J. Fluid Mech. (2017), vol. 827, pp. 1–30.

c Cambridge University Press 2017 1

doi:10.1017/jfm.2017.421

Confinement effects in dip coating

Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 22 Aug 2017 at 05:30:17, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.

Onyu Kim1 and Jaewook Nam1, †

1 Department of Chemical Engineering, Sungkyunkwan University, 2066 Seobu-ro,
Jangan-gu, Suwon-si, Gyeonggi-do 16419, Korea

(Received 11 February 2016; revised 8 June 2017; accepted 12 June 2017)

When a flat plate is withdrawn from a liquid pool, a liquid film is deposited on
the plate. This simple process is called dip coating. In the case of vertically upward
withdrawal, gravity competes with the surface tension and viscous drag, and the
balance between those determine the meniscus shape and hence the film thickness.
Most of the previous studies on dip coating assumed that the pool is sufficiently large
so that the stationary container wall does not affect the film thickness. However, the
cases where the stationary wall affects the entrained film have not been examined
thoroughly so far. In this confined dip coating, the film thickness deviates from that
of unconfined dip coating under the same conditions such as the withdrawal speed
and the physical properties of the liquid. The meniscus in a confined pool is more
curved than that in an unconfined pool owing to wetting on the stationary wall,
which is parallel to the plate. Besides, a channel between the moving plate and the
stationary wall appears; therefore, the flow inside the channel should be included in
an analysis of confined dip coating. In the present study, we analyse the mechanism
that determines the film thickness, both theoretically and numerically.
Key words: capillary flows, coating, thin films

1. Introduction
Perhaps the simplest way to produce a thin liquid film is to deposit a layer
of liquid onto a substrate via withdrawal from a liquid pool. The physics of the
process is known as the drag-out problem or the Landau–Levich problem, after
pioneering studies by Landau & Levich (1942). The corresponding prototypical dip
coating flow is popular not only in laboratory experiments (Scriven 1988) but also
in industry (Schunk, Hurd & Brinker 1997) because of its simplicity. Dip coating
and related methods are sufficiently flexible to produce films on different geometries
(Bretherton 1961; Wilson 1982; Park & Homsy 1984; de Ryck & Quéré 1996) and
can be applied to fluids with various properties (Krechetnikov & Homsy 2005, 2006;
Afanasiev, Münch & Wagner 2007; Dixit & Homsy 2013a,b; Javidi, Pope & Hrymak
https://doi.org/10.1017/jfm.2017.421

2016; Maillard et al. 2016). The study of Landau & Levich (1942) can be also
extended to related systems such as the formation of Marangoni-driven films on a
heated substrate partially immersed in a liquid pool (Münch 2002; Münch & Evans
2005, 2006).

† Email address for correspondence: jaewooknam@skku.edu

 2 O. Kim and J. Nam
In dip coating, the constant film thickness h∗∞ is determined by the balance between
the viscous drag, surface tension, gravity and inertia; and characterised as a function
of the capillary number Ca defined as µus /σ , where µ is the fluid viscosity, us is the
substrate speed and σ is the surface tension. For Newtonian fluids, Landau & Levich
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 22 Aug 2017 at 05:30:17, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.

(1942) determined h∗∞ for low Ca at which gravitational drainage can be negligible.
A number of authors extended the study of Landau & Levich (1942) by including
gravitational drainage (White & Tallmadge 1965; Spiers, Subbaraman & Wilkinson
1974; Wilson 1982), inertia (Kheshgi, Kistler & Scriven 1992; de Ryck & Quéré 1998)
or the normal viscous stress at the meniscus (Spiers et al. 1974; Esmail & Hummel
1975b) in their studies.
Numerical analyses (e.g. finite element (FE) or finite volume computations) have
also been commonly used to determine h∗∞ because no simple film-thickness-predicting
equation covers low to high Ca and to extend the study to complex fluids (Réglat,
Labrie & Tanguy 1993; Jin, Acrivos & Münch 2005; Jenny & Souhar 2009; Campana
et al. 2010; Abedijaberi et al. 2011; Filali, Khezzar & Mitsoulis 2013).
Most of the previous studies on dip coating, including those referred to earlier,
have considered withdrawal of a substrate in the direction opposite to gravity from
an sufficiently large pool. In this unconfined dip coating system, most of the pool
remains virtually static, except in the small region near the moving substrate, where
the flow can be recognised inside the extremely curved meniscus viewed in the far
field (Krechetnikov & Homsy 2006). However, the results of those studies cannot be
directly applied to many practical situations in which the pool is confined, such as a
coating on fibre (Quéré 1999) or laboratory experiments (Ahn et al. 2015). Although
Kheshgi et al. (1992) considered a confined pool in the dip coating problem, his study
considered a meniscus distortion due to an obstacle near the static meniscus.
In the present study, we define the confined pool as a pool where the stationary
container wall parallel to the substrate has an effect on the film thickness and forms
a channel with the substrate. There are, however, no limitations along the withdrawal
direction, as in the usual unconfined dip coating. In a confined dip coating system,
two clearly distinct flow regimes appear depending on the capillary number Ca: the
meniscus-controlled regime for low Ca and the channel-controlled regime for high Ca.
Typical dip coating flows under these regimes are shown schematically in figure 1.
The film thickness is dominantly controlled by the meniscus shape in the meniscus-
controlled regime and by the channel width, which is defined as the distance between
the substrate and the stationary wall, in the channel-controlled regime. Both regimes
exhibit distinct scaling behaviours. We analysed these behaviours theoretically and
numerically. The results of the present study can be used to predict the film thickness
entrained from a confined pool that can be encountered in thin-film productions in
industry or laboratory experiments.

2. Formulation of problem
In dip coating problems, the flow can be divided into three regions: the film
entrainment region (region I), the dynamic meniscus region (region II) and the static
https://doi.org/10.1017/jfm.2017.421

meniscus region (region III), as shown in figure 1. In the film entrainment region, the
deposited liquid film translates along the moving plate. The characteristic length lf for
this region can be easily determined by considering the balance between the viscous
and gravitational forces. In the static meniscus region, the meniscus shape in that
region is determined by the balance between the capillary and gravitational forces.
A simple force balance analysis also yields the characteristic length lc , known as
 Confinement effects in dip coating 3

(a) Moving plate

(b)
I
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 22 Aug 2017 at 05:30:17, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.

Meniscus shape
at high-Ca limit
I

II

Container wall
II

III IV

IV
l

F IGURE 1. Schematic of confined dip coating system for (a) meniscus-controlled regime
and (b) channel-controlled regime. (a) For the flow in the meniscus-controlled regime, i.e.
the region excluding the regions close to the moving plate (I and II), the liquid is virtually
stationary (III and IV). (b) For the channel-controlled regime, the static pool disappears,
and the dynamic meniscus region (II) fills the channel. The dotted line represents the
meniscus shape at the high-Ca limit. See appendix A for a discussion of region IV.

the capillary length. The aforementioned two regions are connected by the dynamic
meniscus region where the dynamic effects due to liquid in motion should be
considered. In this region, the meniscus shape is set by the balance between the
viscous drag, surface tension and gravity, provided inertia is absent. The meniscus
profile along the moving plate determines the constant film thickness h∗∞ . In addition
to the aforementioned three regions, when a pool is confined, the channel region
(region IV) appears far below the meniscus, and the characteristic length of this
region is the channel width l. The flow in this region is fully developed as described
in appendix A.
For unconfined dip coating, Landau & Levich (1942) determined h∗∞ at a low
capillary number Ca, under which gravitational drainage is negligible, as

h∗∞ = 0.9458lc Ca2/3 = 0.9458lf Ca1/6 . (2.1)

√ √
lc ≡ σ /ρg or lf ≡ µus /ρg has been employed as the characteristic length of the
flow, where ρ is the liquid density and g is the gravitational acceleration. These
characteristic lengths imply that gravity has a significant influence on unconfined
dip coating flows. The power-law dependence on Ca varies depending on the choice
https://doi.org/10.1017/jfm.2017.421

of characteristic length. Several authors have chosen lc as the characteristic length

(Campana et al. 2010; Dixit & Homsy 2013a), whereas others have employed lf
(White & Tallmadge 1965; Esmail & Hummel 1975b; Weinstein & Ruschak 2001;
Jin et al. 2005; Benilov & Zubkov 2008), depending on convenience for relevant
analyses. However, we show that the channel width l is the appropriate characteristic
length for confined dip coating flows.
 4 O. Kim and J. Nam
2.1. Dimensionless numbers
The fundamental problem of a confined dip coating system is determining the
dependence of the film thickness h∗∞ on the system parameters (l, us and g), and
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 22 Aug 2017 at 05:30:17, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.

the physical properties of the fluid (ρ, µ and σ ). In the present study, it is assumed
that a substrate is a flat plate and the coating liquid is a pure Newtonian fluid (i.e.
there are no surfactants or particles) that wets the plate completely. For the flow
where the foregoing assumptions are valid, the Buckingham π theorem yields four
dimensionless numbers describing the flow system uniquely. The numbers are the
dimensionless thickness T, the dimensionless channel width L, the capillary number
Ca and the material number m:
s
h∗∞ l µus ρσ 3
T≡ , L≡ √ , Ca ≡ , m≡ . (2.2a−d)
l σ /ρg σ gµ4

Note that three dimensionless numbers are required to define an unconfined dip
coating flow uniquely (Tallmadge & Soroka 1969); however, the newly introduced
parameter, the channel width l, generates the additional dimensionless number
L. Another important number, the Reynolds number Re, can be obtained by the
combination of the other numbers (2.2):

Re ≡ ρus h∗∞ /µ = mTLCa. (2.3)

Here, we limit our study to a Stokes flow (Re  1); T is solely determined by Ca
and L. Accordingly, we set the value of m to 10−5 in FE computations. Under this
condition, we can assume a confined dip coating flow to be a Stokes flow because
we explore the range in Ca up to 102 such that Re ' mCa.

2.2. Governing equations and boundary conditions

We consider a steady flow in a confined dip coating system where the flat plate
moves vertically with a constant velocity v ∗s (the asterisk denotes the corresponding
dimensional variables throughout the present paper), as shown in figure 2, and only
the part of the meniscus closer to the moving plate is examined in the present study.
When the inertia is negligible, the flow is governed by the Stokes equations:

∇ ∗ · v ∗ = 0, (2.4)
∇ ∗ · T ∗ + ρ g = 0, (2.5)

where v ∗ is the velocity, T ∗ is the total stress tensor and g is the gravitational field.
The no-slip conditions at the moving plate and stationary wall are

v ∗ = v ∗s at y∗ = 0, (2.6)
https://doi.org/10.1017/jfm.2017.421

v ∗ = 0 at y∗ = l. (2.7)

The stress balance and the kinematic condition at the meniscus are

nm · T ∗ = σ κ ∗ nm at y∗ = h∗ (x∗ ), (2.8)
nm · v ∗ = 0 at y∗ = h∗ (x∗ ), (2.9)
 Confinement effects in dip coating 5
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 22 Aug 2017 at 05:30:17, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.

A
x

S
x

I II III IV

F IGURE 2. Formulation of the problem. All variables are scaled for the static meniscus
region. A meniscus profile h is represented as a function of the x coordinate. The dotted
curve represents an imaginary static meniscus with a contact angle θc in the absence of
flow. The solid line represents the meniscus shape due to the flow. The meniscus shape
deviates from the shape of the static meniscus in the dynamic meniscus region (region II).
In the inset, κ0 is the curvature at the centre of the imaginary static meniscus and is
discussed in § 3.

where nm is the outward unit normal vector and is given by

 "  ∗ 2 #1/2
dh∗

dh
nm = − ∗ i + j 1+ , (2.10)
dx dx∗

where i, j are the unit vectors in the positive directions of the axes; and κ ∗ is the
curvature of meniscus:
κ ∗ = −∇ ∗ · nm . (2.11)
Here, we set the pressure on the gas side to zero. Finally, the stress tensor T ∗ of a
pure Newtonian fluid is given by

T ∗ = −p∗ I + µ[∇ ∗ v ∗ + (∇ ∗ v ∗ )T ], (2.12)

where p∗ is the pressure and I is the identity tensor.

https://doi.org/10.1017/jfm.2017.421

2.3. Dynamic meniscus region

In this region, the balance among the viscous drag, surface tension and gravity is
important, and the scaling for variables need to be chosen appropriately to examine the
balance. Following Park & Homsy (1984), we non-dimensionalise the Stokes equations
 6 O. Kim and J. Nam
and boundary conditions by the following scales:
σ
(x∗ , y∗ ) ∼ (lCa1/3 , lCa2/3 ), h∗ ∼ lCa2/3 .
(u∗ , v ∗ ) ∼ (us , us Ca1/3 ), p∗ ∼ ,
l
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 22 Aug 2017 at 05:30:17, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.

(2.13a−d)
The variables in these scales are denoted by overbars, and the equations and conditions
in the component form become
∂u ∂v
+ = 0, (2.14)
∂x ∂y
2/3 ∂ u ∂ 2 u ∂p
2

1/3 2 
Ca + 2− − Ca L = 0,
∂x 2
∂y ∂x


(2.15)
4/3 ∂ v 2/3 ∂ v ∂p
2 2 
Ca + Ca − = 0, 

∂x2 ∂y2 ∂y

u = 1, v = 0 at y = 0, (2.16a,b)
 "  2 # 
∂u ∂v ∂u 2/3 ∂v
 
2/3 dh 2/3 dh 
2Ca − + Ca −1 + Ca = 0, 

dx ∂x ∂y dx ∂y ∂x



"  2 #−3/2 
2/3 ∂u 4/3 ∂v 2/3 ∂v d2 h
 
dh 2/3 dh 
p+ Ca + Ca − 2Ca = − 2 1 + Ca


∂y ∂x ∂y

dx dx dx 

at y = h(x), (2.17)

dh
u =v at y = h(x). (2.18)
dx
The importance of gravity in the dynamic meniscus region is represented by
Gd ≡ Ca1/3 L2 , (2.19)
which appears in the first equation of (2.15).

2.4. Static meniscus region

In this region, when the viscous force is negligible, the meniscus shape is determined
by the balance between gravity and surface tension. Again, following Park & Homsy
(1984), we use the following scales:
σ
(x∗ , y∗ ) ∼ (l, l), (u∗ , v ∗ ) ∼ (us , us ), p∗ ∼ , h∗ ∼ l. (2.20a−d)
l
The variables in these scales are denoted by dropping the asterisks and the Stokes
equations and boundary conditions are non-dimensionalised:
∂u ∂v
https://doi.org/10.1017/jfm.2017.421

+ = 0, (2.21)
∂x ∂y
∂ u ∂ 2u ∂p
 2  
2
Ca + 2 − − L = 0,

∂x ∂y ∂x
2


(2.22)
∂ v ∂ 2v ∂p
 2 

Ca + 2 − = 0, 

∂x 2 ∂y ∂y

 Confinement effects in dip coating 7
u = 1, v=0 at y = 0, (2.23a,b)
" #
∂u ∂v ∂u ∂v
   2
dh dh 
2 +− −1 + = 0, 
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 22 Aug 2017 at 05:30:17, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.


dx ∂x ∂y ∂y ∂x
dx



"  2 −3/2 
# at y = h(x),
dh ∂u ∂v ∂v d2 h
 
dh 
p + Ca + − 2Ca = − 2 1 +


dx ∂y ∂x ∂y

dx dx 
(2.24)
dh
u =v at y = h(x). (2.25)
dx

2.5. Channel region

In a confined pool, the flow inside the parallel channel between the moving plate
and the stationary wall becomes significant especially during fast withdrawal. In
this channel region, the pressure gradient is mainly balanced by the viscous drag.
Therefore, the characteristic viscous stress is an appropriate scale for the pressure,
and the remaining variables have the same scales for the static meniscus region:
µus
(x∗ , y∗ ) ∼ (l, l), (u∗ , v ∗ ) ∼ (us , us ), p∗ ∼ . (2.26a−c)
l
The Stokes equations and boundary conditions are then non-dimensionalised as

∂u ∂v
+ = 0, (2.27)
∂x ∂y
∂ u ∂ 2u ∂pc
 2 
L2

+ − − = 0,

∂x2 ∂y2 ∂x

Ca 
(2.28)
∂ v ∂ 2v ∂pc
 2 

+ 2 − = 0, 

∂x 2 ∂y ∂y


u = 1, v=0 at y = 0, (2.29a,b)
u = 0, v=0 at y = 1. (2.30a,b)
The subscript c is added to p to distinguish it from that for the static meniscus region.
Similar to the dynamic meniscus region, the term associated with gravity appears in
the momentum balance equation (2.28):

L2
Gc ≡ . (2.31)
Ca
https://doi.org/10.1017/jfm.2017.421

3. Meniscus-controlled regime
For a low capillary number Ca, the dynamic meniscus region connects the film
entrainment and the static meniscus regions. In this low-Ca regime, the flow rate
through the film entrainment region is determined by the force balance in the
dynamic meniscus region. The meniscus in a confined pool is more curved than
that in an unconfined pool, and the increase in curvature changes the capillarity and
 8 O. Kim and J. Nam
hence affects the force balance in the dynamic meniscus region. Therefore, the film
thickness h∗∞ deviates from that of the unconfined dip coating case.
The radius of curvature of the meniscus is comparable to half the channel width l in
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 22 Aug 2017 at 05:30:17, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.

the static meniscus region. Following Krechetnikov & Homsy (2006), a simple scaling
analysis can be performed with l instead of the capillary length lc in the dynamic
meniscus region:
us σ /(l/2) h∗∞
µ ∗2 ∼ , ∼ l−1 , (3.1a,b)
h∞ ld ld2
where ld is the extent of the dynamic meniscus region. Rearranging these expressions
yields
T ∼ Ca2/3 . (3.2)
In what follows, the prefactor of Ca can be determined.

3.1. Simple model for meniscus-controlled regime

In the low-Ca regime (Ca1/3  1), the entrained film is extremely thin, and hence
ld ∼ (h∗∞ l)1/2 is small. Therefore, the meniscus remains static except for the dynamic
meniscus region.
In this regime, the gravitational drainage Gd is negligible, and the velocity field can
be determined by solving (2.14)–(2.17):
 2  3
y dh
u= − hy + 1. (3.3)
2 dx3
Knowing the velocity field, from the kinematic condition (2.18), the mass conservation
for the dynamic meniscus region can be obtained:
3
!
∂ h d3 h
+ h = 0. (3.4)
∂x 3 dx3

The term inside the parentheses of the preceding equation is the flow rate through
the dynamic meniscus region. This flow rate is constant with respect to x and should
be equal to that through the film entrainment region h∞ . This flow-rate-matching
condition yields a third-order ordinary differential equation of the meniscus profile in
the dynamic meniscus region:

d3 h h∞ − h
3
=3 3
. (3.5)
dx h
In addition, h in the limit to the film entrainment region is the film thickness h∞ :
h → h∞ as x → ∞. (3.6)
https://doi.org/10.1017/jfm.2017.421

It becomes the boundary condition for (3.5), and the equation can be transformed into
the universal form (Bretherton 1961). The second derivative of the meniscus profile h
in the limit to the static pool region can be obtained via numerical integration of the
transformed equation:

d2 h
→ 1.3376h∞ as x → −∞. (3.7)
dx2
 Confinement effects in dip coating 9
Under the same condition (Ca1/3  1), the viscous force is negligible in the
static meniscus region. Therefore, the Stokes equations and boundary conditions
(2.22)–(2.25) for that region can be greatly simplified as
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 22 Aug 2017 at 05:30:17, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.

"  2 #−3/2
d d2 h dh
2
1+ = L2 , (3.8)
dx dx dx

and the integration of the foregoing equation yields

"  2 #−3/2
d2 h dh
2
1+ = L2 x + C, (3.9)
dx dx

where C is an integration constant. Again, we consider only the part of the meniscus
along which d2 h/dx2 > 0. For an unconfined pool, C vanishes in our coordinate system.
However, for a confined pool, C has a finite value because the dimensionless curvature
of a static meniscus does not vanish except for θc = π/2, as shown in figure 2. C is
determined by the pressure profile along the meniscus. The liquid-side pressure p0 at
the centre point Oe of the static meniscus is selected as the pressure datum and is
equivalent to −κ0 , where κ0 is the curvature at point O.
e The liquid-side pressure at
any point A on the meniscus in the inset of figure 2 can be obtained by considering
either the hydrostatic pressure at A or the capillary pressure jump across the curved
meniscus at that point. Both should be identical, and hence C = κ0 . Equation (3.9)
then becomes "  2 #−3/2
d2 h dh
2
1+ = L2 x + κ0 . (3.10)
dx dx
Following the idea of Landau & Levich (1942), we also use the classical curvature-
matching condition to determine the dimensionless film thickness T, i.e. the curvature
of the static meniscus at the ‘upper’ boundary is matched to that of the dynamic
meniscus at the ‘lower’ boundary. For the static meniscus, the slope of the meniscus
vanishes at the matching point:
dh
→0 as h → h∞ . (3.11)
dx
This gives the curvature at that point:
static
d2 h  p
2
= L 2 + (κ0 /L)2 . (3.12)
dx h→h∞


For the dynamic meniscus, equation (3.7) can be rewritten using the variables for the
static meniscus region:
dynamic
d2 h  Ca2/3
https://doi.org/10.1017/jfm.2017.421

= 1.3376 . (3.13)
dx2  h→∞ T
Finally, from the matching between the curvatures (3.12) and (3.13), T is given by
1 0.9458 2/3
T=p Ca . (3.14)
1 + (κ0 /L)2 /2 L
 10 O. Kim and J. Nam
We define the term involving the curvature κ0 in the preceding equation as
1
Cc ≡ p . (3.15)
1 + (κ0 /L)2 /2
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 22 Aug 2017 at 05:30:17, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.

Using Cc , the dimensional film thickness h∗∞ is expressed as

h∗∞ = Cc lc (0.9458Ca2/3 ) or h∗∞ = Cc lf (0.9458Ca1/6 ). (3.16a,b)

The expressions inside the parentheses are identical to (2.1) with the corresponding
characteristic lengths. This identity implies that the previous result of Landau &
Levich (1942) can be corrected by introducing Cc for confined dip coating. Therefore,
we call Cc the confinement-correction factor for the meniscus-controlled regime.
The value of κ0 is required to determine Cc and the complete equation (3.14).
Equation (3.10) can be integrated and rearranged as

dh L2 x2 /2 + κ0 x − 1
=p . (3.17)
dx 1 − (L2 x2 /2 + κ0 x − 1)2

The integration of (3.17) from O

e to S (see figure 2) yields the implicit equation,

2 − k2 h  π  i 2 h π  i L
F , k − F(φ, k) − E , k − E(φ, k) − = 0, (3.18)
k 2 k 2 2
where
2L π + 2θc
k= p (elliptic modulus), φ= , (3.19a,b)
κ02 + 4L2 4

and F and E are the elliptic integrals of the first and second kind, respectively. For a
given L, k can be evaluated from (3.18), and then we can obtain Cc directly instead
of κ0 .
Figure 3 shows the confinement-correction factor Cc as a function of the
dimensionless channel width L for different contact angles θc , and Cc exhibits small-
and large-L limiting behaviours clearly. In the small-L limit, a more confined pool
(smaller L) yields a meniscus with a smaller Cc at a given θc . In the large-L limit, Cc
approaches unity, which implies that confinement effects disappear, i.e. equation (2.1)
is recovered from (3.16). A detailed examination of these two limits can be found in
appendix C.
Interestingly, the power-law dependence of Cc on L is independent of θc at small
L, and hence κ0 depends only on θc . It is related to the influence of gravity on the
meniscus shape and is discussed in § 6.
It
√is particularly interesting to analyse (3.14) in the small-L limit. In this limit, Cc →
L/ 2 cos θc , as explained in § C.1. The resulting equation becomes independent of L
and exhibits the power-law behaviour O(Ca2/3 ):
https://doi.org/10.1017/jfm.2017.421

0.9458
T=√ Ca2/3 . (3.20)
2 cos θc
Therefore, the prefactor of (3.2) depends only on θc in the small-L limit. When the
coating liquid wets the wall almost completely (small θc ), the meniscus is distorted
 Confinement effects in dip coating 11

101
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 22 Aug 2017 at 05:30:17, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.

100

10–1

10–2

10–3 10–2 10–1 100 101

L

F IGURE 3. (Colour online) Confinement-correction factor Cc as a function of dimen-

sionless channel width L. The values are computed from (3.18). The arrow denotes the
direction of increasing θc . The value of Cc decreases as the liquid wets the container wall,
i.e. for lower θc .

further, and a thinner film is entrained. Similar results can be found in the inclined
unconfined dip coating problem (Wilson 1982); the thinner film is located on the side
where the meniscus shape is concave.
The matching condition of Landau & Levich (1942) that yields (3.14) appears to
be irrational. However, matched asymptotic expansions show that this condition is
valid for Ca1/3  1 (Wilson 1982; Park & Homsy 1984; Münch 2002). Equation
(3.14) is equivalent to the leading-order approximation of the asymptotic expansion in
Ca1/3 . Hence, the error behaviour of the equation can be predicted by investigating the
previous study of Wilson (1982), in which unconfined dip coating is discussed. In his
study, when the inclination angle α = 0 (vertical withdrawal), C2 in (30) corresponds
to our C in (3.9) and (3.14) becomes the leading-order term of his equation (39).
In terms of our characteristic length, the first correction is O(Ca) (note that the first
correction is O(Ca1/2 ) in his study because of the different characteristic length from
the present study). Therefore, one might expect the error behaviour of (3.14) to be
O(Ca) for low Ca. However, Wilson (1982) also discussed the non-uniformity in his
asymptotic series that arises from α. We, therefore, can predict the non-uniformity
https://doi.org/10.1017/jfm.2017.421

arising from the contact angle θc , which changes the meniscus profile at the given Ca
and L, in confined dip coating. The error behaviour is briefly discussed in § 5.3.
The approach that leads to (3.14) can also be easily extended to other film-thickness-
predicting equations on the basis of the curvature matching for unconfined dip coating
such as that of White & Tallmadge (1965) and Spiers et al. (1974). See appendix B
for details.
 12 O. Kim and J. Nam
4. Channel-controlled regime
At high Ca, it is expected that the viscous force transferred from the moving plate
is sufficiently large to set the contact line on the stationary wall in motion (i.e. the
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 22 Aug 2017 at 05:30:17, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.

contact line may move upwards along the wall) as with the forced wetting (Blake
& Ruschak 1979); there is no steady-state flow as long as the contact line moves.
However, the contact line would eventually cease to move and be pinned on a wall
that has received special treatments such as an application of surface roughening or at
the sharp corner of a wall that the moving contact line reaches (Oliver, Huh & Mason
1977), as shown in figure 1(b). In what follows, we consider a steady-state flow with
the pinned contact line.

4.1. Limiting behaviour of film thickness in channel-controlled regime

The flow in the channel region is fully developed, as mentioned in § 2, and hence the
Stokes equations with boundary conditions (2.27)–(2.30) can be greatly simplified to
give a velocity field
1 L2
 
dpc
u=1+ + (y2 − y) − y. (4.1)
2 Ca dx
We can obtain the flow rate in units of us l through the channel from the velocity field:
1 1 d
qc = − (pc + Gc x). (4.2)
2 12 dx
The flow rate qc should be equal to that through the film entrainment region. It has
been given by Van Rossum (1958) as
qf = T 1 − 13 Gc T 2 .


(4.3)
The term involving Gc represents the gravitational drainage in the film entrainment
region. The drainage is negligible at high Ca  L2 because Gc T 2 < L2 /Ca in confined
dip coating, where the film thickness is expected to be smaller than the channel width;
i.e. T < 1. Therefore, the dimensionless thickness T is given by
1 1 dpc
T= − . (4.4)
2 12 dx
This result means that T can be determined by considering only the pressure gradient
inside the channel.
Besides, the flow inside the channel at high Ca resembles the coating flow under
a knife or blade. The well-developed theory for knife coatings shows that the film
thickness is half the coating gap for a blade parallel to a moving plate under negligible
surface tension (Ruschak 1985; Coyle 1997). Here, the gap is the distance between the
moving plate and the blade face, which is essentially identical to the channel width l
in confined dip coating. We can therefore expect that dpc /dx will eventually vanish in
the high-Ca limit, from knife coating theory. As a result, the flow becomes an almost
pure Couette flow, and T tends to 1/2 when Ca approaches infinity:
https://doi.org/10.1017/jfm.2017.421

1
T→ 2
as Ca → ∞. (4.5)
This result appears to be similar to that of unconfined dip coating: h∗∞ /lf becomes
independent of Ca in the high-Ca limit in the absence of inertia and surfactants
(Esmail & Hummel 1975b; Weinstein & Ruschak 2001; Jin et al. 2005). However,
the film thickness h∗∞ itself increases without bound with Ca in unconfined dip
coating, whereas h∗∞ is bounded in confined dip coating.
 Confinement effects in dip coating 13
5. Numerical computations
5.1. Finite element model
To solve the Stokes equations with the Galerkin finite element method (G/FEM), we
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 22 Aug 2017 at 05:30:17, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.

need more boundary conditions, in addition to conditions (2.6)–(2.9). These boundary

conditions are represented in the following.
A stationary wall with the no-slip condition should be included in the analysis to
incorporate the confinement effects. Such a wall can be found in early numerical
models for unconfined dip coating (Jenny & Souhar 2009; Filali et al. 2013); however,
there is also an inlet feed port to compensate for the mass loss at the outflow boundary
in those models. The reliability of a numerical model with such an inlet feed would
depend on the location of the feed; the flow through a feed that is not far enough
from the meniscus affects the film thickness when a pool is confined. Besides, the
numerical models with sufficient distance between the meniscus and the feed may be
more computationally expensive than our FE model explained in the following.
The meniscus contacts the stationary wall at an angle θ on the contact line:

nm · nw = − cos θ, (5.1)

where nm and nw are the outward normal unit vectors on the meniscus and the wall,
respectively. The definitions of the normal and tangent vectors and a contact angle in
our FE model are shown in figure 4. When the coating liquid near the stationary wall
remains in static equilibrium, the static contact angle θc is used instead of θ.
Dip coating is a self-metered method in that the flow rate is determined by the force
balance inside the coating flow (Weinstein & Ruschak 2001). Therefore, a velocity
profile should not be prescribed at the inflow or outflow boundary a priori; such a
prescribed velocity profile presets the film thickness h∗∞ .
In this regard, the FEM has an advantage over other methods, because of the
free boundary condition (Papanastasiou, Malamataris & Ellwood 1992). The FEM
can extend the validity of the weak form of the governing equations to synthetic
boundaries (inflow and outflow boundaries), instead of replacing them with unknown
essential or natural boundary conditions, such that

nb · T ∗ = −nb p∗ + nb · µ[∇ ∗ v ∗ + (∇ ∗ v ∗ )T ], (5.2)

where nb is the outward unit normal vector at either the inflow or outflow boundary.
The free boundary condition (5.2), which is typically used for an outflow boundary,
appears not to impose any boundary condition at all; however, it provides an effective
boundary condition (Renardy 1997).
We apply the free boundary condition (5.2) not only to the outflow boundary but
also to the inflow boundary. To the best of the authors’ knowledge, this has never
been attempted before. In our FE model, both the inflow and outflow boundaries are
located sufficiently far from each other so that the velocity fields and the gradient
of the pressure fields are expected to approach asymptotic limits at both ends. This
asymptotic state includes ‘fully developed’ flows (Renardy 1997). Applying the free
https://doi.org/10.1017/jfm.2017.421

boundary condition to the inflow boundary allows the pressure to take any value
at the boundary. It yields the undetermined position of the contact line with the
contact angle prescribed by (5.1). However, at a given Ca and L, both the position
of the contact line and the contact angle should be fixed to have a unique solution
in the frame of reference for our model, but it is impossible to impose them as
boundary conditions in our FE framework simultaneously. Therefore, an additional
 14 O. Kim and J. Nam
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 22 Aug 2017 at 05:30:17, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.

F IGURE 4. Numerical flow domain and corresponding boundary conditions. All boundary
conditions, unit vectors and a contact angle are represented. The free boundary condition
is applied to both the inflow (bottom) and outflow (top) boundaries. The velocity profile
is not prescribed.

condition or constraint is required to find the unique solution. It would seem that
the mesh configuration of the physical domain of interest provides the additional
condition required to determine the position of the contact line implicitly. When the
mesh generation equation (explained in the following) is solved with (2.4) and (2.5)
simultaneously, we can find the unique (steady-state) solution. However, the detailed
mechanism of how the mesh configuration fixes the position of the contact line is
not clear.
As discussed earlier, the physical domain of interest of the confined dip coating flow
should be surrounded by three physical boundaries (the meniscus, the stationary wall
and the moving plate) and two synthetic boundaries (the inflow and outflow planes).
All the boundary conditions are summarised in figure 4.
To find the meniscus shape via standard techniques for boundary value problems,
the set of differential equations and boundary conditions on the unknown, deformable
‘physical’ region R in the x∗ y∗ -plane should be transformed one to one into an
equivalent set on the known, fixed ‘computational’ region G in the ξ η-plane. Here
R and the set on it are transformed by an equation of the form x∗ = M(ξ ), where
https://doi.org/10.1017/jfm.2017.421

x∗ = (x∗ , y∗ ) and ξ = (ξ , η). Any function f (x∗ ) defined on R can be thought of

as a function f (x∗ (ξ )) defined on G as well. The image ξ of x∗ under the inverse
transformation M−1 is a solution to elliptic partial differential equations, which is
similar to that for the diffusional mass transport for the dilute regime,

∇ ∗ · (Dξ ∇ ∗ ξ ) = 0, ∇ ∗ · (Dη ∇ ∗ η) = 0, (5.3a,b)

 Confinement effects in dip coating 15
where Dξ and Dη are mesh diffusivities, which determine the gradients of the
coordinate potentials ξ and η adjusting the nodal distribution inside the ‘physical’
region R. The nodal distribution at the boundaries, prescribed by the functions
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 22 Aug 2017 at 05:30:17, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.

defining their geometries, is specified by a stretching function (Vinokur 1983). The

iso-contour curves of ξ and η define the boundaries of the quadrilateral elements
used to tessellate R, and the intersection points of these curves determine the nodal
positions. The meniscus shape is implicitly determined by the kinematic condition
(2.8). Equations (5.3) have been called the elliptic mesh generation equations, and
they have been discussed in detail by de Santos (1991).

5.2. Solution using the finite element model

The field variables, such as the velocity, pressure and nodal position, are approximated
by a linear combination of the polynomial basis functions. The velocity and nodal
position are approximated by a Lagrangian biquadratic function φi (ξ , η), and the
pressure is approximated by a linear discontinuous function ψk (ξ , η):
n
X n
X m
X
∗
x = x∗i φi (ξ , η), v =
∗
v ∗i φi (ξ , η), ∗
p = p∗i ψi (ξ , η), (5.4a−c)
i=1 i=1 i=1

where n is the number of nodes used for the nodal position and velocity and m is the
number of nodes used for the pressure, and the coefficients x∗i , v ∗i and p∗i are unknown
values to be determined at each node.
The weak forms of (2.4), (2.5) and (5.3) are obtained by multiplying them by the
weighting functions and then integrating over the ‘physical’ region R. The essential
boundary conditions are implemented by replacing the corresponding weighted
residual equation with the desired velocity or node specification. The natural boundary
conditions are implemented through the boundary integrals, which follow from the
divergence theorem. Note that the free boundary condition (5.2) is treated as the
natural boundary condition.
The G/FEM reduces the Stokes and mesh generation equations (2.4), (2.5) and (5.3)
to a set of nonlinear algebraic equations with respect to the basis function coefficients,

R(z, λ) = 0, (5.5)
where z is the solution vector consisting of the FE coefficients of x∗i , v ∗i and p∗i ; and
λ is the parameter vector containing the parameters for the dip coating flow, such as
l, us , θc , ρ and σ . The system of nonlinear equations (5.5) is solved using Newton’s
method:
J R (zi |λ)δzi = −R(zi , λ),

(5.6)
zi+1 = zi + δzi ,
where J R (z) is the Jacobian matrix of R. The solution vector z is accepted as the
solution to the problem when kR(zi )k2 < 5 × 10−9 . The typical value of the solution
https://doi.org/10.1017/jfm.2017.421

update kδzi k2 is of the order of 10−5 at convergence.

We use a direct numerical solver based on the frontal algorithm (Duff, Erisman &
Reid 1989). It appears that there is no zero pivot during the matrix decomposition
step; thus, we can safely conclude that the free boundary condition (5.2) at both the
inflow and outflow boundaries does not yield an ill-posed problem in the resulting
discrete-level matrix system (5.6).
 16 O. Kim and J. Nam
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 22 Aug 2017 at 05:30:17, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.

F IGURE 5. Mesh configuration at Ca = 0.01, L = 5 and θc = 30◦ . 2775 quadrilateral

elements are tessellated into the physical domain. A total of 125 elements and 251 nodes
is used to resolve the curved meniscus shape. The aspect ratio of the mesh configuration
is adjusted for visualisation (d = 6 l).

The initial guess was obtained by replacing the unknown meniscus shape with a
known arc of a circle and substituting the stress balance condition (2.8) with the
kinematic condition (2.9) and Navier’s slip condition with a large slip coefficient β =
108 along the meniscus, β(tm nm : T ∗ ) = tm · v ∗ . Nodal positions were not included in
the unknowns because the normal and tangential unit vectors at the meniscus were
known at this stage; Navier’s slip condition and the kinematic condition (2.9) were
applied to the x∗ - and y∗ -components of the momentum equation (2.5), respectively.
The base solution at Ca = 0.01, L = 5 and θc = 30◦ was computed using the initial
guess. Figure 5 illustrates the mesh configuration of the base solution. Along the
meniscus, a total of 125 elements and 251 nodes were used to resolve the meniscus
shape.
The solutions for different values of the capillary number Ca and the dimensionless
channel width L were computed by using the pseudo-arc-length continuation (Bolstad
& Keller 1986), starting from the base solution. us and l were chosen as the
continuation parameters to adjust Ca and L, respectively. During the continuation
in l, the distance d between the inflow and outflow boundaries should be adjusted
simultaneously to ensure a fully developed flow at the inflow and outflow boundaries.
When the inflow is not fully developed, the accuracy of the flow rate through the
https://doi.org/10.1017/jfm.2017.421

flow domain could be compromised, and hence the film thickness h∗∞ . Unlike other
numerical models (Jin et al. 2005; Campana et al. 2013), we did not restrict the
shape of the free surface at the corner of the element where the outflow boundary
and the free surface are placed. Hence, the free surface slope with respect to the
moving plate is a major indicator of fully developed flow. When the outflow is not
fully developed, the thickness being measured from the moving plate to the free
 Confinement effects in dip coating 17

100

Equation (4.5)
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 22 Aug 2017 at 05:30:17, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.

Equation (3.14)

10–1

10–2

10–3
10–2 10–1 100 101 102
Ca

F IGURE 6. (Colour online) Dimensionless film thickness T versus capillary number Ca

for θc = 30◦ and various L. The curves of T overlap at L < 0.5.

surface along the outflow boundary is larger than the film thickness h∗∞ . According
to our numerical tests, it turns out that d/l = 6 ensures a virtually parabolic velocity
profile and a zero slope of the free surface at the corner at the outflow boundary. We
therefore maintained d/l = 6 during the continuation.
Steady-state solutions are obtained within 3–5 iterations except for the base solution
obtained using the initial guess. To guarantee a high accuracy for h∗∞ , we verified that
the mesh configuration near the highly curved meniscus is properly resolved for every
solution.

5.3. Numerical results

We validated our FE model by comparison with results from other models and
experiments on unconfined dip coating. See appendix D for details.
The results of the FE computations confirm that (3.20) and (4.5) represent the
meniscus- and channel-controlled regimes, respectively. The curves computed at the
dimensionless channel width L < 0.5 overlap and are virtually indistinguishable,
as shown in figure 6. The overlapped curve exhibits power-law dependence and
https://doi.org/10.1017/jfm.2017.421

is independent of L in the low-Ca regime. This dependence is already predicted

by (3.20).
Figure 7 shows the difference between the analytical and numerical results.
Interestingly, the error behaviours are distinct at different contact angles. The
error follows a power-law behaviour  = O(Cam ), where m is unity for θc = 30◦ ,
but changes to m ' 0.75 for θc = 60◦ . This distinction appears to be due to the
 18 O. Kim and J. Nam

10–1
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 22 Aug 2017 at 05:30:17, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.

10–2

10–3 –2
10 10–1
Ca

F IGURE 7. (Colour online) Error  behaviour. The error is defined as the difference
between the analytical and numerical results. The dimensionless film thickness T is
predicted by (3.14) and (3.20) for L = 1 and L = 0.1, respectively. The plot shows the
errors of O(Ca) at θc = 30◦ and O(Ca4/3 ) at θc = 0◦ , as indicated by Wilson (1982)
and Park & Homsy (1984), respectively.

non-uniformity mentioned in § 3.1. A full asymptotic analysis is required to investigate

the non-uniformity; however, it is beyond the scope of the present study.
Figure 8 shows the effect of the contact angle θc on the dimensionless thickness
T. T increases with θc as can be seen in (3.20). However, the dependence of T on
θc decreases as Ca increases, as shown in figure 8(a). This is because the meniscus
shape becomes less sensitive to θc for high Ca, as can be seen from a comparison of
figures 8(b) (Ca = 0.01) and (c) (Ca = 5). The results clearly show that the capillary
effect is crucial in the meniscus-controlled regime discussed in § 3, and the pressure
gradient inside the channel region is crucial in the channel-controlled regime discussed
in § 4.
In the high-Ca regime, the pressure gradient dpc /dx at the inflow boundary and the
gravity Gc monotonically decrease as Ca increases, but Gc decreases faster than dpc /dx
(Gc = o(dpc /dx)), as shown in figure 9(a). Note that dpc /dx does not rapidly vanish
as Ca increases, as shown in figure 9(b). However, we can expect that dpc /dx will
https://doi.org/10.1017/jfm.2017.421

vanish eventually in the high-Ca limit, and the flow will become a Couette flow (see
appendix A for details). The numerical results are consistent with the discussion in
§ 4.1. Interestingly, dpc /dx is independent of L at small L (<0.5) for high Ca, as
shown in figure 9(b), and hence T is also independent of L. This independence is
identical to that found in the low-Ca regime with small L. Finally, all the numerical
results show that l is the critical parameter for both regimes.
 Confinement effects in dip coating 19

(a) 1.0

0.8
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 22 Aug 2017 at 05:30:17, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.

0.6

0.4

0.2

0 30 60 90

(b) 2 (c) 2

1 1

0 0

–1 –1
0 0.5 1.0 0 0.5 1.0
y y

F IGURE 8. (Colour online) Influence of container wall wetting on film thickness and
meniscus shape for L = 0.1 and various Ca. (a) Plot of RT ≡ (T − T0 )/T0 , where T0 is
the value at θc = 0. The effect of the wall wetting on the film thickness diminishes as
Ca increases. (b,c) The coordinates are scaled with l. The arrows denote the direction of
increasing θc . The meniscus shape changes significantly for Ca = 0.01, whereas it changes
slightly for Ca = 5.

6. Characteristic length transition: vanishing gravitational effect

In confined dip coating, the dimensionless thickness T becomes independent of
the dimensionless channel width L for two limit cases: low Ca (meniscus-controlled
regime with small L, § 3), and high Ca  L2 (channel-controlled regime, § 4). For
both limit cases, the gravitational force is overwhelmed by either the capillary or the
viscous force. A comparison of gravity with the surface tension or viscous drag can
https://doi.org/10.1017/jfm.2017.421

aid understanding of these limits.

In the meniscus-controlled regime, the local relative importance of the gravitational
and capillary forces in the dynamic meniscus region can be measured by ρgh∗ /(σ /l).
However, we utilise a more relevant quantity, defined in (2.19):

Gd ≡ Ca1/3 L2 . (6.1)
 20 O. Kim and J. Nam

(a) 10 3

10 2
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 22 Aug 2017 at 05:30:17, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.

101

100

10–1

10–2

10–3
100 101 102

(b) 101

100

10–1

10–2

10–3 0
10 101 102
Ca

F IGURE 9. (Colour online) (a) The ratio of gravitational force Gc to pressure force dpc /dx
and (b) dpc /dx as a function of capillary number Ca for different L at θc = 30◦ . (a) Gc
decays faster than dpc /dx as Ca increases. (b) When the pool is sufficiently confined (L <
0.5), dpc /dx becomes independent of L. Although dpc /dx does not vanish completely, even
at an extremely high Ca of up to 102 , it clearly decreases monotonically.

Note that Gd is an upper bound of ρgh∗ /(σ /l). Gd vanishes when the channel is
confined further (L < 0.5), which reinforces the capillary force due to the highly
https://doi.org/10.1017/jfm.2017.421

curved meniscus. For small L, gravity is negligible in the Stokes equations (2.15)
and (2.22), and T becomes independent of L as shown in (3.20). This means that
the effect of gravity diminishes not only in the dynamic meniscus region but also in
the static meniscus region. It is already signalled in κ0 . For low L, κ0 depends only
on θc as mentioned in § 3.1 and becomes κ0 = 2 cos θc ; therefore, the meniscus shape
becomes an arc of a circle in the low-Ca limit.
 Confinement effects in dip coating 21
The resulting Stokes equations (without gravity) can also be found in the Bretherton
problem in Hele-Shaw cells, where gravity is also insignificant (Park & Homsy 1984).
Furthermore, equation (3.20) with a zero contact angle becomes the leading-order
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 22 Aug 2017 at 05:30:17, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.

approximation for the problem. Note that this contact angle stands for a two-sided
coating on the walls of the cells. The Bretherton problem in Hele-Shaw cells,
therefore, can be regarded as the limiting case of the confined dip coating problem,
and the error behaviour is consistent with O(Ca4/3 ) as indicated by Park & Homsy
(1984), as shown in figure 7. See appendix E for a comparison between confined dip
coating and the Bretherton problems.
In the channel-controlled regime, similar to the meniscus-controlled regime, the
competition between the gravitational and viscous forces can be measured by

L2 ρgl
Gc ≡ = . (6.2)
Ca µus /l
Gc vanishes when the plate withdraws rapidly or when the channel is highly confined
(Ca  L2 ), which enhances the viscous force inside the channel. In the high-Ca limit,
T becomes independent of L. This corresponds to the vanishing of the gravitational
effect.
The independence of T from L for both low- and high-Ca regimes reveals a change
in the number of dimensionless numbers required to describe the confined dip coating
flow uniquely. As discussed in § 2.1, for withdrawal of a flat substrate from a pool
of a pure Newtonian fluid, unconfined dip coating flows require three dimensionless
numbers (or two for inertialess flows: h∗∞ /lc and Ca). Adding the channel width l
to the dimensional analysis increases the number of dimensionless numbers to four
(or three for inertialess flows: T, Ca and L). However, when the gravitational effect
vanishes, the number becomes three again (or two for inertialess flows: T and Ca).
Note that the definitions of T and Ca are independent of g. According to the FE
computations, these vanishing gravitational effects occur over the entire range of Ca
when L is sufficiently small; e.g. L < 0.5 according to our numerical results, as shown
in figure 6.
The change in the number of dimensionless numbers for the system implies a
transition of the system’s characteristic length. The length of the unconfined pool is
the capillary length lc . When the pool is confined, the newly introduced l competes
with lc , and this competition is parameterised by L = l/lc . When the gravitational
effect vanishes, lc is no longer effective, and l becomes the only characteristic length
for confined dip coating flows.

7. Concluding remarks
The unconfined pool assumption is valid only in the absence of a stationary wall
influencing the film thickness h∗∞ . In such a pool, h∗∞ should be independent of the
channel width l; i.e. T is inversely proportional to L, regardless of Ca. This criterion
for an unconfined pool can be expressed as
https://doi.org/10.1017/jfm.2017.421

∂(LT)
= 0. (7.1)
∂L
Otherwise, a pool is confined. It should be emphasised that the pool can be confined
either via small L or via high Ca, and hence it is possible for an unconfined pool
to become confined as Ca increases. In a confined pool, l becomes the appropriate
 22 O. Kim and J. Nam
characteristic length of the system, and two major changes occur: an increase in
the curvature of the meniscus and an intensification in the viscous stress inside
the channel. The increased curvature enhances the capillarity and affects the force
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 22 Aug 2017 at 05:30:17, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.

balance in the dynamic meniscus region. The intensified viscous stress dominates the
(dimensional) pressure gradient in the channel region. These changes affect the flow
rate and hence the film thickness h∗∞ , and we call them the confinement effects in
dip coating.
In confined dip coating, the role of gravity in the force balance diminishes when
Ca1/3  1 and L < 0.5; or Ca  L2 . Consequently, the characteristic length is no longer
determined by the material properties but by the channel width l. The confinement
effects for a dip coating flow are strongly associated with the vanishing gravitational
effect and the increasing importance of the newly introduced characteristic length l.
The dynamic and channel regions are important as they determine the flow rate and
hence the film thickness at low and high Ca, respectively. The thickness-determining
mechanisms in each region correspond to the two regimes of confined dip coating
flows, i.e. the meniscus- and channel-controlled regimes. The criteria for both regimes
cannot be explicitly determined because the threshold Ca of both regimes depends
on the dimensionless channel width L, as shown in figure 6. However, when the
gravitational effect vanishes (L < 0.5), the ranges of Ca . 0.1 and Ca & 10 can be
established for the meniscus- and channel-controlled regimes, respectively.
Between the two regimes, there is a regime where the normal viscous stress
becomes comparable with the capillary pressure in the dynamic meniscus region.
This regime was also observed in unconfined dip coating (Esmail & Hummel 1975a;
Jin et al. 2005). The meniscus shape changes abruptly in the dynamic meniscus
region for low Ca, but the magnitude of the change diminishes as Ca increases, as
shown in figure 10(a). For high Ca > 1, the shape does not change significantly, and
hence the capillary pressure jump. In contrast, the normal viscous stress increases
monotonically and dominates the capillary pressure, as shown in figure 10(b). In this
regime, the flow is no longer unidirectional, and the strong acceleration along the
meniscus generates a non-negligible normal viscous stress in the dynamic meniscus
region. Under these conditions, lubrication-type simple models cannot be used, and
a nonlinear two-dimensional model must be employed to determine the meniscus
shape (Esmail & Hummel 1975b). Analysis of this regime is beyond the scope of
the present study but merits further discussion.
Inertia was not considered in the present study. The importance of inertia depends
on the material number m. The inertia begins to surpass the surface tension at a Weber
number (We = ReCa) close to unity and becomes comparable to the viscous drag
(Campana et al. 2013). For unconfined dip coating, thickening or thinning of the film
occurs above a threshold value of Ca at large Re and low Ca (Tallmadge & Stella
1968; de Ryck & Quéré 1998) or at large Re and high Ca (Esmail & Hummel 1975b),
respectively. In both cases, the maximum thickness scaled with lf exists. Besides, the
meniscus shape in the dynamic meniscus region becomes wavy, and the coating flow
becomes unstable (Esmail & Hummel 1975b; Kamotani, Ostrach & Kizito 1999; Jin
et al. 2005). On the basis of these results, it is expected that thickening or thinning
of the film can also occur in confined dip coating with inertia, depending on Re
https://doi.org/10.1017/jfm.2017.421

(or m) and Ca. However, the maximum thickness occurs at a lower Ca than that in
unconfined dip coating, because the narrow channel causes further acceleration of the
flow near the meniscus at the same Ca. Furthermore, the waviness of the meniscus
shape and hence the instability of the meniscus are also expected to be suppressed,
because of the presence of the stationary container wall. This regime remains to be
analysed.
 Confinement effects in dip coating 23

(a) 5

4
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 22 Aug 2017 at 05:30:17, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.

x 2

–1
0 0.2 0.4 0.6 0.8 1.0
y

(b) 5

0
10–2 10–1 100 101
Ca

F IGURE 10. (Colour online) Meniscus shape change and maximum normal viscous stress
for L = 1 and θc = 30◦ . (a) The coordinates are in units of l. The arrow denotes the
direction of increasing Ca. The meniscus shape changes abruptly in the dynamic meniscus
region for low Ca, but this change diminishes with increasing Ca. (b) τnn m
is the maximum
magnitude of the normal viscous stress on the meniscus. τnn is expressed in units of
m

capillary pressure based on the meniscus shape, which is an arc of a circle with a contact
angle of 30◦ at the wall. τnnm
increases monotonically with Ca. It should be mentioned
that the numerical results obtained with quadratic basis functions for the velocity and
position variables yield wiggles in the profile of the normal viscous stress on the meniscus.
Therefore, the result for the magnitude of τnn m
may be less accurate than those for other
primitive variables, such as the position and velocity.

It is important to mention the non-uniqueness of the film thickness. Wilson (1982)

discussed the possibility of multiple uniform film thicknesses in the non-vertical dip
coating problem where the inclination angle (from the vertical) is larger than a critical
https://doi.org/10.1017/jfm.2017.421

value. Hocking (2001) explained that this non-uniqueness arises, at a low capillary
number Ca, from the presence of the three-phase contact line. Münch & Evans
(2005) and Jin et al. (2005) investigated the non-uniqueness of the film thickness
and the meniscus shape systematically by studying the phase space of an ordinary
differential equation for the shape. Furthermore, Münch & Evans (2006) examined
the interaction between the meniscus and the advancing front in the context of
 24 O. Kim and J. Nam
Marangoni-driven films. In the dip coating problem, multiple solutions were observed
experimentally (Snoeijer et al. 2006, 2008), calculated asymptotically (Benilov et al.
2010) and computed numerically (Gao et al. 2016).
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 22 Aug 2017 at 05:30:17, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.

The present study is limited to complete wetting of the moving plate by the
coating liquid, so that the three-phase contact line at the front of the entrained film is
meaningless from the standpoint of thickness determination. In confined dip coating,
the film thickness h∗∞ for complete wetting is uniquely determined and is thinner
than that in unconfined dip coating. However, for partial wetting, the deposited film
exhibits complex behaviour; a thick film in comparison with the film for complete
wetting can appear at the front of the entrained film in unconfined dip coating.
Snoeijer et al. (2006) and Gao et al. (2016) showed experimental evidence and
numerical predictions regarding the appearance of the thick film, respectively. They
also reported that the thickness of the thick film is determined solely by the physics
of the three-phase contact line; therefore, we speculate that the thick film would not
be significantly affected in confined dip coating.
In the present study, there is no restriction on the liquid supply to the channel.
Therefore, the meniscus reaches a steady state and fixes to the wall. Note that, if the
liquid supply is limited or is unable to keep pace with the removal by the entrained
film, the meniscus cannot be fixed to the wall. In the meniscus-controlled regime,
the meniscus recedes with constant velocity, which is a function of the size of the
container and the flow rate. In the channel-controlled regime, the solution cannot reach
a steady state; the pinned meniscus eventually invades the channel along the wall. It
is very similar to the low-flow limit phenomena induced by a limited supply of liquid
in slot coating flows (Carvalho & Kheshgi 2006). They are signalled by an invasion
of the channel between the coating die and the moving plate by the downstream
meniscus.

Acknowledgements
This research was supported by the Basic Science Research Program (NRF-
2016R1D1A1B03931632) and by the Global Frontier R&D Program of the Center for
Multiscale Energy Systems (2012M3A6A7054861) funded by the National Research
Foundation of Korea (NRF) under the Ministry of Science, ICT and Future Planning,
Korea.

Appendix A. Fully developed flow inside channel

In dip coating, when the plate is dragged from the container, the liquid near the
plate is entrained as a result of liquid viscosity. The thickness of this boundary layer
δ is approximately
µD 1/2
 
δ∼ , (A 1)
ρus
where D is the distance from the starting point of the boundary layer, as discussed
https://doi.org/10.1017/jfm.2017.421

by de Ryck & Quéré (1996). When D is sufficiently long, the flow becomes fully
developed under the meniscus. This means that δ approaches and even surpasses l;
i.e. the viscous stress from the moving plate reaches the container wall. Therefore,
for confined dip coating, a small channel width is sufficient to ensure a fully
developed flow. Within this flow, the viscous force is balanced by the pressure
and the gravitational force.
 Confinement effects in dip coating 25

1.0

0.9
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 22 Aug 2017 at 05:30:17, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.

0.8

0.7

0.6

0.5

0.4

0.3

0.2
FEM results
0.1 Jin et al. (2005)
Lee & Tallmadge (1974)
0
10–4 10–3 10–2 10–1 100 101 102
Ca

F IGURE 11. (Colour online) Comparison of FE computation results with previously

reported values. For the FE model, the liquid pool is sufficiently large (L = 70) that
the film thickness is unaffected by L. The solid line represents the prediction given
by Wilson (1982). The numerical results are in reasonable agreement with the reported
experimental observations (Lee & Tallmadge 1974) in the intermediate Ca range and agree
well with the previous numerical results (Jin et al. 2005) for all examined Ca ranges. This
agreement supports the validity of the free boundary condition (5.2) at both the inflow and
outflow boundaries.

The value of dpc /dx at the inflow boundary can be evaluated from the velocity
profile at that boundary. When the flow is fully developed, the velocity profile is given
by (4.1):
1 L2
 
dpc
u=1+ + (y2 − y) − y. (A 2)
2 Ca dx
Note that the velocity profiles at the inflow boundary from the FE computations are
virtually parabolic and are almost perfectly fitted to (4.1), as shown in figure 12.
This strongly suggests that all the computed flows at the inflow boundary are fully
developed. The pressure gradient is directly obtained from the fitted equation.

Appendix B. White–Tallmadge model for confined dip coating in meniscus-

https://doi.org/10.1017/jfm.2017.421

controlled regime
Following White & Tallmadge (1965), equation (3.7) can be augmented by the
gravitational drainage:
dynamic
d2 h  (Ca − T 2 L2 )2/3
= 1.3376 . (B 1)
dx2 h→∞ T
 26 O. Kim and J. Nam

1.0
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 22 Aug 2017 at 05:30:17, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.

0.8

Couette flow
0.6

0.4

u
0.2

–0.2

–0.4
0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0
y

F IGURE 12. (Colour online) Velocity profiles at inflow boundary for different Ca under
L = 1 and θc = 30◦ . The solid lines (excluding the line indicated by the arrow) are quadratic
functions following equation (4.1), which are fitted to the FE computation results (filled
symbols). The line indicated by the arrow represents the velocity profile of a pure Couette
flow. As Ca increases, the profile approaches that of a pure Couette flow.

The modified White–Tallmadge model for confined dip coating flow can be obtained
in a manner similar to that described in § 3.1:
0.9458
T = Cc (Ca − T 2 L2 )2/3 . (B 2)
L
As shown in figure 13, the corrected equation (B 2) satisfactorily predicts T in the
low-Ca regime. In the thin-film limit, i.e. an extremely low-Ca flow, equation (B 2)
reduces to (3.14) because the gravitational effect vanishes.

Appendix C. Two limits of confinement-correction factor

C.1. Small-L limit
A power series expansion for (3.18) about the point k = 0 of order one in k is

L = k cos θc + O(k3 )
https://doi.org/10.1017/jfm.2017.421

as k → 0, (C 1)

and with the relation (κ0 /L)2 = 4/k2 − 4, the expansion for the confinement-correction
factor about the point k = 0 of order one in k is
k
Cc = √ + O(k3 ) as k → 0. (C 2)
2
 Confinement effects in dip coating 27

100
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 22 Aug 2017 at 05:30:17, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.

10–1

10–2

10–3
10–2 10–1 100 101 102
Ca

F IGURE 13. (Colour online) Dimensionless thickness T obtained from FE computation

results (filled symbols) and confinement-corrected White–Tallmadge equation (B 2) (solid
lines) at θc = 30◦ . The corrected equation predicts T in the low-Ca regime reasonably well;
however, the predictions deviate as Ca increases.

From (C 1) and (C 2), the confinement-correction factor can be expressed as

L
Cc = √ + O(L3 , L2 k3 ) as L → 0 and k → 0. (C 3)
2 cos θc
For small L, Cc is a linear function of L, as shown in figure 3. In addition, the
elliptic modulus k approaches 2L/κ0 and is also a linear function of L, which implies
that k → 0 as L → 0.

C.2. Large-L limit

For large L, equation (3.18) can also be approximated by the Taylor series expansion
at κ0 = 0
1 + sin φ
 
L = 4(sin φ − 1) − 2 ln κ0 . (C 4)
8 cos φ
Equation (C 4) can be rearranged to give
https://doi.org/10.1017/jfm.2017.421

κ0 = 8β(φ) exp(−L/2), (C 5)
where β(φ) = exp[2(sin φ − 1)] cos φ/(1 + sin φ). The confinement-correction factor
can then be approximated as
1
p . (C 6)
1 + 32β 2 exp(−L)
 28 O. Kim and J. Nam
Equation (C 6) becomes unity for large L, because of the vanishing confinement
effects.
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 22 Aug 2017 at 05:30:17, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.

Appendix D. Validation of FE model

The proposed numerical model was tested for the case of large L, for comparison
with the previously reported results for the range L → ∞, i.e. an unconfined pool. To
select an appropriate L value, we performed an arc-length continuation with respect
to l to verify the variation in the h∗∞ value below six significant digits. From the test,
L = 70 was chosen. On the other side, the contact angle at the container wall was
set to π/2 to emphasise an unconfined pool. However, the film thickness should be
independent of the contact angle in an unconfined pool; indeed, a different selection
of the contact angle produces a negligible difference in the film thickness in our FE
model.
Figure 11 shows a comparison between the results from our FE model with L = 70
and other unconfined pool models and experiments. Our predictions exhibit excellent
agreement with the results from Jin et al. (2005), which are regarded as the best high-
accuracy results for inertialess dip coating flow.

Appendix E. Comparison of confined dip coating problem with Bretherton

problem in Hele-Shaw cells at low Ca
For low Ca, confined dip coating flow appears to be similar to that of the Bretherton
problem in Hele-Shaw cells. The flow is caused by a receding meniscus inside the cell,
which is due to two sandwiching walls at a certain gap; hence, thin layers are coated
onto both walls (Park & Homsy 1984). For an infinitesimal depth-to-gap aspect ratio,
the coated film thickness h∗∞ is given by

h∗∞
= 1.3376Ca2/3 . (E 1)
l/2

Note that the gap size is the same as l. In contrast to the confined dip coating problem,
the static meniscus shape in the cells is approximated as an arc of a circle with a
constant curvature of 2/l to leading order O(1) (see Park & Homsy 1984, equations
(4.6) and (4.25a)). Furthermore, the flow in the Hele-Shaw cell has plane symmetry,
but the flow in confined dip coating does not. These discrepancies produce differences
in the prefactors between equations (3.20) and (E 1), except for low L and a zero
contact angle. This exceptional case is discussed in the main text.

REFERENCES

A BEDIJABERI , A., B HATARA , G., S HAQFEH , E. S. G. & K HOMAMI , B. 2011 A computational

study of the influence of viscoelasticity on the interfacial dynamics of dip coating flow.
https://doi.org/10.1017/jfm.2017.421

J. Non-Newton. Fluid Mech. 166, 614–627.

A FANASIEV, K., M ÜNCH , A. & WAGNER , B. 2007 Landau–Levich problem for non-Newtonian liquids.
Phys. Rev. E 76, 036307.
A HN , K., K IM , D., K IM , O. & N AM , J. 2015 Analysis of transparent conductive silver nanowire
films from dip coating flow. J. Coat. Technol. Res. 12, 855–862.
B ENILOV, E. S., C HAPMAN , S. J., M CLEOD , J. B., O CKENDON , J. R. & Z UBKOV, V. S. 2010 On
liquid films on an inclined plate. J. Fluid Mech. 663, 53–69.
 Confinement effects in dip coating 29
B ENILOV, E. S. & Z UBKOV, V. S. 2008 On the drag-out problem in liquid film theory. J. Fluid
Mech. 617, 283–299.
B LAKE , T. D. & RUSCHAK , K. J. 1979 A maximum speed of wetting. Nature 282, 489–491.
B OLSTAD , J. H. & K ELLER , H. B. 1986 A multigrid continuation method for elliptic problems with
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 22 Aug 2017 at 05:30:17, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.

folds. SIAM J. Sci. Stat. Comput. 7, 1081–1104.

B RETHERTON , F. P. 1961 The motion of long bubbles in tubes. J. Fluid Mech. 10, 166–188.
C AMPANA , D. M., U BAL , S., G IAVEDONI , M. D. & S AITA , F. A. 2010 Numerical prediction of the
film thickening due to surfactants in the Landau–Levich problem. Phys. Fluids 22, 032103.
C AMPANA , D. M., U BAL , S., G IAVEDONI , M. D. & S AITA , F. A. 2013 Dip coating of fibers in
the visco-inertial regime: numerical analysis. Ind. Engng Chem. Res. 52, 12646–12653.
C ARVALHO , M. S. & K HESHGI , H. S. 2006 Low-flow limit in slot coating: theory and experiments.
AIChE J. 46, 1907–1917.
C OYLE , D. J. 1997 Knife and roll coating. In Liquid Film Coating, pp. 539–571. Springer.
D IXIT, H. N. & H OMSY, G. M. 2013a The elastic Landau–Levich problem. J. Fluid Mech. 732,
5–28.
D IXIT, H. N. & H OMSY, G. M. 2013b The elastocapillary Landau–Levich problem. J. Fluid Mech.
735, 1–28.
D UFF , I. S., E RISMAN , A. M. & R EID , J. K. 1989 Direct Methods for Sparse Matrices. Oxford
University Press.
E SMAIL , M. N. & H UMMEL , R. L. 1975a A note on linear solutions to free coating onto a vertical
surface. Chem. Engng Sci. 30, 1195–1196.
E SMAIL , M. N. & H UMMEL , R. L. 1975b Nonlinear theory of free coating onto a vertical surface.
AIChE J. 21, 958–965.
F ILALI , A., K HEZZAR , L. & M ITSOULIS , E. 2013 Some experiences with the numerical simulation
of Newtonian and Bingham fluids in dip coating. Comput. Fluids 82, 110–121.
G AO , P., L I , L., F ENG , J. J., D ING , H. & L U , X.-Y. 2016 Film deposition and transition on a
partially wetting plate in dip coating. J. Fluid Mech. 791, 358–383.
H OCKING , L. M. 2001 Meniscus draw-up and draining. Eur. J. Appl. Maths 12, 195–208.
JAVIDI , M., P OPE , M. A. & H RYMAK , A. N. 2016 Investigation on dip coating process by
mathematical modeling of non-Newtonian fluid coating on cylindrical substrate. Phys. Fluids
28, 063105.
J ENNY, M. & S OUHAR , M. 2009 Numerical simulation of a film coating flow at low capillary
numbers. Comput. Fluids 38, 1823–1832.
J IN , B., ACRIVOS , A. & M ÜNCH , A. 2005 The drag-out problem in film coating. Phys. Fluids 17,
103603.
K AMOTANI , Y., O STRACH , S. & K IZITO , J. P. 1999 Experimental free coating flows at high Capillary
and Reynolds number. Exp. Fluids 27, 235–243.
K HESHGI , H. S., K ISTLER , S. F. & S CRIVEN , L. E. 1992 Rising and falling film flows: viewed
from a first-order approximation. Chem. Engng Sci. 47, 683–694.
K RECHETNIKOV, R. & H OMSY, G. M. 2005 Experimental study of substrate roughness and surfactant
effects on the Landau–Levich law. Phys. Fluids 17, 102108.
K RECHETNIKOV, R. & H OMSY, G. M. 2006 Surfactant effects in the Landau–Levich problem. J. Fluid
Mech. 559, 429–450.
L ANDAU , L. & L EVICH , V. G. 1942 Dragging of a liquid by a moving plate. Acta Physiochim.
USSR 17, 42–54.
L EE , C. Y. & TALLMADGE , J. A. 1974 Meniscus shapes in withdrawal of flat sheets from liquid baths.
Dynamic profile data at low Capillary numbers. Ind. Engng Chem. Fundam. 13, 356–360.
M AILLARD , M., B LEYER , J., A NDRIEUX , A. L., B OUJLEL , J. & C OUSSOT, P. 2016 Dip-coating of
https://doi.org/10.1017/jfm.2017.421

yield stress fluids. Phys. Fluids 28, 053102.

M ÜNCH , A. 2002 The thickness of a Marangoni-driven thin liquid film emerging from a meniscus.
SIAM J. Appl. Maths 62, 2045–2063.
M ÜNCH , A. & E VANS , P. L. 2005 Marangoni-driven liquid films rising out of a meniscus onto a
nearly-horizontal substrate. Physica D 209, 164–177.
M ÜNCH , A. & E VANS , P. L. 2006 Interaction of advancing fronts and meniscus profiles formed by
surface-tension-gradient-driven liquid films. SIAM J. Appl. Maths 66, 1610–1631.
 30 O. Kim and J. Nam
O LIVER , J. F., H UH , C. & M ASON , S. G. 1977 Resistance to spreading of liquids by sharp edges.
J. Colloid Interface Sci. 59, 568–581.
PAPANASTASIOU , T. C., M ALAMATARIS , N. & E LLWOOD , K. 1992 A new outflow boundary
condition. Intl J. Numer. Meth. Fluids 14, 587–608.
Downloaded from https://www.cambridge.org/core. Columbia University Libraries, on 22 Aug 2017 at 05:30:17, subject to the Cambridge Core terms of use, available at https://www.cambridge.org/core/terms.

PARK , C.-W. & H OMSY, G. M. 1984 Two-phase displacement in Hele-Shaw cells: theory. J. Fluid
Mech. 139, 291–308.
Q UÉRÉ , D. 1999 Fluid coating on a fiber. Annu. Rev. Fluid Mech. 31, 347–384.
R ÉGLAT, O., L ABRIE , R. & TANGUY, P. A. 1993 A new free surface model for the dip coating
process. J. Comput. Phys. 109, 238–246.
R ENARDY, M. 1997 Imposing ‘no’ boundary condition at outflow: why does it work? Intl J. Numer.
Meth. Fluids 24, 413–417.
RUSCHAK , K. J. 1985 Coating flows. Annu. Rev. Fluid Mech. 17, 65–89.
DE RYCK , A. & Q UÉRÉ , D. 1996 Inertial coating of a fibre. J. Fluid Mech. 311, 219–237.
DE RYCK , A. & Q UÉRÉ , D. 1998 Gravity and inertia effects in plate coating. J. Colloid Interface
Sci. 203, 278–285.
DE S ANTOS , J. M. 1991 Two-phase cocurrent downflow through constricted passages. PhD thesis,
University of Minnesota.
S CHUNK , P. R., H URD , A. J. & B RINKER , C. J. 1997 Free-meniscus coating processes. In Liquid
Film Coating, pp. 673–708. Springer.
S CRIVEN , L. E. 1988 Physics and applications of dip coating and spin coating. MRS Symp. Proc.
121, 717–729.
S NOEIJER , J. H., D ELON , G., F ERMIGIER , M. & A NDREOTTI , B. 2006 Avoided critical behavior in
dynamically forced wetting. Phys. Rev. Lett. 96, 174504.
S NOEIJER , J. H., Z IEGLER , J., A NDREOTTI , B., F ERMIGIER , M. & E GGERS , J. 2008 Thick films of
viscous fluid coating a plate withdrawn from a liquid reservoir. Phys. Rev. Lett. 100, 244502.
S PIERS , R. P., S UBBARAMAN , C. V. & W ILKINSON , W. L. 1974 Free coating of a Newtonian
liquid onto a vertical surface. Chem. Engng Sci. 29, 389–396.
TALLMADGE , J. A. & S OROKA , A. J. 1969 The additional parameter in withdrawal. Chem. Eng.
Sci. 24, 377–383.
TALLMADGE , J. A. & S TELLA , R. 1968 Some properties of the apparent water paradox in entrainment.
AIChE J. 14, 838–840.
VAN ROSSUM , J. J. 1958 Viscous lifting and drainage of liquids. Appl. Sci. Res. 7, 121–144.
V INOKUR , M. 1983 On one-dimensional stretching functions for finite-difference calculations.
J. Comput. Phys. 50, 215–234.
W EINSTEIN , S. J. & RUSCHAK , K. J. 2001 Dip coating on a planar non-vertical substrate in the
limit of negligible surface tension. Chem. Eng. Sci. 56, 4957–4969.
W HITE , D. A. & TALLMADGE , J. A. 1965 Theory of drag out of liquids on flat plates. Chem.
Engng Sci. 20, 33–37.
W ILSON , S. D. R. 1982 The drag-out problem in film coating theory. J. Engng Maths 16, 209–221.
https://doi.org/10.1017/jfm.2017.421