For more than 200 years, scientists of virtually all disciplines—Optics, Physics, Chemistry, and Biology—have enjoyed an intense fascination with molecular chirality and the dissymmetric interaction of chiral systems with circularly polarized light.1–7 In physics, dissymmetries are manifested at the most fundamental subatomic material level.8–10 In biology and chemistry, molecular chirality plays a key role in biochemical synthesis,11–15 molecular recognition,16–30 and self-assembly17, 31–49 processes in nature. More recently, studies of supramolecular chiral assemblies in natural systems have inspired new directions in chiral photonics.50 A great wealth of data on the dissymmetric chiroptical response on ensembles of structurally identical chiral systems has been obtained from spectroscopic tools such as optical rotatory dispersion (ORD)51–55 and circular dichroism (CD),2, 56 fluorescence detected circular dichroism (FDCD).29, 57–71 Combined with detailed theoretical analysis, these tools lead ultimately to assignment of absolute chiral configurations from spectroscopic data.72–82 However, an important aspect of this interaction that remains poorly understood—in particular for condensed phase systems—is the role of local molecular environment and configurational fluctuations that could result in a significant heterogeneity in the chiroptical response of isolated molecular systems.83
The idea of molecular chirality and its manifestation through a dissymmetric interaction with polarized light is introduced in chemistry courses at the introductory level. An object is said to be chiral if the object and its mirror image are not superimposable; that is, the object cannot be mapped onto its mirror image by any combination of rotations and translations.1 For chiral molecular systems, this is manifested in a nonresonant light interaction through a different bulk polarizability (refractive index) for right- and left-circularly polarized radiation. In resonant molecular transitions for chiral materials, the origin of the dissymmetric response derives from a combination of the electric transition dipole, and the magnetic transition dipoles; thus, the product of electric and magnetic transition dipoles has a specific handedness, giving rise to absorption cross sections for light of right or left circular polarization. Both the sign and magnitude of this dissymmetric response—measured as either a differential absorbance or phase-retardance (polarization rotation angle) for resonant or nonresonant interactions, respectively—is captured in the dissymmetry parameter, “g.”
The theory of light-matter interactions with chiral materials has been very well established;1 here, we summarize very briefly some points relevant to single-molecule chiroptical spectroscopy. For a transition between initial and final states | i >, and | f >, the Rotatory Strength Tensor, Rf←i, is defined as
where the quantities and represent the electric (E1), magnetic (M1) transition dipole moments, respectively (i.e., ), and is the electric quadrupole (E2) tensor. Equation 1 indicates that there are two contributions to the overall rotatory strength tensor: The (symmetric) tensor product , which describes electric dipole/electric quadrupole coupling, and the (antisymmetric)  scalar product that describes the electric dipole/magnetic dipole coupling. For observation along the z-axis, the rotatory strength can be expressed as
In general, we can define a rotatory strength for any arbitrary observation direction along the unit vector with orientation defined by polar angles θ and ϕ,84
giving the dissymmetry parameter (for a specific orientation angle) as,
where D( ) is the dipole strength perpendicular to the observation angle.
For isotropic samples, contributions from quadrupole interactions vanish, giving a rotatory strength that appears a product of electric and magnetic dipole matrix elements, thus reducing to the familiar dissymmetry expression,
giving rise to a range of dissymmetry values of −2 < g < +2. [made sure all notation is consistent]
In solution or unoriented solid phases, approximating the rotatory strength (hence the dissymmetry parameter) in an isotropic limit is usually quite good since all orientations are sampled. For isolated molecules in crystals56, 85–88 or polymer-supported films—the typical experimental format used for single-molecule spectroscopy—the fixed (or strongly restricted) molecular orientation implies that both E1-M1 and E1-E2 interactions contribute to the measured dissymmetry. In FDCD, or Circularly Polarized Luminescence (CPL)89–101 where an (ensemble) fluorescence signal is used as a reporter of total absorption, it was recognized early on that the isotropic limit fails when the rotational diffusion time of the molecule is long compared to the fluorescence lifetime, as is the case in viscous solvents.68, 70, 71 In such a case, contributions from electric dipole/electric quadrupole interactions can strongly modify or distort the measured dissymmetry. It is interesting to point out that work from Kahr and coworkers have shown that achiral species (even water!) with a well-defined orientation may show a dissymmetric chiroptical response.102
In addition to orientation effects, molecular solvation and local environment are known to strongly modify the chiroptical response of an isolated molecule, yet remain poorly understood.103–105 In the elegant cavity ring-down polarimetry (CRDP) experiments by Vaccaro and coworkers,52, 54, 106–113 solvation effects were observed to strongly modify the gas-phase optical rotatory dispersion of small chiral molecules. These, and recent theoretical efforts by Kongsted,114–121 Crawford,103, 122–128 Autsbach,51, 129–133 and others, have pointed to the importance of the solvation environment in perturbing the electronic structure of the molecule, thus strongly affecting the chiroptical response.
The role of extrinsic factors is also clearly evidenced in condensed-phase measurements of dissymmetry in CPL from fluorescent chiral molecules.92, 96 Venkataraman and Riehl investigated the CPL for the chiral heterohelicenes in bulk solution, and obtained dissymmetery parameters of magnitude ≈0.001. In some larger molecular systems such as oligo(meta-phenylene ethynylene)s, Moore and coworkers have measured considerably larger ensemble- averaged g-values of ≈0.2,134, 135 and recently Bunz and Neher have reported average g-values as large as 0.37in polyethylene ethynylene systems with chiral side-chains.136, 137 With a similar system, Swager and coworkers have demonstrated circular polarized luminescence and circular dichroism from chiral polyphenylene vinylene systems.138 In an interesting contrast, Tew and coworkers report that their oligo(ortho-phenylene ethynylene)s—shown by 2D-NMR to have a tertiary structure similar to Moore's meta isomers—showed a net chiroptical response of nearly zero.139–142 At the present time, it is difficult to reason the differences in the magnitude of the chiroptical responses for these chemically and structurally similar systems. Thus, there is a need to understand the factors that contribute to the “inhomogeneous broadening” (i.e., differences in local molecular environment, orientational, and configurational fluctuations.) of the molecular chiroptical response.
Over the past 15 years, advances in instrumentation and experimental methodologies in single-molecule imaging and spectroscopy have provided an astounding wealth of information on dynamical processes in condensed phases.143–147 Using single-molecule imaging techniques, researchers are now able to explore, with unprecedented clarity, details of individual molecular behavior, structure, and the coupling with local environment. The advent of ultrahigh precision time-to-digital converters, photon-counting avalanche photodiodes, and enhanced-sensitivity/high- speed charge-coupled device (CCD) imaging detectors coupled with high-numerical aperture microscopy platforms now provide access to a number of single-molecule fluorescence observables such as fluorescence lifetime,60, 144, 148–150 “blinking” behavior,151–163 spectral and spatial diffusion,143, 158, 164–192 and polarization anisotropy.149, 150, 193–201 In recent years, important new methodologies in single-molecule imaging have emerged including sub-diffraction limit spatial imaging, integration of scanning probe microscopies with fluorescence probes, multi-photon imaging, fluorescence resonant energy transfer (FRET), and molecular orientation determination by defocused emission pattern measurements.202–214 Over the past 3 years, work in our group has focused on application of single-molecule imaging techniques to chiral fluorophores to begin to address these questions. In this mini-review, we summarize various aspects of single- molecule chiroptical spectroscopy (SMCS) and some of the interesting manifestations of chiroptical behavior at the single-molecule level.
How Does Single Molecule Spectroscopy Work?
Single molecule spectroscopy (SMS) was first developed by Moerner and coworkers in the late 1980s to investigate the nature of inhomogenous broadening of absorption spectra of dopant fluorophores in cryogenic organic crystals.215, 216 Later adapted to room-temperature systems by Xie,201 and others, using a variety of elegant near- andfar-field optical approaches, SMS has been applied toa vast number of biochemical and material sciences applications. Conventional single-molecule fluorescence imaging works on the principle of detecting large numbers of (105–108) fluorescence photons from a single probe molecule; spectral discrimination of the fluorescence is readily achieved through the large (typically tens of nm) Stokes shift in emission relative to the excitation wavelength.217 This is conveniently implemented on research-grade microscope platforms using high (≥1.3) numerical aperture objectives for efficient photon collection, high-sensitivity CCD cameras (or avalanche photodiodes) for efficient photon detection, along with some means for immobilizing the probe molecules (vis. Polymer-supported thin films, or biotin-streptavidin linkage) on a transparent substrate.
Quantitative information on single-molecule behaviors can be obtained on time-scales as short as 10–100 μs. Assuming an excitation flux large enough to saturate the absorption, the photon production rate from a single molecule approaches that of the spontaneous emission rate (≈100 MHz); assuming overall photon detection efficiencies of ≈50%, one can detect a few thousand fluorescence photons on a 100-μs time interval. With current time-tagged/time-resolved photon-counting techniques coupled with pulsed-excitation, fluorescence lifetimes can accurately be determined within that nominal 100-μs time interval.218, 219 Of course, the stream of fluorescence photons arriving at the detector from a single molecule is not continuous, but is frequently interrupted through access to an excited electronic state that cannot relax to the ground state by radiative decay. This phenomenon is seen in virtually all single-molecule (or quantum dot) systems, is referred to as fluorescence intermittency, or “blinking.” The mechanism of this process and associated time-scales vary, but for organic systems, it is generally accepted that blinking originates from an excited singlet to triplet intersystem crossing. Radiative decay from the triplet state is nominally forbidden, thus the molecule may remain “dark” for extended periods of time (typically on the order of a few tens to hundreds of microseconds). In addition to short-time instabilities, irreversible photochemical bleaching ultimately limits the total number (≈108–1010)
The choice of a single-molecule probe for SMCS must therefore be based on multiple spectral and stability criteria. First and foremost, because of spectral sensitivity constraints of most commercially available detectors, the probe candidate must have robust and stable fluorescence in the visible range of the electromagnetic spectrum (650–350 nm). Ideally, molecular systems should show a minimum of blinking to avoid obscuration of information on short timescales. Additionally, the total integrated fluorescence signal that can be extracted from a single molecule is ultimately limited by the average number of excitation-emission cycles that the molecule undergoes before irreversible photobleaching. In the design of a single-molecule chiroptical probe, where we are particularly interested in discerning dissymmetries in fluorescence excitation or circularly polarized luminescence, photostability on both short and long timescales are critical to the success of the experiment.
In earlier work, one of us (Venkataraman) demonstrated the synthesis of a new class of heterohelicenes that could be easily functionalized to build stable helical structures for electronic or optical applications.220 This new synthetic scheme has enabled new fluorescent chiral probes that can be used to probe a variety of single-molecule chiroptical phenomena. The molecular frame also has a number of locations where the helicene can be attached to various molecular scaffolds, surfaces, or nanoparticles. Figure 1 shows the structures of these molecules based on triarylamines, along with bulk thin-film circular dichroism spectra. We have characterized the absorption and emission spectra of these molecules in various solutions, as well as solid films. The absorption maxima (λmax) of 1, 2, and 3 are 442, 446, and 460 nm, with extinction coefficients (ε) of 22.6 × 103, 17.7 × 103, and 14.4 × 103 cm−1 M−1, respectively. When excited at these wavelengths, 1, 2, and 3 emit at 460, 468, and 486 nm. In addition, we are able to isolate the atropisomers of 1 and 2 and verify the diastereomeric purity by proton NMR. In previous work, Riehl and Venkataraman were able to show that these molecules show a weak circular polarization in fluorescence; careful polarization analysis from solution phase bulk measurements of P1 and M1 enantiomers in solution (see Figure 3b) give (ensemble averaged) g-values of +0.0009 and −0.0011.221
Single-Molecule Studies of M2/P2 Heterohelicenes
As pointed out in the introductory section, the chiroptical signature of a single molecule is defined to a large extent by the chiral axis orientation with respect to the optical detection axis. In sample formats (e.g. solution phase) where the molecule samples all possible orientations on the timescale of the measurement, the off-diagonal elements of the Rotatory Strength tensor vanish, and the dissymmetry is defined by the familiar product of electric-dipole/magnetic dipole terms (eq. 6). For immobilized species required for single-molecule fluorescence imaging, orientational diffusion is either strongly hindered or completely nonexistent. In this case, off-diagonal elements of R (which may be orders of magnitude larger than the diagonal elements) define the orientation-dependent chiroptical dissymmetry. This can be seen by viewing R as a surface plot (see Figure 2) defined by the molecular-frame coordinates, where Z is the chiral axis of the molecule; the dissymmetry can appear either “positive” or “negative” depending on how the molecular frame is oriented with respect to the optic axis (and solid-angle of detection). In this simulation, the elements of R were determined using Time-dependent Density Functional theory at the B3LYP/6-31G(d) level of theory using a single optimized molecular geometry. Thus, it is highly desirable to have a priori (or independent) knowledge of the molecular orientation to make a meaningful comparison with theoretical calculation. Efforts aimed at determining the sensitivity of observation-angle dissymmetry with respect to level of theory and molecular conformational fluctuation is currently in progress.
Emission Pattern Imaging
Emission pattern imaging of single molecule fluorescence with a small (50–100 nm) defocusing has become apopular and well-established technique for probing transition dipole orientation in condensed phases.209, 213, 214, 222, 223 The dipole emission pattern imaging technique readily recovers transition dipole moment orientation in the two polar angles θ, an ϕ by comparison with semiclassical electrodynamics simulation. The defocused radiation (antenna) patterns associated with single-molecule fluorescence are modeled using a point dipole electromagnetic source under a defocusing aberration, and the system interfaces (sample/microscope slide/immersion oil).224, 225 For a point dipole emitter, the emission intensity has a sin2(θ) distribution about the dipole axis, and this spatial anisotropy is mapped onto a distinct spatial pattern uniquely defined by the dipole orientation angles, θ and ϕ. To extract the orientation, one matches the experimentally obtained defocused single-molecule fluorescence image with a simulated image using the polar angles (and defocusing) as fitting parameters.213
The situation for chiral fluorophores is more complicated, but still tractable using a similar semiclassical approach. We have previously shown that defocused images from single helicene molecules are qualitatively different from those of single dipole emitters, manifested as a strong distortion in the radiation pattern that breaks the bilateral symmetry normally observed in emission patterns from linear dipole systems. Chiral chromophores have, in general, a transition moment possessing a finite electric quadrupole and magnetic dipole character. For a system with a g-value of order 0.1, this “effective” dipole (due to the quadrupole) can be a significant fraction (on the order of 1%) of the size of the “ordinary” transition electric dipole. It is therefore reasonable to anticipate a radiation pattern from a chiral source that would be qualitatively different from that of a pure linear dipole molecule.226 We have carried out preliminary simulations of the defocused image obtained for the case of two coherently added electric dipoles (Kirkwood model) having a relative phase shift. This is meant to roughly approximate the image pattern obtained from a long time integration of the radiation from a source with both an electric dipole and electric quadrupole character. More detailed calculations of these images for accurate electric quadrupole and magnetic dipole source characteristics are underway, and are expected from these considerations to yield defocused images that may be amenable to orientation analysis.
Dissymmetry in Fluorescence Excitation from Single Helicene Molecules
In a typical single-molecule fluorescence excitation experiment, the helicene concentration is adjusted to give ≈20 fluorescent spots in the illumination area (20 μm diameter). Linear-polarized laser excitation (457, 440, 405 nm) was passed through a broadband quarter waveplate to produce right- or left-circularly polarized light as desired. The purity in circular polarization, checked by retroreflection of the laser beam at the sample plane on the microscope, was determined to be >98° (the limit of our experimental accuracy). Once mounted on the microscope, the helicene-doped polymer films were illuminated with ≈100–200 μW of laser power, and the fluorescence was imaged with a high- sensitivity CCD camera (Roper Scientific PhotonMax). The single-molecule fluorescence count information was analyzed and background-subtracted using standard image-processing algorithms, and the fluorescence counts (after background subtraction) are recorded as a function of frame number (referred to as intensity trajectories) which can be correlated with a time, polarization or other externally controlled parameter. The helicene intensity trajectories show several of the hallmarks of single-molecule fluorescence including on-off blinking on sub-second timescales, and discrete photobleaching with a characteristic 1/e survival time (for ≈100 μW illumination power) of ≈20 sec.
A number of control experiments were performed to verify that the signals we measured did not arise from experimental artifacts. First, we looked the variation of fluorescence intensity as a function of right/left circular polarization from dye-doped polymer nanospheres. Figure 4 shows the histogram of apparent dissymmetry values obtained from ≈ 100 different nanospheres along with the distribution of dissymmetry parameters from P2 and M2 Helicene molecules excited with 457 nm radiation. The distribution of g-parameters for the nanospheres is symmetric about g = 0, characterized by a fullwidth- half maximum of ≈ 0.3. Similar measurements on single (achiral) linear dipole molecules (DiI18) showed a distribution similar to the nanosphere sample (symmetric about g = 0), with a similar width.
For both sets of molecules, the integration time was set to 1 sec (long enough to average out much of the blinking), and 5 to 10 frames were acquired for a given laser polarization. Only molecules that survived at least 1.5 right/left excitation cycles were included in the analysis so that we could obtain (a) information on fluctuations of fluorescence intensity for a given excitation polarization, and (b) reproducibility of the mean fluorescence signals for different excitation polarizations. In virtually all the single-molecule intensity trajectories that we examined, the dissymmetry in fluorescence excitation appeared to be constant for the duration of the photochemical lifetime, although both positive and negative dissymmetry parameters were observed for a given diastereomer. The distributions of dissymmetry parameters obtained for M2 and P2 helicenes have been published previously. Briefly, these distributions (in particular with 457 nm excitation) are qualitatively different from those of the dye-doped nanosphere controls which are not symmetric about g = 0, and show a substantial width (∼90% of the probability density contained within g = ±1. Recently, Cohen and coworkers have suggested that the breadth of our distribution was related to a linear polarization artifact introduced by the dichroic filter.227 We emphasize that great care went into characterizing the polarization fidelity at the sample, and that a linear dichroism is not responsible for the signals we observe. We speculate that the major difference between Cohen's work and ours derives from the way in which the molecules are assayed. To follow single-molecule intensity trajectories over several 10s of seconds, the molecular subpopulation represented in our dissymmetry distributions necessarily reflects the highest photostability.
Dissymmetry in Fluorescence Emission from Single Helicene Molecules
In addition to fluorescence excitation measurements, we have recently initiated studies of the single- molecule chiroptical response in emission, or circularly polarized luminescence (CPL). In order to avoid any artifacts in CPL induced by transmission through the dichroic filter, we imaged the luminescence (and subsequent polarization analysis) without a dichroic filter. Figure 5 shows a schematic of the experimental configuration, where the sample was illuminated from above the microscope objective, directing the beam down through the collection objective, enabling a well-defined circular polarization in excitation to be used without a dichroic filter. Typically we used ≈100 μW of cw laser power; most of the excitation light is blocked from the detector using a high rejection ratio holographic notch filter (OD at notch wavelength >6), followed by a bandpass filter centered at 510 nm (FWHM 40 nm). To resolve the circularly polarized components in the fluorescence emsission, we used a λ/4 waveplate with fixed orientation coupled with a Wollaston prism. In this way, the single-molecule dissymmetry in emission can be determined on a frame-by-frame basis with a single multichannel detector with no moving parts in the detection optics.
Figure 6 shows some recently obtained data on dissymmetries in circular polarized emission from single (P2) helicenes using the experimental configuration described above. Single P2 molecules were isolated on a zeonex thin film, and excited with ≈100 μW of (unpolarized) 457-nm radiation. Imaging through the Wollaston prism resolves spatially the right (blue) and left (red) circularly polarized components of the emitted fluorescence. Each camera exposure thus constitutes a separate experiment in which we can determine the dissymmetry parameter on a frame-by-frame basis. The intensity trajectories for both RCP and LCP emission for three different P2 molcules is shown in Figure 6 along with the corresponding dissymmetry parameter trajectories. The error bars were established from characterization of the source noise of the sample. Interestingly, in these examples of P2 fluorescence using unpolarized excitation, we observed some single-molecule signals in which the dissymmetry parameter fluctuates between positive and negative values. In addition, we have shown that dissymmetry parameters of order unity can be seen with high signal to noise ratio in fluorescence emission. We are currently exploring the excitation polarization dependence of emission dissymmetries.
In this article, we have summarized our work on single-molecule radiation pattern imaging, fluorescence excitation, and circular-polarized emission dissymmetries using the bridged triaryl amine helicenes as prototype probe molecule. We believe there is a large molecular sample space yet to be explored that will reveal new information on the hidden heterogeneities of the chiroptical response in condensed phase. In addition, our future work will allow direct extraction of molecular frame orientation via emission pattern imaging and comparison with simulations. This, in turn, will provide the necessary components for a direct comparison with high-level theoretical calculations, and new insights into the role of local molecular environment and solvation on the single-molecule chiroptical response.