Danish scientists have harnessed the power of light beams to manipulate 3D arrangements of particles at the click of a mouse button. Rob van den Berg reports.

FromOpto & Laser EuropeMay 2005

Architects of the microscopic world
A good description of Jesper Glückstad would be an architect of the microscopic world. Over the past two years, he and his co-workers Peter John Rodrigo, Vincent Daria and Ivan Perch-Nielsen have been developing machinery to manipulate micrometre-sized particles into 3D structures at the click of a mouse button. What's more, the research team, from the Risø National Laboratory in Roskilde, Denmark, do this using nothing but focused beams of light.

The team's apparatus opens up promising new possibilities in the field of materials science and biomedical research - assembling photonic crystals, seeding crystal growth or sorting cell colonies are just a few examples. When OLE spoke to Glückstad, he was about to leave for the US to meet with scientists from the National Institute of Health, who are keen to apply his technique to cancer and cell research.

Although the concept of the optical trapping of particles has been around for some time (Alan Ashkin and his co-workers from Bell Labs in the US demonstrated optical tweezers in the mid-eighties), Glückstad's approach allows many particles to be manipulated simultaneously in realtime.

Instead of using a single laser beam to trap a particle, which has the disadvantage that it can escape if the beam is not tightly focused, the Danish team uses a pair of laser beams to move each particle. "A particle can be trapped between two counter-propagating beams that have their waists slightly separated along the optical axis," explained Glückstad.

This dual-beam approach, which Glückstad fondly refers to as an "optical elevator", has several benefits over conventional optical tweezers and traps. "Optical tweezers require objective lenses with a very high numerical aperture (NA) to generate a sufficiently strong gradient force. If such a lens is used away from the design plane, the focus is smeared out and the particle is lost," said Glückstad.

"Also, when working with live cells, it is better not to focus too tightly in order not to damage the organelles inside the cell," he continued. "We can work with much simpler optics and can also move them [the particles] over a much larger depth, even out of view of the microscope lens."

By changing the relative powers of two opposing, orthogonally polarized laser beams, Glückstad and his colleagues can vary the axial position of a trapped particle by tens of micrometres. And, thanks to a highly efficient phase-and polarization encoding scheme, they can generate many simultaneous elevators to trap and manipulate up to 80 particles with less than 1 W of near-infrared laser light.

The power of polarization

The apparatus works as follows: an expanded TEM00 mode from a Ti:sapphire laser at 830 nm illuminates a spatial light modulator (SLM) and becomes encoded with a 2D phase-distribution. "This phase distribution can be represented by a spatial pattern of up to 256 levels on a computer monitor, where each grey level represents a different phase-delay between 0 and 2 pi," explained Glückstad.

"This is then projected onto the surface of a spatial polarization monitor (SPM), which acts as a variable waveplate," he continued, "and endows each beam with varying components of s-polarized and p-polarized light." Finally, a polarizing beamsplitter separates the orthogonal (s and p) components of the beams before they are directed to the sample via microscope objectives.

The beauty of the process is that it all occurs rapidly and with almost no optical loss. The system is aligned in such a way that the s and p beams are focused along a common optical axis, but have a slight axial separation in order to perform the trapping. Controlling the relative intensities of the two beams adjusts the position at which a particle finds its equilibrium. In principle, this could be done with any two counter-propagating beams, but, as Glückstad explained: "We use orthogonal polarizations because the beams should not show any interference. We do not want them to see each other."