Supporting Information. Holographic plasmonic nano-tweezers for. dynamic trapping and manipulation

Transcription

1 Supporting Information Holographic plasmonic nano-tweezers for dynamic trapping and manipulation Preston R. Huft, Joshua D. Kolbow, Jonathan T. Thweatt, and Nathan C. Lindquist * Physics Department, Bethel University, St Paul, MN 55112, USA * n-lindquist@bethel.edu, Equal Contribution Substrate Fabrication. The silver Bull s Eye structures were fabricated with the dimensions that showed the optimal plasmon intensities in COMSOL. The process is shown in Figure S1. A homemade electron beam lithography (EBL) system built around a Hitachi SU1510 scanning electron microscope and a LabVIEW TM data acquisition system was used to expose a 150 nm thick layer of polymethylmethacrylate (PMMA) (Microchem, MA, USA), diluted to 3% percent in Anisol. The film was spin coated at 4000 rpm for 60 seconds on a 1 cm 2 silicon chip substrate. The PMMA coated chip was then baked for 60 seconds on a hotplate at 70 C. A typical dose of 5000 µc cm -2 was used to overexpose the PMMA, leaving durable PMMA Bull s Eyes of inner ring radii ranging between 300 nm and 700 nm and ring periods ranging between 450 nm and 490 nm spaced 10 µm apart on the silicon. After developing the chip in MIBK for 30 seconds and then acetone for 2 seconds, the chip was rinsed in isopropanol and dried with a filtered

2 stream of air. A 300 nm layer of Ag was then deposited onto the substrate via thermal deposition at a rate of ~40 nm per minute. The silvered side of the chip was adhered to a glass slide with a UV curing optical adhesive (Norland, NJ, USA). Due to the poor adhesion of the silver to the silicon, the silver can be peeled from the template in a process known as template stripping, 1,2 creating a smooth patterned metal film. After curing, the glass chip was separated from the silicon chip with a razor blade, stripping the silver from the PMMA pattern and leaving behind an inverse imprint of the Bull s Eyes. The silicon/pmma template could then be reused multiple times. Figure S1: Fabrication procedure. (a) A film of PMMA was spun onto a clean silicon wafer and patterned via electron beam lithography. The PMMA was over-exposed to create a durable and reusable template. (b) A thick silver film was then deposited onto the template and (c) peeled off with adhesive and a glass microscope slide, exposing the Bull s Eye pattern. (d) SEM images of the PMMA template and (e) the resulting silver Bull s Eye pattern. (f) Zoomed-out SEM image showing an array of nanostructures. Panels (e) and (f) show the same SEM images as in figures 1c,d of the manuscript.

3 Trapping Experiments. Plasmonic tweezing experiments were done with 200 nm diameter fluorescent polystyrene microspheres (Bang Labs, IN, USA) suspended in water. The beads were chosen to fluoresce with a 532 nm laser (Laser Quantum, Stockport, UK). The stock solution was diluted 1:300 with water. Roughly 4 µl of the diluted solution was dispensed onto a fresh silver substrate and covered with a glass coverslip sealed with a clear acrylic nail polish. The substrate was placed into an inverted microscope setup (Nikon, Japan) with a 100x oil immersion objective (NA = 1.25). The 532 nm illuminating laser had a power of ~10 mw and the trapping laser had a total measured power after the SLM of approximately ~100 mw. Since a 532 nm laser was used for illuminating the beads and a 660 nm laser was used to illuminate the Bull s Eye patterns, a band pass filter (Semrock, NY, USA) was used to visualize only the fluorescence from the beads between 532 nm and 660 nm. Figures S2, S3, and S4 show frames from short sample videos of the bead motion and trapping experiments. Figure S2: Frame from a sample trapping experiment. In this case, the bead is trapped in the center of the Bull s Eye without any holographic motion. Please see the accompanying video.

4 Figures S3: Frame from a sample trapping experiment. In this case, the bead is trapped in the center of the Bull s Eye and undergoes sinusoidal motion. Notice that the motion is sub-diffraction-limited and is nearly imperceptible in the compressed video. The subtle bead motion becomes apparent, however, with centroid fitting and particle localization. Please see the accompanying video. Figures S4: Frame from a sample trapping experiment. In this case, the trapped bead again undergoes sinusoidal motion. As before, the subtle bead motion becomes apparent with centroid fitting and particle localization. Please see the accompanying video. Particle Localization. The beads were imaged with an electron multiplied (EM) CCD camera (Andor, Belfast, UK) with a typical frame exposure time of 50 ms. A movie was recorded of an experiment and then exported to an open-source tracking software called

5 rapidstorm. 3 While the software was originally intended for Stochastic Optical Reconstruction Microscopy (STORM) data analysis, it can also be used for particle localization. Briefly, the rapidstorm software loads the movie files and analyzes them frame-by-frame for bright spots above a certain threshold. These bright spots, themselves the diffraction-limited point-spread-function of the imaging system, can be fit with a two-dimensional Gaussian function and the centroids can be localized to within ~5 nm. The rapidstorm software then output a file that was further analyzed with custom LabVIEW TM software for visualization, cropping, and tracking. Data was also fit, analyzed, and plotted in MATLAB TM. Several experiments illuminated the Bull s Eyes in a brightfield mode with a white LED (Thorlabs, NJ, USA). These experiments were then able to overlay the super-resolved bead locations onto the microscope images. Hologram Generation. The amplitude of the hologram was imaged directly on the CCD from the microscope whereas the phase of the hologram was imaged with a wavefront meter (Thorlabs). To manipulate the focusing of the plasmon beams in the center of the silver Bull s Eye patterns, a hologram needed to be generated for the SLM (Hamamatsu) that would both direct the amplitude of the laser beam as well as its phase according to equation (1) in the manuscript. Unfortunately, the SLM is capable of only adjusting the phase of the incident beam. Typical iterative techniques based on the work of Gerchberg and Saxton 4 can efficiently determine the phase pattern to be displayed on the SLM in order to generate an arbitrary amplitude pattern in the focal plane of the microscope objective. However, the phase is left to vary in the focal plane since the amplitude is fixed. To overcome this limitation, in our case the desired hologram was

6 modeled as a ring of point sources, each with an arbitrary phase. Similar iterative algorithms are sometimes used to create 2D or 3D arrays of focused points. 5-7 As the iterative algorithm proceeded, the amplitude of the light in the focal plane was fixed as the ring of point sources and the phase was allowed to vary as usual except for directly over the point sources themselves, which were fixed by equation (1) given in the main text. The point sources were set close enough to merge into a single ring, now with an arbitrary phase distribution around the circumference. Numerical Simulations. Two-dimensional simulations were performed in MATLAB TM by using equations (1) and (2). The field intensity produced by a ring of point sources given by equation (2) was plotted for the various phase profiles given in equation (1). Three-dimensional simulations were performed in COMSOL TM Multiphysics (COMSOL, Inc., MA, USA). The mesh size was variable from 10 nm near the center and edges to 50 nm in the bulk. Material properties were used from default COMSOL TM libraries, originally taken from published data. 8 The illumination profile consisted of a ring of 12 focused Gaussian beams, each with unity amplitude (i.e., E 0 = 1 V/m), the superposition of which gave a single ring roughly identical in size to the experimentally produced holograms. The phase of each individually focused Gaussian beam was directly altered according to equation (1). When comparing forces produced by the plasmonic traps to those of a conventional optical tweezers, the incoming hologram illumination had the same total power as a single diffraction limited Gaussian beam. References. 1. Nagpal, P.; Lindquist, N. C.; Oh, S. H.; Norris, D. J. Science 2009, 325,

7 2. Lindquist, N. C.; Johnson, T. W.; Norris, D. J.; Oh, S. H. Nano Lett. 2011, 11, Wolter, S.; Löschberger, A.; Holm, T.; Aufmkolk, S.; Dabauvalle, M.-C.; van de Linde, S.; Sauer, M. Nat. Methods 2012, 9, Gerchberg, R. W.; Saxton, W. O. Optik 1972, 35, Dufresne, E. R.; Spalding, G. C.; Dearing, M. T.; Sheets, S. A.; Grier, D. G. Rev. Sci. Instrum. 2001, 72, Sinclair, G.; Leach, J.; Jordan, P.; Gibson, G.; Yao, E.; Laczik, Z. J.; Padgett, M. J.; Courtial, J. Opt. Exp. 2004, 12, Whyte, G.; Courtial, J. New Journal of Physics 2005, 7, Johnson, P. B.; Christy, R. W. Phys. Rev. B 1972, 6,