Cement and Concrete Research 77 (2015) 69–75

Contents lists available at ScienceDirect

Cement and Concrete Research

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

Influence of curing temperature on the process of hydration of

supersulfated cements at early age
C. Angulski da Luz a,⁎, R. D.Hooton b
a
Universidade Tecnológica Federal do Paraná Departamento de Engenharia Civil Campus Pato Branco, 85501970 Pato Branco (PR), Brazil
b
Department of Civil Engineering, University of Toronto, 35 St. George St, Toronto, Ontario, Canada M5S 1A4

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

Article history: In this paper, the influence of curing temperature was studied for supersulfated cements made with two slags
Received 5 March 2015 having different chemical compositions. Supersulfated cements (SSC) made with low-alumina slag developed
Accepted 3 July 2015 lower porosity, higher compressive strength and degree of hydration at higher temperatures. SSC made with
Available online 24 July 2015
high-alumina slag resulted in higher strengths and presented a more complex mechanism of hydration that
was strongly influenced by the solubility of anhydrite.
Keywords:
Blast-furnace slag
© 2015 Elsevier Ltd. All rights reserved.
Supersulfated cement
Temperature
Hydration

1. Introduction
Relative to other building materials, concrete has a low embodied
energy and carbon footprint, however, due to the enormous quantity
of concrete placed each year, the resulting total embodied energy and
carbon footprint is quite large. The concrete industry contributes around
5% to total global CO2 emissions and approximately 90% of this associat-
ed embodied energy and carbon is due to the commonly used Portland
cement binder [1,2].
While the concrete industry commonly makes use of some industrial
by-products as supplementary cementitious materials to reduce cement
content, these replacements are typically in the range of 15–30%. This
paper focuses on supersulfated cements (SSC), where Portland cement
contents are reduced by 85–90% using ground granulated blast-
furnace slag (GBFS) and calcium sulfate.
Supersulfated cements (SSC) were used in the 1950–1960s in
Europe, particularly for mass concrete applications. Later changes in
the iron manufacturing processes yielded slags that no longer fulfilled
the minimum Al2O3 content requirement (13%), leading to its use in
blends with Portland cement (GBFS-cement) [2].
Recently, the European standard for supersulfated cement was su-
perseded by the norm EN 15743/2009, which no longer requires a min-
imum Al2O3 content for the GBFS. The requirements “consist of at least
two-thirds by mass of the sum of calcium oxide (CaO), magnesium
oxide (MgO) and silicon dioxide (SiO2). The remainder contains
⁎ Corresponding author. Tel.: + 55 46 3220 2561.
E-mail addresses: angulski@utfpr.edu.br (C. Angulski da Luz), d.hooton@utoronto.ca
(H. R. D.).
aluminum oxide (Al2O3) together with small amounts of other com-
pounds. The ratio by mass of (CaO + MgO)/(SiO2) shall exceed 1.0” [3].
SSC are primarily comprised of blast furnace slag, (80–85%), calcium
sulfate (10–20%) and an alkaline activator, which is often Portland ce-
ment, though in a relative small quantity (around 5%). The small pro-
portion of Portland cement initially hydrates, providing calcium
hydroxide (CH), which promotes dissolution of the slag. The dissolved
aluminum, calcium and silicon ions from the GBFS react with the
added calcium sulfate (CS) forming ettringite (C 6 AS3 H 32 ) and calcium
silicate hydrate (C–S–H) phases, as demonstrated in this simplified
Eq. 1 [4].
C 5 S3 A þ CH þ 3CS þ 34H → C 6 AS3 H32 þ 3CSH ð1Þ
However, the complete series of reactions is complex and not well
understood. The hydraulic activity of slag plays a very important role
on the process of hydration of SSC and depends on the amorphous
phase, lime and alumina contents [5–8]. Gruskovnjak et al. [8] verified
that the optimum chemical composition of the slag is more important
to its reactivity than the amount of the amorphous phase. They also
found that minor hydrate phases, such as AFm phases or hydrotalcite,
can also occur as product of hydration.
Besides requiring less energy and producing lower greenhouse gas
emissions in its production, supersulfated cement (SSC) can provide re-
sistance to sulfate attack and lower heat of hydration [8–10]. However,
the main disadvantage is that the strength development of SSC is slow
and strongly dependent on temperature [6], which is also typically ob-
served in Portland cement blends with high slag contents. GBFS blends
are not therefore used in applications where the environment tempera-
ture is low or where high early age strength is required. The drawback in

http://dx.doi.org/10.1016/j.cemconres.2015.07.002
0008-8846/© 2015 Elsevier Ltd. All rights reserved.
70 C. Angulski da Luz, H. R. D. / Cement and Concrete Research 77 (2015) 69–75

Table 1 and SiO2 contents, but both slags meet the EN 15743 limits (Table 1).
Chemical composition of slag (A and B) and cement. X-ray diffraction (XRD) showed that both slags had a glass structure
Oxides Slag A Slag B EN 15743 Cement and slag B also contained minor amounts of calcite and anhydrite
SiO2 39.37 32.56 – 20.09
added during grinding (Fig. 1). Fig. 2 shows that the slags had very sim-
Al2O3 8.15 13.74 – 5.35 ilar grain size distribution, both with a d50 of 9 μm. A Canadian type GU
Fe2O3 0.63 0.41 – 2.30 (CEM I) [13] Portland cement (Table 1) was used as the alkaline
CaO 38.90 42.27 – 62.49 activator.
MgO 11.08 6.04 – 2.50
To obtain anhydrite (CaSO4) for the SSC, gypsum (CaSO4·2H2O) was
SO3 2.63a 2.00a – 4.04
Na2O 0.29 0.19 – 1.05 burned for 1 h at 650 °C [5] and the mineralogical composition was con-
K2O 0.35 0.31 – firmed by XRD (Fig. 3).
P2O5 0.01 0.03 – – The SSC was composed of 85% slag, 10% CaSO4 and 5% GU cement. To
TiO2 0.35 0.59 – – study the influence of temperature, two initial curing regimes were
Mn2O3 0.57 0.30 – –
CaO + MgO + SiO2 89.4 80.9 ≥66.7 –
investigated:
Free lime 1.12
(CaO + MgO)/SiO2 1.3 1.5 ≥1.0 – - Initial Curing Regime 1: cured at 23 °C and 95%RH for 24 h.
a - Initial Curing Regime 2: cured immediately after casting at different
Added SO3 from gypsum.
temperatures of 13, 23, 38, 50 °C and 95%RH for 24 h.
- After the initial curing period, the specimens were demolded and
immersed in lime water solution at different temperatures of 13,
the use of GBFS concrete is that its strength development is considerably 23, 38 and 50 °C up to 7 days of age.
slower under standard 20 °C curing conditions [11]. Barnett et al. [11]
investigated the effect of curing temperature on the development of
strength of Portland cement mortars containing GBFS. They observed Both slags A and B were studied under Initial Curing Regime 1. Initial
that the early-age strength was higher at higher temperatures and Curing Regime 2 was conducted only for slag B (high-alumina). For the
much more sensitive at higher levels of GBFS, related to its higher acti- compressive strength tests, X-ray diffraction (XRD), differential and
vation energy. In relation to supersulfated cement (SSC), no studies thermogravimetric analysis (DTA and TGA), and mercury intrusion
have been noted concerning the influence of temperature on its process porosimetry (MIP), mortars and pastes were studied (Table 2).
of hydration. Moreover, unlike slag, the solubility of calcium sulfate, Compressive strength of 50 mm mortars cubes (1:2.75:0.485—
present in SSC up to 20% [3], decreases at high temperatures [12]. cement:sand:water) was determined at 7 days, as per ASTM C109 [14].
This contribution highlights research conducted to investigate the Pastes prepared for MIP, XRD and DTA/TGA analysis at 7 days, were
effects of temperature on the process of hydration, compressive crushed into particles smaller than 9.5 mm, immersed in acetone for 2 h
strength and porosity of supersulfated cement (SSC). Two slags with dif- and dried by filtration to remove the pore fluid. These samples were
ferent chemical compositions (mainly the Al2O3 content) were used to crushed again to obtain particles smaller than 75 μm. These powders
evaluate the performance of SSC under low temperature (lower than were submitted to XRD analysis using CuKα radiation, over a 2θ range
20 °C) and higher temperatures typically experienced at early ages or of 3–70°, using steps of 0.05° and 1 s counting time, and to DTA/
during thermal curing. TGA in nitrogen at 10 °C/min. For the MIP, the samples were mea-
sured using a Quantachrome Autoscan 60 porosimeter applying
2. Materials and methods pressure from 0 to 414 MPa, assuming a contact angle of 140°. Calo-
rimetry was performed using a TAM Air isothermal calorimeter, on
Two slags were used to produce the supersulfated cement (SSC). pastes made from 20 g of SSC with w/c 0.4, over a hydration period
Slag A had relatively lower amounts of Al2O3 and CaO and higher MgO of 7 days.

Fig. 1. X-ray diffraction patterns of slags A and B.

 C. Angulski da Luz, H. R. D. / Cement and Concrete Research 77 (2015) 69–75 71

Fig. 2. Grain size distribution of slags A and B.

3. Results similar: 19 MPa, 18 MPa and 16 MPa respectively. Under Curing Regime
1, the temperature of 50 °C resulted in the highest strength, whereas
3.1. Compressive strength under Curing Regime 2, 50 °C resulted in the lowest strength.

Analysis of the results showed that the compressive strengths for
mortars made with slag A (low-alumina) were always much lower
than for slag B (high alumina) as was also observed by Gruskovnjak
et al. [8] (Fig. 4). For slag A, the higher the curing temperature, the
higher the compressive strength. At 50 °C, the strength was almost
twice that at 13 °C. For slag B, under Curing Regime 1, the mortars
cured at 13 °C exhibited higher strength than those cured at 23 °C, but
the mortars cured at 38 °C and 50 °C also exhibited higher compressive
strengths than at 23 °C. However, when the mortars were exposed to
different temperatures immediately after casting (Curing Regime 2),
the values changed significantly. Under these conditions, the lower
the temperature, the higher the compressive strength. At 13 °C, the
strength was 28 MPa, while those cured at 23, 38 and 50 °C were very
3.2. Calorimetric measurements
The pastes made with slag A and B were tested by isothermal calo-
rimetry for up to 7 days of hydration (Fig. 5). The heat flow rate of pastes
prepared with slag A (low alumina), for all temperatures, was much
lower than that for pastes made with slag B, as also noted in
Gruskovnjak et al. [8]. For slag A, lower and delayed activity was ob-
served, with a long induction period and a wide main hydration peak.
At 13 °C the heat flow rate was very low.
For both slags, the higher curing the temperature, the higher the
heat flow rate, the shorter the induction period and the narrower the
main hydration peak. In addition, temperatures higher and lower than
23 °C influenced the main peak similarly for both slags: at 13 °C the

Fig. 3. X-ray analysis of gypsum (G) and anhydrite (A) used.

72 C. Angulski da Luz, H. R. D. / Cement and Concrete Research 77 (2015) 69–75

Table 2 Table 3
Tests, samples and method of curing. Influence of temperature on the heat flow rate of pastes made with slag A and B.

Tests Slag A Slag B Sample Temperature Main peak (h) Main peak at T °C (h)/main
peak at 23 °C (h)
Compressive strength Curing Regime 1 Curing Regime 1 and 2 Mortar
XRD and DTA/TG Curing Regime 1 Curing Regime 1 and 2 Paste w/c 0.4 Slag A Slag B Slag A Slag B
MIP Curing Regime 1 Curing Regime 1 Paste w/c 0.4
13 °C 83 33 2.52 2.54
23 °C 33 13 1.00 1.00
38 °C 10 5 0.30 0.38
50 °C 7 4 0.21 0.31

indicating that at this temperature the release of heat due the reactions
was lower than that to keep the system at 50 °C.

3.3. Products of hydration

Fig. 4. Compressive strength of mortar cubes made with slags A (Curing Regime 1) and B
(Curing Regime 1 and 2) at 7 days.
main peak occurred 2.5 h later than at 23 °C, and at 38 and 50 °C the
main peak occurred 0.2–0.4 h earlier than at 23 °C (Table 3). For both
slags, the total heat release curves showed an early gain of energy at
13 °C and a loss of energy at 38 and 50 °C due the storage temperature
of material (23 °C). Also for both slags, after 20 h, at 38 °C, the accumu-
lative heat remained practically constant while at 50 °C it dropped
The pastes made with slags A and B under Curing Regime 1 were an-
alyzed using X-ray diffraction at 7 days of hydration (Fig. 6). For slag A,
at 13 °C, the anhydrite was all consumed and gypsum was the main
crystalline product of hydration. As the curing temperature increased,
less anhydrite was consumed and less gypsum was formed. For all cur-
ing temperatures, the main peak of ettringite was very similar, but
seemed a little more intense at 13 and 23 °C curing.
For slag B under Curing Regime 1, lower curing temperatures also
increased the consumption of anhydrite; however, it was not all con-
sumed during 13 °C curing, as observed with Slag A. For all tempera-
tures, the main peaks of ettringite and gypsum were very similar.
For the pastes made with slag B under Curing Regime 2 at 24 h, for
13 °C curing, the diffraction patterns showed that gypsum was the
main product and the formation of ettringite was very weak. At 23 °C,
both ettringite and gypsum were identified. At 38 and 50 °C, only
ettringite was identified and the peaks of gypsum were very weak. At
7 days, for 13 °C curing, the anhydrite and gypsum were practically all
consumed. The higher the curing temperature, the lower was the

Fig. 5. Isothermal calorimetric measurements: heat flow rate and total heat of SSC made with slags A and B.
 C. Angulski da Luz, H. R. D. / Cement and Concrete Research 77 (2015) 69–75
Fig. 6. X-ray analysis of pastes: slag A (Curing Regime 1 at 7 days) and slag B (Curing Regime 1 at 7 days and Curing Regime 2 at 24 h and 7 days).
E: ettringite, G: gypsum, A: anhydrite, C: calcite.
consumption of anhydrite and the lower was the formation of ettringite.
At 13 °C, a significant difference in XRD patterns was observed from 24 h
up to 7 days: high consumption of anhydrite and formation of ettringite.
The DTA/TGA curves are shown in Fig. 7 for the pastes made with
slags A and B. For slag A Curing Regime 1 at 7 days cured at 13 °C, the
presence of ettringite (AFt) and gypsum was confirmed by peaks at
140 °C and 160 °C respectively, and was also observed by XRD. The
peak of C-S-H appeared to be present at 118 °C but it overlapped with
the Aft peak due to the similar temperatures at which they occur [4,8].
For higher curing temperatures, the DTA curves were similar, as were
the phases observed by XRD. The TGA curves showed a total mass loss
between 14.5% and 17.4%, consistent with the compressive strength
test: the higher the curing temperature, the higher the total mass loss,
indicative of a higher degree of hydration and higher compressive
strength values at higher temperatures (Table 4).
For slag B and Curing Regime 1, the DTA curves confirmed the pres-
ence of ettringite and appeared to indicate C–S–H at 118 °C. For all cur-
ing temperatures, the pastes presented similar total mass losses
between 17.5% and 18.2% (Table 4) showing better hydration than
slag A, but not totally consistent with the compressive strength data.
Even for slag B, when the pastes were cured at different tempera-
tures just after demolding (Curing Regime 2), at 24 h and 13 °C curing,
the DTA curves showed the presence of ettringite and gypsum as the
main products of hydration. When the temperature increased, the for-
mation of ettringite also increased, while gypsum decreased, exactly
as observed in the XRD analysis. At 24 h, the higher the temperature,
the higher the total mass loss, showing higher degrees of hydration.
At 7 days, for all curing temperatures, no gypsum was identified and
only Aft was observed as a product of hydration, confirming the XRD
analyses. The total mass loss was lower for 38 and 50 °C curing tem-
peratures, indicating lower degrees of hydration at higher curing
temperatures.
73
3.4. Porosimetry by MIP
For the pastes prepared with slag A under Curing Regime 1, the MIP
porosity at 7 days confirmed the compressive strength results: the
higher the curing temperature, the lower the porosity (by cumulative
intrusion of Hg) and the smaller the threshold pore diameters (Fig. 8).
For the pastes made with slag B under Curing Regime 1 at 7 days, the
MIP porosities were not completely consistent with the compressive
strength results: the paste cured at 13 °C showed higher porosity (by
cumulative intrusion of Hg) and larger threshold pore diameters than
at 23 °C. However, for specimens cured at 38 and 50 °C, the porosity
was lower (but very similar) than that at 23 °C, confirming the higher
mortar compressive strengths. The MIP data also confirmed a lower po-
rosity and smaller threshold pore diameter for the pastes made with the
high-alumina slag (slag B).
4. Discussion
For the low-alumina slag, which was only studied under Curing Re-
gime 1, the mortars cured at higher temperatures reached higher com-
pressive strength. Pastes cured at higher temperatures exhibited lower
porosity and a higher degree of hydration. The formation of ettringite
was weak and was affected by the curing temperature. At a lower curing
temperature (13 °C), the consumption of anhydrite and the formation of
gypsum increased, while at higher temperatures (38 and 50 °C) the
opposite was observed. The DTA analysis indicated higher C–S–H for-
mation at higher temperatures, which would explain the better me-
chanical performance.
The high-alumina slag (slag B) SSC cured under the same conditions
as slag A (Curing Regime 1) also exhibited the highest compressive
strengths and lower porosity when cured at higher temperatures.
74 C. Angulski da Luz, H. R. D. / Cement and Concrete Research 77 (2015) 69–75

Fig. 7. DTA/TG analysis of pastes: slag A (Curing Regime 1 at 7 days) and slag B (Curing Regime 1 at 7 days and Curing Regime 2 at 24 h and 7 days).
CSH: calcium silicate hydrates, Aft: ettringite, G: gypsum, Ht: hydrotalcite, Afm: monosulfate, Cc: calcite.

However, when cured at 13 °C, the mortars exhibited higher compres-

Table 4 sive strength than at 23 °C, a higher consumption of anhydrite, and a
Total mass loss (%) of pastes made with slag A and B at 800 °C. slightly higher formation of ettringite and gypsum. When the mortars
Temperature Slag A Slag B were cured at different temperatures immediately after casting (Curing
Curing Regime 1 Curing Regime 1 Curing Regime 2
Regime 2), it was found that at 24 h, pastes cured at lower temperatures
promoted higher consumption of anhydrite and formation of gypsum,
7d 7d 24 h 7d
but lower formation of ettringite and a lower degree of hydration. How-
13 °C 14.5 17.5 7.2 15.5 ever, at 7 days, the total consumption of anhydrite and gypsum (to form
23 °C 15.9 18.2 10.4 18.2
ettringite) was observed only for the pastes cured at 13 °C. The highest
38 °C 16.9 17.9 12.5 13.0
50 °C 17.4 17.6 11.2 13.3
degree of hydration was observed at lower temperatures (13 and 23 °C),
explaining the higher compressive strength at lower temperatures.

Fig. 8. MIP of pastes prepared with slags A and B at 7 days (Curing Regime 1).
 C. Angulski da Luz, H. R. D. / Cement and Concrete Research 77 (2015) 69–75
For both slags, curing at higher temperatures significantly inhibited
the consumption of anhydrite. Curing at 38 and 50 °C immediately
after casting reduced the dissolution of anhydrite and prevented its re-
action with slag, reducing the compressive strength. In contrast, curing
at 38 and 50 °C after 24 h, allowed a higher dissolution of anhydrite and
seemed to promote the formation of C–S–H, improving compressive
strengths. Curing at low temperature (13 °C) promoted the total disso-
lution of anhydrite (under Curing Regime 2), the formation of gypsum
and its reaction with slag, improving the compressive strength and pro-
viding the highest strength: 27 MPa at 7 days. The solubility of anhydrite
played a very important role in the process of hydration.
5. Conclusions
In this paper, the influence of temperature on the process of hydra-
tion of supersulfated slag cements was studied. Two methods of curing
were used: Initial Curing Regime 1 (24 h at 23 °C followed by 6 days at
either 13, 23, 38 or 50 °C) and Initial Curing Regime 2 (placed immedi-
ately after casting at 13, 23, 38 and 50 °C for 7 days) for both low-
alumina and high-alumina slags. Based on the analyses of the results,
it is possible to draw the following conclusions:
- The SSC made with high-alumina slag resulted in higher compres-
sive strengths at 7 days than SSC made with low-alumina slag, for
all temperatures,
- For low-alumina slag, high temperature curing increased the me-
chanical performance up to 25% but did not increase the formation
of ettringite. The hypothesis is that the temperature only increased
the formation of CSH;
- For high-alumina slag, the thermal curing strongly influenced the
mechanical performance (up to 45%) at both high and low tempera-
tures. Both temperature and the start time of the thermal curing pro-
cess strongly influenced the process of hydration, increasing or
decreasing the compressive strength values. A better compressive
strength development of an SSC maintaining a low temperature dur-
ing the first hours can be achieved.
- The performance of SSC made with high-alumina slag under differ-
ent curing temperatures showed a complex system of hydrates,
mainly due to the availability of sulfate of calcium (related to solubil-
ity of anhydrite) to form ettringite;
- High and low environmental temperatures (38 and 13 °C) during the
first 24 h can significantly influence the mechanisms of hydration
75
and consequently the mechanical performance of supersulfated
cements.
Acknowledgments
The authors wish to thank CNPq (National Counsel of Technological
and Scientific Development) (483661/2013-9) and CAPES (Coordina-
tion for the Improvement of Higher Education Personnel) (3818-11-0)
in Brazil for its support as well as the Natural Sciences and Research
Council of Canada (386860-14).
References
[1] J.S. Damtoft, J. Lukasik, D. Herfort, D. Sorrentino, E.M. Gartner, Sustainable develop-
ment and climate change initiatives, Cem. Concr. Res. 38 (2) (2008) 115–127.
[2] M.C.G. Juenger, F. Winnefeld, J.L. Provis, J.H. Ideker, Advances in alternative cemen-
titious binders, Cem. Concr. Res. 41 (2011) 1232–1243.
[3] EN 15743, Supersulfated Cement—Composition, Specification and Conformity
Criteria, European Committee for Standardization (CEN), Brussels, Belgium, 2009.
[4] Z. Grounds, D.V. Nowell, F.W. Wilburn, The influence of temperature and different
storage conditions on the stability of supersulphated cement, J. Therm. Anal. 41
(1994) 687–699.
[5] D.K. Dutta, P.C. Borthakur, Activation of low lime high alumina granulated blast fur-
nace slag by anhydrite, Cem. Concr. Res. 20 (1990) 711–722.
[6] J. Bijen, E. Niel, Supersulphated cement from blast furnace slag and gypsum avail-
able in the Netherlands and neighbouring countries, Cem. Concr. Res. 11 (1981)
307–322.
[7] V.P. Mehrotra, A.S.R. Sai, P.C. Kapur, Plaster of Paris activated supersulfated slag ce-
ment, Cem. Concr. Res. 12 (1982) 463–473.
[8] A. Gruskovnjak, B. Lothenbach, F. Winnefeld, R. Figi, S.-C. Ko, M. Adler, U. Mäder, Hy-
dration mechanisms of super sulphated slag cement, Cem. Concr. Res. 38 (2008)
983–992.
[9] Z. Grounds, D.V. Nowell, F.W. Wilburn, Resistance of supersulphated cement to
strong sulfate solutions, J. Therm. Anal. Calorim. 72 (2003) 181–190.
[10] E. Erdem, H. Olmez, The mechanical properties of supersulfated cement containing
phosphogypsum, Cem. Concr. Res. 23 (1993) 115–121.
[11] S.J. Barnett, M.N. Soutsos, S.G. Millard, J.H. Bungey, Strength development of mortars
containing ground granulated blast-furnace slag: effect of curing temperature and
determination of apparent activation energies, Cem. Concr. Res. 36 (2006) 434–440.
[12] C. Comel, E. Karmazsin, M. Murat, Theoretical interpretation of the existence at 35 °C
of an absolute maximum of the hydration of calcium sulphate hemihydrate,
Thermochim. Acta 31 (1979) 71–77.
[13] Canadian Standards Association - CSA Standards A3001-08, Cementitious Materials
Used in Concrete., Ontario: Canada, 2008.
[14] American Society for Testing and Materials – ASTM, Standard Specification for Com-
pressive Strength of Hydraulic Cement Mortars (Using 2-in. or [50-mm] Cube Spec-
imens), ASTM C109/C, West Conshohocken, PA, United States, 2008.