1
$\begingroup$

Studying about fermionic commutation relations, the convention I'm following is to consider a set of creation (destruction) operators $\hat{a}_{i}^{\dagger}\left(\hat{a}_{i}\right)$ with $i=1,...,n$ acting on some Hilbert space $V$ where canonical commutation relations are satisfied. Then it is said that the number operators $\hat{n}_{i}=\hat{a}_{i}^{\dagger}\hat{a}_{i}$ form a mutually commuting set of Hermitian matrices and thus there exist $2^{n}$ states $\left|\psi\right\rangle $ which are simultaneous eigenstates of all $n$ of these number operators, spanning a vector space $W$. These notes read that $W_{\perp}$ is the orthocomplement of $W$ in $V$.

My confusion is in distinguishing between $V$, $W$, and $W_{\perp}$. Can you help me clear this out? A short answer is enough, maybe with a very simple example QM system?

$\endgroup$

1 Answer 1

1
$\begingroup$

The vector subspace $W$ will be spanned by states that can be labeled as follows: \begin{align} |\vec k\rangle = |k_1, k_2, \dots, k_n\rangle \end{align} where $k_i = 0$ tells you that there is no fermion present in "mode" $i$ and $k_i = 1$ tells you that there is one fermion present in "mode" $i$. There are $2^n$ such states because there are $2^n$ possible binary sequences of length $n$. These states are defined by acting an appropriate composite of creation operators on the ground state $$ |k_1, k_2, \dots, k_n\rangle = (a_1^\dagger)^{k_1}(a_2^\dagger)^{k_2}\cdots (a_n^\dagger)^{k_n}|0, 0, \dots, 0\rangle $$ where the ground state is defined as the state that is annihilated by every one of the annihilation operators $a_i$. It can be shown that the states $|\vec k\rangle$ are orthogonal; $\langle \vec k| \vec \ell\rangle = \delta_{\vec k, \vec \ell}$. If $W$ is a subspace of Hilbert space $V$ having dimension larger than $W$, then $W_\perp$ will be the set of all states $|\psi\rangle\in V$ that are orthogonal to every state in $W$; $\langle \psi|\vec k\rangle = 0$ for all sequences $\vec k$. This might happen if, for example, one considers a scenario in which the system actually has more than $n$ modes. Say there are actually $N$ possible modes a given fermion can occupy, so in reality the states in $V$ would be written as $$ |k_1, k_2, \dots, k_n, k_{n+1}, \dots, k_N\rangle. $$ Suppose, further, than one defines $$ |k_1, k_2, \dots, k_n, 0, \dots, 0\rangle = |k_1, k_2, \dots, k_n\rangle $$ so that $W$ is a subspace spanned only by the states in which there are no fermions in the higher modes $n+1, n+2, \dots N$. Then any state where there is a fermion in at least one of the higher modes will be in $W_\perp$. As an example, the following state will be in $W_\perp$: $$ |\underbrace{0, 0, \dots, 0}_\text{first $n$ slots}, 1, 0, \dots, 0\rangle. $$

$\endgroup$
3
  • $\begingroup$ Ah ok, so $\hat{a}_{i}^{\dagger}\left(\hat{a}_{i}\right)$ with $i=1,...,n$ act on $V$ but don't span $V$, right? $\endgroup$
    – Bella
    Commented Mar 18, 2018 at 19:16
  • $\begingroup$ @Bella Yes, those are operators on $V$ not states, so that can't span $V$. Moreover in the context you describe, they can act on the ground state in various combinations to generate states that span the subspace $W$, but there might be other, orthogonal states that they cannot generate in this way. $\endgroup$ Commented Mar 18, 2018 at 19:21
  • $\begingroup$ Ok, simple enough, I got it! That subtle difference was escaping me. I was thinking that they were generators for all of the basis states. Thanks $\endgroup$
    – Bella
    Commented Mar 18, 2018 at 19:26

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.