UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Spectroelectrochemical characterization of ultrathin organic films deposited on electrode surfaces Casanova Moreno, Jannu Ricardo 2014

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

Item Metadata


24-ubc_2014_september_casanovamoreno_jannu.pdf [ 27.63MB ]
JSON: 24-1.0167244.json
JSON-LD: 24-1.0167244-ld.json
RDF/XML (Pretty): 24-1.0167244-rdf.xml
RDF/JSON: 24-1.0167244-rdf.json
Turtle: 24-1.0167244-turtle.txt
N-Triples: 24-1.0167244-rdf-ntriples.txt
Original Record: 24-1.0167244-source.json
Full Text

Full Text

SpectroelectrochemicalCharacterization of Ultrathin OrganicFilms Deposited on Electrode SurfacesbyJannu Ricardo Casanova MorenoB.Sc., Universidad Nacional Autonoma de Mexico, 2006A 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)June 2014© Jannu Ricardo Casanova Moreno 2014AbstractThin organic layers deposited on electrodes are ubiquitously proposed for a variety of surface-relatedapplications. The quality of these layers is usually assessed by analytical methods that average themeasured signal over a large area compared to the molecular scale. This work outlines the use of in-situ fluorescence microscopy as a characterization method by analyzing three examples of such layers.First, a heterogeneous Langmuir layer was physically adsorbed to a gold electrode via the Langmuir-Schaefer method and the effects of the substrate analyzed by comparing the adsorbed layer with thepredecessor floating film. Through the use of a dimer-forming fluorophore, substrate mediated conden-sation was suggested.Second, the reductive desorption of self-assembled monolayers (SAMs) from microelectrodes wasused to investigate the movement of the released thiolate molecules. Once in solution, these moleculeswere found to follow a buoyant movement, consequence of the high local concentration of H2 resultingfrom the simultaneous reduction of water under the conditions employed.Finally, a DNA SAM system that has been previously suggested as a biosensing platform was in-vestigated for heterogeneity. It was found that the substrate crystallography had a significant effect onthe density and efficiency of potential driven change in conformation of the immobilized probes. Fur-thermore, a deconvolution method is proposed in order to correct for the effect of the electrode chargingtime constant on the measurements of the kinetics of the DNA conformation change.Overall, the performed experiments show that in-situ fluorescence microscopy is a useful techniqueto analyze distance dependent phenomena involving these deposited layers. Moreover, the couplingbetween electrochemistry and fluorescence allows not only to monitor but also to drive changes in thelayers, creating systems capable of studying the dynamics of the deposited films.The inclusion of the proposed technique as a characterization tool during the development of systemsbased on ultrathin organic films could improve the understanding of the influence of the depositionconditions on the film quality, helping to attain the necessary robustness to make the proposed systemsactually achieve their proposed applications.iiPrefaceThe totality of the experimental work described in this thesis was performed by the author. The experi-mental design as well as the data analysis were performed in collaboration with Dan Bizzotto (supervi-sor).Parts of this thesis have been included in two publications:• Chapter 5. Casanova-Moreno, J. & Bizzotto, D. Electrochemistry and in situ fluorescence mi-croscopy of octadecanol layers doped with a BODIPY-labeled phospholipid: Investigating an ad-sorbed heterogeneous layer Journal of Electroanalytical Chemistry, 2010, 649, 126 - 135, pub-lished by Elsevier, available at http://www.sciencedirect.com/science/article/pii/S1572665710000676.• Chapter 6. Casanova-Moreno, J. R. & Bizzotto, D. What Happens to the Thiolates Created byReductively Desorbing SAMs? An in Situ Study Using Fluorescence Microscopy and Electro-chemistry Langmuir, 2013, 29, 2065-2074, published by the American Chemical Society, freelyaccessible online at http://pubs.acs.org/articlesonrequest/AOR-YZzab8BWKq4sdU4qqUuf.For both publications, the author elaborated a first version of the manuscript that was then repeatedlyedited by Dan Bizzotto and the author.While not the main focus of this thesis, the research presented here made use of the followingcontributions:• The fluorescently labeled alkylthiol BODIPY-C10SH employed in Chapter 6 was synthesized byArnold Kell and Mark S. Workentin at the University of Western Ontario as reported previously [1].• The process to obtain single crystal bead electrodes used in Chapter 7 was optimized in the Biz-zotto laboratory by Zhinan Yu.• The electron microscopy, energy-dispersive X-ray spectroscopy and electron backscatter diffrac-tion images included in Figures 6.6 and 7.8 were obtained by Nidal Alshwawreh.iiiTable of ContentsAbstract . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iiPreface . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . iiiTable of Contents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . ivList of Tables . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . viiList of Figures . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . viiiNomenclature . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xxvAcknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xxviDedication . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . xxviii1 Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.1 Identifying the Problem . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11.2 Scope of the Thesis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 32 Background Theory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52.1 Fundamentals of Electrochemistry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 52.2 Electrochemical Techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 212.3 Fluorescence . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 323 Literature Review . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 413.1 Electrochemistry . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 413.2 Other Techniques . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 443.3 In-situ Fluorescence Microscopy . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 493.4 Perspective . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 51ivTable of Contents4 Experimental . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 524.1 Reagents . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 524.2 Materials . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 544.3 Instruments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 564.4 Spectroelectrochemical Experiments . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 575 On the Effect of the Substrate in Langmuir-Schaefer Deposition . . . . . . . . . . . . . . 595.1 Langmuir Layers . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 595.2 Transfer onto a Solid Substrate . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 625.3 Characterization of Deposited Langmuir Layers . . . . . . . . . . . . . . . . . . . . . . . . 635.4 Scope . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 655.5 Experimental . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 665.6 Results and discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 675.7 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 806 The Fate of Reductively Desorbed Self Assembled Monolayers . . . . . . . . . . . . . . . 826.1 Chemisorption and Self Assembled Monolayers . . . . . . . . . . . . . . . . . . . . . . . 826.2 Reductive Desorption . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 856.3 Macroelectrodes vs Microelectrodes . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 896.4 Experimental . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 926.5 Results . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 966.6 Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1096.7 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1147 Characterization of Electrically "Switchable" DNA Layers for Biosensing . . . . . . . . 1157.1 Deoxyribonucleic Acid (DNA) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1157.2 DNA Sensors . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1187.3 The “switching” DNA Construct . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1237.4 Average vs Space Resolved Measurements . . . . . . . . . . . . . . . . . . . . . . . . . . 1267.5 Objective . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1277.6 Experimental . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1277.7 Influence of the Crystallographic Orientation on the Fluorescence Response . . . . . . 1327.8 On the Deconvolution of the Electrochemical Time Constant on DNA “Switching” Kinetics 140vTable of Contents7.9 Conclusions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1518 Concluding Remarks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1548.1 Summary . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1548.2 Future Work . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1568.3 Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 158References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 159AppendicesA List of Reagents Used . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 176B Flat-field Image Correction Procedure . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 177C Uncorrected and Contrast Enhanced Images Corresponding to Chapter 5 . . . . . . . . 179D Diffusion of Physically Adsorbed DNA from Modified Electrodes . . . . . . . . . . . . . . 181E Comparison of Objective Performance . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183F MCH/DNA-Modified Bead Images at Different Focal Planes . . . . . . . . . . . . . . . . . . 184viList of Tables2.1 Compilation of thermodynamic properties of Au surfaces of different crystallographic ori-entations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182.2 Summary of the electrochemical techniques employed in this work, including an exampleof the typical response for a simple electrochemical cell where faradaic reactions aremostly absent. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 222.3 Compendium of CDC representations and typical impedance responses of electrical com-ponents and simple circuits. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 293.1 General characteristics of the most common analytical techniques for deposited organiclayers. NA = Not applicable . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 454.1 Filters employed for fluorescence measurements . . . . . . . . . . . . . . . . . . . . . . . 577.1 List of names and abbreviations of the nucelobases and their corresponding nucleotides. 1177.2 Select examples of detection using DNA in conductive surfaces. . . . . . . . . . . . . . 1217.3 EIS fitting results for electrolytes containing 10mMand 100mM KNO3 using the (CSt[RS(Qint[RintCint])]). R refers to the added resistor. Percentage error values are included for eachvalue. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1477.4 EIS fitting results for electrolytes containing 10 mM and 100 mM KNO3 using the (CSt[RS(Cint[RintQint])]). R refers to the added resistor. Percentage error values are included foreach value. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 147A.1 List of the general reagents employed. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 176viiList of Figures1.1 Different methods to deposit organic monolayers on solid substrates. . . . . . . . . . . . 21.2 Some examples of defects in deposited monolayers causing heterogeneity. . . . . . . . 32.1 Different models of the electrical double layer, including their potential variation with dis-tance away from the electrode and equivalent circuits. For clarity, only a few solventmolecules are presented, but the reader should understand the remaining blank spaceis also filled with solvent. Rs, CH and Cd and stand for the solution resistance and thecapacitances of the Helmholtz layer and the diffuse part of the double layer respectively. 72.2 Double layer structure with specific anion adsorption including its potential profile andequivalent circuit. Rs and Rad are the solution and adsorption resistances; CiH, CoH andCd and stand for the capacitance of the inner and outer Helmholtz layers as well as thediffuse part of the double layer. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102.3 a) Schematic of the theoretical electrocapillary curves for electrode surfaces covered withsolvent and anion molecules, as well as the thermodynamically most stable configuration.b) Experimental electrocapillary curves for Hg in contact with solutions of various salts 112.4 Representation of the electrical double layer, potential profile, and equivalent circuit forthe case of adsorption of an organic monolayer. . . . . . . . . . . . . . . . . . . . . . . . . 122.5 a) Schematic of the stability region of a deposited monolayer as the overlapping of twodifferent electrocapillary curves. b) Experimental data for the system comprised of 1-octanol in 0.1 mol kg-1KCl on a mercury electrode. Concentrations of 1-octanol vary forthe different curves, being 0, 0.025, 0.0500, 0.0750, 0.100, 0.200, 0.300, 0.400, 0.490,0.600, 0.700, 0.900 and 1.05 mmol kg-1for curves 1 to 13 . . . . . . . . . . . . . . . . . . 132.6 a) Procedure to elaborate a stereographic projection. b) The full stereogram for cubiclattices. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15viiiList of Figures2.7 a) Schematics of a truncated fcc crystal displaying the low index surfaces {111} (red),{110} (blue), and {100} (green). b, c and d) Two-dimensional hexagonal (b), square (c)and rectangular (d) lattices showing their Wigner–Seitz cells. . . . . . . . . . . . . . . . 162.8 Variety of surface planes resulting from truncating an fcc crystal in increasing order ofatomic roughness (left to right, top to bottom) . . . . . . . . . . . . . . . . . . . . . . . . . 172.9 Schematics of electron density spreading (a) and smoothing (b) in a square lattice. Theblack line represents the limits of the electronic density. Notice the dependence of theelectron smoothing on the surface crystallography. . . . . . . . . . . . . . . . . . . . . . . 182.10 Schematic of a three-electrode electrochemical cell. Black lines represent electronic con-ductors while blue ones correspond to ionic conduction. The dashed line symbolizes thephysical boundaries of the electrolyte containing vessel. Encircled variables representmeasuring devices. Rs and R are the solution and uncompensated resistances respec-tively. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 202.11 Cyclic voltammograms of a ethanolic 50 mM solution of LiClO4 in the absence (solid line)and presence of 1 mM ferrocene (dashed line). WE: graphite,  = 100 mV/s . . . . . . 232.12 Relationship between two periodic signals with the same frequency. a) The variation ofpotential and current as a function of time; b) The phasor representation of the samesignals. The potential signal has been taken as the real (Re) axis. . . . . . . . . . . . . . 262.13 Schematic of the three-electrode electrochemical cell illustrating the resistive and (stray)capacitive elements in the absence of the “fourth electrode” Resistances arise both fromsolution conduction and faradaic processes at the electrodes interface. . . . . . . . . . . 312.14 Normalized fluorescence spectra of AlexaFluor 488. The solid and dashed lines representthe excitation and emission spectra, respectively. . . . . . . . . . . . . . . . . . . . . . . . 322.15 Near field interactions of a fluorophore with a metal. Left: induced plasmons create largecharge separation that can escape as photons. Center: at closer metal-fluorophore dis-tances, the charge separation is smaller and no wavevector matching is possible. Right:upon the addition of a higher refractive index material, wavevector matching is possibleand the plasmons radiate to the distal side of the surface. . . . . . . . . . . . . . . . . . 372.16 Simplified schematic of an inverted epi-fluorescence microscope. Inset: detailed view ofa filter assembly set. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 39ixList of Figures3.1 Differential capacitance measurements of a mercury electrode in a 1 M KNO3 aqueoussolution in the absence (dashed line) and saturated with (solid lines) octyl alcohol. Thetwo different solid lines were obtained at 0.24 and 10 kHz as indicated. . . . . . . . . . . 433.2 Selected examples of spatially-resolved organic monolayer characterization. . . . . . . 483.3 Basic setup to conduct in-situ fluorescence experiments . . . . . . . . . . . . . . . . . . . 504.1 Structures of the BODIPY labeled molecules employed . . . . . . . . . . . . . . . . . . . 534.2 Structure of the Alexa Fluor 488 oligonucleotide modification . . . . . . . . . . . . . . . . 544.3 Cell used for electrochemical experiments. . . . . . . . . . . . . . . . . . . . . . . . . . . . 554.4 Schematic of the potential perturbation employed in a basic in-situ fluorescence experi-ment . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 585.1 Compression isotherms for octadecanol, DPPC and POPC Langmuir layers at 20.5 ±0.5°C . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 605.2 Fluorescence images of a DPPC monolayer containing 2 mol % of the fluorescently la-beled phospholipid 1-acyl-2-[N-(7-nitro-2,1,3-benzoxadiazol- 4-yl)amino caproyl] phos-phatidylcholine floating on a water subphase (a) and deposited on an alkylated coverslip(b). The fluorescent probe does not partition into the condensed phase domains whichappear dark compared to the fluid surrounding phase . . . . . . . . . . . . . . . . . . . . 635.3 Desorption model proposed for 1-octadecanol in Au. Schematics of the state of the mono-layer are overlaid on the experimentally determined capacitance . . . . . . . . . . . . . . 655.4 Collection of fluorescence images (top: dimer and bottom: monomer) of the floating layerof octadecanol with 1mol% of BODIPY-HPC. The intensity scale is useful for order ofmagnitude comparisons only and shows the linear mapping of intensity onto gray scalelevel. The scale bar represents 50 µm for all images. . . . . . . . . . . . . . . . . . . . . . 705.5 Fluorescence decays for dimer (a) and monomer (b) emission from a floating monolayercomposed of C18OH and BODIPY-HPC in a 99:1 mole ratio. Vertical lines represent thestandard deviation for three replicate experiments. Fluorescence images for dimer (c ande) and monomer (d and f) emission. The outlined region was first exposed to light for 80sec (left column); then the whole region was exposed for 24 seconds more, resulting inthe images shown in the right column. The scale bar is 50 µm. . . . . . . . . . . . . . . . 73xList of Figures5.6 Comparison of fluorescence images of the (a) floating layer before deposition onto theAu{111} surface and the fluorescence images of the transferred layer at two potentials(b) -0.60V/SCE and (c) -0.80V/SCE. The scale bars are 50 µm for all images and clearlyreflect the fact that the floating layer expands upon adsorption. . . . . . . . . . . . . . . . 745.7 Dimer andmonomer fluorescence images recorded as potential is stepped to -0.80 V/SCE(a). The top portion of the images corresponds to the dimer fluorescence while the bottomportion is the corresponding monomer fluorescence. The outlines are a guide to theregions analyzed. The intensity of the dimer and monomer fluorescence for three regionsof interest (ROI) as a function of potential is given in (b) and (c) respectively; the ratio ofred to green fluorescence is shown in (e) (ROIs are shown in the inset). The capacitanceas a function of potential is given in (d) (200 Hz, 5 mV rms). The scale bar is 50 µm. . 765.8 Fluorescence of a floating monolayer composed of C18OH and BODIPY-HPC in a 99:1mole ratio. The zoomed regions are contrast enhanced to reveal the structures of interest.The images were background subtracted (rolling ball of 1000 pixels, unsharp mask of 20pixel radius weighting of 0.6). The scale bar is 50 µm. . . . . . . . . . . . . . . . . . . . . 785.9 A series of fluorescence images of a desorbed layer composed of C18OH and BODIPY-HPC in a 99:1 mole ratio. The top image is the dimer and bottom is the monomer fluo-rescence and the potential is indicated in the image. The scale bar is 50 µm. . . . . . . 795.10 Dimer(a) and monomer(b) fluorescence intensity (ΔF/FL) variations with the applied po-tential in the three different regions of interest (ROI) depicted in the inset. c) Capacitancemeasurements of the bare (dotted line) and modified (solid line and open circles) Au elec-trode during the application of a step-base potential perturbation in the negative direction(200Hz, 5mV rms). d) Red to green intensity ratio as a function of potential for the regionsof interest shown in the inset . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 816.1 Cyclic voltammograms of a gold electrode in a 0.5 M ethanolic KOH solution in the pres-ence (a) and absence (b) of dodecanethiolate. Dotted line in (a) corresponds to contin-uous cycling whereas the solid line corresponds to the first cycle of a SAM prepared bythe OCP method with immersion time of at least 2 h in 1 mM dodecanethiolate. =100mV/s. The arrow indicates the initial scan direction. . . . . . . . . . . . . . . . . . . . . . 856.2 Proposed mechanisms for reductive desorption of alkylthiol SAMs. . . . . . . . . . . . . 88xiList of Figures6.3 Sequence of in-situ fluorescence images during the reductive desorption of SAMs frommacroelectrodes. a) Selective desorption of the fluorescently labeled alkylthiol BODIPY-C10-SH from gold surfaces with {111} character. b) Mixed monolayer composed of thi-olated fluorescently labeled DNA and mercaptohexanol. The colored traces indicate themovement path of fluorescent aggregates, one of which is followed in the yellow circle.The width of the frames for (a) and (b) are 1 mm and 320 µm respectively. . . . . . . . 906.4 Diagram of the employed microelectrodes. Low background fluorescence was attainedby using borosilicate glass, fluxless solder, nail polish and white PTFE tape. . . . . . . . 916.5 Optical microscopy images of three microelectrodes. . . . . . . . . . . . . . . . . . . . . 926.6 The surface of a glass pipette after been heated by the Pt/Ir filament. a) Photography;black deposits can be seen near the tip. b) Scanning electron microscopy image c)Energy-dispersive X-ray spectroscopy mapping analysis. . . . . . . . . . . . . . . . . . . 936.7 Experimental setup for the study of reductive desorption of SAMs from microelectrodes.The small counter electrode is shown. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 946.8 Variation of the capacitance and fluorescence intensity upon the reductive desorption ofa fluorophore labeled alkylthiol SAM. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 976.9 a) Fluorescence images at selected potentials (shown in the images) during the reductivedesorption. The dashed outlined region corresponds to the Au surface, while the contin-uous line represents the resulting ellipse after applying a threshold (0.35). Video 6.1 inthe accompanying disc shows this desorption experiment. The length of the major andminor axes (b) as well as the displacement of the center of the ellipse (c) are plotted fora smaller subset of employed potentials (d). Symbols in (c) represent raw data while thedotted line results from smoothing with a Savitzky-Golay filter . . . . . . . . . . . . . . . . 996.10 Simulation of an irreversible SAM desorption from a 2 µm radius hemispherical microelec-trode. Top panel shows the cross section of the fluorescence signal at different times. Thecorresponding regions in the electrode plane (x,y) resulting from thresholding the imagesto an intensity value of 20 arb. units (shown as a dotted line) are shown in the bottompanel. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1016.11 Variation of the length of the axes (both major and minor have the same length) in asimulation of a SAM desorption. Threshold has been set to an intensity value of 20 arb.units. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 102xiiList of Figures6.12 Desorption experiments performed with the CE in close proximity (30 µm) to the WE.Lateral view schematic (a) and bottom viewmicrograph (b) of the system. c) Fluorescenceimages at two selected values of potential, for two different In addition experiments. Theblack circle represents the edge of the gold electrode, the white line represents the outlineof the CE. Dashed lines are given as reference for the angular coordinate used. . . . . 1036.13 Displacement (a), speed (b) and angle of movement (c) during the desorption for the twoexperiments shown in Fig. 6.12c. The filled and empty symbols correspond to the topand bottom series of images, respectively. The potential perturbation is shown in (d).Symbols represent raw data while lines result from smoothing with a Savitzky-Golay filter.Error bars on the displacement are smaller than the symbol size; error bars on the speed(shown in the figure) were estimated by determining the sensitivity of the center of theellipse on the threshold values used in image analysis. . . . . . . . . . . . . . . . . . . . . 1046.14 a) Schematic of the system employed to investigate the effect of gravity. b) Fluorescenceimages at E = -2.0 V of two independent desorption experiments. The direction of thearrow points towards the higher point of the glass sheath. A video corresponding to thedesorption experiment shown in the bottom of (b) is included as Video 6.3 in the accom-panying disc. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1056.16 Variation of the terminal speed of the plume of desorbed thiolates as a function of the limitnegative potential employed. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1066.15 Variations in the minor (a) and major axes (b) as well as speed (c) of the center of theellipse fitted to the fluorescent plume on three desorption experiments following the po-tential perturbations shown in (d). Symbols represent raw data while lines result fromsmoothing using Savitzky-Golay filters. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1076.17 Temperature difference (T-T0) finite element simulations corresponding to the maximumobserved current density, 30 mA/cm2 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1106.18 a) Fluorescence images from a desorption experiment on a Au microelectrode (left andcenter) and from an experiment in which H2was bubbled though an empty glass pipettetreated in the same way the microelectrode was. b) Open circuit potential of two macro-scopic wires of Au and Pt/Ir in a 1 mM KOH solution. Initially the solution is degassedwith Ar; at the time indicated with the vertical line H2is bubbled into the solution. . . . . 113xiiiList of Figures7.1 Structure of deoxyribonucleic acid (DNA). Left panel shows the chemical structure of itsdifferent components. A nucleotide is shown delimited by the dotted line. A schematicof the three-dimensional double helix arrangement of these components is shown in theright. The pink spheres represent the phosphate-sugar backbone and the bars symbolizethe bases following of the same color coding as in the left panel. . . . . . . . . . . . . . . 1167.2 Comparison of DNA one (a), two (b) and three (c) dimensional DNA sensing platforms. 1207.3 DNA conformation control through the use of the electrode potential. When the electrodeis negatively charged the DNA is repelled from its surface (left) while positive chargeattracts the DNA towards the electrode (right). . . . . . . . . . . . . . . . . . . . . . . . . . 1237.4 Relationship between applied potential (top left) and fluorescence response (left bottom)from an electrode-immobilized 48-mer double stranded DNA labeled with the Cy3 fluo-rophore in the distal end (right). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1257.5 Variation of the fluorescence modulation amplitude with applied AC potential frequencyfor SAMs containing fluorescently labeled DNA. Two different DNA lengths (24 and 48base pairs) are compared. Buffer electrolyte (pH = 7.3) is composed by 10 mM Tris and50 mM NaCl. Disc electrode diameter is 2 mm. . . . . . . . . . . . . . . . . . . . . . . . . 1267.6 a) Fluorescence microscopy image of a DNA / MCH layer deposited on a gold bead. b)Image created by averaging the values of all the pixels in (a) to represent the informationobtained with spatial averaging techniques. Images are monochromatic and displayed infalse color according to the scale shown in the right side . . . . . . . . . . . . . . . . . . . 1277.7 Schematic of the system employed. One port of the microscope is connected to a CCDcamera to detect images while the other port directs the light towards a photomultipliertube (PMT). The photocurrent is transformed to a voltage by a SR570 current preamplifierand the resulting frequency-dependent modulated signal is used to compute the transferfunction H(ƒ ). The electrochemical impedance Z(ƒ ) is calculated from the measuredpotential and current oscillations. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1307.8 Characterization of the Au bead. a) Brightfield micrograph. b) Scanning electron micro-graph. c) EBSD mapping of a 6 × 6.5 µm are inside the facet outlined in (a). d) Colorcoded stereographic triangle for an fcc material. . . . . . . . . . . . . . . . . . . . . . . . 1337.9 Fluorescence image of the MCH/DNA-SAM modified gold bead electrode. These imagesare the result of merging a collection of fluorescence images at different focal points pro-viding an extended depth of fiel . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 135xivList of Figures7.10 Fluorescence imaging measurements during potential step experiments. a) The potentialperturbation imposed on the interface and the resulting changes in capacitance and flu-orescence. Capacitance values are shown for both Eb and Es parts of the perturbation.The relative increase in fluorescence is shown for five different regions of interest shownin (b). b) The maximum relative change in fluorescence (ΔF/Fb) due to the potentialperturbation with the an overlaid crystallographic triangle as per Fig. 7.9. . . . . . . . . 1367.11 a) An image of region of interest B as seen through the water immersion objective show-ing the extent of illumination and the resulting fluorescence collection that is passed tothe PMT. b) Electrical current (top) and fluorescence intensity (bottom) measured withpotential cycling between the potential limits used. An indication of the potential pertur-bation regions used in the fluorescence FRA spectroscopy measurements is shown. Theelectrochemical current signal has been smoothed with a Savitzky-Golay filter . . . . . 1387.12 Hcell(ƒ ) measured for different ROIs indicated in Fig. 7.10b. The magnitude (a) andphase (b) for this transfer function is shown, along with the magnitude normalized at 30Hz (c and d). Red dashed line in (d) is aids to visualize the cut-off frequency. . . . . . . 1397.13 a) and b) Fluorescence FRA measurements of a single ROI under the following set ofconditions: ƒ = 30 - 100 kHz, EDC = +150 mV, EAC = 200 mV p-p, 10 mM Tris + 10 mMKNO3. Values for CSh are indicated in the figure. c) and d) Electrochemical impedancemeasurements under the same set of conditions except EAC = 5 mV rms. Symbols aremeasured data, lines are fit to a (C[R(C[RQ])]) equivalent circuit. . . . . . . . . . . . . . . 1437.14 Equivalent circuits used to fit the EIS data. a) the general (CSt[RSZint]) circuit. b and c)The two options used for Zint resulting in indistinguishable impedance response. . . . 1457.15 The magnitude and phase for each of the transfer functions Hcell(ƒ ) and Hint(ƒ ). Alsoincluded is including the value of EAC,int(ƒ )/EAC,cell(ƒ ). on a secondary (right hand) yaxis. Subfigures (a) and (b) correspond to the system as described in Fig. 7.7 while anextra 2.46 kΩ resistor has been added to the WE connection in (c) and (d). [KNO3 ] = 10mM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1467.16 The magnitude and phase for each of the transfer functions Hcell(ƒ ) and Hint(ƒ ). Alsoincluded is including the value of EAC,int(ƒ )/EAC,cell(ƒ ). on a secondary (right hand) yaxis. Subfigures (a) and (b) correspond to the system as described in Fig. 7.7 while anextra 1.00 kΩ resistor has been added to the WE connection in (c) and (d). [KNO3 ] =100 mM . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 149xvList of Figures7.17 Fluorescence FRA as measured (open symbols) and corrected (filled symbols) for poten-tial across the interface (solid lines). Results for 10mMand 100mM KNO3 concentrationsin the absence and the presence of the extra R. Fluorescence FRA magnitude has beennormalized using its value at 30 Hz. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1507.18 Correction applied to Rant’s data. Panels (a) and (b) show the correction for digoxigenin-labeled DNA in the absence and presence of sheep immunoglobulin G (IgG) antibody.The potential across the interface is calculated using Rant’s fit to an [RC] circuit with R= 1.6 kΩ and C = 9.6 nF. Uncorrected and corrected FAC(ƒ ) values are shown in panels(c) and (d) respectively, evidencing the increase in the frequency response shift after IgGbinding if the correction is used. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1528.1 Schematic of the systems investigated through in-situ fluorescence microscopy. a) DNAconformational switching (Chapter 7) b) Physical desorption (Chapter 5) c) Chemical des-orption (Chapter 6). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 155B.1 Block diagram of the flatfield correction image procedure . . . . . . . . . . . . . . . . . . 178C.1 Unenhanced verison of Figure 5.7. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179C.2 Contrast enhanced version of Figure 5.7 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 179C.3 Unenhanced verison of Figure 5.9 . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 180C.4 Contrast enhanced version of Figure 5.9 . . . . . . . . . . . . . . . . . . . . . . . . . . . . 180D.1 Fluorescence images of an MCH/DNA layer during a step potential experiment as de-scribed in Section 7.6.3. This layer was analyzed without the overnight soaking step inbuffer. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 182D.2 Average fluorescence intensity (bottom panel) of the whole frame of the images shown inFig. D.1 upon the application of the potential perturbation shown in the top panel. . . . 182E.1 Comparison of brightfield and fluorescence images of a MCH/DNA layer deposited on apolished gold electrode, using a 50× (NA = 0.5) dry objective and a 40× (NA = 0.8) waterimmersion objective. Bars under the images show the minimum and maximum intensityvalues used in the lookup table. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 183F.1 Selected images at different focal points used to construct Figures 7.9a and b. . . . . . 184xviNomenclature4E fourth electrode, page 31µTAS micro total analysis system, page 157A area , page 6 acceleration, page 108AC from alternating current; electrical signals that change periodically with time , page 24AES Auger electron spectroscopy, page 46AFM atomic force microscopy, page 47θ acceleration in the plane of the electrode, page 108B buoyant force, page 108BODIPY fluorescent dye based on the 4,4-difluoro-4-bora-3a,4a-diaza-s-indacene core, page 52C capacitance, page 6CCD charge-coupled device, page 57Cd capacitance corresponding to the diffuse part (i.e. ≥ 2) of the double layer of anelectrode, page 8CDC Circuit Description Code, page 28Cdiff differential capacitance, page 26Cdl capacitance of the double layer of an electrode, page 8CE counter electrode, page 19CH capacitance corresponding to the Helmholtz layer (i.e 0≤ ≤ 2) of an electrodedouble layer, page 8xviiNomenclatureC net capacitance of the th branch of the net surface reaction impedance, page 27CiH capacitance corresponding to the inner Helmholtz layer (i.e 0≤ ≤ 1) of an elec-trode double layer, page 9Cint integral capacitance, page 27CN coordination number of an atom in a crystal, page 15CoH capacitance corresponding to the outer Helmholtz layer (i.e 1 ≤  ≤ 2) of anelectrode double layer, page 9Cp specific heat capacity, page 109CPE constant phase element, page 28CSh shunt capacitor used in the so called fourth electrode, page 31CSt stray capacitance, page 31CV cyclic voltammetry, page 21d distance between two plates in a capacitor, page 6DC from direct current; electrical signals that do not change periodically with time or doso in a very slow fashion, page 21dmin resolution of an imaging system, page 40DPPC dipalmitoylphosphatidylcholine, page 61dsDNA double stranded DNA, page 118E electric potential, page 20E phase containing electric potential, page 25E0 Amplitude of a sinusoidal potential wave, page 24E0′formal potential of an electrode, page 23EAC AC potential perturbation, page 24EDC DC potential perturbation, page 24xviiiNomenclatureEb base potential in a spectroelectrochemical experiment, page 57EDAC 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide hydrochloride, page 119EIS electrochemical impedance spectroscopy, page 27Ef potential of the final step in a spectroelectrochemistry experiment, page 57Es potential of a given step in a spectroelectrochemistry experiment, page 57F Faraday’s constant, page 23F fluorescence intensity, page 34ƒ frequency, page 24ƒc cut-off frequency of a switching DNA frequency response, page 124fcc face centered cubic Bravais lattice, page 15FL intensity of excitation light leaked through the filters, page 66FRET Förster resonance energy transfer, page 34G gas phase of a Langmuir layer, page 60g acceleration of gravity, page 108HPC 1-hexadecanoyl-sn-glycero-3-phosphocholine, page 53HPLC high performance liquid chromatography, page 54i phase containing electrical current, page 25 electrical current, page 200 amplitude of a sinusoidal current wave, page 24AC DC current response, page 24DC average intensityl in an AC fluorescence measurement, page 137DC DC current response, page 24IgG immunoglobulin G, page 151xixNomenclatureIm electrical current 90° out of phase with the potential perturbation, page 26IR infrared radiation, page 44Re electrical current in phase with the potential perturbation, page 26RF impulse response function, page 100IUPAC International Union of Pure and Applied Chemistry, page 19AC AC current response, page 24J current density, page 21j imaginary unit (j2 = −1), page 28k wavevector of a propagating wave, page 35k Boltzmann constant, page 6kd heterogeneous desorption rate constant, page 100knr non-radiative decay rate of the excited state of a fluorophore, page 33kr radiative decay rate of the excited state of a fluorophore, page 33KSV Stern-Volmer quenching constant, page 34LE liquid expanded phase of a Langmuir layer, page 60LIA lock-in amplifier, page 26m parameter of the constant phase element impedance equation, page 31MCH 6-mercapto-1-hexanol, page 124MMA mean molecular area, page 59n number of electrons exchanged in an electrochemical reaction, page 23n0 number concentration of the ions in a z : z electrolyte, page 6NA numerical aperture of an objective, page 40Nbs number of molecules that undergo an absoprtion process, page 33xxNomenclatureNnr number of molecules that undergo a non-radiative decay process, page 33NR neutron reflectivity, page 46Nr number of molecules that undergo a radiative decay process, page 33OCP open circuit potential, page 83[Ox] concentration of the oxidized from of a redox pair, page 23[P] concentration of electrochemically created product, page 100P pressure, page 13p-p peak to peak amplitude of a wave, page 131PL persistence length, page 118PMT photomultiplier tube, page 129POPC 1-palmitoyl-2-oleoylphosphatidylcholine, page 61PTFE polytetrafluoroethylene, page 54PZC potential of zero charge, page 6Q parameter of the constant phase element impedance equation, page 31q electric charge, page 25QCM quartz crystal microbalance, page 120qe charge of the electron, page 6qm excess charge in the metal side of the electrode interface, page 6qs excess charge in solution side of the electrode interface, page 6Qu fluorescence quencher, page 34QY quantum yield of a fluorophore, page 33R electrical resistance, page 20r radial distance from the electrode surface, page 100xxiNomenclatureR0 Förster’s radius , page 35RC circuit composed by a resistor in series with a capacitor, page 27RE reference electrode, page 19[Red] concentration of the reduced from of a redox pair, page 23RG ratio of the outside sheath diameter to conductor diameter in a microelectrode,page 92Rg molar gas constant, page 23RHS resistive heat source, page 109rms root mean square, page 24ROI region of interest, page 74R uncompensated solution electrical resistance, page 20S0 the ground electronic state of a molecule, page 32S1 first excited singlet electronic state of a molecule, page 32SAM self assembled monolayer, page 2SCE saturated calomel reference electrode, page 55SEIRAS surface enhanced infrared absorption spectroscopy , page 44SELEX Systematic Evolution of Ligands by Exponential Enrichment, page 122SEM scanning electron microscopy, page 93SERS surface enhanced Raman spectroscopy, page 46SIMS secondary ion mass spectrometry, page 47ssDNA single stranded DNA, page 116STM scanning tunneling microscopy, page 47t time, page 23T absolute temperature, page 6xxiiNomenclatureT0 absolute temperature in the bulk of the solution, page 110TC tilted condensed phase of a Langmuir layer, page 60Tris 2-amino-2-hydroxymethyl-propane-1,3-diol, page 128UAM uniformly accelerated motion, page 108UC untilted condensed phase of a Langmuir layer, page 60UV ultraviolet light, page 118V volume, page 55 potential scan rate, page 25WE working electrode, page 19 distance normal from the electrode surface , page 82 distance from the electrode at which the outer Helmholtz plane is located, page 5XPS X-ray photoelectron spectroscopy, page 46Z electrical impedance, page 25z charge of the ions in a symmetric electrolyte, page 6ZIm imaginary part of the electrical impedance, page 28ZRe real part of the electrical impedance, page 28Zs net surface reaction impedance, page 27φ phase difference between an applied AC potential and the produced AC current,page 24ϕ0 potential at the electrode surface with respect to the bulk solution, page 6ϕ2 potential at a distance 2 away from the electrode surface, the so called outer Hel-moltz plane, page 8α charge transfer coefficient, page 100β angle of a light wave measured from the normal to an interface , page 35xxiiiNomenclatureδ Debye length, page 8ΔF potential dependent change in fluorescence intensity, page 67ΔGmix Gibbs free energy of mixing, page 70ε relative permittivity of a material, page 6ε0 permittivity of vacuum, page 6 work function of a metal, page 17γ interfacial tension, page 9γ0 interfacial tension of a surfactant-free interface, page 61 surface excess of species , page 9mx maximum amount of adsorbate at the surface of the electrode, page 42H(ƒ ) potential modulated fluorescence FRA transfer fucntion, page 131κT thermal conductivity, page 109λ wavelength of light, page 32λAbsmwavelength of maximum absorption, page 66μ chemical potential, page 13µE microelectrode, page 90μ˜ electrochemical potential of species , page 9η refractive index, page 35 surface pressure, page 60e equilibrium spreading pressure, page 61ρ density, page 109ρ0 density of the bulk solution, page 108ρb density in the buoyant region, page 108xxivNomenclatureκE electrical conductivity, page 94σ surface charge density, page 11Θ surface coverage, page 9θ tilt angle between the horizontal and the electrode surface, page 108τE electrochemical time constant, page 140τF fluorescence lifetime in the presence of a quencher, page 34ω angular frequency (in radians) of AC signals, page 24∅ diameter, page 176xxvAcknowledgementsI acknowledge that the work presented here was carried out in unceded native (Musqueam) territory.Since colonization can only be avoided by not colonizing, my stay here must therefore be only temporary.This work would not have been possible without the help of many people to which I am grateful:• Dan, your sincere interest in your students learning as many diverse skills as possible makes youa good supervisor. Your openness to other points of view and moral integrity, however, makes youan excellent person. I am glad to have met you in both capacities.• Financial support from the people of Mexico, through the Consejo Nacional de Ciencia y Tec-nología (National Council of Science and Technology, CONACYT) scholarship number 207929, isgreatly acknowledged. I recognize the effort this support entails, and truly hope that my traininghere eventually translates into a contribution to a better quality of life for such people.• The technical staff of the UBC Chemistry Department was invaluable in developing the equipmentused here. In particular I would like to recognize the glassblowing skills of Brian Ditchburn whosededicated efforts made possible the fabrication of the highly demanding cells we require.• When research took us away from our area of expertise, guidance in such fields was much ap-preciated. Janine Mauzeroll, Andrzej Baranski and Hua-Zhong Yu must be recognized for theiradvice on microelectrode fabrication, Joule heating and DNA handling, respectively.Being an international student made this not only an academic stage, but also a life experiment (withexcellent results, I must add). Therefore, some special thanks are required in that respect:• Aya (and T&D), your company along this path was one of the most important things that happenedto me in Vancouver and shaped to a great extent my time here. Your creativity and imagination(capable of creating a world of their own) and ability to listen were always with me, making melaugh or sharing my cries. I learned so much from you and shared experiences that made megrow as a person. I truly enjoyed the time by your side in three different continents. I can onlythink that such a nice person will enjoy a nice life.xxviAcknowledgements• Past and present lab members have filled many positions in my life. Colleagues and friends,Robin, Jeffrey, Amanda, Landis, Isaac, Maria, Caroline and Kaylyn need to be thanked both fortheir technical support and their fun company. Amanda, in particular, must be recognized for hersupport in anything related to computers.• Geographical origin put us, the “Latinidos” (Paloma, Emmanuel, Roxana, Montse and Javier) to-gether, proving that if there is an identity I can claim as my own it is the Latin American one. Yourescued me from my deepest loneliness and I will be forever grateful. Montse, your inquisitivepersonality and sensitivity have led you to an arduous path; watching your transformation remindsme of myself. Niusha, you know deep inside you are one of us.• Fortunately for me, all graduate students in the Department are required to TA. Thanks to RobinStoodley for letting me enjoy the teaching in his lab and for welcoming the feedback I had to offer.Thanks to the hundreds of students I taught and that helped me feel useful. Special thanks go toMokit, by far the most hardworking student I have met and who has become a good friend.• Family and friends (in short, my life) from Mexico were always present, from there or whereverthey are, following me as I follow them, reminding me who I am and where do I belong. Soonerthan later, our paths will cross again.• Without the UBC Outdoor Club (VOC) grad school would have probably been shorter, but defi-nitely not as fun. Thanks to the hundreds of people that were always keen to teach or learn newskills. The friendly yet challenging, crazy yet respectful atmosphere of this club captivated me fromthe beginning. Special mention goes to Julien, Steph, Clemens, Diego and Isabel; you did whatseemed impossible, make the mountains an even more beautiful place.• A part of me always lived in this other Vancouver that most people do not seem to notice. But tothe people that did, and shared it with me, thank you from the bottom of my heart. Leonor, you areone of the nicest persons I have ever met and your friendship, along with those of Vale, Emiliano,Azul, Daniel, Uriel, Mandeep, Carlos, Juve, Neto and the like, was crucial in surviving my firstyears. Chance, Adam, Damian, you showed me a new style of doing things and, together withthe Arenas-Sandoval family, adopted me. Thanks to the international migrants for sharing theirstories. Erie and Beth, you gave me the warm feeling that the Philippines and Mexico are closerthan I always thought.xxviiDedicationTo the future students of the Bizzotto lab and elsewhere, in the hopes that this work serves as referencematerial for your own endeavors.xxviiiChapter 1Introduction1.1 Identifying the ProblemIn the search for new technologies, people have always looked to expand the range of the sizes of objectsthey can control and engineer. An example is that of a variety of deposited layers with small thickness thatcover solid substrates. When these layers posses a thickness of less than 100 nm they are generallycalled thin films, and when the thickness drops another order of magnitude (<10 nm) they are oftentermed ultrathin films [2]. By being deposited on substrates, these layers allow for the modification ofthe characteristics of both materials, essentially creating surfaces with particular characteristics that arenot possible in the bulk of either the components of the substrate or the film.If the structure and composition of these films are carefully chosen, the properties of the interfacecan be designed to fulfill a desired function and hence, a new research area of so-called functionalsurfaces has emerged. The applications of these surfaces vary widely, from optic and electronic coat-ings to corrosion prevention to magnetic storage. These surfaces have recently found many uses inbioengineering. In the biomaterials area, surface modification has been employed to prevent adhesionof proteins and cells, create low friction surfaces and control drug or DNA delivery, among others [3].Biosensing has also exploited ultrathin films either using them as membranes that are selectively perme-able to the desired analyte [4] or as substrates which support immobilized biologically selective receptormolecules. Biophysicists have employed deposited lipid bilayers to mimic the cell membrane structureand study its interactions with external stimuli, such as the presence of drugs [5].Under some circumstances, these functional materials can go beyond static systems and certain per-turbations (e.g. light or electricity) used to drive processes on them. By selecting conducting materialsas substrates, it is possible to electrochemically drive processes and dynamically control the propertiesof these layers. For this reason, metals are a common choice of substrate. Noble metals in particular arepreferred due to their relative inert chemistry facilitating the control of unwanted reactions and facilitatingthe preparation procedures.11.1. Identifying the ProblemLangmuir-Blodgett Langmuir-Schaefer Self assemblyFigure 1.1: Different methods to deposit organic monolayers on solid substrates.Organic molecules containing hydrophobic moieties are a popular choice for surfacemodification dueto the relative stability of the deposited layers in aqueous solutions originating from their Van der Waalsinteractions and subsequent poor solubility in water. These molecules are generally composed of longalkyl chains with an active “head” group that favors a preferred orientation when they are placed in aninterface, creating layers of monomolecular thickness (monolayers). Their desired functional propertiesdepend on the substituents on the chains, usually located at the terminus opposed to the substrate.Early deposition processes were developed in the first half of the twentieth century and started withthe formation of a monolayer of amphiphilic molecules in an air | water interface (Langmuir layer). Thefloating layer was subsequently transferred on to a solid substrate through perpendicular emmersion(Langmuir-Blodgett technique, Fig. 1.1 left) or parallel contact with the floating monolayer (Langmuir-Schaefer technique, Fig. 1.1 center). In these layers, the main interaction between the substrate and theadsorbate are Van der Waals forces and the layer is said to be physisorbed. During the 1980s anotherapproach was introduced taking advantage of the fact that specific chemical interactions between theadsorbate’s head and the substrate can cause molecules to spontaneously arrange in an ordered layerwhen the substrate is immersed in a solution of the desired adsorbate (self assembledmonolayers, SAM,Fig. 1.1 right) [6, 7]. Such layers are covalently bonded to the substrate and are said to be chemisorbed.Characterization of these layers and their processes is routinely performed using a variety of optical[8, 9], electrochemical [9] and scanning probe microscopy [10] methods. With the exception of the latter,the information obtained from such techniques usually reflects the average characteristics of the layerin an area large compared to the molecular scale. As a consequence, processes that involve the spatialdistribution of the molecules cannot be fully and accurately characterized.21.2. Scope of the Thesisaggregatelower density absencesideal monolayersubstrateFigure 1.2: Some examples of defects in deposited monolayers causing heterogeneity.The most straightforward example of such problems is the heterogeneity in the density of depositedadsorbates. Because of the relevance of ultrathin films in technology development, most of the pub-lished work in the area relate with the applications of these layers and use oversimplified models of theirstructure and characteristics [11, 12]. From this point of view, the deposited monolayers are implicitlyassumed to present homogeneous molecular density and order (Fig. 1.2). Without evidence to backthis simplistic view, most of the articles published in their use take this assumption for granted. Sys-tem idealization is followed by devaluation or nullification of results that do not match with the expectedresponse, without dedicated efforts to understand the underlying causes of such deviation. These unex-pected responses can be caused, for example, by absence of molecules in certain regions of the layeror an excess of molecules that aggregate in a disorganized fashion. These differences in molecule or-ganization can be due to interactions between the adsorbed molecules or be influenced by the presenceof the substrate. Furthermore, these problems grow bigger as the most advanced applications requiremore complex systems, for example by fabricating layers with more than one component or when multi-ple layers are sequentially deposited. In particular the addition of large biomolecules can aggravate theproblem due to their complex structure and interactions.It is the view of the author that this lack of certainty on the quality of the deposited layers is oneof the factors contributing to the fate of many newly proposed systems that end up in proof-of-conceptexperiments and never achieve enough reliability to be successfully applied. It is desirable and neces-sary then to examine these processes using techniques able to report on the spatial distribution of themolecules affecting the overall quality of the deposited layers and their intended responses.1.2 Scope of the ThesisThe main objective of this work is to employ fluorescence microscopy techniques to image organicultrathin layers deposited in metal surfaces. Due to the micrometer-scale spatial resolution obtained31.2. Scope of the Thesisby the use of an optical imaging method, differences in response of different areas of the layer can beevaluated. By controlling the potential of the metallic substrate, a variety of processes can be driven andthe spatial dependent response evaluated.1.2.1 Specific ObjectivesIn order to prove the generality of the approach proposed, the method was used to interrogate a varietyof systems comprising different combinations of substrates, adsorbates and space-dependent phenom-ena. For each of these systems the following objectives were considered:• To deliberately create a heterogeneous Langmuir layer and examine the changes in morphologyand aggregation state upon deposition on a metallic surface using the Langmuir-Schaefer tech-nique.• To investigate the fate of desorbed molecules resulting from the desorption of a self assembledmonolayer. To describe the forces behind the movement of these molecules and investigate pos-sible ways to control its directionality.• To evaluate heterogeneity and improve the signal interpretation of the response of a general inter-est self assembled monolayer. In this case the system proposed is a sensing DNA layer.4Chapter 2Background TheoryThe goal of this chapter is to give the reader enough information on the principles of the techniquesemployed to understand their application throughout this work.2.1 Fundamentals of ElectrochemistryOne of the pillars of this work is the electrochemical control of the interfaces and as such it requires aconsiderable review of theory. In this section, the structure of the electrode | solution interface as wellas the function of the different electrodes in an electrochemical cell will be described.2.1.1 The Electrode InterfaceFundamental to any electrochemical study is the concept of an electrode surface, it being defined asa charged interface. In the most common approach, this interface is composed by a metal in contactwith a solution. In this particular case, the density of charge in either side of the interface is different.The charge carriers in the metallic side of the interface (electrons) are concentrated in the surface ofthe metal, while the charge carriers in the solution side of the interface (ions) are distributed in a morediffuse way in the solution. This arrangement of electrons and ions is known as the electrical doublelayer and its structure will depend in the composition of the interface as detailed next.Non-adsorbing ElectrolyteThe simplest case of an electrical double layer is that one in which only solvent molecules are in contactwith the metal surface. A first approach to describe the structure of this layer was proposed by Helmholtzin 1853. In his model, excess ions counterbalancing the charge in the metal are aligned at a fixeddistance (termed 2 for reasons that will become obvious later) from the surface of the electrode (Fig2.1 left).52.1. Fundamentals of ElectrochemistryIn order to simplify the understanding of real systems, electrochemists usually look for circuits thatadequately represent their electrochemical behavior using electrical components. Such circuits areknown as equivalent circuits. Helmoltz’s model essentially behaves as a two plate capacitor with theexcess charge in solution (qs) being present in one layer near the surface and having equal magnitudebut opposite sign to the charge in the metal surface (qm). According this model, the area-normalizedcapacitance C isCA=εε0d=εε02(2.1)where A is the electrode area, d is the distance between the two plates (or in this case, between themetal surface and the center of the aligned ions), ε0 is the vacuum permittivity and ε is the relativepermittivity of the medium between the plates. According to this model the potential across the interfacedecreases linearly. Although this model explained the broad characteristics of the system, it failed toreproduce the experimentally observed decrease in capacitance near the potential at which the chargeat the metal in contact with the electrolyte equals zero. This potential value is termed potential of zerocharge (PZC) [13]. A further improvement of the model was needed.Gouy and Chapman independently proposed a model in which the ions balancing the electrodecharge were not aligned at a single distance from the electrode but actually formed a diffuse layer thatextended several nanometers into solution (Fig. 2.1 center). In this case the potential drops non linearlyand the rate of potential change (the electric field) varies depending on the concentration of ions in thatparticular section of the solution. As a result, higher polarization potentials and higher ionic strengthproduce more compact layers. The capacitance for this model for a symmetric(z : z) electrolyte can beshown to be [14]CA=€2εε0z2q2en0/kTŠ1/2cosh(zqeϕ0/2kT) (2.2)where qe is the charge of the electron, z is the charge of the ion in multiples of qe, n0 is the numberconcentration of the ions, k is Boltzmann constant, T is the absolute temperature and ϕ0 the potentialat the electrode surface (relative to the bulk of the solution). This model indeed predicted a minimumfor capacitance at the PZC but a continuous parabolic increase as the potential increased or decreasedfarther from the PZC, as compared with the flattening observed experimentally.62.1.FundamentalsofElectrochemistry+   +   +   +   +   +   +   +   +   +   +   +   +   +   +   +   +   +   +   +   +   +   +   +   +   +   ___ ______+++x2+   +   +   +   +   +   +   +   +   +   +   +   +   +   +   +   +   +   +   +   +   +   +   +   +   +   ________+++++   +   +   +   +   +   +   +   +   +   +   +   +   +   +   +   +   +   +   +   +   +   +   +   +   +   metalsolvated anionHelmholtz Gouy-Chapman Gouy-Chapman-Sternsolvated cationsolvent molecule__ ______Potential+ ++bulk solution bulk solution bulk solutionDistancex2CH Rs Cd Rs Cd RsCHFigure 2.1: Different models of the electrical double layer, including their potential variation with distance away from the electrode and equivalentcircuits. For clarity, only a few solvent molecules are presented, but the reader should understand the remaining blank space is also filled withsolvent. Rs, CH and Cd and stand for the solution resistance and the capacitances of the Helmholtz layer and the diffuse part of the double layerrespectively.72.1. Fundamentals of ElectrochemistryStern provided the next step by considering that, in a non adsorbing electrolyte, there is no chargebetween the electrode and the closest distance that the hydrated ions can get to the electrode. Theclosest of these non interacting ions are located at 2 where its concentration (and thus charge excess)is the highest. At larger separations, less excess ions are present until finally the concentration reachesthe same value as in the bulk. The so called outer Helmholtz plane, 2 is a well defined distance andusually a few ångströms away from the electrode surface [15]. Characterizing the thickness of the diffuselayer is more challenging since it gradually decreases over a range of distances. A characteristic lengthtermed Debye length, δ, has been defined asδ=‚εε0kT2z2q2en0Œ1/2(2.3)and considered as the maximum distance at which the charge in the electrode exerts an influence ona charged particle in solution. At distances greater than δ the charge is screened. The Debye lengthvaries with electrolyte concentration and usually takes values of tens of ångströms achieving valuesgreater than 100 Å only at low electrolyte concentrations (~ 10−4 M) [14].Electrically, this can be modeled as two capacitors in series: one for 0≤ ≤ 2 and the other onefor  ≥ 2 (Fig 2.1 right). Since the inverse of the capacitance of two capacitors in series equals thesum of the inverse of each capacitor1Cdl=1CH+1Cd(2.4)where Cdl, CH and Cd are the total capacitance, the capacitance of the Helmholtz layer, and the capac-itance of the diffuse part of the double layer, respectively. For a non adsorbing (z : z) electrolyte (e.g.NaF in Hg and HClO4 in Au) this equation takes the form [14]ACtot=2εε0+1€2εε0z2q2en0/kTŠ1/2cosh(zeϕ2/2kT)(2.5)By comparison with equations 2.1 and 2.2 one can clearly see that each one of these terms representthe respective capacitance from each one of the layers, the main difference being that in the term forthe diffuse part of the layer, ϕ0, the potential at the surface of the electrode has been replaced by ϕ2,the potential at the distance = 2.A window into the energetics of the metal | electrolyte interface can be gained by measuring itsinterfacial tension, since by definition this is equivalent to the Gibbs free energy per unit area of interface.This has been traditionally done in mercury electrodes due to the convenience of having a liquid metal,82.1. Fundamentals of Electrochemistrybut the theory is equally applicable to solid ones, as long as surface stress is not induced as a resultof an elastic deformation [16]. Under these conditions, a plot of the interfacial tension (γ) vs potential(called electrocapillary curve) presents a maximum value at the PZC decreasing in a parabolic way atboth more positive and more negative potentials. This can be explained by considering Gibbs equation[17]− dγ=∑dμ˜ (2.6)where  and μ˜ are the surface excess and electrochemical potential of species  respectively. Equation2.6 establishes that for a species with positive surface excess (i.e. that tends to be segregated towardsthe surface), an increase in the electrochemical potential causes a decrease in the interfacial tension.Remembering that charge in a metal can only exist in its surface, the reason for such positive surfaceexcess becomes apparent.Ionic AdsorptionThe previous discussion assumed that only non adsorbing ions were present in solution. In reality, twokinds of ions could be identified based on their interaction with the metal surface as a consequence fromtheir distance from it. The closest ions have lost their hydration shell and interact directly with the metaleither through van der Waals forces or covalent binding; they are separated by a distance 1 and aresaid to be specifically adsorbed, since such interaction is specific to certain ion / metal combinations.When specific adsorption is present, charge exists between 0≤ ≤ 2. The presence of this chargeinfluences the potential profile away from the electrode surface. The potential at 1 will depend onthe concentration of adsorbed ions and, when the adsorbed charge is greater than qm (known as su-perequivalent adsorption), this potential can have opposite sign compared to the potential of the metal.In this case, shown in Figure 2.2, the adsorbed anions will attract a diffuse layer of cations increasingthe potential to the value in the bulk [18]. The equivalent circuit becomes more complicated since theHelmoltz layer is now divided in two separate slabs. The inner Helmholtz layer consists in the solutioncontained at distances 0≤ ≤ 1 and the outer Helmholtz layer for 1 ≤ ≤ 2. The composition ofthese layers is different and consequently so are their capacitance values (CiH and CoH, respectively).Furthermore, usually full coverage (Θ=1) by ions is not attained and thus a parallel equivalent circuit ofion-covered and solvent-covered areas equivalent circuits is necessary [19].92.1. Fundamentals of Electrochemistryx2x1+   +   +   +   +   +   +   +   +   +   +   +   +   +   +   +   +   +   +   +   +   +   +   +   +   +   bulk solution DistanceCiH CoH Cd Rsmetalsolvated anionspecificallyadsorbed anionsolvated cationsolvent molecule____Potential++++++++____Figure 2.2: Double layer structure with specificanion adsorption including its potential profile andequivalent circuit. Rs and Rad are the solution andadsorption resistances; CiH, CoH and Cd and standfor the capacitance of the inner and outer Helmholtzlayers as well as the diffuse part of the double layer.The potential at which ions start adsorbing onthe metal surface can be analyzed from a thermo-dynamic standpoint. Let us consider the systemdepicted in Fig. 2.2 consisting of specifically ad-sorbing anions and non adsorbing cations. Thespecific adsorption of anions is facilitated by theelectrostatic attraction that the positively chargedmetal exerts on them. One can imagine that ifthe charge in the electrode was negative instead,the electrostatic repulsion to the anions would pre-vent them from getting in contact with the metal.In other words, at positive potentials the lowestenergy state of the surface can be attained withthe presence of anions in the surface whereasat negative potentials a water-covered surface ismore energetically favored. This is schematizedin Fig 2.3a through two independent electrocapil-lary curves: one for a water covered electrode andanother one for anions adsorbed on it at full cover-age. It can be seen how the latter curve presentsa maximum at potentials more negative than inthe solvent covered case due to the fact that whenthe charge on the metal is zero, the potential in it is set by the adsorbed anions. Ignoring kinetic consid-erations, the system will adopt the smaller energy state (smaller interfacial tension) at each potential andions are predicted to adsorb at potentials more positive than the crossing of the two curves. In realitythe coverage of the anions would not go from 0 to 1 at once. Instead it gradually changes creating acurve with a maximum that falls in the area between the two but not necessarily following the curves tothe crossing point. Fig 2.3b shows measured electocapillary curves for a variety of anions with differentadsorption strengths. Notice how at negative potentials all the curves overlap indicating no adsorption,but at positive potentials strongly adsorbing ions like − produce a more asymmetric curve comparedwith OH− whose adsorption is small.102.1. Fundamentals of ElectrochemistryNaBr KCNSKINaClKOHCa(NO )3 2solvent covered(Θ=0)most stable configurationfully anion covered(Θ=1)a bγ (mN m )-1γ (mN m )-1300350400-1.2 0 0.6-0.6300350400-1.2 0 0.6-0.6E - PZCΘ=0 (V) E - PZCΘ=0 (V)Figure 2.3: a) Schematic of the theoretical electrocapillary curves for electrode surfaces covered withsolvent and anion molecules, as well as the thermodynamically most stable configuration. b) Experi-mental electrocapillary curves for Hg in contact with solutions of various salts. Figure b was adaptedwith permission from [18]. Copyright 1947 American Chemical Society.Molecular AdsorptionUpon adsorption of an organic layer, the Gouy-Chapman-Stern two capacitor model is still applicable butsome adjustments are necessary. First, 2 is increased by the presence of the adsorbed layer [20, 21]since these layers are usually impermeable to ions (Fig 2.4). Layers with alkyl backbones containing 2to 21 carbon atoms have a thickness ranging from 5 to 35 ångströms. Thus it can be said that in general,the presence of an adsorbed organic layer will increase the value of 2 compared to the non-adsorbingcase. In addition, the dielectric constant for long alkyl chains is usually in the range of 2 to 3 in contrastwith the value of 80 for water [22]. The combined effect of the change in these two parameters, asseen in Eq. 2.5, is significant decrease (approximately one order of magnitude) in the capacitance ofelectrodes with adsorbed organic monolayers. As a consequence, a larger potential drops across theHelmholtz layer.The organic monolayers can only exist adsorbed on the electrode surface in a relatively narrow rangeof potentials. In the same way as it was explained for the case of ionic adsorption, this stability rangecan be thought of as the potential range in which a electrocapillary curve for an electrode covered with amonolayer has a lower interfacial tension compared to the solvent covered one (Fig. 2.5) The possibleshift in the PZC of the curve corresponding to the monolayer-covered electrode can arise not only asa consequence of any formal charges on the adsorbate but also be caused by dipoles present in it[23]. The smaller slope observed in the region where the layer is stable is also an indication of the lowcapacitance, since the slope of the γ vs E curve equals the surface charge density σ (with opposite112.1. Fundamentals of Electrochemistry+   +   +   +   +   +   +   +   +   +   +   +   +   +   +   +   +   +   +   +   +   +   +   +   +   +   _________+++x2 bulk solutionCd RsCHFigure 2.4: Representation of the electrical double layer, potential profile, and equivalent circuit for thecase of adsorption of an organic monolayer.122.1. Fundamentals of Electrochemistrysolvent coveredmost stable configurationmonolayer covereda b3754004250 0.6-0.6γ (mN m )-1E - PZCΘ=0 (V)Figure 2.5: a) Schematic of the stability region of a deposited monolayer as the overlapping of twodifferent electrocapillary curves. b) Experimental data for the system comprised of 1-octanol in 0.1 molkg-1KCl on amercury electrode. Concentrations of 1-octanol vary for the different curves, being 0, 0.025,0.0500, 0.0750, 0.100, 0.200, 0.300, 0.400, 0.490, 0.600, 0.700, 0.900 and 1.05 mmol kg-1for curves 1to 13. Fig. b reprinted from [25] with permission from Elsevier.sign)σ = −δγδEμ,T,P(2.7)and the derivative of this one with respect to potential is the capacitance [24]CA=δσδEμ,T,P(2.8)where P and μ are the pressure and chemical potential respectively.Due to the difficulties associated with the measurement of the interfacial tension in solid electrodes, itis more frequent to extract equivalent information measuring the capacitance. Although Fig. 2.4 depictsa neutral adsorbate, this treatment is a general one and is applicable regardless of the charge of theadsorbate and its interaction with the substrate (i.e. physisorption [25] vs chemisorption [26]).2.1.2 Anisotropy and Electrochemistry in Crystalline SystemsAlmost all metals are crystalline in their solid state [27]. As a consequence, the orientation of the surfacewith respect to the crystalline lattice will expose surfaces with different atomic arrangements. In practice,electrode surfaces then can be classified according to their crystal homogeneity as mono-, multi- or poly-crystalline. In monocrystalline surfaces all the electrode area exposed to the electrolyte consists of thesame crystallographic orientation. This is usually achieved by first creating a single crystal bead through132.1. Fundamentals of Electrochemistrymelting a piece of Au and slowly cooling it down. The molten gold can be the end of a wire [28, 29] ora discrete drop inside a graphite crucible [30]. After the selected orientation has been aligned, theelectrode is polished flat to expose only the desired orientation. If the cooling process is acceleratedslightly, different nucleation sites appear during the solidification process and several large crystallinedomains exist with grain boundaries between them. Similar polishing will yield flat surfaces with morethan one crystallographic orientation. A related result can be obtained if a single crystal bead is notpolished. In this case, even if all the atoms are ordered in a single continuous lattice in the bulk, thecurvature of the surface will expose different orientations. Due to the lower surface energy of the lowerindex planes (see below), some flat facets form, most noticeably for the {111} orientation. Finally, if thecooling process occurs fast, a high number of nucleation sites form yielding crystal domains that can beeven too small to be resolved optically; these are polycrystalline electrodes.In order to simplify the presentation of the three-dimensional spatial relationship between differentcrystallographic orientations, a two dimensional stereographic projection is frequently employed. Theprocedure to create such a projection is represented in Fig. 2.6a. In this technique, the different orienta-tions are located in the surface of a reference sphere. A point in the surface of this sphere is chosen asthe point of projection (labeled as the south pole in Fig. 2.6a) and lines are drawn from this point to thelocation of the desired orientations in the reference sphere. The projection is achieved by plotting thepoints where these lines intersect a plane (called projection plane) perpendicular to the line that crossesboth the point of projection and the center of the reference sphere (the line joining the N and S polesin the figure). The position of the desired orientations and the distances between them will change inthe stereographic projection if a different point of projection is selected. As a consequence, distancesbetween the different points do not reflect actual distances in the surface of the sphere. Despite thisapparent distortion, the stereographic projection is an invaluable tool in crystallography as a visual aidfor the representation of crystalline orientation dependent phenomena.The stereographic projections are dependent on the symmetry of the system studied. Fig. 2.6bdepicts the stereogram for a crystal with a cubic lattice. The symmetry symbols for the low-index planeaxes have been included. It can be seen that certain planes posses equivalent symmetry. These planes(written between round brackets) also have similar atomic arrangement and thus they are said to forma family of planes (written between curly brackets). The (111),€111Š,€111Šand€111Šplanes, forexample, are part of the {111} family of planes. Because of this reason plane families, rather thanplanes, will be used throughout this work.The energetics of a given surface will be affected by its atomic arrangement. Gold crystallizes in142.1. Fundamentals of Electrochemistrya bFigure 2.6: a) Procedure to elaborate a stereographic projection. b) The full stereogram for cubic lattices.Fig. a reproduced from [31] with kind permission from Springer Science and Business Media; Fig. bcourtesy of Dr. H. K. D. H. Bhadeshia.a face centered cubic (fcc) Bravais lattice, which means the different low index crystallographic planefamilies {100}, {110} and {111} present square, rectangular and hexagonal symmetry, respectively (Fig2.7a). A useful representation of the crystalline structure is through the division of the whole crystalinto polyhedra that fill the entire space. These are called Wigner–Seitz cells [33]. A two-dimensionalanalogy is shown in Figs. 2.7 b, c and d for hexagonal, square and rectangular lattices, respectively. Inthis very simplified system, the electron density inside of each cell is thought to correspond to a particularnucleus.The atomic roughness of a surface resulting from the truncation of the lattice will depend on the trun-cation angle. This is exemplified in the 2D square lattice shown in Figure 2.7c, where the horizontal andvertical surfaces are rougher than the diagonal ones. In the same way, three-dimensional crystals willpresent different roughness depending on the surface crystallographic orientation. Figure 2.8 presentsa selection of fcc surfaces ordered from smoother to rougher. It is particularly noteworthy that higherindex planes consist in steps of the lower index planes.The difference in the surface atomic roughness between crystallographic plane families entails dif-ferences in surface free energy, as a consequence of the unequal number of “broken” bonds that theatoms in the surface have in comparison with their bulk counterparts. The larger the number of theseincomplete bonds the higher the energy of that surface. For a fcc structure, the coordination number(CN) in the bulk is 12, whereas for the topmost layer of atoms in the {111}, {100} and {110} surfaces itis reduced to 9, 8 and 7 respectively. Furthermore, the {110} surface is so “open”, that even the secondlayer of atoms has one broken bond (CN = 11). When the energy of a surface is too high, the atoms152.1. Fundamentals of Electrochemistrya bc dFigure 2.7: a) Schematics of a truncated fcc crystal displaying the low index surfaces {111} (red), {110}(blue), and {100} (green). b, c and d) Two-dimensional hexagonal (b), square (c) and rectangular (d)lattices showing their Wigner–Seitz cells. The illustration in (a) was created using Mathematica code byCorcoran [32].162.1. Fundamentals of Electrochemistry{111} {100} {311}{110} {210}{211}{310}Figure 2.8: Variety of surface planes resulting from truncating an fcc crystal in increasing order of atomicroughness (left to right, top to bottom)might shift position to decrease such energy. If this displacement changes the ordering of the atoms inthe surface layer, it is termed reconstruction [34].The electrochemistry of gold surfaces possessing different crystallographic orientation is affecteddue to the differences in surface energy. To better understand this, let us consider the work function() of a metal. This quantity represents the energy required to remove one electron from the bulk ofthe metal to a “point just outside” its surface. This point is chosen close enough to the surface so thatthe electrostatic (Coulombic) potential is essentially independent from distance, but far enough fromthe substrate so that the induced image interactions do not affect the potential energy of a test charge[35]. This distance depends on the size of the electrode but is ~10−5 to 10−3 cm for a 1 cm diameterelectrode [36]. The work function is composed of several energy contributions:• The chemical potential reflects the difference in energy due to the electron being surrounded bythe Au environment compared to vacuum and since it is a bulk property, it does not depend on thecrystallographic orientation.• An “electron density spreading” (also known as e- spillover) double layer is formed due to theelectrons of the last atomic layer that are not being as tightly bound as the ones in the bulk. Asa result, they extend into the vacuum beyond the space occupied by their corresponding positive172.1. Fundamentals of Electrochemistrya–+–+b+–Figure 2.9: Schematics of electron density spreading (a) and smoothing (b) in a square lattice. The blackline represents the limits of the electronic density. Notice the dependence of the electron smoothing onthe surface crystallography.Crystal orientation Theoretical surfaceenergy (eV)Experimental workfunction (eV)PZC (V vs. SHE){111} 0.72 5.26 0.56{100} 0.88 5.22 0.33{110} 1.31 5.20 0.19{210} 0.11Reference [38] [39] [13, 40]Table 2.1: Compilation of thermodynamic properties of Au surfaces of different crystallographic orien-tations. Work function data was obtained through angle-resolved photoemission spectroscopy. Thereis another commonly cited set of data [22, 38, 41] (obtained through photolelectric measurements) thatshows a different trend. PZC was obtained in 0.05 to 0.2 M NaF.nuclear countercharges. This layer is positive closer to the metal and negative farther into thevacuum. Moving a negative charge from a positive to a negative side of an electric field requireswork, so this term contributes an increase of the work function [37]. This is a surface effect but,since the spread occurs comparably in all crystalline faces (Fig. 2.9a), it does not lead to anisotropyin the work function.• An “electron density smoothing” double layer is formed if the electron density surface is not per-fectly flat. Instead of retaining the shape of the Wigner–Seitz cells, the electronic density of thesurface layer of atoms rearranges to decrease the sharp points and creates a smoother surfacewith smaller energy. This creates regions of negative potential closer to the metal and regions ofpositive potential farther from it. Since the electron is negatively charged, moving from the nega-tive to the positive ends of the electric field is energetically favored and this potential drop reducesthe work function [37]. The (atomically) rougher surfaces will present more smoothing and hencea more pronounced effect (Fig. 2.9b). As a result, {111} > {100} > {110}.The important implication that emerges from this analysis is that although the potential inside the182.1. Fundamentals of Electrochemistrybulk of the electrode is homogeneous, the potential expressed at the surface varies with the surfacecrystallography. Indeed, this difference in potential can be observed, for example, in the potential ofzero charge (PZC) [42]. Its variation with the gold crystallography is notorious in the selected valuesincluded in Table 2.1. While the smoothest surface {111} presents the highest value, the presence ofroughness decreases it as much as ~0.45 V for the roughest one {210}. Although a linear relationship isgenerally found between the PZC and the work function, certain effects such as surface reconstructionand interactions with the solvent or ions in it can modify this dependence. In conclusion, one can expectdifferences in any potential driven process due to the influence of the substrate crystallography.2.1.3 The Electrochemical CellAn electrochemical cell consists of at least two electrodes, andmore typically, three. These are a workingelectrode1 (WE) whose processes we are interested in, a reference electrode (RE) used as a point ofconstant known potential, and a counter electrode (CE) whose function is to pass current to complete thecircuit with the WE without modifying the composition (and thus the potential) of the reference electrode(Fig. 2.10).In order to be able to establish a potential at the working electrode, the potentiostat’s power supply willvary the potential between the working and the counter electrodes until the potential difference betweenthe working and reference electrodes achieves the desired value. This potential difference, as well asthe current, are measured (represented in Fig. 2.10 by the encircled variables) and stored. In solution,electrical current is carried by ions; at neutral pH, the ion concentration in water is small (2×10−7M). Ifthe only ions present besides H+and OH-are the ones corresponding to an electroactive species whoseprocesses we are trying to measure, these ions will present significant migration towards the working andcounter electrodes. Thus, the equations describing the electrochemical response get more complicatedcompared to the case where diffusion is the only mass transport mechanism involving this species. Tosolve this problem a higher concentration of a supporting electrolyte is added, so that these ions carrymost of the charge in solution, reducing the migration of the electroactive species.The pass of current through the circuit has an undesirable consequence. The solution part of the cellpresents a resistance since the ions in solution are not as mobile as the electrons in the metal. Basedon the position of the reference electrode in the cell (Fig. 2.10), this resistance can be divided in solution1The International Union of Pure and Applied Chemistry (IUPAC) recommends the use of the term “working electrode” onlywhen the pass of current through it creates a significant change in the bulk concentration [43]. Although that is never the casein the work presented here, this term has been preferred over the recommended “indicator electrode” due to widespread usageamong the electrochemistry community.192.1. Fundamentals of ElectrochemistryRsRuworking electrodecounter electrodereference electrodeiEpower supplyFigure 2.10: Schematic of a three-electrode electrochemical cell. Black lines represent electronic con-ductors while blue ones correspond to ionic conduction. The dashed line symbolizes the physical bound-aries of the electrolyte containing vessel. Encircled variables represent measuring devices. Rs and Rare the solution and uncompensated resistances respectively. Adapted from [14] Copyright (2000), withpermission from Wiley.resistance (Rs) and uncompensated resistance (R). According to Ohm’s lawE= R (2.9)a potential difference E is created when an electric current  passes through a resistance R. It becomesclear then that the measurable potential difference EWE−RE, is composed of the potential expressedat the electrode interface (as shown in Figures 2.1, 2.2 and 2.4) and the potential “lost” in the solution,commonly referred to as R drop. As a result, in order to calculate the potential difference actually appliedto the WE it is necessary to subtract the term R from the total applied potential EE−RE. It can be seenthat the larger the current and the resistance are, the more the true applied potential will differ from thedesired value. Although several geometric and electronic methods to correct this problem have beendevised [44], the easiest way to reduce it is by increasing the conductivity of the solution.Besides providing ionic conductivity, the electrolyte in question is preferred to be inert with the givenelectrode to be used. Depending on the experiment to be performed, this inertness can be seen only asa lack of faradaic reactions or, in some cases, as an absence of specific adsorption. Last but not least,it is important for the electrolyte to not react with whatever substances we are interested in and haveappropriate stability once in solution.202.2. Electrochemical Techniques2.2 Electrochemical TechniquesDuring the course of this work a number of electrochemical techniques were employed to characterizethe electrode interfaces. Table 2.2 presents a summary of such techniques, followed by the detaileddescription of each one of them.2.2.1 Cyclic VoltammetryCyclic voltammetry (CV) is the most popular of the DC electrochemical techniques.2 In this techniquepotential is swept linearly in a continuous manner between two limiting potential values while the currentflowing through the WE is measured. The system response is typically recorded as a plot of current as afunction of potential (a voltammogram). For comparison purposes, it is common practice to replace theordinate with a variable normalized with respect to the area of the electrode (A), termed current density(J)J=A(2.10)The typical response of this technique for two different systems is shown in Fig. 2.11. In the absenceof a faradaic reaction (blue line), the current flowing is only due to the charging / discharging of the dou-ble layer capacitor. If the capacitance remains relatively constant throughout this potential range, thecorresponding current response will also remain constant until the limit potential is reached. Upon in-version of the potential scan direction, the current abruptly changes sign and asymptotically approachesits steady value. The value of this steady current depends on the capacitance of the electrode andthe sharpness of the current change in sign is determined by the solution resistance as well as someelectronic filters in the potentiostat.When an electroactive species is present, a second source of current has to be taken into account.This faradaic current arises as the electroactive species exchanges electrons with the electrode. Forfast reactions, enough current flows to satisfy the concentration ratio of oxidized and reduced speciesat the surface of the electrode given by the Nernst equationE= E0’+RgTnFln[Ox]=0[Red]=0(2.11)2Although DC stands for “direct current”, the term has been used historically to name electrical signals that do not changeperiodically with time or do so in a very slow way212.2.ElectrochemicalTechniquesTechniquenamePotential perturbation program Typical current response Processed dataCyclicvoltammetrytEtiEiACvoltammetry /differentialcapacitancetEtiECElectrochemicalimpedancespectroscopytEtiZZReImTable 2.2: Summary of the electrochemical techniques employed in this work, including an example of the typical response for a simple electrochemicalcell where faradaic reactions are mostly absent.222.2. Electrochemical Techniques−2−1012−0.5 0.0 0.5 1.0J (mA/cm2)E (V vs. Ag|AgCl)x10Figure 2.11: Cyclic voltammograms of a ethanolic 50 mM solution of LiClO4 in the absence (solid line)and presence of 1 mM ferrocene (dashed line). WE: graphite,  = 100 mV/swhere E0′is the formal potential of the electrode, n is the number of electrons exchanged, F and Rg arethe Faraday (9.648533 x 104 Cmol-1) andmolar gas (8.314 Jmol-1K-1) constants [45], and [Ox]=0 and[Red]=0 are surface concentrations of the oxidized and reduced forms of a redox pair, respectively.As a consequence, at any given potential, current will flow through the system to try to achieve theconcentrations corresponding to that potential. However, when E− E0′is large enough, the surfaceconcentration of the reactant can go to zero limiting the attainable values of [Ox]=0 / [Red]=0 andthe current that flows through the WE. At these values, the current is said to be mass transfer limited.If the solution is not stirred, diffusion is the main transport mechanism and the current drops with acharacteristic 1/pt dependence (where t is time). Consequently, the general shape of the plots ofcurrent as a function of potential (voltammograms) shows a maximum in current; this is, peaks arepresent.The simplest case involving a faradaic current is a reversible reaction of a reagent homogeneouslysolubilized in the liquid phase. In this case, shown in Fig. 2.11 as a green line, the anodic and cathodicpeaks are symmetric and separated by approximately 60/n mV at 25 °C [46]. More complex systemspresent a wide variety of shapes depending on, for example, reversibility of the reaction or availability ofthe reagents and products (e.g. reactants in the bulk vs surface confined) [47].232.2. Electrochemical Techniques2.2.2 AC Voltammetry and Differential CapacitanceBesides the use of techniques where a DC voltage is applied and the corresponding DC current ismeasured, alternating current (AC) techniques provide extra information by superimposing an alternatingelectric potential (EAC) to its DC value (EDC)E= EDC+ EAC (2.12)and recording both the direct and alternating components of the produced current, DC and AC respec-tively= DC+ AC (2.13)One of these techniques, AC voltammetry, employs a potential ramp similar to the one used in cyclicvoltammetry but with the addition of a single frequency sinusoidal potential perturbationEAC = E0 sin(2piƒ t) = E0 sin(ωt) (2.14)where E0 and ƒ are its amplitude and frequency (in Hz). The angular frequency (in radians/s) is denotedby ω. Assuming that the system is linear (i.e. that ∝ E), the AC potential perturbation will produce asinusoidal AC current response AC of the same frequency, amplitude 0 and shifted in phase by φAC = 0 sin(2piƒ t+ φ) = 0 sin(ωt+ φ) (2.15)It is important that the potential and current variations with time caused by the AC wave are muchlarger than the variation in their respective DC values [48]dEDCdtddtEACrms= 21/2piE0ƒ (2.16)dDCdtddtACrms= 21/2pi0ƒ (2.17)where rms stands for root mean square. This is necessary for two reasons. First, it maintains a steadyDC value during the length of the AC acquisition, which means that the system will not change during themeasurement time. Second, if the DC component variation is comparable with the corresponding ACone, higher harmonics of the DC variation show as low frequency noise in the AC measurements. For242.2. Electrochemical Techniquesthese reasons, slower potential scan rates () are used in AC voltammetry compared to CV. Of course,there is no reason to prevent this technique from being applied at =0, i.e. in potential step experiments.To ensure that the system remains linear (as assumed for the derivation of Eq. 2.15) small pertur-bation amplitudes, E0, have to be used. These values usually range between 5 and 10 mV.To understand the working principle of AC techniques it is necessary to realize that Ohm’s law (Eq.2.9), the way we expressed it in section 2.1.3 is in reality a particular case of a much broader behavior.In this particular case, current changes instantly and linearly as a response to the potential applied andis the behavior of an ideal resistor. There are some other electrical elements in which this is not thecase. Consider for example a capacitor; its fundamental equation establishes that the amount of storedcharge (q) is proportional to the capacitance and the potential difference across the capacitorq= CE (2.18)If we differentiate with respect to time on both sides of the equationdqdt≡ = CdEdt(2.19)we can see that in this case the current is not proportional to the potential but rather to the rate ofchange of this potential with respect to time. This has enormous implications in AC techniques, sincesubstituting Eq. 2.14 in Eq. 2.19= Cddt{E0 sin(ωt)}=ωCE0cos(ωt) (2.20)By comparison with Eq. 2.15 and realizing that cos(ωt) = sin(ωt+ pi/2) it is evident then that 0 =ωCE0 and φ= pi/2. This means that even in this case, the current response can be represented by aproportionality with the potential if we consider the phase shift. It is necessary then to redefine a moregeneral form of Ohm’s law asE= iZ (2.21)where the bold typeface has been used to indicate the quantities contain phase information unlike in Eq.2.9 where they do not. Z is the impedance, a more general form of opposition to electron movement thanresistance, since it is not limited to systems that produce currents in phase with the potential perturbation.For convenience it is useful to represent the phase containing quantities as numbers in the complex252.2. Electrochemical Techniquescurrentpotentialϕa bϕsignaltimeReImFigure 2.12: Relationship between two periodic signals with the same frequency. a) The variation ofpotential and current as a function of time; b) The phasor representation of the same signals. Thepotential signal has been taken as the real (Re) axis.plane. Furthermore, since we are generally interested in the relationship between periodic waves (Eand  for example) the phasor representation is commonly employed to visualize them as vectors in thecomplex plane (Fig. 2.12) [14]. Taking the E perturbation as the positive real axis, the terms “real” and“imaginary” are frequently used to refer to the in-phase (Re) and 90° out of phase (Im) componentsof the current, respectively. It is important to note, however, that both currents are real in the physicalsense of the word (i.e. they are charge movement). Of course the actual current response phase mayfall between 0 and 90°, but in that case it can be decomposed into a linear combination of real andimaginary currents.Experimentally, AC voltammetry data is usually measured by using a dual channel lock-in amplifier(LIA). One of the channels is set to measure the current in phase with the potential perturbation andthe other one to measure the current 90° out of phase with the perturbation (i.e. its quadrature). Al-though AC voltammetry can be used to probe the kinetics of faradaic reactions, in this work it will beused to determine the interfacial capacitance of an electrode. When dealing with real electrodes, asopposed to ideal electric components it is convenient to make a distinction on how the capacitance ismeasured. Differential capacitance (Cdiff) corresponds to the slope of the charge vs potential curve at agiven potential [18]Cdiff = −dqdEμ(2.22)where the subscript μ indicates that the chemical potential (and hence the composition) remains con-262.2. Electrochemical Techniquesstant. Integral capacitance (Cint), on the other hand, is defined as the total charge in the electrode ata given potential divided by the difference between such potential and the PZC (which hence must beknown)Cint = −qE− PZC(2.23)In an ideal capacitor the q vs E curve is a straight line, and thus differential and integral capacitancehave the same value. In real electrodes, this is not the case (as a consequence of transitions betweendifferent thermodynamic states as seen in Figs. 2.3 and 2.5) and the distinction is necessary [18].In the more general case the differential capacitance, is composed not only of the double layercapacitance but also of contributions from surface reactions in parallel with it [19] (collectively termed netsurface reaction impedance, Zs). Provided that an equivalent circuit is known that adequately representsthe electrochemical behavior using electrical components the value of the double layer capacitancecan be extracted from the measurements. In the absence of surface reactions, a circuit composed bya resistor in series with a capacitor (named an RC circuit) can be assumed, where the componentsrepresent the solution resistance and the double layer capacitance (Cdl) respectively. In this case, thedifferential capacitance equals Cdl and can be calculated byCdiff =2Re+ 2ImωE0Im(2.24)On the other hand, if surface reactions are present, Eq. 2.24 is still valid in the low frequency limit undersome circumstances (if Zs is composed of parallel branches of capacitors or RC circuits) but Cdiff nolonger equals Cdl but ratherCdiff = Cdl+∑C (2.25)where C is the net capacitance of the th branch of Zs [49]. It can be seen here that a limitation of ACvoltammetry is that for all but the simplest circuits it is not possible to determine the value of each oneof the capacitors composing Cdiff. This limitation comes from the use of a single frequency and, as willbe seen in the next section, other techniques have been developed to overcome it.2.2.3 Electrochemical Impedance SpectroscopyThe basic principle behind the measurements in electrochemical impedance spectroscopy (EIS) is thesame as in AC voltammetry, the main difference being that the latter generally uses a single frequencyat varying DC potential whereas the former commonly employs multiple frequencies at a single DC272.2. Electrochemical Techniquespotential value. As explained above, an AC sinusoidal potential perturbation is applied and the resultingAC current measured. From the real and imaginary currents, corresponding impedance values can becalculated throughZ≡ ZRe+ jZm =ReE02Re+ 2m+ jmE02Re+ 2m(2.26)where j is the imaginary unit (j2 = −1). As expected, impedance is a complex quantity. The real partZRe is nothing but the usual resistance (as in Eq. 2.9) and the imaginary part Zm has been termedreactance.The great advantage of EIS over other techniques is that it not only provides quantitative informationon the characteristics of the electrode (e.g. capacitance) but also gives qualitative information on thecharacteristics of the electrical system [50]. This is due to the predictable variations of the impedance asa function of frequency for different electric components and combinations thereof. Thus, by finding anequivalent circuit that present the same frequency response as the measured impedance, a descriptionof the electrochemical cell is possible.In order to simply and unambiguously represent an equivalent circuit, the Circuit Description Code(CDC) was proposed by Boukamp in 1986 [51]. Originally it was composed of letters to represent theelectrical components and nested parentheses to indicate their connections. A further modificationcommonly employed (and adopted in this text) uses two different brackets for series and parallel con-nections [52]. In this approach, elements enclosed in square brackets are connected in series whilethose in parenthesis are connected in parallel. Some examples of circuit components as well as somesimple combinations of them are shown in Table 2.3. This table also includes the symbol for ageneralized impedance [53]; this is, one from which no specific characteristics are given. Two commonways of plotting impedance data have been used. Nyquist plot relates the values of the real and imag-inary parts of the impedance at different frequencies. Note that although the independent variable isfrequency, it is not explicitly included in the plot. The other form to present impedance data consists oftwo plots, both of them using log ƒ as the abscissa and the magnitude or the phase of the impedanceas the ordinate.Although most of the components in Table 2.3 present ideal behavior, the real electrode systemssometimes do not. In these cases, the impedance response is better represented by the inclusion of aconstant phase element (CPE) whose impedance is described byZ=1Q(jω)m(2.27)282.2.ElectrochemicalTechniquesTable 2.3: Compendium of CDC representations and typical impedance responses of electrical components and simple circuits.Componentor circuitCDC Symbol Z Nyquist plot Bode plotGeneralizedimpedanceZ not definedResistor R RZRe-Z0Imlog flog |Z| ϕ900(º)Capacitor C 1jωCZRe-Z0Im ωlog flog |Z| ϕ900(º)ConstantphaseelementQ1Q(jω)mZRe-Z0Im ωlog flog |Z| ϕ900(º)292.2.ElectrochemicalTechniquesComponentor circuitCDC Symbol Z Nyquist plot Bode plot[RC] R+ 1jωCZRe-Z0Im ωlog flog |Z| ϕ900(º)[RQ] R+ 1Q(jω)mZRe-Z0Im ωlog flog |Z| ϕ900(º)(RC) R1+RjωCZRe-Z0Imωlog flog |Z| ϕ900(º)[R(RC)]R1 R2R1+R21+R2 jωCZRe-Z0Imωlog flog |Z| ϕ900(º)302.2. Electrochemical TechniquesFigure 2.13: Schematic of the three-electrode electrochemical cell illustrating the resistive and (stray)capacitive elements in the absence of the “fourth electrode”. Resistances arise both from solution con-duction and faradaic processes at the electrodes interface. Reprinted from [61] Copyright (2001), withpermission from Elsevier.where the parameter Q ≥ 0 and 0 ≤m ≤ 1. Note that in the extreme cases where m = 0 andm= 1 equation 2.27 reduces to the resistor and capacitor cases respectively. Usually the exponent mis closer to 1 than to 0 and the system is viewed as a non ideal or lossy capacitor. There have beenmany proposed explanations for this behavior, the most common being that solid metallic electrodes(especially polycrystalline ones) present atomic scale inhomogeneities that cause deviations from theideal capacitor [54, 55]. By itself, the parameter Q is meaningless and circuit specific corrections areneeded to extract a capacitance value [56, 57].In fact, one of the biggest risks of misuse of EIS is to fit an equivalent circuit to the experimentalimpedance response without knowing the underlying physical meaning of the components [58, 59]. Evenwhen all the components of the electrochemical cell are ideal, deviations from the expected response canoccur because of the measurement system itself [60]. As pointed out by Fletcher [61], these deviationsarise due to the coupling of the solution and electrode interface resistances (R1, R2 and R3 in Fig. 2.13)with the stray capacitances (CSt) between the electrodes (C4, C5 and C6 in Fig. 2.13). In particular thelarge resistance of the reference electrode (R2) can create large phase shifts at high frequencies. Thiscan be partially alleviated by immersing a Pt or Au wire in solution and connecting it to the RE node witha small (~0.1 µF) shunt capacitor (CSh) [62, 63], providing an alternate path for the oscillatory (AC) flowof electrons at high frequencies. This approach is usually referred to as the “fourth electrode” (4E).312.3. Fluorescence020406080100400 500 600 700Intensity (normalized)λ (nm)excitationemissionFigure 2.14: Normalized fluorescence spectra of AlexaFluor 488. The solid and dashed lines representthe excitation and emission spectra, respectively. Data obtained from the manufacturer website [65]2.3 FluorescenceIn addition to electrochemistry, fluorescence was used throughout the body of this work as the preferredcharacterization method. The fundamentals of this technique, together with the specifics of microscopywill be revised in what follows.2.3.1 Fluorescence SpectroscopyFluorescence is the process by which a molecule releases absorbed energy through the emission ofa photon without a change in the electronic spin of the molecule. In fluorescence spectroscopy somecharacteristic of this emitted light (usually intensity) is measured as a function of the wavelength λ.The excitation and emission spectra (solid and dashed lines in Fig. 2.14 respectively) represent theabsorption and emission processes and are the mirror image of each other as a result of the Franck-Condon principle [64] as explained below.The emitted photons are the result of the relaxation of the molecule from the lowest vibrational stateof the first excited singlet electronic state (S1) to one of many vibrational states of the ground electronicstate (S0). The relative number of molecules that relaxes to each of those vibrational states dependsof the overlap in the vibrational wave functions between the initial and final energy states. As a result,frequencies corresponding to differences in energy states with more overlap will show greater fluores-322.3. Fluorescencecence intensity. Notably, an overlap in the excitation and emission spectra indicates that the moleculehas enough thermal energy such that some of the electrons are not only being excited from the lowestvibrational level of S0 but also from higher energy ones. Changes in the molecular structure can modifythe energy difference between these states and in consequence the fluorescence characteristics of amolecule.During an excitation-relaxation cycle for a population of fluorophores, only a number of the moleculesthat absorbed radiant energy will re-emit it. Some others will decay in a variety of non-radiative paths.Nbs =Nr+Nnr (2.28)where Nbs, Nr and Nnr are the number of molecules that undergo the absorption, radiative and non-radiative decays, respectively. Each of the latter two values are proportional to the number of moleculesin the excited state and the rate constant for the given processNr =Nexckr (2.29)Nnr =Nexcknr (2.30)where kr and knr are the radiative and non-radiative decay rates.The quantum yield (QY) of a fluorophore is defined as the ratio of the number of emitted photonsto the number of absorbed photons. Since in conventional fluorescence spectroscopy one photon isabsorbed or emitted per every molecule that undergoes the process, this is equivalent toQY≡NrNbs(2.31)Substituting equations 2.28, 2.29 and 2.30 into 2.31, quantum yield can be expressed in terms of therate constants [64]QY=krkr+ knr(2.32)2.3.2 Fluorescence QuenchingThe fluorescence intensity can be diminished (quenched) by numerous different processes, some ofwhich will be outlined below.332.3. FluorescenceStatic QuenchingWhen a fluorophore in its ground state interacts with another molecule, the quencher (Q), it can forma complex that does not posses the same spectral characteristics of the free fluorophore [66]. It can be,for example, that the resulting complex absorption and emission spectra shift to different wavelengthsor that the complex is totally non-fluorescent or even non-absorbing. The variation of the fluorescenceintensity with the quencher concentration can be described by the Stern-Volmer equation [64]F ([Q] = 0)F= 1+ KSV [Q] (2.33)in which F ([Q] = 0) and F are the fluorescence intensities without and with quenching respectively,KSV is called the Stern-Volmer quenching constant and [Q] is the quencher concentration. Since thedecrease in fluorescence is due to the smaller number of molecules excited, the remaining excited onesdo not present a difference in their fluorescence lifetime.Collisional QuenchingThis quenching mechanism is a particular case of a class known as dynamic quenching, in which thereduction in fluorescence is caused by an interaction of the excited state of a fluorophore with a quenchermolecule. These interactions favor the return of the excited fluorophore to the ground state without pho-ton emission. As the name suggests, contact between the fluorophore and the quencher is a requirementfor collisional quenching. A wide variety of molecules and metal ions act as quenchers, molecular oxy-gen being one of the best cases known. This type of quenching can also be described by Stern-Volmerequation. Furthermore, since quenching occurs through an additional depopulation of the excited statethrough non-radiative means, in this case the fluorescence lifetime is similarly reducedF ([Q] = 0)F=τF ([Q] = 0)τF(2.34)where τF ([Q] = 0) and τF are the fluorescence lifetimes in the absence and presence of the quencherrespectively [64].Förster Resonance Energy Transfer (FRET)This is another dynamic quenching process in which a fluorophore transfers energy to another moleculevia dipole-dipole coupling. A first requirement for the process to occur is that the absorption spectrum342.3. Fluorescenceof the acceptor molecule overlaps with the emission spectrum of the donor molecule. Although theacceptor molecule does not need to be fluorescent, usually a second fluorophore is employed whichin turn increases its energy into the excited state and then emits a longer wavelength photon to relaxback to its ground state. Second, energy will only be transferred if the donor transition dipole presentsa non-zero component in the direction of the acceptor transition dipole. The greater this component is,the more efficient the FRET process will be; in other words, FRET will be nonexistent for perpendiculardipoles and maximum for collinear ones. Finally, the efficiency of the FRET process depends on thedistance between the donor and the acceptor molecules. The Förster’s radius (R0) is defined as theseparation at which the FRET efficiency equals 50% [67]. Since FRET takes place over moleculardistances (up to 10 nm) it has been used to estimate protein-protein interactions in living cells [68].Quenching by a Metallic SurfaceIn order to understand the interactions of a fluorophore with a metallic surface it is necessary to reviewthe concept of wavevector matching. A wavevector (k) is a quantity used to describe the rate of changeof a periodic signal in space. It can be thought of as a spatial analogue of the angular frequency and isrelated to the wavelength byk=2piλ(2.35)The wavevectors are also related to the refraction index (η) of a given material since the higher therefractive index of a material is, the slower the waves will travel in it. Since the frequency does notchange with the medium, this means more closely spaced waves and thus longer wavevectorsk= k0η (2.36)where k0 is the wavevector in vacuum.When light travels across an interface, the electric field must be continuous in order for the wave tocross from one medium (1) to another (2). In the plane of the interface this condition is represented bythe projection of the vectorsk1 sinβ1 = k2 sinβ2 (2.37)where β is the angle of wave propagation measured from the normal to the interface for phases 1 and2. The change in direction of propagation occurs as a consequence of the need to keep constant thevalue of the components of the wavevectors in the plane of the interface. Substituting Eq. 2.36 in Eq.352.3. Fluorescence2.37 one obtains the familiar Snell’s lawη1 sinβ1 = η2 sinβ2 (2.38)If η1 >η2, at higher angles of incidence the value of the refracted angle predicted by Eq. 2.38 is largerthan 90°. Under this condition the projection of the incident wavevector on the plane of the interface islarger than the magnitude of the wavevector in medium 2. As a result, there is no angle that can satisfyEq. 2.37 and the light does not propagate into medium 2. Instead, it reflects back into medium one in aphenomenon known as total internal reflection.The same concept can be applied to the metal | solution interface, although the equations get morecomplicated due to the complex nature of the refractive index of metals. For this discussion it is con-venient to think of a fluorophore as an oscillating dipole. When a fluorophore is placed near (<1 µm)a metallic surface, several phenomena can occur depending on the metal-fluorophore separation andthe thickness of the metal. At long distances (≥500 nm) the fluorophore can interact with the metal onlyby far-field radiation. Under these conditions, the required wavevector for propagation in the metal (thesurface plasmon polariton wavevector) is longer than the magnitude of the wavevector in the medium 1.Thus, regardless of the angle of incidence, wavevector matching is not possible, light is reflected backand it interferes with the radiation emitted from the fluorophore. As a result the lifetime oscillates as afunction of distance from the surface [69].At shorter distances (up to ~500 nm), however, non-radiative near-field interactions can occur, inwhich the fluorophore induces electronic oscillations in the metal, creating collective electron movement(plasmons) to which the energy is transferred. The closer the fluorophore, the more closely spaced thecharge separation created by these oscillations. If the oscillations are separated enough, wavevectormatching at the interface is possible and the energy transferred to the plasmons can be recovered asfar-field radiation (Fig. 2.15, left). At very short metal-fluorophore separations (tens of nanometers),the oscillations are very closely spaced, wavevector matching is not possible and the plasmon cannotradiate. The energy instead decays as heat, or in other words, the fluorescence is quenched (Figure2.15, center) [70].Under certain circumstances wavevector matching can occur even for close fluorophore-metal dis-tances, and fluorescence is enhanced. In thin metal films, for example, the plasmons can radiate ifa high refractive index prism is placed on the distal side of the film (Figure 2.15, right). Fluorescenceenhancement may also be observed at nanoparticles. In this case, the main factors determining if the362.3. FluorescenceThin lm+ +–+ +– + +–Metalη1η2η2 > η1 Figure 2.15: Near field interactions of a fluorophore with a metal. Left: induced plasmons create largecharge separation that can escape as photons. Center: at closer metal-fluorophore distances, thecharge separation is smaller and no wavevector matching is possible. Right: upon the addition of ahigher refractive index material, wavevector matching is possible and the plasmons radiate to the distalside of the surface. Adapted from [70], Copyright (2005), with permission from Elsevier.metal will enhance or quench fluorescence are the size and dielectric constant of the particle, as wellas the wavelength of the excitation light [70].PhotobleachingThe most irreversible decrease in fluorescence arises from the photochemical degradation of the fluo-rophore. In this process, the molecule in its excited state undergoes a chemical reaction and is trans-formed to a non-fluorescent form. Due to its chemical nature, the rate of photobleaching depends on thestructure of the fluorophore. Moreover, although photobleaching can be a unimolecular process [71],its rate is accelerated when reactive species are present in the medium. Molecular oxygen, for exam-ple, can interact with the excited state of the fluorophore generating reactive oxygen species capableof damaging the fluorophore [72]. For this reason it is frequent to degas samples when fluorescence isused quantitatively.2.3.3 Fluorescence MicroscopyFluorescence microscopy makes use of optical filters placed in the illumination and imaging light paths,both of which go through the same objective in an episcopic setup (Fig. 2.16). The particular combinationof filters for a given fluorophore are mounted in a filter assembly set (also known as filter cube) as shown372.3. Fluorescencein the inset of Figure 2.16.A simplified diagram for a wide-field episcopic-illumination infinity-corrected inverted fluorescencemicroscope is shown in Figure 2.16. The illumination is provided by an arc lamp (Hg or Xe) and collimatedthrough the use of collector lenses. A field diaphragm is used to adjust the size of the illumination regionin the specimen plane. Once at the filter assembly set, the light first passes through the excitationfilter, whose function is to remove light of wavelengths other than the fluorophore absorption band. Thefiltered light then reaches a dichroic mirror, whose main characteristic is to reflect excitation light whiletransmitting the fluorescence. The excitation light is reflected towards the objective and focused onto thespecimen. Upon excitation, the fluorophore in the sample emits light in all directions and some of thosephotons are collected by the same objective used for excitation. The fluorescence is transmitted throughthe dichroic mirror while most of the excitation light reflected by the sample is directed back towards thelamp. The fluorescence light goes through a bandpass (emission) filter to allow only wavelengths thatcorrespond to the emission of the desired fluorophore and to further reduce remaining excitation lightthat could have made it through the dichroic (so called leakage). In typical biological samples using anoil immersion objective the incident to reflected light ratio can be 100:1 [73]. In metallic samples though,it becomes much larger and the role and quality of the filters become crucial. The parallel rays comingfrom the filter cube then pass through the tube lens that focuses it into either the intermediate imageplane of the objective or on a CCD camera.The resolution of an optical microscopy system can be limited by several factors. Optical non-idealities, such as spherical and chromatic aberrations, can decrease the instrument performance. Thelatter one is of particular relevance to fluorescence microscopy. Chromatic aberration is caused by thefact that light of varying wavelengths experience a different refractive indices when passing through theobjective lens materials. As a result, the focal plane for each wavelength is different, thus preventing theuser from achieving an in-focus image for all the wavelengths. Another consequence which has moresevere implications for confocal than for widefield microscopy arises from the fact that the excitation andthe collection volumes are not the same [75]. A common approach to reduce this problem is throughthe use of compound objective lenses where the second (or third and fourth) materials compensate forthe wavelength dispersion.Even with ideal optics, the resolution of optical microscopy is ultimately limited by diffraction of light.If a point light source is to be imaged, the light has to go through an aperture in the objective (and thecondenser diaphragm) in order to make it to the sample. As the light passes through this aperture, itdiffracts; smaller apertures create a more pronounced effect. Because of the finite size of the aperture,382.3. FluorescenceCCD camerafrom light sourcecollectorlenstubelenseld diaphragmuorescence lter cubeobjectiveeyepieceillumination pathimaging pathspecimento eyepieceand camerato objectiveexcitationlterfrom light sourceemission lterdichroicmirrorFigure 2.16: Simplified schematic of an inverted epi-fluorescence microscope. Inset: detailed view of afilter assembly set. Adapted from [74] and [73] with permission from Olympus and Chroma TechnologyCorp.392.3. Fluorescencelight passing through different points in the aperture will interfere with each other creating a diffractionpattern. This pattern, commonly known as Airy disk, is composed of concentric circles of maxima andminima in light intensity. It is important to note that any light sources smaller than the Airy disk will stillappear with the size of the Airy disk in the image. By the Rayleigh criterion, two objects can be said tobe resolved if the central maximum of one of the Airy disks is far enough away from the other disk tooverlap with its first minimum [76]. The resolution of a given imaging system dmin depends both on thewavelength λ of the light to be imaged as well as on the numerical aperture (NA) of the objective used[77]dmin =λ2NA(2.39)and is usually on the order of hundreds of nanometers.40Chapter 3Literature ReviewThe characterization of organic monolayers deposited on solid substrates has used practically everyknown analytical technique. While it is not the objective of this chapter to review each of them, the mostcommon and/or relevant cases will be described. Examples of results obtainedwith layers representativeof an adsorbate type will be discussed, realizing that these are general examples.3.1 ElectrochemistrySince the main focus of this work is the analysis of organic monolayers or multilayers under the influenceof potential, characterization using electrochemical techniques will be discussed first.As explained in section 2.1.1 an adsorbed organic layer can only exist on the surface of an elec-trode in a limited potential range. Compared to an electrode without an adsorbed layer in contact withelectrolyte (henceforth referred to as bare), the magnitude of the currents observed in cyclic voltamme-try is small. If the potential is scanned past the stability limits, peaks in the current are observed afterwhich current values comparable with the bare electrode are obtained. This is an indication that thelayer has been desorbed. The appearance of the peaks in current is expected in the cases of faradaicdesorption/adsorption reactions, for example in the reductive or oxidative desorption of self-assembledmonolayers. However, peaks also appear in the case of physisorbed layers [78], where electron transferbetween the substrate and the adsorbate is absent.The nature of these peaks is better explained using differential capacitance experiments. In thestability potential region, the capacitance is small due to the low dielectric constant of the adsorbedorganic molecules (see section 2.1.1). Outside of that window, the electrolyte-covered electrode surfaceis more stable and high dielectric constant solvent molecules displace the adsorbate. Similarly to the CVresults, the capacitance vs potential plot displays peaks at potentials near the stability boundary. In thisregion, a variation in the potential changes the surface composition, creating areas covered with organicmolecules, while other areas are devoid of them. Under these conditions the change in the electrode413.1. Electrochemistrycharge depends not only on the capacitive charging due to a change in potential but also on the onecaused by a change in the surface concentration [79]dq=∂q∂EdE+∂q∂Ed (3.1)When substituting in the definition of differential capacitance (Eq. 2.22) we obtainCdiff =∂q∂E+∂q∂E∂∂Eμ(3.2)In the absence of faradaic reactions, the peaks are purely the result of the charge that must flowto/from the interface to keep the potential at the desired value after a change in the interfacial dielectricconstant. Since this behavior differs from purely capacitive charging, it is termed pseudocapacitance.When using AC techniques at high frequencies these peaks are significantly reduced (Fig. 3.1) sincethe oscillation in potential is too fast for the layer to adsorb/desorb appreciably and thus no changein composition occurs,€∂∂EŠμ→ 0 [79]. At the limit of infinite frequency, the second term in equation3.2 vanishes and the resulting capacitance value has been termed “constant coverage capacitance”,“high frequency capacitance”, or “infinite frequency capacitance”. Conversely, the zero frequency limitcapacitance has called “thermodynamic capacitance”. Pseudocapacitance peaks are not exclusive todesorption/adsorption processes and can be seen also during phase transitions of the adsorbed layer[80].Capacitance can also be a useful measurement of the coverage of an electrode. When the electrodeis only partially covered with the layer, the circuit can be represented as two capacitors in parallel.Since capacitors in parallel can be added directly, it follows that the capacitance of the electrode canbe described as a weighted average of the capacitance values for electrodes completely covered withadsorbate (CΘ=1) and with electrolyte solution (CΘ=0) [82]:C=ΘCΘ=1+ (1−Θ)CΘ=0 (3.3)where the surface coverage Θ is defined as the fraction of electrode covered with adsorbate of a givenε. For a system incapable of forming multilayers, surface coverage can also be expressed in terms ofthe maximum amount of adsorbate in the surface (mx):Θ≡mx(3.4)423.1. Electrochemistry0102030405060−1.0 −0.5 0.0 0.5C/A (µF/cm2)E − PZC Θ=0 (V)BareC8OH @ 240 HzC8OH @ 10 KHzFigure 3.1: Differential capacitance measurements of a mercury electrode in a 1 M KNO3 aqueoussolution in the absence (dashed line) and saturated with (solid lines) octyl alcohol. The two differentsolid lines were obtained at 0.24 and 10 kHz as indicated. Data from [81].Solving Eq. 3.3 for Θ yields an equation to calculate the coverage if the capacitance values of the bareand totally covered electrodes are known:Θ=C−CΘ=0CΘ=1−CΘ=0(3.5)Capacitance is not only a function of the dielectric constant of the medium but also of its thickness (re-call Eq. 2.1) As expected, multilayers further decrease the capacitance. Such systems can be obtainedby sequential depositions via Langmuir-Blodgett or Langmuir-Schaefer methods [78] or by matchingfunctional groups in self assembled monolayers.Electrochemistry can also be indirectly used to inform of the quality of a layer. By adding a redoxactive molecule in solution, a bare electrode will show faradaic currents corresponding with the oxidationor reduction of such molecule. The deposition of an organic layer can act as a barrier to the electroactivespecies which, unable to reach the electrode does not undergo an electrochemical reaction, and faradaiccurrent is not observed. This approach is commonly employed as a crude indication of the number ofdefects in the deposited layers; more defective layers producing higher faradaic currents.433.2. Other Techniques3.2 Other TechniquesA variety of other methods of analysis have been used in the study of deposited layers. According to theirworking principle they can be broadly divided in two categories. In the first one, that will be generallyreferred here as spectroscopy, a beam of particles is directed onto the sample. The interaction betweenthis beam and the sample releases more particles (not necessarily of the same nature) which are thendetected. The second kind is scanning probe microscopy (SPM), in which a sharp tip is rastered acrossthe surface in close proximity to the sample. The general characteristics of the techniques detailed inthis section are summarized in Table 3.1. When the substrate is an electrode, it is sometimes possiblerecord the measurements while the system is under electrochemical control. This variation is termedin-situ (as opposed to the conventional ex-situ approach).3.2.1 Spectroscopic MethodsOptical methods (i.e. spectroscopic methods where the particles are photons) have been primarily usedfor the determination of the composition and conformation of the layers. Vibrational spectroscopy isinherently informative of the functional groups present in the adsorbed layer. Furthermore, it is alsoinfluenced by the presence of a surface, since some modes are enhanced and some others decreasedbased on their orientation with respect to the surface plane. As a result, a variety of infrared (IR) spec-troscopy methods that make use of polarized light have been widely employed to calculate the orien-tation of the adsorbed molecules on the substrate [83, 84, 85, 86]. Through in-situ measurements, itis possible to determine the differences in orientation as a function of the applied potential. The mainchallenge of these observations is the presence of an aqueous electrolyte that is highly absorbent inthe IR region. In order to minimize interference, complex configurations are used in which only a verythin layer of electrolyte exists between the optical window and the electrode surface. The drawback ofsuch configuration is the increase of the resistance aggravating the R drop problem (see section 2.1.3)in experiments where a current flows through the WE. Potential-induced chemical changes in the layercan also be detected through the appearance or disappearance of signals corresponding to specificfunctional groups [80]. Although IR imaging has been reported [83, 87, 88], the nanometer thickness oforganic monolayers results in very small signals if the sampling area is reduced to achieve good spatialresolution, making this technique impractical for such systems. Even with the use of surface enhancedinfrared absorption spectroscopy (SEIRAS) the spatial resolution is in the order of tens of micrometers[89].443.2.OtherTechniquesTechnique Obtained information Samplingdepth (nm)Verticalresolution(nm)Suitable forimaging?LateralresolutionSuitable forin-situstudies?ReferencesInfrared spectroscopy(IR)chemical compositionmolecular tilting> 500 NA rarely 9-63 µm yes [83]Raman spectroscopy chemical compositionmolecular tilting> 500 NA yes 1-300 µm yes [90, 91, 92]X-ray photoelectronspectroscopy(XPS)chemical compositionlayer thickness0.5-8 0.2 yes 10-1000 µm no [93]Auger electronspectroscopy(AES)elemental composition 0.5-8 yes 50 nm no [94]Neutron reflectivity(NR)layer thicknesssolvent fraction~500 2 no ~3 cm yes [80]Secondary ion massspectrometry(SIMS)chemical composition yes 0.2-10 µm no [95]Atomic force microscopy(AFM)layer thicknessmolecular orderterminal polarity~0.1 0.1 yes ~ 30 nm yes [10]Scanning tunnelingmicroscopy(STM)layer thicknessmolecular orderelectronic density~0.1 0.01 yes ~ 0.1 nm yes [10]Fluorescence microscopy domain segregation > 500 ~300-1000(confocal)yes ~0.3-1.0 µm yes [96, 97]Table 3.1: General characteristics of the most common analytical techniques for deposited organic layers. NA = Not applicable453.2. Other TechniquesAs an alternative to IR, Raman spectroscopy has been used for similar purposes [98], providingvibrational information without interference from water [99]. Moreover, the signal-to-noise ratio can beimproved through the use of surface enhanced Raman spectroscopy (SERS). The signal enhancementcan be explained by a highly localized electric field generated by nanoscale roughness in the metallicsubstrates [100]. Raman is more suitable for microscopy (Fig. 3.2a) than IR due to the higher diffraction-limited resolution, a consequence of the shorter wavelength (usually visible), as well as to the enhance-ment created by the surface. In fact, most often the resolution is limited not by diffraction but rather bythe dimensions of the laser illumination spot. The drawback of SERS is the convolution of the substrateand adlayer characteristics in the resulting signal. Any heterogeneity observed, for example, can becaused by spatial differences in the distribution of the adsorbate or by variations in the enhancementproperties of the substrate [92].Higher energy photons can also be used to probe the chemical structure of the layers. X-ray pho-toelectron spectroscopy (XPS) is widely used for chemical composition analysis for both organic andinorganic thin films. In this technique a focused X-ray beam is directed towards the sample from whichelectrons are ejected. The kinetic energy of these photoemitted electrons is informative of the elemen-tal identity of the atoms as well as of the chemical environment of the atom, such as oxidation state,bonding to different atoms and different bonding order. By measuring the electron yield at two differentemission angles, the order of moieties of a monolayer containing different atoms (for example in thehead compared to the tail) can be deduced as well as the layer thickness [101]. Although XPS can beused for imaging, its spatial resolution is not very high considering the time requirements to scan overthe sample. For this reason, it is sometimes used as a one dimensional line scan technique [93].Particles other than photons can also be used as a probe beam to interrogate the surfaces. Electrons,for example, are used in Auger electron spectroscopy (AES) to obtain elemental composition information.Although the energy of the Auger electrons is also dependent on the chemical environment, the chemicalshifts are smaller and less well characterized than the corresponding XPS ones. However, the spatialresolution is superior for AES than for XPS, making it more suitable for elemental mapping (Fig. 3.2b)[102]. High vacuum is necessary in both XPS and AES to prevent the emitted electrons from interactingwith gas molecules, limiting the applicability of these techniques to ex-situ analysis.Neutron reflectivity (NR) is another useful technique to determine the thickness of an organic layer.Neutrons produced in nuclear reactors or spallation sources are directed, at grazing angle, towards thesample, where they become scattered by the nuclei of the atoms in the sample. The presence of thethin layer creates interference patterns in the reflectivity profile. By fitting this profile to a model, the463.2. Other Techniquesfitting parameters can be interpreted in terms of the thickness of the layer, the scattering length densityand the volume fraction of solvent in the layer. The latter parameter is of great importance as it informson the quality of the packing. Higher solvent content possibly indicates defective (patchy) layers. Ifthe length of the molecule is known, the thickness can also be used to calculate a mean molecular tiltangle with respect to the surface normal [80]. Because neutrons interact only weakly with matter, in-situexperiments are possible. However, and for the same reason, large samples in the order of 10 cm2 arerequired to obtain a good signal preventing the use of this technique for imaging purposes.Finally, even ion bombardment can be used to interrogate the surfaces. When these (primary) atomicions collide with the surface they release secondary ions that are analyzed in a mass spectrometer.This technique, known as secondary ion mass spectrometry (SIMS), has the great advantage of beingchemically general and providing highly detailed chemical structure information. On the other hand,however, the need to desorb some of the material in the layer for mass analysis necessarily makes ita destructive method of analysis. Due to the low amount of material in the surface and the wide rangeof masses that need to be measured, time-of-flight mass analyzers are preferred. Similar to Auger,scanning the probe beam allows imaging across the surface (Fig. 3.2c) [95, 103] but the need forvacuum precludes in-situ experiments.3.2.2 Scanning Probe MicroscopyBy measuring the vertical deflection of a cantilever tip onto which a sharp tip is attached, atomic forcemicroscopy (AFM) is capable of differentiating between layers of different heights and morphologies.Furthermore, while dragging the tip on the surface, friction creates horizontal twisting in the cantilever.This lateral friction is dependent on the adhesion forces between the tip and the substrate, which in turnis determined by the functional groups at the end of the monolayer [106]. Therefore, this technique isuseful for differentiating domains with dissimilar composition (Fig. 3.2d) [104] even if their heights arethe same [107]. The contrast between the different functionalities can be enhanced by modifying thetip with a specific functionality. In the current state, however, a lack of a widely recognized frictionalcalibration method [108] limits friction measurements, making comparisons only valid when the samespecific conditions (and equipment) are used.Scanning tunneling microscopy (STM) is another probe technique in which the tunneling currentbetween an electrically conductive substrate and the tip is measured. If the tip is set at constant heightSTM yields information on the local density of states of the material. On the other hand, this techniquecan also be used to map topography if the piezoelectric feedback circuit is set to maintain constant473.2. Other TechniqueseScanning tunneling microscopybare covered25 nm48 µm460 µm15 µm80 µmdtopography lateral forceAtomic force microscopycSecondary ions mass spectrometryS AubSH- HSO4-Auger electron spectroscopya1440 cm-1Surface enhanced Raman spectroscopy1335 cm-1Figure 3.2: Selected examples of spatially-resolved organic monolayer characterization. a) SERS imag-ing of a photopatterned p-nitrothiophenol (signal at 1335 cm-1) SAM. Exposure to UV illumination cre-ates a photoproduct with a signal at 1440 cm-1. b) AES map of a HS(CH2)11Cl SAM on Au strips.c) SIMS images of an alkylthiol SAM irradiated with UV light through a grid, creating alkylsulfonate. d)Topography and lateral force AFM images of alkylthiol SAMs terminated in methyl (dark stripes) andhexaethylene glycol (bright stripes) groups. e) STM image of a Au{111} surface partially covered with3,4,9,10-perylene tetracarboxylic dianhydride; Au reconstruction is visible in the background. The ap-proximate length corresponding to the height of each pair of images is indicated in their left margin.Images a-e adapted from [90], [102], [103], [104] and [105] respectively, with kind permission from theAmerican Chemical Society and Springer Science and Business Media.483.3. In-situ Fluorescence Microscopycurrent flow between the substrate and the tip. STM’s very high spatial resolution made possible theacquisition of images (Fig. 3.2e) in which distinct domains of ordered molecules coexist [109] and evensingle molecule defects have been imaged [110]. Phase transitions in the order of the layers has alsobeen observed by performing scans at varying temperatures [111].Since there are no issues with beam permeability through the media, both SPM techniques aresuitable for in-situ studies. In fact, STM is by default an electrochemical technique. The limitations SPMtechniques include the relatively small field of view (rarely over 100×100 µm2) and the lack of chemicalspecificity. More troubling is the uncertainty of the effects of the tip on the layer. This is specially true forsystems in which the molecules are too mobile [111]. In this situation, imaging is impossible since thetip pushes the molecules around. Such is the case, for example, of tethered DNA. In order to be able toresolve single molecules, DNA has to be pinned to the surface using electrostatic interactions causedby chemical agents [112] or applied electric fields [113].3.3 In-situ Fluorescence MicroscopyAn in-situ spectroelectrochemical technique based on fluorescence has been developed as anothertool to characterize adsorbed organic layers onto metal electrode surfaces. Taking advantage of thefluorophore quenching when placed near (in the nm range) a metallic surface (section 2.3.2) [114, 70],potential controlled desorption of the deposited layer from an electrode surface can be monitored bymeasuring fluorescence intensity. This method was initially conceived in the Lipkowski laboratory [78]using a fluorescent analog of octadecanoic (stearic) acid transferred to the electrode surface via theLangmuir-Schaefer method. Spectral resolution was provided through the use of monochromators forwavelength selection and photomultiplier tube for detection. It was demonstrated that at desorptionpotentials, a significant increase in fluorescence could be detected with the characteristic spectrum ofthe employed dye.The same principle has been developed further to include laterally resolved microscopic measure-ments. Figure 3.3 shows the basic setup required to conduct said experiments. A glass cell possessinga thin glass window (250 µm) at its bottom is placed atop of an inverted fluorescence microscope. Thiscell houses all the electrodes placed in such a way that the bottom of the working electrode is imagedby the CCD camera. Electrochemical experiments such as cyclic voltammetry, differential capacitanceand impedance spectroscopy can then be performed while simultaneously recording the fluorescenceresponse. Note that the attainment of spatial resolution comes at a price of decreasing the temporal493.3. In-situ Fluorescence MicroscopyCCD cameralampWECERE250 µm thick windowFigure 3.3: Basic setup to conduct in-situ fluorescence experiments. Adapted from [74] and [115], withkind permission from Olympus America Inc. and Bioanalytical Systems, Inc.resolution of the measurements. While a photomultiplier tube rise time is a few nanoseconds, a camerarequires exposure times in the order of seconds.The ideal systems for this analysis technique are layers made of naturally fluorescent or fluorophore-labeled molecules. Alternatively, non-fluorescent layers can be spiked with a small amount of a labeledmolecule that mixes into the layer. Due to the widespread use of fluorescence in biological microscopy,a wide variety of labeled probes are commercially available suitable for a wide range of environmentpolarity. Chemical specificity is not one of the strengths of fluorescence microscopy. However, manyfunctional group-specific reagents have been developed to detect molecules of interest. Finally, thenumber of different fluorophores that can be detected in the same sample rarely exceeds three, due tothe need of large spectral separation stemming from the use of relative wide bandpass filters.Previous work in this laboratory has been performed on a variety of physi- and chemisorbed mono-layers. Rather than enumerating all the different systems here, such results will be described in detail503.4. Perspectivein the specific chapter where they have more relevance.3.4 PerspectiveIt is clear from the description presented above that full characterization of a deposited layer cannotbe attained using a single analytical technique. Several of them (IR, XPS, NR) lack practically attain-able micrometer lateral resolution and are most commonly used to report on the average value of thedesired parameters over a large illuminated area. Spectroscopic imaging techniques using massive par-ticles (AES, SIMS) provide higher spatial resolution and chemical specificity, but the strong interactionof electrons and ions with matter imposes the requirement of high vacuum, making in-situ experimentsimpossible. The use of photons, on the other hand, provides relatively easy access to in-situ measure-ments because their path is not as strongly impeded. Furthermore, compared to STM, optical techniquesare less disruptive, and there is no need for the layer to remain static. While diffraction limits the res-olution of optical techniques, micron scale information can still provide relevant enough information inheterogeneity of the adsorbed layers and in their desorption processes.In this context, in-situ fluorescence microscopy appears as a useful technique to interrogate pro-cesses that result in distance dependence of layers adsorbed on electrodes. Furthermore, unlike SERS,SPM, and electrochemistry itself, the acquisition of useful information is not limited to the presence ofthe molecules immobilized or in close proximity to the substrate. In fact, the opposite is true, since moresignal is produced when the molecules are farther from the surface. Therefore, it is the opinion of theauthor that the most significant strength of the presented approach is the study of dynamic processesotherwise inaccessible to other techniques.51Chapter 4ExperimentalThe reagents, instrumentation and electrochemical and spectroelectrochemical cells employed are de-scribed here. A description of the in-situ fluorescence procedure is provided as a general overview. Thepeculiarities of the experiments for each particular system, however, will be detailed in their respectivechapters.4.1 Reagents4.1.1 WaterThe use of ultrapure water is of the utmost importance for the success of the experiments described. Thewater employed was purified in the laboratory with a Millipore system. The majority of the water usedwas produced by a Elix S connected in series with a Milli-Q Gradient A10 system. Latter experimentswere performed with water produced by a Milli-Q Integral 5 system. Both of these systems producedwater with a resistivity >18.2 MΩ ·cm and a value of total organic carbon between 1 and 3 ppb.4.1.2 General ReagentsDuring the course of this work, several substances were employed as substrates, surface modifyingagents, buffer components, electrolyte and solvents. These compounds are listed together in AppendixA with some of their properties.4.1.3 Fluorophore-labeled moleculesTwo different types of fluorophores were used to carry out the spectroelectrochemistry experiments. Aclass of dyes based on the 4,4-difluoro-4-bora-3a,4a-diaza-s-indacene core, commercialized under thetrade name BODIPY have become popular due to their insensitivity to changes in the environment (e.g.524.1. ReagentsNNBFF OOOOOPOOO NNNBFFSHBODIPY-HPCBODIPY-C10-SHFigure 4.1: Structures of the BODIPY labeled molecules employedpH). The absorption and emission wavelengths can be tuned by changing the core substituents; increas-ing the length of electron delocalization is reflected in a red shift in the spectra. For availability reasons,two green-emitting dyes were used: BODIPY FL and BODIPY 493/503. The former was labeling a 1-hexadecanoyl-sn-glycero-3-phosphocholine (HPC) based lipid and was acquired fromMolecular Probes(BODIPY-HPC, catalog number D-3792, Fig. 4.1 top). The latter one was part of a custom synthesizedmolecule, termed BODIPY-C10SH, which includes a ten carbon alkyl chain and a thiol terminal group(Fig. 4.1 bottom) [1]. Since the difference in the wavelength of maximum absorption and emission forthese molecules is less than 10 nm, the same set of filters was employed (as described in section 4.3.2)and both dyes will simply be referred to as BODIPY for the rest of the manuscript.These fluorophores are known to form two types of dimers [116]: DI which is not-fluorescent (Försterradius R0 = 58 Å) and DII (R0 = 42 Å) which is believed to be a ground state dimer and acts as a FRETacceptor with the excited state of the monomer as the donor. The resulting emission at 630 nm is red-shifted from the monomer emission [117, 118]. The DII dimer will be simply referred to as the BODIPYdimer in the rest of the text.A disadvantage of the BODIPY dyes is that photobleaching occurs at a relatively fast rate. A morestable fluorophore was used in experiments where photostability was more important than the aggre-gation information provided by the BODIPY dimerization. Alexa Fluor dyes are a series of sulfonated534.2. MaterialsOSO3SO3NH2NH2COONHOOPOOligonucleotideOOFigure 4.2: Structure of the Alexa Fluor 488 oligonucleotide modificationderivatives of common rhodamine or coumarine based dyes [119]. The negative charge imparted by thesulfonate group reduces the tendency of aggregation of the dyes as well as increasing their solubility inwater. On the other hand, Alexa Fluor dyes are more expensive than their non-sulfonated counterparts.For this work Alexa Fluor 488 was chosen because of its similar spectral characteristics compared tothe BODIPY dyes used, allowing for the employment of the same set of filters. This dye was used tolabel DNA as described in the following section (Figure 4.2).4.1.4 DNAThe DNA sequence was custom synthesized by Eurogentec North America and used without further pu-rification. The doublymodified (thiol and fluorophore) 30-mer sequence 5’-(thiol C6)-CTG-TAT-TGA-GTT-GTA-TCG-TGT-GGT-GTA-TTT-(AlexaFluor 488)-3’ was obtained as a powder and purified via reverse-phase and ion-exchange dual high performance liquid chromatography (HPLC) by the manufacturer.4.2 MaterialsVolumes of 5 mL and higher were measured with class A volumetric flasks. Smaller volumes were mea-sured using GlobalTown Accupette II adjustable pipettes. Gilson Diamond pipette tips were autoclavedat 122 °C for 40 min to remove enzymes that degrade DNA (DNases). Volumes of 5 mL or larger ofsolutions were contained in the same volumetric flasks used in their preparation. Smaller volumes werecontained in Bio Plas siliconized microcentrifuge tubes. These tubes are certified to be free of DNasesand RNases. Glass and polytetrafluoroethylene (PTFE) material was washed by soaking in a mixture544.2. Materialsof 1:1 volume of concentrated sulfuric and nitric acid and heating to 75-80 °C for 3-5 h. Material wassubsequently rinsed with ultrapure water and stored overnight filled or soaked in water to remove acidthat would otherwise leach from the glass into the working solutions.4.2.1 Working ElectrodesAlthough ultrathin layers can be deposited in a large variety of substrates, gold was selected in thiswork for a few reasons. First, the employed substrate has to be conductive since we need to control itspotential accurately. Any metal would comply with this requirement; however, the accessible potentialrange is limited by reactions involving the solvent or the electrode material itself. Coinage metals posseshigh oxidation potentials enabling the observation of various processes associated with the depositedlayer without significantly modifying the substrate. However, some other coinage metals like Pt bettercatalyze solvent (H2O) decomposition reactions. Added to it’s relatively low cost and melting point, theprevious arguments make gold a popular choice as substrate for ultrathin film studies.The gold substrates employed in this work varied in size, shape and crystallography depending onthe desired information to be obtained from a particular experiment. Their fabrication and characteristicswill be presented in their respective chapters.4.2.2 Electrochemical CellsWECEREFigure 4.3: Cell used for electrochemical experi-ments.Two general custom-made glass cell designswere employed depending on the experiment.When electrochemistry alone was to be mea-sured, a large glass cell with a typical volume(V) of ~70 mL was employed. This cell was con-nected to another reservoir with a ground glass fit-ted salt bridge. The second reservoir contained aBeckman Coulter 511100 saturated calomel ref-erence electrode (SCE) dipped in saturated KClsolution. The working electrode is fit from the topusing a B14/23 ground glass joint. Smaller B5/13joints are used to introduce the counter electrodeas well as Ar gas connections. Before the exper-554.3. Instrumentsiments, the electrolyte solution was bubbled with Ar for 15 min to remove dissolved O2 and yield an Arsaturated solution. During the experiments, solution bubbling was stopped but Ar flow was maintainedabove the solution surface.The spectroelectrochemical measurements were performed in cells with a 250 µm thick glass bottomoptical window. A general design of such cells is presented in Fig. 3.3. However, some variation in thedesign was made to adapt for the particular needs of each experiment as will be described in theircorresponding chapters. The volume of electrolyte in the cells ranged from 2.5 to ~50 mL. In all thesecells an overhanging reservoir for the reference electrode was attached via a ground glass joint andfilled with the working electrolyte. The solution in the RE compartment would be contaminated with Cl−which was prevented from flowing back to the electrolyte in the cell through the use of a PTFE stopcock.The choice of reference electrode was made considering both the cell design as well as the workingsolutions. Although a smaller Ag|AgCl reference electrode (BASi RE-6) is ideal because of its size, itcannot be used when Tris buffer solutions were employed, since Tris forms a precipitate with Ag blockingthe frit glass at the end of the electrode. In this case a conventional SCE electrode was used.In most experiments, the counter electrode (CE) was a Pt coil. For some other experiments, as notedin the text, a smaller Au CE was employed.4.3 Instruments4.3.1 ElectrochemistryTwo models of potentiostats were used in these experiments; an analog FHI ELAB and a digital EcoChemie Autolab PGSTAT30 equipped with the FRA2 and ECD modules for impedance and low currentmeasurement respectively. Also, two models of lock-in amplifiers (LIA) were used: an EG&G 5208 anda SRS SR830.As described in section 2.2.2, capacitance measurements were performed by feeding a 200 Hzsinusoidal potential wave produced by the lock-in amplifier (~ 5 mV rms) to an external input of thepotentiostat and adding it to the DC potential perturbation desired. Themeasured current is then fed backto the LIA that separates it into its in-phase and quadrature components and capacitance is calculatedusing Eq. ??.564.4. Spectroelectrochemical ExperimentsFluorophore Excitation filter Dichroic mirror Emission Filter ManufacturerBODIPY monomer(chapter 5)450-490 nm 505 nm 515-550 nm OlympusBODIPY dimer(chapter 5)450-490 nm 570 nm 613-695 nm OlympusAlexa Fluor 488,BODIPY monomer(chapter 6)450-490 nm 495 nm 500-550 nm ChromaTable 4.1: Filters employed for fluorescence measurements4.3.2 MicroscopyFluorescence imaging was performed using an Olympus IX-70 inverted microscope equipped with 5×(LMPlanFl, dry, NA = 0.13), 10× (UPlanFl, dry, NA = 0.3), 40× (LUMPlanFLN, water immersion, NA =0.8), and 50× (LMPlanFl, dry, NA = 0.5) Olympus objectives. The diffraction limited resolution was 0.8and 0.5 μm for the 10×and the 50×objectives, respectively as calculated using Eq. 2.39 and 500 nmlight. Three monochromatic charge-coupled device (CCD) cameras were used: SPOT RT (DiagnosticInstruments, 1080 × 1520 pixels), Evolve 512 (Photometrics, 512 × 512 pixels) and ST-7XMEI (SBIGAstronomical Instruments, 765 × 510 pixels). The cameras were controlled either through the manu-facturer software, µManager [120] or in-house built Labview routines. An Olympus illuminator equippedwith an Osram XBO xenon bulb or an X-Cite eXacte mercury light source were used to provide illumi-nation. The filter sets used are listed in Table 4.1. Brightfield (i.e. white light) images were used as areference to relate the position of the fluorescence signals with the structure of the electrodes. The ex-posure times are indicated in the description of each experiment. All the images were processed usingImageJ [121].4.4 Spectroelectrochemical ExperimentsAlthough the study of each particular system required different experiments, a common spectroelec-trochemical experiment procedure was employed. Therefore, it is explained here rather than in eachchapter.In a typical in-situ fluorescence experiment fluorescence imaging and capacitance measurementswere performed during potential step perturbations (Fig. 4.4). The potential of the WE was switchedbetween a base (Eb) and step (Es) potential values. The Eb was kept constant while Es was variedfromEb to a final potential (Ef) in constant decrements. For certain experiments, an extra step to Eb574.4. Spectroelectrochemical ExperimentsEbEfEsEbEfEtwithout steps to Ebwith steps to Ebexposureimagetransfercapacitancemeasurement+–Figure 4.4: Schematic of the potential perturbation employed in a basic in-situ fluorescence experimentwas added in between two step potential values. One fluorescence image was taken while the potentialwas held at each Eb and Es value. The waiting time between potential steps was made longer than thegreatest exposure time to allow for the capacitance measurement and image transfer. The fluorescenceimages were saved as image stacks and analyzed as described in each section.58Chapter 5On the Effect of the Substrate inLangmuir-Schaefer DepositionLayers composed of lipids have been employed as models for biological membranes, either floatingon a solution or deposited in a substrate. In particular, layers of mixed lipid composition have recentlyreceived more attention in an attempt to understand the relation between the composition and function ofcomplex surfactant mixtures like lung surfactant [122, 123], ocular tear films [124] and lipid rafts [125].These complex mixtures are useful models for studying the phase behavior and the heterogeneousnature of the lipid membranes. In this chapter the in-situ fluorescence microscopy method is used tocharacterize the effects of the substrate when a floating octadecanol-based Langmuir layer is depositedthrough the Langmuir-Schaefer method.5.1 Langmuir LayersMolecules containing hydrophobic and hydrophilic moieties are said to be amphiphilic due to their dualpolarity characteristics. Depending on the structure of thesemolecules, they organize in water in differentways. Lipid molecules that have a single alkyl chain, for example, tend to form micelles while those withtwo hydrophobic tails are sterically hindered in micellar form and form bilayers instead.When carefully placed on the water surface, amphiphiles organize themselves in a oriented mono-layer, with the polar head pointing towards the water and the hydrophobic tail pointing towards the gasphase. These floating films have been termed Langmuir layers, after Irving Langmuir who worked withthem at the beginning of the twentieth century [126]. The order state of the monolayer depends on thestructure of the amphiphile as well as the surface molecule density and the temperature. For a givenamphiphile at constant temperature, compression experiments can be performed by changing the areaavailable per molecule (Fig. 5.1). This area is calculated from the size of the water surface and thenumber of molecules added to that surface, and is thus termed mean molecular area (MMA).595.1. Langmuir Layerstilted condensed(TC)liquid expanded(LE)gas(G)untilted condensed(UC)bucklingTC + GTCUCC  OH18fractureTCDPPCTC + LELELEPOPCLE + G020406015 20 25Surface pressure (mN/m)Mean molecular area (Å2)20 40 60 80 100 120Mean molecular area (Å2)OHOOPOOOOOONOOPOOOOOONFigure 5.1: Compression isotherms for octadecanol, DPPC and POPC Langmuir layers at 20.5 ±0.5°C. Data was obtained from [132], [133] and [134] respectively.At a given temperature, an amphiphile can present different phases depending on the MMA. Whenthe density of molecules on the surface is low (large area per molecule) the molecules do not interactwith each other and are said to be in a two-dimensional gas (G) phase. An increase in the moleculessurface density (achieved for example by movement of a trough barrier or reduction of the volume ina drop or bubble [127, 128]) forces the molecules to interact more, achieving a liquid expanded (LE)phase in which no order can be detected and both tilt angle and bond orientation (cis vs trans) arerandom. Further compression yields a region where the liquid expanded phase coexists with an orderedcondensed phase in which the tails are tilted with respect to the water surface; therefore giving it thename tilted condensed (TC). In this phase the chains acquire an all-trans conformation and a uniform tiltangle. At smaller molecular areas this condensed phase undergoes a transformation by aligning the tailsperpendicular to the water surface, yielding an untilted condensed (UC) phase [129]. Further reduction inthe area available results in buckling and fracture of the monolayer forming three-dimensional structureslike multilayer islands, folds or vesicles [130, 131].As seen in Fig. 5.1, compression causes an increase in the surface pressure () defined as the605.1. Langmuir Layersdecrease in interfacial tension caused by the presence of the amphiphile [43]= γ0− γ (5.1)where γ0 and γ are the interfacial tension values without and with amphiphile respectively. Physically,this pressure arises from intermolecular repulsive forces that oppose the compression of the layer.Not all the aforementioned phases are necessarily present in every compression isotherm. The vs MMA curves give some information about which phases are present for a given amphiphile under theconditions at which the experiment was performed. For aliphatic alcohols like 1-octadecanol (C18OH),the liquid expanded phase is nonexistent and the gas phase directly transitions to the tilted condensedphase [135, 127], which shows a linear increase in  upon compression. In saturated phospholipids,the presence of a second chain increases the degrees of freedom for packing. As a consequence, theintermediate LE phase is observed and its transition to the TC one yields a coexistence region observedas a plateau in the isotherm [128]. Note that for the example shown here, dipalmitoylphosphatidylcholine(DPPC), layer fracture occurs before the UC phase can be observed. Furthermore, if at least one of thephospholipid chains is unsaturated, as in the case of 1-palmitoyl-2-oleoylphosphatidylcholine (POPC),the “kink” in the chain causes the molecule to occupy a larger “footprint”, also increasing the numberof possible orientations. Packing then becomes less efficient and the layer cannot condense at roomtemperature. The presence of the LE phase is generally observed as a continuous curve in the isotherm[128].A particular value of  is specially relevant. When a crystal of pure amphiphile is placed on a cleangas | liquid interface, it will release amphiphile molecules as a result of the stress created by the interfacialtension gradient across the surface (Marangoni effect) [136]. This presence of amphiphiles in the surfacewill in turn reduce its interfacial tension, until eventually the stress will be balanced by the cohesiveforces inside the crystal. At this point, the crystal will stop “dissolving” and coexist in equilibrium withthe monolayer [137, 138, 139]. Therefore this value of surface pressure is termed equilibrium spreadingpressure, e.Note that according to thermodynamics, a monolayer should not be compressible to values of >e, since the material in the monolayer should favor the bulk phase. However, due to kinetic con-siderations, it is possible to achieve metastable states if fast compression rates are used [140]. For1-octadecanol and POPC in water e are 34.3 mN/m [141] and 46.8 mN/m [140] respectively. Lookingback at the isotherm, this means that at room temperature and e C18OH layers appear rigid while615.2. Transfer onto a Solid SubstratePOPC ones are more fluid. For DPPC the situation is more complex since the crystals tend to sink intothe aqueous subphase [142, 143].Upon more detailed inspection, the relatively simple phase description of the system turns complex.For 1-octadecanol, for example, small changes in the slope of the compression isotherms have revealeda variety of slightly different phases for each of the condensed phases. It should be noted that the useof the word “condensed” as opposed to “solid” is deliberate. The reason is that although the layers arerigid they are mostly mesophases, in that they possess short range translational order and long rangeorientational order. In other words, they present characteristics of both solid and liquid layers in differentdegrees. A true solid 2D crystal (i.e. possesing long range translational order) can only be obtainedbelow 2 °C for C18OH. The description of the characteristics of each phase is outside of the scope ofthis thesis and the interested reader is referred to the relevant literature [144].The phase transitions can also be seen through the use of microscopy techniques. Brewster anglemicroscopy is capable of distinguishing areas with different amphiphile surface density. In this techniquep-polarized light illuminates the interface from above at a certain angle at which total refraction andno reflection occurs. The value of this angle depends on the refractive indices of the gas and liquidmedia. The presence of a layer in the interface with a different refractive index breaks the total refractioncondition and the reflected light can be detected from the denser domains [145].Fluorescence microscopy has also been routinely employed to probe structure of monolayers atthe air | solution interface [124, 146, 96]. The inclusion of a fluorophore can be achieved by addinga low proportion (typically ~1 mole %) of a labeled molecule or through the use of fluorophore-taggedmembrane components (e.g. proteins) [96]. Contrast is usually obtained due to the ability of the labeledmolecules to partition among the different phases of the layer. While simple intensity measurements arethe most common method, more advanced spectroscopic techniques using fluorescence resonanceenergy transfer (FRET) and fluorescence lifetime in conjunction with microscopy have also been usedto characterize Langmuir films [147, 148, 149, 146].5.2 Transfer onto a Solid SubstrateThe transfer or deposition of the floating Langmuir layer onto solid substrates creates layers supportedby physisorption. This weak interaction between adsorbate and substrate consists entirely of Van derWaals forces. The energy of these interactions is on the order of 1 kJ/mol and usually does not involvea defined chemical bond with the substrate [34].625.3. Characterization of Deposited Langmuir Layers50 µmFigure 5.2: Fluorescence images of a DPPC monolayer containing 2 mol % of the fluorescently labeledphospholipid 1-acyl-2-[N-(7-nitro-2,1,3-benzoxadiazol- 4-yl)amino caproyl] phosphatidylcholine floatingon a water subphase (a) and deposited on an alkylated coverslip (b). The fluorescent probe does notpartition into the condensed phase domains which appear dark compared to the fluid surrounding phase.Reproduced with permission from [151]; copyright 1984 National Academy of Sciences.Literature reports a variety of effects of the deposition procedure on the characteristics of the layer[150]. Fluorescence comparison of these supported layers with their floating counterparts has shownthat in some cases, the transfer results in supported layers that are closely correlated to the floating layer(Fig. 5.2) [151]. Nonetheless, distortion of the shape of the different domains has been observed whenusing the Langmuir-Blodgett technique[152]. Furthermore, evidence of substrate-mediated condensa-tion during deposition of phospholipid layers has been observed through AFM, with the presence ofsmall condensed regions at surface pressures where only liquid phases are stable [153, 154, 152, 155].Further proof of this condensation has come with X-ray diffraction data that shows the same condensedphase for heneicosanoic acid layers deposited from different phases [156]. Finally, it has also been ob-served that the vertical emmersion required for the Langmuir-Blodgett technique can create instabilitiesin the meniscus creating stripes of different composition [157, 154].5.3 Characterization of Deposited Langmuir LayersSPM is a popular tool for the characterization of the deposited layers. However, as explained before, inorder to obtain meaningful results using AFM or STM the perturbation of the layer by the SPM tip mustbe negligible. This is not always true, for example, on a metal surface at potentials close to values wheredisplacement of the lipid from the electrode surface by water is possible. Optical microscopy methodssuch as fluorescence imaging, though lower in resolution, can still play an important role since the layer635.3. Characterization of Deposited Langmuir Layersremains unaltered upon measuring. Nonetheless, it is possible that the inclusion of the dye modifiesthe structure of the layer. This influence, however, would likely be observed on both the floating andadsorbed states, thus still allowing a study of the deposition process itself.In-situ fluorescencemicroscopy has been employed previously [158] to study the adsorption and des-orption of layers of water-insoluble surfactants onto and from electrode surfaces. Octadecanol was firstemployed due to the extensive study of its physisorption on gold electrodes. In this case the surfactantwas not fluorescent, and a small amount (3%) of lipophilic fluorophore was added on the assumptionthat it would represent the behavior of the layer itself. Several potential-induced phase transitions ofthe layer were evidenced by pseudocapacitance peaks in AC voltammetry measurements. An increasein fluorescence, however, was only observed in the transition occurring at the most negative potential.This indicates that the transitions at higher potentials do not involve a significant increase in the metal-fluorophore separation, i.e. the layer is not desorbed. The images presented some slight heterogeneity,interpreted as due the presence of aggregates or collections of them. Interestingly, even at potentials atwhich the layer is desorbed, the aggregates did not change position for extended periods of time (morethan two hours) [159, 97]. While this surfactant is known to form condensed layers at room temperatureit was not expected for the structure to remain stable once desorbed. It was clear that although not incontact with the electrode, an adhesion force prevented the surfactant from diffusing into solution. Evenwhen using a related unsaturated adsorbate (cis-9-octadecen-1-ol) in the hope of obtaining more fluidlayers, the structure remained stable [160].Similar experiments have been performed using the lipid 1,2-dioleoyl-sn-glycero-3-phosphocholine(DOPC) adsorbed on a mercury drop. This doubly unsaturated lipid is known to form liquid Langmuirlayers. Upon the start of the desorption, the fluorescence features became freely mobile. Interestingly,once the layer was fully desorbed the movement ceased [97]. Interpretation of such data was difficultdue to the complications arising from a mobile metallic surface and the effects of the potential in thismovement.All the mentioned observations, together with scattered light measurements, formed the basis of adesorption model (Fig. 5.3). In it, the desorption proceeds through an intermediate porous state wherethe layer remains close to the electrode. At very negative potentials, true desorption exists and thepresence of aggregates is proposed. The readsorption occurs through a different mechanism, explainingthe observed hysteresis both in capacitance and fluorescence. Partial fusion of the micellar aggregatesis followed by a spreading to reform the low-capacitance layer. This model is in agreement with previousneutron reflectivity work in other water-insoluble films physisorbed onto Au. Those studies show that on645.4. ScopeFigure 5.3: Desorption model proposed for 1-octadecanol in Au. Schematics of the state of the mono-layer are overlaid on the experimentally determined capacitance. Reprinted from [97] Copyright (2004),with permission from Elsevier.average, when desorbed, the water content of the layers increases and, in some cases, the desorbedlayer remains close to the electrode separated by a ~10 Å layer of solvent [161, 162].5.4 ScopeIn this chapter, a study into the relationship between inhomogeneities observed in the floating layerand the resulting physisorbed layer will be presented. The adsorbed layers consist mainly of octade-canol and include the fluorescent lipophillic tag BODIPY-HPC which is a BODIPY labeled derivative ofan alkylphosphocholine (Section 4.1.3). The emission properties of the BODIPY monomer and dimershould depend on the state of the surrounding monolayer which makes this probe particularly useful forreporting on the structure or nature of the monolayer in which it is embedded. This characteristic haspreviously been exploited in analysis of a BODIPY thiol derivative self assembled monolayer (SAM) inwhich the existence of aggregates was demonstrated [163]. Electrochemical methods and in-situ fluo-rescence microscopy is used here to characterize the different organizational forms in which lipid-likemolecules exist in the floating monolayer or multilayers and the subsequent changes they undergo whentransferred to an electrode and during potential dependent desorption. The floating layers were preparedto exhibit strong heterogeneity and are therefore ill-defined mono or multilayers which were transferredonto a gold electrode. The contrast provided by the created heterogeneity will enable characterization655.5. Experimentalof the deposited layer which will be desorbed using negative charge. The structure and nature of theoctadecanol layer will be compared to the floating layer from which it was formed and its characteristicsas a function of the electrode potential will be discussed.5.5 Experimental5.5.1 Floating or Langmuir monolayerA 1 mg/mL solution of 1-octadecanol (C18OH) in chloroform was prepared to which a 1 mol % of zwitteri-onic tail labeled phospholipid BODIPY-HPC (Fig. 4.1) was added. This was injected onto the surface ofa deaerated electrolyte solution of 50 mM KClO4 prepared with ultrapure water. The cell was constantlypurged with Ar which enabled the chloroform to evaporate in less than 5 minutes.Fluorescence images of the Langmuir layer were acquired by equipping the inverted epifluorescencemicroscope with two fluorescence cubes, one for monomer emission (450-490 nm excitation filter, 505nm dichroic mirror, 515-550 emission filter) and one for dimer emission (same excitation, 570 nm dichroicfilter, 613-695 nm emission filter). A limitation of the current imaging setup, is that fluorescence fromthe D dimer can only be observed via a FRET process with a nearby excited state monomer sincethe excitation wavelengths used will not be absorbed by the D dimer (λAbsm = 577 nm) [164]. Thefluorescence images were recorded with a SPOT RT camera with variable exposure times. Sequentialimages were taken in two second intervals.Each image was flat-field corrected by dividing each image by the blurred (Gaussian radius = 100 pix-els) average of images taken with the bare Au electrode from which the dark current has been subtracted(scheme in Appendix B). The resulting images represent the ratio of total detected light compared to theexcitation light that leaked through the filters (F/FL). A smoothing filter was then applied to the imagesto reduce the noise, improving the image quality especially in the case of the deposited layers. This filterreplaces the value of each pixel by the mean of the pixels contained in a 3×3 square centered aroundit. The monomer and dimer channels were brought into registry with the ImageJ plugin Align3_TP [165]and color coded for easier visualization (green for monomer emission and red for dimer emission).5.5.2 Deposition and Analysis of the Adsorbed Layer on Au{111}A gold {111} single crystal electrode was flamed annealed with a butane torch and rinsed with ultrapurewater. This electrode was then put in contact with floating monolayer and pulled upwards to create a665.6. Results and discussionhanging meniscus. This served as the working electrode in a three electrode electrochemical cell alongwith a saturated calomel (SCE) reference electrode and a Pt coil counter electrode. Cyclic voltammetry(CV) ( = 20 mV/s) and differential capacitance ( = 5 mV/s, ƒ = 200 Hz, assuming a RC circuit) mea-surements were performed in the potential range -0.10 < E < 0.15 V vs SCE. The frequency was chosento be large enough so as to allow for the lock-in amplifier signal to stabilize within the exposure time ofthe imaging experiments. These measurements served to characterize the adsorbed layer. Spectro-electrochemical experiments were performed as described in Section 4.4. The Eb was kept constantat +0.1 V while Es was varied from +0.15 V to -0.8 V in 50 mV decrements; the waiting time betweenpotential steps was made to be 1 second longer than the greatest exposure time. One fluorescenceimage was taken while the potential was held at each Eb and Es value. Exposure times for monomerand dimer fluorescence were 4 and 7.5 s when using the 10× objective, and 5 s and 9 s when using the50× objective, respectively. Dimer and monomer fluorescence images had to be acquired in separateimaging cycles, dimer first, monomer second and saved separately as image stacks.Image analysis consisted of subtracting the least intense image from a sequence from all the imagesin that particular stack in order to eliminate the contribution from the background which consisted mainlyof reflected illumination light leaking through the filters. The resulting images were flat-field correctedas described in section 5.5.1. These images, labeled ΔF/FL, represent the change in fluorescenceexpressed as a fraction of the leaked light. The corrected images were treated ten consecutive timeswith the smoothing filter also described in section 5.5.1. The fluorescence images are displayed withan intensity to color mapping that maximizes the image contrast (by creating a black that is not zero)without manipulating the linearity of the relationship as shown in the intensity bar at the bottom of allimage series. The colored contrast enhanced and unenhanced versions of select potential step - imageseries (specified in the Results section) are given in Appendix C.5.6 Results and discussionThe characterization of the floating layer is presented first. We found that an increase in the apparentsurface concentration resulted in quite different phase behavior. The lower surface concentration usedwas determined to have a theoretical mean molecular area (MMA) of 5 Å2 if all molecules remainedon the aqueous surface. This is significantly smaller than the ~18-19.5 Å2 fracture MMA reported foroctadecanol monolayers (Fig. 5.1) [144, 132]. It is known from past work [166, 167, 97] that a substantialamount of material is lost to the glass walls and glass tubing that are part of the electrochemical cell.675.6. Results and discussionIf just a small excess is added compared to the amount of amphiphile required to form a monolayer,no crystal would be observed. In the present setup approximately a quarter of the material is lost tothe walls, and a substantial amount remains as a crystal floating on the surface. This is required inthe employed method since the film pressure is equal to the equilibrium spreading pressure. It must benoted that the ratio of the two amphiphiles may not be the same in the deposition solution and in thefloating layer close to the remaining crystal. Therefore, care was taken to conduct the experiments inareas of the layer far (millimeters away) from the crystal. The experiments with a larger apparent surfaceconcentration have 2.5× the amount of material on the surface. The change in the phase behavior dueto the increased material on the surface was important as it provides contrast for the comparison ofthe floating and adsorbed layers. The monolayer adsorption of octadecanol onto Au{111} has beeninvestigated in the past with little evidence of strong phase segregation, presenting a relatively uniformfluorescence [158, 159]. In this study, a small fraction of lipophillic dye that does not mix well withoctadecanol was used to induced phase segregation resulting in a strongly heterogeneous organic layerwhich was ideal for the present goal. Therefore, the BODIPY-HPC dye used is not simply just a probeof the octadecanol layer as it influences its layer characteristics both in the floating and adsorbed states.5.6.1 Low Surface ConcentrationLangmuir or Floating MonolayerThe Langmuir layer was characterized by measuring fluorescence images due to both monomer anddimer emission. In many cases these pair of images cannot be overlaid due to the movement of thelayer during the time needed to take the two images. Figure 5.4 is a collection of dimer and monomerfluorescence images taken of the floating organic layer. A variety of structures were observed thatcould be caused by a collapsed monolayer or the existence of multilayers. The fluorescence imagesin Figure 5.4a and b clearly show striations that are characteristic of a collapsed film [137, 168]. Inthese images and in others presented in this work, three distinct fluorescent regions can be seen: re-gions showing mainly dimer fluorescence (bright red) with little monomer fluorescence; regions showingmainly monomer fluorescence (bright green) with little red fluorescence; and regions of both monomerand dimer fluorescence at intermediate intensities (co-localized red and green fluorescence). Regionswhere neither green nor red fluorescence were observed also exist, believed to be composed of pureoctadecanol excluding any fluorophore, though the non-fluorescent dimer (DI) may exist in this regionor any other and will contribute to the decrease in monomer fluorescence .685.6. Results and discussionRegions of highest dimer emission intensity are correlated with the regions of lower monomer in-tensities and regions of high monomer intensity are correlated with regions of low dimer fluorescence.It is important to remember that excited state BODIPY monomers are required in order to FRET withthe dimer resulting in the observed red fluorescence. The excited state monomer may fluoresce ratherthan FRET if it is not close enough to a BODIPY D dimer which can explain the presence of greenfluorescence observed in red fluorescent regions. The green or monomer regions indicate that theBODIPY-labeled lipid molecules were not close enough to create dimers which then resulted in onlygreen emission. The red or dimer regions have sufficient BODIPY labeled lipid in close proximity sothat dimers can form from some of the monomers. The creation of dimers would be facilitated by afluid-like layer where diffusion would result in an increased probability of dimer formation, while their ab-sence could be explained by the existence of a more rigid solid-like region which prevents the formationof dimers by casual interaction. For the sake of simplicity this region is referred to as solid-like in thismanuscript, just as a description of its mobility characteristics and not with respect to its molecular or-der. This interpretation is further emphasized when examining the morphology of the floating layer usinghigher magnification (50×). Figure 5.4c and d show all three fluorescent regions. The solid-like regionsare characterized by intense monomer fluorescence as described before. The red dimer-rich regions, oncloser inspection reveal circular regions with reduced dimer fluorescence that are either not-fluorescentor composed of a mixture of monomer and dimer. These round regions appear to be phase segregatedfrom their surroundings, they appear to have a solid-like nature, and no change in their shape couldbe determined with time. These small circular regions are clearly observed to randomly move aboutwithin the red regions (see videos 5.1 and 5.2 in the accompanying disc) within the time scale of themeasurement. Such movement displays a range of velocities: not moving near the border and travelingat 0.8 µm/sec for those immersed in the dimer rich region. A detailed analysis of the velocity dispersionis not included since it does not add into the discussion of the effects of the substrate on the structureof the deposited layers.In some cases (Fig. 5.4d) the monomer images do not show high content of monomer inside thecircular regions. Instead dark round regions are surrounded by a green border suggesting that certainexclusion of the fluorophore can occur within this regions. However, some fluorophore might diffuse infrom the red region, becoming diluted by the octadecanol, forming monomer and emitting green fluores-cence. The size of these regions range from 10 to 50 µm in radius in the images shown. Also observedwere larger and smaller circular regions not shown in this selection of images. This distribution of featuresize does not change significantly (e.g. by Ostwald ripening) within the time needed for measurement695.6. Results and discussion0 3000 3040 10012 2735 1756 205 1502 8.5a b c dFigure 5.4: Collection of fluorescence images (top: dimer and bottom: monomer) of the floating layerof octadecanol with 1mol% of BODIPY-HPC. The intensity scale is useful for order of magnitude com-parisons only and shows the linear mapping of intensity onto gray scale level. The scale bar represents50 µm for all images. Images (a) and (b) were acquired with the 10× objective and (c) and (d) with the50× objective. Reprinted from [170] Copyright (2010), with permission from Elsevier.(on the order of a few minutes). While ripening has been reported for mixed composition Langmuirlayers [169], this process is usually much slower (days) that the time window employed here. In somecases, small red features were observed surrounded by large monomer-rich regions (Fig. 5.4c). Thesered features do not move within the green areas suggesting that the monomer-rich regions are solid (thiscan be seen in video 5.3 provided in accompanying disc).In summary, the floating layer displays a variety of regions with different fluorescence characteristics.The existence of these different regions may be caused by segregation of the dye and 1-octadecanol,creating domains with different relative amounts of dye and thus changing their monomer:dimer ratio andconsequently their fluorescence characteristics. In this context, dark regions observed in both monomerand dimer fluorescence images would be regions made up of only 1-octadecanol. This segregationeffect has been observed for immiscible mixture of lipids [171] for mole fractions greater than the 1 mol% employed in this work. A second possibility is the coexistence of different 1-octadecanol phases. Inthis case, the addition of the dye may favor the co-existence of liquid expanded and condensed phaseswith the dye partitioning between these phases. Such behavior has been previously exploited to studyphase transition in floating monolayers [146] but has not been observed in previous experiments on1-octadecanol layers using other dyes [160, 158]. Whether a binary amphiphile mixture will be miscibleor segregate ultimately depends on the sign of the Gibbs free energy of mixing (ΔGmix). This quantity is705.6. Results and discussioncomposed by the ideal (ΔGdmix) and the excess (ΔGemix) free energy of mixing defined, for a particularvalue of , as [172]ΔGmix = ΔGdmix+ΔGemix (5.2)ΔGdmix = RgT [χ1 ln(χ1)+ χ2 ln(χ2)] (5.3)ΔGemix =ˆ 0[A12− (A1χ1+A2χ2)] d (5.4)where χ and A are the mole fraction and molar area of the pure component ; A12 is the area of themixed film. As can be seen, the ideal term represents the entropic term and is always negative. Thus,the excess term, enthalpic in nature, becomes the determining factor in the miscibility of the layer com-ponents. In particular, the relative value of the molar area of the mixed film compared with the areaof the ideally mixed layer reflects the repulsion or attraction interactions of the layer components. Al-though the relevant values of surface pressures and molecular areas can be obtained from compressionexperiments, they were not performed due to the high amount of dye-labeled lipid required.Another possible explanation for the observed fluorescence behavior could be a collapse of themonolayer due to the high concentration of C18OH and BODIPY-HPC creating regions with multilayersthat would present higher fluidity and dye concentration, thus favoring dimer formation. Examples of thissituation have been observed in collapsed lipid layers coupled with poly-electrolyte anions [173]. Thevariety of regions observed for the floating layer could be thus explained by either one or a combinationof these possible explanations and would depend on the way the layer was formed.The fluorescence images of the floating layer suffer from photobleaching which was observed for allfluorescent regions and characterized by differing kinetics (Fig. 5.5a and b). A clear indication of the rateof decay is given in the fluorescence images for the dimer and monomer (Fig. 5.5 c and e, and d and frespectively) by comparing the region within the outline to the region outside. The outlined region wasinitially exposed for 80 s; then the whole field of view was illuminated and the fluorescence image taken(the fluorescence image presented was recorded after a delay of 24 s). The dimer fluorescence imageclearly shows severe photobleaching of the exposed region, while no obvious evidence of intensity dif-ference is observed for the monomer emission. A bi-exponential decay in the green fluorescence wasobserved for all regions due to photobleaching or FRET (with either the D or D dimer). These decaysin the monomer fluorescence were strongly dependent on the nature of the region analyzed: slow de-cay for predominantly monomer containing regions, while a faster decay for dimer or mixed compositionregions. The decay in red or dimer fluorescence observed for regions of mixed monomer and dimer715.6. Results and discussioncomposition can be fit to a bi-exponential decay as well. This can be explained by photobleaching ofthe monomer (reducing the amount of FRET) and/or by dimer decomposition. For dimer rich regions,an initial increase in fluorescence was observed before decaying. This increase can be understood byrealizing that in these regions, most of the dye is present as dimers and dimer fluorescence can beonly observed via FRET with the monomer (at the excitation wavelength used). Initially the limiting fac-tor is the amount of monomer available to act as donor. If we assume that the energy transferred viaFRET transforms dimers into monomers either by separating the dimer into monomers (destabilization)or through destruction of one of the BODIPY molecules in the dimer (photodecomposition), the amountof monomer increases yielding higher dimer fluorescence intensities. However the decreasing dimerconcentration eventually becomes the limiting factor. After this time, the fluorescence intensity wouldbe determined by the available amount of dimer and thus decrease exponentially. This dynamic photo-bleaching is problematic for the quantitative comparison of dimer and monomer fluorescence intensity,thereby limiting our analysis of the dynamics of the floating layer.The photobleaching observed for the floating layer will also be important in the analysis of the layeradsorbed onto the gold electrode, but only when the layer is desorbed or separated from the metalsurface since when adsorbed, the lifetime of the excited state is significantly shortened compared to thefloating layer. While the definition of photobleaching implies that the layer has undergone a chemicaltransformation, most of the characteristics of the layer are given by the lipid structure of the probe (Fig.4.1). A change in the chemical structure on the fluorescent moiety is not expected to strongly modify itssurface activity.Adsorption onto Au(111)Langmuir-Schaefer deposition of this heterogeneous layer onto the Au(111) electrode was characterizedby electrochemical measurements and in-situ fluorescence microscopy. In these experiments, the float-ing layer was imaged while the electrode surface was placed in contact with the floating layer, enablinga correlation between the structures in the floating layer to those that are revealed once the adsorbedlayer is desorbed at the negative potentials. Figure 5.6 compares the floating layer just before depositionon to the electrode surface at 0.00 V/SCE and the fluorescence images taken at two potentials: initialstages of desorption at -0.60 V/SCE and at the desorption potential (-0.80 V/SCE) acquired during thepotential step - imaging procedure. The spacial resolution is poor, but a low magnification was requiredfor a larger field of view to ensure finding the regions that correspond between the floating layer and thedesorbed fluorescence images.725.6. Results and discussion0. Int. (norm)MixedDimer-richMonomer-richDimer-rich + mixed0. Int. (norm)0 2 0 4 0 6 0 8 0t / sabcdefFigure 5.5: Fluorescence decays for dimer (a) and monomer (b) emission from a floating monolayercomposed of C18OH and BODIPY-HPC in a 99:1 mole ratio. Vertical lines represent the standard devi-ation for three replicate experiments. Fluorescence images for dimer (c and e) and monomer (d and f)emission. The outlined region was first exposed to light for 80 sec (left column); then the whole regionwas exposed for 24 seconds more, resulting in the images shown in the right column. The scale bar is50 µm. Reprinted from [170] Copyright (2010), with permission from Elsevier.Matching features are evident when comparing the images of the floating layer (Fig 5.6a) and the layertransferred to the electrode (Fig 5.6b & c) and one obvious example is outlined in the images. This featureis not strongly distorted but becomes larger upon adsorption onto the electrode (scale bar is 50 µm for allimages). A sampling across the surface of the electrode yields a variety of expansion ratios ranging from1.3 to 2× the size measured for the floating layer. Other features present in the floating layer are alsocorrelated with the adsorbed layer (similarly increased in scale), but show clear distortions in their shape.Scratches on the electrode surface are also evident and may contribute to the observed distortion. Thisexpansion may indicate that the floating layer is composed of multilayers while adsorption onto the goldelectrode forces the layer to deposit in a monolayer or as a thinner multilayer. In addition, the striationsdo not remain oriented straight but appear to be mis-shapen or deformed when adsorbed onto thegold surface. The fluorescence images of the floating and deposited layers are significantly different inthe overall fluorescence intensity, the adsorbed layers being more than two orders of magnitude lessintense due to metal mediated quenching. The floating and adsorbed layers also show similar regionsof fluorescence: monomer rich (green), dimer rich (red) and a region of mixed composition containingboth monomer and dimer (lower intensity green and red). The potential dependence of the desorptionof these regions differs as can be seen in comparing the monomer and dimer images at -0.60 and -0.80V. The dimer fluorescence is quite weak at -0.60 V while the monomer fluorescence from the areas that735.6. Results and discussionare close to the red regions is observed. At -0.80 V/SCE, the dimer fluorescence is clearly correlatedwith the dimer regions in the floating layer, while the monomer fluorescence appears to come mainlyfrom regions associated with the red fluorescent regions, but not from regions where only monomerfluorescence was observed. This is an example of the difficulty in quantifying the green fluorescencedue to both monomer photobleaching and conversion of dimer to monomer upon illumination.solid | liquidgas | liquidFigure 5.6: Comparison of fluorescence imagesof the (a) floating layer before deposition onto theAu{111} surface and the fluorescence images of thetransferred layer at two potentials (b) -0.60V/SCEand (c) -0.80V/SCE. The scale bars are 50 µm forall images and clearly reflect the fact that the float-ing layer expands upon adsorption. Reprinted from[170] Copyright (2010), with permission from Else-vier.Closer examination of the adsorbed layer andthe potential dependent fluorescence was doneusing a 50×objective. At this magnification, theshift in the electrode surface after deposition wasmuch larger than the field of view, and thereforea direct correlation with the floating layer beforedeposition was not possible. An indirect ben-efit of this shift is that the adsorbed layer hasyet to be exposed to excitation light and shouldbe less affected by photobleaching or photo-decomposition. Nevertheless, the same generalregions observed for the floating layer can beidentified on the fluorescence images taken dur-ing desorption shown in Figure 5.7. The regionof interest (ROI) at the top left is weakly fluo-rescent and represents the octadecanol rich re-gion, the solid-like border is the monomer rich re-gion (ROI3) which surrounds a mixed monomer-dimer region (ROI2) that contains a dimer rich oval(ROI1). These fluorescence regions are outlinedfor convenience. Examination of the potential dependence of the fluorescence intensity for the three dif-ferent fluorescing regions in the image was performed. Figure 5.7d shows the capacitance values of theinterface at the Es potential during the potential steps in the negative direction. It can be seen that at E =+0.15 V/SCE the electrode capacitance has decreased below the value measured in the absence of theorganic layer. The minimum capacitance (5 µF cm-2) is lower than typically observed for an octadecanolcoated Au {111} surface reported previously [97] which indicates deposition of more than a monolayeron average. The capacitance remains relatively constant until a potential of -0.10 V/SCE where an in-745.6. Results and discussioncrease to a plateau at 10 - 13 µF cm-2 is observed, indicating a potential driven phase change and thepossible creation of holes in the adsorbed layer [97, 160].This layer is stable until a potential of -0.55 V/SCE. At these more negative values the organic layeris displaced from the surface by water as made evident by the increase in the capacitance value to avalue similar to that of the bare Au electrode (dotted line). Capacitance measurements represent theaverage behavior of the organic layer on the electrode. The fluorescence intensity measured in theregions characteristic of monomer and dimer fluorescence can be also compared as a function of thepotential. The same relationship between the dimer rich and monomer rich regions observed for thefloating layer is also true for the transferred layer desorbed from the electrode surface. The potentialdependence of the fluorescence is shown in Figure 5.7b and c. Before interpreting these changes, it isimportant to point out that the quenching efficiency of the gold surfacemay have a strong dependence onthe emission wavelength [174, 158]. It appears that the green fluorescence is less efficiently quenchedand therefore does not need to be as far from the gold surface before fluorescence can be detected ascompared to the red emission. That said, we anticipate that the difference in the phase (red - fluid; green- solid) will also be important for determining the potential dependence (see [160] for an example).The first region to display an increase in the fluorescence as the potential is scanned negatively isROI2, the mixed region. Both green and red fluorescence were observed to increase at -0.40 V/SCEas was the red to green ratio. Both dimer and monomer in ROI2 start to desorb from the electrodeat this potential. The dimer rich region (ROI1) closely follows the response seen for ROI2: a similarincrease in fluorescence intensity and in the red/green ratio with potential. This indicates that theseregions are similar in terms of their fluorescence-potential characteristics, but show a slight differencein the potential where fluorescence is first observed, which may be due to the difference in the phaseor molecular organization of the adsorbed layer. This difference is also evident by noticing the dark ovalin the red and green images persists until more negative potential than the surrounding mixed regionagain suggesting a difference in characteristics of the adsorbed layer. Fluorescence from the monomerrich region ROI3 also increases at this potential and as expected the red/green ratio is much lower. Thered to green ratio is biased due to the sequential nature of the data collection since the green imageshave experienced photobleaching during the collection of the dimer fluorescence. A comparison can bestill useful as we assume that the degree of photobleaching was consistent throughout the image as themonomer was shown to photobleach slowly.Also clear is the decrease in the red/green ratio at the most negative potentials which can be un-derstood through the photobleaching process observed for the floating layer. As the desorbed layer is755.6. Results and discussion0102030(C/A) / PF cm-2Bare AuAu | organic layer0.' dimerROI 3ROI 1ROI' monomer-0.8 -0.6 -0.4 -0.2 0.0 0.2E / V vs SCE0.00.40.8log ('F/F L dim / 'mon)-0.8 -0.6 -0.4 -0.2 0.0 0.2E / V vs SCEF/F LF/F LF/F Lbcdea0 1.68 0.11 0.60Figure 5.7: Dimer and monomer fluorescence images recorded as potential is stepped to -0.80 V/SCE(a). The top portion of the images corresponds to the dimer fluorescence while the bottom portion is thecorresponding monomer fluorescence. The outlines are a guide to the regions analyzed. The intensityof the dimer and monomer fluorescence for three regions of interest (ROI) as a function of potential isgiven in (b) and (c) respectively; the ratio of red to green fluorescence is shown in (e) for E ≤ −0.3 V(ROIs are shown in the inset). At more positive potentials the ratio is dominated by noise due to the lowintensities. The capacitance as a function of potential is given in (d) (200 Hz, 5 mV rms). The scalebar is 50 µm. Unenhanced and contrast enhanced fluorescence images are available in Appendix C.Reprinted from [170] Copyright (2010), with permission from Elsevier.765.6. Results and discussionexposed to excitation light, the dimer becomes destabilized and monomer is created, decreasing thedimer fluorescence and increasing the monomer fluorescence. This characteristic is observed for allregions of interest, the largest change for the dimer rich region.The layer adsorbed onto the gold electrode from the floating layer seems to display the same hetero-geneity observed for the floating layer. The combination of fluorescence intensity and electrochemicalperturbation can extract some characteristics for the layer desorbed from the electrode surface. Un-fortunately, at this surface concentration, we were unsuccessful in capturing the most interesting het-erogeneous regions observed on the floating layer - the dimer fluid region containing the round greenor dark features. Increasing the surface concentration of the floating layer increased the probability ofobserving these types of features.5.6.2 High surface concentrationFloating monolayerMore material was added to the electrolyte surface so as to increase the portion of the surface coveredby the dimer-rich red region which contained the phase separated circular mixed composition features.The amount of C18OH + BODIPY-HPC added to the electrolyte surface was 2.5× the amount used forthe low concentration (sec. 5.6.1). This is much more than a monolayer worth of lipid even consideringthe substantial losses which suggests that the floating layer is composed of multilayers of material. Moredimer-rich fluid regions were observed which also supports the notion of floating multilayers. An exampleof the fluorescence observed from the floating layer before deposition onto the electrode surface is givenin Figure 5.8. The zoomed in regions are contrast enhanced to show the layer structure and are notlinear representations of intensity. Dimer-rich fluid regions containing green dots which are composedof both monomer and dimer are observed. In the example presented, the fluid region is bounded by twomonomer rich solid regions. The bottom right green region contains dimer rich dots, similar to what wasobserved previously. The floating layer is quite heterogeneous and a significant fraction of the floatinglayer was composed of these types of features. The green dots in the red matrix do not change size andmove around within the fluid red region as expected. These dots are on average 10 µm in diameter andare round (circularity of > 0.9; a perfect circle is 1). The circular shape is a result of the minimization ofthe line tension between the two phases. The anti-correlation of the red and green regions is clear inthe expanded and enhanced image of the dimer rich region. This is quite similar to the region shown inFigure 5.4c.775.6. Results and discussionFigure 5.8: Fluorescence of a floating monolayer composed of C18OH and BODIPY-HPC in a 99:1 moleratio. Left and right panels correspond to dimer and monomer images respectively. The zoomed regionsare contrast enhanced to reveal the structures of interest. The images were background subtracted(rolling ball of 1000 pixels, unsharp mask of 20 pixel radius weighting of 0.6). The scale bar is 50 µm.Reprinted from [170] Copyright (2010), with permission from Elsevier.Desorbed monolayerThe heterogeneous floating layer was deposited onto the Au{111} electrode surface. A correlation be-tween the floating and adsorbed layer was again not possible due to the shift in the electrode positionout of the field of view. Therefore the adsorbed layer has not yet been exposed to the excitation light.The fluorescence images recorded are shown in Figure 5.9 as a function of potential. The dimer andmonomer fluorescence intensities are shown in Figure 5.10a and b respectively. The capacitance isshown in Figure 5.10c and the red to green fluorescence intensity ratio is given in Figure 5.10d for theregions of interest (ROI) shown in the inset images. The dimer fluorescence - potential profile was mea-sured first immediately followed by the monomer fluorescence - potential series. The capacitance of thecoated electrode displayed the same minimum as observed for the layer deposited from the low surfaceconcentration (Fig. 5.7d). The increase in capacitance occurs at less negative potentials which we spec-ulate is due to complicated phase changes occurring in the adsorbed multilayer. The layer is assumedto be desorbed at the most negative potential, but capacitance alone is not especially helpful. Somedifferences in the capacitance are noted, but the detailed study of the potential dependent changes inthe adsorbed film relies upon the in-situ fluorescence measurements. The monomer fluorescence setof images show a monomer rich region on the left side (labeled ROI3) and a markedly low fluorescencefeatureless region on the right. The dimer fluorescence images reveal more interesting features, likelydue to the increase in the number of molecules introduced to the electrolyte surface, increasing theprobability of dimer formation. The monomer-rich region was dark in the dimer fluorescence image as785.6. Results and discussionFigure 5.9: A series of fluorescence images of a desorbed layer composed of C18OH and BODIPY-HPC in a 99:1 mole ratio. The top image is the dimer and bottom is the monomer fluorescence andthe potential is indicated in the image. The scale bar is 50 µm. Unenhanced and contrast enhancedfluorescence images are available in Appendix C. Reprinted from [170] Copyright (2010), with permissionfrom Elsevier.expected. The border region, composed of a mixture of monomer and dimer, appears fluorescent atboth emission wavelengths. The dimer-rich region clearly comprised most of the image as shown bythe dimer fluorescence image at -0.70 V/SCE. The general features of the desorbed layer are in corre-spondence with what was described previously. Unique to this set of images, we observe small dimerrich features (one example is outlined as ROI2) that exist in the dimer region but become visible at lessnegative potentials than the general dimer features, but potentials similar to the mixed region. Thesesmall features, highlighted in the contrast enhanced image shown (inset to Fig 5.10d) are reminiscentof the small mixed composition features observed in the floating layer (Fig 5.8) but are distorted formingellipsoids upon adsorption similar to the distortion observed in Figure 5.6.The potential dependent changes in the dimer and monomer fluorescence are shown in Fig. 5.10aand b for three ROIs. The green fluorescence is generally the same for all regions except that themonomer-rich region (ROI3) is more intense. Importantly, at -0.40V/SCE the difference between ROI3and the other ROIs is more prevalent. For potentials positive of -0.30 V/SCE, the monomer fluorescenceintensities are not representative of the adsorbed layer and are characteristic of the background due tothe fluorescence from regions not in the field of view and from the small amount of dye that becamedissolved in the electrolyte. This background undergoes photobleaching which results in a decreasingintensity contribution during the potential imaging measurements becoming a small fraction of the signalat potentials negative of -0.40 V/SCE. A similar influence of the background is observed in Fig. 5.7d forpotentials above -0.20 V/SCE. The magnitude of this background and photobleaching time are variablefrom one experiment to another. An example of the raw data measured for the images shown in Fig.795.7. Conclusions5.10 is given in Appendix C. The monomer fluorescence does reach a maximum at the most negativedesorption potential which can be explained by photobleaching of the desorbed molecules.The dimer fluorescence changes significantly with potential and is different for the three regions.The dimer fluorescence in the dimer-rich region (ROI1) increases to a maximum at -0.60 V/SCE andthen decreases due to the photodecomposition of the dimer. The monomer-rich regions (ROI3) alsoshow some dimer fluorescence, expected for this high surface concentration, which increases to a max-imum intensity (at -0.80 V/SCE) similar to that of the photodecomposed dimer-rich region. The smalldimer-rich regions (represented by ROI2) show a combined behavior. These regions become noticeableat the same potential as the monomer-rich region, with a clear increase above the dimer-rich fluores-cence around -0.40 V/SCE as seen in the monomer fluorescence for ROI3. Their dimer fluorescenceincreases with the dimer-rich surroundings, eventually decaying in a similar manner. The red/green ratioalso shows a similar effect where the small regions respond to potential as the monomer rich regions,then follow the dimer rich fluorescence potential response which may be due to the influence from theoverall increase in red intensity. These red dots appear to behave as regions of a mixed compositionwithin a dimer-rich phase, similar to the features observed for the floating layer before deposition (Fig5.8). Adsorption of this heterogeneous floating layer onto the electrode surface enhances the creationof dimers from these regions of mixed composition. Some of the dye molecules in monomer form seemto aggregate creating dimers induced by the electrode presence but still maintaining the potential de-pendent characteristics of their original phase.5.7 ConclusionsA heterogeneous floating (Langmuir) layer purposely created from a 99:1 mole mixture of octadecanoland BODIPY-HPC was analyzed. This floating layer presents different regions as observed by fluores-cence microscopy. The monomer and the DII dimer emission of the BODIPY moiety revealed a varietyof structures including phase separated fluid regions (dimer rich) and solid regions (monomer rich).The boundary between these phases were composed of a mixture of monomer and dimer resulting ina mixed fluorescence response. A significant monomer photobleaching and dimer destabilization orphotodecomposition due to FRET was observed for the floating layer making quantitative analysis achallenge.These floating layers were deposited onto a Au(111) electrode surface and characterized using ca-pacitance measurements along with in-situ fluorescence microscopy. A direct comparison between the805.7. Conclusions1.00102030(C/A) / PF cm-2Bare AuAu | organic layer0.'F/FL dimerROI 3ROI 1ROI monomer-0.8 -0.6 -0.4 -0.2 0.0 0.2E / V vs SCE-0.8 -0.6 -0.4 -0.2 0.0 0.2E / V vs SCE( ) / ()'F/FL'F/FL dim'F/FL monabcdFigure 5.10: Dimer(a) and monomer(b) fluorescence intensity (ΔF/FL) variations with the applied po-tential in the three different regions of interest (ROI) depicted in the inset. c) Capacitance measurementsof the bare (dotted line) and modified (solid line and open circles) Au electrode during the application ofa step-base potential perturbation in the negative direction (200Hz, 5mV rms). d) Red to green intensityratio as a function of potential for the regions of interest shown in the inset. Adapted from [170] Copyright(2010), with permission from Elsevier.floating layer just before deposition and the resulting deposited layer (imaged at desorption potentials)was possible. Although the general structure of the layer remained the same, an expansion of 1.6×(S.D. = 0.35) and distortion in the fluorescent features was observed. This was explained by the factthat the floating layer contained more than a monolayer of material or was multilayer in nature which al-lowed it to spread over the solid surface. A similar variety of structures or phases were observed to existon the electrode which were found to display different potential dependent changes in fluorescence.The monomer and dimer fluorescence intensity and their ratio were used to characterize the variousadsorbed phases. An example is presented that illustrates the influence of the solid substrate on thecharacteristics of the adsorbed layer. A region in the floating layer that appeared to be mixed in compo-sition was transformed into a dimer rich region that had potential dependent characteristics of both typesof regions. This method is useful for probing the heterogeneity of organic layers adsorbed on electrodesurfaces and their relationship with the original floating layers. The molecular phase characteristics ofthe adsorbed organic layer can be probed using fluorescent probes that form dimers and can participatein FRET but are complicated by the influence of the solid substrate and the photobleaching that occurs.81Chapter 6The Fate of Reductively Desorbed SelfAssembled MonolayersChemisorbed layers offer a more stable option compared to their Langmuir counterparts. In order toremove these films from the electrode surface, the adsorbate undergoes a chemical reaction modifyingits structure. In this chapter, the fate of those molecules once in solution is explored.6.1 Chemisorption and Self Assembled MonolayersThe formation of covalent bonds between some substrate and adsorbates enables the fabrication ofmore stable deposited monolayers. Favored by substrate-adsorbate interaction energies of ~40-100kJ mol-1 [175, 176] these layers form spontaneously and thus are termed self-assembled monolayers(SAMs). Two typical examples are organosilanes on oxidized Si surfaces (including glass) and thiolson precious metals [177]. This latter case has been attracting significant attention since it is an idealsystem for electrochemical control, because the potential of the metallic substrate can be governed atwill.The most common substrate is undoubtedly gold, due to its low reactivity towards most functionalgroups. This means that the tail groups can include a wide variety of functionalities without competingfor the substrate against the thiols. It has been reported that well ordered SAMs can be formed usingthiols, dithiols and sulfides as starting materials. However, the great majority of works deal with thiolsdue to the shorter time required to deposit the SAM.Before starting a deposition the substrate has to be pretreated. Usually the Au has to be cleanedfrom impurities, even if they are electroinactive, since their presence can block the deposition of theSAM [14]. This step is critical for the quality of the deposited layer. Many different variations of theseexist but most follow the same process:1. When working with bulk electrodes (commercially available disk electrodes, for example) polishing826.1. Chemisorption and Self Assembled Monolayerswith alumina [178, 179] or diamond slurries [179] is common practice. This step aims to removelarge contaminants and to expose a new layer of Au atoms. It is usually omitted in thin film elec-trodes.2. Oxidative treatment of the surface; usually performed chemically or electrochemically. Examplesinclude stepping or sweeping to oxidative potentials in acidic [180, 178] or neutral medium [181],submerging the electrode in piranha solution [182] or treatment in a plasma or ozone chamber.The chemically harsh conditions break down organic contaminants, but also creates a thin layerof gold oxide and roughening of the surface.3. Reductive treatment. Can also be performed electrochemically by stepping or sweeping the poten-tial to reductive values [181, 180], or submerging the electrode in ethanol. The reducing conditionstransformed the Au oxide created in the previous step back to metallic Au.Typically the layers are prepared using the open circuit potential (OCP) method. In this approach, aprecleaned substrate is immersed into a solution (liquid or gas) of the thiol adsorbate. It has beenproposed that the thiols attach to the metal spontaneously, according to the surface reaction [176]RSH+A(0)→ ASR+12H2 (6.1)Recent reports, however, point out that hydrogen evolution might not always occur, especially on smoothdefect-free surfaces [183? ]. In reaction 6.1, gold is oxidized while hydrogen is reduced. The fact thatthis reaction involves electron transfer prompted attempts to control the coverage by manipulating thepotential of the substrate. Figure 6.1 presents cyclic voltammograms for a mica supported gold electrodein an ethanolic solution of dodecanethiolate in the presence of KOH [184]. Under these conditions, thefollowing potential dependent equilibrium exists at the surface of the electrodeRS− +A(0)Š ASR+ e− (6.2)Upon continuous cycling (Fig. 6.1a, dotted line) an oxidation peak at ~-0.9 V vs Ag | AgCl | saturatedKCl is observed. This current corresponds to the forward (adsorption) process of Eq. 6.2. On thenegative going (cathodic) scan, the reverse (desorption) reaction appears as a cathodic peak at ~-1.025 V. Using this system, Porter et. al. proved that by carefully selecting the potential value theycould produce incomplete layers [184]. They claim that the equilibrium is attained fast enough so thatcoverage is controlled thermodynamically rather than kinetically. The degree of coverage was estimated836.1. Chemisorption and Self Assembled Monolayersby integrating the charge under the reductive desorption curves (described in more detail in the followingsection). Their results suggested that for hexanethiolate an applied potential of -0.60 V vs Ag | AgCl |saturated KCl for >1 min and <3 min was enough to produce a monolayer with the same coverage as theones produced using the OCP method. Thus, they concluded this was probably a complete monolayer.In contrast to these findings, and using the thiol rather than the thiolate, Ma and Lennox suggestedthat more positive potentials (~+0.2 V vs Ag|AgCl) were required to create a SAM of good enoughquality to block electron transfer from K3Fe(CN)6 in solution [185]. The discrepancy might arise fromthe difference in quantification method for determining a “complete” monolayer. Notably, the monolayersthey obtained at positive potential values blocked the K3Fe(CN)6 even better than the ones obtainedby the open circuit method, suggesting that by holding the potential at values where the reaction isforced to completion, less defects are formed. In similar experiments using thiolated DNA, holding thepotential at +0.3 V vs Ag | AgCl, layers of similar DNA density as the ones produced at OCP methodwere prepared. These layers however, required much less time to achieve said coverage [182]. Lennoxet. al. also pointed out an advantage of controlled potential deposition over the open circuit potentialmethod in conditions where a monolayer of mixed composition is desired. By using such relatively highpotential values kinetics become more important than thermodynamics, thus the composition of theresulting SAM will more closely reflect the composition of the deposition solution. Interestingly, Wanget al. reported observing an improvement on the deposition rate at potentials much more negative thanthe reduction of the dodecanethiol [186]. While the mechanism was not clear, they speculated thatremoval of some adventitious contamination was the main reason behind the accelerated deposition.This seemed plausible especially since another group tried to replicate their results, unsuccessfully[180].Self assembled monolayers have been used to immobilize molecules with a great variety of function-alities, by substituting the last carbon in the chain (the ω position) with the desired functional group. Inorder to reduce the number of surface immobilized impurities which are hard or impossible to remove,only three types of ω substitutions are commonly employed: hydroxyl, amino and carboxyl. Thesemoieties can then undergo quantitative reactions for further modifications. A good review on the partic-ularities of carrying out reactions on a SAM can be found in [187].846.2. Reductive Desorption50 µAab-1.2 -1.0 -0.8 -0.6CurrentPotential (V vs Ag | AgCl | saturated KCl) Figure 6.1: Cyclic voltammograms of a gold electrode in a 0.5 M ethanolic KOH solution in the presence(a) and absence (b) of dodecanethiolate. Dotted line in (a) corresponds to continuous cycling whereasthe solid line corresponds to the first cycle of a SAM prepared by the OCP method with immersion timeof at least 2 h in 1 mM dodecanethiolate. =100 mV/s. The arrow indicates the initial scan direction.Adapted with permission from [184]. Copyright 1992 American Chemical Society.6.2 Reductive DesorptionIn a similar way as explained in the previous chapter (for physisorbed layers), applying very positive ornegative potentials results in desorption of the adsorbed monolayer. However, in this situation a faradaiccurrent flows according to the reactions [188, 189]ASR+ e− → A(0)+RS− (6.3)ASR+ 2H2O→ A(0)+RSO−2 + 3e− + 4H+ (6.4)ASCH2R+ 14OH− → A(0)+RCO−2 + SO2−4 + 11e− + 8H2O (6.5)Of particular interest is the reductive desorption process (Eq 6.3) since it is the most reversibleone,with the sulfur remaining in the -2 oxidation state as opposed to oxidizing it to +2 or +6. Uponapplication of reductive potentials to the metallic substrate, the sulfur-metal bond is reduced therebyeffectively removing the SAM from the metal surface [190, 191, 192, 193], creating thiolate molecules.Equation 6.3 can be rewritten more formally to emphasize that the solvent and the surface have an effectin this displacement equilibrium [194]:856.2. Reductive DesorptionRS(srf,M(hk)) + H2O(q) + ne−(M(hk)) ŠRS−(q) + H2O(srf,M(hk)) (6.6)where “surf ” refers to the species at the metallic surface with a (hkl) orientation and “aq” refers to thespecies in solution. This process has been studied for short and long alkyl-thiols [195, 9, 196], aromaticthiols [194, 197, 198], and thiolated DNA [194], among many other compounds. Through integrationof the charge under the reduction portion of the cyclic voltammogram, this process has been routinelyemployed to quantify the coverage of a SAM [85]. Furthermore, reductive desorption has been usedto regenerate the clean metal surface after using the SAM for sensing [199] or controlled nanoparti-cle growth [200] and for the controlled release of biomolecules, nanoparticles [201]and adhered cells[192]. Although most often the reductive potential is applied to the electrode by means of a potentiostat,reductive conditions can also be achieved by adding a reducing agent to the solution [178].The stability of a SAM is dependent on the balance between metal-adsorbate, adsorbate-adsorbateand solvent-adsorbate interactions [194, 202]. An example can be observed in the difference in the ca-thodic peak potential of the two curves in Fig. 6.1a. The solid curve results from the desorption of a layerthat has been deposited for several hours [184, 189]. The dotted line, on the other hand, correspondsto incomplete layers deposited during the potential cycling (~8 s) [189]. The shift in the potential canbe explained by the poor adsorbate-adsorbate interactions in the latter case. Less well packed alkylchains translate in less stabilization of the layer and thus less energy (less negative potential) is neces-sary to desorb it. The potential required to remove the layer also depends on parameters such as theadsorption site (terrace vs step edge) [203, 204], the metal surface crystallinity [195] and the alkyl chainlength [205]. A careful selection of reductive potential then enables the selective removal of either onlyone component of a multi-component segregated layer [206] or of all the thiols from a particular surfacefeature [207, 1]. Subsequent backfilling with another thiol enables the creation of surfaces otherwiseinaccessible by traditional self assembly.Although agreement exists in literature regarding the nature of the self assembled monolayer reduc-tive desorption [193, 7, 189, 85], the fate of the desorbed molecules once in solution is only speculativelydescribed. If after reductive desorption, the potential is switched to more positive values, reductivelydesorbed thiolates can oxidize the Au at the working electrode to re-form a SAM layer. The quality orcoverage of this reformed monolayer has been used as an indication of the amount of thiolates thatremain close to the electrode after reductive desorption. Cyclic voltammetry experiments by Morin etal. have shown that the amount of this monolayer reformation strongly depends on the solubility of the866.2. Reductive Desorptiondesorbed thiols in the surrounding medium [195]. Once desorbed, short alkyl thiols (e.g 1-butanethiol)redeposited only about 20% of the original SAM, while the rest is lost by diffusion into the solution (Fig.6.2a). On the other hand, values of 70% [208] and 90% [209] have been reported for the reformation of1-hexadecanethiol SAMs. Furthermore, some authors claim negligible loss of thiol when using alkylthi-ols with 13, 17 and 18 carbon atoms, studied using chronoamperometry rather than cyclic voltammetry[210, 211]. For SAMs created with 1-nonanethiol, the electrolyte pH was found to have a significant in-fluence on the extent of re-formation of the SAM as higher pH favored the presence of soluble thiolateswhile acidic solutions favor the insoluble protonated form.Using infrared spectroscopy techniques, Morin’s group proposed a model to explain the presence oftwo cathodic peaks and two anodic peaks on the cyclic voltammograms of a hexadecanethiol SAM onAu{111} [86]. Their model (Fig. 6.2b top) suggests a two step mechanism for the desorption of insolublethiolates: first, the reduction of the Au-S bond creates a physisorbed layer of thiolates, similar to theoriginal SAM. Then at more negative potential, the solvent displaces these thiolates forming physisorbedmicellar type species with a large fraction of the electrode surface covered with solvent [86]. This secondstep is similar to the mechanism previously discussed for the desorption of a physisorbed octadecanollayer from Au{111} (Fig. 5.3) [212].Some evidence of the presence of the above mentioned (micellar) aggregates during the reduc-tion of thiol SAMs has been obtained by in-situ scanning tunneling microscopy (STM) imaging of thedesorption process of SAMs composed of 1-propanethiol, 1-hexanethiol, 1-hexadecanethiol as well as3-mercaptopropionic acid [208, 203]. However, the identity of the observed aggregates is still controver-sial, since they could also be gold adatom islands formed after the SAM desorption or even Au islandswith adsorbed thiol aggregates [214]. Cai and Baldelli proposed an alternate view of the structure of thedesorbed layers employing electrochemical impedance measurements and sum frequency generationspectroscopy [213]. They suggest that while short (C10 and C12) alkylthiols diffuse into solution oncedesorbed, hexadecanethiol and octadecanethiol monolayers retain its two dimensional order even atreductive potentials (Fig. 6.2b bottom). Furthermore, they justify the decrease in reoxidation signal bysuggesting the formation of disulfide bonds. However, their measurements were performed ex-situ orin-situ at open circuit potential, preventing the conclusive use of this data to infer the state of the thiolsat potentials where they are reduced.SAMs created from thiol compounds are dynamic, as evidenced by the potential independent trans-formations of the structure of the SAM switching between the well knownp3×p3R30° and the c(4×2)superlattice in a matter of seconds, as well as the continuous creation and disappearance of defects876.2. Reductive DesorptionS S S S S S S SS S S S S S S SS S S S S S S SS S S S S S S SS S S S S S S SS S S S S S S Sab+ eAuAuAuAuAuAuSSSSSSSS SSSSSSSSSSSSSS SS SSSSSSSSSSSSSSSSSS+ e+ e+ eFigure 6.2: Proposed mechanisms for reductive desorption of alkylthiol SAMs. a) Short chain alkylthio-lates diffuse into solution. b) The presence of longer chains prevents fast diffusion enabling the formationof other structures. Top mechanism, proposed by Byloos et. al. involves a physisorbed layer intermedi-ate and the appearance of micellar aggregates at more negative potentials [86]. Alternatively, Cai andBaldelli (bottom) propose the presence of a physisorbed layer even in the reduced state. They alsopropose the presence of S-S bonds [213].886.3. Macroelectrodes vs Microelectrodes[203, 215]. Initially, the desorption process was proposed to involve a nucleation and growth of thesedefects [203, 216, 217], but more recent analysis suggest a better fit to a model in which domains ofthiol shrink from the edges [196].As seen above, most often the study of the desorbed molecules rely upon indirect evidence of theirbehavior as they produce current during desorption or re-adsorption. Only a few studies have probed thenature of the desorbed molecules. These primarily used in-situ techniques which provide an averagepicture of the molecular state. In other words, with the exception of STM, the employed techniques donot offer spatial resolution. Furthermore, while STM imaging is useful in short scales (typically tens ofmicrons), it cannot characterize the long range (hundreds of microns) fate of the desorbed moleculesonce reductively released from the electrode surface. In contrast, in-situ fluorescence microscopy offersgreat opportunities since, once far away from the electrode (tens of nanometers), fluorophore labeledthiolates are strongly emitting enabling to track their movement in the desired length scales.6.3 Macroelectrodes vs MicroelectrodesThe desorption of SAMs has been performed previously in our laboratory. A fluorescently labeled alkylthiol was used to form a SAM that would be fluorescent upon desorption. Despite some readsorptionupon a sweep back to potentials where the layer is stable, most of the thiol was lost through diffusionto solution. With this technique it was for the first time possible to directly observe the effect of thecrystallinity on the desorption potential [1]. Both reductive and oxidative desorption were investigatedand compared. While reductive desorption occurred in a narrow potential range, its oxidative counterpartwas present at a large potential range and close to oxidation potentials for gold. Furthermore, significantroughening of the electrode was observed after the latter procedure [163].These experiments, performed on larger round beads revealed significant movement in the reduc-tively desorbed thiolate molecules. Desorbed alkylthiol SAMs appear as a diffuse "cloud" (Fig. 6.3a)[1, 163] whereas very localized aggregates were observed to skate across the electrode surface forSAMs created from thiolated DNA (Fig. 6.3b) [218]. The origin of this movement has been speculated toinvolve the shape of the electrode surface (i.e. a bead or a planar polished surface) and electrophoreticattraction towards the counter electrode. Bead experiments have shown several limitations that precludeaccurate analysis of the movement of the desorbed molecules:• Multicrystalline electrodes have many coexisting grains with a variety of surface crystallographies,each one possessing its characteristic desorption potential. Monitoring the movement of desorbed896.4. ExperimentalaFigure 6.3: Sequence of in-situ fluorescence images during the reductive desorption of SAMs frommacroelectrodes. a) Selective desorption of the fluorescently labeled alkylthiol BODIPY-C10-SH fromgold surfaces with {111} character. b) Mixed monolayer composed of thiolated fluorescently labeledDNA and mercaptohexanol. The colored traces indicate the movement path of fluorescent aggregates,one of which is followed in the yellow circle. The width of the frames for (a) and (b) are 1 mm and 320 µmrespectively. Adapted with permission from [1] and [218]. Copyright 2004 and 2009, American ChemicalSociety.thiols is complicated due to the interference from the desorbed molecules from neighboring grainsthat may desorb at slightly different potential.• Due to the curvature of the electrode, it is not clear if the directionality of the observed movementis somehow influenced by the gravity (e.g. are the molecules moving left or up?)Microelectrodes (µEs) offer the advantage of being small enough so that the complete electrodesurface can fit in the field of view of an optical microscope, thus creating a more localized source offluorophore molecules. Furthermore, due to their polycrystalline nature, these electrodes do not possesscrystalline domains large enough to be optically resolved. As a result the fluorescence appears evenacross the surface of the electrode. Similar spectroelectrochemical experiments have been reportedby Ghaly et. al. using microband electrodes [219]. They employed different alkylthiol chain lengthsand found that while the behavior of short chains (6 C atoms) could be modeled by simple diffusion,longer chains (11 C) deviated significantly. They speculated that this was primarily due to differences insolubility, but it will be shown that other factors should also be considered.906.4. Experimentalto groundworking electrodePTFE tapenail polishborosilicateglass tubecoaxial cableuxless solderdeposited particles (not to scale)Figure 6.4: Diagram of the employed microelectrodes. Low background fluorescence was attained byusing borosilicate glass, fluxless solder, nail polish and white PTFE tape. Reprinted with permissionfrom ([220]). Copyright (2013) American Chemical Society.916.4. Experimental100 µm 350 µm250 µm b)a) c)Figure 6.5: Optical microscopy images of three microelectrodes. a) Perspective image composed of16 images taken at different focal positions stitched together with the ImageJ plugin “Extended depthof field” [221, 222] b) and c) Transverse images of the tip of electrodes with RG values of 10 and 36respectively. Adapted with permission from ([220]). Copyright (2013) American Chemical Society.6.4 Experimental6.4.1 Microelectrode Fabrication and CharacterizationMicroelectrodes were fabricated by sealing a 25 µm diameter gold wire inside a borosilicate pipette(Sutter) using a Sutter P-87 pipette puller equipped with a Pt/Ir heating filament. While applying vacuum,the filament was heated enough to soften the glass but not to melt the Au. Once a good seal was attainedbetween the gold and the glass (as inspected with amicroscope) the construct was snapped in themiddleforming two relatively symmetrical electrodes. An electrical connection was achieved by melting a smallamount of flux-less solder (Kester) to the exposed core of a coaxial wire (Precision Instruments) andsoldering to the gold wire using a heat gun. The use of fluxless solder was essential to achieve a lowfluorescence background. A mono audio connector was used to connect both the core and the shield ofthe coaxial wire to the working electrode (WE) and ground connections respectively (Fig. 6.4) to reducenoise.The tip of the electrode was manually polished at an angle close to normal to the electrode axis (Fig.6.5a) with a Narishige EG-40 microgrinder. To reduce the risk of contamination, the microlelectrodeswere not encased in any other material (e.g. epoxy) during polishing. Slight changes in orientation ofthe electrode during the different polishing cycles created the presence of facets with small differencesin angle. Because of the angle at which the electrodes were polished and the small size of the tip somechipping was sometimes observed at the glass edge. The ratio of the outside diameter to Au diameter(RG value) ranges from 10 to 36 (Fig. 6.5b and c).926.4. Experimental100 µmSEM Pt Ir Sib)a)c)Figure 6.6: The surface of a glass pipette after been heated by the Pt/Ir filament. a) Photography; blackdeposits can be seen near the tip. b) Scanning electron microscopy image c) Energy-dispersive X-ray spectroscopy mapping analysis. Adapted with permission from ([220]). Copyright (2013) AmericanChemical Society.Brightfield imaging shows that upon pulling, the sides of the pipette became darkened in an area nearthe tip (Fig 6.6a). Closer examination reveals the presence of particles deposited on the electrode glasssheath. Scanning electron microscopy (SEM) images were obtained with a Hitachi S-2300 electronmicroscope equipped with an Advanced Analysis Technologies energy dispersive X-ray spectrometer(EDX) after a very thin film (less than 10 nm) of Au was vapor deposited on the electrode construct toavoid charging of the glass (6.6b). Furthermore, EDX mapping analysis (6.6c) revealed the presenceof Pt and Ir particles on the glass surface suggesting that the origin of these deposits is the heatingfilament in the pipette puller.6.4.2 Electrochemical and in-situ Fluorescence MethodsThe spectroelectrochemical measurements were performed in the glass cell shown in Fig. 6.7. Forsome experiments, as noted in the text and shown in said figure, a small CE was employed, which wasfabricated in the same way as the WE but with a Au wire protruding approximately 150 µm from theglass surface. The electrolyte was 1 mM KOH. All potentials reported are with respect to Ag|AgCl andwere not corrected for the R drop in solution. According to Newman [223], the resistance of an inlaiddisk can be calculated byR=12κEØ(6.7)936.4. ExperimentalCCD cameraEXFO lampmicroma-nipulatorWECERE250 µm thick windowFigure 6.7: Experimental setup for the study of reductive desorption of SAMs from microelectrodes.The small counter electrode is shown. Adapted from [74] and [115] with permission from Olympus andBioanalytical Systems, Inc.where κE is the electrical conductivity and Ø is the electrode diameter. For a 25 µm diameter disk inthe employed electrolyte R=737 kΩwhich could only create significant R drops at the most negativepotentials where currents larger than 100 nA are produced.The working microelectrode was positioned from the top of the cell using a polytetrafluoroethylene(PTFE) 14/23 joint. The tilt angle was measured optically. A Narishige ONW-131 micromanipulatorwhich allows movements as small as 1 µm was employed to control the position of the counter electrodein the case where the small CE was used.The optimized deposition procedure is as follows. Before every desorption, the microelectrode wascleaned by submerging it into piranha solution (hydrogen peroxide and sulfuric acid) for 5 minutes, fol-lowed by potential cycling in 0.5 M H2SO4(between -0.4 and 1.4 V / Ag|AgCl). The electrode was946.4. Experimentalthen rinsed in ultrapure water (Milli-Q, Millipore), and then HPLC grade methanol (Fisher Scientific).The surface was then modified by introducing the electrode tip into an approximately 1 mM solution ofBODIPY-C10SH, in a 1:1 mixture of CHCl3and MeOH for 10 minutes. The electrode was then suc-cessively rinsed with a mixture of HPLC grade chloroform and methanol, methanol and finally ultrapurewater.The potential perturbation program consisted of a series of potential steps starting at -0.4 V sequen-tially decreasing in 25 mV steps to a given final potential (Eƒ ). The time between images was either180 ± 8 ms or 253±3 ms (for exposures times of 25 and 100 ms) which results in an effective sweeprate of 140 and 100 mV/s respectively. One image was acquired at every potential step. The imageswere analyzed with the ImageJ software [121]. All the images corresponding to a single desorptionexperiment were collected into an image stack. Correction for drift in the position of the electrode wasperformed using the Image Stabilizer plugin for ImageJ [224]. A Gaussian blur filter (radius = 2 px) wasapplied; the minimum values for each pixel in a stack (minimum projection) was used to create an imagewhich was subtracted from all the other images. In this way, a new stack representing the difference influorescence intensity with respect to the minimum was created. An anisotropic diffusion filter (a1 = 0.5,a2 = 0.9, 20 iterations) [225] and another Gaussian blur filter were applied to decrease the noise in theimage. Selected desorption sequences experiments (as noted in the figure captions) are included asvideos in the accompanying disc.6.4.3 Partition Coefficients Measurement / EstimationThe octanol / water partition coefficient of the BODIPY labeled alkylthiol was measured by fluorescencespectroscopy. Ten microliters of a solution in n-octanol were placed in contact with 800 µL of waterand stirred for 10 s in a Vortex mixer. After letting the mixture rest for 10 min the aqueous phase wasremoved and diluted 1:10 in methanol. The original solution in n-octanol was diluted with 1:10 water/ methanol mixture until a similar fluorescence signal was acquired. The partition coefficients for thehomologous series of alkylthiols containing 1 to 20 carbon atoms was predicted using the ChemBio-Draw software (CLogP). It was found that this fluorescently labeled alkylthiol presents an octanol/waterpartition coefficient comparable to 1-heptanethiol and can be considered moderately soluble in water.956.5. Results6.5 Results6.5.1 Fluorescence Imaging of Reductive Desorption from a Au µelectrodeThe Au microelectrode surface was modified with BODIPY-C10SH and placed in the spectroelectro-chemical cell, so that it’s surface was parallel to the focal plane of the inverted microscope. The changesin fluorescence and capacitance were measured as the potential was stepped to more negative poten-tials. The data presented is from layers that displayed capacitance values commensurate with highcoverage SAMs which resulted in significant fluorescence signals. Not all layers prepared were of suffi-cient quality for data analysis, but the trends observed are consistent.Figure 6.8a shows the typical changes in the electrode capacitance upon the application of a poten-tial perturbation that progressively becomes more negative (Fig. 6.8c). Initially the capacitance of theelectrode was 0.10 nF as compared to 0.26 nF for the bare uncoated Au (filled circles) indicating thepresence of the low dielectric SAM on the electrode surface. It is worth mentioning that the expectedcapacitance for a bare Au electrode with this exposed geometric area (4.9×10−6 cm2) is calculated tobe 0.09 nF. This discrepancy can be partially explained via surface roughness. However, another veryimportant factor is the stray capacitance created by the Au wire inside the glass immersed in the elec-trolyte. This becomes important for microelectrodes due to the small ratio of Au surface area exposedto the solution to the total Au/glass surface in the electrolyte solution. It is known that the capacitance ofa SAM covered surface is at least one order of magnitude smaller than the bare Au surface, suggestingthat the capacitance measured for the thiol coated surface is dominated by the stray capacitance. Asthe electrode potential becomes more negative (Fig. 6.8c), both the capacitance (Fig. 6.8a) and thefluorescence intensity (Fig. 6.8b) from the microelectrode increase as a consequence of the reductivedesorption of the BODIPY-C10SH SAM, indicating its removal. Once desorption is complete, the ca-pacitance is similar to that for the uncoated Au surface. It is important to note that at potentials morenegative than -1.4 V, the differential capacitance measured does not represent the double layer capaci-tance values because of the current due to H2evolution caused by reduction of water. The capacitanceis calculated assuming the electrochemical system can be represented as a series RC circuit which isnot accurate once faradaic currents are present. In spite of this artifact, the comparison between theSAM covered and uncoated surfaces clearly shows the loss of the layer at potential values between-1.05 V and -1.4 V.Coincident with the capacitance increase is an increase in fluorescence intensity, a result of theincreasing separation between the reductively desorbed fluorescent thiolate and the metal so that fluo-966.5. Results0.−0.6 −1.0 −1.4 −1.8C (nF)E (V vs. Ag|AgCl) (mean grayscale)SAM coveredBare Au−1.8−1.4−1.0−0.60 5 10 15E (V)t (s)abcFigure 6.8: Variation of the capacitance (a) and fluorescence intensity (b) upon the application of thepotential perturbation shown in (c). Adapted with permission from ([220]). Copyright (2013) AmericanChemical Society.976.5. Resultsrescence quenching is no longer efficient. At the more negative potentials, the decrease in fluorescenceresults from diffusion and the accompanying dilution of fluorophore in addition to photobleaching. Thereductive desorption potentials appear to be more negative than typically reported in literature [194, 1]possibly due to the uncompensated R drop in the 1 mM electrolyte solution used or the slow diffusion offluorescent thiolate from the electrode surface. Low electrolyte concentrations are necessary to studythe effect of the migration of the negatively charged thiolates (see next section). The R drop remainssmall (< 25mV) because of the use of microelectrodes and the small currents (<100 nA) to potentialsas negative as -1.5V. At potentials more negative than this the current flow increases and the R dropcan be more significant, though the reductive desorption process is complete. In addition, the rate atwhich the thiolates diffuse away from the metal electrode is slow compared to the changes in potentialwhich results in a lag of the fluorescence signal compared to the changes in the coverage. Separatemeasurements done in higher electrolyte concentrations (not shown) on larger electrodes with slowerpotential changes show a good correspondence with the literature reported values for desorption.Fluorescence images of the interface during the reductive desorption are shown in Figure 6.9a. It isclear that in this case the fluorescence, and therefore the thiolate molecules, become radially distributedaway from the gold electrode. Image analysis of this process was done by using a constant intensitythreshold for all images, which defines the edge of the diffusing plume of thiolates. Fitting an ellipse to thisthresholded area (shown as a white continuous line in the images) enables characterization of the shapeof the plume via the lengths of the major and minor axes. In this case, the thresholded fluorescenceimage has circular symmetry since the major and minor axes are of equal length throughout the seriesof potential steps. The initial increase in the size of the plume corresponds to an increasing number offluorophores desorbing from the electrode. As the fluorescently labeled thiolates start diffusing away,their local concentration decreases thereby causing a subsequent reduction of fluorescence intensity.The analysismethod usedwill reflect this as a decrease in the length of the axes (modeling of this processis detailed in the following section). Moreover, the center of the ellipse does not change position (within1.5 µm) from its initial location. These results suggest that hemispherical diffusion is the dominant masstransport mechanism as expected for a microelectrode under these conditions.6.5.2 Modeling of the Diffusion Effect on the Fluorescent PlumeModeling of the fluorescence plume radially diffusing from the microelectrode is based on the work byBaur et al. [228] They model the diffusion of material to/from a hemispherical microelectrode (not a disk)as a function of time and distance using a convolution method. This method relies upon the response of986.5. Results-1.20 V-1.35 V-1.50 V-1.65 V-1.80 V-1.875 V25 µm01020304050−1.4 −1.6 −1.8Axis length (Pm)E (V vs. Ag|AgCl)Major axisMinor axis0. (Pm)−1.8−1.6−1.410 12 14E (V)t (s)50 µm ab)c)d)0.80.1Figure 6.9: a) Fluorescence images at selected potentials (shown in the images) during the reductivedesorption. The dashed outlined region corresponds to the Au surface, while the continuous line repre-sents the resulting ellipse after applying a threshold (0.35). Video 6.1 in the accompanying disc showsthis desorption experiment. The length of the major and minor axes (b) as well as the displacement ofthe center of the ellipse (c) are plotted for a smaller subset of employed potentials (d). Symbols in (c)represent raw data while the dotted line results from smoothing with a Savitzky-Golay filter [226, 227].Error bars on the displacement are smaller than the symbol size. Adapted with permission from [220],Copyright (2013) American Chemical Society.996.5. Resultsthe system to an arbitrary change in conditions. The characteristic impulse response function is definedas the response of the system to an infinitely fast spike (delta function) change in the conditions anddepends on time and geometry (e.g. position away from the electrode). Usually, it is not practical toperform such an experiment and a simpler approach is used based on the fact that the derivative of theresponse to a step function is equal to the impulse response function.The concentration of the product of an electrode reaction [P] as a function of time t and radialdistance from the electrode surface r can be calculated using a convolution of the concentration at theelectrode surface [P] (r = 0, t) and the impulse response function (RF)[P] (r, t) = [P] (r = 0, t)∗ RF(r, t) (6.8)where * denotes the convolution operator, equivalent to taking the Fourier transform of the functions,multiplying them in frequency domain and taking the inverse transform [228, 229]. This has been imple-mented in MATLAB, using the rate of irreversible desorption of a SAM (Eq. 14.3.20 in the work by Bardet al [14]) during a potential sweep as a quantity proportional to the concentration at the surface of themicroelectrode[P] (r = 0, t)∝ kdmxexpRgTαFkd(6.9)where kd, mx, α and  are the heterogeneous desorption rate constant, the maximum amount of thioladsorbed at the surface, the charge transfer coefficient, and the potential sweep rate, respectively.The concentration as a function of distance and time was determined after coarsely estimating thevalues of the diffusion coefficient and the rate constant for the reaction (1× 10−5 cm2/s and 25 s-1respectively); a value of 0.5 was chosen for α. The accuracy of these parameters is not crucial becausethe objective of this model is to examine the general fluorophore distribution and not to extract values ofdisplacement or speed to compare with experimental results.The plume of thiolate is generated over a small range of potentials and the resulting plume of materialis calculated. The conversion to fluorescence intensity requires integration of the concentration of thedesorbed product over all the distance normal to the plane of the glass surrounding the electrode. It isassumed that the fluorescence intensity is proportional to the total number of molecules. This generatesfluorescence intensity values as a function of in-plane distance from the center of the electrode. Sincethis is an axially symmetric system, generating an image of the fluorescence in the electrode plane asa function of time is straightforward. These “images” (included as Video 6.2 in the accompanying disc)1006.5. Results050100t = 1.15 s t = 1.80 s t = 2.45 s t = 3.10 s t = 3.75 sFluor. intensity (arbitrary units)0100200y distance / µm0 100 200x distance / µm 0 100 200x distance / µm 0 100 200x distance / µm 0 100 200x distance / µm 0 100 200x distance / µmFigure 6.10: Simulation of an irreversible SAM desorption from a 2 µm radius hemispherical microelec-trode. Top panel shows the cross section of the fluorescence signal at different times. The correspondingregions in the electrode plane (x,y) resulting from thresholding the images to an intensity value of 20arb. units (shown as a dotted line) are shown in the bottom panel. Adapted with permission from ([220]).Copyright (2013) American Chemical Society.were then compared with the experimental images and clearly show a similar trends in the size of thefluorescent plume. The top panel in Figure 6.10 presents a cross section of the simulated fluorescencesignal plotted in arbitrary units. The initial increase in fluorescence appears as a sharp peak that laterbecomes broadened due to diffusion. If a threshold intensity value of 20 arb. units is employed, theresulting circular region (shown in the bottom panel of Fig. 6.10) first becomes larger as the peak growsbut then become smaller and finally disappear as the broadening is so large that the intensity drops belowthe threshold value. The variation of the in the size of the major and minor axes (equal to each other)of the fluorescence plume is shown in Figure 6.11. Changing the sweep rate and diffusion coefficientaffects the current - time profile and the resulting plume size (not shown).6.5.3 Influence of the counter electrode positionPrevious experiments using fluorescence imaging to monitor reductive desorption have suggested acorrelation between the direction of movement of the desorbed species with the position of the counterelectrode[218, 163]. Using a microelectrode enabled a more controlled experiment in which another Auwire of approximately 150 µm in length was used as the counter electrode. As shown in Fig. 6.12a and1016.5. Results050100Axis length / µm1.0 1.5 2.0 2.5 3.0 3.5 4.0t / sFigure 6.11: Variation of the length of the axes (both major and minor have the same length) in a simu-lation of a SAM desorption. Threshold has been set to an intensity value of 20 arb. units. Reproducedwith permission from ([220]). Copyright (2013) American Chemical Society.b the counter electrode was positioned ~30 µm away from the edge of the microelectrode (WE) creatingan asymmetric electric field enhanced by the low concentration of electrolyte (1 mM) used.Figure 6.12c shows the fluorescence images at two selected values of potential during two indepen-dent experiments. When examining the top row, the first image at -1.5 V is comparable to the imageat -1.5 V shown in Fig. 6.8d revealing an essentially symmetric fluorescence distribution centered atthe gold electrode. Closer inspection of the fluorescence images reveals an imperfect ellipse with asmall deformation directed towards the counter electrode, something not observed in Fig. 6.8d wherethe counter electrode was positioned more than 10 mm away. However, at more negative potentials,the fluorescent plume of desorbed molecules reversed direction. This was quantified by measurementof the displacement and speed of the center of the thresholded fluorescent signal (Fig. 6.13a, b and c,filled circles) which shows an initial increase in the displacement at a speed of approximately 1 µm/sdirected towards the CE (indicated as an angle between 0 and 90°), and then at a later time, the fluores-cent plume reverses direction (~-150°) and moves away from the counter electrode at a speed of about4 µm/s.The same experiment (on a newly created SAM layer) was repeated using the same orientation ofthe CE (bottom row of Fig. 6.12c). Once again at small negative values of potential, the fluorescent1026.5. ResultsWECECEWE250 µm0.80.1E = -1.8  VE = -1.5  VRun 1Run 20°90°-90°a)c)b)Figure 6.12: Desorption experiments performed with the CE in close proximity (30 µm) to the WE.Lateral view schematic (a) and bottom view micrograph (b) of the system. c) Fluorescence images attwo selected values of potential, for two different In addition experiments. The black circle representsthe edge of the gold electrode, the white line represents the outline of the CE. Dashed lines are given asreference for the angular coordinate used. Reproduced with permission from ([220]). Copyright (2013)American Chemical Society.1036.5. Resultsb)a)c)d)024681012−1.4 −1.6 −1.8Displacement (µm)E (V vs. Ag|AgCl)Run 1Run 21234567Speed (µm/s)−150−100−50050100150Angle of movement / °−1.8−1.6−1.410 12 14E (V)t (s)Figure 6.13: Displacement (a), speed (b) and angle of movement (c) during the desorption for the twoexperiments shown in Fig. 6.12c. The filled and empty symbols correspond to the top and bottomseries of images, respectively. The potential perturbation is shown in (d). Symbols represent raw datawhile lines result from smoothing with a Savitzky-Golay filter. Error bars on the displacement are smallerthan the symbol size; error bars on the speed (shown in the figure) were estimated by determining thesensitivity of the center of the ellipse on the threshold values used in image analysis. Reproduced withpermission from ([220]). Copyright (2013) American Chemical Society.1046.5. Resultsplume initially moved towards the CE. Unlike the previous case, the direction of this movement was notreversed, but became constant at ~35° at more negative potentials. These measurements clearly indi-cate that the position of the CE somewhat influences the direction of the desorbed molecule movement,but is not the major controlling factor. Moreover, these results clearly show that there is another influencethat dominates the direction of the desorbed thiolate movement.6.5.4 Tilting the microelectrode~ 3°Thiol-uorescent dye construct Microscopeobjectivetilt directiontilt directiontilt direction250 µmb)a)Figure 6.14: a) Schematic of the system employedto investigate the effect of gravity. b) Fluorescenceimages at E = -2.0 V of two independent desorptionexperiments. The direction of the arrow points to-wards the higher point of the glass sheath. A videocorresponding to the desorption experiment shownin the bottom of (b) is included as Video 6.3 in the ac-companying disc. Reproduced with permission from([220]). Copyright (2013) American Chemical Soci-ety.The movement of the desorbed thiolates wasagain investigated using the same procedure as inFig. 6.8, but with the electrode surface purposelytilted in different directions, approximately 3° withrespect to the focal plane of the microscope(same as the horizontal plane, see Fig. 6.14a).The counter electrode was positioned far away(more than 10 mm) and remained in a fixed po-sition during these experiments. In all casesthe fluorescent plume of the desorbed thiolateswas found to move quickly (when compared tothe previous measurements) towards the portionof the electrode that was situated higher, inde-pendent of the position of the counter electrode(Fig. 6.14b). Furthermore, the shape of the fluo-rescence plume was initially circular, but becameelliptical. The direction of the movement clearlyillustrates that the main contribution towards themovement of the desorbed molecules is somehow related to gravity. Possible explanations for this ob-servation can be due to the intrinsic buoyancy of the thiolate molecules or aggregates, or differences inthe density of the electrolyte near the electrode produced by e.g. resistive heating or gas evolution.Further experiments were performed using the microelectrode tilted at 8.8 degrees from level. Thepotential was stepped negatively from -0.4V to different final negative values (-1.3, -1.5 and -1.7 V) at aneffective sweep rate of 140 mV/s. Fluorescence images were also collected at the final potential. Thefluorescence images were analyzed as described previously and the speed of the center of the plume1056.5. Resultsas a function of potential are shown in Fig. 6.15c.The experiment terminating at the least negative potential (-1.3 V) showed the smallest movementof the plume of desorbed thiolates, with the center of the plume moving less than 1 µm during theexperiment. The speed of the center of the plume is small (less than 1 µm/s) and in no particulardirection. The lengths of the major and minor axes which define the edge of the plume undergo muchlarger changes . The start of the desorption of the fluorescent molecules results in a predictable increasein the fluorescence intensity and therefore the size of the plume. At longer times, the length of the axesdecrease as the molecules diffuse away into the electrolyte. Most importantly, the major and minoraxes change in the same manner which clearly shows that no distortion in the shape of the plume wasobserved; it is circular throughout the measurements and is clearly characteristic of the symmetricaldiffusion of the desorbed molecules from the microelectrode at this low potential of -1.3 V even thoughthe electrode surface was tilted. / µm/s-2.0 -1.8 -1.6 -1.4E  / V vs Ag|AgClfFigure 6.16: Variation of the terminal speed of theplume of desorbed thiolates as a function of thelimit negative potential employed. Reproduced withpermission from [220]. Copyright (2013) AmericanChemical Society.The measurements done at the two more neg-ative potentials are clearly different as the plumemoves consistently upwards (i.e. up the inclinedglass surface). The position of the center of theplume shifts with time/potential, moving approx-imately 20 µm. The speed of the center of theplume reaches values of 5 - 10 µm/s. In bothcases, the major and minor axes are similar withinthe time of the experiment, but if the potential of -1.7 V ismaintained for 10 s the plume becomes el-liptical with an aspect ratio of 1.2 (data not shown).In a separate set of experiments with tilt angle of 5.7±1.0° the same potential perturbation was appliedbut Efwas varied from -1.5 to -1.9 V vs Ag|AgCl. The terminal speed for each experiment is plotted asa function of potential in Fig. 6.16. It can be seen that as the limiting potential becomes more negativethe terminal speeds starts to plateau. Unfortunately, time limited the number of points to be acquiredin a day and daily variations in the experiments preclude the direct comparison between speed valuesobtained in different days. However, the trend is consistent with a plateau in speed.Overall for conditions where the CE is separated far from the WE, the shape of the plume can bedescribed as a sum of three distinct effects: i) hemispherical diffusion that, similar to the images in Fig.6.8d, causes a radially symmetric change in the diameter of the plume; ii) buoyancy type of influence that1066.5. Results0.050.0100.0150.0200.0250.0300.0350.0Minor / mMinor / mMinor / µm0.050.0100.0150.0200.0250.0300.0350.0Major / mMajor / µmMajor / m0246810Speed / m/sSpeed / m/sSpeed / m/sSpeed / m/sSpeed / µm/sSpeed / m/s-1.8-1.6-1.4-1.2E / V3.0 6.0 9.0t / sE / VE / Va)b)c)d)Figure 6.15: Variations in the minor (a) and major axes (b) as well as speed (c) of the center of the ellipsefitted to the fluorescent plume on three desorption experiments following the potential perturbationsshown in (d). Symbols represent raw data while lines result from smoothing using Savitzky-Golay filters.Reproduced with permission from [220]. Copyright (2013) American Chemical Society.1076.5. Resultsdisplaces the plume towards higher “elevations” (which is seen as lateral movement); and iii) desorptionwhich may occur over a range of potentials introducing newly desorbed thiolates into the plume closestto the electrode making it appear elongated.The presence of such an strong influence from buoyancy was unexpected and intriguing. Suchmovement can be driven by some volume force created by differences in density. Therefore, to betterunderstand this force, it is necessary to calculate the magnitude of volume force needed to cause theobserved speed values.6.5.5 Relationship Between Buoyant Force and Lateral SpeedSince gravity influences in the movement of the desorbed thiolates, a simple model was employed tocalculate the speed of the solution moving upwards due to differences in density. The buoyant volumeforce B isB= −g(ρb− ρ0) (6.10)where g is the acceleration of gravity, ρb the density in the buoyant region and ρ0 the density of thebulk solution. The problem was treated as a uniformly accelerated motion (UAM), being the acceleration()=Bρb(6.11)The acceleration in the plane of the electrode, θ, was then calculated considering the tilt angle betweenthe horizontal and the electrode surface θθ = sinθ (6.12)The in-plane speed, |r˙|θ, at a distance r = 80 µm from the origin of the movement can be obtained withthe standard UAM formula [230]|r˙|θ =Æ2θr (6.13)Finally, the horizontal component of this speed, |r˙|, was calculated since only movement parallel tothe focal plane can be tracked with our current microscopy setup|r˙| = |r˙|θ cosθ (6.14)1086.6. Discussion6.5.6 Finite Element Method Simulations of Joule HeatingThe observed current in the desorption experiments will unavoidably result in some ohmic (Joule) heatingof the solution near the tip of the electrode. Simulations of this temperature change were carried outusing the Comsol Multiphysics software. A 2D axially symmetric model was built with coordinates r(radial coordinate) and x (axial coordinate). An electrolyte solution box of dimensions 0≤r ≤5 mm and-5≤≤5 mm was built, x = 0 being the surface of the electrode. The potential and current distributionsacross the different materials were calculated according to the time independent form of the electricalcontinuity equation− ∇ · (κEE) = 0 (6.15)where κE is the electrical conductivity and E is the electric potential. The temperature distribution ofthe system was then calculated by solvingρCpδTδt− ∇ · (κT∇T) = RHS (6.16)where ρ is the density, Cp is the specific heat capacity of the material, κT is the thermal conductivityand RHS is the resistive heat sourceRHS= κE |∇E|2 (6.17)The values of the potentials were adjusted to match the highest current density measured in theexperiments (30 mA/cm2). It can be seen in Fig. 6.17 that the highest temperature is not on the surfaceof the electrode but 10 µm away, in agreement with the findings by Baranski and Boika regarding theeffect of Au as a heat sink [231].6.6 DiscussionMovement of the plume of reductively desorbed thiolates is directed upwards. The potential dependenceof the plume velocity can help identify the driving force behind this process. As mentioned before, at verynegative potentials, the reduction of water produces molecular hydrogen which is somewhat soluble inthe electrolyte (1.6× 10−3 g H2/kg) [232]. The change in electrolyte composition near the electrodesurface can produce differences in density due to the substitution of Ar (used for purging the electrolyte)for electrogenerated H2. This difference in density was employed to calculate the upward or buoyantforce on the H2saturated solution. This upward velocity was then used to calculate the velocity of1096.6. Discussion00100 GlassAu Electrolyte-100 100 200Radial distance / µmAxial distance / µm T-T0300 06 × 10-5 KFigure 6.17: Temperature difference (T-T0) finite element simulations corresponding to the maximumobserved current density, 30 mA/cm2. Reproduced with permission from [220]. Copyright (2013) Amer-ican Chemical Society.solution on the surface of the tilted electrode. In this analysis friction losses due to movement alongthe glass surface were not considered (details for these calculations are provided in section 6.5.5). Thedifference in density between water saturated with Ar and that saturated with H2(2.81× 10−2kg/m3)[233] is sufficient to drive advective movement at speeds of approximately 90 µm/s along the plane ofthe electrode surface. This velocity is much larger than what was measured, but it is still reasonable toconclude that the gradient in density due to the H2content of the electrolyte (clearly not saturated) nearthe electrode surface causes the upward movement of electrolyte which as a consequence influencesthe motion of the desorbed thiolates.It is also possible that small H2bubbles, which are inherently buoyant, may be responsible for theobserved movement. The first appearance of a clearly visible bubble in these images occurs at -1.9V, while the onset of the fast movement across the electrode is located at -1.7 V. Optically resolvablebubbles are approximately 1 µm in diameter. The presence of smaller bubbles is not likely as they areunstable due to the higher solubility of the gas due to the higher pressures created by the curvature ofthe bubble surface. Calculations of bubble dissolution time following the work of Duncan [234] showthat even in the case of continuous formation of H2(saturation condition) it would require less than0.01 s to completely dissolve a 1 µm diameter H2bubble. Although the presence of smaller bubbles1106.6. Discussion(nanobubbles) has been reported by the atomic force microscopy community, their stability seems to bedependent on the presence of the surface [235, 236]. While these nanobubbles may be active in themechanism of hydrogen evolution, their existence is not the cause of the observed long range (hundredsof microns) movement away from the surface.Two other alternative explanations include resistive (Joule) heating and the intrinsic buoyancy of thedesorbed molecules / aggregates. Both possibilities can be addressed considering the terminal veloc-ities observed for the -1.5 and -1.7 V measurements shown in Fig. 6.15. Significant changes in thetemperature of the solution near the electrode surface are found to occur with the passage of largeAC currents.[237, 231] However, the small DC currents produced here (less than 0.2 µA) are shown toproduce differences in temperature of less than 1 mK both through finite element method simulations(as shown in Fig. 6.17) and by using the calculation given by Baranski [231]. These calculations showthat the maximum density change is only 1.15× 10−5 kg/m3, more than 2000× smaller than in thecase of dissolved H2resulting in velocities of at least one order of magnitude smaller than observed.Furthermore, the observed movement of the plume of thiolates does not scale with the calculated dif-ference in temperature, as the velocity plateaus at potentials more negative than -1.7 V (Fig 6.16). Itis more likely that the concentration of dissolved hydrogen is close to saturation resulting in a constantvelocity, while the temperature should be able to significantly increase[231] resulting in a continuouslyincreasing plume speed which is not observed. Finally, the idea of intrinsically buoyant molecules orsmall aggregates is also not likely since changes in potential to values more negative than the reductionpotential should not influence their speed. In summary, this work suggests that a difference in densitydue to dissolved hydrogen is the source for the buoyant movement observed. Further proof of this couldbe obtained by changing the rate of H2production by varying the pH or the electrode material. However,another possibility is that the increase in the OH– concentration, produced at the same rate as H2causesthe difference in density. Further experimental evidence to differentiate between these two mechanismscould be obtained by saturating the initial solution with H2thereby allowing the concentration of or OH–to change while keeping the [H2] constant. Nevertheless, the upward movement of electrolyte contain-ing dissolved H2is further supported by an unusual set of measurements detailed below which showthe removal of the SAM from particles on the glass surface.6.6.1 Unusual Fluorescence from the µE Glass SheathElectrochemical generation of dissolved H2produced a rather unexpected result. As shown in Fig-ure 6.6, the µEs used have Pt/Ir particles on the walls of the insulating glass sheath. The microelectrode1116.6. Discussiontip is placed into the SAM deposition solution, thereby exposing and modifying both the Au surface andthe Pt/Ir metallic features on the glass wall. The reductive desorption scan (Fig. 6.18a left and center)shows the appearance of a fluorescent plume from the gold surface and at more negative potentialsa ring of fluorescence which starts from the glass walls (Video 6.4 in accompanying disc). This fluo-rescence ring is clearly due to the removal of BODIPY-C10SH SAM from the electrically isolated Pt/Irparticles resulting in an increase in fluorescence as quenching by the Pt/Ir metal surface is no longerefficient. Their removal must be due to a reductive process [238], but clearly not a directly electrochem-ically controlled one. The electrogenerated dissolved H2is postulated to be the reducing agent whichmoves upward due to its buoyancy. The dissolved H2polarizes the Pt/Ir particles negatively, reduc-ing the metal-sulphur bond and releasing the fluorophore from the surface. The presence of dissolvedH2without creation of H2bubbles is certainly enhanced with microelectrodes due to the hemisphericaldiffusion of the H2away from the surface, delaying the creation of a critical concentration needed to nu-cleate bubble formation. As shown previously (vide supra), this dissolved H2creates a locally buoyantelectrolyte region which then moves upwards reducing the thiol SAM formed on the Pt/Ir particles. Inter-estingly, the fluorescence was not observed if the glass surface was coated in gold via vapor deposition.This can be explained by monitoring the open circuit potential (OCP) for macroscopic Pt/Ir and Au wiresimmersed in electrolyte solution initially saturated with Ar and during bubbling with H2(Fig. 6.18b). TheAu electrode potential did not change significantly, but the Pt/Ir surface adopted a very negative potential(~-800 mV/Ag|AgCl) in agreement with literature values reported for Pt electrodes [239]. This differenceis a result of the higher exchange current density for Pt as compared to Au. Such potential values havebeen shown to be sufficient to at least partially remove the thiol from Pt surfaces[238].Further proof of this phenomena was achieved with the creation of a micropipette in the same fashionas themicroelectrodes, but with a hollow interior. Fluorescence imaging of themicropipette after creatingthe fluorescent SAM on the Pt/Ir particles was performed while H2was injected through the bore in thepipette. Figure 6.18a (right) shows the increase in fluorescence due to the increase in the dissolved H2content in solution and the reductive removal of the thiol from the Pt/Ir particles on the glass surface. Inaddition to detailing another method for chemically removing thiol from Pt surfaces, this also implies themovement of dissolved H2upwards, suggesting that the electrolyte saturated with H2would be morebuoyant than electrolyte saturated with Ar further supporting the mechanism proposed for the movementof the desorbed thiolates.This mechanism explains recent observations by Shepherd et al. in which they notice a difference inthe fraction of oxidatively re-adsorbed thiolate depending on the experimental electrode setup.[107] A1126.6. DiscussionAr satH2 bubblingAua)−1.0−0.8−0.6−0.4− 40 80 120 160E (V vs. Ag|AgCl)t (s)Pt/Irb)Hollow pipette with H2Au microelectrode-1.8 V250 µm-1.4 V100 µmFigure 6.18: a) Fluorescence images from a desorption experiment on a Au microelectrode (left andcenter) and from an experiment in which H2was bubbled though an empty glass pipette treated in thesame way the microelectrode was. b) Open circuit potential of two macroscopic wires of Au and Pt/Ir ina 1 mM KOH solution. Initially the solution is degassed with Ar; at the time indicated with the verticalline H2is bubbled into the solution. Reproduced with permission from [220]. Copyright (2013) AmericanChemical Society.1136.7. Conclusionsgreater fraction of readsorbed thiolates was observed if the electrode surface was in a hanging meniscusconfiguration as compared to when the electrode is submerged. The movement of the thiolates from thesubmerged surface is facilitated by the advective movement described here (in part), while this upwardmovement is not possible in the hanging meniscus configuration.6.7 ConclusionsReductive desorption of thiol SAMs from a Au microelectrode was characterized using in-situ fluores-cence microscopy revealing the fate of the desorbed thiolates once departed from the electrode surface.At moderately negative potentials, when desorbed from the electrode surface, the thiolates radially dif-fuse away from the microelectrode. For more negative potentials, the creation of soluble H2changes thedensity of the electrolyte near the electrode surface and carries the thiolate away from the electrode sur-face, upwards. The velocity of the plume of thiolate was found to depend on the amount of H2created,and therefore on the value of the potential. These observations clearly explain the extent of oxidativere-adsorption after desorption typically observed in CV experiments. The creation of dissolved H2wasalso found to be sufficiently reductive for removing thiol SAMs from a Pt or Pt/Ir surface. Understandingthe forces that determine the movement of these thiolates open the possibility of controlling their motionunder specific electrochemical conditions.114Chapter 7Characterization of Electrically"Switchable" DNA Layers forBiosensingBiosensors are a particular class of analytical devices in which a change in the value of a measurablesignal results from a specific interaction between the analyte and a biological molecule acting as therecognition element [240]. Among these, antibodies, enzymes and nucleic acids have been employeddue to their high specificity. Immobilization of these molecules on an electrode surface provides both asupport and a possible transduction mechanism. However, this integration between the electrode andbiomolecule requires extensive characterization if a reliable performance (necessary for clinical applica-tions) is expected from the sensor. This chapter describes the use of in-situ fluorescence microscopy tobetter understand the structure of DNA modified electrodes. A new method for correctly analyzing theresponse of a potential driven DNA sensor is presented.7.1 Deoxyribonucleic Acid (DNA)Deoxyribonucleic acid (DNA) is an oligomeric chain of individual units called nucleotides. Figure 7.1shows a DNA segment in which one of the nucleotides is outlined by a dotted line. Each nucleotideis formed by a nucleobase, a 2’-deoxyribose sugar and a phosphate group. The bases are dividedinto purine based (guanine and adenine) and the pyrimidine derivatives (cytosine and thymine). Ina nucleotide unit, the base and the phosphate are bound to the sugar through its 1’ and 5’ carbonatoms respectively [241]. Note that atoms in the 2’-deoxyribose are appended the “prime” symbol todifferentiate them from the atoms in the bases. When synthesized in nature, a ligase enzyme forms aphosphodiester bond between the hydroxyl group on the 3’ carbon in the growing DNA chain and the1157.1. Deoxyribonucleic Acid (DNA)ON NNOHNNNNONHHN NOONNNNNHHN NNNONHHNNNOHN NOONNNNNHHOOPOOOOOPOOOOOPOOOOOOPOOOOPOOOOOPOOOOOPOOOOHOOOPOONNON NNNNHHOPOOOOO5’3’Figure 7.1: Structure of deoxyribonucleic acid (DNA). Left panel shows the chemical structure of itsdifferent components. A nucleotide is shown delimited by the dotted line. A schematic of the three-dimensional double helix arrangement of these components is shown in the right. The pink spheresrepresent the phosphate-sugar backbone and the bars symbolize the bases following of the same colorcoding as in the left panel.phosphate in the 5’ carbon of the nucleotide to be added. As a result, in one of the ends of the chain theC3’ carbon is not bonded to another nucleotide while in the other end the phosphate in the C5’ carbonatom is free. Accordingly, a nucleotide chain is said to have a 3’ end and a 5’ end (Fig. 7.1 left) [242].At neutral (physiological) pH the phosphate group is deprotonated, making DNA a polyanion.Each of the nucleotides has a name derived from the corresponding nucleobase and a one letterabbreviation (listed in Table 7.1). Following this nomenclature, a chain of oligonucleotides is representedby a series of these letters usually starting from the 5’ end [242]. For example, the right hand side chainof the DNA shown the Fig. 7.1 can be represented as CTGTA...A chain of nucleotides is known as a single strand of DNA (ssDNA). In nature, DNA usually forms aduplex with two chains interacting with each other through hydrogen bonds between the bases. Due to1167.1. Deoxyribonucleic Acid (DNA)Group Nucleobase Nucleobase structure Nucleotide AbbreviationPurineadenine NNNHNNH2adenosine Aguanine NHNNHNONH2guanosine GPyrimidinecytosineNNHNH2Ocytidine CthymineNHNHOOthymidine TTable 7.1: List of names and abbreviations of the nucelobases and their corresponding nucleotides.1177.2. DNA Sensorsboth shape and hydrogen bonding properties, these interactions are base specific, so a guanine can onlybind a cytosine through three bonds while an adenine can only bind a thymine through two bonds (Fig.7.1 left) [241]. The structure for DNA was elucidated in 1953 by Watson and Crick, using Chargaff’sresults [243] among others, and remains essentially valid until today (they missed the third hydrogenbond in the G·C pair) [244]. They described a double helical model with a spacing of 3.4 Å betweenconsecutive bases and a helix rotation period of 10 bases (Fig. 7.1 right). This duplex is usually termeddouble stranded DNA (dsDNA).One of the main differences between single and double stranded DNA is its rigidity. This can becharacterized by a quantity termed persistence length (PL). In simple terms, this quantity can be definedas the distance at which a polymer ceases to behave as a rigid rod. Consequently, when the length ofthe polymer is shorter than the persistence length it can be considered rigid, while polymer chains longerthan the persistence length will be flexible. For DNA, ionic strength has a significant influence on thepersistence length, since at low ionic strengths repulsion between adjacent phosphate groups stretchesthe DNA, increasing PL [245, 246]. Nevertheless, at a given value of ionic strength there is a significantdifference in PL between single and double stranded DNA [247, 245, 246]. For example, at a ionicstrength of 0.02 mol L-1ssDNA presents a persistence length of ~2.5 nm (approximately 6 bases) [245]whereas the value for dsDNA is ~52 nm (153 base pairs) [246].Since the DNA bases have aromatic character, DNA can be quantified using UV absorption spec-trophotometry. The nucleobases absorb with a maximum at 260 nm, and absorbance at this wavelengthis commonly employed to quantify the amount of DNA. However, several common contaminants suchas phenol or protein can absorb at this wavelength; thus it is better practice to measure the absorbanceat more than one wavelength and use the absorbance ratio as an indicator of purity. Nevertheless, it isimportant to realize that only significant amounts of contaminant will change the absorption appreciablyso small amounts of contaminant may go undetected [248].7.2 DNA SensorsResearch on DNA immobilized on an electrode surface for detection purposes has been conductedsince the 1990s [249]. The relative ease of attachment of one of the ends of a DNA strand through theuse of sulfur - metal chemistry has resulted in a continuous increase in the number of publications thatuse DNA for detection [250]. The most common way to immobilize the DNA is by synthesizing a DNAmolecule with a “linker” hydrocarbon chain (typically containing 6 carbon atoms) terminating in a thiol1187.2. DNA Sensorsgroup. This molecule can then readily form self assembled monolayers on a Au surface. Interestingly,it has been reported that the choice of the DNA end to which the “linker” is attached results in differentmorphologies of the SAM as examined by atomic force microscopy [251]. Monolayers formed by linkingthe DNA through its 5’ terminus appeared very homogeneous in height while the ones linked throughthe 3’ end presented regularly spaced low height structures with high boundaries between them. It wasspeculated that the difference could be due to the way the linker interacts with the DNA chain itself.The drawback of using presynthesized linker-DNA constructs is the low yield associated with themodification reaction to include the thiol, thus increasing synthesis costs. An alternative approach, pro-posed by Zhao et al. is to first deposit an underlying SAM with hydroxyl terminations and then usea condensation reaction between the SAM hydroxyl group and the 5’-phosphate end of the DNA inthe presence of 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide hydrochloride (EDAC) to immobilize un-modified DNA [252]. This method thus reduces DNA modification costs while enabling coverages up to~85%.The common characteristic to all the systems described here is the use of DNA as the biorecognitionelement (probe). The spatial complexity of the initial DNA conformation can be used to broadly dividethe sensing platforms into three categories [253]:• One-dimensional: The DNA strand is initially “free”, forming dsDNA upon the addition of the com-plimentary target molecules (Fig.7.2a).• Two-dimensional: Portions of the probe DNA strand are self complementary creating a stablestem-loop “hairpin” structure, which is destabilized upon the interaction with target molecule (Fig.7.2b)[254, 255].• Three dimensional: This category encompasses any of the many more complex structures thatcan be formed using DNA. Two examples include the naturally occurring G-quadruplex structures[256] and purposely formed tetrahedral DNA structures that separate the probe strand it from themetal surface (Fig.7.2c) [253].Since the fundamental phenomenon required to use the DNA as a recognition element is the hy-bridization of two complementary strands, it is necessary to look at the factors that affect the efficiencyof this process. Of particular importance is the density of ssDNA strands on the sensing surface. Peter-son et. al. have shown that if the strand density is as high as 12×1012 strands/cm2, the hybridizationefficiency is reduced to ~15% compared to a value of 73% for a density of 2×1012strands/cm2 [182].This response is explained in terms of the necessary space for the complementary strand to access1197.2. DNA SensorsFigure 7.2: Comparison of DNA one (a), two (b) and three (c) dimensional DNA sensing platforms.Adapted from [253] with permission from Nature Publishing Groupthe probe strand and create a stable duplex. If the layer is too dense, neighboring DNA molecules canhinder this process.The signal transduction mechanisms also vary, going from mechanical to spectroscopic and ofcourse electrochemical. Table 7.2 lists selected examples of the different sensing techniques strate-gies using DNA as a biorecognition element. Mechanical transduction systems are generally based onthe detection of mass differences using a quartz crystal microbalance (QCM) . Although very sensi-tive, significant care should be put in effectively passivating the surface of the electrode not covered bythe DNA; otherwise non specific adsorption can become problematic, since the nature of the detectionsystem (mass) is not exclusive to the target molecule and relies on specific interactions / binding events.The substrate being an electrode, it is intuitive to think that many strategies have been devised totake advantage of the electrochemical techniques as quantification tools. Earlier approaches for elec-trochemical DNA detection were using the redox properties of the DNA itself [257, 258, 259], resultingin non-sequence specific response. Later, detection made use of small redox-active molecules that in-teract with particular structures present in double stranded DNA. The most common classes are minorgroove binders that coil around the helical structure and intercalators that insert themselves in betweentwo stacked bases [260, 261]. Sequence specificity arises from the fact that dsDNA is only created ifboth the probe strand and its complementary target are present. [262, 263, 264]Recently, the most common strategy is to covalently label the distal end of the DNA with a redoxactive molecule. Electron transfer from the redox reporter is either enhanced or decreased as a result ofa change in conformation associated to the hybridization event. Two different mechanisms for electrontransfer are possible. If the redox probe is close enough to the electrode the electrons might tunnelthrough the passivating SAM [265].1207.2.DNASensorsTechnology Detection principle DetectionlimitSequencespecificReferenceDNA electrochemistryElectrochemical reoxidation of a guanine reductionproduct<16 pM no [257]Minor groove bindersElectrochemically active molecules preferentially bind inthe minor groove of double strand DNA formed uponhybridization of probe and target strands.66.4 aM * yes [262]IntercalatorsElectrochemically active molecules preferentiallyintercalate within the stacked bases of double strand DNAformed upon hybridization of probe and target strands.4 pM yes [263]Redox labeled hairpin DNAUpon hybridization with complimentary strand the redoxprobe moves far from electrode surface diminishingelectrochemical signal10 pM yes [254, 255]Fluorophore labeled hairpinDNAUpon hybridization with complimentary strand thefluorescent probe moves far from electrode surfacediminishing the quenching by the metal10 nM * yes [266, 267]Fluorophore labeled electricallyswitchable DNADouble stranded DNA switches conformation faster andproduces higher fluorescence than ssDNA.<10 pM yes [268, 269]Table 7.2: Select examples of detection using DNA in conductive surfaces. * Not a formal detection limit, just minimum concentration used.1217.2. DNA SensorsIf on the other hand the distance between the electrode and the redox probe is large, charge “hopping”through the bases (particularly guanine) becomes the most favorable route [270]. One of the most wellknown examples using the tunneling approach is the work of Plaxco et al [254, 255, 265]. They attach amethylene blue or ferrocene molecule at the end of a “hairpin” DNA probe. In the absence of the targetmolecule the redox reporter is close to the electrode and electron transfer is efficient. Upon hybridizationwith the complementary strand the distance between the reporter molecule and the electrode increasesthereby making electron transfer less efficient and a decrease in the current is observed. It should benoted that the presence of pinhole defects on the SAM could have an effect on the observed current.The use of “hairpin” DNA has also extended to spectroscopic techniques based on the labeling ofthe DNA distal end with a fluorophore. Initially low fluorescence is observed due to the quenching by themetallic surface. The change in conformation from “hairpin” to linear when dsDNA is formed results in anincrease in separation between the fluorophore and the metal [266]. The resulting effect is a decreasein the quenching and a concomitant increase in fluorescence. A different approach, proposed by Rantet. al. combines potential control and fluorescence measurements [271, 272, 268, 273, 274, 275, 269].This technique makes use of the change in conformation associated with the electrostatic interactionsbetween the electrode and the charged DNA backbone. By labeling the DNA with a fluorophore, mod-ulation in potential translates in a modulated fluorescence signal. The details of this method will beexplained in Section 7.3.Although initially devised as amethod to detect complimentary DNA strands only, it was soon realizedthat it was possible to use DNA to detect proteins and small molecules through the use of aptamers [276].These are fragments of DNA that show large specificity for binding towards a non-nucleic acid target(e.g. small molecules and ions). The process to obtain an aptamer for a particular target molecule is ascreening method named Systematic Evolution of Ligands by Exponential Enrichment (SELEX) [277].In this method, an initial mixture of a large number of random DNA fragments is placed in contact withthe target molecules. DNA fragments that present higher affinity for this target will bind to it and can beeffectively separated from molecules with non-specific binding. Polymerase chain reaction amplificationis used to produce large amounts of the DNA with large affinity. The process is repeated ~3-5 cycles toget a highly selective aptamer.Besides similar strategies to the ones described above, aptamers offer unique possibilities them-selves. For example, if a site specific to a small molecule is placed in the middle of a double strandedchain labeled with an electroactive molecule, electron transfer through the stacked base pairs will beeither disrupted or enhanced when the small molecule binds to the aptamer site [276]. A change in the1227.3. The “switching” DNA Constructmeasured current will then be an indication of a binding event.7.3 The “switching” DNA ConstructAs mentioned in Section 7.1, DNA is negatively charged at physiological pH due to its phosphate back-bone. As a result, DNA immobilized on an electrode surface can be manipulated to some extent bychanging the charge on the electrode surface. When the charge is negative, the DNA molecules will berepelled, making them “stand up” from the surface (Fig. 7.3 left). Conversely, when a positive chargeis accumulated on the electrode, the DNA molecules will be attracted towards it and will “lay flat” on thesurface (Fig. 7.3 right).Figure 7.3: DNA conformation control through theuse of the electrode potential. When the electrode isnegatively charged the DNA is repelled from its sur-face (left) while positive charge attracts the DNA to-wards the electrode (right). Reproduced from [274]with kind permission from Springer Science andBusiness MediaThis phenomenon was first observed by Kel-ley et al. in 1998 using in situ (i.e. under po-tential control) atomic force microscopy [278]. Intheir results, they observed a height value of~55 Å for self assembled monolayers of 15-base-pair double stranded helices at potentials nega-tive of +0.45 V vs a Ag quasi-reference electrode,whereas a value of 20 Å consistent with the diam-eter of dsDNA was obtained at E > +0.45 V vs AgQRE. These values are consistent with the DNAduplex length and diameter respectively, thus sug-gesting the orientations are normal and parallel tothe surface for more negative and more positivepotentials respectively. More recently, Josephsand Ye took this technique one step farther, achieving single molecule resolution [113] . Their mainfinding is the effect of the underlying mercaptohexanol SAM in the switching performance. More pre-cisely, the presence of defects in the layer produced areas with higher electric field. As a result the DNAwas attracted towards these defects with more strength at potentials positive from the PZC becomingstrongly immobilized. In order to create these defects it was necessary to first apply a potential of +0.6V vs Ag|AgCl.Since the DNA molecules respond to electric field, only the charges that are within the Debye length(see section 2.1.1 for a detailed explanation on this concept) will be effectively repelled when the elec-1237.3. The “switching” DNA Constructtrode is negatively charged. This has implications in the conformation change of ssDNA compared todsDNA. In the first case, the molecules are flexible due to their short persistence length. As a result,only the bases closest to the electrode will be affected by the electric field and the conformation of theremaining bases will not significantly change (Fig. 7.3 bottom). On the other hand, the much larger PLvalue for dsDNA means that short double stranded DNA will behave essentially as a rigid rod. Still, onlythe bases within the Debye length will be repelled, but in this case the torque they create is enough tomove the distal end of the DNA strand farther than in the single stranded case (Fig. 7.3 top).In order to have maximum switching efficiency, it is important that neighboring DNA strands do notinteract with each other causing steric hindrance. Thus, a low DNA coverage is a definite requirement ifthese layers are to be used for sensing purposes. Different strategies have been published to achievesuch low coverage layers:1. Short exposure times to a thiolated DNA solution followed by a longer exposure time to 6-mercapto-1-hexanol (MCH) [271, 272].2. Formation of a relatively dense layer of thiolated DNA, filling the spaces in between strands withMCH and subsequent partial reductive desorption in an MCH rich medium [279, 269, 268].3. Exposure to MCH solution followed by thiolated DNA solution [218].If a fluorophore is placed at the distal end of the DNA molecule, then the fluorescence intensity willchange with the DNA conformation due to the difference in distance between the Au and the fluorophore.As mentioned in Section 2.3.2, when the fluorophore is closer to the metal surface its fluorescenceintensity will be smaller. As a result modulation of the potential of the electrode creates a correspondingmodulation in the fluorescence signal (Fig 7.4).Sinusoidal potential perturbations have been used by Rant et al to extract kinetic information on the“switching” of these systems. By varying the frequency of the applied AC potential, it is possible to gofrom low frequency regimes in which the DNA switches completely between the lying down and standingup states to high frequency regimes in which the switching can no longer occur efficiently (inset in Fig.7.5). At these high frequencies the DNA remains essentially static regardless of the potential change[268]. In order to acquire such measurements a fast acquisition system is required (frequencies usuallyreach 100 kHz) comprising a photomultiplier tube and some frequency discerning device, such as a lock-in amplifier or a frequency response analyzer. To characterize this frequency-dependent response, Rantdefines the cut-off frequency (ƒc) as the frequency at which the fluorescence signal amplitude equalshalf of its value at low frequencies. Due to the drag of the solution surrounding the DNA, the cut-off1247.3. The “switching” DNA ConstructFigure 7.4: Relationship between applied potential (top left) and fluorescence response (left bottom)from an electrode-immobilized 48-mer double stranded DNA labeled with the Cy3 fluorophore in thedistal end (right). Reprinted from [272] with permission from Biophysical Society.frequency is influenced by the hydrodynamic radius of the “switching” molecule. Thus, modification ofthe labeled DNA strand alters the response, shifting the cut-off frequency. This effect has been exploitedto construct surfaces that have a different response when ssDNA binds its complementary DNA strand[268] and, in a more elaborate setup, when dsDNA labeled with an epitope protein tag binds to itsrespective antibody [273].Besides using this technology to determine the concentration of a given analyte, Rant has proposedits use to determine the size of the bound target molecules [273]. A shift to lower values in the cut-offfrequency is observed as the hydrodynamic diameter of the target molecule increases. This trend is ex-plained through the added friction that the large target molecule creates during the switching. Althoughinitially very promising, this technique has failed to distinguish differences in target molecule size undercertain conditions (e.g. high viscosity), and a different approach has been suggested. Instead of using afrequency dependent analysis, a potential step is performed and the time resolved evolution of the fluo-rescence signal is measured, providing separate information on the “lifting” and “lying down” processes[275]. By fitting the time-dependent fluorescence to a model, the effective hydrodynamic diameter ofthe bound molecule can be extracted [280]. The precision of the size determination is improved sincenow both rates (going up and down) are measured independently, whereas in the frequency responseanalysis approach the slowest time constant will limit the overall response.1257.4. Average vs Space Resolved MeasurementsFigure 7.5: Variation of the fluorescence modulation amplitude with applied AC potential frequency forSAMs containing fluorescently labeled DNA. Two different DNA lengths (24 and 48 base pairs) arecompared. Buffer electrolyte (pH = 7.3) is composed by 10 mM Tris and 50 mM NaCl. Disc electrodediameter is 2 mm. Adapted from [268]; copyright 2007, National Academy of Sciences.7.4 Average vs Space Resolved MeasurementsThroughout their published work, Rant et al have used fluorescence measurements spatially averagedover spot sizes in the order of hundreds of microns to characterize their layers [279, 269, 268, 275, 272,271]. As such, this technique will be insensitive to differences in the behavior for different regions on theelectrode (Fig. 7.6). Most often in literature, the DNA SAM is prepared by exposing the clean surfaceto a buffered solution of thiol-modified DNA (1-10 µM) for a specified time period (up to several hours),rinsing with buffer and then immersing in MCH solution (1 mM for 1 h) to remove the non-specificallyadsorbed DNA. Through the use of in-situ fluorescence microscopy it has been shown that followingsuch procedure can result in very heterogeneous DNA SAM surfaces (Fig. 7.6a) [218]. Notably, regionsof high intensity (“hot spots”) were observed and interpreted as DNA aggregates. By changing theorder of the sequential deposition process (DNA is preceded by MCH) used to prepare the DNA SAM,significantly less heterogeneity can be obtained, along with much lower surface concentrations. It wasnot clear, however, which factors determined the remaining heterogeneity and what modifications wereneeded to further reduce it or to eliminate the “hot spots”.1267.5. Objective650004875032500162500150 µm a bFigure 7.6: a) Fluorescence microscopy image of a DNA / MCH layer deposited on a gold bead. b)Image created by averaging the values of all the pixels in (a) to represent the information obtained withspatial averaging techniques. Images are monochromatic and displayed in false color according to thescale shown in the right side. Figure (a) adapted with permission from [218]. Copyright 2009 AmericanChemical Society7.5 ObjectiveIn this chapter, further spatial characterization of mixed MCH / DNA layers will be accomplished throughthe use of spectroelectrochemical techniques. The implementation of a frequency response analy-sis method using fluorescence microscopy will be described. This enables frequency dependent fluo-rescence measurements from small regions (90 µm in diameter) of the electrode surface, providing acharacterization method of the molecular arrangement and environment of the SAM. First, the effect ofthe substrate crystallinity on the density and “switching” efficiency of the DNA will be discussed. Thesensitivity of this approach to different regions of the surface modified by a DNA SAM will be demon-strated. Later on, an attempt to decouple the kinetics of the DNA switching from the electrochemicaltime constant will be undertaken. Highlighted are some of the challenges in measuring these modulatedfluorescence signals using low electrolyte concentrations. A procedure for accurately determining thefrequency response curve of the DNA orientation changes with potential on the electrode surface isoutlined.7.6 Experimental7.6.1 Au bead electrode preparationTheworking electrodes weremulti-crystalline Au beads created through themelting of a 0.5mmdiameterAu wire with subsequent cooling. The rate of cooling determines the number of crystalline domains1277.6. Experimentalpresent in a given bead. This procedure results in slightly deformed spheres (Ø≈2.3 mm) with opticallysmooth regions as well as some facets of particular crystalline orientations as well as surface defects.This substrate should represent the smoothest types of surfaces (polished gold discs) that are typicallyused in the preparation of DNA-SAMs.7.6.2 Layer PreparationThe MCH-DNA modified surfaces were prepared as reported previously [218]. Briefly, a Au bead wasimmersed in a 1 mM solution of 6-mercapto-1-hexanol (MCH) in a pH = 7.5 buffer solution composed of10 mM 2-amino-2-hydroxymethyl-propane-1,3-diol (Tris) and 100 mM NaCl for 90 min. After rinsing withthis same buffer and ultrapure water (Millipore, USA) the bead was immersed in a solution containing 10mM Tris, 100 mM NaCl, 500 mM MgCl2 and 1µM DNA for 24 hours. The DNA sequence employed was(from 5’ to 3’) HS-C6-CTG-TAT-TGA-GTT-GTA-TCG-TGT-GGT-GTA-TTT-AlexaFluor 488. This sequencewas chosen such that no secondary structures were stable. Furthermore, the oligonucleotide was longenough to create a separation from the surface that yields a significant fluorescence change but shorterthan the persistence length of dsDNA (future hybridization experiments are planned). Upon immobiliza-tion, the electrodes were rinsed again with Tris / NaCl and ultrapure water. The samples were storedovernight in 10 mM Tris, 100 mM NaCl, 500 mM MgCl2before analysis to remove the non-specificallyadsorbed DNA. If this step is omitted, a fluorescent plume can be observed diffusing away from theelectrode surface during initial potential step experiments (Appendix D).It is important to note that, unlike in previous DNA work carried on in this lab where dithiol modifiedDNA was reduced, purified and the used immediately, this work made use of thiol modified DNA. Thismay have consequences on the DNA coverage and is currently under investigation.7.6.3 In-situ Fluorescence ImagingThe modified surface was characterized using in-situ fluorescence microscopy as described previously[218, 220, 170]. Briefly, the custom spectroelectrochemical cell is used with an inverted microscope(Olympus IX-70) configured as shown in Figure 7.7. The working electrode was positioned in the cen-ter of the cell, 2-5 mm from the window. The electrolyte was 10mM TRIS, 10 mM or 100 mM KNO3pH=7.5, prepared in ultrapure water. The reference electrode was a saturated calomel electrode (SCE)connected via a salt bridge and the counter electrode a Pt wire. The filters used in the epi-fluorescencesetup were from Chroma and specific for the AlexaFluor 488 fluorophore (dichroic mirror 495 nm, ex-1287.6. Experimentalcitation 450-490 nm and emission 500-550 nm). Fluorescence images were acquired with the SBIGCCD camera through a 5×, 10× or 40× objective. The latter objective is a water immersion one, and itsshort working distance (3.3 mm) required the bead to be lowered close to the window, exposing someof the electrode stem to solution. For this objective, the region between the window and objective wasoccupied by MilliQ water.The interface was electrochemically characterized using differential capacitance measurements withan AC perturbation (5 mV rms, 200 Hz) superimposed on a DC potential, applied using a FHI ELABpotentiostat. The measured current was input into a Stanford Research Systems SR830 DSP lock-inamplifier, the real and imaginary components measured and the capacitance value calculated assuminga series RC circuit. Coordination of the fluorescence image acquisition with the application of potentialsteps and measurement of the capacitance was performed using a LabView program. The potentialwas switched between a constant base potential (Eb) and a continually decreasing (negative going) steppotential (Es) with a fluorescence image (3 sec exposure) acquired at each step. The time spent at eachpotential was 7.1 s, and the change in the Es potential was -25 mV. Taking the difference between theimage measured at Es and that measured at the previous Eb, the change in fluorescence intensity (ΔF)as a function of the applied potential was determined for each pixel. Dividing by the image at Eb enablesthe calculation of the relative change in fluorescence across the surface ΔFFb =F(Es)−F(Eb)F(Eb). Besidesproviding a means to compare the “switching” efficiency between different areas of the electrode, thiscalculation helps to correct for variations in lamp intensity over time. Thesemeasurements are effectivelyrecorded in a low frequency (0.14 Hz) regime which may be considered the DC response.This procedure effectively results in a low frequency (0.14 Hz) measurement which may be consid-ered the DC response.7.6.4 Non-imaging Spectroelectrochemical MeasurementsOnce different regions of interest (ROI) were identified from the microscopy images, the non-imaginganalysis for these ROIs was accomplished using the 40× water immersion objective. The illuminationaperture was reduced in size which allowed illumination and therefore collection from only an area of90 µm in diameter. The collected light was filtered, and directed through two collimated pinholes into aNewport 77348 photomultiplier tube (PMT) with rise time = 2.2 ns.For the FRA experiments it is desirable to obtain a linear response of the fluorescence with poten-tial. In order to find this E range, cyclic voltammetry experiments were performed. Furthermore, theseexperiments also served to determine the potential regions of stable response to ensure that desorption1297.6. ExperimentalWECE4E REEAC(f)CShiAC(f)iAC(f)FAC(f)EAC(f)Z(f) =EAC(f)FAC(f)H(f) =150 µmCCD cameraAutolab FRASR570  i-EPMTFigure 7.7: Schematic of the system employed. One port of the microscope is connected to a CCDcamera to detect images while the other port directs the light towards a photomultiplier tube (PMT). Thephotocurrent is transformed to a voltage by a SR570 current preamplifier and the resulting frequency-dependent modulated signal is used to compute the transfer function H(ƒ ). The electrochemicalimpedance Z(ƒ ) is calculated from the measured potential and current oscillations. Adapted from [74]and [115] with permission from Olympus and Bioanalytical Systems, Inc.1307.6. Experimentaldoes not occur. The potential was cycled between +0.35 V and -0.3 V at a rate of 100 mV/s while con-tinuously recording the PMT signal. In this case the potential was applied with an Autolab PGSTAT 30potentiostat equipped with a scan generator module.The fluorescence response to an applied AC potential perturbation was measured with the use of theAutolab PGSTAT30 equipped with the FRA2module over a wide frequency range (30 Hz - 100 kHz). Thepotential perturbation generated by the potentiostat was superimposed onto a DC potential and appliedto the working electrode. The fluorescence signal (DC and modulated) from the PMT was conditionedwith a Stanford Research Systems SR570 low-noise current preamplifier (low pass 1 MHz, 12 dB rolloff, high bandwidth). After biasing the output so as to remove the majority of the DC component of theintensity, the resulting voltage (proportional to the light intensity) was fed into the FRA2 module in thepotentiostat. In addition to the fluorescence FRA response, a conventional electrochemical impedancespectrum (EIS) was acquired (using the same frequency range) immediately after the optical measure-ments with either a 5 mV rms or 200 mV peak to peak (p-p) potential perturbation centered at the sameDC potential. Fitting of the EIS data was performed with the Autolab Nova software. The electrolyteconcentration used was varied from 10 mM to 100 mM. The low ionic strength case (10 mM) introducessignificant problems in the control of the WE potential at higher frequencies (>10 kHz) when using a saltbridge. A “fourth electrode”[62] was introduced to prevent artifacts resulting from the high resistance inthe salt bridge. This was implemented by connecting a Pt wire in the electrolyte solution to the referenceelectrode connector through various small valued capacitors depending on the measurement. Specificson the influence of this capacitor on the measurement is explored further in the results section. The useof the “high speed” bandwidth option in the Autolab PGSTAT30 provided unfiltered signals up to 100kHz.The resulting fluorescence FRA spectrum is output as a transfer functionH(ƒ ) =FAC(ƒ )EAC(ƒ )(7.1)as the fluorescence response (PMT signal) divided by the potential perturbation, both complex quantitiesresulting in a complex transfer function. The magnitude and phase are calculated in the usual manner.In addition, the average fluorescence signal (i.e. the DC value of fluorescence intensity) was determinedand used to normalize H(ƒ ) creating a relative fluorescence FRA spectrum H(ƒ )FDC , allowing comparisonwith the analyzed fluorescence images. Care was taken to ensure the potential never deviated from theregion of stability for these DNA SAMs (-0.4V to 0.4V / SCE). .1317.7. Influence of the Crystallographic Orientation on the Fluorescence Response7.7 Influence of the Crystallographic Orientation on theFluorescence ResponseAs the size of biosensing devices needed for point-of-care applications continuously decrease, the re-quirements for the sensing surface itself have favored research in miniaturization of these layers. Onemain difficulty that arises with a reduced sensor size is the risk of creating a heterogeneous surface suchthat the distribution of the biorecognition element is no longer statistically similar across different devices[11, 12]. In particular, a large number of SAM based sensors use polycrystalline Au substrates. As such,it is necessary to investigate the heterogeneity of the sensing surfaces to determine the effects of thecrystallographic orientation on the biorecognition element coverage and thus in the sensor response.This section describes the use of spectroelectrochemical techniques to probe for heterogeneity in a DNA“switching” sensing platform.7.7.1 Characterization of the BeadsSlow cooling during the fabrication of the gold bead electrodes favors the formation of large crystallinedomains. This is evidenced by the appearance of flat facets that are observable both in brightfield anelectron microscopy images (Fig. 7.8a and b respectively). Electron backscatter diffraction (EBSD) wasused to index the crystallographic orientation of the gold beads. Although EBSD is a well establishedtechnique , problems arose due to the curvature of our electrodes. The surface to analyze should beat an angle of ~70º with respect to the incidence normal to maximize the contrast of the backscatterdiffraction pattern . Since our bead is curved, the angle at which electrons come in contact with theelectrode surface changes from one location to the next one, making the backscatter efficiency notoptimal in most of the surface. As a result, our EBSD experiments were limited to small areas inside theflat facets, where the angle can be controlled to obtain enough electrons for a good indexing procedure.Figure 7.8c shows the EBSD orientation map for an area of approximately 6 × 6.5 µm inside of theoutlined facet in panel (a). When comparing the color (representative of the crystal orientation) with thestereographic triangle for an FCC material (Fig. 7.8d) the low index plane that more closely correspondsto the observed color is a {111} orientation. Based on this assignment, and knowing the symmetry ofan FCC crystal the rest of the low index planes in the same crystalline domain can be elucidated.1327.7. Influence of the Crystallographic Orientation on the Fluorescence Response800 µm adcbFigure 7.8: Characterization of the Au bead. a) Brightfield micrograph. b) Scanning electronmicrograph.c) EBSD mapping of a 6 × 6.5 µm are inside the facet outlined in (a). d) Color coded stereographictriangle for an fcc material.1337.7. Influence of the Crystallographic Orientation on the Fluorescence Response7.7.2 Fluorescence Imaging MicroscopyThe Au bead electrode prepared with the MCH/ssDNA SAM was placed in the spectroelectrochemicalcell and the surface was imaged using a low magnification objective (5×). Since the bead is round, notall of the surface is in focus, so images at different focal planes were collected andmerged together usingthe Extended Depth of Field plugin for ImageJ [221, 222]. This fluorescence image is shown in Figure7.9a. As stated in the previous section, the most noticeable facets have been identified as {111} surfacesby EBSD. The other regions are more difficult to assign through EBSD, but the surface symmetry of thefluorescence led to the obvious conclusion that the four fold symmetric region is {100}. Furthermore,a small facet is observed in this area in agreement with it being the second lowest in surface energy(Table 2.1). The differences in the fluorescence intensity reflect the energetics of the process used tocreate the SAM. MCH is first adsorbed to a clean gold surface, covering all regions with a variety ofsurface coverages, depending on the surface symmetry. The ssDNA is allowed to compete with theMCH for the surface, and the extent of incorporation will depend on the energetics of the thiol exchangeprocess. Those regions to which MCH is more strongly adsorbed will not be effectively exchangedwith a large amount of DNA. From the fluorescence image, it appears that a number of regions havesignificantly more or less DNA than the low index planes. Zooming into one of the crystalline domains,the stereographic triangle for a fcc crystal can be overlayed using the low index planes as a guideline(Figure 7.9b). Assuming all the fluorophores are similarly distant from the Au surface, the fluorescenceimages are a map of the relative strengths of the MCH-Au and ssDNA-Au interactions.In addition to giving access to surface specific information regarding the extent of DNA coverage, thismicroscopic approach to analyzing the MCH/ssDNA surface allows for mapping of the efficiency of thepotential induced changes in the DNA orientation. As mentioned, the reorientation of DNA on a metalsurface can be manipulated by either charging the surface positively, attracting the negatively chargedDNA phosphate backbone, decreasing the distance the fluorophore is from the metal thereby decreas-ing its fluorescence, or oppositely, charging the electrode negatively, repelling the DNA and increasingthe fluorescence measured. A series of essentially DC fluorescence imaging measurements were per-formed during step changes in potential. The fluorescence and capacitance were measured during aseries of potential steps, from a base potential Eb to various Es potentials, each 25 mV more negativethan the last. An example of this series of measurements is shown in Figure 7.10a. As expected thefluorescence increases as the Es potential becomes more negative. A small change in capacitance atEs is observed. It is important to note that the change in the Eb capacitance show very little changeindicating that the layer is not being irreversibly desorbed. The largest change in fluorescence occurs1347.7. Influence of the Crystallographic Orientation on the Fluorescence Response100110100110410310311211210320111111facet111facet500 µm 250 µm a bFigure 7.9: Fluorescence image of the MCH/DNA-SAMmodified gold bead electrode. These images arethe result of merging a collection of fluorescence images at different focal points providing an extendeddepth of field [221, 222]. Note that small artifactual speckles are created by this merging process. Asample of original images is included in Appendix F. The images are pseudocolored to match the emis-sion wavelength of AlexaFluor 488. a) Image of the whole bead at 5× magnification. The low indexorientations are indicated in the figure. b) One of the crystalline domains observed at 10× magnifica-tion. Superimposed onto the image are the surface crystallographic features extracted from the cubicstereogram (Fig. 2.6).at the most negative Es potential value. Taking the difference between the images measured at Es andthat measured at Eb just before then dividing by the Eb image enables the calculation of the relativechange in fluorescence across the surface (ΔF/Fb). The changes observed range from 0 to 20% ofthe fluorescence measured at the positive potential. The extent of DNA “switching” strongly dependson the coverage of DNA [271] which is also evident here. The most intense regions shown in Figure7.9b present the lowest relative increase in fluorescence ({100} and {311}) while the converse is true,regions like {310} and {111} have the largest relative “switching”. The ΔF/Fb shown here has impor-tant consequences in the measurement of the extent of DNA “switching” for the surfaces whose DNAcoverage is manipulated by quick reductive desorption pulses. As we and others have shown [1, 194],the reductive desorption process is strongly surface symmetry dependent, influenced by the PZC of thegold surface. As the {111} surface has the most positive PZC, it would be desorbed first, while the moreopen surfaces, like {210} would be expected to be desorbed later. Changes to the DNA coverage willbe represented non-uniformly across the gold surface, resulting in areas which are {111} being stronglydepleted, compared to other regions. This method for controlling the surface coverage will ultimatelyresult in non-uniform DNA coverage, particularly for the types of multicrystalline gold surfaces (vapordeposited layers on glass/mica) typically used in these measurements/devices.1357.7. Influence of the Crystallographic Orientation on the Fluorescence Response100110100110410310311211210320111111facet111facet500 µm 250 µm a b− (V vs. SCE)3.54.5C/A (µF/cm2 )EbEs0. 100 200 300 400∆F/F b (mean grayscale)t (s)ABCDEABCD E100310311210baFigure 7.10: Fluorescence imaging measurements during potential step experiments. a) The poten-tial perturbation imposed on the interface and the resulting changes in capacitance and fluorescence.Capacitance values are shown for both Eb and Es parts of the perturbation. The relative increase influorescence is shown for five different regions of interest shown in (b). b) The maximum relative changein fluorescence (ΔF/Fb) due to the potential perturbation with the an overlaid crystallographic triangleas per Fig. Fluorescence Frequency Response AnalysisThe extent of DNA reorientation calculated in Figure 7.10b represents the DC response of the inter-face. Rant et.al. used the measure of fluorescence response to an alternating potential perturbationas a method to characterize the DNA modified interface and to distinguish ssDNA from dsDNA afterbinding to its complementary strand [268, 274]. In addition, they showed that differences in the hy-drodynamic radius of the DNA caused by binding of a protein at its distal end produces a significantreduction on the kinetics of the DNA reorientation process [273]. They performed these measurementsafter reductive desorption conditioning of the DNA SAM modified surface, blocking all of the desorbedregions with MCH to prevent non-specific adsorption. Nevertheless, given the results above, a signifi-cant surface heterogeneity towards the response of the layer is to be expected. This section describesthe measurement of the same fluorescence FRA response from small regions (~96 µm in diameter)selected from various regions around the electrode surface. This was accomplished by changing toa 40× water immersion objective with a water drop occupying the space between the objective andthe bottom of the spectroelectrochemical cell. This significantly improved the collection efficiency andspatial resolution (see Appendix E for a comparison between dry and water immersion objectives). Inaddition, to ensure only the region of interest was measured, the opening in the field diaphragm wassignificantly reduced, illuminating only the region of interest through the objective. Since this is an epi-1367.7. Influence of the Crystallographic Orientation on the Fluorescence Responsefluorescence setup, the same objective collects the fluorescence which was then passed through twocollimated pinholes before illuminating a PMT. The resulting alternating fluorescence intensity signal andthe potential perturbation across the cell is processed by the Autolab FRA2 module yielding the transferfunction Hcell(ƒ ) =FAC(ƒ )EAC,cell(ƒ ). To enable comparison with the fluorescence image of relative fluores-cence changes, the transfer function was divided by the average fluorescence signal (FDC), yielding arelative modulated fluorescence changeHcell(ƒ )FDCfor each ROI.Fluorescence FRA spectra were measured on the five different regions shown in Figure 7.10b. Thecircular regions represent the illuminated region from which the measurements were obtained. An ex-ample of these regions is shown in Figure 7.11a for ROI B. Not the whole image is in focus given thatthe region is not in the bottom of the bead and thus not parallel to the focal plane. The resulting flu-orescence intensity cyclic voltammogram is shown in Figure 7.11b along with the measured CV. Thefluorescence signal (given as the PMT photocurrent) increases with a decrease in the DC potential asexpected, which will result in a phase shift of 180° when measuring the transfer function. Also shownis the potential range used in these measurements: 0 to +350 mV; EDC = +175 mV chosen so as toremain in the linear portion of the fluorescence-potential curve. Although the exact description of thefluorescence dependance on potential would require modeling of the DNA movement and a known fluo-rescence quenching curve, a simple evaluation of the experimental response allows to choose the linearportion, which is convenient for the FRA technique. Given that the CV and capacitance measured inthis potential range is relatively constant, it can be reasonably expected that the electrochemical signalsfor this large perturbation will still be in a linear response regime. That said, after the measurement ofthe fluorescence impedance, EIS was also recorded but using the more traditional 5 mV rms amplitude,centered around the same EDC value.Measurement of the transfer function (Hcell(ƒ )) for each of the five regions was accomplished, inaddition to EIS after each set of measurements. The extent of fluorescence modulation is expected tovary significantly for each area as shown in the imaging measurements. For comparison purposes thetransfer function is divided by the average fluorescence intensity which is measured during the FRA.This allows for comparison with the fluorescence image analysis and between regions. The magnitudeand phase of the transfer function for the different ROIs are shown in Figure 7.12a and b respectively.The transfer function shows a roll off with a cut-off frequency at ~1 kHz as in agreement with work byRant (Fig. 7.5) [268]. At low frequencies the phase of the fluorescence response starts at 180°, which isexpected since the fluorescence increases as the potential decreases (or gets more negative). At highfrequencies the phase decreases first to 90° and tending towards zero at the highest frequencies.1377.7. Influence of the Crystallographic Orientation on the Fluorescence Responsea50 µma b350 mV p-p96 µm− (µA/cm2 )−0.3 −0.2 −0.1 0.0 0.1 0.2 0.3 0.4I (µA)E (V vs SCE)Figure 7.11: a) An image of region of interest B as seen through the water immersion objective show-ing the extent of illumination and the resulting fluorescence collection that is passed to the PMT. b)Electrical current (top) and fluorescence intensity (bottom) measured with potential cycling between thepotential limits used. An indication of the potential perturbation regions used in the fluorescence FRAspectroscopy measurements is shown. The electrochemical current signal has been smoothed with aSavitzky-Golay filter [226, 227].In order to graphically compare the cut-off frequency for the different crystallographic orientations theHcell(ƒ ) was divided by the response at 30 Hz, Hcell(30Hz). This normalized signal is shown in Figure7.12c and d. In this plot, ƒc is readily observed as the frequency at which the normalized value ofHcell(ƒ )FDCequals 0.5. Close examination shows that a small but noticeable difference in the cut-off frequencyis observed, ranging from 1300 to 1800 Hz. The significance of this heterogeneity in the frequencyresponse is evidenced by realizing that the cut-off frequency shift caused by hybridization is only ~25 to30% (Fig. 7.5) [268].At this moment the reason behind the observed frequency shift is not clear. Comparing the order ofthe ƒc with the local DNA density (as indicated by both the intensity and “switching” efficiency) for theROIs suggests that the amount of DNA on the surface is not the only determining factor. It is possiblethat differences in charge between regions with different crystalline orientation caused by differences inthe PZC influence the rate of “switching”. However this influence should be visible as a shift along thepotential axis of the F vs E curve and as long as the potential perturbation is changed accordingly toremain in the linear region, that effect should be accounted for. Alternatively, the DNA reorientation isdriven by the modulation of charge in the electrode, which is related to the capacitance of these specificinterfaces (q=CE). Nevertheless, the difference in capacitance for the {100} and {111} interfaces (whichshow the fastest and slowest response times respectively) cannot be that large, and the modulationamplitude used is large enough to ensure that the extent of charging and discharging is similar. Further1387.7. Influence of the Crystallographic Orientation on the Fluorescence Response100110100110410310311211210320111111facet111facet500 µm 250 µm a bABCD Ebc da0. 103 104 105|H cell / FDC| (V−1)f (Hz)ABCDE060120180102 103 104 105H cell/FDC phase (°)f (Hz) 103 104 105|H cell / FDC| normalizedf (Hz) 1300 1600 2000|H cell / FDC| normalizedf (Hz)Figure 7.12: Hcell(ƒ ) measured for different ROIs indicated in Fig. 7.10b. The magnitude (a) andphase (b) for this transfer function is shown, along with the magnitude normalized at 30 Hz (c and d).Red dashed line in (d) is aids to visualize the cut-off frequency.1397.8. On the Deconvolution of the Electrochemical Time Constant on DNA “Switching” Kineticsexperiments are required to elucidate the reason behind the frequency shift.7.8 On the Deconvolution of the Electrochemical Time Constanton DNA “Switching” KineticsIn all the analysis presented in the previous section it has been assumed that the hydrodynamic dragis the main factor limiting how fast the DNA can “switch” its configuration. This might not always be thecase. The movement of the DNA is a response to a change in the charge (and hence in potential) atthe electrode surface. Charge accumulates over a period of time; it is not an instantaneous process.The time taken to charge the electrode depends on its capacitance and on the solution resistance. Thecapacitance depends on the nature of the layer covering the electrode as well as on its area, while theresistance is determined by the electrolyte concentration and the geometry of the electrochemical cell.The charging kinetics of an electrode can be characterized by the electrochemical time constant (τE)defined as the time it takes to charge the electrode to ~63.2% (the actual value is 1− e−1) of the totalcharge. It is clear that if the electrode charging time is comparable or longer than the time needed for theDNA to change its configuration, the frequency dependent fluorescence response will be convoluted withcharging rather than purely reporting on the DNA hydrodynamics. Since the sensors are based on thefrequency response changes caused by the size of the molecule with binding events, electrochemicaleffects are undesirable.Rant et al. have indeed observed order-of-magnitude shifts in the fluorescence frequency responsewhen changing electrolyte concentration and electrode size [268]. To counter the τE problem, theymain-tained the experimental conditions constant to reduce variations in the electrochemical time constant forexperiments to be compared. Furthermore, they have attempted to correct for changes to the electro-chemical time constant by the binding process itself by measuring an electrochemical impedance spec-trum before and after binding. After determining the corresponding electrochemical time constants (τE1and τE2 respectively), the fluorescence cut-off frequency is multiplied by the correction factor τE2/τE1[273]. While both of these approaches help mitigate the effect of the changes of time constant on theexperimental results, they do not eliminate its influence on them.3 As a result, comparison betweenresults performed under different experimental conditions is difficult.It would then be desirable to create a method by which the influence of the electrochemical time3The model for the interpretation of the time resolved measurements mentioned in Section 7.3 does take the charging timeinto account.1407.8. On the Deconvolution of the Electrochemical Time Constant on DNA “Switching” Kineticsconstant is either reduced or eliminated. A straightforward solution to this convolution problem wouldbe to increase the electrolyte concentration in solution. This would decrease the electrochemical timeconstant so that the fluorescence response would be dictated mainly by the DNA hydrodynamics. How-ever, such a condition would result in shielding of the charge on the electrode surface, decreasing theamplitude of the DNA “switching” response [269]. A second possibility is the reduction in the electrodearea, a trend that is clear when comparing the progress of Rant’s publications [268, 273, 280], who usesmicroelectrodes in his most recent publication. This is an attractive possibility given the current sensorminiaturization trends. As discussed in the previous section, however, control of the surface hetero-geneity becomes crucial in such systems. Although both of these approaches ameliorate the influenceof τE, they do not ultimately eliminate it. In this section a methodology is presented that allows for de-convolution of the electrochemistry and hydrodynamics in the measurement of fluorescence responsekinetics due to the potential driven DNA reorientation.The principle of the proposed correction is based on the realization that when the electrode is notcompletely charged, only part of the potential applied to the electrochemical cell (Ecell) would be ex-pressed at the electrode interface (Eint). The remaining potential is dropped across other componentsof the cell, mainly in the solution resistance. Since Eint is responsible for driving the DNA movement,the fluorescence modulation transfer function should be recalculated using the potential that is presentat the electrochemical interface, not the one applied across the whole cell. The output from the FRA2module isHcell(ƒ ) =FAC(ƒ )EAC,cell(ƒ ). Then, if the ratio of the cell potential to the interface potential€EAC,cell(ƒ )EAC,int(ƒ )Šis known, it can be used as a correction factor to define the new transfer functionHint(ƒ )≡FAC(ƒ )EAC,int(ƒ )=FAC(ƒ )EAC,cell(ƒ )·EAC,cell(ƒ )EAC,int(ƒ )=Hcell(ƒ ) ·EAC,cell(ƒ )EAC,int(ƒ )(7.2)It is important to note that all of these quantities are complex numbers and so they have associatedwith them a specific phase for each frequency which is needed to accurately calculate the modifiedtransfer function. For comparison purposes the transfer function is divided by the average fluorescenceintensity measured during the fluorescence impedance (DC), enabling comparison with the image dataand between the ROIs.Once the methodology has been defined the next challenge is to find the value of€EAC,cell(ƒ )EAC,int(ƒ )Šfor eachfrequency employed. This can be accomplished by measuring electrochemical impedance spectra aftereach fluorescence FRA acquisition and fitting the electrochemical response to an equivalent circuit. Inorder for the correction to provide adequate results it is imperative to have good quality impedance data1417.8. On the Deconvolution of the Electrochemical Time Constant on DNA “Switching” Kineticsand an appropriate equivalent circuit. For this reason, substantial efforts were conducted in optimizingthe cell configuration.Fluorescence FRA spectra were measured on the same ROI (using a different electrode from theone used in the previous section) employing a variety of instrumental setups. In EIS measurements, theoptimal configuration of the three electrodes is very important so the measured response (current) isprimarily determined by the interface of interest. For this reason, the electrolyte concentration is typicallylarge, decreasing the solution resistance. In practice, the area of CE is also made significantly largerthan the WE, and the RE is connected to the working solution through a very conductive salt bridge orother configuration. These considerations for proper and accurate EIS measurements are detailed byFletcher [61] with a view to decrease the instrumental artifacts in EIS measurements. As mentioned be-fore, this work requires low electrolyte concentrations to ensure that the electric field extends far enoughinto solution to exert significant influence on the DNA conformation [269]. In addition, the need to isolatethe RE from the working solution so as to decrease any contamination from Cl− on the measurements(which occur over many hours) is critical, especially in the potential region of measurement. This in-troduces a large resistance between the RE and WE (R2 in Fig. 2.13) which will not allow for properpotential measurement and therefore the analyzer is strictly limited to low frequencies. To solve this, ashunt capacitor (CSh) is typically used connected to a Pt wire immersed in the electrolyte, the so-calledfourth electrode [62]. The value of the shunt capacitor is usually determined experimentally to enablemeasurement at higher frequencies. The measurement of the fluorescence FRA under a series of con-ditions as outlined in Figure 7.13a and b. EIS measurements were performed immediately afterwardusing the same experimental configuration, but using a 5 mV rms perturbation. The magnitude andphase of this impedance is shown in Figure 7.13c and d respectively.The fluorescence transfer function shows the expected decrease in signal as the frequency in-creases. The phase of the fluorescence response starts at 180° eventually decreasing at high fre-quencies. The use of the fourth electrode with various CSh is explored next. The changes in the transferfunction with changing the CSh are subtle, but significant. Removing the shunt capacitor significantlylimits the frequency range to 10 kHz, beyond which the potentiostat fails to control the potential appro-priately, resulting in the interface exceeding the safe potential region causing damage to the SAM. Usinga small CSh enables the full frequency range to be measured. A larger CSh enables a smoother transferfunction curve, one that seems more in line with that is expected. Comparing the EIS measurementsperformed immediately afterward, the EIS spectrum obtained with a larger CSh enables the RC nature ofthe interface to become evident. The increase in phase at higher frequencies in EIS is generally used as1427.8. On the Deconvolution of the Electrochemical Time Constant on DNA “Switching” Kinetics100110100110410310311211210320111111facet111facet500 µm 250 µm a bABCD Ebc da. 103 104 105|H cell / IDC| (V−1)f (Hz)no 4thE0.11 nF7.17 nF060120180102 103 104 105H cell / IDC phase (°)f (Hz)103104102 103 104 105|Z| (Ω)f (Hz)−90−60−300102 103 104 105Z phase (°)f (Hz)Figure 7.13: a) and b) Fluorescence FRA measurements of a single ROI under the following set ofconditions: ƒ = 30 - 100 kHz, EDC = +150 mV, EAC = 200 mV p-p, 10 mM Tris + 10 mM KNO3. Valuesfor CSh are indicated in the figure. c) and d) Electrochemical impedance measurements under the sameset of conditions except EAC = 5 mV rms. Symbols are measured data, lines are fit to a (C[R(C[RQ])])equivalent circuit.1437.8. On the Deconvolution of the Electrochemical Time Constant on DNA “Switching” Kineticsa signature of the phase shift problem due to the incomplete compensation for the impedance betweenthe RE and the WE. Since the higher CSh results in a lower phase shift at high frequencies, we usedthis as a signature to determine its optimal value. Increasing the CSh above 7 nF did not change theimpedance or the fluorescence response (data not shown). The magnitude of Z also shows the influ-ence of CSh in the middle frequency range. An artefactual second time constant is observed due to thehigh resistance salt bridge that prevents proper potential control by the potentiostat which is clearly seenin the EIS and most importantly also in the fluorescence measurements. It is clear from this analysisthat the fourth electrode has an influence on the potential expressed at the interface as the fluorescencesignal significantly changes, and in particular the phase of the fluorescence response.Once the optimal value of CS was selected (7.17 nF), an equivalent circuit model of the electrochem-ical cell is necessary to calculate the value of EAC,int(ƒ ). A starting point for the modeling of the electro-chemical cell is a resistor representing solution resistance (RS) in series with a generalized impedancethat represents the interface – at present unknown or not defined – (Zint). The fact that the phase of theimpedance does not decrease to zero even at high frequencies indicates that a stray capacitance CStis also present in the setup. The source of this capacitance has been described by Fletcher [61]. Thegeneral circuit used is shown in Figure 7.14a. To extract the values of RS and CSt from the EIS data, theimpedance of the interface must also be described in terms of an equivalent circuit. In fitting the datato a set of equivalent circuits, three options were tested. Using a CPE alone as the interface does notfit the EIS adequately in the middle frequency range. Including an [RC] component parallel to this CPE(Figure 7.14b) fits the data very well for all the systems measured. We are unable to determine fromEIS fitting alone whether the (Q[RC]) or a (C[RQ]) – shown in Figure 7.14b and c respectively – is theproper representation of the interface as they both gave equally excellent fits (Figs. 7.13c and d includethe simulations with the (C[RQ]) circuit) with very similar RS and CSt values. Therefore, the use of aspecific equivalent circuit for the interface will be limited to the calculation of RS and CSt but, as detailedbelow, will not be used to calculate EAC,int(ƒ ).To determine EAC,int(ƒ ), the measured EIS data was used along with the calculated values of RSand CSt in the following way. The applied potential EAC,cell(ƒ ) is expressed across both arms of theparallel circuit in Figure 7.14a. For the [RSZint] branch, the EAC,cell(ƒ ) is divided between RS and Zint.The potential across the interface (EAC,int(ƒ )) is found using a simple voltage divider argumentEAC,int(ƒ )EAC,cell(ƒ )=ZintRS+Zint(7.3)1447.8. On the Deconvolution of the Electrochemical Time Constant on DNA “Switching” KineticsCStCintCintRS RintRint QintQintZintEintEcella b cFigure 7.14: Equivalent circuits used to fit the EIS data. a) the general (CSt[RSZint]) circuit. b and c)The two options used for Zint resulting in indistinguishable impedance response.It is important to note that EAC,int(ƒ ) is a complex number that has a phase that is different from that ofEAC,cell(ƒ ). Two options exist for determiningZint: to calculate it from the components of the particularequivalent circuit (i.e. Qint, Rint, and Cint) chosen to perform the EIS fit, or to compute its value fromthe measured EIS data using only the values of RS and CSt obtained from the fitting. In the latterapproach, Zint remains as independent as possible from the chosen equivalent circuit, and is thereforethe chosen strategy in this work. In practice, the quality of the fits is so high, that both processes producecomparable results.The fluorescence FRA response for the system as described in Figure. 7.7 is shown in Figure 7.15aand b along with the fraction of the potential dropping across the interface (EAC,int(ƒ )/EAC,cell(ƒ )). Themeasured fluorescence transfer function Hcell(ƒ ) and the transfer function after correction Hint(ƒ ) aresignificantly different in both magnitude and phase. The almost perfect overlap between Hcell(ƒ ) andEAC,int(ƒ )/EAC,cell(ƒ ) indicate that for this set of data, the potential at the interface is clearly controllingthe fluorescence modulation. After deconvoluting the influence of the electrochemical time constant,the orientation change of the DNA appears to be a less obvious function of the frequency. Importantly,the corrected phase equals 180° to 1 kHz, followed by a slight decrease that steepens only when thepotential at the interface EAC,int(ƒ ) drops to ~ 10% of the applied potential. The physical meaning ofthese phase values is not currently understood but recent calibration experiments using a LED instead ofthe DNA modified electrode (not shown) suggest it is created by an artifact of the measurement system.The correction applied to the fluorescence transfer function can be tested by including a resistorbetween the WE lead and the conductor attached to the Au bead electrode. This will act to mimic an in-crease in the solution resistance which should decrease the frequency range where significant potentialis experienced at the interface – essentially increasing the time constant of the electrochemical setup.An advantage of this approach is that it enables the use of exactly the same DNA SAM coated electrodeto be used in exactly the same experimental arrangement. This is important since small changes in the1457.8. On the Deconvolution of the Electrochemical Time Constant on DNA “Switching” Kinetics100110100 410310311211210320111facet111facet500 µm 250 µm a bABCD Ebc da. 103 104 1050.|H/I DC| (V−1) Eint /Ecellf (Hz)Eint/EcellHcellHint060120180102 103 104 105H/I DC phase (°)Eint /EcellH/I DC phase (°)f (Hz) 103 104 1050.|H/I DC| (V−1)f (Hz)060120180102 103 104 105f (Hz)Figure 7.15: The magnitude and phase for each of the transfer functions Hcell(ƒ ) and Hint(ƒ ). Alsoincluded is including the value of EAC,int(ƒ )/EAC,cell(ƒ ). on a secondary (right hand) y axis. Subfigures(a) and (b) correspond to the system as described in Fig. 7.7 while an extra 2.46 kΩ resistor has beenadded to the WE connection in (c) and (d). [KNO3 ] = 10 mM1467.8. On the Deconvolution of the Electrochemical Time Constant on DNA “Switching” Kinetics10 mM % 10 mM + R % 100 mM % 100 mM + R %RS (Ω) 476 0.45 2929 0.150 106 0.41 1096 0.14CSt (nF) 0.182 15 0.021 8.9 0.316 28 0.036 9.9Qint (μF) 0.196 2.8 0.195 1.9 0.201 1.6 0.161 4.3m 0.968 0.28 0.977 0.17 0.986 0.13 0.955 0.65Rnt (kΩ) 2.31 21 7.76 25 0.949 18 0.538 33Cint (nF) 30.0 10 22.6 12 24.5 8.6 61.6 11χ2 1.67× 10−3 2.8× 10−4 6.6× 10−4 1.2× 10−4Table 7.3: EIS fitting results for electrolytes containing 10 mM and 100 mM KNO3 using the(CSt[RS(Qint[Rint Cint])]). R refers to the added resistor. Percentage error values are included foreach value.10 mM % 10 mM + R % 100 mM % 100 mM + R %RS (Ω) 477 0.43 2930 0.15 106 0.37 1096 0.15CSt (nF) 0.184 15 0.021 8.4 0.344 23 0.036 10Qint (μF) 1.08 18 3.36 21 0.548 15 0.851 36m 0.133 2.6 0.150 2.3 0.166 1.3 0.191 3.1Rnt (kΩ) 95.7 6.2 69.3 6.6 60.7 5.3 31.2 19Cint (nF) 0.92069 0.71 0.92196 0.70 0.93998 0.48 0.98751 0.091χ2 1.4× 10−3 2.5× 10−4 4.7× 10−4 1.3× 10−4Table 7.4: EIS fitting results for electrolytes containing 10 mM and 100 mM KNO3 using the (CSt[RS(Cint[RintQint])]). R refers to the added resistor. Percentage error values are included for each value.coverage of DNA or the positions of the other electrodes (e.g. the CE and the 4th electrode) have aninfluence on the measured signals. Furthermore, as mentioned before, the addition of higher concen-trations of electrolyte does not only affect the charging time constant but also the “switching” processitself [269]. The measurements with the added resistance are shown in Figure 7.15c and d. The solutionresistance determined from EIS for the system without the extra resistor was 476 Ω. The added resistorwas 2.46 kΩ which decreased the electrochemical time constant by an order of magnitude. The EISfitting for the system with the extra resistor yielded a total RS of 2.93 kΩ exactly reflecting the artificialincrease in the total resistance. The EIS fitting results are given in Tables 7.3 and 7.4.The uncorrected fluorescence admittance signal shows the decrease in the electrochemical timeconstant due to presence of the additional resistor. As expected, the H(ƒ ) values at low frequency areunchanged. When performing the correction for the voltage across the interface, the two fluorescenceadmittance measurements give very similar results for both the magnitude and phase, confirming theability of this correction procedure to appropriately determine the real fluorescence frequency response.1477.8. On the Deconvolution of the Electrochemical Time Constant on DNA “Switching” KineticsThis corrected signal has been effectively deconvoluted from the electrochemical response. It is impor-tant to point out that the correction is not perfect, failing at frequencies where the voltage across theinterface is <~5% of the applied voltage (frequencies >10 kHz in the case with the added resistor).At high electrolyte concentrations, the fluorescence change is significantly reduced. Whenmeasuredfrom the same spot on the electrode, the fluorescence increase at negative potentials is only ~45% of thatmeasured in 10 mM KNO3 . The extent of potential driven DNA reorientation is strongly attenuated sincemost of the potential has decayed closer to the electrode surface as compared to the measurementsperformed at 10 mM. On the other hand, increasing the conductivity of the electrolyte improves theelectrochemical characterization enabling more of the potential to be expressed at the interface at higherfrequencies. Themeasurement of the fluorescence FRA and the resulting correction are shown in Figure7.16 for the cases without an added resistor (RS = 106 Ω) and with a 1.00 kΩ added resistor resultingin an EIS measured total resistance of 1.10 kΩ. The EIS fitting results for this set of data is shown inTables 7.3 and 7.4.The results clearly show, as seen in the 10mM experiments, that the correction applied to the fluores-cence admittance is able to remove the influence of the added resistor. The limitations in the correctionis also evident in this set of data as increased uncertainty for frequencies > 10 kHz in the case withthe added resistor where the EAC,int(ƒ ) is less than 10% of the EAC,cell(ƒ ). The phase also shows aconsiderable change especially for the higher resistance case, correcting the phase over a significantfrequency range (10 - 10 kHz). The phase remains at 180°, decreasing only at frequencies > 10 kHz.Comparing the four sets of measurements, 10 mM and 100 mM KNO3 without and with addedresistances is possible if the magnitude of the fluorescence response is normalized to that at 30 Hz.Shown in Figure 7.17 are comparisons ofHcell(ƒ ) andHint(ƒ ) magnitude and phase. Also included arethe calculated EAC,int(ƒ )/EAC,cell(ƒ ), shown as lines. The data sets, measured on the same electrodeand on the same 96 µm spot reveal that the fluorescence FRA measurements, when corrected for thepotential at the interface and normalized to the response at 30 Hz, result in effectively the same transferfunction. This response appears not to show a significant decrease in fluorescence modulation below10 kHz. Furthermore, the high electrolyte concentrations, while only showing 40% of the fluorescencemodulation of the 10 mM case, has no significant drop off with frequency until close to 50 kHz. Ifcharacterized by the phase of the fluorescence response, the frequency where the phase crosses 45°is just under 100 kHz. The four sets of data show the system response that have four different timeconstants for charging the interface which span almost two orders of magnitude, while keeping thefluorescence measurement as constant as possible. It is important to note that the electrochemical1487.8. On the Deconvolution of the Electrochemical Time Constant on DNA “Switching” Kinetics100110100 410310311211210320111facet111facet500 µm 250 µm   Eint /EcellH/I DC phase (°)Eint /EcellH/I DC phase (°)a bABCDbc da. 103 104 1050.|H/I DC| (V−1)f (Hz)Eint/EcellHcellHint060120180102 103 104 105f (Hz) 103 104 1050.|H/I DC| (V−1)f (Hz)060120180102 103 104 105f (Hz)Figure 7.16: The magnitude and phase for each of the transfer functions Hcell(ƒ ) and Hint(ƒ ). Alsoincluded is including the value of EAC,int(ƒ )/EAC,cell(ƒ ). on a secondary (right hand) y axis. Subfigures(a) and (b) correspond to the system as described in Fig. 7.7 while an extra 1.00 kΩ resistor has beenadded to the WE connection in (c) and (d). [KNO3 ] = 100 mM1497.8. On the Deconvolution of the Electrochemical Time Constant on DNA “Switching” Kinetics100110100 410310311211210320111facet111facet500 µm 250 µm a bABCD0. 103 104 105|H/IDC| normalized and E int/Ecellf (Hz)10 mM10 mM + R100 mM100 mM + R060120180102 103 104 105H/IDC phase (°)f (Hz)baFigure 7.17: Fluorescence FRA as measured (open symbols) and corrected (filled symbols) for potentialacross the interface (solid lines). Results for 10 mM and 100 mM KNO3 concentrations in the absenceand the presence of the extra R. Fluorescence FRA magnitude has been normalized using its value at30 Hz.fitting errors are equally distributed across the whole frequency range, never exceeding 3% deviation ,too small to create artifacts in the calculated transfer function. The corrected transfer function is thereforethe most accurate representation of the DNA response to potential perturbation, demonstrating that theprocedure used is able to correct the experimental data for the electrochemical time constant differencesthat would be expected when comparing different samples on different days or setups. These resultsalso show that the DNA reorientation rates are rapid and for the low surface concentration interfaceanalyzed does not show substantial modulation decrease except for frequencies above 20 kHz .It is useful to compare the data presented here with published frequency response curves usingan electrode surface modified with DNA of similar length (24-72 bp), but measured in a significantlydifferent experimental arrangement [268, 273]. For their experiments, Rant et al do not always employa RE or salt bridge and the working electrodes possess an area ~10× smaller than in this work. Undersuch conditions they report time constants for the repulsion (negative potentials) and attraction (positivepotentials) that are reflected in ƒc values from 0.5 to 18 kHz. These values are very comparable with theuncorrected response presented in this work and are not similar to the results after correction.In one of Rant’s group publications, impedance spectra accompanied the frequency modulated flu-orescence response for digoxigenin-labeled DNA (Fig. 3 in [273]). In order to prove that the electro-chemical time constant did not significantly change upon binding, they fit the measured impedance to an[RC] circuit. The proposed correction described above was applied to the fluorescence data extracted1507.9. Conclusionsfrom the published graph, using the values of R = 1.6 kΩ and C = 9.6 nF described in the publication.Even using that simple circuit, Figure 7.18a shows that the fluorescence modulation closely follows thepotential drop at the interface. It is not until ƒ>20 kHz that the corrected response starts to decrease. Nophase information is included in this graph since it was not available in the original publication. Upon ad-dition of sheep immunoglobulin G (IgG) antibody a frequency shift is observed in the response slowingdown the “switching” process (Fig. 7.18b) This is reflected in a corrected curve that starts decreasingat ~1 kHz but not as sharply as in the uncorrected case. Figures 7.18c and d show a comparison ofthe response before and after the IgG binding. While a discernible shift is observed even without thecorrection, this difference is significantly amplified after taking into account the actual potential at theelectrode interface.Although Rant’s proposed simple method of normalization of the electrochemical time constant mayhelp reduce its variations due to binding events, the analysis shown above makes evident that thoseresults are still strongly influenced by the electrochemistry of the cell, potentially leading to inaccurateand unreproducible results in even simple arrangements. The inclusion of the correction enables accessto the hydrodynamic-limited “switching” regime decoupling it from the electrochemical time constant. Itmust be noted, however, that large values of τE will be reflected in low amplitudes which will translateinto large noise at the larger frequencies. Despite not being able to separate the upward and downwardmotion processes like the time resolved measurements recently proposed by Rant, it is the view of theauthor that the proposed correction method can help reposition frequency resolved measurements inthe analysis of “switching” DNA. This can be desirable since FRA methods offer a better alternative toinvestigate complex systems where, for example, more than one time constant is present.7.9 ConclusionsThe work presented in this chapter proves that in-situ potential modulated fluorescence from a DNA-containing SAM can be performed on microscopy setups enabling measurements with higher spatialresolution compared with similar studies previously reported. By using a reduced illumination and col-lection area, the frequency response of the DNA conformation change can be monitored with a fast-response PMT on regions of ~96 µm diameter. The trend in relative amplitude of the modulated fluo-rescence shows good correlation between the imaging and FRA experiments for regions with differentcrystallographic orientation. Such observations confirms the inverse relationship between coverage andswitching efficiency. However, the dependency of the cut-off frequency with crystallography is not as1517.9. Conclusions100110100 410310311211210320111facet111facet500 µm 250 µm a bABCDbc da~1.6×~3.3× 104 105|F AC| normalized and E int/Ecellf (Hz)Eint/EcellIAC,cellIAC,int0.000.250.500.751.00103 104 105|F AC| normalized and E int/Ecellf (Hz) 104 105|F AC| normalizedf (Hz)DNADNA + IgG0.000.250.500.751.00103 104 105|F AC| normalizedf (Hz)Figure 7.18: Correction applied to Rant’s data published in [273]. Panels (a) and (b) show the correctionfor digoxigenin-labeled DNA in the absence and presence of sheep immunoglobulin G (IgG) antibody.The potential across the interface is calculated using Rant’s fit to an [RC] circuit with R = 1.6 kΩ andC = 9.6 nF. Uncorrected and corrected FAC(ƒ ) values are shown in panels (c) and (d) respectively,evidencing the increase in the frequency response shift after IgG binding if the correction is used.1527.9. Conclusionsdirect and independent experiments are required to elucidate the causes.Finally, by taking into account the potential drop at the interface of the electrode as opposed tothe one applied to the complete cell, a correction was proposed that deconvolutes the influence of theelectrochemical characteristics of the cell in the measured fluorescence frequency response. Thesemeasurements revealed an important influence of the cell setup and electrolyte concentration in theobserved cut-off frequency of the DNA fluorescence. It is the view of the author that implementation ofthe proposed correction could reopen the door for using the less expensive FRA technique as opposedto the time resolved measurements currently used by Rant.153Chapter 8Concluding Remarks8.1 SummaryElectrode surfaces modified with organic thin films have been widely proposed for diverse applications.While several techniques exist to characterize these layers, some of them require high vacuum, pre-venting their use for in-situ (under potential control) measurements. Fluorescence is a good choice foranalysis of these layers since measurements can be acquired in aqueous solutions with low backgroundinterference. In particular, the distance-dependent fluorescence quenching by thickmetal substrates hasbeen exploited to characterize processes in which the separation between a fluorophore labeled layerand the metallic substrate changes.Throughout the body of this work, three different examples were shown in which the inclusion ofmicroscopy to in-situ fluorescence measurements allowed for spatially resolved measurements of ad-sorbed organic thin layers. The relevance of this capability is made clear in cases where segregation,heterogeneity or displacement of the material in the thin layers occurs. Moreover, the conductive na-ture of the substrate was exploited by using electric potential to drive changes in the adsorbed layer.Such changes ranged from orientation changes in the adsorbed molecules as a result of electrostaticinteractions (Fig. 8.1a), to the desorption of the deposited layer. Depending on the characteristics ofthe layer, this desorption involved the overcoming of physical interactions with the substrate (Fig. 8.1b)or the reductive cleavage of a chemical bond (Fig. 8.1c). Moreover, depending on the solubility of theadsorbate, the desorbed material could remain a few nanometers away from the electrode, or totally lostinto bulk solution.The generality of the employed technique with respect to the substrate was demonstrated by usingsingle-, poly- and multicrystalline electrodes. The area of these electrodes was also variable, span-ning four orders of magnitude from microelectrodes to millimeter sized polished beads. Each of thesesubstrates required a different approach for deposition and, in particular, for cleaning. A variety of ad-sorbate molecules were also employed, including physically adsorbed aliphatic alcohols and a variety1548.1. SummarySSSSSSSS SSSSSSSSSSSSSSS SSSAuOHOHOHOHOHOHOHOH OHOHOHOHOHOHOHOHOH OHOHOHOHOHOHOHOHHOHOHOHOHOHOHOSSa b cAuAuAuAu AuFigure 8.1: Schematic of the systems investigated through in-situ fluorescence microscopy. a) DNA con-formational switching (Chapter 7) b) Physical desorption (Chapter 5) c) Chemical desorption (Chapter6).of thiol-modified molecules, including DNA.In particular, the conclusions obtained for each particular system are:• The influence of the substrate during Langmuir-Schaefer depositions could be observed in two dif-ferent ways. On one hand, features arising from segregated domains were expanded anisotrop-ically when transferred to the substrate in comparison with the precursor floating layer. On theother hand, some particular domains seem to increase their dimer content of the probe. Thisresults agrees with previously reported substrate mediated condensation of layers in “liquid” state.• The fate of alkylthiolates when desorbed from a SAM is influenced by several processes, includingdiffusion, migration and advection. It was shown that under mildly basic conditions, the latter oneis dominant due to the hydrogen evolution concurrent with the reductive thiol desorption. Theseresults explain observations by other groups on differences on reoxidative adsorption dependingon the physical configuration of the cell. Incidentally, it was also observed that the presence ofan electrochemically produced reducing agent (hydrogen in this case) could be used to remotelytrigger thiol desorption on electrically unconnected structures.• Mixed MCH/DNA self assembled monolayers produced by the sequential deposition of MCH andDNA present significantly different DNA surface density depending on the crystallographic orien-1558.2. Future Worktation of the underlying metallic substrate. These differences in density are reflected in the confor-mational switching of the DNA as a response to the charge on the electrode surface. Not only isthe amplitude of the switching affected, but also its kinetics. Since both of these parameters havebeen proposed as bioanalytical signals, their study becomes of the utmost importance speciallyin the commonly used polycrystalline substrates. Furthermore, a deconvolution technique wasproposed to decouple the charging time of the electrode from the DNA hydrodynamics, the latterbeing the property of interest.It should be pointed out that one of the main findings of this work is the low reproducibility in the char-acteristics of the layers from deposition to deposition. Although certain variability is expected, this isusually ignored in the literature. In-situ fluorescence measurements make those observations inevitableand call for better understanding on the factors that influence the adsorption process. Such variationsare not only present between different depositions, but also throughout different regions in the electrode;this is to say, layers are generally heterogeneous. While typical characterization methods like capaci-tance and blocking capabilities towards a redox marker give a general idea of the layer quality, theselow-range heterogeneity cannot be elucidated by such averaging techniques. This has been realized byother groups participating in deposited layer research, and this laboratory currently holds collaborationswith two other institutions in Europe, where in-situ microscopy is being used to assess the quality of thefabricated layers. It is expected that the ability to detect layer imperfections derives into better controlof the layer deposition process improving the robustness necessary for the actual application of thistechnology.8.2 Future WorkEach of the presented examples is a story on its own and could have a long line of derived work. Someof the main ideas and experimental conditions that would help answer some of the remaining questionsare presented next.8.2.1 Influence of the Dye / Fluorescent Label on the Structure of the LayersOne question that always lingers after the performed experiments is how much the needed fluorescentmolecule or moiety itself influences the characteristics of the layer. While it is obvious that the spectro-electrochemical technique presented cannot answer this question on its own (since no signal is obtained1568.2. Future Workin an unlabeled layer), a few modifications can be made to the conditions to assess its influence.In particular, for the physisorbed system described in Chapter 5 the structure of the fluorescentmolecule was significantly different from the octadecanol constituting the majority of the layer. Thesedifferences in structure and their consequential differences in molecular interactions and fluidity resultedin a very heterogeneous segregated layer.As a result of my work with that system, a new project arose in which the BODIPY dye was incorpo-rated in an analog of octadecanol [281]. Experiments performed by my colleagues show that no fluidregions are visible using this dye, confirming the correlation between molecular and layer structures[282]. That said, layers are still far from homogeneous showing significant variation in fluorescencewhich suggests that even relatively similar fluorophore-labeled molecules can produce differences instructures throughout the layer surface.8.2.2 Potential Controlled SAM DepositionAs mentioned in Chapter 6, potential control has been used to control the SAM deposition process[184, 185, 182, 180]. While the layers employed here were entirely prepared at open circuit potential, itwould be very interesting to explore whether the observed heterogeneity can be controlled by varying thedeposition potential. Most of the alkanethiol SAMs are usually prepared from alcoholic solutions. Thus,in order to accurately establish the potential at the working electrode, a suitable reference electrode hasto be prepared for work in those solvents.8.2.3 Integration into Microfluidic DevicesThe use of macroscopic electrodes and electrochemical cells facilitates the investigation of the influenceof the desired parameters in a controlled manner. However, application of these technologies, for exam-ple, for biodiagnostics can benefit enormously from their integration into a micro total analysis system(µTAS) [283]. These devices, also known as lab-on-a-chip, generally consist of microfluidic channelsin which the different stages of the analytical process (e.g. separation, amplification and detection)are carried out. Because of the small dimensions of the channels, the totality of the device (excludingexternal equipment like pumps, light sources and detectors) usually occupies a few square centime-ters. As a consequence, small sample sizes are required, reducing analysis costs. Some examplesof fluorescence-based biodetection in microfluidic chips using SAMs are already available [284]. Fullintegration with electrochemical control and true device miniaturization will depend on the availability of1578.3. Conclusionon-chip electrical control systems as well as light sources and detectors.8.3 ConclusionAs seen above, fluorescence microscopy performed in-situ is an exceptionally useful tool to characterizespatial dependent phenomena in electrodes modified with thin layers. The work described here is justa small sample of all the potential this technique has to offer. It is the belief of the author that thismethodology will help monitor, understand and ultimately improve the quality of the deposited layersto achieve the robustness necessary for the actual application of this technology. Furthermore, thecomplementarity between electrochemistry and fluorescence yields a technique capable not only ofstudying but also of driving dynamic processes in these systems and their components. Only time willtell how much control can be achieved on such a small scale systems but in the author‘s point of view,the future of this field definitively looks exciting.158References[1] Shepherd, J. L.; Kell, A.; Chung, E.; Sinclar, C. W.; Workentin, M. S.; Bizzotto, D. Journal of theAmerican Chemical Society 2004, 126, 8329–8335.[2] ISO, Surface chemical analysis - Vocabulary - Part 1; Standard 18115-1:2013(E), 2013.[3] Vendra, V. K.; Wu, L.; Krishnan, S. Nanotechnologies for the Life Sciences; Wiley-VCH VerlagGmbH & Co. KGaA, 2007.[4] Malem, F.; Mandler, D. Analytical Chemistry 1993, 65, 37–41.[5] Peetla, C.; Stine, A.; Labhasetwar, V. Molecular Pharmaceutics 2009, 6, 1264–1276.[6] Advincula, R. C.; Knoll, W. Functional Polymer Films; Wiley-VCH Verlag GmbH &Co. KGaA, 2011;pp 1–10.[7] Love, J. C.; Estroff, L. A.; Kriebel, J. K.; Nuzzo, R. G.; Whitesides, G. M. Chemical Reviews 2005,105, 1103–1170.[8] Hamoudi, H.; Prato, M.; Dablemont, C. e.; Cavalleri, O.; Canepa, M.; Esaulov, V. A. Langmuir2010, 26, 7242–7247.[9] Porter, M. D.; Bright, T. B.; Allara, D. L.; Chidsey, C. E. D. Journal of the American ChemicalSociety 1987, 109, 3559–3568.[10] Pignataro, B. In Applied Scanning Probe Methods IX ; Tomitori, M., Bhushan, B., Fuchs, H., Eds.;Nano Science and Technolgy; Springer Berlin Heidelberg, 2008; pp 55–88.[11] Sadana, A. In Biosensors: Kinetics of Binding and Dissociation Using Fractals; Sadana, A., Ed.;Elsevier: Amsterdam, 2003; pp 17 – 29.[12] Sadana, A.; Sadana, N. In Handbook of Biosensors and Biosensor Kinetics; Sadana, A.,Sadana, N., Eds.; Elsevier: Amsterdam, 2011; pp 223 – 254.159References[13] Trasatti, S.; Lust, E. In Modern Aspects of Electrochemistry; White, R., Bockris, J., Conway, B.,Eds.; Modern Aspects of Electrochemistry; Springer US, 2002; Vol. 33; pp 1–215.[14] Bard, A. J.; Faulkner, L. R. Electrochemical Methods: Fundamentals and Applications, 2nd ed.;John Wiley & Sons, Inc.: New York, 2000.[15] Hammerich, O.; Lund, H. Organic Electrochemistry, Fourth Edition,; Taylor & Francis, 2000.[16] Valincius, G. Journal of Electroanalytical Chemistry 1999, 478, 40 – 49.[17] Aveyard, R.; Haydon, D. An Introduction to the Principles of Surface Chemistry; Cambridge Textsin Chemistry and Biochemistry Series; University Press, 1973.[18] Grahame, D. C. Chemical Reviews 1947, 41, 441–501.[19] Jovic, V. D.; Jovic, B. M. Journal of Electroanalytical Chemistry 2003, 541, 1–11.[20] Rentsch, S.; Siegenthaler, H.; Papastavrou, G. Langmuir 2007, 23, 9083–9091.[21] Pleith, W. K., W.; Twomey, T. In Adsorption of molecules at metal electrodes; Lipkowski, J.,Ross, P., Eds.; Frontiers of electrochemistry; VCH, 1992; pp 239–283.[22] Haynes,W., Ed.CRCHandbook of Chemistry and Physics, 94th Edition; CRCHandbook of Chem-istry and Physics; Taylor & Francis Limited, 2013.[23] Iwami, Y.; Hobara, D.; Yamamoto, M.; Kakiuchi, T. Journal of Electroanalytical Chemistry 2004,564, 77–83.[24] Buess-Herman, C. In Adsorption of molecules at metal electrodes; Lipkowski, J., Ross, P., Eds.;Frontiers of electrochemistry; VCH, 1992; pp 77–118.[25] Murakami, R.; Sakamoto, H.; Hayami, Y.; Matsubara, H.; Takiue, T.; Aratono, M. Journal of Colloidand Interface Science 2006, 295, 209 – 217.[26] Clasohm, L. Y.; Chen, M.; Knoll, W.; Vinogradova, O. I.; Horn, R. G. The Journal of PhysicalChemistry B 2006, 110, 25931–25940.[27] Michealides, A.; Scheffler, M. In Surface and Interface Science: Concepts and Methods; Wan-delt, K., Ed.; Wiley-VCH: Berlin, 2012; Vol. 1; pp 13 – 72.[28] Clavilier, J.; Faure, R.; Guinet, G.; Durand, R. Journal of Electroanalytical Chemistry and InterfacialElectrochemistry 1979, 107, 205 – 209.160References[29] Bilgic, S. Commun. Fac. Sci. Univ. Ank. Series B 1999, 67–84.[30] Wittels, M. C.; Sherrill, F. A.; Stiegler, J. O. Zeitschrift fűr Kristallographie - Crystalline Materials1963, 119, 42–52.[31] Borchardt-Ott, W. Crystallography: An Introduction; Springer, 2011.[32] Corcoran, S. G. Basic Structure, Plane Orientation, and Facetting in the Face-Centeredand Body-Centered Cubic Crystal Structures. http://demonstrations.wolfram.com/BasicStructurePlaneOrientationAndFacettingInTheFaceCenteredA/.[33] Burzlaff, H.; Zimmermann, H. In International Tables for Crystallography Volume A: Space-groupsymmetry; Hahn, T., Ed.; International Tables for Crystallography; Springer Netherlands, 2002;Vol. A; pp 742–749.[34] Oura, K. Surface science: an introduction; Springer: Berlin; New York, 2003.[35] Bockris, J.; Reddy, A. Volume 2 Modern Electrochemistry; Springer US, 1973; pp 623–843.[36] Bockris, J.; Khan, S. Surface Electrochemistry; Springer US, 1993; pp 59–210.[37] Smoluchowski, R. Phys. Rev. 1941, 60, 661–674.[38] Skriver, H. L.; Rosengaard, N. M. Phys. Rev. B 1992, 46, 7157–7168.[39] Chiarotti, G. In Springer Handbook of Condensed Matter and Materials Data; Martienssen, W.,Warlimont, H., Eds.; Springer Berlin Heidelberg, 2005; pp 979–1030.[40] Lecoeur, J.; Andro, J.; Parsons, R. Surface Science 1982, 114, 320 – 330.[41] Michaelson, H. B. Journal of Applied Physics 1977, 48, 4729–4733.[42] Trasatti, S. Journal of Electroanalytical Chemistry and Interfacial Electrochemistry 1983, 150, 1 –15.[43] MacNaught, A.; Wilkinson, A.; of Pure, I. U.; Chemistry, A.Compendium of Chemical Terminology:Iupac Recommendations; IUPAC Chemical Data Series; Blackwell Science, 1997.[44] Myland, J. C.; Oldham, K. B. Analytical Chemistry 2000, 72, 3972–3980.[45] Mohr, P. J.; Taylor, B. N.; Newell, D. B. Journal of Physical and Chemical Reference Data 2012,41, 043109–1 – 043109–83.161References[46] Nicholson, R. S. Analytical Chemistry 1965, 37, 1351–1355.[47] Wang, J. Analytical Electrochemistry; John Wiley & Sons, Inc., 2006; pp 29–66.[48] Garland, J.; Pettit, C.; Roy, D. Electrochimica Acta 2004, 49, 2623 – 2635.[49] Pettit, C.; Goonetilleke, P.; Roy, D. Journal of Electroanalytical Chemistry 2006, 589, 219 – 231.[50] Park, S.; Yoo, J. Analytical Chemistry 2003, 75, 455A–461A.[51] Boukamp, B. A. Solid State Ionics 1986, 20, 31 – 44.[52] NOVA Impedance Spectroscopy Tutorial. Metrohm Autolab B.V.[53] Sedra, A.; Smith, K.Microelectronic Circuits, 6th ed.; Oxf Ser Elec Series; Oxford University Press,Incorporated, 2010.[54] Kerner, Z.; Pajkossy, T. Electrochimica Acta 2000, 46, 207–211.[55] de Levie, R. Journal of Electroanalytical Chemistry and Interfacial Electrochemistry 1989, 261, 1– 9.[56] Hsu, C. H.; Mansfeld, F. Corrosion 2001, 57, 747–748.[57] Brug, G.; van den Eeden, A.; Sluyters-Rehbach, M.; Sluyters, J. Journal of ElectroanalyticalChemistry 1984, 176, 275–295.[58] Barsoukov, E.; Macdonald, J. Impedance Spectroscopy: Theory, Experiment, and Applications;Wiley, 2005.[59] MacDonald, M. A.; Andreas, H. A. Electrochimica Acta 2014, 129, 290 – 299.[60] Fafilek, G. Solid State Ionics 2005, 176, 2023–2029.[61] Fletcher, S. Electrochemistry Communications 2001, 3, 692 – 696.[62] Mansfeld, F.; Lin, S.; Chen, Y. C.; Shih, H. Journal of The Electrochemical Society 1988, 135,906–907.[63] Mansfeld, F.; Kendig, M. W.; Tsai, S. Corrosion 1982, 38, 570–580.[64] Lakowicz, J. R. Principles of Fluorescence Spectroscopy, 3rd ed.; Springer, 2006.162References[65] Fluorescence Spectra Viewer. http://www.lifetechnologies.com/ca/en/home/life-science/cell-analysis/labeling-chemistry/fluorescence-spectraviewer.html?icid=fr-af488-5.[66] Weber, G. Trans. Faraday Soc. 1948, 44, 185–189.[67] Mayor, S.; Bilgrami, S. In Evaluating Techniques in Biomedical Research; Zuk, D., Ed.; Cell Press:Cambridge, MA, 2007; pp 43–50.[68] Chen, H.; III, H. L. P.; Ikeda, S. R. Journal of Biomedical Optics 2007, 12, 054011–9.[69] Chance, R. R.; Miller, A. H.; Prock, A.; Silbey, R. Journal of Chemical Physics 1975, 63, 1589–1595.[70] Lakowicz, J. R. Analytical Biochemistry 2005, 337, 171–194.[71] Hirschfeld, T. Appl. Opt. 1976, 15, 3135–3139.[72] Zheng, Q.; Jockusch, S.; Zhou, Z.; Blanchard, S. C. Photochemistry and Photobiology 2014, 90,448–454.[73] Reichman, J. Handbook of Optical Filters for FluorescenceMicroscopy. Chroma Technology Corp,2010.[74] Olympus, Research Inverted Microscope IX71/IX81. Brochure, http://www.olympusamerica.com/files/seg_bio/seg_research_ix71-ix81_bro.pdf.[75] Dunn, K.; Wang, E. BioTechniques 2000, 28, 542–550.[76] Murphy, D. B.; Davidson, M. W. Fundamentals of Light Microscopy and Electronic Imaging; JohnWiley & Sons, 2012.[77] Lipson, A.; Lipson, S.; Lipson, H. Optical Physics; Cambridge University Press, 2010.[78] Bizzotto, D.; Lipkowski, J. J. Electroanal. Chem. 1996, 409, 33–43.[79] Parsons, R. In Advances in Electrochemistry and Electrochemical Engineering; Delahay, P., To-bias, C., Eds.; Advances in Electrochemistry and Electrochemical Engineering; Wiley, 1961;Vol. 1.[80] Zawisza, I.; Burgess, I.; Szymanski, G.; Lipkowski, J.; Majewski, J.; Satija, S. Electrochimica Acta2004, 49, 3651 – 3664.163References[81] Grahame, D. C. Journal of the American Chemical Society 1946, 68, 301–310.[82] Damaskin, B. B. Russian Chemical Reviews 1965, 34, 752.[83] Tolstoy, V.; Chernyshova, I.; Skryshevsky, V.Handbook of Infrared Spectroscopy of Ultrathin Films;Wiley, 2003.[84] Liu, H.-B.; Venkataraman, N. V.; Spencer, N. D.; Textor, M.; Xiao, S.-J. ChemPhysChem 2008, 9,1979–1981.[85] Walczak, M. M.; Popenoe, D. D.; Deinhammer, R. S.; Lamp, B. D.; Chung, C.; Porter, M. D.Langmuir 1991, 7, 2687–2693.[86] Byloos, M.; Al-Maznai, H.; Morin, M. The Journal of Physical Chemistry B 1999, 103, 6554–6561.[87] Wilhelm, P.; Chernev, B.; Pőlt, P. Macromolecular Symposia 2005, 230, 105–109.[88] Haap, W. J.; Walk, T. B.; Jung, G. Angewandte Chemie International Edition 1998, 37, 3311–3314.[89] Morschl, R.; Bolten, J.; Bonnefont, A.; Krischer, K. The Journal of Physical Chemistry C 2008,112, 9548–9551.[90] Yang, X. M.; Tryk, D. A.; Ajito, K.; Hashimoto, K.; Fujishima, A. Langmuir 1996, 12, 5525–5527.[91] Seto, K.; Furukawa, Y. Journal of Raman Spectroscopy 2012, 43, 2015–2019.[92] Člupek, M.; Prokopec, V.; Matějka, P.; Volka, K. Journal of Raman Spectroscopy 2008, 39, 515–524.[93] Evans, S. D.; Flynn, T. M.; Ulman, A.; Beamson, G. Surface and Interface Analysis 1996, 24,187–192.[94] Turner, N. H.; Dunlap, B. I.; Colton, R. J. Analytical Chemistry 1984, 56, 373R–416R.[95] Frisbie, C. D.; Wollman, E. W.; Martin, J. R.; Wrighton, M. S. Journal of Vacuum Science & Tech-nology A 1993, 11, 2368–2372.[96] McConnell, H. M. Annual Review of Physical Chemistry 1991, 42, 171–195.[97] Bizzotto, D.; Yang, Y.; Shepherd, J. L.; Stoodley, R.; Agak, J.; Stauffer, V.; Lathuillière, M.;Akhtar, A. S.; Chung, E. Journal of Electroanalytical Chemistry 2004, 574, 167–184.164References[98] Mino, T.; Saito, Y.; Yoshida, H.; Kawata, S.; Verma, P. Journal of Raman Spectroscopy 2012, 43,2029–2034.[99] Hassan, N.; Holze, R. Russian Journal of Electrochemistry 2012, 48, 401–411.[100] Zhao, Y.; Liu, X.; Lei, D. Y.; Chai, Y. Nanoscale 2014, 6, 1311–1317.[101] Iwai, H.; Hammond, J. S.; Tanuma, S. Journal of surface analysis 2009, 15, 264–270.[102] Whitesides, G. M.; Laibinis, P. E. Langmuir 1990, 6, 87–96.[103] Tarlov, M. J.; Burgess, D. R. F.; Gillen, G. Journal of the American Chemical Society 1993, 115,5305–5306.[104] Zhou, Y.; Fan, H.; Fong, T.; Lopez, G. P. Langmuir 1998, 14, 660–666.[105] Lei, S.; Feyter, S. In STM and AFM Studies on (Bio)molecular Systems: Unravelling theNanoworld; Samorì, P., Ed.; Topics in Current Chemistry; Springer Berlin Heidelberg, 2008; Vol.285; pp 269–312.[106] Noy, A.; Frisbie, C. D.; Rozsnyai, L. F.; Wrighton, M. S.; Lieber, C. M. Journal of the AmericanChemical Society 1995, 117, 7943–7951.[107] Smith, S. R.; Han, S.; McDonald, A.; Zhe, W.; Shepherd, J. L. J. Electroanal. Chem. 2012, 666,76 – 84.[108] Munz, M. Journal of Physics D: Applied Physics 2010, 43, 063001.[109] Stawasz, M. E.; Sampson, D. L.; Parkinson, B. A. Langmuir 2000, 16, 2326–2342.[110] Vericat, C.; Vela, M. E.; Salvarezza, R. C. Phys. Chem. Chem. Phys. 2005, 7, 3258–3268.[111] Couto, M. S.; Liu, X. Y.; Meekes, H.; Bennema, P. Journal of Applied Physics 1994, 75, 627–629.[112] Abel, G. R.; Josephs, E. A.; Luong, N.; Ye, T. Journal of the American Chemical Society 2013,135, 6399–6402.[113] Josephs, E. A.; Ye, T. Journal of the American Chemical Society 2012, 134, 10021–10030.[114] Chance, R. R.; Prock, A.; Silbey, R. Adv. Chem. Phys 1978, 37, 1–65.[115] Electrode Polishing and Care. Bioanalytical Systems, Inc: 2701 Kent Avenue West Lafayette In-diana 47906, 2001.165References[116] Mikhalyov, I.; Gretskaya, N.; Bergstrom, F.; Johansson, L. B.-A. Phys. Chem. Chem. Phys. 2002,4, 5663–5670.[117] Tleugabulova, D.; Zhang, Z.; Brennan, J. D. The Journal of Physical Chemistry B 2002, 106,13133–13138.[118] Bergstrom, F.; Mikhalyov, I.; Hagglof, P.; Wortmann, R.; Ny, T.; Johansson, L. B. Journal of theAmerican Chemical Society 2002, 124, 196–204.[119] Panchuk-Voloshina, N.; Haugland, R. P.; Bishop-Stewart, J.; Bhalgat, M. K.; Millard, P. J.; Mao, F.;Leung, W.-Y.; Haugland, R. P. Journal of Histochemistry & Cytochemistry 1999, 47, 1179–1188.[120] Edelstein, A.; Amodaj, N.; Hoover, K.; Vale, R.; Stuurman, N. Current Protocols in MolecularBiology; John Wiley & Sons, Inc., 2010.[121] Abramoff, M.-P. R. S., M.D. Biophotonics International 2004, 11, 36–42.[122] Keating, E.; Rahman, L.; Francis, J.; Petersen, A.; Possmayer, F.; Veldhuizen, R.; Petersen, N. O.Biophysical Journal 2007, 93, 1391–1401.[123] Zasadzinski, J. A.; Ding, J.; Warriner, H. E.; Bringezu, F.; Waring, A. J. Current Opinion in Colloid& Interface Science 2001, 6, 506 – 513.[124] Petrov, P.; Thompson, J.; Rahman, I. A.; Ellis, R.; Green, E.; Miano, F.; Winlove, C. ExperimentalEye Research 2007, 84, 1140–1146.[125] Pike, L. J. J. Lipid Res. 2009, 50, S323–328.[126] Langmuir, I. Journal of the American Chemical Society 1917, 39, 1848–1906.[127] Kwok, D. Y.; Tadros, B.; Deol, H.; Vollhardt, D.; Miller, R.; Cabrerizo-V ílchez, M. A.; Neu-mann, A. W. Langmuir 1996, 12, 1851–1859.[128] Crane, J. M.; Putz, G.; Hall, S. B. Biophysical Journal 1999, 77, 3134 – 3143.[129] Kaganer, V. M.; Möhwald, H.; Dutta, P. Rev. Mod. Phys. 1999, 71, 779–819.[130] Baoukina, S.; Monticelli, L.; Risselada, H. J.; Marrink, S. J.; Tieleman, D. P. Proceedings of theNational Academy of Sciences 2008, 105, 10803–10808.[131] Ybert, C.; Lu, W.; Mőller, G.; Knobler, C. M. The Journal of Physical Chemistry B 2002, 106,2004–2008.166References[132] Drummond, C. J.; Elliott, P.; Furlong, D.; Barnes, G. T. Thin Solid Films 1992, 210-211, Part 1, 69– 72.[133] Cruz, A.; Worthman, L.-A.; Serrano, A. G.; Casals, C.; Keough, K. M. W.; Pérez-Gil, J. EuropeanBiophysics Journal 2000, 29, 204.[134] Miñones, J., J.; Conde, O.; Iribarnegaray, E.; Casas, M.; Dynarowicz-£a¯tka, P. Trends in Colloidand Interface Science XVII; Progress in Colloid and Polymer Science; Springer Berlin Heidelberg,2004; Vol. 126; pp 55–59.[135] Bois, A. G.; Ivanova, M. G.; Panaiotov, I. I. Langmuir 1987, 3, 215–217.[136] Bikerman, J. Kolloid-Zeitschrift und Zeitschrift für Polymere 1963, 191, 33–35.[137] Gaines, G. L. Insoluble monolayers at liquid-gas interfaces; Interscience monographs on physicalchemistry; Interscience Publishers: (New York), 1966; pp xiv, 386.[138] Jalal, I.; Zografi, G. J. Colloid Interface Sci. 1979, 68, 196–198.[139] Adamson, A. W. Physical Chemistry of Surfaces; John Wiley and Sons: New York, 1990; pp1–777.[140] Smith, E. C.; Crane, J. M.; Laderas, T. G.; Hall, S. B. Biophys J 2003, 85, 3048–3057.[141] Smith, R. D.; Berg, J. C. Journal of Colloid and Interface Science 1980, 74, 273 – 286.[142] Bois, A.; Albon, N. Journal of Colloid and Interface Science 1985, 104, 579 – 582.[143] Rosoff, M. Vesicles; Surfactant Science; Taylor & Francis, 1996.[144] Lawrie, G.; Barnes, G. Journal of Colloid and Interface Science 1994, 162, 36 – 44.[145] Hoenig, D.; Moebius, D. The Journal of Physical Chemistry 1991, 95, 4590–4592.[146] Nag, K.; Keough, K. Biophysical Journal 1993, 65, 1019–1026.[147] de Almeida, R. F.; Loura, L. M.; Prieto, M. Chem. Phys. Lipids 2009, 157, 61 – 77.[148] Nag, K.; Fritzen-Garcia, M.; Devraj, R.; Panda, A. K. Langmuir 2007, 23, 4421–4431.[149] Edidin, M. Annu. Rev. Biophys. Biomol. Struct. 2003, 32, 257–283.[150] Motschmann, H.; Möhwald, H. In Handbook of Applied Surface and Colloid Chemistry; Holm-berg, K., Ed.; John Wiley & Sons, 2002; Vol. 1-2; Chapter 5, pp 79–98.167References[151] McConnell, H. M.; Tamm, L. K.; Weis, R. M. Proceedings of the National Academy of Sciences ofthe United States of America 1984, 81, 3249–3253.[152] Leporatti, S.; Brezesinski, G.; Mőhwald, H. Colloids and Surfaces A: Physicochemical and Engi-neering Aspects 2000, 161, 159–171.[153] Spratte, K.; Riegler, H. Langmuir 1994, 10, 3161–3173.[154] Spratte, K.; Chi, L. F.; Riegler, H. EPL (Europhysics Letters) 1994, 25, 211–217.[155] Sanchez, J.; Badia, A. Thin Solid Films 2003, 440, 223 – 239.[156] Shih, M. C.; Peng, J. B.; Huang, K. G.; Dutta, P. Langmuir 1993, 9, 776–778.[157] Moraille, P.; Badia, A. Langmuir 2003, 19, 8041–8049.[158] Bizzotto, D.; Shepherd, J. L. Advances in Electrochemical Science and Engineering; Wiley-VCHVerlag GmbH, 2008; pp 97–126.[159] Shepherd, J. L.; Bizzotto, D. The Journal of Physical Chemistry B 2003, 107, 8524–8531.[160] Shepherd, J. L.; Bizzotto, D. Langmuir 2006, 22, 4869–4876.[161] Burgess, I.; Zamlynny, V.; Szymanski, G.; Schwan, A.; Faragher, R.; Lipkowski, J.; Majewski, J.;Satija, S. Journal of Electroanalytical Chemistry 2003, 550-551, 187–199.[162] Burgess, I.; Li, M.; Horswell, S.; Szymanski, G.; Lipkowski, J.; Majewski, J.; Satija, S. BiophysicalJournal 2004, 86, 1763–1776.[163] Musgrove, A.; Kell, A.; Bizzotto, D. Langmuir 2008, 24, 7881–7888.[164] Bergstrom, F.; Mikhalyov, I.; Hagglof, P.; Wortmann, R.; Ny, T.; Johansson, L. B. JACS 2002, 124,196–204.[165] Parker, J. A. Stack Alignment (Align3_TP). http://www.med.harvard.edu/JPNM/ij/plugins/Align3TP.html.[166] Bizzotto, D.; Noël, J. J.; Lipkowski, J. J. Electroanal. Chem. 1994, 369, 259–265.[167] Bizzotto, D.; Lipkowski, J. Prog. Colloid Polym. Sci. 1997, 103, 201–215.[168] Ries, H.; Swift, H. Langmuir 1987, 3, 853.[169] Hu, Y.; Meleson, K.; Israelachvili, J. Biophysical Journal 2006, 91, 444 – 453.168References[170] Casanova-Moreno, J.; Bizzotto, D. Journal of Electroanalytical Chemistry 2010, 649, 126 – 135.[171] Rice, P. A.; McConnell, H. M. Proceedings of the National Academy of Sciences of the UnitedStates of America 1989, 86, 6445–6448.[172] Eftaiha, A. F.; Brunet, S. M.; Paige, M. F. Journal of Colloid and Interface Science 2012, 368, 356– 365.[173] Yun, S.; Ahn, K.; Kim, M. W. EPL (Europhysics Letters) 2005, 70, 555–561.[174] Chen, Y.; Munechika, K.; Ginger, D. S. Nano Lett. 2007, 7, 690–696.[175] Vos, J. G.; Forster, R. J.; Keyes, T. E. Interfacial Supramolecular Assemblies; John Wiley & Sons,Ltd, 2003; pp 87–152.[176] Ulman, A. Chemical Reviews 1996, 96, 1533–1554.[177] Nuzzo, R. G.; Allara, D. L. Journal of the American Chemical Society 1983, 105, 4481–4483.[178] Yuan, M.; Zhan, S.; Zhou, X.; Liu, Y.; Feng, L.; Lin, Y.; Zhang, Z.; Hu, J. Langmuir 2008, 24,8707–8710.[179] Yang, Z.; Gonzalez-Cortes, A.; Jourquin, G.; Viré, J.-C.; Kauffmann, J.-M.; Delplancke, J.-L.Biosensors and Bioelectronics 1995, 10, 789 – 795.[180] Collman, J. P.; Hosseini, A.; Eberspacher, T. A.; Chidsey, C. E. D. Langmuir 2009, 25, 6517–6521,PMID: 19379005.[181] Hoogvliet, J. C.; Dijksma, M.; Kamp, B.; van Bennekom, W. P. Analytical Chemistry 2000, 72,2016–2021.[182] Peterson, A. W.; Heaton, R. J.; Georgiadis, R. M. Nucleic Acids Research 2001, 29, 5163–5168.[183] Rzeźnicka, I. I.; Lee, J.; Maksymovych, P.; Yates, J. T. The Journal of Physical Chemistry B 2005,109, 15992–15996.[184] Weisshaar, D. E.; Lamp, B. D.; Porter, M. D. Journal of the American Chemical Society 1992, 114,5860–5862.[185] Ma, F.; Lennox, R. B. Langmuir 2000, 16, 6188–6190.[186] Wang, J.; Jiang, M.; Kawde, A. M.; Polsky, R. Langmuir 2000, 16, 9687–9689.169References[187] Chechik, V.; Crooks, R. M.; Stirling, C. J. M. Advanced Materials 2000, 12, 1161–1171.[188] Yang, D.; Al-Maznai, H.; Morin, M. The Journal of Physical Chemistry B 1997, 101, 1158–1166.[189] Widrig, C. A.; Chung, C.; Porter, M. D. Journal of Electroanalytical Chemistry 1991, 310, 335–359.[190] Xu, J.; Li, H.-L. Journal of Colloid and Interface Science 1995, 176, 138 – 149.[191] Finklea, H. O. Electroanalytical Chemistry: A Series of Advances, Vol 19 1996, 19, 109–335.[192] Sun, K.; Jiang, B.; Jiang, X. Journal of Electroanalytical Chemistry 2011, 656, 223 – 230.[193] Schneider, T.W.; Buttry, D. A. Journal of the AmericanChemical Society 1993, 115, 12391–12397.[194] Doneux, T.; Steichen, M.; Rache, A. D.; Buess-Herman, C. Journal of Electroanalytical Chemistry2010, 649, 164 – 170.[195] Yang, D.-F.; Wilde, C. P.; Morin, M. Langmuir 1996, 12, 6570–6577.[196] Doneux, T.; Nichols, R. J.; Buess-Herman, C. Journal of Electroanalytical Chemistry 2008, 621,267–276.[197] Ramírez, P.; Andreu, R.; Calvente, J. J.; Calzado, C. J.; López-Pérez, G. Journal of Electroana-lytical Chemistry 2005, 582, 179 – 190.[198] Doneux, T.; Steichen, M.; Bouchta, T.; Buess-Herman, C. Journal of Electroanalytical Chemistry2007, 599, 241 – 248.[199] Kim, Y.; Kim, H. J.; Lee, M. H.; Kang, Y.; Yang, Y.; Kim, H.; Kim, J. S. Chem. Commun. 2010, 46,8448–8450.[200] Orive, A. G.; Grumelli, D.; Vericat, C.; Ramallo-Lopez, J. M.; Giovanetti, L.; Benitez, G.; Az-carate, J. C.; Corthey, G.; Fonticelli, M. H.; Requejo, F. G.; Creus, A. H.; Salvarezza, R. C.Nanoscale 2011, 3, 1708–1716.[201] Mali, P.; Bhattacharjee, N.; Searson, P. C. Nano Letters 2006, 6, 1250–1253.[202] Schreiber, F. Progress in Surface Science 2000, 65, 151 – 257.[203] Vericat, C.; Andreasen, G.; Vela, M. E.; Martin, H.; Salvarezza, R. C. The Journal of ChemicalPhysics 2001, 115, 6672–6678.170References[204] Vericat, C.; Andreasen, G.; Vela, M. E.; Salvarezza, R. C. The Journal of Physical Chemistry B2000, 104, 302–307.[205] Zhong, C.-J.; Porter, M. D. Journal of Electroanalytical Chemistry 1997, 425, 147 – 153.[206] Smith, R. K.; Lewis, P. A.; Weiss, P. S. Progress in Surface Science 2004, 75, 1 – 68.[207] Lemay, D. M.; Shepherd, J. L. Electrochimica Acta 2008, 54, 388 – 393.[208] Hobara, D.; Miyake, K.; Imabayashi, S.-i.; Niki, K.; Kakiuchi, T. Langmuir 1998, 14, 3590–3596.[209] Yang, D.-F.; Wilde, C. P.; Morin, M. Langmuir 1997, 13, 243–249.[210] Pesika, N. S.; Stebe, K. J.; Searson, P. C. Langmuir 2006, 22, 3474–3476.[211] Laredo, T.; Leitch, J.; Chen, M.; Burgess, I. J.; Dutcher, J. R.; Lipkowski, J. Langmuir 2007, 23,6205–6211.[212] Bizzotto, D.; Lipkowski, J. Progress in Surface Science 1995, 50, 237–246.[213] Cai, X.; Baldelli, S. The Journal of Physical Chemistry C 2011, 115, 19178–19189.[214] Vericat, C.; Vela, M. E.; Benitez, G.; Carro, P.; Salvarezza, R. C. Chem. Soc. Rev. 2010, 39,1805–1834.[215] Arce, F. T.; Vela, M. E.; Salvarezza, R. C.; Arvia, A. J. The Journal of Chemical Physics 1998,109, 5703–5706.[216] Vela, M. E.; Martin, H.; Vericat, C.; Andreasen, G.; Hern ández Creus, A.; Salvarezza, R. C. TheJournal of Physical Chemistry B 2000, 104, 11878–11882.[217] Hatchett, D. W.; Uibel, R. H.; Stevenson, K. J.; Harris, J. M.; White, H. S. Journal of the AmericanChemical Society 1998, 120, 1062–1069.[218] Murphy, J. N.; Cheng, A. K. H.; Yu, H.-Z.; Bizzotto, D. Journal of the American Chemical Society2009, 131, 4042–4050, PMID: 19254024.[219] Ghaly, T.; Wildt, B. E.; Searson, P. C. Langmuir 2010, 26, 1420–1423.[220] Casanova-Moreno, J. R.; Bizzotto, D. Langmuir 2013, 29, 2065–2074.[221] Forster, B.; Van De Ville, D.; Berent, J.; Sage, D.; Unser, M. Microscopy Research and Technique2004, 65, 33–42.171References[222] École Polytechnique Fédérale de Lausanne, BIG > Extended Depth of Field.http://bigwww.epfl.ch/demo/edf/, 2011; http://bigwww.epfl.ch/demo/edf/.[223] Newman, J. Journal of The Electrochemical Society 1966, 113, 501–502.[224] Li, K. The image stabilizer plugin for ImageJ. http://www.cs.cmu.edu/~kangli/code/Image_Stabilizer.html, 2008.[225] Tschumperle, D.; Deriche, R. Pattern Analysis and Machine Intelligence, IEEE Transactions on2005, 27, 506 –517.[226] Savitzky, A.; Golay, M. J. E. Analytical Chemistry 1964, 36, 1627–1639.[227] Press, W. Numerical Recipes 3rd Edition: The Art of Scientific Computing; Cambridge UniversityPress, 2007.[228] Baur, J. E.; Miller, H. M.; Ritchason, M. A. Analytica Chimica Acta 1999, 397, 123 – 133.[229] Engstrom, R. C.; Wightman, R. M.; Kristensen, E. W. Analytical Chemistry 1988, 60, 652–656.[230] Halliday, D.; Resnick, R.; Walker, J. Fundamentals of Physics Extended, 10th Edition; Wiley GlobalEducation, 2013.[231] Baranski, A. S.; Boika, A. Anal. Chem. 2012, 84, 1353–1359.[232] Crozier, T. E.; Yamamoto, S. Journal of Chemical & Engineering Data 1974, 19, 242–244.[233] Bignell, N. J. Phys. Chem. 1987, 91, 1687–1690.[234] Duncan, P. B.; Needham, D. Langmuir 2004, 20, 2567–2578, PMID: 15835125.[235] Tyrrell, J. W. G.; Attard, P. Physical Review Letters 2001, 87, 176104.[236] Craig, V. S. J. Soft Matter 2011, 7, 40–48.[237] Boika, A.; Baranski, A. S. Anal. Chem. 2008, 80, 7392–7400.[238] Williams, J. A.; Gorman, C. B. The Journal of Physical Chemistry C 2007, 111, 12804–12810.[239] Piskarev, I.; Ushkanov, V.; Aristova, N.; Likhachev, P.; Myslivets, T. Biophysics 2010, 55, 13–17.[240] Thévenot, D. R.; Toth, K.; Durst, R. A.; Wilson, G. S. Biosensors and Bioelectronics 2001, 16,121–131.172References[241] Watson, J. D.; Baker, T. A.; Bell, S. P.; Gann, A.; Levine, M.; Losick, R. Molecular biology of thegene, 5th ed.; Pearson/Benjamin Cummings: San Francisco, 2004.[242] Neidle, S. Principles of Nucleic Acid Structure; Academic Press: New York, 2008; pp 20 – 37.[243] Chargaff, E.; Zamenhof, S.; Green, C. Nature 1950, 165, 756–757.[244] Watson, J. D.; Crick, F. H. C. Nature 1953, 171, 737–738.[245] Tinland, B.; Pluen, A.; Sturm, J.; Weill, G. Macromolecules 1997, 30, 5763–5765.[246] Baumann, C. G.; Smith, S. B.; Bloomfield, V. A.; Bustamante, C. Proceedings of the NationalAcademy of Sciences 1997, 94, 6185–6190.[247] Brinkers, S.; Dietrich, H. R. C.; de Groote, F. H.; Young, I. T.; Rieger, B. The Journal of ChemicalPhysics 2009, 130, 215105.[248] Sambrook, J.; Russell, D. Molecular cloning: a laboratory manual, 3rd ed.; Molecular Cloning: ALaboratory Manual; Cold Spring Harbor Laboratory Press, 2001; Vol. 3.[249] Drummond, T. G.; Hill, M. G.; Barton, J. K. Nat Biotech 2003, 21, 1192–1199.[250] Corlan, A. D. Medline trend: automated yearly statistics of PubMed results for any query. 2004;http://dan.corlan.net/medline-trend.html, Accessed: 2013-06-14.[251] Sam, M.; Boon, E. M.; Barton, J. K.; Hill, M. G.; Spain, E. M. Langmuir 2001, 17, 5727–5730.[252] Zhao, Y.-D.; Pang, D.-W.; Hu, S.; Wang, Z.-L.; Cheng, J.-K.; Dai, H.-P. Talanta 1999, 49, 751 –756.[253] Pei, H.; Zuo, X.; Pan, D.; Shi, J.; Huang, Q.; Fan, C. NPG Asia Mater 2013, 5, e51–.[254] Fan, C.; Plaxco, K. W.; Heeger, A. J. Proceedings of the National Academy of Sciences 2003,100, 9134–9137.[255] Xiao, Y.; Lai, R. Y.; Plaxco, K. W. Nat. Protocols 2007, 2, 2875–2880.[256] Bochman, M. L.; Paeschke, K.; Zakian, V. A. Nat Rev Genet 2012, 13, 770–780.[257] Paleček, E.; Boublíková, P.; Jelen, F. Analytica Chimica Acta 1986, 187, 99 – 107.[258] Paleček, E. Electroanalysis 1996, 8, 7–14.173References[259] Ferapontova, E. E. Electrochimica Acta 2004, 49, 1751 – 1759.[260] Egli, M.; Flavell, A.; Pyle, A. M.; Wilson, W. D.; Haq, S. I.; Luisi, B.; Fisher, J.; Laughton, C.;Allen, S.; Engels, J. In Nucleic Acids in Chemistry and Biology; Blackburn, G. M., Gait, M. J.,Loakes, D., Williams, D. M., Eds.; The Royal Society of Chemistry, 2006; pp i–470.[261] Armitage, B. A. In DNA Binders and Related Subjects; Waring, M., Chaires, J., Eds.; Topics inCurrent Chemistry; Springer Berlin Heidelberg, 2005; Vol. 253; pp 55–76.[262] Hashimoto, K.; Ito, K.; Ishimori, Y. Analytical Chemistry 1994, 66, 3830–3833.[263] Sun, X.; He, P.; Liu, S.; Ye, J.; Fang, Y. Talanta 1998, 47, 487 – 495.[264] Millan, K. M.; Mikkelsen, S. R. Analytical Chemistry 1993, 65, 2317–2323.[265] Uzawa, T.; Cheng, R. R.; White, R. J.; Makarov, D. E.; Plaxco, K. W. Journal of the AmericanChemical Society 2010, 132, 16120–16126.[266] Du, H.; Disney, M. D.; Miller, B. L.; Krauss, T. D. Journal of the American Chemical Society 2003,125, 4012–4013.[267] Du, H.; Strohsahl, C. M.; Camera, J.; Miller, B. L.; Krauss, T. D. Journal of the American ChemicalSociety 2005, 127, 7932–7940.[268] Rant, U.; Arinaga, K.; Scherer, S.; Pringsheim, E.; Fujita, S.; Yokoyama, N.; Tornow, M.; Abstre-iter, G. Proceedings of the National Academy of Sciences 2007, 104, 17364–17369.[269] Kaiser, W.; Rant, U. Journal of the American Chemical Society 2010, 132, 7935–7945.[270] Bixon, M.; Giese, B.; Wessely, S.; Langenbacher, T.; Michel-Beyerle, M. E.; Jortner, J. Proceed-ings of the National Academy of Sciences 1999, 96, 11713–11716.[271] Rant, U.; Arinaga, K.; Fujita, S.; Yokoyama, N.; Abstreiter, G.; Tornow, M. Langmuir 2004, 20,10086–10092, PMID: 15518498.[272] Rant, U.; Arinaga, K.; Tornow, M.; Kim, Y. W.; Netz, R. R.; Fujita, S.; Yokoyama, N.; Abstreiter, G.Biophysical Journal 2006, 90, 3666 – 3671.[273] Rant, U.; Pringsheim, E.; Kaiser, W.; Arinaga, K.; Knezevic, J.; Tornow, M.; Fujita, S.;Yokoyama, N.; Abstreiter, G. Nano Letters 2009, 9, 1290–1295.174[274] Rant, U. Bioanalytical Reviews 2012, 4, 97–114.[275] Rant, U.; Kaiser, W.; Hampel, P. A.; Niemax, J.; Langer, A.; Knezevic, J. Apparatus and Methodfor Evaluating Characteristics of Target Molecules. 2012.[276] Fahlman, R. P.; Sen, D. Journal of the American Chemical Society 2002, 124, 4610–4616.[277] Iqbal, S. S.; Mayo, M. W.; Bruno, J. G.; Bronk, B. V.; Batt, C. A.; Chambers, J. P. Biosensors andBioelectronics 2000, 15, 549 – 578.[278] Kelley, S. O.; Barton, J. K.; Jackson, N. M.; McPherson, L. D.; Potter, A. B.; Spain, E. M.;Allen, M. J.; Hill, M. G. Langmuir 1998, 14, 6781–6784.[279] Arinaga, K.; Rant, U.; Knežević, J.; Pringsheim, E.; Tornow, M.; Fujita, S.; Abstreiter, G.;Yokoyama, N. Biosensors and Bioelectronics 2007, 23, 326–331.[280] Langer, A.; Hampel, P. A.; Kaiser, W.; Knezevic, J.; Welte, T.; Villa, V.; Maruyama, M.; Svejda, M.;Jähner, S.; Fischer, F.; Strasser, R.; Rant, U. Nat Commun 2013, 4, –.[281] Musgrove, A.; Bridges, C. R.; Sammis, G. M.; Bizzotto, D. Langmuir 2013, 29, 3347–3360.[282] Musgrove, A. Electrochemically controlled interaction of liposomes with a solid-supported octade-canol bilayer. Ph.D. thesis, University of British Columbia, 2013.[283] Manz, A.; Graber, N.; Widmer, H. Sensors and Actuators B: Chemical 1990, 1, 244 – 248.[284] Wang, J.; Aki, M.; Onoshima, D.; Arinaga, K.; Kaji, N.; Tokeshi, M.; Fujita, S.; Yokoyama, N.;Baba, Y. Biosensors and Bioelectronics 2014, 51, 280 – 285.175Appendix AList of Reagents UsedName Formula State ofmatterPurity Manufacturer2-amino-2-hydroxymethyl-propane-1,3-diol(Tris)NH2C(CH2OH)3 solid 100% UltrapurebioreagentJ.T. Bakerargon Ar gas >99.998% Praxair6-mercapto-1-hexanol HS(CH2)6OH liquid 99% Aldrichgold wire (∅=0.5 mm) Au solid 99.95% Goodfellowgold wire (∅=25 µm) Au solid 99.95% Alfa Aesargold wire (∅=25 µm) Au solid 99.99% World PrecisionInstrumentshydrochloric acid HCl liquid 37% Sigma Aldrichnitric acid HNO3 liquid ACS Fisher Scientific1-octadecanol C18H38O solid >99.5% (GC) Sigma-Aldrichpotassium hydroxyde KOH solid 99.99% Sigma-Aldrichpotassium nitrate KNO3 solid 99.7% (ACS) Fisherpotassium perchlorate KClO4 solid >99% (purissACS)Sigma-Aldrichsulfuric acid H2SO4 liquid ACS Sigma-AldrichTable A.1: List of the general reagents employed.176Appendix BFlat-field Image Correction ProcedureAll the images in Chapter 5 were flat-field corrected through the use of a bare Au brightfield image anda dark current image acquired with the light source off. Figure B.1 illustrates the procedure employedto this end. The main goal of this correction is to compensate for the uneven illumination of the lampwhich is especially important when the lamp is located directly behind the microscope and not coupledthrough an optic fiber.177Appendix B. Flat-field Image Correction Procedurebare Au brightfieldsubtractdivideGaussblur100 pxdark currentfluorescenceFigure B.1: Block diagram of the flatfield correction image procedure178Appendix CUncorrected and Contrast EnhancedImages Corresponding to Chapter 5The fluorescence images that are presented in Figure 5.7 are presented below in an unenhanced format(Fig C.1) and in a strongly contrast enhanced format (Fig C.2). The latter form was created using localcontrast enhancement (unsharp mask of 25 pixels and a weight of 0.6). This treatment does not allowfor a simple interpretation of the intensity, but rather is used for particle or feature size measurements.0 1.68 0 0.6Figure C.1: Unenhanced verison of Figure 5.7. Reprinted from [170] Copyright (2010), with permissionfrom Elsevier.Figure C.2: Contrast enhanced version of Figure 5.7. Reprinted from [170] Copyright (2010), with per-mission from Elsevier.The same set of image treatments (Gaussian blur of 4 pixels, unsharp mask of 250 pixels, weight-179Appendix C. Uncorrected and Contrast Enhanced Images Corresponding to Chapter 5ing of 0.6) result in unenhanced and contrast enhanced versions of Figure 5.9; Fig C.3 and Fig. C.4respectively.Figure C.3: Unenhanced version of Figure 5.9. Reprinted from [170] Copyright (2010), with permissionfrom Elsevier.Figure C.4: Contrast enhanced version of Figure 5.9. Reprinted from [170] Copyright (2010), with per-mission from Elsevier.180Appendix DDiffusion of Physically Adsorbed DNAfrom Modified ElectrodesIf the deposited MCH/DNA layers are analyzed immediately after being removed from the DNA solutionand rinsed with buffer and water, a fluorescence plume can be observed (Fig. D.1) moving from thecenter to the left of the image at potentials too high to be consistent to reductive desorption of the SAM.It is more likely that this fluorescence arises from physically adsorbed DNA that is released as the electricfield causes the tethered layer to change its configuration. At the most negative potentials an increase influorescence is observed due to the usual “switching” as described in Section 7.7.2. These two differentsources of increase in fluorescence can clearly be seen in a plot of fluorescence intensity as a functionof time (Fig. D.2).181Appendix D. Diffusion of Physically Adsorbed DNA from Modified Electrodes500 μm+0.35 V +0.25 V +0.15 V+0.05 V -0.05 V -0.15 V-0.25 V -0.35 V -0.45 VFigure D.1: Fluorescence images of an MCH/DNA layer during a step potential experiment as describedin Section 7.6.3. This layer was analyzed without the overnight soaking step in buffer.− (V vs. SCE)1100130015000 100 200 300 400Fluorescence intensityt (s)Figure D.2: Average fluorescence intensity (bottom panel) of the whole frame of the images shown inFig. D.1 upon the application of the potential perturbation shown in the top panel.182Appendix EComparison of Objective PerformanceIn order to compare the performance of dry (i.e. intended to be used with air as the medium betweenthe objective and the specimen) and water immersion objectives, a sample was prepared by depositingan MCH/DNA layer in a gold bead that had been polished flat. Brightfield and fluorescence images ofthe same area of the electrode are shown in Figure E.1. The increase in the image quality due to thebetter collection efficiency is notorious for the brightfield images, and outstanding in the fluorescenceones.50× dry objectiveBright€eldFluorescence40× water immersion objective113 214 467 21741037 1228 1960 354275 μmFigure E.1: Comparison of brightfield and fluorescence images of a MCH/DNA layer deposited on apolished gold electrode, using a 50× (NA = 0.5) dry objective and a 40× (NA = 0.8) water immersionobjective. Bars under the images show the minimum and maximum intensity values used in the lookuptable.183Appendix FMCH/DNA-Modified Bead Images atDifferent Focal Planes500 μmab250 μmFigure F.1: Selected images at different focal points used to construct Figures 7.9a and b.184


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



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"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items