Consider an $N$-level Hamiltonian with energies $\omega_1...\omega_N$ with coupling drives at frequencies $f_{i,j}$ which couple the $i$ and $j$-th levels (not necessarily resonantly, so $f_{i,j} \neq \omega_j - \omega_i$).

This Hamiltonian looks something like

$$H = \sum_{i}^{N} \omega_i \vert i\rangle \langle i \vert + \sum_{i,j} \Omega_{i,j}\cos(f_{i,j} t) \vert i\rangle \langle j \vert + h.c.$$ Or as a matrix

$$H= \begin{bmatrix} \omega_1 & \Omega_{1,2}\cos(f_{1,2} t) & \Omega_{1,3}\cos(f_{1,3} t) & \dots\\ \Omega^*_{1,2}\cos(f_{1,2} t)& \omega_2& \Omega_{2,3}\cos(f_{2,3} t) & \dots \\ \Omega_{1,3}^*\cos(f_{1,2} t) & \Omega^*_{2,3}\cos(f_{2,3} t)&\omega_3 & \dots\\ \vdots & \vdots &\vdots & \ddots \end{bmatrix}$$

The rotating wave approximate to $H$, denoted $H_{RWA}$ would involve first replacing the $\cos(f_{i,j} t)$ terms with $e^{i f_{i,j} t}$

$$H\approx \tilde{H} = \begin{bmatrix} \omega_1 & \Omega_{1,2}e^{if_{1,2} t} & \Omega_{1,3}e^{if_{1,3} t}& \dots\\ \Omega^*_{1,2}e^{-if_{1,2} t}& \omega_2& \Omega_{2,3}e^{if_{2,3} t} & \dots \\ \Omega_{1,3}^*e^{-if_{1,2} t} & \Omega^*_{2,3}e^{-if_{2,3} t}&\omega_3 & \dots\\ \vdots & \vdots &\vdots & \ddots \end{bmatrix}$$

The second step, which actually involves no approximation and is for convenience, is to come up with an appropriate unitary matrix $U$ which gets rid of all (or almost all?) time dependence with

$$H_{RWA}= U^{\dagger} \tilde{H}U + i(\partial_t U^{\dagger}) U$$


Is there an algorithm for going through the frequencies appearing in the Hamiltonian $H$ and coming up with the appropriate rotating frame transform $U$ and obtain $H_{RWA}$ from $H, \tilde{H}$?

It is easy enough to do the RWA when you have two levels, but it is unclear to me how to do it systematically for larger $N$.

  • $\begingroup$ You're confusing two separate concepts: 1) the rotating frame and 2) the rotating wave approximation. Your $U$ is not a "rotating wave transform" (I've never heard that term before). Rather, it's just a change of frame. $\endgroup$
    – DanielSank
    Sep 3, 2023 at 2:29
  • $\begingroup$ @DanielSank thank you. Perhaps I used the wrong term, but is it not true that you have two ingredients: replace the real sinusoid with one one complex exponential, then a unitary transform to get rid of the time dep. My question is about getting $U$ properly $\endgroup$
    – KF Gauss
    Sep 3, 2023 at 2:46

1 Answer 1


A partial answer to your question is the following. Consider a unitary transformation of the form $$ U(t) = \text{diag}\left(e^{ix_1t},\,e^{ix_2t},\, ... \right), $$ the Hamiltonian, written in the rotating frame, is $U^{\dagger}(t) H U(t) + iU^{\dagger}(t)\partial_t U(t)$. A diagonal element at position $i$ is time independent and given by $\omega_i-x_i$; while an off-diagonal element at position $ij$ (with $j>i$) is time dependent, and it reads $$ \Omega_{ij} e^{i(f_{ij} - x_i + x_j)t}. $$ In order to make all of these terms time independent, we get a set of equation of the form $$ x_i - x_j = f_{ij}. $$

This is a set of (at most) $N(N-1)/2$ equations with only $N$ variables, which means that it can't be solved in general. However, under specific circumstances, it can be solved. In fact notice that if one of the couplings $\Omega_{ij}$ vanishes, then we have one less equation to solve, so the number of independent equations is $N(N-1)/2 - n$, where $n$ is the number of couplings that are set to 0.

For example, if $N=2$, then $N(N-1)/2 = 1$ and we have one equation and two variables: $x_1 - x_2 = f_{12}$, which has infinite solutions, for example $x_1 = f_{12}$ and $x_2=0$. Another example is the case $N=3$ with, say, $\Omega_{12}=0$. In this case, the equation $x_1-x_2=f_{12}$ doesn't hold and we just get $$ \left\{ \begin{array}{c} x_1 - x_3 = f_{13} \\ x_2 - x_3 = f_{23} \end{array} \right. $$ which has infinite solutions, for example $x_3=0$, $x_2 = f_{23}$ and $x_1 = f_{13}$.

Of course this is not the most general answer to your question, but at least it paves the way to a first attempt of solution. Hope this helps! :)

  • $\begingroup$ Thank you! There was a typo in my equation, the time derivative for $U$ was in the wrong place, which changes a few of the signs. Anyways, so would you think that in general it should not be possible to find a $U$ which gets rid of all time dep of the RWA form, even if you allow for off diagonal terms? $\endgroup$
    – KF Gauss
    Sep 3, 2023 at 14:00
  • $\begingroup$ Or is there some theorem that RWA Hamiltonians of the form in my question $\tilde{H}$ can always be written in a time dep form after a unitary transformation? $\endgroup$
    – KF Gauss
    Sep 3, 2023 at 14:37
  • $\begingroup$ Unfortunately I don't know what happens when allowing off diagonal terms in $U(t)$. However the trick shown here can be used in some of the most relevant real life situations. $\endgroup$
    – Matteo
    Sep 3, 2023 at 15:02

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.