Cement and Concrete Research 63 (2014) 111

Contents lists available at ScienceDirect

Cement and Concrete Research

journal homepage: http://ees.elsevier.com/CEMCON/default.asp

Effect of clinker grinding aids on ow of cement-based materials

Joseph J. Assaad a,, Camille A. Issa b,1
a
Holderchem Building Chemicals, P.O. Box 40206, Lebanon
b
Lebanese American University, P.O. Box 36, Lebanon

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

Article history: Grinding aids (GAs) are increasingly used during cement production to reduce energy consumption and/or
Received 9 December 2013 optimize clinker factor. This paper seeks to assess the effect of such additions on variations in ow of cement
Accepted 17 April 2014 pastes, including static yield stress (0) and viscosity (). Grinding tests were performed at xed specic energy
Available online 9 May 2014
consumption (Ec) or Blaine cement neness. For xed Ec, tests have showed that the increase in cement neness
resulting from the addition of higher GA concentration leads to reduced ow and increased 0 and values.
Keywords:
Grinding (A)
Conversely, cement ground for xed Blaine neness exhibited an improvement in owability together with
Rhelogy (A) reduction in 0 and values. This was related to a dispersion effect of cement agglomerates in the presence of
Fineness (A) GA molecules. Special emphasis is placed throughout this paper regarding the effect of GAs on ASTM C465
Hydration (A) requirements.
Clinker (D) 2014 Elsevier Ltd. All rights reserved.

1. Introduction or hardened states. The setting and hardening properties of cement

containing GAs are well documented in literature. For instance,
Grinding aids (GAs) are incorporated during comminution of Ramachandran reported that TEA retards hydration of C3S and -C2S
clinker to reduce electrostatic forces and minimize agglomeration of together with some changes in morphology and microstructure of the
cement grains [13]. Their chemical basis mostly includes ethanol- hydration products [6]. The hydration of C3A was accelerated in the
amines such as monoethanolamine (MEA), diethanolamine (DEA), presence of TEA due to the accelerated formation of hexagonal alumi-
and triethanolamine (TEA) as well as glycols such as propylene glycol nate hydrate and its transformation to a cubic form [6]. Heren and
(PG), monoethylene glycol (MEG), and diethylene glycol (DEG). Olmez found that the addition of increased ethanolamine concentra-
Because of their highly organic polar nature, GAs are preferentially tions alters cement hydration and leads to retardation in setting times
adsorbed on surfaces formed by the fracture of electrovalent bonds in the order of TEA N DEA N MEA [7].
(Ca\O and Si\O), thus reducing surface energy forces. Such additions Triisopropanolamine (TIPA), which is an amino-alcohol and belongs
are commonly used to increase cement neness and compressive to the group of alkanolamines, was found to change hydration reactions
strength for given specic energy consumption (Ec) [1,2]. This could and particularly increase cement strengths. Perez et al. reported that
be particularly the case when producing cement possessing increased TIPA remains in the interstitial paste solution (not adsorbed to the cement
Blaine neness necessary for high early strength requirements surface, as the TEA) and forms iron complexes to accelerate hydration of
(i.e., ASTM C150 Type III cement [4]). Nevertheless, with today's con- C3S and C4AF [8]. Besides the enhancement of alite and felite hydration,
straints regarding the reduction of usable energy, GAs are more and Ichikawa et al. presented evidence that TIPA also promotes the hydration
more used to reduce Ec for given neness, thereby leading to savings of limestone and densies the interfacial transition zone (ITZ) between
in electrical energy and improved mill productivity [1,3,5]. Typical GA hydrated cement paste and sand or aggregate particles [9]. Sandberg
dosage rates used during grinding of clinker vary from 0.01% to 0.15% and Doncaster reported that the strength increase resulting from the
of the manufactured cement mass. presence of TIPA is not solely dependent on an ITZ mechanism [10].
After the grinding process, GAs may not preserve their original mol- The strength gain observed in hydrated cement pastes suggested that
ecule structures; however, they do remain adsorbed onto the cement TIPA is capable to enhance mechanical properties without any paste-
particles to entail variations of cement properties whether in the fresh aggregate ITZ, regardless of the cement type and age of testing [10].

2. Context and objectives of this project

Corresponding author. Tel.: +961 3 437786; fax: +961 5 921118.
E-mail address: [email protected] (J.J. Assaad). Limited data exits in literature pertaining to the effect of GAs on ow
1
Tel.: +961 3 022682; fax: +961 9 547254. of cement-based materials. In fact, current literature has extensively

http://dx.doi.org/10.1016/j.cemconres.2014.04.006
0008-8846/ 2014 Elsevier Ltd. All rights reserved.
2 J.J. Assaad, C.A. Issa / Cement and Concrete Research 63 (2014) 111

treated different aspects of cement rheology as well as the factors and provide a discussion pertaining to the effect of GAs on ASTM C465
mechanisms governing the aptitude to ow [11,12]; however, refrained requirements, and corresponding variations in rheology. Relevant
to consider the effect of GAs added during cement production. Aiad et al. parameters including water demand, setting time, and compressive
were among the few researchers who studied the direct effect of GAs on strength were evaluated. Such data can be of particular interest to
the rheology of cement pastes [13]. The authors found that viscosity is cement manufacturers and concrete technologists as well as standardiz-
highly dependent on the type and dosage rate of ethanolamine used, ing committees dealing with specications for GAs.
whereby a decrease in viscosity was noticed following the sequence of
TEA N poly-TEA N MEA. This was related to the number of O\H groups 3. Experimental investigation
in the ethanolamine molecules that are adsorbed on the surface of ce-
ment grains, causing different repulsive forces and leading to variations 3.1. Materials
in uidity levels. However, it is important to note that the tests carried
out by Aiad et al. [13] cannot be conclusive as the ethanolamines were Industrial clinker used for the production of ASTM C150 Type I
added as post-additions to the cement (i.e., not during the grinding cement [4], ground granulated blast furnace slag meeting the require-
process) and at concentrations varying from 0.1% to 2% of cement ments of ASTM C989 Grade 80, and gypsum materials were employed
weight (i.e., substantially higher than in real situations). in this study. Their chemical compositions are presented in Table 1.
In dry cement, Jolicoeur et al. reported that PG resulted in improved The (C3S + C2S)/(C3A + C4AF) ratio of clinker used is equal to 3.14,
uidity due to the repulsive forces created between the powder parti- indicating high grindability requiring relatively reduced amount of
cles [14]. This was affected by the additive molecular weight following energy for given cement neness [20]. The relative hardness of the
the sequence of trimer (TPG), dimer (DPG), and monomer (PG). At clinker, slag, and gypsum determined according to the Mohs hardness
increased concentrations, however, it is hypothesized that the GA poly- scale were around 5.5, 6, and 2, respectively.
mers adsorbed onto the surface tend to make the clinker grains sticky, Two commercially available GAs were tested. The amine-based GA is
thus causing agglomeration of the system [14,15]. Anna et al. compared commonly used as a grinding aid and strength enhancer in the cement
the Z-potential of clinker containing TEA with others ground with industry. It had 68% active chemicals when determined by the Karl
polycarboxylate (PC) or poly-naphthalene sulfonate (PNS) concrete Fischer method and specic gravity of 1.09. The Thin-Layer Chromatog-
superplasticizers [16]. The authors found that the TEA uidifying raphy analysis of this GA revealed the presence of TIPA (C3H9NO) and
mechanism for the dry cement system lies between the steric hindrance TEA (C6H15NO3) in the proportions of around 58 to 42%, respectively,
associated with PC polymers and electrostatic interaction of PNS of the total active chemicals. Its pH was found equal to 7.2, indicating
with the positive charges of cement grains [16]. While characterizing that this GA has been acidied (given that the pH of TIPA and TEA are
GAs and their impact on cement performance, Katsioti et al. noted an generally in the order of 10.5 1). This GA is completely soluble in
improvement in workability of cement pastes containing TIPA [17]. water and has a brown color, viscosity of 125 cP, and ammonia-like
This was related to the breaking down of cement agglomerates and odor.
balance modication between inter-particle forces. The second GA used in this study is glycol-based composed of DEG
Besides the lack of documentation in literature, the variations in ow (C4H10O3) and MEG (C2H6O2). It is commonly referred to as grinding
and rheology of cement-based materials due to the incorporation of GAs aid and pack-set inhibitor in the cement industry. Its active chemicals
are not abided by any standard specication or testing protocol. For determined by the Karl Fischer method, pH, and specic gravity were
example, the NCHRP report 607 neglected the effect of such additions equal to 72%, 7.8, and 1.107, respectively. It is highly soluble in water,
on the rheology of cementitious materials, and recommended only odorless, and has a brown color and viscosity of 85 cP.
some chemical and physical tests to be performed [18]. ASTM C465
[19] dened acceptance criteria for GAs based on a series of chemical 3.2. Production of cement used for testing
and physical tests (no rheological tests are specied) for determining
whether such additions dramatically affect cement properties pre- A 50-liter grinding mill connected to an electric counter for monitor-
scribed in ASTM C150 [4]. The most relevant physical requirement of ing Ec was used (Fig. 1). The Ec in kWh/ton was determined as (Tc
ASTM C465 includes the water demand needed to achieve normal con- 1000) / (mass of ground mix in kg MF), where Tc refers to the
sistency for cement containing GA, which should not increase by more amount of electricity in kWh measured from the electric counter, and
than 1% from that required by the corresponding control cement. The MF is the constant mill factor taken equal to 3 by the grinding mill man-
setting times of cement ground with GA should not vary by more than ufacturer. The mill's drum diameter, width, and rotational speed were
1 h or 50%, whichever is lesser, from those obtained by the control 400 mm, 400 mm, and 50 rpm, respectively. It contained a total of
cement [19]. ASTM C465 species that the mortar compressive strength 80 kg steel balls among which 36 kg has 20-mm diameter and 44 kg
should not decrease by more than 5% from the value resulting from a has 30-mm diameter. Prior to grinding, the clinker, gypsum, and slag
similar mortar made with the corresponding control cement [19]. materials were crushed and sieved so that all particles are smaller
The rst objective of this paper is to assess the effect of GAs on var- than 10 mm. The gypsum and slag were dried to constant mass at
iations in ow, including static yield stress and viscosity, of cement 45 and 105 C, respectively, prior to use. All grinding tests were
pastes prepared with different water-to-cement ratios (w/c). Grinding conducted using 7 kg of a mix composed by 90% clinker, 5% gypsum,
tests were performed in two ways, i.e. xed Ec or Blaine cement ne- and 5% slag.
ness. Commercially available amine and glycol-based GAs were used First, a mix ground without GA at 42 kWh/ton was tested and con-
at various concentrations. The second objective of this paper is to sidered in this project as being the reference or control cement. Then,

Table 1
Chemical compositions of clinker, slag, and gypsum.

SiO2, % Al2O3, % Fe2O3, % CaO, % MgO, % SO3, %

Clinker 20.6 6.35 4.5 64.1 1.86 0.22

C3S = 54.6%; C2S = 17.4%; C3A = 9.2%; C4AF = 13.7%; Loss on ignition = 1.15%; Na2Oeq = 0.39%; Free lime = 0.26%; Specic gravity = 3.14
Slag 34.5 12.1 0.75 41.2 9.05 2.4
Loss on ignition = 0.21%; Moisture content = 0.04%; Na2Oeq = 0.66%; Slag activity index with cement at 28 days = 86.4%; Specic gravity = 2.94
Gypsum 2.7 0.55 0.4 31.5 1.5 43.2
Free water (T b 45 C) = 0.03%; Combined water (T b 230 C) = 15.6%; Carbon dioxide = 3.7%
 J.J. Assaad, C.A. Issa / Cement and Concrete Research 63 (2014) 111 3

values given in this paper refer to the individual percentages retained

on the 90-m and 38-m sieve, respectively. The residues on the 106-
m sieve were in the range of 0.2% to 3.5%, depending mostly on the
Ec used (note that all particles retained on this later sieve were not
included in the cement mix used for subsequent testing). The heat of
hydration (H) was determined according to ASTM C186 Test Method
by measuring the difference between the heat of solution of dry cement
and that of hydrated cement in a mixture of nitric and hydrouoric acids
[22]. The H values were determined after 3, 7, and 28 days of hydration
periods; however, given that similar trends are obtained, only the 7-day
results will be reported in this paper.

3.3.2. Tests on cement pastes

All pastes were batched with a laboratory mixer using water cooled
to constant temperature of 20 3 C. Water was rst introduced in the
mixer followed gradually by the ground cement over 2 min. After a rest
period of 30 s, the mixing was resumed for 1 additional minute. The
ambient temperature and relative humidity during testing were main-
tained at 23 3 C and 55 5%, respectively. The water demand
required to achieve normal consistency was determined following
ASTM C187 Test Method [23]. Using the same cement paste, the Vicat
initial and nal setting times were then determined as per ASTM C191
Test Method [24] (for clarity reasons, only the nal set values are
reported in this paper).
The effect of GA types and concentrations on ow properties was
evaluated using cement pastes prepared at 0.48 and 0.42 w/c. These
w/c were selected in order to produce pastes with different consistency
levels ranging from highly owable to relatively cohesive. The ow was
Fig. 1. Photo of the grinding mill used for testing. determined as the average diameter of the paste after spreading on a
horizontal surface using an ASTM C230 mini-slump cone [25]. A Brook-
eld rotational viscometer connected to a data logger was used to eval-
GAs were introduced and two series of grinding tests were performed uate static yield stress (0) and viscosity (). The vane used consisted of
at xed or varying Ec values. In the rst series, the Ec was xed at four blades arranged at equal angles around the main shaft; it measured
42 kWh/ton and GAs added at pre-selected concentrations varying 24 mm in height and 12 mm in diameter.
from low to high levels. As will be discussed later, the high levels were Right after mixing, the cement paste was poured in a cylindrical
considered being reached when the water demand, setting time, bowl and allowed to rest for 1 min prior to measuring the 0 value.
and/or compressive strength of cement ground with GAs exceeded The testing protocol consisted on subjecting the paste to a very low
the ASTM C465 limitations [19]. In the second series of grinding tests, rotational speed of 0.3 rpm and recording the changes in torque as a
the pre-selected GA concentrations were added to the mix; but this function of time. The 0 was determined in accordance to Nguyen and
time, the Ec was reduced during grinding in a way to maintain the ne- Boger [26], by considering the maximum torque registered that
ness equivalent to the control cement, i.e. Blaine of 3460 100 cm2/g. indicates the initiation of ow. Following the determination of 0, the
Again, as will be discussed later, the optimum decrease in Ec was vane was stopped and specimen stirred to mitigate the formation of
achieved when cement properties exceeded ASTM C465 limitations preferential shear planes due to particle orientation. The paste was
[19]. At the end of grinding, the temperature of the charge was found then allowed to rest for 1 min for viscosity measurements. The testing
to increase from ambient (i.e., 23 C) to around 33 to 37 C. protocol consisted on maintaining the vane impeller at relatively high
rotational speed of 60 rpm and recording the decay in viscosity as a
3.3. Testing equipment and procedures function of time. Such plots are commonly referred to as thixotropic
breakdown curves, and are characterized by a peak initial viscosity
3.3.1. Tests on powder cement (Initial) which corresponds to the initial structural condition, and
Following grinding, the cement neness was determined using the thereafter decay with time towards an equilibrium value (Equil)
Blaine apparatus, as per ASTM C204 Test Method [21], and by mechan- where balance between occulation and deocculation is reached
ical sieving on 106, 90, and 38 m mesh openings. The R-90 and R-38 [12,27]. It is important to note that the total elapsed time from the initial

Table 2
Repeatability of testing for cement containing 0.11% amine-based GA ground at Ec of 35.7 kWh/ton.

Blaine, cm2/g R-38, % Water demand, % Final set, min 28-Day compression, MPa w/c = 0.42

0, Pa Initial, Pas Equil, Pas

Test no. 1 3495 27 28.4 250 53 31.4 13.2 4.1

Test no. 2 3465 28.8 28 285 54.5 28.7 11.2 3.8
Test no. 3 3255 29.9 26.1 285 48.8 26.3 10.1 3.4
Mean 3405 28.57 27.5 273.33 52.1 28.8 11.5 3.77
St. deviation 130.77 1.46 1.23 20.21 2.95 2.55 1.57 0.35
COV, % 3.84 5.12 4.47 7.39 5.67 8.86 13.67 9.32
4 J.J. Assaad, C.A. Issa / Cement and Concrete Research 63 (2014) 111

Table 3
Properties when cement is ground at constant Ec of 42 kWh/ton.

Control mix Mixtures ground with amine-based GA Mixtures ground with glycol-based GA

GA dosage, % of mass 0 0.06 0.11 0.13 0.14 0.06 0.1 0.11 0.12
H at 7 days, J/g 278 336 387 406 443 297 320 341 362
Blaine, cm2/g 3460 3695 4180 4335 4410 3575 3885 4025 4060
R-38, % 28.3 20.5 14 10.6 7.7 23 14.9 11.2 15.1
Water demand, % 27.25 27.65 28.15 28.3 28.6 27.6 28.15 28.35 28.55
Final set, min 235 220 180 175 150 220 195 195 170
w/c = 0.48 Flow, mm 195 190 180 170 160 205 190 180 175
0, Pa 9.2 10 12.8 14 16 8.8 10.8 12 13.2
Initial, Pas 5.7 6.7 7.1 7.5 9.3 5.4 6 6.6 7.6
Equil, Pas 2.3 2.4 2.5 2.9 3.4 2.2 2.4 2.6 2.8
w/c = 0.42 Flow, mm 150 145 140 130 125 155 145 140 130
0, Pa 30.4 36 44.8 47.2 57.6 31.2 36 38.4 40.8
Initial, Pas 14.3 17.2 20.1 22 25.2 13.9 17 17 19.1
Equil, Pas 4.1 4.3 4.7 5.9 6.8 4 4.6 4.8 5.2
7-d compression, MPa 35.6 38.2 42 42.1 43.9 37 39 39.4 40.1
28-d compression, MPa 47.2 51 56.1 57.7 59 49.4 51.2 53.7 54.6

Notes: Grinding time was xed at 33.4 min, corresponding to Ec = 42 kWh/ton.

For clarity reasons, results obtained for cement ground with 0.03% and 0.09% amine-based GA are not added (same applies for glycol-based GA at 0.03% and 0.08% dosages).

contact of cement with water until the ow measurements was around the 0, Initial, and Equil responses, respectively, may be attributed to
15 min. variations in Blaine neness that leads to different restructuring rates
during the rest period prior to shearing. Also, connement and yielding
3.3.3. Tests on mortars of paste within the rotating vane impeller may lead to some discrepan-
The compressive strength was determined as per ASTM C109 Test cies in rheological measurements [12,26].
Method [28]. The 50-mm cubes were cured in water until the testing
age after 7 and 28 days.
4.2. Grinding tests realized at xed Ec value
4. Test results and discussion
The ow including 0, Initial, and Equil along with the other cement
4.1. Repeatability of testing properties determined following clinker grinding at xed Ec are sum-
marized in Table 3. As expected, cement neness reected by the Blaine
Several cement mixtures were ground 2 to 3 times in order to and sieve residues substantially improved with the addition of higher
evaluate repeatability of testing determined following clinker grinding. concentration of either amine or glycol-based GA. For example, the
Typical responses determined on cement ground with 0.11% amine- Blaine increased from 3460 cm2/g for the control cement to 3695 and
based GA at Ec of 35.7 kWh/ton are given in Table 2. The coefcient of 4335 cm2/g when the amine-based GA was used at a rate of 0.06% and
variation (COV) is taken as the ratio between standard deviation and 0.13%, respectively. The corresponding R-38 values decreased from
mean values, multiplied by 100. Acceptable repeatability was obtained 28.3% to 20.5% and 10.6%, respectively. A slight increase in sieve residues
for the Blaine, R-38, water demand, setting time, and compressive was noticed for cement ground with glycol-based GA at high concentra-
strength responses that resulted in COV of 3.8%, 5.1%, 4.5%, 7.4%, and tion of 0.12%, which may be due to increased stickiness of the powder
5.7%, respectively. The increase in COV up to 8.9%, 13.7%, and 9.3% for cement grains [14,15].

29.0
Water demand, % of cement mass

Permissible 1% increase in
Ec = 42 kWh/ton water, as per ASTM C465
28.5

28.0

27.5

27.0
0 0.03 0.09 0.11 0.13 0.14 0.03 0.08 0.1 0.11 0.12

Blaine ,cm2/g 3460 3505 3885 4180 4335 4410 3495 3600 3885 4025 4060

Ref. Amine-based GA Glycol-based GA

Fig. 2. Variations in water demand and Blaine for cement ground with various concentrations of amine and glycol-based GAs, at xed Ec.
 J.J. Assaad, C.A. Issa / Cement and Concrete Research 63 (2014) 111 5

18
Ec = 42 kWh/ton 4410*
16 w/c = 0.48 * Typical Blaine

Static yield stress, Pa

4060*
values, cm2/g

4180*

4025*
14

3885*

3600*
12

3460*

3505*

3495
*
10

8
0 0.03 0.06 0.09 0.11 0.13 0.14 0.03 0.06 0.08 0.1 0.11 0.12

Flow, mm 195 200 190 180 180 170 160 195 205 200 190 180 175

Ref. Amine-based GA Glycol-based GA

Fig. 3. Variations in 0 and ow for 0.48 w/c pastes prepared using cement ground at xed Ec.

4.3. Determination of permissible GA concentrations at xed Ec permissible concentration of amine-based GA is slightly lower than
0.13%, and also slightly lower than 0.11% for the glycol-based GA (Fig. 2).
The variations in water demand needed to achieve normal consis-
tency for cement ground with various GAs at xed Ec are plotted in 4.4. Flow of cement ground at xed Ec
Fig. 2 (the Blaine values are also shown). Compared to 27.25% water re-
quired for the control cement, the addition of greater GA concentrations The values of 0 are directly affected by the cement used for testing,
is shown to require gradual increases in water, until exceeding the 1% as is typically shown in Fig. 3 for pastes prepared with 0.48 w/c (the var-
permissible ASTM C465 limitation. This can be directly related to iations in ow are also shown). For example, 0 increased from 9.2 Pa for
the higher specic cement surface area that necessitates additional the control cement to 12.8 and 16 Pa when the cement is ground with
amounts of water for proper lubrication. Consequently, as expected, 0.11% and 0.14% amine-based GA, respectively. Such increase was up
the increased cement neness led to reduced setting times due to to 10.8 and 13.2 Pa when the cement is ground with 0.1% and 0.12%
accelerated hydration reactions together with increased compressive glycol-based GA, respectively. The Initial and Equil values followed
strength (Table 3) [29,30]. For example, the setting time decreased similar variations as those observed for 0, i.e. increased with increasing
from 235 min for the control cement to 180 and 195 min with the addi- cement neness due to higher GA concentration (Table 3).
tion of amine or glycol-based GA, respectively, at dosage rates of 0.11%. The increase in 0 and values can primarily be attributed to a
The corresponding compressive strength after 28 days increased from physical effect related to higher internal friction and inter-particle
47.2 MPa for the control cement to 56.1 and 53.7 MPa, respectively. It links resulting from cement mixtures possessing increased Blaine
is to be noted that the setting dropped sharply to 150 min, i.e. below values. The increase in Blaine reects higher packing density, thus lead-
the ASTM C465 limitation of 23560 = 175 min, for the cement having ing to a decrease in ow (as shown in Fig. 3) and increase in shear
a Blaine of 4410 cm2/g and containing 0.14% amine-based GA (Table 3). strength resistance of the plastic material [12,27]. Concurrent with
Based on above data, it can be concluded that both water demand this phenomenon, the increase in rheological properties can be related
required for normal consistency and setting times are key factors to the nucleation and growth of early hydrates that become more
inuencing allowable GA dosage rates at xed Ec. Thus, the maximum important when cement is ground to a ner level. Although the degree

Table 4
Properties when cement is ground with amine-based GA for xed Blaine of 3460 100 cm2/g.

Grinding time, min 33.4 32.1 30.7 29.5 28.4 27.7 27.2
Ec, kWh/ton 42 40.4 38.6 37.1 35.7 34.8 34.2
GA dosage, % of mass 0 0.03 0.06 0.09 0.11 0.13 0.14
H at 7 days, J/g 278 272 290 284 318 281 256
Blaine, cm2/g 3460 3430 3495 3360 3405 3375 3340
R-90, % 8.6 8.1 8.2 9.3 9.8 9.5 11.6
R-38, % 28.3 25.8 22 30.1 28.6 32.8 36.7
Water demand, % 27.25 27.3 27.45 27.2 27.5 27.05 26.9
Final set, min 235 240 235 255 275 290 330
w/c = 0.48 Flow, mm 195 190 200 195 200 205 210
0, Pa 9.2 10 8.8 7.6 7.2 7.6 7.2
Initial, Pas 5.7 5.8 5.3 5.4 5.2 4.9 4.7
Equil, Pas 2.3 2.2 2.2 2.4 2.1 2.0 1.9
w/c = 0.42 Flow, mm 150 150 160 155 160 165 165
0, Pa 30.4 28.8 29.6 28 28.8 26.4 25.6
Initial, Pas 14.3 14.2 13.3 11.9 11.5 10.1 9.3
Equil, Pas 4.1 4.1 3.9 3.7 3.8 3.7 3.5
7-d compression, MPa 35.6 35.9 38.1 38 39 36.9 37.8
28-d compression, MPa 47.2 48.8 49.7 49.6 52.1 50.3 46.2
6 J.J. Assaad, C.A. Issa / Cement and Concrete Research 63 (2014) 111

Table 5
Properties when cement is ground with glycol-based GA for xed Blaine of 3460 100 cm2/g.

Grinding time, min 33.4 32.4 31.2 30.5 29.8 29.3 28.9
Ec, kWh/ton 42 40.7 39.2 38.4 37.5 36.9 36.3
GA dosage, % of mass 0 0.03 0.06 0.08 0.1 0.11 0.12
H at 7 days, J/g 278 282 269 302 270 254 249
Blaine, cm2/g 3460 3415 3475 3520 3515 3400 3345
R-90, % 8.6 7.8 8.4 9 8.9 10.5 14.2
R-38, % 28.3 25.2 24 29.1 34.3 35 41.6
Water demand, % 27.25 27.05 27.2 27.1 27.3 26.85 26.5
Final set, min 235 255 265 260 300 320 355
w/c = 0.48 Flow, mm 195 210 210 225 220 230 230
0, Pa 9.2 8.8 8 7.6 7.6 6.8 6.4
Initial, Pas 5.7 5.3 4.8 4.5 4.4 4.2 4.3
Equil, Pas 2.3 2.2 2.1 2.1 1.9 1.7 1.7
w/c = 0.42 Flow, mm 150 145 160 175 170 175 180
0, Pa 30.4 28.8 25.6 21.6 22.4 18.4 17.6
Initial, Pas 14.3 14.4 12.8 11.1 12.0 9.8 8.8
Equil, Pas 4.1 4 3.6 3.3 3.2 3.3 2.9
7-d compression, MPa 35.6 34.8 37 36.7 35.3 32.9 32.5
28-d compression, MPa 47.2 48 47 49.5 45 43.1 42

of hydration is close to zero during the dormant period, such nucleation was ensured at Ec of 35.7 and 36.9 kWh/ton for cement ground with
can build a network of bridges between hydrated cement particles amine or glycol-based GA, respectively.
together with increased level of inter-particle links [31]. The decrease in Ec provided during the grinding process led to an
It is important to note that GA molecules adsorbed onto the cement increase in R-38 and R-90 values, particularly at high GA concentrations
particles can also play a role during the dormant period of hydration to (Tables 4 and 5). Hence, R-38 increased from 28.3% for the control
alter 0 and measurements. However, it is rather difcult to quantify cement to 36.7% and 41.6% when the amine or glycol-based GA was
this role for cement mixtures ground at xed Ec, given that the resulting used, respectively, at a rate of 0.14% or 0.12%, respectively. The corre-
variations are primarily affected by a physical effect. The inuence of GA sponding Ec dropped from 42 kWh/ton to 34.2 and 36.3 kWh/ton,
molecules on 0 and will be further discussed when tests are realized respectively. Nevertheless, it is important to note that a certain amount
at constant Blaine. of ne material was produced during grinding of those later mixtures,
which allowed maintaining the targeted Blaine neness, as can be
4.5. Grinding tests realized for constant Blaine neness seen in the particle size distribution curves determined using a laser
diffraction analyzer (Fig. 4). The wet-packing densities of those mix-
The various cement properties determined following clinker grind- tures determined according to Wong and Kwan [32] were quite similar
ing at xed Blaine neness of 3460 100 cm2/g using either amine or (i.e., around 0.62 0.005), indicating that Blaine neness is mostly
glycol-based GA are summarized in Tables 4 and 5, respectively. Regard- affected by the packing density of cement. Conversely, the sieve
less of GA type, the addition of increased concentration led to consecu- residues are functions of the maximum particle size and directly affected
tively reduced Ec values. For example, Ec decreased from 42 kWh/ton by the amount of grinding energy applied [1].
for the control cement to 37.1 and 34.2 kWh/ton with the use of 0.09%
and 0.14% amine-based GA, respectively, corresponding to 11.7% and 4.6. Determination of permissible GA concentrations for xed Blaine
18.6% reduction in energy consumption, respectively. For given GA
concentration, higher decreases in Ec were achieved with the use of As can be seen in Fig. 5, the nal setting time remained almost
amine-based GA. For example, at a rate of 0.11%, the targeted Blaine unchanged at low to relatively moderate GA dosages, but then started

100

Control cement
Ec = 42 kWh/ton
80 Wet-packing density = 0.615 Blaine = 3460
Percentage passing, %

100 cm2/g
60 Cement + 0.14% amine GA
Ec = 34.2 kW/ton
Wet-packing density = 0.623

40

Cement + 0.12% glycol GA

20
Ec = 36.3 kWh/ton
Wet-packing density = 0.620

0
0.1 1 10 100 1000
Particle size, m

Fig. 4. Typical particle size distribution (PSD) curves (Note: PSD in laboratory mills may differ from industrial ones due to separators and ventilation).
 J.J. Assaad, C.A. Issa / Cement and Concrete Research 63 (2014) 111 7

380
Blaine = 3460 +/ - 100 cm2/g
Various Ec values
Permissible increase in

Final setting time, min

340 setting, as per ASTM C465
(235 + 60 = 295 min)

300

260

220
0 0.03 0.06 0.09 0.11 0.13 0.14 0.03 0.06 0.08 0.1 0.11 0.12

Ref. Amine-based GA Glycol-based GA

Fig. 5. Variations in setting time for cement ground with various concentrations of amine and glycol-based GAs, for xed Blaine.

increasing sharply at higher dosage rates until exceeding the ASTM The variations in (Compression) determined after 7 and 28 days
C465 limitation of 235 + 60 = 295 min. For example, in the case of for mortars made with cement ground for xed Blaine neness
glycol-based GA, the setting time remained stable at lower GA concen- are plotted in Fig. 6 (the H values are also shown). The (Compression)
trations, and then increased up to 300 and 355 min at a rate of 0.1% is calculated as the strength determined using cement containing
and 0.12%, respectively. Teoreanu and Guslicov suggested the existence GA minus the one determined using control cement, divided by
of an optimum GA concentration that can be correlated with the the strength of control cement, and multiplied by 100. From low to rel-
achievement of a continuous absorption monolayer onto the micro- atively moderate concentrations, the increase in strength was much
fractured cement particles [2]. Beyond such optimum, the adsorbed more pronounced for mortars prepared using cement ground with
monolayer may partly block the hydration reactions and lead to retar- amine-based GA, as compared to those containing the glycol-based
dation in setting times [7,14], just like what happens with the use of one. For given Blaine, this can mainly be attributed to the presence of
water reducers and superplasticizers in concrete mixtures. TIPA that strengthens the CSH compounds and densies the interfa-
The GA adsorption onto the cement grains and saturation of charges cial transition zone between the cement paste and sand particles
are expected to result in improved owability due to the breaking down [9,10]. This is corroborated by the relative increase in H values from
of cement agglomerates and modication of their inter-particle forces 278 J/g for the control cement to 318 and 254 J/g for cement containing
[13,17]. Hence, a decrease in water demand required to achieve normal amine or glycol-based GA, respectively, at a dosage of 0.11%. At high
consistency was observed for cement mixtures ground with amine or concentrations, the hydration reactions can be altered by the adsorbed
glycol-based GAs for xed Blaine neness (Tables 4 and 5; note that GA molecules [6,7,14], thus leading to decreased strength (Fig. 6).
ASTM C465 does not specify any limitation in case a decrease in water Such decrease was particularly pronounced in the case of mortars
is encountered). For example, the water decreased from 27.25% for the prepared with cement ground with glycol-based GA. Hence, at dosage
control cement to 27.1% and 26.5% when the glycol-based GA was rates of 0.11% and 0.12%, the (Compression) dropped below the
used at a concentration of 0.08% and 0.12%, respectively. ASTM C465 limitation of 5% (Fig. 6).

15
After 7 days Blaine = 3460 +/ - 100 cm2/g
(Compression), % of control

10 After 28 days Various Ec values

-5
Permissible 5% decrease in
-10 compression, as per ASTM C465

-15
0 0.03 0.06 0.09 0.11 0.13 0.14 0.03 0.06 0.08 0.1 0.11 0.12

H, J/g 278 272 290 284 318 281 256 282 269 302 270 254 249

Ref. Amine-based GA Glycol-based GA

Fig. 6. Variations in compressive strength (and H values) for mortars made using cement ground with various concentrations of amine and glycol-based GAs, for xed Blaine.
8 J.J. Assaad, C.A. Issa / Cement and Concrete Research 63 (2014) 111

11 35
Blaine = 3460 +/ - 100 cm2/g
Various Ec values

Static yield (w/c = 0.48), Pa

w/c = 0.48

Static yield (w/c = 0.42), Pa

10
30
w/c = 0.42
9
25
8

20
7

6 15
0

0.03

0.06

0.09

0.11

0.13

0.14

0.03

0.06

0.08

0.1

0.11

0.12
Flow, mm 150 150 160 155 160 165 165 145 160 175 170 175 18 0
(w/c = 0.42)
Ref. Amine-based GA Glycol-based GA

Fig. 7. Variations in 0 and ow for 0.42 and 0.48 w/c pastes prepared using various cement mixtures ground for xed Blaine.

Based on results plotted in Figs. 5 and 6, it can be concluded that the at a rate of 0.11% and 0.14%, respectively. As earlier explained, given that
maximum permissible concentration at xed Blaine neness for the the organic GA molecules arrange their dipoles to saturate the charges
amine-based GA is slightly higher than 0.13%, and around 0.1% for the of the newly formed grains, this would create repulsive forces between
glycol-based GA. neighboring cement particles and result in improved owability and
lower 0 values (as shown in Fig. 7).
4.7. Flow of cement ground for xed Blaine The variations in Initial and Equil values for 0.42 w/c pastes plotted
in Fig. 8 are quite similar to those shown in Fig. 7 for 0. Nevertheless,
The variations of 0 determined on 0.42 and 0.48 w/c pastes using for given GA concentration, it is interesting to note that cement ground
cement ground with different concentrations of amine and glycol- with amine-based GA exhibited lower owability and higher 0 and
based GAs are illustrated in Fig. 7 (variations in ow for 0.42 w/c cement values than those registered when the glycol-based GA is used (Figs. 7
pastes are also shown). Unlike the trends of Fig. 3 obtained for cement and 8). For example, Initial increased from 9.8 to 11.5 Pas when the gly-
possessing different neness levels, the 0 appears to decrease when col or amine-based GAs were used, respectively, at a rate of 0.11% in
the cement is ground for xed Blaine using increased GA concentra- 0.42 w/c pastes (the corresponding ow values were 175 and 160 mm,
tions. For example, at w/c of 0.42, a decrease from 30.4 Pa for the control respectively). When cement comes into contact with water, it is the
cement to 28.8 and 25.6 Pa is noted when the amine-based GA was used aluminate phases (C3A and C4AF) that react rst to form a gel based on

18 4.5
Initial viscosity Blaine = 3460 +/ - 100 cm2/g
w/c = 0.42
Equilibrium viscosity
16 Various Ec values
Equilibrium viscosity, Pa.s

4.0
Initial viscosity, Pa.s

14

3.5

12

3.0
10

8 2.5
0

0.03

0.06

0.09

0.11

0.13

0.14

0.03

0.06

0.08

0.1

0.11

0.12

Ref. Amine-based GA Glycol-based GA

Fig. 8. Variations in Initial and Equil for 0.42 w/c pastes prepared using cement ground with various concentrations of amine and glycol-based GAs, for xed Blaine.
 J.J. Assaad, C.A. Issa / Cement and Concrete Research 63 (2014) 111 9

30 the effect of GAs on variations in ow, two indices were established as

Ec = 42 kWh/ton
Initial viscosity: follows:
Viscosity at 60 rpm, Pa.s

25 Various Ec values y = 0.467 x (R2 = 0.94)

0 cement containing GA0 control cement
20 0 ; %  100
48 data points are
0 control cement
used for each plot h i h i
15
Initial Equil cement containing GA Initial Equil control cement
Equilibrium viscosity: ; % h i
10 y = 0.086 x + 1.447 Initial Equil control cement
(R2 = 0.96)
5
Two thresholds of 50% and 25% based on the range of data
0 obtained from this experimental program were placed for these indices.
0 10 20 30 40 50 60
The former value of 50% reects a wide relative tolerance in variations
Static yield stress at 0.3 rpm, Pa
of ow due to the addition of GAs. Conversely, the 25% can be consid-
ered as a rigorous requirement, so as to minimize changes in fresh
Fig. 9. Relationships between 0 with respect to Initial and Equil values for all tested pastes.
cement properties resulting from eventual variations in ow.
The relationship between (0) and () for all tested cement
pastes at xed or various Ec is plotted in Fig. 10, along with the 25%
and 50% threshold regions. Only 3 data points among those fullling
complex sulfoaluminate hydrates. This gel exerts a barrier effect and gov- ASTM C465 requirements are shown to deviate from the 50% thresh-
erns the mass ow between the inner part of the cement grain and pore old region. Those points correspond to the cement ground at xed Ec
water, thus controlling the rheological behavior and hydration of the with amine-based GA at high concentrations of 0.11% and 0.13%. Never-
silicate phases [16,30]. Given that TEA and TIPA have been identied to theless, when the threshold is reduced to 25%, a signicant number
rapidly react with the aluminate phases [6,8] (especially when present reaching around 35% of the data points fullling ASTM C465 require-
at relatively high percentage in the tested clinker (see Table 1)), this ments fall outside the region. This indicates that the 25% allowable
may increase viscosity of the interstitial phase including gel structuration variations in ow become prevalent over the physical ASTM C465
and formation of colloidal crystals between connected cement grains. requirements for water demand, setting time, and/or compressive
This therefore explains the reduction in owability and increase in 0 strength. In other words, the maximum permissible GA concentrations
and measurements for pastes prepared with cement ground with determined as per ASTM C465 should be revised to reect acceptable
amine-based GA, as compared to those containing glycol-based GA. variations in ow. In the case of GAs tested in this study, the maximum
amine and glycol-based GA dosages should thus be in the order of 0.09%
and 0.11%, respectively, at xed Ec. The corresponding dosages should
4.8. Effect of GAs on ow and comparison with ASTM C465 limitations be around 0.13% and 0.08%, respectively, for cement ground for xed
Blaine neness.
Good relationships with high correlation coefcients (R2) can be As shown in Figs. 11 and 12, the prediction of (0) and () indices
established between 0 with respect to Initial and Equil, as shown can well be realized from the variations in Blaine and R-38 values. The
in Fig. 9 for all pastes prepared with cement ground at xed or (Blaine) is calculated as the Blaine determined using cement contain-
variable Ec. Practically, this indicates that the effect of GAs on 0 and ing GA minus the one determined using control cement, divided by the
measurements is independent from the rotational speed or testing Blaine of control cement, and multiplied by 100. The (R-38) is calculat-
protocol used. Also, a fairly good correlation exists between 0 and ed similarly. Hence, for example, to maintain the (0) and () indices
ow values for all tested pastes at xed or variable Ec, given as: within a region of 25%, the (Blaine) and (R-38) values should vary
Flow (mm) = (2 0, Pa) + 216.2 having R2 of 0.85. Hence, the from around 5% to +15% and from 40% to +50%, respectively. It is
higher the ow, the lower the 0 values. In order to better quantify to be noted that the relationship between the variations in Blaine and

100 10 data points rejected by

ASTM C465
75 38 data points fulfilling ASTM
C465 requirements y = 0.92 x
50
R = 0.87
(Viscosity), %

25

-25

-50
Threshold Threshold
-75 of 25%
of 50%
-100
-100 -75 -50 -25 0 25 50 75 100
(Static yield stress), %

Fig. 10. Relationship between (0) and () for all tested pastes with respect to the proposed 25% and 50% threshold regions.
10 J.J. Assaad, C.A. Issa / Cement and Concrete Research 63 (2014) 111

100
96 data points:
75 y = 2.82 x - 12.1
R = 0.84
50

Flow indices, %
25

-25 (Static yield stress)

-50 (Viscosity)

-75
-15 -5 5 15 25 35
(Blaine), %

Fig. 11. Relationship between (Blaine) with respect to ow indices.

R-38 values is given as: (Blaine) = 0.25 (R-38) + 1.3, having R2 When grinding is realized at various Ec to maintain xed Blaine ne-
of 0.76. ness, a gradual increase in setting time was noted with the use of higher
GA concentration. The corresponding allowable dosages as per ASTM
5. Conclusions C465 for the amine and glycol-based GA were slightly higher than
0.13% and around 0.1%, respectively. The water demand tended to
Based on the above results, the following conclusions can be decrease when higher GA dosages were used, because of the repulsive
warranted: forces created through GA adsorption onto the cement particles.
An improvement in owability together with reduction in 0 and
At xed Ec, water demand required for normal consistency increased measurements were observed when grinding is realized for xed
progressively with the GA concentration due to higher cement ne- Blaine neness. This was related to a dispersion mechanism of
ness, until exceeding the ASTM C465 limitation. The maximum per- cement agglomerates due to the GA molecules adsorbed onto the ce-
missible GA concentrations as per ASTM C465 are slightly lower ment particles. Compared to amine-based GA, the use of glycol-based
than 0.13% or 0.11% for the amine or glycol-based GA, respectively. GA was found more efcient to disperse the cement agglomerates;
Cement mixtures ground at xed Ec with additional GA dosages re- however, resulted in sharp drops in compressive strengths.
sulted in reduced setting times and improved compressive strength. The (0) and () are suitable indices to quantify variations in ow
Regardless of the GA type, cement mixtures ground at xed Ec result- due to the addition of GAs. Only 3 data points among those fullling
ed in increased 0 and measurements. This was attributed to a phys- ASTM C465 requirements were found to deviate from a threshold
ical effect related to higher packing density and increased formation set at 50%. However, when the threshold is reduced to 25%, a
rates of early hydrates that increase inter-particle links within the signicant number reaching around 35% was no longer complying
suspension. The rate of 0 and increase was directly affected by to ASTM C465. This suggests that the variations in ow become prev-
the cement neness, which in its turn, is dependent on the GA alent over the physical requirements such as water demand, setting
concentration. time, and compressive strength.

100
(Static yield stress)
75
(Viscosity) 96 data points:
y = 0.006 x 2 - 0.61 x - 14.2
50 R = 0.83
Flow indices, %

25

-25

-50

-75
-80 -60 -40 -20 0 20 40 60
(R-38), %

Fig. 12. Relationship between (R-38) with respect to ow indices.

 J.J. Assaad, C.A. Issa / Cement and Concrete Research 63 (2014) 111 11

Acknowledgments hydration, 12th Int Congress on the Chemistry of Cement, International Publishers,
Canada, July 2007, pp. 16.
[15] A. Bravo, T. Cerulli, M. Giarnetti, M. Magistri, Grinding aids: a study on their mech-
This project is funded by the University Research Council of the anism of action, 11th International Congress on the Chemistry of Cement, 2003,
Lebanese American University (LAU), Byblos, Lebanon. The authors pp. 11001110.
[16] Anna, B., Tiziano, C., Mariagrazia, G., and Matteo, M., Grinding aids: A study on their
wish to acknowledge the experimental support provided by the Labora- mechanism of action, http://www.mapei.com/dam/Pdf/ConferencesGrinding.pdf,
tory personnel of Holderchem Building Chemicals, Baabda, Lebanon. 2004, pp. 15.
[17] M. Katsioti, P.E. Tsakiridis, P. Giannatos, Z. Tsibouki, J. Marinos, Characterization of
various cement grinding aids and their impact on grindability and cement perfor-
References mance, Constr. Build. Mater. vol. 23 (2009) 19541959 (No. 5).
[18] NCHRP Report 607, Specications and protocols for acceptance tests on processing
[1] J.J. Assaad, S. Asseily, J. Harb, Effect of specic energy consumption on cement additions in cement manufacturing, Executive Committee, Transportation Research
neness incorporating amine and glycol-based grinding aids, Mater. Struct. 42 (8) Board 2008, Washington, USA, 2008, ISBN 978-0-309-09934-9. 196.
(2009) 10771087. [19] ASTM C465, Standard Specication for Processing Additions for Use in the Manufac-
[2] I. Teoreanu, G. Guslicov, Mechanisms and effects of additives from the dihydroxy- ture of Hydraulic Cements, ASTM Int, 2010. (Document number ASTM C465-10).
compound class on cement grinding, Cem. Concr. Res. 29 (1) (1999) 915. [20] M. Tokyay, Effect of chemical composition of clinker on grinding energy require-
[3] C.J. Engelsen, Quality improvers in cement making state of the art, SINTEF ment, Cem. Concr. Res. 29 (4) (1999) 531535.
Building and Infrastructure, COIN Project Report 2, ISBN: 978-82-536-1071-9, [21] ASTM C204, Standard Test Methods for Fineness of Hydraulic Cement by Air-
2008, pp. 124. Permeability Apparatus, ASTM Int, 2011. (Document number ASTM C204-11).
[4] ASTM C150, Standard Specication for Cement, ASTM Int, 2012. (Document number [22] ASTM C186, Standard Test Method for Heat of Hydration of Hydraulic Cement,
ASTM C150-12). ASTM Int, 2005. (Document number ASTM C186-05).
[5] A.T. Albayrak, M. Yasar, M. Gurkaynak, I. Gurgey, Investigation of the effects of [23] ASTM C187, Standard Test Method for Amount of Water Required for Normal
fatty acids on the compressive strength of the concrete and the grindability of the Consistency of Hydraulic Cement Paste, ASTM Int, 2011. (Document number
cement, Cem. Concr. Res. 35 (2005) 400404. ASTM C187-11e1).
[6] V.S. Ramachandran, Hydration of cement role of triethanolamine, Cem. Concr. [24] ASTM C191, Standard Test Methods for Time of Setting of Hydraulic Cement by Vicat
Res. 6 (5) (1976) 623631. Needle, ASTM Int, 2008, (Document number ASTM C191-08).
[7] Z. Heren, H. Olmez, The inuence of ethanolamines on the hydration and mechanical [25] ASTM C230, Standard Specication for Flow Table for Use in Tests of Hydraulic
properties of cement, Cem. Concr. Res. 26 (5) (1996) 701705. Cement, ASTM Int, 2008. (Document number ASTM C230-08).
[8] J.P. Perez, A. Nonat, S. Pourchet, M. Garrault, C. Canevet, Why TIPA leads to an [26] D.Q. Nguyen, D.V. Boger, Direct yield stress measurement with the vane method,
increase in the mechanical properties of mortars whereas TEA does not, ACI J. Rheol. 29 (3) (1985) 335347.
Mater. J. vol. 217 (2003) 583594 (SP217-38). [27] J. Assaad, K.H. Khayat, H. Mesbah, Assessment of thixotropy of highly owable and
[9] M. Ichikawa, M. Kanaya, S. Sano, Effect of triisopropanolamine on hydration and self-consolidating concrete, ACI Mater. J. 100 (2) (2003) 99107.
strength development of cements with different character, in: A.B. Amarkai, A.B. [28] ASTM C109, Standard Test Method for Compressive Strength of Hydraulic Cement
Goteborg (Eds.), Proceedings of the 10th International Congress on the Chemistry Mortars (Using 2-in. or [50-mm] Cube Specimens), ASTM Int, 2012. (Document
of Cement, Goteburg Publisher, Gothenburg, Sweden, 1997, (10 pp.). number ASTM C109-12).
[10] P. Sandberg, F. Doncaster, On the mechanism of strength enhancement of cement [29] S. Ahmad, L.L. Qureshi, Variation in neness of cement and its effects on other
paste and mortar with triisopropanolamine, Cem. Concr. Res. 34 (6) (2004) 973976. properties of cement and concrete, 29th Conference on Our World in Concrete &
[11] G.H. Tattersall, The rheology of portland cement pastes, Br. J. Appl. Phys. 6 (1955) Structures, 2004, (August, Singapore, http://cipremier.com/100029012, 9 pp.).
165167. [30] D.P. Bentz, O.M. Jensen, K. Hansen, J. Olesen, H. Stang, C. Haecker, Inuence of
[12] P.F.G. Banll, Rheology of fresh cement paste and concrete, Rheology Reviews, The cement particle size distribution on early age autogenous strains and stresses in
British Society of Rheology, 2006. 61130. cement-based materials, J. Am. Ceram. Soc. 84 (1) (2001) 129135.
[13] I. Aiad, A.A. Mohammed, S.A. Abo-El-Enein, Rheological properties of cement pastes [31] N. Roussel, G. Ovarlez, S. Garrault, C. Brumaud, The origins of thixotropy of fresh
admixed with some alkanolamines, Cem. Concr. Res. 33 (2003) 913 (No. 1). cement pastes, Cem. Concr. Res. 42 (1) (2012) 148157.
[14] C. Jolicoeur, S. Morasse, J. Sharman, A. Tagnit-Hamou, F. Slim, M. Page, Polyol-type [32] H.H.C. Wong, K.H. Kwan, Rheology of cement paste: role of excess water to solid
compounds as clinker grinding aids: inuence of powder uidity and on cement surface area ratio, J. Mater. Civ. Eng. 20 (2008) 189197.