Work extraction from single-mode thermal noise by measurements: How important is information?

TL;DR

This paper investigates how information obtained through measurements can be used to extract work from a single quantum thermal oscillator, analyzing efficiency, power limits, and optimal measurement strategies without relying on traditional heat baths.

Contribution

It introduces a minimal quantum setup to study work extraction from thermal noise using measurements, highlighting the advantages of observation and feedforward strategies over non-measurement approaches.

Findings

Work extraction is more efficient with measurement-based feedforward.

Non-selective measurements can also enable work extraction without information gain.

Optimized homodyne detection enhances work extraction from thermal noise.

Abstract

Our goal in this article is to elucidate the rapport of work and information in the context of a minimal quantum mechanical setup: A converter of heat input to work output, the input consisting of a single oscillator mode prepared in a hot thermal state along with few much colder oscillator modes. We wish to achieve heat to work conversion in the setup while avoiding the use of a working substance (medium) or macroscopic heat baths. The core issues we consider, taking account of the quantum nature of the setup, are: (i) How and to what extent can information act as work resource or, conversely, be redundant for work extraction? (ii) What is the optimal way of extracting work via information acquired by measurements? (iii) What is the bearing of information on the efficiency-power tradeoff achievable in such setups? We compare the efficiency of work extraction and the limitations of power in our minimal setup by unitary (reversible) manipulations and by different, generic, measurement strategies of the hot and cold modes. For each strategy the rapport of work and information extraction is found and the cost of information erasure is allowed for. The possibilities of work extraction without information acquisition, via non-selective measurements, are also analyzed. Overall, we present, by generalizing a method based on optimized homodyning that we have recently proposed, the following insight: extraction of work by observation and feedforward (WOF) that only measures a small fraction of the input, is clearly advantageous to the conceivable alternatives. Our results may become a basis of a practical strategy of converting thermal noise to useful work in optical setups, such as coherent amplifiers of thermal light, as well as in their optomechanical and photovoltaic counterparts.