Construction and Building Materials 167 (2018) 44–54

Contents lists available at ScienceDirect

Construction and Building Materials

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

Heat-treated palm oil fuel ash as an effective supplementary

cementitious material originating from agriculture waste
Belal Alsubari a,b, Payam Shafigh b,c,⇑, Zainah Ibrahim a, Mohd Zamin Jumaat a
a
Department of Civil Engineering, Faculty of Engineering, University of Malaya, 50603 Kuala Lumpur, Malaysia
b
Department of Building Survey, Faculty of Built Environment, University of Malaya, 50603 Kuala Lumpur, Malaysia
c
Center for Building, Construction & Tropical Architecture (BuCTA), Faculty of Built Environment, University of Malaya, 50603 Kuala Lumpur, Malaysia

h i g h l i g h t s

 T-POFA was used up to 70% to replace cement in the paste and investigate its microstructure.
 Utilization of T-POFA in paste showed an improvement in the paste microstructure.
 The quantity of Ca(OH)2 was determined for both filler effect and pozzolanic reaction.

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

Article history: Palm oil fuel ash (POFA) is generated from the combustion of oil palm waste. This ash is considered a dis-
Received 27 March 2017 turbance to the environment since it is dumped without being profit return or recycled, thus causing
Received in revised form 3 January 2018 environmental problems and human health hazards. However, it has been proven that POFA is a poz-
Accepted 23 January 2018
zolanic material that can be utilized as cement replacement in concrete. If POFA is heated at high tem-
Available online 9 February 2018
perature, it exhibits better pozzolanic performance. In this study, the treatment processes as well as
microstructure and pozzolanic characteristics of POFA are considered. Ordinary Portland cement was
Keywords:
replaced with 0, 30, 50, and 70% treated POFA (T-POFA) in a paste. The hydration products of different
Palm oil fuel ash
Heat treatment
hardened pastes were checked by means of simultaneous thermal analysis (STA), X-ray diffraction
Microstructure (XRD), and field emission scanning electron microscopy (FESEM) in order to assess the effect of TPOFA
Pozzolanic reaction on the paste microstructure. According to the test results, the physical properties and chemical compo-
Cement replacement sition of T-POFA greatly enhanced due to heat treatment and the milling process. It was observed that the
calcium hydroxide content decreased in pastes containing T-POFA upon hardening, which is evidence of
its consumption by the pozzolanic reaction. The FESEM images displayed needle-like and reticular C–S–H
phases in the control paste, while floc-like and fibrous-like C–S–H phases well-connected to each other
were observed in the pastes containing T-POFA.
Ó 2018 Elsevier Ltd. All rights reserved.

1. Introduction During the early age of reaction, calcium hydroxide (Ca(OH)2) is

produced due to the reaction of tricalcium silicate (C3S) and dical-
The chemical reactions between cement and water from which cium silicate (C2S). The resulting calcium hydroxide in solid paste
hydration products form, involve different microstructure develop- phase is present mainly in the form of portlandite. It creates mas-
ment stages [1]. Over time, this leads to cement paste stiffening, sive, plate-like hexagonal crystals and forms about 20–25% of the
solidification and hardening [2] as shown in Fig. 1 [3]. solid paste volume [4]. In addition, ettringite is produced during
early hydration due to the reaction of tricalcium aluminate (C3A)
with gypsum in the presence of water [5]. More heat is produced
during C3S hydration, which contributes to strength at an early
Abbreviations: POFA, Palm oil fuel ash; G-POFA, Ground palm oil fuel ash; T-
POFA, Treated palm oil fuel ash; U-POFA, Ungrounded POFA. age (2–3 h to 14 days). However, C2S hydration occurs more slowly
⇑ Corresponding author at: Department of Building Survey, Faculty of Built and is responsible for the long-term strength of Portland cement
Environment, University of Malaya, 50603 Kuala Lumpur, Malaysia. concrete. Over a longer hydration period more C–S–H are pro-
E-mail addresses: [email protected] (B. Alsubari), [email protected] duced, comprising 50–60% of the volume of all solid phases [4].
(P. Shafigh).

https://doi.org/10.1016/j.conbuildmat.2018.01.134
0950-0618/Ó 2018 Elsevier Ltd. All rights reserved.
 B. Alsubari et al. / Construction and Building Materials 167 (2018) 44–54 45

Fig. 1. Different stages in the hydration process of cement [3].

Fig. 2. Original POFA.

Calcium silicate hydrates (C–S–H) are the main binding phases in (OH)2, it leads to more uniform, less permeable and more durable
all Portland cement-based structures [6]. concrete microstructure [13–15].
Because supplementary cementitious materials (SCMs) contain Palm oil fuel ash (POFA) is an agro-waste byproduct that is
high amounts of amorphous silica, they are utilized as mineral recently recognized as a pozzolanic material [16,17]. The Malay-
admixtures in pastes, mortars and concretes to improve their prop- sian Palm Oil Board (MPOB) estimated that the total amount of
erties, prolong the service life of structures and reduce cost and POFA generated annually is approximately 3 million tons [18].
CO2 emissions [7–9]. In cement blended with SCMs two main types POFA is dumped in landfills, which could cause environmental
of chemical reactions occur: cement hydration and the pozzolanic problems. Ground POFA (G-POFA) and treated POFA (T-POFA) are
reaction of the SCM [10]. The pozzolanic reaction occurs when the proven to contain large amounts of glassy phases [19,20]. Both
amorphous silica in SCM reacts chemically with Ca(OH)2 that is POFA types have been used to make blended cement pastes. Jatu-
produced from cement hydration earlier to form hydration prod- rapitakkul et al. [21] studied the filler effect and pozzolanic reac-
ucts, mainly C–S–H [11,12]. One of the most commonly utilized tion of ground river sand and G-POFA. The results revealed that
SCMSs in concrete is fly ash [7]. When fly ash reacts with Ca the compressive strength of mortar due to the filler effect of
46 B. Alsubari et al. / Construction and Building Materials 167 (2018) 44–54

Table 1
Mix proportions of pastes containing T-POFA.

Mix Water to binder ratio Cement (kg/m3) T-POFA (kg/m3) Sand (kg/m3)
Group No. Percentage replacement Code
1 0 Control 0.35 480 0 0
2 30 T-POFA30 0.35 336 144 0
Sand-30 0.35 336 0 144
3 50 T-POFA50 0.35 240 240 0
Sand-50 0.35 240 0 240
4 70 T-POFA70 0.35 144 336 0
Sand-70 0.35 144 0 336

100
OPC G-POFA T-POFA
90

Comulative percentage (%)

80
70
60
50
40
30
20
10
0
0.0 0.1 1.0 10.0 100.0 1000.0
Particle size (µm)

Fig. 4. Particle size distribution of OPC, G-POFA and T-POFA.

Fig. 3. Difference in color of G-POFA and T-POFA. (For interpretation of the
references to colour in this figure legend, the reader is referred to the web version of
this article.)
Table 2
Physical characteristics of OPC, G-POFA, and T-POFA.
ground river sand was almost constant during the 7–90 days per-
Properties OPC G-POFA T-POFA
iod. However, the pozzolanic reaction of G-POFA improved with
Color Grey Dark grey Grey
finer G-POFA particles and greater replacement levels and age of
Passed from sieve No. 325 (45 mm) (%) 92 96 100
mortar. Surface area, BET (m2/g) 3.05 4.94 7.4
Altwair et al. [20] investigated the effect of T-POFA on the Median particle size, d50 (mm) 21 16 13
hydration products of hardened pastes containing different T-
POFA/cement ratios ranging from 0.0 to 0.8. Thermogravimetric
analysis, X-ray diffraction (XRD), and scanning electron micro-
Table 3
scopy (SEM) were used to analyze the effect of T-POFA on the
Chemical composition of OPC, G-POFA, and T-POFA (%).
hydration products after 90 days of water curing. The results of
their study indicated that the Ca(OH)2 amount decreased when Oxide Composition OPC G-POFA T-POFA

the T-POFA content increased. However, the effect of pozzolanic Silica dioxide (SiO2) 17.60 59.17 69.02
reactivity was not considered at early ages in this study. Therefore, Aluminum trioxide (Al2O3) 4.02 3.73 3.9
Iron oxide (Fe2O3) 4.47 6.33 4.33
the aim of this study is to investigate pozzolanic reactivity of T-
Calcium oxide (CaO) 67.43 5.80 5.01
POFA and microstructure characteristics of paste containing T- Magnesium oxide (MgO) 1.33 4.87 5.18
POFA at the early and later ages. Sodium oxide (Na2O) 0.03 0.18 0.18
Potassium oxide (K2O) 0.39 8.25 6.9
Sulfur trioxide (SO3) 4.18 0.72 0.41
2. Materials and methods SiO2 + Al2O3 + Fe2O3 – 69.23 77.25
Loss on ignition (LOI) 2.4 16.1 1.8
2.1. Cement
face area [17,30]. The machine was timed to run for 16 h using
Ordinary Portland cement (OPC) with strength grade of 42.5 an electric motor at a speed of 33 rpm to grind about 6–8 kg. The
MPa was used in this research. high loss on ignition (LOI) content in the ground POFA (G-POFA)
was reduced by placing it in an electrical furnace at 600 °C for 2
2.2. Palm oil fuel ash collection and preparation h. The T-POFA was then ground again for 2 h to obtain fine particle
size.
Raw POFA was collected from a palm oil factory located in Physical and chemical composition tests were carried out to
Selangor state, Malaysia, and is shown in Fig. 2. The raw POFA evaluate the physical and microstructural characteristics of POFA
was placed in an electrical oven at 105 ± 5 °C for 24 h to remove before and after heat treatment. The particle size distribution,
moisture. Coarse residues were removed by sieving the dried POFA specific surface area and median particle size (d50) were measured
through a 300 mm sieve. The sieved POFA was ground in a Los using a particle size analyzer. In addition, the color change due to
Angeles machine to reduce the particle size and increase the sur- heat treatment was visually evaluated and reported. X-ray Fluores-
 B. Alsubari et al. / Construction and Building Materials 167 (2018) 44–54 47

Fig. 5. FESEM Images of: (a) U-POFA, (b) G-POFA and (c) T-POFA.

Fig. 6. X-ray Diffraction (XRD) analysis for U-POFA, G-POFA and T-POFA.

cence (XRF) was used to measure the chemical composition and W/B ratio was kept constant at 0.35. In this ratio the paste had
LOI of G-POFA and T-POFA. Field emission scanning electron micro- good workability. The pastes were cast into small molds of 50 
scopy (FESEM) was also utilized to assess the changes in POFA 50  50 mm3 and then demolded after 24 h. The tests were con-
morphology due to heat treatment. ducted after 7 and 28 days of curing to assess the hydration prod-
ucts at early and late ages. After removal from water, the samples
2.3. Microstructural characteristics were immediately immersed in acetone solution for 24 h to stop
the hydration process [31,32].
To evaluate the reactions between cement and T-POFA, three T- Testing was conducted on powder samples obtained by grind-
POFA-cement pastes and a control paste were prepared, as shown ing the hardened pastes using grinding bowl. The fine powder
in Table 1. Besides, three sand-cement pastes were prepared to was used in the simultaneous thermal analysis (STA) test with a
check the difference in Ca(OH)2 content when sand was used thermal analyzer machine (STA 6000 Perkin Elmer). The changes
instead of T-POFA at the same replacement level. A very fine min- in mass involved the dehydration of hydration products. STA was
ing sand with a particle size of 75% less than 45 mm was used. The performed on samples subjected to 30 °C to 1000 °C with a heating
48 B. Alsubari et al. / Construction and Building Materials 167 (2018) 44–54

110 0.5

0
100
-0.5

Derivative weight DW, (%/min)

90
Mass loss TGA, (%)

-1

80 -1.5

-2
70

TGA-control-7 TGA-control-28 -2.5

60 DW-control-7 DW-control-28
-3

50 -3.5
114
146
176
209
242
274
307
341
373
405
437
470
502
534
567
599
631
662
694
726
758
789
821
853
883
914
945
977
30
51
82

Temperature, (˚C)

Fig. 7. TGA and DTG curves of hydrated powder for control pastes after 7 days and 28 days of curing.

110 0.5

0
100
-0.5

Derivative weight DW, (%/min)

90
Mass loss TGA, (%)

-1

80 -1.5

-2
70
-2.5
60 TGA-T-POFA-30-7 TGA-T-POFA-30-28
DW-T-POFA-30-7 DW-T-POFA-30-28
-3

50 -3.5
121
154
187
222
258
293
328
363
398
432
467
502
536
571
605
639
673
708
742
775
809
844
877
910
943
977
30
53
86

Temperature, (˚C)

Fig. 8. TGA and DTG curves of hydrated powder for T-POFA30 (30% cement replacement) pastes after 7 days and 28 days of curing.

rate of 10 °C/min and 20 ml/min N2. STA is a method of measuring distributions for OPC, G-POFA and T-POFA are shown in Fig. 4. This
the mass change of a sample as a function of the rise in tempera- figure also indicates that G-POFA and T-POFA had median particle
ture [22,23]. The test conditions were perfectly identical for the sizes of about 16 mm and 13 mm, respectively, which is finer than
TGA and DSC signals (same atmosphere, gas flow rate, sample OPC with a median particle size of 21 mm. Reducing the median
vapor pressure, heating rate, thermal contact to the sample cru- particle size increased the rate of T-POFA dissolution and increased
cible and sensor, radiation effects, etc.) [24]. the pozzolanic activity, thus improving the physical and mechani-
XRD testing was conducted on paste powder. The specimens cal properties of the concrete at early and later ages [25,26].
were scanned by an X-ray diffractometer with a 2theta angle from The surface area of POFA was analyzed using a single point BET
5°–90°. To assess the sample morphology, a small piece of crushed (Brunauer, Emmett and Teller). The surface area of POFA increased
hardened paste with average diameter of nearly 1 cm was analyzed from 4.94 m2/g for G-POFA to 7.4 m2/g (increment of roughly 52%)
by field emission scanning electron microscope (FESEM). for T-POFA as shown in Table 2. The grinding by the Los Angeles
abrasion machine caused the surface area of POFA to increase.
The higher specific surface area of POFA indicates a potential
3. Results and discussion increase in pozzolanic reactions, which would enhance the con-
crete properties mostly in the early stages of hydration [26].
3.1. Physical properties and chemical composition of T-POFA XRF was performed to determine the chemical compositions of
OPC, G-POFA and T-POFA. The chemical compositions of the bin-
Following heat treatment, the color of G-POFA changed from ders are shown in Table 3, which indicates that silica dioxide
dark grey to grey as shown in Fig. 3. This can be attributed to the (SiO2), which is the main component of POFA, increased from
reduction of LOI through the heating treatment. The particle size 59.2% to 69%. The increment of silica content was attributed to
 B. Alsubari et al. / Construction and Building Materials 167 (2018) 44–54 49

110 0.5

0
100
-0.5

Derivative weight DW, (%/min)

90
Mass loss TGA, (%)

-1

80 -1.5

-2
70
-2.5
TGA-T-POFA-50-7 TGA-T-POFA-50-28
60 DW-T-POFA-50-7 DW-T-POFA-50-28
-3

50 -3.5
146

884
115

176
209
242
274
307
341
373
405
437
470
502
534
567
599
631
663
694
726
758
789
821
853

915
946
977
30
51
82

Temperature, (˚C)

Fig. 9. TGA and DTG curves of hydrated powder for T-POFA50 (50% cement replacement) pastes after 7 days and 28 days of curing.

110 0.5

0
100
-0.5

Derivative weight DW, (%/min)

90
Mass loss TGA, (%)

-1

80 -1.5

-2
70

TGA-T-POFA-70-7 TGA-T-POFA-70-28 -2.5

60 DW-T-POFA-70-7 DW-T-POFA-70-28
-3

50 -3.5
774
121
154
187
222
258
293
328
363
397
432
466
501
536
570
605
639
673
707
741

808
842
876
909
942
976
30
53
86

Temperature, (˚C)

Fig. 10. TGA and DTG curves of hydrated powder for T-POFA70 (70% cement replacement) pastes after 7 days and 28 days of curing.

Table 4
Summary of mass loss of calcium hydroxide according to STA testing at 7 and 28 days of curing.

Mix Mass loss of Ca(OH)2 by STA (% mass) Quantity of Ca(OH)2 over time
Group No. Percentage replacement level Code 7 day curing 28 day curing
1 0 Control 6.17 7.11 Increased
2 30 T-POFA30 4.81 4.32 Decreased
Sand-30 3.92 4.50 Increased
3 50 T-POFA50 2.78 1.82 Decreased
Sand-50 3.73 4.51 Increased
4 70 T-POFA70 2.35 1.56 Decreased
Sand-70 2.41 2.96 Increased

the removal of the loss of ignition (LOI) of G-POFA due to the heat could be concluded that the overall chemical properties of POFA
treatment. Hence, the sum of SiO2, Fe2O3, and Al2O3 increased from improved via heat treatment.
69% to 77%. In addition, one of the benefits of heat- treating POFA is Additionally, FESEM was carried out to check the changes in
the significant reduction in LOI content, here from 16.1% to 1.8%. It POFA morphology before heat treatment (G-POFA) and POFA after
50 B. Alsubari et al. / Construction and Building Materials 167 (2018) 44–54

110 0.5

0
100
-0.5

Derivative weight DW, (%/min)

90
Mass loss TGA, (%)

-1

80 -1.5

-2
70

TGA-Sand-30-7 TGA-Sand-30-28
-2.5
60 DW-Sand-30-7 DW-Sand-30-28
-3

50 -3.5
121
153
186
221
256
291
327
363
398
432
467
502
536
571
606
640
674
708
742
776
810
844
877
911
944
977
30
53
86

Temperature, (˚C)

Fig. 11. TGA and DTG curves of hydrated powder for sand-30 (30% cement replacement) pastes after 7 days and 28 days of curing.

110 0.5

0
100

Derivative weight DW, (%/min)

-0.5
90
Mass loss TGA, (%)

-1

80 -1.5

-2
70
TGA-Sand-50-7 TGA-Sand-50-28 -2.5
60 DW-Sand-50-7 DW-Sand-50-28
-3

50 -3.5
214

585
118
150
181

248
282
317
352
385
418
452
485
519
552

619
651
684
717
750
782
815
848
880
912
944
977
30
53
85

Temperature, (˚C)

Fig. 12. TGA and DTG curves of hydrated powder for sand-50 (50% cement replacement) pastes after 7 days and 28 days of curing.

heat treatment (T-POFA) as shown in Fig. 5. The FESEM results of ful for evaluating the mass loss steps accurately. As seen in these
G-POFA and T-POFA at the same magnification demonstrate that figures, there are three main endothermic zones. The first weight
the particles were angular and irregular in shape and had a porous loss occurred in the range from 30 to 250 °C and was attributed
texture. In addition, it is observed there was no agglomeration of to the the evaporation of moisture, calcium silicate hydrate (C–
particles after heat treatment. The G-POFA and T-POFA samples S–H) and ettringite.
also underwent XRD analysis to check their silica structure. Fig. 6 The second zone, ranging from approximately 300 °C to 550 °C,
illustrates that major phase of G-POFA and T-POFA was SiO2 (Q = was attributed to the dehydration of calcium hydroxide Ca(OH)2.
Quartz), while C = Cristobalite and potassium aluminum phos- The Ca(OH)2 losses for all samples are summarized in Table 4. This
phate (K3Al2(PO4)3) were minor phases. table indicates that the Ca(OH)2 loss from pastes containing T-
POFA decreased with increasing the replacement level. Also, the
3.2. T-POFA-cement hydration products and microstructure mass changes decreased with prolonged curing time. At 28 days
of curing, the control paste (control-28) had the highest Ca(OH)2
3.2.1. Simultaneous thermal analysis (STA) mass loss of about 7.1%, whereas the paste containing 70% T-
The STA technique is useful for estimating the changes in POFA (T-POFA70-28) had the lowest mass loss of about 1.6%. More-
cement-based material products to calculate their behavior upon over, the DTG curve peak of Ca(OH)2 for the pastes containing T-
exposure to fire. Figs. 7–10 display typical curves for STA at 7 days POFA decreased with higher T-POFA replacement levels and also
and 28 days with the respective mass losses. The curves consist of with curing time. The Ca(OH)2 reduction in the cement paste con-
mass loss (TGA) and derivative thermogravimetric (DTG). TGA is taining T-POFA indicates its consumption by the pozzolanic reac-
used to measures weight changes in relation to temperature tion to form C–S–H. Moreover, the reduction is due to cement
changes, whereas DTG highlights the point at which the weight dilution, hence reducing the formation of Ca(OH)2 from cement
loss is most apparent, provides the decomposition rate and is help- hydration.
 B. Alsubari et al. / Construction and Building Materials 167 (2018) 44–54 51

110 0.5

0
100
-0.5

Derivative weight DW, (%/min)

90
-1
Mass loss TGA, (%)

80 -1.5

-2
70
-2.5
60 TGA-Sand-70-7 TGA-Sand-70-28
DW-Sand-70-7 DW-Sand-70-28 -3

50 352 -3.5

538
129
164
200
237
275
314

389
426
464
501

575
612
649
685
722
758
795
831
867
903
939
975
30
56
92

Temperature, (˚C)

Fig. 13. TGA and DTG curves of hydrated powder for sand-70 (70% cement replacement) pastes after 7 days and 28 days of curing.

Fig. 14. XRD patterns of hydrated powder pastes containing T-POFA at 7 days.

The third mass loss occurred at approximately 600 °C and that the Ca(OH)2 peaks for blended cement are lower than for the
beyond, indicating the complete decomposition of C–S–H and cal- control mortar. They reasoned that it was due to the diluting effect
cium carbonate (CaCO3). It should be noted that the TGA and DTG of cement and pozzolanic reaction of the cementitious materials.
curves for pastes with T-POFA, especially at high replacement
levels (50% and 70%), exhibited more stable behavior during the
temperature increase compared to the control cement paste spec- 3.2.2. Determining the filler and pozzolanic effects
imens. This phenomenon was most likely due to the high amount In order to determine the filler and pozzolanic effects of T-POFA
of high-density C–S–H in these specimens, which were not affected on paste, the Ca(OH)2 quantity was obtained through STA testing of
by high temperature exposure [27]. The results were in agreement specimens containing T-POFA and sand. The cement and water
with previous findings reported on the effect of POFA fineness on contents were kept the same in both specimen types as can be seen
blended cement paste microstructure [28]. Yen et al. [29] used high in Table 1. Figs. 11–13 show three main endothermic zones in the
volumes (50% and 70%) of ternary blended sewage sludge ash pastes containing sand and cement. The first weight loss occurred
(SSA), coal combustion fly ash (FA), and granulated blast furnace between 30 and 250 °C and was attributed to the the evaporation
slag (GGBFS) as cement replacements in mortar. The authors found of moisture, calcium silicate hydrate (C–S–H) and ettringite. The
52 B. Alsubari et al. / Construction and Building Materials 167 (2018) 44–54

Fig. 15. XRD patterns of hydrated powder pastes containing T-POFA at 28 days.

Fig. 16. FESEM of control paste fracture surface at 7 and 28 days.

Fig. 17. FESEM of fracture surface of paste containing 30% T-POFA at 7 and 28 days.
 B. Alsubari et al. / Construction and Building Materials 167 (2018) 44–54 53

Fig. 18. FESEM of fracture surface of paste containing 50% T-POFA at 7 and 28 days.

Fig. 19. FESEM of fracture surface of paste containing 70% T-POFA at 7 and 28 days.

second weight loss occurred at 350 °C to 550 °C and the third took gite formed in the early stages of hydration. The quartz detected in
place at temperatures above 600 °C. the pastes containing T-POFA did not represent one of the hydra-
The STA results clearly signify that the Ca(OH)2 mass increased tion products. It was potentially a result of the high silica (SiO2)
with increasing curing time in paste containing sand, whereas the content in T-POFA.
Ca(OH)2 decreased in paste with T-POFA, as summarized in Table 4. As seen in Figs. 14 and 15, the Ca(OH)2 peak values are lower for
According to this table, the Ca(OH)2 mass differed when sand and pastes with T-POFA over time. An explanation is that during early
T-POFA were used at the same cement substitution levels. For curing (at 7 days), the pozzolanic reaction of T-POFA was relatively
example, at 50% cement replacement, the mass of Ca(OH)2 in weaker than the hydration reaction of cement. However, at 28 days
sand-50 increased from 3.7% at 7 days to 4.5% at 28 days, while of curing, the reduced Ca(OH)2 in the paste indicates it was con-
for T-POFA-50 the mass loss decreased from 2.8% at 7 days to sumed in the pozzolanic reaction. It should also be noted from
1.8% at 28 days. The difference between replacing cement with Figs. 14 and 15 that the amount of non-hydrated C3S decreased
pozzolanic material and replacing it with filler that had no poz- with increasing curing time. The XRD patterns of the pastes with
zolanic reactivity is evident. The increasing Ca(OH)2 amount in and without T-POFA confirm the STA results of Ca(OH)2 gradually
the pastes with sand with prolonged curing time was attributed decreasing with increasing T-POFA content.
to the continuation of the cement hydration reaction while sand
only acted as a filler. However, in the cement pastes blended with
3.4. Field emission scanning electron microscopy (FESEM)
T-POFA, the decreasing Ca(OH)2 with increasing curing age indi-
cates that it was consumed by the pozzolanic reaction of T-POFA.
FESEM was used to investigate the fractured surface
microstructure for the control cement paste and pastes containing
3.3. X-ray diffraction (XRD) 30%, 50% and 70% T-POFA as cement replacement at 7 and 28 days
of water curing. As can be seen in Fig. 16, the control paste after 7
The XRD patterns of hydrated pastes with and without T-POFA days was characterized by early C–S–H phases. In this paste,
at 7 and 28 days of curing are presented in Figs. 14 and 15. The XRD hexagonal plate-like calcium hydroxide (portlandite) formed
patterns show that a remarkable amount of calcium hydroxide inside the pores and the matrix. The formation of these phases
(portlandite) formed early on in curing, especially for the control was a result of the hydration of tricalcium (C3S) and dicalcium
and paste containing 30% T-POFA (T-POFA30-7). Moreover, ettrin- (C2S) silicates, whereas ettringite formed due to the reaction of tri-
54 B. Alsubari et al. / Construction and Building Materials 167 (2018) 44–54

calcium aluminate (C3A) with gypsum. Needle-like ettringite con- [4] P. Mehta, P.J.K. Mehta, P.J. Monteiro, Concr. Microstruct. Proper. Mater. (2006).
[5] P.K. Mehta, Concr. Microstruct. Proper. Mater. (1986).
tributed to the stiffening of the paste and some of the early-age
[6] L. Skinner, S. Chae, C. Benmore, H. Wenk, P. Monteiro, Nanostructure of calcium
strength. By continuing curing to 28 days, the matrix became den- silicate hydrates in cements, Phys. Rev. Lett. 104 (19) (2010) 195502.
ser and the calcium hydroxide was embedded inside the paste [7] V. Malhotra, High-performance high-volume fly ash concrete, Concr. Int. 24 (7)
matrix. (2002).
[8] H. Mohammed Owaid, R.B. Hamid, M.R. Taha, A review of sustainable
When T-POFA was used, well-connected floc-like and fibrous- supplementary cementitious materials as an alternative to all-portland
like C–S–H phases formed, filling the voids in the paste, especially cement mortar and concrete, Aust. J. Basic Appl. Sci. 6 (9) (2012).
later on as shown in Figs. 17–19. It is also noted that with the [9] B. Lothenbach, K. Scrivener, R. Hooton, Supplementary cementitious materials,
Cem. Concr. Res. 41 (12) (2011) 1244–1256.
inclusion of T-POFA, the formation of portlandite diminished and [10] A. Elahi, P. Basheer, S. Nanukuttan, Q. Khan, Mechanical and durability
the remaining ettringite was embedded within the matrix. With properties of high performance concretes containing supplementary
increasing the amount of T-POFA and curing time, the matrix cementitious materials, Const. Build. Mater. 24 (3) (2010) 292–299.
[11] V. Papadakis, S. Antiohos, S. Tsimas, Supplementary cementing materials in
became denser, the fractured surface was almost completely cov- concrete: Part II: a fundamental estimation of the efficiency factor, Cem. Concr.
ered by C–S–H and the voids were filled. The calcium hydroxide Res. 32 (10) (2002) 1533–1538.
reduction and enhanced microstructure of pastes containing T- [12] H. Toutanji, N. Delatte, S. Aggoun, R. Duval, A. Danson, Effect of supplementary
cementitious materials on the compressive strength and durability of short-
POFA occurred due to the consumption of Ca(OH)2, which reacted term cured concrete, Cem. Concr. Res. 34 (2) (2004) 311–319.
with the silica content of T-POFA through pozzolanic reaction to [13] P. Chindaprasirt, S. Homwuttiwong, C. Jaturapitakkul, Strength and water
form C–S–H. In addition, some T-POFA particles appear unreacted permeability of concrete containing palm oil fuel ash and rice husk–bark ash,
Construct. Build. Mater. 21 (7) (2007) 1492–1499.
or partially reacted, meaning that the pozzolanic reaction is ongo-
[14] A. Neville, Chloride attack of reinforced concrete: an overview, Mater. Struct.
ing and the remaining particles that had no role in this reaction 28 (2) (1995) 63.
acted as micro-filler, filling the gaps and enhancing the paste [15] P.E. Stutzman, L. Centeno, Compositional Analysis of Beneficiated fly Ashes,
structure. National Institute of Standards and Technology, 1995.
[16] V. Sata, C. Jaturapitakkul, K. Kiattikomol, Influence of pozzolan from various
by-product materials on mechanical properties of high-strength concrete,
4. Conclusions Construct. Build. Mater. 21 (7) (2007) 1589–1598.
[17] M. Safiuddin, M. Abdus Salam, M.Z. Jumaat, Utilization of palm oil fuel ash in
concrete: a review, J. Civ. Eng. Manage. 17 (2) (2011) 234–247.
Based on the results obtained from this study, the physical [18] C. Chandara, K.A.M. Azizli, Z.A. Ahmad, S.F.S. Hashim, E. Sakai, Heat of
properties and chemical composition of POFA were significantly hydration of blended cement containing treated ground palm oil fuel ash,
Construct. Build. Mater. 27 (1) (2012) 78–81.
improved via heat treatment and grinding. Substituting of OPC
[19] C. Chandara, E. Sakai, K.A.M. Azizli, Z.A. Ahmad, S.F.S. Hashim, The effect of
with T-POFA influenced the paste microstructure and pore solu- unburned carbon in palm oil fuel ash on fluidity of cement pastes containing
tion. As T-POFA content increased from zero to 70%, the calcium superplasticizer, Construct. Build. Mater. 24 (9) (2010) 1590–1593.
[20] N.M. Altwair, M.A.M. Johari, S.F.S. Hashim, Strength activity index and
hydroxide content was reduced. The decreasing calcium hydroxide
microstructural characteristics of treated palm oil fuel ash, Int. J. Civ.
content in pastes containing T-POFA upon hardening is evidence of Environ. Eng. 11 (5) (2011) 85–92.
its consumption by pozzolanic reaction. Microstructural analysis [21] C. Jaturapitakkul, J. Tangpagasit, S. Songmue, K. Kiattikomol, Filler effect and
by FESEM showed needle-like and reticular C–S–H phases in the pozzolanic reaction of ground palm oil fuel ash, Const. Build. Mater. 25 (11)
(2011) 4287–4293.
control paste, while floc-like and fibrous-like C–S–H phases were [22] A. Trník, L. Scheinherrová, T. Kulovaná, R. Černý, Simultaneous differential
found in the pastes containing T-POFA, showing they were well- scanning calorimetry and thermogravimetric analysis of portland cement as a
interconnected. These phases had a connecting role between function of age, Int. J. Thermophys. 37 (1) (2016) 1–9.
[23] W. Wongkeo, P. Thongsanitgarn, A. Ngamjarurojana, A. Chaipanich,
hydrated and un-hydrated compounds by growing from the parti- Compressive strength and chloride resistance of self-compacting concrete
cle surface towards the adjacent matrix, which led to paste containing high level fly ash and silica fume, Mater. Des. 64 (2014) 261–269.
microstructure enhancement. Moreover, even after 28 days the [24] J. Van Humbeeck, Simultaneous thermal analysis, Handb. Therm. Anal.
Calorim. 1 (1998) 497–508.
blended pastes had some unreacted T-POFA particles, indicating [25] E.C. Arvaniti, M.C. Juenger, S.A. Bernal, J. Duchesne, L. Courard, S. Leroy, et al.,
an existing potential and ongoing pozzolanic reaction in this sup- Physical characterization methods for supplementary cementitious materials,
plementary cementitious material. Mater. Struct. 48 (11) (2015) 3675–3686.
[26] I. Celik, The effects of particle size distribution and surface area upon cement
strength development, Powder Technol. 188 (3) (2009) 272–276.
Acknowledgement [27] R.K. Ibrahim, R. Hamid, M.R. Taha, Fire resistance of high-volume fly ash
mortars with nanosilica addition, Const. Build. Mater. 36 (2012) 779–786.
[28] W. Kroehong, T. Sinsiri, C. Jaturapitakkul, P. Chindaprasirt, Effect of palm oil
The authors gratefully acknowledge the financial support from
fuel ash fineness on the microstructure of blended cement paste, Const. Build.
the University of Malaya Postgraduate Research Grant (grant no. Mater. 25 (11) (2011) 4095–4104.
PG170-2015B) and Fundamental Research Grant Scheme (FRGS) [29] C.L. Yen, D.H. Tseng, Y.Z. Wu, Properties of cement mortar produced from
(grant no. FP004-2014B). mixed waste materials with pozzolanic characteristics, Environ. Eng. Sci. 29 (7)
(2012) 638–645.
[30] A.A. Awal, M.W. Hussin, The effectiveness of palm oil fuel ash in preventing
References expansion due to alkali-silica reaction, Cem. Concr. Compos. 19 (4) (1997)
367–372.
[1] H. Taylor, Cem. Chem. (1997). [31] E.F. Irassar, V.L. Bonavetti, M. Gonzalez, Microstructural study of sulfate attack
[2] J.W. Bullard, H.M. Jennings, R.A. Livingston, A. Nonat, G.W. Scherer, J.S. on ordinary and limestone Portland cements at ambient temperature, Cem.
Schweitzer, et al., Mechanisms of cement hydration, Cem. Concr. Res. 41 (12) Concr. Res. 33 (1) (2003) 31–41.
(2011) 1208–1223. [32] Y. Liao, X. Wei, G. Li, Early hydration of calcium sulfoaluminate cement
[3] W. Franus, R. Panek, M. Wdowin, SEM investigation of microstructures in through electrical resistivity measurement and microstructure investigations,
hydration products of portland cement, 2nd International Multidisciplinary Construct. Build. Mater. 25 (4) (2011) 1572–1579.
Microscopy and Microanalysis Congress, Springer, 2015. pp. 105–112.

View publication stats