We next illustrate a non-trivial application of the LB method
to a situation where fluid flow is crucially important. Fig.
14 shows the
phase separation of a critical composition (
A =
B
, MA = MB )
blend
for the same quench as shown in Fig. 11.
The upper and lower boundaries in the figure are energetically
"neutral" (neither fluid preferentially wets
the surface.) A hydrodynamic "stick" boundary condition is imposed at
the walls. The top and bottom walls move at velocities uw and -uw such
that the dimensionless shear
rate equals,
tps = (2uw / d)tps = 0.18,
where d is
the spacing between the walls. The boundary condition is
periodic in the direction parallel to the translating planes. We observe
in the top view that the phase separation pattern
appears to have a "string-like" form at the boundary, but that the structure
within the film is actually more complicated. The phase separation in the plane
perpendicular to the flow is remarkably undisturbed by the flow and closely resembles
a two-dimensional phase separation pattern in the absence of shear. As time
proceeds, the "penetration depth" of the surface-induced
"string-like" structures in Fig.
14 increases and the phase separation morphology
ultimately coarsens
to a state where the fluid interface lies perpendicular to the shearing
planes and parallel to the flow direction. The surface-induced
"structuring" of the blend morphology would be accentuated in the early stages of phase
separation if one of the blend components has a preferential attraction
to the shearing boundaries.
This surface interaction evidently has
a symmetry breaking effect, since the fluid interface was
found to lie parallel to the interface of the shearing boundaries
at long times, when the
boundary interaction was modified in this fashion. This illustrative
calculation shows that the surface interaction can have a large influence on
the ultimate alignment of a phase separated fluid under shear. It is also
apparent that the interpretation of optical and scattering data on sheared
phase separated fluids is complicated by the existence of gradients in
the composition and structure of the fluid. Clearly, these observations
warrant a thorough investigation of the many parameters that seem
relevant to the phenomena
(quench depth, surface energy, viscosity, molecular weight mismatch,
roughness of shearing surface, steady and oscillatory shear, etc.). We
note that many experiments have recently reported "string-like"
structures in sheared phase-separating fluids using both light scattering and
optical microscopy techniques [48,103,104] and "string-like"
structures have also been reported in
two-dimensional LB phase separation simulations [105]. The case of two
dimensions is somewhat special, however, since the Taylor-Tomotika
[106] instability is suppressed in two dimensions [107].
G =
0.537,
= 8.5
and
tp = 0.18.
An illustration of phase separation in an off-critical blend under
steady shear also provides important insights into the kinds of phase
separation morphologies that can be expected experimentally.
Fig. 15 shows an off-critical blend (15-85)
for a dimensionless shear rate,
tps = (2uw / d)tps = 0.58.
Initially, long narrow filaments formed and eventually they
broke apart into droplets due to the well-know Taylor-Tomotika instability
[106].
The droplets (Fig. 16) then became elongated and tilted at approximately 45
degrees relative to the shear plane, as predicted in the limit of a low
concentration of dispersed droplets [108].
However, at a later stage of phase separation (Fig.
17), the droplets coalesce to form undulating string
structures that seem to persist indefinitely in a "dynamic string
state"
(a video of our simulation can be found in reference [109]).
The correlated motions of the strings suggest that the hydrodynamic
interactions between the strings and/or between the
strings and the boundary of the
sheared fluid seem to be playing an important role in the stabilization
and formation of the extended string structures. Subsequent experiments
have indicated a similar string formation phenomenon
in an off-critical blend sheared at low shear rates in a parallel plate geometry having a
narrow gap relative to the droplet size [110].
A = 0.15). The
green regions represent the transition to a high localized phase fraction of
fluid B,
B =
0.85. Note the incipient Taylor-Tomitaka instability in some of the
fluid "strings". The quench depth
G = 0.287,
= 16.4, and the dimensionless shear rate
tp =
0.56.
= 27.6).
= 56.8).