With electronic detection hardware becoming more and more sensitive with improvements in technology, it is now possible to study the behavior of single molecules. In the past, spectroscopists studied ensembles of molecules. They would shine light on a specific section of their sample chamber, exciting a large number of molecules simultaneously. Then, the observed fluorescence, absorption, circular dichroism, etc., would be a single value representing the average property of all the molecules that were selected. Although a large amount of information certainly can be obtained that way, there are certain questions you cannot ask simply by looking at averages ...

Lets say you know a molecule is in two states, state A, and state B. You have a spectroscopic technique that can differentiate between A and B. You measure your sample and find that its 50% A and 50% B. Fine, thats an important fact. But what you don't know is whether a particular molecule is jumping back and forth between A and B and spends 50% of its time in one or the other, or whether there are molecules locked into one state permanently, such that a molecule that is A stays in A, and B stays in B. If you could follow a single molecule for an extended period of time, you could then say whether it is stuck in one state or whether it is oscillating between two.

Most instruments that are set up to do single molecule fluorescence measurements consist of a confocal microscope where the light source is a laser, instead of a lamp. This allows you to pick a single molecule in the field of the microscope and track it through several excitation-relaxation cycles and see if it moves, whether it changes properties, etc. This has been used to study protein folding, polymerase motion along DNA, catalytic enzyme activity, and a number of other biological phenomena.

There are a number of drawbacks to single molecule work presently. One is that fluorophores tend to bleach after a while. You cant follow one molecule forever. Eventually the fluorescent reporter dies and the experiment is over. Certain dyes are less susceptible to photobleaching than others. Most experiments currently use rhodamine or fluorescein because they have a high quantum yield and are relatively stable. Additionally, because you are collecting data on one molecule, rather than an ensemble of billions or more molecules, the noise is much higher. The longer you collect, the more sure you can be of extracting useful data. Single molecule work is going on in other fields besides spectroscopy, including people who use laser traps to measure mechanical forces on specific molecules.