STED microscopy resolves nanoparticle assemblies

We demonstrate the ability of stimulated emission depletion (STED) microscopy, a far-field fluorescence imaging technique with diffraction-unlimited resolution, to reveal the spatial order of fluorescent nanoparticles. Unlike its confocal counterpart, here STED microscopy resolves the arrangements of densely packed 40 nm beads, supramolecular aggregates in a cell membrane, and colloidal nanoparticles. Both raw and linearly deconvolved data disclose unprecedented details of both biological and non-biological nanopatterns. Abbe’s discovery of the diffraction barrier has lead to the popular notion that a far-field light microscope cannot resolve spatial structures that are smaller than about half of the wavelength of light [1]. In consequence, this insight triggered the invention of electron, x-ray, and scanning nearfield optical microscopy (SNOM) [2, 3]. For many years after its invention, the latter seemed to be the only option to overcome the diffraction barrier in the optical domain [4]. To this end, SNOM employs either a tiny aperture or a nanometre-sized tip to confine the light-object interactions to subwavelength dimensions. Although it can accommodate most optical contrast modes, SNOM suffers from a number of drawbacks. While aperture-based SNOM entails feeble signals, its (plasmon and nonlinearly enhanced) tip counterpart requires a tight control of the tip-sample distance for artefact-free operation [5, 6]. In any case, SNOM is undeniably confined to the imaging of surfaces. For many applications in biology and chemistry, it is certainly more desirable to break the diffraction barrier in the far-field. This formidable problem has been addressed several times in the 20th century, but effective methods have emerged only recently [7, 8]. Stimulated emission depletion (STED) microscopy overcomes the resolution-limiting role of diffraction for fluorescent samples. The idea underlying this concept can also be expanded to other contrast modes, provided they involve a reversible saturable optical transition [9, 10]. Nonetheless, being unrivalled in specificity and sensitivity, fluorescence is the most important non-invasive imaging 1 Author to whom any correspondence should be addressed. New Journal of Physics 8 (2006) 106 PII: S1367-2630(06)24131-4 1367-2630/06/010106+8$30.00 © IOP Publishing Ltd and Deutsche Physikalische Gesellschaft 2 Institute of Physics DEUTSCHE PHYSIKALISCHE GESELLSCHAFT mode at the micron and submicron scale. Here, we demonstrate the power of STED microscopy to image the arrangement of densely packed fluorescent nanoparticles. The key element of STED microscopy is a saturated depletion of the fluorescent state of the marker molecule [7, 11], whereby the depletion is accomplished with a focal intensity distribution featuring a local zero, e.g. a doughnut. The typical optical cross-sections and spontaneous relaxation times of a fluorophore dictate that STED is implemented optimally with picosecond excitation pulses that are swiftly followed by red-shifted 100–200 ps pulses for de-excitation [7, 11, 12]. Under these conditions, the population of a typical fluorescent marker decreases nearly exponentially with the energy of the stimulating pulse [13]. Oversaturating STED with sufficiently bright pulses confines the remaining fluorescence to the local zero. The net result is a subdiffraction-sized fluorescent focal spot [12]. Scanning this spot across the sample and registering the fluorescence renders subdiffraction images. This spot is tantamount to the effective point-spread-function (PSF) of the STED microscope, hEff( r), giving the probability that a fluorescent molecule located at r contributes to the measured signal. With hExc( r) and hSTED( r) describing the excitation PSF and STED–PSF, respectively, we obtain hEff( r) = hExc( r) exp[−σ hSTED( r) max]. (1) The PSFs are normalized to unity and σ is the cross-section for stimulated emission at the used wavelength. The flux max gives the number of photons per area per pulse found at the local maximum bordering the zero, i.e. at the doughnut crest. The presence of a local ‘zero’, hSTED(0) = 0, implies that the fluorescence emission at r = 0 is increasingly suppressed with increasing max, whereas that from r = 0 remains unaffected. Thus, the STED–PSF confines the effective PSF to a region that is much narrower than the excitation spot. In fact, equation (1) makes evident that hEff( r) can be narrowed down to a profile of arbitrarily small full-width-halfmaximum (FWHM), although both the FWHM of hExc( r) and hSTED( r) are limited by diffraction. Calculations show that the FWHM of a STED-microscope [9, 10, 14], r, follows r ∼= λ 2n sin α √ 1 + ζ , (2) with ζ ≡ maxσ denoting the ‘saturation factor’ of the depletion. The expression n sin α gives the numerical aperture (NA) at which the focal doughnut is generated. In our experiments, we employed a pulsed laser diode (Picoquant, Berlin, Germany) for excitation, emitting∼100 ps pulses at λExc = 470 nm.The pulses were synchronized with yellow STED pulses (λ = 580–620 nm) from a frequency-doubled optical parametric oscillator (OPO) system operating at 80 MHz. Featuring a spectral width of ∼6 nm, the ∼250 fs OPO pulses were downchirped in a single-mode optical fibre to ∼200 ps duration. The planar wavefront leaving the fibre was subjected to a helical phase delay P(φ) = exp(iφ), with 0 < φ < 2π being an angle centred around the optic axis. The phase delay was realized using a spatial phase modulator (Hamamatsu, Hamamatsu City, Japan) located optically conjugate to the entrance pupil of the lens. The objective lens (NA = 1.4 oil immersion, HCX PL APO, 100×, by Leica, Mannheim, Germany) transformed this wavefront into a doughnut in the focal plane. Figure 1 depicts the excitation spot (a), measured with a scattering 80 nm gold bead, together with the helical phase ramp (b) and the doughnut (c) for λ = 585 nm; the latter serves as the STED–PSF hSTED( r). Figure 1(e) shows the intensity profile hSTED(y = 0), disclosing a central minimum of 160 nm FWHM and hSTED(0) = 0.015. Circular polarization of the pulses ensured New Journal of Physics 8 (2006) 106 (http://www.njp.org/) 3 Institute of Physics DEUTSCHE PHYSIKALISCHE GESELLSCHAFT