Fuel 133 (2014) 332–340

Contents lists available at ScienceDirect

Fuel
journal homepage: www.elsevier.com/locate/fuel

Effect of blast furnace slag grades on fly ash based geopolymer waste

forms
Hui Xu, Weiliang Gong ⇑, Larry Syltebo, Kevin Izzo, Werner Lutze, Ian L. Pegg
Vitreous State Laboratory, The Catholic University of America, 620 Michigan Ave NE, Washington, DC 20064, USA

h i g h l i g h t s

 Slags of grades 80, 100 and 120 investigated on fly ash based geopolymer waste forms.
 Concentrated Hanford radioactive waste similant consisting of more than 20 chemicals.
 Higher grade slag not definitely leading to higher strength or shorter setting times.
 More hydration heat for higher grade slag confirmed by calorimetric study.
 Reactivity of higher grade slag found to be better exploited at enhanced Si/Ca ratios.

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

Article history: Ground granulated blast furnace slags (GGBFSs) of grades 80, 100 and 120 were investigated for high
Received 6 February 2014 waste loading fly ash based geopolymer waste forms. Samples were prepared at a fixed fly ash/GGBFS
Received in revised form 7 May 2014 mass ratio of 5/3, using an activating solution prepared from concentrated Hanford secondary waste
Accepted 8 May 2014
(HSW) simulant. The fresh pastes were subjected to isothermal conduction calorimetry and Vicat setting
Available online 23 May 2014
time measurements, and the cured waste forms were characterized by compressive strength test, XRD
and SEM/EDS analyses, as well as the TCLP leaching test. The results show that GGBFS of higher grade
Keywords:
generated more hydration heat, yet not definitely led to higher compressive strength or shorter setting
Furnace slag
Fly ash
times, suggesting that the GGBFS grading index established for cement industry may not be simply intro-
Geopolymer duced to geopolymer application. It was also found that the reactivity potential of high grade GGBFS in fly
Waste form ash based geopolymer might be better exploited at enhanced SiO2/Al2O3 and SiO2/CaO ratios. Except for
Hanford secondary waste rhenium, which is not regulated in TCLP, all heavy metals and hazardous elements in the HSW simulant
were effectively immobilized by the geopolymer waste forms. However, effect of different GGBFS on
heavy metals and hazardous elements fixation depended on different metals and elements.
Ó 2014 Elsevier Ltd. All rights reserved.

1. Introduction As described by Davidovits, the fundamental structure of geopoly-

Geopolymers are a class of synthetic alkali-activated alumino-
silicate inorganic polymers (AIPs) featuring a predominantly
X-ray amorphous three-dimensional network [1]. The term geo-
polymer was coined by Joseph Davidovits three decades ago [2].
Abbreviations: GGBFS, ground granulated blast furnace slag; HSW, Hanford
secondary waste; AIPs, aluminosilicate inorganic polymers; C–A–S–H, aluminium-
modified calcium silicate hydrate; N–A–S–H, sodium aluminosilicate hydrate; OPC,
ordinary Portland cement; ILAW, immobilized low-activity waste; HEPA, high-
efficiency particulate air; PNNL, Pacific Northwest National Laboratory; ICC,
Isothermal Conduction Calorimeter; ANSI/ANS, American National Standards Insti-
tute/American Nuclear Society.
⇑ Corresponding author. Tel.: +1 (202) 319 5393; fax: +1 (202) 319 4469.
E-mail address: gongw@vsl.cua.edu (W. Gong).
mers consists of SiO4 and AlO4 tetrahedrons linked alternatively
(–SiO4–AlO4–, or –SiO4–AlO4–SiO4–, or –SiO4–AlO4–SiO4–SiO4–)
by sharing all oxygen atoms between two tetrahedral units. The
negative charges on tetrahedral AlO4 are balanced by alkalis cat-
ions (typically Na+ and/or K+) provided by alkaline activating solu-
tion. Theoretically, any aluminosilicate material can be used for
geopolymer synthesis.
The geopolymerization of some industrial wastes such as coal
ashes has attracted increasing interests since the past two decades
[3–7]. Fly ash is generated during combustion of coal in thermal
power plants and essentially contains SiO2 and Al2O3 along with
other compounds, including CaO, Fe2O3, MgO, MnO, etc. [8,9]. Fly
ash has become an important raw binder material for geopolymer
due to its high SiO2 and Al2O3 contents, wide availability in large

http://dx.doi.org/10.1016/j.fuel.2014.05.018
0016-2361/Ó 2014 Elsevier Ltd. All rights reserved.

quantities, sound property of the hardened products, low cost in
comparison to those of traditional calcined natural clays, as well
as significant environmental benefits including natural resource
preservation, low energy consumption and reduced CO2 emission
[10–13]. However, former studies have also revealed that, gener-
ally, the dissolution of fly ash at room temperature had not com-
pleted before the final hardened structure was formed [5,6,14].
The relatively low reactivity of fly ash usually led to slow setting
of the fresh paste and poor compressive strength of the hardened
material [15,16]. In many cases, for fly ash based geopolymers,
an elevated curing temperature of 40–85 °C is needed to favor a
better compressive strength gain [17,18].
In order to compensate the disadvantages of fly ash in geopoly-
merization, ground granulated blast furnace slag (GGBFS) has been
incorporated into fly ash based geopolymers [19,20]. GGBFS is an
amorphous by-product of the steel industry with a latent hydraulic
reactivity, which can be catalyzed by proper activators to form
cementitious materials [21,22]. The incorporation of calcium-rich
GGBFS into fly ash based geopolymers may improve the setting
time and compressive strength of the geopolymer by forming alu-
minium-modified calcium silicate hydrate (C–A–S–H) gel in addi-
tion to the sodium aluminosilicate hydrate (N–A–S–H) gel
(geopolymer gel) [23,24] and compacting the microstructure [25].
On the other hand, GGBFS has been recognized as one of the
major cementitious materials in ordinary Portland cement (OPC)
concretes [26]. For the cement industry, an ASTM C989 standard
has been established, which classifies GGBFS into three strength
grades, i.e. 80, 100 and 120 [27]. According to ASTM C989, the
strength grades of GGBFS are determined by their slag activity
index
Slag activity index; % ¼ SP=P  100
where SP represents average compressive strength of the slag-refer-
ence cement cubes consisting of 50 wt% slag and 50 wt% Portland
cement, and P is the average compressive strength of the reference
cement cubes of 100 wt% Portland cement. The ASTM C989 stan-
dard is briefed in Table 1.
Although incorporation of different amounts of one GGBFS into
fly ash and/or metakaolin based geopolymers has received inten-
sive investigation [19,20,25,28], the implication of the GGBFS grad-
ing index to geopolymer synthesis remains poorly understood.
Besides, to the authors’ best knowledge, no study on effect of
GGBFSs of different grades on chemical durability of fly ash based
geopolymer waste forms has ever been published.
The present study, therefore, investigated the effect of GGBFSs
of grades 80, 100 and 120 on fly ash based geopolymer waste
forms. A concentrated Hanford secondary waste (HSW) stream
simulant containing 5 mol/L sodium and spiked with 100 mg/L
NaReO4 (as an analogue for radionuclide 99Tc) was employed to
prepare the alkaline activating solution. The geopolymer waste
forms were cured at room temperature and characterized by iso-
thermal conduction calorimetry, setting time and compressive
strength measurements, X-ray diffractography (XRD), scanning
electron microscopy (SEM), energy dispersive X-ray spectrum
Table 1
ASTM C 989 standard for the classification of different grade slag.
Day Grade Minimum slag activity index%
index type
Average of last five
consecutive samples
7 days Grade 80 – –
Grade 100 75
Grade 120 95
28 days Grade 80 75
Grade 100 95
Grade 120 115
H. Xu et al. / Fuel 133 (2014) 332–340 333
(EDS), as well as the toxicity characteristic leaching procedure
(TCLP) leach test [29].
2. Materials and methods
2.1. Raw materials
The major raw binder materials used in this study are class F fly
ash from Brandon Shores Power Plant provided by Separation
Technologies LLC, GGBFS grades 80 and 120 provided by Lafarge
North America Inc., and GGBFS grade 100 provided by Holcim US
Inc. An amorphous silica fume, obtained from Norchem Inc., was
employed at 8.67 wt% of the sum of the fly ash and GGBFS as par-
tial replacement of the fly ash and GGBFS in some samples. A Type
5A 8  12 mesh molecular sieve (Zeolite 5A), obtained from Delta
Enterprises, was ground to fine powders and used as an additive
at 1 wt% of the paste. Chemical compositions of the raw materials
as analyzed by X-ray fluorescence (XRF) are given in Table 2. Par-
ticle size distributions of the fly ash and GGBFSs are shown in
Fig. 1.
2.2. Hanford secondary waste simulants
Liquid Hanford secondary waste (HSW) simulant S1 was used to
prepare the alkaline activating solution for the fly ash based geo-
polymer waste forms. The HSW simulant S1 represents the base-
line for the immobilized low-activity waste (ILAW) off-gas
caustic scrubber effluent downstream of the high-efficiency partic-
ulate air (HEPA) filters at the Hanford Site, WA. The composition of
HSW simulant S1, as provided by Pacific Northwest National Labo-
ratory (PNNL), indicates a 1 mol/L sodium concentration [30]. In
this study, the sodium concentration for HSW simulant S1 was
increased to 5 mol/L. It is assumed that radionuclide technetium-
99 will occur in the HSW waste stream in its most soluble form,
i.e. pertechnetate TcO 4 [30]. Therefore, rhenium was spiked as
NaReO4 into the simulant as an analogue for technetium-99 in its
soluble form 99TcO 4 . The recipe for a 10 L batch of the HSW simu-
lant S1 prepared in this study is presented in Table 3.
2.3. Geopolymer preparation
The activating solution was prepared from the HSW simulant S1
without additional water. Under mechanical stirring, sodium
hydroxide was dissolved into the simulant. After cooled to room
temperature, the solution was mixed with solid binder materials
on a Lancaster K-Lab mixer for several minutes. Most of the
obtained paste was cast into 5.08  10.16 cm cylindrical plastic
molds, whereas a small portion of the fresh paste was subjected
to the setting time test. The molded samples were immediately
placed on a vibrating table for about two minutes to remove
entrapped air bubbles. Then, the samples were sealed with lids
and cured at room temperature for designated ages.
For all the GGBFSs, two formulations, A and B, were designed to
prepare the fly ash based geopolymer waste forms. While formula-
tion A is a simple fly ash-GGBFS system containing 1.0 wt% addi-
tive of zeolite 5A, formulation B replaces 8.67 wt% of the sum of
Any individual the fly ash and GGBFS by soluble silica fume. However, the fly
sample ash/GGBFS mass ratios were kept constant at 5/3 throughout the
study. A summarized recipe for the geopolymer waste forms of
70 the two formulations is presented in Table 4. Each recipe was tai-
90
lored for a 6.2 kg batch of fresh geopolymer paste, usually yielding
70 fourteen to fifteen cylindrical samples.
90
The nomenclature used for geopolymer sample ID is given in
110
Table 5.
334 H. Xu et al. / Fuel 133 (2014) 332–340

Table 2
Chemical compositions of raw materials as determined by X-ray fluorescence (wt%).

Composition Ground blast furnace slag Fly ash Silica fume Zeolite 5A
Grade 80 Grade 100 Grade 120
Al2O3 9.59 7.26 9.97 27.25 0.16 26.30
BaO 0.15 0.06 0.11 0.10 0.00 0.01
CaO 37.38 36.48 40.72 0.70 0.45 10.96
Fe2O3 0.55 0.36 0.48 3.06 0.07 1.01
K2O 0.54 0.42 0.44 3.26 0.43 0.54
MgO 11.20 12.69 8.22 0.82 0.58 2.76
MnO 0.77 0.47 0.40 0.02 0.02 0.02
Na2O 0.32 0.00 0.20 0.00 0.00 3.91
PbO 0.00 0.00 0.00 0.01 0.00 0.00
SO3 1.71 1.15 1.12 0.35 0.15 0.01
SiO2 36.89 40.11 37.10 60.41 96.18 42.43
LOIa <0.01 0.38 0.43 1.24 1.39 11.43
a
LOI: loss on ignition.

were determined on Vicatronic Model E044N Automatic Recording

Apparatus (MATEST, Italy) as per ASTM C 191 standard. Cured
waste form samples were subjected to compressive strength mea-
surement on a CM-4000-SD instrument (Test Mark Industries,
USA) at the curing ages of 1, 3, 7, 14, 21 and 28 days. The compres-
sion machine was calibrated against the NIST traceable standards
by a service contractor. The crushed 28-day samples were collected
to prepare the required particle size fraction for the TCLP leaching
test. The cured samples were also characterized by XRD, SEM and
EDS analyses. XRD pattern was obtained on a Thermo ARL X’tra
X-ray Diffractometer. A Jeol JSM-5910LV Scanning Electron Micro-
scope equipped with Oxford 7324 apparatus was employed to per-
form the SEM/EDS analyses.

3. Results and discussion

Fig. 1. Particle size distributions of fly ash and ground granulated blast furnace
slags of grades 80, 100 and 120.
3.1. Setting time

Table 3 Setting time is an important parameter in practice for cementi-

Recipe for a 10 L batch of Hanford secondary waste (HSW) stream simulant S1 tious pastes, which determines the time available for transport,
containing 5 mol/L sodium and 100 mg/L rhenium.
placing and compaction of the materials. In this study, the initial
Order of Analyte Target Reagent Assay Target and final setting times of the fresh geopolymer pastes were mea-
addition mol/L mass (g) sured on a Humboldt vicat apparatus. The results are given in
1 Ag 1.30E05 AgNO3 0.999 0.032 Table 6.
1 As 6.96E05 Na2HAsO47H2O 1.000 0.326 The major factors influencing geopolymer setting times may
1 Al 1.88E01 Al(OH)3 0.990 221.838 include the chemical and phase compositions of raw binder mate-
1 Cd 3.14E06 Cd(NO3)24H2O 0.985 0.015
1 CO2 4.56E02 Na2CO3 1.000 72.472
rials, nature of alkaline activating solution, water-to-geopolymer
3
1 Cr 4.06E04 Na2Cr2O72H2O 0.990 1.832 solids ratio, fineness of the solids, as well as curing temperature
3
1 PO4 1.37E02 Na2HPO47H2O 0.990 55.759 [4,31]. Since the class F fly ash, alkaline activating solution,
1 Si 3.76E03 Na2SiO3 0.980 7.021 water-to-geopolymer solids ratio and curing temperature were
1 SO2 8.82E03 Na2SO4 0.990 18.971
4
kept constant throughout this work, the compositions and fineness
1 NO 3 6.56E01 NaNO3 0.980 851.318
1 NO 2 2.40E02 NaNO2 0.980 25.332 of the GGBFS are believed to be responsible for the differences in
1 Cl 4.50E02 NaCl 0.980 40.233 the setting times. Chindaprasirt has reported that higher reactive
1 F 1.11E03 NaF 0.990 0.682 contents and finer particle size of the raw binder material led to
1 K 1.17E03 KNO3 0.990 1.785 faster reaction and setting of the geopolymer pastes [32]. For fly
1 Pb 1.80E05 Pb(NO3)2 0.990 0.090
1 TOC 1.67E01 Na2C2O4 0.980 343.056
ash-furnace slag based geopolymers, Kumar has observed that
1 TOC 2.05E02 (COOH)22H2O 0.980 39.440 the reaction at 27 °C was dominated by activation of the GGBFS,
2 H2O 5.29E+01 H2O 1.000 14288.127 whereas that at 60 °C was due to combined interaction of fly ash
3 OH 7.94E01 NaOH 0.987 482.784 and GGBFS [25].
4 Tc 2.69E04 HReO4 0.562 1.781
As presented in Table 6, for the fly ash based geopolymer sam-
Spiking level: Re = 100 mg/L Total 16447.521 ples of both A and B formulations, GGBFS grade 120 established the
shortest initial and final setting times, whereas GGBFS grade 100
2.4. Methods of analysis held the longest initial and final setting times. As revealed by the
results of XRF analysis (Table 1), the GGBFS grade 120 contains
Heat evolutions of fresh geopolymer waste form pastes were the highest calcium contents (40.90 wt% as oxide); GGBFS grade
measured at 25 °C on a TAM Air Model 3116 Isothermal Conduc- 80, the second (37.38 wt% as oxide); and GGBFS grade 100, the
tion Calorimeter (ICC). Initial and final setting times of the pastes lowest (36.62 wt% as oxide). Thus the obtained setting times are
 H. Xu et al. / Fuel 133 (2014) 332–340 335

Table 4
Recipe for fly ash based geopolymer waste forms of formulations A and B.

Order of addition Chemical and material Assay Target weight (g)

Formulation A Formulation B
1 HSW simulant S1a 1.000 1618.2 1610.3
2 NaOH 0.990 320.3 320.0
3 Furnace slag 1.000 1582.2 1445.9
3 Fly ash 1.000 2636.9 2409.9
3 Zeolite 5A 1.000 62.1 62.1
4 Silica fume filler 1.000 0.0 365.7
a
HSW simulant S1 containing 5 mol/L sodium and spiked with 100 mg/L rhenium.

Table 5 developed the highest 28-day compressive strength of 74.4 MPa,

Nomenclature for geopolymer sample ID.
Sample ID GGBFS grade
GEO80A 80
GEO100A 100
GEO120A 120
GEO80B 80
GEO100B 100
GEO120B 120
in well accordance with the findings of Chindaprasirt and Kumar.
In addition to the highest calcium content, the GGBFS grade 120
has the finest particle size (Fig. 1), which may also contribute to
the shortest setting times [31,32]. For each GGBFS, it can be found
from Table 6 that formulation B gave longer setting times than
formulation A, with the final setting time of sample GEO120B as
the only exception. For geopolymers, increased setting times have
been linked to a high SiO2/Al2O3 molar ratio by De Silva [33].
Therefore, the longer setting times of formulation B may be attrib-
uted to the higher molar ratios of Si/Al (as shown in Table 5), which
resulted from the replacement of small portions of fly ash and
GGBFS by silica fume.
3.2. Compressive strength
The results of compressive strength test at the curing ages of 1,
3, 7, 14, 21 and 28 days are illustrated in Fig. 2. For formulation A,
the geopolymer samples with GGBFS grades 80, 100 and 120
revealed 1-day compressive strength of 3.5, 5.3 and 16.0 MPa,
respectively. The 1-day compressive strength of sample GEO120A
is 4.6 times of that of GEO80A, suggesting that, for fly ash based
geopolymer, high grade GGBFS favors early compressive strength
development. However, after 14 days of curing, the impressive dif-
ferences among the compressive strengths started to decrease. At
the curing age of 28 day, samples GEO80A, GEO100A and GEO120A
evolved compressive strengths of 38.8, 37.7 and 35.7 MPa, imply-
ing that the high reactivity of high grade GGBFS was not fully
exploited for samples of formulation A.
However, the case for formulation B is different except for the
1-day compressive strengths. With the presence of additional reac-
tive silica source provided by the silica fume, sample GEO120B
which is more than 40% higher than 49.7 MPa for GEO80B and
Formulation 52.5 MPa for GEO100B. In addition, all formulation B samples dis-
A
played better strength development than their formulation A coun-
A terparts. The 28-day compressive strengths of samples GEO80B,
A GEO100B and GEO120B are 28.1%, 39.2% and 108.4% higher than
B those of GEO80A, GEO100A and GEO120A, respectively.
B
Although GGBFS grade 100 was supposed to be more reactive
B
than GGBFS grade 80, no significant difference between the com-
pressive strengths of samples GEO100B and GEO80B was observed
at all curing ages, partly due to the similar CaO contents and par-
ticle sizes of the two slags as shown in Fig. 1 and Table 2.
It is well accepted that the easiness of alkali-activation and
hydration of slag strongly depend on the chemistry and mineralogy
of the slag, including the particle size and degree of amorphicity,
along with the nature and concentration of the alkaline activator
as well the curing conditions [34–37].
It is also known that the incorporation of calcium-rich GGBFS
into fly ash based geopolymers may improve the compressive
strength by forming C–A–S–H gel in the geopolymer matrix
[24,25]. Therefore, the highest calcium content (Table 2) in GGBFS
grade 120 may play a major role in the best compressive strength
development for sample GEO120B. Besides, the finest particle size
of GGBFS grade 120 (Fig. 1) may also be responsible for the com-
pressive strength gain via improving the reactivity and compacting
the microstructure [34,25,38]. For the differences between samples
of formulations A and B, the presence of a reasonably high content
of soluble silica in geopolymer paste has been observed to benefit
the compressive strength evolvement [39,40]. Accordingly, the

Table 6
Setting times of fly ash based geopolymer pastes.

Sample ID Initial setting time (min) Final setting time (min)

GEO80A 111 191
GEO100A 136 271
GEO120A 109 155
GEO80B 134 247
GEO100B 141 327
GEO120B 110 144
Fig. 2. Compressive strength of fly ash based geopolymer waste forms.
336 H. Xu et al. / Fuel 133 (2014) 332–340

significantly enhanced compressive strength of formulation B sam-

ples may be attributed to the addition of the silica fume, which
acted as a reactive filler and provided additional soluble silica in
the geopolymer pastes. From the compressive strength point of
view, all geopolymer samples investigated in this study have met
the waste form acceptance criteria of at least 3.54 MPa [30].

3.3. Isothermal conduction calorimetry

Geopolymerization is a complex process consisting of several

steps such as dissolution, precipitation, reorganization and poly-
condensation, which are largely coupled and occur concurrently
[25]. In order to record the early stage of the reaction progress in
terms of heat evolution, six small batches of fresh geopolymer
pastes were prepared within 30 min, and the equilibration time
required after placement of the samples in the calorimeter, typi-
cally 45 min, was significantly shortened to 3 min. The reaction Fig. 4. Effect of different ground granulated blast furnace slags on heat evolution of
time has been adjusted for each sample starting from the time of fly ash based geopolymer waste forms with silica fume filler.
mixing and the results of the calorimetric test are shown in Figs. 3
and 4. be better exploited by addition of reactive silica source, which is
It is well known that, at room temperature, geopolymerization supported by the compressive strength data (Section 3.2). It was
of fly ash is extremely slow and the calorimetric response is typical also found that the addition of silica fume extended the times for
of a dissolution–precipitation reaction [5,6,25]. Thus the early the heat flow peaks to occur, suggesting longer setting times for
stage heat evolution of alkali-activated fly ash and GGBFS mixtures samples with silica fume filler, which is in consistence with the
is believed to be majorly contributed by hydration of GGBFS. Vicat test results (Table 6) and the report by De Silva [33].
Besides, it should be kept in mind that the heat flow curve repre-
sents several overlapping chemical reactions [25,41]. It was 3.4. X-ray diffractography
reported by Shi and Day [42] that, for alkali-activated slag at
25 °C, a small initial peak appeared just after the addition of the The XRD patterns of the major raw materials and geopolymer
activator solution, which might followed immediately by an addi- waste forms are shown in Figs. 5 and 6, respectively.
tional initial peak. The initial peak was attributed to the wetting In Fig. 5, a diffuse halo peak is seen for each of the XRD pattern,
and dissolution of slag grains and adsorption of some ions onto indicating that all the major raw materials are predominately
the surface of slag grains, whereas the additional initial peak was X-ray amorphous. No distinct crystalline peak was observed for
mainly linked to the formation of the primary C–A–S–H gel [42]. the GGBFS grades 100 and 120. For the GGBFS grade 80, a minor
However, in this study, we failed to catch the initial peaks, if any, level of calcium silicate (Ca2SiO4) (PDF # 00-033-0303) was found,
due to the time needed to prepare 6 samples one by one and the which was also identified by Tossavainen et al. and Luxán et al. in
time required for the calorimeter to equilibrate. different furnace slags [43,44]. The XRD pattern of the fly ash
Both Figs. 3 and 4 show higher heat flow peaks for higher grade reveals two major crystalline phases, i.e. quartz (SiO2) (PDF # 00-
GGBFS, indicating that higher grade GGBFS is more reactive in geo- 046-1045) and mullite (Al6Si2O13) (PDF # 00-015-0776). Both of
polymerization of fly ash-GGBFS blends. However, when compar-
ing Figs. 3 and 4 with Fig. 2, it can be found that high heat of
hydration is not definitely connected with high compressive
strength, which was also observed by Winnefeld et al. [41]. As
shown in Figs. 3 and 4, the heat flow peaks for samples with silica
fume filler are higher than those for samples without silica fume
filler, indicating that the reactivity of fly ash-GGBFS blend may

Fig. 5. X-ray diffraction patterns of class F fly ash, ground granulated blast furnace
Fig. 3. Effect of different ground granulated blast furnace slags on heat evolution of slags and silica fume. Q = quartz (SiO2); M = mullite (Al6Si2O13); and C = calcium
fly ash based geopolymer waste forms without silica fume filler. silicate (Ca2SiO4).
 H. Xu et al. / Fuel 133 (2014) 332–340
Fig. 6. X-ray diffraction patterns of geopolymer waste forms. Q = quartz (SiO2);
M = mullite (Al6Si2O13); C = cancrinite (Na6Ca2Al6Si6O24(CO3)22H2O); and N = nitr-
atine (NaNO3).
the phases are typical for class F fly ash [18,23,24]. A considerable
amount of amorphous material in the fly ash is indicated by the
broad diffuse hump at 15–30 deg (2h). The amorphous nature of
the silica fume is confirmed by the broad diffuse hump centered
at around 21 deg (2h). For all the GGBFSs, the similar hump at
the same 2h region suggests that the major amorphous phases in
the three GGBFSs may possibly be the same.
As shown in Fig. 6, the XRD patterns for all the geopolymer
waste forms are quite similar. No substantial difference was
observed among the XRD patterns for all the six samples with slags
of different grades. The amorphous nature of geopolymer is
revealed by the broad diffuse humps at 15–38 deg (2h). The crys-
talline phases of the geopolymer waste forms include quartz and
mullite, which are the major crystals present in the fly ash
(Fig. 5). By comparing the XRD patterns in Figs. 5 and 6, two new
crystalline phases, i.e. cancrinite (Na6Ca2Al6Si6O24(CO3)22H2O)
(PDF # 00-046-1332) and nitratine (NaNO3) (PDF # 00-036-
1474), were identified in all the geopolymer waste form samples.
The formation of cancrinite was also observed in an alkali-
activated metakaolin-slag waste form by Chen et al. [45] and in
an alkali-activated fly ash by Singh et al. [46]. It is also a reflection
of the notable amount of sodium carbonate in the HSW stream
simulant in this study as shown in Table 3. The presence of nitra-
tine in all the geopolymer waste form samples investigated is
attributed to the remarkable inventory of sodium nitrate in the
waste simulant (Table 3).
It is, however, difficult to identify the N–A–S–H and C–A–S–H
gels formed with XRD techniques due to their amorphous and
semi-crystalline nature, respectively [47,48].
3.5. SEM micrographs and EDS spectra
SEM micrographs of all the geopolymer samples are shown in
Fig. 7. Fig. 7b, d and f shows a generally more condense gel struc-
ture for formulation B in comparison to that for formulation A in
Fig. 7a, c and e. This suggests that samples of formulation B may
develop higher compressive strengths than those of formulation
A [7], which is consistent with the results of compressive strength
test (Fig. 2). Besides, Fig. 7 displays a very similar heterogeneous
337
binding matrix with particles of various shapes and sizes for all
the geopolymer samples.
Take Fig. 7f for example, the light gray angular particle and the
gray spherical particle marked in Fig. 7f were identified as unre-
acted residues of the GGBFS grade 120 and fly ash, respectively
[20]. Both of the unreacted GGBFS and fly ash are distinct from
the porous binder [23]. Heterogeneously distributed cavities were
found both inside and outside the remnant fly ash particles, indi-
cating a more porous microstructure for fly ash geopolymer
[24,49].
The EDS spectrum of the GGBFS particle (not shown herein)
revealed 36.3 wt% CaO, 37.5 wt% SiO2, and 8.8 wt% Al2O3 at the
testing spot, which are in acceptable agreement with the results
of the elemental analysis in Table 2 (40.9 wt% CaO, 37.3 wt% SiO2,
and 10.0 wt% Al2O3). For the fly ash pellet marked in Fig. 7f, the sil-
icon and aluminium contents obtained by EDS analysis are
56.5 wt% SiO2 and 30.9 wt% Al2O3, which are close to 61.2 wt%
SiO2 and 27.6 wt% Al2O3 in Table 2, respectively. The bright sphere
partially shown in Fig. 7f, containing 76.4 wt% Fe as indicated by
the EDS data, is an iron-rich particle, which may be attributed to
the ferrite spinels present in the fly ash [24].
Fig. 8 presents the SEM micrograph and EDS spectra of sample
GEO120B, showing the heterogeneous nature of the binder gel.
The results of EDS analyses for spectra 1 and 2 were summarized
in Table 7. As shown in Fig. 8 and Table 7, the EDS spectrum 1
revealed a molar ratio of MT/MAl = 0.18, where MT = MNa + MK + 2-
MCa, indicating the formation of N–A–S–H gel. For the spectrum
2, a much higher MT/MAl of 4.38 was found, suggesting the pres-
ence of a hybrid of N–A–S–H and C–A–S–H gels in the geopolymer
matrix. The N–A–S–H gel area of spectrum 1 exhibits a notably
different morphological feature compared to the N–A–S–H and
C–A–S–H hybrid gel area of spectrum 2, which is associated with
the different mineralogy and shapes of fly ash and GGBFS particles,
and in particular the differences in solubility and reaction product
gel nature between fly ash geopolymer and alkali-activated slag
precursors [24].
3.6. Toxicity characteristic leaching procedure (TCLP)
The TCLP leaching test on the fly ash based geopolymer waste
forms was conducted after 28 days of curing under room temper-
ature, and the results are presented in Table 8.
As shown in Table 8, the concentrations of Cd, Ag and As in the
TCLP leachates are all lower than their respective detection limits,
except for that of Cd from sample GEO80B, which just reached its
detection limit. Thus Cd, Ag and As were effectively immobilized
by any of the geopolymer waste forms. In both A and B formula-
tions, Cr was best stabilized by samples with GGBFS grade 120
and Pb was best fixed by samples with GGBFS grade 100. For for-
mulation A, sample GEO100A revealed best fixation on Pb, Ba
and Re. However, in formulation B, only Pb was best fixed by sam-
ple GEO100B. The relatively high Pb concentration in the TCLP
leachate was attributed to the relatively high Pb inventory in the
raw material (fly ash). For all the waste form samples, the best fix-
ation of Cr and Re were achieved by GEO120A and GEO100A,
where as Pb and Ba were best immobilized by GEO100B and
GEO120B, respectively. In general, the effect of GGBFS of different
grades on immobilization of the hazardous elements and Re by the
fly ash based geopolymer waste forms differed for different ele-
ments. Besides, the replacement of small portion of fly ash and
GGBFS by amorphous silica fume also played a role on the hazard-
ous elements and Re fixation through altering the host matrices of
the waste forms [50]. However, the use of the silica fume did not
give rise to an overall advantage on fixation of the hazardous ele-
ments and Re. Except for Re, which is not regulated in TCLP and
better determined by the American National Standards Institute/
338 H. Xu et al. / Fuel 133 (2014) 332–340

Fig. 7. SEM image of fly ash based geopolymer waste forms for a. GEO80A, b. GEO80B, c. GEO100A, d. GEO100B, e. GEO120A, and f. GEO120B.

Fig. 8. SEM image and EDS spectra of fly ash based geopolymer waste form GEO120B, showing the coexistence of N–A–S–H gel and C–A–S–H gel in one geopolymer matrix.

American Nuclear Society (ANSI/ANS)-16.1 leaching test [30], all any of the fly ash based geopolymer waste forms as per the US
the hazardous elements have been effectively immobilized by EPA limit.
 H. Xu et al. / Fuel 133 (2014) 332–340 339

Table 7 [3] Gong W, Lutze W, Chen C, Pegg IL. Reactivity of fly ash in strongly alkaline
Summary results of EDS analyses for geopolymer waste form sample GEO120B solution. US Patent; 2011, 0052921 A1.
(atomic %). [4] Chindaprasirt P, Rattanasak U. Utilization of blended fluidized bed combustion
(FBC) ash and pulverized coal combustion (PCC) fly ash in geopolymer. Waste
Elements Spectrum 1 Spectrum 2 Manage 2010;30:667–72.
[5] Chen C, Gong W, Lutze W, Pegg IL, Zhai J. Kinetics of fly ash leaching in strongly
Al 16.14 5.57
alkaline solutions. J Mater Sci 2011;46:590–7.
Ca 0.27 8.04 [6] Chen C, Gong W, Lutze W, Pegg IL. Kinetics of fly ash geopolymerization. J
Fe 0.76 0.52 Mater Sci 2011;46:3073–83.
K 1.34 1.19 [7] Xu H, Li Q, Shen L, Zhang M, Zhai J. Low-reactive circulating fluidized bed
Mg 0.72 1.37 combustion (CFBC) fly ashes as source material for geopolymer synthesis.
Na 0.99 7.20 Waste Manage 2010;30:57–62.
O 57.58 50.50 [8] Li F, Zhai J, Fu X, Sheng G. Characterization of fly ashes from circulating
Si 22.20 25.62 fluidized bed combustion (CFBC) boilers cofiring coal and petroleum coke.
Total 100.00 100.00 Energy Fuel 2006;20:1411–7.
[9] Li Q, Xu H, Fu X, Chen C, Zhai J. Effects of circulating fluidized bed combustion
(CFBC) fly ashes as filler on the performances of asphalt. Asia Pac J Chem Eng
2009;4:226–35.
[10] Duxson P, Provis JL, Lukey GC, van Deventer JSJ. The role of inorganic polymer
technology in the development of ‘green concrete’. Cem Concr Res
Table 8
2007;37:1590–7.
TCLP leaching test results for heavy metals, hazardous elements and rhenium. [11] Andini S, Cioffi R, Colangelo F, Grieco T, Montagnaro F, Santoro L. Coal fly ash as
raw material for the manufacture of geopolymer-based products. Waste
Sample ID Concentration in TCLP leachate (mg/L)a
Manage 2008;28:416–23.
Cd Cr Pb Ag As Ba Re [12] Li F, Li Q, Zhai J, Sheng G. Effect of zeolitization of CFBC fly ash on
immobilization of Cu2+, Pb2+, and Cr3+. Ind Eng Chem Res 2007;46:
GEO80A <0.03 0.17 0.84 <0.07 <0.20 2.20 0.56 7087–95.
GEO100A <0.03 0.18 0.70 <0.07 <0.20 1.40 0.46 [13] Chindaprasirt P, Jaturapitakkul C, Chalee W, Rattanasak U. Comparative study
GEO120A 0.03 0.11 1.00 <0.07 <0.20 1.50 0.51 on the characteristics of fly ash and bottom ash geopolymers. Waste Manage
GEO80B <0.03 0.16 0.67 <0.07 <0.20 1.75 0.50 2009;29:539–43.
GEO100B <0.03 0.17 0.54 <0.07 <0.20 2.16 0.81 [14] Somna K, Jaturapitakkul C, Kajitvichyanukul P, Chindaprasirt P. NaOH-
GEO120B <0.03 0.14 0.77 <0.07 <0.20 1.24 0.67 activated ground fly ash geopolymer cured at ambient temperature. Fuel
EPA limitb 1.00 5.00 5.00 5.00 5.00 100 N/A 2011;90:2118–24.
[15] Puertas F, Martı´ nez-Ramı́rez S, Alonso S, Vázquez T. Alkali-activated fly ash/
a
Note: concentration below detection limit is presented as <0.XX, where 0.XX is slag cements: strength behaviour and hydration products. Cem Concr Res
the detection limit value. 2000;30:1625–32.
b
N/A: not available. [16] Kirschner A, Harmuth H. Investigation of geopolymer binders with respect to
their application for building materials. Ceram-Silik 2004;483:117–20.
[17] Bakharev T. Geopolymeric materials prepared using Class F fly ash and
4. Conclusions elevated temperature curing. Cem Concr Res 2005;35:1224–32.
[18] Palomo A, Grutzeck MW, Blanco MT. Alkali-activated fly ashes: a cement for
Effect of ground granulated blast furnace slags (GGBFSs) of dif- the future. Cem Concr Res 1999;29:1323–9.
[19] Li Z, Liu S. Influence of slag as additive on compressive strength of fly ash-
ferent grades, i.e., grades 80, 100 and 120, on fly ash based geopoly- based geopolymer. J Mater Civ Eng 2007;19:470–4.
mers waste forms was investigated for fixation of a concentrated [20] Escalante Garcia JI, Campos-Venegas K, Gorokhovsky A, Fernández A.
Hanford secondary waste (HSW) stream simulant. The results Cementitious composites of pulverised fuel ash and blast furnace slag
activated by sodium silicate: effect of Na2O concentration and modulus. Adv
show that higher grade GGBFS generated more hydration heat,
Appl Ceram 2006;105:201–8.
suggesting higher reactivity for higher grade GGBFS in geopoly- [21] Shi C. Steel slag-its production, processing, characteristics, and cementitious
merization. However, the incorporation of GGBFS of higher grade properties. J Mater Civ Eng 2004;16:230–6.
in fly ash based geopolymer may not definitely lead to shorter set- [22] Lecomte I, Henrist C, Liégeois M, Maseri F, Rulmont A, Cloots R. (Micro)-
structural comparison between geopolymers, alkali-activated slag cement and
ting time and higher compressive strength. The potential of higher Portland cement. J Eur Ceram Soc 2006;26:3789–97.
grade GGBFS for compressive strength development could be bet- [23] Bernal SA, Provis JL, Walkley B, San Nicolas R, Gehman JD, Brice DG, et al. Gel
ter exploited with the presence of additional amorphous silica nanostructure in alkali-activated binders based on slag and fly ash, and effects
of accelerated carbonation. Cem Concr Res 2013;53:127–44.
source. The setting time and compressive strength of fly ash based [24] Ismail I, Bernal SA, Provis JL, San Nicolas R, Hamdan S, van Deventer JSJ.
geopolymer incorporated with GGBFS are believed to be deter- Modification of phase evolution in alkali-activated blast furnace slag by the
mined by the coexistence of sodium aluminosilicate hydrate incorporation of fly ash. Cem Concr Compos 2014;45:125–35.
[25] Kumar S, Kumar R, Mehrotra SP. Influence of granulated blast furnace slag on
(N–A–S–H) gel (geopolymer gel) and aluminium-modified calcium the reaction, structure and properties of fly ash based geopolymer. J Mater Sci
silicate hydrate (C–A–S–H) gel, which is supported by the XRD/EDS 2010;45:607–15.
results. Although the effect of GGBFS of different grades on chem- [26] Osborne GJ. Durability of Portland blast-furnace slag cement concrete. Cem
Concr Compos 1999;21:11–21.
ical durability of the fly ash based geopolymer waste forms differed [27] American Society for Testing and Materials, ASTM, C989-99.
for different hazardous elements, the results of the TCLP leaching [28] Bernal SA, Provis JL, Rose V, Mejía de Gutiérrez R. Evolution of binder structure
test indicate that all geopolymer waste forms investigated in this in sodium silicate-activated slag-metakaolin blends. Cem Concr Compos
2011;33:46–54.
work have met the US EPA limits for immobilization of toxic heavy
[29] EPA Method 1311: Toxicity characteristic leaching procedure (TCLP). 1992.
metals and hazardous elements. [30] Gong W, Lutze W, Pegg IL. DuraLith alkali-aluminosilicate geopolymer waste
form testing for Hanford secondary waste. PNNL-20565; 2011.
[31] Chindaprasirt P, Chareerat T, Hatanaka S, Cao T. High-strength geopolymer
Acknowledgments using fine high-calcium fly ash. J Mater Civ Eng 2011;23:264–70.
[32] Chindaprasirt P, Homwuttiwong S, Sirivivatnanon V. Influence of fly ash
Lafarge North America Inc., Holcim US Inc., Separation Technol- fineness on strength, drying shrinkage and sulfate resistance of blended
cement mortar. Cem Concr Res 2004;34:1087–92.
ogies LLC and Norchem Inc. are gratefully appreciated for providing [33] De Silva P, Sagoe-Crenstil K, Sirivivatnanon V. Kinetics of geopolymerization:
the furnace slags, Class F fly ash and silica fume, respectively. role of Al2O3 and SiO2. Cem Concr Res 2007;37:512–8.
[34] Ben Haha M, Lothenbach B, Le Saout G, Winnefeld F. Influence of slag
chemistry on the hydration of alkali-activated blast-furnace slag – Part I: effect
References of MgO. Cem Concr Res 2011;41:955–63.
[35] Ben Haha M, Lothenbach B, Le Saout G, Winnefeld F. Influence of slag
[1] Provis JL, van Deventer JSJ. Geopolymerisation kinetics. 1. In situ energy- chemistry on the hydration of alkali-activated blast-furnace slag – Part II:
dispersive X-ray diffractometry. Chem Eng Sci 2007;62:2309–17. effect of Al2O3. Cem Concr Res 2012;42:74–83.
[2] Davidovits J. Mineral polymers and methods of making them. US Patent; 1982, [36] Bernal SA, San Nicolas R, Myers RJ, Mejía de Gutiérrez R, Puertas F, van
4349386. Deventer JSJ, et al. MgO content of slag controls phase evolution and structural
340 H. Xu et al. / Fuel 133 (2014) 332–340

changes induced by accelerated carbonation in alkali-activated binders. Cem
Concr Res 2014;57:33–43.
[37] Shi C, Day RL. Selectivity of alkaline activators for the activation of slags. Cem
Concr Aggr 1996;18:8–14.
[38] Li C, Sun H, Li L. A review: the comparison between alkali-activated slag
(Si+Ca) and metakaolin. Cem Concr Res 2010;40:1341–9.
[39] Kovalchuk G, Fernández-Jiménez A, Palomo A. Alkali-activated fly ash: effect of
thermal curing conditions on mechanical and microstructural development –
Part II. Fuel 2007;86:315–22.
[40] Xu H, Li Q, Shen L, Wang W, Zhai J. Synthesis of thermostable geopolymer from
circulating fluidized bed combustion (CFBC) bottom ashes. J Hazard Mater
2010;175:198–204.
[41] Winnefeld F, Leemann A, Lucuk M, Svoboda P, Neuroth M. Assessment of phase
formation in alkali activated low and high calcium fly ashes in building
materials. Constr Build Mater 2010;24:1086–93.
[42] Shi C, Day RL. A calorimetric study of early hydration of alkali-slag cements.
Cem Concr Res 1995;25:1333–46.
[43] Tossavainen M, Engstrom F, Yang Q, Menad N, Lidstrom Larsson M, Bjorkman
B. Characteristics of steel slag under different cooling conditions. Waste
Manage 2007:1335–44.
[44] Luxán MP, Sotolongo R, Dorrego F, Herrero E. Characteristics of the slags
produced in the fusion of scrap steel by electric arc furnace. Cem Concr Res
2000;30:517–9.
[45] Chen S, Wu M, Zhang S. Mineral phases and properties of alkali-activated
metakaolin-slag hydroceramics for a disposal of simulated highly-alkaline
wastes. J Nucl Mater 2010;402:173–8.
[46] Singh DN, Jha B, Srinivas K. Determination of crystallinity of alkali activated fly
ash by XRD and FTIR Studies. Constitutive Model Geomater, Springer Series
Geomech Geoeng 2013:477–81.
[47] Puertas F, Palacios M, Vázquez T. Carbonation process of alkali-activated slag
mortars. J Mater Sci 2006;41:3071–82.
[48] Li Q, Xu H, Li F, Li P, Shen L, Zhai J. Synthesis of geopolymer composites from
blends of CFBC fly and bottom ashes. Fuel 2012;97:366–72.
[49] Fernández-Jiménez A, Palomo A, Criado M. Microstructure development of
alkali-activated fly ash cement: a descriptive model. Cem Concr Res
2005;35:1204–9.
[50] Van Jaarsveld JGS, Van Deventer JSJ, Lorenzen L. Factors affecting the
immobilization of metals in geopolymerized fly ash. Metall Mater Trans B
1998;29:283–91.