Early-Age Volume Change and Hydration of Expansive Cements

Full text

Proc. of the 10th fib International PhD Symposium in Civil Engineering
July 21 to 23, 2014, Université Laval, Québec, Canada
Early-Age Volume Change and Hydration of Expansive
Cements
Piyush Chaunsali and Paramita Mondal
Department of Civil and Environmental Engineering,
University of Illinois,
Newmark Civil Engineering Building, 205 N. Mathews Ave., Urbana, IL – 61801 (USA)
Abstract
Shrinkage cracking is a predominant deterioration mechanism in concrete structures with high
surface-to-volume ratio. As a way to reduce the shrinkage cracking, expansive cements were
developed, which utilize the early-age expansion to induce the necessary compressive stress for
counteracting the tensile stress developed due to drying shrinkage. The goal of the current study was
to investigate the early-age volume change and hydration of a calcium sulfoaluminate (CSA)-based
cement. The effect of w/cm ratio (0.34 and 0.44), curing condition (sealed and lime water) and
mineral admixture (Class C FA, Class F FA and silica fume) was examined on the expansion
characteristics of CSA-based cement. The results highlight that the slow hydration of ye’elimite at
low w/cm ratio resulted in reduced expansion. The lime water curing singificantly increased the
expansion. Furthermore, the incorporation of mineral admixtures altered the expansion characteristics
by modifying the material stiffness and hydration of ye’elimite.
Keywords: Shrinkage, Expansive Cements, Calcium Sulfoaluminate, X-ray Diffraction
1 Introduction
Concrete deterioration by shrinkage led to the development to expansive cements based on calcium
sulfoaluminate (CSA) clinker [1]. The formation of ettringite crystals during the early-age hydration
of CSA-based cement induces the compressive stress in concrete which helps counteract the tensile
stress developed during shrinkage. The CSA-based cement also provides an environmentally-friendly
alternative to portland cement as its production results in lower carbon foot-print. The CSA clinker is
manufactured at lower kiln temperature (~1250 °C) than that used for portland cement clinker. The
lower demand of limestone as a starting material also reduces the CO2 foot-print. Additionally, the
porous nature of CSA clinker decreases the energy demand for its grinding. Therefore the use of
CSA-based concrete promises a sustainable and durable concrete.
CSA clinker mainly constitutes ye’elimite (C4A3Ŝ), C2S and C4AF. Calcium sulfate is then added
to CSA clinker for desired level of expansion. The main hydration reactions in CSA-based cement are
shown below:
C4A3Ŝ + 18 H C3A.CŜ.12H + 2AH3 (1)
C4A3Ŝ + 2 CŜH2 + 34 H C3A.3CŜ.32H + 2AH3 (2)
C4A3Ŝ + 6 CH + 8 CŜH2 + 74 H 3 C3A.3CŜ.32H (3)
C2S + 2H CH + C-S-H (4)
where C = CaO, A = Al2O3, Ŝ = SO3, H = H2O
It is the reaction between ye’elimite and calcium sulfate which results in the formation of ettringite.
According to the reaction (2), a w/cm ratio of 0.78 is required for the complete hydration of ye’elimite
which leads to higher water demand for the hydration of CSA cement than portland cement. In ab-
sence of calcium sulfate, monosulfate (AFm) is formed instead of AFt. This could be the case when
the calcium sulfate is consumed before complete hydration of ye’elimite. Among the existing theories
on the expansion mechanism, the crystallization pressure has been regarded as a prevailing mecha-
nism [2-5]. Some researchers have also suggested the swelling of fine ettringite crystals as the main
expansion mechanism [6].
The various factors of physico-chemical nature affect the expansion of CSA-based cement. The
water-to-cementitious (w/cm) ratio, CŜ/ye’elimite ratio, pore structure features and material stiffness
are among the important ones [7, 8]. No simple relation between the expansion and ettringite content

10th fib International PhD Symposium in Civil Engineering
has been found. The pore size distribution affects the expansion potential in a way that the growth of
ettringite in smaller pores results in crystallization pressure necessary for the expansion [5].
Mineral admixtures are used to improve the durability of conventional portland cement concrete.
However, there are only a few studies on the influence of pozzolans on the expansion characteristics
of CSA-based cement [9]. The goal of current study was to investigate the effect of Class ‘C’ fly ash,
Class ‘F’ fly ash and silica fume on the hydration and expansion characteristics of CSA-based cement.
Additionally, the effect of w/cm ratio and curing condition is also examined on expansion.
2 Materials and Methods
The study involved using Type I portland cement (OPC), an expansive admixture, Class ‘C’ FA,
Class ‘F’ FA and silica fume (SF). A calcium sulfoaluminate-based admixture, manufactured by CTS
Company, was used in the current study. The expansive admixture had main phases as C4A3Ŝ, CaSO4,
C2S and C4AF. Table 1 shows the chemical composition of all raw materials.
Table 1. Chemical composition of raw materials
Materials
SiO2
Al2O3
Fe2O3
CaO
MgO
SO3
Na2Oeq
LOI
Type I
20.93
4.45
2.72
63.28
3.03
2.44
0.52
1.98
Exp. Adm.*
7.70
7.00
1.17
50.07
0.08
26.04
0.56
2.10
‘C’ FA
37.76
19.43
5.33
25.56
4.09
2.23
1.07
0.58
‘F’ FA
59.08
22.43
8.39
1.59
1.06
0.2
2.07
2.99
SF
93.00
0.70
0.50
0.70
0.70
--
1.00
6.00
* Calcium sulfoaluminate-based admixture
Cement paste samples were prepared at w/cm ratio of 0.34 and 0.44. CSA-based admixture, Class
C FA, Class F FA and SF were used as 15%, 15%, 15% and 5% replacement (by mass) of total ce-
mentitious material, respectively. Table 2 shows the mix proportions used in this study. A water
reducing admixture (Glenium 7500) was used to prepare samples at w/cm ratio of 0.34 to improve the
workability.
Table 2. Mixture proportions
Mix Proportions
OPC
(wt. %)
Exp. Adm.
(wt. %)
‘C’ FA
(wt. %)
‘F’ FA
(wt. %)
SF
(wt. %)
OPC
100
0
0
0
0
OPC+K
85
15
0
0
0
OPC+K+‘C’FA
70
15
15
0
0
OPC+K+‘F’FA
70
15
0
15
0
OPC+K+SF
80
15
0
0
5
The experimental program included the measurement of unrestrained length change using corru-
gated tube according to ASTM C1698. Since the previous studies have shown that the fluid-to-solid
transition takes place at around final set time, the length measurements were also started at the same
time. To examine the effect of curing, companion prismatic samples (25.4 mm x 25.4 mm x 285 mm)
were prepared and demolded after 24 hours before immersing in saturated lime water. The deforma-
tion occurring between the final set time and 24 hours (determined using corrugated tube) was added
to the length change measurement of prismatic samples.
X-ray diffraction (XRD) and thermogravimetric (TG) analysis were performed to monitor the hy-
dration progress of CSA-based cement paste. The samples were initially ground, washed in acetone
and vacuum dried in desiccator for 24 hours before XRD and TG analysis. Additionally, compressive
strength of cubes (size: 50 mm x 50 mm x 50 mm) was determined after 7 days.

Preparation of papers for the 10th fib International PhD Symposium in Civil Engineering
3 Results and Discussions
The expansion characteristics of OPC and expansive cement pastes (OPC+K) were monitored in
saturated lime water, and are presented in Fig 1. The incorporation of CSA-based admixture signifi-
cantly increased the expansion of portland cement. At the w/cm ratio of 0.44, most of the expansion
was complete in 8-10 days whereas a steady growth was observed at the w/cm ratio of 0.34. The
expansion of CSA-based cement depends on the hydration of ye’elimite which results in formation of
ettringite. Slow hydration of ye’elimite appears to have resulted in steady rate of expansion observed
at low w/cm ratio. X-ray diffraction analysis (Fig. 2 (a) after 7 days shows the difference in peak
intensity of ye’elimite. For semi-quantitative analysis of ettringite content, TG analysis was utilized
considering the weight loss in the range of 70 to 120 °C [5]. Figure 2 (b) showed reduced weight loss
(from 70-120 °C) of 6.95% at w/cm – 0.34 compared to 7.61% at w/cm – 0.44. It appears that the
expansion was reduced at low w/cm ratio possibly due to slow hydration of ye’elimite which also
resulted in lower ettringite content. This highlights the importance of high water demand for the
complete hydration of CSA-based cement. Additionally, higher material resistance at low w/cm ratio
is expected to reduce the expansion to some extent.
Fig. 1 Unrestrained expansion of OPC and CSA-based cement paste in saturated lime water
curing
(a) (b)
Fig. 2(a) X-ray diffraction pattern, and (b) derivative thermogravimetric analysis of OPC+K ce-
ment paste at two w/cm ratios after 7 days of lime water curing
Curing condition has been found to have significant influence on expansion as the moist curing
increases the expansion potential [6]. Fig. 3 compares the expansion characteristics in sealed and
0 4 8 12 16 20 24 28 32
Curing in Lime Water (days)
0
1000
2000
3000
4000
5000
6000
7000
8000
Expansion (microstrain)
OPC(0.44)
OPC+K (0.44)
OPC+K(0.34)
0100 200 300 400 500 600 700
Temperature (
o
C)
0
0.05
0.1
0.15
0.2
0.25
Derivative Weight (%/
o
C)
w/cm - 0.44
w/cm - 0.34
6 9 12 15 18 21 24 27
2-Theta (degrees)
E
E
CH
EEY
w/cm - 0.44
w/cm - 0.34

10th fib International PhD Symposium in Civil Engineering
saturated lime water curing. It is evident that the expansion is significantly increased when the sam-
ples undergo lime curing. Although autogenous shrinkage occurs during the hydration of portland
cement (which constitutes 85% of total cementitious binder), the difference in the magnitude of ex-
pansion between sealed and lime water curing does not appear to have solely derived from it. The
sealed curing and the lime water curing did not exhibit difference in terms of weight loss in range of
70-120 °C which is mainly due thermal decomposition of ettringite (Fig. 4). The OPC+K cement
paste undergoing sealed curing and lime water curing showed 8.80% and 8.84% weight loss, respec-
tively. Previous studies have shown that small crystals of ettringite which are formed in presence of
lime tend to swell by imbibing the water [6]. The increase in expansion is approximately two-fold
which could be either be attributed to swelling of CSA hydration products [6] or increase in crystalli-
zation pressure due to higher water activity [4].
Fig. 3 Effect of curing regime on expansion characteristics of OPC+K sample at w/cm – 0.44
Fig. 4 Thermogravimetric analysis of OPC+K sample after 7 days at w/cm – 0.44
Figure 5 shows the expansion characteristics of expansive cement pastes incorporating mineral
admixtures. It is evident that all samples except the one with Class ‘C’ FA achieved the maximum
expansion in 7-8 days whereas Class ‘C’ FA-based mixture stopped expanding beyond 2 days. X-ray
diffraction analysis of hydrated cement pastes showed that the ye’elimite was consumed faster in the
mixture with Class ‘C’ FA. As shown in Fig. 6 (a), ye’elimite could not be detected at 4 days in the
cement paste with Class ‘C’ FA. The C3A and free lime content of Class ‘C’ FA appears to have
resulted in faster consumption of ye’elimite. Also, the CSA-based cement with SF had noticeable
amount of unhydrated ye’elimite even after 14 days as shown in Fig. 6 (b). A previous study has also
shown that the addition of SF suppresses the dissolution of ye’elimite by lowering the pore solution
alkalinity [9].
0 2 4 6 8 10 12 14
Sealed Curing (days)
0
1000
2000
3000
4000
5000
6000
7000
8000
Expansion (microstrain)
Lime Curing
Sealed Curing
0200 400 600 800
Temperature (
o
C)
70
75
80
85
90
95
100
Weight (%)
0
0.1
0.2
0.3
0.4
0.5
Derivative Weight (%/
o
C)
Lime Curing
Sealed Curing
TG
DTG

Preparation of papers for the 10th fib International PhD Symposium in Civil Engineering
Fig. 5 Expansion characteristics of CSA-based cement pastes in presence of mineral admixtures
(w/cm – 0.44)
The incorporation of mineral admixtures altered the expansion of CSA-based cement pastes. Class
‘F’ FA resulted in the maximum expansion whereas SF reduced the expansion. The samples incorpo-
rating the mineral admixtures had same amount of expansive component to begin with, as shown in
Table 2. Thermal analysis also showed (Fig. 7) that the mass loss in range of 70-120 °C was not
significant to have caused the noticeable change in expansion. This highlights the importance of
material stiffness which seems to have played a role in governing the expansion. The 7-day compres-
sive strength of OPC+K, OPC+K+‘F’FA, OPC+K+‘C’FA and OPC+K+SF pastes was 37.24 MPa,
24.82 MPa, 35.87 MPa and 41.89 MPa, respectively. The lowest strength of Class ‘F’ FA paste can
indirectly be related to lowest material stiffness which led to the highest expansion [10]. For SF-based
paste, increased stiffness, incomplete hydration of ye’elimite due to self desiccation and lower pH [9]
appeared to have reduced the expansion.
(a) (b)
Fig. 6 X-ray diffraction patterns of CSA-based cement paste after: (a) 4 days, and (b) 14 days
at w/cm – 0.44 (E – Ettringite, CH – Portlandite, Y – Ye’elimite)
0 4 8 12 16 20 24 28
Curing in Lime Water (days)
0
1000
2000
3000
4000
5000
6000
7000
8000
Expansion (microstrain)
OPC+K
OPC+K+'C'FA
OPC+K+'F'FA
OPC+K+SF
6 9 12 15 18 21 24 27
2-Theta (degrees)
Y
CH
E
EEE
OPC+K
OPC+K+'C'FA
OPC+K+'F'FA
OPC+K+SF
6 9 12 15 18 21 24 27
2-Theta (degrees)
Y
CH
E
EEE
OPC+K
OPC+K+'C'FA
OPC+K+'F'FA
OPC+K+SF

10th fib International PhD Symposium in Civil Engineering
Fig. 7 Thermogravimetric analysis of CSA-based cement pastes after 14 days at w/cm – 0.44
4 Conclusions
The current study investigated the early-age expansion characteristics of CSA-based cement. The
results can be summarized as below:
 The expansion of CSA-based cement was reduced at low w/cm ratio due to slow hydra-
tion of ye’elimite. It highlights the importance of using high w/cm ratio for practical ap-
plication of these cements.
 The effect of lime water curing was significant enough to have caused a two-fold in-
crease in expansion compared to sealed curing.
 The mineral admixtures were found to have altered the expansion of CSA-based cement
by modifying the material stiffness and hydration of ye’elimite.
References
[1] Klein, A.: Calcium Sulfoaluminate and Expansive Cements Containing Same. U.S. Patent
3,155,526 (1964).
[2] Bentur, A. and Ish-Shalom, M.: Properties of Type K Expansive Cement of Pure Compo-
nents II. Proposed Mechanism of Ettringite Formation and Expansion in Unrestrained Paste
of Pure Expansive Component. Cement and Concrete Research 4 (1974) No. 5, pp. 709-721.
[3] Cohen, M.D. and Richards, C.W.: Effects of the Particle Sizes of Expansive Clinker on
Strength-Expansion Characteristics of Type K Expansive Cements. Cement and Concrete
Research 12(1982) No. 6, pp. 717-725.
[4] Flatt, R.J., and Scherer, G.W.: Thermodynamics of Crystallization Stresses in DEF. Cement
and Concrete Research 38 (2008) No. 3, pp. 325-336.
[5] Bizzozero, J., Gosselin, C. and Scrivener, K. L. : Expansion Mechanisms in Calcium Alumi-
nate and Sulfo-aluminate Systems with Calcium Sulfate. Cement and Concrete Research 56
(2014) pp. 190-202.
[6] Mehta, P. K.: Mechanism of Expansion Associated with Ettringite Formation. Cement and
Concrete Research 3 (1973) No. 1, pp. 1-6.
[7] Lobo, C. L.: A study of Type K Expansive Cement Paste and Concrete, and the Influence of
Silica fume. PhD Thesis, Purdue University, 1991.
[8] Chen, I. A., Hargis, C. W. and. Juenger, M. C. G.: Understanding Expansion in Calcium
Sulfoaluminate-Belite Cements. Cement and Concrete Research 42 (2012) No. 1, pp. 51-60.
[9] Lobo, C. and Cohen, M. D.: Hydration of Type K Expansive Cement Paste and the Effect of
Silica Fume: II. Pore Solution Analysis and Proposed Hydration Mechanism. Cement and
Concrete Research 23 (1993) No. 1, pp. 104-114.
[10] Hoff, G. C.: Use of Expansive Cements in Large Sections of Grout and Mortar. ACI Special
Publication 1973, pp. 299-340.
0100 200 300 400 500 600
Temperature (
o
C)
60
70
80
90
100
Weight (%)
0
0.1
0.2
0.3
0.4
Derivative Weight (%/
o
C)
OPC
OPC+K
OPC+K+'C'FA
OPC+K+'F'FA
OPC+K+SF
Ettringite
AFm CH
TG
DTG