Chad Orzel asks what experimental setup he should use for an undergrad photoelectric effect lab:
…we have two different set-ups for doing a photoelectric effect experiment. One of these is a PASCO apparatus with the phototube wired to a circuit inside an actual black box. You shine light into the tube, press a button, and the output of the box rises to the stopping potential for that frequency in a more-or-less exponential manner. This gives very nice results, often within 1% of the accepted value of Planck’s Constant.
The other is an old-school lab, using a homemade monochromator and a phototube with an external voltage generator supplying the stopping potential. For each color of light, the students watch the output of the phototube on an oscilloscope, measure the output voltage for a handful of applied voltages, and extrapolate to find the stopping potential. This is much closer to the way the experiments were originally done, but it also tends to give results that differ from the accepted value by 20-30%.
Your answer would depend a lot on what you think the purpose of the lab should be. My view, like many of the commenters at Uncertain Principles, is that the purpose of labs is to let students learn how to conduct experiments. By this I don’t mean how to use the specific equipment involved (though it’s useful to do so), but how to calculate and justify experimental errors, how to explain why your data is evidence for/against a model, general principles about the points at which to take data when one is calibrating equipment versus when one is taking the actual measurements being used to test the model, etc. From this perspective, the old-school setup seems like a clear winner — it seems doable but still challenging enough, methodologically, to test the students’ experimental skills. Chad, however, says that
…the purpose of the lab is to show that experimental measurements of the photoelectric effect agree well with the Einstein model. The more complicated version doesn’t really add to that understanding, and in fact, the complication tends to obscure the physics. Students spend so much time fretting over the experimental details that they lose track of what it’s supposed to show.
You can argue that they’re learning lab skills in the process, but I’m not all that impressed. The only really useful thing they get out of it is how to use an oscilloscope, and there are other ways to teach that. There’s some fuzzy data-selection heuristic stuff going on in deciding exactly what to use as the stopping potential for any given point, but it’s hard to explain that in such a way that they don’t leave the lab thinking “it’s ok to fiddle with the data to get something closer to the target value.” That’s not only not what we’d like them to learn, but is actively harmful.
The thing is, I don’t see how doing the experiment with the PASCO black-box detectors would “show that experimental measurements of the photoelectric effect agree well with the Einstein model.” Suppose I was skeptical that Einstein’s model has been experimentally validated. Would I be convinced that it has by doing the PASCO experiment? No, because I don’t know that the apparatus accurately converts the energies it measures into stopping potentials, and that the values output by the apparatus actually are those of the stopping potentials. It’s natural to suspect that what the black box is doing isn’t what my instructor claims it’s doing, since the instructor has a vested interest in telling lies-to-children about what the box does. (This suspicion may not be justified. But we can see that inferring the model’s goodness from the black box experiment involves an extra epistemic step, so it isn’t obviously unreasonable to be less convinced by the black box version than by the old school version.)
To summarise:
- From a teaching-the-methodology point of view, the ‘historical’ experiment wins for me.
- From a convincing-students-the-model-is-right point of view, the historical experiment wins too, because the black box experiment isn’t any more convincing to a skeptic of the model.
Incidentally, the comments to Chad’s post highlighted to me how terrible my undergrad physics labs were. A few commenters, including Chad, say that they give the students the equipment in bits and leave them to figure out how to put them together in order to conduct the experiment. Except for introductory labs in which we did extremely simple experiments like measuring g using inclined ramps and shit, I don’t remember having to do any major assembly work for my labs. And what little assembly work we had to do would be laid out in painstaking detail in the lab manual.
Posted by Ponder Stibbons