6
$\begingroup$

In a book, it says, Fock space is defined as the direct sum of all $n$-body Hilbert Space:

$$F=H^0\bigoplus H^1\bigoplus ... \bigoplus H^N$$

Does it mean that it is just "collecting"/"adding" all the states in each Hilbert space? I am learning 2nd quantization, that's why I put this in Physics instead of math.

$\endgroup$
  • 1
    $\begingroup$ Are you asking about what a "direct sum" is or are you asking what the physical motivation to take that direct sum is? $\endgroup$ – ACuriousMind Dec 3 '16 at 15:44
  • 1
    $\begingroup$ en.wikipedia.org/wiki/Direct_sum but you probably read this , and the wikipedia page looks a bit unsure of itself.... $\endgroup$ – user108787 Dec 3 '16 at 16:47
16
$\begingroup$

Suppose you have a system described by a Hilbert space $H$, for example a single particle. The Hilbert space of two non-interacting particles of the same type as that described by $H$ is simply the tensor (aka direct) product

$$H^2 := H \otimes H$$

More generally, for a system of $N$ particles as above, the Hilbert space is

$$H^N := \underbrace{H\otimes\cdots\otimes H}_{N\text{ times}},$$

with $H^0$ defined as $\mathbb C$ (i.e. the field underlying $H$).

In QFT there are operators that intertwine the different $H^N$s, that is , create and annihilate particles. Typical examples are the creation and annihilation operators $a^*$ and $a$. Instead of defining them in terms of their action on each pair of $H^N$ and $H^M$, one is allowed to give a "comprehensive" definition on the larger Hilbert space

$$\Gamma(H):=\mathbb C\oplus H\oplus H^2\oplus\cdots\oplus H^N\oplus\cdots$$

known as the Fock Hilbert space of $H$.

From a physical point of view, the general definition above of Fock space is immaterial. Identical particles are known to observe a definite (para)statistics that will reduce the actual Hilbert space (by symmetrisation/antisymmetrisation for the bosonic/fermionic case etc...).

$\endgroup$
  • 1
    $\begingroup$ Superb answer! I wish they write the QFT textbooks like this. $\endgroup$ – omehoque Feb 17 at 11:54
13
$\begingroup$

Great answers, but just for completeness maybe it will be illustrative to have an example.

Suppose your $H^1$ contains some single-particle states $|a\rangle$, $|b\rangle$, etc. The Fock space removes the limitation on being a single particle, and is composed of $H^0$ (which is 1-dimensional), $H^1$, $H^2 = H \otimes H$, etc. This allows states like

  • the vacuum state, let's call it the empty ket $|\rangle$,
  • all single particle states, $|a\rangle, |b\rangle, \ldots$,
  • all two-particle states, $|aa\rangle, |ab\rangle, |ba\rangle, \ldots$ (NB that this construction deems them distinguishable),

but most importantly

  • any superposition of the above, like $\frac{e^{i\pi/4}}{\sqrt2}|\rangle + \frac12 |a\rangle - \frac12|aab\rangle\otimes\left(\frac1{\sqrt2}|a\rangle + \frac i{\sqrt2}|b\rangle\right)$.

This space is inherently infinite-dimensional even if you start with something small like a qubit. If you want to imagine the result with the help of a basis, simply concatenate the lists of the basis states of all the components:

$$\{|\rangle, |0\rangle, |1\rangle, |00\rangle, |01\rangle, |10\rangle, |11\rangle, |000\rangle, |001\rangle, \ldots\}$$


In the most trivial setting the single particle does not really have any distinct states, so $H^1$ is 1-dimensional. It still makes sense to pick a fiducial state $|{}\circ{}\rangle \in H^1$ and construct the Fock space with basis

$$\{|\rangle =: |0\rangle, |{}\circ{}\rangle =: |1\rangle, |{}\circ{}\circ{}\rangle =: |2\rangle, |{}\circ{}\circ{}\circ{}\rangle =: |3\rangle, \ldots\},$$

an example of a state might be, say, a coherent state

$$|\alpha\rangle = \sum_{n=0}^{\infty} \frac{\alpha^n}{\sqrt{e^{|\alpha|^2}n!}} |n\rangle$$

and you have a nice example of why people can speak of excitations as of "phonons" in a harmonic oscillator even though there's just a single particle oscillating!

$\endgroup$
3
$\begingroup$

Yes, it does. You build a "large" Hilbert space from the "small" ones, if you like.

$\endgroup$

protected by Qmechanic Dec 3 '16 at 21:52

Thank you for your interest in this question. Because it has attracted low-quality or spam answers that had to be removed, posting an answer now requires 10 reputation on this site (the association bonus does not count).

Would you like to answer one of these unanswered questions instead?

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