A Brief Overview of Optical Trapping
Optical trapping is a powerful technique employed in the rapidly expanding field of single-molecule biophysics, and, in particular, single molecule manipulation. Minuscule forces (on the order of piconewtons) may be exerted on biological materials, and displacements of less than a nanometer can be measured using optical tweezers. The fundamentals of this technology are briefly outlined below.
An optical trap consists of a laser beam (single mode, typically TEM00) focused down to a diffraction limited spot using a lens of high numerical aperture, usually a microscope objective. Near-infrared lasers (a popular variety is an Nd-YAG laser, with wavelength of 1064 nm) are utilized to reduce damage to the biological substrate and minimize local heating of the sample buffer. The tight focusing creates a region of high intensity which falls off with a very steep gradient over several hundred nanometers. A dielectric particle near the beam focus will become trapped in all three dimensions due to this Gaussian intensity profile. (A simple way of thinking about gradient forces is to consider the particle drawn to the region in the beam of greatest intensity.)

Illustration of a trapped microsphere. The forces due to the laser light are balanced in all three dimensions and the particle remains stationary at the stable trapping center, save for undergoing Brownian Motion.
If a molecule is attached to the particle and exerts a force on it, the particle will be pulled from the trap center. A restoring force will be exerted by the trap, in a way very much analogous to a Hookeian spring.

Cartoon depicting the analogy between a particle displaced from the stable optical trapping center and a Hookeian spring.
This change in momentum of the bead is equal and opposite to the change in momentum of the laser light; in other words, the light is deflected by a small angle. This angular deflection is detected downstream with a Position Sensitive Detector.

Light being deflected by the bead; as the bead is pulled off-center, the trap exerts a restoring force on it. The resulting change in momentum manifests itself as a deflected beam; downstream, this deflection can be detected with great accuracy.
The trapping potential near the stable center is nearly harmonic; hence, calibration of the spring constant is a straightforward matter, even in the axial dimension. Although there are several methods by which one can accurately calibrate the trap, the simplest is to monitor the particle's Brownian fluctuations and employ the equipartition theorem directly to extract the spring constant. This method presumes a precise position calibration of the setup, which is straightforward using a stuck particle translated through the trap with a piezo-electric stage. For more detail about these techniques, see the reference links above.
For more information, check out the Wikipedia entry on Optical Trapping.

