Mechanisms in Surface Enhanced Raman Scattering by Matthias Meyer A thesis submitted to the Victoria University of Wellington in fullfilment of the requirements for the degree of Doctor of Philosophy in Physics Victoria University of Wellington The MacDiarmid Institute for Advanced Materials and Nanotechnology 2009 Abstract This thesis focusses on a number of topics in surface enhanced Raman scattering (SERS). The aim of the undertaken research was to deepen the general understanding of the SERS effect and, thereby, to clarify some of the disputed issues, among them: What is the origin of the enhancement? What is the physical or chemical effect of ‘salt activation’ in SERS systems? Can we observe single-molecules using SERS? Can we determine the ab- sorbate’s orientation on the surface? In part one (chapters 1-3), as a general introduction, I start with a short overview of the Raman effect and its relation to other molecular spectro- scopic effects (such as fluorescence, Rayleigh scattering, etc... ). Follow- ing these basic remarks, the surface enhancement mechanisms underlying SERS are explained (as a largely electromagnetic field enhancement) and are investigated theoretically on the canonical model of a nanoscopic dimer of silver spheres. The second part (chapter 4) reports on the experimental investigation (elec- tron microscopy, in-situ Raman measurements) of a typical real SERS sys- tem: Lee & Meisel silver colloids. An emphasis is put on the self-limiting aggregation kinetics which is observed in such systems after salt addi- tion. This is also investigated and rationalised by means of Monte-Carlo simulations which are footed on empiric theoretical considerations for the interaction potential. Part three (chapter 5) contains a discussion of the early attempts on single- molecule SERS and points out the shortcomings of the previously used ultra-low concentration approach. In response, an improved and more rig- orous approach is presented: Bi-Analyte SERS. Examplary applications of the technique are discussed. Within these experiments the capability of the technique to prove/disprove (with statistical soundness) single-molecule sensitivity in any SERS system is demonstrated, and single-molecule en- hancement factors are derived. The last part (chapter 6) presents computational studies based on density- functional theory and its use in the context of Raman spectroscopy and SERS. Of particular interest here were the Raman tensors, their visual representation appropriate in the SERS case, their relation to the relative intensities of Raman peaks, and their modification when the photon en- ergy approaches the electronic resonance of the molecule. Last, but not least, a conclusion chapter is presented, where I highlight what has contributed by the thesis to the general understanding of the SERS effect. Acknowledgement There was no time for scholarly details, and, besides, I have always believed that a man can fairly be judged by the standards and taste of his choices in matters of high-level plagiarism. Hunter S. Thompson It is no secret that a PhD thesis is not the product of a single person. It is much more the collaborative work of many, written up by a single author. Accordingly, this thesis was made possible only through the work and lasting support of many people. A few stand out: most notably, Pablo Etchegoin and Eric Le Ru, who provided a great working environment by being open-minded, patient, hard-working and yet constantly approach- able and helpful colleagues. The entire Raman group carries this spirit, and, in addition to Pablo and Eric, I wish to also express particular grat- itude to Andrew Preston, Chris Galloway, Evan Blackie, Joe Trodahl and Ben Ruck, who all made, in one way or another, significant contributions to this thesis. The list of reasons documenting why it is a good idea to do your PhD in this group is far too long to exhaust here. I also wish to thank, for their continuous support, the people of the MacDi- armid Institute and the School of Chemical and Physical Sciences: Margaret Brown, Paul Callaghan, Richard Blaikie, John Spencer, John Lekner, Alan Rennie, and many more (including entire soccer teams!), too numerous to fit into this small frame. 3 I am grateful to the financial support of the MacDiarmid Institute and Victoria University of Wellington, without which this work would have been impossible. Finally, I am forever indebted to my family and friends who I left behind in my hometown Berlin, for their encouragement, understanding and sac- rifice. Their unfathomable love and trust remains untouched, despite my arguably rare attempts to bridge the geographical distance via phone, or even email. Contents 1 Introduction 9 2 Molecular spectroscopy 13 2.1 Molecularstatesandtransitions. 13 2.2 TheRamaneffect ......................... 18 2.3 Experimentalsetupandmaterials. 28 3 Surface-enhancementmechanisms 35 3.1 Electromagneticsurfaceenhancement . 35 3.1.1 Enhancementfactors . 36 3.1.2 Dielectricfunctionofnoblemetals . 40 3.1.3 Solvingtheelectromagneticproblem. 46 3.1.4 Localfieldofaplanarsurface . 48 3.1.5 Localfieldofadimerofspheres . 50 3.2 Chemicalenhancement-controversy. 62 4 Self-limiting colloidal aggregation in SERS-active solutions 66 4.1 NoblemetalcolloidsinSERS . 66 5 4.1.1 Introduction........................ 67 4.1.2 Lee&Meiselsilvercolloids . 69 4.1.3 Somepuzzlingobservations. 72 4.2 DLVOtheory............................ 73 4.2.1 ScreenedCoulombrepulsion . 75 4.2.2 vanderWaalsattraction . 76 4.2.3 ApplicationtoLee&Meiselcolloids . 80 4.3 Electronmicroscopyofsinglecolloids . 86 4.3.1 Investigatedsamples. 88 4.3.2 Singlecolloidsanddimers. 88 4.3.3 Aggregates-smallandlargestructures . 89 4.3.4 Rod-likecolloidalstructures. 93 4.4 In-situ observation of colloidal aggregation . .. 95 4.4.1 Thebuild-upandorigin of SERSsignalsinliquid . 95 4.5 MonteCarlosimulations . 105 4.5.1 Formulatingthemany-bodyproblem . 105 4.6 Conclusion.............................115 5 Single-molecule SERS and Bi-Analyte SERS 116 5.1 EarlySM-SERSclaims . 116 5.1.1 Theultra-lowconcentrationapproach . 117 5.1.2 Poisson distribution - quantised intensities? . 120 5.2 Recent SM-SERS efforts......................124 5.3 Rigorous proof of SM-SERS using two analytes . 126 5.3.1 TheBi-AnalyteSERSidea . 126 5.3.2 Discretemodelhot-spots . 127 5.3.3 Prerequisitecriteria. 130 5.3.4 BiASERS-experiment . 131 5.3.5 Caveats ..........................140 5.3.6 BiASERSapplications . 143 5.4 Conclusion.............................150 6 Quantum chemistry 151 6.1 Introduction ............................152 6.2 Physicalbackground . 153 6.3 Computationalmodels . 156 6.4 ApplicationsofDFT . 157 6.5 ObtainingtheRamantensors . 164 6.6 Ramantensorvisualisation . 165 6.7 Ramantensorsofbenzotriazole . 170 6.8 Connectiontomolecularorientation . 176 6.9 Movingtowardsresonance . 179 6.10 Conclusion.............................186 7 Conclusions 187 A Molecular footprints 191 B BiASERS - from Pascal’s triangle to Mie theory 195 C Artwork 201 D Publications 210 Chapter 1 Introduction I don’t know anything, but I do know that everything is interesting if you go into it deeply enough. Richard Feynman I started the work for this thesis in early 2005, in a newly formed group1 which had joined the field roughly one year earlier. After an initial phase of learning the physical foundations of the Raman effect, and after get- ting accustomed to the experimental realities of surface-enhanced Raman spectrocopy (SERS), the primary research goal was as simple as profound: develop and improve the understanding of the SERS effect. This research undertaken to achieve this goal was mostly, but not solely, experimental in nature. A variety of techniques were employed: optical spectroscopy and microscopy, electron microscopy; also, suitable statistical data analysis software was developed and computational methods were used to predict and understand results. This broad range is reflected in the topics covered throughout the chapters of this thesis, justifying the rather non-specific title “Mechanisms in Surface-Enhanced Raman Spectroscopy”. 1At this point, the initial group consisted solely of Pablo Etchegoin and Eric Le Ru. 9 CHAPTER 1. INTRODUCTION Motivation SERS is a powerful tool for physics, chemistry, biochemistry and related fields. There are two key reasons for this: (i) SERS inherits its high molec- ular specificity from the Raman effect [1, 2], which allows detailed analytic chemistry with the ability to distinguish miniscule (even isotopic) changes in molecules and (ii) SERS provides large enhancements of the Raman sig- nal, thereby compensating for the small cross section, i.e. low intensity, of Raman scattering. The Raman effect is understood [3], but the signal en- hancement provided by a complex mixture of metal nanoparticles with dye and salt raises a whole plethora of questions and problems to investigate. Clearly, in retrospective, it is always easy to formulate the questions which correspond to the findings and insights gained through one’s work. How- ever, a number of general questions emerged early, prior to any serious investigation, and mostly pertaining to the real, experimental system: • What do the colloids actually look like? • Which effects constitute the enhancement in SERS? • Why do SERS samples require activation using salt? • Is single molecule SERS a reality? These basic questions provided the initial momentum for this work and ultimately were the driving force leading to many different, sometimes unexpected, results. This thesis addresses these questions