Strange photoelec. exp. result

[turin]

For some reason, something just came to my mind about a strange result I got when doing a photoelectric effect experiment early this spring. I had "monochromatic" LEDs (I am most suspicious about them) shining into a hole in a "box" that contained the photo sensitive "vacuum tube thing" (the thing that contains the photo emissive electrode in glass). If I shined the LED directly into the hole (i.e. the plane of the junction in the LED ~parallel to the plane of the hole), then I get the photo current to vanish at some voltage repeatably. But, I found that, keeping that voltage, if I tilt the incidence to some value ~20^o off or so, then the the photo current would spike way up, and it would not respond anywhere at all as sensitively to a change in voltage. When I did set the voltage to eliminate the resulting photo current, and then left the tilt, the setting was repeatable. This occurred for all four of the colours (visibly red, blue, green, and yellow) and one UV LED. The voltage settings were repeatable after dismantling the setup and rerunning the experiment the next day.

So, after that long-winded exposition, does anyone have any suggestions for an explanation? I suspect the diodes, but the "box" has a lot of control circuitry that makes me suspicious as well. I may be able to dig up the data (but I doubt it).

 

[Chen]

Are you sure it's not because the tilting changed the intensity of the light?

 

[turin]

Are you sure it's not because the tilting changed the intensity of the light?

Thanks, Chen. No, I'm not sure. Do you suggest that
a) the emission from the LED may have a greater intensity at ~20^o or that,
b) in tilting the apparatus, I somehow may have altered the amount of current through the diodes?

I can buy possibility a), but I don't think b) would survive the repeatability.

 

[Chen]

I'm sorry, it's not the intensity of the light that changes (because that's constant and depends only on the source), it's the amount of energy you provide that changes. I'll explain below...

 

[Chen]

We had a light bulb in a black closed box, with a small openning over which we placed a filter. Then we placed that openning against the hole in the photoelectric unit, so that no external light can come in. Now if I tilted the light source (while somehow keeping the system covered from external light), the illuminance E = I/r² wouldn't be the same anymore, because some parts of the emitter would be closer to the source than others. As a result, it is possible that those parts of the metal emitter would get more energy than they did earlier (if they are now closer to the source than they were before), and a result the electrons emitted by those parts will have more energy and be able to go further before being stopped by the stopping voltage. I think that could explain the "spike" you saw in the current.

The other possible explanation is that by tilting the system, you exposed it to light from outside... but I don't think you did that.

 

[Njorl]

The output of an LED can be very complicated. The light is essentially generated in a planar optical waveguide. The light propogates in different modes. WHile most light will be in the fundamental mode, some will be in higher order modes. The different modes will have different output coupling.

There is also a little bit of stimulated emission. It acts in some ways like a single pass laser. This causes the frequency distribution to be spikey rather than normal. I'd be surprised if you could resolve the spikes with a photoelectric effect apparatus though.

Njorl

 

[turin]

Thanks again, Chen. I don't think that expalins it, though. It's not your fault, it's just that I didn't indicate two more critical features in my description.

1. The experiment was conducted in "pitch darkness" (except a small light to illuminate the readout). In order to make sure, I completely opened up the aperture and took a reading, being liberal with the gauge illumination - null. Only when one of the diodes was on did the reading yield a non-trivial value.

2. Tilting the diode put it considerably further from the aperture. The tilt was entirely in a horizontal plane, not tipping over towards the apparatus like a leaning tower, but like angled parking spaces vs straight parking spaces. The diodes are part of a "diode box" in which they are situated significantly away from the edges and slightly recessed, so tilting the box put required moving the individual diode itself further from the aparatus.

3. The third point I want to re-emphasize is repeatability of these results. At first they seemed just an anomoly because I sumbled on them by accident. I intended to run the experiment several time cycling through the diodes. I began to develop a carelessness about how I situated the diode box, since I had a log amp that reduced the critical nature of zeroing out the photo current reading. It was in this carelessness that I got "a round of bad data." But, upon analysing the other data, I found a pattern, and decided to test the effect of box positioning a bit more carefully. I then found the ~20^o orientation to be significant and ran it through several cycles of each diode and even ran the experiment at a later date.

The different modes will have different output coupling.

I don't follow. To what is the output coupling? Are you saying that there is somehow feedback from the photo electric unit that depends on frequency and tilt? How does this feedback get fed back with such dependence? The apparatus is dull black. Can the reflections from it still be so significant?

There is also a little bit of stimulated emission.

That's interesting. I never thought of it, but, now that you mention it, I would be surprised it it did not occur.