High-Resolution Cryogenic Spectroscopy of Single Molecules in Nanoprinted Crystals

Abstract

We perform laser spectroscopy at liquid helium temperatures (T = 2 K) to investigate single dibenzoterrylene (DBT) molecules doped in anthracene crystals of nanoscopic height fabricated by electrohydrodynamic dripping. Using high-resolution fluorescence excitation spectroscopy, we show that zero-phonon lines of single molecules in printed nanocrystals are nearly as narrow as the Fourier-limited transitions observed for the same guest-host system in the bulk. Moreover, the spectral instabilities are comparable to or less than one line width. By recording super-resolution images of DBT molecules and varying the polarization of the excitation beam, we determine the dimensions of the printed crystals and the orientation of the crystals' axes. Electrohydrodynamic printing of organic nano- and microcrystals is of interest for a series of applications, where controlled positioning of quantum emitters with narrow optical transitions is desirable.

Keywords: nanocrystal; nanoprinting; quantum emitter; single molecule; single-photon source; spectroscopy.

Figure 1

Electrohydrodynamic nanodripping for fabrication of organic crystals. (a) Sketch of the setup utilized for printing organic crystals. The setup integrates a low-resolution side channel and a high-resolution iSCAT channel for real-time imaging of printed structures during fabrication. (b) An example of the side-view for facilitating the coarse alignment of the micropipette (tip) above the sample substrate. The dashed rectangle marks the region of the printed array of nanocrystals (NCs). (c) An example of the iSCAT image of a printed array of Ac NCs. (d) Atomic force microscope image of printed AC microcrystals. (e) Atomic force microscope image of a printed AC NC. Color bars indicate the measured height.

Figure 2

Spatial and spectral distribution of DBT molecules in printed organic crystals. (a) Fluorescence excitation spectrum from 1 μm² of a printed DBT:Ac microcrystal recorded at 2 K. The inset shows an example of a high-resolution spectrum of a single-molecule 00ZPL. The gaps in the scanning range are due to interruptions in the laser cavity lock. The two main insertion sites of DBT in Ac are marked with red and turquoise regions. (b) Spectra of two molecules located within a diffraction-limited spot. Red curves show the data points used to localize individual molecules. (c) Probability density function (PDF) of lateral positions for the two DBT molecules in (b). The location of each emitter is represented by a Gaussian spot. (d) Super-resolution image of 589 molecules between 378.5 THz and 382 THz, i.e., the red region in (a). The embedded emitters mark the crystal’s shape and boundaries. (e) Super-resolution image of 11197 molecules between 382 THz and 385 THz, i.e., the turquoise region in (a). (f) Distribution of frequency detuning and distance between molecular pairs for emitters with less than 150 nm lateral separation.

Figure 3

Spectral properties of printed DBT:Ac nanocrystals. (a) Distribution of emitters in an array of printed nanocrystals. Each DBT molecule is shown as a point. Color code shows the resonance frequency. The scale bar shows 4 μm. (b) Fluorescence of DBT molecules in 100 NCs as a function of excitation frequency. (c) Line width distribution of 150 DBT molecules embedded in the printed Ac NCs.

Figure 4

Saturation behavior and spectral stability of DBT molecules in printed nanocrystals. (a) Detected fluorescence as a function of laser detuning from the resonance of a single molecule, recorded at excitation powers P = (0.17, 1.6, 6.6, 95) nW. Solid lines show Lorentzian fits. (b) Saturation of the fluorescence count rate and power broadening at high pump powers. Lines show theoretical fits with saturation power P_sat = 3.6 nW and maximum fluorescence rate of F_∞ = 257 kcps and an intrinsic line width of γ₀ = 41 MHz. (c) Spectrum of a single molecule recorded at low excitation powers over 1 h. (d) Histogram of the Lorentzian fit centers for 1600 scans in (c). The standard deviation of the measured values is σ_f = 26 MHz. (e) Spectra of 4 molecules recorded over 20 min. (f) Resonance frequency fluctuations σ_f due to spectral diffusion for 15 DBT molecules in different NCs. The red line indicates the median at 48 MHz. The box represents the range between 25 and 75 percentile of the population.

Figure 5

(a) Arrangement for determining the transition dipole moment orientation. A quarter-wave plate (QWP) and a half-wave plate (HWP) are used to adjust the excitation polarization θ_exc. ϕ indicates the angle of the transition dipole. (b) Fluorescence signal of a single molecule for different θ_exc. The dashed line is a fit to a cos² function. (c) i: Simultaneous mapping of ϕ by frequency and polarization sweeps using wide-field illumination of an array of printed nanocrystals. Each molecule is represented by a short line, centered at its spatial coordinates and aligned parallel to its transition dipole angle ϕ. The color of the lines encodes the corresponding resonance frequency. The scale bar shows 4 μm. ii-iv: Close-ups for three exemplary NCs. The underlying image illustrates the PDF of the lateral position of the emitter, similar to Figure 2e. The scale bars are 200 nm. (d) Distribution of the angles between transition dipole moments ϕ in one NC. Dashed line shows a fit corresponding to two possible alignments of molecules, parallel or separated by 29° (see Supporting Information). (e) Spatial distribution of 261 DBT molecules with respect to the center of their corresponding NCs. The most likely orientation of DBT is set to the horizontal axis and can be associated with the b-axis of Ac.