I'm reading these notes about second quantization. In section 1.4 the author introduces many-particle wavefunctions. But I can't understand how basis are defined here.

I know that if $\{\chi_i | i=1, \dots, N\}$ are single-particle wavefunctions (let's choose fermions) then

$$ \Psi = \frac{1}{\sqrt{N!}} \sum_\sigma \mathrm{sgn} (\sigma) \prod_{j=1}^N \chi_j (\sigma(j)) = \frac{1}{\sqrt{N!}} \left| \begin{matrix} \chi_1(\mathbf{x}_1) & \chi_2(\mathbf{x}_1) & \cdots & \chi_N(\mathbf{x}_1) \\ \chi_1(\mathbf{x}_2) & \chi_2(\mathbf{x}_2) & \cdots & \chi_N(\mathbf{x}_2) \\ \vdots & \vdots & \ddots & \vdots \\ \chi_1(\mathbf{x}_N) & \chi_2(\mathbf{x}_N) & \cdots & \chi_N(\mathbf{x}_N) \end{matrix} \right| \tag{1}$$

is a valid N-electron state. Here $\sigma$ denotes some permutation and $\mathrm{sgn} (\sigma)$ is its signature which ensures the antisymmetry of $\Psi$. This is widely known as Slater determinant.

Now, my question is, how does one choose basis states accordingly? I think that $(1)$ denotes only one state. One would need to find $N-1$ more states that are orthogonal. The notes I mentioned define many-particle wavefunction in some obscure way $(1.113)$ and then use the antisymmetrization procedure on those states $\Psi_j$ to yield a tensor product which they call basis. So I am confused about this explanation.

  • $\begingroup$ Your counting is not quite right. Suppose there are $N$ electrons and $M$ single-particle states which are orthonormal, and $M\geq N$. You take any $N$ different states out of $M$, and construct a Slater determinant. All these states form a complete and orthonormal basis for the many-particle system. $\endgroup$ – Meng Cheng Jul 25 '15 at 14:50

Consider a $N$-dimensional vector space $V$ and let $\{\chi_1, \cdots, \chi_N\}$ be basis of $V$.

Next focus attention on the anti symmetric space $(V\otimes\cdots \otimes V)_A$ where $V$ occurs $M\leq N$ times.

A basis of $(V\otimes\cdots \otimes V)_A$ can be constructed out of $\{\chi_1, \cdots, \chi_N\}$ making use of the projector $$A: V\otimes\cdots \otimes V \to (V\otimes\cdots \otimes V)_A$$ defined as the unique linear extension of $$A (v_1\otimes \cdots \otimes v_M) := \frac{1}{\sqrt{M!}}\sum_{\sigma \in P_M} \mbox{sgn}(\sigma)\: v_{\sigma^{-1}(1)} \otimes \cdots \otimes v_{\sigma^{-1}(M)}\:.$$ Above $P_M$ is the symmetric group of $M$ elements.

The said basis of $(V\otimes\cdots \otimes V)_A$ ($M$ times) is made of the elements for $i_1,\ldots, i_M \in \{1,\ldots,N\}$ $$\chi_{i_1}\wedge \cdots \wedge \chi_{i_M} := A(\chi_{i_1}\otimes \cdots \otimes \chi_{i_M})\mbox{ where} \quad i_1 < i_2< \ldots < i_M\:.$$ In view of the last constarint, these elements are ${N \choose M}$ so $1$ for $N=M$. If instead the constraint $i_1 < i_2< \ldots < i_M$ is dropped $\chi_{i_1}\wedge \cdots \wedge \chi_{i_M}$ results to be the zero vector or a vector already counted up to the sign.

Notice that $A (v_1\otimes \cdots \otimes v_M)$ is nothing but the Slater determinant of the $M$ elements $v_k$.

If $V=H$ is a ($N$-dimensional) Hilbert space and $V\otimes\cdots \otimes V$ is equipped with the analogous induced structure, $(V\otimes\cdots \otimes V)_A$ turns out to be a closed subspace, $A$ an orthogonal projector and, starting form an orthonormal basis $\{\chi_1, \cdots, \chi_N\}$ of $H$, the elements $\chi_{i_1}\wedge \cdots \wedge \chi_{i_M}$ form an orthonormal basi $(H\otimes\cdots \otimes H)_A$ as well. The result easily extends to the case of an infinte-dimensional Hilbert space.

| cite | improve this answer | |
  • $\begingroup$ Say $M = N$. Then there is exactly one element $\chi_{i_1} ∧ ... ∧ \chi_{i_N}$. Shouldn't there be $N$ elements in a basis? $\endgroup$ – Minethlos Jul 25 '15 at 15:08
  • $\begingroup$ No, there is only one element. This is not a basis of the whole space, just of the space of anti symmetric elements which is, in that case, one-dimensional. $\endgroup$ – Valter Moretti Jul 25 '15 at 15:09
  • $\begingroup$ That is the reason why $n$-forms in n-dimensional manifolds are, roughly speaking, scalars and can be used to define volume measures. $\endgroup$ – Valter Moretti Jul 25 '15 at 15:12

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.