You are on page 1of 9

Construction and Building Materials 47 (2013) 1201–1209

Contents lists available at SciVerse ScienceDirect

Construction and Building Materials


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

Setting and mechanical properties of alkali-activated fly ash/slag


concrete manufactured at room temperature
N.K. Lee, H.K. Lee ⇑
Department of Civil and Environmental Engineering, Korea Advanced Institute of Science and Technology, Guseong-dong, Yuseong-gu, Daejeon 305-701, South Korea

h i g h l i g h t s

 The setting time was decreased with increasing amounts of slag, water glass and NaOH.
 The proper slag content in the mixture was determined to be 15–20% considering setting, slump and strength.
 The mean pore size of alkali activated fly ash/slag mortar was smaller than that of OPC.
 The increased mesopore volume resulted in decreases in the elastic modulus and long-term compressive strength.

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

Article history: This study aims to investigate the setting and mechanical properties of alkali-activated fly ash/slag con-
Received 14 September 2012 crete manufactured at room temperature. It also examines to what extent the slag in the alkali-activated
Received in revised form 3 May 2013 fly ash/slag mixture improves the mechanical properties of the mixture under room-temperature curing
Accepted 27 May 2013
conditions. A series of tests of the compressive strength, elastic modulus, splitting tensile strength, flow,
Available online 5 July 2013
setting time, and porosity of the alkali-activated fly ash/slag concrete were carried out. The test results
showed that the setting time decreased as the amount of slag and the concentration of the NaOH solution
Keywords:
increased. The proper slag content in an alkali-activated fly ash/slag mixture was determined to be 15–
Alkali activation
Fly ash/slag blended concrete
20% of total binder by weight considering the setting time and compressive strength of the alkali-acti-
Setting time vated fly ash/slag concrete cured at room temperature. In addition, the modulus of elasticity and splitting
Mechanical property tensile strength of the alkali-activated fly ash/slag concrete were slightly lower than those of ordinary
Room temperature manufacturing concrete as predicted by the ACI code and Eurocode 2. The total porosity of the alkali-activated fly
ash/slag mortar was similar to that of ordinary cement mortar, whereas the mean pore size tended to
be smaller than that of ordinary cement mortar.
Ó 2013 Elsevier Ltd. All rights reserved.

1. Introduction reaction of aluminosilicate materials (e.g., fly ash, metakaolin) with


an alkaline solution under ambient room conditions [12]. Accord-
With growing concern over environmental problems related to ing to previous research, when pozzolanic materials such as fly
global warming, there have been many efforts in the construction ash and metakaolin, which are rich in Si and Al, react with an alkali
field to reduce CO2 emissions. It was reported that Portland cement activator, good mechanical and physical properties can be obtained
production worldwide accounts for approximately 7% of the total [12]. Geopolymer has diverse advantages, such as a high mechan-
greenhouse gas emitted into the atmosphere [24], with about 1 ical strength of 62–66 MPa after curing at 60–75 °C for 24 h [28],
ton of CO2 emitted into the atmosphere during the production of good resistance to high temperatures [20], and chemical resistance
1 ton of ordinary Portland cement [25]. In particular, a large quan- to sulphates [2]. Based on these factors, alkali-activated geopoly-
tity of CO2 is generated through the fuel combustion and calcina- mer can be considered as a suitable material to replace ordinary
tion of carbonate in clinker kilns [34]. Thus, research geared Portland cement (OPC) due to its good environmental friendliness
towards the development of new alternatives to ordinary Portland and use of advanced technology.
cement is necessary. The proper alkali activator and appropriate curing temperature
The use of geopolymer is an alternative approach to reduce the are important factors to increase the compressive strength of geo-
use of Portland cement. Geopolymer is generally formed through a polymer. Fly ash-based geopolymer needs to be cured at a rela-
tively high temperature, above 60 °C [31,3,27] to obtain good
⇑ Corresponding author. Tel.: +82 42 350 3623; fax: +82 42 350 3610. mechanical properties as fly ash has low reactivity at room
E-mail address: leeh@kaist.ac.kr (H.K. Lee).
temperature. Accordingly, elevated-temperature curing is, in

0950-0618/$ - see front matter Ó 2013 Elsevier Ltd. All rights reserved.
http://dx.doi.org/10.1016/j.conbuildmat.2013.05.107
1202 N.K. Lee, H.K. Lee / Construction and Building Materials 47 (2013) 1201–1209

particular, a critical factor. Palomo et al. [26] concluded that an in- have been very little studied. The purpose of this paper is to inves-
crease in the curing temperature accelerated the activation of fly tigate the properties of alkali-activated fly ash/slag blended con-
ash with an alkali activator. In other work, van Jaarsveld et al. crete manufactured at room temperature through various
[31] studied the compressive strength change versus the curing experiments which assess the flow, setting time, compressive
age at various temperatures and showed that the compressive strength, strength development with time, elastic modulus, split-
strength increased by 100% when the curing temperature ranged ting tensile strength and porosity. The degree to which the slag
from 50 to 80 °C relative to that at 30 °C. In addition, Hardjito added to alkali-activated fly ash/slag concrete improves the
et al. [16] studied the compressive strength of a geopolymer mor- mechanical properties is also investigated here.
tar cured at 65, 70, and 80 °C for 24 h and showed that the early
compressive strength increased with the curing temperature. 2. Experimental program
These results show that curing temperature plays an important
role in the geopolymerization process of fly ash-based geopolymer. 2.1. Materials
However, high-temperature curing runs counter to the motiva-
In the present study, F-class fly ash and slag were used instead of Portland ce-
tion for not using Portland cement, i.e., reducing CO2 emissions, as ment. The chemical compositions and physical properties of the fly ash and slag are
it consumes more energy compared to room-temperature curing. listed in Table 1.
In addition, this type of curing process may diminish the degree NaOH with a purity level of 98% and water glass (Korean Industrial Standards,
of industrial field applicability. Therefore, the realization of the KS 3-grade; SiO2 (29%), Na2O (10%), H2O (61%), specific gravity 1.38 g/mL) were
used as alkali activators. Distilled water was used to dissolve the NaOH solid, and
room-temperature curing of geopolymer is critical for various the alkali activator was prepared by mixing NaOH solution with water glass. A solu-
commercial uses. tion of 85% phosphoric acid (H3PO4) was used as a set-retarding admixture to retard
Several studies have focused on adding slag containing CaO to the setting time of the alkali-activated fly ash/slag concrete. River sand and crushed
fly ash-based geopolymer in an effort to improve the compressive gravel were used as fine aggregate and coarse aggregate materials, respectively. The
physical properties of the sand and gravel are listed in Table 2.
strength of geopolymers cured under ambient room temperatures.
Several studies involving the use of calcium have reported that cal-
2.2. Mix proportion
cium has a positive effect on the compressive strength of the geo-
polymer [32,35]. Yip et al. [34] found that the form of added Ca2+ The mix proportion of the alkali-activated fly ash/slag paste is provided in Ta-
plays a significant role in improving the mechanical properties of ble 3 and denoted with specific codes. The labels ‘S’ , ‘M’, and ‘H’ represent the slag,
the geopolymer. It was reported that the simultaneous formation molarity of the NaOH solution and the phosphoric acid, respectively. The first num-
ber, ‘10’, ‘15’, ‘20’, ‘25’, or ‘30’, refers to the percentage of the slag replacement for
of a geopolymeric gel and calcium silicate hydrate (CSH) gel within
the fly ash by weight. The second number, ‘4’, ‘6’, or ‘8’, refers to the molarity
a single binder is possible [35]. Yip et al. [36] examined seven dif- (mol/l) of the NaOH solution. The third number, ‘0.5’, ‘1.0’, and ‘1.5’, refers to the ra-
ferent calcium-based materials to investigate the role of calcium in tio of water glass to NaOH solution by weight. The fourth number, 0.5–2.5, refers to
the geopolymerization process, finding that the effects of different the ratio of phosphoric acid to total binder (fly ash + slag) weight expressed as a
calcium silicates on geopolymerization depend significantly on the percentage.
Table 3 shows the mix proportions of alkali-activated fly ash/slag paste for the
crystallinity of the calcium-based materials and on the alkalinity of
setting time tests. The ratio of alkali activator to total binder (fly ash + slag) by
the alkali liquid used. weight was 0.38 for all specimens, and NaOH solutions with three different molar-
In a previous study of geopolymer mixed with fly ash and gran- ities (4 M, 6 M and 8 M) were chosen for the setting time tests. The mass ratios of
ulated blast furnace slag (GBFS), Kumar et al. [22] conducted the water glass to NaOH solution were 0.5, 1.0, and 1.5. The percentages of the slag
replacement for the fly ash by weight were 10%, 15%, 20%, 25% and 30%.
experiments to investigate the isothermal conduction calorimetry,
Because the 8 M NaOH solution caused the rapid setting of the alkali-activated
heat of hydration, microstructure, and compressive strength of fly ash/slag paste, the molarities of the NaOH solution were set to 4 M and 6 M and
geopolymer. In their study, they added GBFS at 5–50% of the total the mass ratios of the water glass to NaOH solution were 0.5 and 1.0 for compres-
binder by weight in an effort to improve the low reactivity of alka- sive strength tests. The mix proportions of alkali-activated fly ash/slag concretes
li-activated fly ash. Temuujin et al. [30] studied the influence of
CaO and Ca(OH)2 on the mechanical properties of fly ash-based Table 1
geopolymers and found that an addition of calcium to fly ash im- Properties of the binder materials.
proved the mechanical properties of the specimens cured at ambi- Fly ash (FA) Blast furnace slag
ent temperatures, whereas the mechanical properties of the
Chemical composition (%)
specimens cured at 70 °C instead showed a decrease. Lloyd et al. CaO 2.60 56.10
[23] revealed that fly ash/slag blended inorganic polymer cement SiO2 46.00 21.00
activated by silicate had a denser microstructure than that acti- Al2O3 33.00 17.00
vated by hydroxide. Recently, in alkali-activated granulated blast Fe2O3 10.50 0.62
SO3 – 0.77
furnace slag/metakaolin blends, the addition of metakaolin to the
slag mixture resulted in an increase in the total setting time, a Physical properties
Specific gravity (S.S.D) 2.38 2.90
reduction of the heat release and a slight reduction in the final Surface area (m2/kg) 290 (Blaine) 485 (Blaine)
mechanical strength of the mortars [4]. In geopolymers based on
Particle size (lm)
fly ash/slag blends, a moist curing condition led to a higher 10% 2.74 2.45
strength of the binders and a less porous system compared to a 50% 19.96 12.12
dry curing condition, as open air curing allowed water to evapo- 90% 95.33 32.89
rate, which reduced the water availability for the geopolymer ma-
trix to grow [17]. Thus, studies of the characteristics of fly ash-
based geopolymer concrete with slag, which contains a consider- Table 2
Properties of the aggregates.
able amount of CaO, may have significant implications with respect
to strength enhancement and the utilization of industrial by- Specific gravity Absorption
products. (%)
O.D. (oven dry S.S.D (saturated surface dry
To date, most studies on geopolymer mixed with fly ash and condition) condition)
slag have focused on micro-structural analysis, characterization, Sand 2.54 2.55 0.08
and the compressive strength of pastes and mortars. Mechanical Gravel 2.68 2.70 0.56
properties of geopolymer concrete mixed with fly ash and slag
N.K. Lee, H.K. Lee / Construction and Building Materials 47 (2013) 1201–1209 1203

Table 3
Mix proportion of alkali-activated fly ash/slag paste for the setting time tests.

Specimens Ratio of solution to binder Ratio of H2O to binder NaOH solution molarity (M) NaOH:water glassa Binder fly ash:slaga H3PO4 (%)
b
S20-8M0.5 0.38 0.28 8 1:0.5 80:20 0
S20-6M0.5b 0.38 0.28 6 1:0.5 80:20 0
S20-6M1.0b 0.38 0.27 6 1:1.0 80:20 0
S20-6M1.5b 0.38 0.26 6 1:1.5 80:20 0
S20-4M0.5b 0.38 0.30 4 1:0.5 80:20 0
S20-4M1.0b 0.38 0.28 4 1:1.0 80:20 0
S20-4M1.5b 0.38 0.27 4 1:1.5 80:20 0
S10-4M0.5b 0.38 0.30 4 1:0.5 90:10 0
S30-4M0.5b 0.38 0.30 4 1:0.5 70:30 0
S20-6M0.5c 0.38 0.28 6 1:0.5 80:20 0
S20-6M1.0c 0.38 0.27 6 1:1.0 80:20 0
S20-4M0.5c 0.38 0.30 4 1:0.5 80:20 0
S20-4M1.0c 0.38 0.28 4 1:1.0 80:20 0
S20-4M0.5c-H0.5 0.38 0.30 4 1:0.5 80:20 0.5
S20-4M0.5c-H1.0 0.38 0.30 4 1:0.5 80:20 1.0
S20-4M0.5c-H1.5 0.38 0.30 4 1:0.5 80:20 1.5
S20-4M0.5c-H2.0 0.38 0.30 4 1:0.5 80:20 2.0
S20-4M0.5c-H2.25 0.38 0.30 4 1:0.5 80:20 2.25
a
Mass ratio.
b
Specimens manufactured at 17 °C.
c
Specimens manufactured at 28 °C.

manufactured with four different types of alkali activators are listed in Table 4. The according to ASTM C 469-10. In addition, the pore size distribution of the alkali-
mix proportions were also determined according to the percentage (10–30%) of the activated fly ash/slag mortar was measured by the MIP method in accordance with
slag replacement for the fly ash by weight. The mass ratio of fine aggregate to bin- ASTM D 4284-07 using an Autopore VI machine by Micromeritics Corp. Mortar frag-
der (fly ash + slag) and the mass ratio of coarse aggregate to binder were 2 and 2.65, ments except for coarse aggregate materials were used for the MIP test. Splitting
respectively. tensile strength tests were conducted after 28 days of curing according to ASTM C
496-04.

2.3. Experimental details


3. Results and discussion
The method of producing alkali-activated fly ash/slag concrete is as follows.
First, fly ash, slag and fine aggregate were dry-mixed for 2 min to ensure homoge-
neity of the mixture. The mixer was stopped and the mixture was activated by add- 3.1. Effects of the alkali activator, slag, and H3PO4 solution on the
ing the alkali activator containing the NaOH solution and water glass, and this was setting of alkali-activated fly ash/slag paste
then mixed for an additional 2 min. The prepared coarse aggregate was subse-
quently mixed for 2 min. A slump test of fresh alkali-activated fly ash/slag concrete
was conducted following the ASTM C143 specification. The fresh alkali-activated fly
Setting time tests were performed in order to ascertain the ef-
ash/slag concrete prepared as described above was immediately cast into three fects of molarity of NaOH solution. The replacement ratio of the
molds with a 100 mm  200 mm cylinder for each mixture. slag for the fly ash by weight and the ratio of water glass to NaOH
All specimens were cured at a temperature of 20 °C and humidity level of 60% in solution were 20% and 0.5 by weight, respectively.
a room held at a constant temperature and humidity levels. After 1 day, all speci-
As shown in Fig. 1, the setting times of the 4 M specimens were
mens were removed from their molds and air-conditioned cured in a conditioning
room until the day of testing. an initial time of 55 min and a final time of 160 min, while those of
The setting time of the alkali-activated fly ash/slag paste was measured using a the 6 M specimens were an initial time of 50 min and a final time
Vicat needle according to ASTM C 191-08. Setting time tests were conducted under of 114 min. Those of the 8 M specimens were faster than any of the
two conditions, 17 °C and 28 °C, as the room temperature greatly affects the setting other specimens (their initial and final times were 10 min and
time and workability of the alkali-activated fly ash/slag concrete.
Compressive strength tests were conducted using a universal testing machine
50 min, respectively). Fig. 2 shows the effects of water glass on
according to ASTM C 39 at 3, 7, 14, 28 and 56 days. The static modulus of elasticity the setting time of the alkali-activated fly ash/slag paste. In the
was calculated based on the stress corresponding to 40% of the ultimate load 4 M specimens, as the amount of water glass increased, both the

Table 4
Mix proportion of alkali-activated fly ash/slag concrete.

Specimens Ratio of solution Ratio of H2O to NaOH solution Ratio of water glass Alkali solution Binder (kg/ Sand Gravel H3PO4
to binder binder molarity (M) to NaOH (kg/m3) m3) (kg/m3) (kg/m3) (kg/m3)
Fly Slag
ash
S20-6M0.5 0.56 0.42 6 0.5 215 293 88 760 1005
S20-6M1.0 0.56 0.40 6 1.0 215 293 88 760 1005
S20-4M0.5 0.56 0.44 4 0.5 215 293 88 760 1005
S20-4M1.0 0.56 0.42 4 1.0 215 293 88 760 1005
S10-4M0.5 0.56 0.44 4 0.5 215 344 34 760 1005
S15-4M0.5 0.56 0.44 4 0.5 215 316 63 760 1005
S20-4M0.5- 0.56 0.44 4 0.5 215 293 88 760 1005 8.5
H2.25
S25-4M0.5- 0.56 0.44 4 0.5 215 272 108 760 1005 8.5
H2.25
S30-4M0.5- 0.56 0.44 4 0.5 215 253 126 760 1005 8.5
H2.25
1204 N.K. Lee, H.K. Lee / Construction and Building Materials 47 (2013) 1201–1209

slag paste was very sensitive to the room temperature during the
manufacturing process.
Fig. 3 shows the setting times with the replacement ratio of the
slag for the fly ash. A higher replacement ratio of the slag for the fly
ash led to a faster setting time, as an increase in the CaO content,
which is the main chemical component of slag, accelerated the
hydration reaction of the mixture, as noted in a previous study
[34].
In Fig. 4, when phosphoric acid (H3PO4) was added to the mix-
ture up to 2.0% of the total binder (fly ash + slag) amount by
weight, the initial time increased slightly and the final time de-
creased. In contrast, when the ratio of phosphoric acid to binder
was 2.25% by weight, the initial and the final setting times in-
creased to 53 min and 90 min, respectively. Phosphoric acid had
a positive effect on retarding the setting of the alkali-activated
Fig. 1. Effect of the NaOH solution molarity on the setting time (solution/binder:
0.38, slag/fly ash: 0.2 by weight). fly ash/slag paste. Gong and Yang [14] found from SEM observa-
tions that phosphate in alkali-activated slag paste promoted a
new needle-like phase formation, which formed a membrane cover

Fig. 2. Effect of the ratio of water glass to NaOH solution by weight on the setting
time (solution/binder: 0.38, slag/fly ash: 0.2 by weight). Fig. 3. Effect of the slag on the setting time (solution/binder: 0.38, water glass/
NaOH solution: 0.5 by weight).

initial and the final times decreased. In the 6 M specimens, when


the ratio of the water glass to NaOH solution by weight was 1.0,
the initial and the final times were longer compared to when it
was 0.5 or 1.5. The 4 M and the 6 M specimens both had the fastest
setting times when the ratio of water glass to NaOH solution by
weight was 1.5. Thus, the setting time decreases with an increase
in the amount of water glass except for the 6M1.0 specimen. The
cause of the exception in the case of the 6M1.0 specimen is not
clear. However, it can be said that the setting properties would
be influenced in accordance with the interaction between the
NaOH solution and the water glass.
Previous studies reported that the alkali activator used in the
geopolymer is greatly affected by the temperature [3]. Accordingly,
in the present study, two different room temperatures, 17 °C and
28 °C, were considered. In Table 5, the initial setting times of the
4M0.5 specimen were less than 20 min at 28 °C and 60–80 min
at 17 °C. From these results, the setting time was found to be faster Fig. 4. Effect of the H3PO4 solution on the setting time (solution/binder: 0.38, slag/
at 28 °C than at 17 °C, indicating that the alkali-activated fly ash/ fly ash: 0.2 by weight).

Table 5
Setting time test results.

S20-6M0.5a S20-6M1.0a S20-6M1.5a S20-4M0.5a S20-4M1.0a S20-4M1.5a S10-4M0.5a S30-4M0.5a S20-8M0.5a


Initial setting (min) 50 75 45 55 40 32 105 31 10
Final setting (min) 114 130 80 160 85 60 180 91 50
S20-6M0.5b S20-6M1.0b S20-4M0.5b S20-4M1.0b S20-4M0.5b-H0.5 S20-4M0.5b-H1.0 S20-4M0.5b-H1.5 S20-4M0.5b-H2.0 S20-4M0.5b-H2.25
Initial setting (min) 18 15 25 15 15 15 14 22 53
Final setting (min) 43 44 60 35 75 70 60 50 90
a
Specimens manufactured at 17 °C.
b
Specimens manufactured at 28 °C.
N.K. Lee, H.K. Lee / Construction and Building Materials 47 (2013) 1201–1209 1205

on the surface of the slag particles. They concluded that the forma- solid in the mixture may degrade the compressive strength of alka-
tion of this new phase may cause the phosphate to retard the alka- li-activated fly ash/slag concrete.
li-activated slag paste. Chang et al. [6] stated that the phosphoric
acid reacted with calcium ions to produce Ca3(PO4)2 which corre- 3.3. Effect of slag on the compressive strength of alkali-activated fly
spondingly reduced the concentration of calcium ions. Therefore, ash/slag concrete
the formation of C–S–H slowed down. Similarly, the calcium from
slag in the present study was consumed to react with phosphoric Fig. 5 shows the compressive strength of the specimens with
acid, which retarded the hydration of alkali-activated fly ash/slag different replacement ratios of the slag for the fly ash. As shown
paste. in Fig. 5a, as the replacement ratio of the slag for the fly ash in-
creased from 10% to 15%, the compressive strength increased from
3.2. Effect of alkali activator on the compressive strength of alkali- 15.5 MPa to 23.0 MPa at 28 days. It was found in previous work
activated fly ash/slag concrete that 100% fly ash-based geopolymers without slag had very low
reactivity at room temperature [26,31]. Accordingly, a high-tem-
Table 6 shows the compressive strength of alkali-activated fly perature curing method was selected to increase the compressive
ash/slag concrete with time. Although larger amounts of NaOH so- strength of geopolymer. In the present study, the compressive
lid and water glass were added to the 6M1.0 specimen compared strength was improved by adding slag, which contains a large
to the other specimens, the compressive strength was lower than amount of CaO, into the mixture. This arose because the presence
that of the 4M0.5 and the 4M1.0 specimens. In addition, the com- of CaO improved the strength of the geopolymer by forming an
pressive strength of the 6M1.0 specimen at 28 days (26.6 MPa) was amorphously structured Ca–Al–Si gel, as also reported in a previ-
lower than that at 14 days (27.4 MPa), while the compressive ous study [34]. Kumar et al. [22] also suggested that the increase
strengths of the 4M0.5 and the 4M1.0 specimens at 28 days in the compressive strength with the addition of slag could result
(29.9 MPa and 30.7 MPa) were equal to or higher than those at from the formation of gel phases (C–S–H and A–S–H) and the com-
14 days (29.9 MPa and 27.4 MPa). According to a previous study, pactness of the microstructure. They concluded that the alkali acti-
geopolymer activated with NaOH solution alone showed a de- vation of slag is predominant in a fly ash/slag mixture at an
crease in its compressive strength compared to geopolymer acti- ambient temperature (27 °C), whereas the interaction between
vated with sodium silicate (Criada et al. [11]). Lloyd et al. [23] fly ash and slag is predominant at 60 °C [22]. Lloyd et al. [23]
also found from microstructural observations that geopolymer showed from a microanalysis that the calcium within the alkali-
activated with a silicate-based solution showed a more homoge- activated fly ash/slag binder is essential to the high early strength
neous microstructure than that activated with a NaOH-based solu- development, although the exact mechanism was not well under-
tion, which may improve the compressive strength. However, this stood in their study. Buchwald et al. [5] showed that the addition
does not correlate well with the results of the present study. The of slag into a slag/metakaolin mixture activated by a sodium
increase in the water glass content resulted not in an increase in hydroxide solution accelerated the condensation reaction of the al-
the compressive strength here. The reason for this is not clear, kali-activated blends compared to both single phases. Thus, the re-
but the compressive strength may decrease due to the high shrink- sults of the present study correlate well with these previous
age amount resulting from an increase in the alkali-activator con- results.
tent. Particularly, an increase in the water glass content from The S30 specimen showed compressive strengths of 14.9 MPa
6M0.5 to 6M1.0 resulted in a significant reduction in the compres- and 26.6 MPa at 3 and 7 days, respectively, as shown in Fig. 5b.
sive strength from 37.4 MPa to 26.6 MPa, while an increase in the These values were higher than that of the S20 and the S25 speci-
water glass content from 4M0.5 to 4M1.0 led to a slight increase in mens. However, the compressive strength of the S30 specimen
the compressive strength from 29.9 MPa to 30.7 MPa. It can be said was lower at 14 days (24.1 MPa) than that of the S30 specimen
from these results that high amounts of water glass and NaOH at 7 days, and that of the S30 specimen at 28 days (28.0 MPa)

Table 6
Summary of test results.

Specimens NaOH solution Ratio of water glass to Ratio of slag to (fly Slump Compressive strength (MPa)
molarity (M) NaOH ash + slag) (mm)
3 Days 7 Days 14 Days 28 Days 56 Days
S20-6M0.5 6 0.5 0.2 200 34.6 37.4
(2.51) (4.04)
S20-6M1.0 6 1.0 0.2 250 27.4 26.6
(0.33) (1.48)
S20-4M0.5 4 0.5 0.2 205 29.9 29.9
(0.15) (4.11)
S20-4M1.0 4 1.0 0.2 200 27.5 30.7
(2.06) (1.03)
S10-4M0.5 4 0.5 0.1 220 2.2 (0.87) 10.4 13.1 15.5 18.4
(0.38) (0.52) (1.26) (1.13)
S15-4M0.5 4 0.5 0.15 240 5.7 (0.62) 12.9 19.9 23.0 25.5
(0.06) (0.11) (0.28) (0.25)
S20-4M0.5- 4 0.5 0.2 200 3.4 (0.14) 10.8 27.6 30.6 29.5
H2.25 (0.57) (2.35) (1.49) (0.20)
S25-4M0.5 4 0.5 0.25 160 17.98 27.15 29.15 39.01 40.96
(0.50) (2.60) (2.06) (1.64) (0.20)
S25-4M0.5- 4 0.5 0.25 160 3.8 (0.14) 18.3 24.9 27.8 27.3
H2.25 (1.95) (3.32) (2.67) (2.72)
S30-4M0.5- 4 0.5 0.3 160 14.9 26.6 24.1 28.0 19.8
H2.25 (2.10) (2.82) (0.57) (0.57) (5.20)

Values in parentheses are standard deviations.
1206 N.K. Lee, H.K. Lee / Construction and Building Materials 47 (2013) 1201–1209

Fig. 6. Surface crack observed in alkali-activated fly ash/slag concrete with a 30%


slag content.

(a) S10 and S15 specimens


amounts of slag in less than 28 days [33]. They stated that a limited
addition of slag in the metakaolin/slag blends improved the
mechanical properties.
The compressive strengths of the S20 and the S25 specimens at
56 days were slightly lower than those of the S20 and the S25 spec-
imens at 28 days, respectively. This may be related to the surface
dry condition of the alkali-activated fly ash/slag concrete cured
in a laboratory (air curing), as previously reported by Collins and
Sanjayan [8].

3.4. Effect of H3PO4 addition on the compressive strength development

Table 6 shows the compressive strength development of the


S25 specimens with/without H3PO4 (phosphoric acid). The addi-
tion of H3PO4 was an effective means of retarding the setting of
the alkali-activated fly ash/slag concrete, though the compressive
(b) S20, S25, and S30 specimens with added H3PO4
strength of the specimen with H3PO4 was slightly lower compared
to the specimen without H3PO4.
Chang et al. [6] found that the addition of phosphoric acid to al-
kali-activated slag paste caused an increase in the setting time, a
reduction of the compressive strength at an early age and an in-
crease in the amount of drying shrinkage. According to Chang
et al. [6] and Gong and Yang [14], a significant amount of calcium
dissolved from slag reacted with PO4 in phosphoric acid, resulting
in the formation of Ca3(PO4)2, while the formation of C–S–H was
limited by the formation of Ca3(PO4)2. Similarly, the specimens
with H3PO4 showed reduced compressive strength in the present
study.

3.5. Comparison of compressive strength of alkali-activated fly ash/


slag concrete with that of ordinary concrete predicted according to the
ACI code and Eurocode 2
(c) S40 specimens with/without added H3PO4
Fig. 7 shows the normalized compressive strength, which is de-
Fig. 5. Compressive strength development of alkali activated fly ash/slag concrete: fined as the compressive strength measured at 3, 7, 14, 28 and
(a) S10 and S15 specimens; (b) S20, S25, and S30 specimens with added H3PO4; and
(c) S25 specimens with/without added H3PO4.
56 days divided by the compressive strength measured at 28 days.
These test results are compared with the predicted compressive
was lower than that of the S20 specimen (30.6 MPa). These results strength of ordinary concrete according to the ACI 209R code [1]
indicate that an increase in the amount of slag does not always in- and Eurocode 2 [13] to show the different characteristics of the al-
crease the compressive strength. The S30 specimen in particular kali-activated fly ash/slag concrete compared to OPC concrete. It
showed unstable compressive strength development, as depicted should be noted that the prediction equations from the two codes
in Fig. 5b. Only the S30 specimen displayed cracking, while the are used for OPC concrete.
S10, S15, S20, and S25 specimens did not show surface cracks. This The ACI 209R code provides the following equation for the com-
may be due to crack evolution caused by the shrinkage of the alka- pressive strength development of concrete with time [1]:
li-activated fly ash/slag concrete, as observed in Fig. 6. In previous t
studies of alkali-activated slag concrete, it was reported that drying fc0 ðtÞ ¼ ðfc0 Þ28 ð1Þ
a þ bt
shrinkage was caused by the formation of an unstable gel phase
[21] and a very dry condition of the surface concrete [10]. In addi- Here ðfc0 Þ28 is the 28-day compressive strength and t in days is the
tion, it was found from physical observations that cracks developed age of concrete. The constants a and b are functions of the type of
within hardened metakaolin/slag blends containing substantial cement used and the type of curing employed, respectively [1].
N.K. Lee, H.K. Lee / Construction and Building Materials 47 (2013) 1201–1209 1207

than the measured compressive strengths of the specimens with-


out H3PO4 at 3, 7, and 14 days; the prediction by the ACI code
was in good agreement with the measured compressive strengths
of the specimens without H3PO4 at 3 and 7 days. The measured
compressive strengths of the specimens with H3PO4 were slightly
lower than the predictions by Eurocode 2 at 3 and 7 days and were
similar to the prediction by the ACI code, apart from the value at
3 days.
It was previously reported that the fly-ash-based geopolymer
concrete cured above 60 °C for a few hours or 1 day attains high
early compressive strength due to the fast hydration reaction
[15]. In addition, other studies have shown that alkali-activated
slag concrete obtains high early compressive strength and a fast
setting time due to the existence of a large amount of CaO in the
(a) S10, S15 and S20-H2.25 specimens mixture [8,18]. However, high early compressive strength was
not obtained in the present study due to the use of low tempera-
ture curing (20 °C) and the small amount of slag (10–30% of the to-
tal binder by weight) in the mixture, although the compressive
strength continued to increase with time, as shown in Fig. 5a and
b. The results obtained in the present study can be later utilized
to predict the shrinkage and creep properties of alkali-activated
fly ash/slag concrete.

3.6. Elastic modulus

Fig. 8 shows the measured elastic modulus of alkali-activated


fly ash/slag concrete. The predicted elastic modulus of OPC con-
crete by the CEB-FIP model code [7] and ACI 318 code (2008) are
also shown in the figure. The modulus of elasticity of normal-
weight concrete can be estimated from the CEB-FIP model code
(b) S25 and S25-H2.25 specimens [7] as
1=3
Fig. 7. Comparison between the measured compressive strength development of Ec ¼ 0:85  2:15  104 ðfc =10Þ Þ ð3Þ
alkali activated fly ash/slag concrete with added H3PO4 and the predicted
compressive strength of ordinary concrete according to the ACI 209R code and where Ec is the modulus of elasticity of concrete at 28 days
Eurocode 2: (a) S10, S15, S20-H2.25 specimens; and (b) S25 and S25-H2.25 (MPa) and fc is the average compressive strength at 28 days
specimens.
[7].According to ACI Building Code 318-08, the modulus of elastic-
ity of concrete (unit weight between 1500 and 2500 kg/m3) can
The constants a and b are set to 4 and 0.85, corresponding to ce- also be determined as follows (ACI 318, 2008):
ment type I and moist curing, respectively.
Ec ¼ wc1:5  0:043fc01=2 MPa ð4Þ
In Eurocode 2, the equation for estimating the compressive
strength development of concrete with time is given as [13]: In this equation, Ec is the static modulus of elasticity (MPa) and wc is
(   0:5 ) the unit weight (kg/m3) (ACI 318, 2008). As shown in Fig. 8, the
0 28
fc ðtÞ ¼ Exp s 1  ðfc0 Þ28 ð2Þ elastic modulus of most specimens were 20–40% lower than the
t predictions by the two codes. The predicted elastic modulus of nor-
mal-weight concrete by the ACI and CEB-FIP model codes were
where ðfc0 Þ28 is the compressive strength at 28 days. The constant s is
a coefficient that depends on the cement type [13]. Here, s is 0.25,
corresponding to Class N (normal early strength) for OPC concrete.
The predicted compressive strengths according to Eqs. (1) and (2)
are shown in Fig. 7.
Fig. 7a shows a comparison of the measured compressive
strength of the S10, the S15, and the S20 specimens with the pre-
dictions in accordance with ACI code and Eurocode 2. The mea-
sured three-day compressive strength of the S10 specimen was
slightly lower than the predicted values, while the measured com-
pressive strengths at 7 and 14 days were similar to the predictions
and the measured compressive strengths after 28 days were
slightly larger than the predictions. The measured compressive
strengths of the S15 and the S20 specimens were lower than the
predicted values at 3, 7, and 14 days, while the measured compres-
sive strength of the S15 specimen was slightly higher than the pre-
dictions at 56 days.
Fig. 8. Comparison between the measured elastic modulus of alkali activated fly
The measured compressive strengths of the specimens with ash/slag concrete and the predicted elastic modulus of ordinary concrete according
H3PO4 and without H3PO4 are compared with the predicted values to the CEB-FIP model and ACI codes ([1] Sofi et al. [29], and [2] Fernández-Jiménez
in Fig. 7b. The predictions by Eurocode 2 are shown to be greater et al. [19]).
1208 N.K. Lee, H.K. Lee / Construction and Building Materials 47 (2013) 1201–1209

shown to be greater than that of alkali-activated fly ash/slag


concrete.
Based on these test results, a prediction equation for the elastic
modulus of alkali-activated fly ash/slag concrete is proposed here,
as shown below (see Fig. 8).
p
3
ffiffiffiffi
Ec ¼ 5300  fc MPa ð5Þ
In previous works, the elastic modulus of alkali-activated fly ash
concrete was compared with prediction models for high-strength
types of concrete [29]. The results showed lower values than the
predictions by the two codes, as shown in Fig. 8. However, it is
not fair to compare directly the elastic modulus present in this
study to the test results by Sofi et al. [29], as the elastic modulus
in their study was obtained with a high compressive strength of
more than 50 MPa.
The elastic modulus of normal-strength alkali-activated fly ash Fig. 9. Incremental pore volume versus pore size distribution of alkali activated fly
concrete reported by Fernández-Jiménez et al. [19] is also shown in ash/slag mortar.
Fig. 8 for comparison. Their results were also lower than the pre-
dictions by the two codes, showing a similar trend with the elastic the elastic modulus and the long-term compressive strength, as
modulus of the alkali-activated fly ash/slag concrete measured in depicted in Figs. 5b, 6 and 8.
the present study. However, it cannot be concluded that the alkali
activation of the fly ash/slag mixture reduced the elastic modulus 3.8. Splitting tensile strength
compared to OPC, as the elastic modulus of concrete mainly de-
pends on the type and content of the coarse aggregate. The measured splitting tensile strengths were compared with
the predicted splitting tensile strength of the OPC concrete accord-
ing to the ACI 318-08 code and Eurocode 2. The relationship be-
3.7. Porosity
tween the mean tensile strength fctm and the compressive
strength fc is expressed as follows [13]:
The apparent densities of the mortar specimens used for the
MIP test were 2.46 (OPC), 2.42 (S15), 2.40 (S20) and 2.38 (S25). fctm ¼ 0:30fc2=3 ; f c 6 50 MPa ð6Þ
The maximum applied intrusion pressure during the test was sus-
tained at about 46,800 psi. The tensile strength fctm can be calculated from the measured split-
Table 7 summarizes the porosities measured by the MIP meth- ting tensile strength fct,sp as follows [13]:
ods. The porosity of the S30 specimen was not measured since sur- fctm ¼ 0:9f ct;sp MPa ð7Þ
face cracking on the specimen may have changed its porosity. The
porosities of the S15, S20, and S25 specimens measured by the MIP The average splitting tensile strength fct can also be calculated as
method were 19.80%, 20.44% and 20.81%, respectively. These val- follows (ACI 318, 2008):
ues were 0.6–1.6% higher than that (19.26%) of the OPC specimen.
qffiffiffiffi
fct ¼ 0:56 fc0 MPa ð8Þ
As the amount of slag in the mixture increased, the porosities also
increased slightly. However, it is hard to state that the influence of It is observed in Fig. 10 that most measured splitting tensile
the slag content on the porosity is significant, since the differences strengths are lower than the predictions by the two codes. Based
in porosity may be in the range of the error of the MIP test and only on these results, the following prediction equation for the splitting
one specimen per each MIP test was used. tensile strength of alkali-activated fly ash/slag concrete is
Fig. 9 shows the relationship between the incremental pore vol- proposed:
ume and pore size distribution of the S15, S20, S25, and OPC mor- qffiffiffiffi
tars. The mean pore sizes of the S20 and the S25 specimens were fct ¼ 0:45  fc0 MPa ð9Þ
smaller than that of the OPC specimens, while the mean pore size
of the S15 specimen was larger. The volumes of the smaller pores
(<20 nm) of the S20 and S25 specimens were larger than those of
the OPC and S15 specimens. These results show that as the amount
of slag increased, the pore volume in the region of the mesopores
also increased. Collins and Sanjayan [9] reported that a higher mes-
opore content in water glass-activated slag concrete led to higher
capillary stress and greater shrinkage during drying. Likewise, high
shrinkage due to the increased number of small pores caused mi-
cro-cracks in the alkali-activated fly ash/slag concrete, decreasing

Table 7
Measured density and porosity of alkali-activated fly ash/slag mortar and concrete.

OPC S10 S15 S20 S25


Apparent density of concrete (kg/ 2530 2510 2460 2520 2570
m3)
Porosity (MIP method, %) 19.26 – 19.80 20.44 20.81
Compressive strength (MPa) 26.22 15.5 23.0 30.6 27.8
w/c 0.44 Fig. 10. Comparison between the measured splitting tensile strength of alkali
Binder: sand:gravel (by vol.) 1: 1.9: 2.5 activated fly ash/slag concrete and the predicted splitting tensile strength of
ordinary concrete according to the Eurocode 2 and the ACI code.
N.K. Lee, H.K. Lee / Construction and Building Materials 47 (2013) 1201–1209 1209

The results of a previous study [29] showed that the splitting tensile [3] Bakharev T. Geopolymeric materials prepared using class F fly ash and elevated
temperature curing. Cem Concr Res 2005;35(6):1224–32.
strength of alkali-activated fly ash/slag concrete was 0.48 times the
[4] Bernal SA, Provis JL, Rose V, de Gutierrez RM. Evolution of binder structure in
square root
pffiffiffiffi of its compressive strength on average. The prediction
pffiffiffiffi sodium silicate-activated slag–metakaolin blends. Cement Concr Compos
(0:48 fc0 ) proposed by Sofi et al. [29] is similar to that (0:45 fc0 ) 2011;33(1):46–54.
proposed in the present study, as shown in Fig. 10. It should be [5] Buchwald A, Tatarin R, Stephan D. Reaction progress of alkaline-activated
metakaolin-ground granulated blast furnace slag blends. J Mater Sci
noted however that further research on the mechanical properties 2009;44(20):5609–17.
of alkali-activated fly ash/slag concrete is required to propose a gen- [6] Chang JJ, Yeih W, Hung CC. Effects of gypsum and phosphoric acid on the
eral relationship between the compressive strength and the split- properties of sodium silicate-based alkali-activated slag pastes. Cement Concr
Compos 2005;27(1):85–91.
ting tensile strength, as only a small number of specimens were [7] CEB-FIP model code. Comite Euro-International Du Beton, 1990 p. 39.
used in the present study. [8] Collins F, Sanjayan JG. Early age strength and workability of slag pastes
activated by NaOH and Na2CO3. Cem Concr Res 1998;28(5):655–64.
[9] Collins F, Sanjayan JG. Effect of pore size distribution on drying shrinkage of
4. Concluding remarks alkali-activated slag concrete. Cem Concr Res 2000;30(9):1401–6.
[10] Collins F, Sanjayan JG. Microcracking and strength development of alkali
This paper has presented the results of an experimental study activated slag concrete. Cement Concr Compos 2001;23(4–5):345–52.
[11] Criada M, Fernández-Jiménez A, de la Torre AG, Aranda MAG, Palomo A. An
conducted to evaluate the setting and mechanical properties of al- XRD study of the effect of the SiO2/Na2O ratio on the alkali activation of fly
kali-activated fly ash/slag concrete manufactured at room temper- ash. Cem. Concr. Res 2007;37(5):671–9.
ature. The following conclusions can be drawn from the results [12] Davidovits J, Comrie DC, Paterson JH, Ritcey DJ. Geopolymeric concretes for
environmental protection. ACI Concr Int 1990;12(7):30–40.
presented here. [13] European Standard. Eurocode 2: design of concrete structures Part 1: general
rules and rules for buildings, Ref. No. prEN 1992–1-1, July 2002, p. 27.
(1) The setting times of alkali-activated fly ash/slag paste [14] Gong C, Yang N. Effect of phosphate on the hydration of alkali activated red
mud–slag cementitious material. Cem Concr Res 2000;30(7):1013–6.
decreased as the amounts of slag and water glass and the [15] Hardjito D. Studies on fly ash-based geopolymer concrete. PhD thesis of curtin
molarity of the NaOH solution increased. university of technology; 2005.
(2) When the molarity of the NaOH solution was 4 M and the [16] Hardjito D, Cheak CC, Ing CHL. Strength and setting times of low calcium fly
ash based geopolymer mortar. Mod Appl Sci 2008;2(4):3–11.
ratio of water glass to NaOH solution by weight was 0.5, [17] Izquierdoa M, Querol X, Phillipart C, Antenucci D, Towler M. The role of open
the alkali-activated fly ash/slag paste had an initial time of and closed curing conditions on the leaching properties of fly ash–slag-based
55 min and a final time of 160 min at a room temperature geopolymers. J Hazard Mater 2010;176(1–3):623–8.
[18] Fernández-Jiménez A, Puertas F. Setting of alkali-activated slag cement.
of 17 °C.
Influence of activator nature. Adv Cem Res 2001;13(3):115–21.
(3) The compressive strengths of the alkali-activated fly ash/slag [19] Fernández-Jiménez AM, Palomo A, Lopez-Hombrados C. Engineering
concrete at 28 days increased with the amount of slag, properties of alkali-activated fly ash concrete. ACI Mater J
except when the amounts of slag were 25% and 30% of the 2006;103(2):106–12.
[20] Kong DLY, Sanjayan JG. Damage behavior of geopolymer composites exposed
total binder weight. The compressive strength of the S30 to elevated temperatures. Cement Concr Compos 2008;30(10):986–91.
specimen at 56 days decreased due to crack evolution [21] Kutti T, Malinowski R. Influence of the curing conditions on the flexural
caused by the shrinkage of the alkali-activated fly ash/slag strength of alkali activated blast furnace slag mortar. Nordic Concr Res
1982:17.1–17.10.
paste. Accordingly, the appropriate replacement ratio of [22] Kumar S, Kumar R, Mehrotra SP. Influence of granulated blast furnace slag on
the slag for the fly ash by weight in the alkali-activated fly the reaction, structure and properties of fly ash based geopolymer. J Mater Sci
ash/slag mixture is 15–20% considering the setting time, 2010;45(3):607–15.
[23] Lloyd RR, Provis JL, van Deventer JSJ. Microscopy and microanalysis of
workability and development of compressive strength. inorganic polymer cements. 1: remnant fly ash particles. J Mat Sci
(4) Current predictions by the ACI code and Eurocode 2 were 2009;44(2):608–19.
shown to be greater than the elastic modulus and splitting [24] Malhotra VM. Introduction: sustainable development and concrete
technology. ACI Concr Int 2002;24(7):22.
tensile strength of alkali-activated fly ash/slag concrete. [25] Mehta PK. Reducing the environmental impact of concrete. ACI Concr Int
The mean pore sizes of the alkali-activated fly ash/slag mor- 2001;23(10):61–6.
tars measured by the MIP method were smaller than that of [26] Palomo A, Grutzeck MW, Blanco MT. Alkali-activated fly ashes – a cement for
the future. Cem Concr Res 1999;29(8):1323–9.
OPC specimens. This indicates that high shrinkage due to the
[27] Perera DS, Uchida O, Vance ER, Finnie KS. Influence of curing schedule on the
increased number of small pores caused micro-cracks in the integrity of geopolymers. J Mat Sci 2007;42(9):3099–106.
alkali-activated fly ash/slag concrete, resulting in decreases [28] Rowles M, ÓConnor B. Chemical optimization of the compressive strength of
in the elastic modulus and the long-term compressive aluminosilicate geopolymers synthesized by sodium silicate activation of
metakaolinite. J Mater Chem 2003;13(5):1161–5.
strength. [29] Sofi M, van Deventer JSJ, Mendis PA, Lukey GC. Engineering properties of
inorganic polymer concretes (IPCs). Cem Concr Res 2007;37(2):251–7.
[30] Temuujin J, van Riessen A, Williams R. Influence of calcium compounds on the
mechanical properties of fly ash geopolymer pastes. J Hazard Mater
Acknowledgements 2009;167(1–3):82–8.
[31] van Jaarsveld JGS, van Deventer JSJ, Lukey GC. The effect of composition and
This research was supported by a grant (Code 11-Technology temperature on the properties of fly ash- and kaolinite-based geopolymers.
Chem Eng J 2002;89(1–3):63–73.
Innovation-F04) from Construction Technology Innovation Pro- [32] Xu H, van Deventer JSJ. The geopolymerisation of aluminosilicate minerals. Int
gram (CTIP) funded by Ministry of Land, Transportation and Mari- J Miner Process 2000;59(3):247–66.
time Affairs (MLTM) of Korean government. The authors would like [33] Yip CK, Lukey GC, van Deventer JSJ. Microstructure and properties of slag/
metakaolinite geopolymeric materials. In: 28th Conference on our world in
to thank Mr. Jung-Gook Jang and Mr. Sam Na at KAIST for their help concrete & structures. Singapore; 28–29 August 2003.
on experiments. [34] Yip CK. The role of calcium in geopolymerisation. PhD thesis of University of
Melbourne, Austrailia; 2004.
[35] Yip CK, Lukey GC, van Deventer JSJ. The coexistence of geopolymeric gel and
References
calcium silicate hydrate at the early stage of alkaline activation. Cem Concr Res
2005;35(9):1688–97.
[1] ACI 209R-92. Prediction of creep, shrinkage, and temperature effects in [36] Yip CK, Lukey GC, Provis JL, van Deventer JSJ. Effect of calcium silicate sources
concrete structures, In: ACI Manual of Concrete. Practice Part 1: materials on geopolymerisation. Cem Concr Res 2008;38(4):554–64.
and general properties of concrete; 1994.
[2] Bakharev T. Durability of geopolymer materials in sodium and magnesium
sulfate solutions. Cem Concr Res 2005;35(6):1233–46.

You might also like