Submerged superhydrophobic (SHPo) surfaces are well known to transition from the dewetted to wetted state over time. Here, a theoretical model is applied to describe the depletion of trapped air in a simple trench and re-arranged to prescribe the conditions for infinite lifetime. By fabricating a microscale trench in a transparent hydrophobic material, we directly observe the air depletion process and verify the model. The study leads to the demonstration of infinite lifetime (> 50 days) of air pockets on engineered microstructured surfaces under water for the first time.Environmental fluctuations are identified as the main factor behind the lack of a long-term underwater SHPo state to date.The stability of the air layer on superhydrophobic (SHPo [1]) surfaces fully submerged in water is of critical importance because their key anticipated applications, such as drag reduction [2][3][4][5] and anti-biofouling [6], are under water. Unfortunately, the air film on the underwater SHPo surface, often called a plastron, is fragile [2,7]. Over time, the air initially trapped on the SHPo surface diffuses away into the surrounding water, collapsing the air-water interface and causing a transition from the dewetted to wetted state [8]. Recent experimental studies showed that the lifetime of the underwater SHPo state is influenced by various environmental parameters [7,[9][10][11]. However, most of the studies only reported statistical information, such as average wetting time; more direct knowledge such as air-depletion dynamics or the effect of roughness geometries is needed to design the SHPo surface to be more robust against wetting. While some underwater insects boast a long-term or even indefinite (tested up to 120 days [12]) plastron, all artificial SHPo surfaces retained their plastron for much shorter length of time (mostly less than 2 an hour; rarely for days [7,9,[13][14][15] This paper aims to understand the air depletion process on submerged SHPo surfaces and to determine if an indefinite plastron is achievable. In order to systematically study the effect of geometric parameters of the surface structures for an intended goal, SHPo surfaces made of regular structures would be far more informative than those of random structures that produce only statistical data. Furthermore, for drag-reduction application in particular, SHPo surfaces with parallel trenches [3,4,17,18] have been found to outperform those with random structures [19,20], making SHPo surfaces with trenches a good candidate to study. Since multiple trenches are in parallel and isolated from each other, a single trench can represent the whole SHPo surface as far as the stability of the trapped air is concerned. The potential over-estimation of the depletion speed on single trench compared with the parallel trenches due to the edge effect [21,22] is considered minor to our goal. Importantly for our purpose instead, a sample with single trench would allow clear images of one air-water meniscus; in comparison, multiple menisci in multiple trench...