Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Generators of SU(2)

  1. Feb 22, 2013 #1
    I am confused with SU(2). How do you prove that the generators are pauli matrices (1/2 sigma)? I would appreciate any link or reference with details of all algebraic steps, taking one from the conditions of unitarity and det=1 to the well know exponential form.

    In general, how do you find generators of arbitrary special unitary group?

  2. jcsd
  3. Feb 22, 2013 #2


    User Avatar
    Science Advisor

    Ballentine, ch7, should provide at least part of what you're looking for.
  4. Feb 23, 2013 #3
    For SU(2),there will be 22-1=3 generators,which are represented by 2×2 matrices.By imposing the condition of unitarity and det=1,one can show that they will be pauli matrices(apart from an overall factor) like done here in a simple way
  5. Feb 24, 2013 #4
    Thanks for replies, I also found it in Arfken.

    But I still wonder, what is the SYSTEMATIC procedure of finding out the generators of arbitrary SU(N)? Do any algorithms exist?

    For example I don't quite understand how to systematically obtain Gell-Mann matrices.
  6. Feb 24, 2013 #5
    The Gell-Mann matrices are a basis for the space of traceless hermitian 3x3 matrices. For SU(N) in general, there are N^2-1 generators. If you look over the Gell-Mann matrices for SU(3) and the Pauli matrices for SU(2) you can probably generalize to come up with a systematic way to enumerate N^2-1 linearly independent traceless hermitian NxN matrixes for any N.
  7. Feb 24, 2013 #6
    How to do that?
  8. Feb 24, 2013 #7
    It's easy just from the definition of the lie algebra of su(n). You need x*=-x and tr(x)=0. Think of the lie algebra as a vector space and show that the pauli spin matrices span it for su(2). So they are the generators. That is to say any 2x2 matrix such that x*=-1 and tr(x)=0 can be spanned by a linear combination of Pauli matrices.
  9. Feb 24, 2013 #8
    I think you're also getting confused between SU(2) and su(2). The Pauli matrices span the lie algebra su(2). The unitary condition on SU(2) imposes the antihermitian on su(2) x*=-x and det U= 1 imposes the tr (x)=0 condition.
  10. Feb 24, 2013 #9
    So there is no rigorous algorithm for general su(n) case?
  11. Feb 25, 2013 #10
    http://en.wikipedia.org/wiki/Genera...eneralized_Gell-Mann_matrices_.28Hermitian.29 It can be shown that these span the lie algebra su(n). There are rigorous algorithms, but that is a subject unto itself. For su(n+1) it's pretty simple since the dimension of the vector space can be shown to be n(n+2) so you need that many linearly independent basis vectors to span it. So take for example su(3) you need 2(2+2)=8 linearly independent antihermitian traceless matrices to span it. So the algorithm is one that can create that many of those kind of matrices for every n.
  12. Feb 25, 2013 #11
    Awesome entry on wiki. Thanks! That sounds like what I was really looking for.
  13. Feb 25, 2013 #12
    Just one more question. How do you prove the commutation relations for generators?

    I understand that one can pick up the generators given by your link and explicitly prove the commutations and then, because it should be independent of representation, state that it's the general form. But nevertheless how to prove them without reference to explicit form of the generators?
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook