Setting Up a Photoelectric Effect Lab Experiment
I spent a week getting my undergrad lab to consistently measure stopping potentials across multiple wavelengths. The photoelectric effect is straightforward in principle — light hits a metal surface and ejects electrons if the photon energy exceeds the material's work function — but the practical execution is full of small things that silently ruin your data. Here is how I actually made it work. You will need a monochromatic light source, preferably a tunable laser or a set of narrow-bandpass filters paired with a stable lamp. A photomultiplier tube or a simple photocell connected to a picoammeter works for detecting current. The key component is the ability to apply a variable retarding potential between the emitter and collector plates. I used a precision voltage source that could sweep from 0 to about -3 volts in 10 millivolt steps. The first thing that catches everyone out is contact potential difference. The metal of your emitter and the metal of your collector are almost never the same material, and even if they are, surface oxidation creates a small built-in voltage that shifts your stopping potential reading by roughly 0.1 to 0.3 volts depending on how long the chamber has been sitting open to air. I ended up characterizing my contact potential by running a sweep with no light at all and noting where the tiny leakage current crossed zero, then subtracting that offset from every measurement afterward. Without doing this, your extracted work function values will be consistently wrong by a margin large enough to invalidate the whole experiment.
Another thing nobody mentions in the textbook version is that the intensity of your light source must stay constant across all wavelengths you test. If you swap filters and your lamp output drifts, your saturation current changes and you might misinterpret it as a kinetic energy effect. I placed a calibrated photodiode right next to the sample holder and monitored its reading continuously. When the intensity drifted more than two percent between filter changes, I adjusted the source voltage to bring it back. The actual procedure runs like this. First, pump down the chamber to below 10^-5 torr. Any residual gas causes ionization noise that shows up as a jittery current reading and occasionally corrodes your emitter surface over time. Second, clean the emitter — usually sodium or potassium coated on a glass substrate for teaching labs, though I prefer a freshly cleaved piece of zinc for better stability. Third, start at the shortest wavelength you plan to measure and work down. Record the stopping potential for at least three different intensities at each wavelength. If the stopping potential changes when you change intensity, something is wrong — either the voltage source has too much ripple or your chamber is not vacuumed well enough. Plot stopping potential against photon frequency and the slope gives you Planck's constant divided by the elementary charge. The intercept on the frequency axis is the threshold frequency, and multiplying that by Planck's constant and the charge gives you the work function. With a decent setup, you should land within five percent of the accepted value for your chosen metal. My best run came in at 4.31 electron volts for zinc, compared to the accepted 4.33. The worst run, from a day when I neglected to check the lamp intensity, was off by twelve percent because the filter changed shifted the spectral output in a way I did not catch.
If you are working with a classroom setup that does not allow vacuum, you can still observe the qualitative behavior — electrons only emit above a threshold frequency regardless of intensity — but you will not get quantitative results because ambient gas scatters the electrons and the contact potential issue becomes much harder to control. In that case, the exercise is really just about confirming the existence of the threshold and the independence from intensity. Don't try to extract Planck's constant from a non-vacuum setup. It won't work, and you will waste two lab sessions chasing errors that have no physical basis. One edge case that cost me half a day: I once used a blue LED as my short wavelength source and got a stopping potential that was roughly half of what the theory predicted. Turns out the LED had a spectral width of about 30 nanometers, and a significant portion of the emission was in the green region where the photon energy was below the threshold. The "current" I was measuring was actually coming from the small fraction of photons above threshold, but the stopping potential was being set by those higher energy photons while most of the optical power was wasted below it. Switching to a narrow bandpass filter centered at 450 nanometers with a 10 nanometer bandwidth fixed the problem immediately. If you are using any broadband or wide-spectrum source, always check the spectral profile with a spectrometer or at least a calibrated photodiode array before trusting your data.
Get the Full Details
