Efficient illumination for microsecond tracking microscopy

Abstract

The possibility to observe microsecond dynamics at the sub-micron scale, opened by recent technological advances in fast camera sensors, will affect many biophysical studies based on particle tracking in optical microscopy. A main limiting factor for further development of fast video microscopy remains the illumination of the sample, which must deliver sufficient light to the camera to allow microsecond exposure times. Here we systematically compare the main illumination systems employed in holographic tracking microscopy, and we show that a superluminescent diode and a modulated laser diode perform the best in terms of image quality and acquisition speed, respectively. In particular, we show that the simple and inexpensive laser illumination enables less than 1 μs camera exposure time at high magnification on a large field of view without coherence image artifacts, together with a good hologram quality that allows nm-tracking of microscopic beads to be performed. This comparison of sources can guide in choosing the most efficient illumination system with respect to the specific application.

Figure 1. Laser illumination: effect of the current modulation.

A) Beads of 1 m diameter stuck on the glass surface are illuminated by the free running laser biased above threshold (at a current of 96 mA). The coherent noise severely degrades the image quality. B) Same field of view as in A), with bias current modulated by a sinusoidal signal of 80 mA, 2 MHz. Exposure time  =  20 s.

Figure 2. Optical spectra of the illumination sources considered and effect of laser current modulation.

A) Optical spectrum of the free running laser diode as a function of the DC bias current. The figure is divided between laser emission (above the threshold current of 60 mA) and amplified spontaneous emission (ASE) below threshold. B) Spectrum of the modulated laser as a function of the modulation frequency (laser DC current: 120 mA, AC modulation: 120 mA, square wave). Each optical spectrum is integrated over a 0.5 s time-window. The color code is the same in A and B. C) Normalized spectra of the SLD, LED, and white lamp (whose spectrum is flat in the visualized region).

Figure 3. Scattering patterns from m bead.

In each panel (A–F) we show, for each source, the projection of the 3D hologram on the axial and lateral direction in the left panel, and its Fourier representation in the right panel. The inset shows the real image obtained at axial position m. The grey and color scale is the same for all the panels. Exposure times: A) and B) 2 s; C) 3 ms; D) 70 s; E) 15 ms; F) 6 ms.

Figure 4. Quantifying fringe visibility and image noise.

A) Example showing the results of the angular average algorithm used to extract signal, fringe visibility and local image noise from the raw images. The raw image shown is obtained with the unmodulated laser at m. The signal is obtained averaging the raw image rotated in 100 steps around its center (determined with sub-pixel resolution). The image noise is obtained subtracting the signal from the raw image. B) Fringe visibility (defined in the radial intensity profile by the difference of the second maximum with the first minimum), C) noise (defined as the standard deviation of the image noise), and D) visibility-to-noise ratio are shown for the different sources at different axial z-positions.

Figure 5. Comparison of the illumination sources.

The six sources are compared in terms of attainable exposure time and relative image quality. The latter is quantified relatively to the SLD source, which gives the best image quality. In the vertical axis we plot the average distance of the images obtained with each source from the images obtained with the SLD (see Methods).

Figure 6. Comparison of the modulated laser and SLD illumination.

Image range (maximum and minimum pixel values) are shown as a function of the exposure time for the modulated laser (blue) and the SLD (red) illuminating the same object (one 1 m bead out of focus). The insets show example images obtained at the exposure time indicated; their grey levels are all fixed within the interval (0, 255) to show under-exposure and saturation. The two sources were focused to illuminate evenly the same field of view, and delivered maximum intensity (laser: 120 mW, sinusoidal modulation of 3 at 2 MHz; SLD: 5 mW).

Figure 7. Tracking at 1 s exposure.

Z trajectories of a stuck bead tracked while moving the objective in steps of 100 nm. The traces are vertically offset for clarity. Acquisition rate: 3000 fps, exposure time: 1 s.