PDA

View Full Version : Unitary matrices


Niles
Feb15-10, 04:45 PM
Hi guys

I have been sitting here for a while thinking of why it is that for a unitary matrix U we have that UijU*ji = |Uij|2. What property of unitary matrices is it that gives U this property?

Best,
Niles.

quasar987
Feb15-10, 07:03 PM
This is true of any matrix: U^*=\overline{U}^T. So U^*_{ji}=\overline{U_{ij}}. And of course, for any complex number z, we have z\overline{z}=|z|^2

Niles
Feb16-10, 02:13 AM
By an asterix I meant complex conjugation, so
(U^\dagger )_{ij} = (U_{ji})^*
. Is it still valid then?

quasar987
Feb16-10, 06:52 AM
It is not valid in this case. Take for example the rotation matrix

cos(t) -sin(t)
sin(t) cos(t)

It is orthogonal, hence unitary. But for any sin(t) different from 0, we have U_{12}U^*_{21}=-\sin^2(t)\neq |\sin(t)|^2=|U_{12}|^2.

Niles
Feb17-10, 07:54 AM
Hmm, I have a problem then. I have a transformation


\mathbf{m} = S\mathbf{a},


which has the components


m_i = \sum_j S_{ij}a_j.


Now I want to find the Hermitian conjugate (I denote this by a dagger, and complex conjugation is denoted by an asterix), and we have


\mathbf{m}^\dagger = \mathbf{a}^\dagger S^\dagger,


which has the components


m^\dagger_i &= \sum_j a_j^\dagger (S^\dagger)_{ij} \\
&=\sum_j a_j^\dagger (S^*)_{ji}.


My teacher says the last step is wrong, but I cannot see why. Can you help me spot the error?

quasar987
Feb17-10, 09:00 AM
Your mistake is when you say

m^\dagger_i &= \sum_j a_j^\dagger (S^\dagger)_{ij}

According to the definition of matrix multiplication, correct is

m^\dagger_i &= \sum_j a_j^\dagger (S^\dagger)_{ji}

Niles
Feb17-10, 09:21 AM
Ahh, I see it now. Of course the column has to be fixed, not the row. Thanks.