Electromagnetism textbooks

[Falgun]

I have always been interested in learning more about electromagnetism after going through Resnick Halliday Krane 5th edition. Upon reading a few ( read quite a lot) of E&M book threads, I have come to realize that the following texts are often pitched as alternatives to each other:

1. Griffiths (4th edition)
2. Purcell & Morin (3rd edition)

My question is why not use them together instead of using one of them as an introduction and one as a follow up?
I understand that Purcell is a bit less advanced than Griffiths but by very less. And Purcell covers more topics it seems. So wouldn't it be repetitive to read them one by one ?

Is the organisation of topics/approach taken so different that they can't be used together?

That is, if I use both would I have to use 1 as a main text and jump around the other book?
Can I read both of them side by side, chapter by chapter without getting confused?

 

[hutchphd]

If the subjects are not presented in lockstep order this will be confusing? There are glossaries in books for this reason. You should follow the development of a particular book I suppose (or visit something like MIT Open Courseware for guidance which is never a bad idea)
I do not know Purcell but taught from Griffiths and like it. I would also suggest Feynman's lectures for light reading.
One source of confusion shifting texts may occur if some are
SI and some are cgs units. I usually got (get) bewildered at some point

 

[jasonRF]

I have always been interested in learning more about electromagnetism after going through Resnick Halliday Krane 5th edition.

By 'going through' I assume that means you have solved a bunch of the homework problems. Let us know if that isn't correct.

Anyway, do you already know special relativity? If not, then you are not ready to read Purcell and Morin. Even if you do know relativity I would vote for using Griffiths, and only bothering with Purcell if you really want a second resource. My opinion may be biased since I took a class based on the 2nd edition of Purcell and I found the book to be difficult to learn from; the 3rd edition may be significantly improved. I later purchased the first edition of Griffiths and instantly wished that the course had used it instead.

If you will be purchasing Griffiths, I would recommend getting an earlier edition since they are usually less expensive. However, if you are purchasing Purcell I would recommend the third edition, since it at least has more example problems so may be a little better for students. It also uses SI units, so equations would look the same as they do in Griffiths (the 2nd edition of Purcell uses cgs).

EDIT. sorry for the rant. To get back to the question, of course it is possible to try and simultaneously read both. The chapters of course do not align but that is obvious from the tables of contents, but you should be able to sort that out. Whether or not that will make for a better learning experience depends completely on you - your learning style, how disciplined you are, how much time you have available to spend, etc.

jason

 

[robphy]

...I have come to realize that the following texts are often pitched as alternatives to each other:

1. Griffiths (4th edition)
2. Purcell & Morin (3rd edition)

Depending on the institution,
then they could be alternatives to each other.

As an undergraduate, we used Purcell in the first intermediate-EM course and Feynman Lectures and the prof's notes in the second.
At my graduate institution, the physics majors had Purcell as their intro-EM course and Griffiths as their intermediate-course. I think there was a second intermediate-EM course.
When I last taught EM at a small liberal arts college, I used Griffiths with some material drawn from Purcell
for an intermediate course. Since there was only one intermediate course, I had to be selective... it wouldn't be practical (and would be expensive) to require both texts.

My question is why not use them together instead of using one of them as an introduction and one as a follow up?
I understand that Purcell is a bit less advanced than Griffiths but by very less. And Purcell covers more topics it seems. So wouldn't it be repetitive to read them one by one ?

Is the organisation of topics/approach taken so different that they can't be used together?

That is, if I use both would I have to use 1 as a main text and jump around the other book?
Can I read both of them side by side, chapter by chapter without getting confused?

I think the two texts can complement each other.
Roughly speaking,
Purcell tries to develop more intuition (which is appropriate for an advanced intro course).
Griffiths develops more mathematical methods with physical intuition.
Their mathematical prerequisites are different.

Of course, you can read corresponding sections to see different views of the same topic
but their chapters don't line up.

 

[MidgetDwarf]

I prefer Kipp: Fundamentals of Electricity and Magnetism, to Purcell. Then follow up with Griffiths. Kipp is a much gentler book than Purcell.

 

[Falgun]

By 'going through' I assume that means you have solved a bunch of the homework problems. Let us know if that isn't correct.

Anyway, do you already know special relativity? If not, then you are not ready to read Purcell and Morin. Even if you do know relativity I would vote for using Griffiths, and only bothering with Purcell if you really want a second resource. My opinion may be biased since I took a class based on the 2nd edition of Purcell and I found the book to be difficult to learn from; the 3rd edition may be significantly improved. I later purchased the first edition of Griffiths and instantly wished that the course had used it instead.

If you will be purchasing Griffiths, I would recommend getting an earlier edition since they are usually less expensive. However, if you are purchasing Purcell I would recommend the third edition, since it at least has more example problems so may be a little better for students. It also uses SI units, so equations would look the same as they do in Griffiths (the 2nd edition of Purcell uses cgs).

EDIT. sorry for the rant. To get back to the question, of course it is possible to try and simultaneously read both. The chapters of course do not align but that is obvious from the tables of contents, but you should be able to sort that out. Whether or not that will make for a better learning experience depends completely on you - your learning style, how disciplined you are, how much time you have available to spend, etc.

jason

I am working through Morin's Classical Mechanics now . As for Resnick Halliday & Krane, I have solved every single problem. So relativity won't be a problem. Also I have single variable/multivariable/vector calculus, ODEs, Linear algebra and an introduction to tensors under my belt.

 

[Falgun]

If the subjects are not presented in lockstep order this will be confusing? There are glossaries in books for this reason. You should follow the development of a particular book I suppose (or visit something like MIT Open Courseware for guidance which is never a bad idea)
I do not know Purcell but taught from Griffiths and like it. I would also suggest Feynman's lectures for light reading.
One source of confusion shifting texts may occur if some are
SI and some are cgs units. I usually got (get) bewildered at some point

This thread was partially inspired by looking at an MIT course on OCW. In that particular course they used Purcell, Griffiths and Feynman! Also in 3rd edition of Purcell, Morin added a lot of problems and changed from cgs to SI. From looking at the assigned readings I found that the professor didn't follow any of the texts linearly and jumped around A LOT. One option would always be to follow the MIT course but I was kinda hoping that I could read both the books simultaneously chapter by chapter. Is that even possible or will I be hopelessly confused afterwards?

 

[Falgun]

I was wondering whether I should create a poll for choosing between Griffiths and Purcell as a first text on E&M. That is sure to bring a lot of divided opinions.

 

[robphy]

I was kinda hoping that I could read both the books simultaneously chapter by chapter. Is that even possible or will I be hopelessly confused afterwards?

Unlike many introductory physics texts (many of which seem like clones with variations of each other),
the more advanced books follow their own chapter arrangement and depth, in tune with the particular author's strengths and interests.

(There's a joke among physics teachers that one could close one's eyes and
open a typical introductory physics textbook to some chapter in the middle...
and guess with some accuracy what that chapter is about.)

 

[marcusl]

I was wondering whether I should create a poll for choosing between Griffiths and Purcell as a first text on E&M. That is sure to bring a lot of divided opinions.

This has been discussed over and over here—just do a search for the threads.

 

[Falgun]

This has been discussed over and over here—just do a search for the threads.

Yes but not a specific poll. But I get that it would be somewhat redundant.

 

[vanhees71]

I would not use Purcell (the Berkeley physics course volume on electromagnetism). It's pretty confusing in an attempt to avoid the necessary math. The "relativity first/early" approach is not bad though. For that I'd rather recommend M. Schwartz, Principles of Electrodynamics. Also Griffiths is very good. As mentioned already above, also the Feynman Lectures (vol. II) give a lot of additional insight.

Concerning the always annoying question of units, I'm a bit undecided. On the one hand SI units are a nuissance, because they introduce the conversion factors $\mu_0$ and $\epsilon_0$ in order to make the introduction of an extra charge (current) unit, the Coulomb (Ampere), consistent. From a relativistic point of view this leads to the idiosyncratic feature of making the electric and magnetic components of the field-strengths tensors, aka $\vec{E}$ and $\vec{B}$, of different dimension. That's why I'd prefer CGS units and there of course the rationalized Heaviside-Lorentz units over the non-rationalized Gaussian units in a course teaching E&M as a relativistic field theory (which is the most elegant, simple and comprehensive way at a more advanced level; in Germany that would be the E&M lecture within the theoretical-physics course, usually taught in the 3rd semester).

On the other hand the SI units are the units used internationally everywhere in experimental physics and engineering. Also the dimensional analysis to check equations is more simple and straight-forward than in CGS units. So I think they should be used in an introductory lecture about E&M (in Germany that's the 2nd semester lecture in the experimental-physics course). To switch between unit systems (SI, Gaussian, Heaviside-Lorentz, natural units) you have to learn anyway simply, because all these units are still used in physics textbooks and scientific papers. The same holds for the different conventions in relativity (east- vs. west-coast, different sign conventions for the curvature and Ricci tensors in GR etc.).

 

[paralleltransport]

I agree with vanhess71. Purcell seems like overly wordy and wishy washy, spending more time "describing" E&M than just teaching you exactly what E&M is.

I would just use griffiths. It's clean, has good examples and problems and can be read cover to cover without any big gaps. I actually think griffiths cover more material than purcell: purcell is a more confusing less systematic treatment of E&M than griffiths at a slightly less mathy level.

Regarding the relativity first treatment of purcell, I actually found that more confusing. Griffiths will teach you relativity in its full blown 4-vector notation including maxwell field tensors which is absolutely foundational to how we understand E&M nowadays. This is the case where just learning the material is easier than circling around it with words. When I study modern things and need a quick refresh I go back go griffiths to remember things. I sold my purcell textbook years ago :)

 

[robphy]

Just to elaborate a point in my first reply above….

Purcell was written as an introductory physics text (just like Kleppner and Kolenkow for intro mechanics)—that is, an advanced-level PHY 102 course.

Griffiths was written as an intermediate (typically junior or senior level) text.

Griffiths was not written as an alternative to the introductory text of Purcell.They are comparable as intermediate texts … but only for places that did not use Purcell (or similar) as an introductory PHY 102 text… which is typical, like in my undergraduate institution.

The issue is whether such students are ready for Griffiths…. Or if they need Purcell (or parts of Purcell) first
….with the likely constraint in smaller institutions that there is only one intermediate-level course.
(One could use parts of Purcell alongside Griffiths.)

 

[vanhees71]

Really? I always considered Purcell as supposed to be not at the introductory (first experimental-physics course lecture on electromagnetism) but at the next level (first theoretical-physics course lecture). For both purposes I don't think it's a good text (it's not a good text at all), but that's of course a personal opinion only, because it obscures the physics by avoiding to introduce the appropriate mathematical tools for its intended purpose to present electromagnetism in a relativity-first approach and instead making many confusing words.

 

[robphy]

When I was a grad student/teaching assistant at UChicago,
the physics majors took the "honor-physics sequence", which had Kleppner for intro Mechanics and Purcell for intro EM. (At UChicago, they had three concurrent sequences of calculus-based physics: the 120s, 130s, and 140s. I'm not sure what they have today.)

From Purcell's preface [whose first edition was from 1965]

This revision of "Electricity and Magnetism," Volume 2 of the
Berkeley Physics Course, ...

(On the Berkeley Physics Course series...)
This is a two-year elementary college physics course for students
majoring in science and engineering. The intention of the writers has
been to present elementary physics as far as possible in the way in
which it is used by physicists working on the forefront of their field.
We have sought to make a course which would vigorously emphasize
the foundations of physics. Our specific objectives were to introduce
coherently into an elementary curriculum the ideas of special
relativity, of quantum physics, and of statistical physics.

This course is intended for any student who has had a physics
course in high school. A mathematics course including the calculus
should be taken at the same time as this course.

From Griffith's preface [whose first edition was from 1981]

This is a textbook on electricity and magnetism, designed for an undergraduate course at the junior or senior level.

To further suggest that these are intended for different audiences
and thus should not be compared as comparable texts:

Both texts mention the Laplace Equation,​

but only Griffiths discusses the Legendre polynomials.​

(Note:
The standard introductory calculus-based texts for scientists and engineers (e.g. Giancoli and Serway) do not mention the Laplace Equation.
A newer intro calc-based text that does mention the Laplace equation is Chabay and Sherwood's Matter and Interactions.)

Clearly, Purcell is an advanced-level introductory (PHY 102) calculus-based textbook.
It's not fair to compare it with Griffiths.

 

[paralleltransport]

When I was a grad student/teaching assistant at UChicago,
the physics majors took the "honor-physics sequence", which had Kleppner for intro Mechanics and Purcell for intro EM. (At UChicago, they had three concurrent sequences of calculus-based physics: the 120s, 130s, and 140s. I'm not sure what they have today.)

From Purcell's preface [whose first edition was from 1965]From Griffith's preface [whose first edition was from 1981]To further suggest that these are intended for different audiences
and thus should not be compared as comparable texts:

Both texts mention the Laplace Equation,​

but only Griffiths discusses the Legendre polynomials.​

(Note:
The standard introductory calculus-based texts for scientists and engineers (e.g. Giancoli and Serway) do not mention the Laplace Equation.
A newer intro calc-based text that does mention the Laplace equation is Chabay and Sherwood's Matter and Interactions.)

Clearly, Purcell is an advanced-level introductory (PHY 102) calculus-based textbook.
It's not fair to compare it with Griffiths.

Hi Robphy,

The author of the thread already had halliday resnick, in other words an introductory treatment of E&M. The next logical step is an "intermediate" treatment of E&M with full blown laplace equation, boundary value etc... This is what Griffiths is. BTW, note that MIT advanced E&M course for freshmen used griffiths. The prerequisite for a lot of students taking it was multivariable calculus.

Purcell would be an awkward thing to use at an intermediate level because it's sophisticated for freshmen who never saw E&M, but for people who already had let's say physics C in high school, it was too dumbed down.

 

[robphy]

Hi

Hi Robphy,

The author of the thread already had halliday resnick, in other words an introductory treatment of E&M. The next logical step is an "intermediate" treatment of E&M with full blown laplace equation, boundary value etc... This is what Griffiths is. BTW, note that MIT advanced E&M course for freshmen used griffiths. The prerequisite for a lot of students taking it was multivariable calculus.

Purcell would be an awkward thing to use at an intermediate level because it's sophisticated for freshmen who never saw E&M, but for people who already had let's say physics C in high school, it was too dumbed down.

Hi paralleltransport.

I know these books.
As an undergraduate at Stony Brook, I had Purcell as my first junior-level EM course.
(I had Halliday&Resnick and Tipler as my intro calc-based textbooks.)
I had Feynman as my second junior-level EM course.
(Later, in graduate courses, I used Ohanian and Jackson.)

I was a teaching assistant for the UChicago physics majors when they used Purcell.

In the past, as a professor at two small liberal arts colleges, I used Griffiths three times.
As a professor at a small state university, I will use Griffiths next semester.
Since at these small institutions, they did not use Purcell as their intro course,
I have used aspects of Purcell in the beginning alongside the first chapter of Griffiths.

 

[paralleltransport]

Hi

Hi paralleltransport.

I know these books.
As an undergraduate at Stony Brook, I had Purcell as my first junior-level EM course.
(I had Halliday&Resnick and Tipler as my intro calc-based textbooks.)
I had Feynman as my second junior-level EM course.
(Later, in graduate courses, I used Ohanian and Jackson.)

I was a teaching assistant for the UChicago physics majors when they used Purcell.

In the past, as a professor at two small liberal arts colleges, I used Griffiths three times.
As a professor at a small state university, I will use Griffiths next semester.
Since at these small institutions, they did not use Purcell as their intro course,
I have used aspects of Purcell in the beginning alongside the first chapter of Griffiths.

I see.

Did you have to supplement purcell for your junior level E&M course? Things like multipole expansion, boundary value problem, and even relativistic E&M (e.g. faraday tensor) i think just treated more straightforwardly in griffiths. For someone who already knew maxwell equation (from resnick/halliday), i'd want to learn something new, and not a rehash of freshmen E&M with some sophisticated comments sprinkled in.

In my opinion, if one knows div/grad curl + resnick halliday, griffiths is the logical next step. The only redeaming feature I liked about purcell was the problems were pretty cool, a bit like kleppner/kolenkow style.

 

[robphy]

I see.

Did you have to supplement purcell for your junior level E&M course? Things like multipole expansion, boundary value problem, and even relativistic E&M (e.g. faraday tensor) i think just treated more straightforwardly in griffiths. For someone who already knew maxwell equation (from resnick/halliday), i'd want to learn something new, and not a rehash of freshmen E&M with some sophisticated comments sprinkled in.

In my opinion, if one knows div/grad curl + resnick halliday, griffiths is the logical next step. The only redeaming feature I liked about purcell was the problems were pretty cool, a bit like kleppner/kolenkow style.

I don’t recall doing any boundary value problems in that Purcell class… but I did BVPs in a math methods class i took at about the same time. I also had a pde class before I transferred to Stony Brook for my junior year. The 2nd EM class used Feynman and we were mainly concerned with radiation.

If I recall correctly, Halliday and Resnick treated maxwell’s equations in the integral form. (Maybe there was a comment about the differential form with Div grad and curl… but I doubt it.)

While Griffiths (in terms of syllabus content) is the next logical choice after any intro-EM course and a Vector Calculus course, in my experience, students have not necessarily mastered the intro content to dive straight into Griffiths. So, some Purcell seems appropriate (especially with Griffiths Ch 1)… for the students I have worked with.

Certainly there may be students that are well prepared and can move quickly. That’s great. But I have to also teach students that need some help to get on board with Griffiths.

While Griffiths is better than Purcell for relativity (which is understandable because Purcell is an advanced PHY 102 textbook), both lack the spacetime-geometric viewpoint. So, for relativistic electromagnetism, I have to look beyond Griffiths (although Griffiths provides a good start).

 

[paralleltransport]

I don’t recall doing any boundary value problems in that Purcell class… but I did BVPs in a math methods class i took at about the same time. I also had a pde class before I transferred to Stony Brook for my junior year. The 2nd EM class used Feynman and we were mainly concerned with radiation.

If I recall correctly, Halliday and Resnick treated maxwell’s equations in the integral form. (Maybe there was a comment about the differential form with Div grad and curl… but I doubt it.)

While Griffiths (in terms of syllabus content) is the next logical choice after any intro-EM course and a Vector Calculus course, in my experience, students have not necessarily mastered the intro content to dive straight into Griffiths. So, some Purcell seems appropriate (especially with Griffiths Ch 1)… for the students I have worked with.

Certainly there may be students that are well prepared and can move quickly. That’s great. But I have to also teach students that need some help to get on board with Griffiths.

While Griffiths is better than Purcell for relativity (which is understandable because Purcell is an advanced PHY 102 textbook), both lack the spacetime-geometric viewpoint. So, for relativistic electromagnetism, I have to look beyond Griffiths (although Griffiths provides a good start).

I see that make sense.

It always bugged me why introductory E&M texts don't write lorentz transformation with cosh and sinh to reinforce the analogy with rotations and lorentz transformation. The ugly square roots really obscure the similarilies. The relativity first treatment of E&M ideally would come from something like landau vol.2 (this is actually how I learned about 4 vectors and index manipulation), but that requires one to know landau vol. 1 (lagrangian formulation) which many students don't by the time they do intermediate E&M.

 

[marcusl]

I also had Purcell as the text for freshman honors E&M, Schwartz for intermediate E&M and Jackson in grad school. Griffiths hadn’t written his text at that time.