(514f) Molecular Dynamics Simulations of Vapor Bubble Growth and Detachment On A Heating Surface
AIChE Annual Meeting
2012
2012 AIChE Annual Meeting
Separations Division
Nucleation and Growth II
Wednesday, October 31, 2012 - 2:36pm to 3:00pm
Nucleate boiling is an
attractive alternative for systems that require large heat transfer from a hot
solid due to the large latent heats that are accessible at relatively low wall
superheat above the fluid's saturation temperature. Based on our previous
simulations of conductively-driven quasi-static vapor bubble growth in
an axisymmetric, cylindrical cell comprised of solid and liquid phases of
finite thicknesses with fluid motion and heat transfer obeying small Reynolds,
Peclet, Capillary and Bond numbers, we found that the appearance and the motion
of three-phase contact line (CL) is a critical element, central to physics
needed to explain the large heat enhancement in nucleate boiling. However, our
simulations also suggested that different ad hoc models for the
poorly-understood motion of the CL as the vapor bubble grows do have a
significant effect on vapor bubble deformation and detachment when the Bond
number is no longer small. Therefore, we have used molecular dynamics (MD) to
simulate a nanoscale version of our three-phase system. Molecular interaction
physics is the only input, to this calculation and we then observe the
resulting CL motion. Because MD not only includes heat transfer, but also fluid
flow, this simulation removes many of the restrictions of our earlier continuum
calculation for this nano-sized system. Since terrestrial gravity is utterly
negligible at the nanoscale, we introduce a fictitious body force to mimic
relevant Bond numbers. In our MD simulations, we study not only the
heterogeneous birth of a vapor nucleus on the solid-fluid interface at constant
pressure by virtue of heat transfer through the solid and its achievement of
stability, but also its subsequent growth, deformation and detachment.
Our MD simulation domain of
interest is a cuboid, which is composed of the fluid region sandwiched between
two parallel solid walls. The fluid phase is made of argon atoms interacting
via a 12-6 Lennard-Jones (LJ) potential. The atoms of two solid walls are
tethered via springs to their lattice positions and also interact with both the
fluid and each other via LJ potentials. By considering the effects such as the
wettability of the solid surface, we find the appropriate interaction parameters
between the solid and fluid atoms in LJ form, which are critical parameters for
being able to observe nucleation and heat transfer. Periodic boundary
conditions are applied in 4 vertical boundaries of the domain. After achieving
thermal equilibrium of all the phases in the simulation domain (solid walls and
liquid argon) at the uniform reduced temperature of T=0.75 (the fluid
saturation temperature), we start to expend the top wall gradually at constant
reduced temperature T=0.75 to some specified pressure at which no vapor bubble
appears. Next, we instantaneously increase the temperature at the bottom layer
of the bottom wall while keep the original lower temperature at the top wall,
both maintained by thermostating.
This maintains a temperature gradient through the fluid phase.
Meanwhile, the pressure at the top wall is kept fixed by allowing the top
surface to freely slide up and down. Our simulations show that after some vapor
patches appear and disappear on the solid surface randomly in space and time,
one of these patches successfully grows into a stable vapor bubble, whose
growth and CL motion we then trace.
As noted, by
artificially applying a uniform body force that creates a relevant Bond number,
we have successfully nucleated and grown a vapor bubble to a moderate size and
observed bubble necking and detachment. We found that, despite the existence of
temperature slip between solid and liquid, the vapor bubble volume still grows
at t3/2, t being time, as in our earlier conduction-only continuum
calculations. This scaling is also in agreement with numerous experimental
observations. Furthermore, our temperature profiles show that vapor bubble
surface and the three-phase CL appear to be at the liquid saturation
temperature, which confirms the assumptions made in our continuum calculations'
boundary condition at liquid-vapor interface. By further increasing this
artificial uniform body force (i.e., Bond number) and keeping it constant, we
find the body force indeed deforms the bubble and causes it to detach from the
surface. During this process, we reset/raise the temperature at the top wall to
avoid vapor condensation at the top of vapor bubble as the bubble grows,
detaches and rise into regions of lower temperature, without changing the magnitude
of the uniform body force. These simulations result in, among other things, the
time evolution of the CL motion. The radius of CL initially expands with the
growth of the bubble at low Bond number. Then, as the bubble grows to a size
where the body force becomes significant, the CL contracts sharply but
continuously as the bubble deforms until detachment. In addition, we also
consider the effects of different parameters, such as those that influence the
temperature slip between solid and liquid, the magnitude of the body force, the
temperature of the bottom wall and the thickness of the bottom wall on vapor
bubble growth and detachment.
See more of this Group/Topical: Separations Division