12.5 Change of Phase: Evaporation and Condensation 12.5.1 Interfacial Conditions We now consider the case of an evaporating (condensing) thin film of a simple liquid lying on a heated (cooled) plane surface held at constant temperature ϑ 0 which is higher (lower) than the saturation tem- perature at the given vapor pressure. It is assumed that the speed of vapor particles is sufficiently low, so that the vapor can be considered an incompressible fluid. The boundary conditions appropriate for phase transformation at the film interface z ϭ h are now for- mulated. The mass conservation equation at the interface is given by the balance between the liquid and vapor fluxes through the interface j ϭ ρ υ (v υ Ϫ v i ) и n ϭ ρ f (v f Ϫ v i ) и n, (12.79a) where j is the mass flux due to evaporation; ρ υ and ρ f are, respectively, the densities of the vapor and the liquid; v υ and v f are the vapor and liquid velocities at z ϭ h; and v i is the velocity of the interface. Equation (12.79a) provides the relationship between the normal components of the vapor and liquid velocities at the interface. The tangential components of both of the velocity fields are equal at the interface: (v f Ϫ v υ ) и t m ϭ 0, m ϭ 1, 2. (12.79b) The boundary condition that expresses the stress balance and extends Equation (12.4b) to the case of phase transformation reads [Delhaye, 1974; Burelbach et al., 1988] j(v f Ϫ v υ ) Ϫ (T Ϫ T υ ) и n ϭ 2H ~ σ (ϑ)n Ϫ ∇ s σ , (12.80a) where T υ is the stress tensor in the vapor phase and temperature dependence of surface tension is accounted for. The energy balance at z ϭ h is given by [Delhaye, 1974; Burelbach et al., 1988] j L ϩ υ 2 υ ,n Ϫ υ 2 f,n ϩ (k th ∇ϑ Ϫ k th, υ ∇ ϑ υ ) и n ϩ 2 µ (e f и n) и v f,r Ϫ 2 µ υ (e υ и n) и v υ ,r ϭ 0, (12.80b) where L is the latent heat of vaporization per unit mass; k th, υ , µ υ , ϑ υ are, respectively, the thermal con- ductivity, viscosity, and the temperature of the vapor; v υ ,r ϭ v υ Ϫ v i , v f,r ϭ v f Ϫ v i are the vapor and liq- uid velocities relative to the interface, respectively; υ υ ,n ϭ v υ ,r и n, v f,n ϭ v f,r и n are the normal components of the latter; and e f , e υ are the rate-of-deformation tensors in the liquid and the vapor, respectively. In Equation (12.80b) the first term represents the contribution of the latent heat, the combination of the second and the third terms represents the interfacial jump in the momentum flux, the combination of the fourth and the fifth terms represents the jump in the conductive heat flux at both sides of the interface, while the combination of the last two terms is associated with the viscous dissipation of energy at both sides of the interface. Since ρ υ / ρ f ϽϽ 1, typically of order 10 Ϫ3 , it follows from Equation (12.79a) that the magnitude of the normal velocity of the vapor relative to the interface is much greater than that of the liquid. Hence, the phase transformation causes large accelerations of the vapor at the interface where the back reaction, called the vapor recoil, represents a force exerted on the interface. During evaporation (condensation) the troughs of the deformed interface are closer to the hot (cold) plate than the crests, so they have greater evaporation (condensation) rates j. The dynamic pressure at the vapor side of the interface is much larger than that at the liquid side, ρ υ υ 2 υ ,n ϭ ϾϾ ρ f υ 2 f,n ϭ . (12.81) j 2 ᎏ ρ f j 2 ᎏ ρ υ 1 ᎏ 2 1 ᎏ 2 Physics of Thin Liquid Films 12-35 © 2006 by Taylor & Francis Group, LLC Momentum fluxes are thus greater in the troughs than at the crests of surface waves. Vapor recoil is a destabilizing factor for the interface dynamics for both evaporation (j Ͼ 0) and condensation (j Ͻ 0) [Burelbach et al., 1988]. Scaled with j 2 , see Equation (12.84), the vapor recoil is only important for appli- cations where very high mass fluxes are involved. Vapor recoil generally exerts a reactive downward pressure on a horizontal evaporating film. Bankoff (1961) introduced the effect of vapor recoil in the analysis of the film boiling. In this analysis the liquid overlays the vapor layer generated by boiling and leads to the Rayleigh–Taylor instability of an evaporating liquid–vapor interface above a hot horizontal wall. In this case the vapor recoil stabilizes the film boiling because the reactive force is greater for the wave crests approaching the wall than for the troughs. To obtain a closure for the system of governing equations and boundary conditions, an equation relating the dependence of the interfacial temperature ϑ i and the local pressure in the vapor phase is added [Plesset and Prosperetti, 1976; Palmer, 1976; Sadhal and Plesset, 1979]. Its linearized form is ~ Kj ϭ ϑ i Ϫ ϑ s ϵ ∆ϑ i , (12.82) where ~ K ϭ 1/2 , ϑ s is the absolute saturation temperature, ˆ α is the accommodation coefficient, R υ is the universal gas constant, and M w is the molecular weight of the vapor [Palmer, 1976; Plesset and Prosperetti, 1976; Burelbach et al., 1988]. Note that the absolute saturation temperature ϑ s serves now as the reference temperature instead of ϑ ∞ in the normalization, Equation (12.49). When ∆ϑ i ϭ 0, the phases are in thermal equilibrium with each other, and in order for net mass transport to take place, a vapor pressure driving force must exist, given for ideal gases by kinetic theory [Schrage, 1953]. The latter is represented in the linear approximation by the parameter ~ K [Burelbach et al., 1988]. Departure from ideal behavior is addressed in the parameter ~ K by the presence of an accommodation coefficient ˆ α depending on interface/molecule orientation and steric effects which represents the probability of a vapor molecule sticking upon hitting the liquid–vapor interface. The set of the boundary conditions Equations (12.80) can be simplified to what is known as a “one- sided” model for evaporation or condensation [Burelbach et al., 1988] in which the dynamics of the liq- uid are decoupled from those of the vapor. This simplification is possible because of the assumption of smallness of density, viscosity, and thermal conductivity of the vapor with respect to the respective prop- erties of the liquid. The vapor dynamics are ignored in the one-sided model, and only the mass conser- vation and the effect of vapor recoil stand for the presence of the vapor phase. The energy balance Equation (12.80b) becomes Ϫk th ∇ϑ и n ϭ j L ϩ , (12.83) suggesting that the heat flux conducted to the interface in the liquid is converted to latent heat of evapo- ration and the kinetic energy of vapor particles. The stress balance at the interface Equation (12.80a) is reduced and now rewritten explicitly for the components of the normal and tangential stresses as Ϫ Ϫ T и n и n ϭ 2H ~ σ (ϑ), T и n и t ϭ ∇ s σ и t. (12.84) In Equation (12.84) the j 2 -term stands for the contribution of vapor recoil. Finally, the remaining bound- ary conditions Equations (12.79) and (12.82) are unchanged. j 2 ᎏ ρ υ j 2 ᎏ ρ υ 2 1 ᎏ 2 2 π R υ ᎏ M w ϑ s 3/2 ᎏ ˆ αρ υ L 12-36 MEMS: Introduction and Fundamentals © 2006 by Taylor & Francis Group, LLC The procedure of asymptotic expansions outlined in the beginning of this chapter is used again to derive the pertinent evolution equation. The dimensionless mass balance Equation (12.13) is modified by the presence of the non-dimensional evaporative mass flux, J ϭ jdL/k th (ϑ 0 Ϫ ϑ s ) EJ ϭ (ϪH τ Ϫ UH ξ Ϫ VH η ϩ W)(1 ϩ H 2 ξ ) Ϫ1/2 , (12.85a) or at leading order of approximation H τ ϩ Q ξ (x) ϩ Q η (y) ϩ EJ ϭ 0, (12.85b) where Q (x) ( ξ , η , τ ) ϭ ͵ H 0 Udς, Q (y) ( ξ , η , τ ) ϭ ͵ H 0 V dς are the components of the scaled volumetric flow rate per unit width parallel to the wall. The parameter E in Equation (12.85) is an evaporation number E ϭ , which represents the ratio of the viscous time scale t v ϭ d 2 /v to the evaporative time scale, t e ϭ ρ d 2 L/k th (ϑ 0 Ϫ ϑ s ) [Burelbach et al., 1988]. The dimensionless versions of Equations (12.82) and (12.83) are: KJ ϭ Θ at ς ϭ H, Θ ς ϭ ϪJ at ς ϭ H, (12.86) where K ϭ ~ K . In the lower equation in Equation (12.86) the kinetic energy term is neglected. For details refer to Burelbach et al. (1988). Equations (12.18), (12.19), (12.53), and (12.86) pose the problem whose solution is substituted into Equation (12.85b) to obtain the sought evolution equation. The general dimensionless evolution Equation (12.21) will then contain an additional term EJ, which arises from the mass flux because of evaporation and condensation now expressed via the local film thickness H. A different approach to theoretically describe the rate of evaporative flux j in the isothermal case is known in the literature [Sharma, 1998; Padmakar et al., 1999]. This approach is based on the extended Kelvin equation that accounts for the local interfacial curvature and the disjoining and conjoining pressures, both entering the resulting expression for the evaporative mass flux j. It was shown by Padmakar et al. (1999) that their evaporation model admits the emergence of a flat adsorbed layer remaining in equilibrium with the ambient vapor phase, and thus in this state the evaporation rate from the film vanishes. This adsorbed layer, however, is usually several molecular spacings thick, which is beyond the resolution of continuum theory. 12.5.2 Evaporation/Condensation Only We first consider the case of an evaporating or condensing thin liquid layer lying on a rigid plane held at constant temperature. Mass loss or gain is retained, while all other effects are neglected. Solving first Equation (12.53) along with boundary conditions Equations (12.51a) and (12.86) and eliminating the mass flux J from the latter yields the dimensionless temperature field and the evaporative mass flux through the interface Θ ϭ 1 Ϫ , J ϭ . (12.87) 1 ᎏ H ϩ K ς ᎏ H ϩ K k th ᎏ dL k th (ϑ 0 Ϫ ϑ s ) ᎏᎏ ρν L Physics of Thin Liquid Films 12-37 © 2006 by Taylor & Francis Group, LLC An initially flat interface will remain flat as evaporation or condensation proceeds. If surface tension, thermocapillary, and convective thermal effects are negligible (i.e., M ϭ S ϭ ε RP ϭ 0), it will give rise to a scaled evolution equation of the form H τ ϩ ϭ 0, (12.88) where E ෆ ϭ ε Ϫ1 E, positive in the evaporative case and negative in the condensing one. K, the scaled inter- facial thermal resistance, is equivalent to the inverse Biot number B Ϫ1 . On the physical grounds, K 0 represents a temperature jump from the liquid surface temperature to the uniform temperature of the saturated vapor ϑ s . This jump drives the mass transfer. The conductive resistance of the liquid film is pro- portional to H, and the total thermal resistance, assuming infinite thermal conductivity of the solid, is given by (H ϩ K) Ϫ1 . For a specified temperature difference ϑ 0 Ϫ ϑ s Equation (12.88) represents a volu- metric balance whose solution, subject to the initial condition H ( τ ϭ 0) ϭ 1, is H ϭ ϪK ϩ [(K ϩ 1) 2 Ϫ 2E ෆ τ ] 1/2 . (12.89) In the case of evaporation E ෆ Ͼ 0 and when K 0, the film vanishes in a finite time τ e ϭ (2K ϩ 1)/2E ෆ , and the rate of disappearance of the film at τ ϭ τ e is finite Έ τ ϭ τ e ϭ Ϫ . For K 0, the value of dH/d τ remains finite, because as the film thins the interface temperature ϑ i , nom- inally at its saturation value ϑ s , increases to the wall temperature. If K ϭ 0 however, the problem becomes singular. In this case the thermal resistance vanishes, and the mass flux will increase indefinitely if a finite temperature difference ϑ 0 Ϫ ϑ s is sustained. The speed of the interface at rupture becomes infinite as well. Burelbach et al. (1988) showed that the interfacial thermal resistance K ϭ 10 for a 10 nanometers thick water film. Since K is inversely proportional to the initial film thickness, K Ϸ 1 for d ϭ 100 nanometers, so that H/K Ϸ 1 at this point. However, H/K Ϸ 10 Ϫ1 at d ϭ 30 nanometers, so that the resistance to con- duction is small compared to the interfacial transport resistance. Shortly after, van der Waals forces become appreciable. 12.5.3 Evaporation/Condensation, Vapor Recoil, Capillarity, and Thermocapillarity The dimensionless vapor recoil gives an additional normal stress at the interface determined by the j 2 -term in Equation (12.84), Π ˆ 3 ϭ Ϫ 3 – 2 E ෆ 2 D Ϫ1 J 2 , where D is a unit-order scaled ratio between the vapor and liquid densities D ϭ ε Ϫ3 . This stress can be calculated using Equation (12.87). The resulting scaled evolution equation for an evap- orating film on an isothermal horizontal surface neglecting the thermocapillary effect and body forces is obtained using the combination of Equations (12.21) and (12.88) with Π 1 ϭ 0, Σ ξ ϭ 0 [Burelbach et al., 1988]: H τ ϩ ϩ ΄ E ෆ 2 D Ϫ1 3 H ξ ΅ ξ ϩ S(H 3 H ξξξ ) ξ ϭ 0. (12.90) Since usually t e ϾϾ t v , E ෆ can be a small number and can be used as an expansion parameter for slow eva- poration compared to the non-evaporating base state [Burelbach et al., 1988] appropriate to very thin evaporating films. 1 ᎏ 3 H ᎏ H ϩ K E ෆ ᎏ H ϩ K ρ υ ᎏ ρ 3 ᎏ 2 E ෆ ᎏ K dH ᎏ d τ E ෆ ᎏ H ϩ K 12-38 MEMS: Introduction and Fundamentals © 2006 by Taylor & Francis Group, LLC Taking into account van der Waals forces and thermocapillarity, the complete evolution equation for a thin heated or cooled film on a horizontal plane surface was givenbyBurelbach et al. (1988) in the form H τ ϩ ϩ Ά΄ AH Ϫ1 ϩ E ෆ 2 D Ϫ1 3 ϩ KM ෆ P Ϫ1 2 ΅ H ξ · ξ ϩ S(H 3 H ξξξ ) ξ ϭ 0 (12.91) with M ෆ ϭ ε M.Here the first term represents the rate of volumetric change; the second one the mass loss/gain; the third, fourth, and fifth ones the attractive van der Waals, vapor recoil, and thermocapillary terms, all destabilizing; while the sixth term describes the stabilizing capillary force. This was the first full statement of the possible competition among various stabilizing and destabilizing effects on a horizontal plate, with scaling making them present at the same order. Other effects such as gravity may be included in Equation (12.91). Joo et al. (1991) extended the work to an evaporating (condensing) liquid film drain- ing down a heated (cooled) inclined plate. Oron and Bankoff (1999) studied the two-dimensional dynamics of an evaporating ultrathin film on acoated solid surface when the potential Equation (12.31d) was used. Three different types of the evolu- tion of avolatile film were identified. One type is related to low evaporation rates associated with rela- tively small E ෆ Ͼ 0 when holes covered by a liquid microlayer emerge, and the expansion of such holes is governed mainly by the action of the attractive molecular forces. These forces impart the squeeze effect to the film and, as a result of this, the liquid flows away from the hole. In this stage the role of evapora- tion is secondary. Figure 12.9 displays such an evolution of avolatile liquid film. Following the nucleation of the hole and during the process of surface dewetting, one can identify the formation of a large ridge, or drop, on either side of the trough. The former grows during the evolution of the film until the drops at both ends of the periodic domain collide. A further recession of the walls of the dry spot leads to the formation of a single large drop that flattens and ultimately disappears, according to Equation (12.89). The stages of the film evolution shown in Figure 12.9(a) are very similar to that sketched in Figure 3 of Elbaum and Lipson (1995). This type of evolution also resembles the results obtained by Padmakar et al. (1998) for the isothermal film subject to hydrophobic interactions and to evaporation driven by the dif- ference between the equilibrium vapor pressure and the pressure in the vapor phase. Such films thin uni- formly to a critical thickness and then spontaneously to dewet the solid substrate by the formation of growing dry spots when the solid was partially wetted. In the completely wetted case, thin liquid films evolved to an array of islands that disappeared by evaporation to a thin equilibrium flat film. Two other regimes corresponding to intermediate and high evaporation rates were discussed in Oron and Bankoff (1999). An important phenomenon was found in the last stage of the evolution of an evaporating film where the latter finally disappears by evaporation: prior to that the film equilibrates, so that its disappearance is practically uniform in space. The film equilibration is caused by the “reservoir effect,” which is driven by the difference in disjoining pressures and manifests itself by feeding the liquid from the large drops into the ultrathin film that bridges between them. Oron and Bankoff (2001) studied the dynamics of condensing thin films on a horizontal coated solid surface. In the case of arelatively fast condensation, where the initial depression of the interface rapidly fills up because of the enhanced mass gain there, the film equilibrates and grows uniformly in space according to Equation (12.89). Note that E ෆ Ͻ 0. When condensation is relatively slow, the evolution of the film exhibits several distinct stages. The first stage, dominated by attractive van der Waals forces, leads to the opening of a hole covered by a microlayer, as shown in the first three snapshots of Figure 12.10(a). This is accompanied with continuous condensation with the highest rate of mass gain attained in the microlayer region corresponding to the smallest thickness H in Equation (12.87). However, opposite to the evaporative case [Oron and Bankoff, 1999], where the “reservoir effect” arising from the difference between the disjoining pressures causes feeding of the liquid from the large drops into the microlayer and film equilibration, in the condensing case the excess liquid is driven from the microlayer into the large drops. This effect is referred to as the “reversed reservoir effect.” The thickness of the microlayer remains nearly constant because of local mass gain by condensation compensating for the impact of the reverse reservoir effect. The first stage of the film evolution terminates in the situation where the size of the hole 1 ᎏ 3 H ᎏ H ϩ K H ᎏ H ϩ K E ෆ ᎏ H ϩ K Physics of Thin Liquid Films 12-39 © 2006 by Taylor & Francis Group, LLC is the largest. The receding of the drops stops due to the increase of the drop curvature and buildup of the capillary pressure that comes to balance with the squeeze effect of the attractive van der Waals forces. From this moment the hole closes driven by condensation, as shown in Figure 12.10(a, b). Once the hole closes, the depression fills up rapidly, the amplitude of the interfacial disturbance decreases, and the film tends to flatten out. The film then grows uniformly in space following the solution Equation (12.89) with negative E ෆ . Oron (2000c) studied the three-dimensional evolution of an evaporating film on a coated solid surface subject to the potential Equation (12.31d). The main stages of the evolution repeat those mentioned pre- viously in the case of a non-volatile film in the section on isothermal films, except for the stage of disap- pearance accompanied by the reservoir effect. Because of the reservoir effect, the minimal film thickness decreases very slowly during the stage of film equilibration. 12.5.4 Flow on a Rotating Disc Reisfeld et al. (1991) considered the axisymmetric flow of an incompressible viscous volatile liquid on a horizontal, rotating disk. The liquid was assumed to evaporate because of the difference between the 12-40 MEMS: Introduction and Fundamentals 1.5 1.0 (a) (b) 0.5 0.0 0.08 0.07 0.06 0.05 0.04 0.03 0.02 0.01 0.00 HH FIGURE 12.9 The evolution of a slowly evaporating film: (a) the initial and intermediate stages of the film evolu- tion, and (b) the final stage of the evolution. The curves in both graphs correspond to the interfacial shapes in con- secutive times (not necessarily equidistant). (Reprinted with permission from Oron and Bankoff (1999).) © 2006 by Taylor & Francis Group, LLC vapor pressures of the solvent species at the fluid–vapor interface and in the gas phase. This situation is analogous to the phase two of spin coating process. The analysis is similar to what is done in the section on isothermal films, but now with an additional parameter describing the process of evaporation, which for a prescribed evaporative mass flux j is defined as E ϭ . Using the procedures outlined in the section on isothermal films, one obtains at leading order the fol- lowing evolution equation H τ ϩ E ϩ r Ά r 2 H 3 ϩ SrH 3 ΄ (rH r ) r ΅ r · r ϭ 0. (12.92) 1 ᎏ r 1 ᎏ 3 2 ᎏ 3 3j ᎏ 2 ερ U 0 Physics of Thin Liquid Films 12-41 36.0 10.0 8.0 6.0 4.0 2.0 0.0 32.0 28.0 24.0 20.0 16.0 12.0 8.0 4.0 0.0 HH (a) (b) FIGURE 12.10 The evolution of a slowly condensing film on the horizontal plane. (a) The curves from the bottom to the top correspond to consecutive times (not necessarily equidistant). (b) The curves from the left to the right cor- respond to consecutive times (not necessarily equidistant). The flat curve corresponds to the interface at a certain time after which the film grows uniformly in space according to Equation (12.89). In the graph (a), the dashed lines rep- resent the location of the solid substrate H ϭ 0. (Reprinted with permission from Oron and Bankoff (2000).) © 2006 by Taylor & Francis Group, LLC Equation (12.92) models the combined effect of local mass loss, capillary forces and centrifugal drainage, none of which describe any kind of instability. For most spin coating applications S is very small, and the corresponding term may be neglected, although it may be very important in planarization studies where the leveling of liquid films on rough surfaces is investigated. Therefore, Equation (12.92) can be simplified H τ ϩ E ϩ r(r 2 H 3 ) r ϭ 0. (12.93) This simplified equation can then be used for further analysis. Looking for flat basic states H ϭ H( τ ), Equation (12.93) is reduced to the ordinary differential equation which is to be solved with the initial condition H(0) ϭ 1. In the case of E Ͼ 0, both evaporation and drainage cause thinning of the layer. Equation (12.93) describes the evolution in which the film thins monotonically to zero thickness in a finite time in contrast with an infinite thinning time by centrifugal drainage only. Explicit expressions for H(τ) and for the time of film disappearance are given in Reisfeld et al. (1991). In the condensing case E Ͻ 0 drainage competes with condensation to thin the film. Initially the film thins due to drainage until the rate of mass gain because of condensation balances the rate of mass loss by drainage. At this point the film interface reaches its steady location H ϭ |E| 1/3 . The cases where inertia is taken into account are con- sidered in Reisfeld et al. (1991), where linear stability analysis of flat base states is given. Experiments with volatile rotating liquid films [Stillwagon and Larson, 1990] showed that the final stage of film leveling was affected by an evaporative shrinkage of the films. Therefore, they suggested sep- arating the analysis of the evolution of evaporating spinning films into two stages with fluid flow domi- nating the first stage and solvent evaporation dominating the second one [Stillwagon and Larson, 1992]. 12.6 Closing Remarks In this chapter the physics of thin liquid films is reviewed and various examples of their dynamics relevant for MEMS are presented, some of them with reference to the corresponding experimental results. The examples discussed examine isothermal, non-isothermal with no phase changes, and evaporating and condensing films under the influence of surface tension, gravity, van der Waals, and centrifugal forces. The long-wave theory has been proven to be a powerful tool for the research of the dynamics of thin liquid films. However, there exist several optional approaches suitable for a study of the dynamics of thin liquid films. Direct numerical simulation of the hydrodynamic equations (Navier–Stokes and continuity) [Scardovelli and Zaleski, 1999] mentioned briefly in the introduction represents one of these options. A variety of methods were developed to carry out such simulations: techniques based on Finite Elements Method (FEM) [Ho and Patera, 1990; Salamon et al., 1994; Krishnamoorthy et al., 1995; Tsai and Yue, 1996; Ramaswamy et al., 1997], techniques based on the boundary-integral method [Pozrikidis, 1992, 1997; Newhouse and Pozrikidis, 1992; Boos and Thess, 1999], surface tracking technique [Yiantsios and Higgins, 1989], and others. Another optional approach is that of molecular dynamics (MD) simulations [Allen and Tildesley, 1987; Koplik and Banavar, 1995, 2000]. Refer directly to these works for more detail. A new approach treating the film interface as a diffuse rather than a sharp one, as presented in this chapter, was recently developed [Pismen and Pomeau, 2000] and applied to various physical situations [Pomeau, 2001; Pismen, 2001; Bestehorn and Neuffer, 2001; Thiele et al., 2001a, b; 2002a, b; 2003]. Lastly, new frontiers in the investigation of the dynamics of thin liquid films were recently discussed in the special issue of “European Physical Journal E, Vol. 12(3), 2003”. An attempt was made to bridge between numerous theoretical and experimental results in order to explain the main mechanism(s) liable to rupture of a film. Open questions, controversial approaches, and contradictory conclusions were all in the focus of the discussion [Ziherl and Zumer, 2003; van Effenterre and Valignat, 2003; Morariu et al., 2003; Kaya and Jérôme, 2003; Bollinne et al., 2003; Sharma, A., 2003; Thiele, 2003; Stöckelhuber, 2003; Richardson et al., 2003; Müller-Buschbaum, 2003; Green and Ganesan, 2003; Oron, 2003; Manghi and Aubouy, 2003; Reiter, 2003]. 1 ᎏ 3 2 ᎏ 3 12-42 MEMS: Introduction and Fundamentals © 2006 by Taylor & Francis Group, LLC Acknowledgments It is a pleasure to express my gratitude to A. Sharma, G. Reiter, and U. Thiele for reading the manuscript and sharing their comments and thoughts with me. A. Sharma, S. Herminghaus, M. F. Schatz, and E. Zussman are acknowledged for providing me with their experimental results used in this chapter. References Adamson, A.W. (1990) Physical Chemistry of Surfaces, Wiley, New York. Allen, M.P., and Tildesley, D.J. (1987) Computer Simulation of Liquids, Clarendon, Oxford. Babchin, A.J., Frenkel, A.L., Levich, B.G., and Sivashinsky, G.I. (1983) “Nonlinear Saturation of Rayleigh–Taylor Instability,” Phys. Fluids 26, pp. 3159–61. Bankoff, S.G. (1961) “Taylor Instability of an Evaporating Plane Interface,” AIChE J. 7, pp. 485–7. Bankoff, S.G. (1990) “Dynamics and Stability of Thin Heated Films,” J. Heat Transfer Trans. ASME 112, pp. 538–46. Bankoff, S.G., and Davis, S.H. (1987) “Stability of Thin Films,” Physicochem. Hydrodyn. 9, pp. 5–7. Becerril, R., Van Hook, S.J., and Swift, J.B. (1998) “The Influence of Interface Profile on the Onset of Long-Wavelength Marangoni Convection,” Phys. Fluids 10, pp. 3230–2. Bestehorn, M., and Neuffer, K. (2001) “Surface Patterns of Laterally Extended Thin Liquid Films in Three Dimensions,” Phys. Rev. Lett. 87, pp. 046101-1–046101-4. Bischof, J., Scherer, D., Herminghaus, S., and Leiderer, P.(1996) “Dewetting Modes of Thin Metallic Films: Nucleation of Holes and Spinodal Dewetting,” Phys. Rev. Lett. 77, pp. 1536–9. Bollinne, C., Cuenot, S., Nysten, B., and Jonas, A.M. (2003) “Spinodal-Like Dewetting of Thermo- dynamically-Stable Thin Polymer Films,” Eur. Phys. J. E 12, pp. 389–96. Boos, W., and Thess, A. (1999) “Cascade of Structures in Long-Wavelength Marangoni Instability,” Phys. Fluids 11, pp. 1484–94. Brochard-Wyart, F., and Daillant, J. (1990) “Drying of Solids Wetted by Thin Liquid Films,” Can. J. Phys. 68, pp. 1084–8. Brusch, L., Kuhne, H., Thiele, U., and Bar, M. (2002) “Dewetting of Thin Films on Heterogeneous Substrates: Pinning Versus Coarsening,” Phys. Rev. E 66, pp. 011602-1–011602-5. Burelbach, J.P., Bankoff, S.G., and Davis, S.H. (1988) “Nonlinear Stability of Evaporating/Condensing Liquid Films,” J. Fluid Mech. 195, pp. 463–94. Burelbach, J.P., Bankoff, S.G., and Davis, S.H. (1990) “Steady Thermocapillary Flows of Thin Liquid Layers. II. Experiment,” Phys. Fluids A 2, pp. 322–33. Burgess, J.M., Juel, A., McCormick, W.D., Swift, J.B., and Swinney, H.L. (2001) “Suppression of Dripping from a Ceiling,” Phys. Rev. Lett. 86, pp. 1203–6. Cahn, J.W. (1961) “On Spinodal Decomposition,” Acta Metall. 9, pp. 795–801. Chandrasekhar, S. (1961) Hydrodynamic and Hydromagnetic Stability, Clarendon, Oxford. Chou, F C., and Wu, P Y. (2000) “Effect of Air Shear on Film Planarization During Spin Coating,” J. Electrochem. Soc. 147, pp. 699–705. Davis, S.H. (1983) “Rupture of Thin Films,” in Waves on Fluid Interfaces, R.E. Meyer, ed., Academic Press, New York, pp. 291–302. Davis, S.H. (1987) “Thermocapillary Instabilities,” Annu. Rev. Fluid Mech. 19, pp. 403–35. Deissler, R.J., and Oron, A. (1992) “Stable Localized Patterns in Thin Liquid Films,” Phys. Rev. Lett. 68, pp. 2948–51. Delhaye, J.M. (1974) “Jump Conditions and Entropy Sources in Two-Phase Systems. Local Instant Formulation,” Int. J. Multiphase Flow 1, pp. 395–409. Dzyaloshinskii, I.E., Lifshitz, E.M., and Pitaevskii, L.P. (1959) Zh Eksp Teor Fiz 37, pp. 229–41; [(1960) “Van der Waals Forces in Liquid Films,” Sov. Phys. JETP 10, pp. 161–170]. Edwards, D.A., Brenner, H., and Wasan, D.T. (1991) Interfacial Transport Processes and Rheology, Butterworth-Heinemann, Boston. Physics of Thin Liquid Films 12-43 © 2006 by Taylor & Francis Group, LLC Elbaum, M., and Lipson, S.G. (1994) “How Does a Thin Wetted Film Dry Up?,” Phys. Rev. Lett. 72, pp. 3562–5. Elbaum, M., and Lipson, S.G. (1995) “Pattern Formation in the Evaporation of Thin Liquid Films,” Israel J. Chem. 35, pp. 27–32. Fermigier, M., Limat, L., Wesfreid, J.E., Boudinet, P., and Quilliet, C. (1992) “Two-Dimensional Patterns in Rayleigh–Taylor Instability of a Thin Layer,” J. Fluid Mech. 236, pp. 349–83. Frenkel, A.L., Babchin, A.J., Levich, B.G., Shlang, T., and Sivashinsky, G.I. (1987) “Annular Flow Can Keep Unstable Flow from Breakup: Nonlinear Saturation of Capillary Instability,” J. Colloid Interface Sci. 115, pp. 225–33. Gau, H., Herminghaus, S., Lenz, P., and Lipowsky, R. (1999) “Liquid Morphologies on Structured Surfaces: from Microchannels to Microchips,” Science 283, pp. 46–9. de Gennes, P.G. (1985) “Wetting: Statics and Dynamics,” Rev. Mod. Phys. 57, pp. 827–63. Goldmann, L.S. (1969) “Geometric Optimization of Controlled Collapse Interconnections,” IBM J. Res. Dev. 13. Green, P.F., and Ganesan, V. (2003) “Dewetting of Polymeric Films: Unresolved Issues,” Eur. Phys. J. E 12, pp. 449–54. Grotberg, J.B. (1994) “Pulmonary Flow and Transport Phenomena,” Annu. Rev. Fluid Mech. 26, pp. 529–71. Hammond, P.S. (1983) “Nonlinear Adjustment of a Thin Annular Film of Viscous Fluid Surrounding a Thread of Another within a Circular Cylindrical Pipe,” J. Fluid Mech. 137, pp. 363–84. Herminghaus, S., Fery, A., Schlagowski, S., Jacobs, K., Seeman, R., Gau, H., Moench, W., and Pompe, T. (2000) “Liquid Microstructures at Solid Interfaces,” J. Phys. Condens. Matter 12, pp. A57–A74. Herminghaus, S., Gau, H., and Moench, W. (1999) “Artificial Liquid Microstructures,” Adv. Mater. 11, pp. 1393–5. Hirasaki, G.J. (1991) “Thin Films and Fundamentals of Wetting Phenomena” in Interfacial Phenomena in Petroleum Recovery, N.R. Morrow, ed., Marcel Dekker, New York. Ho, C M., and Tai, Y C. (1998) “Micro-Electro-Mechanical Systems (MEMS) and Fluid Flows,” Annu. Rev. Fluid Mech. 30, pp. 579–612. Ho, L W., and Patera, A.T. (1990) “A Legendre Spectral Element Method for Simulation of Unsteady Incompressible Viscous Free-Surface Flow,” Comp. Meth. Appl. Mech. Eng. 80, pp. 355–66. Huppert, H.E. (1982) “The Propagation of Two-Dimensional and Axisymmetric Viscous Gravity Currents over a Rigid Horizontal Surface,” J. Fluid Mech. 121, pp. 43–58. Huppert, H.E., and Simpson, J.E. (1980) “The Slumping of Gravity Currents,” J. Fluid Mech. 99, pp. 785–99. Hwang, C C., Chang, S H., and Chen, J L. (1993) “On the Rupture Process of Thin Liquid Film,” J. Colloid Interface Sci. 159, pp. 184–8. Hwang, C C., Lin, C K., and Uen, W Y. (1997) “A Nonlinear Three-Dimensional Rupture Theory of Thin Liquid Films,” J. Colloid Interface Sci. 190, pp. 250–2. Ichimura, K., Oh, S K., and Nakagawa, M. (2000) “Light-Driven Motion of Liquids on a Photoresistive Surface,” Science 288, pp. 1624–6. Israelachvili, J.N. (1992) Intermolecular and Surface Forces, 2nd ed., Academic Press, London. Ivanov, I.B. (1988) Thin Liquid Films: Fundamentals and Applications, Marcel Dekker, New York. Jacobs, K., Herminghaus, S., and Mecke, K.R. (1998) “Thin Liquid Polymer Films Rupture via Defects,” Langmuir 14, pp. 965–9. Jain, R.K., and Ruckenstein, E. (1974) “Spontaneous Rupture of Thin Liquid Films,”J. Chem. Soc. Faraday Trans. II 70, pp. 132–47. Jameel, A.T., and Sharma, A. (1994) “Morphological Phase Separation in Thin Liquid Films,” J. Colloid Interface Sci. 164, pp. 416–27. Joo, S.W., Davis, S.H., and Bankoff, S.G. (1991) “Long-Wave Instabilities of Heated Films: Two Dimensional Theory of Uniform Layers,” J. Fluid Mech. 230, pp. 117–46. Kargupta, K., Konnur, R., and Sharma, A. (2000) “Instability and Pattern Formation in Thin Liquid Films on Chemically Heterogeneous Substrates,” Langmuir 16, pp. 10243–53. 12-44 MEMS: Introduction and Fundamentals © 2006 by Taylor & Francis Group, LLC [...]... calculation in the longitudinal direction, the local electric field E is shown to scale as the inverse of the cross-section area of the electrolyte across the capillary By Equation (13.19), © 2006 by Taylor & Francis Group, LLC 1 3-1 0 MEMS: Introduction and Fundamentals FIGURE 13.5 Electrokinetically driven bubble speed as a function of a concentration-normalized electric field for the KCl electrolyte... concentrations The unnormalized data scatter over 5 decades and are collapsed by the theory the electrokinetic velocity scales the same way However, the flow rate scales as the electrokinetic velocity times the cross-section area and is independent of the cross-section area The flow rate behind the bubble is the same as the flow rate in its surrounding film As a result, there is no back pressure buildup, and the. .. inversely as the square root of the bulk electrolyte concentration In the presence of a tangential electric field E, there is a net body force on the electrolyte that scales as Eρ This body force vanishes in the neutral bulk but accelerates the ions in the double layer to large speeds These streaming ions drag the entire fluid body in the capillary along with them The body force is concentrated in the. .. away from the double layer This asymptote is called the electrokinetic velocity: ε0εζcE uc ϭ Ϫ ᎏ µ (13.19) The constants ε0 and ε are the dielectric permittivities that we have omitted in the previous scaling arguments Because the electrokinetic velocity is flat away from the thin double layer (see Figure 13.4), the flow rate scales as the cross-section area, or R2 For pressure-driven flow, the flow... laid onto the stagnant film The velocity at the rigid interface vanishes, and the parabolic velocity of Equation (13.1) becomes: σ ∂3h y2 yh u(y) ϭ Ϫ ᎏ ᎏ ᎏ Ϫ ᎏ µ ∂x3 2 2 (13.13) The flow rate q is then corrected by the factor of 4 due to the interface traction The same correction yields a factor of 4 to the left-hand side of the Bretherton equation, Equation (13.3) Simply scale Ca by 4 and the same... 1 The total viscous dissipation due to the flow at the caps is the integral of µ times the normal gradient of the flow field at the wall over the transition length This is the capillary pressure required to balance the dissipation Using the previous scaling, this capillary pressure is of the order (σ/R)Ca2/3 Due to the asymmetry of the two caps, this capillary pressure is different at the two caps The. .. electric potential driving forces One fluid should wet the channel or capillary walls while the other is dispersed in the form of bubbles Due to the small channel dimension, the bubbles usually have a free radius larger than the channel radius — it is typically difficult to generate colloid-size bubbles smaller than the channel This chapter addresses several fundamental issues in the transport of these... behind the tip, and the resulting capillary pressure gradient drives fluid into the annular film The reverse happens near the back cap to pick up the stagnant liquid laid down by the front cap Unlike the usual symmetric Stokes flow, the flow around the two caps are not mirror images of each other in this free-surface problem If they were reflectively symmetric, the net pressure drop across the bubble... flow, the flow rate scales as the square of the area, or R4 The electrokinetic velocity is independent of R, whereas the velocity of pressure-driven flow scales as the second power of R Electrokinetic flow is much less efficient than pressure-driven flow, but electrokinetic flow is easier to scale up and down in microfluidic designs Unfortunately, the same flat electrokinetic velocity profile now serves... Structure of Thin Liquid-Crystalline Films at NematicIsotropic Transition,” Eur Phys J E 12, pp 361–6 © 2006 by Taylor & Francis Group, LLC 13 Bubble/Drop Transport in Microchannels 13.1 Introduction 1 3-1 13.2 Fundamentals 1 3-2 13.3 The Bretherton Problem for Pressure-Driven Bubble/Drop Transport 1 3-4 Corrections to the Bretherton Results for Pressure-Driven Flow Hsueh-Chia Chang University . various examples of their dynamics relevant for MEMS are presented, some of them with reference to the corresponding experimental results. The examples discussed examine isothermal, non-isothermal with. interfacial jump in the momentum flux, the combination of the fourth and the fifth terms represents the jump in the conductive heat flux at both sides of the interface, while the combination of the last. Equation (12. 53) along with boundary conditions Equations (12. 51a) and (12. 86) and eliminating the mass flux J from the latter yields the dimensionless temperature field and the evaporative mass flux