Synthesis of waste-derived glass-ceramics from MSWI fly ash and EAF dust Kinetics of nucleation and ...
[ Pobierz całość w formacie PDF ]
Synthesis of Waste-Derived Glass-Ceramics from
MSWI Fly Ash and EAF Dust: Kinetics of Nucleation
and Crystallization
Tien-Chun Chu,
a
Kuen-Sheng Wang,
a
Kae-Long Lin,
b
Chang-Ching Chien,
a
and Jung-Hsing Chen
a
a
Graduate Institute of Environmental Engineering, National Central University, Chung-Li, Taoyuan, Taiwan, Republic of China
b
Department of Environmental Engineering, National Ilan University, I-Lan 26047, Taiwan, Republic of China; kllin@niu.edu.tw
(for correspondence)
Published online in Wiley Online Library (wileyonlinelibrary.com). DOI 10.1002/ep.11647
zation of various industrial [1–4] and municipal wastes [5–10].
This approach has gradually become an important method of
producing environmentally friendly materials while simulta-
neously improving the reuse and recycling of hazardous
inorganic wastes into materials with numerous favorable
properties. These properties include high chemical stability,
durability [11–13], and mechanical strength [14–18], making
them effective construction and decoration materials [7].
Glass-ceramics are composites of a residual glassy matrix
phase and a fine-grained crystal phase [19, 20]. They are
formed by the controlled crystallization of suitable parent
glasses. Crystallization usually involves nucleation and subse-
quent crystal growth, which must be accurately controlled to
ensure that the glass-ceramics thus formed have particular
properties, including crystallinity, mechanical strength, and
chemical resistance. In the glass-crystallization process, the
temperature-dependent nucleation rate (I) and crystal-growth
rate (U) must be optimized to maximize the formation of sta-
ble nuclei and efficiently grow high-yield crystals [21, 22].
This optimization can be achieved by adjusting the thermal
history of glass [23–27]. Thermal history significantly influen-
ces the rate curves for nucleation and crystal growth, particu-
larly if these curves overlap extensively. In this case, glass
crystallization can occur in a single-stage thermal treatment,
and nucleation and crystal growth are coupled. Controlling
the quality of glass-ceramics formed from inorganic solid
waste is more difficult than controlling the quality of glass-
ceramics formed from pure chemicals. Quality is regulated
by optimizing the kinetic mechanisms of nucleation and
crystallization in the waste treatment process. The initial com-
position of the parent glass is complex and varies among
sources of waste, strongly affecting crystallization reactions
and kinetic behavior. Thus, crystallization kinetics are partic-
ularly important in converting waste into glass-ceramics.
Information on glass crystallization is critical for future
research on the activation energy (E
c
) of crystallization and
the mechanisms of crystal nucleation and growth. In the liter-
ature, the nonisothermal crystallization kinetics of glassy
materials have been extensively elucidated by differential
thermal analysis (DTA) using the Kissinger [28, 29] and
Ozawa methods [30, 31]. However, only a few studies investi-
gate the crystallization behavior and kinetics of glass derived
from industrial wastes [32–37]. Most of these studies have
considered only coal fly ash or blast furnace slag and have
not provided significant information on other wastes.
The kinetics of crystal nucleation and growth critically
affect the conversion of waste into high-quality glass-
ceramics. This study investigates the effects of prenucleation
treatment on the kinetic behavior of a waste-derived glass-
ceramic synthesized from a mixture of municipal solid waste
incinerator fly ash and electric arc furnace dust. The crystal-
lization kinetic parameters of annealed and prenucleated
glass samples were determined by differential thermal analy-
sis under nonisothermal conditions. Prenucleation experi-
mental results revealed a temperature and time of maximum
nucleation of 700
C and 30 min, respectively. Crystallo-
graphic and microstructural analyses revealed that the main
crystalline phases were melilite and augite, with an equiaxed
grain morphology, and were embedded in the glassy matrix.
The activation energies (E
c
) for the crystallization of the
annealed and prenucleated glass samples, determined using
modified Kissinger and Ozawa equations, were in the ranges
of 367.4–395.2 and 199.8–214.6 kJ/mol, respectively. The
Avrami constant (n) was 1.8 for the annealed glass and 1.5
for the prenucleated glass. These results confirm a significant
difference between the E
c
values of the annealed glass and
prenucleated glass, suggesting that mixed ash-based glass is
suitable for use in the two-stage crystallization thermal treat-
ment adopted in this study.
2012 American Institute of Chemi-
cal Engineers Environ Prog, 00: 000–000, 2012
Keywords: nucleation, glass-ceramic, crystallization
kinetics, incinerator fly ash, differential thermal analysis
8
INTRODUCTION
Landfills are commonly used for disposing municipal solid
waste incinerator (MSWI) fly ash and electric arc furnace
(EAF) dust. From the viewpoint of life cycle analysis, the
recycling of MSWI fly ash and EAF dust for glass-ceramics
production and reused as raw materials. Increasing demand
for natural resources and the scarcity of acceptable solid
waste disposal sites are motivating many local governments
to consider resource recovery as an alternative and would
now seem to be economically feasible. In recent years, sub-
stantial effort has been made to elucidate the conversion of
slag-derived glass into useful glass-ceramic composites. This
conversion can be carried out by vitrification and recrystalli-
2012 American Institute of Chemical Engineers
Environmental Progress & Sustainable Energy (Vol.00, No.00) DOI 10.1002/ep
Month 2012
1
Therefore, the goal of this study is to evaluate the nuclea-
tion and crystallization behavior of a waste-derived glass-
ceramic synthesized by recycling MSWI fly ash and EAF dust.
The kinetics of the nucleation and growth of crystals in the
synthesizing process were analyzed as follows: (a) nuclea-
tion-rate curves were plotted to determine the optimal nucle-
ation conditions using the prenucleation approach and (b)
the parameters of the kinetics of crystallization for annealed
and prenucleated glass samples were determined by noniso-
thermal DTA using Ozawa and Matusita and Sakka methods
[38] to identify the activation energy of crystallization and the
crystallization mechanisms.
MATERIALS AND METHODS
MSWI Fly Ash, EAF Dust, and Waste Glass Preparation
The fly ash used in this study was collected from the
cyclone of a mass-burning incinerator located in the northern
part of Taiwan. The incinerator, capable of processing 1350
metric tons of local municipal solid waste per day, is
equipped with air pollution control devices consisting of a
cyclone, a semidry scrubber system, and a fabric bag-house
filter. EAF dust and waste glass were collected from a steel-
making factory in the middle of Taiwan and from a waste
resource factory located in the northern part of Taiwan. In
total, 200 kg of fly ash and 100 kg of EAF dust were obtained
from the incineration plant and steel-making factory, respec-
tively. The MSWI fly ash, EAF dust, and waste glass were ho-
mogenized, oven dried at 1058C for 24 h, and pulverized in a
ball mill until particles could pass through a #100 mesh sieve,
after which the chemical composition was characterized.
Figure 1. DTA thermogram of the annealed glass scanned at
a heating rate of 10
C/min. [Color figure can be viewed in the
online issue, which is available at wileyonlinelibrary.com.]
8
Differential thermal analysis: The thermal behavior of the
annealed glass was determined by DTA (Seiko SSC 5000)
using a bulk glass sample. The glass transition temperature
(T
g
) and crystallization temperature (T
p
)werethusdeter-
mined. Figure 1 shows a typical DTA thermogram of
annealed glass scanned at a heating rate of 108C/min. The
value of T
g
is 6408C, and a sharp exothermic peak covers
the temperature range of 750–8508C. The maximum crystal-
lization peak temperature was 8008C. According to the lit-
erature [33], the temperature at maximum nucleation rate is
typically between the glass transition temperature and the
crystallization onset temperature. Accordingly, the annealed
glass samples were prenucleated at various temperatures
(640–7508C) for various durations (10–60 min) before DTA
scanning at a fixed heating rate (258C/min). To determine
the optimal nucleation conditions for annealed glass, the
effects of the nucleating conditions on the exothermic peak
temperature and the exothermic peak area in the DTA runs
were also studied. The kinetic parameters of the crystalliza-
tion of annealed and prenucleated glass samples, including
the Avrami constant (n) and the activation energy, were
calculated from DTA data. In the nonisothermal DTA
experiments, the glass samples (35 mg) were heated at
rates of 10, 20, 25, and 308C/min and Al
2
O
3
served as the
reference material in the DTA apparatus. The exothermic
peak temperatures of crystallization from the DTA curves
were used to evaluate the kinetic parameters.
Crystal structural observation: A Hitachi S-3500N scanning
electron microscope (SEM) was used for SEM observation
and crystal structural determination.
Preparation of Slag from MSWI Fly Ash, EAF Dust, and
Waste Glass by Comelting Treatment
The initial parent glass was produced from a mixture of
60 wt % fly ash, 20 wt % EAF dust, and 20 wt % waste glass
cullet. Waste glass cullet was used to provide sufficient glass
network former (SiO
2
) to enhance the glass-forming capacity
of the mixture. One hundred grams of well-mixed sample
was melted in an alumina crucible at 14008C for 1 h in an
electrically heated furnace and then rapidly quenched in
water to form a granulated slag. The resulting slag was
crushed and remelted twice under the same conditions to
ensure good homogeneity. The refined melt was subse-
quently quenched by pouring onto a stainless steel plate at
room temperature to obtain waste-derived glass. To eliminate
the thermal residual stresses, the slag-derived glass samples
were annealed at 580
8
C for 1 h and then slowly cooled to
room temperature (25
C). These glass samples are hereafter
referred to as annealed glass.
8
Analyses
RESULTS AND DISCUSSION
This study presents the following chemical and physical
analyses of the ash, waste glass, and the glass-ceramic
samples:Chemical composition: X-ray fluorescence (XRF)
was performed with an automated RIX 2000 spectrometer.
The specimens were prepared for XRF analysis by mixing
0.4 g of the sample and 4 g of 100 Spectroflux at a dilu-
tion ratio of 1:10. Homogenized mixtures were placed in
Pt–Au crucibles before being treated for 1 h at 10008Cin
an electrical furnace. The homogeneous melted sample
was recast into glass beads 2 mm in thick and 32 mm in
diameter.
Mineralogy: XRD analysis was carried out using a Siemens
D-5000 X-ray diffractometer with CuKa radiation and 2y
scanning, ranging between 58 and 808 (2y). The XRD
scans were run in 0.058 steps, with 1-s counting time.
Characterization of MSWI Fly Ash, EAF Dust,
and Waste Glass
Table 1 presents the composition of the MSWI fly ash,
EAF dust, and waste glass. XRF analysis demonstrates that
the major components of the MSWI fly ash were CaO
(39.7%), Cl
2
(32.3%), and SiO
2
(6.9%). The main components
of the EAF dust were Fe
2
O
3
(32.3%), ZnO (30.1%), CaO
(8.7%), and SiO
2
(5.3%). The main components of the waste
glass were SiO
2
(67.4%), Na
2
O (11.1%), and CaO (7.6%).
Prenucleation of Annealed Glass
As reported by Marotta et al., the optimal nucleation tem-
perature and duration for producing glass-ceramics can be
2 Month 2012
Environmental Progress & Sustainable Energy (Vol.0000, No.0000) DOI 10.1002/ep
(1/T
p
2 1/T
p
0
) as a function of nucleation rate (I
v
) yields a
straight line. Therefore, as the number of nuclei increases, T
p
in DTA shifts to a lower temperature, and (1/T
p
2 1/T
p
0
)
increases.
Figure 2a shows the DTA thermograms of the annealed
glass after prenucleation at various temperatures for 30 min.
Based on these DTA results, the exothermic peak of the crys-
tallization annealed glass shifts to a significantly lower tem-
perature after prenucleation treatment. However, the exother-
mic DTA peaks gradually disappeared at prenucleation tem-
peratures exceeding 7208C. Prenucleation temperatures of
730 and 7508C did not produce an exothermic peak. These
prenucleation temperatures may have been too high, causing
the nuclei to begin growing into crystallites and leading to
complete crystallization of the samples before the DTA pro-
cess. The XRD analysis of both prenucleated samples demon-
strates the presence of crystalline species.
This study investigates the relationship between the pre-
nucleation temperature and the change in the peak area in
the DTA curves by considering the peak area ratio (A/A
0
).
This ratio is the volume fraction of crystals in the pre-
nucleated glass divided by that in the annealed glass, as
determined by DTA measurements. A and A
0
represent the
areas under the exothermic peaks of crystallization in the
prenucleated samples and the annealed sample, respectively.
If the ratio of these peak areas significantly exceeds unity for
a particular prenucleation temperature, then the glass sam-
ples can generate many nuclei. To determine the temperature
that yields the maximal nucleation rate, (1/T
P
2 1/T
p
0
) and
A/A
0
are plotted as functions of prenucleation temperature
(T
n
). Figure 2b shows that (1/T
p
2 1/T
p
0
) increased with pre-
nucleation temperature up to 7208C but fell abruptly to zero
beyond this temperature. Theoretically, the nucleation rate
should be maximal at 7208C, but the peak area ratio as a
function of prenucleation temperature was inconsistent with
this expectation. This finding is related to the fact that A/A
0
suddenly fell from 1.5 to less than 1 at a prenucleation tem-
perature of 7208C, suggesting that some crystallization had al-
ready occurred before reaching this temperature. Further-
more, A/A
0
reached its highest value at a prenucleation tem-
perature of 7008C, at which the value of (1/T
p
2 1/T
p
0
) was
high. Therefore, the experiments in this study used a temper-
ature of 700
determined using the DTA approach [32, 39, 40]. In the pre-
nucleation experiment in this study, the glass samples were
heated under various prenucleation conditions before DTA
scanning at a fixed heating rate. This study also examines the
effects of nucleating conditions on the exothermic peak tem-
perature and exothermic peak area during the DTA runs. The
curves of nucleation rate against temperature can be plotted
using the Marotta method [32, 39]:
R
1
þ constant
;
E
c
1
T
p
lnðIvÞ¼
T
p
(1)
where I
v
is the nucleation rate, E
c
is the activation energy for
crystallization, R is the gas constant, and T
p
0
and T
p
are the
exothermic peak temperatures of crystallization of the
annealed samples and prenucleated samples, respectively. In
the DTA analysis, the exothermic DTA peaks vary with the
prenucleation conditions. This is because the concentration
of nuclei in the glass depends on the thermal histories of the
prenucleation process. The Marotta equation reveals that,
for a given glass composition and heating rate, the curve of
Table 1. Chemical composition of MSWI fly ash, EAF dust,
and waste glass cullet.
Chemical
composition
(wt %)
MSWI
fly ash
Waste
glass
EAF dust
SiO
2
6.91
5.28
67.38
Al
2
O
3
2.81
0.99
1.63
Na
2
O
4.34
2.14
11.06
K
2
O
5.28
3.69
0.56
MgO
1.61
1.78
0.93
CaO
39.68
8.69
7.63
TiO
2
0.59
0.08
0.02
Fe
2
O
3
1.16
30.93
0.80
ZnO
0.96
30.07
0.1
MnO
0.08
1.99
—
NiO
0.03
0.05
—
P
2
O
5
0.91
0.23
0.03
SO
3
2.70
0.45
—
Cl
2
32.33
—
—
C to maximize the nucleation rate. Varying the
prenucleation time at this prenucleation temperature from 10
to 60 min shifted T
p
, as revealed by DTA analysis. As
8
LOI
7.69
7.36
—
Figure 2. (a) DTA thermograms of the annealed glasses after prenucleated at different temperatures for 30 min and (b) plots
of (1/T
P
2 1/T
P
0
) and A/A
0
versus prenucleation temperatures (T
n
).
Environmental Progress & Sustainable Energy (Vol.0000, No.0000) DOI 10.1002/ep
Month 2012
3
Figure 3. (a) DTA thermograms of the annealed glasses after heated at optimum prenucleation temperature (7008C) for 10–60
min and (b) plots of 1/T
P
and A/A
0
versus prenucleation time (t
n
).
Table 2. The glass transition, T
g
, and crystallization peak
temperature, T
p
, of both annealed and prenucleated glasses
at the different heating rates.
Annealed
glass
Prenucleated
glass
Crystalli-
zation
peak shift,
DT
p
*
Heating
rate, a
(8C/min)
T
g
(8C)
T
p
(8C)
T
g
(8C)
T
p
(8C)
10
640
797
657
766
31
20
651
824
661
799
25
25
656
845
665
805
40
30
660
860
667
814
46
*
T
p
(prenucleated glass)
2
T
p
(annealed glass).
Figure 4. DTA thermograms of the studied glass scanned at
heating rates of (1) 108C/min, (2) 208C/min, (3) 258C/min,
and (4) 308C/min. [Color figure can be viewed in the online
issue, which is available at wileyonlinelibrary.com.]
Table 2 shows the glass transition temperature, T
g
, and the
crystallization peak temperature, T
p
, of both annealed and
prenucleated glass at various heating rates. As displayed in
Figure 4, both T
g
and T
p
peaks shift to a higher temperature
as the heating rate increases, for all glass samples. This find-
ing is consistent with the literature [25, 26, 33–36]. At higher
heating rates, the glass crystallizes at a higher temperature.
This may be because slow heating rates allow sufficient time
for thermal soaking in the nucleation stage. Table 2 shows
that a large crystallization peak shift (DT
p
) occurs at all heat-
ing rates. The difference between the crystallization peak
temperatures of the annealed and prenucleated glass is sub-
stantial. However, prenucleation treatment shifts the exother-
mic peak of glass crystallization to a significantly lower tem-
perature, indicating that more stable nuclei were formed in
prenucleated glass than in annealed glass.
In nonisothermal methods, the Ozawa [30] and Matusita
and Sakka equations [38] are used to determine the kinetic
parameters of the crystallization of glass from DTA data. This
makes it possible to estimate important kinetic parameters,
including crystallization activation energy (E
c
) and those
associated with the mechanism of crystal growth, using the
aforementioned equations from the variation in the tempera-
ture of the exothermic DTA peak as a function of heating
rate. The classical Kissinger method has been extensively
described above, according to the Marotta method, a higher
concentration of nuclei (N) in the glass corresponds to a
higher 1/T
P
. Figures 3a and 3b demonstrate that the 1/T
P
val-
ues obtained from the DTA curves did not vary significantly
from 10 to 40 min. However, A/A
0
reached its maximum at
30 min, suggesting that
the optimal nucleation time for
annealed glass was 30 min.
Nonisothermal Crystallization Kinetics of Annealed
and Prenucleated Glass
This study uses nonisothermal DTA methods to investigate
the effects of prenucleation treatment on the crystallization
behavior of mixed ash-based glass. DTA experiments were
conducted on the annealed and prenucleated glass samples,
and the relationship between the heating rates and the peak
temperature in DTA was discussed. Figure 4 shows the DTA
thermograms of the studied glass scanned at heating rates of
(1) 108C/min, (2) 208C/min, (3) 258C/min, and (4) 308C/min.
4 Month 2012
Environmental Progress & Sustainable Energy (Vol.0000, No.0000) DOI 10.1002/ep
Figure 5. The Ozawa plots of ln(
2
ln(1
2
x))T versus ln
a
for annealed and prenucleated glass.
Figure 6. Crystallization activation energy plots of annealed and prenucleated glass obtained by modified Kissinger equation.
adopted to estimate the activation energy for crystallization
[28, 29]. A plot of ln(a/T
p
2
) as a function of 1/T
p
2
is linear,
with a gradient that yields an estimate of the activation
energy (E
c
). However, several studies [33–35, 38] have dem-
onstrated that the activation energy for crystallization
obtained using the Kissinger model is incorrect if numerous
nuclei form during the DTA scans. Therefore, it cannot be
directly applied to the crystallization of amorphous materials.
Two modified kinetic models (modified Kissinger and modi-
fied Ozawa) with the corrected activation energy for crystalli-
zation have been used, and these have been described by
Matusita and Sakka [38]. They suggested that the crystal
growth dimension, m, should be considered when applying
the Kissinger and Ozawa equations, and the activation
energy can be estimated using
ln
a
the dimensionality of crystal growth, and n is an Avrami con-
stant, which reflects the crystallization mechanism [41, 42]. In
special cases in which surface crystallization predominates,
n 5 m 5 1 and Eqs. 2 and 3 reduce to the original Kissinger
and Ozawa equations.
Ozawa developed a relatively simple method for elucidat-
ing the mechanism of crystallization from nonisothermal DTA
data [30]. Based on the crystallization exothermic peak in
DTA, the Avrami constant (n) can be calculated using Oza-
wa’s equation:
d ln
½
ln
ð
1
x
Þg
T
¼n
f
(4)
dðln aÞ
where x is the volume fraction that crystallized at a fixed tem-
perature T when heated at a heating rate of a. Accordingly, x
is the ratio of the fractional area at a specific temperature T to
the total area under a crystallization peak. Figure 5 plots
ln(2ln(1 2 x))
T
as a function of lna for annealed and pre-
nucleated glass. According to Ozawa’s analysis, a plot of
ln(2ln(1 2 x))
T
against lna yields a straight line. From the
slopes of these plots, an Avrami (n) value of 1.8 (close to 2) is
obtained for the annealed glass and 1.5 for the prenucleated
glass. According to Matusita and Komatsu’s classification [41],
these values indicate that bulk crystallization dominates in
¼
mE
c
n
T
p
RT
p
þconstant ðModifiedKissingermethodÞ
(2)
mE
c
RT
p
þconstant ðModifiedOzawamethodÞ
;
n
lnða
Þ¼
(3)
where a represents the heating rate, T
p
is the temperature at
which the crystallization peak is highest, E
c
is the activation
energy for crystallization, R is the gas constant, m represents
Environmental Progress & Sustainable Energy (Vol.0000, No.0000) DOI 10.1002/ep
Month 2012
5
[ Pobierz całość w formacie PDF ]