See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/275670721

Liquid flow in polyurethane foams for filtration applications: a study on their

characterization and permeability estimation

Article  in  Journal of Porous Materials · June 2015

DOI: 10.1007/s10934-015-9948-2

CITATIONS READS

11 294

4 authors:

Subhashini Gunashekar Krishna M. Pillai

University of Wisconsin - Milwaukee University of Wisconsin - Milwaukee
16 PUBLICATIONS   100 CITATIONS    119 PUBLICATIONS   2,142 CITATIONS   

SEE PROFILE SEE PROFILE

Benjamin C. Church Nidal Abu-Zahra

University of Wisconsin - Milwaukee University of Wisconsin - Milwaukee
26 PUBLICATIONS   899 CITATIONS    67 PUBLICATIONS   384 CITATIONS   

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Mechanical properties of Cast Alumina-forming austentic alloys View project

Investigation of the effect of micro structure on wicking on using Lattice Boltzmann Method View project

All content following this page was uploaded by Krishna M. Pillai on 07 June 2016.

The user has requested enhancement of the downloaded file.

Liquid flow in polyurethane foams
for filtration applications: a study on
their characterization and permeability
estimation

Subhashini Gunashekar, Krishna

M. Pillai, Benjamin C. Church & Nidal
H. Abu-Zahra

Journal of Porous Materials

ISSN 1380-2224
Volume 22
Number 3

J Porous Mater (2015) 22:749-759

DOI 10.1007/s10934-015-9948-2

1 23
Your article is protected by copyright and all
rights are held exclusively by Springer Science
+Business Media New York. This e-offprint is
for personal use only and shall not be self-
archived in electronic repositories. If you wish
to self-archive your article, please use the
accepted manuscript version for posting on
your own website. You may further deposit
the accepted manuscript version in any
repository, provided it is only made publicly
available 12 months after official publication
or later and provided acknowledgement is
given to the original source of publication
and a link is inserted to the published article
on Springer's website. The link must be
accompanied by the following text: "The final
publication is available at link.springer.com”.

1 23
 Author's personal copy
J Porous Mater (2015) 22:749–759
DOI 10.1007/s10934-015-9948-2
Liquid flow in polyurethane foams for filtration applications:
a study on their characterization and permeability estimation
Subhashini Gunashekar1 • Krishna M. Pillai2 • Benjamin C. Church1
Nidal H. Abu-Zahra1
Published online: 22 March 2015
Ó Springer Science+Business Media New York 2015
Abstract Open-cell polyurethane foams have been in-
vestigated for various filtration applications to treat in-
dustrial byproducts and waste water. Flow-related
morphological characterization of three types of poly-
urethane foams: (1) control, unmodified polyurethane
foam, (2) flexible polyurethane foam obtained by chain
extension, and (3) rigid polyurethane foam obtained by
chain extension, is presented in this study. Analysis of
micrographs, using histograms based on pore size, revealed
that the pore size varied in the foams based on their
composition. The permeability (K) of the foams for liquid
flow, measured by a falling-head permeameter was in the
range of 0.45 9 10-11–2.62 9 10-11 m2. In comparison to
the theoretical permeability models, the experimental val-
ues were smaller by two orders of magnitude. The theo-
retical prediction of the permeability improved after
considering effective porosity and replicated the trend seen
in the experiments, but the large gap between the predic-
tion and experiments remained. The failure of the theore-
tical models points to the need for developing more
accurate permeability models after including 3-D pore in-
terconnections and swelling of solid matrix.
Keywords Open-cell polyurethane foam  Permeability 
Porosity  Porous medium  Darcy’s law
& Krishna M. Pillai
krishna@uwm.edu
1
Materials Science and Engineering Department, University of
Wisconsin - Milwaukee , 3200 N Cramer St,
Milwaukee 53211, WI, USA
2 purposes. However, the ease of processing has allowed the
Mechanical Engineering Department, University of
Wisconsin – Milwaukee, 3200 N Cramer St, Milwaukee,
WI 53211, USA
•
1 Introduction
In recent years, the concept of a truly tailor-made polymer
has become a reality as a result of better understanding of
the polymer structure–property relationship, newer poly-
merization techniques, and availability of new low-cost
monomers. The endless possibilities of the structure–
property–processing relationship, have led to fascinating
applications of these materials [1]. Foamed plastics are
resins with a sponge-like cellular structure. They usually
contain a dispersed gaseous phase encapsulated by solid
matrix. The difference in the gaseous-phase and solid-
phase properties allows one to tailor make foams with
unique phase-combinations that have desired properties for
specific applications. The gaseous voids dispersed in the
polymer matrix alter the structure, properties, and mor-
phologies of the solid polymer in numerous ways. For
example, the cellular pore structure in foams can be used to
regulate the fluid-flow velocity, dissipate disturbances, and
increase the mass transfer area [2, 3]. The main applica-
tions of foams stem from the modification of physical
properties of the solid matrix by the presence of a gas as
continuous or discontinuous phase [4]. The density reduc-
tion obtained due to the dispersion of gas bubbles inside the
polymeric matrix gives the foam its characteristic proper-
ties, including insulation and mechanical strength. The
amount of gas dispersed depends on the cell size distri-
bution. Density reduction also leads to structural changes in
foams. The final properties of the foam depend on various
factors such as expansion ratio, cell size distribution, and
open cell content [4, 5].
Polyurethane foams are most widely used for insulation
tailoring of these foams for other applications. The flow of
fluids through open-cell foams has been investigated
123
 Author's personal copy
750
extensively for a variety of engineering applications [6].
Open-cell polyurethane foams have shown to exhibit a
reasonable capacity for ion exchange and hence are suit-
able as ion exchange media for the removal of heavy-metal
ions [7]. The synthesis and characterization of chain-ex-
tended polyurethane foams as a filtration medium to re-
move lead ions from water has been discussed in detail by
the authors elsewhere [8, 9]. The cell structure and per-
meability are important parameters that describe the suit-
ability of a porous medium for flow-through applications
and hence a thorough understanding of these parameters
for the functionalized polyurethane foam is essential for its
end-use. In their review of the geometry-based theoretical
models of cellular plastics, Edouard et al. [10] observed
that geometric modelling of cellular foams is significantly
more difficult than that of conventional packed beds or
even honeycomb like materials due to the more intricate
geometry of foam structures. Hu and Edouard developed a
new model [11] based on the packing of regular pentagonal
dodecahedrons which was quite close to the real structure
of a unit cell of the solid foam. Pichat, has described an-
other method by considering only simple relationships
based on the porosity and cell diameter to describe the
specific surface area, pressure drop and external solid–fluid
mass transfer [12].
In the last 150 years, several models have been devel-
oped to characterize fluid flow in porous media on the basis
of macroscopically measurable flow quantities. However,
the first of these models can be traced back to Darcy’s
publication in 1856 [13]. Darcy established the well-known
Darcy’s law which states that the pressure drop per unit
length for a flow through a porous medium is proportional
to the product of the fluid velocity and the dynamic vis-
cosity, and inversely proportional to the permeability (K).
Since Darcy’s law is applicable for relatively slow moving
flows, other models based on this have been developed ever
since [14]. A model proposed by Hilyard and Collier [15]
for flow of air shows that the permeability of the reticulated
open-cell foams varies with the square of mean cell di-
ameter (D2). The permeability of the reticulated open cell
aluminum foams determined by this model was found to be
50–200 9 10-9 m2 for 2–4 mm mean diameter of the
cells; respectively [16]. Ahmed et al. have described an
analytical method to determine the air-flow permeability of
solid foams theoretically. The modeling procedure de-
scribed by them is based on a modified cubic lattice con-
sidering only the physical principles and geometrical
considerations with no adjustable parameters in order to
reconcile the theoretical work with the experimental data
[17]. They validated this model on slim and fat foams to
determine the limit values of permeability and concluded
that the pressure drop cannot be determined accurately by
this model as the foams will be in an intermediate form
123
J Porous Mater (2015) 22:749–759
during operation depending on the flow velocity and por-
osity of the foam [17]. Mills has described the use of
computational fluid dynamics to determine and compare
the air-flow permeability for a range of open celled poly-
urethane foams and concluded that the foam permeability
was a function of the area of the largest hole in the cells
[18]. However this computation was done for laminar air-
flow through a wet Kelvin foam model with regular cell-
hole sizes. This limitation would make it difficult to apply
this method to determine the permeability for water-flow
through other foams with irregular cell-hole sizes. Gent and
Rush [19], studied the effect of laminar and turbulent flows
along with the effects of deformation on permeability of
several open-celled foam materials concluded that for
normal urethane foams, the permeability values were much
smaller and were not related to the cell size. This deviation
was attributed to the presence of membranes and apertures
in the foam which govern the flow resistance instead of the
cell size as described by other researchers. In another
heuristic model reported by Antohe et al. [20], the per-
meability of aluminum foams was found to be 5 9 10-10–
20 9 10-10 m2, irrespective of the mean diameter of the
cells [16]. Dawson et al. [6] developed an improved rela-
tion between the permeability, porosity and compressive
strain of open-cell foams filled with a Newtonian liquid
subjected to compressive strains. For water flow, the per-
meability of the polyurethane foam was found to be in the
range of 3.2–4.9 9 10-9 m2 [6].
From literature, it can be concluded that experimental
determination or theoretical estimation of permeability
using conventional or modified methods and models are
limited by the type of foam under consideration. Several
researchers have considered uniformly sized open-celled
foams which makes it easier to define the cell structure
using geometric models. However, custom-made foams
targeting a particular application, may not always have the
uniform cell structure which limits the usability of con-
ventional methods to determine the flow properties like
permeability and pressure drop of the foam. Thus, research
in this area is very important to develop other methods and
models that can be used to characterize all types of foams.
This paper, demonstrates one such case where the em-
pirical methods do not agree with the experimental deter-
mination of permeability for water-flow through a novel
polyurethane foam that is considered for water filtration
applications.
This study focuses on the structural characterization of
functionalized polyurethane foams using statistical analysis
of their micrographs. The measurements are used to de-
termine the foam permeability using empirical methods,
which are further compared with the predictions from the
theoretical permeability models using the foam porosity.
The observations from this study lead to some unusual
 Author's personal copy
J Porous Mater (2015) 22:749–759
conclusions on the validity of the theoretical permeability
models, developed for particulate porous media, when
applied to the polyurethane foams. It has to be noted that
the terms ‘pores’ and ‘cells’ are used interchangeably in
this work.
The paper is organized as follows. The materials, the
details of the equipment used to characterize the poly-
urethane foam, the procedure for polyurethane foam syn-
thesis and experimental determination of permeability are
discussed in Sect. 2. Section 3 discusses the use of his-
tograms to characterize the pore size distribution in the
foams and compares the permeability values obtained from
experiments to the permeability values obtained from
empirical formulae. This section also discusses the effect of
permeability on the pressure drop of the foam. The main
results and conclusions are summarized in Sect. 4.
2 Experimental details
2.1 Materials for polyurethane foam synthesis
The materials used to synthesize the chain extended poly-
urethane foam are: (1) Polypropylene glycol (PPG), with a
number average molecular weight of 1200, was obtained
from Sigma Aldrich, USA. It was dried in a vacuum oven
at 70 °C for 24 h prior to its use. This forms soft segments
in the synthesized polyurethane and contributes to the
flexibility of the final product. (2) Toluene Diisocyanate
(TDI) containing 80 % of the 2,4 isomer and 20 % of the
2,6 isomer was obtained from Alfa Aesar, USA. This forms
hard segments in polyurethane and contributes to the
rigidity of the final product. (3)N, N-Bis(2-hydroxyethyl)-
2-aminoethanesulfonic acid (BES) with a purity of 99 %
was obtained from Alfa Aesar, USA. This was the chain
extender used to link the linear polymer chains and func-
tionalize the polymer matrix. (4) Dimethyl Sulfoxide
(DMSO) with a purity of 99.9 % was obtained from Alfa
Aesar, USA and was used as a solvent to dissolve the chain
extender. (5) Dibutyltin dilaurate (DBTL) with a purity of
95 % pure was obtained from Sigma Aldrich, USA and
was used as a catalyst to speed up the polyurethane reac-
tion. (6) Polysiloxane surfactant was obtained from Sigma
Aldrich, USA to aid the foaming process. (7) Ultra high
pure Nitrogen gas was obtained from Airgas, USA to carry
out the reaction in an oxygen free environment.
2.2 Functionalized polyurethane foam synthesis
Polyurethane is formed by the addition reaction of the –OH
groups from a polyol (PPG) with the –OCN groups of an
isocyanate (TDI). The linear polymer thus formed can be
linked using a chain extender (BES). The choice of the
751
chain extender determines the availability of functional
groups for various applications. Water was used as a
blowing agent as it reacts with the unreacted –OCN groups
in the polymer to initiate the foaming reaction which is
caused by the formation of a carbamic acid intermediate
that liberates carbon-dioxide gas. The experimental set up
and the detailed procedure to produce chain extend-
ed/functionalized polyurethane foams are discussed else-
where [8, 9]. Of all the compositions analyzed by the
authors, three foam compositions were chosen for this
study based on the physical and structural characteristics of
the foams. The compositions and the physical character-
istics of the foam samples are summarized in Table 1.
The control polyurethane sample (F0), did not contain
any chain extender and hence was not functionalized. It
was mainly used as a reference to compare the flexible (F1)
and rigid (F2) functionalized foam samples. The number of
moles of isocyanate (TDI) in the foam composition alters
the hard segments in polyurethane making the foam either
rigid or flexible [21]. Additionally, Shore-A durometer
(ASTM D2240) was used to determine the hardness of the
foam samples to quantitatively differentiate the foams
synthesized by each formulation. The control, unfunction-
alized polyurethane foam sample, F0, had a hardness of 12
Shore-A. The functionalized polyurethane foam samples,
F1 and F2, measured 7 and 41 Shore-A, respectively. These
readings allow us to characterize F1 as flexible and F2 as
rigid.
2.3 Density measurement
Foam samples from the core were cut into cubes and their
dimensions and weights were recorded to determine the
density according to the standard test method, ASTM D
3574-01. This test method is used for the flexible poly-
urethane foams in slab, bonded or molded forms, and is
based on the weight to volume ratio of the foam specimens.
A circular or square shaped sample from the core of the
molded foam is weighed and its volume is determined by
measuring the dimensions of the sample using a ruler. The
density is then determined by dividing the mass of the
sample with its measured volume and the results are re-
ported in relevant units [22].
2.4 Porosity and pore size measurement
An AutoPore IV 9500 V1.09 porosimeter was used to de-
termine the porosity using the mercury intrusion method.
However, the results were inconclusive due to the reaction
of the foam samples with mercury (see ‘‘Appendix’’).
Hence another method based on the void volume was used
to determine the porosity of the samples. A polyurethane
polymer with the composition of F0 was cast without
123
 Author's personal copy
752
Table 1 Foam types and their composition
Samples PPG (g) TDI (g) BES (g) DMSO (g)
F0 50 18.3 0 0
F1 50 18.3 6.7 15.7
F2 50 27.4 6.7 15.7
foaming and was allowed to set in a glass mold. The
sample was demolded and its density was determined ac-
cording to the ASTM D 3574-01 method described above.
Porosity is defined as the ratio of the volume of pores in
a sample to the total sample volume. The latter was de-
termined by dividing the mass of the foam sample by the
density of the polymer used. The difference in the sample
and polymer volumes yields the void or pore volume of the
foam sample, which is used to determine the porosity of the
sample. The porosity, e, was hence determined using the
equation:
Vsample  Vpolymer
e¼ ð1Þ
Vsample
Here, Vsample is the volume of the foam sample and Vpolymer
is the volume of the un-foamed polymer.
Pore size was measured with a Keyence VHX digital
microscope. Two fields of view were obtained from each
foam sample without regard to orientation. Pore size was
measured by fitting a circle to pores in each image and
taking the diameter of the fitted circle. Histograms based
on pore diameters (i.e., the diameters corresponding to the
largest circles fitted inside various pores in the micrograph)
were generated from two different fields of view for each
foam sample.
2.5 Open-cell content
Flow in foams happens through the interconnected open
cells only, hence it is important to determine the ratio of the
open cells vis-a-vis the total number of cells in the foam.
The open-cell content of the foam samples were evaluated
according to the standard test method C described in
ASTM D 6266. AccuPyc II 1340 FoamPyc v1.07 instru-
ment was used for this analysis. This method uses nitrogen
gas to fill the sample chamber and the difference in gas
volume to the sample volume is used as a measure of the
open cell volume in the foam sample.
2.6 Permeability measurement
A falling-head permeameter set-up, shown in Fig. 1, was
used to determine the permeability of the foam samples. A
graduated cylinder with open ends was fitted to a burette
123
J Porous Mater (2015) 22:749–759
DBTL (g) Surfactant (g) Water (g) Foam type
2 0.25 6 Control
2 0.25 6 Flexible
2 0.25 6 Rigid
stand and a cylindrical foam sample with the same di-
ameter as the graduated cylinder was placed inside it. To
ensure that the wall effect is nullified, the space between
the sample and cylinder walls were sealed using a silicon
caulking agent. Water was run through the foam several
times before starting the experiment to ensure there were
no leaks through the foam walls and all the water was
flowing through the sample only. The main principle be-
hind this method is that a certain volume of water is al-
lowed to run through the sample, and the time taken for this
flow-through the foam sample is directly related to the
permeability of the sample; the higher the permeability, the
lower is the time [23]. The column was filled with water to
the maximum height and the time required for the head to
come down to a final, lower height was recorded.
The permeability, K was then determined by the
equation:
laL h1
K¼ ln ð2Þ
qgAt h2
Here l is the viscosity and q is the density of water at room
temperature, g is the acceleration due to gravity, a is the
cylinder cross-section area, A is the cross-section area of
the foam sample (in this case, a is equal to A as the foam
sample is placed within the graduated cylinder), L is the
length of the foam sample, t is the time of flow, h1 is the
initial height of water in the column, and h2 is the final
height of water in the column [24].
3 Results and discussion
3.1 Density of foams
In general, the density of foam samples increased gradually
as the chain extender (BES) and the isocyanate (TDI)
content increased in the foam formulations. As shown in
Fig. 2, the non-functionalized foam F0 had a density of
0.16 g/cm3, while the density of the functionalized flexible
and rigid foam samples, F1 and F2, were 0.23 and 0.36 g/
cm3; respectively. The increase in the density of the rigid
sample F2 is due to the presence of excess TDI in the foam
composition which forms more hard segments in the
polymer backbone during the synthesis of polyurethane.
 Author's personal copy
J Porous Mater (2015) 22:749–759
Fig. 1 Falling head
permeameter. a Schematic
diagram. b Actual set-up
0.40
0.35
0.30
Density (g/cc)
0.25
0.20
0.15
0.10
0.05
0.00
F0 F1 F2
Foam Sample
Fig. 2 A pictorial comparison of the density of foam samples
3.2 Porosity of foams
Based on the volumes of the polymer and the foam sam-
ples, the porosities of F0, F1 and F2 were estimated to be
89, 85 and 90 %, respectively. The porosity of the non-
functionalized (F0) and the rigid functionalized (F2) foams
seem to be similar in comparison to the flexible function-
alized foam (F1) sample. This suggests two possibilities:
the first is that the solvent DMSO, which is used to dissolve
the chain extender BES in F1, acts as a plasticizer, thus
improving the flexibility of the foam compared to the
sample F0, which does not contain any DMSO or BES in
its formulation [8]. On the other hand, higher amounts of
TDI present in the formulation of F2 sample seems to
lower the plasticizing effect of DMSO [25] by rendering
the foam more rigid with increased porosity. Several re-
actions between DMSO & diisocyanate’s like TDI have
been reported which prove the above observation [26].
Additionally, higher amounts of DMSO can affect the foam
structure as it is a strong organic solvent capable of dis-
solving polyurethane [27].
753
3.3 Open-cell content
The open-cell content of non-functionalized foam sample,
F0, was determined to be 89 %. This shows that the molar
ratio and the type of polyol (PPG) and isocyanate (TDI)
used in the formulation are suitable for the synthesis of
open-cell foams. Functionalizing the foam using the chain
extender BES did not seem to affect the open cell content
as F1 foam samples showed an open-cell content of 91 %.
However, increasing the molar ratio of TDI to PPG seemed
to affect the open-cell content in the foams. As a surprising
discovery, the open-cell content of rigid, F2 foam samples,
were found to be 70 %. It appears that higher amounts of
isocyanate groups in the foam formulation can lead to
excessive foaming followed by foam collapse during
molding; thus reducing the number of open cells in the
foam [28]. Hence, the open-cell content of foam may be
related to its initial composition. On the other hand, one
could also speculate that the rigid nature of the F2 foam
sample may have restricted the flow of gas through it,
which in turn caused the creation of a lower number of
open-cell pores in the foam. However, in order to establish
the validity of either of these plausible mechanisms, further
research is needed.
3.4 Foam structure and pore size analysis
The optical micrographs of the foam samples show that the
structure is different compared to the completely open cells
of reticulated foams due to the presence of cell membranes
between cells [29] as seen in Fig. 3a–c. From the micro-
graphs, it can also be seen that the foam contains several
large pores, with smaller pores distributed on the walls of
the larger ones. Moreover, the pore distribution was not
uniform in the foams and the pores seemed to be denser at
the core. In polyurethane foam production, the foaming and
gelling reactions are critical for cell nucleation, growth,
123
 Author's personal copy
754
Fig. 3 Optical micrographs of a non-functionalized foam F0, b functionalized flexible foam F1, and c functionalized rigid foam F2
distribution, and collapse [3]. The foam composition,
polyurethane chemistry, and foaming process are the main
factors which determine the kind of foam structure [2, 3,
29].
The two sections of F0 foam (Fig. 4a) show a very broad
and a very narrow pore size distribution curves. This
indicates that F0 sample has a non-uniform pore size dis-
tribution. In comparison, two different sections of F1 foam
sample (Fig. 4b) have similar bell-shaped curves indicating
a more uniform pore size distribution.
If the pore size distribution of these two foam samples
were compared based on the process parameters, one could
say that the lack of BES and DMSO in F0 sample may have
led to a non-uniform pore size distribution. In terms of the
Fig. 4 Pore-size-distribution
histograms of a non-
functionalized foam, F0.
b Functionalized flexible foam,
F1. c Functionalized rigid foam,
F2
123
J Porous Mater (2015) 22:749–759
physical properties, the foam was also slightly rigid com-
pared to the F1 foam sample. The flexibility of the foam
may be another reason for the observed differences in pore
size distributions.
On studying the foam compositions for an explanation,
it can be seen that the solvent DMSO used in F1 (Table 1)
acts as a plasticizer imparting flexibility to the foam. The
lack of DMSO in F0 may have restricted the distribution of
pores during foaming, thus leading to a slightly rigid ma-
trix and a non-uniform pore distribution in the foam. The
F2 foam sample had a very rigid structure and, compared to
F1, contained the same amounts of BES and DMSO, but
higher amounts of TDI, in its composition. However, from
the bell-curves shown in Fig. 4c, the pore-size distributions
 Author's personal copy
J Porous Mater (2015) 22:749–759
look similar to that of F0. From the foam composition
given in Table 1, it is clear that the presence of more TDI
enhances the production of CO2 gas during foaming, thus
leading to the formation of more pores. However, some of
these pores are entrapped with no room for expansion due
to the rigorous reaction, while others expand and collapse
eventually [2, 3]. While this is a major factor for the
rigidity in the foam, the presence of BES and DMSO, or
the lack thereof in appropriate ratios, seem to influence the
pore size distribution to some extent.
If a histogram was generated for any foam using the pore
data for multiple cross-sections, the distribution data will be
more reasonable in determining the average pore size of the
foam samples. Histograms generated by combining the
measurements from two different fields of view for each of
the foams are shown in Fig. 5. The average pore size deter-
mined by this method is summarized in Table 2. Comparing
the combined histograms of each foam sample alongside each
other shows that the pore size distribution of the foam samples
does not vary significantly and the average pore diameters of
the foam samples are in the range of 0.4–0.5 mm.
3.5 Permeability and fluid flow through foams
As the foams tested in this study are intended for water
filtration applications, the permeability, K, is an important
method to characterize a porous medium because it is
perhaps the most important property to predict fluid flow in
a porous medium [30]. Several theoretical models are
available to predict the K of porous media. Of these, many
were developed to analyze air-flow through open-cell and
partially open-cell polyurethane foams [16], and were also
used for water flow predictions through these foams as
discussed in the introduction section. The theoretical per-
meability models are often a function of the mean cell
diameter and porosity of the compressed, fully open-cell,
polyurethane foams. However, there are some deviations
from these model predictions when applied to low density
foams subjected to compressive strains [6, 14].
The K of the functionalized and non-functionalized
polyurethane foams considered in this study was
F0 F1 F2
0.30
Pore Fraction
0.20
0.10
0.00
0 0.2 0.4
0.6 0.8 1
1.2 1.4 1.6
1.8 2 2.2 2.4 2.6 F0
2.8 3
Pore Size (mm)
Fig. 5 The combined pore-size histograms of F0, F1 and F2 foams
755
Table 2 Pore diameters based on the combined pore-size distribution
histograms of the three foams
Foam sample Cell diameter (mm)
Average Median
F0 0.49 0.31
F1 0.40 0.28
F2 0.45 0.36
determined using a falling-head permeameter experiment.
It has been shown that K varies as the square of the mean
cell diameter, D, for open-cell polyurethane foams [15], i.e.
K / D2 ð3Þ
Most of the models developed on permeability incor-
porate this primary relationship. As shown in Table 3, the
experimental K values of the chain extended polyurethane
foams are in the range of 0.45 9 10-11–2.62 9 10-11 m2.
The permeability values obtained from literature range
from 0.6 9 10-9 to 1.65 9 10-9 m2 for slow-recovery
polyurethane foams [15, 16]. The experimentally measured
values of the permeability are lower by two orders of
magnitude compared to the values in literature for poly-
urethane foams. However, the literature values for the
permeability are for air flow. (Due to fluid slippage at very
small length-scales, the air-flow permeability is usually
higher than the fluid-flow permeability for any given por-
ous medium [23].)
To compare the experimental values to the conventional
theoretical models for liquid flow, some well-known
models that are predominantly used to determine K for
liquid (water, oil, etc.) flow through porous media, which
are based on the mean particle diameter (D) and porosity
(e) of the porous medium, were chosen. These models have
been found to be suitable for particulate porous media and
are of the general form:
K ¼ D2 /ðeÞ ð4Þ
Here the porosity function, /(e) is a function determined by
the theory used to predict liquid flow in porous media.
Though several such theoretical models are available in
literature, the equations developed by Kozeny and Carmen
and Duplessis & Masliyah described in [30] were consid-
ered in this study. The corresponding porosity functions for
these two models are:
1 e3
Kozeny and Carmen: /ðeÞ ¼ ð5Þ
180 ð1  eÞ2
Duplessis and Masliyah: /ðeÞ
e½1  ð1  eÞ0:33 ½1  ð1  eÞ0:66 
¼ ð6Þ
63ð1  eÞ1:33
123
 Author's personal copy
756
Table 3 Properties of the functionalized polyurethane foams
Foam Open Porosity Effective Avg. cell Permeability,
cell (%) (%) porosity eeff dia. (mm) Kexperimental (m2)
(%)
F0 90 89 80 0.49 2.62 9 10-11
-11
F1 91 85 77 0.40 1.08 9 10
F2 70 90 63 0.45 0.45 9 10-11
Since the foam-based porous media are not made of
particles but cells, the mean cell diameter was used for D to
determine the theoretical permeability. The average cell
diameter was obtained from the statistical analysis of the
micrographs (see Fig. 5; Table 2). The experimental and
theoretical K values along with other foam properties are
summarized in Table 3. The theoretical K values obtained
from the particulate media models are close to the ones
reported in literature [6, 16] for Polyurethane foams,
however the former (like the latter) deviate from the ex-
perimental K values by two orders of magnitude.
The permeability shown in Eq. 3 is directly proportional
to the square of particle (cell) diameter. The theoretical
permeability listed in Table 3 follow this relationship with
the cell diameters for F0, F1 and F2. Foam F1, with the
lowest cell diameter, has the lowest theoretical perme-
ability; whereas, foam F0, with the highest cell diameter,
exhibits the highest theoretical permeability. However, this
trend is not reflected by the actual measured permeability:
although F0 does have the highest permeability, F2 shows
the lowest permeability rather than F1.
In order to explain this anomalous behavior, another pa-
rameter that pertains to the interconnectivity of the cells or
pores has to be considered. Open-cell content (Sect. 3.3),
gives the percentage of the pores that are interconnected and
will take part in the forced flow of a fluid through the sponges.
The values listed in Table 3, show that the percentages of
interconnected pores are higher (around 90 %) in foams F0
and F1, and lower (around 70 %) in foam F2. If we define an
effective porosity, eeff, as the ratio of volume of intercon-
nected pores (open cells) to the total volume, then eeff is the
product of the open cell % and the porosity. The new effective
permeability Keff, using the two theoretical models were ob-
tained by using eeff in Eqs. 5 and 6 as listed in Table 3. After
this modification, the trends in the theoretical permeability
follow the trend seen in the experimentally measured per-
meability: F0 has the highest permeability, while F2 has the
lowest permeability. Hence it is important to take into account
the interconnectivity of pores while estimating any theoretical
permeability. However, despite this improvement, the theo-
retical permeability values are higher than the experimental
permeability values by two orders of magnitude.
123
J Porous Mater (2015) 22:749–759
Permeability, Ktheoretical (m2) Effective permeability, Keff
(m2)
Kozeny and Duplessis and Kozeny and Duplessis and
Carmen Masliyah Carmen Masliyah
77.8 9 10-9 25.3 9 10-9 17.1 9 10-9 6.9 9 10-9
-9 -9 -9
24.2 9 10 8.9 9 10 7.7 9 10 3.3 9 10-9
82.0 9 10-9 25.7 9 10-9 2.1 9 10-9 1.0 9 10-9
Typically, the theoretical models are capable of pre-
dicting the permeability within an order of magnitude for
the particulate and fibrous porous media. The large de-
viation between the experimental results and the theoretical
predictions for the functionalized foams comes as a sur-
prise and is a cause of serious concern. The possible causes
for such a big divergence are speculated to be:
1. The theoretical models consider porosity and the
average cell diameter, and do not take into account
the cell connectivity during water flow through the
foams. The small pores connecting the large cells may
be responsible for the unusually low permeability
values observed experimentally in the polyurethane
foams. (This is because the maximum pressure drops
occur in the smallest cross-sections of the flow
channels.) A form of 3-D network model, where the
larger cells are connecting to their neighbors through
the smaller pores, will be able to predict the perme-
ability values more accurately.
2. As discussed in the following section (Sect. 3.6), the
foams seem to be swelling when exposed to water and
that may lead to the disappearance of many pores
during the permeability experiments. Since the foam
micrographs shown in Fig. 3a–c were taken from dry
foam samples, the mean cell diameter and porosity
values listed in Table 3 may be higher than the actual
values during the flow of water through the foams.
Let us discuss a consequence of the lower-than-expected
permeability of the foam samples. The pressure drop of the
foam can be determined using Darcy’s equation [13, 16]:
KADP
Q¼ ð7Þ
gL
where, DP is the pressure drop applied to a parallelepiped
shaped piece of foam, Q is the volumetric flow-rate through a
cross-section of the geometry that is normal to the flow di-
rection, A is the corresponding cross-section area, L is the
length of the foam in the parallelepiped, and g is the viscosity
of water. For a given flow rate, this equation relates the
permeability to the applied pressure drop inversely. Due to
the lower-than-expected experimental K values, larger
 Author's personal copy
J Porous Mater (2015) 22:749–759
Fig. 6 3D surface micrographs of (left) dry and (right) wet foam F1
applied pressures are required for higher (practical) flow
rate—applications. Such high pressures may deform the
foam structure during the proposed filtering operation, thus
lowering the efficiency of the considered polyurethane
foams.
3.6 Characterization of foam surfaces using surface
microscopy
To determine the effect of conditioning (soaking) on the
structure of the foams, 3D micrographs of the uncondi-
tioned (dry) and conditioned (wet) samples of F1 foam
were obtained using a Keyence Microscope (Fig. 6). The
micrographs show the distribution of pores and some pore-
connectivity at the surface of the foam. Of all the three
foams considered in this study, F1 was chosen for the 3D
surface analysis as it had intermediate permeability values
and the highest density of open cells.
A couple of important conclusions can be drawn from this
investigation: (a) it is clear from Fig. 6 that wetting the foam
with water leads to the disappearance of a large percentage of
pores due to swelling of the matrix material. This observation
may explain the low permeability values seen in the foams;
(b) the open-cell polyurethane foams have shown to be more
permeable due to their open mesh-like skeletal structure
which contains a high percentage of void space [31]. Hence,
one can attribute the low permeability observed in the func-
tionalized foams to its skeletal structure which does not
contain high void space as seen from the 3D micrographs.
4 Summary and conclusions
Statistical analysis of the micrographs of cross-sections
using normal distribution curves has proved to be a good
method to characterize the structure and cell size
757
distribution of functionalized open cell polyurethane
foams. Several models that predict the permeability of
open-cell foams are available; however, most of them
either focus on air flow through the foams or are based on
reticulated open-cell foams. The chain-extended poly-
urethane foams discussed in this study are not reticulated
but have an open-cell content ranging between 70 and
90 % based on the foam composition. Due to the fact that
the experimentally-measured permeability is two orders-
of-magnitude smaller than the theoretically-predicted
permeability, none of the available theoretical perme-
ability models (including the particulate porous-media
models by Kozeny-Carman and Duplessis-Masliyah) can
be directly applied to predict the permeability of these
foams. The multiplication of the open cell percentage
with the overall porosity results in effective porosity,
which is the ratio of interconnected pores (that permit
liquid flow through them on the application of an external
pressure) to the total volume. Use of the effective porosity
instead of the overall porosity not only brings the theo-
retical models for permeability closer to the experimental
value, but also brings in the same trend as seen in the
experiments. This clearly means that the effective poros-
ity is the right porosity for estimating the foam
permeability.
The low permeability of these foams indicates a high
pressure drop for any practical water/liquid flow rate. Such
high pressure drops will result in high stresses in the solid
‘skeleton’ of the matrix, thus leading to a possibility of
deformation and compression of the tested polyurethane
foams during their proposed use as water filter. Also, the
two orders-of-magnitude difference between the theoretical
and experimental permeability points to a strong need to
develop better permeability models by (a) developing 3-D
network models to incorporate the large pressure drops
anticipated in smaller pores connecting bigger cells, and by
123
 Author's personal copy
758
(b) incorporating the effect of swelling of the solid (poly-
urethane) matrix when exposed to the flowing liquid
(water).
Appendix
Mercury porosimetry was attempted to estimate the por-
osity of the foams. In this method, the sample is dipped in
a pool of mercury and the latter is gradually pressurized.
As a result, the non-wetting mercury is progressively
forced into the pores of the porous (sponge) sample with
larger pores getting filled up before the smaller ones. The
total volume of mercury thus pushed into the porous
medium equals the total pore volume. The porosity is
estimated by dividing the pore volume by the sample
volume. Due to the relation between mercury pressure
and pore size, it is also possible to estimate the pore size
distribution in the sample.
The porosity measured by mercury porosimetry showed
that the porosity of F0 foam sample was 65.3 % while the
pore size varied from 100 to 20 lm. Samples F1 and F2
returned considerably lower porosities values of 13.4 and
10.2 %, respectively. All of the mercury test samples were of
similar size and geometry; visual examination suggested
similar porosity levels. A mercury intrusion test should result
in a penetrometer stem volume change between 33 and 90 %
[32]. Samples F1 and F2 showed a penetrometer stem vol-
ume change of 2 and 3 % respectively indicating that mer-
cury did not enter the samples as pressure was applied during
the test. The mercury test assumes that mercury is non-
wetting on the sample surface; such that upon initial filling of
the sample penetrometer, the mercury comes into contact
with the sample but does not penetrate into the structure. The
low porosity numbers obtained from the F1 and F2 samples
suggest that the functionalized foam increases the wettability
of mercury by reacting with its sulphonic groups [33], thus
making the mercury porosimetry method unsuitable for
porosity measurements with these samples.
References
1. R.O. Ebewele, Polymer Science and Technology (CRC Press,
Technology and Engineering, Boca Raton, 2000), pp. 3–4
2. A.H. Landrock, Handbook of Plastic Foams Types, Properties,
Manufacture and Applications (Elsevier, Technology & Engi-
neering, Park Ridge, 1995), pp. 1–2
3. S.T. Lee, N.S. Ramesh, Polymeric Foams: Mechanisms and
Materials (CRC Press, Technology & Engineering, Boca Raton,
2004), pp. 1–4
4. S.K. Annapragada, Mechanism of Foaming on Polymer-Paper-
board Composites, PHD Thesis (Georgia Institute of Technology,
Atlanta, 2007)
123
J Porous Mater (2015) 22:749–759
5. R.W. Dyson, Engineering Polymers, Springer, Science, ch 4,
Polymer Foams, 1990, pp. 81–85
6. M.A. Dawson, J.T. Germaine, L.J. Gibson, Permeability of open-
cell foams under compressive strain. Int. J. Solids Struct. 44(16),
5133–5145 (2007)
7. R. McBryer, Polyurethane Foams Formulation and Manufacture
(Technomic Publishing, Lancaster, 1999), p. 347
8. S. Gunashekar, N. Abu-Zahra, Synthesis of functionalized poly-
urethane foam using BES chain extender for lead ion removal
from aqueous solutions. J. Cell. Plast. 1–18 (2014). doi:10.1177/
0021955X14559255
9. S. Gunashekar, N. Abu-Zahra, Characterization of functionalized
polyurethane foam for lead ion removal from water. Int. J. Polym.
Sci. (2014). doi:10.1155/2014/570309
10. D. Edouard, M. Lacroix, C. Pham, M. Mbodji, C. Pham-Huu,
Experimental measurements and multi-phase flow models in solid
SiC foam beds. AIChE J. 54(11), 2823–2832 (2008)
11. T.T. Huu, M. Lacroix, C.P. Huu, S.D. Edouard, Towards a more
realistic modeling of solid foam: use of pentagonal dodecahedron
geometry. Chem. Eng. Sci. 64(24), 5131–5142 (2009)
12. P. Pichat, Photocatalysis and Water Purification: From Funda-
mentals to Recent Applications, John Wiley & Sons—Tech-
nology & Engineering, Hoboken 2013, part II, ch-6
13. H. Darcy, Les Fontaines Publiques de ville de Dijon (Dalmont,
Paris, 1856)
14. K. Boomsma, D. Poulikakos, The effects of compression and
pore size variations on the liquid flow characteristics in Metal
foams. ASME J Fluids Eng. 124, 263–272 (2002)
15. N.C. Hilyard, P. Collier, A structural model for air flow in flex-
ible PUR foams. Cell. Polym. 6(6), 9–26 (1987)
16. N.J. Mills, Polymer Foams Handbook: Engineering and Biome-
chanics Applications and Design Guide, Butterworth-Heinemann;
2007; pp 177–179, 483–485
17. J. Ahmed, C. Pham-Huu, D. Edouard, A predictive model based
on tortuosity for pressure drop estimation in ‘slim’ and ‘fat’
foams. Chem. Eng. Sci. 66, 4771–4779 (2011)
18. N.J. Mills, The wet kelvin model for air flow through open-cell
polyurethane foams. J. Mater. Sci. 40, 5845–5851 (2005)
19. A.N. Gent, K.C. Rusch, Permeability of open-cell foamed ma-
terials. J. Cell. Plast. 2(1), 46–51 (1966)
20. B.V. Antohe, J.L. Lage, D.C. Price, R.M. Weber, Experimental
determination of permeability and inertia coefficients of me-
chanically compressed aluminum porous matrices. ASME J
Fluids Eng. 119, 404–412 (1997)
21. M. Chanda, S.K. Roy, Plastics Technology Handbook, 3rd Edn,
(CRC Press, Boca Raton 1998); Ch 2, pp 279–285
22. ASTM D 3574—Standard Test Methods for Flexible Cellular
Materials—Slab, Bonded, and Molded Urethane Foams, ASTM
Vol. 08.02 Plastics (II): D3222 D5083
23. J. Bear, Dynamics of Fluids in Porous Media (American Elsevier,
New York, 1972)
24. T.W. Lambe, Soil Mechanics (Hoboken, John Wiley & Sons,
Technology & Engineering, 1969), p. 281
25. N.M.K. Lamba, K.A. Woodhouse, S.L. Cooper, Polyurethanes in
Biomedical Applications (CRC Press, Medical, Boca Raton,
1997), pp. 74–79
26. D.C. Allport, D.S. Gilbert, S.M. Outterside, MDI & TDI: Safety,
Health and the Environment (Wiley, Hoboken, 2003), p. 298
27. R. Vignes, DMSO-A new clean, unique, superior solvent,
American Chemical society (2000), http://www.ac-limoges.fr/
physique-chimie/IMG/pdf/vignes-acs.pdf. Accessed Feb 2014
28. M. Szycher, Szycher’s Handbook of Polyurethanes (CRC Press,
Boca Raton, 1999), pp. 217–218
29. K. Ashida, Polyurethane and Related Foams: Chemistry and
Technology (CRC Press Technology & Engineering, Boca Raton,
2006), p. 79
 Author's personal copy
J Porous Mater (2015) 22:749–759
30. R. Masoodi, K.M. Pillai, Wicking in Porous Materials: Tradi-
tional and Modern Modeling Approaches (CRC Press Tech-
nology & Engineering, Boca Raton, 2012), pp. 80–81
31. Design News report: Open-pore Polyurethane foam and felt
(1964), http://scienceservice.si.edu/pages/090005b.htm. Accessed
July 2014
View publication stats
759
32. Micrometrics, Mercury Intrusion Porosimetry theory. http://www.
micromeritics.com/Repository/Files/Mercury_Porosimetry_Theory_
poster.pdf. Accessed Mar 2014
33. T. Braun, A.B. Farag, Gas-solid and liquid-liquid contact via a
new geometry for the solid phase. Anal. Chim. Acta 99, 1–36
(1978)
123