You are on page 1of 8

Construction and Building Materials 79 (2015) 65–72

Contents lists available at ScienceDirect

Construction and Building Materials


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

Rheological evaluations of interground and blended cement–limestone


suspensions
Kirk Vance a,⇑, Aashay Arora a, Gaurav Sant b, Narayanan Neithalath a
a
School of Sustainable Engineering and the Built Environment, Arizona State University, Tempe, AZ 85287, United States
b
Laboratory for the Chemistry of Construction Materials (LC2), Department of Civil & Environmental Engineering, University of California, Los Angeles, CA 90095, United States

h i g h l i g h t s

 Rheological effects of limestone in cementitious suspensions brought out.


 Reduction in yield stress with limestone addition; no changes in plastic viscosity.
 Lower apparent yield stress when limestone is interground than blended.
 Viscoelastic storage modulus and apparent yield stress well related.

a r t i c l e i n f o a b s t r a c t

Article history: This paper reports a comparative study of the rheological properties of suspensions composed using
Received 4 August 2014 ordinary portland cement (OPC), interground with, and blended with limestone. Two different inter-
Received in revised form 22 October 2014 ground portland limestone cements (PLCs) and three blended limestone cements were examined. The
Accepted 24 December 2014
blended mixtures were prepared using the same limestone content (by replacement) as the interground
Available online 21 January 2015
mixture, and blended to match the particle size distribution (PSD) of the plain OPC and interground PLCs.
This methodology was used to separate the influence of packing and surface area from the influence of
Keywords:
the limestone incorporation technique (i.e., intergrinding or blending). The results indicate that the inclu-
Rheology
Cement
sion of limestone decreases the yield stress due to reduced van der Waals forces between limestone-to-
Limestone OPC particles, as compared to OPC-to-OPC particle interactions. The plastic viscosity is noted to be inde-
Interground pendent of the limestone replacement level, and dependent solely on the volumetric solid loading and
Fly ash PSD of the solids in the suspension. Intergrinding of limestone is shown to have beneficial effects on
the yield stress compared to PSD-matched OPC–limestone blends, due to coarser OPC fraction in the
interground PLC as compared to the blended mixture. Finally, via oscillatory shear experiments, the
use of a stress plateau identification technique is demonstrated to be an accurate means to identify
the yield stress in cementitious suspensions all suspensions.
Ó 2015 Elsevier Ltd. All rights reserved.

1. Introduction which the PLCs available in the market fall into, and (ii) mechanical
mixing or blending of the OPC with limestone (blended portland
There continues to be interest in the use of fine limestone as a limestone cement, BPLC). As limestone is a softer material than
partial replacement for ordinary portland cement (OPC) in con- the clinker, intergrinding results in limestone particles that are
crete. This is due to benefits linked to: low environmental impact, finer than the OPC particles [9,10], while blending can be tailored
improved particle packing and filler effects resulting from provi- to produce a system where the particle size distributions (PSDs)
sion of additional surface for reactions, and reactions with the alu- of the limestone is similar, finer, or coarser than OPC. Several stud-
minate phase(s) [1–8]. There are two main strategies for replacing ies [11–14] have investigated the influence of limestone on the
OPC with limestone: (i) intergrinding the limestone with the clin- rheological performance of cement suspensions. These studies
ker in a ball mill (interground portland limestone cement, IPLC), to have investigated the influences of: fineness, superplasticizers,
and limestone content on rheology. However, in these studies
limestone replacement was carried out by intergrinding or using
⇑ Corresponding author. Tel.: +1 602 663 1778.
limestone that is coarser or finer than the OPC. Thus the influence
E-mail addresses: kevance@asu.edu (K. Vance), aaror24@asu.edu (A. Arora),
of limestone in these systems is difficult to discern as it is
gsant@ucla.edu (G. Sant), Narayanan.Neithalath@asu.edu (N. Neithalath).

http://dx.doi.org/10.1016/j.conbuildmat.2014.12.054
0950-0618/Ó 2015 Elsevier Ltd. All rights reserved.
66 K. Vance et al. / Construction and Building Materials 79 (2015) 65–72

confounded with the effects of changes in particle surface area and face charges on cement grains indicate that the net van der Waals
particle packing. effect is attractive. Thus, above the coagulation limit, as limestone
The study presented in this paper separates the effects of replaces portland cement in the suspension, the contacts between
changes in particle surface area and particle packing to explore OPC particles are decreased, and the influence of limestone–
the effects of limestone incorporation technique: blending versus cement and limestone–limestone contacts become more prevalent,
intergrinding limestone, on paste rheology. BPLC suspensions are and interparticle attractive forces would be expected to decrease
prepared with limestone which is PSD-matched to OPC to mini- along with the apparent yield stress.
mize or eliminate the above-mentioned confounding effects. Rhe-
ological characterization of cementitious suspensions is 2. Experimental program
commonly carried out using strain/stress growth experiments
from which parameters such as the apparent yield stress and plas- 2.1. Materials
tic viscosity are extracted. The apparent yield stress is indicated by
Two as-manufactured IPLCs were used in this study, one conforming to ASTM
a non-zero stress at a zero strain rate and the plastic viscosity is a C1157 [26] and the other to ASTM C595 [27] with the former being the finer of
measure of the increase in strain rate required for a unit increase in the two. The control OPC used in this study was a Type I/II OPC, conforming to ASTM
stress. This work applies a new stress plateau identification tech- C150 [28], and was the parent OPC used to generate the three different BPLC mix-
nique presented in [15,16] to determine the apparent yield stress, tures. Limestone of four different median particle sizes (0.7, 3, 10, and 15 lm) was
blended with the pure OPC to form the BPLC mixtures. Each of the BPLC mixture has
and extract the plastic viscosity from the linear portion of the shear a particle size distribution that matches either that of the OPC, the ASTM C1157 PLC
rate-shear stress response. In an effort to facilitate discussions on or ASTM C595 PLC. The purity of the limestone used is greater than 95% CaCO3. In
the material response, which is the focus of the paper, a brief the figures and discussions in this paper, the ASTM designation (for example, C 595)
review of the factors that influence the rheological response of par- is used for the interground and the corresponding blended systems. For the size
matched BPLC mixtures, the notation ‘=’ is used in the figures to refer to the distri-
ticulate suspensions, apparent yield stress in particular, follows.
bution that the blend is matched to, i.e., OPC + LS = OPC refers to the OPC–limestone
blend which matches the particle size distribution of OPC. The oxide compositions
and physical characteristics of these materials are presented in Table 1.
1.1. Apparent yield stress and its determinants in particulate The particle size distributions (PSDs) of the OPC and the as-obtained IPLCs,
suspensions determined using dynamic light scattering, are shown in Fig. 1(a) and (b), respec-
tively. BPLCs were prepared to match the PSDs of the two IPLCs and the parent
The plastic viscosity of a suspension is strongly influenced by OPC, using the same replacement level of OPC by limestone as in the IPLCs (11.1%
by mass; 12.7% by volume, as provided by the supplier). Thus, three blended cement
the solid loading and particle separation, as it is a measure of resis-
mixtures were created. A least-squares fitting and error minimization procedure
tance to flow once it has commenced. The apparent yield stress, on was adopted to obtain the quantities of limestone of each size required to match
the other hand, is primarily influenced by particle effects, such as the PSD curves of the OPC or IPLCs. The differential PSDs of these blended systems
jamming, particle spacing, surface area and roughness [14,17,18], as compared to the OPC and the corresponding IPLC systems are presented in
Fig. 1(c). The best possible matches were made by blending chosen amounts of
and interparticle forces (van der Waals and electrostatic repulsive
OPC and the four limestone particle sizes such that the limestone volume fraction
forces) [19,20]. The interaction potential between particles in a col- in the blend is 12.7%, by minimizing the least squares error. It is notable that the
loidal suspension is governed by DLVO theory, which combines the blended mixtures were able to reasonably match the size distribution of the OPCs
effects of van der Waals and electrostatic forces [21,22]. Though or the IPLCs except for the case of the C1157 IPLC. The C1157 IPLC is notably finer
cementitious suspensions have larger particles than would gener- than the blend. However, Fig. 1(c) shows that the finer fraction in the C1157 IPLC
and the corresponding BPLC mixture are very similar. Since the limestone is ground
ally be considered colloidal, DLVO theory is likely still applicable
finer than the clinker in the intergrinding process, it is important to match the finer
as the influence of Brownian motion is minimal due to the relative fraction appropriately to ensure comparability, which has been accomplished here.
size of the particles present in the suspension. Electrostatic repul-
sive forces originate from the adsorbed ions in the diffuse portion 2.2. Experimental procedure
of the electric double layer (EDL) [21]. The EDL constitutes an inner
stern layer of tightly held ions that carry a charge that is opposite The suspensions were prepared at a constant volumetric water-to-solids ratio
(w/s = 1.42), which corresponds to a mass based w/s = 0.45 for OPC suspensions,
to that on the particle surface and the more weakly held ions in the
and a slightly higher mass-based w/s for blends of limestone due to the lower spe-
diffuse outer layer having the same charge as the particle surface. cific gravity of this material. A constant volumetric ratio was used as the flow
Electrostatic repulsive effects are proportional to: exp[- behavior of concentrated suspensions is influenced by the volume fraction of solids
j(r  2a)] [23,24], where r is the distance from the center of the [29,30]. The specimens were mixed in accordance with ASTM C1738 [31], using a
particle, j is the Debye length, and a is the particle diameter. low shear rate of 5000 RPM for initial mixing, followed by a 30 s mixing period at
a high shear rate of 12,000 RPM, a two minute covered rest period, and finally a
Van der Waals forces, on the other hand have been shown to be 90 s mixing at the high shear rate.
proportional to 1/r2, and thus do not decay as rapidly (with dis- Two different rheological experiments were conducted: one, a typical shear rate
tance) as the electrostatic repulsive force. Further, it is noted that ramp study similar to the one developed in [16], using the parallel plate configura-
van der Waals forces between the same materials are always tion, and two, a stress growth, low-amplitude oscillatory shear study using a con-
centric cylinder geometry. The typical shear rate ramp study allows for the
attractive, while between two different materials they may be
determination of yield stress and plastic viscosity of the suspensions while the
repulsive. Above a certain particle concentration, termed the coag- stress growth study provides the viscoelastic parameters of the suspension. Both
ulation limit, the effect of the van der Waals attractive forces dom- studies were carried out using a TA Instruments AR 2000EX dynamic shear rheom-
inate over the electrostatic repulsive forces, and it has been shown eter, with all rheological parameters (stress, strain rate, storage modulus, and loss
that cementitious suspensions are often above this limit [24,25]. modulus) being extracted using the TA Instruments Trios software package. The
shear rate ramp study was conducted using a 50 mm parallel plate configuration
Though portland cement grains are polycrystalline, the occurrence with the Peltier plate conditioned to a surface temperature of 25 ± 0.1 °C. The upper
of coagulation in portland cement suspensions combined with the surface of the plate geometry was serrated to a depth of 1 mm and the lower surface
likelihood that electrostatic effects are repulsive due to similar sur- of contacting plate (i.e., Peltier cover plate) was serrated to a depth of 0.15 mm to

Table 1
Chemical composition and physical characteristics of the materials.

Material SiO2 (%) Al2O3 (%) Fe2O3 (%) CaO (%) MgO (%) SO3 (%) D50 (lm)
OPC (ASTM C150) 19.60 4.09 3.39 63.21 3.37 3.17 12.4
IPLC (ASTM C595) 18.57 3.80 2.99 63.87 2.93 3.15 11.3
IPLC (ASTM C1157) 18.57 3.84 3.11 63.59 3.38 3.23 8.3
K. Vance et al. / Construction and Building Materials 79 (2015) 65–72 67

Fig. 1. Particle size distribution (PSD) of (a) OPC and C 1157 and C 595 IPLCs, and (b) limestones used in this study, and (c) differential PSDs of IPLCs (solid lines) compared to
the corresponding BLPCs (dashed lines).

prevent slip at the shearing surface. Slip causes local particle concentration changes Rheological parameters (apparent yield stress and plastic viscosity) for the
near the shearing surface, potentially resulting in significant underestimation of the shear strain rate growth experiment were extracted using a method presented pre-
rheological properties [32,33]. The gap between the upper and lower plates was set viously [16]. The yield stress is determined as the stress value indicated by the
to 2 mm, based on prior studies [14,16]. The oscillatory shear experiments were stress plateau, and the plastic viscosity is determined as the slope of the best fit line
performed using a concentric cylinder configuration with a constant gap of over the linear portion of rheological data above a strain rate of 5/s. This method
1.07 mm between the outer surface of the bob and inner surface of the cylinder. was shown to be more consistent than using any of the commonly used rheological
The contact surfaces were grit-blasted to minimize slip. models (such as Bingham or Herschel-Bulkley) for cementitious suspensions. A typ-
The experimental procedure used in the shear rate ramp experiment consists of ical diagram showing the parameter extraction methodology is presented in Fig. 3.
a step-up pre-shear to homogenize the sample, followed by a ramp-up and ramp-
down of shear rates. Data acquisition was completed during only the final ramp-
up and ramp-down phases of the study, wherein only data from the down-ramp
3. Results and discussions
portion was used for the extraction of rheological parameters. A graphical overview
of the experimental sequence is presented in Fig. 2. The shear rates used in this
study ranged from 0.005/s-to-100/s, with three points investigated per decade. At 3.1. Rheological response of PSD-matched OPC–limestone blends
each step, data is collected every second. Each step was terminated once a steady
state has been achieved, as indicated by a difference in torque of less than 5% over This section describes the influence of different levels of PSD-
5 consecutive readings, with a maximum step duration of 10 s. The typical duration
of the rheological experiment was 3 min, while the total duration of the experiment
matched limestone replacement on the rheological behavior of
from water addition to completion was approximately 7 min. This experimental OPC suspensions. The PSD-matching approach allows the effects
scheme minimizes the experimental duration; as cement hydration is a time- of particle packing and arrangement to be discounted, thereby
dependent process, the shortest possible experiment duration is necessary to justify facilitating an unbiased investigation of the nature of limestone
the assumption that minimal dissolution and hydration occurs during the experi-
in these suspensions. Fig. 4 shows the shear rate-shear stress and
ment. The oscillatory shear experiment consisted of a stress growth experiment
beginning at a stress of 0.001 Pa and ending at 50 Pa, at a constant oscillatory fre- shear rate-apparent viscosity relationships (flow curves) for the
quency of 1 Hz, as this stress range was found to adequately capture the linear vis- suspensions constituted of OPC and the OPC–limestone blend
coelastic regime in the suspensions investigated. Oscillatory shear experiments (where the cumulative PSD of the blend matches that of the
were conducted so as to minimize the effects of hydration by minimizing the exper- OPC), where 11.1% of OPC (by mass) is replaced by PSD-matched
iment duration. Thus they were conducted over a period of approximately 3 min
with an overall experiment duration from the time of water addition, including
limestone.
mixing, of approximately 6 min. The rheological performance of OPC suspensions has been
shown to be strongly influenced by solid loading, fineness and
the size distribution of particles [14,17], and the surface interaction
between cement particles. As discussed previously, van der Waals
forces exist between cement particles, and above the coagulation
limit, will dominate over electrostatic repulsive forces [19,34,35].
Aqueous suspensions of calcium carbonate at a high pH have sim-
ilarly been shown to develop large attractive van der Waals forces,
indicative of coagulation [20].
Immediately noticeable from Fig. 4 is the fact that the inclusions
of PSD-matched limestone in an OPC suspension induce negligible
change in the plastic viscosity (as noted by the near-parallel lines),
and a reduction in the apparent yield stress of about the same
magnitude as the volumetric limestone replacement level in the
system. The minimal change in plastic viscosity with the incorpo-
ration of limestone suggests that these surface interaction effects
are less significant in the flow of suspensions once flow has been
initiated. This concept can be intuitively confirmed by considering
the negligible influence of superplasticizers on the plastic viscosity
of portland cement suspensions [36]. The adsorption of superplast-
Fig. 2. The rheological procedure for shear rate ramp study. icizers on the surface of cement particles results in a decrease in
68 K. Vance et al. / Construction and Building Materials 79 (2015) 65–72

Fig. 3. The rheological procedures for shear rate ramp study.

supporting the idea that interparticle spacing is unchanged, and


thus surface area remains relatively constant. Thus, it is proposed
that, as PSD-matched limestone replaces OPC in the suspension,
the magnitude of the van der Waals forces between limestone
and cement are lower than those between the cement particles.
The replacement of portland cement by limestone, which reduces
the particle contacts between cement grains1, thus acts to decrease
the net interparticle attractive forces in the suspension, thereby
decreasing the apparent yield stress.
To further investigate the influence limestone replacement on
the rheological properties of portland cement suspensions, a study
was performed at additional PSD-matched limestone contents of
23% and 33% by volume (20% and 30% by mass) while maintaining
the same volumetric water-to-solids ratio of 1.42. Fig. 5(a) depicts
the apparent yield stress and plastic viscosity of these suspensions
as a function of the volumetric PSD-matched limestone replace-
ment level. Noted from this figure is a similar influence as seen
Fig. 4. Representative flow curves for OPC and OPC–limestone suspensions (OPC in Fig. 4, with a constant plastic viscosity regardless of the lime-
replacement by 11.1% by mass of PSD-matched limestone). Solid lines denote the stone content (within the range of replacement levels studied)
stress and the dotted lines, the apparent viscosity. and a continuing decrease in apparent yield stress as the limestone
content is increased (attributed to a reduction in cement–cement
contacts and an increase in cement–limestone contacts which have
interparticle attractive forces, decreasing the apparent yield stress lower associated interparticle forces). This supports the idea that
but resulting in minimal changes in the plastic viscosity. The slight plastic viscosity is more heavily influenced by the particle size dis-
influence of superplasticizers on plastic viscosity of OPC suspen- tribution and the surface area of particles present in the suspen-
sions is attributed to the destruction of agglomerations which sion. Even as more limestone replaces cement in the suspension,
releases water and enhances fluidity [11]. Thus, plastic viscosity, the composite PSD and the specific surface areas remain relatively
which is a measure of the resistance to flow as the strain rate is unchanged, resulting in unchanged plastic viscosity. The apparent
increased, is more heavily influenced by the nature of the suspend- yield stress, on the other hand, is more strongly influenced by
ing fluid, and the solid fraction and size distribution of the particles interparticle forces. These results support the discussions above,
present [17,37]. As the particle size distribution and solid loading wherein the van der Waals forces between adjacent limestone
are equivalent in the mixtures considered here, the plastic viscos- and cement particles are significantly lower than those between
ity is not significantly influenced by the inclusion of limestone. cement particles, and thus increasing the limestone content in
The apparent yield stress, on the other hand, is a measure of the the suspension at a constant solid loading with an unchanged
resistance to initiate flow, and is thus more strongly influenced by PSD results in a continued reduction in interparticle forces, and
dispersion and friction forces as well as particle jamming effects. subsequent reductions in the apparent yield stress.
The rheology of OPC suspensions has been shown to depend signif- To investigate these effects, the yield stress was normalized by
icantly on the interparticle spacing of particles [17], which would the mass of cement in the suspension in Fig. 5(b) and plotted as a
have a direct effect on the magnitude of the forces listed above. function of the volumetric limestone replacement ratio. The nor-
Increasing the spacing by the inclusion of a higher volume of water
or a reduction in surface area more significantly decreases the elec- 1
A computer simulation on a 200 lm  200 lm  200 lm 3D representative
trostatic repulsive forces (as compared to van der Waals dispersion volume element packed with particles and water, corresponding to the chosen (w/s)V,
forces), thereby decreasing the interparticle attractive forces showed that the interparticle spacing was unchanged due to the incorporation of
size-matched limestone. Further, when 12.7% of cement is replaced by limestone,
between cement particles, and resulting in a reduction in the
cement–cement particle contacts reduce to 75.25% (from 100% for a pure OPC
apparent yield stress. Further, it is noted the plastic viscosity system), with 23.06% being cement–limestone particle contacts and 1.69% being
remains relatively constant despite OPC replacement by limestone, limestone–limestone particle contacts.
K. Vance et al. / Construction and Building Materials 79 (2015) 65–72 69

Fig. 5. Influence of the content of PSD-matched limestone on determined rheological properties of the suspension.

malized yield stress is found to remain constant at low limestone


contents, but increases as the replacement level increases beyond
a certain value. At low OPC replacement levels, OPC–OPC and
OPC–limestone particle contacts would likely dominate. The
unchanged normalized yield stresses at lower replacement levels
supports the idea that the repulsive van der Waals forces between
limestone and OPC particles is likely negligible as compared to the
attractive force between OPC particles. However, as the limestone
content increases further, limestone–limestone contacts become
prevalent. It has been shown that suspensions of limestone result
in similar van der Waals attractive forces as suspensions of cement
[20], and when the limestone–limestone contacts in the suspen-
sion become more prevalent, the normalized apparent yield stress
increases. It is likely, that if the limestone content were increased
further, an inflection point in the continuation of Fig. 5(a) would
be found, as increasing limestone–limestone contacts as compared
to limestone–OPC contacts would eventually result in an increase Fig. 6. Yield stress and plastic viscosity of the OPC, IPLC, and the corresponding
in the apparent yield stress. BPLC suspensions Several important observations can be made from Fig. 6: (i) the
plastic viscosity is approximately constant regardless of whether the mixture is
blended or interground, with the exception of the C1157 IPLC suspension, (ii) the
C1157 IPLC suspension shows both higher apparent yield stress and plastic
viscosity, and (iii) the C595 BPLC suspension has a higher apparent yield stress as
3.2. Comparing the rheological parameters of IPLC and the
compared to the corresponding interground PLC suspension.
corresponding BPLC suspensions

3.2.1. Yield stress and plastic viscosity tions on the influence of the fineness of limestone powder on OPC
Rheological studies were performed on the IPLCs and PSD- suspension rheology which has been reported in detail in [14].
matched BPLCs as presented in Section 2. The blends were propor- The third observation, i.e., of the C595 IPLC suspension having a
tioned to match the PSD of the parent OPC and both IPLCs (Fig. 1) lower apparent yield stress as compared to the equivalent blended
with the same volumetric OPC replacement level as the IPLCs suspension, can be explained by considering the relative size distri-
(12.7% by volume or 11.1% by mass). Fig. 6 shows the rheological butions of limestone and OPC in the blend. The similarities in PSDs
parameters of the IPLCs and the corresponding BPLCs. of the blended and the interground systems (Fig. 1(c)) combined
The first two observations can be explained as follows. The con- with similar volume fractions of limestone leads to the expectation
stant plastic viscosity across all suspensions except the C 1157 IPLC that the apparent yield stress and plastic viscosity in these suspen-
is due to the similar PSDs and specific surface areas. As discussed sions would be equivalent. However, intergrinding limestone with
previously, similar PSDs and specific surface areas yield similar portland cement clinker will result in a finer limestone phase and a
plastic viscosities if all other experimental variables are held con- coarser cement phase in the blended cement due to the relative
stant, due to the constant water film thicknesses around the parti- softness of limestone as compared to portland cement clinker [9].
cles in these suspensions. Achieving a reasonable PSD match with The results from the PSD-matched OPC–limestone blends in Sec-
the finer C1157 IPLC was not possible with the sizes of limestone tion 3.1 leads to the inference that the interparticle attractive
used in this study, as mentioned in Section 2. The finer interground forces between limestone and cement are lower than those
mixture has a larger surface area than the corresponding blended between the cement grains, resulting in a reduction in yield stress
mixture which is coarser, resulting in a decrease in water film as PSD-matched limestone replaces OPC. Further, at the same
thickness and subsequent increase in plastic viscosity. The replacement level it is noted that the C595 IPLC is finer than the
increased fineness also decreases particle separation, which OPC-equivalent OPC–limestone blend as indicted in Fig. 1c, which
increases the interparticle attractive forces, and consequently the would be expected to increase the yield stress of the C595 IPLC sus-
apparent yield stress of the C1157 IPLC suspension as compared pension as noted in Fig. 6. Similarly, as both the C595 IPLC and the
to the other suspensions. These results are similar to the observa- corresponding BPLC have comparable PSDs, it would be expected
70 K. Vance et al. / Construction and Building Materials 79 (2015) 65–72

Fig. 7. Relationship between shear stress and storage (solid lines) or loss (dashed lines) moduli of suspensions: (a) pure OPC and the OPC–limestone blend with same PSD as
OPC; (b) C595 IPLC and BPLC; and (c) C1157 IPLC and BPLC.

that the yield stress of these two suspensions would be equivalent, perfectly elastic solid. The elastic portion of the response is domi-
however they are noted to be different. The primary difference nated by interparticle contacts and interparticle forces. Fig. 8
between the blended and interground systems, as discussed previ- shows the phase angles for all the suspensions evaluated here,
ously, is the particle size distributions of the limestone and OPC in and it can be noticed that the trend is consistent with the observa-
the mixture. Thus, the reduction in yield stress in the C595 IPLC tions in Fig. 6, with the C1157 IPLC suspension showing the lowest
suspension as compared to the corresponding BPLC suspension is phase angle, and the PSD-matched OPC–limestone blend equiva-
likely attributed to the coarser cement particles in the IPLC mix- lent to the pure OPC showing the highest. Comparing the pure
ture. This coarser OPC fraction would result in a decrease in the OPC suspension to the PSD-matched blend corresponding to OPC,
cement particle contacts as compared to the equivalent blended there is a noted increase in the phase angle for the blended suspen-
mixture, which subsequently reduces the interparticle attractive sion. As the PSDs are similar and solid loading is the same in both
forces and decreases the apparent yield stress. these suspensions, this effect can be attributed solely to changes in
the interparticle forces in these mixtures. This result also follows
3.2.2. Observations from oscillatory shear stress ramp experiments the inferences from the discussions on limestone–cement disper-
Small amplitude oscillatory stress growth experiments were sion forces presented earlier, where the inclusion of a relatively
performed on suspensions of IPLC and BPLC suspensions. Oscilla- small fraction of PSD-matched limestone results in a more fluid
tory stress growth experiments are used to extract the storage behavior, attributed to a reduction in interparticle attractive forces
and loss moduli, which are the elastic and viscous components of in the particle network due to lower limestone–cement dispersion
a viscoelastic material respectively, as a function of applied stress. forces. The lower phase angle in the C595 BPLC suspension as com-
The typical shape of the moduli-shear stress relationships is a hor- pared to the corresponding IPLC suspension can similarly be attrib-
izontal modulus plateau, followed by a significant reduction over a uted to the particle sizes of cement and limestone as discussed
rather small range of stress, indicative of structural breakdown previously. Finally, the lower phase angle in the C1157 IPLC sus-
[38], and followed by a plateau at a lower modulus level. Fig. 7 pension as compared to all other suspensions can be attributed
shows the relationships between the shear stress and storage/loss to the finer particle size distribution in this suspension, which
moduli for the OPC, the IPLC, and the corresponding BPLC would act to decrease particle spacing and increase particle con-
suspensions. tacts thereby increasing the interparticle attractive forces and
Comparing Fig. 7(a)–(c), it is noted that the trends in the storage allowing the suspension to behave in a more solid-like manner.
and loss moduli of the suspensions in the linear viscoelastic regime Comparisons of the viscoelastic response to the flow response
(i.e., defined as the regime over which the moduli are independent from the strain growth experiments indicate a very strong correla-
of stress [39], i.e., see the plateau region) can be expressed as tion between the storage modulus and the yield stress as indicated
OPC < C595 < C1157, for both the as-obtained cements (OPC and
the IPLCs), and the PSD-matched BPLCs. The increase in moduli
corresponds with an increase in the fineness of the cement or the
blend, where the OPC is least fine and the C1157 PLC is the finest
as indicated in the PSD curves presented in Fig. 1. Further, this
trend is consistent with the apparent yield stress results shown
in Fig. 6.
To allow for further discussions on the viscoelastic nature of
these suspensions, it is useful to examine the viscoelastic phase
angle d, defined as:

G00
¼ tan d ð1Þ
G0
The phase angle represents the ratio of the viscous to elastic
moduli of the material, with a lower value representing a more
solid, or elastic behavior, and a higher value indicating a more
liquid or viscous response. Thus, a phase angle of 1 would repre-
sent a Newtonian fluid and a phase angle of 0 would represent a Fig. 8. Viscoelastic phase angle (d) for investigated suspensions.
K. Vance et al. / Construction and Building Materials 79 (2015) 65–72 71

Fig. 9. (a) Relationship between yield stress and storage modulus in the linear viscoelastic regime, and (b) effect of median particle size on yield stress and storage modulus.

in Fig. 9(a). This relationship is one that is expected, as a higher primary dependence on the component volume fractions and size
storage modulus would indicate a more elastic behavior, which distribution of the particles. It is further noted that interground
in turn indicates a more structured network with higher interpar- PLC suspensions demonstrated lower apparent yield stresses than
ticle attractive forces, and consequently a higher apparent yield the blended suspensions of the same PSD and limestone replace-
stress. The good correlation between the storage modulus and ment ratio at the levels studied. This is a result of the softer lime-
yield stress also confirms the accuracy of the determination of stone particles being ground finer than the clinker in the
apparent yield stress from the plateau of shear stress-shear rate interground mixtures and consequently the coarser OPC particles
relationship at lower shear rates rather than the use of simplified resulting in reduced cement particle contacts (and thus reduced
models [16]. dispersion forces). The finer IPLC system in this study showed
Fig. 9(b) illustrates the influence of median particle size (D50) on the opposite behavior as above, a phenomena attributed the finer
both the storage modulus and yield stress of the suspensions. A rel- PSD of the IPLC suspension as compared to the size matched BPLC
atively strong correlation is noted in this figure; however there is suspension. A very good correlation was obtained between the vis-
some degree of scatter. As the median particle size increases, the coelastic storage modulus and the apparent yield stress, while both
water film thickness on the surface of the suspended particles the storage modulus and yield stress showed reasonable inverse
would increase, resulting in increased particle separation and a correlations with the median particle size.
reduction in the interparticle attractive forces. These results are
consistent with those in [14,17], which indicate that the rheologi- Acknowledgements
cal properties of suspensions are strongly influenced by the sizes of
the suspended particles. However, it needs to be remembered that The authors gratefully acknowledge the National Science Foun-
the viscoelastic and yield stress response are also influenced by the dation (CMMI: 1068985) and Arizona State University for the sup-
particle interactions in these suspensions. Thus, the scatter in this port of this research. The materials for this research were provided
figure is likely attributed to changes in interparticle forces in the by Lehigh Cement and Omya A.G. and these corporations are
particle network resulting from the replacement of OPC with lime- acknowledged. K.V. also acknowledges the Dean’s Fellowship from
stone. As described in detail previously through deductions from the Ira A. Fulton Schools of Engineering at Arizona State University
experimental observations, limestone–cement particle interaction (ASU). This research was conducted in the Laboratory for the Sci-
effects have been illustrated to be lower than cement–cement ence of Sustainable Infrastructural Materials (LS-SIM) at ASU and
interaction, which justifies the changes in the yield stress and vis- the authors gratefully acknowledge the support that has made this
coelastic response of these binder systems despite comparable laboratory possible. The contents of this paper reflect the views of
PSDs. the authors who are responsible for the facts and accuracy of the
data presented herein, and do not necessarily reflect the views
4. Conclusions and policies of the funding agency, nor do the contents constitute
a standard, specification, or a regulation.
This paper has presented the rheological influence of limestone
in interground and blended OPC suspensions. OPC–limestone References
blends were prepared by using limestone particles of four different
median sizes to ensure that the blends have the same PSDs as the [1] Bonavetti V, Donza H, Menéndez G, Cabrera O, Irassar EF. Limestone filler
cement in low w/c concrete: a rational use of energy. Cem Concr Res
parent OPC or the IPLCs. This helped eliminate the effects associ- 2003;33:865–71.
ated with the influences of particle packing and surface areas. With [2] Caldarone MA, Zemajtis JZ. Effect of use of limestone on various properties of
increasing amounts of limestone in the blends, yet keeping the portland cement-Part II. Skokie (IL): Portland Cement Association; 2008.
[3] Bentz D, Irassar EF, Bucher B, Weiss WJ. Limestone fillers conserve cement:
composite PSD the same as that of the parent OPC, the apparent part 1: an analysis based on powers model. Concr Int 2009;31:41–6.
yield stress was found to decrease because of the increasing inter- [4] Hawkins P, Tennis P, Detwiler R. The use of limestone in portland cement: a
particle spacing between the cement grains. This indicates that the state-of-the-art review. EB227 Portland Cem Assoc, Skokie IL 2003;44.
[5] De Weerdt K, Haha MB, Le Saout G, Kjellsen KO, Justnes H, Lothenbach B.
van der Waals forces between limestone and cement are signifi-
Hydration mechanisms of ternary Portland cements containing limestone
cantly lower in magnitude than those between the cement parti- powder and fly ash. Cem Concr Res 2011;41:279–91.
cles. The determination of viscoelastic parameters of the [6] Lothenbach B, Le Saout G, Gallucci E, Scrivener K. Influence of limestone on the
suspensions (storage and loss moduli, and the phase angle) also hydration of Portland cements. Cem Concr Res 2008;38:848–60.
[7] Vance K, Aguayo M, Oey T, Sant G, Neithalath N. Hydration and strength
validated these observations. The plastic viscosity on the other development in ternary portland cement blends containing limestone and fly
hand remained invariant with limestone content, reinforcing its ash or metakaolin. Cem Concr Compos 2013;39:93–103.
72 K. Vance et al. / Construction and Building Materials 79 (2015) 65–72

[8] Kumar A, Oey T, Falla GP, Henkensiefken R, Neithalath N, Sant G. A comparison [24] Neubauer CM, Yang M, Jennings HM. Interparticle potential and sedimentation
of intergrinding and blending limestone on reaction and strength evolution in behavior of cement suspensions: effects of admixtures. Adv Cem Based Mater
cementitious materials. Constr Build Mater 2013;43:428–35. 1998;8:17–27.
[9] Tsivilis S, Voglis N, Photou J. A study of the intergrinding of clinker and [25] Yang M, Neubauer CM, Jennings HM. Interparticle potential and sedimentation
limestone. Miner Eng 1999;12:837–40. behavior of cement suspensions: review and results from paste. Adv Cem
[10] Opoczky L. Grinding technical questions of producing composite cement. Int J Based Mater 1997;5:1–7.
Miner Process 1996;44:395–404. [26] ASTM C1157/C1157M-11. Standard performance specification for hydraulic
[11] Sakai E, Masuda K, Kakinuma Y, Aikawa Y. Effects of shape and packing density cement. West Conshohocken, PA: ASTM International; 2013.
of powder particles on the fluidity of cement pastes with limestone powder. J [27] ASTM C305-13. Standard practice for mechanical mixing of hydraulic cement
Adv Concr Technol 2009;7:347–54. pastes and mortars of plastic consistency. West Conshohocken, PA: ASTM
[12] Mikanovic N, Jolicoeur C. Influence of superplasticizers on the rheology and International; 2013.
stability of limestone and cement pastes. Cem Concr Res 2008;38:907–19. [28] ASTM C150/C150M-12. Standard specification for portland cement. West
[13] Vikan H, Justnes H. Rheology of cementitious paste with silica fume or Conshohocken, PA: ASTM International; 2012.
limestone. Cem Concr Res 2007;37:1512–7. [29] Chong JS, Christiansen EB, Baer AD. Rheology of concentrated suspensions. J
[14] Vance K, Kumar A, Sant G, Neithalath N. The rheological properties of ternary Appl Polym Sci 2003;15:2007–21.
binders containing portland cement, limestone, and metakaolin or fly ash. Cem [30] Jeffrey DJ, Acrivos A. The rheological properties of suspensions of rigid
Concr Res 2013;52:196–207. particles. AIChE J 1976;22:417–32.
[15] Barnes HA. The yield stress—a review or ‘‘pamsa qei’’—everything flows? J [31] ASTM C1738-11a. Standard practice for high-shear mixing of hydraulic
Non-Newton Fluid Mech 1999;81:133–78. cement paste. West Conshohocken, PA: ASTM International; 2011.
[16] Vance, Kirk. Early age characterization and microstructural features of [32] Mannheimer RJ. Effect of slip on the flow properties of cement slurries. Annu.
sustainable binder systems for concrete. PhD Dissertation. Arizona State Meet. Pap. Div. Prod. 1983
University, 2014. [33] Saak AW, Jennings HM, Shah SP. The influence of wall slip on yield stress and
[17] Bentz DP, Ferraris CF, Galler MA, Hansen AS, Guynn JM. Influence of particle viscoelastic measurements of cement paste. Cem Concr Res 2001;31:205–12.
size distributions on yield stress and viscosity of cement–fly ash pastes. Cem [34] Lesko S, Lesniewska E, Nonat A, Mutin J-C, Goudonnet J-P. Investigation by
Concr Res 2012. atomic force microscopy of forces at the origin of cement cohesion.
[18] Erdoğan ST, Martys NS, Ferraris CF, Fowler DW. Influence of the shape and Ultramicroscopy 2001;86:11–21.
roughness of inclusions on the rheological properties of a cementitious [35] Gmira A, Zabat M, Pellenq R-M, Van Damme H. Microscopic physical basis of
suspension. Cem Concr Compos 2008;30:393–402. the poromechanical behavior of cement-based materials. Mater Struct
[19] Flatt RJ. Dispersion forces in cement suspensions. Cem Concr Res 2004;37:3–14.
2004;34:399–408. [36] Banfill PFG. Rheology of fresh cement and concrete. Rheol Rev 2006;2006:61.
[20] Mikanovic N, Khayat K, Pagé M, Jolicoeur C. Aqueous CaCO3 dispersions as [37] Grzeszczyk S, Lipowski G. Effect of content and particle size distribution of
reference systems for early-age cementitious materials. Colloids Surf high-calcium fly ash on the rheological properties of cement pastes. Cem
Physicochem Eng Asp 2006;291:202–11. Concr Res 1997;27:907–16.
[21] Verwey EJW, Overbeek JTG. Theory of the stability of lyophobic colloids. J [38] Asaga K, Roy DM. Rheological properties of cement mixes: IV. Effects of
Colloid Sci 1955;10:224–5. superplasticizers on viscosity and yield stress. Cem Concr Res
[22] Levine S, Dube GP. Interaction between two hydrophobic colloidal particles, 1980;10:287–95.
using the approximate Debye-Hückel theory. I. General properties. Trans [39] Barnes HA. Non-Newtonian I of, Mechanics F. A handbook of elementary
Faraday Soc 1939;35:1125–40. rheology. Univ. of Wales, Institute of Non-Newtonian Fluid Mechanics; 2000.
[23] Hunter RJ, White LR. Foundations of colloid science. Clarendon Press; 1987.