August 2012
Spotlight Summary by Richard Bowman
Dynamical hologram generation for high speed optical trapping of smart droplet microtools
At first glance, it would seem that the closest computer games get to the physics laboratory is distracting grad students from their research. However, the technology that enables us to fight ever more realistic monsters has huge potential when applied to scientific problems. The challenge of rendering realistic 3D graphics is one that requires an enormous amount of processing power, and thanks to the demand for smooth, high quality games, it is now possible to pick up a pocket supercomputer with hundreds of processing cores for a couple of hundred pounds, in the form of a consumer graphics card.
Graphics cards are finding their way into a number of scientific endeavours, from analysis of x-ray diffraction data to complicated simulations. One such area is computer-generated holography, where diffraction patterns are calculated to steer and re-shape laser beams, in experiments ranging from quantum optics to optical tweezers. Often the patterns are applied using liquid crystal micro displays, allowing the hologram to be dynamically updated. This, however, requires fast calculation of the holograms: a problem to which the graphics processor is uniquely well suited. Graphics-card calculation is particularly appropriate as these micro-displays are usually connected to the computer as secondary monitors.
The optical tweezers community realised the potential of this approach early---the first example being the work of Reicherter and Haist at the University of Stuttgart in 2005. As the power of graphics cards has increased and toolkits such as CUDA and OpenCL have made them more accessible, a number of implementations have been made freely available, for example from the groups of Goksör (Gothenburg), Di Leonardo (Rome) and Padgett (Glasgow).
Lanigan and co-workers have used this technology not with the conventional nematic liquid crystal micro display, but with a ferroelectric display. This has a much faster response time (tens of microseconds rather than tens of milliseconds) but displays only binary holograms. However, displaying these binary holograms in rapid succession allows them to multiplex up to 24 beams to create a number of independent optical traps---without using the more complicated multi-trap patterns found in holographic optical tweezers. The combination of ferroelectric micro displays and GPU acceleration is a good one, and it has the potential to be very fast. The downside of binary ferroelectric modulators is their efficiency: while a microdisplay optimised for the laser wavelength used would help, the theoretical maximum for a binary grating is less than half (as first and minus first orders have identical power).
As with conventional holographic tweezers, to make this system user friendly the graphics card processing was coupled to user-friendly LabVIEW interface. The aim here is to make the system more accessible to non-specialists, indeed the interdisciplinary author list of this paper is evidence of the system escaping the confines of an optics laboratory.
The motivation for developing this tool is the manipulation of "smart droplet microtools", surfactant-coated beads that allow material to be injected and extracted from live cells, without destroying the cells. In the paper, this concept is proved by making contact with both synthetic vesicles and cancer cells: in both cases, contact is verified by dye being transferred between the object being probed and the smart droplet. This technique might allow, in the future, the same cell to be analysed at different times without interrupting its development---and it is certainly another step towards making micro-organisms as accessible as those on larger length scales.
You must log in to add comments.
Graphics cards are finding their way into a number of scientific endeavours, from analysis of x-ray diffraction data to complicated simulations. One such area is computer-generated holography, where diffraction patterns are calculated to steer and re-shape laser beams, in experiments ranging from quantum optics to optical tweezers. Often the patterns are applied using liquid crystal micro displays, allowing the hologram to be dynamically updated. This, however, requires fast calculation of the holograms: a problem to which the graphics processor is uniquely well suited. Graphics-card calculation is particularly appropriate as these micro-displays are usually connected to the computer as secondary monitors.
The optical tweezers community realised the potential of this approach early---the first example being the work of Reicherter and Haist at the University of Stuttgart in 2005. As the power of graphics cards has increased and toolkits such as CUDA and OpenCL have made them more accessible, a number of implementations have been made freely available, for example from the groups of Goksör (Gothenburg), Di Leonardo (Rome) and Padgett (Glasgow).
Lanigan and co-workers have used this technology not with the conventional nematic liquid crystal micro display, but with a ferroelectric display. This has a much faster response time (tens of microseconds rather than tens of milliseconds) but displays only binary holograms. However, displaying these binary holograms in rapid succession allows them to multiplex up to 24 beams to create a number of independent optical traps---without using the more complicated multi-trap patterns found in holographic optical tweezers. The combination of ferroelectric micro displays and GPU acceleration is a good one, and it has the potential to be very fast. The downside of binary ferroelectric modulators is their efficiency: while a microdisplay optimised for the laser wavelength used would help, the theoretical maximum for a binary grating is less than half (as first and minus first orders have identical power).
As with conventional holographic tweezers, to make this system user friendly the graphics card processing was coupled to user-friendly LabVIEW interface. The aim here is to make the system more accessible to non-specialists, indeed the interdisciplinary author list of this paper is evidence of the system escaping the confines of an optics laboratory.
The motivation for developing this tool is the manipulation of "smart droplet microtools", surfactant-coated beads that allow material to be injected and extracted from live cells, without destroying the cells. In the paper, this concept is proved by making contact with both synthetic vesicles and cancer cells: in both cases, contact is verified by dye being transferred between the object being probed and the smart droplet. This technique might allow, in the future, the same cell to be analysed at different times without interrupting its development---and it is certainly another step towards making micro-organisms as accessible as those on larger length scales.
Add Comment
You must log in to add comments.
Article Information
Dynamical hologram generation for high speed optical trapping of smart droplet microtools
P. M. P. Lanigan, I. Munro, E. J. Grace, D. R. Casey, J. Phillips, D. R. Klug, O. Ces, and M. A. A. Neil
Biomed. Opt. Express 3(7) 1609-1619 (2012) View: Abstract | HTML | PDF