FROM NANO TO MICRO TO MACRO SCALES IN BOILING

V. K. DHIR, H. S. ABARAJITH, AND G. R. WARRIER

Henry Samueli School of Engineering and Applied Science
University of California, Los Angeles
Los Angeles, CA 90095-1592
Ph: (310) 825-8507, Fax: (310) 206-4061
Email: vdhir@seas.ucla.edu

1. Introduction

Boiling, being the most efficient mode of heat transfer is employed in various energy conversion
systems and component cooling devices. The process allows accommodation of high heat fluxes at low
wall superheats. At the same time the process is very complex and its understanding imposes severe
challenges. Prior to inception of boiling, at low heat fluxes, the heat transfer is controlled by convection
(natural or forced). At higher heat fluxes, the heat transfer is controlled by bubble dynamics. Initially
during partial nucleate boiling, discrete bubbles form on the heater surface. At moderately high heat
fluxes, bubbles start merging laterally as well as vertically. The lateral and vertical merger of bubbles
indicates the transition from partial nucleate boiling to fully developed nucleate boiling.
In the past, several attempts have been made to model bubble growth and bubble departure processes
on a heated wall. Lee and Nydahl [1] calculated the bubble growth rate by solving the flow and
temperature fields numerically. They used the formulation of Cooper and Llyod [2] for the micro layer
thickness. However they assumed a hemispherical bubble and wedge shaped microlayer and thus they
could not account for the shape change of the bubble during growth.
Zeng et al. [3] used a force balance approach to predict the bubble diameter at departure. They
included the surface tension, inertial force, buoyancy and the lift force created by the wake of the
previously departed bubble. But there was empiricism involved in computing the inertial and drag forces.
The study assumed a power law profile for growth rate with the proportionality constant exponent
determined from the experiments.
Mei et al. [4] studied the bubble growth and departure time assuming a wedge shaped microlayer.
They also assumed that the heat transfer to the bubble was only through the microlayer, which is not
totally correct for both subcooled and saturated boiling. The study did not consider the hydrodynamics of
the liquid motion induced by the growing bubble and introduced empiricism through the shape of the
growing bubble. Welch [5] has studied bubble growth using a finite volume method and an interface
tracking method. The conduction in the solid wall was also taken in to account. However, the microlayer
was not modeled explicitely.
In 1994, Sussman et al. [6] presented a level-set approach for computing incompressible two-phase
flow. By keeping the level set as a distance function, the interface was easily captured by the zero level-
set. The calculations, for air bubbles in water and falling water drops in air, yielded satisfactory results.
Though the level-set method is easy to use, the numerical discretization of the level-set formulation does
not satisfy mass conservation, in general. Chang et al. [7] introduced a volume correction step to the
level-set formulation in 1996. By solving an additional Hamilton-Jacobi equation to steady state, the mass
was forced to be conserved. In 1999, Son et al. [8] developed a model for growth of an isolated bubble on
a heated surface using complete numerical simulation. The model, based on Sussman’s level-set method,
captures the bubble interface and offers many improvements over previously published models. It yields
the spatial and temporal distribution of the wall heat flux, the microlayer contribution and the interface
heat transfer. In this model a static contact angle was used both for the advancing and receding phases of
the interface. However, the numerical results agreed well with data from experiments. One possible

197

S. Kakaç et al. (eds.), Microscale Heat Transfer, 197 – 216.

© 2005 Springer. Printed in the Netherlands.
198

reason could be that the constant contact angle used in the numerical studies represents the average value
of the advancing and receding contact angles (static contact angle) and the bubble is symmetrical in pool
boiling case. Also, the time over which the receding contact angle prevails is much shorter than that for
the advancing contact angle. In 2001, Son [9] modified Chang et al’s formulation and included the
volume correction formulation into the boiling heat transfer model.
Singh and Dhir [10] have obtained numerical results for low gravity conditions by exercising the
numerical simulation model of [8], when the liquid is subcooled. The computational domain was divided
in to two regions viz. micro and macro regions. The interface shape and velocity and temperature field in
the liquid in the macro region were obtained by solving the conservation equations. For the micro region,
lubrication theory was used, which included the disjoining pressure in the thin liquid film. The solutions
of the micro region and macro region were matched at the outer edge of the micro layer.
Abarajith and Dhir [11] studied the effects of contact angle on the growth and departure of a single
bubble on a horizontal heated surface during pool boiling under normal gravity conditions. The contact
angle was varied by changing the Hamaker constant that defines the long-range forces. They also studied
the effect of contact angle on the microlayer and macrolayer heat transfer rates.
In spite of all the advances in the computational techniques for solving boiling problems, the one
variable that has still not been modeled correctly is the contact angle (static or dynamic). In all the
previous studies, the contact angle is specified as an input for the simulations. The reason the contact
angle has not been modeled is that it depends on the physical phenomena occurring close to or at the solid
surface, which by its very nature, occurs at very small length scales (of the order of nano to micrometers).
In this work, the contact angle is related to the Hamaker constant.

2. Mathematical Development of Model

The three-dimensional model discussed in this paper is an extension of the two-dimensional model
developed earlier by [8]. The model is used to study both single dynamics and multiple bubble merger
during subcooled and saturated pool nucleate boiling. The computational domain is divided into two
regions, namely, the micro region and the macro region as shown in Fig 1. The micro region is a thin film
that lies between the solid wall and bubble whereas the macro region consists of the bubble and it’s
surrounding. Both the regions are coupled through matching of the shape at the outer edge of the micro
layer and are solved simultaneously. Microlayer modeling covers length scales from nano to micrometers,
whereas the macro region includes the length scales from micrometers to millimeters and above.

2.1 MICRO REGION

A two-dimensional quasi-static model is used for the micro region and no azimuthal variations are
considered. As such, the solution for the microlayer thickness is obtained in the radial direction from the
center of the bubble base. This solution is assumed to be valid for all the azimuthal positions. This
assumption is still applied during the multiple bubble merger process when the bubble shapes are not
symmetrical, such that no cross flow occurs in the circumferential direction. Furthermore, the length of
the micorlayer is assumed to remain constant (though varies with contact angle) throughout the bubble
growth.
The equation of mass conservation in micro region is written as,

q 1 w G
Ul .rudz , (1)
r wr ³0

h ffg
 199

where q is the conductive heat flux from the wall, defined as kl (Twall Tiint ) with G as the thickness of the
G
thin film. Lubrication theory has been used ([12], [13] and [14]). According to the lubrication theory, the
momentum equation in the micro region is written as,

pl
wp w 2u
P , (2)
wr wz 2

where pl is the pressure in the liquid. The heat conducted through the thin film must match that due to
evaporation from the vapor-liquid interface. By using a modified Clausius-Clapeyron equation, the energy
conservation equation for the micro region yields,

kl (Twall Tiint ) ª (p º
v )Tv .
hev «Tint Tv  l » (3)
G «¬ Ul h ffg »¼

The evaporative heat transfer coefficient is obtained from kinetic theory as,
1/ 2
ª M º Uv h 2ffg
hev 2« » andd Tv s ( pv ) . (4)
¬ 2S RTv ¼ Tv

The pressure of the vapor and liquid phases at the interface are related by,
A q2 , (5)
pl pv K  03 
G 2 U v h 2ffg

where A0 is the dispersion constant. The second term on the right-hand side of equation (5) accounts for
the capillary pressure caused by the curvature of the interface, the third term is for the disjoining pressure,
and the last term originates from the recoil pressure. The curvature of the interface is defined as,

1 w ª wG § G · º» .
2

K r / 1 ¨ ¸
(6)
r wr « wr © wr ¹ ¼»
¬

Combining the mass conservation (Eq. (1)), momentum conservation equation (Eq. (2)), mass
balance and energy conservation, (Eq. (3)) and pressure balance (Eq. (5)) along with Eq. (6) for the
curvature for the micro-region results in a set of three nonlinear first-order ordinary differential Eqs. (7),
(8) and (9), as derived in [15],

wG r G r ((1 G r2 ) (1 2 3/ 2
) ª Ul h ffg § q A q2 º
  r
« Tint Tv   03  2 »
, (7)
wr r V «¬ Tv © hev ¹ v h ffg ¼
»

wTint qG r 3Tv heev P*

  , (8)
wr N l  heevG (N l  hevG ) Ul2 h ffg rG 2

w> @ rq ,
 (9)
wr h ffg
200

3
where the mass flow rate in the thin film, *  rG U wpl .
3P wr
Equations (7), (8) and (9) can be simultaneously integrated using a Runge-Kutta scheme, when
boundary conditions at r R0 are given. In the present case, the interface shape obtained from micro and
macro solutions is matched at the radial location R1 . As such, this is the end point for the integration of
the above set of equations. The radius of the dry region beneath a bubble, R0 , is related to R1 by the
definition of the apparent contact angle, tan 00.5
5 /( 1 0)
.
The boundary conditions for the film thickness at the end points are given as,

G G0 G 0 0 at r R0
, (10)
G h / 22,, G rr 0 at r R1

where, G 0 is the interline film thickness at the inner edge of micro-layer, r = R0, and is calculated by
combining Eqs. (3) and (4) and requiring that Tint Twwall at r R0 with h being the spacing of the three-
dimensional grid for the macro-region. For a given Tiint,0 at r R0 , a unique vapor-liquid interface is
obtained.

z=Z

z
Mac

Liquid

Vapor
Tsat
x

y Wall
x=X

Go Micro Region
z
G h M
x
r =Ro Wall

Figure 1 Computational domain for nucleate boiling showing details of micro and macro region.
 201

2.2 MACRO REGION

For numerically analyzing the macro region, the level set formulation developed in [8] for nucleate
boiling of pure liquids is used. The interface separating the two phases is captured by a distance function,
I , which is defined as a signed distance from the interface. The negative sign is chosen for the vapor
phase and the positive sign for the liquid phase. The discontinuous pressure drop across the vapor and
liquid, caused by the surface tension force, is smoothed into a numerically continuous function with a G -
function formulation (see [6] for details). The continuity, momentum, and energy conservation equations
for the vapor and liquid in the macro region are written as,

Ut  ’ ˜ 0, (11)

G G G G
U p u uT g ET (T Ts ) g (t ) K H, (12)

U c p Tt T N’T for
f 0, (13)

T Ts ( pv ) for H 0. (14)

The fluid density, viscosity and thermal conductivity of the fluid are defined in terms of the step function,
H , as,

U Uv  ( l v )H , (15)

P 11 Pv 1 ( l
1
v
1
)H , (16)

N 11 N l 1 H , (17)

where, H , is the Heaviside function, which is smoothed over three grid spaces as described below,

­
° 1 if I 1.5h
°° . (18)
H ® 0 if I 1.5
.5h
°
°0.5
I ª 2SI º
sin 2 ) if | I | 1.5h
/(2
°°̄ 3 ¬ 3h ¼»

The mass conservation Eq. (11) can be rewritten as,

G
’ ˜u  /U , (19)

The term on right hand side of Eq. (19) is the volume expansion due to liquid-vapor phase change. From
the conditions of the mass continuity and energy balance at the vapor-liquid interface, the following
equations are obtained,
G
m U , (20)
202

m ’T / h fg , (21)

G G
where m is the evaporation rate, and uint is the interface velocity. If the interface is assumed to advect in
the same way as the level set function, the advection equation for density at the interface can be written
as,
G
Ut  uint ˜’U 0. (22)

Using Eqs. (18), (20) and (21), the continuity equation (Eq. (19)) for the macro region can be rewritten as,
G
G m
’ ˜u ˜’U . (23)
U2

The vapor produced as a result of evaporation from the micro region is added to the vapor space through
the cells adjacent to the heated wall, and is expressed as,

§ 1 dV · m mic
¨ ¸ G H (I ) , (24)
© Vc ddt ¹ mic Vc Uv

where, Vc is the volume of the control volume and m

 mic is the evaporation rate from the micro-layer
which can be expressed as,

R1 Nl ( )
m mic ³ w in
int
rddr . (25)
R0 h ffgG

The bubble expansion due to the vapor addition from the micro layer is smoothed at the vapor-liquid
interface by the smoothed delta function as given in [6],

G H (I ) / I. (26)

In the level set formulation, the level set function, I , is used to keep track of the vapor-liquid
interface location as the set of points where I 0 , and it is advanced by the interfacial velocity while
solving the following equation,
G
It uint ˜’I . (27)

To keep the values of I close to that of a signed distance function, | I | 1, I is reinitialized after every
time step,

wI u1 I0
(1 | I |) , (28)
wt I02  h 2

where, I0 is a solution of Eq. (27) and u1 is the characteristic interface velocity, which is set to unity.
 203

In these numerical simulations, the independent variables are: (i) wall superheat, (ii) liquid
subcooling, (iii) system pressure, (iv) thermophysical properties of test fluid, (v) contact angle, (vi)
gravity level, (vii) thermophysical properties of the solid and surface quality (conjugate problem), and
(viii) heater geometry.

3. Details of Computations

Figure 1 shows the computational domain used in the simulations. Details of the micro and macro
regions are also shown in Fig. 1. In these simulations, the gravity vector is oriented in the –z direction
(i.e., the bubble is growing on an upwards facing heated surface). The simulations are carried out on a
uniform grid ('x = 'y = 'z).

3.1 BOUNDARY CONDITIONS

The boundary conditions for the pool boiling simulations are as follows:
u 0 0, 00, x 0, Ix 0, at x 0
u x vx wx 0 0, x 0, I x 0, at x X
u v w 0 0, w , I z cos M , at z 0 (29)
u z wz v 0 0, z 0, Iz 0, at z Z
u y v y wy 0 0, y 0, I y 0, at y 0
uy vy wy 0,
0 y 0, Iy 0, at y Y

where M is the static contact angle.

3.2 SOLUTION PROCEDURE

For the numerical calculations, the governing equations for micro and macro regions are
nondimensionalized by defining the characteristic length, l0 , the characteristic velocity, u0 , and the
characteristic time, t0 as,

l0 /[ g ( )]; u0 ggl0 ; t0 l0 / u0 . (30)

In performing these simulations, the wall temperature is assumed to be constant and the
thermodynamic properties of the individual phases are assumed to be insensitive to the small changes in
temperature and pressure. The assumption of constant property is reasonable as the computations are
performed for low wall superheat range.
The simulations were performed assuming that the flow is laminar. Additionally, the contact angle is
assumed to be known. The initial velocity is assumed to be zero everywhere in the domain. The initial
fluid temperature profile is taken to be linear in the natural convection thermal boundary layer and the
thermal boundary layer thickness, G T , is evaluated using the correlation for the turbulent natural
convection on a horizontal plate as, G T 7.14(Q lD l / E T )1/ 3 .
The governing equations are numerically integrated by following the procedure of Son et al. [8].
1) The value of A0 , the Hamaker (dispersion) constant is initially guessed for a given contact angle.
This initial guess can be obtained from Molecular Dynamics simulation results (if available).
2) The macro layer equations are then solved to determine the value of R1 (radial location of the
vapor-liquid interface at G h / 2. )
204

3) The micro layer equations are subsequently solved with the guessed value of A0 , to determine
the value of R0 (radial location of the vapor-liquid interface at G G 0 . )
4) The apparent contact angle is then calculated from tan M 00.5
5 /( 1 0
) and steps 1 - 4 are
repeated for a different value of A0 , if the values of the given and the calculated apparent contact
angles do not match.

4. Experiments

In order to validate the results of the numerical simulations, nucleate boiling experiments need to
be performed under defined conditions. The dynamics of single and multiple bubbles were experimentally
studied by Qiu et al. ([16] and [17]) using a polished silicon wafer as the test surface. The wafer is 10 cm
in diameter and 1 mm thick. Single and multiple cavities (with sizes varying from 10 to 3 Pm) were
etched on the wafer using standard microfabrication techniques. Thin-film strain gage heaters were
attached to back surface of the wafer. By energizing the heaters individually, any number of cavities
could be activated. Thermocouples attached to the bottom were used to measure the temperature in the
vicinity of the cavity. The silicon wafer heater assembly was placed in the experimental apparatus shown
in Fig. 2(a). Figure 2(b) shows the details of the test wafer. A CCD camera (up to 1220 frames/sec) was
used to capture the boiling process.

(a) (b)

Figure 2 (a) Experimental apparatus (b) details of test heater.

 205

5. Results and Discussion

5.1 HAMAKER CONSTANT

As mentioned earlier, the value of A0 , the Hamaker constant (dispersion constant), is found by
iteration so as to match the bubble shape at the outer edge of the microlayer with that of the macrolayer,
for a given contact angle. Figure 3 shows the variation of the dispersion constant, A0 , with contact angle
for two fluids: water and PF5060. The dispersion constant, A0 changes from negative to positive value at
around 18˚ indicating the change to attractive nature between the liquid and wall. The value of the
dispersion constant A0 does not vary much between water and PF5060, for the same contact angle and
'Tw = 8 ˚C.

Figure 3 Variation of the dispersion constant (Ao) with contact angle (Fluids: water and PF5060,
'Tw = 8 oC, 'Tsub = 0 oC).

5.2 SINGLE BUBBLE

Figure 4(a) shows the variation of bubble departure diameter with wall superheat for boiling of
saturated water at one atmosphere pressure. Both the bubble diameter and bubble growth period increases
with wall superheat. Also shown in Fig. 4(a) is the experimental data of Qiu et al. [16]. Good agreement
between the experimental and numerically predicted bubble departure diameters is observed, though the
bubble growth time is slightly over predicted. Figure 4(b) shows the variation of bubble departure
diameter and bubble growth period with liquid subcooling, for water ('Twall = 8 oC). Both the bubble
departure diameter and bubble growth period increase with increasing liquid subcooling. The contribution
of the various heat transfer mechanisms (microlayer, evaporation around the bubble boundary, and
condensation) as a function of time are shown in Fig. 5, for boiling of subcooled water. The condensation
around the bubble is zero in the initial stages of bubble growth (up to 32 ms), when the bubble is still
smaller than the thermal boundary layer. Once the bubble diameter becomes larger than the thermal
boundary layer, the condensation rate increases (shown in Fig.5 as a negative value).
The bubble growth history for two fluids with different contact angles (water and PF5060) is
shown in Fig. 6. In general, the lower the contact angle, the smaller is the bubble departure diameter and
206

the bubble growth time. The corresponding evaporative heat transfer rates from the micro and macro
layers are shown in Fig. 7. The microlayer evaporation rate increases with increasing contact angle
because the bubble base area and interfacial area increases with increasing contact angle. A corresponding
increase in the evaporation rate from the macrolayer is also observed. The area-averaged Nusselt number
for multiple bubble growth and departure cycles are plotted in Fig. 8. It is seen that the microlayer
contributes about 20% of the total heat transfer rate. Also, it takes about 10 to 12 cycles before quasi-
static conditions are achieved.

3.5 'Tw = 9 oC
Equivalent Diameter, mm

2.5

1.5
'Tw = 7 oC
1 Lift off
0.5 Saturated water
0
0 10 20 30 40 50
Time, ms

(a)

3.5

3
Equivalent Diameter, mm

2.5

'Tsub = 3 oC
1.5 'Tsub = 1 oC

1
Lift off
0.5

0
0 20 40 60 80 100 120 140 160
Tim e, m s

(b)

Figure 4 Comparison of numerical simulations with experimental data (a) Effect of wall
superheat, (b) effect of liquid subcooling (fluid: water, I = 54o, g = 1.0ge).
 207

The scaling of the bubble departure diameter and the bubble growth time with gravity level is
shown in Fig. 9. The comparison of the experimental data of Qiu et al. [16] with the numerical
predictions are shown in Fig. 10. It can be seen that the bubble departure diameter scales as g -0.5, while the
bubble growth time scales as g –1.05, for water. Numerical simulations are in general agreement with the
observed behavior. The data set that lie well below the single bubble curve corresponds to situations in
which bubbles departed after merger. The reason for this will be clear later from the numerical results for
merged bubbles.

0.6 o
Wall Superheat : 8.0 C
o
0.5 Total Liquid Subcooling : 1.0 C
Test Liquid : Water
o
Contact Angle : 54
0.4
Total
0.3 Evaporative
Evaporation Condensation
Q,W

0.2 Microlayer

0.1
Microlayer
0

-0.1
Condensation
-0.2
0.0 10.0 20.0 30.0 40.0 50.0 60.0 70.0 80.0
Time, ms

Figure 5 Contribution of the various heat transfer mechanisms during subcooled pool nucleate boiling.

4
Equivalent Diameter, mm

3.5
3
Sat. Water
2.5 g = 1.0ge
'Tw = 10 oC
2 I= 54o
1.5
1 Sat. PF5060
g = 1.0ge
0.5 'Tw = 19 oC
I= 10o
0
0 10 20 30 40
Time, ms
Figure 6 Comparison of bubble departure diameter and bubble growth time for water and PF5060.
208

(a) (b)

Figure 7 The variation of heat transfer rates with time for various contact angles (a) from micro layer
and (b) from macro region (Fluid: water, p = 1.01 bar, 'Tw = 8 qC, 'Tsub = 0 qC).

Figure 8 Variation of Nusselt number with time for various bubble growth cycles (fluid: water, 'Tw =
6.2 oC, 'Tsub = 0.0 oC, g = 1.0ge, I = 38o).
 209

o
'Tsub= 0 C
o
0.1
Tw-Ts= 8 C
Bubble Diameter, m

-4
g/ge=10
0.01

-2
g/ge=10
0.001
g/ge=1

0.0001
0.001 0.01 0.1 1 10 100 1000
Time, s

Figure 9 Scaling of equivalent bubble diameter with gravity (Fluid: water).

(a) (b)

Figure 10 Comparison of experimental data with numerical prediction for various gravity levels
(a) bubble departure diameter and (b) bubble growth time.
210

5.3 TWO BUBBLE MERGER

During nucleate boiling, increasing the wall superheat results in the increase in the bubble release
frequency and in the number of nucleation sites that become active. As a result, merger of bubbles both
normal and along the heater surface can occur which results in the formation of vapor columns and
mushroom type bubbles. Qualitative comparison of the numerical and experimental bubble shapes during
the merger of two bubbles in the vertical [17] and lateral [18] directions is shown in Figs. 11 and 12,
respectively. From these figures it can be seen that there is very good agreement between the observed
and predicted bubble shapes. A comparison of the bubble growth rate during lateral merger of two
bubbles is shown in Fig. 13. The predicted growth rate, time of merger, time of departure, and departure
diameter are in good agreement with the experimentally obtained values. Figure 14 shows a similar
comparison for lateral bubble merger under low-gravity conditions.

Figure 11 Comparison of numerical and experimental bubble shapes during vertical merger [17] (fluid:
water, 'Tw = 10 oC, 'Tsub = 0.0 oC, g = 1.0ge, I = 38 o).

Figure 12 Comparison of the experimental data from Mukherjee and Dhir [18] and numerical bubble
shapes during two bubble merger (fluid: water, 'Tw = 5.0 oC, 'Tsub = 0.0 oC, g = 1.0ge, spacing = 1.5
mm).
 211

3.5

Experim ental
3 Num erical
Bubble Equivalent Diameter, mm

2.5

2 Water
I = 54o
ǻTw = 5 oC
1.5
ǻTsub = 0 oC
Spacing = 1.5 mm
1

0.5

Merger Lift-offf
0
0 10 20 30 40 50 60 70
Tim e , m s

Figure 13 Comparison of numerically predicted bubble growth with experimental data of Mukherjee and
Dhir [18] for saturated water at earth normal gravity.

t = 0.5 s t = 2.5 s t = 2.9 s

t = 3.05 s t = 3.15 s t = 3.2 s

Figure 14 Comparison of experimental and numerical bubble shapes during the merger of two bubbles at
low gravity (fluid: water, 'Tw = 5 oC, 'Tsub = 3 oC, g = 0.01ge, I = 54o, spacing = 7 mm).
212

5.4 THREE BUBBLE MERGER

Figures 15 shows the bubble shapes for three bubbles located at the corners of an equilateral
triangle for microgravity conditions (fluid: water, 'Tw = 7 oC, 'Tsub = 0.0 oC, g = 0.01ge, I = 54o, spacing
= 6 mm). These simulations were carried out in a computational domain of 40 mm × 40 mm × 80 mm.
The symmetry conditions imposed at the wall (four side walls and the to wall), given in Eq. (29), were
replaced by no-slip boundary conditions (u = v = w = 0). All other boundary conditions remain the same.
From Fig. 15 it can be seen that the bubbles begin to merge at t = 0.5 sec. Thereafter, the merged
bubble grows as a single bubble and finally lifts off at t = 4.2 sec. Figure 16(a) shows the bubble growth
rate comparison of the three bubble merger process with that for a single bubble. It can be seen that the
merged bubble lifts off at a much smaller diameter compared to the single bubble. The growth period for
the merged bubble is also smaller than that for a single bubble. Figure 16(b) shows the net force acting on
the vapor mass for the three bubble merger case and the single bubble case. The force acting downward is
negative while the force acting upward is positive. It is found that during bubble merger an additional
vertical force (which we call the “lift force”) is induced by the fluid motion. At about 2.5 seconds when
the force changes sign and the merged bubble starts to detach, the single bubble is still experiencing a
negative force and continues to grow. The difference between the two at 2.5 seconds is designated as the
“lift force” and this additional “lift force” causes the merged bubble to lift off earlier. The bubble merger
process also increases the heat transfer rate as shown in Fig. 17. This is due to the increase in interfacial
area and the fluid motion induced by bubble merger.

t = 0 sec t = 0.8 t = 3.5 sec

t = 0.2 t = 1.0 sec t = 3.8 sec

t = 0.3 t = 2.0 sec t = 4.2 sec

t = 0.5 t = 3.0 sec t = 4.3 sec

Figure 15 Growth, merger and departure of three bubble in a plane (fluid: saturated water, g = 0.01ge, I
= 54o).
 213

25

Three Bubble Diameter

Equivalent Diameter, mm

20
Single Bubble Diameter
e

15

Single Bubble- Base

10 Diameter

5
Three Bubble- Base
Diameter
Lift-off Lift-off
0 Merger
er
0 2 4 6 8
Time, sec

(a)

1.50E-03

1.00E-03 Three Bubble Merger

Normal Force, N

Single Bubble
5.00E-04

0.00E+00
0 1 2 3 4 5 6 7 8
-5.00E-04
Lift-force
-1.00E-03
Time, sec

(b)

Figure 16
6 Comparison of (a) bubble growth history and (b) normal force for single and three bubble
merger cases.
214

3.00

2.50 Three bubbles

Heat Transfer Rate, W 2.00

1.50

1.00

0.50
Single bubble
0.00
0.00 2.00 4.00 6.00 8.00 10.00
Time, sec

Figure 17
7 Comparison of heat transfer rates for single and three bubble merger cases.

6. Summary

x Numerical simulations of the bubble dynamics during pool nucleate boiling have been carried out
without any approximation of the bubble shapes. The effect of microlayer evaporation is included. By
focusing on the micro and macro regions, the length scales from nano to micro to macro have been
connected.
x Effects of wall superheat, liquid subcooling, contact angle and level of gravity on bubble growth
process, bubble diameter at departure and growth period have been quantified.
x  Bubble mergers normal to the heater and along the heater leading to the formation of vapor columns
and mushroom type bubbles have been studied.
x  The merger process is highly nonlinear. A “lift force” leading to premature departure of bubbles from
the heating surface after merger has been identified. 

Acknowledgements

This work received support from NASA under the Microgravity Fluid Physics Program.

NOMENCLATURE

A0 , Hamaker constant, [J]; Ja, Jacob number, (UlCp'Tw/hfg);

Cp , specific heat, [J/(kg K)]; k, thermal conductivity, [W/mK];
D, diameter of the bubble, [m]; K, interfacial curvature, [1/m];
g, gravitational acceleration, [m/s2]; l0, characteristic length scale, [m];
h, grid spacing for the macro region; M, molecular weight, [g];
hev, evaporative heat transfer coefficient,
G
m, evaporative mass rate vector at interface,
[W/(m2 K)]; [kg/(m2 s)];
hfg, latent heat of evaporation, [J/kg]; m micro , evaporative mass rate from micro layer,
H, step function;
 215

[kg/s]; D, thermal diffusivity, [m2/s];

p, pressure, [bar]; Et , coefficient of thermal expansion, [1/K];
q, heat flux, [W/m2]; G, liquid thin film thickness, [m];
r, radial coordinate, [m];
GT , thermal layer thickness, [m];
R, radius of computational domain, [m];
R, universal gas constant, [J/mol K]; G H (I ) ,
smoothed delta function, [m];
R0 , radius of dry region beneath a bubble, M, apparent contact angle, [deg.];
[m]; I, level set function;
R1 , radial location of the interface at y = h/2, T, dimensionless temperature, (T -
[m]; Tsat)/(Tw - Tsat);
t, time, [s]; P, viscosity, [Pa.s];
t0, characteristic time, [s];
Q, kinematic viscosity, [m2/s];
T, temperature, [oC];
U, density, [kg/m3];
U, velocity in r direction, [m/s];
G V, surface tension, [N/m];
uint , interfacial velocity vector, [m/s];
*, mass flow rate in the micro layer,
u0, characteristic velocity, [m/s]; [kg/s];
Vc, volume of a control volume in the micro Subscripts
region, [m3]; f, fluid;
v, velocity in y direction, [m/s]; int, interface;
w, velocity in z direction, [m/s]; l, liquid;
x, coordinate, [m];
r, w/wr;
X, length of computation domain in x
sat, saturation;
direction, [m];
s, solid;
y, coordinate, [m];
t, w/wt;
Y, length of computation domain in y
v, vapor;
direction, [m];
w, wall;
z, vertical coordinate normal to the heating
wall, [m]; y, w/wy;
Z, height of computational domain, [m]; z, w/wz;
Greek symbols

References

1. Lee, R.C and Nyadhl, J.E. (1989) Numerical calculation of bubble growth in nucleate boiling from
inception to departure, Journal of Heat Transfer, Vol. 111, pp. 474-479.
2. Cooper, M.G. and Lloyd, A.J.P. (1969) The microlayer in nucleate boiling, International Journal of
Heat and Mass Transfer, Vol. 12, pp. 895-913.
3. Zeng, L.Z., Klausner, J.F. and Mei, R. (1993) A unified model for the prediction of bubble
detachment diameters in boiling systems-1. Pool boiling, International Journal of Heat and Mass
Transfer, Vol. 36, pp. 2261-2270.
4. Mei, R., Chen, W. and Klausner, J. F. (1995) Vapor bubble growth in heterogeneus boiling-1. growth
rate and thermal fields, International Journal of Heat and Mass Transfer, Vol. 38, pp. 921-934.
5. Welch, S.W.J. (1998) Direct simulation of vapor bubble growth, International Journal of Heat and
Mass Transfer, Vol. 41, pp. 1655-1666.
6. Sussman, M., Smereka, P and Osher, S. (1994) A level set approach for computing solutions to
incompressible two-phase flow, Journal of Computational Physics, Vol. 114, pp. 146-159.
7. Chang, Y.C., Hou, T.Y., Merriman, B., and Osher, S. (1996) A level set formulation of Eulerian
interface capturing methods for incompressible fluid flows, Journal of Computational Physics, Vol.
124, pp. 449–464.
216

8. Son, G., Dhir, V.K., and Ramanujapu, N. (1999) Dynamics and heat transfer associated with a single
bubble during nucleate boiling on a horizontal surface, Journal of Heat Transfer, Vol.121, pp.623-
632.
9. Son, G. (2001) Numerical study on a sliding bubble during nucleate boiling, KSME International
Journal, Vol. 15, pp. 931–940.
10. Singh, S. and Dhir, V.K. (2000) Effect of gravity, wall superheat and liquid subcooling on bubble
dynamics during nucleate boiling, Microgravity Fluid Physics and Heat Transfer (editor: Dhir, V.K.),
Begell House, New York, pp.106-113.
11. Abarajith, H.S. and Dhir, V.K. (2002) Effect of contact angle on the dynamics of a single bubble
during pool boiling using numerical simulations, Proceedings of IMECE2002 ASME International
Mechanical Engineering Congress & Exposition, New Orleans.
12. Stephan, P., and Hammer, J. (1994) A new model for nucleate boiling heat transfer, Wärme- und
Stoffübertragung, Vol.30, pp. 119-125.
13. Lay, J.H., and Dhir, V.K. (1995) Numerical calculation of bubble growth in nucleate boiling of
saturated liquids, Journal of Heat Transfer, Vol. 117, pp.394-401.
14. Wayner, P.C. (1992) Evaporation and stress in the contact line region, Proceedings of the
Engineering Fundamentals Conference on Pool and Flow Boiling, ASME, pp. 251-256.
15. Bai, Q., and Dhir, V.K. (2001) Numerical Simulation of Bubble Dynamics in the Presence of Boron in
the Liquid, Proceedings of IMECE’01, New York, NY.
16. Qiu, D.M., Dhir, V.K., Hasan, M.M., Chao, D., Neumann, E., Yee, G., and Witherow, J. (1999)
Single Bubble Dynamics During Nucleate Boiling Under Microgravity Conditions, Engineering
Foundation Conference on Microgravity Fluid Physics and Heat Transfer, Honolulu, HI.
17. Son, G, Ramanujapu, N, and Dhir, V.K. (2002) Numerical simulation of bubble merger process on a
single nucleation site during pool nucleate boiling, Journal of Heat Transfer, Vol. 124, pp. 51-62.
18. Mukherjee, A. and Dhir, V.K. (2004) Numerical and experimental study of bubble dynamics
associated with lateral merger of vapor bubbles during nucleate pool boiling, In Press, Journal of
Heat Transfer.