Comment on "Carnot efficiency at divergent power output" (and additional discussion)

In a recent Letter [EPL, 118 (2017) 40003], Polettini and Esposito claimed that it is theoretically possible for a thermodynamic machine to achieve Carnot efficiency at divergent power output through the use of infinitely-fast processes. It appears however that this assertion is misleading as it is not supported by their derivations as demonstrated below. In this Comment, we first show that there is a confusion regarding the notion of optimal efficiency. We then analyze the quantum dot engine described in Ref. [EPL, 118 (2017) 40003] and demonstrate that Carnot efficiency is recovered only for vanishing output power. Moreover, a discussion on the use of infinite thermodynamical forces to reach Carnot efficiency is also presented in the appendix.