Evaporation-induced cavitation in nanofluidic channels

Abstract

Cavitation, known as the formation of vapor bubbles when liquids are under tension, is of great interest both in condensed matter science as well as in diverse applications such as botany, hydraulic engineering, and medicine. Although widely studied in bulk and microscale-confined liquids, cavitation in the nanoscale is generally believed to be energetically unfavorable and has never been experimentally demonstrated. Here we report evaporation-induced cavitation in water-filled hydrophilic nanochannels under enormous negative pressures up to −7 MPa. As opposed to receding menisci observed in microchannel evaporation, the menisci in nanochannels are pinned at the entrance while vapor bubbles form and expand inside. Evaporation in the channels is found to be aided by advective liquid transport, which leads to an evaporation rate that is an order of magnitude higher than that governed by Fickian vapor diffusion in macro-and microscale evaporation. The vapor bubbles also exhibit unusual motion as well as translational stability and symmetry, which occur because of a balance between two competing mass fluxes driven by thermocapillarity and evaporation. Our studies expand our understanding of cavitation and provide new insights for phase-change phenomena at the nanoscale.nanobubbles | confined fluids | confined water | bubble dynamics | bubble formation

Fabrication of Centimeter Long, Ultra-Low Aspect Ratio Nanochannel Networks in Borosilicate Glass Substrates

“…However, the comphant nature of the adhesive materials limits the aspect ratio compared to glass-glass or glass-silicon devices. The lowest previously reported aspect ratio for glassglass devices was 0.00025 [27], which is an order of magnitude lower than previously reported 0.005 for PDMS-glass devices [28].…”

“…The measured depth of the nanochannels ranged from 16.0nm±0.1nm to 227nm±5nm. Therefore, aspect ratio for the bonded devices ranged from 0.0075 for 227 nm channels down to 0.0005 for 16 nm channels compared to previous lowest report of 0.005 using PDMS-stamp and stick bonding [28]. Design 2 featured a completely nanochannel network.…”

“…The cross section image shows that after bonding the channel depth varies from 62 to 77 nm at different locations along the channel width; however, it is known that the action of cutting glass to measure nanochannel depths can alter the edge profile leading to a variable cross section image [28]. Therefore, a more reliable measurement for channel depth is the AFM trench depth and then channel operation verification by a fiow measurement.…”

“…The channel height varies at different points in the channel from 62 to 77 nm, presumably due to cutting the nanochannels for imaging[28]. Microscope images of nanochannels of various depths filling with isopropyl alcohol in a capillary fill.…”

Nanofluidic transport governed by the liquid/vapour interface

“…The membranes were placed on a droplet (~7 μL) of the photoresist for 2 h to fill the nanopores and then baked on a hot plate at 100 °C for 50 min to evaporate the solvent. The photoresist was then etched by air plasma in a plasma cleaner (Harrick Plasma Inc.) 13 at 7.16 W and ~0.7 mTorr. The membranes were then briefly washed with an aqueous solution of 5wt% phosphoric acid for about 1 min, dried with nitrogen, and exposed to vapour of perfluorodecyltrichlorosilane (Gelest) overnight in a vacuum desiccator.…”