Is Initial Coherence a Problem in Experiments? In experiments, or optical setups, is the initial coherence of lasers something to be concerned about? You can calculate interference patterns, but those generally rely on the phase shift between the two sources being zero, right? In a laboratory, how do you know two lasers will be in phase?
 A: In general this is not a problem if you are careful, but if you are not then it can certainly ruin your experiment and wash out any possible interference. The problem is not quite ensuring that the lasers be in phase, since at any given time the two sources will have some definite phase relationship with each other: the problem is making sure that they stay in phase.
If the lasers were perfect sinusoidal sources, it would not be a problem. You might have a harder time controlling the relative phase between the two, but you would always observe interference between them.
However, real lasers are not perfect and they have some amount of phase slippage: because they are not perfectly sinusoidal sources, if you wait long enough, the phase will have slipped a little bit, either forwards or backwards. This slippage is noise and its magnitude and direction is essentially random; it is caused among other things by a small amount of spontaneous emission on top of the stimulated (lasing) emission in the gain medium. Since this slippage is always there, the only thing you can do is quantify it, and the relevance measure is the coherence time of the laser. This is the time $\tau$ you must wait for the laser to bear, at time $t+\tau$, no definite phase relationship to its state at time $t$.
For normal experiments the easiest thing is to use a single laser and a beam splitter, so that your two sources are effectively phase-locked. The only thing you need to ensure is that the delays inside your interferometer are not too great, so that you don't end up comparing the state of your laser between a time $t$ and a time later than $t+\tau$.
If you use two independent lasers, however, you must ensure that the whole experiment, including the capture of the interferogram, occurs in a time shorter than $T$. The reason for this is that, while you will always have interference as there is always some phase relation, if the relative phase between the sources slips by $\pi$ then the zones of constructive and destructive interference will switch places. If this happens too often - if the phase slips too quickly - and you end up taking a photograph over an exposure time longer than $T$, then you will not capture the dark fringes as the light fringes will cover them before you're done. If $\tau$ is really short, then you will need to be very, very careful with your measurement to ensure you capture the coherence.
A: Most of the time, experiments use a single laser with a beam splitter, so they get two beams that come from the same source. This has the great advantage that it is cheaper to buy laser plus beam splitter than two lasers.
If you have two independent lasers with a different phase, it is the same as a single beam split and travel paths of non equal length. In a interference pattern, it usually appears as a lateral shift.
