Supporting Information
A Tunable, Strain-Controlled Nanoporous MoS2 Filter for Water Desalination Weifeng Li1, Yanmei Yang1, Jeffrey K. Weber2, Gang Zhang*3, Ruhong Zhou*1,2,4 1. School for Radiological and Interdisciplinary Sciences (RAD-X) and Collaborative Innovation Center of Radiation Medicine of Jiangsu Higher Education Institutions, Soochow University, Suzhou, China, 215123 2. Computational Biological Center, IBM Thomas J. Watson Research Center, Yorktown Heights, NY 10598, USA 3. Institute of High Performance Computing, A*STAR, Singapore, 138632 4. Department of Chemistry, Columbia University, New York, NY 10027, USA *E-mail: (G.Z.)
[email protected]; (R.H. Z.)
[email protected] 1.
Two Chloride Translocation Events Observed Over the Course of Piston-Driven Simulations For our water flow simulations, we obtained 35 trajectories in total (5 strain cases, each under 7 piston pressures). In the extreme case of the 12%-strained MoS2 filter, two Cl- translocation events were observed (one each under piston pressures of 10 MPa and 100 MPa). As illustrated in Fig. S1, the z-coordinates of the two Cl- ions pass through the MoS2 filter at 23.3 ns (under 10 MPa) and 29.2 ns (under 100 MPa) after binding near the nanopore center for a short while. It bears emphasizing that the saltwater region contains 30 Cl- ions; the occurrence of only one translocation event in each case still suggests this nanoporous MoS2 filter has good salt filtering capabilities.
Figure S1. The z-coordinates of the Cl- ions that have successfully passed through the MoS2 filter under 10 MPa and 100 MPa, respectively. The MoS2 nanopore center is at z = 0; translocation points are highlighted by red arrows. 2.
Gate Size and Deformation under Pressure We have measured two aspects of the deformations of the nanopore under applied RO pressure: (i) the change of nanopore radius (deformation inside the MoS2 pore); and (ii) the deviation of membrane surface in the water flow direction (deformation of the surface, dz, defined as the deviation in the z direction of Mo atoms at the nanopore edge from the idealized flat surface). The definitions are illustrated in Fig. S2a. First, the gate opening is mostly dependent on the strain strength. As demonstrated in the Fig. S2b, the radius of nanopore is almost linearly dependent on the strain strength, from 0.34 nm (free MoS2) to 0.47 nm (12% strained MoS2). Secondly, the applied RO pressure will slightly enlarge both the pore radius (increase ~0.001nm under 12% strain) and dz (less than 0.001nm) as shown in Fig. S2c and S2d. Moreover, the dz reaches a plateau after about 40 MPa under the current settings. For industrial applications, the typical operating RO pressure is usually several MPa, which is at the lower end in our simulations. Thus, the pressure effect on the nanopore structure is mostly negligible.
Figure S2. (a) Definitions of the pore radius and structure deformation in the water flow direction (dz) due to the pressure; (b) the pore radius with respect to strain; (c) pore radius of MoS2 under 12% strain with respect to piston pressure and (d) deformation of MoS2 under 12% strain in the water flow direction with respect to piston pressure. 3.
Calculation of 2D-Projected Water Flux The calculation of 2D-projected water flux illustrated in Fig. 3 in the main text is comprised of 3 steps: 1) The water molecules that passed through the MoS2 filter were identified. 2) The coordinates of the water molecules in step 1 were monitored, and (x, y) coordinates were tabulated when each water molecule was at its closest position in relation to the nanopore center. 3) The 2D distribution of the (x, y) coordinates was calculated and plotted. 4.
Umbrella Sampling Simulation for Potential of Mean Force Profiling In order to probe the filtration mechanism of nanoporous MoS2, we calculated potential of mean force profiles for water or salt ions passing through the MoS2 nanopore with an umbrella sampling method.1-3 For the present system, the z-component of the distance between a molecule and the center of geometry (COG) of the nanopore was selected as the principle collective variable. The value was sampled from -1.0 nm (sea water side) to 1.0 nm (pure water side) with a window resolution of 0.1 nm. In order to enhance the sampling efficiency, the x and y components of the distance were constrained in a range of |dx|