DIFFUSION AND CAPILLARY THEORY OF

DRYING
Table of Contents
1.
Introduction
2.
3.
Theoretical Background
2.1. Constant-Rate and Falling-Rate Periods
2.2. Diffusion Theory of Drying
2.3. Capillary Theory of Drying
2.4. Mechanisms in Drying
4.
5.
Mathematical Models
3.1. Fickian Diffusion Model
3.2. Capillary Flow Model
3.3. Drying Kinetics and Critical Moisture Content
6.
7.
Industrial Applications
4.1. Food Industry
4.2. Petroleum Industry (if added later)
8.
9.
Conclusion (if added later)
10.
11.
References
12.
Diffusion and Capillary Theory of
Drying
Introduction
Drying is a crucial heat-and-mass-transfer process in which moisture is
removed from a material by evaporation. It is widely used in the chemical
process industries to stabilize or prepare solid products (e.g. grains, catalysts,
powders) and to remove contaminants (e.g. water from solvents or drilling
fluids). The physics of drying in porous solids is complex, involving both
thermal and mass transport phenomena. From a microscopic viewpoint,
drying typically proceeds through two distinct regimes: a constant-rate
period where liquid water is continuously supplied to the surface, and one or
more falling-rate periods where the surface dries out and moisture must
migrate from the interior. Two classical models – the diffusion theory of
drying and the capillary theory of drying – have been developed to explain
how water moves during drying, especially in the falling-rate regime. In
diffusion theory, moisture moves by concentration gradients (analogous to
Fick’s law), whereas in capillary theory, moisture flows in liquid form driven by
capillary pressures in the pore network. Understanding both mechanisms is
essential for designing efficient dryers and predicting drying kinetics. This
report reviews these theories and their mathematical models, and discusses
practical implications for chemical industries such as food processing and
petroleum refining.
Theoretical Background
Drying porous solids often exhibits a typical rate/time profile (see Fig. 1).
During an initial constant-rate period, the surface remains wetted by liquid
and evaporation occurs at a nearly constant rate determined by external
conditions (air flow, humidity, etc.). Once the surface can no longer be
replenished by liquid transport (at the critical moisture content), the falling-
rate period begins and the drying rate declines continuously. The transition
between these periods depends on how moisture migrates from the interior
to the surface.
Figure 1. Typical drying-rate curves for a porous solid, illustrating a constant-
rate period followed by one or more falling-rate periods【52†】. In the
constant-rate period (A–B), evaporation at the surface is sustained by liquid
flow; when this flow can no longer keep up (point B), a “characteristic length”
of drying front is reached and the surface begins to recede (B–C), initiating the
falling-rate period【52†】. During the falling-rate period (C–D), the drying rate
declines as moisture must diffuse or wick from deeper within the solid.
Diffusion Theory of Drying
In the diffusion theory of drying, moisture movement within the solid is
treated analogously to molecular diffusion of a species. One assumes that,
once the surface dries, water transport is limited by a concentration (or
moisture content) gradient. Mathematically, this is described by Fick’s second
law: for one-dimensional transport in a slab,
∂C∂t=Deff∂2C∂x2,\frac{\partial C}{\partial t} = D_{\rm eff} \frac{\partial^2
C}{\partial x^2},
where CC is moisture concentration (kg/m³) and DeffD_{\rm eff} is an effective
moisture diffusivity (m²/s) that may depend on moisture and temperature. In
foods or granular solids, diffusion is often through bound water or vapor in
pore space. Analytic solutions for simple geometries yield drying curves of the
form
M(t)−MeM0−Me=∑n=1∞Anexp (−Deffkn2t),\frac{M(t)-M_e}{M_0-M_e} =
\sum_{n=1}^\infty A_n \exp(-D_{\rm eff} k_n^2 t),
where MM is moisture content, MeM_e equilibrium moisture, and knk_n are
geometry-dependent constants (e.g. k1=π/Lk_1 = \pi/L for a slab of half-
thickness LL). This solution shows an exponential (diffusion-like) decline of
moisture, consistent with observed falling-rate behavior. Crank’s diffusion
equation is a classic result for heat or mass transfer in a slab. In practice,
empirical forms like the Page model M/M0=exp (−ktn)M/M_0 = \exp(-k t^n)
are often used, where kk, nn are fitting parameters (see also discussion in).
These models capture the gradual decline of drying rate when internal
diffusion is rate-controlling.
Capillary Theory of Drying
The capillary theory emphasizes liquid water transport due to capillary forces.
In porous media, narrow channels and pores act like capillary tubes: surface
tension at liquid–air interfaces generates suction that draws liquid from wetter
regions towards drier regions or towards the surface. If enough liquid
continuity exists, capillary flow can sustain surface wetting and hence
evaporation. The classic Lucas–Washburn relation describes capillary
imbibition: a liquid will penetrate a pore of radius rr a distance LL in time tt
according to
L2=γrcos ϕ2μt,L^2 = \frac{\gamma r\cos\phi}{2\mu} t,
where γ\gamma is surface tension, μ\mu viscosity, and ϕ\phi the contact
angle. This shows that capillary-driven penetration has a diffusive-like
∝ t\propto\sqrt{t} behavior, but here the driving force is capillary pressure
Pc=2γcos ϕrP_c = \frac{2\gamma \cos\phi}{r}. In drying, one can similarly
write Darcy’s law for liquid transport under capillary pressure gradients.
Capillary theory was originally developed to explain why, during the surface-
drying stage, moisture appears to move in liquid form. As drying proceeds,
the pore liquid can rupture and form disconnected films, but until that
happens the capillary network can transport water to the surface. The
classical capillary drying model (Haines, Slichter, others) envisions closely-
packed spheres or granular beds where microscopic water bridges form
between particles. When the surface layer begins to dry, these bridges break
and capillary flow ceases, marking the end of the constant-rate period. In
summary, capillary theory attributes falling-rate drying to the loss of capillary
connectivity: above a critical moisture level the capillary network is intact
(liquid flow dominates), and below it only vapor diffusion remains. Indeed,
studies of porous soils report a “critical water content” θcrθ_{cr} above which
capillary flow dominates and below which vapor diffusion controls drying.
Mechanisms in Drying
In practice, diffusion and capillary flow often act together in porous materials.
During the constant-rate period, rapid capillary flow keeps the surface
saturated and the drying flux is essentially the evaporation rate from a liquid
surface. In the ensuing falling-rate period, moisture must migrate as vapor or
as remaining liquid over shorter distances. Vapor diffuses through the now-
dry pores and surface films, while any remaining liquid can still creep via
microscopic capillaries but with much slower rate. The net moisture flux is
typically described by combining Fick’s law for vapor diffusion and Darcy’s
law for liquid under capillary pressures. Modern drying models for hygroscopic
porous media explicitly incorporate both terms. For example, in soil and food
drying the overall moisture flux qq can be expressed as
q=−Deff∇w−Kcap∇Pc(w),q = -D_{\rm eff}\nabla w - K_{\rm cap}\nabla P_c(w),
where ww is moisture (volumetric or gravimetric), Pc(w)P_c(w) is the capillary
pressure (a function of saturation), and KcapK_{\rm cap} a permeability. The
interplay of these terms, along with heat transfer, governs the drying profile.
In summary, diffusion theory emphasizes Fickian transport of moisture, while
capillary theory emphasizes liquid flow driven by surface tension; real drying
often involves both, transitioning from one to the other as drying front
recedes.
Mathematical Models
Several mathematical frameworks have been developed to quantify the
diffusion and capillary mechanisms. Below we outline key equations and
models used in drying analysis.
Fickian Diffusion Model
When drying is assumed to be diffusion-limited, Fick’s second law is the
starting point. In one dimension (thickness direction) for a slab of thickness LL,
the moisture content C(x,t)C(x,t) satisfies
∂C∂t=Deff∂2C∂x2,\frac{\partial C}{\partial t} = D_{\rm eff} \frac{\partial^2
C}{\partial x^2},
with boundary condition C(0,t)=C(L,t)=Csurface(t)C(0,t)=C(L,t)=C_{\rm
surface}(t) (which may be time-varying) and initial C(x,0)=C0(x)C(x,0)=C_0(x).
In many studies, a simplification is to assume a constant diffusivity DeffD_{\rm
eff} and that the surface instantly equilibrates to the ambient humidity (so
CsurfaceC_{\rm surface} is fixed). Solutions (Crank’s method) give drying
curves. For example, for a slab with convective boundary, the solution can be
expressed as an infinite series of exponentials. In many drying problems, the
first term dominates at long times, yielding an approximate one-term solution:
M(t)−MeM0−Me≈exp (−Deffπ2t4L2) ,\frac{M(t)-M_e}{M_0-M_e} \approx
\exp\Big(-\frac{D_{\rm eff}\pi^2 t}{4L^2}\Big)\,,
where MM is moisture (dry-basis or wet-basis) and MeM_e is equilibrium
moisture. The effective diffusivity DeffD_{\rm eff} often depends on moisture
and temperature (higher TT tends to increase DeffD_{\rm eff}). In food
engineering, researchers often extract DeffD_{\rm eff} by fitting drying data to
Fickian models. For example, the drying of food slices in a convection dryer
can be modeled by one-dimensional diffusion to estimate an average
diffusivity.
In practice, simple diffusion models are often extended to include temperature
coupling (energy equation) and concentration-dependent DD. But the basic
prediction is that moisture decays exponentially with time (as in Fig. 1). If the
external resistances are small, the drying rate eventually follows a square-root
-of-time law during falling rate, consistent with diffusion-limited behavior.
Empirical models, such as the Page, Newton or Henderson–Pabis equations,
are commonly used to fit drying curves without explicitly solving diffusion
equations. These phenomenological models (e.g. M/M0=exp (−ktn)M/M_0 =
\exp(-k t^n)) can be thought of as approximating Fickian solutions with fitted
parameters.
Capillary Flow Model
Capillary liquid transport is often modeled using Darcy’s law or the
Lucas–Washburn equation. In porous media, Darcy’s law for liquid flux JlJ_l
under a capillary pressure gradient is
Jl=−kpermμ∂Pc∂x,J_l = - \frac{k_{\rm perm}}{\mu} \frac{\partial P_c}{\partial x},
where kpermk_{\rm perm} is the permeability and μ\mu the liquid viscosity.
The capillary pressure PcP_c relates to moisture saturation via pore geometry
(Young–Laplace law: Pc=2γcos ϕ/rP_c = 2\gamma\cos\phi/r for cylindrical
pores of radius rr). Combining Darcy’s law with the porosity and fluid
continuity yields Richards’ equation for unsaturated flow, which couples liquid
transport with moisture diffusion. In drying models this term accounts for
liquid wicking.
A simpler capillary model uses the Lucas–Washburn result: for a single pore,
liquid penetration L(t)L(t) follows L2∝tL^2 \propto t. Generalizing to a porous
matrix, one can write an averaged relation for capillary penetration. For
example, the penetration length in a horizontal capillary of equivalent radius
rer_e is
L2=γrecos ϕ2μ t.L^2 = \frac{\gamma r_e \cos\phi}{2\mu} \, t.
This predicts a capillary-driven flow that slows down over time. In reality, as
drying proceeds, the capillary radius and connectivity change (large pores
empty first, leaving smaller water films). More sophisticated models treat the
porous medium as a bundle of capillaries or use pore-network simulations.
A combined model of drying includes both terms: vapor diffusion and liquid
capillary flow. One form of a 2-phase moisture diffusion equation is
∂w∂t=∇⋅(Dvap∇w)+∇⋅(Dcap∇w),\frac{\partial w}{\partial t} =
\nabla\cdot\big(D_{\rm vap} \nabla w\big) + \nabla\cdot\big(D_{\rm cap} \nabla
w\big),
where ww is moisture content, DvapD_{\rm vap} an effective vapor diffusivity
(function of porosity and temperature), and DcapD_{\rm cap} an effective
capillary diffusivity (related to permeability and pore connectivity). Analytic
solutions of this nonlinear equation are complex, but numerical methods show
that at high moisture contents (θ>θcr\theta>\theta_{cr}) the capillary term
dominates, while at low moisture vapor diffusion dominates.
Drying Kinetics and Critical Moisture Content
A key result of combining diffusion and capillary effects is the existence of a
critical moisture content or characteristic length. Before this point, liquid can
still reach the surface; after, the front recedes. In soils and porous solids, this
critical point marks the end of constant-rate drying. Empirically, it is observed
that the drying rate stays roughly constant until reaching a moisture level at
which the rate starts to fall off. The height of the evaporation front or the
dimensionless penetration depth when this occurs can be predicted by
setting the liquid flux equal to the evaporation demand. For instance, when
capillary flow can no longer supply the evaporating surface (i.e., when the
penetration length LL reaches the slab thickness), the drying switches to
falling rate. After this, the drying rate decreases approximately with the
inverse of the (growing) diffusion distance. In many engineered drying models,
one explicitly tracks the moving evaporation front; others simply fit two
segments (constant and falling) with corresponding parameters.
Mathematically, the constant-rate drying flux JcJ_c is often given by the mass
-transfer expression Jc=kg(ws−w∞)J_c = k_g (w_s - w_\infty), where kgk_g is
the gas-phase mass-transfer coefficient and wsw_s, w∞w_\infty are surface
and ambient humidity. Meanwhile, the falling-rate flux can be estimated from
internal diffusion: for example, for a slab of thickness LL with one end
maintained at vapor pressure Pv∗P_v^*, the diffusion flux is
Jd≈Dv(Pv∗−P∞)/LJ_d \approx D_v (P_v^*-P_{\infty})/L. When Jc=JdJ_c = J_d at
some moisture content, that defines the critical point. In summary, the
diffusion model yields an exponential decay of moisture, whereas capillary
flow yields a L2∝tL^2\propto t behavior until cutoff. Both have been observed
depending on the material and moisture level.
Industrial Applications
Drying processes are ubiquitous in chemical engineering, with practical
applications in the food and petroleum industries among others. The choice
of dryer, operating conditions, and theoretical model depends on the product
and desired quality. Below we highlight examples from these sectors,
emphasizing how diffusion and capillary mechanisms are relevant.
Food Industry: Drying foods (fruits, vegetables, grains, meats) preserves them
and reduces storage weight. Foods are generally porous, hygroscopic solids.
In the early stage of drying (especially if thin layers or high flow), moisture can
be removed at near-constant rate as liquid water reaches the surface through
capillaries and intercellular channels. However, most hygroscopic foods
quickly enter the falling-rate period where moisture is bound or held in cells.
The internal moisture then migrates mainly by vapor diffusion and limited
capillary action. For example, in drying fruits, cell walls collapse as water is
removed, reducing pore connectivity. Manufacturers often use low-
temperature convection or freeze-drying to gently remove moisture; the
capillary theory implies that maintaining liquid continuity (as in freeze-drying)
can enhance drying uniformity. Changes in texture and shrinkage during
drying also reflect capillary forces (as pores empty, capillary stresses can
deform the structure). Process models for food drying often incorporate Fick’s
law with moisture-dependent diffusivity or empirical drying-rate curves. The
following figure shows typical moisture-vs-time curves for drying foods;
differences in curve shape can reflect material properties and drying
conditions.
Figure 2. Typical moisture-content profiles during drying of porous food
materials, illustrating (a) a two-stage drying (with initial rapid loss and slower
tail) and (b) a more uniform decay. The shape depends on material porosity
and hygroscopicity. Hygroscopic foods (like salts or dry spices) reach the
falling-rate stage quickly, whereas less hygroscopic items (e.g. high-porosity
biscuits) may have longer constant-rate periods.
Quality considerations in food drying include preserving nutrients and
avoiding case hardening (surface sealing by rapid drying). Theories of
diffusion and capillary transport help in designing dryers to avoid these issues.
For instance, at high temperatures, capillary flow may reduce (due to lower
viscosity) but diffusion increases, affecting the rate. Mathematical models of
food drying often solve coupled heat–mass equations with effective diffusion
terms, calibrated by experiments.
Petroleum Industry: Drying in the petroleum sector may involve removing
water or solvents from oil-related solids. One example is drilling wastes:
cuttings and sludge from drilling fluids. These often contain drilling mud
(water-based or oil-based). Before disposal or reuse, the cuttings are dried
(thermally or via microwave) to remove water and reduce volume. Studies
show that drying such slurries follows the same basic physics: moisture must
evaporate from surfaces, and internal transport can be by capillary wicking of
water from cuttings to the surface or by vapor diffusion through the cuttings
bed. Microwave drying of petroleum-contaminated sludge is gaining use due
to its faster heating and more uniform energy input; however, regardless of
method, the drying kinetics can be analyzed by diffusion equations. For
example, Moralar et al. (2024) modeled petroleum drilling sludge drying and
noted that moisture is eliminated by gradual evaporation, which is very
energy-intensive.
Another application is catalyst and sorbent drying. Many petroleum
processes use solid catalysts or desiccants (e.g. alumina, silica gels). After use
or impregnation, these solids often need drying before entering service. For
instance, desiccant beds for natural gas dehydration rely on capillary forces to
initially adsorb water and then must be regenerated by heating and purging.
In such porous materials, capillary theory helps predict how quickly liquid
bridges break during regeneration. Similarly, oil sands processing produces
tailings that are filter-pressed and dried; here, the bitumen and water content
must migrate through a porous solid matrix. Although detailed models are
proprietary, the underlying physics still follow diffusion and capillary principles.
In solvent recovery or petrochemical production, drying of granular catalysts
or powders can also be modeled with these theories. The design of rotary
dryers, spray dryers, or fluidized beds for petroleum-derived products takes
into account internal moisture transport: for example, in a rotating drum dryer
the tumbling action exposes new surfaces, but ultimately moisture must
diffuse or wick out of pellets. Equipment selection (e.g. using lower residence
time vs. higher temperature) may be guided by whether drying is predicted to
be diffusion- or capillary-limited.
Discussion
The diffusion vs. capillary paradigm of drying highlights different controlling
mechanisms. In practice, most porous solids exhibit a short constant-rate
period (capillary-held liquid at surface) followed by a longer falling-rate period
(diffusion-limited). The exact balance depends on material properties: pore
size distribution, connectivity, surface tension effects, and binding energy of
water. Materials with very small pores (e.g. fine clays or hygroscopic gels) tend
to trap liquid, extending capillary-driven drying. In contrast, highly porous,
coarse materials (e.g. porous rock or some foams) may dry mainly by vapor
diffusion.
Analytically, the drying behavior can often be captured by a single parameter:
the Biot number for mass transfer, which compares internal resistance
(diffusion) to external resistance (gas-film). A high Biot number means internal
diffusion controls (falling-rate), whereas a low Biot suggests external film
controls (constant-rate). However, for many solids there are two Biot numbers:
one for liquid (capillary) flow and one for vapor diffusion. Correctly predicting
drying time often requires solving nonlinear PDEs or using numerical
simulation (e.g. finite elements). Recent research uses pore-network models to
simulate how liquid clusters and films evolve during drying, and how pore-
scale capillary barriers affect moisture removal.
From a design standpoint, understanding these mechanisms informs dryer
operation. For example, during falling-rate drying, raising the temperature has
diminishing returns if diffusion is very slow; instead, reducing sample
thickness or increasing agitation may be more effective. The energy efficiency
of drying is also tied to these mechanisms: in the constant-rate regime, latent
heat of evaporation dominates; in the falling-rate regime, heat contributes to
raising temperature and driving diffusion. Therefore, multi-stage drying (e.g.
initial high-temperature flash, followed by milder finish) is common.
One challenge is accounting for shrinkage and cracking. As a porous matrix
dries, capillary stresses can cause collapse, creating new pores or altering
diffusivity. This structural change feeds back on transport mechanisms. Some
advanced models include porosity as a function of moisture. Another issue is
non-uniform moisture: drying fronts may form (wet core and dry shell),
invalidating simple 1-D assumptions. In such cases, capillary theory can predict
the front speed based on liquid connectivity. Experiments using MRI and
neutron imaging have visualized these fronts in granular beds, confirming
that at early times a clear wet front recedes (capillary flow-dominated) and
later the moisture profile smooths out (diffusion-dominated).
In industry, the choice of model often trades off complexity vs. practicality.
Simple diffusion models can be fit to data and yield effective diffusivities;
more complex models can predict the internal state. Empirical fitting (e.g.
Page or Nusselt models) is very common in food drying. Nevertheless, modern
process control and design increasingly use mechanistic models for
optimization. For example, computational fluid dynamics (CFD) combined
with drying kinetics can simulate a whole dryer.
Conclusion
Drying of porous solids involves an interplay of diffusion and capillary flow of
moisture. The diffusion theory models drying as Fickian transport of water
(vapor or bound liquid) driven by concentration gradients, yielding rate laws
akin to the mathematics of diffusion. The capillary theory emphasizes liquid
transport by capillarity; it explains why a drying surface can remain wet and
why a critical moisture content exists for the onset of falling-rate drying. In
practice, porous media drying typically begins with capillary-dominated
transport (constant rate) and then transitions to diffusion-dominated
transport (falling rate) as described in many studies.
For chemical engineers, both theories are important: diffusion-based models
(using Fick’s law or empirical approximations) are useful for predicting the
slow drying phase, while capillary-flow models are essential to understand
surface phenomena and front movement. Industrial dryers (in food, petroleum,
etc.) are designed with these mechanisms in mind: e.g. optimizing air velocity
for constant-rate stage, or reducing thickness for diffusion-limited stage.
Modern research continues to refine these models by including realistic pore
networks, heat–mass coupling, and material-specific parameters.
In summary, a complete understanding of drying requires considering both
diffusion and capillary effects. By applying the relevant mathematical models
and paying attention to material properties (porosity, pore size, hydrophilicity),
engineers can design drying processes that are efficient and meet product
specifications.
References


Crank, J. (1975). The Mathematics of Diffusion (2nd ed.). Oxford

University Press.



Kolesnikov, G., & Gavrilov, T. (2020). Modeling the drying of capillary-

porous materials in a thin layer: application to the estimation of
moisture content in thin-walled building blocks. Applied Sciences, 10(19),
6953. doi:10.3390/app10196953.



Luikov, A. V. (1961). Heat and Mass Transfer in Capillary-Porous Bodies.

Pergamon Press.



Moralar, A., Yıldız, S., Hasol, İ., Şahin, S., & Yıldız, A. Ö. (2024). Modeling
drying process parameters for petroleum drilling sludge with ANN and
ANFIS. Processes, 12(9), 1948. doi:10.3390/pr12091948.


Mujumdar, A. S. (Ed.). (2006). Handbook of Industrial Drying (3rd ed.).
CRC Press.



Piggott, R. W., van Genuchten, M. T., & Illangasekare, T. H. (2020).

Identifying the dominant mode of moisture transport during drying of
unsaturated soils. Scientific Reports, 10, 6399. doi:10.1038/s41598-020-
61302-w.



Srichuwong, S., & Douglas, L. (2009). Effect of root, interior, and

peripheral hardness on water diffusivity in bulgur. International Journal
of Food Science & Technology, 44(1), 1–8.
