Negative pressure characteristics of an evaporating meniscus at nanoscale
© Maroo and Chung; licensee Springer. 2011
Received: 25 July 2010
Accepted: 12 January 2011
Published: 12 January 2011
This study aims at understanding the characteristics of negative liquid pressures at the nanoscale using molecular dynamics simulation. A nano-meniscus is formed by placing liquid argon on a platinum wall between two nano-channels filled with the same liquid. Evaporation is simulated in the meniscus by increasing the temperature of the platinum wall for two different cases. Non-evaporating films are obtained at the center of the meniscus. The liquid film in the non-evaporating and adjacent regions is found to be under high absolute negative pressures. Cavitation cannot occur in these regions as the capillary height is smaller than the critical cavitation radius. Factors which determine the critical film thickness for rupture are discussed. Thus, high negative liquid pressures can be stable at the nanoscale, and utilized to create passive pumping devices as well as significantly enhance heat transfer rates.
The physical attributes of phenomenon associated with the nanoscale are different from those at the macroscale due to the length-scale effects. In nature, transport processes involving a meniscus are usually associated with nano- and micro-scales. Capillary forces are of main importance in micro- and macro-scale fluidic systems. However at nanoscale, disjoining forces can become extremely dominant. These disjoining forces can cause liquid films to be under high absolute negative pressures. A better insight into negative liquid pressures can be gained from the phase diagram of water, which shows the stable, metastable, and unstable regions . Usually in such cases cavitation is observed, i.e., vapor bubbles form when a liquid is stretched. However, for the formation of a spherical vapor bubble, a critical radius of cavitation R c (defined as : ) has to be achieved. Thus, if the radius of a bubble is greater than R c, it will grow unrestricted. No cavitation will occur if any dimension of the liquid film is smaller than R c, and the liquid can exist in a metastable state.
Briggs, in 1955, heated water in a thin-walled capillary tube, open to atmosphere, up to 267°C for about 5 s before explosion occurred, and concluded that during the short time before explosion occurs, the water must be under an internal negative pressure . It has only been recently shown, through experiments that water can exist at extreme metastable states at the nanoscale. Water plugs at negative pressures of 17 ± 10 bar were achieved by filling water in a hydrophilic silicon oxide nano-channel of approximate height of 100 nm . The force contribution in water capillary bridges formed between a nanoscale atomic force microscope tip and a silicon wafer sample was measured, and negative pressures down to -160 MPa were obtained . Important consequences of the negative liquid pressures include the ascent of sap in tall trees , achieving boiling at temperatures much lower than saturation temperatures at corresponding vapor pressure , and liquid flow from bulk to evaporating film regions during heterogeneous bubble growth [8, 9].
Molecular dynamics is a vital tool to simulate and characterize the importance of disjoining force effects on the existence of negative pressures in liquids at the nanoscale. It can also provide means to compare the strength of disjoining and capillary forces at such small scales, which has not yet been possible via experiments. Although negative liquid pressure has been experimentally shown for water, it should theoretically exist in other liquids as well. With this aim, we simulated two cases of nanoscale meniscus evaporation of liquid argon on platinum wall using molecular dynamics simulation. To the best of our knowledge, this is the first study to show the existence of negative liquid pressures via molecular simulations.
The above potential form is employed for both Ar-Ar and Ar-Pt interaction with the following values: σ Ar-Ar = 3.4 × 10-10 m, ε Ar-Ar = 1.67 × 10-21 J, σ Ar-Pt = 3.085 × 10-10 m, ε Ar-Pt = 0.894 × 10-21 J.
The cutoff radius is set as r cut = 4σ Ar-Ar The force of interaction is calculated from the potential function by
All the boundaries in x and y directions are periodic. The width of the periodic boundary above the upper walls in the x-direction is restricted to the width of the opening. Any argon atom which goes above the upper walls does not interact with the wall atoms anymore. The top boundary in the z-direction is the mirror boundary condition where the argon atom is reflected back in the domain without any loss of energy, i.e., the boundary is adiabatic and elastic in nature. The 'fluid-wall thermal equilibrium model' is used to numerically simulate heat transfer between wall and fluid atoms [11, 12]. The algorithm used to calculate the atomic force interactions is the linked-cell algorithm. The equations of motion are integrated in order to obtain the positions and velocities of the atoms at every time step. The integrator method used here is the Velocity-Verlet method. Liquid atoms are distinguished from vapor atoms based on the minimum number of neighboring atoms within a certain radius . Vapor pressure is evaluated as defined elsewhere , which has been previously verified by the authors . The simulation process is divided into three parts: velocity-scaling period, equilibration period, and the heating period. During the velocity-scaling period (0-500 ps), the velocity of each argon atom is scaled at every time step so that the system temperature remains constant. This is followed by the equilibration period (500-1000 ps) in which the velocity-scaling is removed and the argon atoms are allowed to move freely and equilibrate. The wall temperatures during these two steps are the same as the initial system temperature. At the start of the heating period (1000-3000 ps), heat is transferred to the meniscus from the platinum wall region and evaporation is observed. Two cases are simulated in this study:
After the equilibrium period, the temperature of platinum wall underneath the opening (shown in red color in Figure 1) is simulated to be 130 K while the rest of the wall (shown in blue color in Figure 1) is kept at the initial temperature of 90 K.
After the equilibrium period, temperatures of all walls are simulated to 130 K.
where δ' and δ" are, respectively, the first and second derivatives of film thickness with respect to x-position. Equation 4, although a macroscopic formula, serves as a good approximation . The variation of meniscus thickness is determined in the x-z plane at different time intervals. The meniscus, formed from liquid argon atoms, is divided into square bins of dimension 1σ Ar-Ar × 1σ Ar-Ar and the number of atoms in each square is determined. A check is performed from the Pt wall in the positive z-direction such that if the number of atoms in a bin falls below 0.5 times the average number density, an interface marker is placed at the center of that bin. Interface markers are placed to determine the meniscus interface using this procedure, and a fourth-order polynomial fit of these markers is used to obtain the function δ(x).
The disjoining pressure quickly goes down to near-zero values as the meniscus thickness increases away from the non-evaporating film region. The capillary pressure variation is shown in Figure 3c, d for Case I and Case II, respectively. The capillary pressure is zero in the non-evaporating region as the non-evaporating film has a flat interface. A capillary pressure gradient exists in the meniscus region. Capillary pressure reaches negative values at the edge of the meniscus due to curvature effects and is a result of the simulation domain studied here. Comparing the disjoining and capillary pressure values, it is seen that disjoining forces dominate in nanoscale ultra-thin films, as related by Equation 3, while capillary forces become prominent with increase in film thickness and curvature.
Figure 4 also provides insight into the factors which determine the stability of such films. The difference between the normalized function values for Π = R c and Π = δ(x) is smaller for Case I than Case II, which implies that the tendency for the liquid film to rupture is higher for Case I. The following question arises: what is the critical thickness δ cr at which these liquid films would rupture, i.e., cavitate? This can be determined from the definitions of critical cavitation radius and augmented Young-Laplace equation, i.e., δ cr = -2γ/[P v - P c(δ cr) - P d(δ cr)]. In the case of non-evaporating films, which form during heterogeneous bubble growth, this equation can be simplified by assuming P c = 0 due to the planar nature of the film. Using Equation 3 where the repulsive term can be neglected as σ for liquid-solid interaction is generally smaller than δ by an order of magnitude, the following equation can be derived: 6πP v δ 3 cr + 12πγδ 2 cr A = 0, which can be solved analytically to determine the critical thickness for rupture. It can be seen that δ cr is a function of the vapor pressure, substrate temperature (indirectly via the liquid-vapor surface tension term), and substrate-liquid interaction (embedded in the Hamaker constant A). Premature rupture of non-evaporating film during bubble growth can lead to significant increase in pool boiling heat transfer and delaying the critical heat flux limit.
Negative pressure in liquids has been a point of interest over past several decades. An attempt has been made in this work to study and quantify the components of negative pressures in evaporating nano-menisci using molecular dynamics simulation. The disjoining and capillary pressures are evaluated in an evaporating meniscus at the nanoscale. Disjoining forces significantly dominate the capillary forces for ultra-thin films at the nanoscale. Liquid pressure in the meniscus is calculated using the augmented Young-Laplace equation. The center of the meniscus is found to be under high absolute negative pressures. It is shown that cavitation cannot occur as the critical cavitation radius is larger than the thickness of the meniscus. The factors determining the critical film thickness required for rupture are discussed. This property of sustaining high negative pressures at the nanoscale can be engineered to provide passive transport of liquid, and applied in power devices to attain significantly higher heat rejection rates, which is one of the major bottlenecks in achieving next generation computer chips, nuclear reactors, and rocket engines. This study serves as a first step toward understanding pressure characteristics in capillaries at the nanoscale using molecular simulations, with water nano-capillaries being the most intriguing and a near future goal.
We acknowledge the partial support by Andrew H. Hines, Jr./Progress Energy Endowment Fund.
- Angell CA: Approaching the limits. Nature 1988, 331: 206. 10.1038/331206a0View ArticleGoogle Scholar
- Fisher JC: The fracture of liquids. J Appl Phys 1948, 19: 1062. 10.1063/1.1698012View ArticleGoogle Scholar
- Briggs LJ: Maximum superheating of water as a measure of negative pressure. J Appl Phys 1955, 26: 1001. 10.1063/1.1722122View ArticleGoogle Scholar
- Tas NR, Mela P, Kramer T, Berenschot JW, van den Berg A: Capillarity induced negative pressure of water plugs in nanochannels. Nano Lett 2003, 3: 1537. 10.1021/nl034676eView ArticleGoogle Scholar
- Yang SH, Nosonovsky M, Zhang H, Chung KH: Nanoscale water capillary bridges under deeply negative pressure. Chem Phys Lett 2008, 451: 88. 10.1016/j.cplett.2007.11.068View ArticleGoogle Scholar
- Tyree MT: Plant hydraulics: The ascent of water. Nature 2003, 423: 923. 10.1038/423923aView ArticleGoogle Scholar
- Nosonovsky M, Bhushan B: Phase behavior of capillary bridges: towards nanoscale water phase diagram. Phys Chem Chem Phys PCCP 2008, 10: 2137.View ArticleGoogle Scholar
- Potash M Jr, Wayner PC Jr: Evaporation from a two-dimensional extended meniscus. Int J Heat Mass Transfer 1972, 15: 1851. 10.1016/0017-9310(72)90058-0View ArticleGoogle Scholar
- Maroo SC, Chung JN: Heat transfer characteristics and pressure variation in a nanoscale evaporating meniscus. Int J Heat Mass Transfer 2010, 53: 3335–3345. 10.1016/j.ijheatmasstransfer.2010.02.030View ArticleGoogle Scholar
- Stoddard SD, Ford J: Numerical experiments on the stochastic behavior of a Lennard-Jones gas system. Phys Rev A 1973, 8: 1504. 10.1103/PhysRevA.8.1504View ArticleGoogle Scholar
- Maroo SC, Chung JN: Molecular dynamic simulation of platinum heater and associated nano-scale liquid argon film evaporation and colloidal adsorption characteristics. J Colloid Interface Sci 2008, 328: 134. 10.1016/j.jcis.2008.09.018View ArticleGoogle Scholar
- Maroo SC, Chung JN: A novel fluid-wall heat transfer model for molecular dynamics simulations. J Nanoparticle Res 2010, 12: 1913–1924. 10.1007/s11051-009-9755-2View ArticleGoogle Scholar
- Weng J, Park S, Lukes JR, Tien C: Molecular dynamics investigation of thickness effect on liquid films. J Chem Phys 2000, 113: 5917–5923. 10.1063/1.1290698View ArticleGoogle Scholar
- Maroo SC, Chung JN: Nanoscale liquid-vapor phase-change physics in nonevaporating region at the three-phase contact line. J Appl Phys 2009, 106: 064911. 10.1063/1.3225992View ArticleGoogle Scholar
- Kohonen MM, Christenson HK: Capillary condensation of water between rinsed mica surfaces. Langmuir 2000, 16: 7285–7288. 10.1021/la991404bView ArticleGoogle Scholar
- Zhang R, Ikoma Y, Motooka T: Negative capillary-pressure-induced cavitation probability in nanochannels. Nanotechnology 2010, 21: 061057.Google Scholar
This article is published under license to BioMed Central Ltd. This is an Open Access article distributed under the terms of the Creative Commons Attribution License (http://creativecommons.org/licenses/by/2.0), which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.