Derivation of the density of states?

patric44
Messages
308
Reaction score
40
Homework Statement
i have a question concerning the derivation of the density of states
Relevant Equations
g(E) =sqrt(2)/pi^2*(m/hbar^2)^(3/2)*sqrt(E)
hi guys
i have a question about the derivation of the density of states , after solving the Schrodinger equation in the 3d potential box and using the boundary conditions ... etc we came to the conclusion that the quantum state occupy a volume of ##\frac{\pi^{3}}{V_{T}}## in k space
and to count the total number of quantum states its easier to count them in a shell with thickness dk then integrate , so we have :
$$N_{shell} =\frac{\frac{1}{8}4\pi*k^{2}dk}{\frac{\pi^{3}}{V_{T}}}$$
then integrating to get the total number of states give us :
$$N_{T} =\frac{V_{T}}{(2\pi^{2})}\frac{1}{3}k^{3}$$
and translating that expression in terms of the energy
$$N_{T} =\frac{V}{3}(\frac{\sqrt(2)}{\pi^{2}})(\frac{m}{\hbar^{2}})^{3/2}E^{3/2}$$
now dividing by V and E to get the number of quantum states per unit volume and energy give us :
$$g(E) =\frac{1}{3}(\frac{\sqrt(2)}{\pi^{}2})(\frac{m}{\hbar^{2}})^{3/2}E^{1/2}$$
but that expression doesn't look similer to the standerd one in textbooks
$$g(E) =(\frac{\sqrt(2)}{\pi^{}2})(\frac{m}{\hbar^{2}})^{3/2}E^{1/2}$$
what i am doing wrong here ?
 
Physics news on Phys.org
Isn't ##g(E)## rather what you called ##N_{\text{shell}}## expressed in terms of ##E=\hbar^2 k^2/(2m)##? Note that you also have to express also ##\mathrm{d} k## in terms of ##\mathrm{d} E##. With this you get the textbook result, i.e.,
$$g(E)=\frac{\sqrt{m^3 E}}{\sqrt{2} \hbar^3 \pi^2}.$$
 
vanhees71 said:
Isn't ##g(E)## rather what you called ##N_{\text{shell}}## expressed in terms of ##E=\hbar^2 k^2/(2m)##? Note that you also have to express also ##\mathrm{d} k## in terms of ##\mathrm{d} E##. With this you get the textbook result, i.e.,
$$g(E)=\frac{\sqrt{m^3 E}}{\sqrt{2} \hbar^3 \pi^2}.$$
i am a little bit confused here , isn't ##N_{\text{shell}}## is just the number of quantum states in the shell dk and the density of states is defined as the number of quantum sates per unit volume per unit energy , so i cannot just say that ##\frac{N_{\text{shell}}}{V}## is ##g(E) ## since i need to get the total states by integration then divide by ##V## then divide by ##E## ? , replacing k and dk in ##N_{\text{shell}}## expression gives me :
$$ \frac{V}{2}*\frac{\sqrt{2}}{\pi^{2}}*(\frac{m}{\hbar^2})^{3/2}\sqrt{E}dE $$
which is not slimier to the expression i need :
$$g(E) = \frac{\sqrt{2}}{\pi^{2}}*(\frac{m}{\hbar^2})^{3/2}\sqrt{E}$$
am i misinterpreting something here ?
 
What's meant is by definition
$$g(E)=\frac{1}{V_T} \frac{\mathrm{d} N_{\text{shell}}}{\mathrm{d}E}=\frac{1}{V_T} \frac{\mathrm{d} N_{\text{shell}}}{\mathrm{d}k} \frac{\mathrm{d} k}{\mathrm{d} E}.$$
Your ##N_{\text{shell}}## should be rather ##\mathrm{d} N_{\text{shell}}##.

Your additional factor 2 may be from a degeneracy of states, e.g., if you refer to the em. field in a cavity where for (almost all) ##\vec{k}## you have ##g=2## distinct polarization states.
 
  • Informative
Likes patric44
vanhees71 said:
What's meant is by definition
$$g(E)=\frac{1}{V_T} \frac{\mathrm{d} N_{\text{shell}}}{\mathrm{d}E}=\frac{1}{V_T} \frac{\mathrm{d} N_{\text{shell}}}{\mathrm{d}k} \frac{\mathrm{d} k}{\mathrm{d} E}.$$
Your ##N_{\text{shell}}## should be rather ##\mathrm{d} N_{\text{shell}}##.

Your additional factor 2 may be from a degeneracy of states, e.g., if you refer to the em. field in a cavity where for (almost all) ##\vec{k}## you have ##g=2## distinct polarization states.
oh i think i got it now , so the density of states is just the number of energy states per unit volume confined to that specific dk shell , and the extra degeneracy will cancel the extra ##\frac{1}{2}## that i don't want , isn't that right ?
i am sorry but why is it defined that way ? its a bit confusing , isn't it just more intuitive to define the density of states as the total states in the 1/8 sphere per unit volume per unit energy ?!
 
That would give you the average density whereas you want the density for a given ##E##.
 
  • Like
  • Informative
Likes vanhees71 and patric44
Back
Top