Let $\widehat{a}^{+}_{i}$ and $\widehat{a}_{i}$ be the usual bosonic creation and annihilation operators. Consider $$\widehat{q}_{i} = \sqrt{\frac{\hbar}{2m_{i}w_{i}}}(\widehat{a}_{i}+ \widehat{a}^{+}_{i}).$$ We want to calculate time evolution of $\widehat{q}_{i}(\tau)$ in the interaction picture: $$\widehat{q}_{i}(\tau) = e^{i \widehat{H}_{0} \tau/\hbar} \widehat{q}_{i} e^{-i \widehat{H}_{0} \tau/\hbar},$$ where $$\widehat{H}_{0} = \sum_{i} \frac{1}{2m_{i}} \widehat{p}_{i}^{2} + \frac{1}{2} m_{i} w_{i}^{2} \widehat{q}_{i}^{2},$$ where $m_{i}$ and $w_{i}$ is the mass and natural frequency of the $i$th harmonic oscillator, and $\widehat{q}_{i}$ and $\widehat{p}_{i}$ are the canonical position and momentum operator.

I don't understand why the result is $$\widehat{q}_{i}(\tau) = \sqrt{\frac{\hbar}{2m_{i}w_{i}}}(\widehat{a}_{i} e^{-i w_{i} \tau/\hbar} + \widehat{a}_{i}^{+} e^{i w_{i} \tau/\hbar}). $$


Compute the time dependence of $a$ in the interaction picture first (no hats here, because it's too much typing). Let $a_I$ denote the lowering operator in the interaction picture. Define the following symbols:

  1. $H_i \equiv \hbar \omega \left( a_i^{\dagger}a_i + 1/2 \right)$

  2. $H_0 \equiv \sum_i H_i$

  3. $K_i \equiv H_i/\hbar$.

  4. $K \equiv H_0/\hbar$

  5. $\bar{a}_i \equiv e^{i K \tau}a_ie^{-i K \tau}$, this is the thing you would probably notate as $\hat{a}_i(\tau)$.

Now we compute the time dependence of $\bar{a}_j$

$$\begin{align} \bar{a}_j &= e^{i K \tau} a_j e^{-i K \tau} \\ \dot{\bar{a}}_j &= e^{i K \tau} a_j (-i K) e^{-i K \tau} + (i K)e^{i K \tau} a_j e^{-i K \tau}\\ &= i \left[ K, e^{i K \tau} a_j e^{-i K \tau} \right] \\ &= i \left[ K, \bar{a}_j \right] \quad (*) \end{align}$$

Suppose we guess that $\bar{a}_j = e^{-i \omega_j \tau}a_j$. Plugging this into $(*)$ we get

$$ \begin{align} \dot{\bar{a}}_j &= i \left[ K, a_j \right] \\ -i\omega_j \bar{a}_j &= \frac{i}{\hbar} \sum_i \left[ H_i, \bar{a}_j \right] \\ &= \frac{i}{\hbar}\sum_i e^{-i\omega_j \tau} \left[ \hbar \omega_i \left( a_i^{\dagger} a + 1/2 \right), a_j\right] \\ 1/2\text{ commutes with }a_j \qquad &= i \sum_i \omega_i e^{-i\omega_j \tau} \left[ a_i^{\dagger} a_i, a_j\right] \\ \text{commutation relation}\qquad &= i\sum_i\omega_ie^{-i \omega_j \tau}(-\delta_{ij}a_i) \\ &= -i \omega_j e^{-i\omega_j \tau} a_j \\ &= -i \omega_j \bar{a}_j \end{align} $$ which is consistent with our guess. Therefore, we have proved that $\bar{a}_j = e^{-i\omega_j \tau}a_j$. By complex/hermetian conjugation you find $\bar{a}^{\dagger}_j = e^{i\omega_j \tau} a_j^{\dagger}$.

The expression for $q_i$ is now obtained by writing $q$ in terms of $a$ and $a^{\dagger}$, as you already did, and then substituting the time dependence we just computed.

| cite | improve this answer | |
  • $\begingroup$ Here $H_{0}$ is actually the sum of the terms $\hbar w (a_{i}^{+} a_{i} + 1/2)$... $\endgroup$ – nchar Sep 12 '14 at 15:33
  • $\begingroup$ @nchar: Yeah, but the existence of multiple modes doesn't change anything in the calculation. $\endgroup$ – DanielSank Sep 12 '14 at 19:28
  • $\begingroup$ @ncha: Oh sorry, didn't realize you were OP. Do you need an explanation of how to deal with the indices on $a_i$? $\endgroup$ – DanielSank Sep 12 '14 at 19:29
  • $\begingroup$ @nchar: Added more detail, in particular showing how the commutator takes care of all those extra indices for you. $\endgroup$ – DanielSank Sep 12 '14 at 21:13

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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