Department of Aerospace Engineering

Boiling and Condensation

29 April 2020

1 Boiling Heat Transfer

When evaporation of liquid occurs at a solid-liquid interface, it is termed boiling. The process occurs
when the temperature of the surface Ts exceeds the saturation temperature Tsat corresponding to the liquid
pressure. Heat is transferred from the solid surface to the liquid, and the appropriate form of Newton’s law
of cooling is
q̇s = h(Ts − Tsat ) = h ∆Te (1)
where ∆Te = Ts − Tsat is called the excess temperature. The process is characterized by the formation of
vapour bubbles, which grow and subsequently detach from the surface. Vapour bubble growth and dynamics
depend, in a complicated manner, on the excess temperature, the nature of the surface, and thermophysical
properties of the fluid, such as its surface tension. In turn, the dynamics of vapour bubble formation affect
liquid motion near the surface and therefore strongly influence the heat transfer coefficient.
Boiling may occur under various conditions. For example, in pool boiling the liquid is quiescent and its
motion near the surface is due to free convection and to mixing induced by bubble growth and detachment.
In contrast, for forced convection boiling, fluid motion is induced by external means, as well as by free
convection and bubble-induced mixing. Boiling may also be classified according to whether it is subcooled
or saturated. In subcooled boiling, the temperature of most of the liquid is below the saturation temperature
and bubbles formed at the surface may condense in the liquid. In contrast, the temperature of the liquid
slightly exceeds the saturation temperature in saturated boiling. Bubbles formed at the surface are then
propelled through the liquid by buoyancy forces, eventually escaping from a free surface.

1.1 Dimensionless parameters

In our of convection heat transfer, we nondimensionalized the governing equations to identify relevant dimen-
sionless groups. This approach enhanced our understanding of related physical mechanisms and suggested
simplified procedures for generalizing and representing heat transfer results.
Since it is difficult to develop governing equations for boiling and condensation processes, the appropriate
dimensionless parameters can be obtained by using the Buckingham pi theorem. For either process, the
convection coefficient could depend on the difference between the surface and saturation temperatures, ∆Te ,
the body force arising from the liquid-vapour density difference, g(ρl − ρv ), the latent heat h f g , the surface
tension σ , a characteristic length L, and the thermophysical properties of the liquid or vapour: ρ , c p , k, µ .
That is,

h = h ∆Te , g(ρl − ρv ), h f g , σ , L, ρ, c p , k, µ (2)
Since there are 10 variables in 5 dimensions (m, kg, s, J, K), there are (10 - 5) = 5 Π-groups, which can be
expressed in the following forms:

ρg(ρl − ρv )L3 c p ∆Te µc p g(ρl − ρv )L2

 
hL
= f , , , (3)
k µ3 hfg k σ

1
 Figure 1: Different boiling regimes in pool boiling

or, defining the dimensionless groups,

ρg(ρl − ρv )L3
 
Nu = f , Ja, Pr, Bo (4)
µ3
The Nusselt and Prandtl numbers are familiar from our earlier single-phase convection problems. The new
dimensionless parameters are the Jakob number Ja, the Bond number Bo, defined as

c p ∆Te g(ρl − ρv )L2

Ja = and Bo =
hfg σ
and a nameless parameter that bears a strong resemblance to the Grashof number. This unnamed parameter
represents the effect of buoyancy-induced fluid motion on heat transfer. The Jakob number is the ratio of
the maximum sensible energy absorbed by the liquid (vapour) to the latent energy absorbed by the liquid
(vapour) during boiling process. In many applications, the sensible energy is much less than the latent
energy and Ja has a small numerical value. The Bond number is the ratio of the buoyancy force to the
surface tension force. In subsequent sections, we will delineate the role of these parameters in boiling and
condensation.

2 Pool Boiling
Pool boiling provides a convenient starting point for the discussion of mechanism of heat transfer in boil-
ing systems. Despite that fact that it has been extensively studied an its mechanism is reasonably well
understood, it is not possible to predict theoretically its heat transfer characteristics.
S. Nukiyama in 1934 was the first to identify different regimes of pool boiling and established experimen-
tally the characteristics of pool boiling phenomena. He immersed nichrome and platinum wires into a body
of saturated water and initiated boiling on the surface of the wires by passing electric current through it.
Nukiyama noticed that boiling takes different forms, depending on the value of the excess temperature ∆Te .
Four different boiling regimes are observed: natural convection boiling, nucleate boiling, transition boiling,
and film boiling (figure 1). These regimes are illustrated on the boiling curve in figure 2, which is a plot of
boiling heat flux versus the excess temperature. Although the boiling curve given in this figure is for water,
the general shape of the boiling curve remains the same for different fluids.

Free convection boiling

Free convection boiling is said to exist if ∆Te ≤ ∆Te,A , where ∆Te,A ≈ 5◦ C. In this regime the energy transfer
from the wire surface to the saturated liquid takes place by free convection and can be predicted by the

2
 Figure 2: Nukiyama’s boiling curve for saturated water at atmospheric pressure.

methods discussed earlier. The surface is only few degrees above the saturation temperature of the liquid,
but the free convection current produced in the liquid are sufficient to remove the heat from the surface.

Nucleate boiling
Nucleate boiling exists in the range ∆Te,A ≤ ∆Te ≤ ∆Te,C , where ∆Te,A ≈ 30◦ C. This regime, where the bubbles
are formed on the surface, two distinct regions can be identified. The region between A and B, bubbles start
to form at the favoured sites on the surface of the wire, but as soon as the bubbles are detached from the
surface, they are dissipated in the liquid. In region between B and C, the nucleation sites are numerous and
the bubble generation rate is so high that continuous column of vapour appear. As a result, very high heat
fluxes are obtainable in this region. In practical applications, the nucleate boiling regime is most desirable
because large heat fluxes are obtainable with small temperature differences. In the nucleate boiling regime,
the heat flux flux increases rapidly with increasing temperature difference until the peak heat flux is reached.
The location of this peak heat flux is called burning point or the critical heat flux. As soon as the peak heat
flux is exceeded, an extremely large temperature difference is needed to realize the resulting heat flux. Such
high temperature differences may cause the burning up, or melting away, of the wire.

Transition boiling
The region corresponding to ∆Te,C ≤ ∆Te ≤ ∆Te,D , where ∆Te,D ≈ 120◦ C, is termed transition boiling or
unstable film boiling. Bubble formation is now so rapid that a vapour film or blanket begins to form on
the surface. At any point on the surface, conditions may oscillate between film and nucleate boiling, but
the fraction of the total surface covered by the film increases with increasing ∆Te . Because the thermal
conductivity of the vapour is much less than that of the liquid, h (and q̇s ) must decrease with increasing
∆Te .

Film boiling

3
Film boiling exists for ∆Te ≥ ∆Te,D . At point D of the boiling curve, referred to as the Leidenfrost point,
the heat flux is a minimum, q̇s,D = q̇min , and the surface is completely covered by a vapour blanket. Heat
transfer from the surface to the liquid occurs by conduction and radiation through the vapour. As the
surface temperature is increased, radiation through the vapour film becomes more significant and the heat
flux increases with increasing ∆Te . Figure 1(d) illustrates the nature of the vapour formation and bubble
dynamics associated with film boiling.

3 Pool Boiling Correlations

Boiling regimes discussed above differ considerably in their character, and thus different heat transfer relations
need to be used for different boiling regimes.

3.1 Free convection regime

In the free convection boiling regime (∆Te ≤ 5◦ C), boiling is governed by free convection currents, and heat
transfer rates in this case can be determined accurately using free convection relations presented earlier.
They are in the form
Nu = f (Gr, Pr) (5)
Once the heat transfer coefficient h is obtained, the heat flux for the free convection regime is determined
from
q̇s = h(Ts − Tsat ) (6)
Several correlations are discussed to determine the free convection heat transfer coefficient h for various
geometries, such as vertical, inclined, and horizontal plates and vertical and horizontal cylinders.

3.2 Nucleate boiling regime

Nucleate boiling regime involves two separate processes: the formation of bubbles on the surface, which
is referred to as the nucleation, and the subsequent growth and motion of these bubbles. The analysis of
nucleate boiling requires prediction of the number of surface nucleation sites and the rate at which bubbles
originate from each site. The type and the condition of the heated surface also affect the heat transfer.
These complications made it difficult to develop theoretical relations for heat transfer in the nucleate boiling
regime, and we had to rely on relations based on experimental data. Yamagata et al. were the first to show
the influence of nucleation sites on the heat rate and to demonstrate that q̇s is approximately proportional
to ∆Te3 . It is desirable to develop correlations that reflect this relationship between the surface heat flux
and the excess temperature. The most widely used correlation for the rate of heat transfer in the nucleate
boiling regime was proposed in 1952 by Rohsenow, and expressed as
g(ρl − ρv ) 1/2 c pl (Ts − Tsat ) 3
   
q̇s = µl h f g (7)
σ Cs f h f g Prln
where
q̇s = nucleate boiling heat flux, W/m2
µl = viscosity of liquid, Pa s
h f g = latent heat of vaporization, J/kg
g = gravitational acceleration, m/s2

4
 ρl , ρv = density of liquid and saturated vapour, kg/m3
σ = surface tension of liquid-vapour interface, N/m
c pl = specific heat of liquid, J/kg◦ C
Ts = surface temperature of heater, ◦ C
Tsat = saturation temperature of fluid, ◦ C
Cs f = constant, determined from experimental data
Prl = Prandtl number of saturated liquid
n = 1.0 for water and 1.7 for other liquids
All properties are for the fluid should be evaluated at Tsat . In equation (7) the exponent n and the coefficient
Cs f are the two parameters to adjust the correlation for the liquid-surface combination. The surface tension
at the liquid-vapour interface for water and for some other fluids is given in the Heat Transfer Data Book.
Experimentally determined values of the constant Cs f is given in Heat Transfer Data Book for various liquid-
surface combinations. These values can be used for any geometry since it is found that the rate of heat
transfer during nucleate boiling is essentially independent of the geometry and orientation of the heated
surface.
If equation (7) is rewritten in terms of a Nusselt number based on an arbitrary length scale L, it will be
in the form
NuL ∝ Ja2 Pr1−3n Bo1/2
Comparing with equation (4), we see that only the first dimensionless parameter does not appear. If the
Nusselt number is based on the characteristic bubble diameter, the expression reduces to the simpler form
NuDb ∝ Ja2 Pr1−3n
The condition of the heater surface greatly affects heat transfer, and the Rohsenow equation given above
is applicable to clean and relatively smooth surfaces. The results obtained using the Rohsenow equation can
be in error by ±100% for the heat transfer rate for a given excess temperature and by ±30% for the excess
temperature for a given heat transfer rate. Therefore, care should be exercised in the interpretation of the
results.
Recall from thermodynamics that the latent heat of vaporization (or enthalpy of vaporization) h f g of a
pure substance decreases with increasing pressure (or temperature) and reaches zero at the critical point.
Noting that h f g appears in the denominator of the Rohsenow equation, we should see a significant rise in
the rate of heat transfer at high pressures during nucleate boiling.

Critical heat flux for nucleate boiling

In the design of boiling heat transfer equipment, it is extremely important for the designer to have a knowledge
of the maximum heat flux in order to avoid the danger of burnout. The maximum (or critical) heat flux
in nucleate pool boiling was determined theoretically through dimensional analysis by S. S. Kutateladze in
Russia in 1948 and N. Zuber in the United States in 1958 using hydrodynamic stability analysis, and is
expressed as
ρg(ρl − ρv ) 1/4
 
q̇max = Ccr ρv h f g (8)
ρv2
where
q̇max = critical heat flux, W/m2
h f g = latent heat of vaporization, J/kg

5
 g = gravitational acceleration, m/s2
ρl , ρv = density of liquid and saturated vapour, kg/m3
σ = surface tension of liquid-vapour interface, N/m
Ccr = constant, depends on surface geometry
which is independent of surface material and is weakly dependent upon the hot surface geometry through
the leading constant, Ccr . For large horizontal cylinders, for spheres, and for many large finite surfaces, use
of a leading constant with the value Ccr = π/24 ≈ 0.131 (the Zuber constant) agrees with experimental data
to within 16%. For large horizontal plates, a value of Ccr = 0.149 gives better agreement with experimental
data. The properties in equation (8) are evaluated at the saturation temperature. Equation (8) applies when
the characteristic length of the hot surface, L, is large relative to the bubble diameter, Db . However, when
the hot surface is small, such that the Confinement number,
r
σ
Co = 2
gL (ρl − ρv )
is greater than approximately 0.2, a correction factor must be applied to account for the small size of the
surface. Exhaustive experimental studies by Lienhard and his coworkers indicated that the value of Ccr is
about 0.15. Specific values of Ccr for different heater geometries are listed in Heat Transfer Data Book.
Note that ρv increases but σ and h f g decrease with increasing pressure, and thus the change in q̇max
with pressure depends on which effect dominates. The experimental studies of Cichelli and Bonilla indicate
that q̇max increases with pressure up to about one-third of the critical pressure, and then starts to decrease
and becomes zero at the critical pressure. Also note that q̇max is proportional to h f g , and large maximum
heat fluxes can be obtained using fluids with a large enthalpy of vaporization, such as water.

Minimum heat flux

Minimum heat flux, which occurs at the Leidenfrost point, is of practical interest since it represents the
lower limit for the heat flux in the film boiling regime. Using the stability theory, Zuber (1958) derived the
following expression for the minimum heat flux for a large horizontal plate,

σ g(ρl − ρv ) 1/4
 
q̇min = Cρv h f g (9)
(ρl + ρv )2
where the properties are evaluated at the saturation temperature. The constant, C = 0.09, has been experi-
mentally determined by Berenson. This result is accurate to approximately 50% for most fluids at moderate
pressures but provides poorer estimates at higher pressures. A similar result has been obtained for horizontal
cylinders.

3.3 Film boiling regime

At excess temperatures beyond the Leidenfrost point, a continuous vapour film blankets the surface and
there is no contact between the liquid phase and the surface. Bromley (1950) developed a theory for the
prediction of heat flux for stable film boiling on the outside of a horizontal cylinder. The heat flux for film
boiling on a horizontal cylinder or sphere of diameter D is given by
"
#1/4
ρv g(ρl + ρv )kv3 h f g + 0.68c pv ∆Te
hc = 0.62 (10)
µv D∆Te

6
where

hc = average heat transfer coefficient, W/m2

µv = viscosity of liquid vapour, Pa s
h f g = latent heat of vaporization, J/kg
g = gravitational acceleration, m/s2
ρl , ρv = density of the liquid and saturated vapour, kg/m3
c pv = specific heat of saturated vapour, J/kg◦ C
kv = thermal conductivity of saturated vapour, W/mK
∆Te = Ts − Tsat = excess temperature, ◦ C
D = outside diameter of tube, m

The vapour properties are to be evaluated at the film temperature T f = (Ts + Tsat )/2, which is the aver-
age temperature of the vapour film. The liquid properties and h f g are to be evaluated at the saturation
temperature at the specified pressure.
Equation (10) has been derived by assuming that heat transfer across the vapour film is by pure conduc-
tion, it does not include the radiation effects. Treating the vapour film as a transparent medium sandwiched
between two large parallel plates and approximating the liquid as a blackbody, radiation heat transfer can
be determined from  4 4

Ts − Tsat
hr = εσ (11)
Ts − Tsat
Here ε is the emissivity the heating surface and Ts is the absolute surface temperature. Bromley suggested
the average heat transfer coefficient hm can be

hm = hc + 0.75 hr (12)

which correlates experimental data well. Note that the gravitational acceleration g, whose value is approxi-
mately 9.81 m/s2 at sea level, appears in all of the relations above for boiling heat transfer. The effects of
low and high gravity (as encountered in aerospace applications and turbomachines) are studied experimen-
tally. The studies confirm that the critical heat flux and heat flux in film boiling are proportional to g1/4 .
However, they indicate that heat flux in nucleate boiling is practically independent of gravity g, instead of
√
being proportional to g, as dictated by equation (7).
Note also that the analogy between film boiling and film condensation does not hold for small surfaces
with high curvature because of the large disparity between vapour and liquid film thicknesses for the two
processes. The analogy is also questionable for a vertical surface, although satisfactory predictions have
been obtained for limited conditions.

4 Forced Convection Boiling

In previous sections we considered pool boiling on a heated surface immersed in a quiescent mass of liquid
In pool boiling, fluid flow is due primarily to the buoyancy driven motion of bubbles originating from the hot
surface. In contrast, for forced convection boiling, flow is due to a directed (bulk) motion of the fluid, as
well as to buoyancy effects. Forced convection boiling is also referred to as flow boiling. Conditions depend
strongly on geometry, which may involve external flow over hot plates and cylinders or internal (duct) flow.
Internal, forced convection boiling is commonly referred to as two-phase flow and is characterized by rapid
changes from liquid to vapor in the flow direction.

7
4.1 External forced convection boiling
For external flow over a hot plate, the heat flux can be estimated by standard forced convection correlations
up to the inception of boiling. As the temperature of the plate is increased, nucleate boiling will occur,
causing the heat flux to increase. If vapor generation is not extensive and the liquid is subcooled, Bergles
and Rohsenow [24] suggest a method for estimating the total heat flux in terms of components associated
with pure forced convection and pool boiling.
Both forced convection and subcooling are known to increase the critical heat flux q̇max for nucleate
boiling. Experimental values as high as 35 MW/m2 (compared with 1.3 MW/m2 for pool boiling of water at
1 atm) have been reported [25]. For a liquid of velocity V moving in cross flow over a cylinder of diameter D,
Lienhard and Eichhorn [26] have developed the following expressions for low and high velocity flows, where
properties are evaluated at the saturation temperature.
"  1/3 #
ρv h f gV 4
Low velocity: q̇max = 1+ (13)
π We
" #
ρv h f gV (ρl /ρv )3/4 (ρl /ρv )1/2
Hight velocity: q̇max = + (14)
π 169 19.2We1/3

where We is the weber number defines as the ratio of inertia to the surface tension force and has the form
ρvV 2 D
We = (15)
σ
The high and low velocity regions, respectively, are determined by whether the heat flux parameter q̇max /ρv h f gV
is less than or greater than [(0.275/π)(ρl /ρv )1/2 + 1]. In most cases, equations (13) and (13) correlate q̇max
data within 20%.

4.2 Internal forced convection boiling: Two-phase flow

Internal forced convection boiling, commonly referred to as two-phase flow, is much more complicated in
nature than in the pool boiling in a quiescent liquid. This so because there is no free surface for the vapor
to escape, and thus both the liquid and the vapor are forced to flow together. The two-phase flow in a
tube exhibits different flow boiling regimes, depending on the relative amounts of the liquid and the vapor
phases. This complicates the analysis even further.
Consider flow development in a vertical tube that is subjected to a constant surface heat flux, through
which fluid is moving in the upward direction, as shown in figure 3. Heat transfer to the subcooled liquid
that enters the tube is initially by single-phase forced convection and may be predicted using the correlations
discussed earlier. Farther down the tube, the wall temperature exceeds the saturation temperature of the
liquid, and vaporization is initiated in the subcooled flow boiling region. This region is characterized by large
radial temperature gradients, with bubbles forming adjacent to the hot wall and subcooled liquid flowing
near the center of the tube. The thickness of the bubble region increases farther downstream, and eventually,
the core of the liquid reaches the saturation temperature of the fluid. Bubbles can then exist at any radial
location, and the time-averaged mass fraction of vapor in the fluid, X , exceeds zero at any radial location.
This marks the beginning of the saturated flow boiling region. Within the saturated flow boiling region, the
mean vapor mass fraction defined as R
ρ u(r, x) X dAc
X = A (16)
ṁ
increases and, due to the large density difference between the vapor and liquid phases, the mean velocity of
the fluid, um , increases substantially.

8
 Figure 3: Various flow regimes encountered in flow boiling in a tube under forced convection.

The first stage of the saturated flow boiling region corresponds to the bubbly flow regime. As X increases
further, individual bubbles coalesce to form slugs of vapor. This slug-flow regime is followed by an annular-
flow regime in which the liquid forms a film on the tube wall. This film moves along the inner surface of the
tube, while vapor moves at a larger velocity through the core of the tube. Dry spots eventually appear on the
inner surface of the tube and grow in size within a transition regime. Eventually, the entire tube surface is
completely dry, and all remaining liquid is in the form of droplets that travel at high velocity within the core
of the tube in the mist regime. After the droplets are completely vaporized, the fluid consists of superheated
vapor in a second single phase forced convection region. The increase in the vapor fraction along the tube
length, along with the large difference in the densities of the liquid and vapor phases, increases the mean
velocity of the fluid by several orders of magnitude between the first and the second single-phase forced
convection regions.
The local heat transfer coefficient varies significantly as X and um decrease and increase, respectively,
along the length of the tube, x. In general, the heat transfer coefficient can increase by approximately
an order of magnitude through the subcooled flow boiling region. Heat transfer coefficients are further
increased in the early stages of the saturated flow boiling region. Conditions become more complex deeper
in the saturated flow boiling region since the convection coefficient, defined in equation (1), either increases or
decreases with increasing X , depending on the fluid and tube wall material. Typically, the smallest convection
coefficients exist in the second (vapour) forced convection region owing to the low thermal conductivity of
the vapor relative to that of the liquid.

9
 The following correlation has been developed for the saturated flow boiling region in smooth circular
tubes:  0.1  0.7
h ρl 0.16 0.64 q̇s
= 0.6683 X (1 − X) f (Fr) + 1058 (1 − X)0.8Cs, f (17a)
hsp ρv Ghfg
or  0.45  0.7
h ρl 0.72 0.08 q̇s
= 1.136 X (1 − X) f (Fr) + 667.2 (1 − X)0.8Cs, f (17b)
hsp ρv Ghfg
0 < X ≤ 0.8
where G = ṁ/Ac is the mass flow rate per unit cross-sectional area. In utilizing equation (17), the larger
value of the heat transfer coefficient, h, should be used. In this expression, the liquid phase Froude number
is
(G/ρl )2
Fr =
gD
and the coefficient Cs, f depends on the surface-fluid combination, with representative values given in Data
Book. Equation (17) applies for horizontal as well as vertical tubes, where the stratification parameter,
f (Fr), accounts for stratification of the liquid and vapor phases that may occur for horizontal tubes. Its
value is unity for vertical tubes and for horizontal tubes with Fr ≥ 0.04. For horizontal tubes with Fr ≤ 0.04,
f (Fr) = 2.63 Fr0.3 . All properties are evaluated at the saturation temperature, Tsat . The single-phase
convection coefficient, hsp , is associated with the liquid forced convection region of figure 3 and is obtained
from equation the Dittus-Boelter equation with properties evaluated at Tsat . Because Dittus-Boelter equation
is for turbulent flow, it is recommended that equation (17) not be applied to situations where the liquid
single-phase convection is laminar. Equation (17) is applicable when the channel dimension is large relative
to the bubble diameter, that is, for Confinement numbers,
r
σ
Co = 2
gD (ρl − ρv )

In order to use equation (17), the mean vapor mass fraction, X , must be known. For negligible changes in
the fluid’s kinetic and potential energy as well as negligible work, we may write
π q̇s D x
X(x) = (18)
ṁ h f g

where the origin of the x-coordinate, x = 0, corresponds to the axial location where X begins to exceed zero,
and the change in enthalpy, ut + pv, is equal to the change in X multiplied by the enthalpy of vaporization,
h f g.
Correlations for the subcooled flow boiling region and annular as well as mist regimes are available in
the literature. For constant heat flux conditions, critical heat fluxes may occur in the subcooled flow boiling
region, in the saturated flow boiling region where X is large, or in the vapor forced convection region. Critical
heat flux conditions may lead to melting of the tube material in extreme conditions. Additional discussions
of flow boiling are available in the literature. Extensive databases consisting of thousands of experimentally
measured values of the critical heat flux for wide ranges of operating conditions are also available.

5 Condensation Heat Transfer

Condensation occurs when the temperature of a vapour is reduced below its saturation temperature. This
is usually done by bringing the vapor into contact with a solid surface whose temperature Ts is below the

10
saturation temperature of the vapour. The steam condensers for the power plants are typical examples of
the application of condensing of steam.
Two distinct forms of condensation are observed: film condensation and dropwise condensation. In
film condensation, the condensate wets the surface and forms a liquid film on the surface that slides down
under the influence of gravity. The thickness of the liquid film increases in the flow direction as more vapor
condenses on the film. This is how condensation normally occurs in practice. In dropwise condensation, the
condensed vapor forms droplets on the surface instead of a continuous film, and the surface is covered by
countless droplets of varying diameters (figure 4).
In film condensation, the surface is blanketed by a liquid film of increasing thickness, and this “liquid
wall” between solid surface and the vapor serves as a resistance to heat transfer. In dropwise condensation,
however, the droplets slide down when they reach a certain size, clearing the surface and exposing it to vapor.
There is no liquid film in this case to resist heat transfer. As a result, heat transfer rates that are more than
10 times larger than those associated with film condensation can be achieved with dropwise condensation.
Therefore, dropwise condensation is the preferred mode of condensation in heat transfer applications.

Figure 4: Modes of condensation on a surface.

Reynolds number for condensate flow

As was the case in forced convection involving a single phase, heat transfer in condensation also depends
on whether the condensate flow is laminar or turbulent. Again the criterion for the flow regime is provided
by the Reynolds number, which is defined as

Dh ρl Vl 4Ac ρl Vl
Re = = (19)
µl Pµl

where Vl is the average velocity of the condensate film and Dh is the hydraulic diameter for the condensate
flow, given by
4Ac cross-sectional area for condensate flow
Dh ≡ = 4×
P wetted perimeter
The Reynolds number at the lowest part of the condensing surface can be expressed in a more convenient
for as
4Ṁ
Re = (20)
Pµl

11
where Ṁ = ṁx=L is the mass flow rate of condensate at the lower part of the condensing surface. The wetted
perimeter depends on the geometry of the condensing surface

b

 for vertical or inclined plate of width b
P= 2L for horizontal surface of length L


πD vertical tube of outside diameter D

Experiments have shown that the transition form laminar to turbulent flow condensation takes place at a
Reynolds number about 1800.

5.1 Film condensation on a vertical plate

Despite the complexities associated with film condensation, useful results may be obtained by making
assumptions that originated with an analysis by Nusselt. Consider the condensation of vapour on a vertical
plate, as shown in figure 5. Here x is the axial coordinate measured downward along the plate, and y is the
coordinate normal to the condensing surface. The condensate thickness is represented by δ = δ (x). The
following assumptions are made for the analysis.

Figure 5: Boundary layer effects related to film condensation on a vertical surface. (a) Without approximation. (b)
With assumptions associated with Nusselt’s analysis, for a vertical plate of width b.

1. The plate is maintained at a uniform temperature Ts that is less than the saturation temperature Tsat
of the vapour.
2. The fluis properties are constant.
3. The vapour is stationary or has a low velocity, and so it exerts no drag on the motion of the condensate.

12
 4. The downward flow flow of condensate under the action of gravity is laminar.
5. The flow velocity associated with the condensate film is low, as a result the advection terms are
negligible.
6. Heat transfer across the condensate layer is by pure conduction, hence the liquid temperature distri-
bution is linear.

The x-momentum equation for the film can be written with ρ = ρl and ν = νl for the liquid. The pressure
gradient is obtained under free stream conditions and is
d p∞
= ρv g
dx
since the free stream density is the vapor density. From the fifth approximation, momentum advection terms
may be neglected, and the x-momentum equation may be expressed as

∂ 2u
µl = −g(ρl − ρv ) (21)
∂ y2
Integrating once and applying boundary condition of the form ∂ u/∂ y|y=δ = 0, we obtain

∂u g(ρl − ρv )
= (δ − y) (22)
∂y µl
where δ is the thickness of condensate layer (film thickness) at the position x. Integrating once more and
using the no-slip boundary condition u(0) = 0, we obtain an expression for velocity profile:

y2
 
g(ρl − ρv )
u(y) = yδ − (23)
µl 2

The mass flow rate of the condensate at a location x, where the boundary layer thickness (film thickness) is
δ , is determined from Z Z δ
ṁ(x) = ρl u dA = b ρl u dy (24)
Ac 0
where Ac is the cross-section area (normal to x-direction) of the condensate and b is the width of the plate.
Substituting the u(y) relation from equation (23) into equation (24) gives

gbρl (ρl − ρv )δ 3
ṁ(x) = (25)
3µl
whose derivative with respect to x is
d ṁ gbρl (ρl − ρv ) 2 dδ
= δ (26)
dx µl dx
which represents the rate of condensation of vapor over a vertical distance dx. The rate of heat transfer from
the vapor to the plate through the liquid film is simply equal to the heat released as the vapor is condensed
and is expressed as
d Q̇ = h f g d ṁ (27)
where h f g is the latent of condensation. The amount of heat released d Q̇ over the area dAs = bdx must be
transferred across the condensate layer of thickness δ by conduction, according to assumption 6. Therefore,
(Tsat − Ts )
d Q̇ = kl (bdx) (28)
δ

13
where kl is the thermal conductivity of the liquid. Equating equations (27) and (28), we obtain

d ṁ kl b (Tsat − Ts )
= (29)
dx hfg δ

Equating equations (26) and (29) to each other and separating the variables, we obtain the following differ-
ential equation for the thickness of the condensate layer:

dδ µl kl (Tsat − Ts ) 1
= (30)
dx gρl (ρl − ρv )h f g δ 3

Integrating from x = 0 where δ = 0 (the top of the plate) to any x location of interest on the surface, the
liquid film thickness at any location x is determined to be
 1/4
4µl kl (Tsat − Ts )x
δ (x) = (31)
gρl (ρl − ρv )h f g

Since we have established the relation for the thickness of the condensate later, the local heat transfer
coefficient hx for the condensation is determined from the definition

(Tsat − Ts ) kl
hx (Tsat − Ts ) = kl → hx = (32)
δ δ (x)
or
1/4
gρl (ρl − ρv )h f g kl3

hx (x) = (33)
4µl (Tsat − Ts )x
The average heat transfer coefficient over the entire plate over the length 0 ≤ x ≤ L is determined from its
definition by substituting the hx relation and performing the integration. It gives
Z L
1 4
hL = hx dx = hL (34)
L 0 3

where hL is the local heat transfer coefficient h(x) at x = L. Introducing equation (33) into equation (34),
we obtain
1/4
gρl (ρl − ρv )h f g kl3

hL = 0.943 (35)
µl (Tsat − Ts )L
The average Nusselt number then has the form
1/4
gρl (ρl − ρv )h f g L3

hL L
NuL = = 0.943 (36)
kl µl kl (Tsat − Ts )

The physical properties in equations (33), (35), and (36) should be evaluated at the film temperature
T f = (Tsat + Ts )/2. The vapor density ρv and latent heat of vaporization h f g should be evaluated at Tsat .
Equation (35), which is obtained with the simplifying assumptions stated earlier, provides good insight
on the functional dependence of the condensation heat transfer coefficient. However, it is observed to under-
predict heat transfer because it does not take into account the effects of the nonlinear temperature profile
in the liquid film and the cooling of the liquid below the saturation temperature. Both of these effects can
be accounted for by replacing h f g by h∗f g as recommended by Rohsenow:

h∗f g ≡ h f g + 0.68 c pl (Tsat − Ts ) = h f g (1 + 0.68 Ja) (37)

14
where c pl is the specific heat of the liquid at the average film temperature and Ja = c pl (Tsat − Ts )/h f g is the
Jakob number. With this modification, the average heat transfer coefficient for laminar film condensation
over a vertical flat plate of height L is determined to be
#1/4
gρl (ρl − ρv )h∗f g kl3
"
hL = 0.943 0 < Re < 30 (38)
µl (Tsat − Ts )L

where the heat transfer coefficient has units of (W/m2 K) and the Reynolds number is given by

4Ṁ
Re = (39)
Pµl

where the mass flow rate Ṁ = ṁ|x=L . Using the relation

Q̇ = h f g Ṁ = hL As (Tsat − Ts ) (40)

the expression for Reynolds number may be rewritten as

4Q̇ 4As hL (Tsat − Ts )

Re = = (41)
Pµl h f g Pµl h f g

where As is the surface area of the plate on which the condensation occurs and P = b for vertical surfaces.
This relation is convenient to use to determine the Reynolds number when the condensation heat transfer
coefficient or the rate of heat transfer is known.
The Reynolds number can also be expressed in terms of film thickness δL = δ (x = L). To obtain such a
relation we substitute equation (25) into equation (39):

4gρl (ρl − ρv )δL3

Re = (42)
3µl2

An approximate expression for the heat transfer coefficient hL can be obtained for the case ρv  ρl (that
is, ρl − ρv ≈ ρl ). Noting that δL = kl /hL , the Reynolds number becomes
3 3
4gρl (ρl − ρv )δL3 4gρl2
 
kl 4g kl
Re = ≈ = (43)
3µl2 3µl2 hL 3νl2 3hL /4

where νl is the kinematic viscosity of liquid condensate. Then the average heat transfer coefficient hL in
terms of Re becomes

1.47 kl g 1/3
 
hL ≈ (44)
Re1/3 νl2
[0 < Re < 30, ρv  ρl ]

The average Nusselt number then has the form

hL (νl2 /g)1/3 1.47

NuL = = 0 < Re < 30 (45)
kl Re1/3
The results obtained from the theoretical relations above are in excellent agreement with the experimental
results.

15
 Figure 6: Flow regimes during film condensation on a vertical plate.

Wavy laminar flow on a vertical plate

At Reynolds numbers greater than about 30, it is observed that waves form at the liquid-vapor interface
although the flow in liquid film remains laminar (figure 6). The flow in this case is said to be wavy laminar.
The waves at the liquid-vapor interface tend to increase heat transfer. But the waves also complicate the
analysis and make it very difficult to obtain analytical solutions. Therefore, we have to rely on experimental
studies. The increase in heat transfer due to the wave effect is, on average, about 20%, but it can exceed
50%. The exact amount of enhancement depends on the Reynolds number. Based on his experimental
studies, Kutateladze (1963) recommended the following relation for the average heat transfer coefficient in
wavy laminar condensate flow,
 1/3
Re kl g
hL = (46)
1.08 Re1.22 − 5.2 νl2
[30 < Re < 1800, ρv  ρl ]

The average Nusselt number then has the form

hL (νl2 /g)1/3 Re
NuL = = 30 < Re < 1800 (47)
kl 1.08 Re1.22 − 5.2

A relation for the Reynolds number in the wavy laminar region can be determined by substituting the hL
relation in equation (46) into the Re relation in (41) and simplifying. It yields (for ρv  ρl ),
"  1/3 #0.82
3.7 L kl (Tsat − Ts ) g
Re = 4.81 + (48)
µl h∗f g v2l

Turbulent flow on a vertical plate

At a Reynolds number of about 1800, the condensate flow becomes turbulent. Several empirical relations
of varying degrees of complexity are proposed for the heat transfer coefficient for turbulent flow. Again
assuming ρv  ρl for simplicity, Labuntsov (1957) proposed the following relation for the turbulent flow of

16
condensate on vertical plates:
 1/3
Re kl g
hL = −0.5
(49)
8750 + 58 Prl (Re0.75 − 253) νl2
[Re > 1800, Prl ≥ 1, ρv  ρl ]

The average Nusselt number then has the form

hL (νl2 /g)1/3 Re
NuL = = −0.5
Re > 1800, Pr, ≥ 1 (50)
kl 8750 + 58 Prl (Re0.75 − 253)

The physical properties of the condensate are again to be evaluated at the film temperature T f = (Tsat +Ts )/2.
The Re relation in this case is obtained by substituting the hL relation (49) into the Re relation in (41),
which gives
" #4/3
0.069 L kl Prl0.5 (Tsat − Ts ) g 1/3
 
0.5
Re = − 151 Prl + 253 (51)
µl h∗f g v2l
Graphical representation of the foregoing correlations for the wave-free laminar, wavy laminar, and turbulent
flow are provided in figure 7, and the trends have been verified experimentally by Gregorig et al. for water
over the range 1 < Re < 7200.

Figure 7: Nondimensionalized heat transfer coefficients (Nusselt numbers) for condensation on vertical plates.

Laminar flow on a vertical tube

Equation (38) for vertical plates can also be used to calculate the average heat transfer coefficient for laminar
film condensation on the outer surfaces of vertical tubes provided that the tube diameter is large relative to
the thickness of the liquid film (D  δL ). As a check, the thickness of the liquid film at the bottom of the
tube δL may be calculated by using δL = kl /hL in order to see the condition D  δL is satisfied or not.

17
5.2 Film condensation on an inclined plate
The correlations developed for laminar film condensation on a vertical flow can be readily extended for film
condensation on an inclined plate, making an angle θ with the vertical, as illustrated in figure 8. The results

Figure 8: Film condensation on an inclined plate.

for the local and the average heat transfer coefficients are given respectively as
#1/4
gρl (ρl − ρv )h∗f g kl3
"
hx (x) = cos θ (52)
4µl (Tsat − Ts )x
and #1/4
gρl (ρl − ρv )h∗f g kl3
"
hL = 0.943 cos θ (53)
µl (Tsat − Ts )L
This approximation gives satisfactory results especially for θ ≤ 60◦ .

5.3 Film condensation on horizontal tubes and tube banks

The analysis of heat transfer for condensation on the outside surface of a horizontal tube is more complicated
than that for a vertical surface. Nusselt’s analysis for laminar film condensation may be extended to horizontal
tubes and the average heat transfer coefficient is determined to be
#1/4
gρl (ρl − ρv )h∗f g kl3
"
hD = 0.729 (54)
µl (Tsat − Ts )D

where D is the outside diameter of the tube and the heat transfer coefficient has units of (W/m2 K). It may
be noted that equation (54) can easily be modified for a sphere of diameter D by replacing the constant
0.729 by 0.815.
A comparison of the heat transfer coefficient relations for a vertical tube of height L [equation (38)] and
a horizontal tube of diameter D [equation (54)] yields
 1/4
hvert D
= 1.29 (55)
hhorz L
Setting hvert = hhorz gives L = 1.294 D = 2.77D, which implies that for a given Tsat − Ts , the average heat
transfer coefficients for a vertical tube of length L and a horizontal tube of diameter D become equal when
L = 2.77 D. Considering that the length of a tube in any practical application is several times its diameter,
it is common practice to place the tubes in a condenser horizontally to maximize the condensation heat
transfer coefficient on the outer surfaces of the tubes.

18
Horizontal tube banks
Condenser design generally involves horizontal tubes stacked on top of each other as shown in figure 9. The
average thickness of the liquid film at the lower tubes is much larger as a result of condensate falling on

Figure 9: Film condensation on (a) a single horizontal tube and (b) a vertical tier of horizontal tubes with a
continuous condensate sheet.

top of them from the tubes directly above. Therefore, the average heat transfer coefficient at the lower
tubes in such arrangements is smaller. Assuming the condensate from the tubes above to the ones below
drain smoothly, the average film condensation heat transfer coefficient for N tubes each of diameter D, in
a vertical tier can be expressed as
#1/4
gρl (ρl − ρv )h∗f g kl3
"
 1 
hD N -tubes = 0.729 = 1/4 hD 1-tube (56)
µl (Tsat − Ts )ND N

This relation does not account for the increase in heat transfer due to the ripple formation and turbulence
caused during drainage, and thus generally yields conservative results.

5.4 Film condensation inside horizontal tubes

Condensers used for refrigeration and air-conditioning systems generally involve vapor condensation inside
horizontal or vertical tubes. Heat transfer analysis of condensation inside tubes is complicated by the fact
that it is strongly influenced by the vapor velocity and the rate of liquid accumulation on the walls of the
tubes as illustrated in figure 10. At high vapor velocities [figure 10(a)], the flow is characterized by two-

Figure 10: Condensate flow in a horizontal tube with high and low vapor velocities.

phase annular conditions. The core of the annulus is occupied by the vapor and its diameter decreases as

19
the condensate thickness increases in the direction of the flow. At low vapor velocities [figure 10(b)], the
condensate flow is from the upper portion of the tube to the bottom of the tube.
Chato (1962) recommends the following condensation at low vapour velocities inside the tubes:
#1/4
gρl (ρl − ρv )h0f g kl3
"
hD = 0.555 (57)
µl (Tsat − Ts )D

where
3
h0f g ≡ h f g + c pl (Tsat − Ts ) (58)
8
This result has been developed for the condensation at low-vapour Reynolds number such that
ρvVv D
Rev ≡ < 35, 000 (59)
µv
where the Reynolds number of the vapor is to be evaluated at the tube inlet conditions using the internal
tube diameter as the characteristic length. Heat transfer coefficient correlations for higher vapor velocities
are given by Rohsenow (1973).

5.5 Dropwise condensation

Dropwise condensation, characterized by countless droplets of varying diameters on the condensing surface
instead of a continuous liquid film, is one of the most effective mechanisms of heat transfer, and extremely
large heat transfer coefficients can be achieved with this mechanism.
In dropwise condensation, the small droplets that form at the nucleation sites on the surface grow as
a result of continued condensation, coalesce into large droplets, and slide down when they reach a certain
size, clearing the surface and exposing it to vapor. There is no liquid film in this case to resist heat
transfer. As a result, with dropwise condensation, heat transfer coefficients can be achieved that are more
than 10 times larger than those associated with film condensation. Large heat transfer coefficients enable
designers to achieve a specified heat transfer rate with a smaller surface area, and thus a smaller (and
less expensive) condenser. Therefore, dropwise condensation is the preferred mode of condensation in heat
transfer applications.
Dropwise condensation has been studied experimentally for a number of surface-fluid combinations. Of
these, the studies on the condensation of steam on copper surfaces has attracted the most attention because
of their widespread use in steam power plants. P. Griffith (1983) recommends these simple correlations for
dropwise condensation of steam on copper surfaces:

hdrop = 51, 104 + 2044 Tsat 22◦ C ≤ Tsat ≤ 100◦ C (60a)

◦
hdrop = 255, 510 Tsat ≥ 100 C (60b)
where the heat transfer coefficient has units of (W/m2 K). The very high heat transfer coefficients achievable
with dropwise condensation are of little significance if the material of the condensing surface is not a good
conductor like copper or if the thermal resistance on the other side of the surface is too large.

Problems for Practice

The following problems are worked out examples taken for various books. You are encouraged to solve the
problems without referring to the solutions given in the books. If you are stuck during the attempt to solve
a problem, you may refer to the relevant book.
Note: Use ‘Heat Transfer Data Book’ to obtain properties of fluids and solids.

20
 1. A heated brass plate is submerged in a container of water at atmospheric pressure. The plate tem-
perature is 110◦ C. Calculate the heat transfer per unit area of plate. (Holeman – Example 9.3)

2. The bottom of a copper pan, 0.3 m in diameter, is maintained at 118◦ C by an electric heater. Estimate
the power required to boil water in this pan. Estimate the critical heat flux. (Bergman, Lavine,
Incropera, and DeWitt – Example 10.1)

3. Water is to be boiled at atmospheric pressure in a mechanically polished stainless steel pan placed
on top of a heating unit. The inner surface of the bottom of the pan is maintained at 108◦ C. If the
diameter of the bottom of the pan is 30 cm, determine (a) the rate of heat transfer to the water and
(b) the rate of evaporation of water. (Cengel and Ghajar – Example 10.1)

4. Water in a tank is to be boiled at sea level by a 1 cm-diameter nickel plated steel heating element
equipped with electrical resistance wires inside. Determine the maximum heat flux that can be attained
in the nucleate boiling regime and the surface temperature of the heater in that case. (Cengel and
Ghajar – Example 10.3)

5. Water is boiled at atmospheric pressure by a horizontal polished copper heating element of diameter
D = 5 mm and emissivity ε = 0.05 immersed in water. If the surface temperature of the heating wire
is 350◦ C, determine the rate of heat transfer from the wire to the water per unit length of the wire.
(Cengel and Ghajar – Example 10.3)

6. Water at 5 atm flows inside a tube of 1-in [2.54 cm] diameter under local boiling conditions where
the tube wall temperature is 10◦ C above the saturation temperature. Estimate the heat transfer in a
1.0 m length of tube. (Holeman – Example 9.4)

7. A vertical square plate, 30 by 30 cm, is exposed to steam at atmospheric pressure. The plate
temperature is 98C. Calculate the heat transfer and the mass of steam condensed per hour. (Holeman
– Example 9.1)

8. The outer surface of a vertical tube, which is 1 m long and has an outer diameter of 80 mm, is
exposed to saturated steam at atmospheric pressure and is maintained at 50◦ C by the flow of cool
water through the tube. What is the rate of heat transfer to the coolant, and what is the rate at
which steam is condensed at the surface? (Bergman, Lavine, Incropera, and DeWitt – Example 10.3)

9. Saturated steam at atmospheric pressure condenses on a 2 m-high and 3 m-wide vertical plate that is
maintained at 80◦ C by circulating cooling water through the other side (Fig. 1030). Determine (a)
the rate of heat transfer by condensation to the plate and (b) the rate at which the condensate drips
off the plate at the bottom. (Cengel and Ghajar – Example 10.4)

10. The condenser of a steam power plant operates at a pressure of 7.38 kPa. Steam at this pressure
condenses on the outer surfaces of horizontal tubes through which cooling water circulates. The
outer diameter of the pipes is 3 cm, and the outer surfaces of the tubes are maintained at 30◦ C.
Determine (a) the rate of heat transfer to the cooling water circulating in the tubes and (b) the rate
of condensation of steam per unit length of a horizontal tube. (Cengel and Ghajar – Example 10.6)

Resources
1. Bergman, T. L., Lavine, A., Incropera, F. P., and DeWitt, D. P, Introduction to Heat Transfer, 6th
ed., John Wiley (2011).

21
2. Cengel, Y. A. and Ghajar, A. J., Heat and Mass Transfer: Fundamentals & Applications, 6th ed.,
McGraw-Hill (2019).

3. Holman, J. P., Heat Transfer, 10th ed., McGraw-Hill (2010).

4. Ozisik, M. N., Heat Transfer: A Basic Approach, McGraw-Hill (1985).

22