Journal of Colloid and Interface Science 581 (2021) 627–634

Contents lists available at ScienceDirect

Journal of Colloid and Interface Science

journal homepage: www.elsevier.com/locate/jcis

Elastic plastic fracture mechanics investigation of toughness of wet

colloidal particulate materials: Influence of saturation
George V. Franks a,⇑, Mitchell L. Sesso b, Matthew Lam a, Yi Lu a, Liqing Xu a
a
Department of Chemical Engineering, University of Melbourne, Parkville, Vic 3010, Australia
b
Department of Engineering, School of Engineering and Mathematical Sciences, College of Science, Health and Engineering, La Trobe University, Vic 3086, Australia

highlights
graphicalabstract
● Elastic Plastic Fracture
Mechanics toughness of wet
particulate materials.
● Toughness (JIC) of micron sized
particulate materials increases with
saturation.
● Plastic deformation constitutes more
than 99% of energy dissipated in
fracture.
● The elastic part of the fracture
toughness (KIC) also increases with
saturation.
● KIC contributes less than 1% of the
total energy of fracture.

articleinfo abstract
Article history:
Received 15 May 2020 Hypothesis: Previous use of linear elastic fracture mechanics to estimate toughness of wet particulate
Revised 16 July 2020 materials underestimates the toughness because it does not account for plastic deformation as a dissipa-
Accepted 29 July 2020 tion mechanism. Plastic deformation is responsible for the majority of energy dissipated during the frac-
Available online 2 August 2020 ture of wet colloidal particulate materials. Plastic deformation around the crack tip increases with
saturation of the particulate body. The toughness of the body increases with increasing saturation.
Keywords: Experiments: Elastic plastic fracture mechanics using the J-integral approach was used for the first time
Cracking to measure the fracture toughness (JIC) of wet micron sized alumina powder bodies as a function of
Fracture mechanics satura- tion. The samples were prepared by slip casting. The saturation was controlled by treatment in a
Wet particulate materials
humid- ity chamber. The elastic modulus (E) and the energy dissipated by plastic flow (A pl) were
Plastic deformation
measured in uniaxial compression. The critical stress intensity factor (K IC) was measured using a
Toughness
diametral compres- sion sample with a flaw of known size. The fracture toughness (J IC) was calculated
from these measured quantities and the geometry of the specimen.
Findings: Elastic plastic fracture mechanics was used for the first time to quantitively account for plastic
deformation of wet particulate materials. The linear elastic fracture mechanics approach previously used
accounted for less than 1% of the total energy dissipated in fracture. Toughness (JIC) was found to
increase with increasing saturation due to plastic deformation that increased with saturation level. The
improved understanding of toughness as a function of saturation will aid in providing quantitative
analysis of cracking in drying colloidal films and bodies.
© 2020 Elsevier Inc. All rights reserved.

⇑ Corresponding author
E-mail address: gvfranks@unimelb.edu.au (G.V. Franks).

https://doi.org/10.1016/j.jcis.2020.07.142
0021-9797/© 2020 Elsevier Inc. All rights reserved.
628 G.V. Franks et al. / Journal of Colloid and Interface Science 581 (2021) 627–634

1. Introduction
ter in two half molds with half cylindrical depressions. These two
halves were held together by rubber bands and placed upon a
Fracture of particulate materials is important in a range of nat-
plas- ter board. Modelling compound was used to secure the two
ural and industrial drying phenomena including mud cracking as
halves to the plaster board underneath. Stainless steel cylinders
well as drying of paint, coatings, ceramics and sewage sludges [1–
were placed on top of the plaster mold to provide a reservoir for
6]. Although drying cracking has been studied extensively [6– 23]
the slip. The set-up is shown in Fig. S1a in the supplementary
there is little information about the fracture toughness of wet
material. The mold produced slip cast samples of 1 cm diameter
particulate materials [23–38] and even less about the role of the
and 2 cm length. The samples were cast for 1 h. The top of the
degree of saturation on the fracture toughness. The previous work
stainless-steel cylin- ders was covered with cling wrap. After
has assumed linear elastic fracture mechanics although plas- ticity
casting, the top surface of the sample was cleaned up to produce
at the crack tip has been considered [16,23,38]. Table 1 pre- sents a
a flat surface by removing the excess material before demolding
timeline of the advances in understanding the role of plasticity and
the samples. Diametral com- pression samples with flaws for
saturation on the fracture of wet particulate materi- als. The
fracture toughness testing were slip cast by pouring the
current paper builds on this previous work and advances our
suspensions into aluminium disk-shaped molds with a diameter
previous work [38] where we investigated the critical stress
34.5 mm placed on plaster of Paris boards as shown in Fig. S1b
intensity factor (KIC, fracture toughness) via a linear elastic fracture
in the supplementary material. Flaws (length (2ao) = 10.35 mm)
mechanics (LEFM) approach on dry and nearly saturated particu-
were cast through the thickness of the speci- mens using
late bodies. Although the LEFM approach would normally apply
elongated diamond shaped aluminium inserts. The width of the
to brittle materials only, we applied a modification of the model
flaw at its widest point (at the centre of the disk) is
to account for non-linearities by assuming a process zone is pre-
1.25 mm. The amount of suspension cast was such that a consoli-
sent and implementing an iteration procedure to determine its size
dated disk thicknesses of approximately 8–10 mm resulted. The
[39]. In the current work, we use an elastic plastic fracture
casting time was 6 h. The mold was covered with cling wrap to pre-
mechanics (EPFM) approach, the J-integral, to determine the
vent water loss from the top surface of the casting. Further details
toughness (JIC) of particulate materials for the first time. This
are provided elsewhere [38]. The mechanical tests on the dry sam-
approach accounts for the energy dissipated by plastic deformation
ples were conducted after drying in an oven for 24 h at 80 °C. Sam-
around the crack tip. Previously, only linear elastic fracture
ples to be tested wet were placed in a humidity-controlled
mechanics was used which did not accurately account for the plas-
chamber as described below.
tic deformation of the material. The use of EPFM used in the pre-
The samples to be tested wet were stored in a humidity cham-
sent paper quantitively accounts for the plastic deformation
ber. A rectangular fish-tank was used as the chamber. Air was bub-
within the sample which as we will show, constitutes the majority
bled through water in a one-armed Erlenmeyer flask then into the
of energy dissipated in fracture of wet particulate materials. In the
chamber. The water was maintained at 50 to 70 °C. A humidity
EPFM approach, JIC is analogous to the critical strain energy release
probe within the chamber was used to control the pump. The
rate (G) in LEFM. In the previous work [38] we used an ultrasonic
pump switched on when the humidity dropped below the
technique to measure the elastic modulus of the dry and nearly
setpoint and switched off when above the setpoint. The
saturated powder bodies. The modulus values obtained from stress-
apparatus is shown in Fig. S2 in the supplementary material.
strain measurements are typically about an order of magni- tude
The set point was gener- ally either 50% or 90%. Samples left in
lower than those obtained from sound velocity measure- ments.
the chamber at 50% humidity reached a saturation level of
The reason for this difference is most likely due to the higher
between about 30 to 60% after 1 to 3 days. Samples left in the
frequency of the sonic measurements relative to the mechanical
chamber for 2 to 5 days at 90% humidity reached saturation
tests. The literature reports modulus for dry powder compacts via
levels between about 70 and 95%. The relative density and
ultrasound on order of 1 GPa [40], and wet powder compacts via
saturation for the samples were measured via the weight
compression on order of 100 MPa [41–43] similar to our results.
difference method, where the volume of water contained within
Due to this discrepancy, in the current work we measure the elastic
the wet bodies was determined by weighting the samples
modulus and the area under the stress–strain curve in uniaxial
after removal from the humidity chamber, then after drying for a
compression as a function of saturation. We mea- sure the critical
minimum of 24 hrs at 80 °C. Archimedes’ method using wax coat-
stress intensity factor (K IC) using diametral com- pression samples
ing of the powder compacts was used to determine the relative
with flaws of known size over a wide range of saturation values.
density of the dried cast cylinders as described in prior work
These measurements enable us to calculate the toughness (JIC) of
[48]. Using the relative density and the weight loss during drying,
the wet particulate material as a function of saturation.
it was possible to determine the fraction of saturation.

2. Experimental and calculations 2.2. Fracture toughness determination

2.1. Sample preparation JIC is determined as described in ASTM E1820-17a, Standard

Test Method for Measurement of Fracture Toughness [49]. Our
Aqueous suspensions were prepared from high purity alpha- approach is based on the assumption that one half of the diametral
alumina powder (AKP-15, Sumitomo Chemical Co. Ltd, Tokyo compression geometry used in our work is equivalent to one of the
compact tension specimens in ASTM E1820 section A2, as shown in
Fig. 1. The plane stress limit is valid for the disk geometry used.
JIC is given by Eq A2.4 to 2.6 in ASTM E1820 for plane stress as
Japan) with a mean particle diameter of 0.7 lm [45]. . Σ
2
K
2 gpl
Suspensions JIC ð1Þ
were prepared containing 0.25 vol fraction of solids (/), which I Apl E þ
¼ 1 C— m
were dispersed at pH 4.0 (±0.1) using analytical grade HNO3 fol- B Nb o
lowed by ultrasonic dispersion (Misonix Sonicators S-4000, New- plaster of Paris molds. The molds were produced by preparing plas-
town CT, USA) to create a well dispersed sample.
Cylindrical samples for uniaxial compression tests to determine
modulus and area under the stress strain curve were slip cast into
 The first term is the elastic component of J IC and the second
term is the plastic component of JIC. Our previous approach [38]
only considered the elastic component. The critical stress
intensity
factor (KIC) is determined from the diametral compression test
as described below. The Poisson’s ratio (m) is assumed to be
0.3. The
 G.V. Franks et al. / Journal of Colloid and Interface Science 581 (2021) 627–634 629

Table 1
Development of understanding of the role of plasticity and saturation in fracture toughness of wet particulate materials.

Authors and year Material Particle Solids Elastic Fracture Fracture Toughness Approach to investigating fracture of
size volume modulus Strength (MPa) wet particulate materials
% (MPa)

Zarzycki, [37] 1988 Silica gels Generally 23 0.05 to KIC 10-4 to 10-5 MPa.m—1/2 KIC measurement of wet silica colloidal
wet on order 10 gels using a single notched beam in
of 10 nm tension.
Kamiya et al, [44] 1995 Si3N4 dry 170 nm 35 to 50 0.2 to 1 Diametral compression of dry granules.
Strength only considered, not
toughness.
Franks and Lange, Al2O3 wet 230 to 48 to 60 5 to 10c Uniaxial compression of saturated
[28,45] 1996, 1999 590 nm powder compacts. Plastic to brittle
transition observed.
Balzer et al., [42] 1999 Al2O3 wet 500 to 55 to 61 10 to 175 0.02 to 0.25c Measurement of strength and modulus
Hesselbarth et al., 600 nm of saturated powder compacts in
[43] 2001 compression.
Carneim and Green, Al2O3 dry 50 to 70 400 to 0.01 to 0.8 Diametral compression test of dry
[40] 2001 2000 powder compacts for tensile strength.
Elastic modulus measurement of dry
powder compacts by ultrasound.
Prabhakaran et al., [41] Al2O3 wet 340 nm 53 1 to 100 0.08 to 0.47c Strength and modulus measurement of
2008 saturated powder compacts in
compression. No toughness
measurements.
Holmes et al., [16] 2008 Al2O3 wet 350 nm 30 Flow around the crack tip is important.
before Discusses ‘‘resistance to cracking” but
drying toughness not explicitly discussed.
Yu and Lange, [27] Al2O3 wet 250 nm 62 3.3 wet Simple strength measurements of dry
2010 and dry 1.6–1.9 dry and wet particulate materials
demonstrate that wet is stronger.
Kitsunezaki, [29–30] Theoretical Modelling the role of plasticity on
2010, 2011 modelling fracture of wet particulate materials
suggests plasticity acts to resist crack
growth, but no mechanical properties
measured.
Sarkar and Al2O3 wet 340 nm 53 0.4 to 1.7 Relates fracture stress to flaw size in
Tirumkudulu, [46] before drying colloidal films.
2012 drying
Goehring et al., [23] Poly 58 to 64 to 74 130 to KC/E = 0.11 (lm)½ Plasticity around the crack tip is
2013 styrene 99 nm 230 GC = 2 J/m2 important in fracture. Suggestion that
latex wet decreasing yield stress will increase
toughness. Estimated combination of
KIC and E using crack tip opening
approach. The energy dissipated in
fracture was more than 10 times the
energy dissipated elastically. Estimated
the critical strain energy release rate
(G) but did not measure toughness
independently.
Birk-Braun et al., [47] Silica and Silica 7 to 10 KC/E 0.05 to 0.15 (lm)½ Estimate of combination of toughness
2017 poly 22 nm before KC 0.01 to 0.3 MPa.m—1/2 (KIC) and modulus using crack opening
styrene Poly drying displacement. Modulus of wet powder
latex wet styrene compact estimated from theory.
latex 100 Investigated role of particle size.
to 450 nm
Sesso and Franks, [38] Al2O3 wet 700 nm 67 to 69 5700 to 0.19 wet to 0.04 KIC 0.015 to 0.035 MPa.m—1/ Independent measurement of
2017 and dry 7100 wet dry (note with 2
wet toughness (KIC) and modulus,
2900 to macroscopic KIC 0.003 to 0.006 MPa.m—1/ quantitative consideration of the
5000 dry flaws) 2
dry plastic zone to calculate the critical
GIC 0.1 to 0.17 J.m—2 wet strain energy release rate (GIC) using
GIC 0.005 to 0.01 J.m—2 dry Linear Elastic Fracture Mechanics.
Investigated dry and nearly saturated
conditions. Modulus measured by
ultrasound.
Current paper Al2O3 wet 700 nm 59 to 72 150 wet 0.11 wet (>50% KIC 0.03 MPa.m—1/2 wet KIC First application of Elastic Plastic
and dry to 234 sat) to 0.03 dry 0.0038 MPa.m—1/2 dry Fracture Mechanics using the J-integral
dry (note with JIC 1000 to 2000 J/m2 wet to quantitively determine toughness
macroscopic (>50% sat (JIC). Investigation of nearly the full
flaws) JIC 60 to 100 J/m2 dry range of saturation from 0% to 98%.

Missing values in the Table were not reported in the paper.

c denotes strength in compression rather than tension.

630 G.V. Franks et al. / Journal of Colloid and Interface Science 581 (2021) 627–634
Ic f I pffi ffiffiffi ffiffioffiffi
K ¼r F p a ð3Þ

Fig. 1. Schematic illustration showing the correspondence between dimensions of
a) the compact tension specimen in ASTM E1820 [49] and b) the diametral
compression specimen used in this work where R is equivalent to W. Both the
tensile load (F) on the compact tension specimen and the compressive load (P) on
the diametral compression specimen produce a tensile stress ( rt) applied to the
flaw tip.
modulus (E) is determined from the uniaxial compression test as
described below. The factors B N, bo and gpl, are determined from
the geometry. BN is thickness of the disk which is usually about 8
to 10 mm. bo is the uncracked ligament (R – a o). R (the radius of
the diametral compression sample) is assumed to be equal to W
in ASTM E1820 (see Fig. 1). ao is half the flaw length. In this work
(bo = R - ao = 17.25 – 5.175 = 12.075 mm). According to ASTM
E1820, gpl ¼ 2 þ 0:522b
R
o
for this geometry. The area under the load
displacement curve (A pl) is determined from the uniaxial compres-
sion test as described below. A pl is a measure of the energy dissi-
pated by plastic deformation.
Uniaxial compression tests were conducted using the Instron ®
5848 MicroTesterTM with a 2 kN loadcell (Instron – Illinois Tool
Works Inc., Norwood MA, USA). Compression was between flat
stainless-steel platens at a rate of 0.1 mm/min for samples with
greater than 20% saturation and at 0.05 mm/min for samples
with less than 20% saturation. The load and displacement were
con- verted to stress and strain using the dimensions of the
cylindrical sample, 1 cm diameter by 2 cm long. A total of 45
samples with saturation varying between 0% and 86%
saturation were measured. The elastic modulus is taken as the
steepest slope of the loading curve; ignoring the apparent low
slope at the start of the curve due to non-parallel sample
flattening. The area under the load–dis- placement curve (Apl)
was determined by measurement of the area under the stress–
strain curve using the trapezoidal rule and the cylinder diameter
and length.
Diametral compression was carried out on an Instron ® 5848
MicroTesterTM with a 2 kN loadcell (Instron – Illinois Tool Works
Inc., Norwood MA, USA). The disk was loaded in compression
along its diameter parallel to the long axis of the flaw to generate a
ten- sile stress within the sample perpendicular to the loading
direc- tion. At the centre of the disk, the stress is at a maximum;
assuming line contact load, the magnitude of the elastic stress
derived from theory by Hertz are; [50]
2P
rx ¼ pD ð2Þ
t ther investigation but is beyond the scope of this work.
where rx is the stress in the direction perpendicular to the
loading direction, P is the compressive load, D is the diameter
and t the disk thickness. The fracture stress (rf) is taken as the
tensile stress in the direction perpendicular to the loading
direction when the crack propagates from the flaw tip. 21
samples were tested with a range of saturations from 0 to 98%
saturation.
The critical stress intensity factor (K IC) is determined for a pure
mode I type loading condition as follows;
where rf is the failure stress, a o is half the flaw length (5.175 mm),
and FI is the non-dimensional shape coefficient for mode I loading
which is FI = 1.387 for ao/R = 0.3 as in the present geometry [51–
52].
JIC can then be calculated from Eq. (1) using the geometry of the
diametral compression sample and the measured mechanical
properties. The analysis is based on the assumption that the mate-
rial is homogeneous and isotropic and that it is symmetric in com-
pressive and tensile stress–strain behaviour. A particulate material
comprised of colloidal sized particles produced by slip casing from
a well dispersed suspension is expected to be homogeneous on a
macroscopic length scale relevant to the crack extension (many
orders of magnitude larger than the particle size). Objects cast
from well dispersed slurries without additional pressure are
known to be isotropic [53].
3. Results and discussion
3.1. Influence of saturation on stress–strain behaviour
Typical stress–strain curves measured in uniaxial compression
for three different saturation values are shown in Fig. 2. The max-
imum slope of the loading curve (elastic modulus) decreases with
increasing saturation. The point of fracture occurs at the highest
stress shown on the figure for each sample. The unloading modulus
is taken as the same as the initial loading modulus. In many cases
there was no clear yield point; instead, often the rate of increase of
the stress with increasing strain just decreased. Such behaviour is
consistent with micromechanical modelling which suggests col-
loidal particulate bodies yield at very low strains typically on order
of 0.1% [54,55]. The area under the stress-strain curve increases
dramatically with increasing saturation. The modulus of the 45
samples tested are plotted as a function of saturation in Fig. 3.
There is some scattering in the data which appears more signifi-
cant at lower saturations. The scatter is possibly due to slight vari-
ations in the non-uniformity in saturation throughout the sample.
Although the green density of the samples ranges from 2.33 to
2.58 g/cc with average 2.47 g/cc (62% theoretical density (TD)),
there appears to be no correlation between density and modulus
when examined over narrow saturation ranges. (See Fig. S3 in
the supplementary materials.) The data is best fit to a line with
equation (E = —0.844 (% sat) + 233.7) where E is in MPa as shown
as the dashed line in Fig. 3. The modulus obtained from the com-
pressive stress–strain curves are likely to be different from those
measured in our previous work that applied the acoustic method,
[38] this is likely the effect of anisotropy, the inherent difficulty
in measuring saturated materials with the acoustic approach and
the influence of frequencyof the measurement which warrants fur-
Fig. 4 shows the value of A pl (area under the load–displacement
curve as defined in ASTM E1820 Fig. A1.2 using the Basic Test
Method) for the 45 individual uniaxial compression tests. The area
under the stress strain curves was determined using the trape-
zoidal rule, then the area under the unloading curve was sub-
tracted, to obtain the ‘‘toughness” of the sample in MPa. This
area was converted to A pl by multiplying the ‘‘toughness” by the
sample volume to obtain Apl in Pa.m3 (equivalent to N.m). Again,
there is a fair bit of scatter in the data, for these measurements
the data at higher saturation have more scatter. The scatter is likely
due to non-uniform distribution of moisture in the sample and not
due to density variation as shown in Fig. S4 in the supplementary
information. The data is best fit to a line with equation (Apl = 7.8
5 ×10-4 (% sat) + 0.0039) where A pl is in Pa.m3 as shown as the
dashed line in Fig. 4.
 G.V. Franks et al. / Journal of Colloid and Interface Science 581 (2021) 627–634
Fig. 2. Typical stress-strain curves in uniaxial compression for 0% (black solid
line), 68% (red dashed line) and 82% (blue dotted line) saturation. The elastic
modulus decreases as a function of saturation and the area under the stress-strain
curve increases as a function of saturation. (For interpretation of the references to
colour in this figure legend, the reader is referred to the web version of this
article.)
Fig. 3. Plot of elastic modulus from compression experiments verses saturation.
The dashed line is the best linear fit to the data (E = — 0.844 (% sat) + 233.7) with E
in MPa. The quality of the fit is indicated in the inset in the figure. The average green
density of the dried samples was 2.47 g/cc = 62% TD ranging from 2.33 to 2.58 g/cc.
3.2. Influence of saturation on elastic component of toughness (KIC)
The typical load–displacement behaviour for diametral com-
pression samples containing flaws is shown in Fig. 5 for a dry
and highly saturated sample. The crack extends from the flaw at
the highest load shown in the figure. This fracture stress is calcu-
lated from the fracture load and sample geometry using Eq. (2).
Both the fracture load (stress) and displacement (strain) generally
increase with increasing saturation. In all cases the load due to the
self-weight of the sample (0.29 N) was less than 4.2% of the frac-
ture loads which ranged from 6.95 N to over 160 N. The critical
stress intensity factor (KIC) was calculated using Eq. (3) for each
of the 21 samples tested. The plot of K IC as a function of saturation is
shown in Fig. 6. The data is fit to the equation (K IC = (0.0039)*(%
sat)0.41 + 0.0038) where KIC is in MPa.m0.5 as shown as the solid line
631
Fig. 4. Plot of Apl verses saturation. The dashed line is the best linear fit to the
data (Apl = 7.85 × 10-4 (% sat) + 0.0039) with Apl in Pa.m3. The quality of the fit is
indicated in the inset in the figure. The average green density of the dried samples
was 2.47 g/cc = 62% TD ranging from 2.33 to 2.58 g/cc.
in Fig. 6. The data seems to have less scatter than the uniaxial com-
pression results for reasons unknown. The average green density of
the samples was 2.64 g/cc = 66.5% TD ranging from 2.43 to 2.86 g/
cc. These density values are slightly higher than the densities of the
samples tested in uniaxial compression. The slight difference prob-
ably due to the difference in the slip casting mold geometry. The
KIC values increase dramatically with increasing saturation as sat-
uration increases from zero to about 10 or 15%, then increases
more gradually up to 100% saturation.
Sarkar and Tirumkudulu [56] presented an asymptotic
analysis of the deformation field near a crack tip under mode-I
loading con- ditions. The analysis also resulted in predictions of
both an effec- tive modulus and KIC for a saturated colloidal
packing, of which a brief summary is provided in the
supplementary material. The effective modulus (Eeff) was shown
to be a function of the shear modulus of the solid phase (G),
coordination number of particles (M), random close packing
concentration (/rcp) , principle stiffness component (Q —1) and the
11
characteristic strain (s0 ) in the packing as shown in the
supplementary material. The stress intensity factor (K IC) was
also determined from an expression presented by the same
authors using the effective modulus (Eeff) and the surface
tension of water (c = 0.072 N/m). The values of shear modulus
(G) of 156 GPa, coordination number (M) of 6, random close
pack- ing concentration (/rcp) of 0.64 for dispersed Al2O3
suspensions previously reported in the literature [57] were
used to calculate the effective modulus and resulting KIC.
It is important to note that the calculation of the effective mod-
ulus is a function of the characteristic strain in the sample. In our
calculations, we have used the elastic strain limit to determine
the predicted effective modulus (Eeff) which for the nearly fully sat-
urated samples (for example as shown for the 82% saturation sam-
ple in Fig. 2) is about s0 = 0.004. In the calculation of the critical
stress intensity factor (KIC) the total strain in the body when frac-
ture initiates should be used as the characteristic strain which for
the nearly fully saturated samples (for example as shown for the
82% saturation sample in Fig. 2) is about s0 = 0.035.
Table 2 compares the modulus and stress intensity factor from
the predictions based on the analysis from Sarkar and Tirumkudulu
[56] with those determined from experimental techniques used in
this work for the nearly fully saturated case. The predicted modu-
lus and stress intensity factor are in a similar range with the mea-
sured values presented in this work.
632 G.V. Franks et al. / Journal of Colloid and Interface Science 581 (2021) 627–634
Fig. 5. Typical load–displacement curves for diametral compression tests for
ao/R = 0.3 and R = 34.5 mm. The dry sample was 7 mm thick and the wet sample
was 10 mm thick. The solid black line is for 0% saturation and the red dotted line
is for the 88% saturation. Both the fracture load and displacement increase with
increasing saturation. (For interpretation of the references to colour in this figure
legend, the reader is referred to the web version of this article.)
Fig. 6. Plot of KIC verses saturation. The solid line is drawn to guide the eye. The
solid line is the fit of the data to (KIC = (0.0039)*(% sat)0.41 + 0.0038) where KIC is
in MPa.m0.5. The quality of the fit is indicated in the inset in the figure. The
average green density of the dried samples was 2.64 g/cc = 66.5% TD ranging
from 2.43 to
2.86 g/cc.
3.3. Influence of saturation on toughness (JIC)
JIC was then calculated for the 21 diametral compression sam-
ples using their individual KIC values, their geometry and the
mea- sured values of E and Apl according to Eq. (1). Since E and
Apl were measured on different samples than the diametral
compression samples there is no 1 to 1 correspondence between
the saturation in the diametral compression and uniaxial
compression samples. In order to determine the modulus and
Apl to use for each JIC calcu- lation, the values of E and A pl were
determined from the linear equations of best fit to the E and Apl
data at the saturation corre- sponding to the actual saturation of
the individual diametral com- pression sample. The results of JIC
as a function of saturation for these samples is shown in Fig. 7 as
the solid black circles. The toughness (JIC) increases dramatically
with increasing saturation.
In all cases the second term in Eq. (1) (plastic component of JIC)
accounts for more than 99% of J IC. Surprisingly, this is true for even
the dry samples presumably because either micro-cracking or
some local particle rearrangement [54–55] occurs prior to failure
dissipating significant energy. The trend in J IC is linear with respect
to saturation. This is because A pl is the major contributor to J IC in
the calculation and the data fitting for A pl is a line with respect to
saturation. The black short-dashed line is the best linear fit to the
data (JIC = 21.06 (% sat) + 81) with J IC in Pa.m (equivalent to J/m2).
The results of Sesso and Franks [38] were also recalculated using
the improved estimate of the modulus and the value of Apl obtained
in the current work. In this earlier work [38] the flaw size to
specimen diameter was varied from 0.1 to 0.5 rather than fixed at
0.3 as in the current work. The results for flaw sizes ao/R = 0.1,
0.3 and 0.5 measured previously [38] are also visible on Fig. 7 as
red open diamonds for ao/R = 0.1, blue open circles for a o/R = 0.3.
and green open squares for ao/R = 0.5. The data from Sesso and
Franks [38] for ao/R = 0.3 falls only slightly below the black short-
dashed line indicating the fit to the a o/R = 0.3 data calculated in the
current work. Note, the result from ref [38] at ao/R = 0.3 at the
highest saturation is not displayed because it is anomalously low.
The black solid line in Fig. 7 shows the result when a o/R is
extrapolated to zero, which represents the situation when the only
flaws present are inherent in the particle packing. This may be
assumed to be the pristine fracture toughness of the material as
a function of saturation represented as the equation (J IC = 12.97
(% sat) + 60) with JIC in Pa.m. The results are consistent with the
previous work that suggests that wet particulate materials are
stronger [27] and tougher [38] than their dry counterpart. It is
now clear that this increased strength and toughness comes from
the greater degree of plastic deformation observed as saturation
increases. Understanding the influence of saturation on fracture
toughness of wet particulate materials may aid in understanding
when and where cracks occur during drying of suspensions.
4. Conclusions
The fracture toughness (JIC) of micron sized alumina powder
bodies has been measured using elastic plastic fracture
mechanics using the J-integral approach for the first time. Table
1 presents a timeline of the literature which brought the field
from simple frac- ture strength measurements of dry particulate
materials to quanti- tative consideration of the role of plasticity
in the fracture of wet particulate materials presented in the
current paper. Goehring et al., [23] significantly advanced the
field in 2013 when they dis- covered that the plastic component
of the toughness of saturated particulate materials comprised
over 90% of the energy dissipated, although they did not
measure KIC independently. Later, our linear elastic fracture
mechanics analysis of wet particulate materials
[38] quantitively accounted for the plastic deformation zone in
independent measurement of KIC and G.
The approach presented here is based on the J-integral elastic
plastic fracture mechanics which is more appropriate for materials
exhibiting significant plastic deformation. [49] JIC was found to
increase with increasing saturation over the entire range of satura-
tion measured. The data presented in the current work range from
0% to 98% saturation, filling the gap (between 0 and 88%
saturation) left by the previous work [38] and extending the high
saturation range to nearly 100%. The major component of J IC is the
plastic dis- sipation term in all cases consistent with Goehring et al.,
[23]. The plastic dissipation term is responsible for over 99% of the
energy dissipated in fracture. The energy dissipated in plastic flow
(Apl, related to the area under the stress–strain curve) increases
dramat- ically with saturation and is primarily responsible for the
increase of JIC with increasing saturation. K IC also increases with
saturation,
 G.V. Franks et al. / Journal of Colloid and Interface Science 581 (2021) 627–634
Table 2
Comparison of the experimentally measured values of modulus (E) and stress intensity factor (KIC) with those calculated form theory [56].
Model [56]
eo Eeff KIC
[MPa] [MPa·m½]
0.004 175 N/A
0.035 N/A 0.034
Fig. 7. Plot of JIC verses saturation. The black closed circles are the results from the
21 tests performed in the current work. The black short-dashed line is the best
linear fit to the data from the 21 tests with ao/R = 0.3 (JIC = 21.06 (% sat) + 81) with JIC
in Pa.m. The red long-dashed line is the fit for the a o/R = 0.1 data from ref [38], the
blue dash-dot line is the fit for the a o/R = 0.3 data from ref [38] and the green dotted
line is the fit for the data for the a o/R = 0.5 data from ref [38]. The data from ref [38]
for ao/R = 0.3 lies only slightly below the line for the results presented in the current
paper. The black solid line shows the result when the data is extrapolated to a o/R
approaching 0, which can be represented by the equation (J IC = 12.97 (% sat) + 60)
with JIC in Pa.m. (For interpretation of the references to colour in this figure legend,
the reader is referred to the web version of this article.)
but this elastic component is only a minor contributor (<1%) to the
energy dissipated in fracture of wet particulate materials of col-
loidal sized particles.
The next logical steps in understanding the fracture behavior of
wet particulate materials would be to investigate the influence of
solids concentration (relative density of the powder compact)
and investigate the role of changing interparticle potential from
repulsive to attractive by changing the pH relative to the isoelectric
point. Also, comparison of different approaches to measure the
elastic modulus to develop consensus on the most appropriate
approach should be conducted with particular attention to the role
of the frequency (rate) of the deformation. The new knowledge will
aid in understanding and quantitively predicting fracture during
drying colloidal particulate materials such as in mud cracking, a
topic of significant interest [6–23].
CRediT authorship contribution statement
George V. Franks: Conceptualization, Formal analysis,
Funding acquisition, Supervision, Writing - original draft. Mitchell
L. Sesso: Conceptualization, Formal analysis, Supervision,
Writing - review & editing. Matthew Lam: Investigation,
Methodology, Writing - review & editing. Yi Lu: Formal analysis,
Investigation, Methodol- ogy, Writing - review & editing. Liqing
Xu: Investigation, Method- ology, Writing - review & editing.
633
Measured
E KIC
[MPa] [MPa·m½]
149 N/A
N/A 0.030
Declaration of Competing Interest
The authors declare that they have no known competing finan-
cial interests or personal relationships that could have appeared
to influence the work reported in this paper.
Acknowledgements
Thanks to Laura Jukes for help with suspension preparation and
slip casting. Thanks to Mahesh Tirumkudulu, IIT Mumbai, for help
with the predictions of modulus and toughness from theory.
Thanks to the Australian Research Council for funding Discovery
Project DP150102788.
Appendix A. Supplementary data
Supplementary data to this article can be found online at
https://doi.org/10.1016/j.jcis.2020.07.142.
References
[1] L. Goehring, R. Conroy, A. Akhter, W.J. Clegg, A.F. Routh, Evolution of mud-
crack patterns during repeated drying cycles, Soft Matter 6 (2010) 3562–3567.
[2] L. Goehring, S.W. Morris, Cracking mud, freezing dirt, and breaking rocks, Phys.
Today 67 (11) (2014) 39–44.
[3] K.B. Singh, M.S. Tirumkudulu, Cracking in drying colloidal films, Phys. Rev. Lett 98
(2007) 218302.
[4] Y. Itaya, M. Hasatani, R&D Needs-Drying of Ceramics, Drying Technology 14 (6) (1996)
1301–1313.
[5] W. Weiyun, L. Aimin, Z. Xiomin, Y. Yulei, Multifractality analysis of crack images
form indirect thermal drying of thin-film dewatered sludge, Physica A 390 (2011) 2678–
2685.
[6] R.C. Chiu, T.J. Garino, M.J. Cima, Drying of Granular Ceramic Films: I, Effect of
Processing Variables on Cracking Behavior, J. Am. Ceram. Soc. 76 (1993) 2257–
2264.
[7] R.C. Chiu, M.J. Cima, Drying of Granular Ceramic Films: II, Drying Stress and
Saturation Uniformity, J. Am. Ceram. Soc. 76 (1993) 2769–2777.
[8] L.A. Brown, C.F. Zukoski, L.R. White, Consolidation during drying of aggregated
suspensions, AIChE J. 48 (2002) 492–502.
[9] W.P. Lee, A.F. Routh, Why do drying films crack?, Langmuir 20 (23) (2004)
9885–9888.
[10] G.W. Scherer, Theory of drying, J. Am. Ceram. Soc. 73 (1990) 3–14.
[11] M.S. Tirumkudulu, W.B. Russel, Role of Capillary Stresses in Film Formation,
Langmuir 20 (2004) 2947–2961.
[12] P. Wedin, C.J. Martinez, J.A. Lewis, J. Daicic, L. Bergström, Stress development
during drying of calcium carbonate suspensions containing carboxymethylcellulose
and latex particles, J. Colloid Interface Sci. 272 (2004) 1–9.
[13] P. Wedin, J.A. Lewis, L. Bergstorom, Soluble organic additive effects on stress
development during drying of calcium carbonate suspensions, J. Colloid
Interface Sci. 290 (2005) 134–144.
[14] E.R. Dufresne, E.I. Corwin, N.A. Greenblatt, J. Ashmore, D.Y. Wang, A.D. Dinsmore, J.X.
Cheng, X.S. Xie, J.W. Hutchinson, D.A. Weitz, Flow and Fracture in Drying
Nanoparticle Suspensions, Phys. Rev. Lett. 91 (22) (2003) 224501.
[15] E.R. Dufresne, D.J. Stark, N.A. Greenblatt, J.X. Cheng, J.W. Hutchinson, L. Mahadevan,
D.A. Weitz, Dynamics of Fracture in Drying Suspensions, Langmuir 22 (2006)
7144–7147.
[16] D.M. Holmes, F. Tegeler, W.J. Clegg, Stresses and strains in colloidal films during
lateral drying, J. Europ. Ceram. Soc. 28 (2008) 1381–1387.
[17] A. Nishimoto, T. Mizuguchi, S. Kitsunezaki, Numerical study of drying process
and columnar fracture process in granule-water mixtures, Phys. Rev. E 76 (2007)
016102.
[18] W. Man, W.B. Russel, Direct measurements of critical stresses and cracking in thin
films of colloid dispersions, Phys. Rev. Lett 100 (2008) 198302.
[19] H.N. Yow, I. Beristain, M. Goikoetxea, M.J. Barandiaran, A.F. Routh, Evolving
stresses in latex films as a function of temperature, Langmuir 26 (2010) 6335– 6342.
634 G.V. Franks et al. / Journal of Colloid and Interface Science 581 (2021) 627–634

[20] L. Goehring, J. Li, P.-C. Kiatkirakajorn, Drying paint: from micro-scale dynamics K. Prabhakaran, S.P. Tambe, A. Melkeri, N.M. Gokhale, S.C. Sharma, Mechanical properties
to mechanical instabilities, Phil. Trans. R. Soc. A 375 (2017) 20160161. of wet-coagulated alumina bodies prepared by direct coagulation casting using a MgO
[21] A. Schmeink, L. Goehring, A. Hemmerle, Fracture of a model cohesive granular coagulating agent, J. Am. Ceram. Soc. 91 (2008) 3608– 3612.
[42] B. Balzer, M.K.M. Hruschka, L.J. Gauckler, Coagulation Kinetics and Mechanical
material, Soft Matter 13 (2017) 1040–1047.
Behavior of Wet Alumina Green Bodies Produced via DCC, J. Colloid Interface
[22] H. Lama, R. Mondal, M.G. Basavaraj, D.K. Satapathy, Cracks in dried deposits of
Sci. 216 (1999) 379–386.
hematite ellipsoids: Interplay between magnetic and hydrodynamic torques, J. Colloid
[43] D. Hesselbarth, E. Tervoort, C. Urban, L.J. Gauckler, Mechanical properties of coagulated wet
Interface Sci. 510 (2018) 172–180.
particle networks with alkali-swellable thickeners, J. Am. Ceram. Soc. 84[8] (2001)
[23] L. Goehring, W.J. Clegg, A.F. Routh, Plasticity and fracture in drying colloidal
1689–1695.
films, Phys. Rev. Let. 110 (2013) 024301.
[44] H. Kamiya, K. Isomura, G. Jimbo, T. Jun-ichiro, Powder processing for the fabrication of
[24] E.M. Rabinovich, C.R. Kurkjian, N.A. Kopylov, D.A. Fleming, Mechanical
Si3N4 Ceramics: I, Influence of spray-dried granule strength on pore size
strength of participate silica gels, J. Mater. Sci. 26 (1991) 6685–6692.
distribution in green compacts, J. Am. Ceram. Soc. 78[1] (1995) 49– 57.
[25] J.R. Nichols, M.E. Grismer, Measurement of fracture mechanics parameters in silty-clay
[45] G.V. Franks, F.F. Lange, Mechanical behavior of saturated, consolidated,
soils, Soil Sci. 162 (5) (1997) 309–322.
alumina powder compacts: effect of particle size and morphology on the plastic-to-
[26] D.G. Bika, M. Gentzler, J.N. Michaels, Mechanical properties of
brittle transition, Colloids Surf. A: Physicochem. Eng. Asp. 146 (1999) 5–17.
agglomerates, Powder Technol. 117 (2001) 98–112.
[46] A. Sarkar, M.S. Tirumkudulu, Ultimate Strength of a colloidal packing, Soft Matter 8
[27] B. Yu, F.F. Lange, Shape forming alumina slurries via Colloidal IsoPressing, J.
(2012) 303–306.
Europ. Ceram. Soc. 30 (2010) 2795–2803.
[47] N. Birk-Braun, K. Yunus, E. Rees, W. Schabel, A. Routh, Generation of strength
[28] G.V. Franks, F.F. Lange, Plastic-to-brittle transition of saturated, alumina
in a drying film: How fracture toughness depends on dispersion properties,
powder compacts, J. Amer. Ceram. Soc. 79 (12) (1996) 3161–3168.
Phys. Rev. E 95 (2017) 022610.
[29] S. Kitsunezaki, Crack growth in drying paste, Adv. Powder Technol. 22 (2011) 311–
[48] C. Chuanuwatanakul, C. Tallon, D.E. Dunstan, G.V. Franks, Controlling the
318.
microstructure of ceramic particle stabilized foams: influence of contact angle
[30] S. Kitsunezaki, Crack Growth and Plastic Relaxation in a Drying Paste Layer, J. Phys. Soc.
and particle aggregation, Soft Matter 7 (2011) 11464–11474.
Jap. 79 (2010) 124802.
[49] ASTM E1820-17a. Standard Test Method for Measurement of Fracture
[31] A. Sarkar, M.S. Tirumkudulu, Asymptotic analysis of stresses near a crack tip in a
Toughness, ASTM International, West Conshohocken, PA, USA. 2017.
two-dimensional colloidal packing saturated with liquid, Phys. Rev. E 83 (2011)
[50] H.R. Hertz, Ueber die Berührung fester elastischer Körper, J. Für Reine Angew.
051401.
Math. 92 (1882) 156–171.
[32] L. Goehring, W.J. Clegg, A.F. Routh, Wavy cracks in drying colloidal films, Soft
[51] C. Atkinson, R.E. Smelser, J. Sanchez, Combined mode fracture via the cracked
Matter 7 (2011) 7984–7987.
Brazilian disk test, Int. J. Fract. 18 (1982) 279–291.
[33] M.I. Smith, J.S. Sharp, Effects of substrate constraint on crack pattern formation
[52] K. Shetty, A.R. Rosenfield, W.H. Duckworth, Fracture Toughness of Ceramics
in thin films of colloidal polystyrene particles, Langmuir 27 (2011) 8009–8017.
Measured by a Chevron-Notch Diametral-Compression Test, J. Am. Ceram.
[34] H. Cao, D. Lan, Y. Wang, A.A. Volinsky, L. Duan, H. Jiang, Fracture of colloidal
Soc., 68 (1985) C-325-C-327.
single-crystal films fabricated by controlled vertical drying deposition, Phys. Rev. E
[53] J. S. R ee d , Prin cip l es o f Cera mics Pro c es s ing, 2 n d ed . , Wiley, New York,
82 (2010) 031602.
1995.
[35] J.A. Lewis, K.A. Blackman, A.L. Ogden, J.A. Payne, L.F. Francis, Rheological
[54] S. Roy, M.S. Tirumkudulu, Universality in consolidation of colloidal gels, Soft Matter
property and stress development during drying of tape-cast ceramic layers, J. Amer. Ceram.
12 (2016) 9402–9406.
Soc. 79 (1996) 3225–3234.
[55] S. Roy, M.S. Tirumkudulu, Micro-mechanical theory of shear yield stress for
[36] B.J. Ennis, G. Sunshine, On wear as a mechanism of granule attrition, Tribol. Int. 26
strongly flocculated colloidal gel, Soft Matter 16 (2020) 1801–1809.
(1993) 319–327.
[56] A. Sakar, M.S. Tirumkudulu, Asymptotic analysis of stresses near a crack tip
[37] J. Zarzycki, Critical stress intensity factors of wet gels, J. Non-Crystalline Solids
in a two-dimensional colloidal packing saturated with liquid, Phys. Rev. E 83 (2011)
100 (1988) 359–363.
051401.
[38] M.L. Sesso, G.V. Franks, Fracture toughness of wet and dry particulate
[57] K.B. Singh, L.R. Bhosale, M.S. Tirumkudulu, Cracking in Drying Colloidal Films of
materials comprised of colloidal sized particles: role of plastic deformation, Soft
Flocculated Dispersions, Langmuir 25 (2009) 4284–4287.
Matter 13 (2017) 4746–4755.
[39] M.J. Adams, D. Williams, J.G. Williams, The use of linear elastic fracture
mechanics for particulate solids, J. Mater. Sci. 24 (1989) 1772–1776.
[40] T.J. Carneim, D.J. Green, Mechanical properties of dry-pressed alumina green
bodies, J. Am. Ceram. Soc. 84 (2001) 1405–1410.

[41]