Topics They Don't Seem to Teach in Undergrad

[TomServo]

I've noticed that there are certain topics that physics departments don't seem to teach at the undergrad level, except maybe smaller introductory elective courses here and there. It seems like the emphasis is entirely on making E&M and Quantum specialists at the expense of certain other areas. I'm curious as to why this is the case for the following topics:

General relativity
Anything to do with fluid dynamics (it seems completely possible, judging by the curricula I've looked at, that one could get a BS in physics without ever learning about why things float)
Heat transfer (yes, one thermo class is required at most institutions, but this seems almost like an afterthought)

In short, it seems like mechanical and aerospace engineers study a heck of a lot of physics that physics majors themselves gloss over. Why is this? Or am I completely wrong?

 

[Andy Resnick]

You are not wrong, and I would also add 'optics' to your list. I have no idea why the 'canonical' curriculum is defined the way it is. Although the US does not have an accreditation of physics degrees in particular (except for medical physics via CAMPEP), the IOP (UK) accreditation requirements list a specific set of topics for "Core of Physics" on http://www.iop.org/education/higher_education/accreditation/file_43311.pdf, and the topics you mention are conspicuously absent. That is, General Relativity is not considered part of the 'core of physics' according to the IOP!

For that matter, it's not clear why (in the US) 124 +/- credit hours are needed to get a BS degree- one of the national accreditation agencies simply states:

"3-5-202. Education Requirements. The minimum number of credits required for the bachelor’s degree shall be 120 semester hours, 180 quarter hours, or their equivalent, normally earned over a period of eight semesters, 12 quarters, or their equivalent. Transfer and award of credit for appropriate work at other institutions may be granted."
but does not give a reason for the numbers.

To summarize, the current state of the BS degree in the US is based on fairly arbitrary guidelines. To be sure, since our BS Physics degree does not have an accreditation distinct from our parent institution, we have wide latitude in what subjects to cover (given that we have state guidelines which are designed to allow students to transfer with a minimum of problems). If we wanted to, we could petition the faculty senate to change required courses- and most likely, if we were serious about it, the petition would be granted.

 

[f95toli]

*I studied fluid dynamics as an undergrad. But I studied engineering physics.

*Heat transfer is pretty easy at the "theoretical" level (as long as you don't start worrying about real-world applications, and limit yourself to conduction and radiation) and tends to be covered quite well in various courses. I e.g. solved lots of heat transfer problems in my PDE course (as well as in comp phys and math phys courses)

*GR is very complicated if you do it "properly" and it is very unlikely that you will get the mathematical background as part of you undergrad math courses.
Also, remember that there are few applications of GR outside of cosmology whereas QM and EM are used in just about every field of physics (and quite a few fields of engineering). Or in other words: GR might be interesting but it would be an unnecessary course for nearly everyone.

 

[AlephZero]

I wonder if part of the issue with fluid dynamics is that (notwithstanding the huge advances in the 20th century) there are still large parts of the subject that are semi-empirical. If understanding the general solution of the Navier-Stokes equations are still the subject of a Millennium Prize, there is a very big hole at the core of the subject from a purist's point of view.

Of course engineers just have to get on and make the best practical use of what is already known, but that's the basic difference between physics and engineering.

 

[lisab]

When I was a senior physics major, my university decided to offer a class in gravity to junior/senior undergrads. It was touted as a trial class, a beta-version. I had one look at the syllabus and thought, no thanks...one of the best decisions I made that year!

Turns out, the few intrepid students who did take that class were simply blown away - totally unprepared for that level of work. It didn't help that most were taking E&M and QM at the same time, so they just didn't have much time left to commit to it.

The class wasn't offered again, AFAIK.

So I know it's been discussed by at least one school, to broaden the base of undergrad exposure. But it would almost have to come at the expense of narrowing the traditional E&M and QM material - something most physics departments aren't willing to do.

 

[SophusLies]

Statistics/probability and non-awkward communication skills.

 

[D H]

*General relativity.
As others have noted, the math is a bit on the advanced side even for the typical senior physics major. Some colleges let seniors, with permission, take introductory graduate level courses.

*Fluid dynamics.
Yep. Undergrad physics education here pretty much stops at Halliday & Resnik. Physics majors don't have to know what a shock wave is, which is pretty shocking.

*Heat transfer.
And thermo in general. Halliday & Resnik again is about as far many physics majors get in studying thermo. Statistical physics is offered at the undergrad level in many schools but it is optional.

In short, it seems like mechanical and aerospace engineers study a heck of a lot of physics that physics majors themselves gloss over. Why is this? Or am I completely wrong?

It's because the primary goal of undergraduate physics is to educate students so they have a good chance of getting admitted to some graduate physics program. Most graduate physics programs focus on solid state physics, particle physics, or optics. Each of these requires a solid quantum and E&M background, so that's what they teach.

 

[Dembadon]

It probably varies by institution. At mine, "Modern Optics and Photonics (after EM & Quantum prereq)" is required, as well as probability and statistics, but Special and General Relativity are not.

 

[Nabeshin]

I don't really understand the approach that 'GR is too mathy, we cannot teach it to undergraduates.' It seems to me absolutely unacceptable that an undergraduate go through his education and be called a physics major without having learned, at least at a basic level, GR. I see no reason why courses based on Hartle or Schutz shouldn't be required just as much as a course in QM at the undergraduate level. Obviously all physics subjects have many layers of understanding, but strangely many people seem to think that if you're not going to teach GR with a full introduction to differential geometry then it's not even worth teaching.

Nonsense.

 

[Angry Citizen]

Also, remember that there are few applications of GR outside of cosmology

Pardon? GR is the reason why we have GPS satellites that work. And that's just one area that I'm aware of where GR made an enormous impact outside of cosmology. God only knows where else it's been important.

 

[Mépris]

Halliday & Resnik again is about as far many physics majors[/B] get in studying thermo. Statistical physics is offered at the undergrad level in many schools but it is optional.

Seriously? That Fundamentals of Physics text book? So, it's in grad school that students learn the meat of physics - during those two years of the MS course? :S

 

[quantum13]

*General relativity.
As others have noted, the math is a bit on the advanced side even for the typical senior physics major. Some colleges let seniors, with permission, take introductory graduate level courses.

What kind of math is needed for to understand general relativity? I don't understand the big difference between a senior undergrad and a first year grad student. Do grad students take those advanced math courses concurrently with GR or is it assumed they have more time to study?

 

[D H]

Also, remember that there are few applications of GR outside of cosmology

Pardon? GR is the reason why we have GPS satellites that work. And that's just one area that I'm aware of where GR made an enormous impact outside of cosmology. God only knows where else it's been important.

f95toli said very few. You named one. That qualifies as "very few." I'll name one more: high-precision planetary ephemerides over millennia or longer. Now name one more.

Of both those named applications, you don't have to understand general relativity as a taught as a subject to be able to understand the subject application. A weak field approximation as a given works just fine. Besides, physicists don't work on GPS for the most part. It is a solved problem as far as physicists are concerned.


It is important to remember that the primary purpose of an undergrad physics program is to prepare students for a graduate physics program. Television programs such as Nova make it appear that cosmology is the only thing that PhD physicists do. The reality is that very few PhD physicists work in this area.

Seriously? That Fundamentals of Physics text book?

I was thinking more of Halliday & Resnick Physics rather than the dumbed-down Fundamentals of Physics, but yes. A typical undergrad physics has eight or so required physics classes for juniors and seniors, and even fewer if the physics department happens to in a college that offers a bachelors of arts degree as opposed to a bachelors of science. The junior level statistical mechanics course is optional in a some undergrad physics programs, about a third from browsing at various schools. If it is offered at all, the follow-on senior level statistical physics course is optional in a solid majority of such programs.

So, it's in grad school that students learn the meat of physics - during those two years of the MS course? :S

That's true for many subject areas, not just physics. The freshman and sophomore years give a smattering across a broad range of subjects. Juniors and seniors relearn those subjects and build upon them. Then when you get to grad school you find that even the hard subjects you took as a junior and senior were but a simple facade to what is really going on.

A specific example: special relativity. Some aspects of special relativity are so simple that it can be taught in part in high school physics. You'll revist these simpler aspects of special relativity as a freshman or sophomore. Then you'll relearn it at least twice as a junior and senior. It is one of many subjects covered in the upper level undergrad classical mechanics class, it is taught once again in the junior level E&M class, and if you take a high energy physics class you'll hit the subject for a third time. And then as a grad student you'll hit the subject at least twice more. Somewhere along the way you'll derive the relativistic form of Maxwell's equations.

What kind of math is needed for to understand general relativity?

Differential geometry.

I don't understand the big difference between a senior undergrad and a first year grad student.

The big difference is in the diversity of classes, the difficulty of those classes, and the assumed background. Undergrads in the US have to take quite a few classes completely outside of their specialty. How many depends on whether the physics department offers a bachelor of arts versus a bachelor of science degree. Grad school students are completely immersed in their subject matter; they have very few if any required outside classes.

Grad school classes are considerably more challenging than are undergrad classes. Some of your undergrad classmates will not apply to or will not be accepted in a graduate program. The instructors can and do make those grad courses much more intense because they are teaching to the cream of the crop.

And finally, differential geometry almost universally is not a required subject for undergrad physics majors. It is often taught as a part of the graduate level class on general relativity.

Do grad students take those advanced math courses concurrently with GR or is it assumed they have more time to study?

Most undergrads have several unstated minors such as studies of the opposite sex. Grad students can only audit these subjects. They are spending too much time studying.

 

[Andy Resnick]

<snip>

It is important to remember that the primary purpose of an undergrad physics program is to prepare students for a graduate physics program. <snip>

That's part of the problem with Physics degrees- on one hand, we explicitly tell prospective students that a BS degree in Physics is not designed to help them get a job, but we also tell students that a BS degree in physics gives them all kinds of skills that employers want (e.g. problem-solving).

In truth, teaching more thermo, fluids, and optics would likely help undergrads to use their degrees.

 

[D H]

That's part of the problem with Physics degrees- on one hand, we explicitly tell prospective students that a BS degree in Physics is not designed to help them get a job, but we also tell students that a BS degree in physics gives them all kinds of skills that employers want (e.g. problem-solving).

In truth, teaching more thermo, fluids, and optics would likely help undergrads to use their degrees.

I agree. It would also help those who want a physics undergraduate degree but an engineering graduate degree. Hearkening way back to the original post,

n short, it seems like mechanical and aerospace engineers study a heck of a lot of physics that physics majors themselves gloss over.

That is exactly right. Mechanical and aerospace engineers are better versed in classical mechanics, fluid mechanics, and thermodynamics than are most physics majors. This can hurt those who want the broad-based learning and problem solving skills that an undergraduate physics degree offers but want the specialized skills that a graduate engineering degree offers.

 

[Mororvia]

My undergraduate physics (BS) requirements were as follows:
Physics Laboratory for Scientists I and II
Thermodynamics and Modern Physics (sort of a survey course)
Classical Mechanics I
Thermal and Statistical Physics
Electronics
Advanced Laboratory
Quantum Mechanics I
Electricity and Magnetism I

Outside of the first two lines above, most of these were at the junior/senior type of level. Physics majors also had to complete the introductory mechanics and E&M lecture courses experienced by many science/engineering majors, as well as the typical calculus sequence plus 2 more math courses at the junior/senior level. These were usually linear algebra and PDEs.

There was still a good amount of electives to choose from even after completing the university requirements in english, history, etc. Many people took optics, electronics, nuclear physics and solid state as well as more advanced courses in classical and quantum mechanics and E&M. One could even take courses in the engineering department, though you kind of had to fight to get in.

 

[Andy Resnick]

I've been trying to think of alternate approaches to the BS Physics degree; approaches that serve a larger number of students better than the current curriculum. The two I've come up with so far are:

1) get rid of the 3 or 4- credit hour course and replace them with more focused 1- or 2-credit hour courses. This could offer more options to the student, and keep the overall level of content high. Seriously, why spend 8+ credit hours on Physics I and II, only to cover those topics *again* at the 300/400 level? If it's a question of mathematical sophistication, then it's not an issue with the Physics- the Physics department could offer a "mathematical methods for Physicists" class that properly prepares the students.

2) Introduce a 'classical field theory' sequence (2 4-hour classes, for example) that covers continuum mechanics (including fluids), thermodynamics, and general relativity. Electromagnetism could also be included to some degree, but the sequence would replace courses covering thermo, modern, etc.

Again, I want to emphasize that the 'standard' curriculum is a lot more arbitrary than you think, and there isn't a set of rules about what subjects an accredited curriculum has to contain.

 

[twofish-quant]

I've noticed that there are certain topics that physics departments don't seem to teach at the undergrad level, except maybe smaller introductory elective courses here and there.

There are, and there are reasons that certain topics are left out of the core physics program.

General relativity

This tends to get left out because there are large parts of physics in which it's irrelevant.

Anything to do with fluid dynamics (it seems completely possible, judging by the curricula I've looked at, that one could get a BS in physics without ever learning about why things float)
Heat transfer (yes, one thermo class is required at most institutions, but this seems almost like an afterthought)

The reason that tends to get left out is that much of "real world" heat transfer and fluid dynamics turns out to be both empirical and domain specific.

In short, it seems like mechanical and aerospace engineers study a heck of a lot of physics that physics majors themselves gloss over. Why is this? Or am I completely wrong?

It's the knapsack principle. In order to teach one thing, you have to toss something out. Mechanical and aerospace engineers have to deal with heat and fluid transfers, but they don't usually end up studying quantum mechanics at anything more than superficial levels.

 

[D H]

It's the knapsack principle. In order to teach one thing, you have to toss something out.

Exactly. That undergrad knapsack only holds 120 to 130 semester hours, and the physics department only gets to fill a portion of the knapsack. A physics department will adjust the contents of its portion of the knapsack in response to how typical grad school value the courses in that typical undergrad knapsack. If year after year grad schools tend to prefer students who have taken solid state physics over those who haven't, eventually that will make physics of solids a mandatory course. If physics grad schools regularly accept students who don't even have enough fluid mechanics to explain how things float, well eventually that will make fluid mechanics an optional course (if it is offered at all by the physics department).

Fortunately, there are 20 or so technical hours in that knapsack that are completely up to the student. A student who wants to take fluid mechanics or advanced thermodynamics because that student's goal is to pursue a graduate career in engineering can usually find a way to do so.

 

[mal4mac]

I don't really understand the approach that 'GR is too mathy, we cannot teach it to undergraduates.' It seems to me absolutely unacceptable that an undergraduate go through his education and be called a physics major without having learned, at least at a basic level, GR. I see no reason why courses based on Hartle or Schutz shouldn't be required just as much as a course in QM at the undergraduate level. Obviously all physics subjects have many layers of understanding, but strangely many people seem to think that if you're not going to teach GR with a full introduction to differential geometry then it's not even worth teaching.

Nonsense.

I agree. Hartle has written a paper discussing this:

http://arxiv.org/abs/gr-qc/0506075

And why not string theory? You could use a textbook like "A First Course in String Theory" by
Barton Zwiebach.

It's doing a disservice to physics students not teaching them the rudiments of these topics. Even if they don't go on to become physics researchers, they may become physics teachers and pupils are likely to ask, "What's string theory then...". Or they may become science journalists and be asked to write something on these popular topics. And they need some background for dinner party conversations! And... most importantly... they deserve to be taken near to the frontier of the subject if it as at all possible - and Zwiebach and Hartle have shown that it is.

School students should look carefully at the advanced courses offered in Colleges they are thinking of applying to, to see if GR, string theory and other "cutting edge" topics are actually offered - if they want to do these! And check at interview if there are any barriers to doing these courses - I read that at Cambridge only the top 30 students get to do certain "sexy" topics like these - the rest are filtered into solid state, medical physics, etc - anything directly useful to the current military-industrial-state complex, which hasn't much use for strings or GR (yet...)

 

[Mépris]

Is it possible for one to take the more advanced variant of these courses (say, classical mechanics) straight away?

I took it for granted that every physics program had a "Maths for Physicist" kind of course, only that in the US, one could skip this and do the "pure math" variant of this course sequence.

That's part of the problem with Physics degrees- on one hand, we explicitly tell prospective students that a BS degree in Physics is not designed to help them get a job, but we also tell students that a BS degree in physics gives them all kinds of skills that employers want (e.g. problem-solving).

In truth, teaching more thermo, fluids, and optics would likely help undergrads to use their degrees.

Is this not what the Engineering Physics degree is for? Physics + a concentration on a few applied aspects of it? I've seen programs with concentrations on renewable energy, medical physics or with EE courses thrown in.

 

[Andy Resnick]

On a whim, I pulled out my (ahem) 1987 course catalog, compared it to what is the current catalog, and the results are telling. As you may expect, there are changes across the board for all science and engineering majors. However, Physics *by far* has the fewest changes. What better way to communicate to prospective students that Physics is a dead subject?

There could be many reasons for this: engineering programs have always had to be responsive to the changing requirements in industry; among the sciences, Physics students have traditionally considered industrial employment as a secondary option. It would also be incorrect to claim that Biology and Chemistry have somehow undergone more significant changes in their disciplines than Physics, and so their curricula need to reflect those larger changes. Physics has grown just as much (if not more) than the other sciences.

To be fair, my undergrad program now has many fewer required classes and many more electives- this follows my first idea, to offer a wider variety of subjects to the student. This places a higher burden both on the student and advisor to make sure the students ends up with a coherent educational plan- something that has always been done in the case of a minor (or major with a particular concentration).

 

[Angry Citizen]

And why not string theory?

I think it sets an awful precedent to teach theories for which there exists no experimental evidence. Indeed, even its theoretical basis is questionable. As a research topic it is fair game; but as an undergraduate class? You might as well be teaching Intelligent Design.

 

[TomServo]

Is physics at the stage where, like engineering once did, it needs to split? Has it gotten too big, with E&M, quantum, fluid mechanics, thermodynamics, computational, solid state, optical, laser, plasma, astro, classical, relativistic, etc, that saying "I'm a physics major" is more of an umbrella term than an actual major, like saying "I'm an engineering major" means you aren't majoring in "engineering" but a certain kind of engineering?

I mean, from a certain pov, we have a de facto split already, with aerospace and mechanical engineers studying certain areas of physics to a much, much greater extent than most bona fide physicists. From that pov, you could regard aerospace and mechanical engineering majors as experimental physicists in the areas of fluid mechanics, thermodynamics, etc, could you not?

The only way I can see bringing back fluid mechanics, thermodynamics, and at least giving them their due attention on an undergrad-level of understanding for a BS without making the physics BS a six-year degree, is to ditch virtually all GECs to make way for the classes. FWIW, I'm all in favor of ditching virtually all GECs. I would love to be able to at least have the option of trading one more English or art history class to get a better understanding of how water works its mojo.

I have one more question about this:

The reason that tends to get left out is that much of "real world" heat transfer and fluid dynamics turns out to be both empirical and domain specific.

I'm a second-year undergrad so I'm not quite sure what you're saying here. I mean, I understand "empirical" and I think I grasp "domain specific" but could you expand on this? Are you saying that heat transfer and fluid mechanics has too much turbulence or too many atoms or something to be modeled mathematically in a solvable way, so all solutions for a given problem are derived numerically? And if so, since it's still physics, why not at least talk about it?

 

[twofish-quant]

It's doing a disservice to physics students not teaching them the rudiments of these topics.

The problem is that you have to take something out Among the list of things that a physics researcher must know, string theory is pretty low on the list.

Even if they don't go on to become physics researchers, they may become physics teachers and pupils are likely to ask, "What's string theory then...".

Then, answer "I don't really know".

 

[twofish-quant]

Exactly. That undergrad knapsack only holds 120 to 130 semester hours, and the physics department only gets to fill a portion of the knapsack.

And if you increase the science/physics component of the course, then what gets hit are the humanities, which are going to help the student after he gets his Ph.D. and is trying to stay sane looking for work.

 

[twofish-quant]

Is physics at the stage where, like engineering once did, it needs to split? Has it gotten too big, with E&M, quantum, fluid mechanics, thermodynamics, computational, solid state, optical, laser, plasma, astro, classical, relativistic, etc, that saying "I'm a physics major" is more of an umbrella term than an actual major, like saying "I'm an engineering major" means you aren't majoring in "engineering" but a certain kind of engineering?

The problem with that is that:

1) it's a bad into to get overly specialized
2) every field of physics interacts with every other field. You aren't going to be a decent astrophysicist without some exposure to EM, QM, fluid mechanics, themodynamics, solid state, etc. etc.

FWIW, I'm all in favor of ditching virtually all GECs. I would love to be able to at least have the option of trading one more English or art history class to get a better understanding of how water works its mojo.

I'm not. The problem is this is really going to bite you went you start looking for work. If you want to apply your physics knowledge to something non-trivial, then you'll need to under how English and art history work. Just as example, right now I'm trying to figure out how to display multi-dimensional data, and having some very, very basic art history gives me some ideas on how to go about doing that.

Are you saying that heat transfer and fluid mechanics has too much turbulence or too many atoms or something to be modeled mathematically in a solvable way, so all solutions for a given problem are derived numerically?

Not even numerically. If you want to see how much friction there is when you pump water down a pipe, you get a pipe, pump water down it, and measure how much friction there is. We don't really have much of a clue how to calculate it. You can do complex CFD simulations, but in the end there are always a ton of fudge factors that you have to correlate back to experiment.

And if so, since it's still physics, why not at least talk about it?

Because there isn't much to talk about. If you want to see what happens, you look up the number in a table and that's about it. If you want to do a numerical simulation, you load that table into the computer and that's it.

This is in contrast with say calculating the orbitals of the hydrogen atom, in which you can start with some basic principles and come up with some solutions. Also there *are* problems in thermo which you can start with basic principles and then calculate solutions. With fluids and heat transfer, it's mostly "measure it" and put it into a table.

The other thing is that with most of the things that they teach in physics classes, you learn one rule, and you can apply it in a 100 different places. Most heat transfer and fluids problems end up being extremely specific and there isn't a general rule that you can apply to 100 different places.

 

[Andy Resnick]

Is physics at the stage where, like engineering once did, it needs to split? <snip>

That's sort of happening already- at my undergraduate institution, one can major in Physics, Applied Physics, Engineering Physics, Physics with a concentration in: Optics, Microelectronics, Nuclear Science, Computing, Biology and Medicine, Environmental Science, or Space Science. Physics majors can also minor in Astrobiology.

It does increase a risk to the student- early specialization in a field that is no longer relevant upon graduation, but it also increases opportunities for the students who can create a unique niche for themselves.

 

[Andy Resnick]

<snip>Most heat transfer and fluids problems end up being extremely specific and there isn't a general rule that you can apply to 100 different places.

This is clearly false, and possibly reflects an underlying ignorance of the relevant science.

 

[atyy]

I don't really understand the approach that 'GR is too mathy, we cannot teach it to undergraduates.' It seems to me absolutely unacceptable that an undergraduate go through his education and be called a physics major without having learned, at least at a basic level, GR. I see no reason why courses based on Hartle or Schutz shouldn't be required just as much as a course in QM at the undergraduate level. Obviously all physics subjects have many layers of understanding, but strangely many people seem to think that if you're not going to teach GR with a full introduction to differential geometry then it's not even worth teaching.

Nonsense.

You are right. But only because you are young:) Newton's laws of gravity used to be cutting edge. Now, they are high school physics. (I confess the find the non-Gauss's law proof that the gravitational field inside a shell is zero to be difficult.)