2
$\begingroup$

In real space we write basis vector for spinless fermions in binary notation for example if there are M=4 sites in system and N=2 fermions then basis vectors will be: $0011, 0101, 0110, 1001, 1010, 1100$. Hamiltonian in numerical form ($H=-t\sum_{<j,j+1>}(c_j^\dagger c_{j+1}+h.c.)+U\sum_{<j,j+1>}n_jn_{j+1}$) can be written simply using bitwise operations of C/C++, Fortran or MATLAB. One can see hopping part of H is off-diagonal and interaction part is diagonal in real space.

When we work in Fourier space Hamiltonain become $$\tilde{H}=\sum_k\epsilon_k\tilde{c_k^\dagger}\tilde{c_k} + \sum_k\tilde{U_k}\tilde{n_k}\tilde{n_{-k}}$$ with $\epsilon_k=-2t\cos{k}$ and $\tilde{U}_k=\frac{1}{L}\sum_j U(j) e^{-ik.j}$ as explained in this pdf.


What I can't understand is that how do we define our basis vector in fourier space?


My understanding about it:

What I have understood from this so for is that let we have a 1D line from $-\pi$ to $+\pi$ (first brillion zone) on which $k$ points are discreetly define. If we have M=4 and N=2 then set of $k$-points is $-\pi$, $-\frac{\pi}{2}$,$+\frac{\pi}{2}$, $+\pi$
Now considering these 4 points as sites on which fermions can reside our basis vectors can be again given as they were given in real space i.e. $0011, 0101, 0110, 1001, 1010, 1100$.
For simplicity I take limit $U=0$ and calculate Hamiltonian for both real and fourier space case.
REAL SPACE:
$$H_{R}=-t\begin{bmatrix} 0 & 1 & 0 & 0 & -1 & 0 \\ 1 & 0& 1& 1& 0& -1\\ 0 & 1& 0& 0& 1& 0\\ 0 & 1& 0& 0& 1& 0\\ -1 & 0& 1& 1& 0& 1\\ 0 & -1& 0& 0& 1& 0\\ \end{bmatrix}$$ Let t=1 then Eigenvalues=[-2, -2, -4.4e-16, 0, 2, 2] (using MATLAB function eig())

FOURIER SPACE:

$\tilde{c_k^{\dagger}}\tilde{c_k}=\tilde{n_k}=$ number operator in k-space. So our hamiltonian for U=0 should be diagonal with values $$ H_{F}= -2t*diagonal[\cos{(\pi/3)}+\cos{\pi}, \cos{(-\pi/3)}+\cos{\pi}, \cos{(-\pi/3)}+\cos{(\pi/3)}, \cos{(-\pi)}+\cos{\pi}, \cos{(-\pi)}+\cos{(\pi/3)}, \cos{(-\pi)}+\cos{(-\pi/3)}] $$

$$=-t\begin{bmatrix} 1 & 0 & 0 & 0 & 0 & 0 \\ 0 & 1 & 0 & 0 & 0 & 0 \\ 0 & 0 & -2 & 0 & 0 & 0 \\ 0 & 0 & 0 & 4 & 0 & 0 \\ 0 & 0 & 0 & 0 & 1 & 0 \\ 0 & 0 & 0 & 0 & 0 & 1 \\ \end{bmatrix}$$

for t=1 eigenvalues=[-2, 1, 1, 1, 1, 4].


results are not matching, I consider there is any fault in my method of defining basis vectors in k-space. So, please guide be how to properly build basis vectors in k-space.

$\endgroup$
  • $\begingroup$ Where did the $\cos(\pi/3)$ come from? $\pi/3$ is not a valid k-vector... $\endgroup$ – Jahan Claes Aug 6 '17 at 20:53
2
$\begingroup$

I think you've made a couple of mistakes in your allowed k-vectors.

First, the allowed k-vectors are not $-\pi,-\frac{\pi}{2},\frac{\pi}{2},\pi$. The allowed k-vectors are $-\frac{\pi}{2},0,\frac{\pi}{2}, \pi$. In the Brilloin zone, $k=\pi$ and $k=-\pi$ are the same state, so you double counted this state while neglecting $k=0$.

Second, for some reason when you computed $H_F$, you wrote terms like $\cos(\frac{\pi}{3})$ on the diagonal. This is clearly an error, since $\frac{\pi}{3}$ is not an allowed k-value. If you write out $H_F$ more carefully, with the correct k-values, you should get the energies to match like you want.

(Note there could also be an error in your $H_R$, I didn't check it too closely. But fix the k-error and see!)

$\endgroup$
  • $\begingroup$ Thank you so much. $H_R$ is correct I wast taking k-points wrong. $\endgroup$ – Luqman Saleem Aug 7 '17 at 11:22
  • $\begingroup$ Here is your upvote sir. At that time I didn't have privileges to upvote any answer. $\endgroup$ – Luqman Saleem Aug 30 '18 at 14:48

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.