I am reading a paper, and I came across the Green-Kubo formulation, where the conductivity $\sigma$ of charged particles is related to the time correlation function of the $z$-component of the collective ionic current $J_z(t)$:

$$\sigma_{GK} = \frac{1}{V k_B T} \int_0^{\infty} dt \; C_{JJ}(t)$$

where $C_{JJ}(t) = \langle J_z(0) J_z(t) \rangle$ and the collective current along the $z$ direction is $J_z(t) = \sum_{i=1}^N q_i v_{z, i}(t)$. $V$ is the volume of the system and $q_i$ and $v_{z, i}$ are the charge and $z$-component of the velocity of the $i$th charged particle. $\langle ... \rangle$ is an equilibrium ensemble average.

My question is, what is the time correlation function? Is the time correlation function $C_{JJ}(t) = \langle J_z(0) J_z(t) \rangle$? Or is the time correlation function the integral: $\int_0^{\infty} dt \; C_{JJ}(t)$?

  • $\begingroup$ Excellent question Andrew! Just so you know, it's normal not to "sign" questions here because there is a link to your profile in the bottom right of every question that serves as a signature of sorts. So I edited out that part of your post. $\endgroup$
    – David Z
    May 8, 2012 at 6:41

1 Answer 1


Your first answer is the correct one. The temporal auto-correlation function (or the time correlation function) of a fluctuating quantity $A(t)$ is $$C_{AA}(\tau) = \langle A(t) A(t+\tau)\rangle$$ This is a measure of how correlated $A(t)$ is to its value at another time $t+\tau$.

The Green-Kubo formula that connects a response function (the electrical conductivity in your case) to the time-integral of a correlation function (the current correlation function in your case) is a general result of linear response theory.

See this as well.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

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