Position dependent $y$ channel reading in lock-in in a photocurrent microscope setup I am currently building a scanning photocurrent measurement setup. It basically focuses a CW laser down to $50\ um$ size on the sample and scan the laser beam over the sample (a metal) and record the generated dc current. I use a current preamplifier and a lock-in to detect the current with pA resolution (the chopper is chopping the light, so the dc current becomes a $110\ Hz$ quasi AC current). Right now I am facing a weird problem. When I focus the light onto one sample spot, the signal shown in lock-in is good, which means x channel reading is stable and y channel is nearly zero. However, when I just moved the laser to a nearby point, the $y$ channel suddenly becomes nonzero and stabilizes at a certain value. Of course, I can do auto phase and remove $y$ channel reading. Then it requires doing auto phase whenever I need to change position, which is impossible. This happens all the time.
What I want to ask is what are the possible reasons to have this problem?
 A: One effect to consider: How wide is the beam passing through the chopper? As the blades of the chopper cut across the beam, it will first block one side of the beam. When the hole in the chopper moves across, it will then first reveal one side of the beam.
As a result of this, the focussed light spot will not simply turn on and off, but it will also effectively wiggle back and forth slightly. This will happen out of phase with the desired overall chopping motion.
Now, imagine you park your beam on a part of your device that has a position- or angle-dependent photocurrent. As the beam wiggles back and forth, it will induce an out-of-phase photocurrent.
There are a variety of things you can do to test whether this is important, for example seeing how the sign of the weird signal changes on different edges of the device. The reason I suggest this mechanism is that you identified a very tight focus of 50nm, and so I wouldn't be surprised to hear that the unfocussed laser beam is as wide as the gaps in the chopper wheel.
Sometimes an out-of-phase signal is purely artifical like this, and so you shouldn't use auto-phase because that will contaminate the good part of your signal.
(This is not the only explanation for such an observation, by the way! Really you have to dig into your system and look for clues, and carefully think about how your experiment has been set up)
