They fired a laser through a lens, which then went to a mirror that bounced the beam back. The reflected beam interfered with the incoming beam. Meanwhile, the scientists suspended polystyrene spheres in water directly in the path of the laser beams. The beams held the spheres in place vertically, and any pulling or pushing force moved the spheres to the left or to the right.
The effect worked because the particles were small enough to scatter photons of light. The photons got pushed ahead of the particles and since the photons had momentum, the particles got a slight kick backwards.
Grier, who has himself experimented with tractor beams (http://physics.nyu.edu/grierlab/conveyor7c/) , noted that this is the first time anyone has shown that the pulling force was due to the scattering and that you could use it to manipulate objects this way. "It's just a beautifully clear demonstration," he said.
By changing the polarization of the light beam, the scientists could move objects in any direction. Zemanek's team found something else, though: the particles sorted themselves by size, with larger ones going to the left and smaller ones going to the right.
The sorting happens because how the light scattered by the particle depends on its size. At a given wavelength small particles will scatter light more than large ones. The light scatters in several directions, and at certain specific particle sizes, Zemanek said the tiny spheres acted like lenses, with some of the light making a kind of focus on one side.
The areas of focused light created a zone where the light's potential energy was at a minimum. This was a "well" where nearby particles would tend to fall. "It's like the cups for eggs in a carton." Zemanek said. A sphere would fall into the well, and as one was pulled along, it drags another behind it. Meanwhile other spheres fell in line, provided they were the same size.
Zemanek noted that if one were assembling a tiny robot, this would be a good way to move the parts around to where they are needed. Since the movement is done with light, it doesn't mater if the particles are metallic or not, or if they are susceptible to electric fields.
Grier said that the future work should focus on extending the range of the tractor beams to more than the micrometer scales. He noted that right now it seems that the object to be moved has to be smaller than the diameter of the beam. "Whether or not that's right, or how stringent a limit it sets, remains to be seen," he said.
The work was published in the Jan. 20 issue of Nature Photonics.