See

discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/257388947

Determination of cement hydration and

pozzolanic reaction extents for fly-ash cement
pastes

Article in Construction and Building Materials · February 2012

DOI: 10.1016/j.conbuildmat.2011.07.007

CITATIONS READS

40 555

4 authors, including:

Qiang Zeng Kefei Li

Zhejiang University Tsinghua University
48 PUBLICATIONS 474 CITATIONS 111 PUBLICATIONS 858 CITATIONS

SEE PROFILE SEE PROFILE

Patrick Dangla
University of Paris-Est
116 PUBLICATIONS 1,507 CITATIONS

SEE PROFILE

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

Shenzhen-Zhongshan Sea link project, service life design and analysis for prefabricated concrete slab
View project

Poromechanical study of multiphase transition in cement-based porous materials View project

All content following this page was uploaded by Qiang Zeng on 16 October 2017.

The user has requested enhancement of the downloaded file.

 Construction and Building Materials 27 (2012) 560–569

Contents lists available at ScienceDirect

Construction and Building Materials

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

Determination of cement hydration and pozzolanic reaction extents for fly-ash

cement pastes
Qiang Zeng a,b, Kefei Li a,⇑, Teddy Fen-chong b, Patrick Dangla b
a
Civil Engineering Department, Tsinghua University, Beijing 100084, PR China
b
Université Paris-Est, Institute Navier, UMR CNRS–LCPC–ENPC, Champs sur Marne 77420, France

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

Article history: This article intends to quantify the reaction extents of cement grains and fly-ash (FA) particles in blended
Received 15 January 2011 cement pastes during their hardening process. To this aim, a synthetic model for cement hydration extent
Received in revised form 1 June 2011 in blended pastes is established taking into account the dilution effect, local w/c augmentation effect as
Accepted 13 July 2011
well as heterogeneous nucleation effect of FA particles. Then the cement hydration and FA reaction
Available online 19 August 2011
extents of blended cement pastes with two w/b ratios (0.3, 0.5) and four FA contents (0%, 20%, 40%,
60%) are investigated. The non evaporable water content (Wn) and calcium hydroxide content (CH) were
Keywords:
measured by thermal gravity analysis (TGA) and the FA reaction extent was quantified by the selective
Cement paste
Fly-ash
dissolution method. On the basis of these experimental data, the cement hydration extent and FA pozzo-
Hydration extent lanic reaction extent are determined from both experimental approach and model-based approach. From
Non-evaporable water the results, it is observed that local w/c, heterogeneous nucleation effects and FA hydration have compa-
Pozzolanic reaction rable contribution to the total Wn content in blended pastes while the FA hydration dominates in CH con-
tent. Furthermore, the good agreement between the experimental hydration extents and model-based
extents validates the established synthetic model for cement hydration extent in blended pastes.
Ó 2011 Elsevier Ltd. All rights reserved.

1. Introduction strength for cement-based materials. Fly-ash is believed to have

pozzolanic reaction after the age of 3d as blended with cement [3].
Fly ash (FA), a by-product of thermal power industry, has been This pozzonlanic reaction is much slower than cement hydration:
widely used as a supplementary cementitious material (SCM) in con- it is reported that FA reaction extent is just about 25% at the age of
crete technology for its contribution to the fresh concrete workabil- 360d for a blended paste having w/b = 0.5 and FA content of 40%
ity, microstructure formation as well as long term mechanical and [4]. Meanwhile, adding fly-ash into cement–water system promotes
durability properties. The dosage of FA in binder materials is usually the hydration of cement grains by two mechanisms: (1) under same
adapted to different engineering applications and different curing w/b ratio the replacement of cement by FA augments the water to ce-
conditions. For instance, for prestressed concrete and reinforcement ment ratio (w/c) thus promotes the hydration extent of cement by
concrete with early exposure to environment, the fly-ash content, more water availability [5,6]; (2) the FA sphere particles adsorb
mass ratio of fly-ash in binder, is limited to 25% and 30% according Ca2þ and provide precipitation sites for cement hydrates, such as
to practice regulations [1]. For underground environment the FA ettringite (AFT) and calcium hydroxide (CH) [5], thus help to accel-
content can be as high as 50% and this value can reach 70% for roller erate the cement grain hydration process [7–9]. Furthermore, recent
compacted concrete (RCC), used for dam engineering [2]. research reveals that the main hydrate of cement and fly-ash, cal-
Fly-ash is a pozzolana additive, not capable to react with water cium silicate hydrate (C–S–H), adopt two distinct morphologies: a
without an alkaline activator. The role of fly-ash during concrete low density C–S–H at the surface of cement and FA particles and a
hardening process is relatively well understood nowadays. Firstly, high density C–S–H deeper into the particles [10,11].
the spherical geometry of fly-ash particles, acting as lubricating Although the hydration mechanisms of cement-FA-water system
agent, helps to improve the rheology properties of fresh paste before are well understood, the hydration extent quantification for cement
its setting. Secondly, the hydration of fly-ash, named as pozzolanic and FA is not yet clarified, neither by experimental means nor on
reaction, consumes the cement hydrates (portlandite) and fills the theoretical basis. The commonly used experimental methods for
already-formed pore structure by its own hydration product, thus identifying the cement hydration extent include the backscatter
helping to achieve a more compact pore structure and higher electron image analysis (BSE/IA) [12], the thermal gravity analysis
(TGA) [13] and XRD/Rietveld Analysis [12,14,15]. The specific
⇑ Corresponding author. experimental methods to quantify the pozzolanic reaction extent in-
E-mail address: likefei@tsinghua.edu.cn (K. Li). clude the selective dissolution method [16,17], TGA and BSE/IA [18].

0950-0618/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved.
doi:10.1016/j.conbuildmat.2011.07.007
 Q. Zeng et al. / Construction and Building Materials 27 (2012) 560–569 561

Among these methods, TGA can provide reliable measurement on For blended cement pastes incorporating inert fillers or SCM
the amount of non-evaporable water (Wn) and the content of with mass fraction of fp, the w/c ratio can be converted from w/b
calcium hydroxide (CH) in blended cement pastes. According to ratio through,
the hydration mechanisms of cement grains and FA particles, both
1
reactions contribute to Wn but FA reaction consumes CH from w=c ¼ w=b ð2Þ
1  fp
cement hydration. With the chemical equations of cement hydra-
tion and FA reaction the respective reaction extent for cement and Obviously, under a same w/b ratio the blended cement pastes
FA can be determined. (fp > 0), compared to the neat cement paste (fp = 0), have larger
Following this approach, this article investigates the cement cement hydration extent. This hydration extent augmentation,
hydration and FA pozzolanic reaction (hydration) extents on the purely due to the ‘‘dilution effect’’ of additives, can be expressed as,
basis of TGA results on Wn and CH. To help the results interpreta-
Daw=c ¼ ac  a0 ð3Þ
tion, a synthetic model for FA effect on cement hydration extent is
firstly proposed in Section 2. Then materials and experimental where a0 is the hydration extent of neat cement paste. Combining
details are given in Section 3. Afterwards, the hydration extents Eqs. (1) and (3) gives,
of cement grains and FA particles in blended pastes are determined  
in Section 4 from both experimental and model-based approaches,
Daw=c 1  fp
¼ exp b 1 ð4Þ
and the concluding remarks are given in the end. a0 w=b

2. Synthetic model for cement hydration extent 2.2. Effect of heterogeneous nucleation

2.1. Effect of w/c ratio The heterogeneous nucleation effect of FA contributes to the ce-
ment hydration extent by surface absorption and precipitation.
As a known fact, large w/c ratio creates in cement pastes a pore This effect is greatly influenced by the particle size or specific area
structure with high porosity, high connectivity and low strength. of FA. Based on the experimental results, Lawrence et al. [8] ob-
However, the cement hydration extent (degree) is more advanced tained the following expression to account for the heterogeneous
due to more water availability [19]. Classic regression model links nucleation (HN) effect of FA,
the cement hydration extent ac with w/c ratio through the
DaHN w=b
following relation [13], ¼ ð5Þ
  a0 1 þ ðhc =Seff Þht
b
ac ¼ a1 exp  ð1Þ where DaHN is the hydration extent augmentation by HN effect, hc
w=c
is a constant (8300 m2/kg) and ht is assumed to be dependent on
where ac,1 are the cement hydration extent and ultimate hydration curing age. The term Seff is the effective surface area (m2/kg), ex-
extent, the constant b characterizes the sensitivity of hydration ex- pressed as a function of FA specific area Sp, FA mass fraction fp
tent with respect to w/c ratio. In Fig. 1 are summarized the available and weight function e [8]:
results in the literature on the ac  w/c relation, with the regressed
fp
constants a1, b given in Table 1. Seff ¼ Sp eðfp Þ ð6Þ
1  fp
with the weight function e expressed as [8]:
 0:7 "  3:4 #1
1 þ cosðfp  pÞ fp
e¼  1þ ð7Þ
2 36:8

The ht parameter is regressed in this article by an exponential func-

tion in terms of curing age tc,
 
h2
ht ¼ 1  h1 exp  ð8Þ
tc
with h1,2 calibrated as 0.612, 0.370 from data in [8].

2.3. Combined effect of w/c and heterogeneous nucleation

For Portland cement blended with lime stone powder, the total
effect of hydration extent augmentation, compared to the hydra-
tion extent of neat paste a0 with same the w/b ratio, can be written
Fig. 1. Relation between ac and w/c ratio for different curing ages of cement pastes as the sum of Daw/c and DaHN,
[13,20–24].
ac ¼ a0 ½1 þ Daw=c þ DaHN  ð9Þ
Substitution of Eqs. (1) and (5) into Eq. (9) gives,
Table 1  
Parameter regression for ac  w/c relation from data in Fig. 1. ac 1  fp w=b
¼ exp b þ ð10Þ
Curing age Ultimate hydration Constant b Correlation a0 w=b 1 þ ðhc =Seff Þht
(d) extent a1 () (–) coefficient R2 (–)
This synthetic model is validated by the hydration extent measure-
7 0.742 0.113 0.883
ment of cement in blended pastes from Bonavetti et al. [25]. Fig. 2
28 0.899 0.120 0.938
90 0.985 0.127 0.942 shows the predicted ac through Eq. (10) compared to the measure-
ments, the relevant parameters presented in Table 2. It can be seen
562 Q. Zeng et al. / Construction and Building Materials 27 (2012) 560–569

Table 4
Mixture proportioning for blended cement pastes.

Paste w/b (–) Sample Cement Fly-ash Water fp (–)

(kg/m3) (kg/m3) (kg/m3)
Paste I 0.5 PI0 1218 0 609 0%
PIF1 937 234 588 20%
PIF2 677 451 559 40%
PIF3 435 650 543 60%
Paste II 0.3 PII0 1612 0 484 0%
PIIF1 1224 306 459 20%
PIIF2 874 583 437 40%
PIIF3 556 834 417 60%

recommended procedure avoiding the hydrates decomposition [30]. Then the dried
samples were subject to TGA test by a device of 2050/TA type. The Wn values were
obtained from the weight loss of samples between 105 °C and 900 °C, illustrated in
Fig. 3, and calculated as [13,20],
m900  m105
Wn ¼  ðfc  LOIc þ fp  LOIp Þ ð11Þ
Fig. 2. Predicted ac by synthetic model versus cement hydration extent measure- m900
ments from Bonavetti et al. [25].
where m105 and m900 are respectively the sample weights measured at 105 °C and
900 °C, fc,p the mass fractions of cement and FA in binder and LOIc,p the ignition
losses of cement and FA. The subscripts c, p represent the cement and pozzolana
Table 2 without special mention in this article. The maximum temperature in TGA of this
Regression parameters in Eq. (10) for evaluating cement hydration extent in Fig. 2. study was 900 °C, lower than the complete release temperature for non-evaporable
Curing age (d) w/b (–) hc (m2/kg) ht (–) water (1000 °C), but the weight loss higher than 850 °C was found to be insignificant
[31]. The CH values were evaluated by sample weight loss during TGA at the decom-
1 0.25–0.5 8300 0.586 position temperature of CH (about 450 °C), see Fig. 3b and d. The details of CH
3 0.25–0.5 0.459 amount evaluation from TGA curves can be found in [31].
7 0.25–0.5 0.419 A selective dissolution procedure using picric acid and water was adopted to
28 0.25–0.5 0.396 determine the pozzolanic reaction extent of fly-ash, conforming to the guide of
ACI Committee 211 [32]. The principle of selective dissolution is that the minerals
in cement grain, cement hydrates and FA hydrates are all acid-soluble and only the
that very good agreement is obtained, thus this synthetic model is unreacted part of fly-ash particles is insoluble in picric acid [16,17]. In this study,
this assumption was further corrected by a reference test on the residue measure-
capable to describe the incremental effect of non-pozzolanic addi-
ment of picric dissolution on cement grains and FA particles. These values were
tives on cement hydration extent. used as the basis for hydration extent evaluation. Then the powder samples of
blended pastes were subject to picric acid dissolution test and the pozzolanic reac-
3. Materials and experiments tion extent is evaluated by,

Rb  m900 =m0  ð1 þ fp  LOIp þ fc  LOIc Þ  fc Rc

3.1. Materials ap ¼ 1  ð12Þ
m900 =m0  ð1 þ fp  LOIp þ fc  LOIc Þ  fp Rp
A Portland cement and a low calcium fly-ash are retained for this study. The This equation is adapted from the expression from Li et al. [20], taking into account
used cement is PI type according to Chinese standard [26], corresponding to CEM further the ignition losses. In the equation, Rb,p,c represent the dissolution residue of
I in EN197 [27] and Type I in ASTM standard [28]. The chemical composition and sample, unreacted FA and cement respectively, and m0 is the initial sample weight.
physical properties of cement and fly-ash are presented in Table 3. The mineral con-
tents of cement were analyzed, through Bogue’s procedure [29], as C2S (21.38%), C3S
(58.88%), C3A (6.49%), C4AF (8.77%), Gypsum (0.75%) and others (3.73%). The 4. Hydration extents of FA cement pastes
blended cement pastes in study adopt two water to binder ratios (w/b) and four
FA contents (fp), the material proportioning given in Table 4.
4.1. Hydration extents deduced from Wn and CH

3.2. Test methods

For cement–water system, the cement hydration consumes
The TGA method was used to determine the amount of non-evaporable water water and produces calcium hydroxide (CH). Thus both the
(Wn) and calcium hydroxide (CH). Crushed samples from hardened pastes at given chemically bond water (Wn) and CH can represent the hydration
ages (7d, 28d, 90d, 180d) were ground to powder with maximum particle size con- extents. Normally a fully hydrated Portland cement paste can pro-
trolled to 80 lm. Then the powder samples were oven-dried for 48 h at 50 ± 1 °C, a
duce 15–25% CH [13]. As for cement–FA–water system, both ce-
ment hydration and FA reaction contribute to Wn while the FA
reaction consumes the CH from cement hydration. Furthermore,
Table 3
Chemical composition and physical properties of cement and fly-ash.
the presence of FA tends to increase the cement hydration extent
as previously analyzed. So, the total Wn and CH amounts,
Chemical composition/physical properties Cement Fly-ash W tn ; CHt , can be expressed as,
Silica (SiO2, %) 22.93 57.60
Alumina (Al2O3, %) 4.29 21.90 W tn ¼ W cn fc þ W pn fp ð13aÞ
Iron oxide (Fe2O3, %) 2.89 2.70 CHt ¼ CHc fc þ CHp fp ð13bÞ
Calcium oxide (CaO, %) 66.23 7.80
Magnesium oxide (MgO, %) 1.92 1.68 where W cn and W pn are, respectively, the amounts of non-evaporable
Sulfur trioxide (SO3, %) 0.35 0.41
Sodium oxide (Na2O (eq), %) 0.70 1.05
water generated by hydration of unit weight cement and FA, CHc is
Free calcium oxide (CaO (f), %) 0.64 – the CH amount by hydration of unit weight cement and CHp is the
Chloride (Cl, %) 0.006 – CH amount consumed by hydration of unit weight FA, adopting neg-
Loss on ignition (LOI, %) 1.70 7.05 ative values. From cement chemistry, 1 g Portland cement, as com-
Density (g/ml) 3.12 2.06
pletely hydrated, produces 0.22–0.27 g non evaporable water
Specific area (m2/kg) 343 355
[33,34]. Following the stoichiometry analysis of hydration reactions,
 Q. Zeng et al. / Construction and Building Materials 27 (2012) 560–569 563

Fig. 3. TGA results for paste samples at age of 90d: (a) weight loss of PI samples, (b) differential weight loss of PI samples, (c) weight loss of PII samples and (d) differential
weight loss of PII samples.

Fig. 4. Amount of non evaporable water for Paste I and Paste II samples at different ages: filled circle () for measured Wn, solid line (–) for diluted Wn by inert filler fp, dash
line (- -) for increased Wn by local w/c effect, dot line (  ) for increased Wn by both w/c and heterogeneous nucleation effects.
564 Q. Zeng et al. / Construction and Building Materials 27 (2012) 560–569

presented in Appendix A, the following values are obtained: where CHc(1) = 0.204, CHp(1) = 0.853 are estimated by stoichi-
W cn ð1Þ ¼ 0:272; W pn ð1Þ ¼ 0:168. Once the non evaporable water ometry analysis of cement and FA hydrations in Appendix A. Eqs.
W cn ; W pn are determined, the cement hydration extent ac and FA reac- (13)–(15) provide actually an experimental evaluation approach
tion extent ap are, for hydration extents ac,p, and this method is to be used later in
the hydration extent quantification.
W cn W pn
ac ¼ ; ap ¼ ð14Þ Using Eq. (10) and considering the proportionality in Eqs. (14)
W cn ð1Þ W pn ð1Þ and (15), W cn and CHc can also be determined from the known
From the same reasoning, the hydration extents can also be ex- W 0n ; CH0 for neat paste with the same w/b ratio,
pressed in terms of CH content,  
W cn CHc 1  fp w=b
CHc CHp ¼ ¼ exp b þ ð16Þ
ac ¼ ; ap ¼ ð15Þ W 0n CH0 w=b 1 þ ðhc =Seff Þht
CHc ð1Þ CHp ð1Þ
The cement hydration extent ac can thus be determined. Further-
more, as W cn ; CHc are known, the FA hydration extent ap can be eval-
uated from Eq. (13b) as,

W tn  W cn fc CHt  CHc fc
ap ¼ ¼ ð17Þ
W pn ð1Þfp CHp ð1Þfp

Eqs. (16) and (17) provide another way to evaluate the hydration
extents on the basis of the synthetic model. This model-based ap-
proach is to be also used in quantifying the hydration extents.

4.2. Wn and CH results

The measurement of Wn for all samples is illustrated in Fig. 4,

figured out by black filled circles. For neat pastes (PI0, PII0), W 0n in-
creases with curing age. For PI0 (w/b = 0.5), W 0n augments by 60%
from 7d to 90d and by 70% for PII0 (w/b = 0.3). For all ages the
W 0n of PI0 is systematically higher than PII0 due to higher w/b con-
tribution to cement hydration. For the blended pastes PI1-3 and
PII1-3, the total non evaporable content, W tn , increases with curing
age for each FA content fp, and decreases with FA content. The dilu-
Fig. 5. Contribution of pure dilution effect (solid line), local w/c effect (dark gray tion effect of FA, taken as inert fillers, on W tn is illustrated by blue
area), heterogeneous nucleation (light gray area) and FA hydration (grid-filled area) lines in Fig. 4. The dash green lines take into account the effect of
to W tn of PI samples at 28d. local w/c augmentation by FA replacement using the model in

Fig. 6. CH content for Paste I and Paste II samples at different ages: filled circle () for measured CH, solid line (–) for diluted CH by inert filler fp; dash line (- -) for increased
CH by local w/c effect, dot line (  ) for increased CH by both w/c and heterogeneous nucleation effects.
 Q. Zeng et al. / Construction and Building Materials 27 (2012) 560–569 565

Eq. (4). Furthermore, the dash green lines consider the heteroge- line and measured CH contents is due to the FA hydration con-
neous nucleation effect by Eq. (5). A more detailed illustration is gi- sumption CHp.
ven in Fig. 5 for different contributions to W tn . The difference
between the dash green line and the measured W tn can be attrib-
uted to the FA hydration contribution W pn . 4.3. Synthetic effect of FA on blended cement hydration
The CH content of all cement pastes is illustrated in black filled
circles in Fig. 6. It can be seen that the CH content of Paste I sam- As previously described, the synthetic effects of FA in the hydra-
ples is larger than that of Paste II samples due to larger hydration tion process of blended cement pastes include the pure dilution ef-
extent of cement. Different from the observation on W tn values, the fect, local w/c augmentation effect as well as heterogeneous
CH content does not show clear correlation with curing age due to nucleation effect. Taking the pure dilution effect (diluted cement
the opposite contribution of cement hydration and FA hydration to grains by inert fillers) as the reference state, blue lines in Figs. 4
CH content. The pure dilution effect of FA, local w/c effect and the and 6, the incremental amounts of non-evaporable water (DWn)
heterogeneous nucleation effect on CH content are illustrated in and CH content (DCH), compared to the corresponding neat paste,
Fig. 6, respectively, by blue lines, red dash lines and green dash can be splitted into three contributions,
lines. The relative importance of these factors on CH content is pre-
sented in details in Fig. 7. The difference between the dash green DW n ¼ ðDW n Þw=c þ ðDW n ÞHN þ W pn ð18aÞ
DCH ¼ ðDCHÞw=c þ ðDCHÞHN þ CHp ð18bÞ

The third terms at the right side of Eq. (18) represent the contribu-
tion or consumption effect of FA reaction on Wn and CH contents,
illustrated by the area between the red dash line and black circle
line in Figs. 5 and 7. As aforementioned, DWn is directly from mea-
surement (DWn)w/c,HN are evaluated respectively from Eqs. (4), (14)
and Eqs. (5), (14), and W pn is deduced from Eq. (18a). The values of
DWn, ðDW n Þw=c;HN ; W pn are presented in Table 5. The four terms in Eq.
(18a) show slight increment with curing age and the maximum va-
lue of DWn is reached for FA content of 40%. Similar observations
were also obtained in the literature [8,9].
The values of DCH are measured, the terms (DCH)w/c,HN are
evaluated respectively through Eqs. (4), (15) and Eqs. (5), (15),
and CHp is evaluated from Eq. (18b). These values are presented
in Table 6. The values of CHp are negative to denote the CH con-
sumption. At the age of 7d, DCH values are slightly negative for
Paste I samples but positive for Paste II samples, due to the more
advanced FA reaction extent in Paste I samples. At the age of
Fig. 7. Contribution of pure dilution effect (solid line), local w/c effect (light gray 90d, with the advancement of FA reaction the DCH values are all
area), heterogeneous nucleation (grid-filled area) and FA hydration (sum of the negative, meaning the FA pozzolanic reaction has substantially de-
three areas) to CH content of PI samples at 90d.
creased the CH content in hardened pastes.

Table 5
Contribution of fly-ash content to non evaporable water amounts at 7d (28d)[90d].

w/b (–) FA Wn total change w/c effect HN effect FA reaction

fp (%) w/c (–) DWn (103) (DWn)w/c (103) (DWn)HN (103) W pn (103)

0.5 0 0.50 – (–) [–] – (–) [–] – (–) [–] – (–) [–]
20 0.63 18.3(21.2)[36.2] 5.7(6.9)[10.1] 9.9(11.2)[15.7] 4.3(4.9)[16.4]
40 0.83 19.5(25.2)[44.3] 8.7(10.5)[15.7] 7.5(8.5)[11.8] 5.3(9.9)[26.8]
60 1.25 18.6(21.5)[40.1] 8.9(10.8)[16.1] 4.1(4.7)[6.5] 8.8(9.6)[27.5]
0.3 0 0.30 – (–) [–] – (–) [–] – (–) [–] – (–) [–]
20 0.38 16.9(26.8)[36.1] 8.4(12.1)[15.9] 5.2(7.0)[8.6] 5.3(12.6)[18.6]
40 0.50 23.4(47.4)[46.6] 13.1(18.9)[24.9] 3.9(5.3)[6.5] 9.4(23.2)[24.2]
60 0.72 25.3(33.2)[41.0] 13.6(19.7)[26.0] 2.1(2.9)[3.6] 15.5(10.5) [18.4]

Table 6
Contribution of fly-ash content to CH amounts at 7d (28d)[90d].

FA CH total change w/c effect HN effect FA reaction

w/b (–) fp (%) w/c (–) DCH (103) (DCH)w/c (103) (DCH)HN (103) CHp (103)
0.5 0 0.50 – (–) [–] – (–) [–] – (–) [–] – (–) [–]
20 0.63 5.5 (3.9) [12.2] 6.9 (6.9) [8.0] 12.0 (11.3) [12.3] 24.3 (14.3) [32.5]
40 0.83 0.3 (8.3) [18.1] 10.6 (10.6) [12.3] 9.0 (8.5) [9.3] 19.9 (10.8) [39.7]
60 1.25 0.6 (1.1) [32.0] 10.8 (11.0) [12.7] 5.0 (4.7) [5.1] 16.4 (16.8) [49.8]
0.3 0 0.30 – (–) [–] – (–) [–] – (–) [–] – (–) [–]
20 0.38 2.2 (22.6) [11.1] 8.4 (7.1) [10.2] 5.1 (4.1) [5.5] 11.3 (11.4) [26.8]
40 0.50 8.0 (17.4) [1.5] 13.1 (11.1) [15.9] 3.9 (3.1) [4.2] 9.0 (3.1) [21.6]
60 0.72 11.1 (13.2) [33.1] 13.6 (11.7) [16.7] 2.2 (1.7) [2.3] 4.6 (26.6)[52.1]
566 Q. Zeng et al. / Construction and Building Materials 27 (2012) 560–569

4.4. Cement hydration extent The experimental approach consists in measuring the mass losses
of ignition and mass residues by selective dissolution method
The experimental and model-based approaches, established in and quantifying the reaction extent ap by Eq. (12). The obtained
Section 4.1, are used to quantify the cement hydration extent ac. values of ap are presented in Fig. 10. It can be observed that ap of
The experimental approach is to solve directly the equation array Paste I samples is systematically higher than that of Paste II sam-
in Eqs. (13a), (13b) and obtain the hydration extent by Eq. (14) ples. This is due to the fact that Paste I samples, w/b = 0.5, provide
or Eq. (15). The ac solved by experimental approach is presented
in Fig. 8a and b. For both Paste I and II samples, the ac augments
with curing age by hydration process. Not surprisingly, the cement
hydration extent ac of Paste I samples is larger than the corre-
sponding Paste II samples, reflecting clearly the effect of higher
w/b ratio on cement hydration. For Paste I or II samples, the cement
hydration extent ac increases systematically with fly-ash content fp
due to both the local w/c augmentation and heterogeneous nucle-
ation effects.
The model-based approach to determine ac consists in solving
Eqs. (16) and (17) with W 0n ; CH0 measured for neat pastes. The re-
sults are presented in Fig. 9 in terms of the ac values from the
experimental approach. It can be observed that the ac values from
the two approaches are correlated to a linear relationship. This
observation validates actually the synthetic model for local w/c
effect and heterogeneous nucleation effect of FA in Eq. (10).

4.5. FA pozzolanic reaction extent

The pozzolanic reaction extent of FA, ap, in blended pastes can Fig. 9. Cement hydration extent ac by experimental approach versus ac values from
model-based approach.
also be evaluated in experimental and model-based approaches.

Fig. 8. Cement hydration extent solved by experimental approach for Paste I Fig. 10. Pozzolanic reaction extent ap determined by dissolution method for Paste I
samples (a) and Paste II samples (b). samples (a) and Paste II samples (b).
 Q. Zeng et al. / Construction and Building Materials 27 (2012) 560–569 567

more pore solution (CH content) for the later FA pozzolanic reac- Analogue to the single effect of w/c illustrated in Fig. 11, the sin-
tion. The ap values decrease with FA content, e.g. ap at fp = 20% is gle effect of heterogeneous nucleation of FA particles on cement
almost double of ap at fp = 60%. The FA reaction extent evolves sub- hydration is presented in Fig. 12a and b, respectively from Wn
stantially with curing age, i.e. ap = 0.05–0.10 at 7d and ap = 0.2–0.4 and CH values. Same observation is obtained for the predicted
at 90d. Furthermore, the ap values are not sensitive to curing age and measured FA reaction extent. Finally, Fig. 13a and b present
within 28d for high w/b (Paste I) samples. the estimated FA reaction extent, through Wn and CH values, in
The model-based approach is to determine W pn or CHp through terms of measured values. Here both local w/c effect and heteroge-
Eq. (18) and derive ap by Eq. (14) or Eq. (15). If only w/c effect is neous nucleation effect are considered in Eq. (18). Compared with
taken into account in Eq. (18a), the derived ap is presented in Figs. 11 and 12, much better agreement is found in Fig. 13a and b,
Fig. 11a in terms of measured pozzolanic reaction extent. An over- validating again the synthetic model for cement hydration in
estimation of pozzolanic reaction extent is observed, and this is blended pastes with both w/c and heterogeneous nucleation
due to the fact that the physical contribution of FA to the cement effects.
hydration is underestimated. In other words, not only the higher
w/c ratio can promote the cement hydration, but also the cement
5. Conclusions
hydration can be accelerated by the heterogeneous nucleation ef-
fect provided by FA particles. Fig. 11b presents the ap values de-
1, This article intends to quantify the hydration extents for
rived from Eq. (18b) in terms of measured values. Again, only the
cement grains and FA particles in blended cement pastes.
local w/c effect is considered in the prediction. Contrary to the
To aid the interpretation of experimental data, a synthetic
Wn-based estimation in Fig. 11(a), this CH-based estimation
model is firstly established for cement hydration extent with
underestimates the FA reaction extent significantly. The same rea-
the presence of FA. This model incorporates the pure dilu-
son for overestimation from Wn-based estimation is responsible
tion effect, the local w/c augmentation effect and the heter-
for the underestimation by CH-based approach: the underestima-
ogeneous nucleation effect of FA on cement hydration
tion of the cement hydration degree in blended pastes induces also
extent. This model is validated by literature data and used
the underestimation of CH formation.
for cement hydration extent in blended pastes.

Fig. 11. Comparison between the measured FA reaction extents and the predicted Fig. 12. Comparison between the measured FA reaction extents and the predicted
values from Wn-based estimation (a) and CH-based estimation (b) with only local values from Wn-based estimation (a) and CH-based estimation (b) with only
w/c effect. heterogeneous nucleation effect.
568 Q. Zeng et al. / Construction and Building Materials 27 (2012) 560–569

Acknowledgement

The research is supported by China National major fundamental

research Grant (973 Program, No. 2009CB623106).

Appendix A. Stoichiometry of cement and FA hydration

reactions

The hydration process of Portland cement can be summarized

as follows [3]:

C2 S þ 2H ¼ 0:5C3 S2 H3 þ 0:5CH ðA1:aÞ

C3 S þ 3H ¼ 0:5C3 S2 H3 þ 1:5CH ðA1:bÞ
C3 A þ CSH2 þ 10H ¼ C4 ASH12 ðA1:cÞ
C4 AF þ 2CH þ 10H ¼ C6 AFH12 ðA1:dÞ

As presented in Table 3, the gypsum content is very low fS ¼ 0:75,

and fCSH2 =f C3 A ¼ 0:115 < 0:637 [3], one more chemical reaction
can happen as gypsum is consumed completely:
C3 A þ CH þ 12H ¼ C4 AH13 ðA:2Þ
Employing the mass content of minerals listed in Table 3, one
obtains,

W cn ð1Þ ¼ 0:209f C2 S þ 0:237f C3 S þ 0:444f C4 AF þ 0:8f C3 A ¼ 0:272

ðA:3aÞ

CHc ð1Þ ¼ 0:215f C2 S þ 0:487f C3 S þ 0:430f CSH2  0:274f C3 A

 0:306f C4 AF
¼ 0:292 ðA:3bÞ

However, this CHc(1) value from stoichiometry analysis is signifi-

cantly higher than what has been measured for a fully hydrated
Portland cement paste, i.e. 15–25% [13]. This is due to the possible
carbonation of CH during TGA test. From TGA profiles this carbon-
Fig. 13. Comparison between the measured FA reaction extents and the predicted ated quantity is estimated, for decomposition temperature range
values from Wn-based estimation (a) and CH-based estimation (b) with both w/c 550–700 °C, to around 30% of total CH. Thus, the adopted value
and HN effects.
for CHc(1), for experimental evaluation purpose, is retained as
0.204 (0.292  0.70).
The pozzolanic reactions of FA can be expressed as,
2, The TGA method was used to measure the Wn content and
CH content during the hardening process of blended cement 1:5CH þ S ¼ C1:5 SH1:5 ðA:4aÞ
pastes. The selective dissolution method was employed to A þ 4CH þ 9H ¼ C4 AH13 ðA:4bÞ
quantify the pozzolanic reaction extent of FA. The Wn con- A þ CSH2 þ 3CH þ 7H ¼ C4 ASH12 ðA:4cÞ
tent and CH content were measured for the samples (PI,
PII) of two w/b ratios and four FA content levels. With help By the same stoichiometry analysis one can obtain the and W pn ð1Þ
of the established synthetic model, the respective contribu- CHp(1), representing Wn and CH contents as FA hydrates
tions of pure dilution effect, w/c effect, heterogeneous effect completely:
and FA hydration are quantified for the Wn and CH measure-
ment of blended pastes. From the results, it is observed that W pn ð1Þ ¼ 1:588f A cA  1:35f S ¼ 0:168 ðA:5aÞ
during the hardening process the above factors have compa-
rable impact on Wn content while FA hydration dominates CHp ð1Þ ¼ 1:85f S cS  2:907f A cA þ 0:925f S ¼ 0:853 ðA:5bÞ
over other factors for CH content.
where cj=S,A represents the weight fraction of the reactant oxide j in
3, The cement hydration extent is determined through experi-
the pozzolana and it is assumed cS = cA = 0.5 in this study. The
mental approach as well as model-based approach and the
above value of W pn ð1Þ by stoichiometric analysis can be verified
ac values from two approaches are linearly correlated. The by some published results [34]: W pn ¼ 0:167 with 30% FA blended
pozzolanic reaction extent of FA hydration is also determined
with Portland cement at curing age of 360d.
through experimental approach, by selective dissolution
method, and model-based approach. The FA hydration extent
References
is found to be much influenced by curing age, i.e. ap  0.1 for
7d and ap = 0.2  0.4 for 90d. Moreover, the comparison [1] Chinese civil engineering society guide to durability design and construction of
between experimental and model-based results confirms concrete structures (CCES01-2004). Beijing, China: Chinese Building Industrial
that both w/c effect and heterogeneous effect of FA on Publishing; 2005.
[2] Dunstan MRH, Hicks CH. Adaptive construction methodology for the Ghatgher
cement hydration should be taken into account to achieve a Saddle dam India’s first RCC dam. In: Berga et al., editors. Roller compacted
reliable prediction of FA hydration extent in blended pastes. concrete dam. Lisse: Swets & Zeitlinger; 2003. p. 239–47.
 Q. Zeng et al. / Construction and Building Materials 27 (2012) 560–569 569

[3] Papadakis VG. Effect of fly ash on Portland cement systems Part I: low-calcium [19] Powers TC. Structure and physical properties of hardened Portland cement
fly ash. Cem Concr Res 1999;29:1727–36. paste. J Am Ceram Soc 1958;41:1–6.
[4] Hanehara S, Tomosawa F, Kobayakawaa M, Hwang K. Effects of water/powder [20] Li X, Yan PY, A RH. Assessment method of hydration degree of cement in
ratio, mixing ratio of fly ash, and curing temperature on pozzolanic reaction of complex binder based on the calcium hydroxide content (in Chinese). J Chin
fly ash in cement paste. Cem Concr Res 2001;31:31–9. Ceram Soc 2009;37. 1597–61.
[5] Xu A, Sakar SL, Nilsson LO. Effect of fly ash on the micro-structure of cement [21] Pang GS, Chae ST, Chang SP. Predicting model for pore structure of concrete
mortar. Mater Struct 1993;26:414–24. including interface transition zone between aggregate and cement paste. Int J
[6] Berrey EE, Hemmings RT, Cornelius BJ. Mechanisms of hydration reaction in Concr Struct Mater 2009;3:81–90.
high volume fly ash cements and mortars. Cem Concr Compos [22] Tang CW. Hydration properties of cement pastes containing high-volume
1990;12:253–61. mineral mixtures. Comput Concr 2010;7:17–38.
[7] Cyr M, Lawrence P, Ringot E. Efficiency of mineral admixtures in mortars: [23] Atahan HN, Oktar ON, Tasdemir MA. Effects of water–cement ratio and curing
quantification of the physical and chemical effects of fine admixtures in time on the critical pore width of hardened cement paste. Constr Build Mater
relation with compressive strength. Cem Concr Res 2006;36:264–77. 2009;23:1196–200.
[8] Lawrence P, Cyr M, Ringot E. Mineral admixtures in mortars: effect of inert [24] Termkhajornkit P, Nawa T, Kurumisawa K. Effect of water curing conditions on
materials on short-term hydration. Cem Concr Res 2003;33:1939–47. the hydration degree and compressive strengths of fly ash-cement paste. Cem
[9] Cyr M, Lawrence P, Ringot E. Mineral admixtures in mortars: quantification of Concr Compos 2006;28:781–9.
the physical effects of inert materials on short-term hydration. Cem Concr Res [25] Bonavetti V, Donza H, Menendez G, Cabrera O, Irassar EF. Limestone filler
2005;35:719–30. cement in low w/c concrete: a rational use of energy. Cem Concr Res
[10] Li C, Sun H, Li L. A review: the comparison between alkali-activated slag 2003;33:865–71.
(Si + Ca) and metakaolin (Si + Al) cements. Cem Concr Res 2010;40:1341–9. [26] GB 175-2007/XG1-2009 Common Portland cement.
[11] Constantinides G, Ulm F-J. The nanogranular nature of C–S–H. J Mech Phys [27] European Standard EN 197-1. Cement-Part 1: composition, specification and
Solids 2007;55:64–90. conformity criteria for common cements.
[12] Scrivener KL, Fullmann T, Gallucci E, Walenta G, Bermejo E. Quantitative study [28] ASTM C150-00. Standard specification for Portland Cement.
of Portland cement hydration by X-ray diffraction/Rietveld analysis and [29] Bogue RH. Calculation of the compounds in Portland cement. Ind Eng Chem
independent methods. Cem Concr Res 2004;34:1541–7. 1929;1:192–7.
[13] Lam L, Wong YL, Poon CS. Degree of hydration and gel/space ratio of high- [30] Aligizaki KK. Pore structure of cement-based materials: testing, interpretation
volume fly ash/cement systems. Cem Concr Res 2000;30:747–56. and requirement. New York: Taylor & Francis; 2006.
[14] Hoshino S, Yamada K, Hirao H. XRD/Rietveld analysis of the hydration and [31] Marsh BK, Day RL. Pozzolanic and cementitious reactions of fly ash blended
strength development of slag and limestone blended Cement. J Adv Concr cement pastes. Cem Concr Res 1998;18:301–10.
Technol 2006;4:357–67. [32] ACI Committee 211. Guide for selecting proportions for high-strength concrete
[15] Skibsted J, Hall C. Characterization of cement minerals, cements and their with Portland cement and fly ash. ACI 226.4R. ACI Mater J 1993;90:272–83.
reaction products at the atomic and nano-scale. Cem Concr Res [33] Schwarz N, Neithalath N. Influence of a fine glass powder on cement
2008;38:205–25. hydration: comparison to fly ash and modeling the degree of hydration. Cem
[16] Suprenant BA, Papadopoulos G. Selective dissolution of Portland-fly-ash Concr Res 2008;38:429–36.
cements. J Mater Civil Eng 1991;3:48–59. [34] Escalante-Garcia JI. Nonevaporable water from neat OPC and replacement
[17] Li S, Roy DM, Kumer A. Quantitative determination of pozzolanas in hydrated materials in composite cements hydrated at different temperatures. Cem
system of cement or CaðOHÞ2 with fly ash or silic fume. Cem Concr Res Concr Res 2003;33:1883–8.
1985;15:1079–86.
[18] Scrivener KL. Backscattered electron imaging of cementitious microstructures:
understanding and quantification. Cem Concr Comp 2004;26:934–45.

View publication stats