Free Physics (and Math) Books

Challenge: Think of any physics book that is known by its author’s last name.

OK, what is its free replacement?

A variant on this question: How much of the MIT undergraduate physics curriculum can be taught with free books? The only reasonable answer would be all of it, because we’ve had the Web for 30 years now. Sadly, the textbook business is not reasonable.

If people had decided to be useful at any point in the past generation, you could go to and click to download all-the-textbooks-you-need.tgz, but we got MOOCs instead. Not to mention the “open courseware” that too much of the time is just a stack of PowerPoints. Oh, and software that puts kids under surveillance so that a company can monetize their behavior. Because that’s the future we deserved, right?

There are books out there, but they peter out after you get past the first year or so, and a lot is pitched either too low or too high. Either there’s a few chapters in a big “university physics” kind of volume that wouldn’t be enough to fill a whole semester, or there’s a substantial text that’s intended for graduate students. Plenty of times, one finds a totally decent set of lecture notes that whiffs at the last step by not incorporating homework problems. If we really want institutional change, we need (among other things) more drop-in replacements for the books to which physicists habitually turn, so that we can overcome the force of tradition.

In what follows, I go through the MIT course catalogue and provide links and commentary.

Mechanics I: “Newtonian” physics
Yes, yes, $F = ma$ was from Euler, not Newton, the concept of energy wasn’t really codified until du Châtelet, etc. Calling the subject “Newtonian” mechanics is an oversimplification, but a readily-understood one.

  • Moebs, Ling, Sanny et al., University Physics volume 1 (OpenStax).
  • Feynman, Leighton and Sands, The Feynman Lectures on Physics, volume 1. These lectures have a reputation for being difficult to learn out of, if you’re learning for the first time, but inspirational if you have some grasp of the subject already. I personally suspect that their reputation for difficulty is partly undeserved. The troubles Caltech had with the course at the beginning sound like the start-up challenges of every course I’ve ever taken that was being offered for the first time. The big problem, I think, is that the exercises are not smoothly integrated into the text and don’t provide a manageable gradation from easy and bite-sized to demanding and hearty. Of course, much of what was “state of the art” in 1964 is so no longer, and if you’re burnt out on the Feynman-industrial complex, I can appreciate that too.
  • Chakrabarty et al., 8.01 (MIT OpenCourseWare).

The more accelerated version of this course, intended for students who had a stronger math background going in, reached topics that the OpenStax book doesn’t. In particular, it could go as far as solving the Kepler problem, which the Feynman Lectures themselves don’t do.

Calculus I: Single-variable

In my day, this was taught out of Purcell, as opposed to the more advanced course that one could take as an elective, which was based on Griffiths with inflections of Jackson. A drop-in replacement with that level and scope still seems hard to come by.

  • Moebs, Ling, Sanny et al., University Physics volume 2 (OpenStax). Good so far as it goes, but lacks the differential version of the Maxwell equations and the introduction to special relativity that we got.

Calculus II: Multivariable

Waves and Vibrations

  • Lee and Georgi, 8.03sc (MIT OpenCourseWare).


Differential Equations

Quantum I
The first semester of quantum physics was mostly an attempt to teach the subject without linear algebra. If you made it out knowing how to solve the particle-in-a-box and the harmonic oscillator, you were on track.

  • Zwiebach, 8.04 (MIT OpenCourseWare).
  • Feynman, Leighton and Sands, The Feynman Lectures on Physics, volume 3. Oddly, given the reputation for being intimidating, the background presumed by volume 3 puts it more at this level than the second or third semesters of quantum mechanics that follow. However, the lack of exercises relegates this to supplemental reading.

Statistical I
What I think we need here is an OA counterpart roughly comparable to, e.g., Finn’s Thermal Physics. The online resources I’ve managed to turn up so far have pitched at a higher level than the spot we’d need to fill here. I recently taught a thermo course for undergrads in the early part of a physics major, and a lot of online lecture notes seem to rush through in a single chapter what we spent half the semester on.

  • Moebs, Ling, Sanny et al., University Physics volume 2 (OpenStax). Fine for a first encounter with macroscopic, equilibrium thermodynamics, but it stops before the topic of free energy and is lacking on the statistical-physics perspective.

Quantum II

Laboratory I
Good grief, this was brutal. And not for any particularly good reason. On top of the intrinsic difficulties, like lab equipment never working right and having to master a new topic every three weeks as you went from one experiment to the next, this was also most students’ first encounter with the “job skills” of physics: fitting curves to data, doing a literature review, giving a technical talk, writing a technical paper, etc.

A lab course might not have a textbook at all, per se. Instead, most of the reading will be historical papers about whatever pivotal experiment you are trying to replicate, along with manuals for the equipment you are struggling to use.

More than anywhere else on this list, this is where the openness of software becomes a concern. Locking students into doing data analysis with proprietary tools is, on the whole, a bad move. (And if the extent of the instruction is “here’s some MATLAB, good luck”, doing the same but with Python instead is no less friendly.)

Statistical II

Laboratory II
I did better in the second semester of Junior Lab than I did in the first, partly because we had more time for each experiment, but mostly because it took me the whole first semester to figure out what the heck I was doing.

Mechanics II
This is the first encounter with Lagrangian and Hamiltonian methods. I am still looking for a reference on this material at a suitable level that I’m really happy with, but David Tong’s lecture notes come close.

Mechanics III

  • Stewart, 8.09 (MIT OpenCourseWare). The Hamilton–Jacobi language is introduced here in the standard way, i.e., presuming a fairly in-depth knowledge of canonical transformations. Students who have already had to master two whole formalisms are generally too tired for this. However, the path can be eased slightly by explaining how to go directly from the “Newtonian” to the Hamilton–Jacobi presentation.

Math Elective
The physics department requires each student to wander over to the mathematics department for one elective beyond differential equations. Most students fill this slot with linear algebra, I believe, though I went with real analysis for some reason.

Quantum III

  • Harrow, 8.06 (MIT OpenCourseWare).


  • Fitzpatrick, Classical Electromagnetism: An intermediate level course. These lecture notes are split across many small pages, which may be exactly what some readers who aren’t me want. Clear, but terse: the words connecting the equations are kept to a minimum, but not reduced beyond that. Lacks exercises (unless, like me, you have to verify that everything which “follows easily” really does follow easily).
  • Feynman, Leighton and Sands, The Feynman Lectures on Physics, volume 2. More demanding than volume 3 in some respects. All the caveats mentioned above apply here, too.