The intriguing mass transport properties of carbon nanotubes (CNTs) have received widespread attention, especially the rapid transport of water through CNTs due to their atomically smooth wall interiors. Extensive research has been dedicated to the comprehension of various aspects of water flow in contact with CNTs, the most prominent ones being the studies on slip and flow rates. Experimental and computational studies have confirmed an enhanced water flow rate through this graphitic nanoconfinement. However, a quantitative agreement has not yet been attained. These disparities coupled with incomplete knowledge of the mechanisms of water transport at nanoscale regimes are hindering the possibilities to integrate CNTs in numerous nanofluidic applications. In the present review, we focus on the slip and flow rates of water through CNTs and the factors influencing them. We discuss the key sources of discrepancies in water flow rate and suggest directions for future study.

We introduce an analytical method to predict the slip length (L _s) in cylindrical nanopores using equilibrium molecular dynamics (EMD) simulations, following the approach proposed by Sokhan and Quirke for planar channels [39]. Using this approach, we determined the slip length of water in carbon nanotubes (CNTs) of various diameters. The slip length predicted from our method shows excellent agreement with the results obtained from nonequilibrium molecular dynamics (NEMD) simulations. The data show a monotonically decreasing slip length with an increasing nanotube diameter. The proposed EMD method can be used to precisely estimate slip length in high slip cylindrical systems, whereas, L _s calculated from NEMD is highly sensitive to the velocity profile and may cause large statistical errors due to large velocity slip at the channel surface. We also demonstrated the validity of the EMD method in a BNNT-water system, where the slip length is very small compared to that in a CNT pore of similar diameter. The developed method enables us to calculate the interfacial friction coefficient directly from EMD simulations, while friction can be estimated using NEMD by performing simulations at various external driving forces, thereby increasing the overall computational time. The EMD analysis revealed a curvature dependence in the friction coefficient, which induces the slip length dependency on the tube diameter. Conversely, in flat graphene nanopores, both L _s and friction coefficient show no strong dependency on the channel width.

Although the importance of temperature control in nonequilibrium molecular dynamics simulations is widely accepted, the consequences of the thermostatting approach in the case of strongly confined fluids are underappreciated. We show the strong influence of the thermostatting method on the water transport in carbon nanotubes (CNTs) by considering simulations in which the system temperature is controlled via the walls or via the fluid. Streaming velocities and mass flow rates are found to depend on the tube flexibility and on the thermostatting algorithm, with flow rates up to 20% larger when the walls are flexible. The larger flow rates in flexible CNTs are explained by a lower friction coefficient between water and the wall. Despite the lower friction, a larger solid-fluid interaction energy is found for flexible CNTs than for rigid ones. Furthermore, a comparison of thermostat schemes has shown that the Berendsen and Nosé-Hoover thermostats result in very similar transport rates, while lower flow rates are found under the influence of the Langevin thermostat. These findings illustrate the significant influence of the thermostatting methods on the simulated confined fluid transport.