What could be an example of an anti-hermitian matrix $i\hat{H}t$ , which satisfies the matrix exponential $$e^{i\hat{H}t/\hbar}$$ being a unitary matrix?
-
2The matrix exponential is not unitary here. – Angina Seng Aug 04 '19 at 07:29
-
2The matrix is not unitary here because A is not equal to -A^+ – Federico Fallucca Aug 04 '19 at 07:30
-
1@LordSharktheUnknown essentially I am trying to find a 2 x 2 matrix for $i\hat{H}t$ which is anti-hermitian, hence $\hat{H}$ is hermitian, in the exponential matrix $e^{i\hat{H}t/\hbar}$. So that the exponential in the end will be unitary. Essentially this exponential matrix is a time evolution operator from QM... so how can I find a antihermitian matrix for $i\hat{H}t$ so that I can proof the exponential to be unitary? – EPIC Tube HD Aug 04 '19 at 08:07
-
1I'll edit the question to make it more clear! – EPIC Tube HD Aug 04 '19 at 08:07
-
Related. – Rodrigo de Azevedo Mar 13 '21 at 16:27
1 Answers
For any skew hermitian matrix $S$,
$S^\dagger = -S, \tag 1$
we have
$(e^S)^\dagger = e^{S^\dagger} = e^{-S}, \tag 2$
as follows by applying the adjoint operation ${}^\dagger$ term-by-term to the power series
$e^S = \displaystyle \sum_0^\infty \dfrac{S^n}{n!}, \tag 3$
viz
$(e^S)^\dagger = \displaystyle \sum_0^\infty \dfrac{(S^n)^\dagger}{n!} = \sum_0^\infty \dfrac{(S^\dagger)^n}{n!} = e^{S^\dagger}; \tag 4$
therefore, from (2),
$(e^S)^\dagger e^S = e^S (e^S)^\dagger = I, \tag 5$
that is, $e^S$ is unitary.
Now for hermitian $H$,
$H^\dagger = H, \tag 6$
it follows that
$(iHt/\hbar)^\dagger = -iH^\dagger t /\hbar = -iHt/\hbar, \tag 7$
i.e., $iHt/\hbar$ is skew-hermitian; we conclude that $e^{iHt/\hbar}$ is unitary.
It is then a simple matter to present a large host of examples illustrating these arguments; we merely need pick any hermitian $H$ and form $e^{iHt/\hbar}$, for instance, with
$H = \begin{bmatrix} a & bi \\ -bi & a \end{bmatrix} = H^\dagger, \; a, b \in \Bbb R, \tag 8$
$iHt/\hbar = \begin{bmatrix} iat/\hbar & -bt/\hbar \\ bt/\hbar & iat/\hbar \end{bmatrix} = (-iHt/\hbar)^\dagger \tag 9$
is skew-, and so
$e^{iHt/\hbar}= \exp \left ( \begin{bmatrix} iat/\hbar & -bt/\hbar \\ bt/\hbar & iat/\hbar \end{bmatrix} \right) \tag{10}$
is unitary.
- 72,871