxicity studies [37]. Transmission electron microscopy images showthat same dyes which prevent β-sheet formation in the CD spectra alsoreduce Aβ fibril formation [37]. The dyes’ reduction of peptide β-sheet for-mation also correlates with a reduction in the cytoxicity of the peptide [37].674.1.Background:Aβ40aggregationmodulationbyfluoresceinderivativesFigure 4.1: Chemical structures of the six dyes used in Aβ40 aggregation modulation studies [37]. Chemdrawstructures were produced by the author.684.1.Background:Aβ40aggregationmodulationbyfluoresceinderivativesFigure 4.2: CD spectra of Aβ monomer and Aβ aggregates modulated by fluorescein derivatives. (a) CD spectraof Aβ40 monomer and Aβ40 incubated with and without EOB and PHB for 5 days at 37oC. Aβ40 is shown toaggregate in the presence of EOB, while PHB partially prevents aggregation. (b) CD spectra of Aβ40 incubatedwith and without EOY, ROB and ERB for 5 days at 37oC. Incubation with EOY, ROB and ERB is shown toprevent aggregation. (c) CD spectra of Aβ40 incubated with and without FLN 5 days at 37oC. FLN does notprevent aggregation ( c© Creative Commons) [37].694.2. XAS Studies of Aβ40 AggregationA follow-up study from the same group investigated the binding siteof these dyes via fluorescence quenching and competitive sequence-specificantibody binding studies [71]. These studies suggest that the dyes interactwith the first 16-22 residues of the Aβ40 peptide. Fluorescence quenchingresulting from dye-peptide binding is observed for ERB, EOY, PHB andROB in the presence of the Aβ11-22 fragment. Furthermore, ERB, EOY,PHB and ROB were shown to inhibit antibody-Aβ40 binding for the first16 residues of Aβ40 only.The aim of this chapter is to probe the nature of these dye-peptide in-teractions by probing for the presence of XB by XAS and examining thelocation of the binding site with Nuclear Magnetic Resonance (NMR) spec-troscopy.4.2 XAS Studies of Aβ40 AggregationAlthough chlorine, bromine and iodine substituents are all present in thedyes, only Br K-edge XAS is used to probe for XB interactions between thedye and the peptide. To mimic biological environments, the Aβ40 aggrega-tion is studied in a salt buffer solution which includes NaCl. The presenceof Cl– in the samples at 20 times the concentration of the dyes precludesthe use of Cl K-edge XAS. The resulting Cl K-edge XAS spectrum would bedifficult to interpret as it would be an average of all chlorine atoms present,not just the chlorine substituents on the dye. Using Br K-edge XAS toprobe for XB in these systems is also preferred to I L3-edge XAS as thehigher energy Br K-edge is associated with less self-absorption and reducedinstrument noise given the available beamlines: 30-channel detector for BrK-edge beamline compared with the 1-4 channel detector used at I L3-edgebeamline. Br K-edge XAS also allows us to probe one of the most interestingoutcomes of Wong and collaborators: the fact that EOY can inhibit Aβ40β-sheet formation while EOB, with two less bromine substituents, cannot.To investigate these systems, previously published procedures were fol-lowed [72] with the additional step of agitation at 200 rpm during incuba-tion as Aβ40 is found to only aggregate with the addition of this condition(figure 4.3). As found by Wong and collaborators, β-sheet formation is ob-served after 5-day incubation at 37oC for both the peptide alone in thebuffer solution and when incubated with EOB. No β-sheet formation is ob-served when Aβ40 is incubated with EOY and PHB, and the CD data forboth is indicative of random coil structure. Interestingly, this data differsslightly from Wong and collaborators findings. Wong and collaborators’ CD704.2. XAS Studies of Aβ40 Aggregationdata indicates PHB only partially prevents aggregation (figure 4.2) whilethe PHB CD spectrum collected by the author is more similar to that ofEOY, indicating that PHB is just as effective as EOY at preventing aggre-gation (figure 4.3). This similarity between the two different dye systems isrepeatable (figure 4.4).With the repeatability of the CD experiments confirmed, Br K-edge XASanalysis of Aβ40 incubated with EOB, EOY, and PHB were then conducted(figures 4.5, 4.6, and 4.7), with XAS of the dyes alone in buffer solution serv-ing as the negative control and the EOB system, in which β-sheet formationis not prevented, serving as the positive control. Samples were incubated for5 days with daily aliquots removed and flash-frozen for XAS analysis. A lackof significant changes in the pre-edge features of all XAS spectra indicatesthat no XB formation is observed in these systems. CD spectra are presentedfor these samples to illustrate that using a different incubator, daily aliquot-ing, and flash-freezing do not affect peptide aggregation (figure 4.10). Asno change was observed in the XAS data over the course of 5 days, and thetime-dependent secondary structure of Aβ42 in the presence of these dyeshas been extensively studied [72], the time-dependent secondary structureof Aβ40 in the presence of EOB, EOY, and PHB was not investigated.The procedure outlined by Wong and collaborators uses 10 equivalentsof the dyes to prevent β-sheet formation in Aβ40, but any unbound dyewill make changes in the XAS spectra smaller and more difficult to observe.Thus, Br K-edge spectra of Aβ40 in the presence of 0.86 equivalents of EOBand EOY were acquired to minimise spectral effects of unbound dyes presentin the samples. Small decreases in pre-edge feature intensity is observed forboth Aβ40 co-incubation with EOY (figure 4.8) and for the positive control,Aβ40 co-incubation with EOB (figure 4.9). This small decrease observedmay be insignificant due to the limitations of the experiments. However, ifthe decrease in pre-edge feature intensity is attributable to XB formation, itis observed for both systems, and thus XB alone cannot be responsible forEOY’s prevention of Aβ40 β-sheet formation.As we have shown that Cl– can act as a XB electron donor, XAS datafor EOY and PHB in the presence of NBu4Cl was also acquired. Thesestudies investigate whether XB formation with Cl– may prevent XB for-mation between the dyes and Aβ40. While small changes in the edge andpost-edge regions are present in the spectra (figure 4.11), there is no changein the pre-edge feature. The small changes in the post-edge region suggeststhat the solvation environment of the bromine substituents is changing asthe signal in the edge-region arises from scattering. The lack of change inthe pre-edge features indicates no change in the electronics of the bromine714.2. XAS Studies of Aβ40 Aggregationsubstituents. The lack of change in the electronic environment suggests thatthere is no XB between solvated Cl– and the dyes.724.2. XAS Studies of Aβ40 AggregationFigure 4.3: CD spectra for 30μM Aβ40 incubated at 37oC, 200 rpm, for5 days with and without 300μM of EOY, EOB, PHB. The CD spectra ofAβ40 in the absence of the dyes and in the presence of EOB is indicativeof β-sheet formation suggesting that aggregation has occurred. The CDspectra of Aβ40 in the presence of EOY and PHB is indicative of randomcoil suggesting that aggregation has been prevented. These experimentswere performed to confirm Wong and collaborators findings [37]. Due tohigh HT (high tension) voltage values (>500V), data below 205 nm was notincluded. The spectra in this figure are noisier than the rest of the CD datain this thesis due to the age of the light source in the CD instrument. Thelight source was replaced before the collection of the other spectra.734.2. XAS Studies of Aβ40 AggregationFigure 4.4: CD spectra for 30μM Aβ40 incubated at 37oC, 200 rpm, for5 days with and without 300μM of FLN, EOY, PHB. The CD spectra ofAβ40 in the absence of the dyes and in the presence of FLN is indicativeof β-sheet formation suggesting that aggregation has occurred. The CDspectra of Aβ40 in the presence of EOY and PHB is indicative of randomcoil suggesting that aggregation has been prevented. Due to high HT (hightension) voltage values (>500V), data below 205 nm was not included.744.2. XAS Studies of Aβ40 AggregationFigure 4.5: Br K-edge XAS spectra of Aβ40 incubated at 37oC, 200 rpm,with 10x EOY over the course of 5 days. As XAS samples are not recov-erable, each spectrum represents a new aliquot from a stock sample. Eachaliquot was flash frozen in liquid nitrogen after being removed from the bulksolution. The calibration of these spectra was particularly difficult due tothe beamline instability during these samples. These difficulties are evidentin the approximately 0.3eV shifts between the spectra in this figure and toa lesser extent in figures 4.6 and 4.7. Despite this, it is evident that thereis no significant change in the intensity of the pre-edge feature (13473 eV)indicating no XB formation.754.2. XAS Studies of Aβ40 AggregationFigure 4.6: Br K-edge XAS spectra of Aβ40 incubated at 37oC, 200 rpm,with 10x EOB over the course of 5 days. There is no significant change inthe intensity of the pre-edge feature (13473 eV) indicating no XB formation.As XAS samples are not recoverable, each spectrum represents a new aliquotfrom a stock sample. Each aliquot was flash frozen in liquid nitrogen afterbeing removed from the bulk solution. The day 5 sample was lost duringthis process.764.2. XAS Studies of Aβ40 AggregationFigure 4.7: Br K-edge XAS spectra of Aβ40 incubated at 37oC, 200 rpm,with 10x PHB over the course of 5 days. There is no significant change in theintensity of the pre-edge feature (13473 eV) indicating no XB formation. AsXAS samples are not recoverable, each spectrum represents a new aliquotfrom a stock sample. Each aliquot was flash frozen in liquid nitrogen afterbeing removed from the bulk solution. The day 4 sample was lost duringthis process.774.2. XAS Studies of Aβ40 AggregationFigure 4.8: Br K-edge XAS spectra of 1.2 equivalents of Aβ40 incubatedwith EOY at 37oC, 200 rpm, for 2 days. The length of the beamtime atSSRL prevented a longer incubation time. A small decrease in peak intensityis observed. This is also observed in the EOB positive control (figure 4.9),and could be attributed to noise for both experiments. Thus, if this smalldecrease is attributable to XB formation, it is not responsible for EOY’sprevention of Aβ40 β-sheet formation.784.2. XAS Studies of Aβ40 AggregationFigure 4.9: Br K-edge XAS spectra of 1.2 equivalents of Aβ40 incubatedat 37oC, 200 rpm, with EOB for 2 days. The length of the beamtime atSSRL prevented a longer incubation time.A small decrease in peak intensityis observed. This is also observed in the Aβ40 sample incubated with EOY(figure 4.8), and could be attributed to noise for both experiments.794.2. XAS Studies of Aβ40 AggregationFigure 4.10: CD spectra for 30 μM Aβ40 incubated at 37oC, 200 rpm, for5 days with and without 300 μM of EOY, EOB, PHB. These experimentswere conducted to probe whether the changes required for the experimentsconducted at SSRL affected aggregation. Changes include: incubation atSSRL and daily removal from incubation/agitation to flash freeze aliquotsfor XAS analysis. Samples were flash frozen and transported back to UBCfor CD analysis. As expected, the CD data indicates that Aβ40 incubatedwith EOB has β-sheet structure while Aβ40 incubated with EOY and PHBhas random coil structure. The lack of β-sheet formation in the Aβ40 sampleis concerning but as this sample was not analysed by XAS, this is tentativelyattributed to the fact that while the Aβ40 sample was removed from theincubator daily, it was not perturbed by aliquoting by syringe. As β-sheetformation was observed in the EOB positive control, the lack of aggregationin the Aβ40 sample was not investigated further due to the resource-intensivenature of repeating this experiment at SSRL. Due to high HT (high tension)voltage values (>500V), data below 205 nm was not included in this thesis804.2. XAS Studies of Aβ40 AggregationFigure 4.11: Br K-edge XAS spectra of EOY and PHB in presence andabsence of Cl–.814.3. Nuclear Magnetic Resonance Spectroscopic Studies of Aβ Aggregation4.3 Nuclear Magnetic Resonance SpectroscopicStudies of Aβ AggregationAs XAS reveals that the interaction between EOY and Aβ40 is unlikelydue to the presence of XB, the dye-peptide interactions are investigatedby 2D proton Nuclear Magnetic Resonance (1H-NMR) Spectroscopy. BothTotal Correlation Spectroscopy (1H-TOCSY) and Nuclear Overhauser EffectSpectroscopy (1H-NOESY) data was collected. These two experiments arecomplementary as TOCSY NMR reveals which protons are in the sameamino acid, and NOESY NMR reveals which amino acids are neighbours.In correlation NMR spectroscopy, the transfer of magnetisation, or coupling,between nuclei occurs if the two nuclei are connected through bonds. Totalcorrelation NMR spectroscopy (TOCSY) enables the identification of spinsystems, groups of these coupled nuclei where the nuclei in the group arecoupled to at least one other nucleus within the group and no nuclei outsideof the group. In TOCSY NMR of peptides, these spin systems correlate withamino acids. By contrast, in 1H-NOESY NMR, magnetisation is transferredthrough space rather than through bonds. Thus, crosspeaks arise in a 1H-NOESY spectrum if two protons are within approximately 5 A˚.Titration of EOY with Aβ40 results in changes in many NOESY cross-peaks indicative of a binding interaction between EOY and Aβ40. Thesechanges include appearance of peptide-peptide crosspeaks suggesting un-structured regions of the peptide becoming more rigid upon addition ofEOY (figure 4.12). The appearance of NOESY crosspeaks between pro-tons with 7.4 to 7.7 ppm with peptide methyl protons (0.65 - 0.8 ppm) areeither additional peptide-peptide crosspeaks or dye-peptide crosspeaks (fig-ure 4.14). These crosspeaks may be direct evidence of the residues to whichEOY binds. Furthermore, NOESY peptide-peptide crosspeaks appear uponaddition of EOY and then disappear as more equivalents of the dye areadded (figure 4.13) and peptide-peptide and dye-dye crosspeaks disappearupon addition of EOY (figures 4.13 and 4.15). These changes are indicativeof more complex structural changes which require assignments to elucidate.824.3.NuclearMagneticResonanceSpectroscopicStudiesofAβAggregationFigure 4.12: The same region of the NOESY spectra for Aβ40 in the presence of different EOY concentrations(left) and the NOESY spectrum for the apo-peptide (right) is shown. In the right hand-side plot, dark/light blue,dark/light green, dark/light orange and dark/light red corresponds to the +/- peak phases of the NMR spectrumof 0.46 mM Aβ40 in the presence of 0.46mM, 0.67mM, 0.89mM and 1.10mM EOY, respectively. In the left hand-side plot, dark/light red corresponds to the +/- peak phases of the NMR spectrum of 0.46 mM Aβ40. Peakswhich are present in the spectra of dye with peptide but are absent in the apo-peptide spectrum are indicatedwith arrows. The different colours indicate coupled systems.834.3.NuclearMagneticResonanceSpectroscopicStudiesofAβAggregationFigure 4.13: A comparison of various 1H-NOESY spectra illustrating the appearance and then disappearanceof crosspeaks upon mixing EOY with Aβ40. The same region of the NOESY spectra for Aβ40 in the presence ofdifferent EOY concentrations (right) and the NOESY spectra for the apo-peptide (middle) and the dye (left) isshown. In left hand-side plot, dark/light blue, dark/light green, dark/light orange and dark/light red correspondsto the +/- peak phases of the NMR spectrum of 0.46 mM Aβ40 in the presence of 0.46mM, 0.67mM, 0.89mMand 1.10mM EOY, respectively. In middle plot, dark/light blue corresponds to the +/- peak phases of the NMRspectrum of 0.46 mM Aβ40. In right hand-side plot, dark/light blue corresponds to the +/- peak phases of theNMR spectrum of 0.46 mM Aβ40. Peaks which are present in the spectra of dye with peptide but are absent inthe apo-peptide spectrum are indicated with arrows. Pink arrows indicate peaks which appear upon addition of1 equivalent of the dye but disappear as more dye is added. Red arrows indicate peptide peaks which disappearas more dye is added.844.3.NuclearMagneticResonanceSpectroscopicStudiesofAβAggregationFigure 4.14: A comparison of various 1H-NOESY spectra illustrating the appearance of additional crosspeaksupon mixing EOY with Aβ40. The same region of the NOESY spectrum for the apo-peptide (top) and the NOESYspectra for Aβ40 in the presence of different EOY concentrations (bottom) is shown. In the top plot, dark/lightred corresponds to the +/- peak phases of the NMR spectrum of 0.46 mM Aβ40. In the bottom plot, dark/lightblue, dark/light green, dark/light orange and dark/light red corresponds to the +/- peak phases of the NMRspectrum of 0.46 mM Aβ40 in the presence of 0.46mM, 0.67mM, 0.89mM and 1.10mM EOY, respectively. Thecomparison of these spectra shot that many amide/aromatic proton to methyl proton cross-peaks appear uponaddition of the dye.854.3.NuclearMagneticResonanceSpectroscopicStudiesofAβAggregationFigure 4.15: A comparison of various 1H-NOESY spectra illustrating the disappearance of apo-dye and apo-peptide crosspeaks upon mixing EOY with Aβ40. The same region of the NOESY spectra for Aβ40 in thepresence of different EOY concentrations (right) and the NOESY spectra for the apo-peptide (middle) and thedye (left) is shown. In left hand-side plot, dark/light blue, dark/light green, dark/light orange and dark/light redcorresponds to the +/- peak phases of the NMR spectrum of 0.46 mM Aβ40 in the presence of 0.46mM, 0.67mM,0.89mM and 1.10mM EOY, respectively. In middle plot, dark/light red corresponds to the +/- peak phases ofthe NMR spectrum of 0.46 mM Aβ40. In right hand-side plot, dark/light red corresponds to the +/- peak phasesof the NMR spectrum of 0.46 mM Aβ40. Arrows indicate peaks that are present in either the apo-dye spectrumor the apo-peptide spectrum but not in the spectra of the mixture. Two arrows of the same colour indicate theabove- and below-diagonal crosspeaks for the same two protons.864.3. Nuclear Magnetic Resonance Spectroscopic Studies of Aβ AggregationOverlap in the amide region of the 1D spectrum of the peptide promptedthe addition of d3-TFE to the NOESY/TOCSY NMR sample to increasepeak separation in amide region as done in the literature [70, 73, 74]. Whilethe separation of amide peaks aids assignment, the broad solvent peaks ofd3-TFE and water at 3.75 ppm and 4.75 ppm presumably also hides a signif-icant number of Hα crosspeaks. Consequently, the spin systems identified inthe TOCSY spectra are likely incomplete, and each spin systems identifiedcorresponds with a minimum of ten of the forty residues. Thus, assignmentby hand was not feasible. In an effort to solve this, python algorithms werewritten by the author to aid with identifying spin systems in the TOCSYspectrum of the apo-peptide and finding corresponding crosspeaks in theNOESY spectrum (appendices B.2 and B.1). The logic of these programs isdepicted in flowcharts (figures 4.16 and 4.17).Figure 4.16: The algorithm logic for the python program developed forTOCSY analysis. Decisions are in blue.The data from above and below the diagonal in the TOCSY and NOESYspectra was separated in analysis as this provides an informal validation874.3. Nuclear Magnetic Resonance Spectroscopic Studies of Aβ AggregationFigure 4.17: The algorithm logic for the python program developed forNOESY analysis. Decisions are in blue.method for the spin systems and nearest neighbours identified in both datasets.An additional assignment tool was developed by the author which cantake any NMR spectral data published in the Protein Data Bank and gener-ate a TOCSY spectrum for that peptide or protein (figure 4.18, appendix B.3).The program also overlays the user’s spectrum on top of the simulation of theTOCSY spectrum for the TOCSY data. Despite the similar experimentalconditions, there is little overlap between the user’s data and the simulatedTOCSY spectrum [75]. However, this tool can be used to generate a fewpossible assignments including the arginine (R5) residue highlighted in theimage.884.3.NuclearMagneticResonanceSpectroscopicStudiesofAβAggregationFigure 4.18: A screenshot of the TOCSY generator program which can take any NMR spectral data published onthe Protein Data Bank and generate a TOCSY spectrum for that peptide or protein and overlay the users data(black x’s). The data from the Protein Data Bank is displayed with a rainbow colour scheme to help identifywhere the residue lies in the amino acid sequence. A mouse-over tool enables the user to identify a peak of interest(here R5 or arginine 5, the 5th residue in Aβ40, is being pointed to by the cursor).894.3. Nuclear Magnetic Resonance Spectroscopic Studies of Aβ AggregationUsing the distance restraints from the NOESY data, a network graph ofnearby spin systems is generated (figure 4.19). Lowering the constraint from0.007 ppm to 0.006 ppm for two peaks to be considered as the same protonsignificantly reduces the number of distance restraints, providing additionalvalidation for the remaining distance constraints. While the assignmentof these spin systems is ongoing, the only spin systems present which cancorrespond to a glycine residue are 15 and 17 due to the presence of eithera proton with a chemical shift less than 2 ppm or multiple protons withchemical shifts above 6 ppm in the remaining spin systems. Due to thepreponderance of glycine residues in C-terminus half of Aβ40, five glycineresidues in total with a maximum of three non-glycine sequential residuesin the last 18 residues of the peptide, it is likely that the spin systems inthis network graph are in the N-terminus half of the peptide. As these spinsystems correspond to the changing NOESY crosspeaks discussed earlier, thecurrent data suggests that the changes observed in the NOESY spectrumresult from structural changes in and EOY binding to the first half of thepeptide.904.3. Nuclear Magnetic Resonance Spectroscopic Studies of Aβ Aggregation(a)(b)Figure 4.19: Network graphs of spin system distance restraints. The con-straint for two peaks to be correspond to the same proton is (a) 0.006 ppmand (b) 0.007 ppm. Spin systems were generated with the TOCSY pythonprogram (appendix B.1) and the spin system numbers refer to the list of spinsystems in the relevant text file (appendix C). Distance restraints were gen-erated using the NOESY python program (appendix B.2). Line thicknesscorrelates to the NOESY crosspeak intensity.914.4. Conclusion4.4 ConclusionPreviously reported results showing that the aggregation of Aβ40 peptideis modulated by halogenated fluorescein derivatives was verified. Using BrK-edge XAS, we showed that XB interactions are not responsible for theimportance of the halogens in these systems. No XB were found in thesesystems. Although no XB were found in this system, this is nonethelessan illustration that XAS can be used to probe complex biological systems.With the prevalence of halogenated substituents in biological systems, thisuse of XAS could provide insight into other complex systems.As XAS suggests that XB are not responsible for EOY’s ability to bindto Aβ40, 1H-TOCSY and 1H-NOESY NMR experiments were used to fur-ther probe this interaction. The disappearance of apo-peptide and apo-dyepeak in addition to the appearance of peptide-peptide crosspeaks as moreequivalents of EOY are added to the Aβ40 sample are a clear indication of abinding interaction between the dye and the peptide. While the Hα regionof the spectra was obscured by presence of the TFE solvent peak, headwayhas been made on the assignment of these spectra using the author’s pythonprograms. While the assignment of these spectra is ongoing, current NMRdata supports the literature finding that EOY binds to the first 22 residuesof the Aβ peptide.As the NMR data, supported by the literature, suggests that EOY doesbind to Aβ40 while the XAS data suggests there are no XB, the halogensubstituents must be responsible for other electrostatic phenomena whichenable the dye-peptide interaction. One possible explanation is that thehalogen substituents are important in how they change the electronics offluorescein scaffold to which they are attached, for example in increasing theoverall hydrophobicity of the molecule and promoting peptide-dye cation piinteractions. However, the lack of binding between EOB which has two non-halogen electron withdrawing substituents and Aβ40 [71] suggests that thisisn’t the sole role of the halogens. Instead, the region of negative chargesurrounding the positively-charged σ-hole of halogen substituents might in-teract electrostatically with positively-charged residues on the peptide. Suchan interaction in combination with the increased hydrophobicity of the dyecould strengthen pi-pi interactions between the dye and the peptide. We hopeto further elucidate the nature of this interaction once the NMR assignmentsare complete.924.5. Methods4.5 MethodsPreparation of Peptide Thin Films. Aβ40 was purchased from Anaspecas a lyopholised powder. 1.0mg of Aβ40 was dissolved in 1 ml of hexafluo-roisopropanol by sonication for 5 minutes. The solution was stored overnightat room temperature and then aliquoted into 12 vials. Aliquots were driedwith a gentle stream of compressed air perturbing the surface of the liquiduntil a thin film was visible. The thin films were left in a vacuum dessicatorovernight and then stored at -20oC until needed for aggregation experiments.Aggregation of Aβ40 for all CD and XAS experiments. Aβ40 prepa-ration was based on published studies [72]. Aβ40 thin films were reconsti-tuted by sonication for 10 minutes ([Aβ40] = 300 μM) in the following solu-tion: 300 μM CH3CN, 300 μM Na2CO3, 250 μM NaOH. Buffer solution (8.5mM NaCl, 14 μM Na2CO3, 0.85mM NaOH, 6.0% CH3CN) and dye stocksolution (10mM dye in buffer solution) were added to the dissolved Aβ40 toyield a solution with 10 equivalents of dye to peptide (30 μM dye and 30μM Aβ40). An equivalent volume of buffer solution was used in place of dyestock solution when Aβ40 was prepared in absence of dyes. Solutions wereincubated at 37oC and agitated at 200 rpm for 5 days. For the experimentsin figures 4.8 and 4.9, the initial amount of Aβ40 dissolved was increased toyield final concentrations of approximately 350 μM Aβ40 and 300 μM dye.CD data collection. Circular Dichroism (CD) data was collected on aJASCO J-815 CD spectrometer with an air cooled xenon arc lamp from190-250 nm. Due to high HT (high tension) voltage values (>500V), databelow 205 nm was not included in this thesis.XAS data collection and analysis. Br K-edge XAS data was collectedat the SSRL beamline 7-3 under ring conditions 500 mA at 3.0 GeV. Thisbeamline has a 20-pole, 2 T wigglers, 0.8 mrad beam, and a Si (220) dou-ble crystal monochromator that was detuned by 50% intensity to attainharmonic rejection. The incident X-ray intensity (I0), sample absorption(I1), and Br reference absorption (I2) were measured as transmittance usingargon-filled ionization chambers. Signal was detected with a 30-channel Gedetector. All data was measured at 20 ± 15 K within an Oxford InstrumentsCF1208 continuous-flow liquid helium cryostat. SixPack was used to cali-brate, average, and normalise XAS spectra [44]. KBr (E0 = 13474) spectraacquired at the same time as sample data was used as a calibrant for BrK-edge spectra.934.5. MethodsNMR sample preparation. An Aβ40 thin film was dissolved in dissolv-ing solution by sonication for 10 minutes. Buffer solution, deuterated water(D2O), and deuterated trifluoroethanol (d3-TFE) were added according totable 4.1.Aβ40 1 mgDissolving Solution 45 μLBuffer Solution 180 μLD2O 25 μLd3-TFE 250 μLTable 4.1: NMR sample preparation by amount of each component.NMR data collection. 1H-TOCSY (TOtal Correlation SpectroscopY)and 1H-NOESY (Nuclear Overhauser Effect SpectroscopY) experiments werecollected on an Bruker 850MHz instrument. Mixing times were 150 ms and60 ms for 1H-NOESY and 1H-TOCSY, respectively. 64 scans (1H-TOCSY)and 128 (1H-NOESY) were collected over a 8150 x 256 (F2 x F1) time do-main. A water suppression pulse sequence was used for both experiments.944.5. MethodsIn bio, it’s found frequently,That binding is helped by XB,But for amyloid beta,The XAS dataShows there’s no XB to see.95Chapter 5Conclusions and FutureDirectionsThe strength of this work lies in providing evidence of the partially covalentnature of XB and establishing a technique which is able to probe XB in awide variety of systems. In XB involving a halide electron donor, a diagnos-tic pre-edge feature in the XAS spectrum of the XB acceptor is indicativeof the XB. This feature requires the presence of XB as it corresponds to anelectron transition to the antibonding orbital of the XB. The weakness ofthis technique is that it requires a halide XB electron donor. Other commonXB donors, oxygen, sulphur, and nitrogen, for example, are more difficultto study with XAS due to atmospheric attenuation and self-absorption as-sociated with Z<16 elements.In an effort to establish a XB characterisation methodology which is uni-versally applicable, XAS of the XB electron acceptor was also investigated.As XB electron acceptors are, by definition, halogens, the problems associ-ated with XAS of lighter elements are avoided. We showed that XAS of XBelectron acceptors can also characterise XB. Although the XB-associatedchanges in the XAS spectra of XB electron acceptors are not as pronouncedas the changes in the XAS spectra of XB electron donors, diagnostic trendsare still observed. These trends result from electron transitions to the anti-bonding orbital of the covalently bound halogen changing as orbital overlapwith the electron donor changes.This ability to probe XB with XAS is illustrated by the work in Chapters2-4 of this thesis. In the halide trapping systems of chapter 2, XAS showsthat a XB of comparable strength to a HB exhibits charge transfer moresimilar to that of a coordinate bond (section 2.2). The charge transfer toeach halogen decreases as a single electron donor is involved in more inter-molecular interactions (section 2.4). Also in Chapter 2, probing for XB inKBr and I2 frozen solutions by XAS shows that this technique can explorethe stoichiometry of these bonds (section 2.3). Indeed, the Br– charge do-nation to I2 reaches a maximum when I2 is present in 1.3x stoichiometricexcess. These results suggest that while each Br– can be involved in multi-96Chapter 5. Conclusions and Future Directionsple XB, the overall charge donation from Br– cannot surpass 1.3 times thecharge donation of Br– involved in 1 XB.In Chapter 3, the ability of XAS to probe non-crystalline systems enablesus to look for XB in DSSC. DSSC are a particularly challenge for manyanalytical methods as key interactions occur at the solid-liquid interfacebetween the TiO2-bound dyes and the electrolyte. We have used XAS toconfirm the presence of XB in the DSSC and probe the relationship betweenXB covalency and regeneration rate. Although it was necessary to use Cl–in place of the I– electrolyte, our results still suggest that increasing XBcovalency increases dye regeneration rate and consequently overall DSSCefficiency.In Chapter 4, XAS is used to probe another challenging system forXB: the modulation of Alzheimers’-associated Aβ40 peptide aggregation byfluorescein-derivatives. This system again poses unique challenges for anal-ysis as the peptide is expected to aggregate. Like the DSSC, XAS allowsus to probe this systems despite its complex environment. In this system,the XAS data suggests that partially covalent XB are not present in thesesystems. Thus, the importance of halogen substituents in preventing ag-gregation does not arise from halogen bonding. Furthermore, TOCSY andNOESY NMR experiments are used to probe the role of the EOY in inhibit-ing Aβ40 peptide aggregation. This preliminary data supports the currentliterature findings that these dyes interact with the first half of the Aβ40peptide.The use of XAS to probe for XB is in its infancy, and there are manyinteresting systems still to be explored. The author has facilitated twofuture collaborations to be explored by the Kennepohl group with Dr. RobinPerutz and Dr. Marta Mosquera to use this technique to probe XB in othersystems. A recent publication from the Perutz group [76] explores the use offluorides involved in a metal-fluoride coordinate bond as electron donors inXB. This work investigates trends in XB geometry and strength using X-raycrystallography and 19F solid state NMR. The intended collaboration willexplore how charge transfer correlates with bond strength in these systemsby looking at the XAS of both the XB electron acceptors and donors.The collaboration initiated by the author with Dr. Mosquera aims toprobe for the presence of XB facilitating backbonding in nearby coordinatebonds. Recent work by Dr. Mosquera probed XB-based networks involvingruthenium complexes in the presence and absence of XB where the halideligands act as both the electron donor in the XB and in the coordinate bond[77]. In the all of the systems studied, with one exception, the ruthenium-halide bond distance was shown to increase in the presence of XB. This97Chapter 5. Conclusions and Future Directionsincrease in bond length due to the presence of the competing XB interaction.However, an unexpected decrease in the length of the ruthenium-bromidebond length is observed when the bromide ion also acts as an electron donorin a halogen bond with Br2. We propose this may be due to pi backbondingfrom the Ru to the σ∗XB orbital. As the σ∗XB orbital is perpendicular to theRu, it may allow the Br– to act as a pi acceptor ligand in the coordinate bond(figure 5.1). We propose to probe the nature of the interaction with Ru andBr K-edge X-ray absorption spectroscopy. If our hypothesis is correct, thiswill be the first example in the literature of XB facilitating pi backbondingin a coordinate bond involving a ligand which is traditionally σ/pi donor.Initial experiments have already been conducted for this project and theproject will be continued by the author’s co-worker.Figure 5.1: (a) XB network involving Br– and Br2 studied by Mosqueraand collaborators. XB networks involving different combinations of X– andX2 where X = Cl, Br, I were also studied [77]. (b) and (c) Depictions of theσ∗XB to illustrate that this orbital is approximately perpendicular to the Ru-Br coordinate bond. This orientation may facilitate an interaction betweenthe unoccupied σ∗XB and the occupied Rudxz orbital. This interaction mayenable the Br– to act as a pi acceptor in the coordinate bond, resulting in adecrease in Ru-Br bond length. For clarity, Ru and its orbitals have beendepicted in blue and the atoms and orbitals involved in the XB of interestare depicted in green.98Chapter 5. Conclusions and Future DirectionsIn the literature, XB are often described as non-covalent interactions[13, 25, 67]. This perceived lack of covalency means that their applicationsremain structural, i.e. XB in anion recognition [13, 16, 17, 26, 40], molecularself assembly [13, 23], and facilitating binding in biological systems [64–67].Our work in chapter 3 demonstrating the ability of these bonds to facilitatedye regeneration by electron transfer illustrates that the role XB play insystems is also electronic. It is important to consider the charge transferin these bonds to fully realise their applications. These bonds may alreadyplay important roles in catalysis and electron transfer systems without ourrealising, and understanding the role of the XB would allow us to furtheroptimise these systems. To change the perception in the literature we pro-pose the use of an arrow to represent XB. An arrow instead of a dottedline emphasises the charge transfer present in XB and could thus facilitatethe chemistry community to think of these bonds as weak coordinate bondswhich can exhibit σ/pi donor and acceptor characteristics rather than purelyelectrostatic interactions facilitating structural changes. Furthermore, XBdo in fact align with the IUPAC definition of a dative bond as it is a par-tially covalent bond involving in two atoms, “one of which serves as a donorand the other as an electron acceptor” and that “minimum-energy rupture...follows a heterolytic bond cleavage” [6].99Chapter 5. Conclusions and Future DirectionsFor two atoms involved in XB,Electrons are shared partially.Let’s not think too narrowAnd instead use an arrowTo emphasize covalency!100Bibliography[1] Timothy Clark, Matthias Hennemann, Jane S. Murray, and PeterPolitzer. Halogen bonding: the σ-hole. Journal of Molecular Model-ing, 13(2):291–296, 2007.[2] Timothy Clark. Halogen bonds and σ-holes. Faraday Discussions,203:9–27, 2017.[3] Thorsten Glaser, Britt Hedman, Keith O. Hodgson, and Edward I.Solomon. Ligand K-Edge X-ray Absorption Spectroscopy: A directprobe of ligand-metal covalency. Accounts of Chemical Research,33(12):859–868, 2000.[4] Edward I. Solomon, Britt Hedman, Keith O. Hodgson, Abhishek Dey,and Robert K. Szilagyi. Ligand K-edge X-ray absorption spectroscopy:covalency of ligand–metal bonds. Coordination Chemistry Reviews,249(1-2):97–129, 2005.[5] Abhishek Dey, Yonging Jiang, Paul Ortiz De Montellano, Keith O.Hodgson, Britt Hedman, and Edward I. Solomon. S K-edge XAS andDFT calculations on cytochrome P450: Covalent and ionic contribu-tions to the cysteine-Fe bond and their contribution to reactivity. Jour-nal of the American Chemical Society, 131(10):7869–7878, 2009.[6] IUPAC. Compendium of Chemical Terminology, 2nd ed. (the “GoldBook”). Compiled by A. D. McNaught and A. Wilkinson. BlackwellScientific Publications, Oxford (1997). XML on-line corrected ver-sion: http://goldbook.iupac.org (2006-) created by M. Nic, J. Jirat,B. Kosata; updates compiled by A. Jenkins. ISBN 0-9678550-9-8.https://doi.org/10.1351/goldbook. 2014.[7] Pierre Kennepohl and Elizabeth Gillis. CHEM208: Fundamentals ofCoordination Chemistry. University of British Columbia, first edition,2017.101Bibliography[8] Christer B. Aakero¨y, Paul D. Beer, Ngong Kodiah Beyeh, Lee Brammer,Mathieu Branca, David L. Bryce, Janet E. Del Bene, Alison J. Edwards,Mate Erdelyi, Catharine Esterhuysen, Marc Fourmigue´, Pierre Ken-nepohl, Lucia Myongwon Lee, Marta E.G. Mosquera, Jane S. Murray,Chantal L. Mustoe, William T. Pennington, Peter Politzer, Kevin E.Riley, Sergiy V. Rosokha, Steve Scheiner, Mark S. Taylor, Seiji Tsuzuki,Ignacio Vargas-Baca, and Yijue Xu. The halogen bond in solution: gen-eral discussion. Faraday Discussions, 203:347–370, 2017.[9] N.W. Alcock. Secondary bonding to nonmetallic elements. In Advancesin Inorganic Chemistry and Radiochemistry, volume 15, pages 1–58.1972.[10] Peter Politzer, Jane S. Murray, and Timothy Clark. Halogen bond-ing and other σ-hole interactions: a perspective. Physical ChemistryChemical Physics, 15(27):11178, 2013.[11] Robert H. Crabtree. Hypervalency, secondary bonding and hydro-gen bonding: siblings under the skin. Chemical Society Reviews,46(6):1720–1729, 2017.[12] Lee Brammer. Halogen bonding, chalcogen bonding, pnictogen bond-ing, tetrel bonding: origins, current status and discussion. FaradayDiscussions, 203:485–507, 2017.[13] Thomas M. Beale, Michael G. Chudzinski, Mohammed G. Sarwar, andMark S. Taylor. Halogen bonding in solution: thermodynamics andapplications. Chemical Society reviews, 42(4):1667–80, 2013.[14] Pierangelo Metrangolo, Franck Meyer, Tullio Pilati, Giuseppe Resnati,and Giancarlo Terraneo. Halogen bonding in supramolecular chemistry.Angewandte Chemie - International Edition, 47(33):6114–6127, 2008.[15] Benjamin R. Mullaney, Amber L. Thompson, and Paul D. Beer. Anall-Halogen bonding rotaxane for selective sensing of halides in aque-ous media. Angewandte Chemie - International Edition, 53(43):11458–11462, 2014.[16] Matthew J. Langton, Sean W. Robinson, Igor Marques, Vı´tor Fe´lix,and Paul D. Beer. Halogen bonding in water results in enhancedanion recognition in acyclic and rotaxane hosts. Nature Chemistry,6(12):1039–1043, 2014.102Bibliography[17] Nathan L. Kilah, Matthew D. Wise, Christopher J. Serpell, Amber L.Thompson, Nicholas G. White, Kirsten E. Christensen, and Paul D.Beer. Enhancement of anion recognition exhibited by a halogen-bonding rotaxane host system. Journal of the American Chemical So-ciety, 132(34):11893–11895, 2010.[18] Xin Ding, Matti Tuikka, and Matti Haukk. Halogen bonding in crystalengineering. In Recent Advances in Crystallography. InTech, 2012.[19] Arijit Mukherjee, Srinu Tothadi, and Gautam R. Desiraju. Halogenbonds in crystal engineering: Like hydrogen bonds yet different. Ac-counts of Chemical Research, 47(8):2514–2524, 2014.[20] Pierangelo Metrangolo and Giuseppe Resnati. Halogen bonding: Aparadigm in supramolecular shemistry. Chemistry - A European Jour-nal, 7(12):2511–2519, 2001.[21] Angelika Bruckmann, Miguel Pena, and Carsten Bolm. Organocatalysisthrough halogen-bond activation. Synlett, 2008(6):900–902, 2008.[22] Florian Kniep, Stefan H. Jungbauer, Qi Zhang, Sebastian M. Walter,Severin Schindler, Ingo Schnapperelle, Eberhardt Herdtweck, and Ste-fan M. Huber. Organocatalysis by neutral multidentate halogen-bonddonors. Angewandte Chemie - International Edition, 52(27):7028–7032,2013.[23] Pierangelo Metrangolo and Giuseppe Resnati. Chemistry: Halogenversus hydrogen. Science, 321(5891):918–919, 2008.[24] Marc Fourmigue´. Halogen bonding: Recent advances. Current Opinionin Solid State and Materials Science, 13(3-4):36–45, 2009.[25] Peter Politzer, Pat Lane, Monica C. Concha, Yuguang Ma, and Jane S.Murray. An overview of halogen bonding. Journal of Molecular Mod-eling, 13(2):305–311, 2007.[26] Benjamin R. Mullaney, Benjamin E. Partridge, and Paul D. Beer.A halogen-bonding bis-triazolium rotaxane for halide-selective anionrecognition. Chemistry - A European Journal, 21(4):1660–1665, 2015.[27] Sean W. Robinson, Chantal L. Mustoe, Nicholas G. White, AshaBrown, Amber L. Thompson, Pierre Kennepohl, and Paul D. Beer.103BibliographyEvidence for halogen bond covalency in acyclic and interlocked halogen-bonding receptor anion recognition. Journal of the American ChemicalSociety, 137(1):499–507, 2015.[28] Steve Scheiner. Comparison of halide receptors based on H, halogen,chalcogen, pnicogen, and tetrel bonds. Faraday Discussions, 203:213–226, 2017.[29] Anthony C. Legon and D.J. Millen. Angular geometries and other prop-erties of hydrogen-bonded dimers: a simple electrostatic interpretationof the success of the electron-pair model. Chemical Society Reviews,16:467, 1987.[30] J.R. Schmidt and W.F. Polik. WebMO Enterprise, version 13.0, 2013.[31] Sergiy V. Rosokha, Charlotte L. Stern, and Jeremy T. Ritzert. Ex-perimental and computational probes of the nature of halogen bond-ing: Complexes of bromine-containing molecules with bromide anions.Chemistry - A European Journal, 19(27):8774–8788, 2013.[32] Sebastiaan B. Hakkert, Ju¨rgen Gra¨fenstein, and Mate Erdelyi. The 15NNMR chemical shift in the characterization of weak halogen bonding insolution. Faraday Discussions, 203:333–346, 2017.[33] Albert Thompson, David Attwood, Eric Gullikson, Malcolm Howells,Kwang-Je Kim, Janos Kirz, Jeffrey Kortright, Ingolf Lindau, Yan-wei Liu, Piero Pianetta, Arthur Robinson, James Scofield, James Un-derwood, Gwyn Williams, and Herman Winick. X-ray Data Booklet.Lawrence Berkeley National Laboratory, Berkeley, CA, USA, 2005.[34] Susan E. Shadle, Britt Hedman, Keith O. Hodgson, and Edward I.Solomon. Ligand K-Edge X-ray Absorption Spectroscopy as a probeof ligand-metal bonding: Charge donation and covalency in copper-chloride systems. Inorganic Chemistry, 33(19):4235–4244, 1994.[35] Edward I. Solomon. Final comments. Comments on Inorganic Chem-istry, 3(5):319–320, 1984.[36] Fraser G.L. Parlane, Chantal L. Mustoe, Cameron W. Kellett,Sarah J.C. Simon, Wesley B. Swords, Gerald J. Meyer, Pierre Ken-nepohl, and Curtis P. Berlinguette. Spectroscopic detection of halogenbonding resolves dye regeneration in the dye-sensitized solar cell. Na-ture Communications, 8(1):1761, 2017.104Bibliography[37] H. Edward Wong, Jacob A. Irwin, and Inchan Kwon. Halogenationgenerates effective modulators of Amyloid-Beta aggregation and neuro-toxicity. PLoS ONE, 8(2):e57288, 2013.[38] Chantal L. Mustoe, Mathusan Gunabalasingam, Darren Yu, Brian O.Patrick, and Pierre Kennepohl. Probing covalency in halogen bondsthrough donor K-edge X-ray absorption spectroscopy: polyhalides ascoordination complexes. Faraday Discussions, 203:79–91, 2017.[39] Wesley B. Swords, Sarah J.C. Simon, Fraser G.L. Parlane, Rebecca K.Dean, Cameron W. Kellett, Ke Hu, Gerald J. Meyer, and Curtis P.Berlinguette. Evidence for interfacial halogen bonding. AngewandteChemie - International Edition, 55(20):5956–5960, 2016.[40] James M. Mercurio, Richard C. Knighton, James Cookson, and Paul D.Beer. Halotriazolium axle functionalised [2]rotaxanes for anion recog-nition: Investigating the effects of halogen-bond donor and preorgani-sation. Chemistry - A European Journal, 20(37):11740–11749, 2014.[41] Uwe Bergmann, Pieter Glatzel, and Stephen P. Cramer. Bulk-sensitiveXAS characterization of light elements: from X-ray Raman scatteringto X-ray Raman spectroscopy. Microchemical Journal, 71(2-3):221–230,2002.[42] Michael J. Hynes. EQNMR: a computer program for the calculationof stability constants from nuclear magnetic resonance chemical shiftdata. Journal of the Chemical Society, Dalton Transactions, (2):311,1993.[43] Daniel C. Harris. Quantitative chemical analysis. WH Freeman andCompany, 6th edition, 2003.[44] Sam M. Webb. SIXPack a graphical user interface for XAS analysisusing IFEFFIT. Physica Scripta, T115(0001):1011, 2005.[45] Mario Ulises Delgado-Jaime and Pierre Kennepohl. Development andexploration of a new methodology for the fitting and analysis of XASdata. Journal of Synchrotron Radiation, 17(1):119–128, 2010.[46] Mario Ulises Delgado-Jaime and Pierre Kennepohl. Development andexploration of a new methodology for the fitting and analysis of XASdata. Journal of Synchrotron Radiation, 17(1):119–128, 2010.105Bibliography[47] Per H. Svensson and Lars Kloo. Synthesis, structure, and bond-ing in polyiodide and metal iodideiodine systems. Chemical Reviews,103(5):1649–1684, 2003.[48] Sarah J.C. Simon, Fraser G.L. Parlane, Wesley B. Swords, Cameron W.Kellett, Chuan Du, Brian Lam, Rebecca K. Dean, Ke Hu, Gerald J.Meyer, and Curtis P. Berlinguette. Halogen bonding promotes higherdye-sensitized solar cell photovoltages. Journal of the American Chem-ical Society, 138(33):10406–10409, 2016.[49] Michael Gra¨tzel. Photoelectrochemical cells. Nature, 414(6861):338–344, 2001.[50] Thanh Binh Phan, Martin Breugst, and Herbert Mayr. Towards ageneral scale of nucleophilicity? Angewandte Chemie - InternationalEdition, 45(23):3869–3874, 2006.[51] Gerrit Boschloo and Anders Hagfeldt. Characteristics of the io-dide/triiodide redox mediator in dye-sensitized solar cells. Accountsof Chemical Research, 42(11):1819–1826, 2009.[52] Peter Atkins and Julio de Paula. Physical Chemistry: Thermodynamics,Structure, and Change. Oxford University Press, 10th edition, 2014.[53] Ma´te´ Erde´lyi. Halogen bonding in solution. Chemical Society Reviews,41(9):3547, 2012.[54] Thai Thanh Thu Bui, Slimane Dahaoui, Claude Lecomte, Gautam R.Desiraju, and Enrique Espinosa. The nature of halogen...halogen in-teractions: a model derived from experimental charge-density analysis.Angewandte Chemie - International Edition, 48(21):3838–41, 2009.[55] Gabriella Cavallo, Pierangelo Metrangolo, Roberto Milani, Tullio Pilati,Arri Priimagi, Giuseppe Resnati, and Giancarlo Terraneo. The halogenbond, 2016.[56] Mario Ulises Delgado-Jaime, Jay C. Conrad, Deryn E. Fogg, and PierreKennepohl. X-ray absorption methods for the determination of Ru-Clbond covalency in olefin metathesis catalysts: On the normalization ofchlorine K-edges in ruthenium complexes. Inorganica Chimica Acta,359(9):3042–3047, 2006.106Bibliography[57] P. Auffinger, F. A. Hays, E. Westhof, and P. S. Ho. Halogen bonds inbiological molecules. Proceedings of the National Academy of Sciences,101(48):16789–16794, 2004.[58] A. Wojtczak, V. Cody, J.R. Luft, and W. Pangborn. Structure ofrat transthyretin (rTTR) complex with thyroxine at 2.5 A resolution:first non-biased insight into thyroxine binding reveals different hormoneorientation in two binding sites. Acta crystallographica. Section D, Bi-ological crystallography, 57(Pt 8):1061–70, 2001.[59] Karl-Heinz van Pe´e and Susanne Unversucht. Biological dehalogenationand halogenation reactions. Chemosphere, 52(2):299–312, 2003.[60] I. Hendrikje Buss, Revathy Senthilmohan, Brian A. Darlow, NinaMogridge, Anthony J. Kettle, and Christine C. Winterbourn. 3-Chlorotyrosine as a marker of protein damage by myeloperoxidase intracheal aspirates from preterm infants: association with adverse res-piratory outcome. Pediatric Research, 53(3):455–462, 2003.[61] Weijia Wu, Michael K. Samoszuk, Suzy A.A. Comhair, Mary JaneThomassen, Carol F. Farver, Raed A. Dweik, Mani S. Kavuru, Serpil C.Erzurum, and Stanley L. Hazen. Eosinophils generate brominating ox-idants in allergen-induced asthma. Journal of Clinical Investigation,105(10):1455–1463, 2000.[62] Z. Shen, S.N. Mitra, W. Wu, Y. Chen, Y. Yang, J. Qin, and Stanley L.Hazen. Eosinophil peroxidase catalyzes bromination of free nucleosidesand double-stranded DNA. Biochemistry, 40(7):2041–51, 2001.[63] P. Shing Ho, Jeffrey M. Vargason, and Brandt F. Eichman. No Title.Nature Structural Biology, 7(9):758–761, 2000.[64] Rainer Wilcken, Markus O. Zimmermann, Andreas Lange, Andreas C.Joerger, and Frank M. Boeckler. Principles and applications of halo-gen bonding in medicinal chemistry and chemical biology. Journal ofMedicinal Chemistry, 56(4):1363–1388, 2013.[65] Leo A. Hardegger, Bernd Kuhn, Beat Spinnler, Lilli Anselm, RobertEcabert, Martine Stihle, Bernard Gsell, Ralf Thoma, Joachim Diez,Jo¨rg Benz, Jean-Marc Plancher, Guido Hartmann, Yoshiaki Isshiki,Kenji Morikami, Nobuo Shimma, Wolfgang Haap, David W. Banner,107Bibliographyand Franc¸ois Diederich. Halogen bonding at the active sites of Hu-man Cathepsin L and MEK1 Kinase: Efficient interactions in differentenvironments. ChemMedChem, 6(11):2048–2054, 2011.[66] Leo A. Hardegger, Bernd Kuhn, Beat Spinnler, Lilli Anselm, RobertEcabert, Martine Stihle, Bernard Gsell, Ralf Thoma, Joachim Diez,Jo¨rg Benz, Jean-Marc Plancher, Guido Hartmann, David W. Banner,Wolfgang Haap, and Franc¸ois Diederich. Systematic investigation ofhalogen bonding in protein-ligand interactions. Angewandte Chemie -International Edition, 50(1):314–318, 2011.[67] Rainer Wilcken, Xiangrui Liu, Markus O. Zimmermann, Trevor J.Rutherford, Alan R. Fersht, Andreas C. Joerger, and Frank M. Boeck-ler. Halogen-enriched fragment libraries as leads for drug rescue ofmutant p53. Journal of the American Chemical Society, 134(15):6810–6818, 2012.[68] George G. Glenner and Caine W. Wong. Alzheimer’s disease andDown’s syndrome: Sharing of a unique cerebrovascular amyloid fib-ril protein. Biochemical and Biophysical Research Communications,122(3):1131–1135, 1984.[69] J. Hardy. The Amyloid hypothesis of Alzheimer’s Disease: Progressand problems on the road to therapeutics. Science, 297(5580):353–356,2002.[70] Michael G. Zagorski and Colin J. Barrow. NMR studies of amyloid.beta.-peptides: proton assignments, secondary structure, and mecha-nism of an .alpha.-helix .fwdarw. .beta.-sheet conversion for a homol-ogous, 28-residue, N-terminal fragment. Biochemistry, 31(24):5621–5631, 1992.[71] Jacob A. Irwin, H. Edward Wong, and Inchan Kwon. Determiningbinding sites of polycyclic aromatic small molecule-based amyloid-betapeptide aggregation modulators using sequence-specific antibodies. An-alytical Biochemistry, 470:61–70, 2015.[72] Joon Seok Lee, Byung Il Lee, and Chan Beum Park. Photo-inducedinhibition of Alzheimer’s β-amyloid aggregation in vitro by rose bengal.Biomaterials, 38:43–49, 2015.108[73] Heinrich Sticht, Peter Bayer, Dieter Willbold, Sonja Dames, CarolineHilbich, Konrad Beyreuther, Rainer W Frank, and Paul Rosch. Struc-ture of amyloid A 4 4 1-4O ) -peptide of Alzheimer ’ s disease. Bio-chemistry, 298:293–298, 1995.[74] Orlando Crescenzi, Simona Tomaselli, Remo Guerrini, Severo Sal-vadori, Anna M. D’Ursi, Piero Andrea Temussi, and Delia Picone. So-lution structure of the Alzheimer amyloid β-peptide (1-42) in an apolarmicroenvironment. European Journal of Biochemistry, 269(22):5642–5648, 2002.[75] Subramanian Vivekanandan, Jeffrey R. Brender, Shirley Y. Lee, andAyyalusamy Ramamoorthy. A partially folded structure of amyloid-beta(1–40) in an aqueous environment. Biochemical and BiophysicalResearch Communications, 411(2):312–316, 2011.[76] Vargini Thangavadivale, Pedro M. Aguiar, Naseralla A. Jasim, Sarah J.Pike, Dan A. Smith, Adrian C. Whitwood, Lee Brammer, and Robin N.Perutz. Self-complementary nickel halides enable multifaceted compar-isons of intermolecular halogen bonds: fluoride ligands vs. other halides.Chemical Science, 9(15):3767–3781, 2018.[77] Marta E. G. Mosquera, Irene Egido, Carlos Hortelano, Mar´ıa Lo´pez-Lo´pez, and Pilar Go´mez-Sal. Comparison of halogen bonding networkswith Ru complexes and analysis of the influence of the XB interactionson their reactivity. Faraday Discussions, 203:257–283, 2017.109Appendix ABlueprintXAS ParametersIn this appendix are the final averages of all fitting parameters from Blueprint-XAS for 1-Cl, 2-Cl, 3-Cl and 4-Cl and the Dye-X series. As the fits for1-Br, 2-Br, 3-Br and 4-Br were deemed inconclusive, the final fitting pa-rameters for these spectra are not included.General method for setting parameter restrictions in fits:• E0 range/spline lower limit: energy of the edge peak +5 eV• Peak intensity ranges (I1 and I2): limits must be sufficiently large toavoid truncating solution distribution• Energy of peak: ±1-2 eV from visual peak maximum• % Gaussian contribution to shape of peak (G#) 40-60%• Background lower limit: approximately 20 eV range below E0110Appendix A. BlueprintXAS ParametersCoefficient Name Identifier (ID) Average Standard Dev(Spline & edge peak fitting)Degree-Pre-edge M1 2.0 0.6Degree-Spline1 N1 4 1Degree-Spline2 P1 4 1E0 (edge-Spline) E0 2820.9 0.5Width (edge) W1 0.9 0.2Gaussian (edge) G1 0.57 0.07Pre-edge, Max B1 2813 1Spline, Min C1 2823 1Intensity I1 2.3 0.4Position O1 2819.8 0.2Width W2 1.7 0.2Gaussian, fraction G2 0.6 0(Pre-edge peak fitting)Intensity I2 0.9 0.1Position O2 2816.88 0.01Gaussian, fraction G3 0.6 0Width W3 0.56 0.01Table A.1: XAS fit parameters for 1-Cl111Appendix A. BlueprintXAS ParametersCoefficient Name Identifier (ID) Average Standard Dev(Spline & edge peak fitting)Degree-Pre-edge M1 2.0 0.6Degree-Spline1 N1 3.1 0.6Degree-Spline2 P1 4 1E0 (edge-Spline) E0 2823 1Width (edge) W1 1.3 0.4Gaussian (edge) G1 0.6 0Pre-edge, Max B1 2812 2Spline, Min C1 2823 1Intensity I1 5.7 0.9Position O1 2820.4 0.2Gaussian, fraction G2 0.9 0.1Width W2 1.9 0.3(Pre-edge peak fitting)Intensity I2 2.0 0.5Position O2 2817.33 0.03Gaussian, fraction G3 0.51 0.09Width W3 0.72 0.04Table A.2: XAS fit parameters for 2-Cl112Appendix A. BlueprintXAS ParametersCoefficient Name Identifier(ID) Average Standard Dev(Spline & edge peak fitting)Degree-Pre-edge M1 2.0 0.6Degree-Spline1 N1 4 1Degree-Spline2 P1 4 1E0 (edge-Spline) E0 2821.7 0.5Width (edge) W1 0.8 0.2Gaussian (edge) G1 0.56 0.08Pre-edge, Max B1 2812 1Spline, Min C1 2823 1Intensity I1 2.5 0.6Position O1 2820.6 0.2Width W2 1.5 0.3Gaussian, fraction G2 0.6 0(Pre-edge peak fitting)Intensity I2 0.9 0.3Position O2 2818.4 0.2Gaussian, fraction G3 0.44 0.05Width W3 1.0 0.6Table A.3: XAS fit parameters for 3-Cl113Appendix A. BlueprintXAS ParametersCoefficient Name Identifier (ID) Average Standard Dev(Spline & edge peak fitting)Degree-Pre-edge M1 2.0 0.5Degree-Spline1 N1 3.1 0.6Degree-Spline2 P1 4 1E0 (edge-Spline) E0 2820.9 0.2Width (edge) W1 0.8 0.4Gaussian (edge) G1 0.6 0Pre-edge, Max B1 2814 2Spline, Min C1 2823 1Intensity I1 0.42 0.02Position O1 2820.1 0.1Width W2 1.56 0.08Gaussian, fraction G2 0.6 0(Pre-edge peak fitting)Intensity I2 0.11 0.06Position O2 2816.2 0.1Gaussian, fraction G3 0.58 0.05Width W3 0.9 0.9Table A.4: XAS fit parameters for 4-Cl114Appendix A. BlueprintXAS ParametersCoefficient Name Identifier(ID) Average Standard Dev(Spline & edge peak fitting)Degree-Pre-edge M1 2.4 0.9Degree-Spline1 N1 2.6 0.8E0(edge-Spline) E0 2826.9 0.7hwhm W1 4.55 0.7Pre-edge, Max B1 2811 3Spline, Min C1 2828 1Intensity I1 0.18 0.01Position O1 2824.70 0.06hwhm W2 2.78 0.05Gaussian fraction G2 0.92 0.07(Pre-edge peak fitting)Intensity I2 0.0027 0.0004Position O2 2818.7 1E-12hwhm W3 0.72 0.04Gaussian fraction G3 1.0 0.1Table A.5: XAS fit parameters for NOBF4 + Cl–Coefficient Name Identifier(ID) Average Standard Dev(Spline & edge peak fitting)Degree-Pre-edge M1 2.39 0.9Degree-Spline1 N1 2.49 0.9E0(edge-Spline) E0 2825.6 0.7hwhm W1 4.6 0.7Pre-edge, Max B1 2811 3Spline, Min C1 2827 2Intensity I1 0.131 0.005Position O1 2824.40 0.07hwhm W2 2.50 0.03Gaussian fraction G2 0.97 0.07(Pre-edge peak fitting)Intensity I2 0.005 0.001Position O2 2818.70 3E-12hwhm W3 1.00 9E-15Gaussian fraction G3 0.7 0.3Table A.6: XAS fit parameters for Dye-F+. ← Cl–115Appendix A. BlueprintXAS ParametersCoefficient Name Identifier(ID) Average Standard Dev(Spline & edge peak fitting)Degree-Pre-edge M1 2.4 0.9Degree-Spline1 N1 2.6 0.9E0(edge-Spline) E0 2825.40 0.8hwhm W1 4.8 0.2Pre-edge, Max B1 2810 3Spline, Min C1 2827 1Intensity I1 0.33 0.02Position O1 2824.60 0.02hwhm W2 2.07 0.05Gaussian fraction G2 0.60 9E-15(Pre-edge peak fitting)Intensity I2 0.029 0.002Position O2 2818.40 0.03hwhm W3 1.24 0.03Gaussian fraction G3 0.60 2E-12Table A.7: XAS fit parameters for Dye-Br+. ← Cl–Coefficient Name Identifier(ID) Average Standard Dev(Spline & edge peak fitting)Degree-Pre-edge M1 2.6 0.9Degree-Spline1 N1 2.3 0.8E0(edge-Spline) E0 2824.8 0.2hwhm W1 5.7 0.3Pre-edge, Max B1 2811 3Spline, Min C1 2827 2Intensity I1 0.18 0.01Position O1 2824.60 0.02hwhm W2 2.39 0.05Gaussian fraction G2 0.46 0.07(Pre-edge peak fitting)Intensity I2 0.0128 0.0008Position O2 2818.60 0.003hwhm W3 0.68 0.02Gaussian fraction G3 0.4 2E-12Table A.8: XAS fit parameters for Dye-I+. ← Cl–116Appendix BPython ProgramsThis appendix consists of a list of the python programs written by me andused to aid in workup of the Aβ peptide NMR data. The code for theseprograms is reported here with syntax highlighting using the opensourceLaTeX package “minted”. A guide for the syntax highlighting is below.# This is a comment within a program. Comments give the user a# rough idea of how a program works.’’’ 3 apostrophes also denotes a comment ’’’"One or two apostrophes denotes a string of text. This is a string"3 # Numbers and numerical operators are in grayprint "this string" # Purple denotes Python specific commandssome_variable = 3 # Black denotes user defined# variables and their propertiesdef function: # Light green denotes user defined functionsB.1 TOCSY Program - finding spin systemsThe purpose of the TOCSY program is to find all possible amino acid spinsystems using a list of the peaks picked by the user in Topspin software. Theprogram uses the peak list file generated by exporting the peak list fromTopspin. Changing the numerical values for the same pk and same syschanges how close in (ppm) two peaks must be to be considered as twopeaks which correspond to the same proton ( same pk) or to be identifiedas within the same spin system ( same sys).#Extracts Spin Systems from TOCSY Peaksimport sys117B.1. TOCSY Program - finding spin systemswith open("TOCSY_2_2_peak.txt", "r") as file:i = 0pk_mtrx_0 = list()for line in file:if i > 7:peak = [float(num) for num in line.split()]pk_mtrx_0.append(peak)i += 1’’’Change the _same_pk and _same_sys to changehow close in (ppm) two peaks must be to beconsidered as the same peak or to be identified aswithin the same spin system, respectively.’’’_same_pk = 0.011_same_sys_0 = 0.007file1 = open("generated_spins.txt", "w")’’’## Use the following 3 lines for debugging ##for get_pk in pk_mtrx:print get_pk[3],get_pk[4]sys.exit(0)’’’def get_spin_sys(pk_mtrx,_same_sys):spin_sys = list()# Get individual peaks from pk_mtrx as list (get_pk)for i in range(len(pk_mtrx)):get_pk = pk_mtrx[i]pk = [get_pk[3],get_pk[4]]pk_in_sys = list()pk_sys_list = list()# Inputs first peak as first spin systemif i == 0: spin_sys.append(pk)# Get individual spin systems from spin_sys as list(get_sys)elif i > 0:118B.1. TOCSY Program - finding spin systemsnew_sys = Truefor j in range(len(spin_sys)):get_sys = spin_sys[j]# Checks if current peak (get_pk) is in# current spin system (get_sys)if any([abs(get_sys[h]-pk[0]) <= _same_sysfor h in range(len(get_sys))])or any([abs(get_sys[g]-pk[1])<= _same_sys for g in range(len(get_sys))]):pk_in_sys.append(j)pk_sys_list.append(get_sys)new_sys = False# If peak is in 1 or more spin system, this code creates new# spin sytem with peak and all spin systems which include# peakif new_sys == False:get_sys = spin_sys[pk_in_sys[0]]for n in range(len(pk)):if all([abs(get_sys[p]-pk[n]) >= _same_sysfor p in range(len(get_sys))]):get_sys.append(pk[n])if len(pk_in_sys) >= 2:for m in range(1,len(pk_in_sys)):check_sys = spin_sys[pk_in_sys[m]]for n in range(len(check_sys)):if all([abs(get_sys[p]-check_sys[n])>= _same_sys for p inrange(len(get_sys))]):get_sys.append(check_sys[n])#spin_sys.remove(spin_sys[pk_in_sys[m]])for i in range(len(pk_sys_list)):if i > 0:del_sys = pk_sys_list[i]spin_sys.remove(del_sys)spin_sys[pk_in_sys[0]] = get_sys# If new peak is not in any spin system,# create new spin systemif new_sys == True: spin_sys.append(pk)return spin_sys119B.1. TOCSY Program - finding spin systems# Creates matrix of spin systems which contain# too many values (i.e. biologically impossible)def too_long(spin_sys):long_sys = list()for sys in spin_sys_0:if len(sys) > 12: long_sys.append(sys)# "12" = max # of protons in an AA in Abeta40return long_sys# Identify which peaks belong to unreasonably# large (len>12) spin systemsdef extract_pks(pk_mtrx, long_sys, _same_sys):new_pk_mtrx = list()for sys in long_sys:for spin in sys:for pk in pk_mtrx:if any([abs(spin-pk[x]) <= _same_sysfor x in range(3,5)]):new_pk_mtrx.append(pk)return new_pk_mtrx# Prints # of spin system and sorted spin_sys# (from greatest to least), no return valuedef print_sys(spin_sys):sys_count = 0for j in range(len(spin_sys)):spin_sys[j].sort(reverse = True)if len(spin_sys[j]) > 2: sys_count +=1spin_sys.sort()for sys in spin_sys:print sysif len(sys) < 13 and len(sys) > 2:for s in sys:file1.write(str(s)+" ")file1.write("\n")print "number of spin systems which don’t" +" correspond to a single peak:" +str(sys_count)spin_sys_0 = get_spin_sys(pk_mtrx_0,_same_sys_0)print_sys(spin_sys_0)120B.2. NOESY Program - finding neighbouring spin systems# Looks for unreasonably large spin systems,# extracts peaks and finds spin systems for# progressively smaller allowed difference between# two peak valuesnew_spin_sys = list()check_sys = spin_sys_0check_val = _same_sys_0while any(len(sys) > 12 for sys in check_sys) and check_val > 0:check_sys = too_long(check_sys)redo_pks = extract_pks(pk_mtrx_0, check_sys, check_val)check_val -= 0.001check_sys = get_spin_sys(redo_pks, check_val)print len(check_sys)for sys in check_sys: sys.sort(reverse = True)for sys in check_sys:if len(sys) <= 12 and any(nsys == sys for nsysin new_spin_sys) == False:new_spin_sys.append(sys) #appends sys to new_spin_sysiff sys is not in new_spin_sys alreadyprint_sys(new_spin_sys)file1.close()B.2 NOESY Program - finding neighbouring spinsystemsThe purpose of the NOESY program is to read in a file containing the spinsystems identified by the TOCSY program in the previous section, and thendetermine which of those spin systems must be near enough to generatea peak in the NOESY spectrum. To use the spin systems generated bythe previous program, copy/paste the command line output of the TOCSYprogram into a text file named ”generated spins.txt”.#2D NMR data processingimport matplotlib.pyplot as plt’’’’generate_spins.txt is a text file of the spins system printed121B.2. NOESY Program - finding neighbouring spin systemsby the TOCSY program in the previous program’’’with open("generated_spins.txt", "r") as file0:i = 0spin_sys = list()for line in file0:if i > 0:shifts = [float(num) for num in line.split()]spin_sys.append(shifts)i += 1##reads in peaklist generated by Topspin for NOESY datawith open("NOESY_1_1.txt", "r") as file:i = 0pk_mtrx = list()for line in file:if i > 0:peak = [float(num) for num in line.split()]pk_mtrx.append(peak)i += 1’’’The aim of the following code is to find all peaks in theNOESY spectrum which belong to a spin system identifiedin the TOCSY spectrum by doing the following:The following code finds F2 values for NOESY peakswhich correspond to a proton in one of the TOCSY spinsystems identified.The peaks of interest are added to "pks1"Each entry of "pks1":[corresponding TOCSY spin system #,NOESY F2 value, NOESY F1 value]The final value for x is the number of peaksin the NOESY spectrum which correspondto a TOCSY spin system.122B.2. NOESY Program - finding neighbouring spin systems’’’x = 0pks1 = list()for i in range(len(pk_mtrx)):get_pk = pk_mtrx[i]for j in range(len(spin_sys)):get_sys = spin_sys[j]for h in range(len(get_sys)):if abs(get_pk[1]-get_sys[h]) <= 0.01:print j+2,(get_pk[1],get_pk[2])pk = [j+2, get_pk[1],get_pk[2]]pks1.append(pk)x += 1print x’’’The aim of the following code is to find neighbouring spinsystems by identifying all peaks in the NOESY spectrumwith both F2 and F1 values (i.e. chemical shifts)corresponding to spin systems identified in the TOCSYspectrum by doing the following:The code searches through the peaks in "pks1" to identifywhich peaks also have an F1 value (chemical shift)corresponding to an identified spin system.The peaks of interest are added to "pks2"Each entry of "pks2":[F2’s corresponding TOCSY spin system,F1’s corresponding TOCSY spin system,NOESY F2 value, NOESY F1 value]The final value of y is the number of peaks in the NOESYspectrum where both F2 and F1 values correspondto a COSY spin system.’’’neighbours = list()y = 0pks2 = list()123B.3. Generate TOCSY Spectrum Programfor i in range(len(pks1)):get_pk = pks1[i]for j in range(len(spin_sys)):get_sys = spin_sys[j]for h in range(len(get_sys)):if abs(get_pk[2]-get_sys[h]) <= 0.005:get_pk.insert(0, j+2)print get_pkneighbours.append([get_pk[0],get_pk[1]])get_pk = get_pk[1:]y += 1print y’’’The following code graphs the peaks in theNOESY spectrum which give distance restraintinformation about two of the identified spinsystems in the TOCSY data.’’’AA1,AA2 = zip(*neighbours)plt.scatter(AA1,AA2, color = "b")plt.scatter(AA2,AA1, color = "b")plt.show()B.3 Generate TOCSY Spectrum ProgramThis program compares user acquired TOCSY NMR data with NMR datapublished on the Protein Data Bank (PDB). The data from the PDB isreported as chemical shift values for each proton. This program will usethese chemical shift values to generate the expected TOCSY spectrum forthe published values. This code can be used to compare TOCSY NMR datafor any peptide with any available data for the same or a similar peptidepublished on PDB (provided the proton chemical shifts are reported). Thecode is currently generates a TOCSY spectrum for any human Aβ40 peptidePDB data, but to use this for another peptide, simply replace the Aβ40 one-letter amino acid sequence with that of the desired peptide. The file nameswill also need to match the user’s filenames for the PDB chemical shift fileand the TOCSY peaklist exported from Topspin.124B.3. Generate TOCSY Spectrum Program# Generate TOCSY spectra for Abeta40# peptide from BRMB 17764 (published on PDB)import matplotlib.pyplot as pltimport matplotlib.colors as colorsimport matplotlib.cm as cmxfrom collections import defaultdict# Input one letter code sequence for peptide of interest herepeptide = "DAEFRHDSGYEVHHQKLVFFAEDVGSNKGAIIGLMVGGVV"jet = cm = plt.get_cmap("jet")values = range(len(peptide))cNorm = colors.Normalize(vmin=0, vmax=values[-1])scalarMap = cmx.ScalarMappable(norm=cNorm,cmap=jet)print scalarMap.get_clim()tocsy_pts = defaultdict(list)# Opens file downloaded from PDB store info in chem_shiftswith open("PDB_shifts.txt", "r") as file0:i = 0chem_shifts = list()for line in file0:if i > 62:atom_info = [float(value) for value in line.split()]chem_shifts.append(atom_info[5:11])i += 1’’’Generates dictionary ’spin_sys’ with chemicalshifts of each AA (key is 3 letter AA code)’’’spin_sys = defaultdict(list)for i in range(len(chem_shifts)):atom = chem_shifts[i]if len(atom) > 1:key = atom[1]+atom[0]125B.3. Generate TOCSY Spectrum Programif "H" in str(atom[3]):spin_sys[key].append(float(atom[5]))’’’Generates dictionary with all F2,F1 points forTOCSY spectra (key is 3 letter AA code)’’’for key in spin_sys:aa = spin_sys[key]f2f1 = list()for i in range(len(aa)-1):for j in range(len(aa)):if j > i: f2f1.append((aa[i],aa[j]))tocsy_pts[key].append(f2f1)def plot_AA(ax1, aa, key, aa_loc):key_loc = key + str(aa_loc+1)colorVal = scalarMap.to_rgba(values[aa_loc])f2f1 = list()f2f1.append(tocsy_pts[key_loc])F1 = []F2 = []for list_j in f2f1:for list_k in list_j:for point in list_k:F2.append(point[0])F1.append(point[1])ax1.scatter(F2,F1, label = aa + str(i+1),color = colorVal, s = 10+10/(aa_loc+1))ax1.scatter(F1,F2, color = colorVal, s = 10+10/(aa_loc+1))’’’Returns data for making invisible mouse over plot’’’label = aa + str(i+1)xdata_add = F2xdata_add = xdata_add + F1 #adds points below the diagonalydata_add = F1ydata_add = ydata_add + F2 #adds points below the diagonalreturn xdata_add, ydata_add, label’’’Defines mouse over event’’’126B.3. Generate TOCSY Spectrum Programdef hover(event):vis = annot.get_visible()if event.inaxes == ax1:cont, ind = sc.contains(event)if cont:update_annot(ind)annot.set_visible(True)fig.canvas.draw_idle()else:if vis:annot.set_visible(False)fig.canvas.draw_idle()def update_annot(ind):pos = sc.get_offsets()[ind["ind"][0]]annot.xy = posfor n in ind[’ind’]: text = m_label[n]#text = "{}, {}".format(" ".join(list(map(str,ind["ind"])))," ".join([m_label[n] for n in ind["ind"]]))annot.set_text(text)#annot.get_bbox_patch().set_facecolor(cmap(norm(c[ind["ind"][0]])))#annot.get_bbox_patch().set_alpha(0.4)fig, ax1 = plt.subplots()xdata = list()ydata = list()m_label = list()for i in range(len(peptide)):letter = peptide[i]if letter == "A":xtemp, ytemp, ltemp = plot_AA(ax1, "A", "ALA", i)for t in xtemp: xdata.append(t)for t in ytemp: ydata.append(t)for i in range(len(xtemp)): m_label.append(ltemp)elif letter == "C":127B.3. Generate TOCSY Spectrum Programxtemp, ytemp, ltemp = plot_AA(ax1, "C", "CYS", i)for t in xtemp: xdata.append(t)for t in ytemp: ydata.append(t)for i in range(len(xtemp)): m_label.append(ltemp)elif letter == "D":xtemp, ytemp, ltemp = plot_AA(ax1, "D", "ASP", i)for t in xtemp: xdata.append(t)for t in ytemp: ydata.append(t)for i in range(len(xtemp)): m_label.append(ltemp)elif letter == "E":xtemp, ytemp, ltemp = plot_AA(ax1, "E", "GLU", i)for t in xtemp: xdata.append(t)for t in ytemp: ydata.append(t)for i in range(len(xtemp)): m_label.append(ltemp)elif letter == "F":xtemp, ytemp, ltemp = plot_AA(ax1, "F", "PHE", i)for t in xtemp: xdata.append(t)for t in ytemp: ydata.append(t)for i in range(len(xtemp)): m_label.append(ltemp)elif letter == "G":xtemp, ytemp, ltemp = plot_AA(ax1, "G", "GLY", i)for t in xtemp: xdata.append(t)for t in ytemp: ydata.append(t)for i in range(len(xtemp)): m_label.append(ltemp)elif letter == "H":xtemp, ytemp, ltemp = plot_AA(ax1, "H", "HIS", i)for t in xtemp: xdata.append(t)for t in ytemp: ydata.append(t)for i in range(len(xtemp)): m_label.append(ltemp)elif letter == "I":xtemp, ytemp, ltemp = plot_AA(ax1, "I", "ILE", i)for t in xtemp: xdata.append(t)for t in ytemp: ydata.append(t)for i in range(len(xtemp)): m_label.append(ltemp)elif letter == "K":xtemp, ytemp, ltemp = plot_AA(ax1, "K", "LYS", i)for t in xtemp: xdata.append(t)for t in ytemp: ydata.append(t)for i in range(len(xtemp)): m_label.append(ltemp)elif letter == "L":128B.3. Generate TOCSY Spectrum Programxtemp, ytemp, ltemp = plot_AA(ax1, "L", "LEU", i)for t in xtemp: xdata.append(t)for t in ytemp: ydata.append(t)for i in range(len(xtemp)): m_label.append(ltemp)elif letter == "M":xtemp, ytemp, ltemp = plot_AA(ax1, "M", "MET", i)for t in xtemp: xdata.append(t)for t in ytemp: ydata.append(t)for i in range(len(xtemp)): m_label.append(ltemp)elif letter == "N":xtemp, ytemp, ltemp = plot_AA(ax1, "N", "ASN", i)for t in xtemp: xdata.append(t)for t in ytemp: ydata.append(t)for i in range(len(xtemp)): m_label.append(ltemp)elif letter == "P":xtemp, ytemp, ltemp = plot_AA(ax1, "P", "PRO", i)for t in xtemp: xdata.append(t)for t in ytemp: ydata.append(t)for i in range(len(xtemp)): m_label.append(ltemp)elif letter == "Q":xtemp, ytemp, ltemp = plot_AA(ax1, "Q", "GLN", i)for t in xtemp: xdata.append(t)for t in ytemp: ydata.append(t)for i in range(len(xtemp)): m_label.append(ltemp)elif letter == "R":xtemp, ytemp, ltemp = plot_AA(ax1, "R", "ARG", i)for t in xtemp: xdata.append(t)for t in ytemp: ydata.append(t)for i in range(len(xtemp)): m_label.append(ltemp)elif letter == "S":xtemp, ytemp, ltemp = plot_AA(ax1, "S", "SER", i)for t in xtemp: xdata.append(t)for t in ytemp: ydata.append(t)for i in range(len(xtemp)): m_label.append(ltemp)elif letter == "T":xtemp, ytemp, ltemp = plot_AA(ax1, "T", "THR", i)for t in xtemp: xdata.append(t)for t in ytemp: ydata.append(t)for i in range(len(xtemp)): m_label.append(ltemp)elif letter == "V":129B.3. Generate TOCSY Spectrum Programxtemp, ytemp, ltemp = plot_AA(ax1, "V", "VAL", i)for t in xtemp: xdata.append(t)for t in ytemp: ydata.append(t)for i in range(len(xtemp)): m_label.append(ltemp)elif letter == "W":xtemp, ytemp, ltemp = plot_AA(ax1, "W", "TRP", i)for t in xtemp: xdata.append(t)for t in ytemp: ydata.append(t)for i in range(len(xtemp)): m_label.append(ltemp)elif letter == "Y":xtemp, ytemp, ltemp = plot_AA(ax1, "Y", "TYR", i)for t in xtemp: xdata.append(t)for t in ytemp: ydata.append(t)for i in range(len(xtemp)): m_label.append(ltemp)#Plots acquired data over simulated data#User can change file name for data exported from topspin herewith open("Ab40_titration1_2_TOCSY.txt", "r") as file0:i = 0peak_list = list()for line in file0:if i > 7:peak_info = [float(value) for value in line.split()]peak_list.append(peak_info[3:5])i += 1F2 = []F1 = []F2,F1 = zip(*peak_list)ax1.scatter(F2, F1, marker = "x", label = "Real",color = "k", s = 10)ax1.scatter(F1, F2, marker = "x", color = "k", s = 10)’’’Use this code for mouse over labels for Topspin data.Comment out if only want AA labels for PDB data’’’simdata_len = len(xdata)for f in F2: xdata.append(f)for f in F1: xdata.append(f)for f in F1: ydata.append(f)for f in F2: ydata.append(f)130B.3. Generate TOCSY Spectrum Programfor i in range(len(F2)): m_label.append("(" + str(F2[i]) + ","+ str(xdata[simdata_len+i]) + ")")for i in range(len(F1)): m_label.append("(" + str(F1[i]) + ","+ str(xdata[simdata_len+i]) + ")")’’’’’’’’’Creates one ivisible plot for all mouseover events ’’’sc = ax1.scatter(xdata, ydata, marker = "o",color = "#999999", s = 10, alpha = 0)annot = ax1.annotate("", xy=(0,0), xytext=(20,20),textcoords="offset points",bbox=dict(boxstyle="round", fc="w"),arrowprops=dict(arrowstyle="->"))annot.set_visible(False)’’’ Calls mouseover hover function ’’’fig.canvas.mpl_connect("motion_notify_event", hover)plt.gca().invert_yaxis()plt.gca().invert_xaxis()plt.subplots_adjust(left = 0.1, right = 0.7, top = 0.95, bottom = 0.1)plt.title("Simulated TOCSY spectra")plt.xlabel("F2 (ppm)")plt.ylabel("F1 (ppm)")plt.xlim((10,0))plt.ylim((10,0))plt.legend(bbox_to_anchor=(1.5, 1.05),ncol = int(round(len(peptide)/20)))plt.show()131Appendix CGenerated Spin SystemsNote: Chemical shifts are given to 4 decimal places as this is the defaultoutput from the Topspin software.Spin System 1: 1.0781 1.0683 0.6726Spin System 2: 2.143 1.9422Spin System 3: 2.1918 2.1255 1.9275Spin System 4: 2.3868 1.8103 1.698 5.1785 1.1011Spin System 5: 2.7457 1.5174Spin System 6: 2.7808 1.5076Spin System 7: 2.7925 1.2879Spin System 8: 2.8608 1.1853 0.7752Spin System 9: 2.8705 2.0338 1.9948 1.8348 1.7713 1.3855 1.3709 1.17561.117 1.1072Spin System 10 : 3.2899 0.951Spin System 11: 3.3387 0.9705Spin System 12 : 3.4596 3.4498 2.3768Spin System 13 : 3.4888 1.0096Spin System 14 : 3.8653 2.9661 1.3514Spin System 15: 4.0252 1.3221Spin System 16: 4.0428 1.2146Spin System 17: 4.3373 2.9041Spin System 18: 5.1779 1.825Spin System 19: 6.8905 6.6053Spin System 20 : 7.3118 7.1616 3.9295 1.9031 1.751 0.7459 0.7312Spin System 21: 7.3274 7.2156Spin System 22 : 7.4229 6.615Spin System 23 : 7.5205 6.8103Spin System 24 : 7.5751 7.4932 7.1863 3.9832Spin System 25 : 7.7097 7.2981 1.4246 0.7117Spin System 26 : 7.7331 4.1053Spin System 27: 7.7662 3.91 1.9617 0.785Spin System 28 : 7.8189 4.3689 2.8943 2.8846Spin System 29 : 7.8364 7.7994 3.8416132Appendix C. Generated Spin SystemsSpin System 30: 8.0568 4.1248Spin System 31 : 8.3904 7.9632 3.8269Spin System 32 : 8.484 7.6414 3.993133

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            data-media="{[{embed.selectedMedia}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
https://iiif.library.ubc.ca/presentation/dsp.24.1-0372344/manifest

Comment

Related Items