Question about a Problem from Sakurai

  • Thread starter Xyius
  • Start date
  • Tags Tags
    Sakurai
3 replies · 3K views · last reply
Xyius
Messages
501
Reaction score
4
Hello!

I am studying the Sakurai book on Quantum mechanics and I am doing a problem. I have the solutions to the problems to help me understand the material better but I do not understand this solution.

Homework Statement



SEE "Sakurai Problem 1" in attachments

K is the propagator in wave mechanics.


2. Solution

SEE "Sakurai Problem 1" in attachments

There are a few parts of this solution that I do not understand.

1.) In the first part it states that "The probability is.."

[tex]P(Ea')=exp(-βEa')/Z[/tex]

Probability of what? It doesn't actually tell me what the "Partition function" is or means. Isn't the propagator an operator? I thought in order to have a probability you need to have a state in mind.

2.) I also do not understand why the ground state energy is equal to that summation "U=..." in the next line.

3.) I do not understand the first change of variables in the differential, da' = L/2π dk

If anyone could help me understand this, it would be much appreciated! :D
 
Attachments
  • Textbook problem 2.32: define the partition function Z as the integral of d3x' K(x',t;x',0) evaluated at beta = it/hbar. Asks to show the ground-state energy is -(1/Z)(dZ/d beta) as beta goes to infinity, illustrated for a particle in a one-dimensional box.
    sakurai problem 1.png
    10.5 KB · Views: 604
  • Handwritten derivation of a partition function and ground state energy for a particle in a one-dimensional box with periodic boundary conditions, ending with the result 1/2β.
    sakurai problem 2.jpg
    52.8 KB · Views: 587
If beta were an inverse temperature, P would be the probability to find E_a' in a canonical ensemble.
The precise probability interpretation does not matter too much, as all you want to show is that the limit on beta gives you the ground state energy.
No, the propagator isn't an operator but a matrix element of the the time evolution operator between states <x',t'| and |x,0>.
 
If you will see the text,you will find that space integral of propagator with x''=x' in K(x'',t;x',t0) will give you G(t)=Ʃa'exp(-iEa't/h-).This is just the trace of time evolution operator and is independent of representation.Now you have to identify β=it/h-(with t imaginary),and you will identify it as partition function.
 
Thank you for the replies.

DrDu said:
If beta were an inverse temperature, P would be the probability to find E_a' in a canonical ensemble.

I still do not seem to understand. How do they actually obtain the expression for probability? I don't understand the reasoning behind it.

andrien said:
If you will see the text,you will find that space integral of propagator with x''=x' in K(x'',t;x',t0) will give you G(t)=Ʃa'exp(-iEa't/h-).This is just the trace of time evolution operator and is independent of representation.Now you have to identify β=it/h-(with t imaginary),and you will identify it as partition function.

I didn't even know what a partition function was until I just looked it up. I never took statistical mechanics. :\