Photo-induced dark states in
fluorescence spectroscopy –
investigations & applications
ANDRIY CHMYROV
Doctoral Thesis in Physics
Royal Institute of Technology
Stockholm, Sweden 2010
Thesis for the degree of Doctor of Philosophy in Physics to be presented with due
permission of Royal Institute of Technology for public examination and criticism on
Thursday April 29th in FA31, Albanova University Center, KTH, Stockholm
Copyright © Andriy Chmyrov
Stockholm 2010
Applied Physics
Experimental Biomolecular Physics
Royal Institute of Technology
SE-106 91 Stockholm
Sweden
The following papers are reprinted with permission:
Paper I Copyright © 2007 by American Chemical Society
Paper IV Copyright © 2008 by The Royal Society of Chemistry and Owner Societies
Paper VI Copyright © 2009 by American Chemical Society
Paper VII Copyright © 2010 by the authors; licensee MDPI Publishing, Open access
TRITA-FYS 2010:20
ISSN 0280-316X
ISRN KTH/FYS/--10:20--SE
ISBN: 978-91-628-8108-5
Printed by E-Print, Stockholm
Abstract
This thesis focuses on investigations of transient dark states of fluorescent
molecules using spectroscopic techniques. The main purpose is to show and
convince the reader that transient dark states are not always a nuisance, but
also represent an additional source of information. Several studies with fluorescence correlation spectroscopy were performed, all related to non-fluorescent states such as triplet state or isomerized states.
Photobleaching is one of the main problems in virtually all of the fluorescence techniques. In this thesis, mechanisms that retard photobleaching are
characterized. Several compounds, antioxidants and triplet state quenchers,
which decrease photobleaching, are studied, and guidelines for achieving optimal fluorescence brightness using these compounds are presented.
Triplet state quenching by several compounds was studied. Detailed investigations of the fluorescence quencher potassium iodide demonstrated
that for some of fluorophores, except of quenching, there is fluorescence enhancement mechanism present. In agreement with the first publication in
this thesis, antioxidative properties were found to play an important role in
the fluorescence enhancement. Quenching of the triplet state is proposed
as a tool for monitoring diffusion mediated reactions over a wide range of
frequencies.
Specially designed fluorophores combining high triplet yields with reasonable fluorescence brightness and photostability were characterized for
possible applications in novel super-resolution imaging techniques based on
fluorescence photoswitching. Except of benefits for imaging techniques, photoinduced switching to non-fluorescent states could be used for monitoring
molecular diffusion, which was also demonstrated in this thesis.
Studies of the triplet state kinetics of fluorophores close to dielectric interfaces were performed using fluorescence spectroscopy. The analysis of the
triplet state kinetic can provide information about the local microenvironment and electrostatic interactions near dielectric interfaces.
Keywords: fluorescence correlation spectroscopy, triplet state, isomerisation,
photobleaching, quenching, diffusion, total internal reflection, interface
iii
Papers
This thesis is based on the following papers (reprinted with permissions by
the publications):
Paper I
Strategies to Improve Photostabilities in Ultrasensitive Fluorescence Spectroscopy
Jerker Widengren, Andriy Chmyrov, Christian Eggeling, Per-Åke Löfdahl,
and Claus A. M. Seidel
The Journal of Physical Chemistry A, 2007, 111 (3): 429-440
DOI: 10.1021/jp0646325
Contribution by author: The author performed part of the measurements
and data analysis.
Paper II
Iodide as a triplet state promoter and quencher – mechanisms and possible
implications
Andriy Chmyrov, Tor Sandén and Jerker Widengren
Manuscript
Contribution by author: The author performed all of the measurements
and data analysis. The author worked out a major part of the theoretical
framework. TS assisted with initial measurements and data analysis.
Paper III
Nitroxide spin-label quenching of fluorophore’s triplet state as a tool for studying
diffusion mediated reactions in lipid membranes
Johan Strömqvist, Andriy Chmyrov, Sofia Johansson, August Andersson,
Lena Mäler and Jerker Widengren
Manuscript
Contribution by author: The author performed the characterization of the
quencher and the dye and performed solution measurements.
Paper IV
Characterization of new fluorescent labels for ultra-high resolution microscopy
Andriy Chmyrov, Jutta Arden-Jacob, Alexander Zilles, Karl-Heinz Drexhage
and Jerker Widengren
Photochemical and Photobiological Sciences 2008, 7 (11): 1378-1385
DOI: 10.1039/b810991p
Contribution by author: The author performed all of the measurements and
data analysis.
iv
Paper V
Recovery of photo-induced reversible dark states utilized for molecular diffusion
measurements
Andriy Chmyrov, Tor Sandén and Jerker Widengren
Manuscript
Contribution by author: The author performed all of the instrument and
experiment design, measurements, data acquisition and analysis. The author
worked out a major part of the theoretical framework. TS assisted with
discussions, simulations and preliminary measurements.
Paper VI
Triplet-State Investigations of Fluorescent Dyes at Dielectric Interfaces Using
Total Internal Reflection Fluorescence Correlation Spectroscopy
Hans Blom, Andriy Chmyrov, Kai Haßler, Lloyd M. Davis and Jerker
Widengren
The Journal of Physical Chemistry A, 2009, 113 (19): 5554-5566
DOI: 10.1021/jp8110088
Contribution by author: The author together with HB conducted the
measurements. The author performed reference measurements on a confocal
setup, and computer simulations.
Paper VII
Electrostatic Interactions of Fluorescent Molecules with Dielectric Interfaces
Studied by Total Internal Reflection Fluorescence Correlation Spectroscopy
Hans Blom, Kai Haßler, Andriy Chmyrov and Jerker Widengren
International Journal of Molecular Sciences, 2010, 11(2): 386-406
DOI: 10.3390/ijms11020386
Contribution by author: The author together with HB performed a large
part of the measurements and data analysis.
Related publications by author not included in the
thesis:
Paper VIII
Maximizing the Fluorescence Signal and Photostability of Fluorophores by
Quenching Triplet and Radical States
Daniela Pfiffi, Stanislav Kalinin, Denis Dörr, Ralf Kühnemuth, Sebastian
Overmann, Brigitte A. Bier, Andriy Chmyrov, Jerker Widengren, Thomas J.
J. Müller, Klaus Schaper and Claus A. M. Seidel
Manuscript
v
Contents
Abstract
iii
Papers
iv
Chapter 1. Introduction
1
Chapter 2. Fluorescence
5
5
2.1 Fluorescence and photophysics
2.1.1 Historical background
5
2.1.2 Electronic states
7
2.1.3 Absorption and emission spectra
9
2.1.4 Fluorophore structure
11
2.1.5 Triplet state
12
2.1.6 Isomerised state
14
2.1.7 Fluorescence quenching
16
2.1.8 Photobleaching
17
2.2 Fluorescence microscopy
19
2.2.1 Confocal microscopy 19
2.2.2 TIR microscopy
24
2.2.3 Super-resolution microscopy
29
2.3 Fluorescence correlation spectroscopy
33
2.3.1 Confocal FCS
34
2.3.2 TIR FCS
42
2.3.3 TRAST spectroscopy
45
Chapter 3. Photobleaching & additives
49
3.1 Antioxidants n-propyl gallate and ascorbic acid
50
3.2 Triplet state quenchers mercaptoethylamine and cylcooctatetraene 55
3.3 Optimisation of fluorescence output
vi
57
Chapter 4. Triplet state & quenchers
59
4.1 Iodide as triplet state promoter and quencher
60
4.2 Triplet state quenching for monitoring diffusion mediated reactions 66
Chapter 5. Triplet state & applications
71
5.1 Novel fluorescent labels for super-resolution microscopy 72
5.2 TRAST spectroscopy for measuring diffusion
77
Chapter 6. Triplet state & surface effects
83
6.1 Triplet state kinetic rates at dielectric interfaces
84
6.2 Electrostatic interactions at dielectric interfaces
89
Chapter 7. Concluding remarks
95
Acknowledgements
99
References
101
vii
Chapter 1
Introduction
Fluorescence is a wonderful tool which provides great specificity and sensitivity for the investigation, analysis, control and diagnostics in many fields relevant
to physical, chemical, biological and medical sciences. During the past 20 years
there has been a remarkable increase in the use of fluorescence especially in the
biological sciences. Fluorescence right now is a dominant tool used extensively in
biotechnology, flow cytometry, medical diagnostics, DNA sequencing, forensic and
genetic analysis, just to name a few [Lakowicz 2006]. Utilization of fluorescence
for cellular and molecular imaging has also increased dramatically, especially with
development of fluorescent proteins [Tsien 1998], which enable genetic encoding
of a fluorescent label into the proteins of interest. New imaging techniques based on
fluorescence offer single molecule sensitivity and are approaching single molecule
resolution [Hell 2009a]. Fluorescence correlation spectroscopy (FCS) was developed for the characterization of the dynamics of molecular processes in systems at
thermodynamic equilibrium [Magde et al. 1972]. With proper averaging of stochastic single molecule fluctuations, it provides an insight into single molecule dynamics
by measuring statistical fluorescence intensity fluctuations on a macroscopic scale.
Being used most often for measuring diffusion coefficients and concentrations, it
can access any kinetics originating from a molecular process that manifests itself
as a change in the fluorescence intensity [Krichevsky and Bonnet 2002]. Not only
fluorescence emission per se provides useful information, but also its absence constitutes a complimentary or even major source of data. This thesis deals with a non1
Charpter 1. Introduction
fluorescent part of the fluorescence process – photoinduced transient dark states,
such as triplet, isomerised or radical states of fluorescent molecules.
The thesis starts with an extended introduction to different aspects of fluorescence and some of the fluorescence techniques: photophysics, fluorescence microscopy and correlation spectroscopy, focusing mainly on features relevant for the following chapters. Further four chapters present the work behind the seven papers. In
the last chapter the work is summarized and concluded.
Some of the fluorescence techniques were invented too early to be widely applied
immediately – both confocal microscopy and fluorescence correlation spectroscopy
belong to them. Several decades of technological advances were needed, but now
these techniques represent the standard tools for molecular imaging and monitoring of molecular interactions. Nonetheless, further improvements, both in terms of
technology and methodology, are persisting to arise, offering new possibilities and
extending the practical applicability of these techniques. This thesis contributes to
such everlasting efforts.
Despite almost 140 years since the first chemical synthesis of Fluorescein (one
of the first and still one of the most widely used fluorophore) [Pawley 2006], the
development of new fluorescent probes is still a necessity. New fluorescence imaging
and detection methods desire to obtain higher fluorescence output rates in harsher
excitation conditions for longer times. In Chapter 3 (paper I) problems of extension and maximization of fluorescence output are addressed, by the use of carefully
selected additives. In this chapter it is shown that a large part of non-fluorescent
fluorophore radicals can be regenerated back into active fluorescent emitters. Balancing the amount of additives is crucial for the success of such experiments, since
any excess of additives will turn viable fluorophores into undesired dark states.
Fluorescence quenching is widely recognized as a highly versatile tool and applied for studying molecular interactions [Lakowicz 2006]. In Chapter 4 (papers II
and III) the behavior of two common quencher molecules – potassium iodide and
TEMPO is investigated. In paper II the beneficial effects of fluorescence recovery
from transient dark states are observed for several fluorophores in the presence of
iodide. A recipe in the sense of fluorophores and concentration range is proposed in
order to take advantage of such effects. In paper III a quenching mechanism of triplet states of fluorophores is proposed for monitoring diffusion in lipid membranes.
Advantage of the relatively long triplet lifetime is utilized, while still keeping the
high sensitivity of fluorescence as a detection mode.
The presence of various long-lived dark transient states of fluorophores has gen2
Andriy Chmyrov
erally been considered disadvantageous, as they reduce the fluorescence brightness
and render the data analysis more complicated [Rasnik et al. 2006]. Chapter 5
(papers IV and V) reveals a different view on such dark states. In paper IV the
fluorophores were intentionally developed and characterized for having high triplet
yields, while being still reasonably fluorescent and photostable. The combination of
such properties meets the requirements of new super-resolution imaging techniques.
In paper V a concept is described where the transitions to reversible dark states
are monitored for measuring the mobility of the molecules. This method features
a simple instrumentation, possible parallelization and a wide range of accessible
concentrations.
It is well known that the fluorescence properties can be modified close to interfaces [Hellen and Axelrod 1987]. Chapter 6 (papers VI and VII) address the
properties of fluorophores near dielectric interface using the phenomenon of total
internal reflection as an excitation mode. Since the use of the triplet states as readout
parameter is increasing, paper VI provides an insight on photophysical kinetics of
the fluorophore in the proximity of a dielectric interface. In paper VII the electrostatic interactions between dielectric surfaces and fluorophores possessing different
electric charges are investigated. Although negligible from macroscopic point of
view, these small charges and their interactions are vital for understanding dynamics
of molecules close to interfaces.
3
Chapter 2
FLUORESCENCE
2.1 Fluorescence and photophysics
Fluorescence is a form of luminescence, the physical process of emission of light
upon excitation of a molecule. There are several different types of luminescence,
categorized according to the mode of excitation – photoluminescence (including
fluorescence, phosphorescence and delayed fluorescence), chemoluminescence (resulting of a chemical reaction), bioluminescence (by a living organism), electroluminescence, (in response to an electric current passed through it), mechanoluminescence (resulting from any mechanical action on a solid), sonoluminescence (in
response to ultrasound) and others [Valeur 2002]. In this thesis we will be dealing
with photoluminescence – the process of emitting a photon caused by absorption
of a photon(s).
There are two types of photoluminescence: fluorescence and phosphorescence.
Fluorescence is typically a fast process (nanoseconds), involving the radiative relaxation of a molecule from a state with a paired spin to the ground state. Phosphorescence is normally a slow process (milliseconds), which involves the radiative
relaxation from a state of higher spin multiplicity, usually a triplet state.
2.1.1. Historical background
The first reported observation of fluorescence was made by the Spanish physician
5
Chapter 2. Fluorescence
Nicolas Monardes in 1565. He described the wonderful peculiar blue color of an infusion of the wood called Lignum Nephriticum. This wood was further investigated
by Boyle, Newton and others, but the phenomenon was not understood. The first
reported observation of phosphorescence was made in 1602 by an Italian cobbler,
Vincenzo Cascariolo, whose hobby was alchemy [Valeur 2002]. One day he went
for a walk in the Monte Paterno area and he picked up some strange heavy stones.
After calcination with coal, he observed that these stones glowed in the dark after
exposure to light. It was recognized later that the stones contained barium sulfate,
which, upon reduction by coal, led to barium sulfide – a phosphorescent compound.
Despite the fact that first recorded observation of phosphorescence was made
later than that of fluorescence, the term “phosphorescence” is much older (meaning that this phenomenon was observed earlier, but the knowledge about that did
not pass through time). It comes from the Greek word φωσφορος [fōsforos] = light
carrying, (φωV [fōs] = light, φερω [ferō] = to carry). The term phosphor has indeed been assigned since the Middle Ages to materials that glow in the dark after
exposure to light [Valeur 2002]. The term “fluorescence” was introduced by Sir
George Gabriel Stokes, a physicist and professor of mathematics at Cambridge in
the middle of the nineteenth century. In his paper from 1853 he invented the term
fluorescence, from the name of the fluorspar (mineral containing calcium fluoride:
fluorite), which was known to emit light when exposed to solar light beyond the
violet part of the spectrum. Interestingly, fluorescence in this mineral is due to the
presence of small amounts of impurities (mainly europium ions, yttrium and dysprosium), because fluorite itself is not fluorescent.
The year earlier, Stokes reported about the phenomenon of emitting a light following absorption of light [Stokes 1852]. He formed the solar spectrum using a
prism, and while moving a tube filled with a solution of quinine sulfate through the
spectrum he observed a blue glow of the solution in the non-visible part of the spectrum corresponding to UV excitation. Stokes stated that the emitted light is always
of longer wavelength than the exciting light, today known as Stokes’ law. However,
already 10 years before that, the French physicist Edmond Becquerel published a
paper where he reported about wavelength shifts for light emitted by calcium sulfide,
which is phosphorescent. The term “luminescence” (coming from the Latin lumen
= light) was introduced first as luminescenz by the physicist and light historian
Eilhardt Wiedemann in 1888, to describe ‘all those phenomena of light which are
not solely conditioned by the rise in temperature’, as opposed to incandescence (the
emission of light, typically red and infrared, from a hot body due to its high temperature) [Valeur 2002].
6
Andriy Chmyrov
Energy
E1
E0
Nuclear coordinate
Figure 2.1: Franck-Condon principle energy diagram. The potential wells are shown favoring
transitions between the vibrational states n = 0 and n = 2.
2.1.2 Electronic sates
The process of fluorescence can be divided into three main events, all of which
occurs at different timescales, separated by several orders of magnitude. First, the
excitation of a molecule occurs in femtoseconds (10-15 s) by absorption of an incoming photon of suitable energy. As a result, an electron gets promoted from the
highest occupied molecular orbital (HOMO) to the lowest unoccupied molecular
orbital (LUMO). Since the mass of an electron is at least three orders of magnitude
lower than the mass of the nuclei, the transition time is too short for any significant
displacement of nuclei. As a consequence, all electronic transitions in the energydistance plot are vertical [Condon 1926; Franck and Dymond 1926], as it is described by the Franck-Condon principle. This means that an electronic transition
out of the lowest vibrational state of the ground state S0 of the molecule, which is
mainly populated at room temperature conditions according to the Boltzmann distribution, will take place into a higher vibrational level of the first electronic excited
state. Franck-Condon principle is applied equally to absorption and to fluorescence
emission. This is illustrated in figure 2.1.
After excitation, the vibrational relaxation (typically on picoseconds, 10-12 s,
time scale) occurs to the lowest vibrational state of the excited state [Kasha 1950]
7
Chapter 2. Fluorescence
kreISC2
S2
T2
kISC2
kS12 kS21
kT12 kT21
S1
T1
kISC
k02 k01 k10
kIC
kT
S0
Figure 2.2: Jablonski diagram. Electronic (thick horizontal lines) and associated vibrational
energy levels and the most important excitation and de-excitation pathways are shown. Rates
subscripts: XY (where X = 0, 1, S1, T1, S2, T2 and Y = 0, 1, 2) – designate transition from
the initial state X to the final state Y of the same multiplicity; IC – internal conversion; ISC
– intersystem crossing from singlet to triplet state; reISC2 – reverse intersystem crossing from
second excited triplet state; T – triplet relaxation from first excited triplet state, T1, to the
ground state, S0.
(Kasha’s rule – after the American scientist of Ukrainian origin Michael Kasha).
Finally, emission of a longer wavelength photon and return of the molecule to the
ground state occurs in the relatively long time period of nanoseconds (10-9 s).
Several important deviations from this simple scheme can occur, most of which
are presented on Jablonski diagram [Jabłoński 1935] of fluorescence, named after
polish physicist of Ukrainian origin Alexander Jablonski (1898-1980). One of the
many possible variants of this diagram is presented in figure 2.2.
The electronic states are arranged vertically by energy and grouped horizontally
by spin multiplicity (described later). The vibrational ground states of each electronic state are indicated with thick lines, the higher vibrational states with thinner lines.
Typically, a fluorescent molecule at the lowest ground state (S0) is in a singlet
state, which means that molecular orbital is occupied by two electrons with an opposite spin direction. Multiplicity is the quantification of the amount of unpaired
electron spin and is calculated as 2s+1, where s is the number of singly occupied
electrons multiplied by the electron spin projection quantum number mS = -1/2 or
+1/2 [Banwell and McCash 1994]. In the singlet state all of the electrons occupying
molecular orbital are paired (have opposite spins) and the multiplicity is 1. This
arrangement is strictly required by the Pauli Exclusion Principle. Upon excitation
the electron preserves its spin and because of that the multiplicity of higher excited
states (S1, S2) is also 1. The process of excitation in fluorophores is always a singlet8
Andriy Chmyrov
singlet transition. After excitation to higher vibrational sublevels of a higher excited
singlet state (typically S1) a molecule quickly relaxes to the lowest vibrational level
of the higher excited singlet state. Due to a short time (few nanoseconds) for which
the electron resides in S1 state, and because of the energy mismatch for S0-S1 and
S1-S2 transitions, there is typically no further excitation to higher singlet states. From
the lowest vibrational sublevel of S1 state there are different de-excitation pathways
to return to the ground state, S0. The most common one is a transition to one of
higher vibrational sublevels of the ground singlet state. This transition can be radiative (accompanied by emission of a photon) – which is called fluorescence, or nonradiative (no photon will be emitted, excess of energy is released as heat) – which
is called internal conversion. There are a number of other possible non-radiative
transitions – intersystem crossing, intramolecular charge transfer, conformational
change, and different pathways due to intermolecular interactions such as electron
transfer, proton transfer, energy transfer, excimer/exciplex (excited state complex)
formation, photochemical transformations [Turro et al. 2009].
2.1.3 Absorption and emission spectra
Since electronic transitions in fluorophore molecules occur between discrete energy levels, one may expect absorption and emission spectra to be a series of sharp
lines, as it is for atoms (see figure 2.3). However, a peculiarity of the spectra of
organic fluorophores as opposed to atomic and ionic spectra is the width of the
absorption and emission bands, which usually covers several tens of nanometers
[Schäfer 1990]. This becomes clear if one recalls that a typical dye molecule may
have at least fifty (and usually more) atoms, giving rise to at least 150 normal vibrations of the molecular skeleton. Many of these vibrations are closely coupled to the
electronic transitions by the change in electron densities over the bonds constituting
the conjugated chain. After the electronic excitation has occurred, there is a change
in bond length (typically 1%-2%) due to the change in electron density. Since the
bond is lengthened in the excited molecule, atoms will start to oscillate, classically
speaking, around this new position. A molecular skeletal vibration is excited this
way. In general case of a large fluorophore molecule, many normal vibrations are
coupled to an electronic transition. Additionally, collisional and electrostatic perturbations, caused by surrounding solvent molecules, broaden the individual lines of
vibrational states. As a further complication, every vibronic sublevel of every electronic state, including a ground state, has superimposed on it a ladder of rotationally
excited sublevels. These are extremely broadened because of the frequent collisions
with solvent molecules which hinder the rotational movement so that there is quasicontinuum of states superimposed on every electronic level. The population of these
9
Chapter 2. Fluorescence
A
0→2
2←0
Fluorescence
0→5
0→4
0→6
0→3
Absorption
0→1
1←0
5←0
4←0
6←0
3←0
B
λex max. = 525 nm
λem max. = 552 nm
0–0
Energy
(inversely proportional to wavelength)
400
450
500
550
600
650
700
wavelength / nm
Figure 2.3: A. Schematic representation of the absorption and fluorescence spectra corresponding to the energy diagram in figure 2.1. Electronic transitions between the lowest vibrational levels of the electronic states (the 0-0 transition) have the same energy in both absorption and fluorescence. B. Spectrum of a commonly used fluorescent dye, Rhodamine 6G
levels in contact with thermalized solvent molecules is determined by a Boltzmann
distribution. After an electronic transition which leads to a non-equilibrium FranckCondon state, the approach to thermal equilibrium is very fast in liquid solutions
at room temperature. The reason is that a large molecule experiences at least 1012
collisions/s with solvent molecules, so that equilibrium is reached in time on the
order of one picosecond [Schäfer 1990].
Because of all the mentioned reasons, a typical absorption spectrum is practically continuous all over the absorption band (see figure 2.3). The same is true for
the fluorescence emission corresponding to the transition from the electronically
excited state of the molecule to the ground state. Thus, the emission spectrum is
a mirror image of the absorption spectrum displaced towards longer wavelength
[Stokes 1852] (due to partial lost of the energy in vibrational relaxation – Stokes
shift) by reflection at the wavelength of the purely electronic transition. In some
cases, excitation by high energy photons leads to the population of higher electronic
and vibrational levels (S2, S3, …), which quickly dissipate the excess of energy by internal conversion as the fluorophore relaxes to the lowest vibrational level of the first
excited state (see figure 2.2). Because of this rapid relaxation process, emission spectra are generally independent of the excitation wavelength, which is referred as the
Vavilov’s rule [Wawilow 1927] (some fluorophores emit from higher energy states,
but such activity is rare). For this reason, emission is the mirror image of the ground
state to lowest excited state transitions, but not of the entire absorption spectrum,
which may include transitions to higher energy levels [Lakowicz 2006]. In addition,
one needs to compare the symmetry of excitation and emission spectra in a linear
10
Andriy Chmyrov
plot having wavenumber (the reciprocal of wavelength or the number of waves per
centimeter) on the abscissa axis, because wavenumber is directly proportional to the
frequency and quantum energy.
In addition, the polarity of a solvent and the local environment of the fluorophore molecule will generally influence it’s the fluorescence emission spectra. This
property is known, when talking about spectral effects, as solvatochromism or, in
a more general way, as perichromism (peri: around). This dependence is due to the
fact that fluorescence lifetimes (1-10 nanoseconds) are usually much longer than
the time required for relaxation of solvent molecules (10-100 picoseconds). Since
absorption of light occurs on much faster timescale (femtoseconds), the absorption
spectrum is much less dependent from polarity. Solvents of higher polarity shift
emission spectra to longer wavelengths [Lakowicz 2006].
2.1.4 Fluorophore structure
Electronic transitions are characterized by their energies. Working with fluorescence, one is usually interested in transitions which energy differences fall into
the visible region of the electromagnetic spectrum (380-750 nm). Fluorescent molecules are typically organic unsaturated compounds, consisting of hydrocarbons and
their derivatives and possessing at least one double or triple chemical bond. Double
and triple bonds consist of one s (sigma) and one or two π (pi) covalent bonds.
Sigma bonds are the strongest type of covalent chemical bonds and are characterized
by the rotational symmetry of their wavefunctions in respect to the bond direction
[Banwell and McCash 1994]. Sigma bond is formed by two electrons in atomic s
orbitals, one s and one pz orbital or two pz orbital (z is defined as the axis of the
bond). Pi bonds are usually weaker than sigma bonds, and they are formed by two
electrons of atomic p orbitals which are overlapping laterally. Pi bonds are rigid and
do not allow rotation around the bond axis. Organic compounds without double
or triple bonds usually absorb at wavelengths below 160 nm, which not only fall
out of visible region but constitute energies that are higher than dissociation energy
of most chemical bonds. Therefore photochemical decomposition is likely to occur
upon absorption of such high energy photons. Carbon-carbon double –C=C– (or
triple –CºC–) bonds absorb photons with wavelength of 170 nm, which is still
far from the visible region. If two double bonds are separated by a single bond
–C=C–C=C–, the two double bonds are called conjugated. Two conjugated double
bonds absorb at 220 nm, three conjugated double bonds – at 260 nm [Yadav 2004].
By increasing the number of the conjugated bonds, absorption and emission wavelengths increase. Thus, fluorophore molecules absorbing and emitting in the visible
11
Chapter 2. Fluorescence
A
B
C
HN
O
NH
≅
O
C
O
Figure 2.4: Chemical structures for A: benzene, B: anthracene, C: Rhodamine 6G.
range of the spectra possess several conjugated double bonds. For this molecules
the highest occupied molecular orbital (HOMO) is typically a bonding π orbital,
and the lowest unoccupied molecular orbital (LUMO) is typically antibonding π*
orbital. Absorption of a photon typically induces π®π* transition in fluorophores
[Schäfer 1990]. In conjugated systems π orbitals typically extend over the whole
system and allow free electron movement around it, which is called delocalization of
electrons. In most cases the more delocalized π orbitals are, the lower the energy of
the transition is, and the longer the wavelength. In benzene (figure 2.4A, chemical
formula C6H6) π orbitals are completely delocalized above and below of the plane of
carbons, so on the structure figures three alternating double bonds are typically substituted by one ring inside. Compounds with delocalization of electrons in a ring (as
benzene) are characterized by increased chemical stability and called aromatic due
to historical reasons (however only some of them have notable aromas) [Birks 1970].
A single aromatic ring (benzene) absorbs light strongly around 180 nm with
weaker band at 200 nm, the added conjugation of three rings in anthracene (figure
2.4B) shifts absorption to around 250 nm. Substitution of one of the central carbons in anthracene by oxygen makes xanthene, which is the basis of a large family
of dyes which includes Fluorescein, Eosins, and Rhodamines. Xanthene dyes absorb
in the blue to yellow region of the spectra and fluoresce from green to red region,
covering most of the visible range. Introducing additional groups to the xanthene
unit makes Rhodamine 6G (figure 2.4C), which is one of the most well studied
fluorophores because of its use in dye lasers [Duarte and Hillman 1990].
2.1.5 Triplet state
De-excitation via the intersystem crossing pathway is of particular interest in the
context of this thesis. Intersystem crossing is a non-radiative transition between two
iso-energetic vibrational sublevels belonging to electronic states of different multiplicities (see figure 2.2). Typically, intersystem crossing occurs from a singlet state
12
Andriy Chmyrov
T
S
S
T
S
S
crossing
avoiding
T
T
mixed
T
T
Spin-orbit
coupling
Figure 2.5: Left: intersystem crossing is strictly forbidden for lowest vibrational singlet (S) and
triplet (T) state. Right: intersystem crossing is partially allowed when spin-mixing mechanism
is available near the crossing point of the energy curves for the S and T states.
to a triplet state, however, opposite transitions (reverse intersystem crossing) are
also possible, especially from higher excited states [Widengren and Seidel 2000;
Ringemann et al. 2008]. Because the triplet state has a higher multiplicity, it has a
lower energy level than the excited single state, this is referred as Hund’s rule [Banwell and McCash 1994]. Therefore, from the energy point of view this transition
can be possible, because it is dissipative. However, these transitions are inevitably
accompanied by a spin flip and are classically “forbidden” by quantum mechanics,
just as a direct excitation from the ground state S0 into the triplet state T1 is forbidden. Due to spin-orbit coupling (coupling between the orbital magnetic moment
and the spin magnetic moment) there is a mixing of vibrational sublevels of singlet
and triplet states (see figure 2.5). Under these conditions classical prohibition is relaxed and intersystem crossing transition becomes possible. Presence of heavy atoms
[Kasha 1952; McGlynn et al. 1969; Ketsle et al. 1976] (with large atomic number,
like Br, I, Pb) or paramagnetic species [Hoijtink 1960] (including oxygen [Stracke
et al. 1999]) in the structure of the molecule or its surroundings can influence the
efficiency of intersystem crossing.
For some fluorescent molecules this de-excitation is very efficient and occurs
faster than fluorescence. For example in benzophenone the rate of intersystem crossing kISC is 1010 s-1, whereas the rate of fluorescence k10 is 106 s-1, meaning that intersystem crossing occurs at a rate which is 104 faster than fluorescence, making fluorescence improbable. After intersystem crossing and following vibrational relaxation,
the molecule cannot easily return to the excited singlet state because of the energy
difference. Nor can it easily return to the ground state (which is a singlet state), as
13
Chapter 2. Fluorescence
this transition requires another “forbidden” spin flip. Hence, the triplet excited state
usually has a long lifetime (compare to lifetimes of the other excited states), because
it has generally nowhere to which it can easily go. Because of this long lifetime, at
suitable conditions (2-color excitation or high irradiances) there can be an excitation to higher triplet states if photon(s) with the matching energy will be absorbed
[English et al. 2000; Widengren and Seidel 2000; Ringemann et al. 2008]. The molecule will eventually relax back from the excited triplet state T1 to the ground singlet
state S0 either radiationless (via interactions with surrounding molecules by different mechanisms - dissipating energy as heat, energy transfer to another molecule in
triplet state, e.g. oxygen) or by emitting a photon (phosphorescence). Although this
process is again classically “forbidden”, it nevertheless occurs when there is no other
open pathway by which the molecule can dissipate its excess of energy.
2.1.6 Isomerised state
Fluorescent molecules containing double bond between carbon atoms in the
chemical structure typically have an additional de-excitation mechanism involving
cis- trans- isomerisation [McCartin 1965]. The terms cis and trans are from Latin, in
which cis means “on the same side” and trans means “on the other side” or “across”.
Double chemical bonds are rigid and do not allow free rotation around their axis,
but there is a possibility to make a 180° twist around it. This transition changes
the structure of the molecule and, strictly speaking, makes a new entity with different physical properties. Since photoisomerisation takes place in the conjugated
hydrocarbon chain, this signposts that fluorophores belonging to cyanine (carboxycyanine) family will exhibit this relaxation pathway. Cyanine is a non-systematic
name of a synthetic dye family belonging to polymethine group. They were first
synthesized over a century ago, and there are a large number of cyanines reported
in the literature.
Cyanines were originally used to increase the sensitivity range of photographic
emulsions, also as passive modelockers in lasers, as media in CD-R and DVD-R
discs. Cyanine dyes are generally characterized by high extinction coefficients, short
fluorescence lifetimes and relatively low fluorescence quantum yield <0.30 (due to
the large flexibility of long conjugated chain). They cover green to red span of the
visible spectrum and are among the few fluorophores that absorb at far red wavelengths, which define their popularity in biomedical applications due to low autofluorescence and reduced damage in cells. A variety of low-cost, energy efficient,
rugged diode lasers and highly sensitive detectors in the visible-near-IR region also
facilitates the extensive use of long wavelength cyanine dyes. The importance of
cyanine dyes has therefore motivated a large amount of scientific work concerning
14
Andriy Chmyrov
A
O
O
O
N
N
O
hν
N
N
hν
trans-isomer
cis-isomer
B
Energy
N1
kPperp
kNperp
S1
P1
Perp1
k01 k10
kperpN
kperpP
kP01 kP10
S0
kPN
N0
0°
N
P0
90°
180°
θ
P
Figure 2.6: A. Chemical structures for trans- and cis- isomers of Cy5. B. Kinetic scheme normally adopted to model the photophysical behavior of carbocyanine dyes. In the model, N0
and N1 denote the ground singlet and the first excited singlet of the thermodynamically stable
trans- conformation of the dye. P0 and P1 are the corresponding states of the photoisomerised
cis- form. Upon isomerisation the bond angle, denoted by q, in one of the double bonds of the
conjugated hydrocarbon chain connecting the two headgroups of the dye is twisted by 180°.
Photoinduced isomerisation and back-isomerisation take place from N1 and P1 via the partially
twisted intermediate state, Perp1. At Perp1 deactivation to the ground-state takes place, either
to N0 or to P0. The parameters k01 and k10 denote the excitation and deexcitation rate of transisomer, with kP01 and kP10 being the corresponding rate parameters for the photoisomerised
cis- state. kPN is the rate of thermal deactivation of P0 to N0 in the absence of excitation.
their photophysical properties. Among the most popular fluorophores belonging to
cyanine family are Cy3, Cy5 (figure 2.6A) and Alexa Fluor 647.
Absorption and emission spectra of cis- and trans- isomers are usually shifted
in relation to each other, with cis isomer being shifted more to the red region the
visible spectrum [McCartin 1965; Chibisov 1966; Tinnefeld et al. 2001; Huang
et al. 2005]. Except of the difference in absorption/emission spectra, there is significant difference also in fluorescence brightness. Cis- isomer is considered to be
a “dim”, non-fluorescent state. This is due to a larger flexibility of the conjugated
chain, which makes non-radiative internal conversion decays to be the predominant
pathways of de-excitation. Usually, trans- isomers are more stable than the cis- isomers. This is partly due to their shape; the straighter shape of the trans- isomer leads
15
Chapter 2. Fluorescence
to hydrogen intermolecular forces that make the isomer more stable (see figure 2.6).
Transitions between trans- and cis- state are driven by excitation, and shift of the
absorption spectrum leads to difference in excitation rate for different photoisomers.
This defines the steady state proportion between trans- and cis- isomers during constant excitation (which is about 50% in each state for Cy5 at excitation with 594 nm
or 633 nm light). Figure 2.6B presents the Jablonski diagram normally adopted to
model the photophysical behavior of cyanine dyes [Widengren and Schwille 2000].
Triplet state population and intersystem crossing are rather inefficient for most
of cyanines, due to the competition with effective deexcitation via photoisomerisation. A direct populating of the triplet level of cyanine fluorophores occurs under
conditions preventing or limiting the process of trans-cis isomerisation [Chibisov
1976]. Therefore, the triplet energy levels are omitted in figure 2.6B for simplicity.
Different environmental factors influence the rate of isomerisation, among others
– viscosity of the surrounding medium, temperature, solvent polarity and absence
or presence of sterical hindrances [Korobov and Chibisov 1983; Aramendia et al.
1994; Noukakis et al. 1995].
2.1.7 Fluorescence quenching
The term quenching refers to any process which decreases the emitted fluorescence of a given substance. A variety of processes can result in quenching, such as
excited state reactions, energy transfer, complex-formation and quenching due to
collisions with other molecules [Lakowicz 2006]. Typically fluorescence quenching
is perceived as an undesirable process, because it reduces the fluorescence output.
However, it can be taken advantage of as a valuable source of information about the
interactions of a fluorescent molecule with a quencher. There are countless examples
of such applications, to name a few – to study structure and dynamics of proteins
[Eftink and Ghiron 1981; Zhuang et al. 2000; Yang et al. 2003; Chattopadhyay et
al. 2005; Nettels et al. 2007] and interactions with nucleotides [Seidel 1991; Seidel
et al. 1996; Widengren et al. 1997; Bonnet et al. 1998; Eggeling et al. 1998a; Zhu
et al. 2005].
It has been known for a long time that fluorescence is quenched by certain anions [Pringsheim 1949]. The quenching ability strongly depends on the chemical
nature of the anion, with quenching ability being stronger for iodide (I-), bromide
(Br-) and less stronger for chloride (Cl-). As one of the mechanisms of quenching by
anions, a charge transfer reaction could be considered [Drexhage 1990].
Another mechanism – resonance energy transfer [Förster 1948] (often referred as
16
Andriy Chmyrov
FRET – Fluorescence or Förster Resonance Energy Transfer) occurs over distances
up to 10 nm through nonradiative dipole–dipole coupling. FRET efficiency depends on the distance between the donor and the acceptor, the spectral overlap of
the energy donor emission spectrum and the energy acceptor absorption spectrum,
the relative orientation of the donor emission dipole moment and the acceptor absorption dipole moment. This mechanism is being extensively used in many biophysical fields and there are many reviews of various applications published [Clegg
1992; Selvin 2000; Jares-Erijman and Jovin 2003; Roy et al. 2008; Schuler and
Eaton 2008; Clegg 2009; Matthews et al. 2010].
Quenching by the so called heavy atom effect originates due to enhancement
of the intersystem crossing rate to the triplet state. There can be quenching due to
internal [Chandra et al. 1978; Solov’ev and Borisevich 2005] (when quencher is a
part of a molecule itself ) or external [Ketsle et al. 1976; Bryukhanov et al. 1992;
Rae et al. 2003] (with quencher being present in the solution) heavy atom effect.
Mechanistically, it responds to a spin-orbit coupling enhancement produced by a
heavy atom, which decreases the energy difference and facilitates transitions from
the singlet to the triplet state.
Further more, at high concentrations of fluorophores there can be fluorescence
quenching due to aggregation [Jelley 1936; Levshin and Nizamov 1966; Drexhage
1990; Kelkar et al. 1990; Yuzhakov 1992; Bergström et al. 2001; Marmé et al. 2005]
(dimerisation), which is most pronounced in solutions where the solvent consists
of small, highly polar molecules – notably water. Dispersive forces between the
large fluorophore molecules tend to bring them together, which is slightly counteracted by repulsive Coulomb forces if the fluorophores are charged. Upon aggregation fluorescence intensity is reduced and absorption spectrum radically changes
its shape, with an enhancement of the short-wavelength part at the expense of the
long-wavelength part.
Excited state reactions also contribute to quenching of fluorescence [Tomin
2008]. For example, the absorption spectrum of fluorophore Rhodamine 6G in
ethanol is unchanged even at concentrations as high as 10 mM, which suggests that
there is no dimerisation. However, fluorescence at such concentrations is strongly
reduced due to collisions of the excited state molecules with those in the ground
state [Baranova 1965].
2.1.8 Photobleaching
Photobleaching is usually defined as the irreversible decomposition (formation
of non-fluorescent product) of fluorescent molecules in the excited state because
17
Chapter 2. Fluorescence
of their interaction with molecular oxygen (or impurities in the solution) before
fluorescence may occur. Because of its influence as a limiting factor for the performance of, basically, all fluorescence techniques, photobleaching has been extensively
studied by various means [Widengren and Rigler 1996; Eggeling et al. 1998b; Dittrich and Schwille 2001; Eggeling et al. 2005; Eggeling et al. 2006; Yeow et al.
2006; Vogelsang et al. 2008]. The average number of excitation and emission cycles
that occur for a particular fluorophore before photobleaching is dependent upon its
molecular structure and the local environment. Some fluorophores bleach quickly
after emitting only a few photons, while others that are more robust can undergo
thousands or millions of cycles before bleaching occur.
Several mechanisms could lead to photobleaching of a fluorophore, typically
involving triplet states and/or radical states. The triplet state is relatively long-lived
with respect to the singlet state, thus allowing excited molecules a much longer
timeframe for interaction with the environment. The triplet state has two unpaired
electrons and thus has a radical character, reacting with impurities, dissolved oxygen, solvent molecules or other fluorescent molecules to yield various decomposition products. Especially the oxygen concentration is a very important factor in
photobleaching dynamics, because of its involvement both in quenching of the
triplet state [Kawaoka et al. 1967; Wilkinson 1997; Hubner et al. 2001] and in the
production of non-fluorescent compounds. The quenching of the triplet state of the
fluorophore facilitates the relaxation from the excited triplet to the ground singlet
state, resulting in the formation of the higher energy singlet oxygen species. Such
singlet oxygen reacts rapidly with exposed chemical groups in organic fluorophores;
amino acids such as cysteine, histidine, tyrosine, and tryptophan [Davies 2004];
and guanosine in DNA [Sies and Menck 1992]. The oxidized dyes are thereafter no
longer fluorescent, and such oxidative damage impairs the folding and function of
biomolecules. At different experimental conditions oxygen can therefore either enhance or reduce photobleaching affecting the equilibrium between triplet quenching and radical formation. Substituting the oxygen in the solution by another efficient triplet quencher decreases photobleaching and extends the observation times
of the fluorophores [Rasnik et al. 2006; Widengren et al. 2007; Aitken et al. 2008;
Vogelsang et al. 2008].
Under certain circumstances, the photobleaching effect can also be utilized to
obtain specific information that would not otherwise be available. For example,
in fluorescence recovery after photobleaching (FRAP) experiments [Axelrod et al.
1976; Meyvis et al. 1999; White and Stelzer 1999; Lippincott-Schwartz et al. 2001],
fluorophores within a target region are intentionally bleached with excessive levels
18
Andriy Chmyrov
of irradiation. As new fluorophore molecules diffuse into the bleached region of the
specimen (recovery), the fluorescence emission intensity is monitored to determine
the lateral diffusion rates of the target fluorophore. In this manner, the translational
mobility of fluorescently labeled molecules can be ascertained within small (few
micrometers) region of a single cell or section of living tissue.
2.2 Fluorescence microscopy
The most popular application of fluorescence is in optical microscopy. Among
the advantages of fluorescence in microscopy are its specificity and sensitivity, high
spatial and temporal resolution and compatibility with live biological systems. The
high specificity is due to the fact that detected signal (fluorescence) originates only
from the introduced fluorescent labels, with most of the background being efficiently suppressed – ‘you see only what you want to see’. The sensitivity comes from high
brightness of fluorophores and very efficient detectors, sensitive to single photons.
The spatial resolution of optical systems is limited by diffraction of light, however
several microscopy techniques has successfully circumvented this limitation [Hell
and Wichmann 1994; Betzig et al. 2006; Rust et al. 2006; Hell 2007]. Temporal
resolution on the fast time scale is limited so far only by electronics and instrumental development is advancing rapidly. Temporal resolution on the slow time scale is
limited by photobleaching, which can be reduced with various additives. Furthermore, development of fluorescent proteins [Tsien 1998] has greatly facilitated the
use of fluorescence in live cell imaging due to the possibility of genetically encoding
of a fluorescent label into the proteins of interest.
2.2.1 Confocal microscopy
A distinctive characteristic of confocal microscopy is the presence of a pinhole
in the light detection path (see figure 2.7), which reduces out-of-focus background
and decreases depth of field of view, which by-turn makes possible to do optical
sectioning and 3D image recording.
The basic concept of confocal microscopy was originally developed by Marvin
Lee Minsky in the mid-1950s (patented in 1957) when he was a postdoctoral student at Harvard University, working on imaging of neural networks in unstained
preparations of brain tissue [Minsky 1957]. This invention was severely ahead of it’s
time due to lack of intense light sources (lasers were not yet invented) necessary for
fluorescence imaging and the computer data-processing power required to handle
large amounts of data. Unintentionally, shortly after Minsky’s patent had expired,
practical laser scanning confocal microscope designs were translated into working
19
Chapter 2. Fluorescence
focal plane
objective
lens
excitation source
(laser)
dicroic mirror
tube
lens
pinhole
detector
Figure 2.7: Illustration of the principle of a confocal fluorescence microscope. The excitation
light is reflected by a dicroic mirror and is focused by an objective lens with high numerical
aperture into the sample, creating a small excitation volume. Fluorophores inside the excitation volume are excited and emit fluorescence in all directions. Part of fluorescence from the
excitation volume is collected by the objective lens, passes through a pinhole which discriminates out-of-focus signal (dotted lines), and is recorded by a detector.
instruments by several investigators. But it was not until almost 30 year after its invention that the first 3-dimensional images of fluorescently labeled biological samples were demonstrated by the use of laser scanning confocal microscopes [Carlsson
et al. 1985; Carlsson and Åslund 1987; White et al. 1987]. The first commercial
instruments appeared in 1987. During the 1990s, advances in optics and electronics afforded more stable and powerful lasers, high-efficiency scanning mirror units,
high-throughput fiber optics, better thin film dielectric coatings, and detectors having reduced noise characteristics. In addition, improved fluorescent labels started
to be synthesized. Coupled to the rapidly advancing computer processing speeds,
enhanced displays, and large-volume storage technology that emerged in the late
1990s, all prerequisites were created for a virtual explosion in the number of applications that could be targeted with laser scanning confocal microscopy [Pawley 2006].
Even though confocal microscopy provides only a marginal improvement in both
axial and lateral optical resolution compare to conventional widefield optical epifluorescence microscopy, but is able to exclude from the resulting images secondary
fluorescence in areas of the focal plane. The principal sketch of a typical confocal
microscope arrangement is presented in figure 2.7.
20
Andriy Chmyrov
As seen from the sketch in figure 2.7, in confocal microscopy only a single point
of a specimen is illuminated at a time. To receive an image of the whole specimen,
some scanning arrangement needs to be introduced. In the original instrument built
by Minsky the laser beam was kept stationary and the specimen itself was moved on
a scanning stage. This arrangement has the advantage of uniform imaging properties
over the entire image area, which can eliminate most lens defects (like off- axis aberrations and vignetting) that would affect the image. For soft biological specimens,
however, movement of the specimen can cause wobble and distortion, resulting in
a loss of resolution in the image. The speed and accuracy of the movements of the
scanning stage would thus contribute as limiting factors for temporal and lateral
resolution.
Another possibility is to scan the laser beam with a help of a system of mirrors.
This is done typically with computer-controlled galvanometer mirrors, which oscillate around a central position [Paddock 2000]. One mirror oscillates faster, scanning excitation and detection pathway around fast axis (in either lateral or axial
direction). Another mirror oscillates slowly, forming two-dimensional image in a
raster way. A third dimension is added by scanning the objective with stepper motor (typically in axial direction). Several sophisticated methods of scanning has been
developed with time [Saggau 2006], with mechanisms without moving parts for
example using liquid crystal optics [Khan and Riza 2006] or arrays of laser emitting
diodes [Poher et al. 2007]. Compare to specimen scanning, beam scanning is much
less demanding concerning mechanical precision, because the scanning mechanism
can be placed on the image side of the microscope objective. The specimen is not
subjected to any mechanical influence, and its size and weight are of no relevance
to the scanning process. The disadvantage with beam scanning is that the imaging
properties will not be uniform over the image area due to off-axis aberrations and
vignetting.
As early as in the middle of the 19th century, fundamental works of the German
physicist Ernst Karl Abbe laid the foundations of light microscopy. He developed
many important concepts, among them a mathematical description for the resolution limit of the microscope, often referred as Abbe’s resolution limit.
Optical resolution describes the ability of an imaging system to resolve details
in the object that is being imaged. The ability of a lens to resolve details is usually
determined by the quality of the lens but is ultimately limited by diffraction [Hecht
2001]. The point spread function (PSF) describes the response of an imaging system
to a point source. Light coming from a point in the object is diffracted by the lens
aperture and thus forms a diffraction pattern in the image plane, which has a central
21
Chapter 2. Fluorescence
spot and surrounding bright rings, separated by dark nulls. This pattern is known as
the Airy pattern (see figure 2.8A), and the central bright lobe as the Airy disk. The
intensity of the Airy pattern (the Fraunhofer diffraction pattern of a circular aperture) is mathematically given by:
 2 J (r ) 
PSF(r ) =  1 
 r 


2
(1)
2π
r NA , q is the half-focusing angle of the objective lens, r is a radial
λ
space coordinate, l is wavelength of light, NA = n sinq is the Numerical Aperture
of the objective, which characterizes the ability of the objective lens to focus/collect
light, n is the index of refraction of the medium in which the lens is working, J1 is
the Bessel function of the first kind of order one. The airy pattern can be successfully
approximated with the Gaussian function (figure 2.8B).
where ρ =
According to the empirical diffraction limit, known as a Rayleigh criterion, the
images of two different points are regarded as just resolved when the principal diffraction maximum of one image coincides with the first minimum of the other
(figure 2.8C-D). If the distance is greater, the two points are well resolved and if it
is smaller, they are not resolved [Hecht 2001]. Mathematically, this corresponds to
an intensity dip of 26.4 % between the peaks. The radius of the first zero-intensity
fringe of the Airy disc, Rayleigh lateral resolution, is given by:
∆rwidefield =
0.61l
NA
(2)
These considerations for resolution assume that the object is viewed in conventional wide-field microscopy. Likewise, due to reciprocity (illumination and detection is done via the same objective lens) this formula is valid for the minimum size
of diffraction-limited focus of the excitation volume (Figure 2.8E).
In the case of confocal microscopy, the point spread function will be squared:
PSFmicroscope=(PSFlens )2, because the diffraction of illumination source will be combined with the diffraction of the pinhole in the image plane. In the limit of an
infinitely small pinhole, its image will be identical to PSF. Defining the resolution
according to the same Rayleigh criterion, one obtains:
22
Andriy Chmyrov
%
&
'
UDGLDOFRRUGLQDWHLQXQLWVRIO
1RUPDOL]HG36)DPSOLWXGH
$
(
FRR
í
í
RI
LWV
Q
QX
HL
DW
UGLQ
LDO
UDG
D[LDOFRRUGLQDWHLQXQLWVRIO
O
UDGLDOFRRUGLQDWHLQXQLWVRIO
)
D[LDOFRRUGLQDWHLQXQLWVRIO
í
í
í í
UDGLDOFRRUGLQDWH
í í
UDGLDOFRRUGLQDWH
Figure 2.8: A. Airy pattern. B. A radial cross-section through the Airy pattern (solid curve)
and its Gaussian profile approximation (dashed curve) for NA=1.2. C. The Rayleigh criterion
for resolution. D. Detected intensity distribution from two point-sources just resolved by
the Rayleigh criterion. E. The axial intensity distributions for a typical widefield fluorescence
microscope, approximated as the Gaussian function – the excitation volume. F. The axial
intensity distributions for a typical confocal fluorescence microscope – the detection volume.
∆rconfocal =
0.44l
NA
(3)
23
Chapter 2. Fluorescence
The axial resolution, i.e. the resolution along the optical axis of the microscope
or z-axis, is defined using the three-dimensional diffraction image of a point source
that is formed near the focal plane [Born and Wolf 1999]. The PSF along the optical
axis has the following mathematical form:
2
 z 
 sin 

z
PSF( z ) = sinc 2 =  4 
4  z 
 4 
(4)
2π
NA 2 z , and z is the coordinate in the axial direction with its origin
nλ
in the focus.
where ζ =
Using Rayleigh criterion for widefield imaging, one obtains axial resolution limit:
2nl
NA 2
(5)
1.41nl
NA 2
(6)
∆z widefield =
For confocal case:
∆z confocal =
In contrast to the lateral resolution, the axial resolution decreases with the inverse square of the numerical aperture of the objective, NA. The ratio of axial-tolateral resolution is substantially larger than one and is inversely proportional to the
NA of the objective.
2.2.2 TIR microscopy
In conventional widefield and confocal microscopy illumination is done with a
broad cone of light. This is disadvantageous due to inefficiency of light utilization,
photobleaching and undesirable excitation and scattering from out-of-focus regions
– referred as background signal. The introduction of a confocal pinhole significantly
restricts some of mentioned effects, but does not eliminate them completely. Multiphoton excitation microscopy [Denk et al. 1990] goes a step further by restricting the excitation area to an ellipsoid having sub-micron dimensions owing to the
very low probability of the near-simultaneous absorption of multiple photons. This
24
Andriy Chmyrov
θ2
water n2
kr
k′
θC
k
θ
θ
θ1
kr′
glass n1
Figure 2.9: Illustration of total internal reflection. Plane waves are incident on a glass-water
from far away. k ¢ and kr¢ are the wave vectors of incident and totally reflected waves, respectively; k and kr are the wave vectors of incident and refracted waves, respectively; q is an
incidence angle of a totally reflected wave; qC is a critical angle; q1 is an incidence angle of a
refracted wave; q2 is an angle of refraction.
reduces the amount of out of focus photobleaching and secondary fluorescence
produced by fluorophores in the cone above and below of the excitation volume.
Both confocal and multiphoton fluorescence microscopy produce optical sections
of similar size. If there is a need to produce an excitation/detection volume in farfield from the objective lens, there is no alternative to focusing of light with a lens.
Oppositely, if one is interested in studying processes by fluorescence which occurs
directly at the interface of lens-specimen – in near-field, the phenomenon of total
internal reflection could be taken advantage of.
Total internal reflection (TIR) fluorescence microscopy employs the unique
properties of an induced evanescent wave to selectively illuminate and excite fluorophores in a restricted region immediately adjacent to a glass-water interface [Axelrod
et al. 1984]. The basic concept of TIR is simple, requiring only an excitation light
beam traveling at a high incident angle through the solid glass coverslip (see figure
2.9). Refractive index differences between the glass and water regulate how light is
refracted or reflected at the interface as a function of incident angle. Refraction at
the interface is described by Snell’s law:
n1 sin q1 = n2 sin q2
(7)
For a plane wave that is incident from the optically denser medium at the interface, (n1 > n2), at a specific critical angle, the beam of light is totally reflected from
the glass/water interface, rather than passing through and refracting in accordance
with Snell’s Law. This corresponds to refraction at 90 degrees (q2=90°, sinq2=1), and
25
Chapter 2. Fluorescence
the condition for the critical incidence angle arises from Snell’s Law:
sinqC =
n2
n1
(8)
Total internal reflection does not occur suddenly as a new phenomenon at the
critical angle, but a continuous transition is followed from predominant refraction
with a small amount of reflection (at low incidence angles), to total reflection when
the critical angle is exceeded. As the incident angle increases toward the critical
angle value, the transmitted (refracted) beam diminishes in intensity while the reflected beam grows stronger. At all angles greater than the critical angle (q1 > qc ),
total internal reflection takes place, in which essentially all of the light is reflected
back into the first medium (this effect builds the basis for light guidance in optical fibers and waveguides). Even though the light no longer propagates into the
second medium, there is a small amount of penetration of light across the interface,
which then propagates parallel to the surface, creating an electromagnetic field in
the second medium immediately adjacent to the interface [Hecht 2001]. Solving
the Maxwell equations for this case shows that the z component of the wave vector
for the refracted beam becomes imaginary. The intensity is therefore not zero inside
the medium with refractive index n2 but decays exponentially with the distance, z,
to the interface.
I ( z ) = I (0) exp (− z d )
(9)
The penetration depth (d ), at which the intensity decays by a factor 1/e, is dependent upon the wavelength of the incident illumination l0, the angle of incidence q, and the refractive indices of the media at the interface n1 and n2, according
to the equation:
d=
−1 2
λ0 2 2
n1 sin θ − n22 )
(
4π
(10)
At small incidence angles, light waves propagating through the interface to the
lower-refractive index medium are sinusoidal, and have a characteristic period. At
increasing angle, approaching the critical value, the period of refracted rays becomes
longer and the propagation direction becomes more nearly parallel to the interface.
26
Andriy Chmyrov
When the critical angle is achieved, the wave period becomes infinite and the refracted light wavefronts are aligned perpendicular to the interfacial surface. This
field is termed the evanescent field, or evanescent wave, and within a limited region
near the interface, usually less than 200 nanometers, it is capable of exciting fluorophores. In fact, the excitation of fluorescence was used as an early proof for the
existence of the evanescent field [Woods 1934].
The evanescent wave intensity at the surface I(0) is a function of both the
incident angle and the polarization components of the light beam. It is composed of
contributions from parallel (P-pol, I 0P ) and perpendicularly (S-pol, I 0S ) polarized
fields, which are
P
0
I =I
P
4 cos 2 q (2 sin 2 q − n 2 )
n 4 cos 2 q + sin 2 q − n 2
I 0S = I S
4 cos 2 q
1− n2
(11)
(12)
where n = n2 n1 , IP and IS denote incident intensities. Up to a factor of 5 in enhancement of the excitation intensity compare to the incident intensity is possible
using TIR excitation on a glass/water interface. This enhancement factor can be
increased even further on metal films [Burghardt and Thompson 1984; Hellen and
Axelrod 1987]. The maximum enhancement is reached when the beam is incident
at the critical angle and the axial extent of the evanescent field approaches very large
values. Figure 2.10A shows the resulting levels of irradiance due to the evanescent
field for different angles of incidence for P- and S-polarized incident light.
Fluorophores existing in the close proximity of an interface do not emit light
isotropically, unlike those dispersed in bulk solution [Lukosz and Kunz 1977a; Lukosz and Kunz 1977b; Lukosz 1979]. Instead, fluorescence emission is produced
in a complex spatial pattern that is highly dependent upon the orientation of the
fluorophore transition dipoles with respect to the interface geometry [Hellen and
Axelrod 1987; Enderlein et al. 1999]. Characteristics of fluorescence emission close
to an interface between dielectric media can be derived by modeling the fluorophore
by a classical dipole. For a randomly oriented dipole near an interface separating
two dielectric half-spaces the radiated power is canalized along the critical angle and
emission into the medium with higher refractive index is favored. This is illustrated
in Figure 2.10B, where the emission profiles were calculated using a formalism proposed by Mertz [Mertz 2000].
27
Chapter 2. Fluorescence
Enhancement factor I0 / Iinc
A
B
6
5
90 4
3
150
p-polarized
s-polarized
4
120
60
30
2
1
180
3
0
2
1
0
60
210
θC
65
70
75
80
85
90
water 240
330
270
300
glass
Angle of incidence θ (degrees)
Figure 2.10: A. Normalized enhancement factor of the evanescent field for P-polarization
I0P/I P and for S-polarization I0S/I S of the incident light at a glass (n1 = 1.52) and water (n2 =
1.33) interface as a function of the angle of incidence. B. Normalized radiated power displayed
as a polar plot assuming a randomly oriented dipole located at the water/glass interface, z = 0.
The right half-circle (±90°) shows the highly anisotropic emission into the glass, with a maximum around the critical angle (qC ≈ 61°), which can be collected by high numerical-aperture
microscope objective. The normalization is such that radiated power is scaled to that of the
same dipole in the absence of the interface (which has a value of 1 on this polar graph).
From technical instrumental point, two main configurations of total internal reflection fluorescence microscopy (TIRFM) exist [Axelrod 2001], called prism-type
TIRFM and objective-type TIRFM. In prism-type TIRFM a glass prism is used to
achieve the evanescent excitation at the sample. It is placed above the sample, which
is sandwiched between two coverglasses. A water immersion objective lens is placed
below the lower coverglass and immersion oil is used on the interface between the
prism and the upper coverglass to eliminate refractive index change. A laser beam
traveling inside the prism is used to generate an evanescent wave at the interface
between the upper glass-slide and the sample.
In objective type TIRFM, an oil-immersion objective with high NA (≥ 1.4) is
used for the excitation and the collection of emitted light. A laser beam is focused
onto the back focal plane of the objective, which causes the beam to emerge collimated from the objective. This beam is focused off-axis onto the back focal plane,
and therefore it emerges under a certain angle q > 0 out of the objective. If the angle
is large enough, the beam is totally reflected at the coverglass/water interface and an
evanescent field builds up at the surface.
Several variations of setups with these configurations exist, including configurations that allow using a conventional light source instead of a laser, or a waveguide
instead of a prism to generate an evanescent field. Both setups have their advantages
28
Andriy Chmyrov
and disadvantages. Both of these most common configurations were successfully applied to image single fluorescent molecules. The prism-type TIRFM benefits from
decoupling of excitation and emission paths, which offers more flexibility regarding
selection of incidence angle and significantly reduces the background due to the fact
that excitation light does not travels into emission pathway. One major disadvantage of the prism-type excitation is the lower collection efficiency. In objective-type
TIRFM a high NA objective is used that collects light efficiently. Moreover, fluorescence emission in close vicinity of the interface is highly favored into the direction
of the interface (Figure 2.10B). Larger part of this anisotropic emission goes into
the glass and propagates at an angle close to qC. This fluorescence is mostly collected
by high NA objectives in objective-type TIRFM, but totally missed in prism-type
TIRFM (because it goes in the direction opposite to objective lens). It was shown
[Enderlein et al. 1999] that the fraction of collected fluorescence can exceed 60%
for objective-type TIRFM. The presence of a relatively large prism above the sample
also hinders access and hampers sample manipulation in prism-type TIRFM.
2.2.3 Super-resolution microscopy
The resolving power of a microscope is the most important feature of that optical system and it influences the ability to distinguish between fine details within
specimen. As discussed in previous chapters, the fundamental limit of resolution
in light microscopy is imposed by diffraction. An instant glance on the equation 2,
defining the resolution limit, immediately suggests two obvious ways to decrease
it by reducing the wavelength or increasing the numerical aperture of the objective lens. Unfortunately, many problems arise with following these suggestions. The
light with wavelength below 350 nanometers is often phototoxic to live biological
specimens and is strongly absorbed by common optics used for visible light microscopy. Increase of the numerical aperture of the objective lens is presently limited to
the largest technical achievable semi-aperture angle of about 70°, which in combination of immersion oil with high refractive index (n=1.52) gives highest NA values
of about 1.5.
Many different sophisticated techniques were developed to circumvent the diffraction limit in light microscopy. If there is an interest only in surface-bound studies, techniques based on TIR (previous section) significantly reduce axial resolution.
Reducing lateral resolution in surface-bound studies is well achieved in the scanning
near field optical microscope (SNOM) [Synge 1928; Hecht et al. 2000]. It uses a
small sub-diffraction aperture or small tip to illuminate the object, which is scanned
over the sample. Although resolution down to 10 nm is achievable [Hosaka and
Saiki 2006] with SNOM, it is ultimately a surface-bound technique unfeasible for
29
Chapter 2. Fluorescence
most of biological applications, except of membrane studies.
Many ideas have been worked out to address the resolution problem in the farfield as well [Hell 2009b]. In 1956, Toraldo di Francia suggested shrinking the
central focal spot by applying an elaborate phase pattern in the entrance pupil of
the objective lens [Di Francia 1952]. However, the creation of smaller central spot is
accompanied with giant sidelobes making this concept impractical. A redistribution
of the focal energy will always suffer from such side lobe effects.
As early as in 1966 ideas of using different gratings [Lukosz 1966] for illumination and/or the detection pathway for improving the resolution in reflection imaging were expressed. Later, many techniques were developed based on this principle, like laterally modulated excitation microscopy [Heintzmann and Cremer
1999], structured illumination microscopy [Gustafsson 2000], objective-launched
standing-wave total internal reflection fluorescence microscopy [Chung et al. 2006].
Nonlinear relationship between the illumination intensity and the excitation probability in structural illumination were taken advantage of in saturated patterned
excitation microscopy (SPEM) [Heintzmann et al. 2002] and saturated structuredillumination microscopy (SSIM) [Gustafsson 2005].
In addition, some techniques take to the extreme both the illumination aperture
and the detection aperture, like 4Pi microscopy [Hell 1990; Hell and Stelzer 1992]
and I5M microscopy [Gustafsson et al. 1995]. As the name 4Pi suggests (a full
sphere has a solid angle of 4π), the idea is to get as close as possible to illumination
(and detection) from all sides of the sample. This is achieved by combining two
objective lenses opposed to each other along the optical axis direction. I5M microscopy uses similar method combined with incoherent illumination. However, these
techniques significantly improve resolution only in axial direction.
Furthermore, several techniques are able to produce super-resolution images
without breaking the diffraction limit per se – using localization of emitters. The
resolution problem arises when there are several fluorescent molecules in close proximity to each other. If there is only one emitter, it is possible to localize it with very
high accuracy by fitting the obtained fluorescence PSF with an appropriate analytical function and determining the center of it [Thompson et al. 2002]. If the fluorescence from a single molecule is distributed in a Gaussian profile and the background
noise is small compared to the molecular signal, the error in the fitted position is
inversely proportional to the number of detected photons from that molecule. To
decrease the error of localization by two, one needs to detect 4 times more photons
– which inevitably relates to the image acquisition time and limits performance of
such techniques when it comes to imaging of dynamical processes. For localization
30
Andriy Chmyrov
techniques it is therefore crucial to avoid the presence of several emitters inside of
one PSF. This is done differently in various realizations. In photoactivation light microscopy (PALM) [Betzig et al. 2006] this is done by sparse activation of fluorescent
proteins and subsequent intentional bleaching in the end of localization. Stochastic
optical reconstruction microscopy (STORM) [Rust et al. 2006] exploits the unique
property of switching of the cyanine dye, Cy5, between a fluorescent and a dark
state in a controlled and reversible manner by light of different wavelengths [Bates
et al. 2005]. Fluorescence photoactivation localization microscopy (FPALM) [Hess
et al. 2006] relies on the low probability of simultaneous excitation of several molecules under very low excitation irradiances. Photoactivation localization microscopy
with independently running acquisition (PALMIRA) [Geisler et al. 2007] benefits
from the un-synchronization of excitation pulses with the detection read out, which
decreases the possibility of detection fluorescence from several close emitters under sparse excitation. In points accumulation for imaging in nanoscale topography (PAINT) [Sharonov and Hochstrasser 2006] fluorescence is generated by the
binding of fluorescent probes to the structure to be imaged, and switches off by
free diffusion or another dark (bleached) state. Ground-state depletion and singlemolecule return microscopy (GSDIM) [Fölling et al. 2008] stochastically shelves
fluorophores into the triplet state and uses unsynchronized camera as a read-out
mechanism. As a general feature, localization techniques typically produce very nice,
sharp, background-free images (because the background is deliberately rejected if it
does not pass predefined threshold), but suffer from long acquisition times (from
hours to several minutes per image) and sample drifts during imaging.
The most successful super-resolution method so far is stimulated emission depletion (STED) microscopy [Hell and Wichmann 1994]. In STED the fluorescence,
which is created by a focused beam of excitation light, is reduced in space by simultaneously applying a second spot of light shaped into a doughnut that features
a central zero, for molecular de-excitation (see figure 2.11). The role of the deexcitation (STED) beam is to effectively confine molecules to the ground state, thus,
effectively switching off the ability of the fluorophore to fluoresce. Stimulated deexcitation occurs within the nanosecond lifetime of the fluorescent state. Because no
de-excitation occurs at the central zero, the excited state is allowed to spontaneously
decay only in the central region close to the zero. STED microscopy is technically
somewhat complex, demanding precise alignment of optical components for overlapping two beams with high accuracy and stability. Initially it required a system of
several complex and expensive laser sources, however with time variants of STED
microscopy with one pulsed supercontinuum fiber laser [Wildanger et al. 2008] or
continuous lasers [Willig et al. 2007; Moneron et al. 2010] have been developed.
31
Chapter 2. Fluorescence
Normalized PSF amplitude
1
500
0.8
0.6
I
Isat
0.4
0.2
0
-1
-0.5
0
0.5
radial coordinate
50
10
1
Figure 2.11: Effective molecule detection efficiency (MDE) function in STED/GSD techniqes. Black line represents Gaussian MDE without switching doughnut. Gray profiles represent MDE after illumination with a doughnut with local zero at radial coordinate 0 for peak
intensities I0 = 10 (light gray), 50 (gray), 500 (dark gray) times over saturation limit Isat, at
which 50% of the molecules are in a fluorescent state. Left axis represents intensity distribution within MDE. Right axis represents magnitude of switching intensity.
Combination of STED and 4Pi microscopy [Dyba and Hell 2002] allows simultaneous increase of resolution in all dimensions, and viability of two-photon excitation in STED microscopy was demonstrated [Moneron and Hell 2009]. Simultaneous dual color imaging with STED microscopy was first implemented [Donnert et
al. 2007; Meyer et al. 2008] using complex laser system for excitation, and later significantly simplified by using supercontinuum laser source. An ultimate resolution
of 5.8 nm have been demonstrated [Rittweger et al. 2009] by imaging fluorescent
nitrogen vacancies in diamonds. Besides high demands regarding the experimental
setup and the excitation sources (which tends to be soften upon recent technological
developments), STED microscopy requires high irradiances for stimulated emission
– on the order of GW/cm2 (given in part from the short fluorescence lifetime and the
low stimulated emission cross-section). Decreasing either the fluorescence lifetime
or utilization of long lived non-fluorescent states will decrease irradiances needed for
de-excitation of fluorophores outside the central zero point.
To circumvent the necessity of high irradiances for de-excitation, ground state
depletion (GSD) microscopy [Hell and Kroug 1995] was suggested. GSD microscopy is somewhat similar to STED microscopy, although it uses an entirely different mechanism for switching off fluorescence. To allow for much lower intensities,
the fluorophore is switched off by transiently shelving it into the metastable triplet
state, T1 (see figure 2.2). Due to the long triplet lifetime (about 3 to 6 orders of
magnitude longer than fluorescence lifetime), de-excitation irradiances are reduced
by the same factor. Other suitable non-fluorescent state for GSD microscopy is
32
Andriy Chmyrov
photo-isomerised state in cyanine dyes or in photoactivatable fluorescent proteins
[Hell et al. 2003]. Experiments with the asFP595 protein demonstrated [Hofmann
et al. 2005] sub-diffraction resolution imaging with switching intensities as low as
10W/cm2, which are by 6 orders of magnitude lower than those required for STED.
Various variants of utilizing reversible saturable optical (fluorescence) transitions
were generalized under term RESOLFT microscopy [Hell et al. 2004].
2.3 Fluorescence correlation spectroscopy
Fluorescence correlation spectroscopy (FCS) is a fluctuation spectroscopy technique, which makes use of temporal fluctuations in the detected fluorescence signal
under equilibrium conditions in order to obtain information about the processes
that give rise to these fluctuations. Statistical analysis is performed on fluctuations
in the form of a correlation function, which provides information about the characteristic times and the relative weights of different processes which give rise to
fluctuations.
FCS is closely related to other optical fluctuation method – dynamic light scattering (DLS), also known as quasi-elastic light scattering. DLS characterizes molecular motion in terms of optical interference effects from measurements of scattered coherent laser light [Berne and Pecora 1976]. It is an excellent method for
studying molecular transport in highly resolved and fairly concentrated systems, or
in mixtures where one species dominates the scattering. However, it is somewhat
limited to diffusion studies and relatively high concentrations, and is not useful for
measuring the progress of chemical reactions or following the dynamics of specific
molecule at low concentration given the presence of other molecular species at high
concentrations.
In contrast to the other popular fluorescence technique for studying diffusion
and binding kinetics – fluorescence recovery after photobleaching (FRAP) [Axelrod
et al. 1976], FCS is not relying on external perturbation of the system under study,
but instead it harnesses the deviations from the equilibrium. Macro-equilibrium
states often appear to be very dynamic on the molecular level. For example, when
studying diffusion, Brownian motion of the molecules will cause deviations in number of molecules inside the observation volume and smaller the observation volume
– larger these deviations will be. By labeling molecules of interest with a fluorophore,
fluctuations in number of molecules will be translated to fluctuations in emitted
fluorescence signal. Proper detection and analysis of these fluctuations will provide
information about the diffusion properties of the molecule of interest. In fact, all
of the processes which effect fluorescence emission are feasible for monitoring and
33
Chapter 2. Fluorescence
investigation with FCS, for example – translational diffusion [Magde et al. 1972],
diffusion coupled ligand-receptor interactions [Elson and Magde 1974], rotational
motion [Ehrenberg and Rigler 1974], uniform translation and laminar flow [Magde
et al. 1978], singlet-triplet transitions dynamic [Widengren et al. 1994], cis-trans
isomerisation dynamic [Widengren and Schwille 2000], FRET [Widengren et al.
2001], antibunching [Mets et al. 1997], charge transfer [Widengren et al. 1997],
oxidation-reduction [Widengren et al. 2007], just to name a few.
FCS was developed in early 1970s and the first published application was diffusion and binding studies of DNA-ethidium bromide interaction [Magde et al.
1972]. However, the technical difficulties imposed by immaturity of equipment
(excitation sources, fluorescence filters, detectors and electronics) resulted in large
excitation/detection volumes containing many molecules, low relative fluctuations
of a signal which needed to be compensated by very long measurement times (up
to 24 hours) [Magde et al. 1974]. Due to the mentioned shortcomings the technique did not really get exploited up to 1990s, when Rudolf Rigler and co-workers
first used a confocal microscope arrangement combined with FCS to detect single
molecules [Rigler and Widengren 1990]. By reducing the focal volume of excitation
and applying the confocal principle, background is reduced considerably, number
of molecules inside detection volume is reduced, generating larger relative fluctuations, collected and analyzed by more high-performance detectors and electronics.
Besides its confocal implementation, FCS have been combined with many other
modes of excitation. Many experimental variations evolved, like scanning FCS [Petersen 1986], two-photon excitation FCS [Berland et al. 1995], two-color cross
correlation [Schwille et al. 1997], TIR-FCS [Thompson et al. 1981], STED-FCS
[Kastrup et al. 2005], sub-wavelength sized apertures FCS [Leutenegger et al. 2006],
two-focus FCS [Dertinger et al. 2007], inverse FCS [Wennmalm et al. 2009] and
others.
2.3.1 Confocal FCS
A typical confocal FCS setup (figure 2.12) consists of a confocal microscope, in
which fluorescence is directed to a highly sensitive photodetector with high temporal resolution. A laser beam is expanded by a telescope with a pair of lenses, reflected
by a dicroic mirror and directed into a high numerical aperture microscope objective. Different degrees of laser beam expansion are used in order to achieve different sizes of the excitation volume. The objective then focuses the laser beam into a
sample, where fluorophores are excited. The emitted fluorescence is partly collected
by the same objective, transmitted through a dicroic mirror and focused onto a
34
Andriy Chmyrov
focal plane
objective
lens
excitation
filter
excitation source
(laser)
2.0
1.8
<T>
dicroic mirror
τT
tube
lens
correlator
pinhole
G(τ)
1.6
1/<N>
τD
1.4
beamsplitter
cube
APD
1.2
1.0
1E-5
1E-4
1E-3
0.01
0.1
correlation time, ms
1
10
emission
filters
APD
Figure 2.12: Sketch of a typical confocal FCS setup. Characteristic FCS curve exhibiting
correlation component due to diffusion and relaxation term for singlet-triplet transitions included in the lower left corner.
pinhole. The size of the observation volume is defined by the size of the confocal
pinhole and the excitation volume, which can be as low as few femtoliters if the laser
beam is focused to a diffraction limited spot. An important technical constrain is to
match the size of the pinhole and the excitation volume in order to achieve desirable (Gaussian) properties of the detection volume [Hess and Webb 2002]. After
the pinhole, fluorescence is directed to a detector, normally avalanche photodiode
(APD) or photomultiplier tube (PMT). Due to a presence of the undesirable effects of afterpulsing (that is generating a false count after detecting a photon) and
dead-time (inability to detect a second photon immediately after detecting a first
photon) in detectors, typically two detectors are used in FCS. Fluorescence after
the pinhole is re-collimated with a lens, split by a beam splitter and focused onto
the sensitive area of each detector. In order to minimize reflected and Rayleigh and
Raman scattered laser light, thin film interference filters are placed in the beam path
35
Chapter 2. Fluorescence
before detectors. The electrical signal from the detectors is analyzed by a specialized
hardware correlator or recorded by a data collection card.
Fluorescence intensity (number of photons) collected from the detection volume
at certain time t will be proportional to the molecular detection efficiency (MDE)


function E (r ) and the concentration of the emitting molecules C (r , t ) :



I (t ) = q ∫ E (r ) QC (r , t ) d 3r
(13)
where the brightness factor Q depends on the absorption cross section, the fluorescence quantum yield and the fluorescence lifetime and q is a proportionality constant accounting for the overall instrument detection efficiency.
Here an ideal solution of chemical compounds consisting of only one fluo
rescent species is considered. Its local concentration is C (r , t ) , the space-and
time ensemble averaged concentration C = C (r , t ) and the local deviation


dC (r , t ) = C (r , t ) − C . In equation (13) photon shot noise is not included, since
it is not correlated for different emitters and does not contribute to correlation function G(t).
The deviation of the detected photon counts from the mean I = I (t ) is:



d I (t ) = I (t ) − I = q ∫ E (r ) Q dC (r , t ) d 3r
(14)
In FCS experiments, the fluorescence correlation function G (t ) is defined as
time averaged product of the intensity I(t) at time t multiplied by intensity I(t+t) at
time t+t, normalized by the square of average intensity
I (t ) ⋅ I (t + t )
G (t ) =
I (t )
(15)
2
This expression can be rewritten as
I (t ) ⋅ I (t + τ )
I (t )
36
2
=
(I
+ δ I (t )) ⋅ ( I + δ I (t + τ ))
I
2
=1+
δ I (t ) ⋅ δ I (t + τ )
I
2
Andriy Chmyrov
Due to ergodicity of the systems studied by FCS, it’s possible to replace the time
average with the ensemble average:
G (τ ) = 1 +
δ I (0) ⋅ δ I (τ )
I
2
(16)
By substituting equation (14) into equation (16) we obtain
G (τ ) = 1 +
q2
I

2

∫∫ E (r ) E (r ′)Q
2


 
δC (r , 0) δC (r ′, τ ) d 3r d 3r ′
(17)
The observation volume is mathematically represented by the MDE function

E (r ) , which is proportional to the intensity detected from a single emitter as a

function of its position r in the sample volume. It is calculated by multiplying
the excitation intensity Iexc by the collection efficiency function (CEF) [Rigler et al.
1993].



E (r ) = I exc (r ) ⋅ CEF(r )
(18)
The CEF describes the probability of detection a photon as a function of the
position of the emitter [Qian and Elson 1991], and may be expressed as:
1


  
CEF(r ) = ∫ circ (r ′ s0 ) PSF(r ′, r ) d r ′
∆
(19)
where PSF denotes the point spread function of the confocal microscope (Airy pattern) and ∆ is a normalization function. Integration is carried out in the sample
space. The disk function, circ, represents the transfer function of the pinhole projected into the sample space, and the pinhole projection, s0, is the radius of the
confocal pinhole divided by the magnification of the detection system.
In order to simplify analytical expression, the MDE function for confocal detection is frequently approximated by 3-dimensional Gauss function:
 x2 + y2
z 2 

E (r ) = E (0) ⋅ exp −2
−
2


w xy2
w z2 
(20)
37
Chapter 2. Fluorescence
where wxy and wz are the lateral and axial radii of the detection volume. This approximation is rather accurate under most typical experimental conditions and it
produces mathematically simple analytical expression for the correlation function.
Another even more accurate approximation assumes a Gauss-Lorentzian [Marrocco
2007; Blom and Björk 2009] volume, however this analytical expression it is not so
mathematically straightforward in comparison to simpler 3D-Gauss approximation.
In order to obtain a simpler analytical expression for the correlation function
in equation (17), the concentration fluctuation function have to be obtained by
solving the differential equations that govern the molecular dynamics – either diffusion or reaction-diffusion equation (the later for the case of binding, singlet-triplet
transitions, isomerisation or others). Here, an expression of the correlation function
for only pure diffusion will be derived. The derivation of the correlation function for
reaction-diffusion systems is well described in [Elson and Magde 1974].
The diffusion equation is a partial differential equation which is usually written
as:

∂C (r , t )

= D∇2C (r , t )
∂t
(21)
where D is the diffusion coefficient of a molecule in aqueous solution, defined according to Stokes-Einstein relationship:
D=
kBT
6πη Rh
(22)
with h representing the viscosity of the solution, kB is Boltzmann’s constant, T is the
absolute temperature and Rh is the hydrodynamic radius, which could be used to
estimate the molecular weight of a particle. For spherical particles with molecular
weight M more than 1000 Da, Rh µ M 1/3 and consequently D µ M -1/3.
The diffusion equation (21) can be easily solved by applying a Fourier transform,
which yields:
dC (q , t )
dt
38
= −q 2 DC (q , t )
(23)
Andriy Chmyrov

−3 2
−i q r
where C (q , t ) = (2p) ∫ e C (r , t ) d 3r is a Fourier transform of C (r , t ) .
The solution of equation (23) is:
C (q , t ) = C (q , 0) exp (−Dq 2t )
(24)
In the following derivations, the condition of ideality of the chemical solution
will be used: the correlation length is much smaller than the distances between the
molecules. The positions of different molecules of the same species are therefore
uncorrelated, which mathematically may be written as:


 
dC (r , 0) ⋅ dC (r ′′, 0) = C d (r − r ′′)
(25)
where C is the mean square fluctuation of the concentration of the molecules


C (r , t ) in a unit volume, for Poisson statistics being equal to its average C (r , t )
Using equations (24), (25) and taking into account the fact that the Fourier
transform and the ensemble averaging are independent linear operators and thus
can be applied in any order, we can evaluate
−3 2
′
δC (r , 0) δC (r ′, τ ) = (2π) ∫ e iqr δC (r , 0) δC (q , τ ) d 3 q =
−3 2
= (2 π )
−3 2
= (2 π )
∫
∫
′
e iqr exp (−Dq 2 τ ) δC i (r , 0) δCk (q , 0) d 3 q =






−3 2
′
′′
e iqr exp (−Dq 2 τ ) d 3 q ×(2π) ∫ e −iqr δC (r , 0) δC (r ′′,0) d 3r ′′ =
−3
= (2 π )
C
∫
e
  
−iq (r −r ′)

exp (−Dq 2 τ ) d 3 q
(26)
Substitution of equation (26) into equation (17) gives
G (τ ) = 1 +
q2
I
×C
−3
2
(2 π )
∫e
∫∫ d
  
−iq (r −r ′)
3
  

rd 3r ′ E (r ) E (r ′)Q 2 ×

exp (−Dq t ) d q
2
(27)
3


By changing the order of integration – first on r and r ¢ and using the fact that
39
Chapter 2. Fluorescence
Fourier transform of an even function is even, we can write
G (t ) = 1 +
q2
I
2
∫ E (q )
2
Q 2 C exp (−Dq 2 t )d 3 q
(28)
(q ) = (2p)−3 2 e −iqr E (r ) d 3r is the Fourier transform of E (r ) .
where E
∫
(0) = (2p)−3 2 E (r ) d 3r , then obviously E (r ) d 3r = (2p)3 2 E
(0)
Note that E
∫
∫
The average fluorescence intensity áIñ in equation (28) is determined from equation (13):
32
(0) qQ C
I = q ∫ E (r ) Q C d 3r = (2p) E
(29)
From equation (20), the Fourier transform of MDE is
2

 w2
w2 w
(q ) = E (0) xy z exp − xy (q x2 + q 2y ) − w z qz2 
E

 8
8
8 
 (0) = E (0)
E
w xy2 w z
(30)
(31)
8
Combining equations (29), (30) and (31) into equation (28) gives:
−3
G (τ ) = 1 +
(2 π )
C
∫
 
 ω xy2
ω z2 2
2
2
exp −
q
+
q
−
qz − Dq 2 τ  d 3 q
(
x
y)

 4
4
(32)
This expression can be easily integrated, which results in:
−1


1
1 + 4 D τ 
G (τ ) = 1 + 3 2 2

ω xy2 
π ω xy ω z C 
−1 2


1 + 4 D τ 
2 

ω 
(33)
z
In order to have the ability of measuring concentrations, the size of observation
volume needs to be defined. This size is normally defined by the effective volume Veff
40
Andriy Chmyrov
[Thompson 1992]:
Veff º
where Wn ≡
1
E
n
E
(0) ∫
n


W12
W2
(34)
(r ) d 3r
Using equation (20) we get Veff = π 3/ 2 ω xy2 ω z . Defining the average number of
molecules within observation volume as N = Veff C , the characteristic diffusion
time across the detection region as τ D = ω xy2 4D and the axial to radial aspect ratio
of the sampling volume as w = w z w xy , the following expression is deduced:
−1
G (τ ) =
τ 
1 
1 + 
N
τ D 

1 + τ

ω 2τ
−1 2



D
+1 =
1
G D (τ ) + 1
N
(35)
Due to the presence of singlet-triplet transitions in almost all fluorophores and
the fact that triplet state is not fluorescent, these transitions will have an influence
on the correlation curve and need to be taken into account. Singlet-triplet transitions can be treated as a unimolecular reaction, and such reaction-diffusion systems
are described by modifying the diffusion equation (21) into:

m
∂C l (r , t )


= Dl ∇2C l (r , t ) + ∑ K lkC k (r , t )
∂t
k =1
(36)
where the first term accounts for diffusion and the second term describes

chemical changes. C l (r , t ) is the fractional concentration of the molecules in state
Sl (singlet S0, S1 or triplet T1). Coefficients Klk are combined from the chemical rate
constants and the equilibrium concentrations of the species.
Equation (36) can be solved by taking Fourier transform in the similar fashion
as equation (21), taking into account modified variant of equation (25) for the case
of several species, the positions of which are uncorrelated is space:


 
dC i (r , 0) ⋅ dC k (r ′′, 0) = C i dik d (r − r ′′)
(37)
The full derivation of the correlation function for chemical reactions is present41
Chapter 2. Fluorescence
ed in [Elson and Magde 1974], for singlet-triplet transitions in [Widengren et al.
1995] and for trans-cis isomerisation in [Widengren and Schwille 2000].
In the general case of chemical reactions, brightnesses of the states should be
considered [Thompson 1992], but the triplet state is assumed to be totally dark,
and FCS function describing combining Brownian diffusion and triplet kinetics
thus becomes:
G (t ) =
1
GD (t ) 1 −T +T exp (− t t T ) + 1


N (1 −T )
(38)
where T is the space and time averaged fraction of the molecules inside the
observation volume being in the triplet state and tT is the relaxation time of singlettriplet transitions, defined by the rates from figure 2.2 (if only excitations to the first
excited states are considered) [Widengren et al. 1995]:
T=
k01kISC
k01 (kISC + kT ) + k10 kT

k k
t T = kT + 01 ISC
k +k

01
(39)
−1



10 
(40)
2.3.2 TIR-FCS
Total internal reflection fluorescence correlation spectroscopy (TIR-FCS) was
introduced in 1981 [Thompson and Axelrod 1981; Thompson et al. 1981]. Before the implementation of the confocal principle to FCS, evanescent field was the
natural way of confining the observation volume. Initial experimental realizations
were done with the adaptation of prism-type TIRFM and were used for absorption
studies of Rhodamine labeled immunoglobulin on quartz [Thompson and Axelrod 1981]. Although the theoretical framework was established for TIR-FCS, the
technique was rarely used in the following years, presumably due to the large background and the low detection efficiency, implying small signal-to-noise ratio and
poor reliability of the measurements.
Introduction of modern objective based TIR-FCS [Hassler et al. 2005a; Hassler
et al. 2005b] increased significantly the countrates per molecule and signal-to-noise
ratio, facilitating the applicability of the technique. A principal sketch of a typical
objective-based TIR-FCS setup is presented in figure 2.13.
42
Andriy Chmyrov
interface
objective
lens
excitation
filter
excitation source
(laser)
dicroic mirror
tube
lens
correlator
pinhole
APD
beamsplitter
cube
emission
filters
APD
Figure 2.13: Sketch of a typical objective-based TIR-FCS setup. Compare to a confocal FCS
setup, the main difference is in the presence of the third lens in the laser path before the dicroic
mirror. This lens focuses the beam off-axis onto the back focal plane of the objective lens. The
beam emerges collimated from the objective, totally reflects from the interface, returns off-axis
into objective and is finally blocked. The focusing lens and the dicroic mirror are moved in one
block by a linear translator with micrometer screws to adjust the lateral position of the laser
beam entering the objective. In this way, the excitation angle can be adjusted without altering
the optical path length between the focusing lens and the objective.
Derivation of the correlation function for TIR excitation is done analogously
to the confocal case, the only difference being the MDE function. In the following
derivation of the correlation function for Brownian diffusion, the results from the
previous section will be used. MDE function for TIR detection volume could be
approximated by:
 x 2 + y 2 
 

 ⋅ exp − z 
E (r ) = E (0) ⋅ exp −2
2
 h 


w xy 
(41)
43
Chapter 2. Fluorescence
where wxy is the lateral radius of the excitation volume and h is the detected penetration depth of the evanescent field.
Since MDE is technically defined only in the positive half-space (z >0), for the
purpose of the analytical derivation of Fourier-transform equation (41) will be complemented, for negative half-space, by introducing modulus of axial coordinate z.
 x 2 + y 2 
 z

 ⋅ exp − 
E * (r ) = E (0) ⋅ exp −2
2
 h 


w xy 
(42)
According to equation (28), we need to calculate the Fourier transform of MDE:

 ω xy2
ω xy2
h
−
2
2 

0
E * (q ) = E ( ) ⋅
⋅
⋅
exp
q
+
q
(
)
x
y
 8
2 2π (1 + h 2 qz2 )


 * (0) = E (0) ⋅
E
ω xy2
2 2π
⋅h
(43)
(44)
Combining equations (29), (43) and (44) into equation (28) finally gives:
−3
G (τ ) = 1 +
(2 π )
C
−2
2 2
∫ (1 + h qz )
 
 ω xy2
⋅ exp −
q x2 + q 2y ) − Dq 2 τ  d 3 q
(

 4
(45)
This expression can be easily integrated, resulting in:
−1


1
1 + 4 D τ 
G (τ ) = 1 +

4πω xy2 h C 
ω xy2 
 2 Dτ  2 D τ   D τ 


 

 h π + 1 − h 2  w i h  (46)



where w is the complex error function, also known as the Faddeeva function:
w ( x ) = exp (−x 2 ) erfc (−ix )
(47)
and erfc is the complimentary error function. By making substitutions such as
44
Andriy Chmyrov
structure parameter w = h w xy , axial diffusion time t z = h 2 4 D , and calculating
effective volume from equations (42) and (34) to be Veff = 4πω xy2 h , we can rewrite
equation for the correlation function for diffusion in TIR:
−1
ω 2 τ 
1

G (τ ) = 1 +
N
τ z 
 τ

 

1 − τ  w i τ
+
 πτ
 2τ   4τ

z
z
z

 +1

(48)
Taking singlet-triplet transitions into account by substituting the diffusion equation (21) with the reaction-diffusion equation (36), one can derive the autocorrelation function for diffusion+triplet kinetics (full derivation presented in [Hassler
2006]):



1 + T exp − τ  ×
 1 −T
 τ 

T 
−1
2

 w τ   τ

τ  
τ
 
 w i
×1 +
+ 1 −

τ z   πτ z  2τ z   4τ z
G (τ ) = 1 +
1
N



(49)
2.3.3 TRAST spectroscopy
Transient state (TRAST) spectroscopy is a recently developed technique [Sandén
et al. 2007; Sandén et al. 2008] which uses temporal modulation of the excitation
in order to obtain information about photoinduced dark transient states, like the
triplet state. It relies on the fact that fluorescence emission, after excitation, is not a
stochastic but rather a deterministic process (on the timescale of µs), governed by
the kinetic rates of transitions between the electronic states and well described by
a system of differential equations (from figure 2.2, considering only ground singlet
and first excited singlet and triplet states):

S0 (r, t ) −k01 (r, t )
k10
kT   S0 (r , t )






∂    

  S (r, t )
,
−
+
0
r
,
t
k
k
S
r
t
k
=
(
)
(
)
(
)
1
0
1
10
1
ISC
  
∂t    
0
−kT  T1 (r , t )
kISC
T1 (r , t ) 
(50)
In the differential equation system, the excitation rate k01 is given by


k01 (r , t ) = s01 I exc (r , t ) , where s01 is the excitation cross-section (calculated from


the molar extinction coefficient) and k01 (r , t ) = s01 I exc (r , t ) denotes the excitation
45
Chapter 2. Fluorescence


irradiance at a location r within the detection volume. S0 (r , t ) is the probability


for the molecule to reside in the ground singlet state, S1 (r , t ) and T1 (r , t ) – in the
first excited singlet and triplet state, respectively. Assuming that, the fluorophore
concentration can be written as:





C (r , t ) = C (r , t ) S0 (r , t ) + S1 (r , t ) +T1 (r , t )
(51)



so that S0 (r , t ) + S1 (r , t ) +T1 (r , t ) = 1 .
The initial conditions to equation (50) originate from the fact that before the
excitation at time t = 0 all of the fluorophores are regarded to be in ground singlet
state:
S0 (r, 0) 1
    
 S (r , 0) = 0
 1    
T (r , 0) 0
 1
  
(52)
In order to obtain a simpler expression for the solution, one can assume that the
decay of the singlet state by fluorescence or internal conversion is much faster than
either of the processes of intersystem crossing or triplet state decay:
k10  kISC , kT
(53)
Applying the boundary conditions (52) to differential equation (50), taking into
account simplification (53) and assuming the excitation rate to be stationary in time,
the probabilities of occupying the different electronic states for a fluorophore in

space r as a function of time [Widengren et al. 1995] become:

S0 (r , t ) =

(r )
k k
e l1 t +
+ k ) + k10 kT
10 T

(r )
01
ISC
T

(r )

l2 (r )t
01
k (k
k
e

(r )
k01 + k10

S1 (r , t ) =

+

(r )
01
(k

(r )
k01
k10 kISC

(r )
01
+ k10 ) k (kISC + kT ) + k10 kT 



(r )
e l3
t

(r )
k01(r )kT
e l1 t +

(r )
k01 (kISC + kT ) + k10 kT
(54)

(r )


(r )
(k01(r ) ) kISC
k01
l3 t
l2 (r )t
e
+
e



r
( )
+ k10
k01
(k01(r ) + k10 ) k01(r ) (kISC + kT ) + k10kT 


(r )


(r )
(r )
k01(r )kISC
k01
kISC

l1 t
T1 (r , t ) = (r )
e − (r )
e l3 t
k01 (kISC + kT ) + k10 kT
k01 (kISC + kT ) + k10 kT

46
2
Andriy Chmyrov
1
S0
τAB
τT
B
0.5
0
0.06
S1 0.03
0
1
T1
80
Fluorescence Intensity
A
Fexc,1
60
Fexc,2
Fexc,cw
F2
40
20
w2
0.5
0
10-9
10-8
10-7
10-6 10-5
t (s)
10-4
0
10-3
w1
F1
T
2
4
6
Time (µs)
Figure 2.14: A. Relative populations of electronic states achieved with rates k01 = 7.9 µs-1, k10 =
260 µs-1, kISC = 8.0 µs-1, and kT = 0.5 µs-1. B. Fluorescence intensity for pulsed excitation with
two different pulse widths w1 and w2. Fexc1 and Fexc2 are average fluorescence during the pulse,
F1 and F2 are average fluorescence during the measurement, Fexc,cw is corresponding average
fluorescence for non-modulated measurement. Adapted from [Sandén et al. 2007]
with the eigenvalues related to the relaxation modes of the population kinetics:

l1(r ) = 0


l2(r ) = −(k01(r ) + k10 )

(r )
3
l


k (r )k 
= − kT + (r01) ISC 
k01 + k10 

(55)
In (55), the first eigenvalue l1 is zero indicating that the population will approach steady-state as t®∞. The second eigenvalue l2 represents the so called antibunching term (tAB = l2) [Mets et al. 1997], reflecting the time it takes to establish
equilibrium between singlet states. The third eigenvalue l3 is related to the rate at
which buildup of the triplet state population takes place. Visualization of the dynamics of the states defined by equation (54) is presented in figure 2.14A.
In the results of the simulations presented in figure 2.14, it is important to note
that population of first excited state S1 after onset of the excitation initially increases,
and then decreases due to triplet state build up. Since state S1 is the state from which
fluorescence is generated, it can be expected that fluorescence intensity immediately after the excitation will be higher than during the steady-state equilibrium.
However, to detect this extra-fluorescence, high temporal resolution is needed. To
circumvent this problem, pulsed excitation is applied and the excitation is switched
off after certain time – pulse width w. Obviously, for the very short pulses with
47
Chapter 2. Fluorescence
t AB < w < tT , before the triplet state has built up, the average intensity during the
excitation Fexc1 will be higher than the average intensity Fexc2 for the long pulses. If
the time between the excitations in the pulses is short, the molecules in the triplet
state will not yet relax to the ground state (due to finite relaxation time), contributing to the difference in average fluorescence for different pulses (see figure 2.14B).
Since the excitation of all of the molecules inside the observation volume is synchronized, by reading out average intensity for modulated measurement it is possible to
obtain information about transition rates between the states.
The initial implementation of TRAST was realized on a confocal FCS setup
using APDs with high temporal resolution as detectors. However high temporal
resolution is superfluous for this technique and it was demonstrated that detection
with ordinary of-the-shelf CCD camera works equally good for detection, decreasing instrumental demands and allowing parallel, large-scale detection [Sandén et
al. 2007]. TRAST is easily implemented on an ordinary laser scanning confocal
microscope, because scanning is equivalent to temporal modulation, which was successfully demonstrated in [Sandén et al. 2008].
Furthermore, TRAST have been applied for measuring the triplet state rates
of the fluorophores in liposomes [Sandén et al. 2008], for imaging oxygen consumptions during the contractions of a single smooth muscle cell [Geissbühler et
al. 2010] and for monitoring oxidation-reduction reactions on surfaces [Spielmann
et al. 2010].
48
Chapter 3
Photobleaching & additives
Photophysical and photochemical properties of fluorescent molecules are of great
importance for all applications of fluorescence spectroscopy where a high read-out
rate or a high sensitivity is required. In particular, this is important for applications
regarding single molecule detection (SMD) or fluorescence correlation spectroscopy
(FCS), where a low fluorescence signal can not be compensated by an increased fluorophore concentration. In this respect, there is a need to optimize the fluorescence,
both with respect to the total number of photons that can be detected per molecule,
as well as with respect to the fluorescence emission rate itself.
In paper I, various strategies are proposed for improving the fluorescence rate
and decreasing photobleaching of fluorophores. In particular, the effects of the addition of two main categories of anti-fading compounds (see figure 3.1): anti-oxidants (n-propyl gallate, nPG, and ascorbic acid, AA) and triplet state quenchers
(mercaptoethylamine, MEA, and cyclooctatetraene, COT) were investigated and
the relevant rate parameters involved were determined for Rhodamine 6G (Rh6G).
Addition of each of the compound categories yielded significant improvements in
the fluorescence brightness of the monitored fluorescent molecules in FCS measurements. For anti-oxidants, we identify the balance between reduction of photoionized fluorophores on the one hand, and that of intact fluorophores on the other,
as an important guideline for what concentrations to add for optimal fluorescence
generation in FCS and SMD experiments. For nPG/AA this optimal concentration
49
Chapter 3: Photobleaching & additives
A
n-Propyl gallate (nPG)
B
Ascorbic acid (AA)
O
HO
H
HO
O
O
O
HO
HO
HO
OH
OH
C
Mercaptoethylamine (MEA)
H 2N
D
Cyclooctatetraene (COT)
SH
Figure 3.1: Chemical structures of anti-fading compounds investigated. A. n-Propyl gallate. B.
Ascorbic acid. C. Mercaptoethylamine. D. Cyclooctatetraene.
was found to be in the lower micromolar range, which is considerably less than
what has previously been suggested [Longin et al. 1993]. Also, for MEA, a compound known as a triplet state quencher, it is in the end its anti-oxidative properties,
and the balance between reduction of fluorophore cation radicals and that of intact
fluorophores, that defines the optimal added concentration. Interestingly, in this
optimal concentration range, the triplet state quenching is still far from sufficient to
fully minimize the triplet populations.
3.1 Antioxidants n-propyl gallate and
ascorbic acid
nPG as an antioxidant in mM concentrations
According to previous reports on the photobleaching retarding effects of nPG,
relatively high nPG concentrations of around 10 mM should be necessary in order
to obtain a significant effect [Giloh and Sedat 1982; Longin et al. 1993; Gaigalas
et al. 2004]. However, from our FCS measurements nPG was found to strongly
quench the Rh6G molecules at these concentrations. To analyze the nature of the
quenching more closely a series of correlation curves was measured at varying nPG
concentrations and excitation irradiances (figure 3.2A). The measured autocorrelation curves showed a characteristic fluctuation process in the time range of a few
hundred nanoseconds or shorter, which very much resembled that obtained from
the electron transfer induced quenching of deoxy guanosine triphosphate (dGTP)
50
Andriy Chmyrov
4.0
Rh6G-nPG
SC1
3.5
kc01
G(τ)
3.0
2.5
kdiss1
S1
kass1
k01
B
Rh6G
kc10
kdiss0
k10
SC0
kass0
S0
kISC
kT
T1
2.0
1.5
1.0
rate (µs-1)
A
10 mM nPG
3 mM nPG
1 mM nPG
1E-5 1E-4 1E-3 0.01
25
kass1x[nPG]
20
kass0x[nPG]
kdiss0
15
10
5
0
0.1
1
10
correlation time (ms)
0
2
4
6
8
nPG concentration (mM)
10
Figure 3.2: A: FCS curves of R6G in aqueous solution containing 1, 3 and 10 mM of nPG.
Inset: Rate scheme describing the relaxation behavior of the two exponential processes of the
recorded correlation curves, involving singlet-triplet transitions and charge transfer following
hydrophobic complex formation (Rh6G-nPG). B: Extracted rates of association (kass0, kass1)
and dissociation (kdiss) of Rh6G and nPG.
in a previous study [Widengren et al. 1997].
The time-dependent part of the correlation curves was well fitted to the expression of equation (38) with an additional exponential process:
G (t ) =
G D (t )
1 −T − C +T exp (− t t ) + C exp (− t t ) + 1 (56)
T
C 
N (1 −T − C ) 
where GD(t) is defined in the same way as in equation (35).
Similar to dGTP, and because nPG is known as a reducing agent, this additional exponential process can be attributed to nPG acting as an electron donor
with charge-transfer taking place upon hydrophobic interaction between the nPG
and the Rh6G molecules. As for the charge transfer observed for Rh6G and dGTP
[Widengren et al. 1997], a rate scheme as outlined in the inset of figure 3.2A was
used to describe the observed exponential relaxations in the FCS curves. Except of
usual states S0, S1 and T1 (figure 2.2), this scheme includes the possibility of complex
formation between Rh6G and nPG and subsequent electron transfer from nPG to
Rh6G with different rates depending on wherever Rh6G is in its S0 or S1 state.
The rate scheme can be reduced to a three state model, consisting of a Rh6G singlet state, S (S0 and S1), the lowest triplet state, T1, and a complexed state with nPG,
C (SC0 and SC1, see inset figure 3.2A), by recognizing that the intra-singlet state
transitions (i.e., those between S0 and S1, and between SC0 and SC1, respectively) take
51
Chapter 3: Photobleaching & additives
place on a time scale much faster than any of the other dynamic processes involved:
′
kdiss 0
kISC


→S ←


→T1
C←


k
kass′
T
(57)
where kass¢ and kISC¢ denote the effective rates of complex formation between Rh6G
and nPG and of intersystem crossing within the Rh6G molecules, respectively. Due
to very fast deactivation of the excited state of Rh6G-nPG complex, dissociation
from this state (rate kdiss1) can be disregarded in equation (57). The measured amplitudes (T and C) and relaxation times (tT and tC) from the correlation curves are
given from the corresponding to equation (57) system of coupled first order differential equations. This system was solved in a similar way as was done in [Widengren
et al. 1997]. The obtained rate constants of kass0, kass1 and kdiss0 are plotted versus the
concentration of nPG in figure 3.2B. Note the increase of the association rates kass0.
and kass1.
From the results presented in figure 3.2B, it can be concluded, that at millimolar
concentrations nPG does not show an overall beneficial effect on the Rh6G fluorescence due to its strong contribution to both static (kass0) as well as dynamic (kass1)
fluorescence quenching. Although the photobleaching measurements undertaken at
those concentrations indeed show a retardation of the photobleaching. Also under
our experimental conditions we observe no significant shortening of the overall decay times of the correlation curves. However, the fluorescence quenching is much
more pronounced. Addition of nPG in millimolar concentrations therefore does not
seem to provide a useful strategy to reduce photobleaching under FCS conditions.
nPG as an antioxidant in µM concentrations
In contrast, at concentration levels of nPG in the µM range, at which most of
the static and dynamic quenching effects observed at millimolar concentration are
no longer traceable, the retarding effect of nPG was still pronounced in the sense
that addition of increasing amounts of nPG increased the overall decay times of
the correlation curves. In particular, this effect could clearly be seen from the FCS
measurements with the expanded sample volume element, so that the passage time
of the Rh6G molecules are in the order of a few milliseconds (see sigure 3.3A).
These correlation curves recorded with the expanded sample volume were well
fitted to the expression used to describe photobleaching [Eggeling et al. 1998b]:
GB (t ) = G (t ) 1 − B + B exp (−kB t )
52
(58)
Andriy Chmyrov
2.2
n-Propyl Gallate
2.0
G(τ)
1.8
B
20 µM
2 µM
0.5 µM
0 µM
1.6
1.4
kox1(106 s-1)
0.1
koxn(109 s-1)
0.01
kred´(106 s-1)
1E-3
1.2
1.0
1
rate (µs-1)
A
0.01
0.1
1
10
correlation time (ms)
100
1E-4
0.1
1
10
n-PG concentration (µM)
Figure 3.3: A. FCS curves recorded from Rh6G and nPG in aqueous solution using an expanded sample volume. B. The parameters of kox1, kox2, and kred¢ obtained after fitting the
experimental parameters (B, kB) retrieved from each measurement for various concentration
of nPG.
where G(t) is defined as in equation (38)
With increasing concentrations of nPG, the amplitude B and was found to continuously decrease, while the relaxation rate kB was found to increase (figure 3.3A).
Taking into account the ability of nPG to act as an electron donor, as shown above,
this behavior is well in agreement with a bimolecular reaction involving the nPG
and the fluorophore molecules. In this reaction, the fluorescence emission of the
fluorophores can be “switched off” as a consequence of laser irradiation and photoionization. The photoionized Rh6G molecules can then be “switched on” again by
receiving an electron, following a collisional encounter with an nPG molecule. The
major features of this photoionization/reduction process are outlined in figure 3.4A.
From previous studies on photobleaching it was known that photobleaching
quantum yield FB deviates from linear dependence at high excitation irradiances
[Eggeling et al. 1998a]. This increase of FB was found to be taken well into account
by including excitation to higher excited singlet and triplet states from which a subsequent photobleaching reaction may occur. For this reason, the scheme outlined in
figure 3.4A was extended to take into account photoionization from higher excited
states as well. This modified 5-level model is presented in figure 3.4B.
Defining and solving the correspondent system of rate equations and fitting the
experimentally obtained values of B and kB from equation (58) measured at various
concentrations of nPG and various excitation irradiances, one can extract the corresponding oxidation (kox1 and koxn) and reduction (kred) rates, which are presented
in figure 3.3B. While kox1 and koxn remain close to constant and independent of the
nPG concentration, kred¢ (defined as kredx[nPG]) shows a linear dependence.
53
Chapter 3: Photobleaching & additives
A
B
F
Photo-oxidation
reduction
nPG+
e-
.
R+
Sn
Tn
kS1n
kSn1
kISC
S1
nPG
k01
S0
kT1n kTn1
k10
T1
koxn
kox1
.
R+
kT
kred
Figure 3.4: A. Major features of the photoionization/reduction process of R6G in the presence of nPG, viewed on a time scale >> the triplet relaxation time tT . F denotes the fluorescent, non-ionized fluorophore, R+ is the photoionized state that can be regenerated into F by
donation of an electron from nPG. B. Five-state model taking photoionization from both first
and higher excited singlet (S1 and Sn ) and triplet states (T1 and Tn ) into account (extension of
figure 2.2). kox1 and koxn denote the rate coefficients for photo-oxidation from the lower and
higher excited singlet/triplet states, respectively. kred denote the reduction rate of fluorophore
radicals.
The dependence of kox1, koxn, and kred¢ on nPG concentration (figure 3.3B) well
supports the hypothesized model of photo-oxidation and subsequent restoring reduction of the photo-oxidized fluorophores by nPG, taking place via collisional
interactions.
AA as an antioxidant in µM concentrations
To strengthen the support of the proposed model further, we also investigated in
the same manner the well-known antioxidant AA. As for the investigation involving
nPG, a series of correlation curves of Rh6G in water were recorded at various AA
concentrations and excitation intensities using an expanded volume element. The
results are very similar to those obtained after addition of nPG. As shown in figure 3.5, addition of increasing amounts of AA in the sub-millimolar concentration
range to the Rh6G solution increases the overall correlation decay time (compare
figures 3.5 and 3.3A).
From the effects of nPG and AA on the fluorescence of Rh6G, it seems as if the
concentration of antioxidants has to be balanced. The antioxidants then have to be
present at high enough concentrations that the effective rate of reduction can compete with the effective photo-oxidation rate. On the other hand, if present at too
high concentrations, the extent of reduction of fluorescent fluorophore molecules
by the antioxidants can get significant and can generate a strong static as well as
dynamic quenching.
54
Andriy Chmyrov
Ascorbic acid
2.0
G(τ)
1.8
1.6
80 µM
20 µM
8 µM
2 µM
0 µM
1.4
1.2
1.0
0.01
0.1
1
10
correlation time (ms)
100
Figure 3.5: FCS curves of Rh6G in aqueous solution with addition of various concentrations
of ascorbic acid (0-80 µM). Excitation irradiance is 0.7 MW/cm2.
3.2 Triplet state quenchers –
mercaptoethylamine and cyclooctatetraene
Adding MEA into an aqueous solution of Rh6G led to a clear reduction of the
triplet state contribution in the correlation curves (figure 3.6A). Moreover, in addition to the triplet state relaxation (~1 µs range), an additional relaxation process in
the ~10 µs range could be observed (figure 3.6A) with an amplitude that increased
A 2.0
B
MEA
12
1.8
rate (µs-1)
1.4
8 mM
4 mM
2 mM
0 mM
1.2
1E-4
1E-3
0.01
0.1
correlation time (ms)
kT
8
6
4
2
1
0
10
Figure 3.6: A. FCS curves of Rh6G in aqueous solution with addition of MEA at various
concentrations (0-8 mM). Excitation intensity
2.9 MW/cm2 for all correlation curves. B. The
MEA concentration dependence of kISC and kT,
extracted from FCS measurements and the tT
and T, measured at different excitation intensities (10 kW/cm2-3MW/cm2). C. FCS curves
of Rh6G in deoxygenated aqueous solution
with addition of 2 mM MEA at different excitation intensities.
C
2.2
0
2
4
6
8
10
MEA concentration (mM)
12
MEA 2mM, deoxygenated
2.0
1.8
G(τ)
G(τ)
1.6
1.0
kISC
10
1.6
1.4
1.2
1.0
390 kW/cm2
180 kW/cm2
70 kW/cm2
0 mM / 70 kW/cm2
1E-4 1E-3 0.01 0.1
1
10
correlation time (ms)
100
55
Chapter 3: Photobleaching & additives
with the applied excitation intensities. By fitting the correlation curves measured
at different excitation intensities to equation (56) and then fitting all the obtained
T and tT parameters to equations (39) and (40), respectively, the rates for intersystem crossing kISC and triplet state decay kT could be determined [Widengren et al.
1995]. The triplet state rate parameters determined for different concentrations of
MEA are plotted in the figure 3.6B. It can be seen that kT increases linearly with the
MEA concentration while kISC remains essentially unaffected. In addition, a strong
increase of the countrate per molecule (CPM) was observed upon addition of MEA.
However, by increasing the concentrations of MEA above 2-6 mM, no significant further increase in the fluorescence output could be noticed. We attribute this
to the process underlying the additional exponential relaxation process that could be
observed in the FCS curves upon addition of MEA (figure 3.6A). The amplitude of
this additional process was increasing for concentrations above 6 mM, with simultaneous decrease of fluorescence rate (CPM). Therefore, this can be contributed to
MEA-induced reduction of the Rh6G molecules.
In order to strengthen this assumption, measurements with MEA were performed under deoxygenated conditions. Molecular oxygen plays a double role for
fluorophores. It is an efficient triplet state quencher, and collisional triplet state
quenching by oxygen is the main contribution to kT under air saturated conditions [Korobov and Chibisov 1983]. On the other hand, this triplet quenching
process results in an excited singlet state of oxygen (so called singlet oxygen), which
is very reactive and contributes to the photodegradation of fluorophores [Singh et
al. 1992]. As can be seen from figure 3.6C – without oxygen a strong accumulation
of the fluorophores in their triplet states occurs. Upon addition of MEA the triplet
relaxation term is significantly reduced. However, the second relaxation process is
clearly seen, with much higher amplitude than under air-saturated conditions, at
similar excitation irradiances and MEA concentrations. This is well in agreement
with previous reports stating that the major decay pathway of the reduced (nonfluorescent) form of Rh6G is by collisional interactions with the dissolved molecular
oxygen. From this point of view, addition of MEA to deoxygenated solutions has a
stronger relative effect on the triplet state population. However, MEA also promotes
formation of reduced Rh6G, which in the absence of oxygen can be strongly accumulated. Therefore, in deoxygenated as well as in air-saturated solutions, there is a
tradeoff between triplet quenching on the one hand and fluorophore reduction on
the other upon addition of MEA.
Also, FCS measurements on Rh6G with various concentrations of COT were
performed in ethanol. The beneficial effect in terms of reduction the triplet state
56
Andriy Chmyrov
A 2.0
B
COT/EtOH
8
kT
7
1.8
kISC
6
1.4
1.2
1.0
rate (µs-1)
G(τ)
1.6
2.4 mM
0.5 mM
0 mM
1E-4
1E-3
0.01
0.1
correlation time (ms)
5
4
3
2
1
1
10
0
0.0
0.5
1.0
1.5
2.0
COT concentration (mM)
Figure 3.7: A. FCS curves of Rh6G in ethanol with addition of COT at various concentrations (0-2.4 mM). B. COT concentration dependence of kISC and kT, extracted from FCS
measurements at different excitation intensities.
fraction is visible in figure 3.7A. Performing the same analysis as for MEA, linear
dependence of the triplet deactivation rate kT versus COT concentration was observed, with triplet intersystem crossing rate kISC being unchanged. At COT concentrations higher than 3 mM the triplet quenching effect was diminished, probably
because of micelle formation of COT molecules. Below this limit, the addition of
COT enhanced the fluorescence output from the Rh6G molecules. These results
support the view of COT as a potent triplet state quencher of Rh6G, as put forth by
investigators since its use in dye lasers [Pappalardo et al. 1970].
3.3 Optimization of fluorescence output
A summary of the achieved fluorescence countrate per molecule (CPM) of
Rh6G is plotted versus excitation irradiance in figure 3.8A-B. In agreement with
previous findings, addition of the antioxidant nPG in micromolar concentrations
has a strongly beneficial effect on the recorded CPM. However, at increased concentrations the extent of electron donation and quenching of intact fluorophores by the
antioxidants becomes increasingly dominant and reduces the fluorescence brightness of the fluorophores. With MEA, higher countrates could be achieved than
with addition of AA or nPG, which reflects the double impact of MEA – as a triplet
quencher and as antioxidant at the same time. In figure 3.8C CPM values are plotted versus MEA concentration at the excitation intensity where maximum CPM
values were achieved. It is interesting to note that the optimum MEA concentration
is considerably lower than that required for a full triplet quenching (see figure 3.6A).
This indicates that for MEA, its antioxidative properties and the balance between
electron donation to photo-oxidized fluorophores (promoting fluorescence) and the
57
Chapter 3: Photobleaching & additives
B
10
0 µM n-PG
2 µM n-PG
50 µM n-PG
200 µM n-PG
10
100
1000
Excitation irradiance (kW/cm2)
Figure 3.8: CPM of Rh6G measured by
FCS with an expanded volume element versus excitation irradiance and after addition of
various concentrations of nPG (A) and MEA
(B). C. CPM of Rh6G as a function of added
concentration of MEA.
CPM (kHz)
100
100
0 mM MEA
1 mM MEA
4 mM MEA
20 mM MEA
10
1
C
10
100
1000
Excitation irradiance (kW/cm2)
400
350
CPM, kHz
CPM (kHz)
A
300
250
200
Iexc = 0.3 MW/cm2
150
100
50
0
10
20
30
40
MEA concentration (mM)
50
reduction of non photo-oxidized fluorophores (quenching fluorescence) is at least as
important as the extent of triplet quenching.
58
Chapter 4
Triplet state & quenchers
Fluorescence quenching can be used profitably as a valuable source of information about interactions of a fluorescent molecule with a quencher. Among different
varieties of quenching mechanisms (Section 2.1.7) and fluorophore states involved,
the triplet state is especially beneficial due to its long lifetime (microseconds). This
longer lifetime allows a longer time for interactions with local environment, thus
offering greater sensitivity. In this chapter fluorescence quenching by two quenchers
– potassium iodide (paper II) and TEMPO (paper III) is studied in detail.
As demonstrated in Chapter 3, balancing the concentration of additives influencing the fluorescence is of crucial importance. Compounds which are known to
act as fluorescence promoters in certain conditions (like nPG and AA in µM concentrations) can turn out to be the fluorescence quenchers if applied inappropriately
(meaning – in mM concentrations). In paper II, FCS was used to investigate the
effect of potassium iodide (KI) on a set of fluorophores. Iodide is well known for enhancing intersystem crossing, in particular by the so-called heavy atom effect [Birks
1970; Turro et al. 2009]. Previously, it has been shown for several fluorophores that
addition of KI lead to a strong enhancement of the triplet state population [Widengren et al. 1995]. However, in paper II we show that for many fluorophores absorbing/emitting in the blue/green range, addition of KI instead leads to a reduction of
the triplet state population. Apart from a heavy atom effect, triplet state deactivation
was also found to take place by electron transfer between the fluorophore and KI.
59
Chapter 4: Triplet state & quenchers
This electron transfer was found not only to affect the triplet state, but also photooxidized as well as fluorescently viable fluorophore molecules in their singlet states.
In paper III, an approach to study bimolecular interactions in model lipid bilayers and biological membranes based on the triplet quenching mechanism is introduced. It exploits the influence of membrane-associated electron spin resonance
labels TEMPO on the fluorescence signal of a likewise membrane-bound fluorophore marker. Using fluorescence as readout maintains a high detection sensitivity,
without sacrificing the benefit of the long lifetime and environmental sensitivity
of the triplet state. The quenching mechanisms were investigated first in aqueous
solution. Then the TEMPO-induced quenching of Lissamine Rhodamine B (LRB)
was analyzed in unilamellar liposomes, with the fluorophores covalently linked to
lipid headgroups. A two-dimensional model for a diffusion-controlled, bi-molecular
quenching process was applied and found to predict well the diffusion behavior and
the collisional encounter frequencies between the LRB- and TEMPO-labeled lipids.
The proposed approach can be applied to a wide range of molecular interaction
studies in membranes.
4.1 Iodide as a triplet state promoter and
quencher
A range of fluorophores was investigated with respect to the effect of KI on their
triplet state population dynamics. For the various fluorophores, KI was added in
concentrations from 0.5 mM up to 50 mM, and their triplet state kinetics was studied at different excitation irradiances (from 10 to 1000 kW/cm2). From this initial
investigation, it was possible to identify two categories of fluorophores. For the first
group (in order of increase in absorption wavelength: Rhodamine 6G, TMR, Lissamine Rhodamine B (LRB), ATTO 590, Alexa 594, ATTO 610, Alexa 610 and Alexa 633), a prominent increase of triplet state fraction was observed with increasing
KI concentrations. This response is well in agreement with that previously observed
for KI on Rh6G [Widengren et al. 1995]. For the second group of fluorophores (in
order of decrease in absorption wavelength: Rhodamine 123, Rhodamine Green
(RhGr), ATTO 488 and Alexa 488) the effect of KI was the opposite – with increasing KI concentrations a distinct decrease of the triplet state fraction was observed.
To investigate the underlying mechanisms behind the different response to KI, one
fluorophore from each category, Rh6G from the first group and RhGr from the
second group, were selected and investigated more in detail.
A set of FCS curves recorded at various KI concentrations for these two fluoro60
Andriy Chmyrov
0 mM KI
0.2 mM KI
0.5 mM KI
1 mM KI
2 mM KI
4
G(τ)
B
Rhodamine 6G
5
3
Rhodamine Green
2.8
0 mM KI
0.5 mM KI
2 mM KI
10 mM KI
50 mM KI
2.4
G(τ)
A
2.0
1.6
2
1.2
1
1E-5
1E-4
1E-3 0.01
0.1
correlation time (ms)
1
10
1E-5
1E-4
1E-3 0.01
0.1
correlation time (ms)
1
10
Figure 4.1: A. FCS curves of Rh6G measured in aqueous solution with KI. Excitation irradiance 90 kW/cm2. With increasing KI concentration the triplet fraction T is increasing from
0.29 to 0.78, while the triplet relaxation time tT is decreasing from 1.4 µs to 0.35 µs. B. FCS
curves of RhGr measured in aqueous solution with KI in concentrations between 0 and 50
mM. Excitation irradiance 350 kW/cm2. With increasing KI concentration, T is decreasing
from 0.50 to 0.13, while tT is decreasing from 1.3 µs to 0.27 µs.
phores is shown in figure 4.1. From this figure, the different response can be clearly
seen, with a significant increase of the triplet state population, T, of Rh6G and a
corresponding decrease of T for RhGr. For both fluorophores the triplet relaxation
time was observed to significantly decrease with increasing KI concentrations. For
RhGr, a second relaxation process in the time range 5-10 µs could be observed in
the FCS curves at KI concentrations above 5 mM. Increase of the amplitude and
decrease of the relaxation time of this relaxation process versus the applied excitation irradiance was observed in the FCS curves, in a similar way as for the triplet
relaxation term. This suggests that except of an external heavy atom effect influencing the intersystem crossing rate from the singlet to the triplet manifold, there is, at
least for the second group of fluorophores, an additional triplet state deactivation
mechanism present. Such appearance of the second exponential process occurring
in the FCS curves of RhGr upon addition of KI indicates that an enhanced deactivation of the lowest triplet state is accompanied by the build-up of another, more
long-lived non-fluorescent state. As a possible additional deactivation mechanism
of the triplet state of RhGr in the presence of KI, a charge transfer reaction could be
considered [Korobov and Chibisov 1983].
To further investigate the underlying mechanisms of the deactivation of the RhGr
triplet state by KI, and what possible additional effects KI may have on the fluorescence properties of RhGr, FCS measurements on RhGr were performed systematically over a broad range of KI concentrations (1µM-50mM) on the setup with an
expanded excitation/detection volume. The outcome is presented in figure 4.2.
61
Chapter 4: Triplet state & quenchers
A
B
2.4
0 µM KI
2 µM KI
5 µM KI
20 µM KI
0.1
rate
G(τ)
2.0
1.6
1
0.01
1E-3
kox1 (106 s-1)
1E-4
1.2
1E-4 1E-3 0.01 0.1
1
correlation time (ms)
10
100
1E-5
koxn (109 s-1)
kred´ (106 s-1)
0.1
1
KI concentration (µM)
10
Figure 4.2: A. FCS curves of RhGr measured in aqueous solution using an expanded excitation volume. Excitation irradiance 240 kW/cm2. With increasing KI concentration, the amplitude of the second relaxation process R is decreasing from 0.42 to 0.09, while the relaxation
time tR is decreasing from 378 µs to 117 µs. B. The parameters of kox1, koxn, and kred¢ obtained
after fitting the experimental parameters from FCS curves obtained for each concentration of
KI in µM range.
This behavior is similar to that found for the well known antioxidants n-propyl
gallate (nPG) and ascorbic acid (AA) (see Section 3.1). Interestingly, KI is not primarily known as an antioxidant and antifading compound, as it appears to be for
the fluorophores of the second group, but rather the opposite – as a fluorescence
quencher. However, when dissociated in aqueous solution into potassium and iodide ions, it is known to act as a mild reducing agent [Drexhage 1990]. As an electron donor, it participates in a bimolecular reaction with the fluorophore molecules,
which thereby return to the normal state by receiving an electron from the iodide
ions. The expanded volume measurements were analyzed in the same way as described in Section 3.1. The obtained values of the oxidation and reduction rates for
each concentration of KI are plotted in figure 4.2B. While the oxidation rates kox1
and koxn remain virtually constant and independent of KI concentration, the reduction rate, kred, shows a linear dependence.
When gradually increasing the KI concentration from 0.5 mM to 50 mM, it
was observed (figure 4.1B) that the triplet relaxation term in the FCS curves was
decreased and shifted to faster relaxation times (tens of nanoseconds). At the same
time, at KI concentrations above 10 mM an additional relaxation process in the
time range of several microseconds could be observed. In analogy to the behavior observed for other investigated antioxidants (nPG and AA), applied in similar
concentrations (see Section 3.1) this latter process can be attributed to reduction
of non-oxidized fluorophores. Also decrease of the excited state lifetime of RhGr
was observed in presence of mM concentrations of KI. A modified kinetic scheme
incorporating these transitions is presented in figure 4.3.
62
Andriy Chmyrov
Sn
Tn
kS1n kSn1
kred*
.
R-
kISC
kISC-KI
S1
koxn
kT1n kTn1
kox1
T1
kT-KI
k01 k10 k10-KI
kT
kox*
.
R+
kred
S0
Figure 4.3: A modified kinetic scheme of photoinduced processes in presence of KI at mM
concentrations. Compare to figure 3.4B, the influence on intersystem crossing rate due to
external heavy atom effect is included by the rate kISC-KI. Triplet quenching effects is included
by the rate kT-KI. Singlet quenching effect is indicated by k10-KI. Reduction of intact viable
fluorophores by iodide is included by the rate kred*, and KI-independent oxidation of fluorophore anions represented by the rate kox*. R+ indicates cationic fluorophore radical. R- indicates
anionic fluorophore radical.
In order to reduce the complexity of the system, FCS measurements with high
concentrations of KI were again recorded from a diffraction limited observation
volume, with resulting dwell times of the fluorophore molecules in the order of 30
µs. At this fast passage times and such high concentrations of KI, photo-oxidation
of the fluorophores is relatively low and is fully recovered by iodide, therefore the
fraction of fluorophores being in R+ approaches to zero. Consequently, transitions to
and from the photo-oxidized state, R+, can be disregarded in the correlation curves,
B
1/τfl / 109 s-1
10
100
KI concentration (mM)
kred*′ / 106 s-1
1
-1
kox* / 106 s-1
0.1
1
1
C
kT / 10 s
6
rate
rate
1
10
kISC / 106 s-1
rate
A
0.01
1
10
KI concentration (mM)
10
100
KI concentration (mM)
Figure 4.4: A. The dependence of fluorescence deactivation rate from KI concentration extracted from TCSPC measurements. B. KI concentration dependence of kISC and kT, extracted
from FCS measurements. C. The parameters of iodide-induced reduction of fluorophores
kred*¢ (=kred*´[KI]) and iodide-independent oxidation of fluorophore anions kox* obtained after
fitting the experimental parameters from FCS curves obtained for each concentration of KI
in mM range.
63
Chapter 4: Triplet state & quenchers
A 2.8
B
kISC (µs-1)
2.4
2.4
kT (µs-1)
2.0
1.6
1.2
1E-5
rate
G(τ)
2.0
0 mM IB
20 mM IB
50 mM IB
100 mM IB
1E-4
1E-3 0.01
0.1
correlation time (ms)
1.6
1.2
0.8
1
10
0
20
40
60
80
IB concentration (mM)
100
Figure 4.5: A. FCS curves of RhGr measured in ethanol with Iodobenzene. Excitation irradiance 1 MW/cm2. With increasing IB concentration, the amplitude of triplet relaxation term
T increases from 0.25 to 0.45, and the triplet relaxation time tT decreases from 401 ns to 245
ns. B. IB concentration dependence of kISC and kT, extracted from FCS measurements and T
and tT parameters.
and only transitions to and from the reduced radical state R - and the triplet dynamics remain to be considered. The outcome of the analysis is presented in figure 4.4.
From the investigations of the effects of KI on RhGr, we observe that dissolved
I not only promotes the intersystem crossing to the triplet state by an external
“heavy atom” effect, but can also deactivate the T1 state, as well as the excited singlet
states by a charge transfer reaction. In order to demonstrate and more specifically
investigate the isolated “heavy atom” effect for the fluorophores of the second group
(Rhodamine Green - alike), measurements with Iodobenzene (IB) were performed
in ethanol. A set of FCS curves, recorded at various IB concentrations are shown in
figure 4.5A, together with the determined kT and kISC rates in figure 4.5B.
-
When IB is dissolved into ethanol, Iodine is not present as an ion, but rather in
a coupled form. This excludes charge transfer reactions and fluorophores present in
the solution are only expected to be influenced by a pure external heavy atom effect.
Indeed, as can be seen from figure 4.5A, there is a clear increase of the triplet state
fraction with increasing IB concentrations. The intersystem crossing rate was found
to increase linearly with the IB concentration, while the triplet relaxation rate is
close to constant (figure 4.5B).
Given the different effects of KI on the transition rates between the electronic
states of RhGr, which can influence the fluorescence yield in different and opposing
directions, we investigated effects of I- on the fluorescence countrate per molecule
(CPM) of RhGr by adding KI at different concentrations. The outcome is shown
in figure 4.6.
64
Andriy Chmyrov
Counts per molecule (kHz)
350
300
0 mM
0.5 mM
1 mM
2 mM
5 mM
10 mM
30 mM
50 mM
60 mM
80 mM
100 mM
150 mM
200 mM
300 mM
250
200
150
100
50
0
0.0
0.5
1.0
1.5
2.0
2.5
Laser power (mW)
Figure 4.6: CPM of RhGr without and with KI measured by FCS versus laser power. CPM is
increasing significantly for KI concentrations up to 5 mM, but at higher concentrations effect
of fluorescence quenching and KI-induced oxidation is prevailing.
As previously found for MEA, the optimum concentration of KI is considerably
lower than that required for a full triplet quenching of RhGr fluorophore (figure
4.1B). It is rather the balance between antioxidative properties of KI and its strength
of triplet quenching which defines the concentration at which a maximum CPM
can be reached.
A
3.2
0 mM TEMPO
10 mM TEMPO
2.0
1.6
1.6
1.2
1.2
1E-4
1E-3 0.01
0.1
correlation time (ms)
1
0 mM TEMPO
10 mM TEMPO
2.4
G(τ)
2.0
1E-5
Lissamine Rhodamine B
2.8
2.4
G(τ)
B
Rhodamine Green
2.8
10
1E-5
1E-4
1E-3 0.01
0.1
correlation time (ms)
1
10
Figure 4.7: FCS curves of RhGr (A) and LRB (B) measured in aqueous solution with and
without TEMPO. Excitation irradiance 450 kW/cm2. With addition of TEMPO, for RhGr
the amplitude of triplet relaxation term T decreases from 0.48 to 0.07, and the triplet relaxation time T decreases from 1.4 µs to 0.12 µs. Also, the second relaxation process is observed
with the amplitude 0.23 and relaxation time 1.7 µs. For LRB, addition of TEMPO results in
the increased T from 0.32 to 0.57, and the decreased T from 1.7 µs to 0.1 µs.
65
Chapter 4: Triplet state & quenchers
A similar double-sided effect of both triplet quenching and induction, as observed for KI on the fluorophores of the second category, can also be found for other
compounds. Similar to molecular oxygen, transitions to and from the triplet state
of fluorophores can also be enhanced by other paramagnetic species [Korobov and
Chibisov 1983]. Figure 4.7 shows FCS curves monitoring the triplet state population dynamics of the fluorophores RhGr and LRB in the presence of TEMPO choline. TEMPO choline is a widely used label compound in electron spin resonance
spectroscopy [Banwell and McCash 1994]. For fluorophores with excitation wavelength below 560 nm, like TMR, Rh6G and RhGr, a decrease of the triplet state
population is observed upon addition of TEMPO. For fluorophores with excitation
wavelength above 560 nm (such as LRB, ATTO590 and Alexa594) TEMPO induces triplet state build-up. This effect will be used in the following section.
4.2 Triplet state quenching for monitoring
diffusion mediated reactions
Time-correlated single photon counting (TCSPC) measurements of LRB in
aqueous solution show that upon addition of TEMPO in mM concentrations a
small decrease in the fluorescence lifetime of LRB can be observed. The enhancement of k10 by the addition of TEMPO follows to a first approximation a linear relationship k10 = (0.70 ± 0.001)×109+(8.8±0.8) ×109 [TEMPO] M-1s-1. The presence
of 1 mM TEMPO thus yields a relative change of k10 of LRB of about 1-2 percent. A
similar relative change in the fluorescence intensity can be expected, when recorded
under conventional excitation conditions (non-saturating excitation intensities).
Given the paramagnetic properties of TEMPO, a considerably stronger relative effect can be expected on the transition rates to and from the triplet state of LRB.
To investigate the quenching properties of TEMPO on the triplet states of LRB,
a series of FCS measurements was performed in aqueous solution, both under airsaturated and deoxygenated conditions. Under air-saturated conditions, FCS curves
were recorded from LRB sample solutions with varying concentrations of TEMPO
(0 mM to 5 mM), and for each TEMPO concentration, the excitation irradiance
was varied from 20 to 500 kW/cm2. Upon addition of TEMPO an increase in the
triplet amplitude, T, and a drastic shortening of the triplet relaxation time, tT, could
be observed in the correlation curves (see figure 4.8A). The correlation curves could
be well fitted to equation (38) without the need to add any additional exponential
relaxation terms in the fitting process.
For each concentration of TEMPO, the intersystem crossing rate, kISC, could
66
Andriy Chmyrov
A 2.0
1.8
G(τ)
B
Lissamine Rhodamine B
0 mM TEMPO
0.1 mM TEMPO
1 mM TEMPO
5 mM TEMPO
1.6
20
15
kISC (106 s-1)
kT (106 s-1)
10
1.4
5
1.2
1.0
1E-5
1E-4
1E-3 0.01
0.1
correlation time (ms)
1
10
0
0
2
4
6
8
TEMPO concentration (mM)
10
Figure 4.8: A. FCS curves of LRB measured in aqueous solution with and without TEMPO
in concentrations 0-5 mM. Excitation irradiance 370 kW/cm2. With increasing TEMPO concentration the triplet fraction T is increasing from 0.37 to 0.65, while the triplet relaxation
time tT is decreasing from 1.6 µs to 0.07 µs. B. TEMPO concentration dependence of kISC and
kT, extracted from FCS measurements and T and tT parameters.
then be obtained as a global fitting parameter from the excitation irradiance dependence of tT and T using equations 39 and 40. Results of such fit are presented in
figure 4.8. It can be noted that both kT and kISC are influenced by the concentration
of added TEMPO, [TEMPO], displaying a close to linear dependence. Considering the influence of TEMPO on kT and kISC to be a bimolecular reaction, molar
quenching rates coefficients kQISC = 1.6 × 109 M-1s-1 and kQT = 4.0 × 108 M-1s-1, were
determined.
From the quenching coefficients one can see that the effect of TEMPO is about 4
times stronger on the kISC rate, than on the kT, which results in the observed increase
of the triplet state fraction under air-saturated conditions. The opposite effect was
observed in a previous study [Heupel et al. 2000], where the effect of TEMPO on
the triplet state parameters of Rhodamine 6G (Rh6G) was investigated. However,
Rh6G has a lower maximum absorption wavelength than LRB, i.e. a higher lying
first excited singlet state. Since the energy levels of S1 and T1 are typically closely
related, also the T1 state of Rh6G can be expected to lie higher in energy than that
of LRB. Because of this, charge transfer reactions to the ground triplet state of
TEMPO are more likely to occur from the T1 state of Rh6G than from the T1 state
of LRB. This additional de-activation channel of the T1 state should then be reflected in the FCS measurements as a decreased steady-state triplet populations and a
shortening of the triplet relaxation times. Indeed, FCS measurements on Rh6G and
TEMPO showed not only a decrease of the triplet state fraction (in agreement with
results reported in [Heupel et al. 2000]), but also the presence of a second relaxation component in the correlation curves. This additional term can be attributed
67
Chapter 4: Triplet state & quenchers
to formation of Rh6G radicals via their T1 states [Becker et al. 1998]. However, for
LRB the FCS measurements give no evidence of a strong charge-transfer-mediated
quenching of T1. This effect was therefore not included in the analysis.
The FCS measurements were repeated on the same samples under deoxygenized
conditions. Similar to the observation in air-saturated measurements, the triplet relaxation times were found to decrease upon addition of TEMPO. However, in contrast to the air-saturated measurements, the triplet state amplitudes, T, decreased
with increasing [TEMPO]. In air-saturated aqueous solutions most of the deactivation of T1 is due to quenching by molecular oxygen. In the absence of oxygen, kT
is on the order 103s-1. Addition of TEMPO then leads to a much stronger relative
increase of kT, compared to the relative increase it would generate in an air-saturated
solution. With this in mind, and with reference to equation 40, it is clear that the
decrease of T with increasing [TEMPO] found in deoxygenated solutions reflects
the same influence of TEMPO on the kT and kISC rates of LRB, as found under airsaturated conditions.
Also FCS measurements were performed on LRB-labeled liposomes in aqueous
solution, with varying fractions of TEMPO labeled lipids included, both under
air-saturated and under de-oxygenized conditions. The fraction of TEMPO-labeled
lipids in the liposomes, L, was varied from 0 to 8% for the air-saturated measurements, and from 0 to 2% under the de-oxygenized conditions. Following the same
type of analysis as for the measurements in solution, the triplet rates kT and kISC were
determined.
The number of reactions per second (j) between molecules, B, and a single
molecule, A, is equivalent to the flow of molecules B through a volume within a reaction radius, s, around A. For the reactions taking place in a spherical shell, i.e. for
reactions in a membrane, Berg and Purcell provided an approximate expression for
j [Berg and Purcell 1977]. In the special case when all molecules occupy the same
average area, A, then j may be written as:
ϕ=
4π (DA + DB ) P
1.1 A
 1.2 A 

L ln 
 4πLs 2 
−1
(59)
Here, DA and DB are the diffusion coefficients of A and B, respectively, and P is the
probability of a reaction to occur when the reaction partners are within the reaction distance, s, from each other. L is the fraction of B molecules relative the total
number of molecules in the membrane. In this context, A means lipids in the liposome without TEMPO and B means lipids with covalently bound TEMPO.
68
Andriy Chmyrov
40
30
0.08
0.06
0.04
0.02
0 Fraction TEMPO-labeled lipids, L
0 0.002 0.004 0.006 0.008
20
0
0.001 0.002 0.003 0.004 0.005 0.006 0.007 0.008
Fraction TEMPO-labeled lipids, L
20
10
kISC (106 s-1)
1/ϕT (µs)
60
40
B
0.12
0.10
1/ϕISC (µs)
80
kT (106 s-1)
A
1.2
1.0
0.8
0.6
0.4
0.2
0 Fraction TEMPO-labeled lipids, L
0.001 0.002 0.004 0.006 0.008
0
-10
0.001 0.002 0.003 0.004 0.005 0.006 0.007
Fraction TEMPO-labeled lipids, L
Figure 4.9: Average times between two quenching reactions in liposomes in a deoxygenized
solution, leading to deactivation of the T1 state (A) of LRB, and to intersystem crossing from
the S1 to the T1 state (B) of LRB, determined for each fraction L from the difference of the
inverse rates obtained from FCS analysis. Gray lines are fits to equations (59), (60) and (61)
respectively. Insets show rates kT and kISC, respectively, determined from FCS measurements at
different excitation irradiances.
From the kT and kISC rate parameters, determined from FCS measurements,
the average time between two quenching reactions could be determined for each
fraction L from the difference of the inverse rates, in the presence and absence of
TEMPO-labeled lipids:
1
1
=
kT ( L ) − kT (0)
jT
(60)
1
1
=
kISC ( L ) − kISC (0)
jISC
(61)
Here, jT and jISC denote the molecular encounter rates as given by equation
(59), for triplet deactivation from T1 and intersystem crossing to T1, respectively.
kT(L) and kISC(L) denote the rates of kT and kISC at a lipid fraction of L, and kT(0) and
kISC(0) are the same rates in the absence of TEMPO-labeled lipids (L=0).
Results of the analysis are presented in figure 4.9 for the de-oxygenated liposomes measurements. It can be noted, that collision rates jT determined via the
triplet relaxation rate kT give much better correspondence to the model, which reflects higher sensitivity of the triplet deactivation rate to TEMPO in the absence of
oxygen. Another advantage of the de-oxygenation for the proposed approach is that
in the absence of oxygen, the triplet state lifetimes get significantly longer and even
slower molecular interaction frequencies can be followed.
69
Chapter 5
Triplet state & applications
Photoinduced dark states in fluorescence are still largely considered to be a nuisance, because they are always associated with a loss of fluorescence signal. In this
chapter two possible applications are described, which make use of transient states –
as labels for optical nanoscopy (paper IV) and as readout for monitoring molecular
diffusion (paper V).
Photo-induced switching of fluorophores into dark, long-lived states, such as the
triplet state, has recently gained increasing interest, as a means to achieve ultra-high
optical resolution. Out of the multitude of super-resolution techniques (see Section 2.2.3), GSD is especially promising technique for biological live cell application, owing to lower irradiances required for the switching mechanism. For optimal
performance of the GSD strategy, the fluorophores used should be readily photoswitchable into dark states. High triplet state populations of most fluorophores can
be easily achieved by use of certain additives that enhance photo-induced triplet
state generation, or by removal of dissolved oxygen. However, addition of these
chemical compounds or de-oxygenation is not always compatible with biological
applications. Also, dark states often appear as precursor states for photobleaching
(see Section 2.1.8 and Chapter 3), which potentially can limit their usefulness. In
paper IV, a set of fluorophores was investigated with respect to their triplet state
dynamics and photostabilities, under conditions relevant for super-resolution microscopy. Out of the investigation, several fluorophores were found to meet the
71
Chapter 5: Triplet state & applications
requirements for super-resolution microscopy, combining a prominent triplet state
yield with reasonable photostability.
In paper V, recovery of photoinduced reversible dark states is utilized as a readout parameter for molecular diffusion investigations. In TRAST techniques (see
Section 2.3.3) transitions of these states were monitored via the response in the
time-averaged fluorescence to a time-modulated excitation. In the initial TRAST
realizations typically information about the population dynamics of photo-induced
transient states was obtained. In a similar way it is also possible to get information
about diffusion in and out of the detection volume, because these transitions are superimposed on a modulation of the excitation and transitions to dark states – both
of which are well-defined.
5.1 Novel fluorescent labels for superresolution microscopy
To achieve a combination of a high triplet yield in biologically relevant environments (like water, buffers) combined with a low photobleaching quantum yield,
two chemical strategies were devised. The first of them, substitution by sulfur, has
been known to promote triplet state build-up due to the heavy atom effect (see Section 2.1.7). First, a pyronin fluorophore was used as the starting point for chemical modification, where Oxygen was substituted by Sulfur. However, the resulting
Thiopyronin (figure 5.1A) has no coupling functionality. Also it is not stable at
pH-values beyond 8, since it is prone to nucleophilic attack by hydroxyl ions at
A
B
C
R2
R4
N
S
N
H 2N
N
NH2
R5
R4
N
X
N
R6
Figure 5.1: Chemical structures of invesOH
tigated fluorophores. A. Thiopyronin. B.
C
ATTO 465. C. Rhodamine derivatives, for
D
O
Rhodamine 6G: R2 = CO2C2H5, R4 = CH3, R5 =
H, R6 = C2H5, X = O; for TMR: R2 = CO2C2H5, R4 = H, R5
= CH3, R6 = CH3, X = O; for Thiorhodamine: R2 = COOH,
R4 = H, R5 = CH3, R6 = CH3, X = S; for ATTO Thio12: R2
= CON(CH3)(CH2)3COOH, R4 = H, R5 = CH3, R6 = CH3,
X = S. D. Halogenated rhodamine derivatives, for DR25: R1
= H, R2 = CO2C2H5; for MR71: R1 = F, R2 = CO2C2H5; for
JA4: R1 = Cl, R2 = CO2H; for JA98: R1 = Br, R2 = CO2H.
72
R6
R1
N
R1
R1
R1
R2
O
N
R5
Andriy Chmyrov
Thiopyronin
λexc max.
= 563 nm
450
500
λem max.
= 585 nm
τfl = 2.0 ns
550
600
650
Wavelength/nm
700
750
ATTO 465
λexc max.
= 453 nm
350
400
λem max.
= 508 nm
τfl = 2.2 ns
450
Thiorhodamine
λexc max.
= 567 nm
λem max.
= 594 nm
500
550
Wavelength/nm
500
550
600
650
Wavelength/nm
700
750
λem max.
= 652 nm
450
500
λem max.
= 607 nm
τfl = 2.0 ns
550
600
650
Wavelength/nm
700
750
450
500
λem max.
= 608 nm
τfl = 4.0 ns
550
600
650
Wavelength/nm
450
500
550
600
650
Wavelength/nm
700
750
TMR
λexc max.
= 552 nm
λem max.
= 577 nm
450
500
500
550
600
650
Wavelength/nm
700
Figure 5.2: Absorption and emission spectra in water for
Thiopyronin (A), Thiorhodamine (B), ATTO Thio12 (C),
Rhodamine 6G (D), TMR (E), DR 25 (F), MR71 (G),
JA4 (H), JA98 (I), ATTO 465 (J).
750
750
JA4
550
600
650
Wavelength/nm
700
750
JA98
λexc max.
= 584 nm
λem max.
= 601 nm
τfl = 4.0 ns
τfl = 2.1 ns
450
700
λem max.
= 608 nm
τfl = 4.0 ns
λexc max.
= 590 nm
τfl = 3.9 ns
ATTO Thio12
λexc max.
= 580 nm
650
Rhodamine 6G
λexc max.
= 525 nm
τfl = 2.1 ns
450
600
MR71
λexc max.
= 585 nm
450
500
550
600
650
Wavelength/nm
700
750
DR25
λexc max.
= 565 nm
450
500
λem max.
= 589 nm
τfl = 4.0 ns
550
600
650
Wavelength/nm
700
750
the central carbon atom of the chromophore. To render this position inaccessible, a
carboxyphenyl group was introduced. The synthesized thiorhodamine (figure 5.1C)
was found to be much more pH stable. Finally, a NHS-ester coupling group was introduced to permit coupling to biomolecules of interest, resulting in ATTO Thio12.
The second strategy applied, halogenation, is well known to promote triplet state
build-up, as seen for instance in Eosin and Erythrosin. In these fluorophores the
halogen atoms are linked directly to the chromophore unit xanthene body of the
fluorophore, very close to the fluorescently active π-orbital region. However, this
leads to a triplet induction too strong for our purpose. Therefore we investigated
instead a set of Rhodamine derivatives with halogenated carboxyphenyl groups. The
molecular structures for all investigated fluorophores are presented in figure 5.1.
Absorption and emission maxima as well as fluorescence lifetimes are presented in
figure 5.2.
The triplet state properties of the fluorophores were investigated in detail by
FCS. In figure 5.3, FCS curves are shown for Thiopyronin (TP), Thiorhodamine
(TR) and ATTO Thio12 (AT12). A set of FCS curves for Rhodamine 6G (Rh6G),
recorded at comparable excitation intensities, are shown for reference. Comparing
the correlation curves for TP, TR and AT12 with those of Rh6G it is evident that
73
Chapter 5: Triplet state & applications
4 kW/cm2
10 kW/cm2
42 kW/cm2
1.6
1.4
1.2
1.2
1E-4
1
1.0
1E-5
10
D
ATTO Thio12
2.0
1.8
G(τ)
1E-3 0.01
0.1
Correlation time (ms)
4 kW/cm
12 kW/cm2
33 kW/cm2
1.2
1.2
1E-3 0.01
0.1
Correlation time (ms)
1
10
1
10
3 kW/cm2
13 kW/cm2
39 kW/cm2
1.6
1.4
1E-4
1E-3 0.01
0.1
Correlation time (ms)
Rhodamine 6G
2.0
1.4
1.0
1E-5
1E-4
1.8
2
1.6
4 kW/cm2
12 kW/cm2
36 kW/cm2
1.6
1.4
1.0
1E-5
Thiorhodamine
2.0
1.8
G(τ)
G(τ)
1.8
C
B
Thiopyronin
2.0
G(τ)
A
1.0
1E-5
1E-4
1E-3 0.01
0.1
Correlation time (ms)
1
10
Figure 5.3: FCS curves recorded in aqueous solution at different excitation irradiances. A.
Thiopyronin, triplet fraction T is increasing from 0.53 to 0.87, relaxation time tT: 0.68-0.14
µs. B. Thiorhodamine, T: 0.65-0.92, tT: 0.72-0.14 µs. C. ATTO Thio 12, T: 0.67-0.90,
tT: 0.78-0.19 µs. D. Rhodamine 6G, T: 0.04-0.21, tT: 2.95-1.54 µs.
the triplet formation of these dyes is much more prominent. For TP, TR as well
as AT12, the fraction of fluorophores within the excitation volume that are in the
triplet states reaches 90% and more, already at irradiances of 30–40 kW/cm2. The
recorded correlation curves could be well fitted to a model accounting for diffusion
and singlet–triplet transitions – equation (38). CPM of all the fluorophore molecules, being in the fluorescent and non-fluorescent (triplet) states, is significantly
lower for TP, TR and AT12 than for the reference (Rh6G, TMR) fluorophores. On
the other hand, when correcting for the non-fluorescent fraction of the fluorophores
being in their triplet states, the fluorophores show a (singlet state) CPM which is
well comparable to fluorophores such as Fluorescein.
The halogenated rhodamine derivatives (MR71, JA4, JA98 and DR25—a rhodamine analogue without halogens for reference) were studied by FCS in the same
fashion. For these fluorophores no significant effects could be directly observed in
the individual correlation curves on the triplet state parameters T and tT arising
from the halogenation in the phenyl ring.
74
Andriy Chmyrov
A
B
Thiopyronin
1.2
Thiorhodamine
0.9
1.0
0.8
0.8
0.7
T (%)
fitted T
τT (µs)
fitted τT
0.6
0.4
T (%)
fitted T
τT (µs)
fitted τT
0.6
0.5
0.4
0.2
0.3
0.0
C
1.0
0
10 20 30 40 50 60 70
Excitation irradiance (kW/cm2)
1.0
0.2
80
D
ATTO Thio12
0.9
0
2
4
6
8
10 12 14
Excitation irradiance (kW/cm2)
2.0
16
Rhodamine 6G
1.6
0.8
T (%)
fitted T
τT (µs)
fitted τT
0.7
0.6
0.5
1.2
T (%)
fitted T
τT (µs)
fitted τT
0.8
0.4
0.4
0.3
0
2
4 6 8 10 12 14 16 18 20
Excitation irradiance (kW/cm2)
0.0
0
50
100 150 200 250 300
Excitation irradiance (kW/cm2)
Figure 5.4: Relaxation times for singlet–triplet transitions, tT, and the fraction of fluorophores
being in the triplet state, T, determined by FCS at different excitation irradiances. The fitted
expressions for tT and T are indicated by a dashed and a solid line, respectively.
In the same manner ATTO 465 (figure 5.1B) was studied as a possible candidate for having relatively high triplet quantum yield. For this fluorophore and for
irradiances up to 200 kW/cm2 the fraction of the fluorophores being in the triplet
state increased up to 75%, but then dropped to about 40% for higher irradiances
applied. This drop in triplet fraction could be attributed to a high photobleaching quantum yield FB of ATTO 465 and photodestruction of fluorophores at high
excitation irradiances. The relative contribution of fluorescence from the periphery
of the detection volume is increased, where lower excitation irradiances are experienced, and where the relative triplet state population of the fluorophores is lower.
This drop in the observed triplet state population can be reversed by adding an
antioxidant, 30 µM of ascorbic acid in this particular case. This indicates that the
photodestruction mainly is due to photooxidation.
In order to extract the rate coefficients for intersystem crossing, kISC, and triplet
state decay, kT , FCS measurements were performed at different excitation irradiances. Extracted parameter values of T and tT were fitted to equations (39)-(40).
75
Chapter 5: Triplet state & applications
Table 5.1: Fluorescence parameters for studied fluorophores: fluorescence lifetime tfl, fluorescence quantum yield in EtOH q, intersystem crossing rate kISC and triplet relaxation rate kT
Fluorophore
τfl (ns)
kISC (µs-1)
kT (µs-1)
q in EtOH
Thiopyronin
Thiorhodamine
ATTO Thio12
ATTO 465
Rhodamine 6G
TMR
DR25
MR71
JA4
JA98
2.0
2.1
2.0
2.2
3.9
2.1
4.0
4.0
4.0
4.0
99
125
118
56
1.1
0.9
0.5
0.5
0.6
1.0
0.9
0.5
0.5
1.3
0.5
0.5
0.5
0.3
0.4
0.8
0.45
0.34
0.34
0.55
0.95
0.90
0.97
0.90
0.90
0.80
Results are presented in figure 5.4.
The triplet state parameters extracted from the FCS curves for all fluorophores
studied in this section are presented in Table 5.1. From the fitted parameters it is
evident that the high triplet population for the fluorophores TP, TR, AT12 and
ATTO 465 is mainly due to the increase of the intersystem crossing rate kISC. The
rate kT does not seem to be influenced significantly by the chemical modifications
and is believed to be dependent mainly on the concentration of dissolved oxygen.
As a possible explanation for this strong increase of the kISC rate in TP, TR and
AT12 compared to similar traditional rhodamines (Rh6G, TMR), the intramolecular heavy-atom effect may be considered (Section 2.1.7). As a result, spin–orbit coupling will be stronger, promoting transitions from S1 to T1. This explanation would
be in agreement with the literature and observations for TP, TR and AT12 where
the central oxygen atom is replaced by sulfur, whose atomic mass is twice as large.
However, this reasoning does not explain the case of ATTO 465, where the central
oxygen atom is replaced by nitrogen, which has an atomic mass even smaller than
oxygen. Although a convincing theoretical description for the increase in triplet
yield cannot be given, a simple empirical rule can explain both situations. This rule
states that in chromophore unit where the π-electrons can make a loop when oscillating between the end groups, the triplet yield will be higher than in a related compound where this loop is partially blocked [Drexhage 1990]. In the acridine fluorophore ATTO 465 the central nitrogen atom participates strongly in the conjugation
of the chromophoric system making π-electron looping possible and thus giving
rise to a high triplet yield. This effect is relatively small in common rhodamines as
the oxygen atom in the center of the chromophore essentially prevents looping of
π-electrons. A heuristic argument for this effect is that the circulating electrons create an orbital magnetic moment which couples with the spin of the electron. This
76
Andriy Chmyrov
increased spin–orbit coupling is believed to enhance the rate of intersystem crossing,
thus giving rise to a higher triplet yield.
In case of the halogenated fluorophores, a minor trend can be observed, indicating that halogenation by heavy halogen atoms causes a slight increase in kISC. The
reason why this effect is not more prominent is most likely that the halogenated sites
are not in direct contact with the chromophore unit.
Photobleaching properties of the fluorophores were studied with FCS. The fluorophores TP, TR and AT12 are similar in structure and were also found to show very
similar photostabilities, when compared to each other. The photobleaching quantum yield FB of TP was estimated to be 1.3×10-4. It can be noted that the photostability of TP is thus not as good as that of stable rhodamine fluorophores such as TMR
or Rh6G, but still better than that of, for example, coumarins and well comparable
to that of Fluorescein, an extensively used in fluorescence microscopy studies.
5.2 TRAST spectroscopy for measuring
diffusion
As shown recently [Sandén et al. 2007; Sandén et al. 2008], the dependence of the
time-averaged fluorescence on the modulation characteristics of the excitation (e.g.
duration, power and separation of the pulses within an excitation pulse train) can
be used to extract information about the population dynamics of photo-induced,
long-lived transient states of fluorophores. Likewise, if the transient state transition properties of the fluorophore are known it is also possible to get information
about the modulation characteristics from the fluorescence response. This concept
to determine molecular dwell times or diffusion properties can in principle be realized, by utilizing the kinetic properties of a range of transient photo-induced states.
However, trans-cis isomerisation (see Section 2.1.6) offers an additional advantage,
because it is light driven in both directions. Apart from a slow thermal deactivation
by kPN (figure 2.6B), in the absence of excitation the cis-isomer tends to remain
photo-isomerised. The extent of build-up of cis-isomers thus reflects the amount
of excitation irradiation that the molecules have been exposed to, or in other words
how long time they on average have spent in the excitation volume. By modulation, the distribution of the excitation irradiation in time can be adapted so that the
extent of the build-up of the transient state, in our case the cis isomer, is maximally
sensitive to changes in the dwell times of the molecules in the excitation volume.
In order to incorporate diffusion to the kinetic scheme of isomerisation presented in figure 2.6B, some simplifications were applied. Transitions between the
ground (N0 and P0) and the first excited states (N1 and P1) were disregarded, because
77
Chapter 5: Triplet state & applications
kISO′
Trans
kD
Cis
kBISO′
kD
kD
“Bulk”
Figure 5.5: The three-state model used to characterize the fluorescence fluctuations observed
in the modulated excitation experiments presented in this section. The concentration of fluorescent molecules is assumed to be constant, and normalized to unity both inside and outside
of the detection volume, such that: Trans(t) + Cis(t) = 1 and Bulk(t) = 1.
they occur on a very fast timescale of a couple of nanoseconds, and compare to utilized excitation pulse duration of several µs, the steady-state populations are already
achieved. The diffusion process is simplified by the chemical reaction with the rate
of kD of molecular exchange between the molecules inside of the excitation volume
(in Trans and Cis states) and outside of the volume (“Bulk” state). With these simplifications applied, the kinetic scheme could be reduced to the presented in figure 5.5.
By defining and solving the system of linear differential equations corresponding
to the model of Figure 5.5, it is possible to calculate the time-dependence of the
fluorescence and population levels of the isomerised states for square wave excitation
pulse trains with different pulse characteristics. Representative outcomes of the
calculated time development of the fluorescence and the population levels of the
isomerised states are shown in figure 5.6.
As can be seen in figure 5.6, the time required for pulse trains with short excitation
pulse durations, to generate a steady-state ratio between the trans- and cis- isomers
of the fluorescent molecules is longer than that required for longer pulses with the
same repetition rate. Fluorescence intensity F within the pulse is thus higher for the
shorter pulses, if the number of pulses per molecular passage through the excitation
volume is limited. Since fluorescence emission rate is rigorously defined by the
excitation and relaxation rates, it will be only the diffusion time which will lead to
deviations from the predicted averaged fluorescence.
In figure 5.7 simulations of normalized time-averaged fluorescence intensity
(FEXC = F / h, where h = w / T ) are presented for several values of the pulse width
and varying pulse period for stationary and diffusing fluorophore molecules. It can
be seen that for molecules diffusing faster through the excitation volume (shorter
diffusion times) FEXC is higher, because the molecules experience lower number of
excitations per passage. By measuring the average fluorescence within the pulse for
different pulse widths and different pulse periods, it should be possible to extract
diffusion rate kD from such decays.
78
Andriy Chmyrov
A
0.12
B
pulse width = 1.0 µs
pulse width = 5.0 µs
0.1
Trans
F
0.08
0.06
0.04
pulse width = 1.0 µs
pulse width = 5.0 µs
0.8
0.7
0.6
0.02
0
1
0.9
0
10
20
30
40 50 60
Time (µs)
70
80
90
0.5
100
0
10
20
30
40 50 60
Time (µs)
70
80
90
100
Cis
Figure 5.6: A realization of the solution to the C 0.5
0.4
differential equations corresponding to figure
5.5 for relative populations of Trans (subplot
0.3
B) and Cis (subplot C) isomers, and emitted
0.2
fluorescence F (subplot A), which assumed to
pulse width = 1.0 µs
0.1
pulse width = 5.0 µs
be proportional to N1 population of Trans iso0
0
10
20
30
40
50
60
70 80 90 100
mer. During the pulses, a cis-isomer populaTime (µs)
tion is built up, which then decays marginally
due to thermal deactivation in-between the
pulses. Solid lines represent stationary fluorophore and dotted lines - the diffusion-mediated
exchange of the molecules into and out of the excitation/detection volume to kD=1/50 µs-1.
In order to test the proposed model in detail, it is required to perform modulated
measurements on fluorescent molecules with different excitation volume residence
times. Different residence times can either be achieved by coupling fluorophore Cy5
to molecules of different sizes. However, a more convenient and tunable approach
is to vary the residence times by applying different flow speeds in a microchannel
using a syringe pump. The flow speeds were analyzed with FCS using the expression
10.0
9.5
9.0
Fexc
8.5
8.0
7.5
7.0
6.5
pulse width = 1.0 µs
pulse width = 3.0 µs
pulse width = 5.0 µs
6.0
5.5
0
50
100
150
200
250
Pulse period T (µs)
Figure 5.7: Calculated plots, with the time-averaged fluorescence plotted versus the pulse
period T, with all pulses having the same pulse width w of 1.0 µs (blue), 3.0 µs (green) and
5.0 µs (red). Solid lines are calculated for diffusion rate kD=0 µs-1, dash lines for kD=1/300 µs-1,
dash-dot lines for kD=1/200 µs-1 and dot lines for kD=1/100 µs-1.
79
Chapter 5: Triplet state & applications
A 3.2
2.4
2.0
Flow speed (mm/s)
2.8
G(τ)
B
Flow rates:
0 µl/min
50 µl/min
125 µl/min
200 µl/min
400 µl/min
700 µl/min
1200 µl/min
10
1.6
1.2
1
1E-4
1E-3
0.01
0.1
Correlation time (ms)
1
100
Flow rate (µl/min)
10
1000
Figure 5.8: A. FCS curves for Cy5 in presence of laminar flow in microchannel with fits. Excitation irradiance is 16 kW/cm2. Diffusion time tD = 70 µs, flow time tflow varies from 230 µs
at flow rate 50 µs/min to 10 µs at flow rate 1200 µl/min. B. observed linear dependence of the
flow speed V = tflow / w0 versus applied flow rate with the syringe pump.
from [Gösch et al. 2000]. The results are presented in figure 5.8.
At low flow rates (few µl/min), the characteristic flow times tflow , as obtained
from the FCS measurements are lower than the diffusion time tD. tflow is then not
representative for the molecular residence times. To estimate the residence times at
different flow rates, we therefore instead used the integral of the area of the correlation curve representing diffusion and/or flow:
∞
t residence = ∫ GD (t )Gflow (t ) d t
(62)
0
A 1.0
B 1200
1000
Normalized Fexc
0.9
0.8
flow 0 µl/min
flow 40 µl/min
flow 200 µl/min
flow 300 µl/min
0.7
0.6
0
100
200
Pulse period (µs)
300
1/kD
800
600
400
200
0
0.0
0.2
0.4
0.6
τresidence/τD
0.8
Figure 5.9: A. Measured time-averaged fluorescence FEXC from measurement with time-modulated excitation on Cy5 for various flow rates 0-300 µl/min. Pulse width w=2.0 µs, pulse
period T is varied from 5.1 to 300 µs. Curves are normalized to 1 for the longest pulse period
of 300 µs. Dotted lines represent the fit according to the proposed model B. Inverse of the
diffusion rate kD obtained as a result of the fit of the data represented in plot A.
80
1.0
Andriy Chmyrov
−1

τ
where GD (τ ) = 1 + 
 t 
D
−1 2


τ
1 +

 ( z / ω )2 ⋅ τ 
D
0
0
−1
  t 2 
t  
 


and
Gflow (t ) = exp −
⋅ 1+

  t flow  
tD  
A series of excitation-modulated measurements was performed on Cy5 in aqueous solution in a microchannel applying different flow rates. The residence times
of the fluorophore molecules were varied from 130 µs to ~40 µs, measured by FCS
in the same setup, but without excitation modulation, before and after each series
of excitation-modulated measurements. For each adjusted flow rate, FEXC was measured by excitation-modulated measurements, applying different pulse trains varying the pulse widths, w (1-5 µs), and the pulse periods, T (5-300 µs) of the different
pulse trains. A significant increase in FEXC could be observed with increasing T for
pulse widths shorter than the residence times of the fluorophores (figure 5.9A). The
peak power of the pulses was kept constant, corresponding to an excitation irradiance of 24 kW/cm2 in the detection volume, for each pulse train series.
The measured FEXC and its variation with T and w were analyzed by the model
displayed in figure 5.5, having all of the rates fixed to values obtained from FCS
measurements of freely diffusing fluorophore. Only the diffusion rate (kD) was entered as a free parameter. In figure 5.9B, the resulting inverse diffusion rates, 1/kD,
are plotted versus the residence times, tresidence, as determined from the corresponding FCS measurements, displaying a linear relationship between 1/kD and tresidence.
Excitation-modulated measurements were also performed in the same manner
on the fluorescent protein DsRed, which shows a prominent dark state relaxation
component in FCS curves with relative amplitude of about 30% and with relaxation
time in the sub-ms time range. The isomerisation dynamics in DsRed is significantly
more complex than in Cy5. As for many fluorescent proteins a multitude of states
can be expected, with relaxation times found over a broad time range. We therefore did not apply the model from figure 5.5 to evaluate the DsRed measurements.
Nonetheless, the normalized average fluorescence FEXC versus the pulse train period
was found to follow the same increase as in Cy5 measurements. Consequently, this
support the applicability of the proposed method for protein mobility studies in live
cells using fluorescent proteins.
81
Chapter 6
Triplet state & surface effects
In the majority of fluorescence techniques one is interested in maximizing the
fluorescence output per emitter. This can be done in various ways, like using antioxidative additives in order to decrease photobleaching (Chapter 3), decreasing excitation irradiances in order to decrease triplet state build up, introducing modulated
excitation with low duty cycle to allow the fluorophores to relax from the triplet
state before the next excitation will occur. The last two approaches lead to increased
measurement times, which is disadvantageous. As the triplet state plays an important role in the saturation of the fluorescence emission, the finite photodegradation
lifetime of the fluorophore, and the production of highly reactive oxygen species, it
is important to investigate triplet state kinetics under different experimental conditions. In TIR-microscopy the excitation field is inevitably confined to an interface
of two media with different refractive indexes, and the presence of this interface can
influence the photophysical processes of the fluorophore – resulting in a decreased
or an increased fluorescence rate, depending on the type of the interface. Paper VI
addresses specifically the influence of the interface on triplet state rates.
Another property of the interface – electrical charge – is of primary importance
for better understanding of interfacial dynamics when it comes to interactions of
charged particles. Since many of the commonly used fluorophores carry electrical
charge (positive, negative or both), electrostatic interactions will inevitably influence their diffusion, and therefore fluorescence. This effect, which may be consid83
Chapter 6: Triplet state & surface effects
ered undesirable, can be taken advantage of and used for probing the electrostatic
properties of a charged surface. TIR-FCS is a well suitable technique for analysis of
diffusion in the vicinity of the interface. Results of such investigations and analysis,
published in paper VII, provide an increased understanding of how fluorophores are
influenced by the microenvironment of a dielectric surface, and show a promising
approach for characterizing electrostatic interactions at interfaces in general.
6.1 Triplet state kinetic rates at dielectric
interfaces
To determine the rate constants for intersystem crossing, kISC, and the decay of
the triplet-state, kT , the TIR-FCS autocorrelation function is collected at different
excitation intensities – so called “power series”. In figure 6.1, the resulting TIR-FCS
autocorrelation functions for Fluorescein measured at three different excitation intensities are shown. The measurements show contributions from triplet-state kinetics and free diffusion. To highlight the kinetics of the triplet state, G(t) has been
normalized to one at short lag times, t. This makes it clearly visible that the triplet
fraction T increases and triplet relaxation time tT decreases with increasing excitation intensities
1.0
P = 1.65 mW
P = 6.60 mW
P = 19.40 mW
G(τ)
0.8
0.6
0.4
0.2
0.0
1E-5
1E-4
1E-3
0.01
0.1
1
10
Correlation time (ms)
Figure 6.1: Normalized correlation curves of Fluorescein in TRIS buffer (pH = 8.2 and 150
mM NaCl) at three different laser excitation powers and the corresponding TIR-FCS fits using
equation (49.)
The data from figure 6.1 are fitted with the correlation function given by expression (49) to extract the experimental values of T and tT. By plotting the experimentally determined T and tT values as a function of excitation irradiances and simultaneously making nonlinear least square fits to equations (39) and (40), the unknown
kISC and kT rates are determined (shown in figure 6.2).
84
Andriy Chmyrov
A 1.8
B
Fluorescein
1.6
1.4
1.2
1.0
2.4
T
2.0
1.6
1.2
0.6
0.4
0.8
0.2
0.4
0.0
0
4x1016
2x1016
Photons µs-1 cm-2
0.0
0.0
6x1016
D
Rhodamine 110
2.0
T
5.0x1016
1.0x1016
Photons µs-1 cm-2
1.5x1016
Rhodamine 123
1.2
1.2
τT (µs)
0.8
T
τT (µs)
T
0.4
0.4
0.0
0.0
2.0
τT (µs)
1.6
1.6
0.8
ATTO 488
2.8
τT (µs)
0.8
C
3.2
5.0x1016
1.0x1017
Photons µs-1 cm-2
0.0
0.0
1.5x1017
1.0x1016
5.0x1016
Photons µs-1 cm-2
1.5x1016
Figure 6.2: The measured triplet-state parameters, tT (filled squares) and T (filled circles), and
the corresponding weighted and spatially averaged global fits for: Fluorescein (A), ATTO 488
(B), Rhodamine 110 (C) and Rhodamine 123 (D).
To be able to compare the intersystem crossing and the triplet decay rates deduced by TIR-FCS, with and without the glass surfaces, FCS measurements of the
same sample were also performed with a confocal setup, using the original approach
introduced in [Widengren et al. 1995]. Obtained rates for the investigated fluorophores are presented in table 6.1
From table 6.1 it can be seen that the triplet rates kISC and kT measured with
TIR-FCS are slightly higher than the reference rates determined by confocal FCS.
Table 6.1: Obtained values of the triplet rates for Fluorescein (in TRIS buffer, pH=8.2 150
mM NaCl) and ATTO 488, Rhodamine 110, and Rhodamine 123 (in PBS buffer, pH=7.4).
TIR-FCS
Fluorescein
ATTO 488
Rhodamine 110
Rhodamine 123
confocal FCS
kISC (µs-1)
kT (µs-1)
kISC (µs-1)
kT (µs-1)
13.5 ± 1.0
1.7 ± 0.1
1.3 ± 0.1
1.2 ± 0.1
0.66 ± 0.04
0.32 ± 0.01
0.49 ± 0.03
0.58 ± 0.04
11.1 ± 1.0
1.5 ± 0.1
1.0 ± 0.1
1.0 ± 0.1
0.54 ± 0.04
0.25 ± 0.03
0.45 ± 0.04
0.53 ± 0.04
85
Chapter 6: Triplet state & surface effects
A reason for the higher rates at the surface could be the underestimation of the
axial extent of the excitation, h. Also an underestimation of the magnitude of the
evanescent field would lead to an overestimation of the rates. In addition, the values
used for the excitation intensity distribution also depend on the saturation properties of the dye. In confocal FCS, high laser powers often generate distortion of the
emission profile [Hess and Webb 2002], which may lead to underestimation of the
kISC rates. Furthermore, high excitation irradiance opens up additional relaxation
channels, like reverse intersystem crossing that may lower the kT values [Ringemann
et al. 2008]. The confocal reference rates might therefore be slightly underestimated.
The relatively low irradiances used in TIR-FCS do not introduce the problems
just mentioned. However, the fluorescence lifetime, tF , used in calculating the triplet rates should probably be modified slightly. The reason for this is that the interface
can modify the electromagnetic decay channels of the fluorophore [Fort and Grésillon 2008]. For a bare dielectric surface this can lead to a decrease in fluorescence
lifetime by 5-10 % compared to that in bulk solution. To take this into account, the
determined rates for kISC of Fluorescein, ATTO 488, Rhodamine 110, and Rhodamine 123 should be increased about five to ten percent.
When the corresponding measurements and analysis were done for Rhodamine
6G in concentration 50 nM, much higher triplet rates in TIR-FCS were obtained
(kISC = 1.6 ± 0.2 µs-1 and kT = 1.2 ± 0.04 µs-1), compare to reference confocal FCS
values of kISC = 1.1 ± 0.2 µs-1 and kT = 0.49 ± 0.05 µs-1. This indicates that there is
an interface-induced effect which influence the triplet-state kinetics of Rhodamine
6G. To investigate this further we decided to change the concentration of the Rh6G
sample. Figure 6.3 shows a comparison of the TIR-FCS triplet-state rate analysis of
tT and T at sample concentrations of 1 nM and 50 nM in PBS. As it is clearly visible, lower concentration generates longer triplet relaxation times and higher triplet
amplitudes.
Among possible reasons for the concentration-dependence of the triplet rates for
Rh6G, hydrophobicity-induced interactions with the surface opening new efficient
deactivation channels may be considered [Mialocq et al. 1991]. Formation of aggregates is often accompanied by a change in the radiative and radiation-less transition
probabilities. Furthermore, aggregates in the form of adsorbed monomers, dimers,
and trimers have been shown to play an important role in long range energy and
electron-transfer mechanisms for freely diffusing Rh6G molecules [López Arbeloa
et al. 1988]. The strong spin-orbit coupling induced by aggregates through spreading in their singlet energy levels can actually increase the intersystem crossing rates,
as seen above. Shortening of the triplet relaxation times, via triplet-triplet annihila86
Andriy Chmyrov
B
1.6
1.2
1.0
0.8
0.6
0.4
0.2
0.0
0.0
28
24
1.4
Triplet Amplitude (%)
Triplet Relaxation Time (µs)
A 1.8
1 nM
50 nM
1.0x1017
5.0x1016
Photons µs-1 cm-2
1 nM
50 nM
20
16
12
8
4
0
0.0
1.5x1017
Figure 6.3: Comparison of measured triplet
state parameters tT (A) and T (B) for Rh6G in
PBS at two different concentrations (1 nM and
50 nM). C. The corresponding global fits of
the TIR-FCS data for Rh6G at 1 nM, yielding
kISC = 2.4 ± 0.2 µs-1 and kT = 0.66 ± 0.07 µs-1.
C
1.0x1017
5.0x1016
Photons µs-1 cm-2
1.8
1.6
1.5x1017
1 nM
1.4
1.2
1.0
0.8
τT (µs)
0.6
T
0.4
0.2
0.0
0.0
1.0x1017
5.0x1016
Photons µs-1 cm-2
1.5x1017
tion, may additionally be enhanced by surface adsorbed aggregates [Bryukhanov et
al. 1978].
Another dye-solvent effect that is of interest is the well-known property of fluorescence quenching by potassium iodide (see Chapter 4). The effect is shown in
figure 6.4. Addition of 5 mM iodide give rise to large triplet populations and fast
triplet relaxation times, which transform into increased measured triplet rates.
In contrast to the bulk solution case, the extracted kISC and kT rates do not show
a linear dependence over the whole concentration range. The trend of getting similar rates with TIR-FCS and confocal FCS decreased as the iodide concentration is
lowered. This can be attributed to influence from aggregates that enhance the triplet
rates and to changes in saturation and bleaching properties of Rh6G at the surface.
Here it was generally assumed that all Rh6G fluorophores are photochemically
intact and no photobleaching may occur. The saturation and photobleaching properties of the fluorophore under investigation can differ considerably with different
environmental parameters. The ratio of the intersystem crossing rate and the triplet
relaxation rate actually gives an estimate on how prone the dye is to photobleaching.
Addition of potassium iodide greatly increases this ratio, which leads to increased
87
Chapter 6: Triplet state & surface effects
A
B
1.0
1.2
0 mM [KI]
5 mM [KI]
0.8
G(τ)
1.4
0.6
0.4
0.4
0.2
0.9
1E-4
1E-3 0.01
0.1
Correlation time (ms)
1
0 mM [KI]
2 mM [KI]
5 mM [KI]
0.8
0.7
0.6
P = 47.90 mW
0.5
0.4
0.3
0.2
0
0.0
0.0
10
10 20 30 40 50 60 70 80 90 100
Axial distance (µm)
D
Detected Fluorescence / a.u.
C 1.0
Concentration / a.u.
0.2
P = 47.90 mW
1E-5
0.1
T
0.8
0.6
0.0
τT (µs)
1.0
5.0x1016
1.0x1017
Photons (µs-1cm-2)
1.5x1017
70
0 mM [KI]
2 mM [KI]
5 mM [KI]
60
50
40
P = 47.90 mW
30
20
10
0
0.0
0.1
0.2
0.3
0.4
Axial distance (µm)
0.5
0.6
Figure 6.4: A. normalized correlation curves of Rh6G in PBS buffer without and with addition of 5 mM KI, and the corresponding TIR-FCS fits to equation at maximum laser excitation power. B. triplet state parameters for 50 nM Rh6G + 5 mM KI with the corresponding
weighted and spatially averaged global fits, yielding kISC = 24.4 ± 1.5 µs-1 and kT = 0.71 ± 0.04
µs-1. C. Simulations of concentration changes of Rh6G molecules in the axial direction without and with addition of KI (2 mM and 5 mM) at maximum laser excitation power. D. The
axial detected fluorescence profiles in the observation volume. Note the large fluorescence
saturation effects in the vicinity of the interface with high potassium iodide concentrations.
photodestruction probabilities. Computer simulations of how the concentration
of Rh6G molecules in the sample volume is depleted in the presence of bleaching
are shown in figure 6.4C. A strong depletion of molecules within the observation
volume and its surroundings is predicted.
In the vicinity of the surface, high triplet-state buildup of Rh6G due to very
high excitation irradiances is further increased upon addition of potassium iodide.
This inevitably leads to promoted photodestruction of fluorophores at the interface,
which makes the overall concentration depletion even larger. Figure 6.4D shows the
saturation of detected fluorescence in the observation volume when adding potassium iodide. When all of these effects are taken into account, there is a resulting
reduction in the relative fluorescence contributions from Rh6G molecules close to
88
Andriy Chmyrov
Table 6.2: Obtained values of the intersystem crossing rates, kISC, and the triplet relaxation
rate, kT , for Rhodamine 6G with and without KI.
TIR-FCS
[KI] (mM)
0
0.2
0.5
2
5
τf-1 (µs-1)
253
254
257
267
284
confocal FCS
kISC (µs-1)
kT (µs-1)
kISC (µs-1)
kT (µs-1)
1.6 ± 0.2
1.6 ± 0.1
3.5 ± 0.3
10.8 ± 0.7
24.4 ± 1.5
1.2 ± 0.1
0.92 ± 0.06
0.72 ± 0.04
0.66 ± 0.03
0.71 ± 0.04
1.1 ± 20%
1.6 ± 20%
5 ± 20%
15 ± 20%
28 ± 20%
0.49 ± 10%
0.50 ± 10%
0.53 ± 10%
0.61 ± 10%
0.73 ± 10%
the surface compared to those further away. This leads to a bias in the correlation
analysis towards non-interfacial Rh6G molecules, which may explain the similar
triplet rates seen with TIR-FCS and confocal FCS at high potassium iodide concentrations (table 6.2).
6.2 Electrostatic interactions at dielectric
interfaces
When a solid material is in contact with an aqueous solution, a thin interfacial
charge layer is formed, which consist of surface charges and charge balancing counterions. The presence of the surface charges causes a rearrangement of the ions in the
vicinity of the surface. This arrangement of a nonzero net charge at the solid-liquid
interface is usually referred to as the electric double layer. In the electric double
layer, two regions of charge distributions may be identified. Immediately next to the
charged surface, counterions are bound to the surface due to strong electrostatic attraction. Outside of this immobile layer counterions may move, meaning that they
can diffuse in the potential set up by the partially screened surface. The immobile
region is often referred to as the compact layer or the Stern layer, and the mobile
region – as the diffusive layer. The electrostatic potential at the boundary dividing
the compact layer and the diffusive layer is the so-called zeta (z) potential [Delgado
et al. 2005]. A schematic model of the electric double layer on a glass surface is
shown in figure 6.5.
On uncoated glass surface, acidic silanol groups acquire a negative charge (SiO−)
at neutral pH. This negative surface charge density is balanced by counterions attracted to the surface and diffusive ions counterbalancing the potential set up by
the partially screened surface. On uncoated hydrophilic surfaces such as glass, the
surface potential has been shown to be indistinguishable from the zeta potential
[Bouzigues et al. 2008].
89
Chapter 6: Triplet state & surface effects
z (nm)
0
100
Glass surface
-
200
+
+
+
+
+
-
-
+
+
+
+
300
-
Immobile
+
-
-
+
+
+
500
-
+
+
400
-
+
Mobile
Figure 6.5: Schematic model of electric double layer. The negative surface charges (−) are
counterbalanced by bound and free ions (+). The electrostatic surface potential is plotted for
salt concentrations of 100 mM, 10 mM, 1 mM, and 0.1 mM, which gives Debye lengths of
0.95 nm (black curve), 3 nm (blue curve), 9.5 nm (green curve), and 30 nm (orange curve),
respectively. The evanescent penetration depth is for comparison plotted in the red curve.
How rapid the surface potential drops is connected to the Debye-Hückel
parameter [Ohshima 2006], which for dilute electrolytes can be expressed as
κ 2 = 2χe 2 εr ε0 kBT . Here c is the ionic strength of the solution, e is the elementary
charge and the parameters er and e0 are the permittivity of the electrolyte and that
of vacuum, kB is the Boltzmann constant and T is the absolute temperature. The
B
0.5
0.8
0.4
0.6
0.3
G(τ)
G(τ)
A 1.0
0.1 mM salt
0.4
100 mM salt
10 mM salt
0.1 mM salt
0.2
10 mM salt
0.2
0.0
100 mM salt
1E-6
1E-5
1E-4
1E-3
Correlation time (s)
0.01
0.1
0.1
0.0
1E-6
1E-5
1E-4
1E-3
Correlation time (s)
Figure 6.6: A. Autocorrelation curves for the cationic Rhodamine 123 at ionic strengths of
100 mM, 10 mM and 0.1 mM. To visualize the effect of the electrostatic interactions on the
diffusion time, the curves have been normalized. B. The unnormalized TIR-FCS curves used
to visualize changes in concentration of molecules in the vicinity of the glass surface.
90
0.01
Andriy Chmyrov
reciprocal of k correspond to the fundamental length scale of the surface potential,
or in other words to the thickness of the electric double layer. This so-called Debye
screening length, zD, depends on the ionic strength of the solution, c, and can at
room temperature be expressed as κ−1 = z D = 0.3 / χ in units of nanometer. Figure 6.5 shows the surface potential at the different ionic strengths, together with the
evanescent excitation profile given by the z-dependent part of equation (41). The
electrostatic surface potentials are plotted according to Gouy-Chapman theory that
assumes it to decay exponentially with distance to the interface [Ohshima 2006]:
Ψ ( z ) = Ψ 0 exp (− z z D ) .
The influence of the surface potential on the molecular motion is illustrated in
figure 6.6, which shows the TIR-FCS autocorrelation curves for cationic Rhodamine 123 interacting with the negative glass surfaces at different ionic strengths.
Note an increased decay-time (figure 6.6A) with decreasing ionic strength (meaning a less screened surface potential). Figure 6.6B shows how the concentration of
molecules, in the vicinity of the surfaces, increases as the ionic strength decreases
from 100 mM to 0.1 mM. This is reflected in a decrease of the amplitude of the
correlation curve. Both the shift in amplitude, as wells as the shift in decay-time of
the correlation curves, clearly shows that the positively charged rhodamine 123 is
attracted to the negatively charged dielectric surface.
Figure 6.7 shows all the parameters extracted from TIR-FCS measurements for
Rhodamine 123. At low ionic strength cationic fluorophore molecules are attracted
towards the negatively charged surface, which results in increased local concentration (measured via number of molecules N ), increased axial passage times – the
fluorophore spends more time closer to the surface, and increased CPMs and triplet
amplitude – due to on average higher excitation irradiances experienced.
Assuming a Poisson-Boltzmann description for the number of fluorophores (at
thermal equilibrium) in the vicinity of the surface, the one-dimensional concentration distribution at the surface may be approximated as:
C (r ) = C ( z ) = C B × exp(−qe Ψ( z ) / kBT )
(63)
Here q is the charge value of the fluorescent molecule probing the screened glass
surface (q = +1, ±0, −1, −2) and CB is the concentration in the bulk (that is far away
from the surface). In TIR-FCS, the number of molecules is determined exclusively
by the concentration profile, C ( z ) , and the dimension and shape of the detection
volume. While there are no (well-defined) boarders of the detection volume, its ab91
Chapter 6: Triplet state & surface effects
A
B
3.6
3.2
-10
2.4
Ψ (mV)
N, NX
2.8
2.0
1.6
-35
0
20
25
40 60 80 100 120 140
Ionic strength (mM)
D
20
0
20
40 60 80 100 120 140
Ionic strength (mM)
0
20
40 60 80 100 120 140
Ionic strength (mM)
0
20
40 60 80 100 120 140
Ionic strength (mM)
600
550
CPM (kHz)
τz (µs)
-20
-30
0.8
C
-15
-25
1.2
0.4
0
-5
15
10
500
450
5
0
20
E
40 60 80 100 120 140
Ionic strength (mM)
400
F
0.16
2.8
τT (µs)
0.14
T
0.12
0.10
0.08
2.4
2.0
1.6
0.06
0.04
3.2
1.2
0
20
40 60 80 100 120 140
Ionic strength (mM)
Figure 6.7: Parameters extracted from TIR-FCS measurements on Rhodamine 123 as a function of ionic strength. A. The averaged number of molecules, N (triangles) and the ratio of
molecules, Nx=N/NB (pentagons). B. The deduced mean surface potential of the electric double layer. C. The axial passage-time, tz. D. The counts-rate-per-molecule, CPM. E. The triplet
amplitude, T. F. The triplet relaxation time, tT .
solute size is exactly defined and corresponds to the volume containing N molecules
−1
as defined via equation (49), that is C ( z ) ∝ N = [G (0)×(1 −T )] . Similarly, the
bulk values NB µ CB can be approximated with the number of molecules found in
the volume at a “fully” screened surface.
92
Andriy Chmyrov
A
B
6
22
18
4
τz (µs)
N
5
Rhodamine 6G+
(cationic)
3
Rhodamine 6G+
(cationic)
14
10
2
1
C
6
0
20
40
60
80
Ionic strength (mM)
100
0
D
2.4
20
40
60
80
Ionic strength (mM)
100
4.4
2.2
1.8
τz (µs)
N
2.0
Fluorescein-/-2
(anionic/diaonic)
1.6
3.6
Fluorescein-/-2
(anionic/diaonic)
2.8
1.4
1.2
2.0
1.0
0
1.6
60
80
100
0
F
Rhodamine 110+(zwitterionic)
1.4
N
40
Ionic strength (mM)
20
40
60
80
Ionic strength (mM)
100
5.8
Rhodamine 110+(zwitterionic)
5.6
5.4
1.2
τz (µs)
E
20
ATTO 488
(neutral)
1.0
5.2
ATTO 488
(neutral)
5.0
4.8
0.8
0
20
40
60
Ionic strength (mM)
80
100
0
20
40
60
80
100
Ionic strength (mM)
Figure 6.8: The number of molecules N (left) and the axial passage times tz (right) for Rhodamine 6G (top), Fluorescein (middle), Rhodamine 110 and ATTO 488 (bottom) molecules,
in the TIR-FCS probe volume, as a function of ionic strength. For various fluorophores concentrations are different.
Assuming negatively charged glass surfaces, depending on the charge of the fluorescent molecule, three cases of concentration variations can thus be identified and
presented in figure 6.8. No concentration changes at all (N equal to NB ) should be
generated when q is zero. Fluorescent molecules having a positive q-value should be
attracted towards the surface (N larger than NB ), whereas a negative q-value would
93
Chapter 6: Triplet state & surface effects
lead to repelling of molecules by the surface (N smaller than NB ). By measuring
the excess (or deficit) of molecules with TIR-FCS, a mean surface potential (figure
6.7B) may thus be evaluated for each ionic strength via equation (63).
94
Concluding remarks
This thesis focuses on investigations of transient dark states in fluorescent molecules using spectroscopic techniques. The main purpose is to show and convince
the reader that transient dark states are not always a nuisance, but also represent an
additional valuable source of information. Knowledge about origins, properties and
dynamics of transient states could be used as a mean to either utilize these states, or
eliminate them.
In paper I, it was shown that with the right strategy and using on-shelf additives,
it is possible to significantly reduce photodegradation of fluorophores. Thus, higher
signal and longer observation times in fluorescence imaging and spectroscopy can
be achieved. The balance between reduction of ionized fluorophores and reduction
of intact fluorophores should be an important guideline for selecting the concentration ranges to be applied. Adjusting the properties of potent available additives
by introducing new chemical groups with desired effects is the way for improving
performance of the anti-bleaching agents. This would not be possible without the
detailed characterization of the process of fluorescence blinking.
In paper II triplet state quenching was investigated. Potassium iodide, a compound commonly used as a fluorescence quencher, turned out to be a fluorescence
promoter for several fluorophores. The underlying mechanisms were investigated,
and apart from the heavy atom effect, an electron transfer reaction was identified.
This electron transfer reaction was found not only to affect the triplet state, but also
the photo-oxidized as well as the fluorescently viable fluorophore molecules in their
singlet states. The kinetic transition rates and their KI concentration dependence
95
Concluding remarks
were determined, providing information for which fluorophores and in what concentrations KI can be added such that primarily triplet state and other non-fluorescent states of the fluorophores are deactivated, rather than the fluorescently viable
forms of the fluorophores. Based on this information, and when added in balanced
concentrations, KI can be used as a fluorescence promoter and anti-fading compound. In addition, the fact that KI acts differently on different fluorophores and
fluorophore states suggests that KI can be used as a contrast enhancement mechanism in biomolecular dynamics and interaction studies.
In paper III quenching of fluorescence with the nitroxide radical TEMPO was
investigated in aqueous solution. This resulted in the use of this quencher for monitoring bimolecular reactions in lipid membranes. The high environmental sensitivity of the triplet state combined with the excellent detection sensitivity of the
fluorescence read-out meets all prerequisites for exploitation of triplet quenching for
monitoring different processes.
In paper IV special purpose fluorophores were developed and characterized for
possible use in super-resolution imaging techniques based on photoswitching. Due
to the involvement of the triplet state in the photobleaching mechanisms, the fluorophores were required to combine high triplet yield with reasonable photostability.
Of the dyes investigated, the rhodamine and the pyronin dyes with sulfur atom
replacing the central oxygen atom in the xanthene unit were found to meet these
requirements. Thiorhodamine with a coupling functionality became available commercially as ATTO Thio 12.
In paper V photoinduced switching to non-fluorescent states was used for monitoring molecular diffusion. Since fluorescence emission is rigorously defined (on a
timescale of µs) by the kinetic excitation and relaxation rates, factors influencing
fluorescence emission in a systematic way (like Brownian diffusion) could be monitored and measured, given the right model of the underlying kinetics. The proposed
approach provide advantages such as simple instrumentation, scaling possibility and
wide range of accessible concentrations. Feasibility of the proposed method for protein mobility studies in live cells using fluorescent proteins was demonstrated.
In papers VI and VII studies of the triplet state kinetics of fluorophores at dielectric interfaces were performed. These studies are important due to the involvement
of the triplet state in the saturation of fluorescence emission and photodegradation
of the fluorophore. Better understanding of the fundamental photophysical processes is a prerequisite for optimization of conditions for ultrasensitive fluorescence
microscopy. For all of the flurophores studied, higher triplet rates were found. This
was attributed to possible modifications of photophysical properties, occurring at
96
Andriy Chmyrov
dielectric interfaces. Specific interface-induced effects were observed for fluorophore
Rhodamine 6G, dependent on the concentration of the fluorophore and the ionic
strength of the solvent. Also, the analysis of the triplet state kinetic can provide
information about local microenvironment and electrostatic interactions near dielectric interfaces.
97
Acknowledgements
This thesis would not have been possible without the help and support of
colleagues, family and friends.
First, I thank my supervisor Jerker Widengren for giving me the opportunity
to pursue research in the fascinating field of fluorescence, for his support and
everlasting enthusiasm and optimism.
I thank all the members, past and present, of the Experimental Biomolecular
Physics at KTH:
Tor Sandén for the fruitful collaboration, optimism and hospitality,
Hans Blom for all the help in the lab and the constant supply of references,
Kai Haßler for knowledge about TIRF and sport activities,
Evangelos Sisamakis for sharing inexhaustible enthusiasm and great company,
Johan Strömqvist for creative modeling,
Per Thyberg for correlation theory and data analysis support,
Heike Hevekerl for patience and the promising collaboration,
for the good atmosphere: Gustav Persson, Per-Åke Löfdahl, Anders Hedqvist,
Stefan Wennmalm, Sofia Johansson, Thiemo Speilmann, Lei Xu, Daniel
Rönnlund, Håkan Lepp, August Andersson, Shadi Khatibi, Magnus Brändén,
Lina Salomonsson and Yu Ohsugi.
I express my deep gratitude to all of the members of SPOTLIGHT consortium,
from Siegen – Professor Karl-Heinz Drexhage, Jutta Arden-Jacob and Alexander
Zilles; from Göttingen – Professor Stefan Hell, Christian Eggeling, Lars Kastrup
and Stefan Bretschneider; from Turku – Professor Pekka Hänninen.
I thank for short (but very intense) visits and long lasting collaboration with
Düsseldorf University – Professor Claus Seidel, Steffen von Dahlen, Daniela
Pfiffi, Stanislav Kalinin, Suren Felekyan and Ralf Kühnemuth.
Thanks to the members of Cellphysics goup at KTH, and especially to: Athanasia
Christakou, Padideh Kamali-Zare, Jacob Kowalewski and Susanna Rydholm.
Finally, and most of all, I thank to my family for all the love and support, and for
encouragement of my all initiatives.
99
References
1.
2.
3.
4.
5.
6.
7.
8.
9.
10.
11.
12.
Aitken, C. E., R. A. Marshall and J. D. Puglisi (2008). An Oxygen Scavenging System
for Improvement of Dye Stability in Single-Molecule Fluorescence Experiments. Biophysical
Journal 94(5): 1826-1835.
Aramendia, P. F., R. M. Negri and E. S. Roman (1994). Temperature Dependence of Fluorescence and Photoisomerization in Symmetric Carbocyanines.Influence of Medium Viscosity
and Molecular Structure. The Journal of Physical Chemistry 98(12): 3165-3173.
Axelrod, D., D. E. Koppel, J. Schlessinger, E. Elson and W. W. Webb (1976). Mobility
measurement by analysis of fluorescence photobleaching recovery kinetics. Biophysical Journal 16(9): 1055-1069.
Axelrod, D., T. P. Burghardt and N. L. Thompson (1984). Total Internal Reflection Fluorescence. Annual Review of Biophysics and Bioengineering 13(1): 247-268.
Axelrod, D. (2001). Total Internal Reflection Fluorescence Microscopy in Cell Biology. Traffic 2(11): 764-774.
Banwell, C. N. and E. M. McCash (1994). Fundamentals of Molecular Spectroscopy. London, McGraw-Hill.
Baranova, E. G. (1965). On the Concentration Quenching of the Luminescence of Rhodamine 6G Solutions. Optics and Spectroscopy 18: 230.
Bates, M., T. R. Blosser and X. Zhuang (2005). Short-Range Spectroscopic Ruler Based on
a Single-Molecule Optical Switch. Physical Review Letters 94(10): 108101.
Becker, S., I. Gregor and E. Thiel (1998). Photoreactions of rhodamine dyes in basic solvents. Chemical Physics Letters 283(5-6): 350-356.
Berg, H. C. and E. M. Purcell (1977). Physics of chemoreception. Biophysical Journal
20(2): 193-219.
Bergström, F., I. Mikhalyov, P. Hagglof, R. Wortmann, T. Ny and L. B. A. Johansson
(2001). Dimers of Dipyrrometheneboron Difluoride (BODIPY) with Light Spectroscopic
Applications in Chemistry and Biology. Journal of the American Chemical Society 124(2):
196-204.
Berland, K. M., P. T. So and E. Gratton (1995). Two-photon fluorescence correlation
101
References
13.
14.
15.
16.
17.
18.
19.
20.
21.
22.
23.
24.
25.
26.
27.
28.
29.
30.
102
spectroscopy: method and application to the intracellular environment. Biophysical Journal
68(2): 694-701.
Berne, B. J. and R. Pecora (1976). Dynamic Light Scattering. New York, Wiley-Interscience.
Betzig, E., G. H. Patterson, R. Sougrat, O. W. Lindwasser, S. Olenych, J. S. Bonifacino,
M. W. Davidson, J. Lippincott-Schwartz and H. F. Hess (2006). Imaging intracellular
fluorescent proteins at nanometer resolution. Science 313(5793): 1642-1645.
Birks, J. B. (1970). Photophysics of aromatic molecules. London, John Wiley & Sons Ltd.
Blom, H. and G. Björk (2009). Lorentzian spatial intensity distribution in one-photon
fluorescence correlation spectroscopy. Applied Optics 48(31): 6050-6058.
Bonnet, G., O. Krichevsky and A. Libchaber (1998). Kinetics of conformational fluctuations in DNA hairpin-loops. Proceedings of the National Academy of Sciences of the
United States of America 95(15): 8602-8606.
Born, M. and E. Wolf (1999). Principles of Optics: Electromagnetic Theory of Propagation,
Interference and Diffraction of Light, Cambridge University Press.
Bouzigues, C. I., P. Tabeling and L. Bocquet (2008). Nanofluidics in the Debye Layer at
Hydrophilic and Hydrophobic Surfaces. Physical Review Letters 101(11): 114503.
Bryukhanov, V. V., G. A. Ketsle and L. V. Levshin (1978). Phosphorescence of higher aggregates of rhodamine 6G in mixtures of polar and nonpolar solvents. Journal of Applied
Spectroscopy 28(4): 507-509.
Bryukhanov, V. V., B. F. Minaev, A. S. Kusenova and S. G. Karstina (1992). External
heavy atom effect on the intersystem crossing from the lower and higher excited states of rhodamine dyes on a silica surface. Journal of Applied Spectroscopy 56(2): 146-150.
Burghardt, T. P. and N. L. Thompson (1984). Effect of planar dielectric interfaces on
fluorescence emission and detection. Evanescent excitation with high-aperture collection. Biophysical Journal 46(6): 729-737.
Carlsson, K., P. E. Danielsson, R. Lenz, A. Liljeborg, L. Majlöf and N. Åslund (1985).
Three-dimensional microscopy using a confocal laser scanning microscope. Optics Letters
10(2): 53-55.
Carlsson, K. and N. Åslund (1987). Confocal imaging for 3-D digital microscopy. Applied
Optics 26(16): 3232-3238.
Chandra, A. K., N. J. Turro, A. L. Lyons and P. Stone (1978). The intramolecular external heavy atom effect in bromo-, benzo-, and naphthonorbornenes. Journal of the American
Chemical Society 100(16): 4964-4968.
Chattopadhyay, K., E. L. Elson and C. Frieden (2005). The kinetics of conformational
fluctuations in an unfolded protein measured by fluorescence methods. Proceedings of the
National Academy of Sciences of the United States of America 102(7): 2385-2389.
Chibisov, A. K. (1966). Photoisomerization of some dyes by pulsating light irradiation.
Teoreticheskaya i Eksperimental’naya Khimiya 2: 825-829.
Chibisov, A. K. (1976). Triplet states of cyanine dyes and reactions of electron transfer with
their participation. Journal of Photochemistry 6(3): 199-214.
Chung, E., D. Kim and P. T. So (2006). Extended resolution wide-field optical imaging:
objective-launched standing-wave total internal reflection fluorescence microscopy. Optics
Letters 31(7): 945-947.
Clegg, R. M. (1992). Fluorescence resonance energy transfer and nucleic acids. Methods in
Andriy Chmyrov
Enzymology 211: 353-388.
31. Clegg, R. M. (2009). Förster resonance energy transfer-FRET what is it, why do it, and
how it’s done. Fret and Flim Techniques. T. W. J. Gadella. 33: 1-57.
32. Condon, E. (1926). A Theory of Intensity Distribution in Band Systems. Physical Review
28(6): 1182-1201.
33. Davies, M. J. (2004). Reactive species formed on proteins exposed to singlet oxygen. Photochemical and Photobiological Sciences 3(1): 17-25.
34. Delgado, A. V., E. Gonzalez-Caballero, R. J. Hunter, L. K. Koopal and J. Lyklema
(2005). Measurement and interpretation of electrokinetic phenomena - (IUPAC technical
report). Pure and Applied Chemistry 77(10): 1753-1805.
35. Denk, W., J. Strickler and W. Webb (1990). Two-photon laser scanning fluorescence microscopy. Science 248(4951): 73-76.
36. Dertinger, T., V. Pacheco, I. v. d. Hocht, R. Hartmann, I. Gregor and J. Enderlein
(2007). Two-Focus Fluorescence Correlation Spectroscopy: A New Tool for Accurate and
Absolute Diffusion Measurements. ChemPhysChem 8(3): 433-443.
37. Di Francia, G. (1952). Super-gain antennas and optical resolving power. Il Nuovo Cimento (1943-1954) 9(0): 426-438.
38. Dittrich, P. S. and P. Schwille (2001). Photobleaching and stabilization of fluorophores
used for single-molecule analysis with one- and two-photon excitation. Applied Physics BLasers and Optics 73(8): 829-837.
39. Donnert, G., J. Keller, C. A. Wurm, S. O. Rizzoli, V. Westphal, A. Schönle, R. Jahn, S.
Jakobs, C. Eggeling and S. W. Hell (2007). Two-Color Far-Field Fluorescence Nanoscopy.
Biophysical Journal 92(8): L67-L69.
40. Drexhage, K.-H. (1990). Structure and Properties of Laser Dyes. Dye Lasers. F. P. Schäfer.
Berlin, Springer-Verlag: 155-200.
41. Duarte, F. J. and L. W. Hillman (1990). Dye laser principles, with applications, Academic
Press.
42. Dyba, M. and S. W. Hell (2002). Focal Spots of Size l/23 Open Up Far-Field Florescence
Microscopy at 33 nm Axial Resolution. Physical Review Letters 88(16): 163901.
43. Eftink, M. R. and C. A. Ghiron (1981). Fluorescence quenching studies with proteins.
Analytical Biochemistry 114(2): 199-227.
44. Eggeling, C., J. R. Fries, L. Brand, R. Günther and C. A. M. Seidel (1998a). Monitoring
conformational dynamics of a single molecule by selective fluorescence spectroscopy. Proceedings of the National Academy of Sciences of the United States of America 95(4): 15561561.
45. Eggeling, C., J. Widengren, R. Rigler and C. A. M. Seidel (1998b). Photobleaching of
fluorescent dyes under conditions used for single-molecule detection: Evidence of two-step
photolysis. Analytical Chemistry 70(13): 2651-2659.
46. Eggeling, C., A. Volkmer and C. A. Seidel (2005). Molecular photobleaching kinetics of
Rhodamine 6G by one- and two-photon induced confocal fluorescence microscopy. ChemPhysChem 6(5): 791-804.
47. Eggeling, C., J. Widengren, L. Brand, J. Schaffer, S. Felekyan and C. A. M. Seidel
(2006). Analysis of photobleaching in single-molecule multicolor excitation and Forster resonance energy transfer measurements. The Journal of Physical Chemistry A 110(9): 297995.
103
References
48. Ehrenberg, M. and R. Rigler (1974). Rotational brownian motion and fluorescence intensify fluctuations. Chemical Physics 4(3): 390-401.
49. Elson, E. L. and D. Magde (1974). Fluorescence correlation spectroscopy. I. Conceptual
basis and theory. Biopolymers 13: 1-27.
50. Enderlein, J., T. Ruckstuhl and S. Seeger (1999). Highly Efficient Optical Detection of
Surface-Generated Fluorescence. Applied Optics 38(4): 724-732.
51. English, D. S., E. J. Harbron and P. F. Barbara (2000). Probing photoinduced intersystem
crossing by two-color, double resonance single molecule spectroscopy. The Journal of Physical
Chemistry A 104(40): 9057-9061.
52. Fölling, J., M. Bossi, H. Bock, R. Medda, C. A. Wurm, B. Hein, S. Jakobs, C. Eggeling
and S. W. Hell (2008). Fluorescence nanoscopy by ground-state depletion and single-molecule return. Nature Methods 5(11): 943-945.
53. Förster, T. (1948). Zwischenmolekulare Energiewanderung und Fluoreszenz. Annalen der
Physik 437(1-2): 55-75.
54. Fort, E. and S. Grésillon (2008). Surface enhanced fluorescence. Journal of Physics D:
Applied Physics 41(1): 013001.
55. Franck, J. and E. G. Dymond (1926). Elementary processes of photochemical reactions.
Transactions of the Faraday Society 21: 536 - 542.
56. Gaigalas, A. K., L. Wang, K. D. Cole and E. Humphries (2004). Photodegradation of fluorescein in solutions containing n-propyl gallate. Journal of Physical Chemistry A 108(20):
4378-4384.
57. Geisler, C., A. Schönle, C. von Middendorff, H. Bock, C. Eggeling, A. Egner and S. W.
Hell (2007). Resolution of l/10 in fluorescence microscopy using fast single molecule photoswitching. Applied Physics A: Materials Science & Processing 88(2): 223-226.
58. Geissbühler, M., T. Spielmann, A. Formey, I. Märki, M. Leutenegger, B. Hinz, K.
Johnsson, D. Van De Ville and T. Lasser (2010). Triplet Imaging of Oxygen Consumption
during the Contraction of a Single Smooth Muscle Cell (A7r5). Biophysical Journal 98(2):
339-349.
59. Giloh, H. and J. W. Sedat (1982). Fluorescence microscopy: reduced photobleaching of
rhodamine and fluorescein protein conjugates by n-propyl gallate. Science 217(4566):
1252-5.
60. Gösch, M., H. Blom, J. Holm, T. Heino and R. Rigler (2000). Hydrodynamic flow
profiling in microchannel structures by single molecule fluorescence correlation spectroscopy.
Analytical Chemistry 72(14): 3260-3265.
61. Gustafsson, M. G. L., D. A. Agard and J. W. Sedat (1995). Sevenfold improvement of
axial resolution in 3D wide-field microscopy using two objective lenses. Proceedings of SPIE
2412: 147-156.
62. Gustafsson, M. G. L. (2000). Surpassing the lateral resolution limit by a factor of two using
structured illumination microscopy. Journal of Microscopy 198: 82-87.
63. Gustafsson, M. G. L. (2005). Nonlinear structured-illumination microscopy: Wide-field
fluorescence imaging with theoretically unlimited resolution. Proceedings of the National
Academy of Sciences of the United States of America 102(37): 13081-13086.
64. Hassler, K., T. Anhut, R. Rigler, M. Gosch and T. Lasser (2005a). High count rates with
total internal reflection fluorescence correlation spectroscopy. Biophysical Journal 88(1): L1L3.
104
Andriy Chmyrov
65. Hassler, K., M. Leutenegger, P. Rigler, R. Rao, R. Rigler, M. Gösch and T. Lasser
(2005b). Total internal reflection fluorescence correlation spectroscopy (TIR-FCS) with low
background and high count-rate per molecule. Optics Express 13(19): 7415-7423.
66. Hassler, K. (2006). Thesis: Single molecule detection and fluorescence correlation spectroscopy on surfaces. Laboratoire d’Optique Biomedicale. Lausanne, EPFL: 116.
67. Hecht, B., B. Sick, U. P. Wild, V. Deckert, R. Zenobi, O. J. F. Martin and D. W. Pohl
(2000). Scanning near-field optical microscopy with aperture probes: Fundamentals and
applications. The Journal of Chemical Physics 112(18): 7761-7774.
68. Hecht, E. (2001). Optics, Addison Wesley.
69. Heintzmann, R. and C. G. Cremer (1999). Laterally modulated excitation microscopy:
improvement of resolution by using a diffraction grating. Proceedings of SPIE.
70. Heintzmann, R., T. M. Jovin and C. Cremer (2002). Saturated patterned excitation microscopy - a concept for optical resolution improvement. Journal of the Optical Society of
America A 19(8): 1599-1609.
71. Hell, S. and E. H. K. Stelzer (1992). Fundamental improvement of resolution with a
4Pi-confocal fluorescence microscope using two-photon excitation. Optics Communications
93(5-6): 277-282.
72. Hell, S. W. (1990). Double confocal microscope. European patent.
73. Hell, S. W. and J. Wichmann (1994). Breaking the Diffraction Resolution Limit by Stimulated-Emission - Stimulated-Emission-Depletion Fluorescence Microscopy. Optics Letters
19(11): 780-782.
74. Hell, S. W. and M. Kroug (1995). Ground-State-Depletion Fluorescence Microscopy - a
Concept for Breaking the Diffraction Resolution Limit. Applied Physics B-Lasers and Optics 60(5): 495-497.
75. Hell, S. W., S. Jakobs and L. Kastrup (2003). Imaging and writing at the nanoscale with
focused visible light through saturable optical transitions. Applied Physics A: Materials Science & Processing 77(7): 859-860.
76. Hell, S. W., M. Dyba and S. Jakobs (2004). Concepts for nanoscale resolution in fluorescence microscopy. Current Opinion in Neurobiology 14(5): 599-609.
77. Hell, S. W. (2007). Far-field optical nanoscopy. Science 316(5828): 1153-1158.
78. Hell, S. W. (2009a). Microscopy and its focal switch. Nature Methods 6(1): 24-32.
79. Hell, S. W. (2009b). Far-Field Optical Nanoscopy. Single Molecule Spectroscopy in Chemistry, Physics and Biology. A. Gräslund, R. Rigler and J. Widengren. Berlin Heidelberg,
Springer. 96: 365-398.
80. Hellen, E. H. and D. Axelrod (1987). Fluorescence emission at dielectric and metal-film
interfaces. Journal of the Optical Society of America B 4(3): 337-350.
81. Hess, S. T. and W. W. Webb (2002). Focal volume optics and experimental artifacts in
confocal fluorescence correlation spectroscopy. Biophysical Journal 83(4): 2300-2317.
82. Hess, S. T., T. P. K. Girirajan and M. D. Mason (2006). Ultra-High Resolution Imaging by Fluorescence Photoactivation Localization Microscopy. Biophysical Journal 91(11):
4258-4272.
83. Heupel, M., I. Gregor, S. Becker and E. Thiel (2000). Photophysical and photochemical
properties of electronically excited fluorescent dyes: A new type of time-resolved laser-scanning
spectroscopy. International Journal of Photoenergy 1(3): XXIII-XXIV.
84. Hofmann, M., C. Eggeling, S. Jakobs and S. W. Hell (2005). Breaking the diffraction
105
References
barrier in fluorescence microscopy at low light intensities by using reversibly photoswitchable proteins. Proceedings of the National Academy of Sciences of the United States of
America 102(49): 17565-17569.
85. Hoijtink, G. J. (1960). The influence of paramagnetic molecules on singlet-triplet transitions. Molecular Physics: An International Journal at the Interface Between Chemistry
and Physics 3(1): 67 - 70.
86. Hosaka, N. and T. Saiki (2006). 10 nm Spatial Resolution Fluorescence Imaging of Single
Molecules by Near-Field Scanning Optical Microscopy Using a Tiny Aperture Probe. Optical
Review 13(4): 262-265.
87. Huang, Z., D. Ji, S. Wang, A. Xia, F. Koberling, M. Patting and R. Erdmann (2005).
Spectral Identification of Specific Photophysics of Cy5 by Means of Ensemble and Single
Molecule Measurements. The Journal of Physical Chemistry A 110(1): 45-50.
88. Hubner, C. G., A. Renn, I. Renge and U. P. Wild (2001). Direct observation of the triplet
lifetime quenching of single dye molecules by molecular oxygen. The Journal of Chemical
Physics 115(21): 9619-9622.
89. Jabłoński, A. (1935). Über den Mechanismus der Photolumineszenz von Farbstoffphosphoren. Zeitschrift für Physik A Hadrons and Nuclei 94(1): 38-46.
90. Jares-Erijman, E. A. and T. M. Jovin (2003). FRET imaging. Nature Biotechnology
21(11): 1387-1395.
91. Jelley, E. E. (1936). Spectral absorption and fluorescence of dyes in the molecular state [1].
Nature 138(3502): 1009-1010.
92. Kasha, M. (1950). Characterization of electronic transitions in complex molecules. Discussions of the Faraday Society 9: 14-19.
93. Kasha, M. (1952). Collisional Perturbation of Spin-Orbital Coupling and the Mechanism
of Fluorescence Quenching. A Visual Demonstration of the Perturbation. The Journal of
Chemical Physics 20(1): 71-74.
94. Kastrup, L., H. Blom, C. Eggeling and S. W. Hell (2005). Fluorescence Fluctuation Spectroscopy in Subdiffraction Focal Volumes. Physical Review Letters 94(17): 178104.
95. Kawaoka, K., A. U. Khan and D. R. Kearns (1967). Role of Singlet Excited States of
Molecular Oxygen in the Quenching of Organic Triplet States. The Journal of Chemical
Physics 46(5): 1842-1853.
96. Kelkar, V. K., B. S. Valaulikar, J. T. Kunjappu and C. Manohar (1990). Aggregation
characteristics of laser dye rhodamine 6g in aqueous surfactant solutions. Photochemistry
and Photobiology 52(4): 717-721.
97. Ketsle, G. A., L. V. Levshin and V. V. Bryukhanov (1976). Mechanism of the external
heavy atom effect on intersystem crossing in rhodamine dye solutions. Journal of Applied
Spectroscopy 24(5): 573-577.
98. Khan, S. A. and N. A. Riza (2006). Demonstration of a no-moving-parts axial scanning
confocal microscope using liquid crystal optics. Optics Communications 265(2): 461-467.
99. Korobov, V. E. and A. K. Chibisov (1983). Primary Photoprocesses in Colorant Molecules.
Russian Chemical Reviews 52(1): 27-42.
100. Krichevsky, O. and G. Bonnet (2002). Fluorescence correlation spectroscopy: the technique
and its applications. Reports on Progress in Physics 65(2): 251-297.
101. Lakowicz, J. R. (2006). Principles of fluorescence spectroscopy. New York, Springer.
102. Leutenegger, M., M. Gösch, A. Perentes, P. Hoffmann, O. J. F. Martin and T. Lasser
106
Andriy Chmyrov
(2006). Confining the sampling volume for Fluorescence Correlation Spectroscopy using a
sub-wavelength sized aperture. Optics Express 14(2): 956-969.
103. Levshin, L. V. and N. Nizamov (1966). Concentration quenching of the luminescence of
rhodamine dyes in binary solvents. Journal of Applied Spectroscopy 5(6): 531-534.
104. Lippincott-Schwartz, J., E. Snapp and A. Kenworthy (2001). Studying protein dynamics
in living cells. Nature Reviews Molecular Cell Biology 2(6): 444-456.
105. Longin, A., C. Souchier, M. Ffrench and P. A. Bryon (1993). Comparison of anti-fading
agents used in fluorescence microscopy: Image analysis and laser confocal microscopy study.
Journal of Histochemistry and Cytochemistry 41(12): 1833-1840.
106. López Arbeloa, F., P. Ruiz Ojeda and I. López Arbeloa (1988). The fluorescence quenching
mechanisms of Rhodamine 6G in concentrated ethanolic solution. Journal of Photochemistry and Photobiology A: Chemistry 45(3): 313-323.
107. Lukosz, W. (1966). Optical Systems with Resolving Powers Exceeding the Classical Limit.
Journal of the Optical Society of America 56(11): 1463-1471.
108. Lukosz, W. and R. E. Kunz (1977a). Light emission by magnetic and electric dipoles close
to a plane interface. I. Total radiated power. Journal of the Optical Society of America
67(12): 1607-1615.
109. Lukosz, W. and R. E. Kunz (1977b). Light emission by magnetic and electric dipoles close
to a plane dielectric interface. II. Radiation patterns of perpendicular oriented dipoles. Journal of the Optical Society of America 67(12): 1615-1619.
110. Lukosz, W. (1979). Light emission by magnetic and electric dipoles close to a plane dielectric
interface. III. Radiation patterns of dipoles with arbitrary orientation. Journal of the Optical Society of America 69(11): 1495-1503.
111. Magde, D., E. Elson and W. W. Webb (1972). Thermodynamic Fluctuations in a Reacting
System—Measurement by Fluorescence Correlation Spectroscopy. Physical Review Letters
29(11): 705-708.
112. Magde, D., E. L. Elson and W. W. Webb (1974). Fluorescence correlation spectroscopy. II.
An experimental realization. Biopolymers 13(1): 29-61.
113. Magde, D., W. W. Webb and E. L. Elson (1978). Fluorescence correlation spectroscopy. III.
Uniform translation and laminar flow. Biopolymers 17(2): 361-376.
114. Marmé, N., G. Habl and J.-P. Knemeyer (2005). Aggregation behavior of the red-absorbing oxazine derivative MR 121: A new method for determination of pure dimer spectra.
Chemical Physics Letters 408(4-6): 221-225.
115. Marrocco, M. (2007). Fluorescence correlation spectroscopy of diffusion probed with a
Gaussian-Lorentzian spatial distribution. Chemical Physics Letters 449(1-3): 227-230.
116. Matthews, D. R., L. M. Carlin, E. Ofo, P. R. Barber, B. Vojnovic, M. Irving, T. Ng
and S. M. Ameer-Beg (2010). Time-lapse FRET microscopy using fluorescence anisotropy.
Journal of Microscopy 237(1): 51-62.
117. McCartin, P. J. (1965). Observation of Metastable Geometrical Isomers of Cyanines by
Flash Photolysis. The Journal of Chemical Physics 42(8): 2980-2981.
118. McGlynn, S. P., T. Azumi and M. Kinoshita (1969). Molecular spectroscopy of the triplet
state. Englewood Cliffs, N.J., Prentice-Hall.
119. Mertz, J. (2000). Radiative absorption, fluorescence, and scattering of a classical dipole
near a lossless interface: a unified description. Journal of the Optical Society of America B
17(11): 1906-1913.
107
References
120. Mets, Ü., J. Widengren and R. Rigler (1997). Application of the antibunching in dye fluorescence: Measuring the excitation rates in solution. Chemical Physics 218(1-2): 191-198.
121. Meyer, L., D. Wildanger, R. Medda, A. Punge, S. O. Rizzoli, G. Donnert and S. W.
Hell (2008). Dual-Color STED Microscopy at 30-nm Focal-Plane Resolution. Small 4(8):
1095-1100.
122. Meyvis, T., S. De Smedt, P. Van Oostveldt and J. Demeester (1999). Fluorescence Recovery After Photobleaching: A Versatile Tool for Mobility and Interaction Measurements in
Pharmaceutical Research. Pharmaceutical Research 16(8): 1153-1162.
123. Mialocq, J. C., P. Hébert, X. Armand, R. Bonneau and J. P. Morand (1991). Photophysical and photochemical properties of rhodamine 6G in alcoholic and aqueous sodium dodecylsulphate micellar solutions. Journal of Photochemistry and Photobiology A: Chemistry
56(2-3): 323-338.
124. Minsky, M. (1957). Microscopy apparatus patent. United States.
125. Moneron, G. and S. W. Hell (2009). Two-photon excitation STED microscopy. Optics
Express 17(17): 14567-14573.
126. Moneron, G., R. Medda, B. Hein, A. Giske, V. Westphal and S. W. Hell (2010). Fast
STED microscopy with continuous wave fiber lasers. Optics Express 18(2): 1302-1309.
127. Nettels, D., I. V. Gopich, A. Hoffmann and B. Schuler (2007). Ultrafast dynamics of protein collapse from single-molecule photon statistics. Proceedings of the National Academy
of Sciences 104(8): 2655-2660.
128. Noukakis, D., M. Van de Auweraer, S. Toppet and F. C. De Schryver (1995). Photophysics of thiacarbocyanine dye in organic solvent. The Journal of Physical Chemistry 99(31):
11860-11866.
129. Ohshima, H. (2006). Theory of colloid and interfacial electric phenomena. Amsterdam,
Elsevier Academic Press.
130. Paddock, S. (2000). Principles and practices of laser scanning confocal microscopy. Molecular Biotechnology 16(2): 127-149.
131. Pappalardo, R., H. Samelson and A. Lempicki (1970). Long-pulse laser emission from
rhodamine 6G. Quantum Electronics, IEEE Journal of 6(11): 716-725.
132. Pawley, J. B., Ed. (2006). Handbook of Biological Confocal Microscopy, Springer.
133. Petersen, N. O. (1986). Scanning fluorescence correlation spectroscopy. I. Theory and simulation of aggregation measurements. Biophysical Journal 49(4): 809-815.
134. Poher, V., H. X. Zhang, G. T. Kennedy, C. Griffin, S. Oddos, E. Gu, D. S. Elson, M.
Girkin, P. M. W. French, M. D. Dawson and M. A. Neil (2007). Optical sectioning
microscopes with no moving parts using a micro-stripe array light emitting diode. Optics
Express 15(18): 11196-11206.
135. Pringsheim, P. (1949). Fluorescence & Phosphorescence New York, Wiley InterScience.
136. Qian, H. and E. L. Elson (1991). Analysis of confocal laser-microscope optics for 3-D fluorescence correlation spectroscopy. Applied Optics 30(10): 1185-1195.
137. Rae, M., A. Fedorov and M. N. Berberan-Santos (2003). Fluorescence quenching with
exponential distance dependence: Application to the external heavy-atom effect. The Journal
of Chemical Physics 119(4): 2223-2231.
138. Rasnik, I., S. A. McKinney and T. Ha (2006). Nonblinking and long-lasting single-molecule fluorescence imaging. Nature Methods 3(11): 891-893.
139. Rigler, R. and J. Widengren (1990). Ultrasensitive detection of single molecules by
108
Andriy Chmyrov
fluorescence correlation spectroscopy. BioScience. B. Klinge and C. Owman: 180-183.
140. Rigler, R., U. Mets, J. Widengren and P. Kask (1993). Fluorescence Correlation Spectroscopy with High Count Rate and Low-Background - Analysis of Translational Diffusion.
European Biophysics Journal with Biophysics Letters 22(3): 169-175.
141. Ringemann, C., A. Schonle, A. Giske, C. von Middendorff, S. W. Hell and C. Eggeling
(2008). Enhancing fluorescence brightness: Effect of reverse intersystem crossing studied by
fluorescence fluctuation spectroscopy. ChemPhysChem 9(4): 612-624.
142. Rittweger, E., K. Y. Han, S. E. Irvine, C. Eggeling and S. W. Hell (2009). STED microscopy reveals crystal colour centres with nanometric resolution. Nature Photonics 3(3):
144-147.
143. Roy, R., S. Hohng and T. Ha (2008). A practical guide to single-molecule FRET. Nature
Methods 5(6): 507-516.
144. Rust, M. J., M. Bates and X. W. Zhuang (2006). Sub-diffraction-limit imaging by stochastic optical reconstruction microscopy (STORM). Nature Methods 3(10): 793-795.
145. Saggau, P. (2006). New methods and uses for fast optical scanning. Current Opinion in
Neurobiology 16(5): 543-550.
146. Sandén, T., G. Persson, P. Thyberg, H. Blom and J. Widengren (2007). Monitoring
kinetics of highly environment sensitive states of fluorescent molecules by modulated excitation and time-averaged fluorescence intensity recording. Analytical Chemistry 79(9): 33303341.
147. Sandén, T., G. Persson and J. Widengren (2008). Transient State Imaging for Microenvironmental Monitoring by Laser Scanning Microscopy. Analytical Chemistry 80(24): 95899596.
148. Schäfer, F. P. (1990). Principles of Dye Laser Operation. Dye Lasers. F. P. Schäfer. Berlin,
Springer-Verlag: 1-89.
149. Schuler, B. and W. A. Eaton (2008). Protein folding studied by single-molecule FRET.
Current Opinion in Structural Biology 18(1): 16-26.
150. Schwille, P., F. J. Meyer-Almes and R. Rigler (1997). Dual-color fluorescence cross-correlation spectroscopy for multicomponent diffusional analysis in solution. Biophysical Journal
72(4): 1878-86.
151. Seidel, C. A. M. (1991). Nucleic acid base specific quenching of coumarin-120-derivative
in nucleotid-conjugates-photoinduced electron transfer. Proceedings of SPIE.
152. Seidel, C. A. M., A. Schulz and M. H. M. Sauer (1996). Nucleobase-Specific Quenching of Fluorescent Dyes. 1. Nucleobase One-Electron Redox Potentials and Their Correlation with Static and Dynamic Quenching Efficiencies. The Journal of Physical Chemistry
100(13): 5541-5553.
153. Selvin, P. R. (2000). The renaissance of fluorescence resonance energy transfer. Nature Structural & Molecular Biology 7(9): 730-734.
154. Sharonov, A. and R. M. Hochstrasser (2006). Wide-field subdiffraction imaging by accumulated binding of diffusing probes. Proceedings of the National Academy of Sciences
103(50): 18911-18916.
155. Sies, H. and C. F. M. Menck (1992). Singlet oxygen induced DNA damage. Mutation
Research/DNAging 275(3-6): 367-375.
156. Singh, R. J., J. B. Feix and B. Kalyanaraman (1992). Photobleaching of merocyanine 540:
involvement of singlet molecular oxygen. Photochemistry and Photobiology 55(4): 483-
109
References
489.
157. Solov’ev, K. N. and E. A. Borisevich (2005). Intramolecular heavy-atom effect in the photophysics of organic molecules. Physics-Uspekhi(3): 231.
158. Spielmann, T., H. Blom, M. Geissbuehler, T. Lasser and J. Widengren (2010). Transient State Monitoring by Total Internal Reflection Fluorescence Microscopy. The Journal of
Physical Chemistry B 114(11): 4035-4046.
159. Stokes, G. G. (1852). On the Change of Refrangibility of Light. Philosophical Transactions of the Royal Society of London 142: 463-562.
160. Stracke, F., M. Heupel and E. Thiel (1999). Singlet molecular oxygen photosensitized by
Rhodamine dyes: correlation with photophysical properties of the sensitizers. Journal of Photochemistry and Photobiology a-Chemistry 126(1-3): 51-58.
161. Synge, E. H. (1928). A suggested method for extending microscopic resolution into the ultramicroscopic region. Philosophical Magazine Series 7 6(35): 356 - 362.
162. Thompson, N. L. and D. Axelrod (1981). Total internal reflection/ fluorescence correlation
spectroscopy study of antibody surface binding kinetics. Biophysical Journal 33(2): A183.
163. Thompson, N. L., T. P. Burghardt and D. Axelrod (1981). Measuring surface dynamics of
biomolecules by total internal reflection fluorescence with photobleaching recovery or correlation spectroscopy. Biophysical Journal 33(3): 435-454.
164. Thompson, N. L. (1992). Fluorescence correlation spectroscopy. Topics in Fluorescence
Spectroscopy. Volume 1. Techniques. J. R. Lakowicz. New York, Plenum Press: 337-378.
165. Thompson, R. E., D. R. Larson and W. W. Webb (2002). Precise Nanometer Localization
Analysis for Individual Fluorescent Probes. Biophysical Journal 82(5): 2775-2783.
166. Tinnefeld, P., D.-P. Herten and M. Sauer (2001). Photophysical Dynamics of Single Molecules Studied by Spectrally-Resolved Fluorescence Lifetime Imaging Microscopy (SFLIM).
The Journal of Physical Chemistry A 105(34): 7989-8003.
167. Tomin, V. (2008). Studying reactions that proceed from highest excited states of molecules
using fluorescence quenching. Optics and Spectroscopy 105(4): 496-510.
168. Tsien, R. Y. (1998). THE GREEN FLUORESCENT PROTEIN. Annual Review of Biochemistry 67(1): 509-544.
169. Turro, N. J., V. Ramamurthy and J. C. Scaiano (2009). Principles of molecular photochemistry : an introduction. Sausalito, California, University Science Books.
170. Valeur, B. (2002). Molecular Fluorescence: Principles and Applications, Wiley-VCH.
171. Vogelsang, J., R. Kasper, C. Steinhauer, B. Person, M. Heilemann, M. Sauer and P. Tinnefeld (2008). A Reducing and Oxidizing System Minimizes Photobleaching and Blinking
of Fluorescent Dyes. Angewandte Chemie International Edition 47(29): 5465-5469.
172. Wawilow, S. J. (1927). Die Fluoreszenzausbeute von Farbstofflösungen als Funktion der
Wellenlänge des anregenden Lichtes. II. Zeitschrift für Physik A Hadrons and Nuclei
42(4): 311-318.
173. Wennmalm, S., P. Thyberg, L. Xu and J. Widengren (2009). Inverse-Fluorescence Correlation Spectroscopy. Analytical Chemistry 81(22): 9209-9215.
174. White, J., W. Amos and M. Fordham (1987). An evaluation of confocal versus conventional imaging of biological structures by fluorescence light microscopy. The Journal of Cell
Biology 105(1): 41-48.
175. White, J. and E. Stelzer (1999). Photobleaching GFP reveals protein dynamics inside live
cells. Trends in Cell Biology 9(2): 61-65.
110
Andriy Chmyrov
176. Widengren, J., R. Rigler and Ü. Mets (1994). Triplet state monitoring by fluorescence correlation spectroscopy. Journal of Fluorescence 4(3): 255-258.
177. Widengren, J., Ü. Mets and R. Rigler (1995). Fluorescence Correlation Spectroscopy of
Triplet-States in Solution - a Theoretical and Experimental-Study. The Journal of Physical
Chemistry 99(36): 13368-13379.
178. Widengren, J. and R. Rigler (1996). Mechanisms of photobleaching investigated by fluorescence correlation spectroscopy. Bioimaging 4(3): 149-157.
179. Widengren, J., J. Dapprich and R. Rigler (1997). Fast interactions between Rh6G and
dGTP in water studied by fluorescence correlation spectroscopy. Chemical Physics 216(3):
417-426.
180. Widengren, J. and P. Schwille (2000). Characterization of photoinduced isomerization
and back-isomerization of the cyanine dye Cy5 by fluorescence correlation spectroscopy. The
Journal of Physical Chemistry A 104(27): 6416-6428.
181. Widengren, J. and C. A. M. Seidel (2000). Manipulation and characterization of photoinduced transient states of Merocyanine 540 by fluorescence correlation spectroscopy. Physical
Chemistry Chemical Physics 2(15): 3435-3441.
182. Widengren, J., E. Schweinberger, S. Berger and C. A. M. Seidel (2001). Two new concepts to measure fluorescence resonance energy transfer via fluorescence correlation spectroscopy: Theory and experimental realizations. Journal of Physical Chemistry A 105(28):
6851-6866.
183. Widengren, J., A. Chmyrov, C. Eggeling, P.-Å. Löfdahl and C. A. M. Seidel (2007).
Strategies to improve photostabilities in ultrasensitive fluorescence spectroscopy. The Journal
of Physical Chemistry A 111(3): 429-440.
184. Wildanger, D., E. Rittweger, L. Kastrup and S. W. Hell (2008). STED microscopy with
a supercontinuum laser source. Optics Express 16(13): 9614-9621.
185. Wilkinson, F. (1997). Quenching of electronically excited states by molecular oxygen in fluid
solution. Pure and Applied Chemistry 69(4): 851-856.
186. Willig, K. I., B. Harke, R. Medda and S. W. Hell (2007). STED microscopy with continuous wave beams. Nature Methods 4(11): 915-918.
187. Woods, R. W. (1934). Physical optics, Macmillan.
188. Yadav, L. D. S. (2004). Organic Spectroscopy. Dodrecht, Springer Netherlands.
189. Yang, H., G. Luo, P. Karnchanaphanurach, T.-M. Louie, I. Rech, S. Cova, L. Xun and
X. S. Xie (2003). Protein Conformational Dynamics Probed by Single-Molecule Electron
Transfer. Science 302(5643): 262-266.
190. Yeow, E. K. L., S. M. Melnikov, T. D. M. Bell, F. C. De Schryver and J. Hofkens (2006).
Characterizing the Fluorescence Intermittency and Photobleaching Kinetics of Dye Molecules
Immobilized on a Glass Surface. The Journal of Physical Chemistry A 110(5): 1726-1734.
191. Yuzhakov, V. I. (1992). Aggregation of dye molecules and its influence on the spectral luminescent properties of solutions. Russian Chemical Reviews(6): 613.
192. Zhu, P., J.-P. Clamme and A. A. Deniz (2005). Fluorescence Quenching by TEMPO: A
Sub-30 Å Single-Molecule Ruler. Biophysical Journal 89(5): L37-L39.
193. Zhuang, X., T. Ha, H. D. Kim, T. Centner, S. Labeit and S. Chu (2000). Fluorescence
quenching: A tool for single-molecule protein-folding study. Proceedings of the National
Academy of Sciences of the United States of America 97(26): 14241-14244.
111