Poulsen Kasper, Majland Marco, Lloyd Seth, Kjaergaard Morten, Zinner Nikolaj T
Department of Physics and Astronomy, Aarhus University, Ny munkegade 120, 8000 Aarhus C, Denmark.
Department of Mechanical Engineering, Massachusetts Institute of Technology, Cambridge, Massachusetts 02139, USA.
Phys Rev E. 2022 Apr;105(4-1):044141. doi: 10.1103/PhysRevE.105.044141.
Maxwell's demon is the quintessential example of information control, which is necessary for designing quantum devices. In thermodynamics, the demon is an intelligent being who utilizes the entropic nature of information to sort excitations between reservoirs, thus lowering the total entropy. So far, implementations of Maxwell's demon have largely been limited to Markovian baths. In our work, we study the degree to which such a demon may be assisted by non-Markovian effects using a superconducting circuit platform. The setup is two baths connected by a demon-controlled qutrit interface, allowing the transfer of excitations only if the overall entropy of the two baths is lowered. The largest entropy reduction is achieved in a non-Markovian regime and, importantly, due to non-Markovian effects, the demon performance can be optimized through proper timing. Our results demonstrate that non-Markovian effects can be exploited to boost the information transfer rate in quantum Maxwell demons.
麦克斯韦妖是信息控制的典型例子，这对于设计量子设备至关重要。在热力学中，妖是一种智能存在，它利用信息的熵性质在储热器之间对激发进行分类，从而降低总熵。到目前为止，麦克斯韦妖的实现主要限于马尔可夫热库。在我们的工作中，我们使用超导电路平台研究这种妖在多大程度上可以借助非马尔可夫效应。实验装置是由一个由妖控制的三量子比特接口连接的两个热库，只有当两个热库的总熵降低时才允许激发的转移。在非马尔可夫 regime 中实现了最大的熵降低，重要的是，由于非马尔可夫效应，可以通过适当的时机优化妖的性能。我们的结果表明，可以利用非马尔可夫效应来提高量子麦克斯韦妖中的信息传输速率。
