13
$\begingroup$

If I take a periodic wavefunction $\psi\left(\vec{r}\right)$ and then take the Fourier space dispersion of the wave function as defined below $$ \psi(\vec{k})=\iiint_{-\infty}^{+\infty}\psi\left(\vec{r}\right)e^{-\vec{k}\cdot\vec{r}}\mathrm{d}^3\vec{r} $$

Is there a reason for calling $\psi(\vec{k})$ the momentum space representation of the wavefunction? (I understand the fact that the vector space $\vec{k}$ gets quantized in accordance to the formulation, $\vec{k}\cdot\vec{R}=2\pi$, where in $\vec{R}$ is the lattice translation vector periodicity of $\psi(\vec{r})$ in a crystal lattice), but is there some other reason for calling it momentum space?

$\endgroup$
  • $\begingroup$ Maybe this interests you: math.stackexchange.com/questions/13902/… $\endgroup$ – Constantine Black Jan 3 '18 at 16:18
  • $\begingroup$ I think the appearance of the Fourier transform as the momentum representation associated with the location is the mathematical description of the wave nature of particles. $\endgroup$ – hyportnex Jan 9 '18 at 18:29
  • $\begingroup$ Plane waves $e^{i k r}$ are eigenvectors of the momentum operator with eigenvalues $k$. $\endgroup$ – ZachMcDargh Jan 9 '18 at 18:44
22
$\begingroup$

As per my username, I feel it is partially my responsibility to address this question. I said it before and I'll say it again: The Fourier Transform is not an accident. There are countless reasons it has the precise form it has.

Let $F[f]$ denote the Fourier Transform of $f$, and let $\boldsymbol P=-i\boldsymbol \partial$ denote the momentum operator. We have $$ F[\boldsymbol Pf]=\boldsymbol p F[f]\tag1 $$ so that $F$ diagonalises $\boldsymbol P$. Indeed, the plane-wave basis $\mathrm e^{i\boldsymbol p\cdot \boldsymbol x}$ satisfies $$ \boldsymbol P\,\mathrm e^{i\boldsymbol p\cdot \boldsymbol x}=\boldsymbol p\,\mathrm e^{i\boldsymbol p\cdot \boldsymbol x} \tag2 $$ which automatically implies $(1)$, as claimed.

From this we learn that any operation that includes $\boldsymbol P$ becomes trivial if we work with $F[f]$ instead of with $f$ -- if we work in Fourier space. Thus, Fourier space is known as momentum space. Convenient, right?

In a nutshell, $F[\psi]$ is to momentum what $\psi$ is to position. This is a direct consequence of the (formal) fact that $$ F[\langle \boldsymbol x|]=\langle \boldsymbol p|\tag3 $$ which means that both sides agree when they act on $|\psi\rangle$.

$\endgroup$
7
$\begingroup$

but is there some other reason for calling it momentum space?

The canonical commutation relation for the position and momentum operators is (in one dimension)

$$[X, P]|\psi\rangle = (XP - PX)|\psi\rangle = i\hbar|\psi\rangle$$

On the position basis, this is

$$[x,P_x]\psi(x) = (xP_x - P_xx)\psi(x) = i\hbar\psi(x)$$

and it follows that a position basis representation of the momentum operator is

$$P_x = -i\hbar\partial_x$$


A wavefunction with definite momentum $p$ is then

$$\psi_p(x) = e^{\frac{i}{\hbar}px} = \langle x|p\rangle$$

such that

$$P_x\psi_p(x) = -i\hbar\partial_x\,e^{\frac{i}{\hbar}px} = pe^{\frac{i}{\hbar}px} = p\psi_p(x)$$

Now, if we write

$$k = \frac{p}{\hbar}$$

then

$$\psi_p(x) = e^{ikx}$$


The momentum space wavefunction $\psi(p)$ is just the ket $|\psi\rangle$ projected onto the momentum basis:

$$\psi(p) = \langle p | \psi\rangle = \int\mathrm{d}x\, \langle p | x\rangle\langle x | \psi\rangle$$

where I have used the completeness relation

$$1 = \int\mathrm{d}x\,|x\rangle\langle x | $$ but

$$\psi(x) = \langle x | \psi\rangle$$

and

$$e^{-\frac{i}{\hbar}px} = \langle p| x\rangle$$

thus

$$\psi(p) = \int\mathrm{d}x\,\psi(x)\,e^{-\frac{i}{\hbar}px}$$

or, finally

$$\psi(k) = \int\mathrm{d}x\,\psi(x)\,e^{-ikx} $$

$\endgroup$
4
$\begingroup$

Yet another name justification:

In both crystal and de-Broglie formalism, momentum is defined as $p=\hbar k$. This creates a mapping between a pure wave with wavenumber k and a point in momentum space. If a wavefunction is a superposition of, say, three waves, its momentum representation will consist of three points.

Fourier Transform performs the same mapping. The transform "scans" all possible k and furnishes zero all over, except for the pure wave constituents of the waveform.

$\endgroup$

protected by Qmechanic Jan 3 '18 at 16:06

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.