In an earlier post I discussed my philosophy of teaching special relativity. My main idea was that physics professors should keep the “weird stuff” at bay, and start with non-controversial statements; once students are on board, you can push under the grass and show them the seething Lynchian bugs beneath.
Well, what about quantum mechanics? Does the same philosophy apply?
My answer is yes, of course. Don’t start with Schrödinger’s cat. Don’t mention the Heisenberg uncertainty principle, or wave collapse, or the EPR experiment, or Bell’s theorem, or the double slit experiment, or quantum teleportation, or many worlds, or Einstein’s dice. Start with the problems of physics, circa 1900, and how those problems were gradually solved. In working out how physicists were gradually led to quantum mechanics, students will build up the same mental framework for understanding quantum mechanics. At least, that’s how it works in theory.
Now, my perspective is from the point of view of a professor who teaches only undergraduates. I only get to teach quantum mechanics once a year: in a course called Modern Physics, which is sort of a survey course of 20th century physics. (If I were to teach quantum mechanics to graduate students, my approach would be different; I’d probably start with linear algebra and the eigenvalue problem, but that’s a post for another day.) As it is, my approach is historical, and it seems to work just fine. I talk about the evidence for quantized matter (i.e. atoms), such as Dalton’s law of multiple proportions, Faraday’s demonstration in 1833 that charge is quantized, Thomson’s experiment, Millikan’s experiment, and so on. Then I explain the ultraviolet catastrophe, show how Planck was able to “fix” the problem by quantizing energy, and how Einstein “solved” the problematic photoelectric effect with a Planckian argument. Next is the Compton effect, then the Bohr model and an explanation of the Balmer rule for hydrogen spectra…
We’re not doing quantum mechanics yet. We’re just setting the stage; teaching the student all the physics that a physicist would know up until, say, 1925. The big breakthrough from about 1825-1925 is that things are quantized. Things come in lumps. Matter is quantized. Energy is quantized.
The big breakthrough of 1925-1935 is, strangely, the opposite: things are waves. Matter is waves. Energy is waves. Everything is a wave.
So then, quantum mechanics. You should explain what a wave is (something that is periodic in both space and time, simultaneously). Here, you will need to teach a little math: partial derivatives, dispersion relations, etc. And then comes the most important step of all: you will show what happens when two (classical!) wave functions are “averaged”:
ψ1 = cos(k1x – ω1t)
ψ2 = cos(k2x – ω2t)
Ψ(x,t) = (1/2) cos(k1x – ω1t) + (1/2) cos(k2x – ω2t)
Ψ(x,t) = cos(Δk·x – Δω·t) · cos(k·x – ω·t)
where Δk ≡ (k1 – k2)/2, k ≡ (k1 + k2)/2, etc.
[Here I have skipped some simple algebra.]
This entirely classical result is crucial to understanding quantum mechanics. In words, I would say this: “Real-life waves are usually combinations of waves of different frequencies or wavelengths. But such ‘combination waves’ can be written simply as the product of two wave functions: one which represents ‘large-scale’ or global oscillations (i.e. cos(Δk·x – Δω·t)) and one which represents ‘small-scale’ or local oscillations (i.e. cos(k·x – ω·t)).
This way of looking at wave functions (remember, we haven’t introduced Schrödinger’s equation yet, nor should we!) makes it much easier to introduce the concept of group velocity vs. phase velocity: group velocity is just the speed of the large-scale wave groups, whereas phase velocity is the speed of an individual wave peak. They are not necessarily the same.
It is also easy at this point to show that if you combine more and more wave functions, you get something that looks more and more like a wave “packet”. In the limit as the number of wave functions goes to infinity, the packet becomes localized in space. And then it’s simple to introduce the classical uncertainty principle: Δk·Δx > ½. It’s not simple to prove, but it’s simple to make plausible. And that’s all we want at this point.
We’re still not doing quantum mechanics, but we’re almost there. Instead, we’ve shown how waves behave, and how uncertainty is inherent in anything with a wave-like nature. Of course now is the time to strike, while the iron is hot.
What if matter is really made from waves? What would be the consequences of that? [Enter de Broglie, stage right] One immediately gets the Heisenberg relations (really, this is like one line of algebra at the most, starting from the de Broglie relations) and suddenly you’re doing quantum mechanics! The advantage of this approach is that “uncertainty” seems completely natural, just a consequence of being wave-like.
And whence Schrödinger’s equation? I make no attempt to “prove” it in any rigorous way in an undergraduate course. Instead, I just make it imminently plausible, by performing the following trick. First, introduce complex variables, and how to write wave functions in terms of them. Next, make it clear that a partial derivative with respect to x or t can be “re-written” in terms of multiplication:
d ψ /dx → ik ψ
d ψ /dt → –iω ψ
Then “proving” Schrödinger’s equation in a non-rigorous way takes 4 lines of simple algebra:
E = p2/2m
E ψ = (p2/2m)ψ
Now use the de Broglie relations E = ħω and p = ħk…
ħw ψ = (ħ2k 2/2m) ψ
iħ(∂ψ/∂t) = (–ħ2/2m) ∂2ψ/∂x2
There’s time enough for weirdness later. Right now, armed with the Schrödinger equation, the student will have their hands full doing infinite well problems, learning about superposition, arguing about probability densities. As George McFly said, “I am your density.” And as Schrodinger said, probably apocryphally, “Don’t mention my cat till you see the whites of their eyes.”