Abused Photonics Terms: Coherence

Posted on 2008-04-22. Permanent link: http://www.rp-photonics.com/spotlight_2008_04_22.html

Ref.: encyclopedia articles on coherence, coherence length and coherence time; spotlight articles of 2008-02-22 (Launching Light from a Bulb into a Single-Mode Fiber), 2007-06-24 (The Plague of a Narrow Emission Linewidth), 2006-09-22 (Coherence Length of Ultrashort Pulses)

The concept of coherence is sufficiently complex to allow for significant confusion, at least when the corresponding terms are used in a vague way. A very common problem is the distinction between temporal coherence and spatial coherence. It happens quite frequently that authors celebrate the notable feature of laser beams (or other light beams) to be coherent, without telling which kind of coherence they mean. Obviously, this doesn't convey a lot of information. After all, there are optical sources with high spatial coherence, but low temporal coherence (see e.g. the articles on supercontinuum generation or white light sources), and other sources where the opposite holds.

In the context of trains of ultrashort pulses, the temporal coherence cannot be called either high or low. As discussed in the spotlight article of 2006-09-22, the coherence function exhibits a fast decay but also "revivals" for rather long delay times. This shows that coherence cannot be quantified just with one number.

The (ab)use of the term "coherence length" in the context of nonlinear parametric interactions is particularly problematic. Here, a deterministic rather than random mismatch of the optical phases of two beams is interpreted like a loss of coherence. If we would apply that idea to a quasi-monochromatic laser beam, two points on the beam axis with a distance of only half a wavelength would be considered to see mutually incoherent light – which is of course nonsense. It would be clearly better to diagnose coherence whenever two oscillating fields exhibit a predictable phase relationship. Therefore, the mentioned use of the term "coherence length" should actually be called an abuse. I think the only argument against this is that it is frequently done.