Photons - particles of light - are not like the particles we encounter in everyday life, like say, a marble, which is localized in space. Instead, photons are distributed across space. The way in which a photon is distributed, which we can loosely think of as its shape, is characterized by its so-called wave-packet or wave-function. There are many different ways in which photons can be produced experimentally, and the wave-packets of the produced photons varies enormously between these different schemes.

In my previous paper ``Quantum gate characterization in an extended Hilbert space" (discussed in my previous post), we examined the effects of mode-matching upon the operation of linear optics quantum gates. During this research, one observation we made was that the significance of mode-mismatch in a circuit depends very heavily on the wave-packets of the photons used in the experiment. Naturally, every experimentalist wants to minimize the effects of mode-mismatch since it destroys the operation of their gates. This motivates the question ``what type of wave-packets minimize the effects of mode-mismatch?". In this paper - joint work with Tim Ralph and Michael Nielsen - (pre-print available at quant-ph/0505139) we address this question and establish two criteria which minimize such effects:

• Photons should be Gaussian in shape (i.e. they should look like a Bell-curve)
• Photons should be as broad as possible (i.e. the Gaussian curve should be stetched out as much as possible)

From a practical perspective this means that experimentalists should try and employ photon sources which produce photons of this type. In principle this should improve the operation of experimental quantum gates.