1
$\begingroup$

In the book Bose-Einstein Condensation by Pitaevski, Lev; Petrovitch, and Sandro Stringari (Oxford University Press), the Hamiltonian for weakly interacting Bose gas reads as, $$H=\sum\dfrac{p^2}{2m}\hat{a}_p^\dagger \hat{a}_p+\dfrac{1}{2V}\sum_{p_1,p_2,q} V_q \hat{a}_{p_1+q}^\dagger \hat{a}_{p_2-q}^\dagger \hat{a}_{p_1} \hat{a}_{p_2}\tag{1}\label{1}$$ In the book they have assumed $V_q=V_0$ to be independent of $q$. Expanding the second term one can write the following, $$\dfrac{V_0}{2V} \hat{a}^\dagger_0 \hat{a}^\dagger_0 \hat{a}_0 \hat{a}_0+\dfrac{V_0}{2V}\sum_{q\neq0}(\hat{a}^\dagger_q\hat{a}_q+\hat{a}^\dagger_{-q}\hat{a}_{-q}+\hat{a}^\dagger_{q}\hat{a}^\dagger_{-q}+\hat{a}_{q}\hat{a}_{-q})\\ =\dfrac{V_0}{2V} \hat{a}^\dagger_0 \hat{a}^\dagger_0 \hat{a}_0 \hat{a}_0+\dfrac{V_0}{2V}\sum_{q\neq0}(2\hat{a}^\dagger_q\hat{a}_q+\hat{a}^\dagger_{q}\hat{a}^\dagger_{-q}+\hat{a}_{q}\hat{a}_{-q}) $$

However, in a thesis I saw $V_q$ was a function of $q$, although it was not specified what kind of function it was. In that case, the second term was expanded as follows, $$\begin{eqnarray} &&\dfrac{V_0}{2V} \hat{a}^\dagger_0 \hat{a}^\dagger_0 \hat{a}_0 \hat{a}_0+\dfrac{1}{2V}\sum_{q\neq0}[V_q(\hat{a}^\dagger_q\hat{a}_q+\hat{a}^\dagger_{-q}\hat{a}_{-q}+\hat{a}^\dagger_{q}\hat{a}^\dagger_{-q}+\hat{a}_{q}\hat{a}_{-q})]\tag{2}\label{2}\\ &=&\dfrac{V_0}{2V} \hat{a}^\dagger_0 \hat{a}^\dagger_0 \hat{a}_0 \hat{a}_0+\dfrac{1}{2V}\sum_{q\neq0}[V_q(2\hat{a}^\dagger_q\hat{a}_q+\hat{a}^\dagger_{q}\hat{a}^\dagger_{-q}+\hat{a}_{q}\hat{a}_{-q})]\tag{3}\label{3} \end{eqnarray} $$ My question: Is it valid to go from \eqref{2} to \eqref{3} even when $V_q$ may be of indefinite parity(neither purely even nor purely odd)? In short, is the following true if the operators are bosonic? $$\sum_{q\neq0}V_q(\hat{a}^\dagger_q\hat{a}_q+\hat{a}^\dagger_{-q}\hat{a}_{-q})=2\sum_{q\neq0}V_q\hat{a}^\dagger_q\hat{a}_q$$

$\endgroup$
2
$\begingroup$

In the book "Bose-Einstein Condensation" (page 27) the author wrote that $V(\vec{r})$ is a two-body potential.

We can therefore assume that this potential only depends on the distance between particles $|\vec{r}|$ (I don't have in mind another possible kind of two-body potential). So we have $V(\vec{r})=V(-\vec{r})$. And for the Fourier transform (I do it in 1D for simplicity of notations) :

$\begin{align}V_{-q}&=\int\limits_{-\infty}^{+\infty}dxe^{+iqx}V(x)\\ &=\int\limits_{+\infty}^{-\infty}(-dx)e^{-iqx}V(-x) \\&= \int\limits_{-\infty}^{+\infty}dxe^{-iqx}V(x) \\ &=V_{q} \end{align}$

where I did the change of variable $x\rightarrow -x$ in the second line

$\endgroup$

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.