Open Collections

UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Design of quantum dot and smartphone-based luminescent bioassay platforms for prospective point-of-care… Petryayeva, Eleonora 2016

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

Item Metadata

Download

Media
24-ubc_2017_february_petryayeva_eleonora.pdf [ 91.11MB ]
Metadata
JSON: 24-1.0340478.json
JSON-LD: 24-1.0340478-ld.json
RDF/XML (Pretty): 24-1.0340478-rdf.xml
RDF/JSON: 24-1.0340478-rdf.json
Turtle: 24-1.0340478-turtle.txt
N-Triples: 24-1.0340478-rdf-ntriples.txt
Original Record: 24-1.0340478-source.json
Full Text
24-1.0340478-fulltext.txt
Citation
24-1.0340478.ris

Full Text

Design of Quantum Dot and Smartphone-based                                        Luminescent Bioassay Platforms for Prospective                Point-of-Care Diagnostics by ELEONORA PETRYAYEVA Hon.B.Sc., The University of Toronto, 2011 M.Sc., The University of Toronto, 2012 A THESIS SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE AND POSTDOCTORAL STUDIES (Chemistry) THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver) December 2016 © Eleonora Petryayeva 2016  ii  Abstract Smartphones are essential components of daily life. These devices feature built-in cameras and light sources, data storage, and wireless data transmission, making them emerging devices for optical imaging and diagnostic bioassays. To date, the majority of smartphone-based diagnostics have been developed for colourimetric assays, which often suffer from limited multiplexing capability and poor sensitivity. In general, fluorescence-based assays offer greater sensitivity and multiplexing capacity, and in combination with smartphone platform may help to overcome these limitations. This thesis describes research toward the development of smartphone platforms for fluorescence-based bioassays using quantum dots (QDs) and Förster resonance energy transfer (FRET), and addresses two critical challenges: multiplexing and analysis of biological sample matrices. Multiplexing was achieved by matching the built-in RGB (red-green-blue) channels of smartphone cameras with the narrow, bright, and tunable emission of QDs. The QDs provided superior brightness in comparison to traditional fluorescent dyes and proteins, and served as excellent FRET donors in assays that used proteases as model analytes. Up to three-plex assays were demonstrated for the detection of trypsin, chymotrypsin, and enterokinase. The analytical performance of the smartphone-based platform matched that of a bench-top spectrofluorimeter, where the smartphone was a fraction of the cost and size. A smartphone-based platform was also developed for detection of analytes in serum and whole blood. Most clinical samples will take this form and necessitate careful assay design to overcome challenges associated with physical, optical and chemical properties of whole blood. Blood is strongly absorbing, scattering, autofluorescent, and contains high concentrations of proteins and small molecules. A well-thought-out combination of QDs, FRET, and a paper-in-PDMS chip enabled direct, single-step and quantitative fluorescence-based detection of thrombin activity in whole  iii  blood. The research in this thesis is a foundation for the development of novel point-of-care diagnostics assays with consumer electronics that could help enable personalized health care.  iv  Preface Chapter 1 is an adaptation of two published works. Section 1.1 is reproduced in part from Petryayeva, E.; Algar, W. R., Toward point-of-care diagnostics with consumer electronic devices: the expanding role of nanoparticles. RSC Adv. 2015, 5 (28), 22256-22282 with permission from The Royal Society of Chemistry (Copyright 2015 The Royal Society of Chemistry). Sections 1.3.1–1.3.5 are reproduced in part from Petryayeva, E.; Algar, W. R.; Medintz, I. L., Quantum Dots in Bioanalysis: A Review of Applications Across Various Platforms for Fluorescence Spectroscopy and Imaging. Appl. Spectrosc. 2013, 67 (3), 215-252 with permission from the Society of Applied Spectroscopy (Copyright under Creative Commons Attribution - NonCommercial 4.0 International licence). These review articles were written by Eleonora Petryayeva with input and editing from Dr. Algar. Sections 1.2 and 1.3.6 contain unpublished work. Chapter 2 is an adaptation of published work, and is reproduced from Petryayeva, E.; Algar, W. R., A Job for Quantum Dots: Use of a Smartphone and 3D-Printed Accessory for All-In-One Excitation and Imaging of Photoluminescence. Anal. Bioanal. Chem. 2016, 408 (11), 2913-2925, with permission of Springer (Copyright 2016 Springer). Eleonora Petryayeva and Dr. Algar conceived the research. Eleonora Petryayeva designed, completed, and analyzed the experiments and data presented. Pritesh Padhiar in the UBC Department of Chemistry Mechanical Engineering Shop assisted with 3D printing based on the device designs created by Eleonora Petryayeva. The manuscript was co-written by Eleonora Petryayeva and Dr. Algar. Chapter 3 is an adaptation of published work, and is reproduced from Petryayeva, E.; Algar, W. R., Multiplexed Homogeneous Assays of Proteolytic Activity Using a Smartphone and Quantum Dots. Analytical Chemistry 2014, 86, 3195-3202, with permission from The American Chemical Society (Copyright 2014 The American Chemical Society). Eleonora Petryayeva and Dr. Algar conceived the research. Eleonora Petryayeva designed, completed, and analyzed the experiments and data presented. Eleonora Petryayeva wrote the manuscript with input and editing from Dr. Algar.  v  Chapter 4 contains currently unpublished data, with the exception of Section 4.2.4.1.1, which is an adaptation of published work, and is reproduced from Kim, H., Petryayeva, E., Algar, W. R., Enhancement of Quantum Dot Forster Resonance Energy Transfer within Paper Matrices and Application to Proteolytic Assays. IEEE J. Sel. Top. Quantum Electron. 2014, 20 (3), 7300211, with permission from IEEE (Copyright 2014 IEEE). Eleonora Petryayeva and Dr. Algar conceived the research. H. Kim was an undergraduate summer student supervised by Dr. Algar and Eleonora Petryayeva, and, using methods and protocols developed by Eleonora Petryayeva, he prepared the samples used for fluorescence lifetime imaging (FLIM) that appear in this thesis. Olga Solodova was an undergraduate student supervised by Dr. Algar and Eleonora Petryayeva, and she contributed to experiments designed to determine the density of functional groups in cellulose paper samples. These results appear in this thesis in Section 4.2.2.1. Dr. Gethin Owen at the Centre for High-Throughput Phenogenomics, UBC, performed the SEM imaging reported in Section 4.2.3. Dr. Ken Wong at the Interfacial Analysis & Reactivity Laboratory, UBC, performed the XPS analysis of paper samples that appear in Section 4.2.2. For all the remaining sections in this chapter, Eleonora Petryayeva designed, completed, and analyzed the experiments and data presented. Chapter 5 is an adaptation of published work, and is reproduced from Petryayeva, E.; Algar, W. R., Proteolytic Assays on Quantum Dot-Modified Paper Substrates Using Simple Optical Readout Platforms. Anal. Chem. 2013, 85 (18), 8817-8825, with permission from the American Chemical Society (Copyright 2013 The American Chemical Society). Eleonora Petryayeva and Dr. Algar conceived the research. Eleonora Petryayeva designed, completed, and analyzed the experiments and data presented. The manuscript was written by Eleonora Petryayeva, with input and editing from Dr. Algar.  Chapter 6 is an adaptation of published work, and is reproduced from Petryayeva, E.; Algar, W. R., Single-step bioassays in serum and whole blood with a smartphone, quantum dots and paper-in-PDMS chips. Analyst 2015, 140, 4037-4045, with permission from The Royal Society of Chemistry (Copyright 2015 The Royal Society of Chemistry). Eleonora Petryayeva and Dr. Algar conceived the research. Eleonora Petryayeva designed, completed, and analyzed the experiments and data presented. Eleonora Petryayeva wrote the manuscript with input and editing from Dr. Algar.  vi  Table of Contents Abstract	..................................................................................................................................	ii	Preface	...................................................................................................................................	iv	Table	of	Contents	...................................................................................................................	vi	List	of	Tables	........................................................................................................................	xiii	List	of	Figures	........................................................................................................................	xv	List	of	Schemes	...................................................................................................................	xxxi	List	of	Abbreviations	..........................................................................................................	xxxii	Acknowledgements	...........................................................................................................	xxxiv	Chapter	1	 Introduction	........................................................................................................	1	1.1	 Point-of-Care	Diagnostics	........................................................................................................	2	1.1.1	 Consumer	electronic	devices	for	POC	.....................................................................................	2	1.1.2	 The	need	for	point-of-care	diagnostics	...................................................................................	3	1.1.3	 Clinical	tests	and	biomarkers	..................................................................................................	5	1.1.4	 Lateral	flow	assays	..................................................................................................................	8	1.1.5	 Utility	of	consumer	electronic	devices	..................................................................................	10	1.1.6	 Optical	properties	of	nanoparticles	......................................................................................	11	1.1.7	 Bioassays	with	consumer	electronics	and	nanoparticles	......................................................	16	1.1.7.1	 Light	sources	.................................................................................................................................	16	1.1.7.2	 CMOS	image	sensors:	digital	cameras	to	smartphones	................................................................	18	1.1.7.2.1	 Digital	imaging	technology	....................................................................................................	18	1.1.7.2.2	 Growing	analytical	applications	............................................................................................	20	1.2	 Fluorescence	.........................................................................................................................	24	1.2.1	 The	Perrin–Jablonski	diagram	...............................................................................................	24	1.2.2	 Absorption–formation	of	the	electronic	excited	state	..........................................................	26	1.2.2.1	 Absorption	spectrum:	shape	and	intensity	...................................................................................	28	1.2.3	 The	Franck-Condon	electronic	excited	state	.........................................................................	31	1.2.4	 Quantum	yield	and	fluorescence	lifetime	.............................................................................	33	 vii  1.2.5	 Quenching	processes	............................................................................................................	35	1.2.6	 Factors	affecting	fluorescence	..............................................................................................	36	1.2.7	 Fluorescence	measurements	................................................................................................	37	1.2.7.1	 Fluorescence	microscopy	and	imaging	.........................................................................................	39	1.2.8	 Förster	resonance	energy	transfer	........................................................................................	39	1.2.8.1	 Classical	Förster	formalism	...........................................................................................................	41	1.2.8.2	 Measurement	of	FRET	efficiency	..................................................................................................	45	1.2.8.3	 Assumptions	underlying	use	of	FRET	............................................................................................	47	1.3	 Semiconductor	Quantum	Dots	...............................................................................................	49	1.3.1	 What	is	a	quantum	dot?	........................................................................................................	49	1.3.2	 Absorption	and	photoluminescence	.....................................................................................	49	1.3.3	 Quantum	confinement	and	core/shell	structures	.................................................................	51	1.3.3.1	 Surface	states	and	effects	.............................................................................................................	53	1.3.4	 Quantum	dot	materials	.........................................................................................................	54	1.3.4.1	 Synthesis	of	QDs	...........................................................................................................................	56	1.3.5	 Functionalization	of	QDs	.......................................................................................................	56	1.3.5.1	 Interfacial	chemistry	.....................................................................................................................	56	1.3.5.2	 Bioconjugation	of	QDs	..................................................................................................................	60	1.3.6	 Quantum	dots	as	FRET	donors	..............................................................................................	62	1.3.6.1	 Selection	of	QD-dye	FRET	pairs	.....................................................................................................	69	1.4	 Contributions	of	This	Thesis	...................................................................................................	72	1.4.1	 Background	...........................................................................................................................	72	1.4.2	 Thesis	overview	.....................................................................................................................	74	Chapter	2	 Use	of	a	Smartphone	and	3-D	Printed	Accessory	for	All-in-One	Excitation	and	Imaging	of	Photoluminescence	..............................................................................................	78	2.1	 Introduction	..........................................................................................................................	78	2.2	 Results	...................................................................................................................................	80	2.2.1	 Design	of	the	3D-printed	accessory	......................................................................................	80	2.2.2	 Substrates	for	all-in-one	imaging	..........................................................................................	82	2.2.3	 Photoluminescent	materials	.................................................................................................	83	2.2.4	 Detection	of	photoluminescence	..........................................................................................	86	2.2.5	 Effect	of	smartphone	imaging	application	parameters	.........................................................	89	 viii  2.2.6	 Two-colour	imaging	...............................................................................................................	93	2.2.7	 Model	binding	assay	..............................................................................................................	96	2.2.8	 Model	FRET-based	thrombin	assay	.......................................................................................	97	2.2.8.1	 FRET	pairs	......................................................................................................................................	97	2.2.8.2	 Thrombin	assay	.............................................................................................................................	99	2.2.9	 Tandem		QD540a–A610	FRET	conjugate	.............................................................................	101	2.3	 Discussion	...........................................................................................................................	102	2.3.1	 Smartphone	imaging	platform	............................................................................................	102	2.3.2	 QDs	are	ideal	for	smartphone	imaging	...............................................................................	104	2.4	 Conclusion	...........................................................................................................................	108	2.5	 Experimental	Section	...........................................................................................................	109	2.5.1	 Materials	and	reagents	.......................................................................................................	109	2.5.2	 Preparation	of	GSH-coated	QDs	..........................................................................................	110	2.5.3	 3D-printed	apparatus	..........................................................................................................	110	2.5.4	 Binding	assays	.....................................................................................................................	111	2.5.5	 FRET	experiments	................................................................................................................	111	2.5.6	 Data	analysis	.......................................................................................................................	112	2.5.6.1	 Quantum	yields	...........................................................................................................................	112	2.5.6.2	 FRET	calculations	........................................................................................................................	112	Chapter	3	 Multiplexed	Homogeneous	Assays	of	Proteolytic	Activity	Using	a	Smartphone	and	Quantum	Dots	.....................................................................................................................	113	3.1	 Introduction	........................................................................................................................	113	3.2	 Results	.................................................................................................................................	116	3.2.1	 RGB	colour	imaging	of	quantum	dots	.................................................................................	116	3.2.2	 RGB	colour	imaging	of	mixtures	of	quantum	dots	..............................................................	118	3.2.3	 FRET	pairs	and	their	fidelity	................................................................................................	119	3.2.3.1	 Calibration	curves	.......................................................................................................................	123	3.2.4	 Validation	of	proteolytic	assays	with	digital	colour	imaging	...............................................	125	3.2.4.1	 Temporal	stability	of	UV	lamp	and	smartphone	camera	............................................................	125	3.2.4.2	 Red,	green,	and	blue	channel	proteolytic	assays	........................................................................	127	3.2.5	 Multiplexed	proteolytic	assays	with	digital	colour	imaging	................................................	132	3.3	 Discussion	...........................................................................................................................	136	 ix  3.4	 Conclusions	.........................................................................................................................	137	3.5	 Experimental	Section	...........................................................................................................	137	3.5.1	 Materials	and	reagents	.......................................................................................................	137	3.5.2	 Peptide	labeling	...................................................................................................................	138	3.5.3	 Instrumentation	and	data	acquisition	.................................................................................	139	3.5.4	 Conjugation	of	peptides	to	QDs	and	enzyme	assays	...........................................................	140	3.5.5	 Data	analysis	.......................................................................................................................	141	3.5.5.1	 FRET	parameters	.........................................................................................................................	141	3.5.5.2	 RGB	images	analysis	....................................................................................................................	141	3.5.5.2.1	 Normalization	of	progress	curves	.......................................................................................	141	3.5.5.2.2	 Calculation	of	initial	rates	and	specificity	constants	...........................................................	142	Chapter	4	 Characterization	and	Evaluation	of	Chemical	Modification	of	Cellulose	Paper	Substrates	for	Prospective	Point-of-care	Diagnostics	with	Immobilized	Quantum	Dots	.......	143	4.1	 Introduction	........................................................................................................................	143	4.2	 Results	and	Discussion	.........................................................................................................	147	4.2.1	 Selection	of	surface	ligands	.................................................................................................	147	4.2.2	 Characterization	of	surface	chemistry	................................................................................	149	4.2.2.1	 Quantification	of	accessible	functional	groups	...........................................................................	153	4.2.3	 Immobilization	of	quantum	dots	.........................................................................................	158	4.2.3.1	 Photobleaching	...........................................................................................................................	166	4.2.3.2	 Long-term	stability	of	QD-paper	substrates	...............................................................................	167	4.2.4	 Immobilization	of	QD-peptide	bioconjugates	.....................................................................	168	4.2.4.1	 FRET	on	paper	substrates	...........................................................................................................	168	4.2.4.1.1	 Fluorescence	lifetime	imaging	microscopy	(FLIM)	..............................................................	173	4.2.5	 Effect	of	paper	surface	chemistry	on	proteolytic	rates	.......................................................	174	4.3	 Conclusions	.........................................................................................................................	175	4.4	 Experimental	Section	...........................................................................................................	176	4.4.1	 Materials	and	reagents	.......................................................................................................	176	4.4.2	 Preparation	of	paper	substrates	.........................................................................................	177	4.4.2.1	 Preparation	of	paper	substrate	(3)	.............................................................................................	177	4.4.2.2	 Preparation	of	paper	substrate	(4b)	...........................................................................................	178	4.4.2.3	 Synthesis	of	N-(2-aminoethyl)-5-(1,2-dithiolan-3-yl)pentanamide	............................................	180	4.4.2.4	 Preparation	of	paper	substrate	(6b)	...........................................................................................	181	 x  4.4.3	 Quantification	of	functional	groups	....................................................................................	182	4.4.3.1	 Aldehyde	groups	.........................................................................................................................	182	4.4.3.1	 Primary	amine	groups	.................................................................................................................	183	4.4.3.2	 Thiol	groups	................................................................................................................................	184	4.4.4	 Characterization	of	paper	substrates	..................................................................................	184	4.4.5	 Preparation	of	DHLA-coated	QDs	........................................................................................	184	4.4.5.1	 Synthesis	of	DHLA	.......................................................................................................................	184	4.4.5.2	 DHLA	QD	ligand	exchange	..........................................................................................................	185	4.4.6	 Immobilization	of	QDs	and	QD-peptide	conjugates	............................................................	185	4.4.7	 SEM	imaging	of	immobilized	QDs	.......................................................................................	186	4.4.8	 Photobleaching	experiments	..............................................................................................	186	4.4.9	 Fluorescence	lifetime	imaging	microscopy	(FLIM)	..............................................................	186	4.4.10	 Enzyme	assays	...................................................................................................................	187	4.4.11	 Data	analysis	.....................................................................................................................	187	Chapter	5	 Proteolytic	Assays	on	Quantum	Dot-Modified	Paper	Substrates	Using	Consumer	Digital	Cameras	for	Optical	Readout	....................................................................................	189	5.1	 Introduction	........................................................................................................................	189	5.2	 Results	and	Discussion	.........................................................................................................	193	5.2.1	 FRET	pair	and	LED	excitation	...............................................................................................	193	5.2.2	 Assay	design	and	characterization	......................................................................................	193	5.2.3	 Spectral	validation	of	protease	assays	on	paper	substrates	...............................................	195	5.2.4	 Proteolytic	assays	with	digital	colour	imaging	....................................................................	197	5.2.4.1	 Trypsin	and	chymotrypsin	paper-based	assays	..........................................................................	197	5.2.4.2	 Aprotinin	inhibition	assays	with	digital	colour	imaging	..............................................................	202	5.2.4.3	 Multiplexed	proteolytic	assays	with	digital	colour	imaging	........................................................	202	5.2.4.4	 Proteolytic	activation	assays	with	digital	colour	imaging	...........................................................	205	5.2.5	 Assays	with	a	webcam	and	smartphone	camera	................................................................	207	5.3	 Conclusions	.........................................................................................................................	208	5.4	 Experimental	Section	...........................................................................................................	209	5.4.1	 Materials	and	reagents	.......................................................................................................	209	5.4.2	 Preparation	of	paper	substrates	.........................................................................................	211	5.4.3	 Enzyme	assays	.....................................................................................................................	211	 xi  5.4.4	 Instrumentation	and	data	acquisition	.................................................................................	211	5.4.5	 Data	analysis	.......................................................................................................................	214	5.4.5.1	 FRET	parameters	.........................................................................................................................	214	5.4.5.2	 Analysis	of	spectral	data	.............................................................................................................	215	5.4.5.3	 Analysis	of	RGB	data	...................................................................................................................	215	5.4.5.4	 Normalization	of	progress	curves	...............................................................................................	215	5.4.5.5	 Calculation	of	normalized	initial	rates	........................................................................................	216	Chapter	6	 Single-Step	Bioassays	in	Serum	and	Whole	Blood	with	a	Smartphone,	Quantum	Dots	and	Paper-in-PDMS	Chips	............................................................................................	217	6.1	 Introduction	........................................................................................................................	217	6.2	 Results	.................................................................................................................................	220	6.2.1	 Assay	design:	selection	of	QDs	............................................................................................	220	6.2.1.1	 The	effect	of	QD	emission	wavelength	range	.............................................................................	221	6.2.1.2	 The	effect	of	QD	excitation	wavelength	.....................................................................................	221	6.2.1.3	 Selection	of	FRET	pair:	QD630-A647	and	QD650-A680	..............................................................	223	6.2.2	 Assay	design:	immobilization,	reference	spot,	and	sample	chip	.........................................	225	6.2.3	 Assay	design:	readout	platform	..........................................................................................	228	6.2.3.1	 LED	brightness	............................................................................................................................	229	6.2.3.1	 LED	crosstalk	...............................................................................................................................	231	6.2.3.2	 LED	emission	attenuation	by	blood	............................................................................................	232	6.2.4	 Thrombin	assays	..................................................................................................................	232	6.2.4.1	 Blind	assay	for	thrombin	activity	in	blood	..................................................................................	234	6.2.5	 Competitive	binding	assays	.................................................................................................	235	6.3	 Discussion	...........................................................................................................................	237	6.4	 Conclusions	.........................................................................................................................	240	6.5	 Experimental	Section	...........................................................................................................	241	6.5.1	 Materials	and	reagents	.......................................................................................................	241	6.5.2	 Peptide	modifications	.........................................................................................................	242	6.5.2.1	 Dye-labeled	peptides	..................................................................................................................	242	6.5.2.2	 Biotinylated	peptide	...................................................................................................................	242	6.5.3	 Preparation	of	paper	test	strips	..........................................................................................	243	6.5.3.1	 Sizing	and	chemical	reduction	of	cellulose	paper	.......................................................................	243	6.5.3.2	 Immobilization	of	QD-peptide	conjugates	..................................................................................	243	 xii  6.5.4	 Fabrication	of	PDMS/glass	chip	...........................................................................................	245	6.5.5	 Assay	procedures	................................................................................................................	246	6.5.5.1	 Thrombin	assays	.........................................................................................................................	246	6.5.5.2	 Competitive	binding	streptavidin	assay	......................................................................................	247	6.5.6	 Instrumentation	and	data	acquisition	.................................................................................	247	6.5.6.1	 Characterization	of	LEDs	.............................................................................................................	247	6.5.6.2	 Absorbance,	transmittance	and	fluorescence	measurements	...................................................	248	6.5.6.3	 Sample	matrix-dependent	image	acquisition	parameters	.........................................................	250	6.5.7	 Data	analysis	.......................................................................................................................	251	6.5.7.1	 FRET	parameters	.........................................................................................................................	251	6.5.7.2	 Image	and	RGB	data	analysis	......................................................................................................	252	Chapter	7	 Conclusions	and	Future	Work	...........................................................................	253	7.1	 Thesis	Overview	..................................................................................................................	253	7.2	 Significance	.........................................................................................................................	254	7.3	 Future	Work	........................................................................................................................	256	7.3.1	 Improving	sensitivity	...........................................................................................................	256	7.3.2	 Universal	imaging	platform	for	smartphones	.....................................................................	259	7.3.3	 Scope	of	analytes	and	assays	..............................................................................................	260	7.3.4	 Interaction	between	nanoparticles	and	proteases	.............................................................	260	7.3.5	 Interaction	between	nanoparticles,	proteases,	and	solid	substrates	.................................	264	7.4	 Concluding	Remarks	............................................................................................................	265	References	..........................................................................................................................	266	Appendix	I	...........................................................................................................................	304	Appendix	II	..........................................................................................................................	311	Appendix	III	.........................................................................................................................	313	Copyright	Acknowledgements	.............................................................................................	315	  xiii  List of Tables Table 1.1 NP materials for POC/PON diagnostics. .................................................................................... 13	Table 1.2 Properties of common fluorescent dyes used as acceptors in QD-FRET bioassays. ................ 70	Table 1.3 Summary of spectral overlap, Förster distance and crosstalk for QD-Alexa Fluor dyes. ........... 71	Table 2.1 Characteristics of QD materials, R-PE, and fluorescein. ........................................................... 85	Table 2.2 Peptides sequences. ................................................................................................................ 109	Table 2.3 First exciton wavelength and corresponding molar absorption coefficients for QDs. .............. 110	Table 3.1 Properties of the QDs and FRET pairs used for RGB imaging. ............................................... 116	Table 3.2 Photophysical parameters of FRET pairs. ............................................................................... 120	Table 3.3 Relative changes in RGB channel intensity after 1 h. .............................................................. 122	Table 3.4 Comparison of the specificity constants, kcat/Km (M–1 s–1), calculated for one-plex assays between the fluorescent plate reader and smartphone RGB imaging readout formats. .......................... 131	Table 3.5 Peptide substrate sequences. .................................................................................................. 139	Table 3.6  Preparation of stock solutions of QD peptide conjugates. ...................................................... 140	Table 4.1 Apparent XPS surface composition characterizing surface modifications.a ............................. 150	Table 4.2 Summary of paper substrate (4b) pre-treatment procedures evaluated for QD immobilization. ................................................................................................................................................................. 164	Table 5.1 Peptide substrate sequences. .................................................................................................. 210	Table 6.1 Photophysical parameters of QD630-A647 and QD650-A680 FRET pairs. ............................ 224	Table 6.2 Characterization of LEDs as excitation sources for QD630 and QD650. ................................ 230	Table 6.3 Results of blind assays for thrombin in whole blood. ............................................................... 235	Table 6.4 Peptide substrate sequences. .................................................................................................. 242	Table 6.5 Spotted solutions of QD–peptide conjugates. .......................................................................... 245	 xiv  Table 6.6 LEDs used for optimization experiments. ................................................................................ 248	Table 6.7 Acquisition parameters for iPhone and Lapse It Pro application. ............................................ 251	Table 6.8 Acquisition parameters for Thorlab CMOS camera and micro-Manager software. ................. 251	  xv  List of Figures Figure 1.1 An illustration of the convergence of consumer electronic devices and nanoparticles for POC/PON diagnostics. ................................................................................................................................. 3	Figure 1.2 Basic design of a lateral flow immunochromatographic assay (LFIA). The device comprises a sample pad, conjugate pad, detection zone with test (T) and control (C) lines, and an absorbent pad. The sample containing analyte is added to the sample pad and drawn towards the absorbent pad by capillary action. NP-antibody conjugates bind to analyte (antigen) present in the sample and are captured on the test line, whereas NP-antibody conjugates that have not bound antigen are captured on the control line. . 9	Figure 1.3 (A) Size-dependent molar extinction coefficient of Au NPs as a function of wavelength. The inset photographs show solutions of 5, 10, 15 and 20 nm Au NPs at 1 nM concentration and with an optical density of 1 (ca. 90, 10, 3, and 1 nM concentrations for 5, 10, 15 and 20 nm Au NPs, respectively). (B) Size/composition-tunable absorbance and emission of CdSe/ZnS and CdSeS/ZnS QDs. The inset photograph shows samples of different sizes of CdSe/ZnS QDs under UVA (365 nm) illumination. Photograph reprinted with permission from ref. [126]. Copyright 2011 American Chemical Society. (C) Upconversion emission spectra of NaYF4:Yb/Tm (20/0.2 mol %; blue line) and NaYF4:Yb/Er (18/2 mol %; green line) nanoparticles. The inset photographs show samples of these nanoparticles under 980 nm excitation with a diode laser (600 mW). Adopted with permission from ref. [127]. Copyright 2008 American Chemical Society. ....................................................................................................................................... 15	Figure 1.4 Emission spectra of commercial light sources well-suited to POC/PON diagnostics: (A) various colour LEDs emitting in UV-visible region of the spectrum; (B) white-light LED; (C) five common wavelengths of laser diodes (the FWHM > 1 nm is a measurement artifact); and (D) and a UVA lamp or “black light.” ................................................................................................................................................ 18	Figure 1.5 (A) Examples of consumer electronic devices equipped with CMOS cameras: (i) smartphones; (ii) digital cameras; and (iii) wearable devices. The image in (iii) is reproduced with permission from ref. [137]. Copyright 2014 American Chemical Society. (B) Simplified schematic of a CMOS image sensor. (C) Spectral sensitivity of a typical CMOS image sensor without (black) and with RGB colour filters (coloured lines). The typical blocking region of an IR filter is also shown. ................................................. 19	Figure 1.6 (A) Cell phone-based fluorescence imaging of individual NPs and viruses: (i) Front view of the smartphone microscope and a schematic diagram of its components; (ii) Images of 100 nm fluorescent NPs acquired with the cell phone show excellent agreement with SEM images. Adapted with permission from ref. [157]. Copyright 2013 American Chemical Society. (B) Attachment that enables use of a smartphone as a spectrophotometer for fluorescence emission measurements. The key component is a  xvi  transmission diffraction grating. Reprinted with permission from ref. [158]. Copyright 2014 American Chemical Society. ....................................................................................................................................... 21	Figure 1.7 Jablonski diagrams illustrating the processes of absorption and photoluminescence, as well as competing non-radiative relaxation processes. Singlet electronic states are labeled as Sn, and triplet electronic states are labeled as Tn. The electronic states are shown as potential energy wells with superimposed vibrational states, νn. (A) An example of processes involved in a radiative pathway. (i) Absorption from Soνo to S1ν2, followed by (ii) vibrational relaxation to S1νo and (iii) radiative relaxation to the ground state—fluorescence. (B) Processes involved in non-radiative pathways. Absorption from Soνo to S1ν2 (i) or S2ν4 (ii), followed by (iii) vibrational relaxation to S2ν0 and (iv) internal conversion to S1ν7. This process is followed by vibrational relaxation to S1νo and, from this state, an emission of photon can take place as shown in (A, iii) or by (v) internal conversion—a transition to Soν20, followed by (iii) vibrational relaxation to the Soνo. (C) Other non-radiative relaxation pathways: (i) excitation to S1ν2, followed by (ii) vibrational relaxation to S1νo, and then (iii) intersystem crossing to T1ν1. This process is followed by vibrational relaxation to T1νo. From this state molecules return to the ground state either via non-radiative pathways or via radiative processes—delayed fluorescence or phosphorescence. ............ 25	Figure 1.8 Illustration of Franck-Condon principle for an electronic transition from the ground electronic state, So to the excited state, S1. (A) The maximum transition probability is observed for transitions with maximum overlap between initial and final vibrational wavefunctions, based on Franck-Condon factors. The most probable excitation (blue line) is shown into the second vibrational state (ν2*) of S1, and the corresponding most probable relaxation transition is shown to the second vibrational state (ν2) of So (red line). The probability density function is shown in green. (B) A corresponding schematic for absorption and emission spectra. The spectral characteristics of individual transitions are typically only observed in a gas phase or at very low temperatures, broadening (shown as shaded region) is observed in solution at room temperature. As a result of the Franck-Condon principle, the absorption and emission spectra are approximately a mirror image of each other. .............................................................................................. 30	Figure 1.9 Jablonski diagram illustrating the FRET process. The donor is excited from the ground state So to any of the vibrational levels of the first excited state S1, and then relaxes to the lowest vibrational level of S1 via vibrational relaxation. During nonradiative resonance energy transfer, the donor relaxes to the ground state without emitting a photon, while an acceptor is simultaneously excited to a vibrational level of S1. Then acceptor undergoes the vibrational relaxation to the lowest vibrational level of S1, followed by either emission of a photon (i.e. fluorophore acceptor) as it relaxes to the ground state or nonradiative relaxation (i.e. dark quencher acceptor). ............................................................................... 40	Figure 1.10 Schematic representation of the spectral overlap integral, which is shown as the grey area, J, where the emission spectrum of the donor overlaps with absorption spectrum of acceptor. ..................... 43	 xvii  Figure 1.11 Representation of the donor and acceptor transition dipole moments and angles used to calculate the orientation factor, κ2. ............................................................................................................. 44	Figure 1.12 (A) Size-tunable PL of CdSe QDs. The photograph was taken under UV illumination (365 nm). (B) TEM image of a CdSe/ZnS QD. A and B reprinted with permission from ref. [126]. Copyright 2011 American Chemical Society. (C) Size-dependent absorption and fluorescence spectra of CdSe QDs. Reprinted with permission from ref. [193]. Copyright 2010 American Chemical Society. (D) Absorption and PL spectra of ZnxCd1−xSe QDs with Zn mole fractions of (a) x = 0, (b) 0.28, (c) 0.44, (d) 0.55, and (e) 0.67. Reprinted with permission from ref. [194]. Copyright 2003 American Chemical Society. ................................................................................................................................................................... 50	Figure 1.13 Illustration of band gap engineering by selection of core and shell materials. The relative energy of conduction band and valence band edge states between the core and shell determine the localization of the electron and hole, and the nature of the transition associated with exciton recombination, offering an additional means of tuning the optical properties of QDs. (A) Type-I QD with localization of both carriers in the core; (B) Type-II QD with localization of the electron in the shell; (C) Type-II QD with localization of the hole in the shell; (D) Quasi-Type-II QD with localization of the electron in both and the core and shell; and (E) Inverse-Type-I QD with localization of both carriers in the shell. . 52	Figure 1.14 Illustrative overview of the chemistry of core-shell QDs. Coatings for aqueous solubility: (i) amphiphilic polymer coating with carboxyl(ate) groups; (ii) amphiphilic polymer coating with PEG oligomers; (iii) dithiol ligand with a distal PEG oligomer; (iv) dithiol ligand with a distal zwitterionic functionality; (v) dithiol ligand with a distal carboxyl(ate) group. Common R groups include carboxyl, amine, and methoxy, although many others can be introduced (e.g. see vi, x, xi). Methods for conjugating biomolecules of interest (BOI): (vi) biotin-streptavidin binding; (vii) polyhistidine self-assembly to the inorganic shell of the QD; (viii) amide coupling using N-(3-dimethylaminopropyl)-N′-ethylcarbodiimide (EDC) and sulfo-N-hydroxysuccinimide (s-NHS) activation; (ix) heterobifunctional crosslinking using succinimidyl-4-(N-maleimidomethyl)cyclohexane-1-carboxylate (SMCC; structure not shown); (x) aniline-catalyzed hydrazone ligation; and (xi) strain-promoted azide-alkyne cycloaddition. The double arrows are intended to represent conjugation between the functional groups and, in principle, their interchangeability (not reaction mechanisms or reversibility). Not drawn to scale. ................................................................. 58	Figure 1.15 (A) An illustration of the distance dependence of a QD-dye FRET pair with the corresponding FRET efficiency curve as a function of donor-acceptor separation in terms of r/Ro. (drawn not to scale). (B) Absorption and emission spectra for a QD-dye FRET pair. The shaded area indicates qualitative spectral overlap. ......................................................................................................................................... 64	Figure 1.16 (A) Plot of FRET efficiency as a function of donor-acceptor separation in terms of r/Ro for a fixed, a, the number of acceptors per QD donor. The effective enhancement from 50% to 91% with  xviii  increasing a from 1 to 10 is shown at the donor-acceptor separation equivalent to the Förster distance (dashed lines). (B) Plot of FRET efficiency as a function of the number of acceptors, a, per QD donor for a fixed r/Ro value. ....................................................................................................................................... 65	Figure 1.17 Poisson distribution of QD-MBP conjugates. Agarose gel electropherogram of CdSexS1−x/ZnS QD-His5-MBP conjugates assembled from different relative amounts of MBP per QD (indicated at top). The banding is characteristic of a distribution of conjugate valences (indicated at left). ................................................................................................................................................................... 66	Figure 1.18 Overview of main design components and a flowchart of critical optimization parameters. .. 77	Figure 2.1 Smartphone accessory for all-in-one excitation and imaging of PL. (A) 3D rendering of the accessory. The rendering is partially transparent. The two circular features are the filter positions (i, ii) and the light grey features are the reflector supports (iii, iv). Photographs of the 3D-printed accessory (B) without and (C) with the smartphone in position. (D) Schematic of the device, showing the excitation light path. The arrow in panel A indicates the orientation of this view. .............................................................. 81	Figure 2.2 (A) Spectrum of light from the flash module of the smartphone with and without the blue band-pass excitation filter. (B) Transmission spectra of the band-pass excitation filter and long-pass emission filter. (C) False-colour image that characterizes the homogeneity of the combined excitation intensity and collection efficiency over the useful imaging area. The scale bar is 0.5 cm. .............................................. 82	Figure 2.3 (A) Smartphone PL images of PDMS-on-glass microfluidic chips filled with (i) QD525a (0.5 µM), (ii) QD605 (0.2 µM); and (iii) QD630 (0.2 µM). The channel dimensions are ~300 µm. (B) Smartphone PL images of QD525a immobilized on glass beads (0.5 mm dia.). (C) Smartphone PL image of paper substrates with immobilized QD520, QD525a, QD605, QD630. (D) Smartphone PL image of an agarose gel (1.0%) showing the difference in electrophoretic mobility between a QD605, QD605-[Pep(A488)]20 conjugates, and an equivalent amount of Pep(A488). ......................................................... 83	Figure 2.4 Absorption and emission spectra for the QDs, R-PE, and fluorescein. (A) Absorption spectra and (B) normalized PL emission spectra for the QD materials: (i) QD520, (ii) QD525a, (iii) QD540a, (iv) QD550, (v) QD565, (vi) QD585, (vii) QD605, (viii) QD630, (ix) QD650. (C) Absorption spectra (dotted lines) and emission spectra (solid lines) for R-PE and fluorescein. ........................................................... 84	Figure 2.5 Smartphone PL imaging of QDs, fluorescein, and R-PE. (A) Images (i) of the PL from different concentrations of the nine different QD materials, and (ii) plots of G or R channel intensity measured from the smartphone images as a function of QD concentration. The QDs are grouped in the legend according to the imaging channel in which they appear brightest, and the corresponding channel is plotted. The limit of detection (LOD) for the R or G channel intensity is shown in the inset, which zooms in on the lower concentrations. (B) Analogous images (i) and plots (ii) for fluorescein (G channel) and R-PE (both  xix  channels). Trend lines for selected QD materials are shown for comparison. All images were acquired with the following acquisition settings: shutter speed 1/4 s, ISO 2000, and white balance at 4500 Kelvin; these settings were optimized for sensitivity and not colour selectivity. ..................................................... 87	Figure 2.6 R (red circles) and G (gren circles) channel intensities from smartphone PL images for increasing concentrations of the nine QD materials from Figure 2.5. All images were acquired with the following acquisition settings: shutter speed 1/4 s, ISO 2000, and white balance at 4500 Kelvin; these settings were optimized for sensitivity and not colour selectivity. ............................................................... 88	Figure 2.7 (A) Change in G (green circles) or R (red circles) channel intensity in smartphone images for samples of QD525a (G channel) and QD630 (R channel) as the ISO setting is changed. The inset shows the change in the crosstalk in the secondary image channel (as a percentage of the signal in the primary image channel) as the ISO setting is changed. (B) Crosstalk for fluorescein, R-PE, and QD605 as a function of ISO setting. ............................................................................................................................... 90	Figure 2.8 Change in G or R channel intensities in smartphone images for samples of QD525a, QD605, QD630 as the shutter speed is changed at (A) ISO 1000 and (B) ISO 2000. The main panel plots the channel intensities as a function of N, where 1/N is the shutter speed setting. The insets show the value of the exposure time (1/N) in units of seconds. QD525a were imaged in G channel and QD605 and QD630 were imaged in R channel in (A) and (B). ...................................................................................... 91	Figure 2.9 Effect of white balance/colour temperature (Kelvin) setting. (A) Images of solutions of QD525a, QD630, and a mixture of these two QDs (concentrations of each QD unchanged) at different white balance/colour temperature settings. (B) G or R channel intensities and crosstalk levels (secondary channel intensity as a percentage of the primary channel intensity) at ISO 2000, ISO 1000, and ISO 800. ................................................................................................................................................................... 92	Figure 2.10 R/G channel intensity ratio as a function of white balance/colour temperature (Kelvin) setting at ISO 2000, ISO 1000, ISO 800. ............................................................................................................... 93	Figure 2.11 Changes in G or R channel intensity for mixtures of QD520 and QD630, where the concentration of only one QD varied and the concentration of the other was kept constant: (A) 0.1 µM QD630 with variable concentrations of QD520; and (B) 1.0 µM QD630 with variable concentrations of QD520. Control samples of only QD520 or only QD630 are shown for reference. The intensities are shown in panels (i), the corresponding G/R or R/G channels ratios are plotted in panels (ii), and the correlation between the spectrofluorimetric PL intensities for samples of QD520, QD630, and mixtures of QD520 and QD630 are shown in panels (iii). ............................................................................................. 95	 xx  Figure 2.12 Model binding assay with avidin as a target analyte. (A) Schematic of the avidin-binding assay. (B) Smartphone PL images (i) and G or R channel intensities (ii) as a function of avidin concentration. The G channel intensity was corrected for crosstalk from the QD630 signal. .................... 97	Figure 2.13 Normalized absorption (dotted lines) and emission spectra (solid lines) for: (A) QD605 and A647, and (B) QD630 and A680. ............................................................................................................... 98	Figure 2.14 (A) FRET between QD630 and A680: (i) PL emission spectra for QD630-[SubTHR(A680)]N conjugates with increasing N, the average number of A680 acceptors per QD; (ii) corresponding R channel intensity from smartphone images of the same samples. (B) FRET between QD605 and A647: (i) PL emission spectra for QD605-[SubTHR(A647)]N conjugates with increasing N, the average number of A647 acceptors per QD; (ii) corresponding R channel intensity from smartphone images of the same samples. The insets in (i) show the PL ratios (blue) and relative change in QD PL intensity (red) with increasing N, the average number of acceptors per QD. The PL ratios in the insets of (i) are in terms of peak heights (not areas). ............................................................................................................................ 99	Figure 2.15 FRET-based assay for proteolytic activity. (A) Schematic of the QD630-[SubTHR(A680)]16 conjugate and its mechanism of sensing thrombin activity via FRET. (B) R image channel intensity measured for QD630-[SubTHR(A680)]16 conjugates after incubation for 15 min with different concentrations of human thrombin spiked into 50% v/v serum (main panel) and buffer (inset). The QD PL intensity recovers as a result of thrombin activity. .................................................................................................. 100	Figure 2.16 Enhancement of dye PL emission intensity through FRET-sensitization from a QD donor. (A) Schematic of the conjugate. (B) Overlay of the absorption spectra of the QD and A610 with the spectrum of the excitation light. (C) Emission spectrum of the QD540a-[Pep(A610)]8. (D) Graph showing the increase in A610 emission intensity (R channel) with smartphone PL imaging after conjugation to the QD540a, versus control samples of only QD540a and A610 alone. The inset shows representative smartphone images of the tandem conjugates and control samples. ...................................................... 101	Figure 2.17 Sizes of (A) fluorescein, (B) 3.5 nm diameter QD; (C) 6 nm diameter QD, and (D) R-PE. The R-PE structure was generated from PDB entry 1EYX. ............................................................................ 105	Figure 2.18 TEM images of CdSe/CdS/ZnS core/shell/shell and alloyed CdSeS/ZnS core/shell QDs. .. 106	Figure 3.1 (A) Design of a homogeneous multiplexed assay to monitor protease activity via FRET with QD donors. Acceptor (QSY35, QSY9, or A647)-labeled peptide substrates containing a cleavage site for trypsin (TRP), chymotrypsin (ChT), or enterokinase (EK) were assembled on QDs with blue (QD450), green (QD540), and red (QD625) emission. The numbers in the abbreviations for the QDs indicate the approximate wavelength of their emission maximum. (B) Protease activity was measured through the  xxi  recovery of QD PL and the resulting increase in the corresponding RGB channel intensity in digital colour images acquired with a smartphone and a handheld UV lamp for excitation. .......................................... 115	Figure 3.2 False coloured, RGB split digital images of QD samples of various concentrations. Each row in the figure is an RGB channel from a digital colour image of a single microtiter plate row. ...................... 117	Figure 3.3 Colour images of various concentrations of QDs (inset) and the correlation between the measured RGB channel intensities and the integrated PL intensities for QD450 (blue), QD540 (green), and QD625 (red). ..................................................................................................................................... 117	Figure 3.4 (A) QD PL spectra and (B) corresponding colour images for the (a) QD mixture, (b) QD450, (c) QD540, and (d) QD625. (i) QD450 (30 pmol), QD540 (5 pmol), and QD625 (5 pmol). (ii) QD450 (50 pmol), QD540 (10 pmol), and QD625 (10 pmol). (iii) QD450 (75 pmol), QD540 (20 pmol), and QD625 (20 pmol). RGB values shown in panel (B) were calculated by accounting for the crosstalk using eqns. 3.1-3.3 (Section 3.5.5.2, page 141). The dashed lines in panel (A) are spectra for mixtures, and the solid coloured lines are for the individual QDs. ................................................................................................. 118	Figure 3.5 (A) Emission profile for the UV lamp excitation source; emission spectra of the QD450, QD520, QD625, and A647; and approximate cutoff wavelengths for the shortpass and longpass filters placed in front of the smartphone camera lens. Normalized absorption and emission spectra of (B) the QD450–QSY35 FRET pair, (C) the QD540–QSY9 FRET pair, and (D) the QD625–A647 FRET pair. ... 119	Figure 3.6 (A) Calibration curves showing the decrease in QD PL intensity as the number of acceptors per QD increases for QD450–[SubTRP1(QSY35)] (blue), QD540–[SubChT(QSY9)] (green), and QD625–[SubTRP2(A647)] (red) conjugates. (B) Calibration curves for the same samples as in panel (A), showing decreases in RGB channel intensities as the number of acceptors per QD increases. (C) Photograph showing smartphone images of the samples in microtiter plate wells. The average number of acceptors per QD is shown at the top of the photograph. ......................................................................................... 121	Figure 3.7 PL spectra for (A) QD450–[SubTRP(QSY35)] conjugates (blue channel), (B) QD540–[SubChT(QSY9)] conjugates (green channel), and (C) QD625–[SubEK(A647)] conjugates (red channel). The insets show the corresponding FRET efficiency plots, raw data as open circles and Poisson corrected data as solid circles. ................................................................................................................. 121	Figure 3.8 (A) Agarose gel electrophoresis (0.8%, 1×TBE, 100 V, 30 min) of QD450–[SubTRP1(QSY35)]20, QD540, and QD625. The arrows at the left indicate the corresponding bands in the mixture (i). Reference samples of QD540 and QD625 were run in parallel, (ii) and (iii), respectively. The colour image was acquired with a smartphone camera with UV lamp excitation. Note the yellow colour from the overlap between the QD540 and QD625 in lane (i). A faint blue band can be observed above the QD540/QD625 band. (B) Intensity line profiles in the blue channel of gel image shown in panel (A). The QD540 and  xxii  QD625 have some crosstalk in the blue channel of the image, but there is distinct band for the QD450. ................................................................................................................................................................. 123	Figure 3.9 Mixed digest/substrate calibration curves used for the analysis of progress curves in one-plex assays with (A) a fluorescence plate reader and (B) smartphone RGB imaging. The x-axis refers to the number of native peptide substrates per QD (with N–x pre-digested peptide, where N = 12 or 16): (i) QD625–[SubTRP2(A647)]12 conjugates, (ii) QD540–[SubChTQSY9]16 conjugates, (iii) QD450–[SubTRP1(QSY35)]12 conjugates. The data points were fit with the simplest possible polynomial function (linear, quadratic, or cubic). ...................................................................................................................... 124	Figure 3.10 Relative decreases in QD625 PL intensity upon assembly of 4, 8, and 12 equivalents of peptide with various modifications of the C-terminal residue (A488 = Alexa Fluor 488, A555 = Alexa Fluor 555, Cys = cysteine). Note that neither A488 nor A555 is a FRET acceptor for QD625. ......................... 125	Figure 3.11 Time trace of the relative output from UV lamp at 367 nm over 60 min. .............................. 126	Figure 3.12 Time traces for blue and green LED output measured with (A) a spectrometer and (B) a smartphone digital camera. (C) The blue and green LED emission spectrum and the corresponding colour image of the LEDs. ................................................................................................................................... 126	Figure 3.13 Proteolytic digestion of QD625–[SubTRP2(A647)]12 conjugates by TRP and comparison of data acquired with (A) a fluorescence plate reader and (B) a smartphone and RGB imaging: (i) raw PL data; (ii) normalized PL data; (iii) conversion of the normalized PL data to the average number of acceptors per QD and the bulk equivalent concentration of peptide substrate. (C) Comparison of initial proteolytic rates measured from the fluorescence plate reader data and smartphone RGB imaging data. ....................... 128	Figure 3.14 Proteolytic digestion of QD540–[SubChT(QSY9)]16 conjugates by ChT and comparison of data acquired with (A) a fluorescence plate reader and (B) a smartphone and RGB imaging: (i) raw PL data; (ii) normalized PL data; (iii) conversion of the normalized PL data to the average number of acceptors per QD and the bulk equivalent concentration of peptide substrate. (C) Comparison of initial proteolytic rates measured from the fluorescence plate reader data and smartphone RGB imaging data. ....................... 129	Figure 3.15 Proteolytic digestion of QD450–[SubTRP1(QSY35)]12 conjugates by TRP and comparison of data acquired with (A) a fluorescence plate reader and (B) a smartphone and RGB imaging: (i) raw PL data; (ii) normalized PL data; (iii) conversion of the normalized PL data to the average number of acceptors per QD and the bulk equivalent concentration of peptide substrate. (C) Comparison of initial proteolytic rates measured from the fluorescence plate reader data and smartphone RGB imaging data. ................................................................................................................................................................. 130	 xxiii  Figure 3.16 Progress curves for two-plex homogeneous assays of proteolytic activity. (A) RG two-plex assay with QD540–[SubChT(QSY9)]16 (20 pmol) and QD625–[SubTRP2(A647)]12 (30 pmol). (B) GB two-plex assay with QD450–[SubTRP1(QSY35)]12 (100 pmol) and QD540–[SubChT(QSY9)]16 (20 pmol). The conjugates were exposed to the indicated mixtures of TRP and ChT. The ordinate and abscissa scales are the same in each panel of (A) or (B). Corresponding raw data is shown in Figure 3.17. ................... 132	Figure 3.17 Representative unprocessed RGB intensity time traces for two-plex assays with (A) QD540–[SubChT(QSY9)]16 and QD625–[SubTRP2(A647)]12 conjugates upon exposure to various concentrations of TRP and ChT, and (B) QD540–[SubChT(QSY9)]16 and QD450–[SubTRP1(QSY35)]12 conjugates upon exposure to various concentrations of TRP and ChT. Corresponding normalized data is shown in Figure 3.16. ......................................................................................................................................................... 133	Figure 3.18 Three-plex homogeneous proteolytic assays. (A) Representative colour images (brightness enhanced for clarity) for QD450–[SubTRP1(QSY35)]12 (100 pmol), QD540–[SubChT(QSY9)]16 (20 pmol), and QD625–[SubEK(A647)]12 (30 pmol) mixed with various concentrations of TRP, ChT, and EK: (i) reference sample (no proteases); (ii) 1 nM TRP, 1 nM ChT, and 4.5 nM EK; (iii) 8.6 nM TRP, 2 nM ChT, and 1.5 nM EK; (iv) 2.2 nM TRP, 16 nM ChT, and 0.5 nM EK. (B) Progress curves corresponding to (ii), (iii), and (iv). .............................................................................................................................................. 135	Figure 4.1 Stepwise chemical modification of cellulose paper substrates to immobilize QDs via self-assembly. (A) cellulose oxidation was used to generate reactive aldehyde groups (2) that were subsequently reacted to prepare imidazole (3) and thiol (4) functionalized cellulose paper, (B) cellulose paper was modified with 3-aminopropyl triethoxysilane to give primary amine functionalized paper (5), that was subsequently reacted with NHS-ester activated lipoic acid to yield thiol-modified cellulose paper (6). Details of reagents and reaction conditions are given in Schemes 4.2–4.5 (p. 178–182). ................ 148	Figure 4.2 XPS survey analysis for stepwise modification of cellulose paper with (A) imidazole and lipoic amine ligands; (i) full spectra, (ii) N 1s peak observed upon addition of imidazole ligands (3) and (iii) N 1s, S 2s, and S 2p peaks observed on lipoic amine modified paper (4). (B) Chemical modification of paper via silanization with APTES; Full spectra are shown in (i) and a region of interest in (ii). The peaks corresponding to O 1s, C 1s, N 1s, S 2s, S 2p, Si 2s, and Si 2p orbitals are labeled. ............................. 151	Figure 4.3 ATR-FTIR spectra for each step of chemical modification of cellulose paper samples. (1) untreated cellulose, (2) oxidized cellulose, (3) imidazole ligand modified cellulose, (4) lipoic amine modified, (5) APTES modified and (6) cellulose modified with APTES and lipoic acid ligands. ............... 152	Figure 4.4 (A) Absorption spectra for 1,3,5-triphenyl formazan (TTC formazane) in ethanol. (B) Calibarion curve used to determine molar absorption coefficient for TTC formazane (slope = 8590 M–1cm–1, path length b = 0.5 cm) in ethanol. The inset shows the calibration curve fit at lower concentrations of TTC  xxiv  formazan. (C) Experimental data for quantification of aldehyde functional groups in six replicate samples of paper substrate (2), the control sample corresponds to measurements made with untreated cellulose paper (1). All spectra were collected in > 97.5% ethanol. ........................................................................ 155	Figure 4.5 (A) Representative absorption spectra for Ruhemann’s purple collected to construct a calibration curve, shown in panel (B). (C) Digital photographs of paper samples with spotted Ruhemann’s purple solution used for calibration curve and three replicate samples of (5) used for quantification of primary amines. ........................................................................................................................................ 157	Figure 4.6 Representative absorption spectra showing formation of TNB2– product as a result of the reaction of Ellman’s reagent with thiol ligands (A) in paper substrate (4b), and (B) in paper substrate (6b). Control sample measured on untreated paper (1) (black dashed line) and blank sample containing no paper substrate (dashed blue line) are also shown for a reference. ........................................................ 157	Figure 4.7 Background corrected relative PL intensity of chemically modified paper substrates (according to ligands shown in Figure 4.1) after exposure to DHLA630 QDs. All data was normalized to the brightest sample functionalized with imidazole ligands (3). Paper substrates (3), (4b), and (6b) functionalized with surface ligands for specific QD immobilization. Paper substrate (2), (4a), (5), and (6a) retained QDs due to non-specific interactions. Error bars represent the standard deviation of three replicate measurements. The inset shows the corresponding QD PL emission spectra for selected modification steps. Dashed lines are used to show QD PL intensity from paper substrates (2), (4a), and (6a) that are not expected to have strong interactions with QDs (i.e. dithiolane groups in (4a, 6a) versus dithiol groups in (4b, 6b)) and signals are attributed to non-specific QD adsorption. ............................................................................... 159	Figure 4.8 SEM images for cellulose paper modified with (A) imidazole, (B) lipoic amine, and (C) APTES and lipoic acid before (i) and after (ii) exposure to QDs. All scale bars are 200 µm for 100× magnification, 500 nm for 50 000× magnification, and 100 nm for 200 000× magnification. ........................................... 161	Figure 4.9 Effect of buffer composition on optical properties of immobilized QDs in paper substrate (4b). (A) Absorption spectra of immobilized QDs deposited from different buffer solutions. (B) Relative QD PL intensity corresponding to (A) showing that increase in buffer pH and its buffering capacity is accompanied with higher brightness of immobilized QDs. ....................................................................... 163	Figure 4.10 Effect of paper substrate (4b) pre-treatment with different buffers on the density of immobilized QDs and their relative brightness. (A) Absorption spectra collected for immobilized QDs on paper substrates washed with (i) ammonium acetate (pH 4.5), (ii) water, (iii) tris-borate buffer (pH 7.4), and borate buffer (pH 9.2), as noted in Table 4.2. (B) Relative QD absorption from (A) at 610 nm and corresponding PL intensity at 628 nm. Both QD absorption and emission are normalized to sample (i). 164	 xxv  Figure 4.11 Two-photon imaging of immobilized QD630 (A) on paper substrate (3), (B) on paper substrates (4b), and (C) on paper substrates (6b). .................................................................................. 165	Figure 4.12 Photobleaching and photobrightening of QD630 capped with DHLA ligands in solution (black trace) and immobilized in paper substrates (3) (green trace), (4b) (blue trace), and (6b) (orange trace). ................................................................................................................................................................. 166	Figure 4.13 Comparison of storage stability of immobilized DHLA-QD630 on imidazole ligand modified paper (3) and lipoic amine modified paper (4b) in different media (A) buffer, (B) buffer:glycerol (1:1 v/v), and (C) dry samples as function of temperature (4oC and room temperature, RT). The PL intensity of QDs in solution is also shown for reference (dashed line). All data points are normalized to Day 0 and error bars represent the standard deviation of three replicate measurements. ................................................ 168	Figure 4.14 FRET characterization. (A) Solution-phase FRET (i) Representative absorption spectra showing assembly of 0, 4, 8, 12, 16, and 20 A555-labeled peptides per QD, (ii) PL emission corresponding to samples in (i). The inset shows the FRET efficiency as a function of the number of A555 acceptors calculated from quenching of QD PL according to eqn. 1.24, (iii) The A555/QD PL ratio as a function of the number of A555 acceptors. (B) Representative paper-phase FRET on imidazole modified paper (3). (i) Absorption spectra showing paper-phase assembly of QD bioconjugates prepared with 0, 4, 8, 12, 16, and 20 of A555-labeled peptides. (ii) PL spectra corresponding to (i) measured from fully hydrated paper samples, and (iii) PL spectra corresponding to (i) measured from air-dried samples. (C) Same as (B) on lipoic amine modified paper substrate (4b). ................................................................... 171	Figure 4.15 Comparison of A555/QD PL ratio and FRET efficiency for QD bioconjugates immobilized in paper samples (3), (4b), and (6b) shown for hydrated samples (A) and dry samples (B). FRET efficiency was calculated using eqn. 4.4.  Corresponding PL spectra are shown in Figure 4.14. ............................ 172	Figure 4.16 (A) Fluorescence lifetime imaging of paper substrates functionalized with lipoic amine (4b) and modified with QD-A555 bioconjugates at ratios 0, 4, 8, 12, 16, and 20 A555 labeled peptides per QD. (B) Normalized average two-component fluorescence lifetime distributions for the six different conjugates in bulk solution: (i) 0, (ii) 4, (iii) 8, (iv) 12, (v) 16, and (vi) 20 A555 per QD. (C) Average two-component fluorescence lifetime distributions for the same conjugates subsequently immobilized on paper substrates. The data correspond to the images in (A). (D) Comparison of the FRET efficiency calculated from the decrease in QD PL lifetime between bulk solution in (B) and the paper substrates in (C). ........ 174	Figure 4.17 Paper-based proteolytic activity assay with immobilized QDs in paper sample (i) imidazole-modified paper (3); (ii) lipoic amine modified paper (4b); and (iii) APTES-LA modified paper (6b). ........ 175	Figure 5.1 (A) Design of a paper-based assay to monitor protease activity via FRET with QD donors. Acceptor dye-labeled (A555) peptide substrates containing a cleavage site for trypsin (TRP),  xxvi  chymotrypsin (ChT), and enterokinase (EK) were assembled on immobilized QDs. Protease activity was measured through changes in the intensity and colour of PL from the spots of immobilized QDs and peptides. (B) Synthetic steps used for modification of cellulose paper fibers with bidentate thiol surface ligands for immobilization of QDs. (C) Colour images of paper-immobilized QDs and A555-labeled QD-Sub(TRP) peptide conjugates under white light and their PL under UV illumination. .............................. 192	Figure 5.2 Normalized absorption and PL spectra for the QD and A555 FRET-pair. The emission profile for the LED excitation source is also shown. ............................................................................................ 193	Figure 5.3 (A) Three-dimensional PL image of immobilized DHLA-coated QDs on cellulose paper fibers obtained using confocal microscopy. The image dimensions are 775 × 775 × 42 µm. (B) Fluorescence lifetime images of paper substrates modified with green-emitting DHLA-coated QDs and QD-SubTRP conjugates (i.e. QD-A555). The scale bars in (A) and (B) are 100 µm. (C) Fluorescence lifetime distributions for immobilized QDs and QD-SubTRP conjugates (i.e. QD-A555). ........................................ 194	Figure 5.4 (A) Time-dependent PL spectra for paper-immobilized QD-SubTRP conjugates upon exposure to 430 nM TRP over 1 h. PL spectra were collected at 1 min intervals. Protease activity resulted in the simultaneous recovery of QD PL and loss of FRET-sensitized A555 PL. Residual A555 emission above the baseline is indicative of incomplete proteolysis or some limited adsorption of digested peptide fragments to the QDs. (B) Progress curves for the hydrolysis of QD-SubTRP conjugates catalyzed by different concentrations of TRP. ............................................................................................................... 196	Figure 5.5 (A) PL spectra of immobilized DHLA-QDs and QD-SubTRP conjugates (i.e. QD-A555). Dotted lines show the relative spectral response of the digital camera (Moticam 1 [484]) in the G and R channels. (B) Colour digital images of immobilized QD-SubTRP at various time points of protease activity acquired with a digital camera and corresponding pseudo-coloured images in the green and red channel upon R-G-B splitting. (C) The empirical relationship between R/G ratio and PL ratio. (D) R/G progress curves obtained from colour images of immobilized QD-Sub(TRP) spots upon exposure to different concentrations of TRP (see legend in panel E). (E) R/G progress curves in panel D converted to progress curves in terms of PL ratio using the data in panel C. (F) Calibration curve for measuring TRP activity based on the normalized initial rates of digestion. Normalized initial rates were reproducible within 10%. ................................................................................................................................................................. 198	Figure 5.6 (A) Examples of raw R/G ratio progress curves obtained from colour images of immobilized QD-SubChT spots upon exposure to different concentrations of ChT. (B) Normalized R/G progress curves derived from the data in panel A and additional data. (C) PL ratio progress curves derived from the data in panel B and the power relationship between R/G and PL ratios, using relationship shown in Figure 5.5C. (D) Calibration curve for ChT activity based on normalized initial rates derived from the progress curves in  xxvii  panel B. The curve was used to determine the apparent Km. (E) Representative examples of changes in the mean intensity in the G and R channels for QD-SubChT exposed to buffer (blank) and 20 nM ChT. . 201	Figure 5.7 Progress curves showing the inhibition of TRP (860 nM, 20 µg/mL) activity by different concentrations of aprotinin. ...................................................................................................................... 202	Figure 5.8 Two-plex assays of proteolytic activity. Progress curves and corresponding colour images acquired upon exposure of immobilized spots of QD–SubTRP and QD–SubChT to protease mixtures containing (A) 215 nM (5 µg/mL) TRP and 8 nM (0.2 µg/mL) ChT, (B) 86 nM (2 µg/mL) TRP and 80 nM (2 µg/mL) ChT, and (C) 8.6 nM (0.2 µg/mL) TRP and 200 nM (5 µg/mL) ChT. Only the negative control spot with SubTRP is shown (blank). ........................................................................................................... 203	Figure 5.9 Non-normalized R/G ratio progress curves for exposure of immobilized spots of QD–SubTRP and QD–SubChT to protease mixtures containing (A) 215 nM (5 µg/mL) TRP and 8 nM (0.2 µg/mL) ChT, (B) 86 nM (2 µg/mL) TRP and 80 nM (2 µg/mL) ChT, and (C) 8.6 nM (0.2 µg/mL) TRP and 200 nM (5 µg/mL) ChT. Only the negative control spot with SubTRP is shown (blank). The corresponding normalized data is shown in Figure 5.8. ................................................................................................... 204	Figure 5.10 Progress curves and corresponding colour digital images for multiplexed protease assays. Spots of GSH-QDs functionalized with SubTRP, SubChT, and SubEK were tested against a mixture containing TRP 8.6 nM (0.2 µg/mL), ChT 8.0 nM (0.2 µg/mL), and EK 7.6 nM (0.2 µg/mL). Only the negative control spot with SubTRP is shown. ............................................................................................. 205	Figure 5.11 (A) Schematic of cascaded pro-enzyme activation: EK activates pTRP to TRP (step I), which then activates pChT to ChT (step II). The formation of active TRP and ChT is monitored with SubTRP and SubChT, respectively. (B) Progress curves and corresponding colour digital images for direct activation of 195 nM pChT (5 µg/mL) by 215 nM TRP (5 µg/mL). Open circles correspond to the control samples containing only pChT at 5 µg/mL. (C) Progress curves and corresponding colour digital images for cascaded activation of 210 nM pTRP (5 µg/mL) and 195 nM pChT (5 µg/mL) initiated with EK (5 units). Open circles correspond to the control samples containing a mixture of pTRP and pChT at 5 µg/mL concentration each (no EK). ..................................................................................................................... 206	Figure 5.12 Comparison of R/G progress curves acquired with a digital camera, computer webcam, and iPhone (smartphone) upon exposure of spots of immobilized QD-SubTRP and QD-SubChT to a mixture containing 215 nM TRP (5 µg/mL) and 160 nM ChT (4 µg/mL). The camera responses are almost indistinguishable. ...................................................................................................................................... 208	Figure 5.13 Schematics of the instrumental setups used for (A) acquisition of PL spectra, (B) colour digital PL images with either a low-cost digital microscopy camera (Moticam 1) or consumer webcam, and  xxviii  (C) colour digital PL images with a smartphone. For the smartphone imaging, the LED source was powered from three 1.5 V batteries (in series) instead of the USB connected DAQ module. Schematics are not to scale. In each case, a long-pass filter was used to block reflected LED light. The built-in colour filters of the cameras discriminated between QD and dye emission. ....................................................... 213	Figure 6.1 (A) Design of paper test strips to measure thrombin activity via FRET with immobilized QD donors and A647 acceptor dye-labeled peptide substrates containing a cleavage site recognized by thrombin. Protease activity was measured through the recovery QD PL with loss of FRET. (B) Paper test strips with sample and reference spots of immobilized QD-peptide conjugates were (i) enclosed within PDMS/glass sample cells that were then (ii) filled with a biological sample matrix such as serum, diluted blood or whole blood. Note the opacity of the whole blood. (C) Photograph of the setup used for smartphone readout of QD-FRET test strip assays with serum and blood samples. The inset shows the setup with the LED470 illuminating the PDMS/glass sample chip. .......................................................... 219	Figure 6.2 Spectra showing the absorption of a blood sample, the absorption (dashed line) and emission (solid line) spectra associated with the QD630-A647 FRET pair, and the emission spectrum of the LED470 excitation source. The fluorescence spectra are normalized for easy comparison. The absorbance spectra were measured for the following solutions: 0.27 µM QD630, 4.0 µM A647, and 0.05% v/v blood (1 cm path length). The transmission spectrum of a 624/40 bandpass filter used to isolate QD emission prior to the smartphone camera is also shown. ........................................................................ 220	Figure 6.3 Normalized absorption and PL spectra for the (A) QD630 and (B) QD650. The transmission spectra of the bandpass filters are shown as solid black lines and the absorption spectrum of a 0.05% blood sample (path length = 1 cm) is shown as a shaded grey region. ................................................... 221	Figure 6.4 (A) Variation in the intensity of the QD630 and QD650 PL spectra as a function of sample matrix (excitation at 470 nm). (B) Excitation wavelength-dependent attenuation of the QD630 and QD650 PL in serum and whole blood (solid lines). The wavelength-dependent molar absorption coefficients of QD630 and QD650 are also shown (dashed lines). These measurements were done with a fluorescence plate reader. ............................................................................................................................................. 222	Figure 6.5 Normalized absorption and emission spectra for (A) the QD630-A647 FRET pair and (B) the QD650-A680 FRET pair. Absorption spectra are shown as dotted lines and emission spectra are shown as solid lines. ............................................................................................................................................ 223	Figure 6.6 TEM images of (A) QD630 and (B) QD650. .......................................................................... 223	Figure 6.7 Comparison of FRET efficiency between (A) the QD630-A647 and (B) the QD650-A680 FRET pairs in bulk solution, measured with a fluorescence plate reader. Insets in (A) and (B) show the FRET efficiencies calculated from quenching of QD PL according to eqn. 1.24 as a function of the number of  xxix  acceptors per QD (open circles, raw data; closed circles, Poisson corrected data). (C) PL spectra of paper-immobilized QD630-A647 peptide conjugates. The inset shows extent of QD PL quenching as a function of the number of acceptors, measured from spectra acquired with fluorescence plate reader (open circles) and digital images acquired with smartphone (closed circles). .......................................... 225	Figure 6.8 (A) Experimental design to determine effect of the optical path length through blood on measured PL intensity of immobilized QDs. The paper substrates were placed between two glass slides and distance separation was controlled by placing cover slips (n = 1–8) between them. (B) Effect of the optical path length through blood on attenuation of QD630 PL intensity from paper substrates measured with iPhone and CMOS monochrome camera. ........................................................................................ 227	Figure 6.9 Effect of thickness of the paper on the intensity of LED470 light transmitted through chemically modified Whatman chromatography paper No. 4 soaked in buffer. ......................................................... 228	Figure 6.10 Normalized progress curves for thrombin activity in (A, D) buffer, (B, E) serum, and (C, F) whole blood samples, measured via smartphone imaging (A, B, C) and via a USB-CMOS camera (D, E, F). Representative smartphone images are shown for three points in the assays: prior to the addition of sample, immediately after adding sample, and after 30 min. In each image, the spiked thrombin concentrations were (i) 0, (ii) 15.1 NIH U mL–1, (iii) 30.3 NIH U mL–1, (iv) 121 NIH U mL–1, (v) 242 NIH U mL–1, and (vi) 484 NIH U mL–1. ................................................................................................................ 233	Figure 6.11 Progress curves for the detection of thrombin activity in blind samples. Error bars describe standard deviation (n = 5). ........................................................................................................................ 234	Figure 6.12 (A) PL emission spectra corresponding to the increasing number of SAv-A647 binding to immobilized QD630-Pep(biotin) conjugates. The number of SAv per QD were 0, 0.5, 1, 1.5, 2, 2.5, 3, 3.5, 4, 4.5, 5, and 6. (B) QD PL intensity corresponding to (A); spectral data is plotted in terms of the QD peak PL intensity and smartphone data is plotted as the red channel intensity. The other panels show data from paper-based competitive SAv assays in (C) buffer and (D) 50% blood. The data is the relative QD PL intensity as a function of time upon exposure to different concentrations of SAv (analyte) and a fixed concentration of SAv-A647 (competitor). For assays in buffer, the concentration of SAv-A647 was 2.0 µM, whereas the concentration for assays in 50% blood was 1.0 µM. ............................................................ 236	Figure 6.13 Preparation of paper substrates with immobilized QD-peptide conjugates and assembly of glass/PDMS chip containing paper substrate within sample cell. ............................................................ 244	Figure 6.14 Drawing of a positive-relief template used to mould PDMS to form 18 sample cells on a microscope slide-sized chip. .................................................................................................................... 246	 xxx  Figure 6.15 Schematics of the instrumental setups used for (A) acquisition of LED spectra, (B) solution-phase QD PL with LED excitation, and (C) colour digital PL images with a smartphone. For the smartphone imaging, the pulsed LED source was powered from the USB connected DAQ module. Schematics are not to scale. A bandpass filter was used to isolate QD PL and reject reflected LED light and FRET-sensitized emission from dye-acceptor. .................................................................................. 250	Figure 7.1 Representative data for proteolysis by (A) trypsin, (B) thrombin, and (C) plasmin on (i) GSH, (ii) DHLA, and (iii) PEG coated QDs as a function of additional peptide substrate assembled on QDs. A total number of peptides per QD is 10 (blue), 20 (green), 40 (orange), and (60) red on GSH (i) and DHLA (ii) capped QDs and 10 (blue), 20 (green), 30 (orange), and 40 (red) on PEG coated QDs. Maximum peptide loading on QD was determined using gel electrophoresis (data not shown): 60 peptides on GSH and DHLA capped QDs, and 40 peptides on PEG-coated QDs. Corresponding gel electrophoresis (0.7%, TB, pH 8.5) for QDs mixed with an increasing concentration of the protease are also shown; [QD] = 200 nM and protease concentration (from left to right on the image) 0, 0.1, 0.2, 0.4, 0.8, 1.6, 3.2, 6.4, 13, and 25 µM. ...................................................................................................................................................... 263	Figure A1 (A) Determining the quantum yields of green-emitting QD525 and QD540 relative to fluorescein, and (B) determining the quantum yields of red-emitting QD605 and QD625 relative to rhodamine B. ............................................................................................................................................ 312	Figure A2 Screenshot of the custom LabVIEW program used to control LEDs. ..................................... 313	Figure A3 Screenshot of the custom LabVIEW program used to acquire time-based spectra with Green-Wave spectrometer (StellarNet). .............................................................................................................. 314	  xxxi  List of Schemes Scheme 4.1 Colorimetric tests used to estimate the surface densities of accessible functional groups. Aldehyde content was measured in sample (2) with TTC, primary amine content in sample (5) was measured with ninhydrin, and available thiol groups in samples (4b) and (6b) were determined with Ellman’s reagent…………………………………………………………………………………………………..154 Scheme 4.2 Chemical modification of cellulose paper fibers with imidazole-based ligands for immobilization of the QDs. ………………………………………………………………………………………178 Scheme 4.3 Chemical modification of cellulose paper fibers with N-(2-aminoethyl)-5-(1,2-dithiolan-3-yl)pentanamide ligands for immobilization of the QDs. ………………………………………………………179 Scheme 4.4 Synthesis of lipoic acid-NHS. ……………………………………………………………………180 Scheme 4.5 Chemical modification of cellulose paper fibers with lipoic acid ligands for immobilization of the QDs. …………………………………………………………………………………………………………...182 Scheme A1 Synthesis of DHLA-PEG ligands. ………………………………………………………………..304 Scheme A2 Synthesis of DHLA-zwitterionic (ZW) ligands. ………………………………………………….308   xxxii  List of Abbreviations A488 Alexa Fluor 488 A555 Alexa Fluor 555 A610 Alexa Fluor 610 A647 Alexa Fluor 647 A680 Alexa Fluor 680 APTES 3-aminopropyltriethoxysilane ATR-FTIR Attenuated Total Reflectance FTIR Au NP Gold nanoparticle BB Borate buffer BP Bandpass (filter) BSA Bovine serum albumin ChT Chymotrypsin CMOS Complementary metal–oxide–semiconductor Cys Cysteine DCC Dicyclohexylcarbodiimide DCM Dichloromethane DHLA Dihydrolipoic acid DHLA-PEG Polyethylene glycol appended DHLA DHLA-ZW Zwitterion appended DHLA DIC Diisopropylcarbodiimide DIPEA N,N-diisopropylethylamine DMF Dimethylformamide DMSO Dimethylsulfoxide DNA Deoxyribonucleic acid EK Enterokinase FLIM Fluorescence lifetime imaging microscopy FRET Förster resonance energy transfer FTIR Fourier transform infrared FWHM Full-width-at-half-maximum GSH Glutathione HEPES N-2-hydroxyethylpiperazine-N-2-ethane sulfonic acid HOMO Highest occupied molecular orbital ISO International Organization for Standardization LA Lipoic acid LA-NH2 N-(2-aminoethyl)-5-(1,2-dithiolan-3-yl)pentanamide   xxxiii  LED Light emititng diode LOD Limit of detection LUMO Lowest unoccupied molecular orbital MPA Mercaptopropionic acid MS Mass spectrometry NHS N-hydroxysuccinimide NP Nanoparticle PBS Phosphate buffered saline pChT Chymotrypsinogen PDMS Polydimethylsiloxane PEG Polyethylene glycol PL Photoluminescence POC Point-of-care PON Point-of-need pTRP Trypsinogen QD Quantum Dot QSY35 Quencher (λmax = 476) QSY9 Quencher (λmax = 561) R-PE R-phycoerythrin RGB Red-Green-Blue SAv Streptavidin SDS Sodium dodecyl sulfate SEM Scanning electron microscopy Sub Substrate peptide TB Tris-borate TEM Transmission electron microscopy THR Thrombin TMAH Tetramethylammonium hydroxide TTC 2,3,5-triphenyltetrazolium chloride TRP Trypsin UV Ultraviolet UCNP Upconversion nanoparticle WB White balance XPS X-ray photoelectron spectroscopy ZW Zwitterionic   xxxiv  Acknowledgements I would like to thank my advisor Dr. Algar for the continuous support of my research and his guidance in writing this thesis. It has been a memorable and unique experience to be his first graduate student.  I would like to thank my PhD committee: Dr. Blades, Dr. Bizzotto, and Dr. Berlinguette, for their insightful comments and suggestions. I want to extend my gratitude to Dr. Gethin Owen for assistance with SEM imaging, Dr. Saeid Kamal for assistance and training with microscopy imaging, Dr. Ken Wong for assistance with XPS analysis, Dr. Garnet Martens and Brad Ross for TEM training. Special thanks go to Pritesh Padhiar and Kenny Bach, and all the members of the Mechanical Engineering Shop. Special thanks are also extended to Brian Greene, who recently retired from Electronic Shop, Elena Polishchuk at Biological Services Laboratory, and Helen Wright. I also thank Yun Ling for assistance with mass spectrometry and Emily Seo at the Shared Instrument Facility.  I would like to thank all of my colleagues at the Algar lab and students I had the pleasure to mentor: Huyngki Kim and Olga Solodova. I also would like to thank all of my coathors of my publications for the collaborations. I am deeply grateful to the Natural Sciences and Engineering Research Council of Canada (NSERC) and the University of British Columbia for graduate scholarships, awards, and travel grants. Last but not the least, I would like to thank my family: my parents, my husband, and his parents for continuous love, understanding and support throughout my PhD, writing this thesis and life in general.  1 Chapter 1 Introduction  This chapter is an adaptation of published work, and is reproduced in part from:  (1) Petryayeva, E.; Algar, W. R.; Medintz, I. L., Quantum Dots in Bioanalysis: A Review of Applications Across Various Platforms for Fluorescence Spectroscopy and Imaging. Appl. Spectrosc. 2013, 67 (3), 215-252, with permission from The Society of Applied Spectroscopy (Copyright under Creative Commons Attribution - NonCommercial 4.0 International licence). (2) Petryayeva, E.; Algar, W. R., Toward point-of-care diagnostics with consumer electronic devices: the expanding role of nanoparticles. RSC Adv. 2015, 5 (28), 22256-22282, with permission from The Royal Society of Chemistry (Copyright 2015 The Royal Society of Chemistry).  This thesis presents original research towards the development of fluorescence-based point-of-care diagnostic platforms using smartphone and semiconductor quantum dots. Conceptually, this work represents: (1) the application of luminescent nanomaterials for smartphone imaging; (2) the development of new chemistries for the design of paper-based bioassays; and (3) capitalization on the unique optical properties of quantum dots to enable detection of biomarkers in biological fluids. These concepts are largely developed through model protease assays that rely on Förster resonance energy transfer (FRET). This introductory chapter reviews important background information, including an overview of progress in the field of point-of-care (POC) assays, the current role of smartphones and nanomaterials in the advancing POC devices, a physical description of fluorescence and FRET, and a review of the structure, chemistry and optical properties of semiconductor quantum dots.   2 1.1 Point-of-Care Diagnostics 1.1.1 Consumer electronic devices for POC There is a critical need for point-of-care (POC) diagnostics in health care and a parallel need for similar point-of-need (PON) diagnostics in other sectors. Such technology could have a profoundly positive impact on health, wellness and quality-of-life in both the developed and developing worlds. Over the past decade, a rapidly growing trend has been the design of portable, low-cost bioassays that use consumer electronic devices such as smartphones, digital cameras, scanners, and CD/DVD/Blu-Ray disc players for quantitative readout of results. This trend is a new twist on an older concept embodied by many strip reader devices, which, although not a consumer product per se, can provide low-cost and portable readout of assays. The overarching objective of this research is to enable a full range of point-of-care (POC) diagnostic tests that can improve the efficiency and accessibility of health care globally, and eventually help realize personalized medicine [1, 2]. The technical strategies used to address POC applications are often transferable between health care and other sectors that stand to benefit from rapid on-site or field testing; for example, environment, agriculture/aquaculture, food and water quality assurance, and public safety and security, where such tests tend to be referred to as “point-of-need” (PON). Another important trend over the past decade has been the application of nanotechnology to problems of biomedical and analytical importance, where the unique properties of nanoparticles can increase the efficacy of therapies, improve analytical figures of merit in assays, and provide new opportunities for research and development [3-12]. Not surprisingly, the above trends are converging with exciting outcomes.  The general development and applications of POC/PON devices have been widely reviewed [13, 14] including assays for specific classes of analyte (e.g. microbes [15], cancer biomarkers [16], toxins [17]), specific analysis formats (e.g. paper-based assays and devices [18-22], lateral flow assays [23], lab-on-a-chip [24-26], and centrifugal microfluidic devices [24-26]), specific readout devices (e.g. scanners [27], CD/DVD [28], and Blu-Ray [29] drives, smartphones [1, 30-32]), and specialized areas of development (e.g. paper-based assays with nanoparticles [1, 30-32], immunoassays with nanoparticles [33], and microfluidic assays with gold nanoparticles [34]). The convergence of consumer electronic devices with nanoparticle materials for the development of assays and diagnostics that are amenable to POC/PON settings is illustrated  3 graphically in Figure 1.1. The devices of interest include the aforementioned smartphones, digital cameras, scanners, CD/DVD and Blu-Ray disc players, as well as strip readers and, to a limited extent, blood glucose meters. Nanoparticles of interest include gold nanoparticles, quantum dots, upconversion nanoparticles, silica and polymer nanoparticle composites, viral nanoparticles, and carbon nanoparticles—all of which lend themselves to optical readout. The extensive library of nanoparticles that are currently available offers remarkable choice in selecting materials that can maximize the analytical performance of assays with consumer electronic devices, providing exciting opportunities for current and long-term societal impact in the context of POC/PON assays and diagnostics.    Figure 1.1 An illustration of the convergence of consumer electronic devices and nanoparticles for POC/PON diagnostics.  1.1.2 The need for point-of-care diagnostics The aim of POC diagnostic technology is to provide robust, portable, reliable, rapid, inexpensive and simple testing of clinical biomarkers and other analytes. Minimization of the size, cost and operational complexity of the analysis method and instrumentation is integral to this goal, but achieving the best possible analytical figures of merit is not. Rather, it is sufficient to achieve figures of merit that satisfy clinically relevant thresholds and ranges of analyte. It should also be  4 possible to easily transport and store consumables and devices (if any) at points of care without loss of function. Both the U.S. Food and Drug Administration and the World Health Organization have recommended criteria for POC diagnostics [35-37]. POC diagnostic tests are needed in both developed and developing countries, where the current models of health care delivery are unsustainable, albeit for different reasons. In developed countries, advanced diagnostic technologies and services are readily available through centralized laboratories that can be accessed by the public, typically at the direction of physicians or during hospital stays. The demands on these services and their cost are such that the health care expenditures in developed countries are a growing fiscal burden, amounting to 7–9% of the gross domestic product (GDP) for G7 nations in 2012, with a projected increase to an average of 11% for advanced economies by 2050 [38, 39]. Moreover, rural areas and remote areas of developed countries tend to be underserviced compared to urban centres. Travel to urban centres for medical testing creates extra stress for patients and adds further costs; for example, just in 2010-2011, the northern territory of Nunavut, Canada (pop. 33 000), had more than $72M in health care costs that were associated with travel and out-of-territory services [40]. An array of POC diagnostic technologies that are sufficiently simple, rapid, reliable and economical to be deployed in a physician’s office or in patient’s homes would be a tremendous step toward increasing the efficiency of health care in developed countries.  In developing countries, the problem of accessibility to health care is greatly exacerbated, where large populations may have little or no access to even the basic health services of the developed world due to financial limitations, a shortage of skilled personnel, and a lack of infrastructure [35]. The lack of infrastructure is not only with respect to biomedical and clinical equipment, but may also include running water, refrigeration and electricity. In addition to basic medical tests and screening for chronic disease, affordable test kits for infectious diseases can be a life-saving intervention in many developing countries, where millions die every year due to inadequate diagnosis and these tests could help prevent epidemics from turning into pandemics [35]. Considering PON testing, rapid and low-cost methods of analysis for food safety and water quality, counterfeit medicine, and veterinary testing are also needed [19]. Foodborne illnesses are a direct result of ingestion of food contaminated with pathogens such as Salmonella, E. coli O157:H7, and cholera, the latter of which affects 3–5 million people and kills  5 more than 100 000 each year [41]. Non-existent protocols for water testing in rural areas also remains a major health issue associated with diarrheal diseases [42, 43]. Contamination of drinking water supplies with heavy metals (from industrial and mining production; e.g. mercury from gold mining), agricultural pesticides and fertilizers, sewage, and other wastewater contaminants poses both short-term and long-term health hazards. An estimated 50% of hospital patients worldwide suffer from illness associated with contaminated water [44]. PON diagnostic tests for food and water quality, and for early detection and screening of infectious disease, can improve life expectancy, shorten recovery times and reduce treatment costs [35, 45].  1.1.3 Clinical tests and biomarkers To develop a comprehensive array of POC tests, it is necessary to detect a wide range of biomarkers and analytes with often disparate technical requirements. Ideally, these tests would use a common technology for quantitative readout, and would be able to directly analyze blood, urine, sputum, saliva, and sweat samples—all with little or no user intervention and straightforward readout of results. Most current clinical diagnostics do not meet these criteria.  Common classes of analytes in clinical diagnostics include blood gases (e.g. O2, CO2), pH, electrolytes (e.g. Ca2+, Mg2+, Na+, K+, Cl–), transport proteins (e.g. lipoproteins, ceruloplasmin, transferrin, haptoglobin, haemoglobin), metabolites (e.g. glucose, creatinine, urea, lactate), enzymes (e.g. creatine phosphokinase, alkaline phosphatase, aspartate/alanine aminotransferase), vitamins (e.g. beta-carotene; vitamins A, B12, C and 25-hydroxyvitamin D), hormones (e.g. thyroid stimulating, follicle stimulating, testosterone, estrogen), cytokines (e.g. interleukins; tumour necrosis factor), therapeutic drugs (e.g. digoxin, perhexiline, cyclosporine, tacrolimus), drugs of abuse (e.g. amphetamines, barbiturates, benzodiazepines, cannabinoids, opiates), cardiac and inflammatory markers (e.g. C-reactive protein, troponin I, myoglobin), genes (e.g. BRCA1, BRCA2), and infectious agents and pathogens (e.g. influenza, measles). Conventional laboratory procedures for assaying these analytes are often complex. Sample processing can be labourious and require specialized training, and the analyses often utilize instrumentation that is expensive, non-portable, and operated by skilled technicians; for example, spectrophotometric, electrochemical and chromatographic measurements, molecular biology techniques, cell culture and counting, among many other methods [46]. Notable exceptions to the above are lateral flow assays, which are much more amenable to POC applications, as described in Section 1.1.4.  6 The above classes of analyte can be reduced to three basic groups that can address most diagnostic needs: proteins, nucleic acids, and small molecules. Proteins, whether enzymes, antibodies, certain hormones, cytokines, or otherwise, are frequent targets of POC diagnostics. The RCSB protein databank lists > 100 000 entries from various species [47], and more than 18 000 or 92% of gene-encoded human proteins have been catalogued, as well as many proteins from pseudogenes and non-coding RNA [48-51]. Depending on the target, protein biomarker concentrations typically range from picomolar to micromolar in bodily fluids. In the case of enzymes, their activity may also be of interest in addition to their concentration, in which case a product of that enzymatic activity is measured. When concentration is of interest, the biorecognition elements that are the basis for protein-targeting assays are usually antibodies, aptamers, or ligands that selectively bind to the target protein. Certainly, the standard format for protein detection is an immunoassay such as an enzyme-linked immunosorbent assay (ELISA), which has been one of the most prominent clinical laboratory tests over the past 20 years. As biorecognition elements, antibodies have remained indispensible because of their specificity and affinity but have several potential drawbacks, including limited stability, batch-to-batch variability, and high production costs. In many ways, aptamers are preferable biorecognition elements for POC/PON assays because they are more robust and more economically produced [52, 53]; however, aptamers that have affinity comparable to antibodies are not yet known for many target proteins.  Nucleic acid assays are increasingly important for the diagnosis of disease, identification of pathogens, and identification of genetic conditions and predispositions. Genes can be useful biomarkers for organisms and their physiology, and are sometimes preferable biomarkers over the proteins that they encode. Many thousands of genomes from eukaryotes, prokaryotes and viruses have been completely or partially sequenced [54], including the estimated 19 000 protein-coding human genes [55]. The analysis of DNA and RNA directly from bodily fluids is limited by its very small amount (e.g. 102–1011 copies per 1 mL of blood), necessitating multiple preparatory steps prior to analysis (e.g., extraction, purification and amplification). Conventionally, polymerase chain reaction (PCR) is used for amplification of DNA. The thermocycling inherent to this process is not ideal for POC/PON applications, although computer-based thermocycling is possible [56]. Alternatively, there are now amplification  7 techniques that do not require thermocycling and may thus be more compatible with POC/PON applications. As reviewed recently [57], these isothermal techniques include loop-mediated amplification (LAMP), rolling circle amplification (RCA), nucleic acid sequence-based amplification (NASBA), signal mediated amplification of RNA technology (SMART), and nicking endonuclease signal amplification (NENNA). Synthetic oligonucleotides that are complementary to the sequence of target genes are typically used as biorecognition elements, and are robust and relatively inexpensive. The most common small molecule targets for diagnostics are metabolites such as hormones, vitamins, amino acids, sugars, and other small organic molecules. The Human Metabolome Database lists ca. 42 000 entries [58, 59], and many of these can serve as indicators of disease. Similar to proteins, some metabolites are found at high concentrations (> 1 mM), at low concentrations (< 1 nM), and concentrations in between. The most common biorecognition element for metabolites are antibodies, although the small size and similar chemical structures of these analytes often yield limited sensitivity and specificity, including cross-reactivity, that makes immunoassay-based detection challenging. Aptamers are again promising alternatives to antibodies in these assays, but are still limited by the number of aptamers available and their affinity for their small molecule targets. Another diagnostic test of interest is the detection of specific cell types and microorganisms; for example, pathogens that cause disease. Six common types of pathogens include viruses, bacteria, fungi, prions, protozoans and parasites. Infectious diseases contribute to more than 95% of all death in developing countries, and include human immunodeficiency virus (HIV), malaria (Plasmodium parasite), tuberculosis (Mycobacterium tuberculosis), and hepatitis A/B virus [60]. Moreover, at the time of writing, the worst recorded Ebola virus outbreak in history has infected >8 000 people and killed more than 5 000 people, mostly in West Africa [61]. Although pathogen outbreaks in the developed world are relatively rare and generally minor, there are nonetheless recurring instances of contamination of food with Salmonella, Staphylococcus aureus, Listeria monocytogenes and E.coli O157:H7 pathogens [62, 63]. The gold standard methods for detecting pathogens are culture-based assays that provide good sensitivity and selectivity, but require long incubation times that limit rapid responses to outbreaks [64]. An  8 alternative strategy for more rapid analysis is to target protein, nucleic acid, and small molecule biomarkers that are pathogen specific. The above discussion on health-related targets for diagnostics is by no means comprehensive, and there is also significant demand for molecular diagnostic tests beyond health care. For example, public service employees are often tested for illegal recreational drugs, and elite athletes are tested for performance-enhancing substances. Many heavy metals and small molecules such as pesticides and other toxic pollutants are important targets in environmental analysis [65], while rapid monitoring and diagnostic tests are also valuable tools for biofuel production and other non-health areas of the biotechnology sector [66, 67]. Many of the approaches and challenges described for health-related diagnostics are equally applicable to these other sectors, and vice versa. 1.1.4 Lateral flow assays Lateral flow assays have been one of the most successful and POC-amenable formats since the introduction of lateral flow immunochromatographic assays in 1988 by Unipath [68]. This format combines concepts from paper chromatography and immunosorbent assays. It frequently does not require the additional washing steps of the latter and typically needs only 0.1 mL of sample. LFA devices are suitable for the direct analysis of blood samples as plasma components are separated from blood cells within minutes, and are also suitable for the analysis of urine and other bodily fluids. Routinely used, commercially available POC immunochromatographic tests include those for pregnancy (human chorionic gonadotropin level) and ovulation; infectious diseases (e.g. malaria, influenza, HIV); drugs of abuse (e.g. NIDA-5 panel for cannabinoids, cocaine, amphetamines, opiates and phencyclidine); and cardiac biomarkers (e.g. troponin I, creatine kinase-MB, myoglobin). Lateral flow immunochromatographic assay strips consist of a sample application pad, a conjugate pad, a membrane (e.g. nitrocellulose, cellulose), and an absorbent pad, as shown in Figure 1.2 [17, 19, 69]. Reporter antibodies conjugated with a contrast-providing reagent (dye-stained latex beads originally [70]) are deposited but not immobilized on the conjugate pad. A fluid sample is applied to the sample pad and wicks down the length of the test strip. As the sample passes through the conjugate pad, the contrast reagent-reporter antibody conjugates bind  9 to the target analyte. Further along the strip, the target analyte also binds to capture antibodies immobilized in the test zone, resulting in retention of the contrast label. Colour imparted to the test zone by the contrast label indicates the presence of target analyte in the sample. A control zone also tends to be included on the membrane, and this zone contains antibodies that bind to the reporter antibody. The absorbent pad ensures steady wicking of the sample fluid along the test strip. Many variations of this general assay design are possible; common variations include substitution of antibodies with other biorecognition elements, or the use of a competitive assay format rather than a sandwich assay format.   Figure 1.2 Basic design of a lateral flow immunochromatographic assay (LFIA). The device comprises a sample pad, conjugate pad, detection zone with test (T) and control (C) lines, and an absorbent pad. The sample containing analyte is added to the sample pad and drawn towards the absorbent pad by capillary action. NP-antibody conjugates bind to analyte (antigen) present in the sample and are captured on the test line, whereas NP-antibody conjugates that have not bound antigen are captured on the control line.     10 Although dye-stained latex beads are still used to generate contrast in lateral flow assays, many commercial assays now use gold nanoparticles, and additional nanoparticle materials have been investigated as contrast reagents in research toward new or improved diagnostics. There is also an increasing demand for quantitative rather than qualitative results from lateral flow assays, and this quantitation is often possible through analysis of digital images instead of simple visual inspection.  1.1.5 Utility of consumer electronic devices Whether in the developed world or the developing world, modern consumer electronic devices can help address challenges in POC/PON testing in three principal ways: (i) lower equipment and infrastructure costs; (ii) miniaturization and portability; and (iii) data processing, storage and communication. The most common devices, which include scanners, CD/DVD and Blu-Ray players, web cams, cell phones and smartphones, offer the foregoing benefits to different degrees and have different suitability for the developed world versus the developing world. In the developed world, all of the above devices share the benefit of low cost, which arises from their mass production and a highly competitive marketplace. Prices typically range from $10–$1000 depending on the device and much of the developed world already owns one or more of these devices. In the United States, for example, ownership statistics are 80% for DVD/Blu-Ray players, 64% for laptop computers, 57% for desktop computers, 45% for cell phones, and 62% for a smartphone (2013 data) [71]. In the developing world, one wishes to discuss cost in terms of cents rather than dollars; however, it must be recognized that the main role of consumer electronic devices in a POC/PON test will be quantitative readout and data handling, and these devices remain among the best candidates for establishing a frontline of health care infrastructure, particularly if the corresponding consumables for diagnostic tests cost pennies and also support qualitative assessment when these devices are not available. Moreover, these devices are not beyond the reach of the developing world as some people in these countries have easier access to mobile phone technology than they do to clean water [72].  From a technical perspective, scanners offer large-area colour imaging with reproducible positioning and illumination. Many current models of scanners are compact, support wireless communication, and can be fully operated via a USB connection to a laptop or notebook  11 computer. Disc players are common household items and optical drives are widely available as built-in or peripheral components of laptop/notebook computers. Discs also offer a substrate for arraying assay zones (e.g. microarray format) and integrating microfluidic channels, while disc players and drives offer optical readout and spinning motion that provides a centripetal force suitable for driving fluid flow. Cell phones, and later smartphones (the distinction being that smartphones have an operating system), have undergone remarkable technological growth over the last two decades. The first cell phone, the Motorola DynaTAC, became commercially available in 1984. It weighed 790 g and was 25 cm in length with a price of $4 000 (ca. $10 000 in 2016 dollars) [73]. The current generation of smartphones, such as the best-selling Samsung Galaxy and Apple iPhone models, offer immensely greater capabilities at a fraction of the price and a fraction of the size ($600–$1000, 130 g). These capabilities include high-quality built-in cameras, multiple modes of wireless communication (e.g. WiFi®, Bluetooth, LTE), global positioning systems, security features, excellent data storage capacity, processing and graphics power to support software applications (apps), and many hours of battery life.  In the context of POC/PON diagnostics, smartphones are leading candidates to fulfill the role of computers in modern laboratory instrumentation, with lower cost and greater portability than notebook/laptop computers (which have built-in webcams and similar wireless connectivity). Furthermore, although a POC/PON test may be simple enough for a minimally skilled technician or unskilled person to conduct, determining a diagnosis or prognosis from test results may not always be as straightforward. Smartphone apps can potentially automate sample logging and data processing, and store or send results for subsequent interpretation by medical professionals or other highly-skilled personnel, whether locally or across the world. Importantly, these devices are also globally ubiquitous with 1.5 billion mobile telecommunications subscribers in the developed world and 5.4 billion in the developing world [74], albeit that the latter are primarily cell phone users rather than smartphone users. 1.1.6 Optical properties of nanoparticles While consumer electronics can provide a means of assay readout, these devices cannot generate the readout signal or contrast themselves. These signals must come from selective recognition chemistry that is directly or indirectly coupled to a physical process that generates a measurable output. Nanoparticles (NPs) can be used for the generation of these signals and provide  12 enhancements or advantages over molecular reagents. By definition, NPs are particles that are less than 100 nm in their largest dimension [75], although here the definition is stretched to include particles with dimensions of hundreds of nanometers. The small size and molecule-like diffusion of NPs is complemented by large surface area-to-volume ratios, interfaces that can be further functionalized, and, in the case of many NP materials, size-dependent properties that are either not observed with their bulk analogues or are significantly enhanced. Table 1.1 briefly summarizes the key features of some NP materials that are currently used in POC/PON diagnostics or which are promising candidates for future use.  From the standpoint of optical diagnostics, there are many NP materials of interest. At present, the three most common materials are gold NPs (Au NPs), quantum dots (QDs), and lanthanide-based upconversion nanoparticles (UCNPs). Au NPs have plasmon bands in their UV-visible absorption spectrum and exhibit strong light scattering that increases with increasing NP size (Figure 1.3A) [3, 5, 76]. These properties manifest as an intense red colour for assay readout with sensitivity that typically exceeds that of dyes and other materials. Although this red colouration can be seen by the naked eye and provide sufficient sensitivity for many assays, LODs can be improved by orders of magnitude with signal amplification strategies such as silver enhancement, which increase the optical contrast of the test zone. The silver enhancement strategy was popularized by Mirkin and coworkers [77] and relies on the reductive deposition of silver on Au NPs, thereby increasing the nanoparticle size and extinction coefficient, darkening their macroscopic appearance on a white background. The first and compelling demonstration of this strategy was the detection of 50 fM of target DNA with scanner readout [77]. Silver enhancement remains commonplace for POC diagnostic assays with scanners, as well as other consumer electronic devices. Alternatively, amorphous carbon NPs appear black and, like silver-enhanced Au NPs, provide high contrast under white-light illumination [78].      13 Table 1.1 NP materials for POC/PON diagnostics. NP Material a Approx. Size (nm) Optical Readout Features b POC Usage CMOS c SCN OD SR Au NPs Gold 5–200 High optical density; intense red colour; silver amplification High ✔[79-83] ✔[77, 84-90] ✔[91-98] ✔[99-102] Polymer NPs Polystyrene 10–1000 Properties of dopant/cargo molecule/NP (e.g., QDs, lanthanide complexes) High d ✔[103] u u ✔[57, 104, 105] Amorph. Carbon NPs Carbon < 1000 (irregular) High optical density Moderate u ✔[106-109] u u QDs CdSe/ZnS CdSeS/ZnS 3–10 Bright, tunable and spectrally narrow PL; spectrally broad light absorption Moderate ✔[110] -- -- ✔[57, 105, 111-113] UCNPs NaYF4:Yb doped with Eu3+, Tb3+, Ho3+ 20–50 Upconversion PL; spectrally narrow PL Moderate u -- -- ✔[114-116] Silica NPs Silica 10–500 Properties of dopant/cargo molecule/NP  (e.g., Au NPs, lanthanide complexes) Low u u u ✔[99, 117] Viral NPs Protein 10–1000 Properties of dopant/cargo molecule/NP Low u ✔[118] u u Carbon dots Carbon 2–6 Bright, spectrally broad PL Future? u -- -- u Pdots π-conjugated polymers 5–50 Bright PL from very strong light absorption; composites with other optically-active materials Future? u -- -- u Legend: CMOS, device with CMOS image sensor; OD, optical drive for disc player for CDs/DVDs/BRDs; SCN, scanner; SR, strip reader; ✔, current use; u, possible or probable future use. Notes: a Typical materials listed. There are many possible materials for QDs and polymer NPs. b NPs may have properties beyond those listed here. c Includes cell phones, smartphones, digital cameras, wearable technology, etc. d There are many uncited examples of polymer NPs as carriers for dye molecules.  14 QDs exhibit bright, size-dependent photoluminescence (PL) that is easily excited and can be tuned across a wide spectral range (Figure 1.3B) [6, 7]. The optical properties of QDs, which are discussed in more detail in Section 1.3, are generally considered to be superior to those of fluorescent dyes: their light absorption is much stronger and more spectrally broad, and their emission is much more spectrally narrow and resistant to photobleaching. Lanthanide-based upconversion nanoparticles (UCNPs) convert near-infrared (NIR) excitation into visible emission, the colour of which depends on their composition (Figure 1.3C) [9-11]. Upconversion is not possible with fluorescent dyes. With both QDs and UCNPs, their light emission against a dark background provides contrast for assay readout. Further properties of QDs that make them advantageous for POC/PON assays with consumer electronics, and in comparison to fluorescent dyes and other materials, are described in detail in Section 1.3. More generally, the benefits of fluorescence detection in POC/PON assays include potentially greater sensitivity and lower LODs, potentially greater tolerance of sample colouration, and new possibilities for multiplexed analyses. The trade-off is that fluorescence measurements are somewhat more technically demanding, requiring an excitation light source and readout against a dark background (i.e. exclusion of ambient light). Fortunately, POC/PON-amenable light sources are widely available (see Section 1.1.7.1) and 3D printing provides a convenient means of producing ambient light-blocking enclosures or attachments for smartphones or other complementary-metal-oxide-semiconductor (CMOS) image sensor devices. Other NP materials that are of interest in context of POC/PON diagnostics are silica NPs [119-121] and polymer NPs [122, 123]. However, in contrast to the foregoing materials, it is their physical properties that are of interest rather than their optical properties. These NPs, which can have dimensions of hundreds of nanometers, can serve as carriers of molecules or smaller NPs that provide contrast (e.g. Au NPs, QDs, UCNPs). Viral NPs and genome-free virus-like NPs (collectively, VNPs) can also serve as carriers of contrast reagents, with the benefit of being monodisperse, tailorable through genetic engineering and chemical functionalization, and producible at a large scale [124, 125]. Analogous to the original use of latex beads as a carrier for dye molecules, the concept is that many NP contrast reagents can be associated with a single binding event even though there will be no more than one carrier particle per binding event, resulting in greater sensitivity.   15  Figure 1.3 (A) Size-dependent molar extinction coefficient of Au NPs as a function of wavelength. The inset photographs show solutions of 5, 10, 15 and 20 nm Au NPs at 1 nM concentration and with an optical density of 1 (ca. 90, 10, 3, and 1 nM concentrations for 5, 10, 15 and 20 nm Au NPs, respectively). (B) Size/composition-tunable absorbance and emission of CdSe/ZnS and CdSeS/ZnS QDs. The inset photograph shows samples of different sizes of CdSe/ZnS QDs under UVA (365 nm) illumination. Photograph reprinted with permission from ref. [126]. Copyright 2011 American Chemical Society. (C) Upconversion emission spectra of NaYF4:Yb/Tm (20/0.2 mol %; blue line) and NaYF4:Yb/Er (18/2 mol %; green line) nanoparticles. The inset photographs show samples of these nanoparticles under 980 nm excitation with a diode laser (600 mW). Adopted with permission from ref. [127]. Copyright 2008 American Chemical Society.  16 Several other optically-active NP materials are known and, to the best of our knowledge, have yet to be utilized for POC/PON assays with consumer electronic devices. For example, carbon nanotubes [128, 129], graphene oxide [130, 131], carbon dots [132, 133], and nanodiamonds [134, 135] exhibit PL that can be useful for biological imaging and assays; however, it is not clear that the characteristics of this PL (e.g. brightness, spectral range of absorption and emission) is well-suited to readout with consumer electronic devices. On the other hand, semiconducting polymer nanoparticles (Pdots) [136] have exceptionally bright PL that is certainly promising for readout with consumer electronic devices, and the lack of examples to date is likely a product of the novelty of the materials. Beyond optical properties, many NP materials also have magnetic or electrochemical properties that are of interest.  1.1.7 Bioassays with consumer electronics and nanoparticles This section describes consumer electronic components and devices that are being actively developed as platforms for readout of POC/PON diagnostics and assays. Light-emitting electronic components, which are common to all of the assays considered, are first reviewed, then, the basic design elements and functional principles underlying digital imaging utility as a readout platform are described, followed by examples of assays that use this device in combination with NPs for readout of results.  1.1.7.1 Light sources Common light sources for POC/PON assays with NPs include white or coloured light-emitting diodes (LEDs), laser diodes and, to a lesser extent, hand-held ultraviolet (UV) lamps or “black lights.” These light sources permeate the developed world and are available at low price points. LEDs are ubiquitous as indicator lights and display backlights in electronic devices, in traffic signals and signage, and in both decorative and ambient lighting products. Laser diodes are critical components of optical drives/disc players, printers, barcode scanners, manufacturing technology, telecommunication systems, and are also used in medicine and dentistry. Hand-held UV lamps that emit long-wavelength UVA light have been traditionally used for forgery detection (e.g. monetary bills, documents) but are gradually being replaced by LEDs that emit in the same spectral range. All of these light sources can be battery-operated for extended periods, which is a critical consideration for use in POC/PON applications.   17 Of the above light sources, LEDs are the most economical ($0.01–$1.00 typical) and the most amenable to miniaturization (millimeter dimensions). LEDs usually have low operating voltages (3–5 V) and low power consumption (~10–3–10–2 W), although higher-power LEDs (>10–1 W) are available at greater cost than noted above. Low-cost, low-power LEDs are the most relevant to POC/PON applications. The emission from an LED is incoherent and distributed over a relatively wide angular range. Its peak emission wavelength, which may be in the UV, visible or infrared region of the spectrum, is determined by the semiconductor composition of the diode. Spectral full-widths-at-half-maxima (FWHM) are typically in the range of ca. 15–50 nm. Representative examples of some low-cost LED spectra are shown in Figure 1.4A. Whereas colour LEDs (particularly blue and UV wavelengths) are well-suited to readout of photoluminescence, white-light LEDs are well-suited to colorimetric readout. Most white-light LEDs are actually blue LEDs with a phosphor coating that has broadband emission in the green-red region of the spectrum, as shown in Figure 1.4B. Colorimetric readout is also possible with a combination of red, green and blue LEDs. Diode lasers provide more intense illumination than LEDs and have coherent, monochromatic emission (FHWM < 1 nm). From the perspective of POC/PON applications, laser diodes of the type found in laser pointers (~10–3 W) and optical disc drives and players (~10–1 W) are the most relevant. Figure 1.4C shows the emission from laser diodes that are commonly used for excitation of photoluminescence, including violet (405 nm), blue (447 nm), green (532 nm), red (650 nm) and infrared (980 nm) wavelengths. Note that many green laser diodes are actually infrared laser diodes that have been frequency doubled and fitted with an IR-blocking filter (DPSS lasers). Hand-held, battery-operated UVA lights are another light source that is potentially suitable for POC/PON applications. These sources are low-pressure mercury discharge lamps where a phosphor converts the 254 nm emission from mercury to 365 nm emission from the lamp. A coating on the quartz tube absorbs any visible light. Power consumption is typically on the order of a few watts. The principal benefit of these sources is that relatively large areas can be illuminated with spectrally narrow light (FWHM ~15 nm, Figure 1.4D). Although “mini” or “pen” lamps are commercially available, UVA lights are less amenable to miniaturization than LEDs or diode lasers.  18  Figure 1.4 Emission spectra of commercial light sources well-suited to POC/PON diagnostics: (A) various colour LEDs emitting in UV-visible region of the spectrum; (B) white-light LED; (C) five common wavelengths of laser diodes (the FWHM > 1 nm is a measurement artifact); and (D) and a UVA lamp or “black light.”  1.1.7.2 CMOS image sensors: digital cameras to smartphones 1.1.7.2.1 Digital imaging technology Modern CMOS image sensors are compact, provide high image quality, and are widely incorporated into consumer devices such as cell phone and smartphone cameras; webcams; wearable technology (e.g. Google Glass, Sony’s SmartEyeglass); and digital cameras for traditional photography, hobbies (e.g. Raspberry Pi), and recreational activities (e.g. GoPro). A selection of these devices is shown in Figure 1.5A. CMOS technology has also permeated scientific research in the form of microscopy cameras. The primary advantages of CMOS sensors over charge-coupled device (CCD) sensors are full integration of circuitry (which is more amenable to miniaturization), lower power consumption, and faster frame rates. Originally, these advantages were at the expense of image quality; however, improvements in fabrication technology and consumer demands for increased performance from their mobile devices have driven the advancement of CMOS technology to its current pinnacle.   19  Figure 1.5 (A) Examples of consumer electronic devices equipped with CMOS cameras: (i) smartphones; (ii) digital cameras; and (iii) wearable devices. The image in (iii) is reproduced with permission from ref. [137]. Copyright 2014 American Chemical Society. (B) Simplified schematic of a CMOS image sensor. (C) Spectral sensitivity of a typical CMOS image sensor without (black) and with RGB colour filters (coloured lines). The typical blocking region of an IR filter is also shown.  As shown in Figure 1.5B, a CMOS image sensor has two main optical components: a pixel sensor array and optical filters for colour transmission (e.g. Bayer filter) and blocking UV and IR light.  Millions of pixels capture light and convert that light into a voltage proportional to its intensity, where each pixel has its own amplification and digitization circuitry. CMOS sensors have wavelength-dependent sensitivity between ca. 380–1100 nm, as shown in Figure 1.5C. Colour information is obtained by superimposing an array of bandpass filters that transmit either blue, green or red light on the pixel array. The most common filter pattern is the “Bayer mosaic,”  20 which is a repetitive 2 × 2 grid with one red filter, two green filters, and one blue filter per four pixels. This ratio of filters was designed to mimic the human eye’s greater sensitivity to green light [138]. IR and UV filters can be added to block unwanted wavelengths of light from outside the visible spectrum. Electronics and demosaicing algorithms convert the pixel signals into digital colour images. 1.1.7.2.2 Growing analytical applications CMOS-based cameras, especially those in smartphones, have emerged as promising tools for heath care and bioanalysis over the past few years. In particular, smartphones have been suggested to be amenable to POC diagnostics and telemedicine and Martinez et al. were among the first to suggest the use of a smartphone camera for quantitative analysis of colourimetric assays [139].  Numerous smartphone apps and accessories have become available to assist the general public with basic health monitoring; for example, heart rate [140, 141], blood pressure [140], body mass index [140], and detection of ear infections [142] and potential skin cancer [143, 144]. Initial evidence suggests that smartphone technology can support better health outcomes, as demonstrated with apps that promote physical activity and weight loss [145, 146]. Smartphone imaging has also been investigated as a POC/PON readout platform for molecular diagnostics such as immunoassays [147-149], nucleic acid hybridization assays [150], and colorimetric assays for cholesterol [151], food allergens  [152], enzymes [153], and various urinary, salivary and sweat biomarkers [154, 155]. Digital images acquired with a smartphone camera can be analysed to extract quantitative information, most frequently in terms of the grayscale or RGB colour intensities for pixels of interest. These analyses can be done with computer-based image analysis software designed for either scientific research or consumer use (e.g. Adobe Photoshop), including freely available software (e.g. ImageJ), as well as smartphone apps. Analytical and biological applications of cell phone and smartphone cameras also go beyond macroscopic digital photography. For example, when these cameras are combined with additional optics, they can be used for dark-field and bright-field microscopy of cells [156]. Imaging of a single fluorescent polystyrene NP (100 nm diameter) has also been demonstrated by the Ozcan Laboratory using a smartphone (Nokia PureView 808) [157]. The phone was equipped with a high-resolution CMOS sensor (41 MP) and utilized oversampling technology  21 (i.e. pixel binning) that enabled capture of five times more light than a typical zoom camera. A compact attachment to the phone was fabricated using 3D printing technology; it integrated a 405 nm laser diode excitation source (75 mW) powered by three 1.5V batteries (AAA size), a longpass filter to remove scattered excitation light, a 2× magnification lens, and optomechanics for focus adjustment (Figure 1.6A).    Figure 1.6 (A) Cell phone-based fluorescence imaging of individual NPs and viruses: (i) Front view of the smartphone microscope and a schematic diagram of its components; (ii) Images of 100 nm fluorescent NPs acquired with the cell phone show excellent agreement with SEM images. Adapted with permission from ref. [157]. Copyright 2013 American Chemical Society. (B) Attachment that enables use of a smartphone as a spectrophotometer for fluorescence emission measurements. The key component is a transmission diffraction grating. Reprinted with permission from ref. [158]. Copyright 2014 American Chemical Society.  22 Smartphone cameras can also be used as spectrographs [159]. As an example, the Cunningham Laboratory developed a simple transmission grating interface for a smartphone camera (iPhone 4) that enabled acquisition of full fluorescence emission spectra (Figure 1.6B) [158]. The diffraction grating (1200 lines/mm) dispersed fluorescence excited with a green diode laser pointer onto the camera. Due to the built-in UV and IR blocking filters, the smartphone camera-spectrophotometer was sensitive over the spectral range ca. 400–700 nm with a spectral dispersion of ~0.3 nm/pixel. Tests with a molecular beacon assay for microRNA demonstrated a limit of detection (LOD) of 1.3 pM, which was superior to a 3.6 nM LOD obtained with a conventional spectrofluorimeter. The observed enhancement was a combined effect of the greater quantum efficiency of the CMOS sensor in the smartphone versus the photomultiplier tube (PMT) in the spectrofluorimeter (40% vs. 12%, not accounting for PMT amplification), as well as a more than 30 000-fold increase in excitation efficiency. The latter was a result of the greater output power of the diode laser source (~300 mW) versus the xenon lamp in the spectrofluorimeter (10 µW), and a more than 300-fold smaller illumination volume with laser excitation. The Dana Laboratory has demonstrated that a smartphone camera can also be integrated into a confocal Raman system for detection of the surface enhanced Raman scattering spectrum from ethanol [160]. Raman spectra were acquired with green laser excitation (532 nm, 10 mW) by placing a collimator and transmission grating in front of the camera sensor, which had overall sensitivity comparable to CCD and PMT detectors. Observation of blinking events from single molecules diffusing in and out of hot spots on a silver nano-island plasmonic substrate were observed with the smartphone at 30 fps video recording. Smartphones can also serve as platforms for surface plasmon resonance (SPR)-based assays. Preechaburana et al. designed a disposable device that used a smartphone (iPhone 4) display screen as a light source and used its user-facing camera to measure reflectivity [161]. The SPR coupler was made from polydimethylsiloxane (PDMS) and epoxy to gently adhere to the phone’s screen, which in turn displayed a guide for alignment with a red rectangle that provided illumination. Image acquisition was done using a custom app that allowed for control of exposure time and ISO number (i.e. sensitivity level to light). This platform was able to detect β2 microglobulin (β2M), a biomarker for cancer, kidney disease and inflammatory disease, over a clinically relevant range of concentrations with an LOD of 0.1 µg mL–1 [161].  23 In addition to cell phones and smartphones, CMOS-based digital imaging assays run the gamut of technology from conventional digital cameras to new wearable devices. For example, Deiss et al. recently developed low-cost, portable paper-based culture devices for the analysis of antimicrobial susceptibility using a digital photography camera (Canon EOS Rebel T3i) for readout [162], whereas the Ozcan Laboratory demonstrated the use of Google Glass for readout of LFS immunoassay results, identification of sample codes, and transmission, analysis and storage of the results using hands-free voice operation [137].  Many of the CMOS-imaging assays that are being developed utilize a growing array of NP materials. To date, the most common materials include Au NPs for colourimetric detection, as was the case for the Google Glass example noted above, as well as QDs and UCNPs for fluorescence detection. Fluorescence-based assays offer a number of advantages in comparison to colorimetric assays, including improved selectivity and sensitivity. An overview of the fluorescence process and mechanism of resonance energy transfer are described in the following section.           24 1.2 Fluorescence Luminescence is the emission of ultraviolet, visible or infrared radiation by any atom, molecule or lattice. If emission of light is a result of photon absorption, the process is termed photoluminescence (PL). Other types of luminescence include chemiluminescence, bioluminescence, and electroluminescence, as result from a chemical reaction, biochemical reaction, and passage of electrical current, respectively. A particular type of PL where emission of a photon is a result of an electronic transition from an excited state to a ground state without a change in spin multiplicity is termed fluorescence.  1.2.1 The Perrin–Jablonski diagram  Fluorescence is a photon emission process that takes place during molecular relaxation from electronic excited states to the ground state. This process involves transitions between both electronic and vibrational states of a polyatomic fluorescent molecule or fluorophore. The multi-step process of fluorescence is accompanied by a number of competing processes and is best described with introduction of a Jablonski diagram. As shown in Figure 1.7, the electronic states of a molecule are represented by potential energy wells. Within each electronic state, there are multiple vibrational energy states, each of which can be subdivided even further into rotational energy levels. For simplicity and clarity, only a few vibrational states (νn) are depicted in the diagram, and rotational levels are omitted entirely. Electronic states are typically separated by energies on the order of 1–10 eV (ca. 10 000–80 000 cm–1) [163]. The separation between vibrational levels is typically on the order of 0.25 eV (ca. 2 000 cm–1) [163]. Photons of light in the ultraviolet–visible region of the spectrum (200–800 nm, 6.2–1.6 eV) may trigger an electronic transition, as discussed in detail in Section 1.2.2.    25  Figure 1.7 Jablonski diagrams illustrating the processes of absorption and photoluminescence, as well as competing non-radiative relaxation processes. Singlet electronic states are labeled as Sn, and triplet electronic states are labeled as Tn. The electronic states are shown as potential energy wells with superimposed vibrational states, νn. (A) An example of processes involved in a radiative pathway. (i) Absorption from Soνo to S1ν2, followed by (ii) vibrational relaxation to S1νo and (iii) radiative relaxation to the ground state—fluorescence. (B) Processes involved in non-radiative pathways. Absorption from Soνo to S1ν2 (i) or S2ν4 (ii), followed by (iii) vibrational relaxation to S2ν0 and (iv) internal conversion to S1ν7. This process is followed by vibrational relaxation to S1νo and, from this state, an emission of photon can take place as shown in (A, iii) or by (v) internal conversion—a transition to Soν20, followed by (iii) vibrational relaxation to the Soνo. (C) Other non-radiative relaxation pathways: (i) excitation to S1ν2, followed by (ii) vibrational relaxation to S1νo, and then (iii) intersystem crossing to T1ν1. This process is followed by vibrational relaxation to T1νo. From this state molecules return to the ground state either via non-radiative pathways or via radiative processes—delayed fluorescence or phosphorescence.   The simplest model used to describe a three-step fluorescence process is based on a system that comprises a singlet ground (S0), and a singlet excited (S1) states. First, a molecule absorbs a photon of appropriate wavelength and one of the electrons undergoes a transition from the  26 ground state, S0, associated with the highest occupied molecular orbital (HOMO) to an excited singlet state, S1, associated with lowest unoccupied molecular orbital (LUMO). This transition is very rapid and occurs on the femtosecond timescale. In the second step, the electron in the excited state undergoes rapid vibrational relaxation to the lowest vibrational state of S1. As a result, some of the energy deposited into the molecule by the photon is dissipated to the surroundings as heat. Vibrational relaxation takes place on a picosend timescale. The third and final step is the return of electron from energetically unstable excited state, S1 to any vibrational level of singlet ground state, S0, via emission of a photon (i.e. fluorescence). This process takes place on the sub-nano and nanosecond timescale. As a result of the differences in the timescales between fluorescence and vibrational relaxation processes, fluorescence almost always occurs from the lowest vibrational level of the first excited singlet state (S1). This phenomenon, known as Kasha’s rule [164], leads to an emission spectrum that is virtually identical irrespective of the excitation wavelength. All of the processes introduced above, as well as other competing relaxation processes shown in Figure 1.7, are discussed in detail in following sub-sections. 1.2.2 Absorption–formation of the electronic excited state Absorption of a photon by a molecule transforms light energy into electronic potential energy, thus promoting an electron from an occupied molecular orbital to an unoccupied molecular orbital of greater energy. Common types of electronic transitions include n → π* and π → π*, where n, π, and π* are non-bonding, pi bonding, and pi antibonding orbitals, respectively. These transitions are characteristic of conjugated pi electron systems, where the energy of the transition decreases as the extent of the conjugation increases [165]. The functional group of a molecule that absorbs light is known as a chromophore. The probability of light being absorbed by a chromophore is determined by a number of selection rules, as outlined below. I. Resonance condition  Absorption occurs when the frequency of an incident photon matches the natural frequency of the transition from the lowest vibrational level (ν0) of the ground state (S0) to any of the  27 vibrational levels (νm) of the excited state (Sn), as given by eqn. 1.1, where f is the resonant frequency of photon, h is Planck’s constant, and ΔE is the energy of transition. Ephoton = hf = ΔEtransition = E Snνm( )−E S0ν0( )                                         (1.1) II. Photoselection principle  An electronic transition from the HOMO to the LUMO alters the spatial electron distribution, often accompanied by a physical shift of the electron density, making the total electron density more diffuse. The transition moment is a vector that represents the transient dipole of charge displacement during this transition. Chromophores preferentially absorb polarized light with the electric field vector aligned along the transition moment. The probability of absorption is proportional to the square of the scalar product of the electric field vector of incident light, E , and the transition dipole vector of the chromophore, µ , [165]: PAbs ∝ E ⋅µ( )2= E 2µ 2 cos2α                                                    (1.2) where α  is the angle between vectors E  and µ . According to eqn. 1.2, the maximum probability of light absorption occurs when light is polarized parallel to the transition dipole vector, and no absorption occurs when two vectors are perpendicular.  III. Electronic Selection rules These selection rules are based on the conservation of angular momentum during a transition and determine whether transitions are observed (i.e. allowed transitions) or not (i.e. forbidden transitions). The following conditions must be satisfied for allowed transitions [166]: 1. Total orbital angular momentum quantum number, Λ, satisfiesΔΛ = 0,±1 2. Total spin angular momentum quantum number, Σ, satisfies ΔΣ = 0  3. Total angular momentum quantum number, Ω, where Ω = Λ +Σ , is the quantum number for the component of total angular momentum (orbital and spin), satisfies ΔΩ = 0,±1 . 4. Spin does not change: ΔS = 0   28 The rule (4) states that electronic transitions between states of different multiplicities (e.g. between singlet and triplet states) are forbidden. However, these transitions do have a probability of occurring via spin-orbit coupling, a process that causes mixing of the singlet and triplet wavefunctions. Fundamentally, spin-orbit coupling is the result of the interactions of two magnetic moments arising from an electron spinning around a nucleus and about its own axis.  Other selection rules are based on changes in symmetry. The Laporte selection rule states that for centrosymmetric molecules, transitions can only occur if there is a change in parity—a parameter related to the orbital angular momentum summation over all electrons [167]. In practice, molecular vibrations can distort symmetry and Laporte forbidden transitions become partially relaxed [167]. 1.2.2.1 Absorption spectrum: shape and intensity The light-induced transitions of chromophores to their excited states involve simultaneous electronic and vibrational transitions. The electronic levels determine the position of the absorption bands (energy or wavelength) and the vibrational and rotational levels define the band shape. Therefore, the magnitude and the shape of the absorption bands depend on the probability of each separate transition between the ground state and vibrational levels of the excited states. At room temperature, the initial position of an electron in the ground electronic state is assumed to be the lowest vibrational level (νn = 0). The Boltzmann distribution function, eqn. 1.3, describes the occupancy of the vibrational states in a molecule: Nν=n+1Nν=n= e −ΔEvib kbT( )                                                                 (1.3) where N is the number of molecules in the corresponding vibrational state (νn = 0, 1), ΔEvib is the energy difference between two vibrational states, kB is the Boltzmann constant, and T is the absolute temperature. The thermal energy, kBT, is only about 0.025 eV (200 cm–1) at room temperature in comparison to the vibrational energy of 0.12–0.37 eV (1000–3000 cm–1), implying that the population of higher vibrational levels of the ground state is negligible [168].  The Franck-Condon principle states that all transitions from the ground state to the excited state are vertical transitions, as shown in the Perrin-Jablonski diagram (Figure 1.7), occurring without  29 any change in the position of nuclei. This principle is the result of differences in the timescale of the processes: the absorption of light is extremely rapid (10-15 s) in comparison to nuclear motions, which occur 102–103 times slower [165]. The quantum mechanical formulation of this principle is that the intensity of a vibrational transition is proportional to the square of the overlap integral between the vibrational wavefunctions, ψ, of the initial and final state (eqn. 1.4).  This squared integral is known as the Franck-Condon Factor (FCF) and the total optical strength is the product of FCF and the constant electronic interaction term. FCF = ψ final* ⋅ψinitial dτ∫⎡⎣ ⎤⎦vib2                                                      (1.4) The vibrational overlap integral depends on the offset between the equilibrium bond lengths for the ground state and the excited state potential energy wells [168]. As shown in Figure 1.8, as a molecule moves to a vibrational level of the excited state during an electronic transition, this level must be instantaneously compatible with the nuclear position of the initial state. The excitation process is much faster than motions of nuclei, such that absorption is a vertical transition. Once molecule is in the excited state, its electronic configuration changes and the nuclei must move to reorganize to the new electronic configuration via molecular vibration [165]. As a result, there is an offset along the nuclear coordinate axis in the ground state and the excited state potential energy wells. The absorption of light results in the transition of a molecule to a nonequilibrium excited state, known as a Franck-Condon state. The multiple pathways available for a molecule to return to the ground state are discussed in Section 1.2.3.   30  Figure 1.8 Illustration of Franck-Condon principle for an electronic transition from the ground electronic state, So to the excited state, S1. (A) The maximum transition probability is observed for transitions with maximum overlap between initial and final vibrational wavefunctions, based on Franck-Condon factors. The most probable excitation (blue line) is shown into the second vibrational state (ν2*) of S1, and the corresponding most probable relaxation transition is shown to the second vibrational state (ν2) of So (red line). The probability density function is shown in green. (B) A corresponding schematic for absorption and emission spectra. The spectral characteristics of individual transitions are typically only observed in a gas phase or at very low temperatures, broadening (shown as shaded region) is observed in solution at room temperature. As a result of the Franck-Condon principle, the absorption and emission spectra are approximately a mirror image of each other.  The probability of a given transition between the ground state and an excited state is measured with either the molar absorption coefficient (ε, L mol–1cm–1) or the oscillator strength (f) [165]. The oscillator strength is a dimentionless parameter that compares a quantum mechanical transition to the expected fully allowed one. A strong transition is associated with f = 1, whereas quantum-mechanically forbidden transitions can have f values of 10–3 or less [165].   The molar absorption coefficient is linearly proportional to the absorption cross-section, σ (in the units of cm2) of a chromophore, according to eqn. 1.5: σ = 2.303εcn = 2.303εc(Nc /103) = 3.82×10−21ε                                            (1.5)  31 where c is concentration of chromophore, n is the number of molecules per cm3, and N is the Avogadro’s number. A larger molar absorption coefficient and absorption cross-section indicates a greater probability of an electronic transition. For “fully allowed” transitions (i.e. spin rule and Laporte rule), molar absorption coefficients are greater than 105. In cases where transitions are spin allowed but symmetry forbidden, the values fall in the range 102–104 due to the symmetry distortion described above [169]. The relationship between the molar absorption coefficient and oscillator strength, f, is described by eqn. 1.6: f = 4.3×10−5 ε(ν ) dν∫                                                             (1.6) where ν is the frequency in units µm–1. The Beer-Lambert law characterizes the absorption process of a chromophore according to eqn. 1.7, where A is the absorbance, T is the transmittance, λ is the wavelength, Io(λ) is the incident light intensity, and I(λ) is the remaining light intensity after passing through the sample with chromophore concentration, c, and sample path length, b. A λ( ) = − logT λ( ) = log I0 λ( )I λ( )⎛⎝⎜⎜⎞⎠⎟⎟= ε λ( )b c                                                (1.7)  1.2.3 The Franck-Condon electronic excited state Upon excitation of a chromophore with light, it reaches one of the vibrational levels of the excited state (Smνn), as shown on the Perrin-Jablonski diagram in Figure 1.7. Subsequently, a chromophore will undergo several relaxation processes to return to its ground state. These include non-radiative (vibrational relaxation, internal conversion, intersystem crossing) and radiative (fluorescence, phosphorescence) processes.  An electronic excited state is energetically unstable as the chromophore is not in thermal equilibrium with the surrounding medium. As a result, rapid vibrational relaxation on the timescale of 10–13–10–11 s brings a molecule from the higher vibrational levels (n > 0) to the lowest vibrational level (n = 0) of S1 [165]. During this process, chromophores collide with other  32 molecules (i.e. solvent) and transfer some of the energy imparted by the absorption of the photon to the surroundings as heat. Vibrational relaxation is the most efficient process, with a timescale shorter than any other relaxation process, so that a molecule always relaxes to Smν0 state before any other process will take place.  Internal conversion is a non-radiative process that allows for transitions between different electronic states with the same multiplicity, such as from the Smν0 state to the S(m-1)νn state. The internal conversion process is typically described by either direct vibrational coupling (vibronic level overlap) or quantum mechanical tunneling (no direct vibronic overlap but small energy gap). The energy gap between electronic excited states (Sn, n≥2) is small compared to that between the first excited state, S1, and ground state, S0. Furthermore, the numerous vibrational states with a typical energy gap of ca. 0.1 eV, along with superimposed rotational levels (10–3-10–2 eV) can generate approximately isoenergetic levels between two electronic states. Since internal conversion takes place between two isoenergetic levels without energy loss, it is a horizontal transition on the Perrin-Jablonski diagram. Quantitatively, the rate constant for non-radiative transition is described by Fermi’s golden rule, which states that the rate is proportional to the product of two parameters: (1) the density of energetically matching vibrational levels between the initial and final states, and (2) the square of the vibronic coupling term [170]. Empirically, rate constants decrease exponentially as the energy difference between two electronic states increases [170]. This trend is known as the energy gap law. In general, as the energy gap between two electronic states decreases, the density of the vibrational states increases, and the Franck-Condon factors between two states increase, as well as the efficiency of internal conversion. Importantly, because there is a large energy gap between the first excited state, S1, and the ground state, S0, the rate of internal conversion is much slower, and fluorescence becomes a significant competitive relaxation process. The timescale of internal conversion between two excited states is comparable to the timescale of vibrational relaxation (10–13–10–11 s), while internal conversion to the ground state can be on the order 10–9–10–7 s. A direct consequence of the rates of vibrational relaxation and internal conversion is the empirical Kasha’s rule. It states that for polyatomic molecules, irrespective of the electronic excitation, the luminescence is only observed from the lowest vibrational level of the first excited electronic state of a given multiplicity. Exceptions to Kasha’s rule have been observed.  33 For example, for molecules with large energy gaps between S1 and S2, fluorescence corresponding to the S2→S1 transitionis observed (e.g. azulene, cycl[3.3.3]azine) [170]. Alternatively, for molecules in which the oscillator strength of the S0→S1 transition is very small in comparison to S0→S2, a two-level S2→S0 fluorescence transition can be observed (e.g. ovalene) [170]. Fluorescence is the spontaneous emission of radiation by the excited molecule as it transitions from S1ν0 to S0νn. The energy of the emitted photon is equal to the energy difference of the corresponding transition. The timescale of the fluorescence process is 10–10–10–7 s, which corresponds to the duration an excited state molecule remains in the first excited state before emitting photon [165]. The emission of a photon is as fast as the absorption of a photon (10–15 s). Similar to the absorption transition, the emission transition is governed by the Franck-Condon principle, which is represented by an approximate mirror-image relationship between the fluorescence spectra and first absorption band of a fluorophore, as illustrated in Figure 1.8. As a consequence of Kasha’s rule, the Franck-Condon principle, and solvent effects (see page 36), the fluorescence emission spectrum is shifted to longer wavelengths in comparison to the absorption spectrum. The difference between absorption and emission maxima is known as Stokes shift.  Another de-excitation process from the S1ν0 state that competes with fluorescence is intersystem crossing to the excited triplet state, T1νn. This process occurs between two isoenergetic vibrational levels of two electronic states with different multiplicity. Although such transitions are forbidden, spin-orbit coupling (i.e. coupling between spin magnetic moment and orbital magnetic moment) can be sufficiently large for efficient intersystem crossing. Phosphorescence is a radiative transition from the T1νn state to the S0νn state. This transition is forbidden but, similar to the intersystem crossing process, can be observed as a result of spin-orbit coupling. The radiative rates for phosphorescence are very low with a timescale on the order of 10–6-10–3 s. With such a slow process, the high probability of collisions with solvent molecules favors non-radiative relaxation to the ground state via intersystem crossing and vibrational relaxation. 1.2.4 Quantum yield and fluorescence lifetime The fluorescence intensity and fluorescence lifetime of a fluorophore depend on the relative magnitudes of the rates for the radiative and non-radiative processes described in Section 1.2.3.  34 The quantum yield represents the fraction of excited molecules that decay to the ground state, S0, with the emission of photon. The fluorescence quantum yield, Φf, is defined in terms of all rate constants involved in depopulation of the excited state according to eqn. 1.8. Φ f =k fk f + knr∑                                                             (1.8) knr∑ = kic + kisc + kq + ket + kpd +...                                             (1.9) where kf is the fluorescence decay rate, knr is the total non-radiative decay rate grouped according to eqn. 1.9, kic is the rate of internal conversion, kisc is the rate of intersystem crossing, kq is the rate of any fluorescence quenching process, ket is the rate of energy transfer, and kpd is the rate of photodegradation of the fluorophore. Since the quantum yield represents a fraction of fluorophores that relax to the ground state via emission of photon, its value always varies within 0 ≤ Φf < 1. Although quantum yield can be very close to unity, some non-radiative losses occur, as knr is almost never zero [164]. The fluorescence lifetime is the average time a molecule spends in the excited state prior to relaxation to the ground state, and defined as a reciprocal rate constant for the decay of the excited state: τ =1k f + knr∑=Φ fk f                                                        (1.10) The rate equation for depopulation of the excited state, S1, at time, t, following excitation is a first-order process given by eqn. 1.11. −d S1(t)[ ]dt = (kr + knr∑ ) S1(t)[ ]                                               (1.11) Emission is a random event and the response that follows a pulsed excitation (used to instantaneously generate an excited state population) is an exponential decay: I(t) = I0 exp −t τ( )                                                    (1.12)  35 where I(t) is the fluorescence intensity measured at time t, and I0 is the intensity at time, t = 0. The fluorescence lifetime is a statistical average and fluorophores emit randomly throughout the decay.  According to eqn. 1.12, the excited state lifetime is the time at which 63% of the fluorophores have relaxed to their ground state.  1.2.5 Quenching processes Fluorescence is strongly dependent on the local environment and a variety of processes may decrease or quench fluorescence intensity. A wide range of molecules and ions can act as quenchers. The major mechanisms of the interaction between a fluorophore and a quencher include static quenching, dynamic quenching, and combined quenching [164]. Static quenching occurs when a nonfluorescent ground state complex is formed between a fluorophore and a quencher. Static quenching is a concentration-dependent process. Dynamic quenching describes any quenching process that affects the excited state, leading to a change in the excited state lifetime of a fluorophore. Dynamic quenching processes are distance-dependent and are based on collisional quenching, a process of depopulation of the excited state via intermolecular collisions with quencher. Charge transfer, electron transfer, and through space quenching—Förster resonance energy transfer (see Section 1.2.8) are all examples of dynamic quenching. Cases where fluorophore can be quenched by both collisions and complex formation with the same quencher are classified as combined quenching. All three quenching mechanisms described above are often analyzed using the Stern-Volmer equation. The relative changes in fluorescence intensity and lifetime as a function of quencher concentration, the effect of temperature, and changes in absorption and emission spectra can be used to identify the mechanism of quenching [164]. Photobleaching or dye photolysis results in an irreversible loss of dye fluorescence as a result of photochemical modification. Although photobleaching is not a quenching process per se, it can contribute strongly to decreases of fluorescence intensity. During this process, a molecule in the excited state undergoes a permanent structural change and ground state fluorophore is never recovered. Many factors, including local environment and the power of the excitation light may affect the mechanism of photobleaching, and therefore reaction rates [171]. In contrast to the quenching mechanisms discussed above that are reasonably well understood, photodegradation is still poorly understood phenomenon. Among several theories used to explain photobleaching, the  36 main mechanism is attributed to the interactions of excited state fluorophore and molecular oxygen in its triplet state (3O2) generating singlet oxygen [171]. The long-lived triplet state permits excited state molecules to interact with their surrounding for longer time (10–6–10–3 s). The photoionization of fluorophores in the triplet excited state generates highly reactive dye radicals towards solvent and other solutes. Susceptibility to photobleaching varies from dye to dye. Some dyes have shorter lifespans and degrade after emitting a few hundred photons, while other dyes may emit millions of photons before being bleached [171]. Importantly, the intensity of excitation light plays a non-trivial role in all photobleaching mechanisms. 1.2.6 Factors affecting fluorescence The fluorescence intensity, quantum yield, and excited state lifetime in most molecules are extremely sensitive to their local microenvironment. The effects of the temperature, solvent polarity, viscosity, acidity, hydrogen bonding, and presence of quenchers are often the basis for analysis, as they modulate the rate of non-radiative processes [165]. In contrast, the radiative decay rate and natural lifetime are typically intrinsic properties of a fluorophore, and show little to no dependence on the microenvironment [172], with the exception of the environment of plasmonic metal structures [173, 174]. An increase in temperature is always accompanied by a decrease in the quantum yield and the lifetime because of the increase in non-ratiative decay rates. Higher temperatures result in greater diffusion, increased collisions with solvent molecules, more efficient intramolecular vibrations and rotation—all of which promote non-radiative relaxation pathways. High viscosity solvents, as well as low temperatures reduce the number of collisions with excited state molecules, thereby slowing down non-radiative deactivation process and increasing quantum yield.  The role of the solvent and its polarity becomes important not just in its influence on the quantum yield of a fluorophore, but also on the broadening, shape and position of its fluorescence spectra. The latter effects are referred to as solvatochromism, where interactions of solvent molecules with fluorophore can cause bathochromic shifts (i.e. shifts to longer wavelength). Therefore, Stokes shifts are a result of vibrational relaxation to S1ν0, the Franck-Condon factors, and solvent polarity [168]. Solvent molecules surround a ground state fluorophore based on the interactions of their dipole moments. Upon excitation of a fluorophore, the energy difference between the ground state and the excited state produces an increase in its  37 electric dipole moment. The surrounding solvent molecules respond to stabilize an excited state fluorophore. First, an instantaneous electronic polarization of the surrounding solvent molecules, and second, by absorbing excess vibrational energy released as fluorophore relaxes to the lowest vibrational level of the first excited state. Finally, solvent molecules stabilize and lower the energy of the excited state by rearranging around the fluorophore. This process, known as solvent relaxation, reduces the energy difference between excited state and ground state which induces the red-shift of fluorescence emission spectrum [164, 168]. Increased solvent polarity generally leads to a larger stabilization effect, while the polarity of the fluorophore determines its sensitivity to solvent effects.  Another parameter that affects the fluorescence of molecules with basic or acidic substituents is pH. The fluorescence intensity, shape and spectral position of absorption and emission bands may be different for protonated and deprotonated forms of a fluorophore. For example, fluorescein can exist as a cation, neutral molecule, anion, and dianion with pKa’s of 2.08, 4.31, and 6.43, respectively [175]. All these forms have different absorption spectra and molar absorption coefficients (ε437,cation = 53 000, ε434,neutral = 11 000, ε453,anion = 29 000, ε490,dianion = 76 900 M–1cm–1). Furthermore, the most intense quantum yield is observed for dianion (0.93), whereas the anion and cation have quantum yields of 0.37 and 0.18, respectively [175].  In addition to the parameters discussed above that affect fluorescence, and the quenching processes introduced in Section 1.2.5, a variety of other photophysical processes may also affect fluorescence, including collisions with heavy atoms (e.g. halogens) or paramagnetic species (e.g. dissolved oxygen), excimer or exciplex formation, electron transfer, and proton transfer [164]. 1.2.7 Fluorescence measurements Steady-state and time-resolved fluorescence spectroscopy are among the primary tools used to investigate physical, chemical, and biological systems. Some of the common parameters used to gain a physical insight or analytical information include changes in quantum yield, excited state lifetime, number of emitters, emission wavelength, and anisotropy. Steady-state fluorescence measurements are based on constant illumination of the sample with a continuous beam of excitation light. This measurement represents equilibrium conditions, such  38 that the concentration of excited state fluorophores is constant under constant illumination. Typically, steady-state fluorescence measurements are done by (i) measurement of fluorescence intensity integrated over a fixed bandwidth of emission wavelengths, (ii) measurement of fluorescence intensity as a function of emission wavelength with a fixed excitation wavelength (an emission spectrum), and (iii) measurement of fluorescence intensity at a fixed emission wavelength as a function of excitation wavelength (an excitation spectrum). The fluorescence intensity, F, measured by a given spectrofluorimeter for sufficiently dilute samples (εbc < 0.05) is given by eqn. 1.13, where Φf is the quantum yield of a fluorophore, b is the path length of the excitation light through the sample, c is the concentration of the fluorophore in the sample, ε(λ) is the molar absorption coefficient of the fluorophore, Po(λ) is the spectral power density of the excitation light, K(λʹ) is the combined collection and detection efficiency of spectrofluorimeter, and L(λʹ) is the band shape function of the fluorophore emission [176]. F( ʹλ ) = 2.303Φ f bc P0 (λ)∫∫ ε(λ) K( ʹλ ) L( ʹλ ) dλ d ʹλ                              (1.13) Fluorescence measurements are typically performed on the same instrument and under the same excitation conditions as relative changes in intensity. From eqn. 1.13, it can be seen that fluorescence intensity is proportional to the concentration and quantum yield of a fluorophore, while all other terms can be approximated as constant for a given instrument under a specified set of conditions (i.e. excitation wavelength, sample path length).  In time-resolved measurements, the decay in fluorescence intensity is measured as a function of time on a timescale of the excited state lifetime after excitation with a pulsed light source. In contrast to steady-state measurements, time-resolved experiments are not at equilibrium conditions and rather provide information about dynamic changes in the excited state population. Two methods for measuring time-resolved fluorescence are time-domain and frequency-domain spectroscopy. In the time-domain method, the sample is excited with a short pulse of light and the decay in fluorescence intensity is recorded as a function of time. Time correlated single photon counting (TCSPC) is a common approach for time-domain measurements being used for routine lifetime measurements, fluorescence lifetime imaging microscopy (FLIM), and various  39 single molecule studies [164, 177]. In the frequency-domain method, the sample is excited by sinusoidally modulated light at high frequency. The fluorescence response has a similar waveform, but is modulated and phase-shifted from the excitation curve based on the lifetime of the sample emission [164]. 1.2.7.1 Fluorescence microscopy and imaging Fluorescence microscopy provides spatial and intensity information about a fluorescent sample. Similar to the steady-state measurements described above, the sample is illuminated with a fixed bandwidth of excitation light and emission over a fixed bandwidth is acquired. The difference is the method and the detector used to record fluorescence emission—a digital image is created when an optical image formed by the microscope is recorded by a detector. Common imaging detectors include charge-coupled devices (CCDs) and scientific CMOS. When digital image of the sample is reconstructed from individual measurements across the sample (i.e scanning), photomultiplier tubes (PMTs) are often used as detectors. A digital image is a two-dimensional grid of equally sized pixels, where each pixel represents a defined finite sized area in a specific location of the sample. The number of photons detected at each pixel during image acquisition is converted to the intensity value that can be correlated to the concentration of fluorophore depending on the nature of the sample [178]. Fluorescence microscopy is used for relative intensity measurements (e.g. fluorescent staining), co-localization studies, as well as Förster resonance energy transfer (FRET) and fluorescence recovery after photobleaching (FRAP) experiments [178]. 1.2.8 Förster resonance energy transfer Förster resonance energy transfer (FRET) is a mechanism of energy transfer via non-radiative through-space dipole-dipole interactions between a donor molecule its excited state (D*) and an acceptor molecule its ground state (A). The acceptor can be a chromophore that is non-fluorescent (quencher), another fluorophore, or the same type of fluorophore as the donor (homoFRET) [164]. The dipole-dipole interaction (i.e. donor relaxation and acceptor excitation transition moments) does not require molecular contact between the donor and the acceptor; however, close proximity (< 10 nm) is necessary for efficient energy transfer. FRET is a resonant process, indicating that the energy difference between the excited and ground state of the  40 acceptor must be equal to the energy difference between the excited and ground state of the donor. This energetic resonance is expressed by the spectral overlap between the absorption spectrum of the acceptor and the emission spectrum of the donor. In this energy transfer process, there is no net change in the energy of the system and any excess energy is dissipated into vibrational modes in accordance with conservation of energy. The process of FRET is conceptually summarized in eqn. 1.14 and illustrated in Jablonski diagram in Figure 1.9. D*+A→D+ A*                                                      (1.14)   Figure 1.9 Jablonski diagram illustrating the FRET process. The donor is excited from the ground state So to any of the vibrational levels of the first excited state S1, and then relaxes to the lowest vibrational level of S1 via vibrational relaxation. During nonradiative resonance energy transfer, the donor relaxes to the ground state without emitting a photon, while an acceptor is simultaneously excited to a vibrational level of S1. Then acceptor undergoes the vibrational relaxation to the lowest vibrational level of S1, followed by either emission of a photon (i.e. fluorophore acceptor) as it relaxes to the ground state or nonradiative relaxation (i.e. dark quencher acceptor).  41 1.2.8.1 Classical Förster formalism  The quantum mechanical treatment of nonradiative energy transfer is based on Fermi’s golden rule [170]: k = 2π!VD*A−DA*2ρ                                                         (1.15) where k is the rate constant for the radiationless transition, ! is the reduced Planck constant, ρ is the density of final states relevant for the transition (i.e. the density of isoenergetic donor-acceptor states), and V represents an interaction term coupling the wavefunctions of the initial and final states. When the intermolecular distance between donor and acceptor is much larger than the size of the molecules, the dipole-dipole interaction between the transition moments of the donor (D*) and the acceptor (A) can be represented by a dominant Coulombic coupling term, VCoul [179]. The orbital overlap term is negligible and only becomes significant at very short distances (Dexter energy transfer). Therefore, eqn 1.15 can be modified for a FRET process as follows: kFRET =2π!VCoul2ρ =2π!κ µD! "!µA! "!4πε0n2r32ρ                                        (1.16) where µD and µA are the transition dipole moments associated with donor relaxation and acceptor excitation, κ is the orientation factor between these two dipole moments, ε0 is the vacuum permittivity, n is the refractive index, and r is the intermolecular distance between donor and acceptor. Since the interaction between the two dipoles decreases with the third power of distance, the FRET rate constant, kFRET, exhibits an inverse sixth power dependence on the distance between donor and acceptor. In addition, eqn. 1.16 shows the origin of the n–4 and κ2 dependences of energy transfer rate. The expression given in 1.16 can be further represented by experimentally measureable spectral quantities according to eqn. 1.17, known as the Förster equation. The square of the transition dipole moments is proportional to the oscillator strength, which can be represented by either a radiative lifetime for the donor, τD or molar absorption  42 coefficient for the acceptor, εA, according to eqn. 1.6. The density of states, ρ, combined with εA is substituted by the normalized spectral overlap integral, J. kFRET =9(ln10)κ 2 ΦD128π 5NAvn4τ D r6J                                                   (1.17) where ΦD is the quantum yield of the donor, NAv is Avogadro’s number, and τD is the excited state lifetime of the donor in the absence of acceptor. When rate constants for energy transfer and spontaneous decay are equal (kFRET = kr + knr = τD –1), the donor-acceptor distance is known as critical transfer distance (or Förster distance), R0: R06 =9(ln10)ΦDκ 2J(λ)128π 5NAv n4= (8.79×10−28mol) n−4ΦD κ 2 J(λ) (cm6 )               (1.18) The typical range of a Förster distance for a donor-acceptor pair is 20–60 Å [164, 165]. A larger Förster distance implies higher FRET efficiency for a given donor-acceptor separation; however, there is always an upper limit at which the transition dipoles of the donor and the acceptor can no longer interact.  Combining eqn. 1.17 and 1.18, the rate of energy transfer is given by eqn. 1.19. FRET is the relaxation process that competes with all other processes including fluorescence, as was seen in eqn. 1.9. In the absence of an acceptor, kFRET is zero. kFRET =1τ DR0r⎛⎝⎜⎞⎠⎟6                                                                (1.19)    43  Figure 1.10 Schematic representation of the spectral overlap integral, which is shown as the grey area, J, where the emission spectrum of the donor overlaps with absorption spectrum of acceptor.  The spectral overlap integral is derived from the spectral overlap between the donor emission and acceptor absorption and provides a quantitative measure of the possible resonant transitions. The conceptual representation of spectral overlap is illustrated in Figure 1.10. Mathematically, the spectral overlap is defined by eqn. 1.20, where ID is the normalized donor fluorescence intensity (eqn. 1.21) as a function of wavelength, λ and εA(λ) is the wavelength-dependent molar absorption coefficient of acceptor (M-1cm-1 = 103 mol-1cm2) [164]. J(λ) = ID εA (λ) λ 4 dλ∫                                                         (1.20) ID dλ ≡1∫                                                                   (1.21)  FRET occurs as a result of the interaction between the transition dipoles of a donor and acceptor; the magnitude of this interaction, as was noted above, depends on the relative orientation and separation of these two transition dipole moments. The strongest interaction is observed for the parallel orientation (κ2 = 4) and no interaction occurs when two vectors are perpendicular (κ2 = 0) [164]. The magnitude of the orientation factor is determined according to eqn. 1.22, where θT is the angle between the donor and acceptor transition dipoles, and θA and θD are the angles the donor and acceptor dipoles make to the line connecting them, as illustrated in Figure 1.11 [164]. Experimentally, it can be challenging to determine the actual orientation of dipoles for molecules  44 that are free to move or rotate, although fluorescence anisotropy experiments are useful in determining this parameter [164, 180]. For instance, the polarization of two randomly oriented dipoles for donor and acceptor can be measured and correlated to the range of possible orientation factors [180]. Typically, it is assumed that for a dynamic isotropic distribution (random rotation), the orientation factor, κ2, is 2/3, and, for a static isotropic distribution (orientations do not change during timespan of energy transfer), κ2 is 0.476 [164]. In general, variation in κ2 does not result in significant errors for calculated distances [164, 181]. κ 2 = (cosθT −3cosθA cosθD )2                                               (1.22)   Figure 1.11 Representation of the donor and acceptor transition dipole moments and angles used to calculate the orientation factor, κ2.   The FRET efficiency (i.e. quantum yield of energy transfer), E, is strongly dependent on the donor-acceptor separation and the Förster distance, as shown by eqn. 1.23. This relationship applies to an isolated donor-acceptor pair, where a single donor interacts with a single acceptor. For a case where a single donor can interact with multiple acceptors and vice versa, this expression may not necessarily be applicable (see Section 1.3.6 for details). E = kFRETkFRET + kr + knr∑=R06r6 + R06                                                (1.23)  45 The FRET efficiency approaches 100% at distances r < 0.5R0 and approaches 0% when r > 2R0. The separation range 0.5R0 < r < 1.5R0 provides the most sensitive transduction in FRET efficiency and allows FRET to operate as a “molecular ruler” [164]. The transduction is commonly achieved over separation distances 1–10 nm. 1.2.8.2 Measurement of FRET efficiency The characteristic parameters used to quantify FRET include a decrease in donor fluorescence intensity, an increase in FRET-sensitized acceptor fluorescence intensity, a decrease in donor lifetime, a decrease in donor emission anisotropy, and depolarization of the acceptor emission [164]. Experimentally, the FRET efficiency is often determined from measureable changes in fluorescence intensity (F), quantum yield (Φ), or lifetime (τ) between an isolated donor (denoted with a subscript D) and a donor in the presence of acceptor (denoted with a subscript DA), as described by the relationships given by eqn. 1.24 [164]. These approaches rely on a relative measurement by comparing two systems with and without introduction of resonance energy transfer pathway. Therefore, identical excitation conditions must be used when FRET efficiency is determined from the changes in intensity or quantum yield. E =1− FDAFD=1− ΦDAΦD=1− τ DAτ D                                                   (1.24) Calculation of FRET efficiencies from time-resolved experiments through the changes in excited state lifetime of the donor assumes that both lifetimes, τD and τDA, are monoexponential or amplitude weighted average lifetimes [164]. From a practical perspective, the advantage of eqn. 1.24 in comparison to eqn. 1.23 is that it is applicable to FRET pairs where the donor–acceptor ratio is not just 1:1, but a general case 1:a, where a > 0.  A modified form of eqn. 1.23 for a number of acceptors (a > 0) is introduced in Section 1.3.6.  The FRET efficiency can also be determined from changes in the FRET-sensitized acceptor emission accounting for contributions of direct excitation (i.e. acceptor only system). This approach requires a correction factor to account for differences in the molar absorption coefficients of the donor and the acceptor. Two common approaches are based on either (i) an  46 internal reference acquired at an excitation wavelength where donor does not absorb, or (ii) an external reference with a sample containing only acceptor. For a sample with an internal reference, the FRET efficiency is given by: E = FADAAA −FAAAADFAAADD                                                       (1.25) where FAD is the FRET sensitized acceptor intensity with donor excitation, FAA is the acceptor intensity following acceptor excitation, AAA is the acceptor absorbance at acceptor excitation wavelength, and AAD and ADD are the acceptor and donor absorbance, respectively, at the donor excitation wavelength. The expressions given in eqn. 1.24 and 1.25 are often used for characterization of a FRET system by determining the donor-acceptor separation, r. The Förster distance, R0, for a given FRET pair is constant, and therefore measurements of the FRET efficiency can permit measurements of r. Consequently, FRET has been used to quantitatively determine molecular distances in biochemical and molecular biological systems (e.g. to study conformational changes during protein folding or protein interactions in living cells) [182-184].  In contrast, FRET-based bioassays do not generally rely on measuring the donor-acceptor separation, but rather the quantification of the relative concentrations of donor-acceptor pairs is used [185]. The processes of association (e.g. DNA hybridization) or dissociation (e.g. proteolysis) of biomolecules result in either increases or decreases in the number of FRET pairs. These changes are often measured using a FRET ratio, where the donor emission in the absence of acceptor is not required, provided that the quantum yields of the donor and the acceptor are known. The expression given by Eq. 1.26 is valid assuming direct excitation of the acceptor is negligible: FADFDA=ΦAEΦD 1−E( )                                                              (1.26) where FAD is FRET-sensitized acceptor emission and FDA is donor emission.  47 1.2.8.3 Assumptions underlying use of FRET There are several assumptions routinely applied in the analysis of FRET systems, including the ideal dipole approximation, the magnitude of the orientation factor, the refractive index of the media and the effect of spectral broadening on the calculated spectral overlap. The conditions under which these assumptions are valid, and the consequences of their breakdown on the interpretation of the FRET data are introduced in this section.   The derivation of the simple relationship between the energy transfer rate and the relative separation and orientation of donor and acceptor is given by eqn. 1.17, which is based on the assumption that transition densities can be approximated as point dipoles. The representation of the electronic coupling term, V, in Fermi’s golden rule (eqn. 1.15) by the transition dipole-transition dipole interaction energy is a special case of the multipole expansion used by T. Förster [179]. The space around the donor is visualized as a group of electrical oscillators producing an electric field is divided in to four zones. These are the contact or Dexter zone (< 1 nm), the near-field zone (1-10 nm), the intermediate zone (10-1000 nm), and the radiation zone (> 1000 nm) [186]. The Förster formalism is only valid within the near-field zone, as the total electronic coupling can be represented adequately by the Coulombic interactions, V ≈ VCoul. The expansion of the interaction energy into power series (i.e. multipole expansion) results in the main dipole-dipole interaction term for uncharged molecules [187]. The contributions from multipole terms are generally negligible [187]. Therefore, within the ideal dipole approximation, the interactions between molecules can be described by V ≈ VCoul ≈ Vdip-dip. This approximation is valid at donor-acceptor separation down to 20 Å [179].  Within the Dexter zone, the exchange interactions caused by the overlap of molecular orbitals of the donor and the acceptor require an additional coupling term: V = VCoul + Voverlap. Therefore, as molecular separation becomes comparable to the size (i.e. a spatial extent of transition density) of the molecules, Voverlap term becomes significant and the validity of ideal dipole approximation becomes questionable [179]. The orientation factor, which depends on the mutual orientation of the donor and acceptor, represents another facet of Förster formalism. In biological samples, there may be fluctuations in the positions and relative orientations of donors and acceptors, allowing for assumption of a dynamic isotropic distribution of transition moments [164]. Therefore, the value of κ2 is taken as 2/3, an average value over time, which corresponds to the dynamic isotropic limit, where  48 rotations of donor and acceptor are fast in comparison to the excited state lifetime [188]. This assumption can lead to large errors in single molecule measurements, where averaging over many FRET-pairs or over longer times would be unpractical [185]. In ensemble measurements, this assumption may not be strictly valid for all the cases; however, the corresponding error may not be significant enough. While the rate of energy transfer, kFRET, has a linear relationship with orientation factor, κ2, the Förster distance, R0, and the experimentally calculated donor-acceptor separation, r have an inverse sixth power dependence. Given that for a static isotropic distribution found in rigid systems, the orientation factor is assumed to be 0.475, the discrepancy in the R0 or r between static and dynamic distributions is estimated to be 5%. Different donor-acceptor conformations can lead to the orientation factor values in the range 0 ≤ κ2 ≤ 4. In unfavorable cases (0 ≤ κ2 ≤ 2/3) often found in anisotropic systems or systems with reduced fluidity, a two-fold discrepancy can be observed if the assumption is used [188]. In contrast, favorable systems (2/3 ≤ κ2 ≤ 4) produce a maximum error of 35% in the calculation of the distance between donor and acceptor [164]. Fluorescence anisotropy measurements can be used to confirm fast isotropic rotation that lead to unpolarized emission and can be used to estimate limits on orientation factor [180, 189]. Another parameter that often requires an approximation during FRET analysis is the refractive index, n, of the microenvironment surrounding the donor and the acceptor. The extent of the dielectric screening of the Coulomb interaction between the donor and the acceptor exhibits a VCoul ∝ n–2 dependence. In many bioassays the fluorophores are attached to larger biomolecules (e.g proteins) that may have a different refractive index in comparison to the solvent. For convenience, it was proposed that, in applying the Forster formalism to fluorophores in proteins, the refractive index of the solvent may be used instead of the one corresponding to the dielectric screening of the protein [190]. This approach continues to be widely adopted, although there are examples of using a local refractive index (e.g. protein, membrane) rather than the solvent refractive index [191]. The typical range of refractive indices associated with biological media fall within a very narrow range. For example, water has a refractive index of 1.33, lipid membranes a refractive index of 1.46, proteins a refractive index of 1.5 [192], corresponding to less than 10% variation in the Förster distance based on R06 ∝n−4.   49 1.3 Semiconductor Quantum Dots 1.3.1 What is a quantum dot?  QDs are colloidal semiconductor nanocrystals with dimensions between ca. 1-10 nm. Excitons are generated in the nanocrystals upon the absorption of light, and electron-hole recombination leads to luminescence. Although depicted as spheres in most illustrations, QDs are crystalline materials with facets and a lattice structure analogous to the bulk semiconductor material. Depending on its size, each nanocrystal can comprise hundreds to thousands of atoms, a large fraction (>10%) of which are located at the nanocrystal surface (i.e. a high surface area-to-volume ratio). As described in more detail below, most of the QDs used in analytical applications are synthesized as core/shell structures, where the core nanocrystal is overcoated with another semiconductor material to protect and improve its optical properties. The “flagship” QD material is undoubtedly core/shell CdSe/ZnS. 1.3.2 Absorption and photoluminescence  It was the unique photophysical properties of QDs that first generated excitement for biological imaging and analysis. QDs have become renowned for eye-catching photographs (Figure 1.12A) of differentially sized QDs under ultraviolet (UV) illumination that show a bright rainbow of photoluminescence (PL). The bright PL is the result of high quantum yields (Φ = 0.1–0.9) combined with large molar extinction coefficients (105–107 M–1 cm–1). As shown in Figure 1.12B-C, QDs have broad absorption spectra that continuously increase in magnitude from their first exciton peak to shorter wavelengths in the near-UV. QD PL spectra are shifted to slightly longer wavelengths than the first exciton absorption peak, such that an effective Stokes shift >100 nm can be achieved. The PL is also spectrally narrow with an approximately Gaussian profile (FWHM 25–35 nm). The stunning rainbow of QD PL arises from the fact that the peak emission wavelength shifts as a function of nanocrystal size and material. The QD size and PL colour can be selected by controlling the temperature and duration of crystal growth during synthesis. Photographs of the type in Figure 1.12A exemplify the utility of QDs for multiplexed analyses and multicolour imaging: a single light source can excite many colours of QD simultaneously (broad absorption) and each PL contribution can be readily resolved or deconvolved (narrow emission).  50  Figure 1.12 (A) Size-tunable PL of CdSe QDs. The photograph was taken under UV illumination (365 nm). (B) TEM image of a CdSe/ZnS QD. A and B reprinted with permission from ref. [126]. Copyright 2011 American Chemical Society. (C) Size-dependent absorption and fluorescence spectra of CdSe QDs. Reprinted with permission from ref. [193]. Copyright 2010 American Chemical Society. (D) Absorption and PL spectra of ZnxCd1−xSe QDs with Zn mole fractions of (a) x = 0, (b) 0.28, (c) 0.44, (d) 0.55, and (e) 0.67. Reprinted with permission from ref. [194]. Copyright 2003 American Chemical Society.  Other advantageous optical properties of QDs include excited state lifetimes that tend to be longer than those of fluorescent dyes (> 10 ns), superior resistance to photobleaching and chemical degradation (due to the inorganic composition and confinement of the exciton), and two-photon absorption cross sections (103–104 GM) that are orders of magnitude larger than those of fluorescent dyes [195]. QDs are thus excellent probes for tracking dynamic processes over time, and for two-photon imaging of tissues or other complex biological specimens where near-infrared (NIR) excitation mitigates challenges associated with autofluorescence and attenuation of excitation light by strong protein absorbance (e.g. hemoglobin) in the visible region [196, 197].   51 1.3.3 Quantum confinement and core/shell structures The size-dependence of QD PL is the result of quantum confinement. As a bulk material is reduced to nanoscale dimensions, the density of states decreases near the conduction band and valence band edges, resulting in the emergence of discrete excitonic states. The band gap energy further increases with decreasing nanocrystal size as the exciton is confined to smaller dimensions than its Bohr radius (i.e. the preferred distance between the electron and hole). The PL emission wavelength shifts since exciton recombination occurs between the band edge states. For example, bulk CdSe has a bandgap energy of 1.76 eV and a Bohr exciton diameter of 9.6 nm [193], whereas the band gap energy of 2–7 nm CdSe nanocrystals decreases from 2.8 eV to 1.9 eV with PL shifting between 450–650 nm. The range over which the band gap energy and PL wavelength can be tuned by quantum confinement depends on the material of the nanocrystal (vide infra) and its bulk band gap energy. PL emission centered at wavelengths between 380–2000 nm can be obtained with appropriate selection of the semiconductor material and nanocrystal size [198].  While a QD is approximately a physical representation of the particle-in-a-box concept, an important difference is that the core nanocrystal does not provide an infinite potential barrier for confinement of the exciton. Furthermore, the lattice structure of the nanocrystal abruptly terminates at its surface, which can lead to localized “trap” states within the quantum confined band gap. Trap states can sometimes be observed as band gap emission, which appears as a broad peak on the bathochromic side of the expected band edge emission. These states, as well as leakage of the excitonic wavefunction outside the core nanocrystal, promote non-radiative pathways for recombination of the exciton [193]. To improve PL efficiency, the core nanocrystal can be coated with a few layers of a structurally similar semiconductor with a higher band-gap energy, as is the case with widely utilized CdSe/ZnS and CdTe/ZnS QDs. Such an arrangement, where the core band edge states are both intermediate in energy to those of the shell, is referred to as a Type-I heterostructure. This configuration is the most common in bioanalytical applications since it offers the best confinement of the exciton (Figure 1.13A) and the highest rates of radiative recombination (i.e. brighter PL). Confinement is not complete, however, as shell growth is typically accompanied by a 5–10 nm bathochromic shift in the QD PL spectrum.    52  Figure 1.13 Illustration of band gap engineering by selection of core and shell materials. The relative energy of conduction band and valence band edge states between the core and shell determine the localization of the electron and hole, and the nature of the transition associated with exciton recombination, offering an additional means of tuning the optical properties of QDs. (A) Type-I QD with localization of both carriers in the core; (B) Type-II QD with localization of the electron in the shell; (C) Type-II QD with localization of the hole in the shell; (D) Quasi-Type-II QD with localization of the electron in both and the core and shell; and (E) Inverse-Type-I QD with localization of both carriers in the shell.  Other heterostructure configurations are designed to localize the electron and/or hole outside of the core nanocrystal. For example, in Type-II heterostructures (e.g. CdTe/CdSe, CdSe/ZnTe), [199] the electron and hole are localized in the shell and core, respectively, or vice versa. This behaviour arises from an offset between the band edge states of the core and shell (Figure 1.13B). The exciton recombines across the core/shell interface and, consequently, the emission wavelength corresponds to an energy less than the band gap of either the core or shell material. The decreased overlap between the electron and hole wavefunctions also results in lower absorption coefficients and longer PL decay times. Type-II QDs are potential NIR emitters and growth of a second Type-I shell (e.g. CdSe/CdTe/ZnSe) [200] can enhance quantum yields; however, other Type-I and alloyed NIR emitting QDs (e.g. InAs/ZnSe, InAs/CdSe, InAs/InP, Cu:InP/ZnSe and InAsxP1–x/InP/ZnSe) are also being actively developed [201-203]. Quasi Type-II QDs have only a small offset between, for example, the conduction band edge states of the  53 core and shell, such that the electron is delocalized over the whole nanocrystal while the hole is confined to the core (Figure 1.13C) [204]. Inverse (or reverse) Type-I QDs (e.g. CdS/CdSe, ZnSe/CdSe) [205] are designed to localize both the electron and hole into the shell. The band edge states for the shell are both intermediate to those of the core (Figure 1.13D). These configurations also require a secondary Type-I shell (e.g. ZnSe/InP/ZnS) [206] to enhance PL emission. Finally, lattice strain between the core and shell can be used to tune the optical properties of certain QDs. For example, growth of epitaxial shells on ZnS, ZnSe, CdS, or CdSe on small, soft CdTe cores can be used to shift band energies and thus PL emission. Compressive strain in the core increases the energy of its band edge states while synergistic tensile strain in the shell decreases the energy of its band edge states [207]. The effect of growing thicker shells can be large enough to induce Type-II band alignment in a Type-I heterostructure such as CdTe/ZnSe [207]. To date, Type-II QDs have not found widespread use in bioanalytical applications. 1.3.3.1 Surface states and effects The energies of band edge states are not the only determinants of QD PL. Even with growth of a Type-I shell, surface states can still affect the PL of real QDs (i.e. imperfect structures). For example, the “blinking” or fluorescence intermittency of QDs, which is perhaps the second most renowned property after their size-tunable emission, is associated with surface states. Blinking can be observed at the single particle level, has a power law probability distribution, and is a consequence of either (i) charging and discharging of the core nanocrystal, or (ii) trapping of carriers at surface states before they can relax to emissive band edge core states [208]. Auger recombination is the predominant relaxation mechanism in charged QDs, resulting in very efficient PL quenching until the QD core is neutralized. While detrimental in some applications of QDs, the observation of blinking is useful to confirm tracking of a single QD [209, 210] and has enabled super-resolution imaging [211].  In addition to blinking, QDs sometimes exhibit other interesting optical phenomena under high intensity excitation. These phenomena include bluing, photobrightening, and photodarkening, which are observable in the ensemble [212]. Bluing corresponds to an irreversible hypsochromic shift in the band edge emission, and is the manifestation of photooxidative etching of the average nanocrystal size [213]. Brightening, or photoactivation, is an increase in the QD PL intensity  54 under irradiation and is associated with changes in the properties of the QD surface. These changes have been suggested to include the passivation of defect states and dangling bonds [212], or displacement of trapped charges [214, 215], each leading to a decrease in a “dark fraction” of non-luminescent QDs in the ensemble. The extent of photobrightening, as well as the opposite effect, photodarkening, depends on the duration and intensity of irradiation, although the latter seems to be induced at higher irradiation intensities, above-gap excitation energies, and longer irradiation times. The competitive kinetics of photobrightening and photodarkening have been investigated and found to yield different steady state QD PL intensities for different irradiation intensities [215]. The aforementioned dark fraction, which has been observed experimentally via fluorescence coincidence analysis (FCA), is inversely correlated to the ensemble quantum yield [216, 217]. It has been suggested that the mechanism for formation of the dark fraction is analogous to that for blinking behavior [218], albeit that the dark fraction is not a byproduct of blinking over extended timescales [216, 217]. Interestingly, a decrease in the size of the dark fraction is responsible for the apparent increase in the ensemble QD quantum yield that is frequently observed upon “passivation” with adsorbed macromolecules such as proteins [219]. The importance of the above effects in analytical applications of QDs is variable, depending on both the characteristics of the batch of QDs utilized and the spectroscopic parameters of the experiment (e.g. laser power). Ensemble assay methodologies based on one-time measurements at low power excitation tend to be relatively immune, whereas single molecule tracking experiments with high intensity excitation are the most susceptible to these effects. In either case, good or poor quality QDs can make a tremendous difference in the outcome of an experiment. Maintaining continuity in the properties of QD materials is thus an ongoing challenge in the field. 1.3.4 Quantum dot materials  As alluded to above, QDs have been synthesized from a broad range of semiconductor materials. The most popular materials have been CdSe, CdTe and their core/shell analogs, CdSe/ZnS and CdTe/ZnS. This popularity can be attributed to well-established synthetic protocols, emission that can be size-tuned over the visible/NIR region, and, not least of all, commercial availability.  55 Traditionally, emission has been tuned on the basis of core nanocrystal size with these materials, and the role of the Type-I shell has been to passivate dangling bonds on the surface of the core, better confine the exciton (vide supra), and enhance the QD’s optical properties (e.g. the quantum yield can increase by 20–35%) [220, 221]. For this purpose, the growth of a thin shell is important. For example, with CdSe/ZnS QDs, the 12% lattice mismatch between CdSe and ZnS necessitates that growth of the ZnS shell be limited to a few atomic layers before lattice strain detrimentally affects the PL properties [222]. Thicker shells have been desirable to render QDs more robust or prevent blinking [223]. Effective approaches for growing thicker shells and relaxing lattice mismatch have included incorporating a small amount of Cd into the shell material [224], and synthesis of gradient or multi-shell structures (e.g. CdSe/CdS/ZnS) [223, 225, 226]. As an alternative to size-tuning of PL, QD core materials can also be alloyed. The PL emission of ternary alloyed QDs (e.g. CdSexTe1–x, CdSxSe1–x, CdxZn1–xS, Cd1–xZnxSe) can be varied while maintaining a constant size (Figure 1.12C) [194, 227-231], and these materials are also commercially available.  In addition to II-VI semiconductors, other materials used for QD synthesis include III-V (e.g. InP) [232, 233] or group IV (e.g. Si) [234, 235] semiconductors. To some degree, the investigation of alternative materials to CdSe and CdTe has been driven by the perceived toxicity of Cd-based QDs (see refs. [236-238] for a discussion of the complex issue of toxicity; QDs can be used in both toxic and non-toxic capacities). Although synthesis protocols for alternative materials are still being optimized to yield optical properties that match those of CdSe/ZnS and CdTe/ZnS QDs, there has been considerable progress. For example, InP/ZnS QDs [239] (with emission in the 480–750 nm range) and InP/ZnSe/ZnS QDs [240] have been reported with Φ = 0.4–0.6 and a FWHM of 50–60 nm. In addition to the benefits of NIR emission for in vivo applications, QD size plays an important role in determining their fate in vivo. Renal clearance and minimal accumulation in organs (e.g. spleen, kidney, liver, etc.) is observed with nanoparticles < 5.5 nm in hydrodynamic diameter [241]. Recently, Park et al. reported synthesis of highly luminescent CuInxSey/ZnS core/shell QDs (Φ = 0.6) with emission within the NIR biological window at 741 nm, a FWHM of 175 nm, and an average diameter of 5 nm [242]. With the exception of the large FWHM, these QDs are almost ideal for prospective in vivo applications. Some non-Cd QD materials (e.g. InP/ZnS and InGaP/ZnS) are currently available commercially.      56 1.3.4.1 Synthesis of QDs Unfortunately, the laboratory synthesis of high-quality colloidal QDs is still largely restricted to experienced chemists. Despite numerous attempts in the literature to synthesize QDs in aqueous media using convenient air-stable precursors, QDs with narrow FHWM (a function of the distribution of particle size, i.e. monodispersity) and high quantum yields have been almost exclusively obtained though solvothermal methods that use organometallic precursors and non-polar organic solvents at high temperature and under inert atmosphere (i.e. pyrolysis of inorganic precursors) [243-245]. The possible exception is the aqueous synthesis of CdTe QDs, where quantum yields have been reported to reach 82%, but are typically closer to ca. 40% [246-248]. These QDs can also be relatively monodisperse, with FWHM typically in the range of 30–60 nm. 1.3.5 Functionalization of QDs While the optical properties of QD attract the lion’s share of excitement, experts have now come to realize that the surface area of the QD is almost as valuable: A QD can serve as a nanoscale scaffold with physicochemical properties and biological activity that can be tailored through interfacial chemistry and bioconjugation. Functionalization is done in multiple steps, and the design and execution at each step is critical to the efficacy of the QD in its intended application [126, 249-253]. 1.3.5.1 Interfacial chemistry Since most biological applications use core/shell QDs, the inorganic shell is generally the first site for modification. In particular, high quality QDs prepared by solvothermal methods are coated with hydrophobic surfactants and require modification to render them water-soluble for biological applications. As shown in Figure 1.14i-v, there are two well established routes to water soluble QDs: (i) ligand exchange (i.e. replacement of the native surfactants), which yields more compact QDs; and (ii) encapsulation with an amphiphilic polymer (building around the native surfactants), which typically yields brighter QDs. Ideally, the core/shell QD PL properties are insensitive to interfacial chemistry; however, the typical few-atom thick Type-I shells do not fully isolate the nanocrystal core, and the optical properties of QDs are still somewhat affected by adsorbed molecules, pH, temperature, and other properties of the local environment [254]. This sensitivity is a consequence of imperfect confinement of the exciton and/or non-uniform  57 coverage of the shell material on the core [224]. Other important considerations for the hydrophilic modification of QDs include the net charge, colloidal stability (i.e. resistance to aggregation), long-term coating stability (i.e. stable association between the organic coating and inorganic QD), compatibility with bioconjugate chemistries (i.e. for attaching biomolecules of interest), and resistance to the non-specific adsorption of proteins and other biomolecules in a sample matrix (i.e. non-fouling). In the following paragraphs, the chemistry of coating QDs for aqueous dispersion is discussed in more detail, focusing first on the interface exposed to bulk solution, then discussing the interface between the organic coating and the inorganic QD. One of the most widely used methods for dispersing QDs in aqueous solution is to modify their outer surface with anionic carboxylate groups. At sufficiently basic pH and low ionic strength, electrostatic repulsion between QDs affords a stable colloidal suspension; however, efficient charge screening at high ionic strength and/or neutralization of the carboxylates at acidic pH yields insoluble aggregates of QDs [255, 256]. Carboxylate coatings (Figure 1.14i, v) also tend to be prone to the non-specific adsorption of proteins due to their charge. Popular alternatives to carboxylate coatings are those featuring poly(ethylene glycol) (PEG; Figure 1.14ii, iii) oligomers or zwitterionic moieties (Figure 1.14iv). Both PEGylated and zwitterionic coatings offer colloidal stability over broad ranges of pH and ionic strength, and minimal non-specific adsorption for improved biocompatibility [257, 258]. The advantage of zwitterionic coatings over those based on PEG is more compact size [259, 260]; however, PEG oligomers can be modified with a variety of terminal functional groups (e.g. carboxylic acids, amines, hydroxyl, PEG, biotin) with minimal impact on the overall colloidal stability of the QDs [261]. Albeit that numerous well-established protocols for QD ligand exchange (i.e. replacement of native hydrophobic ligands with hydrophilic ligands) have been reported, it remains a challenge to characterize this process, including the number of hydrophobic ligands remaining, the density or number of hydrophilic ligands per QD, and the homogeneity of ligand distribution on QD interface. Batch-to-batch variation in these parameters can have a significant impact on the reproducibility of assays, and the interactions observed between QDs and biomolecules (e.g. non-specific adsorption).  58  Figure 1.14 Illustrative overview of the chemistry of core-shell QDs. Coatings for aqueous solubility: (i) amphiphilic polymer coating with carboxyl(ate) groups; (ii) amphiphilic polymer coating with PEG oligomers; (iii) dithiol ligand with a distal PEG oligomer; (iv) dithiol ligand with a distal zwitterionic functionality; (v) dithiol ligand with a distal carboxyl(ate) group. Common R groups include carboxyl, amine, and methoxy, although many others can be introduced (e.g. see vi, x, xi). Methods for conjugating biomolecules of interest (BOI): (vi) biotin-streptavidin binding; (vii) polyhistidine self-assembly to the inorganic shell of the QD; (viii) amide coupling using N-(3-dimethylaminopropyl)-N′-ethylcarbodiimide (EDC) and sulfo-N-hydroxysuccinimide (s-NHS) activation; (ix) heterobifunctional crosslinking using succinimidyl-4-(N-maleimidomethyl)cyclohexane-1-carboxylate (SMCC; structure not shown); (x) aniline-catalyzed hydrazone ligation; and (xi) strain-promoted azide-alkyne cycloaddition. The double arrows are intended to represent conjugation between the functional groups and, in principle, their interchangeability (not reaction mechanisms or reversibility). Not drawn to scale.  As noted above, there are two main methods for modifying QDs with functional groups such as carboxylic acids and PEG oligomers. The first of these methods is ligand exchange, which  59 involves the replacement of hydrophobic surfactants from QD synthesis with higher-affinity hydrophilic ligands via mass-action. The most common ligands are bifunctional molecules with thiol groups that coordinate to Zn2+ sites on the surface of the QD at one end, and display carboxylate or PEG groups at the other (Figure 1.14iii-v). While thiols will also coordinate to the Cd2+ at the surface of a bare CdSe core, the ZnS shell is less prone to oxidation and Zn2+ has higher binding affinity with basic ligands, improving the coating stability of the final aqueous QDs [226]. Coating stability is also improved by using bidentate ligands with two coordinating thiol groups. For example, an extensive library of bidentate ligands derived from dihydrolipoic acid (DHLA) have been reported, including those appended with functional group-terminated PEG oligomers [256] or compact zwitterionic moieties [260]. The major challenge of ligand exchange with thiols is a reduction in the quantum yield of the QD. Considerable efforts have been made to refine ligand exchange procedures to minimize such effects [262-264]. Commercially available QDs with hydrophobic surfactants are often made water-soluble by ligand exchange with commercially available thiol ligands (e.g. 3-mercaptopropionic acid) [72].  Amphiphilic polymers are a second type of coating that can be applied to QDs, and are designed to have a mixture of hydrophilic groups and hydrophobic alkyl side-chains. The alkyl side-chains interdigitate with alkyl-bearing surfactants from QD synthesis (e.g. trioctylphosphine oxide, TOPO), leaving the hydrophilic groups at the surface of the now water-soluble QDs (Figure 1.14 i, ii). Common chemical strategies for preparing amphiphilic polymers include partial grafting of polyacrylic acid or poly(maleic anhydride) backbones with alkyl amines, where the remaining sites on the backbone are left as carboxylic acids or appended with PEG chains [265-269]. These polymer coatings better retain the original brightness of synthesized QDs since they build an additional layer onto the surface of the QD without altering coordination to the inorganic interface (i.e. less opportunity for forming surface traps). Polymer coatings also provide good long-term coating stability, but typically larger hydrodynamic radii than QDs coated with bifunctional ligands [198]. Water-soluble QDs with amphiphilic polymer coatings are available commercially, as are QDs coated with phospholipids, which interact with the as-synthesized QDs in an analogous fashion. Further details on the diversity of possible coatings for QDs, including those that are less widely utilized or still emerging (e.g. coordinating polymers [270, 271]) can be found in recent reviews [6, 126, 261, 272].  60 1.3.5.2 Bioconjugation of QDs Bioconjugation strategies of QDs can be broadly classified into (i) covalent coupling and (ii) self-assembly/specific recognition; both strategies have been used to couple enzymes, proteins, peptides, antibodies, and oligonucleotides to QDs [6, 250]. It is critical to note that, without suitable bioconjugation, the utility of QDs in bioimaging and bioanalysis will be greatly hindered, regardless of their highly favorable optical properties. Further, irreproducibility in bioconjugation also tends to translate into irreproducibility in experimental results. A key conceptual difference between QDs and fluorescent dyes is that QDs are effectively surfaces that can be modified with many biomolecules at many different sites, whereas fluorescent dyes typically have one reactive group that labels one of many sites on a biomolecule. This difference creates unique challenges for QDs and other nanoparticles, which have been thoroughly reviewed elsewhere [250]. Some of the most general and pragmatic strategies for the bioconjugation of QDs are summarized below and a few these strategies are illustrated in Figure 1.14vi-xi. Covalent conjugation methods provide a new chemical bond between a biomolecule of interest and the ligand or polymer coating of a QD. The robustness of the linkage is a function of both the bond stability and coating stability. The most common chemistry is to couple amine-bearing biomolecules to carboxylated QDs (or the opposite configuration) using amide-bond forming, water-soluble activating reagents such as N-(3-dimethylaminopropyl)-N′-ethylcarbodiimide (EDC) and sulfo-N-hydroxysuccinimide (s-NHS) (Figure 1.14viii) [273]. This approach is an effective “shotgun” method that works well in some applications and poorly in others. With many proteins, this chemistry neither provides good control over the number of proteins conjugated per QD nor their orientation (potentially affecting biological activity). Another common outcome is a fraction of crosslinked aggregates, which tend to result from coupling between the large number of amine and carboxyl groups present on the surface of a protein. EDC chemistry is often most effective with mono-reactive biomolecules, which is the case for many synthetic oligonucleotides and peptides. As an alternative to EDC, some commercial QD suppliers offer bioconjugation kits that target either amine or sulfhydryl groups on biomolecules, and couple via hydrazone ligation (Figure 1.14x) and heterobifunctional crosslinkers with maleimide groups (Figure 1.14ix), respectively [274]. These reactions tend to offer somewhat  61 better control over the final bioconjugates. The liabilities of conventional covalent conjugation methods have generated strong interest in developing highly chemoselective ligation reactions that provide excellent control over nanoparticle bioconjugation [250]. The aforementioned hydrazone ligation [275] is one such example, as is copper-free strain-promoted azide-alkyne cycloaddition (Figure 1.14xi; often called “click chemistry”) [276, 277]. Both of these chemistries have commercially available ‘chemical handles’ that can be used to modify amine-bearing QDs and biomolecules for subsequent ligation [278]. Alternative bioconjugation strategies based on self-assembly and specific recognition take advantage of high-affinity non-covalent interactions to assemble biomolecules of interest to QDs. The best-known example of specific recognition is the tight-binding (femtomolar dissociation constants) between biotin and the avidin family of tetravalent proteins (Figure 1.14vi). Almost any biomolecule can be biotinylated using commercially available kits and reagents, assuming that it is not already sold with a biotin modification. Streptavidin-modified QDs are also available commercially, permitting widespread access to a diverse array of QD bioconjugates. This strategy permits a moderate level of control over the number of biomolecules assembled per QD (conjugate valence) and their orientations; however, there are limitations associated with the heterogeneous attachment of the streptavidin to the underlying QD coating [279]. To date, the bioconjugate method that has provided the best overall control is self-assembly between polyhistidine-appended biomolecules and the ZnS shell of ligand-coated QDs (Figure 1.14vii; nanomolar dissociation constants), which provides excellent control over conjugate valence and orientation [280]. Both expressed proteins and commercially synthesized peptides can be readily obtained with polyhistidine tags. Relatively facile methods have also been developed for chemically ligating these tags to synthetic oligonucleotides [281]. Polyhistidine assembly has also been extended to commercial carboxylate polymer-coated QDs [280]. The primary advantage of polyhistidine self-assembly and biotin-streptavidin is that bioconjugations proceeds almost quantitatively without need for excess reagents and purification. A variety of other self-assembly/recognition methods have been developed, but do not yet enjoy the same widespread use and accessibility; these methods have been reviewed elsewhere [250].   62 1.3.6 Quantum dots as FRET donors A majority of FRET-based bioanalyses utilize QDs as donors for organic dye acceptors, and these configurations have several advantages over more traditional dye-dye FRET pairs. From the Förster formalism, FRET efficiencies for QD-dye FRET pairs scale according to the inverse sixth power relationship with donor-acceptor separation distance, as described in eqn. 1.23. As shown in Figure 1.15A, the sensitivity of the FRET efficiency to donor-aceptor separation is most optimum in a region between 0.5R0 < r < 1.5R0. Some of the important optical and physical properties of QDs and their impact on QD-dye FRET systems are listed below:  • Strong and broad absorption with high molar absorption coefficients in the UV region of the spectrum (ε = 106–107 M–1cm–1) o Strong QD absorption in the UV–blue region allows selection of an excitation wavelength that minimizes direct excitation of the acceptor dye (see Figure 1.15B) o When this attribute is paired with high QD quantum yields, it becomes possible to use lower intensities of excitation radiation without sacrificing signal-to-noise [282, 283]  o Low power excitation minimizes direct excitation of acceptor, and also minimizes the effect of photobleaching of organic dyes [185] • Narrow, symmetric and size-tunable QD PL with the full-width-half-maximum (FWHM) ca. 25-35 nm  o It permits optimization of the spectral overlap integral with only limited crosstalk between donor and acceptor emission  • High quantum yield (10–60%) o Higher quantum yields lead to larger Förster distance (see eqn. 1.18) and observed FRET efficiency  • Large surface area  o Supports modification with multiple acceptor dyes, thus enhancing the rate and efficiency of energy transfer compared to a discrete donor-acceptor pair  A priori, the strong and broad light absorption by QDs also suggests that they would be ideal acceptors; however, efficient and unavoidable direct excitation of the QDs, coupled with their  63 relatively long excited state lifetime, largely negates this advantage when paired with putative fluorescent dye donors (an excited state QD is not a good acceptor). The solution to this challenge has been to pair QDs as FRET acceptors with luminescent lanthanide complexes as donors [284, 285]. Lanthanide ions (e.g. Tb3+, Eu3+) typically have excited state lifetimes on the order of 10–4–10–3 s (cf. 10–9–10–8 s for dyes and 10–8–10–7 s for QDs). As such, directly excited QDs return to their ground state and become good acceptors following a microsecond delay after flash/pulsed excitation, while lanthanide ions remain in their excited state as good donors [281, 284]. Förster distances can reach ~10 nm with lanthanide-QD FRET pairs and >7 nm with QD-dye pairs [281], compared to < 6 nm with conventional dye-dye pairs.  In bioanalytical applications, the great advantage of FRET is the ability to turn QD PL ‘on’ or ‘off’ in response to biorecognition events (e.g. ligand-receptor binding, enzyme activity, DNA hybridization) or other physicochemical stimuli (e.g. pH). Since measured signals are not strictly based on the accumulation of QDs, FRET methods can be applied in the ensemble and down to the level of single particles. Numerous configurations using QDs and FRET have been reported for the detection of metal ions [278, 286], small molecules [219, 287, 288], toxins [289], drugs [290], protease [291, 292] and nuclease [293, 294] activity, hybridization assays [295, 296], immunoassays [297], and pH [298, 299]. In each case, the underlying idea is that a donor/acceptor is added or removed from the vicinity of a FRET-paired QD, either physically (e.g. association or dissociation) or through a change in its resonance (i.e. a large spectral shift).   64  Figure 1.15 (A) An illustration of the distance dependence of a QD-dye FRET pair with the corresponding FRET efficiency curve as a function of donor-acceptor separation in terms of r/Ro. (drawn not to scale). (B) Absorption and emission spectra for a QD-dye FRET pair. The shaded area indicates qualitative spectral overlap.   From the standpoint of the design of QD-FRET bioassays, the ability to exploit the non-trivial surface area of QDs and their numerous sites for bioconjugation offers significant advantages. First, QDs can be assembled with various biomolecules to achieve different functionality (e.g. sensing, targeting, etc.). Second, QDs can be assembled with an increasing number of acceptors tuning the FRET efficiency without changing the donor-acceptor separation distance. The effect of additional acceptors surrounding QDs on the FRET efficiency is described by introducing a modification of eqn. 1.23, where a is the total number of acceptors. E =R0 / r( )6in∑1+ R0 / r( )6in∑≈aR06r6 + aR06                                             (1.27) QDs act as scaffolds, providing the opportunity for a single QD to interact with multiple acceptors. Although, each acceptor may be located at a slightly different relative separation, these variations are commonly neglected and it is assumed that acceptors are located centro-symmetrically around a QD core [282]. A sequential increase in the number of acceptors leads to  65 a non-linear increase in FRET efficiency. The dependence of the FRET efficiency on the acceptor stoichiometry is best illustrated with a FRET efficiency plot as shown in Figure 1.16, where additional interactions between donor and acceptor transition dipoles generates a new scaling relationship (eqn. 1.27).   Figure 1.16 (A) Plot of FRET efficiency as a function of donor-acceptor separation in terms of r/Ro for a fixed, a, the number of acceptors per QD donor. The effective enhancement from 50% to 91% with increasing a from 1 to 10 is shown at the donor-acceptor separation equivalent to the Förster distance (dashed lines). (B) Plot of FRET efficiency as a function of the number of acceptors, a, per QD donor for a fixed r/Ro value.   Quantification of the strong dependence of the FRET efficiency on the number of acceptors requires a precise knowledge of the conjugate valence. Most bioconjugates exhibit a heterogeneous distribution of valencies in accordance with Poisson statistics, as given by eqn. 1.28, where P is the probability of forming a QD conjugate with precisely n acceptors with an ensemble average of N acceptors [300]. With sufficient resolution, a distribution of conjugate valences can be observed using gel electrophoresis, as shown in Figure 1.17.   P(N,n) = e−NN nn!                                                              (1.28)  66  Figure 1.17 Poisson distribution of QD-MBP conjugates. Agarose gel electropherogram of CdSexS1−x/ZnS QD-His5-MBP conjugates assembled from different relative amounts of MBP per QD (indicated at top). The banding is characteristic of a distribution of conjugate valences (indicated at left).  The effect of this distribution on FRET efficiency is estimated by taking into account the existence of all of the subpopulations (n = 0, 1, 2, 3, 4, etc.) and summing up the individual contributions to the FRET efficiency over the entire sample [291]: E(N ) = P(N,n) nR06r6 + nR06n=1∞∑                                                   (1.29) This equation provides a Poisson distribution weighting of the FRET efficiency. The errors in FRET efficiency introduced by assuming a single population model for N < 4 can be very significant, but such errors decrease with increasing number N. The maximum error for N ≥ 4 is ca. 5.6 % irrespective of r/R0 ratio [291]. However, for a case of N = 1 and r/R0 =1, the single population model overestimates FRET efficiency by ca. 36 %. A further decrease in the r/R0 ratio will produce even larger discrepancies [291]. Although a single population can be sufficiently accurate to describe bioconjugate systems with four or more acceptors, eqn. 1.29 should be applied for systems with small N to minimize potential errors in calculated FRET efficiencies. It was noted in Section 1.2.8.3 that the Förster formalism is only applicable in the systems that can be approximated as point dipoles, and that these dipoles are not in close contact. For spherical QDs, the charge carrier wave functions are symmetric and their maxima are located at the center of QD. Therefore, an approximation that the point dipole of the QD is located at its  67 center is generally valid. Theoretical calculations by Allan and Delerue, and Curutcher et al., have determined that the Förster formalism is valid for all direct bandgap semiconductor nanocrystals such as CdSe/ZnS QDs [301, 302]. This model also provides a good approximation even when an acceptor is in a direct contact with the QD surface, contrary to the case of organic fluorophores. Since the QD point dipole is located at the centre of the nanocrystal, the radius of QD defines the minimum possible donor-acceptor separation (i.e. 1–4 nm depending on the QD size, see Section 1.3.2). It is known that the dipole-dipole approximation fails with dyes at length scales on the order of their dimension [179]; however, FRET involving QDs is typically over distances comparable to the size of QDs, such that the approximation does not become invalid in real systems. The transition density of the QD is built up from two quasi-spherical clouds of opposite sign, while the potential it generates can be approximated as a sum of the two potentials created by the point charge located at the centers of these oppositely charged clouds [302]. In CdSe QDs, the positive and negative charges are located at the centre of the nanocrystal and separated by a dipole of ca. 7 Å [302]. As a consequence of the localized dipole at the centre of the QD, the total volume of the QD and its radius are taken into account in estimating donor-acceptor separation distances. In other words, the radius of a nanocrystal imposes a minimum separation distance for any acceptor located above the QD surface. This implication is often taken into consideration upon selection of surface ligands that are sufficiently small in size or length so as to not contribute extensively to donor-acceptor separation. This factor is often a limitation for amphiphilic coatings that offer a higher quantum yield, and protein (streptavidin) coated QDs that allow for easy bioconjugation as noted in Section 1.3.5.  The validity of the Förster formalism with QD donors has also been confirmed by a number of experimental studies. One early study used a dye labeled maltose binding protein conjugated to a QD, which allowed the separation of the dye and core to be smaller than 100 Å [300]. The authors found a good agreement between theoretical and experimental data for the FRET efficiency and spectral overlap. The apparent donor-acceptor distances calculated from efficiency and Förster distances correlated well with the bioconjugate dimensions based on structural data. Two separate studies have observed agreement between measured and predicted FRET efficiencies as a function of variation of donor-acceptor separation using a rigid polypeptide as a variable length linker [303, 304]. The FRET efficiencies determined from steady-state and lifetime measurements followed the Förster formalism and exhibited inverse  68 sixth power dependence (r–6), as indicated by eqn. 1.22 and 1.23. Pons et al. characterized FRET with single molecule spectroscopy and found a good agreement with ensemble measurements by observing a systematic increase in energy transfer efficiency with an increasing number of acceptors [305].  As was mentioned in Section 1.2.8.3, the orientation factor (κ2) is often assumed to be 2/3 for isotropic systems where the donor and acceptor transition dipoles sample a random orientation during the donor excited state lifetime. However, the QD dipole transition has significantly different angular distribution in contrast to organic fluorophores [306]. Organic fluorophores have a fixed emission dipole orientation (nondegenerate or linear emitters), whereas CdSe nanocrystals have been shown to have a degenerate transition dipole oriented isotropically in two dimensions (circular emitter), giving rise to a dark axis (c-axis) in the orthogonal plane that does not couple to the light field [306-308].  The two dimensional degenerate dipole (2D-DD) is a result of equal contributions of two transitions from two different electron-hole fine structure states driven by right- and left-circularly polarized light [308]. This phenomenon suggests that the assumption κ2=2/3 is not necessarily valid and depends on the relative orientation of acceptors to the dark axis. However, it was suggested that errors introduced by this assumption are small and the value 2/3 is a good first approximation for a system with partially random orientation of the QD transition dipole and random and dynamic orientation of an acceptor transition dipole [282]. Another advantage of using QDs as FRET donors is their typically longer lifetimes (>10 ns) in comparison to organic fluorophores. As a consequence, QDs may remain in the excited state, when many acceptors have already returned to their ground state. The result is the formation of a population of available acceptors (even when direct excitation of the acceptor is non-zero). A large majority of fluorescent dyes have lifetimes in the range of 0.5–5 ns, and decay kinetics that are described by mono-exponential decay functions [283]. The PL decay dynamics of QDs are typically more complex in comparison to organic fluorophores and appropriate data analysis is required. QDs exhibit multi-exponential behaviour that can continuously change over the course of measurements [309, 310]. The correlation of decay rates to the emission intensities revealed that at high intensities (longer lifetimes), the decay rate is mono-exponential, while at low intensities (shorter lifetimes) it is multi-exponential. This observation implies that fluctuations in  69 emission intensities are dominated by the dynamics of the non-radiative rate component, knr. Maximum-intensity single exponential decays (radiative, kr) are consistent among different batches of QDs of the same size, as it is an intrinsic parameter of the QD core, while non-radiative decay rates (quality of surface, trap sites, shell coating, ligands) contribute to multi-exponential decay dynamics [309]. In comparison, multi-exponential decays with dyes are observed when attached to organic and biological molecules upon conformational changes or when they sample two different environments (e.g. exposed and shielded from water) [309]. In order to calculate the FRET efficiency from the changes in the excited state lifetime (i.e. according to eqn. 1.24) for a QD FRET donor with a biexponential lifetime, an amplitude-weighted average lifetime is calculated according to eqn. 1.30:  τ av =A1τ1 + A2τ 2A1 + A2                                                         (1.30) where Ai denotes the amplitude and τ i  the lifetime of the ith component.  Despite the numerous advantages that QDs offer as FRET donors in comparison to organic fluorophores, the complex photophysical properties give rise to some drawbacks, which include photoactivation, emission enhancement, intermittency and bluing (these phenomena are described in detail in Section 1.3.3.1). All these process, with the exception of intermittency, introduce a systematic error in determining the FRET efficiency of the system. 1.3.6.1 Selection of QD-dye FRET pairs There are a number of factors that need to be taken into consideration when selecting QD-dye FRET pairs for bioanalytical applications. Three important considerations are (i) direct excitation of FRET acceptor dye, (ii) spectral overlap between the QD donor and the dye acceptor, and (iii) emission crosstalk between the QD donor PL in the dye acceptor PL. Since FRET can only take place between the donor in its excited state and the acceptor in its ground state, selecting an excitation wavelength that preferentially excites the donor with minimal direct excitation of acceptor is crucial. Fortunately, QD donors exhibit broad absorption and can be efficiently excited with any wavelength in the UV-visible region of the spectrum. Over this region, it is also possible to select a wavelength that provides minimal excitation of the dye acceptor. Although  70 this selection criterion is rarely discussed in the literature, it is important to consider, particularly in the designs where the excitation wavelength is restricted by the instrumentation used (i.e. laser, LED).  It is apparent at this point that spectral overlap is an essential condition for FRET to take place. However, there needs to be a careful balance as to how closely QD emission is located to dye emission. Most dyes have small Stokes shifts, such that the greater spectral overlap desired for greater FRET efficiency (i.e. QD PL superimposed with dye absorption) leads to a larger spectral crosstalk, γ from QD PL in dye emission:    γ =FQD,λ=Dye MaxFQD,λ=QD Max×100%                                                   (1.31) Table 1.2 lists some optical properties of selected Alexa Fluor dyes commonly used as FRET acceptors with QDs [311]. Table 1.3 provides information on the calculated spectral overlap (eqn. 1.20) between QD donors (FWHM 30 nm) denoted as QDX, where X corresponds to the wavelength of emission maxima, and these dye acceptors. The Förster distances, calculated according to eqn. 1.18, depend on the quantum yield of QDs, such that for all QD-dye FRET pairs, an increase in QD quantum yield from 0.1 to 0.5 results in ca. 1–2 nm increase in Förster distance. Table 1.2 Properties of common fluorescent dyes used as acceptors in QD-FRET bioassays. Dye λAbs (nm) ε(λmax)  (M–1cm–1) λEm (nm) Quantum Yield A488 496 73 000 519 0.92 A546 556 112 000 573 0.79 A555 555 155 000 565 0.10 A568 578 88 000 603 0.69 A594 590 92 000 617 0.66 A610 612 144 000 628 N/Aa A633 632 159 000 647 N/Aa A647 650 270 000 668 0.33 A660 663 132 000 690 0.37 A680 679 184 000 702 0.36 aN/A = not available. All reported values are obtained from ref. [309].  71 Table 1.3 Summary of spectral overlap, Förster distance and crosstalk for QD-Alexa Fluor dyes.   QD490 QD520 QD540 QD565 QD585 QD605 QD630 QD650 A488 J (×10-9)a 3.02               Ro (ΦQD=0.1)b 4.21        Ro (ΦQD=0.5)b 5.51        γc 14.2%        A546 J (×10-9)a 0.746 3.06 4.98           Ro (ΦQD=0.1)b 3.34 4.22 4.58      Ro (ΦQD=0.5)b 4.36 5.52 5.98      γc 0.06% 1.6% 9.0%      A555 J (×10-9)a 1.87 6.38 8.73           Ro (ΦQD=0.1)b 3.89 4.77 5.03      Ro (ΦQD=0.5)b 5.08 6.24 6.57      γc 0.10% 4.1% 21.7%      A568 J (×10-9)a 0.453 2.35 3.80 6.68 7.18       Ro (ΦQD=0.1)b 3.07 4.04 4.37 4.81 4.86    Ro (ΦQD=0.5)b 4.01 5.28 5.72 6.28 6.36    γc 0.04% 0.07% 0.3% 5.2% 41.4%    A594 J (×10-9)a 0.369 1.67 2.90 5.71 8.48 6.83     Ro (ΦQD=0.1)b 2.98 3.81 4.18 4.68 5.00 4.82   Ro (ΦQD=0.5)b 3.88 4.99 5.47 6.12 6.54 6.31   γc 0.03% 0.05% 0.09% 1.0% 10.0% 66.0%   A610 J (×10-9)a 0.193 1.12 2.19 4.94 8.46 13.9     Ro (ΦQD=0.1)b 2.66 3.57 3.99 4.57 5.00 5.43   Ro (ΦQD=0.5)b 3.48 4.67 5.22 5.98 6.54 7.11   γc 0.03% 0.03% 0.05% 0.3% 2.8% 26.0%   A633 J (×10-9)a 0.0856 0.752 1.42 3.90 6.53 10.7 18.6   Ro (ΦQD=0.1)b 2.33 3.34 3.71 4.39 4.79 5.20 5.70  Ro (ΦQD=0.5)b 3.04 4.37 4.86 5.75 6.26 6.80 7.45  γc 0.03% 0.03% 0.03% 0.06% 0.3% 0.3% 45.0%  A647 J (×10-9)a 0.0889 0.607 1.16 3.27 7.08 12.7 26.4 36.0 Ro (ΦQD=0.1)b 2.34 3.22 3.59 4.27 4.85 5.35 6.04 6.36 Ro (ΦQD=0.5)b 3.06 4.21 4.69 5.58 6.35 6.99 7.90 8.32 γc N/D N/D N/D N/D 0.06% 0.3% 5.2% 41.0% A660 J (×10-9)a 0.572 1.75 2.69 5.05 7.89 11.2 15.6 20.6 Ro (ΦQD=0.1)b 3.19 3.85 4.13 4.59 4.94 5.24 5.54 5.80 Ro (ΦQD=0.5)b 4.17 5.03 5.40 6.00 6.46 6.85 7.24 7.59 γc N/D N/D N/D N/D N/D 0.06% 0.4% 4.1% A680 J (×10-9)a 0.115 0.760 1.25 2.63 4.50 7.38 12.7 20.9 Ro (ΦQD=0.1)b 2.44 3.35 3.64 4.12 4.50 4.89 5.35 5.81 Ro (ΦQD=0.5)b 3.20 4.37 4.75 5.38 5.88 6.39 6.99 7.60 γc N/D N/D N/D N/D N/D N/D 0.1% 1.0% aSpectral overlap, J in units cm6mol–1; bFörster distance in nm; cQD crosstalk calculated using eqn. 1.31.  72 1.4 Contributions of This Thesis This thesis describes original research toward development of smartphone-based POC platforms for “one-step” multiplexed assays. “One-step” implies that no additional pre-treatment and post-washing would be required, and this goal was achieved by using the highly distance-dependent FRET assay format with QDs as donors. The author’s work began in September 2012—mere months after publication of the work by the Ozcan Laboratory demonstrating QD-based detection of E.coli [110]. Design of a smartphone-based platform using combination of QDs and FRET was an original idea of the author, and later served as inspiration for the work published by Noor et al. [312, 313]. The overall objective of this thesis was to develop a proof-of-principle diagnostic platform that exploited the unique properties of QDs and the imaging capability of smartphones. In particular, the research in this thesis emphasizes the multiplexing capabilities of QDs and FRET for protease activity assays and research towards paper-based on-chip assays amenable for direct detection in biological matrices (serum and whole blood) via QD immobilization. 1.4.1 Background In the past five years, there has been a surge in the number of publications that have used smartphones or cellphones in bioassay development. On one hand, this interest was driven by the need to develop POC diagnostics that can enable personalized medicine and increase the efficiency and accessibility of health care, particularly in resource-limited settings such as developing countries. On the other hand, today there are nearly 7 billion mobile subscriptions worldwide, equivalent to 95.5% of the world population. Smartphones have become an essential tool in people’s daily lives. Mass produced, portable and relatively low-cost, these devices provide an ever-expanding set of features including data storage, spatial mapping, temporal tracking, and wireless transmission of information. The role of smartphones for POC diagnostics falls in two main categories: (1) control modules, power sources, and data handling hubs, and (2) optical detectors utilizing their built-in cameras. The latter case was primarily developed for direct imaging applications. In one of the earliest studies, the Fletcher laboratory demonstrated brightfield imaging of sickle red blood cells and fluorescence imaging with LED excitation of Auramine O stained M. tuberculosis-infected sputum samples [314]. In 2011, the Ozcan laboratory demonstrated platform for cell counting [315, 316] that was ultimately suitable to  73 count white blood cells in brightfield mode, red blood cells in fluorescent mode, and hemoglobins in transmission/absorbance mode [317]. Numerous assays based on colorimetric detection have been developed to date [147, 152, 318-321]. For example, on-chip ELISA-based detection of an ovarian cancer biomarker, HE4, in urine was demonstrated by correlating the red pixel intensity of colour images with 3,3’,5,5’-tetramethylbenzidine (TMB) substrate concentration [147]. More recently, the Ozcan laboratory developed a 96-well ELISA using transmission mode smartphone imaging with blue LEDs as an excitation source [320]. This LED light was passed through each well, and then was collected via 96 individual optical fibers, quantifying mumps, measles and herpes simplex virus IgGs. The recent advances in the area of POC diagnostics cannot be fully appreciated without acknowledging progress in the areas of nanomaterials and nanotechnology. Gold nanoparticles became a preferred choice among colour reagents in lateral flow assays. They have large, size-dependent molar absorption coefficient, and are observed as an intense dark red colour, and exhibit pronounced colour changes (from red to blue) upon aggregation. Many Au NP-based assay formats developed previously for qualitative lateral flow assays and solution-based assays were translated into smartphone-enabled quantitative formats [79, 80, 322-324]. The second most popular nanomaterial explored for bioassay development is QDs. The unique optical and physical properties of semiconductor QDs attracted the attention of numerous research groups that develop fluorescence-based assays. From the standpoint of smartphone-based assay development, the main advantages of QDs include (i) their brightness, which permits use of low-power excitation sources; (ii) their narrow emission, which can be matched with the RGB channels of the built-in colour filters of smartphones; (iii) their broad absorption spectra, which permit use of relatively broad-band excitation sources without increasing background in detection channel; and (iv) their amenability to multiplexing. In 2012, the Ozcan Laboratory exploited the brightness of QDs to demonstrate detection of Escherichia coli O157:7 (E. coli) using cell phone imaging [110]. The inner surface of glass capillaries was used as a solid support for the immobilization of anti-E. coli antibodies. The capillaries were placed in a custom-built cell phone attachment that accommodated two sets of UV LEDs for excitation of the red-emitting QDs and a suitable longpass filter to isolate QD emission. E. coli were detected in a sandwich immunoassay with biotinylated secondary antibody and streptavidin-coated QDs. The LOD in buffer and milk samples was ca. 5–10 CFU mL−1 with good specificity. In 2015, the  74 Chan Laboratory demonstrated multiplexed detection using QD barcodes of pathogen DNA, extracted form patient samples, following an amplification step [325].  In addition to approaches that utilize QDs as fluorescent labels, QDs have been shown to be excellent FRET donors. Since the seminal work by Medintz et al. in 2003, QD-FRET assays have been developed for the detection of various analytes [185]. However, until recently, all of these assays required sophisticated laboratory instrumentation for measurements. In 2014, Noor et al. reported digital image (iPad) based detection of paper-based QD-FRET nucleic acid hybridization assay [312]. Amplification of the signals observed by drying paper substrates permitted 450 pmol detection limit with UV lamp excitation. In subsequent work, Noor et al. integrated thermophilic helicase-dependent amplification step to achieve a 37 zmol detection limit with digital imaging [313].   Despite many advances to the field of smartphone diagnostics, as demonstrated by the studies highlighted above, there remains many limitations to translating these technologies in to molecular diagnostics of wider utility. To date, the vast majority of these assays were based on direct imaging of the sample (e.g. cells, bacteria) or commercially available kits. Other challenges associated with POC devices often include poor analytical sensitivity, reproducibility, and lack of multiplexing. 1.4.2 Thesis overview This thesis is divided into seven chapters, including this introductory chapter, and is complemented by the appendices that contain additional experimental methods and details. Published results are noted at the beginning of each chapter. Chapter 2 describes a simple, 3D-printed prototype accessory that enables all-in-one smartphone excitation and imaging of PL. The requirement of an external excitation source for smartphone imaging of fluorescence assays is eliminated by demonstrating that the built-in LED photographic flash of the smartphone can be used for this purpose. Moreover, this chapter shows that the sensitivity of smartphone fluorescence imaging with built-in flash excitation is competitive with imaging using an external UV lamp, and that QDs are superior for smartphone imaging assays over traditional dyes and fluorescent proteins. A proof-of-concept protein  75 binding assay and FRET-based protease activity assay were demonstrated, and the use of QDs as FRET donors can permit imaging of fluorescent dyes that would otherwise not be bright enough to detect using all-in-one smartphone PL excitation and imaging. This work provides a foundation for smartphone-based PL bioassays without external optoelectronic devices, helping to maximize simplicity and robustness while minimizing cost, which are important features for prospective POC diagnostic technologies.   Chapter 3 demonstrates a first step towards the development of multiplexed assay of proteolytic activity with QD peptide conjugates. The spectrally narrow PL emission and strong, broad light absorption of QDs was combined with red-green-blue (RGB) digital color imaging for quantitative, multiplexed homogeneous assays. Alloyed CdSeS/ZnS QDs with emission in the blue, green, and red region of the spectrum were carefully paired with FRET acceptors, such that the changes in QD PL as a result of proteolysis were measured from RGB channel intensities in digital colour images. Importantly, this work shows that quantification of proteolytic rates is identical for smartphone based detection and a sophisticated fluorescence plate reader. This work was recently highlighted by Stevens as an important contribution to nanoparticle based POC diagnostics [326], and, to date, remains the only report of smartphone digital imaging as a readout for multiplexed FRET-based assay with QD donors.  Chapter 4 evaluates three surface chemistries for the immobilization of water-soluble QDs on cellulose paper fibers via self-assembly. These chemistries included:  i. modification of oxidized cellulose with 1-(3-aminopropyl)imidazole to afford substrates modified with imidazole ligands;  ii. modification of oxidized cellulose with N-(2-aminoethyl)-5-(1,2-dithiolan-3-yl)pentanamide, a derivative of lipoic acid, to afford substrates with thiol ligands;  iii. modification with lipoic acid via silanization step with 3-aminopropyltriethoxysilane (APTES) to afford substrates modified with thiol ligands.  These methods yielded a high density of ligands that were successful in immobilizing QDs at high density. It was possible to immobilize QDs coated with a variety of ligands, as well as pre-assembled QD-peptide conjugates. Furthermore, each of the three evaluated chemistries had a different effect on the photobleaching behaviour of QDs, their long-term stability, the density of  76 immobilized QDs, and the enhancement in the rate and efficiency of FRET between immobilized QDs and dye acceptors within paper matrix in comparison to solution-phase FRET. Overall, development of underlying chemistry of cellulose fibers used for immobilization of QDs offers a promising platform for the design of new methods of bioanalysis. Furthermore, it acts as an important component of assay development providing opportunities and challenges in controlling the properties of immobilized QDs and the extent of the interactions between biomolecules and cellulose interface. Understanding the effects of the underlying surface chemistry was a crucial component for development of the solid-phase assays described in Chapter 5 and Chapter 6. Chapter 5 evaluates the suitability of using paper immobilized QDs for FRET-based transduction of protease activity. This chapter represents one of the most significant contributions of this thesis, as it merges many ideas developed to combine smartphone imaging with a “one-step” assay format via QD-FRET. The PL response was acquired with a smartphone and compared with other devices such as a miniature fiber-optic spectrometer ($2000), an educational-grade digital microscopy camera ($250), and a consumer webcam ($40). The diagnostic utility of QD-modified paper substrates with each of the foregoing detection systems was demonstrated through a series of solid-phase proteolytic assays based on FRET for simultaneous detection of three proteases that belong to pancreatic family of enzymes—enterokinase, trypsin, and chymotrypsin. Overall, this chapter makes the idea of smartphone-based QD-FRET assay in complex biological matrix developed in Chapter 6 more feasible.  Chapter 6 outlines design criteria and demonstrates proof-of-concept for an assay format that utilizes smartphone readout for the single-step, FRET-based detection of hydrolase activity in serum and whole blood, using thrombin as a model analyte. This study builds on the work described in Chapter 2-5, which highlights the design components necessary for an assay format that permits direct measurements in whole blood. Essentially, every component was carefully optimized to circumvent challenges of whole blood as a matrix including: the LED illumination, the QD-dye FRET pair, the reference signal, the paper test strips, and the PDMS-on-glass sample chip. Overall, this chapter shows how smartphones and QDs can be integrated to permit assays in serum and whole blood in a format that will ultimately be suitable for many point-of-care diagnostic applications. An overview of the main design components and a summary of  77 guidelines that can be used to “build your own smartphone-based QD-FRET assay platform” are shown in Figure 1.17. Chapter 7 provides a brief summary of the achievements presented in the preceding chapters, as well as an outline for future work. In summary, this thesis presents unique contributions to the areas of nanotechnology, bioanalytical chemistry, and the development of point-of-care diagnostics with consumer electronic devices. Smartphone imaging combined with QDs, FRET and paper substrates was successfully demonstrated for multiplexed bioanalysis, as well as addressing challenges of performing assays in whole blood.   Figure 1.18 Overview of main design components and a flowchart of critical optimization parameters. 78 Chapter 2 Use of a Smartphone