879313

8/13/2019 879313

1/5

Alkali-aggregate reactivity of typical siliceious glass and carbonate

rocks in alkali-activated fly ash based geopolymersDuyou Lu*, Yongdao Liu, Yanzeng Zheng, Zhongzi Xu, Xiaodong Shen

Nanjing university of Technology, 5 Xin Mofan Road, Nanjing, 210009, China

ABSTRACT

For exploring the behaviour of alkali-aggregate reactivity (AAR) in alkali-activated geopolymeric materials and

assessing the procedures for testing AAR in geopolymers, the expansion behaviour of fly ash based geopolymer mortars

with pure silica glass and typical carbonate rocks were studied respectively by curing at various conditions, i.e. 23℃ and

38℃with relative humidity over 95%, immersed in 1M NaOH solution at 80℃. Results show that, at various curing

conditions, neither harmful ASR nor harmful ACR was observed in geopolymers with the criteria specified for OPCsystem. However, with the change of curing conditions, the geopolymer binder and reactive aggregates may experience

different reaction processes leading to quite different dimensional changes, especially with additional alkalis and

elevated temperatures. It suggests that high temperature with additional alkali for accelerating AAR in traditional OPCsystem may not appropriate for assessing the alkali-aggregate reactivity behaviour in geopolymers designed for normal

conditions. On the other hand, it is hopeful to control the dimensional change of geopolymer mortar or concrete byselecting the type of aggregates and the appropriate curing conditions, thus changing the harmful AAR in OPC into

beneficial AAR in geopolymers and other alkali-activated cementitious systems.

Keywords: geopolymer; armorphous silica; alkali-silica reaction; alkali-carbonate reaction; testing method

1. INTRODUCTIONGeopolymers refer to a group of cementitious materials manufactured by reacting alumino-silicate materials (e.g. fly ash

and metakaolin) with an alkali solution [1]. Comparing with traditional Ordinary Portland Cement (OPC), which is an

extremely resource and energy intensive product, those new cementitoius mateials have abundant raw materials

resources, low energy consumption in manufacture and superior mechanical and physical properties. These materialshave gained more and more interest as alternative candidates for cementitious materials in civil engineering and other

applications in waste management, art decoration, etc .

Alkali has duplex effects on the synthesis of geopolymers and their durability. On one hand, alkali cations are considered

as structure-forming element balancing the negative framework charge of tetrahedral aluminum, while the OH-ions arecatalyst in the dissolution of source alunminosilicates. Therefore, sufficient alkali is essential for both the

geopolymerization process and the structure of the final product. On the other hand, the existence of certain amount of

free alkali may have harmful effect on its long term stability. Normally, the alkali content in geopolymers is about

4~6%(Na2Oe.), which is several times higher than that in OPC. Both chemistry theory and experimental results indicatethat the free alkalis do exist in geopolymers [2, 3]. The release of the alkali could result in efflorescence of the

geopolymer binder itself and excess free alkali may also cause possible harmful aggregate reactions, especially when

alkali-susceptible aggregates are available. Therefore, a study on the behavior and mechanism of the interaction of

geopolymeric binder phases with aggregates, especially the alkali-susceptible aggregates in OPC binders, is of paramount importance in the development of geopolymer concretes.

Previous findings of alkali-aggregate reaction (AAR) in alkali-activated slag concretes have shown that the alkali-aggregate reaction behaviour could be significantly different depend on the composition of raw materials, the types of

reactive aggregates and curing conditions, etc [4-6]. Alkali-silica reaction (ASR) induced much lower expansion whilealkali-carbonate reaction (ACR) induced significantly higher expansion than that in traditional OPC system. AAR could

be either destructive or constructive in alkali-activated cementitious materials. For geopolymers, however, the raw

materials, polymerization mechanism and products of geopolymerization are quite different from those in alkali-

activated slag cements and OPC. No systematic research has yet been carried out on the alkali-aggregate reaction

behaviour and its mechanisms in geopolymers. Limited research focused only on the comparative study on AAR induced

Fourth International Conference on Smart Materials and Nanotechnology in Engineering,edited by Jayantha A. Epaarachchi, Alan Kin-tak Lau, Jinsong Leng, Proc. of SPIE Vol. 8793,

879313 · © 2013 SPIE · CCC code: 0277-786X/13/$18 · doi: 10.1117/12.2032185

Proc. of SPIE Vol. 8793 879313-1

wnloaded From: http://spiedigitallibrary.org/ on 12/22/2013 Terms of Use: http://spiedl.org/terms

8/13/2019 879313

2/5

expansion between geopolymers and OPC concretes [7,8]. Most of these findings were based on results in the

accelerated mortar bar method (ASTM C1260), which has been the mostly widely used test with additional alkali andelevated temperature to accelerate AAR in traditional OPC system. In order to explore the mechanisms of AAR in

geoplolymers for futher performance improvement on geopolymers and establish suitable test procedure for assessing

AAR in geopolymeric materials, the reaction behaviour of amorphous silica, a highly ASR reactive aggregate, anddolomitic limestone from Kingston, Canada (CK), which is a typical ACR aggregate, were studied by length

measurement of geopolymer mortars cured at different curing conditions specified in the traditional laboratory tests forAAR in OPC systems.

2. MATERIALS AND EXPERIMENTAL2.1 MaterialsA class F low calcium fly ash (FA) and a calcined superfine metakaolin (MK) were used as primary aluminosilicatesources. Sodium water glass mixed with reagent sodium hydroxide was used as activated alkali solution. The modulus

and solids content of the modulated water glass are 1.4 and 37%, respectively. A highly ASR reactive aggregate, high-

purity recycled amorphous silica glass (QG), a typical highly ACR reactive rock, argillaceous dolomitic limestone (CK)from Kingston, Canada and a non reactive pure dolomite from China (YT) were used as aggregates. The aggregates were

crushed and sieved into 2.5~5.0 mm in size. The chemical compositions of FA, MK and QG are listed in table 1. The

chemical and mineral compositions of YT and CK are listed in table 2.

TABLE 1: Chemical Composition of fly ash, metakaolin and silica glass wt/%

SiO2 Al2O3 Na2O K 2O CaO Fe2O3 MgO SO3 TiO2 Total

FA 50.80 31.34 -- 1.50 3.90 5.93 1.27 1.45 -- 96.28

MK 55.03 43.33 -- 0.46 0.82 -- -- 0.36 100

QG 99.70 0.07 0.07 -- 0.05 0.01 -- 0.02 -- 99.92

TABLE 2: Chemical and Mineral Compositions of carbonate rocks wt%

Chemical Composition Mineral Composition

Al2O3 SiO2 Fe2O3 CaO MgO LOI Dolomite Calcite

YT 0.76 0.75 0.20 29.60 21.76 46.55 97.34 --

CK 3.14 10.16 0.89 40.31 5.51 37.88 25.19 58.30

2.2 ExperimentalGeopolymer binders were prepared by first dry-mixing FA and MK powders, then combining with alkali solutions with a

mechanical mixer. The mass ratio of FA to MK was 65/35. Sodium silicate solution was prepared at molar SiO2/Na2O

ratio of 1.4, the water to solids mass ratio of geopolymer binders was 0.31. For geoplolymer mortars with aggregates, the

geopolymer binder was mixed first, the aggregate, either QG, YT or CK, was then mixed into the binder by mass ratio of

Aggregate/(FA+ MK) =1:1.

For each of the three kinds of curing conditions, 23 ºC and over 95% R.H., 38 ºC and over 95% R.H., immersed in a 1M NaOH solution at 80 ºC, four bars, 20mm×20mm×80mm in size, were cast from the geopolymer binder and the

geopolymer mortars, respectively. All casted molds were covered with a plastic film and cured at 23ºC for 24 h, then

demolded, after the initial length measured then moved into different environments for furthering testing. For specimenscured without additional alkali, i.e., 23ºC and over 95% R.H., 38ºC and over 95% R.H., specimens were sealed with

plastic film aging to test at specific periods. For specimens immersed in a 1M NaOH solution at 80ºC, the bars were

moved into the solution at room temperature after taking the initial readings, then warmed up to and maintained at 80ºC.

The length changes of specimens were measured regularly with a comparator. For the specimen cured at 38ºC, over

95%R.H. and in 80ºC 1M NaOH solution, the bars were cooling down at 23ºC for 6 h before each measurement.



8/13/2019 879313

3/5

3. RESULTS AND DISCUSSION3.1 AAR behaviour in geopolymers cured at mild temperature without additional alkalies

Fig. 1 shows the linear dimensional change curves of geopolymer binder and mortars cured with a relative humidity over

95% at 23ºC and 38℃, respectively.

When cured at 23ºC (Fig.1 a), geopolymer binder without aggregates shrunk in the whole testing period. The process ofshrinkage can be divided into two stages, a rapid shrinkage before 7 days, finishing almost half of the total shrinkage

attained in the testing period, and a stage continuing to increase at a steady rate throughout the testing period. Comparedwith the binder, the dimensional change behaviour of geopolymer mortars was quite different. For mortar with QG, a

slight expansion was observed at the first day, which was contrary to the rapid shrinkage of the binder at early ages. The

expansion then decreased at first few days, dominated by the shrinkage of the geopolymer binder with similar trend. But

the shrinkage of geopolymer mortars with QG was smaller than that of the geopolymer binder at the same age. It

indicates that QG was involved in a very rapid expansive reaction other than geopolymerization at early age, leading to aquick and slight expansion of the mortar. However, the expansion did not last with time suggesting that the glass-induced

expansive reaction and its effect was either suppressed by the following geopolymerization process or the glass

continued to react but formed the non-expansive product with the inter-reaction with geopolymerization.

For mortars with carbonates rocks, CK and YT, however, both mortars gave a similar but smaller shrinkage than the

control and levelled off at an earlier time. There was no obvious differences in shrinkage behaviour between the mortars

with CK and YT.

0 30 60 90 120 150 180

-0.08

-0.06

-0.04

-0.02

0.00

0.02

E x p a n s i o n / %

Age/d

Control

YT

CK

QG

0 30 60 90 120 150 180-0.06

-0.04

-0.02

0.00

0.02

0.04


Age/d

Control

YT

CK

QG

(a) cured at 23℃ with a relative humidity over 95% (b) cured at 38℃ with a relative humidity over 95%

Figure 1 Linear dimensional change curves of geopolymer binder and mortars

When the specimen were cured at 38ºC (Fig.1 b), the geopolymer binder also shrunk throughout the testing periods, but

the shrinkage behaviour was quite different in two aspects compared with Fig.1a. Besides two stages of the developmentof shrinkage, rapid development and slow down stages, the binder cured at 38ºC seemed to reach dimensional

stabilization after 60 days, with a smaller attained shrinkage level. It indicates that increasing the curing temperature

would promote the geopolymerization reaction and shorten the time required for the binder to reach dimensionalstabilization. It was in accordance with literatural results that both the dissolution rate of Si and Al from raw

alunminsilicates and the geopolymerization rate of gel products increased with increasing temperature [9, 10].

Mortar with QG developed a slight expansion at the first day, then shrunk with a similar trend as in the geopolymer

binder. Compared with those cured at 23ºC. it developed a much smaller expansion at the early age. It was probably dueto that, the increase of curing temperature, accelerated the geopolymerization process, while weakening the effect of

glass reaction and its interaction with geopolymerization.

Compared with Fig.1a, mortars with YT and CK in Fig 1b developed quite different curves. Mortar with YT gave a

similar shrinkage at early period and a smaller shrinkage at later period. The trend was in accordance well with the

binder, suggesting YT mainly acted as inert filler, without much chemical effect. Mortar with CK, however, showed a

very different trend from both the control and the mortar with YT. It shrunk very slightly at the first 5 days, and thenkept stable till 56 days. After a slight expansion, it reached a stable dimension after 120 days. The major shift from



8/13/2019 879313

4/5

shrinkage to slight expansion suggests that some expansive chemical reactions occurred to some extent in the system. It

seems different from that in OPC system, in which mortar or concrete with CK would develop harmful expansion,leading to the cracking of specimens.

3.2 AAR behaviour in geopolymers at 80 in 1M NaOH solutionsFigs. 2 and 3 show the linear dimensional change curves of geopolymer binder and mortars with ASR and ACR

aggregates immersed in 1M NaOH solution at 80ºC, respectively.

0 5 10 15 20 25 30

-0.06

-0.05

-0.04

-0.03

-0.02

-0.01

0.00


Age/d

Control

QG

0 5 10 15 20 25 30-0.08

-0.06

-0.04

-0.02

0.00

0.02

0.04


Age/d

Control YT CK

Figue 2 Linear dimensional change curves of geopolymer binder

and mortar with QG immersed in 1M NaOH solution at 80℃

Figure 3 Linear dimensional change curves of geopolymer binder

and mortars with YT and CK in 1M NaOH solution at 80℃

The process of shrinkage development in geopolymer binder can be divided into four stages, rapid shrinkage, continue toshrinkage slowly, stabilization and shrinkage reduction, which were obviously different from those in binders cured at

relative low temperatures and without additional available alkalis (Fig. 1). The binder shrunk rapidly at the first day,

finishing almost of the total shrinkage attained in the testing period. The shrinkage continued to increase at a steady rate

before 7 days, and then the binder arrived at dimensional stabilization before 21 days. With the prolonging of curing

time, the binder started to expand. It indicates that high alkali solution combined with high temperature accelerates thegeopolymerization reaction and promotes the formation of structural stable geopolymer binder. However, with the

extension of curing period, the composition and structure of binder could be modulated by external alkali intervention

and further geopolymerization. Some transformations with dimensional expansion took place in the binder. Compared

with Fig.1, the behaviour of mortar with QG was also different significantly. Contrary to a slight expansion of themortars with QG at early ages, the mortars cured at higher temperature and with additional alkali was dominated by the

shrinkage of the binder in the whole testing period, but with a smaller shrinkage than that in the binder at the same age.The expansive effect at later age was also weaker than that in the binder. Since a lot of gels floating in the curing alkali

solution and the obvious gap between glass particles and binder were observed, glass particles were involved in reaction,

but probably formed low viscosity non-expansive products. The gap between glass particles and binder were acted as a

buffer for the expansion due to the phase transformation in the binder, resulting in a lower expansion than that in binder

at later age.

For mortars with carbonate rocks shown in Fig. 3, the control and the mortar with YT shrunk in the whole testing periodand generally gave similar trends as those in Figs.1 and 2. Mortar with YT developed a smaller shrinkage than that in the

control. Mortar with CK gave a very different behaviour from the control and the mortar with YT, as well as the mortar

with CK cured at 23℃ and 38℃ without additional alkalis. Contrary to the rapid shrinkage of the control and mortar

with YT at early period, mortar with CK expanded slightly before 7 days, then shrunk slightly dominated by theshrinkage of the binder matrix. The binder in Figs.2 and 3 developed different pattern suggesting that the accompanying

ions in the soaking solution may have some impact on the reactions in geoploymer binder and its dimensional stability.

4. CONCLUSIONSBoth QG and CK, the typical highly reactive ASR and ACR aggregates in OPC system, did not show harmful expansionin geopolymers synthesized in present study and cured at various curing conditions specified in different testing



8/13/2019 879313

5/5

procedures for OPC. It suggests that AAR seems not a problem in geoploymers, although it is an issue always being

concerned in alkali-activated cementitious systems.

At various curing conditions tested in present paper, both the geopolymer binder and mortars with QG and CK

experienced different reaction processes, thus leading to quite different dimensional changes, especially with additional

alkalis and elevated temperatures. It suggests, on one hand, that high temperature with additional alkali for accelerating

AAR in traditional OPC system may not appropriate for assessing the alkali-aggregate reactivity behaviour in

geopolymer designed for normal conditions. On the other hand, however, it is hopeful to control the dimensional changeof geopolymer mortar or concrete by selecting the type of aggregates and the appropriate curing conditions, thus

changing the harmful AAR in OPC into beneficial AAR in geopolymers and other alkali-activated cementitious systems.

5. ACKNOWLEDGEMENTSThe supports received from National Natural Science Foundation of China (No.51072080)，the Key Project of Chinese

Ministry of Education (No. 210079) and the Priority Academic Program Development of Jiangsu Higher Education

Institutions (PAPD) are gratefully acknowledged.

REFERENCES

[1] Davidovits, J. “Geopolymers and geopolymeric new materials”. Thermal Analysis (35),429 (1989).[2] Duxon, P, Lukey, GC, Separovic, F, Van Deventer, JSJ., “The effect of alkali cations on aluminum

incorporation in geopolymeric gels”. Industry engineering Chemistry Research (44): 832-839 (2005).[3] Duxon, P, Provis, JL, Lukey, GC, Van Deventer, JSJ, “39K NMR of free potassium in geopolymers”. Industry

engineering Chemistry Research (45): 9208-9210 (2006).

[4] Gifford, PM, Gillott, JE, “Alkali-silica reaction (ASR) and alkali-carbonate reaction (ACR) in activated blastfurnace slag cement (ABFSC) concrete”. Cement and Concrete Research (26): 21-26 (1996).

[5] Krivenko, PV, Mhitaryan, NM, Chirkova, VV, Zgardan, EP, “Durability of alkaline cement concretemade with alkali-reactive aggregates”. In: Malhotra, VM (Editor): Pr oceedings of the fourth CANMET/ACI

international conference on durability of concrete, Sydney, Australia: 587-596(1997).[6] Krivenko, PV, Gelevera, AG, Fedorenko,YV, “The effect of alkali on destructive and constructive processes

during alkali-aggregate reaction”. In: Caijun Shi and Xiaodong Shen (Editor): RILEM PRO 72, Advances in

Chemically-Activated Materials, CAM’2010, Jinan, China: 147-154 (2010).

[7] García Lodeiro, I, Palomo, A, “Alkali-aggregate reaction in activated fly ash systems”. Cement and ConcreteResearch (37): 175-183 (2007).[8] Fernández-Jiménez, A, García Lodeiro, I, “Durability of alkali-activated fly ash cementitious materials”.

Material Science (42): 3055-3065 (2007).[9] Lee, ST, Lee, JH, “Microstructural observations of fly ash-based geopolymer composites with different curing

conditions”. In: Caijun Shi and Xiaodong Shen (Editor): RILEM PRO 72, Advances in Chemically-Activated

Materials, CAM’2010, Jinan, China: 155-161 (2010).

[10] Pavel Rovnaník, “Effect of curing temperature on the development of hard structure of metakaolin-basedgeopolymer”. Construction and Building Materails (24): 1176-1183 (2010).


879313

Documents