Electricity Resonance-Induced Fast Transport of Water through

Simulation process detailed. This material is available free of charge via the Internet at http://pubs.acs.org. PDF. nl500664y_si_001.pdf (115.38 kB)...
0 downloads 0 Views 3MB Size
Letter pubs.acs.org/NanoLett

Electricity Resonance-Induced Fast Transport of Water through Nanochannels Jianlong Kou,*,†,‡ Hangjun Lu,† Fengmin Wu,*,† Jintu Fan,§ and Jun Yao*,‡ †

Institute of Condensed Matter Physics, Zhejiang Normal University, Jinhua 321004, China School of Petroleum Engineering, China University of Petroleum (East China), Qingdao 266580, China § Department of Fiber Science and Apparel Design, Cornell University, Ithaca, New York 14853-4401, United States ‡

S Supporting Information *

ABSTRACT: We performed molecular dynamics simulations to study water permeation through a single-walled carbon nanotube with electrical interference. It was found that the water net flux across the nanochannel is greatly affected by the external electrical interference, with the maximal net flux occurred at an electrical interference frequency of 16670 GHz being about nine times as high as the net flux at the low or high frequency range of (80 000 GHz). The above phenomena can be attributed to the breakage of hydrogen bonds as the electrical interference frequency approaches to the inherent resonant frequency of hydrogen bonds. The new mechanism of regulating water flux across nanochannels revealed in this study provides an insight into the water transportation through biological water channels and has tremendous potential in the design of high-flux nanofluidic systems. KEYWORDS: water transport, nanochannel, electricity resonance, frequency

T

how electrical interference influence water permeation through nanochannels. In this Letter, we report on the use of molecular dynamics simulations for investigating the dynamic behavior of water permeation through a nanochannel under varying electrical interferences. We observed that the electrical interference can greatly affect water permeation and bring unexpected high water net flux. It has been reported that a nanochannel of (6, 6) single wall carbon nanotube (SWCNT) with the diameter of 0.81 nm can share the properties with biological channels.18 Here, a (6, 6) SWCNT of 1.34 nm in length was prepared. The simulation framework is shown in Figure 1. In order to avoid the effect of the water molecules outside the SWCNT on the transport of confined water molecules within the SWCNT, two monolayer graphenes with a pore diameter slightly larger than the diameter of SWCNT are placed at the two ends of SWCNT, making them perpendicular to the SWCNT and dividing the space into three parts. To simulate the water permeation through the SWCNT nanochannel with electrical interference, a periodic vibrational positive charge with the charge magnitudes of 1.0e is imposed at the center outside the SWCNT. The charge vibrates periodically along the direction vertical to the surface of the SWCNT, in accordance with the relation y(t) = y0+Acos(2πf t + φ), where y0 is the center of vibration, A is the amplitude, f is

he transportation of water molecules through nanochannels is fundamental to the understanding of biological activities and the design of novel nanofluidic devices.1−3 When water molecules are confined inside nanoscale pores, they usually display fascinating properties and behaviors, including the unusual phase transition,4,5 the extra fast motion,3,6−9 and the excellent on−off gating behavior,10,11 which are completely different from those in bulk systems. Recently, water permeation through the biological protein channels (Aquaporin, Glpf, and MscS) has been studied.12−17 However, the complexity of biological channels in terms of composition, structure, and water-channel interactions makes it difficult, if not impossible, to elucidate the various mechanisms involved. It has been recognized that some behaviors of water molecules confined in simple nanochannels are very similar to those in complicated biological channels, for example, wavelike density distribution, single-file water chain of concerted dipole orientation, and wet−dry transition;18−21 it is therefore beneficial to investigate the fundamental mechanisms of the biological water channels by using a simple nanochannel as a model. It is well known that charged residues exist in the biological protein channels, and they play an important role in water permeation.12−14 Various designs of water channels with charged residues have been proposed to study the properties and behaviors of water,11,22−28 as a tiny signal may bring significant responses in nanoscale systems.29 Biological channels are actually exposed to vibrational electrical signals due to metabolism. Therefore, it is important to understand © 2014 American Chemical Society

Received: February 20, 2014 Revised: July 8, 2014 Published: July 14, 2014 4931

dx.doi.org/10.1021/nl500664y | Nano Lett. 2014, 14, 4931−4936

Nano Letters

Letter

The average net flux and number ⟨N⟩ of water molecules inside the channel in different frequencies are shown in Figure 2. The net flux is defined as the difference between the number

Figure 1. Snapshot of a typical simulation framework. The cyan spheres are the carbon atoms of the SWCNT and graphite sheets. The single blue sphere at the middle of the SWCNT is a periodic vibrational positive charge(q = +1e). The water molecules are depicted by spheres with oxygen in red and hydrogen in white.

Figure 2. Net flux and average number of water molecules inside the SWCNT as a function of vibrational frequency for A = 0.9 Å.

the vibrational frequency, and φ is the initial phase. In this study, we choose A = 0.9 Å and φ = 0. The value of the amplitude was chosen based on the prior knowledge that when the distance between the vibrational charge and the SWCNT is less than 0.2 Å, almost no water molecules can permeate through the channel, and when the distance is greater than 2.0 Å, the nanochannel is likely completely opened.11 To keep the whole system electrically neutral, a negative charge with the same charge magnitude is assigned close to the lower boundary of the system. The molecular dynamics simulations were performed using Gromacs-4.0.730 in the NVT ensemble. The TIP3P water model31 was used. The system temperature was kept at a constant of 300 K and the dynamics process was conducted to allow the system to exchange heat, which was maintained by using a V-rescale thermostat.32 The carbon atoms were modeled as uncharged Lennard-Jones particles. The carbon atoms of SWCNT at the inlet and outlet and those on the two monolayer graphenes were fixed, whereas the other atoms of the SWCNT were flexible. The periodic boundary conditions were employed in all directions. In order to produce a unidirectional flow, an additional acceleration of 0.01 nm ps−2 was applied to each water molecule along z direction33 and a hydrostatic pressure difference of 15 MPa was imposed between the two ends of the SWCNT. (For detailed simulation process, see Supporting Information) Without a pressure gradient, the water molecule moves randomly and average net flux is almost zero. With a pressure gradient, water molecules possess higher probability of moving along the direction of the applied pressure gradient than that of the opposite direction, resulting in a nonzero average net flux along the pressure gradient direction. The net flux is defined in below. We performed a series of molecular dynamics simulations at different electrical interferences to study the influence of external charged residues on water flow. Here, we changed the vibrational frequency ( f = 50 GHz to 100 000 GHz) of charge to simulate the electrical interference. For each system with a different frequency, the simulations were performed for 115 ns, and the results of the last 110 ns of the simulation were collected for analysis.

of water molecules leaving from one end and the other (again having entered from the opposite end) per nanosecond. Interestingly, the average net flux and number ⟨N⟩ are sensitive to the frequency. It clearly shows that the average net flux does not change in ranges of low frequency (