Giant Thermal Rectification from Polyethylene Nanofiber Thermal Diodes

TL;DR

This paper demonstrates that polyethylene nanofiber thermal diodes can achieve extremely high thermal rectification with small temperature differences, enabling efficient phononic computing through phase-dependent thermal conductivity.

Contribution

The study introduces a novel polyethylene nanofiber thermal diode design with unprecedented rectification factors driven by phase-dependent thermal conductivity, operable at small temperature differences.

Findings

Thermal rectification factors up to 1.20 achieved.

High dimensionless rectification factors (12-25) surpass other diodes.

Device operates efficiently with temperature differences less than 20°C.

Abstract

The realization of phononic computing is held hostage by the lack of high performance thermal devices. Here we show through theoretical analysis and molecular dynamics simulations that unprecedented thermal rectification factors (as large as 1.20) can be achieved utilizing the phase dependent thermal conductivity of polyethylene nanofibers. More importantly, such high thermal rectifications only need very small temperature differences (< 20 oC) across the device, which is a significant advantage over other thermal diodes which need temperature biases on the order of the operating temperature. Taking this into consideration, we show that the dimensionless temperature-scaled rectification factors of the polymer nanofiber diodes range from 12 to 25 - much larger than other thermal diodes (< 8). The polymer nanofiber thermal diode consists of a crystalline portion whose thermal conductivity is highly phase-sensitive and a cross-linked portion which has a stable phase. Nanoscale size effect can be utilized to tune the phase transition temperature of the crystalline portion, enabling thermal diodes capable of operating at different temperatures. This work will be instrumental to the design of high performance, inexpensive and easily processible thermal devices, based on which thermal circuits can be built to ultimately enable phononic computing.