Some microstructure properties at early age of ettringite binder based on rich C12A7 calcium aluminate cement

Journal of Science and Technology in Civil Engineering NUCE 2018. 12 (3): 44–50

SOME MICROSTRUCTURE PROPERTIES
AT EARLY AGE OF ETTRINGITE BINDER BASED ON
RICH C12 A7 CALCIUM ALUMINATE CEMENT
Nguyen Ngoc Lama,∗
a

Faculty of Building Materials, National University of Civil Engineering,
55 Giai Phong road, Hai Ba Trung district, Hanoi, Vietnam

Article history:
Received 24 January 2018, Revised 04 April 2018, Accepted 27 April 2018

Abstract
The mineral composition of calcium aluminate cements is traditionally based on CA (monocalcium aluminate
- CaO · Al2 O3 ). Recently, a new cement with the main compound of C12 A7 (Mayenite) has been developed for
rapid hardening binder. This cement is used in conjunction with a sulfate binder to form a new type binder called
ettringite binder due to the high quantity of ettringite in the hydration product, opened new possibilities for
mortar and concrete formulations. This paper focuses on some microstructure characteristics of the ettringite
binder based on a C12 A7 rich cement and a hemihydrate at early age. Some important characteristics of this
binder were found, such as: short setting time (about 40–50 minutes), rapid expansion just after initial setting
time, rapid evolution of porosity and bound water during the first 5 hours of hydration. The correlation between
bound water and porosity of hardened binders was also found in this paper.
Keywords: ettringite binder; early-age; setting; C12 A7 ; hemihydrate.
c 2018 National University of Civil Engineering

1. Introduction
Within the last few decades, the number of bridges, roads, houses damaged or degraded has
been being increased [1, 2]. Therefore, rapid hardening repair materials are in high demand for
these applications as they allow for minimizing traffic delays, road closures and timesaving, etc. . .

[2, 3]. Calcium aluminate cement (CAC), whose mineral composition is traditionally designed around
monocalcium aluminate - CaO · Al2 O3 (CA), is a rapid hardening binder and usually used for special
applications where high early strength and increased durability are desired. Their setting time is
close to that of OPC, typically around 3 hours, but their hardening rate is in the range of 10 MPa to
20 MPa (compression) per hour from setting. This rapidity is compatible with applications that require
compressive strength from 10 MPa to 30 MPa after 4h to 6h, such as industrial floor repairs [4–6]. For
applications requiring higher rapidity, the hydration has to be accelerated. The most common way to
accelerate hydration of CAC is adding lithium salts (Li2 CO3 or Li2 SO4 for example) [7–9]. Recently,
∗

Corresponding author. E-mail address: (Lam, N. N.)

44


Lam, N. N. / Journal of Science and Technology in Civil Engineering

a new cement with C12 A7 (12 CaO · 7 Al2 O3 ) as a major mineral compound, has been developed.
Increasing the amount of the C12 A7 phase of the CAC is another way to accelerate the hydration rate
since this phase is more active than CA and contributes greatly to the setting time of the cement.
Too much of this phase can cause flash setting in the mortar and concrete, however, its percentage is
typically regulated in the manufacturing process. A new blended cement systems incorporating both
CAC and calcium sulfate (C$H x ) have been developed to utilize the rapid hardening characteristics of
rich C12 A7 CAC but at a reduced cost because of the ratio of CaO/Al2 O3 in C12 A7 increased closer to
the one of ettringite [10–13], which leads to the quick formation of ettringite with a higher content.
The hydration of an ettringite binder containing calcium aluminate cement (CAC) and calcium
sulfate (C$H x ) induces ettringite (C6 A$3 H32 ) and aluminum hydroxide (AH3 ) as follows [14–16]:
C12 A7 + 12 C$H x + (137−12x) H −−−→ 4 C6 A$3 H32 + 3 AH3

(1)

In order to have more understandings about the hydration of ettringite binder consisting rich
C12 A7 cement, some important characteristics were investigated such as setting time, chemical shrinkage, endogenous shrinkage, porosity and pore distribution by Mercury intrusion porosimetry (MIP)
and bound water by DTA-TG analysis. In addition, the relationship of these properties is also discussed.
2. Materials and test methods
2.1. Materials
The binder used in this research consists of 75% by weight of the rich C12 A7 calcium aluminate
cement and 25% by weight of hemihydrate C$H0.5 . The amount of C12 A7 , CA, C3 A, C4 A3 $, C2 S
and Ferrite in CAC was 57.2%, 2.1%, 0.4%, 0.3% and 18.1% by weight, respectively, which were
determined by the Rietveld quantitative phase analysis. This binder has a water/binder ratio of 0.3.
2.2. Test methods
After mixing, the Vicat penetration according to EN 196-3 and chemical shrinkage as per ASTM
C1608 of the pastes were determined.
The endogenous shrinkage of the paste was measured as described in detail in [17]. A rubber
membrane containing binder paste was submerged in water. The change of volume of the cement
paste was measured by the amount of liquid displaced by the immersed sample, typically by measuring its weight change. This method is also referred as the buoyancy method.
The binder paste was cast in small closed plastic bottles. After being cured for 2h, 3h, 5h, 10h,
24h, the solid binder then was crushed and immediately immersed in the acetone solution in further
two days to stop hydration. After that, the pieces of samples with a size of about 1 cm3 were placed
in a desiccator to remove the acetone and ensure that no further hydration could be taken place. The
specimens then were used for pore structure analysis or bound water analysis.
The pore structure of specimens was determined by mercury intrusion porosimetry. This measurement was performed with the Micromeritics Auto Pore IV. The specimen was placed in a glass
tube and filled with a non-wetting liquid (mercury) under vacuum conditions with a pressure of less
than 50 µm/Hg. The glass tube with the specimen and mercury was subsequently placed in a highpressure analysis port. The high-pressure analysis port utilized oil to continue pressing mercury into
the specimen, with a pressure ranging from 14.7 psi to 60,000 psi, and the intrusion mercury volume
was recorded at each pressure point.
45


The binder paste was cast in small closed plastic bottles. After being cured for 2h, 3h, 5h, 10h, 24h, the solid binder then was

crushed and immediately immersed in the acetone solution in further two days to stop hydration. After that, the pieces of samples with a
size of about 1cm3 were placed in a desiccator to remove the acetone and ensure that no further hydration could be taken place. The
specimens then were used for pore structure analysis or bound water analysis.
The pore structure of specimens was determined by mercury intrusion porosimetry. This measurement was performed with the
Micromeritics Auto Pore IV.Lam,
The specimen
was placed
in a glassand
tube
and filled with
a non-wetting
liquid (mercury) under vacuum
N. N. / Journal
of Science
Technology
in Civil
Engineering
conditions with a pressure of less than 50 µm/Hg. The glass tube with the specimen and mercury was subsequently placed in a highpressure
port. The
high-pressure
analysis
utilized into
oil to continue
mercury
into µm)
the specimen,
with a pressure
Theanalysis
hardened
binders

were
alsoport
ground
small pressing
powder
(< 50
to determine
theranging
bound
from 14.7 psi to 60,000 psi, and the intrusion mercury volume was recorded at each pressure point.

water in binder by heating samples to 1000˚C at a heating rate of 10˚C/min. The content of bound
The hardened binders were also ground into small powder (<50µm) to determine the bound water in binder by heating samples to
water
in
binder was calculated based on the weight of samples at 30˚C and 300˚C as presented in the
1000°C at a heating rate of 10°C/min. The content of bound water in binder was calculated based on the weight of samples at 30°C and
Fig.
1:as presented in the Fig. 1:
300°C
Temperature,°C
0

100

200

300

400

500

600

700

800

900 1000

20
Bound water

10
0

DTA(µV)

-10
-20
-30
-40
-50

-60
-70
-80
2h

3h

5h

7h

24h

Figure 1. Calculation scheme of bound water in binder
Figure 1. Calculation scheme of bound water in binder

3. Results and discussion
Vicat penetration
of paste with time
3.3.1.Results
and discussion
The setting time presented by the Vicat penetration is an indicator presenting the liquid - solid transition, which is very
important
in assessing
practicalofconstruction
operations,
3.1.
Vicat
penetration
paste with
time such as finishing, sawcutting and curing, etc.... The results of Vicat penetration
with time of the binder consisting of 75% rich C12A7 cement and 25% hemihydrate are shown the Table 1:

The setting time presented byTable
the1.Vicat

an paste
indicator presenting the liquid - solid
Setting penetration
time of ettringiteis
binder
transition, which
is very important
in assessing
construction
operations,
such(mm)
as finishing,
Time (minutes)
Vicat penetration
(mm) practical
Time
(minutes)
Vicat penetration
sawcutting and curing,
etc. . . The results
of Vicat penetration
consisting of
0
40
42 with time of the binder
25
40
22
75% rich C12 A7 5cement and 25% hemihydrate
are shown the43Table 1.

10
15
20

Time (minutes)
25



0
5
10
15
20
25
30
35
38
39
40
41

30
35

40

44

Table 1. 40

Setting time of ettringite binder
paste
45
40

46

Vicat penetration
(mm)
40
40
40
40
40
40
40
40
40
39
36
33
29

Time (minutes)
47

40

48
42

49
43
44
45
46
47
48
49
50
51
52
53

40

20
19
17

Vicat penetration
(mm)
13
9
6

25
22
20
19
17

13
9
6
3
1
0.5
0.5

2

The research results showed that after 38 minutes of hydration, the setting process of paste was
started. At this moment, the hydration products, especially ettringite and AH3 , were generated more
46


The research results showed that after 38 minutes of hydration, the setting process of paste was started. At this mome
hydration products, especially ettringite and AH3, were generated more and more and begin to create a skeleton structure in the
With progress of hydration process, hydration products were crystallized and enlarged, and this leads to the liquid-solid transition pr
3.2. Chemical shrinkage, autogenous shrinkage

The correlation between autogenous shrinkage and chemical shrinkage during the early ages is depicted in Fig. 2. Overa
observed that this relationship is not linear to chemical shrinkage, and that no simple relationship exists between them. It is caused
change of the macroscopic volume that occurs concurrently with chemical shrinkage, can be observed either expansively or contrac
Lam, N. N.The
/ Journal
Civil
Engineering
expansionofofScience
ettringite and
binderTechnology

paste is due tointhe
formation
of ettringite but this shrinkage reaction is accompanied by ch
Lam, N. N./ Journal of Science and Technology in
Civil Engineering
shrinkage.

39

and more and
begin to create a50skeleton structure3 in the paste. With progress
of hydration process,
Time,minutes
39
(a)
200to the
400 liquid-solid
600
800
1000
1200
1400
hydration products
were crystallized
and enlarged,1 and this0leads
transition
process.
36
51

40
41

33

52

0.5

29
53
3.2. Chemical
shrinkage, autogenous
shrinkage 0.5

100

Shrinkage(mm3/gbinder)

38

Settingperiod

80

search results showed that after 38 minutes of hydration, the setting process of paste was started. At this 60
moment, the
ducts, especially ettringite and AH3, were generated more and more and begin to create a skeleton structure in the paste.
40 process.
of hydration process, hydration products were crystallized and enlarged, and this leads to the liquid-solid transition

The correlation between autogenous shrinkage and chemical shrinkage during the early ages is
depicted in Fig. 2. Overall, it is observed that this relationship is not linear to chemical shrinkage, and
Shrinkage
shrinkage, autogenous shrinkage
20
that no simple relationship exists between them. It is caused
by the change of the macroscopic volume
orrelation between autogenous shrinkage and chemical shrinkage during the early ages is depicted in Fig. 2. Overall, it is
that occurs concurrently with chemical shrinkage, can be 0observed either expansively or contractively.
this relationship is not linear to chemical shrinkage, and that no simple relationship exists between them. It is caused by the
macroscopic volume
that occurs
concurrently
chemical shrinkage,
canpaste
be observed
either to
expansively
or -20
contractively.
The
expansion
ofwith
ettringite
binder
is due
the formation
of ettringite but this shrinkage reaction
Expansion

n of ettringite binder paste is due to the formation of ettringite but this shrinkage reaction is accompanied by chemical
EndogenousShrinkage
ChemicalShrinkage
is accompanied by chemical shrinkage.
Time,minutes

(a)
0

200

400

600

800

(b)
1000

1200

0

Settingperiod

80
60
40

Shrinkage

20
0

-20

40

60

(b)
0

20

40

80

80

100 120 140 160 180 200

Settingperiod

60
40

Shrinkage

20
0
-20

Expansion

EndogenousShrinkage

Shrinkage(mm3/gbinder)

20

100

Shrinkage(mm3/gbinder)

Shrinkage(mm3/gbinder)

100

100

Time,minutes

1400

Lam, N. N./ Journal of Science and Technology in Civil Engineering
Expansion

ChemicalShrinkage

EndogenousShrinkage
ChemicalShrinkage
skeleton, while the continuing chemical
shrinkage stresses induce
a strain in the binder as autogenous shrinkage. The
expand after the initial setting time, but expansion rate will not increase any more after 300 minutes (5 hours) of hydra
Figure
2. Chemical
shrinkage
of binder
that theand
binderendogenous
after
5h has
sufficientlyshrinkage
strength toand
resist
the expansion
stress caused
by ettringite formation.
Figure
2. Chemical shrinkage
ofendogenous
binder
Time,minutes
a) During the firstshrinkage
24h (1400 minutes)
b) During the first 200 minutes

3.3.
Pore
distribution
60 a)
80During
100 120 the
140 first
160 24h
180
(1400
b) During
the40first
200
minutesthe binder is very fluid and all volume change of
In the200
first stageminutes);
before initial setting
time, about
minutes
of hydration,
Theshrinkage.
experimental
results of
specimens
determined
MIP methods
and high
were plasticity
shown in the
Figs.

paste is attributed to chemical
It should
bethe
noted
that thewere
liquid
stage thebybinder
paste with
does
not 3-5.
cau
porosity
Figs. 4-5 present
the relationship
between mercury
intrusions
andshrinkage
pore sizes.as long as there is insuf
stress in the paste. Intotal
Fig.
2, the and
autogenous
shrinkage
curve is coinciding
with the
chemical
binder stiffness to resist the forces. This means that autogenous shrinkage is nearly equal to chemical shrinkage for a short period
Settingperiod
the initial setting time. In the next stage, at about 53 minutes after hydration, the binder paste begins to stiffen and forms an
50

7h

25

dV/dlog(r)

Accumulated pore volume [%]

3.3. Pore distribution

Totalporosity,%

80 the first stage before initial setting time,
In
about
60 40 minutes of hydration, the binder is very
fluid40and all volume change of binder paste is at40
Shrinkage
tributed to chemical shrinkage. It should be noted
20
that the liquid stage the binder paste with high
30
0
plasticity
does not cause any stress in the paste. In
20
Fig.-202, the autogenous shrinkage curve is coincidExpansion
EndogenousShrinkage
ChemicalShrinkage

ing with the
chemical shrinkage
as long as there
10
isFigure
insufficient
binder
stiffness
to
resist
the forces.
2. Chemical shrinkage and endogenous shrinkage of binder
a) During
the first 24h
(1400autogenous
minutes)
b)
During the firstis
200nearly
minutes
This
means
that
shrinkage
0
0
5
10
15
20

25
first stage before initial setting time, about 40 minutes of hydration, the binder is very fluid and all volume change of binder
equal
to
chemical
shrinkage
for
a
short
period
beuted to chemical shrinkage. It should be noted that the liquid stage the binder paste with high plasticity does not cause any
Time,hours
aste. In Fig. 2, thefore
autogenous
is coinciding
with theIn
chemical
as long as there is insufficient
theshrinkage
initialcurve
setting
time.
the shrinkage
next stage,
s to resist the forces. This means that autogenous shrinkage is nearly equal to chemical shrinkage for a short period before
at stage,
about
5353minutes
hydration,
binder

ing time. In the next
at about
minutes afterafter
hydration,
the binder pastethe
begins
to stiffen and forms Figure
an initial 3. Total porosity of binder with time
Figure 3. Total porosity of binder with time
paste begins to stiffen and forms an initial skele3
ton, while the continuing chemical shrinkage stresses
induce a strain in the binder as1.2 autogenous
50
2h
2h
shrinkage. The binder begins to expand after the45 initial setting time, but expansion
rate will not
in1
3h
3h
40
crease any more after 300 minutes (5 hours) of hydration. This proves that the binder after 5h has
35
5h
5h
0.8
sufficiently strength to resist the expansion stress30caused by ettringite formation.

24h

20
15

0.6

7h

24h
0.4

The experimental results of the specimens were
determined by MIP methods and were shown in
10
0.2
5
the Figs. 3–5. Fig. 3 shows the total porosity and Figs. 4–5 present the relationship between mercury
0
0.001

47

0.01

0.1

1

10

100

Pore diameter [µm]
Figure 4. Accumulated pore volume of binder for various
curing
time of hardened binder for
Figure 4. Accumulated pore
volume
different curing time


0
0.001

0.01

0.1

1

Pore diameter [µ
Figure 5. Pore size distribution of binder w
Figure 5. Pore size distribution of hardene
different curing time

The results in the Fig. 5 show that total pore volume of samples decreased at different curing time, from 44.34%
at 5h, but with a slight decrease from 5h to 24h due to the rapid hydration rate of CAC and hemihydrate. In figure 4


00

00

55

10
10

1515

2020

2525

Time,hours
Time,hours

Lam, N. N. / Journal of Science and Technology in Civil Engineering
Figure
Figure3.3.Total
Totalporosity
porosityofofbinder
binderwith
withtime
time

intrusions and pore sizes.
1.2
1.2

2h2h

45
45
40
40

3h3h

11

35
35

5h5h

0.8
0.8

30
30

7h7h

25
25

dV/dlog(r)
dV/dlog(r)

Accumulated pore

pore volume
volume [%]
[%]
Accumulated

50
50

24h
24h

20
20
15
15

10
10

3h3h
5h5h
7h7h
24h
24h

0.4
0.4
0.2
0.2

55
00
0.001
0.001

0.6
0.6

2h2h

0.01
0.01

0.1
0.1

11

1010

100
100

Pore
Porediameter
diameter[µm]
[µm]
Figure4.4.Accumulated
Accumulatedpore
porevolume

volumeofofbinder
binderfor
forvarious
various
Figure
Figure
4. Accumulated
pore
of hardened
curing
timevolume
curing
time
Figure
Accumulated
pore
volume
hardened
binderfor
for
Figure
4.4.Accumulated
pore
volume
ofofhardened
binder
binder
for different
curing
different

curingtime
time time
different
curing


00
0.001
0.001

0.01
0.01

0.1
0.1

11

1010

100
100

Pore
Porediameter
diameter[µm]
[µm]
Figure5.5.Pore
Poresize
sizedistribution

distributionofofbinder
binderwith
withtime
time
Figure
Figure
5.5.Pore
Pore
sizedistribution
distribution
of
hardened
binder
Figure5.
Poresize
size
distribution
hardened
binder
for
Figure
ofofhardened
binder
for
fordifferent
different
curing
time
differentcuring
curingtime

time

Theresults
resultsininthe
theFig.
Fig.5 5show
showthat
thattotal
totalpore
porevolume
volumeofofsamples
samplesdecreased
decreasedatatdifferent
differentcuring
curingtime,
time,from
from44.34%
44.34%atat2h2htoto32.32%
32.32%
The
The
results
indecrease
the from
Fig.
55htoshow
that
total
pore
volume

of
samples
decreased
at different
curing
5h,but
butwith
with
slightdecrease
from5h
to24h
24hdue
due
therapid
rapid
hydration
rateofof
CAC
andhemihydrate.
hemihydrate.
figure
andfigure
figure
atat5h,
a aslight
totothe
hydration
rate
CAC
and

InInfigure
4 4and
5,5,it itis is
easily
observed
that
bothcapillary
capillary
pore32.32%
(>0.01µm)at
and
gelbut
porewith
(<0.01µm)
increased
significantly
during
the
first5due
5hours
hours
hydration
easily
observed
both
(>0.01µm)
and
gel
pore
(<0.01µm)

increased
significantly
during
the
first
hydration
time,
fromthat
44.34%
at 2hpore
to
5h,
a slight
decrease
from
5h to
24h
toofof
the
rapid
and
thatmost
mostpore
pore
diameters
thespecimens
specimens
aredistributed
distributed
between

0.01µm
µmtoit
to2is
2µm.
µm.
and
that
diameters
ofofthe
are
between
hydration
rate
of CAC
and
hemihydrate.
In Figs.
40.01
and
5,
easily observed that both capillary pore
Thepore
poresize
size
distribution
differential
curve
obtainedbybytaking
takingthe
theslope

slopeofof
thepore
poresize
sizedistribution
distribution
curve
the
LogDifferential
Differential
The
distribution
differential
the
curve
the
Log
(> 0.01
µm)
and
gel pore
(< 0.01curve
µm)isisobtained
increased
significantly
during
the
first
5 hours
of
hydration

and
Intrusionagainst
againstpore
poresizes
sizesininFig.
Fig.5.5.The
Thepeaks
peaksininFig.
Fig.5 5represent
representthe
thepore
porediameters
diameterscorresponding
correspondingtotothe
thehigher
higherrate
rateofofmercury
mercury
Intrusion
that most pore diameters of the specimens are distributed between 0.01 µm to 2 µm.
intrusionper
perchange
changeininpressure.
pressure.These
Thesepeaks
peaksare
arecalled
called“threshold”
“threshold”pore
porediameters.

diameters.The
Thepeak
peakatataround
around1 1µm
µmcould
couldbebefound
foundononthethecurve
curve
intrusion

The pore size distribution differential curve is obtained by taking the slope of the pore size distribution curve the Log Differential Intrusion against pore sizes in Fig. 5. The peaks in Fig. 5 represent
44

the pore diameters corresponding to the higher rate of mercury intrusion per change in pressure. These
peaks are called “threshold” pore diameters. The peak at around 1 µm could be found on the curve
of sample at 2h and 3h and more finer peak (around 0.2 µm) at later age. This may be caused by the
enlargement of gel pore due to the hydration evolution.
3.4. Bound water in hardened binder
The evolution of bound water in hardened binder determined by DTA-TG method is shown in
Fig. 6 and the total porosity as a function of bound water is shown in Fig. 7.
It can be observed that the amount of bound water increases rapidly from 2h to 5h due to the rapid
hydration of the binder and then decreased. The binder also exhibits the same behavior as porosity
characteristic. This is not surprising since the binders are dominated by rapid formation of ettringite.
This further implies that bound water can represent the hydration of ettringite binder at early age. It
is confirmed by comparing bound water with porosity of hardened binder paste in the Fig. 7.
The results of the bound water versus its porosity in Fig. 7 show a linear relation. Any interpolation point at porosity can find the value of bound water. This comes with no surprise since the
hydration products, which are proportional to bound water, form the more and more and fill into the
pores and decreases the total porosity of system.
4. Conclusions
From the tested results of the microstructure properties of quick hardening binder based on rich

C12 A7 calcium aluminate cement, some conclusions can be withdrawn as below:
48


sampleatat2h
2hand
and3h
3hand
and more
more finer
finer peak
age.
This
may
be be
caused
by the
enlargement
of gelofpore
due todue
theto th
ofof
sample
peak (around
(around0.2
0.2µm)
µm)atatlater
later
age.
This

may
caused
by the
enlargement
gel pore
hydrationevolution.
evolution.
hydration
3.4.Bound
Boundwater
waterininhardened
hardened binder
binder
3.4.
The
evolution
of
bound
water in
byby
DTA-TG
method
is shown
in Fig.
6 and
porosity
as a as
The evolution of bound water
in hardened
hardenedbinder

binderdetermined
determined
DTA-TG
method
is shown
in Fig.
6 the
andtotal
the total
porosity
function
of
bound
water
is
shown
in
Fig.
7.
function of bound water is shown in Fig. 7.

Lam, N. N. / Journal of Science and Technology in Civil Engineering
5050

20
20

4040

Porosity,%

Porosity,%

Boundwater,%

Boundwater,%

25
25

15
15
10
10
55

00

R²=0.9963
R²=0.9963

2020
1010

00

10
10

20

20

00
0 0

30
30

Time,hours
Time,hours



3030

10 10

20 20

30 30

Boundwater,%
Boundwater,%

Figure6.6.Evolution
Evolution of
of bound
bound water
in
with

7. Relation between
total porosity
and bound
water
Figure
water
inbinder
binder
withtime
time Figure
Figure
Figure 6. Evolution
of bound
water
in binder
Figure7.7.Relation
Relationbetween
betweentotal
totalporosity
porosityand
andbound water

with time

bound water

canbe
beobserved
observed that
that the

the amount
rapidly
from
2h 2h
to 5h
to the
hydration
of theofbinder
and an
ItItcan
amount of
of bound
boundwater
waterincreases
increases
rapidly
from
to due
5h due
to rapid
the rapid
hydration
the binder
thendecreased.
decreased. The
The binder
binder also
also exhibits
exhibits the
characteristic.

ThisThis
is not
surprising
sincesince
the binders
are ar
then
the same
samebehavior
behaviorasasporosity
porosity
characteristic.
is not
surprising
the binders
dominated
byrapid
rapid
formation
of ettringite.
ettringite.containing
This
implies
bound
water
canaluminate
represent
the
hydration
ettringite

binderbinder
at early
dominated
formation
Thisfurther
further
implies
bound
water
can
represent
the
hydration
of ettringite
at earl
- by
The
setting
time of
binder
the
richthat
Cthat
cement
andofhemihydrate
12 A
7 calcium
age.It Itisisconfirmed
confirmedby
bycomparing

comparing bound
bound water
binder
paste
in the
Fig.Fig.
7. 7.
age.
waterwith
withporosity
porosityofofhardened
hardened
binder
paste
in the

takes place earlier when compared to that of OPC or standard CAC, only 40–50 minutes after mixing
Theresults
resultsof
ofthe
the bound
bound water
water versus
a linear
relation.
AnyAny
interpolation
pointpoint
at porosity
can find

The
versusits
itsporosity
porosityininFig.
Fig.7 7show
show
a linear
relation.
interpolation
at porosity
canthe
find th
with
water.
valueofofbound
boundwater.
water. This
This comes
comes with
products,
which
are are
proportional
to bound
water,
form form
the more
value
with no
no surprise

surprisesince
sincethe
thehydration
hydration
products,
which
proportional
to bound
water,
the mor
-and
The
paste
begins
to expand
rapidly
just
after the initial setting time. This expansion
and
more
and
fillbinder
intothe
thepores
pores and
and
decreases
system.
and
more

fill
into
decreasesthe
thetotal
totalporosity
porosityofof
system.
period prolongs until 5h of hydration.
Conclusions
4. 4.
Conclusions
- The porosity and bound water varied dramatically during the first 5 hours of hydration. There is
Fromthe
thetested
testedresults
resultsof
of the
the microstructure
microstructure properties
binder
based
on rich
C12A7
calcium
aluminate
cement,
some som
From
propertiesofofquick
quickhardening

hardening
binder
based
on rich
C12A7
calcium
aluminate
cement,
also
a good
conclusions
canbe
becorrelation
withdrawn as
asbetween
below: bound water and porosity and these two parameters showed the same
conclusions
can
withdrawn
below:
development at early-age.

Thesetting
setting time
time of
of binder
binder containing
aluminate
cement
andand

hemihydrate
takestakes
placeplace
earlierearlier
when whe
- -The
containing the
therich
richC12A7
C12A7calcium
calcium
aluminate
cement
hemihydrate
comparedtotothat
thatofofOPC
OPCor
or standard
standard CAC,
mixing
with
water.
compared
CAC, only
only40-50
40-50minutes
minutesafter
after
mixing
with

water.

References
- Thebinder
binderpaste
paste begins
begins to expand
initial
setting
time.
This
expansion
period
prolongs
until until
5h of5h
hydration.
- The
expand rapidly
rapidlyjust
justafter
afterthe
the
initial
setting
time.
This
expansion
period
prolongs

of hydration

- Theporosity
porosityand
bound water
water varied
dramatically
during
the
first
5 hours
of hydration.
There
is also
a good
correlation
between
- The
bound
varied
dramatically
during
the
first
5needs
hours
of hydration.
There
is
also

a good
correlation
betwee
[1]
World
Bankand
(2016).
Vietnam
2016:
Rapid flood
damage
and
assessment.
World
Bank
report.
boundwater
waterand
andporosity
porosityand
and these
these two
development
at early-age.
bound
two parameters
parametersshowed
showedthe
thesame
same

development
at early-age.

[2] Texas Department of Transportation (2004). Standard specifications for construction and maintenance of
highways, streets, and bridges.
[3] Bizzozero, J. and Scrivener, K. L. (2014). Hydration and microstructure of rapid-strength binders based
References
References
on OPC accelerated by early ettringite formation. In Proceedings of the International Conference on
WorldBank
Bank(2016).
(2016).
Vietnam 2016:
rapid
flood
assessment.
World
Bank
report.
[1][1]
World
Vietnam
rapidFrance.
flooddamage
damageand
andneeds
needs
assessment.
World
Bank

report.
Calcium
Aluminates,
Avignon,
[4]
Fryda,
H.,
Berger,
S.,
Bordet,
F.,
Andreani,
P.
A.,
Martinet,
A.,
and
Brigandat,
P. (2014).
Ultra
fast
[2]
Texas
Department
of
Transportation
(2004).
Standard
specifications
for

construction
and
maintenance
of highways,
streets,
and an
[2] Texas Department of Transportation (2004). Standard specifications for construction and maintenance
of
highways,
streets,
bridges. hydration opening new application fields: a comparison of different calcium aluminate technologies. In
bridges.
Proceedings
of the
Conferencemicrostructure
on Calcium Aluminates,
Avignon,
France.
Bizzozero
Scrivener
K. International
of of
rapid-strength
binders
based
on OPC
accelerated
by early
[3][3]Bizzozero
J.,J.,Scrivener

K.
L. (2014).
(2014). Hydration
Hydrationand
and microstructure
rapid-strength
binders
based
on OPC
accelerated
by earl
[5]
Onishi,
K.
and
Bier,
T.
A.
(2010).
Investigation
into
relations
among
technological
properties,
hydration
ettringite
formation.
CALCIUM
ALUMINATES

Proceedings
of
the
International
Conference
2014,
Avignon,
France.
ettringite formation. CALCIUM ALUMINATES - Proceedings of the International Conference 2014,
Avignon,
France.
kinetics and early age hydration of self-leveling underlayments. Cement and Concrete Research, 40(7):
FrydaH.,
H.,Berger
BergerS.,
S., Bordet
Bordet F., Andreani
Brigandat
P. P.
(2014).
Ultra
fast fast
hydration
opening
new application
[4][4]Fryda
Andreani P.A.,
P.A.,Martinet
MartinetA.,
A.,and

and
Brigandat
(2014).
Ultra
hydration
opening
new applicatio
1034–1040.
fields:a acomparison
comparison of
of different
different calcium
ALUMINATES
- Proceedings
of the
fields:
calcium aluminate
aluminate technologies.
technologies.CALCIUM
CALCIUM
ALUMINATES
- Proceedings
of International
the Internationa
[6] Le
Bihan,
T. (2010).
Etude du comportement des chapes autonivelantes en ciment sulfo-alumineux: outils
Conference
2014,

Avignon,
France.
Conference
2014,
Avignon,
France.
exp´erimentaux et de mod´elisation. PhD thesis, INSA Lyon, 2010, in French.
5
[7] Klaus, S. R., Neubauer, J., and Goetz-Neunhoeffer, F. (2013). Hydration kinetics of CA2 and CA–
investigations performed on a synthetic calcium aluminate cement. Cement and Concrete Research, 43:
62–69.
[8] Damidot, D., Rettel, A., and Capmas, A. (1996). Action of admixtures on fondu cement: Part 1. Lithium
and sodium salts compared. Advances in Cement Research, 8(31):111–119.
[9] Damidot, D., Rettel, A., Sorrentino, D., and Capmas, A. (1997). Action of admixtures on fondu cement:

49


Lam, N. N. / Journal of Science and Technology in Civil Engineering

[10]
[11]
[12]

[13]
[14]
[15]
[16]

[17]

II. Effect of lithium salts on the anomalous setting time observed for temperatures ranging from 18 to
35˚C. Advances in Cement Research, 9(35):127–134.
Martin, I., Patapy, C., and Cyr, M. (2014). Parametric study of binary and ternary ettringite-based systems.
In Proceedings of the International Conference on Calcium Aluminates, Avignon, France.
Scrivener, K. L. and Capmas, A. (1998). LEA’s Chemistry of Cement and Concrete, chapter 13: Calcium
Aluminate Cements. Elsevier Science & Technology Books - London.
Le Saout, G., Lothenbach, B., Taquet, P., Fryda, H., and Winnefeld, F. (2014). Hydration study of a
calcium aluminate cement blended with anhydrite. In Proceedings of the International Conference on
Calcium Aluminates, Avignon, France.
Andreani, P. and Touzo, B. (2014). Mineral composition and hydration of a C12 A7 rich binder. In
Proceedings of the International Conference on Calcium Aluminates, Avignon, France.
Lamberet, S. (2005). Durability of ternary binders based on Portland cement, calcium aluminate cement
and calcium sulfate. PhD thesis, Ecole Polytechnique Federale de Lausanne, Switzerland.
Georgin, J. F. and Prud’homme, E. (2015). Hydration modelling of an ettringite-based binder. Cement
and Concrete Research, 76:51–61.
Bayoux, J., Bonin, A., Marcdargent, S., Verschaeve, M., and Mangabhai, R. J. (1990). Study of the
hydration properties of aluminous cement and calcium sulphate mixes. Calcium aluminate cements - E.
& F. N. Spon, Chapman & Hall, London.
Lura, P. and Jensen, O. M. (2007). Measuring techniques for autogenous strain of cement paste. Materials
and Structures, 40(4):431–440.

50