Large deviations of heat flow in harmonic chains
Anupam Kundu, Sanjib Sabhapandit, Abhishek Dhar
Abstract
Open-access reader
Anupam Kundu, Sanjib Sabhapandit, Abhishek Dhar
Abstract
Open-access reader
We consider heat transport across a harmonic chain connected at its two ends to white-noise Langevin reservoirs at different temperatures. In the steady state of this system the heat $Q$ flowing from one reservoir into the system in a finite time $\tau$ has a distribution $P(Q,\tau)$. We study the large time form of the corresponding moment generating function $ \sim g(\lambda) e^{\tau\mu (\lambda)}$. Exact formal expressions, in terms of phonon Green's functions, are obtained for both $\mu(\lambda)$ and also the lowest order correction $g(\lambda)$. We point out that, in general a knowledge of both $\mu(\lambda)$ and $g(\lambda)$ is required for finding the large deviation function associated with $P(Q,\tau)$. The function $\mu(\lambda)$ is known to be the largest eigenvector of an appropriate Fokker-Planck type operator and our method also gives the corresponding eigenvector exactly.
OpenAlex reports 70 citations for this work. Citation counts describe recorded attention and do not establish research quality.
A contribution statement is not available in the OpenAlex record.
Method details are not available in the OpenAlex metadata.
Findings are not separately available in the OpenAlex metadata.
Limitations are not available in the OpenAlex metadata.
Application details are not available in the OpenAlex metadata.
We consider heat transport across a harmonic chain connected at its two ends to white-noise Langevin reservoirs at different temperatures. In the steady state of this system the heat $Q$ flowing from one reservoir into the system in a finite time $\tau$ has a distribution $P(Q,\tau)$. We study the large time form of the corresponding moment generating function $ \sim g(\lambda) e^{\tau\mu (\lambda)}$. Exact formal expressions, in terms of phonon Green's functions, are obtained for both $\mu(\lambda)$ and also the lowest order correction $g(\lambda)$. We point out that, in general a knowledge of both $\mu(\lambda)$ and $g(\lambda)$ is required for finding the large deviation function associated with $P(Q,\tau)$. The function $\mu(\lambda)$ is known to be the largest eigenvector of an appropriate Fokker-Planck type operator and our method also gives the corresponding eigenvector exactly.
Key concepts: Lambda, Eigenvalues and eigenvectors, Physics, Order (exchange), Operator (biology), Mathematical physics, Chain (unit), Second moment of area