Achievable Information-Energy Exchange in a Brownian Information Engine through Potential Profiling

The information engine extracts work from a single heat bath using mutual information obtained during the operation cycle. This study investigates the influence of potential shaping in a Brownian information engine (BIE) in harnessing the information from thermal fluctuations. We designed a BIE by considering an overdamped Brownian particle inside a confined potential and introducing an appropriate symmetric feedback cycle. We find that the upper bound of the extractable work for a BIE with a monostable centrosymmetric confining potential, with a stable state at the potential center, depends on the bath temperature and the convexity of the confinement. A concave confinement is more efficient than a convex one for an information-energy exchange. For a bistable confinement with an unstable center and two symmetric stable basins, one can find an engine-to-refrigeration transition beyond a certain barrier height related to the energy difference between the energy barrier and the stable basins. Finally, we use the concavity-induced gain in information harnessing to device a BIE in the presence of a multistable potential that can harvest even more energy than monostable confinement.