PROPERTIES OF LIGHTWEIGHT CONCRETE MANUFACTURED WITH FLY ASH, FURNACE BOTTOM ASH, AND LYTAG

77
PROPERTIES OF LIGHTWEIGHT CONCRETE
MANUFACTURED WITH FLY ASH, FURNACE
BOTTOM ASH, AND LYTAG
Yun Bai, Ratiyah Ibrahim, and P.A. Muhammed Basheer
Queen’s University, Belfast, U.K.
Abstract
Fly ash (FA), furnace bottom ash (FBA) and Lytag (LG) were used in the current
study to replace ordinary portland cement (OPC), natural sand (NS) and coarse
aggregate (CA), respectively, and thereby to manufacture lightweight concrete
(LWC). Two control mixes containing no replacement materials were designed with
a 28-day compressive strength of 20 N/mm
2
 and 40 N/mm
2
. For each compressive
strength, three different mixes, viz. (a) 100%OPC+100%NS + 100%CA, (b)
100%OPC + 100%FBA + 100%LG and (c) 70%OPC + 30%FA + 100%FBA +
100%LG, were manufactured with slump in the range of 30 ~ 60 mm. The density,
compressive strength, pull-off surface tensile strength, air permeability, sorptivity
and porosity of the concretes were investigated.
The results indicated that it is possible to manufacture lightweight concrete with
density in the range of 1560-1960 kg/m
3
 and 28-day compressive strength in the
range of 20-40 N/mm
2
 with various waste materials from thermal power plants.
However, the introduction of FBA into concrete would cause detrimental effect on
the permeation properties of concrete. With part of OPC replaced with FA, the
strength decreased, but the permeability of the resulting concrete improved.  
1.  Introduction
Lightweight concrete (LWC) has been successfully used since the ancient Roman
times and it has gained its popularity due to its lower density and superior thermal
insulation properties [1]. Compared with normal weight concrete (NWC), LWC can
significantly reduce the dead load of structural elements, which makes it especially
attractive in multi-storey buildings. However, most studies on LWC concern “semilightweight” concretes, i.e. concrete made with lightweight coarse aggregate and
natural sand. Although commercially available lightweight fine aggregate has been
used in investigations in place of natural sand to manufacture the “total-lightweight” 78  International Workshop on Sustainable Development and Concrete Technology
concrete [2, 3], more environmental and economical benefits can be achieved if
waste materials can be used to replace the fine lightweight aggregate.
Lytag is one of the most commonly used lightweight aggregates, which is
manufactured by pyro-processing fly ash (FA), while FA and furnace bottom ash
(FBA) are two waste materials from coal-fired thermal power plants. They are,
respectively, lighter than traditional coarse aggregate, OPC and natural sand. The
previous investigations carried out by the authors on using FBA from a thermal
power plant in Northern Ireland as a sand replacement material indicated that FBA
could be a potential fine aggregate in NWC for certain applications [4, 5]. However,
the application of FBA in structural LWC is not well defined. Therefore, the current
study investigates the possibility of manufacturing structural LWC with FA, FBA
and Lytag
2. Experimental Program
2.1  Materials
The cement used was the Class 42.5N portland cement supplied by Blue Circle, U.K.,
complying with BS 12: 1991 [6].
For the control mixes, the coarse aggregate used was 10 mm crushed basalt and the
fine aggregate used was medium graded natural sand complying with BS 882: 1992
[7]. Both materials are from the local sources in Northern Ireland. They were oven
dried at 40
o
C for 24 hours and cooled to 20
o
C before using in the manufacture of
concrete. The FA and FBA used were supplied by Kilroot Power Station in Northern
Ireland, U.K. The FBA was dried firstly in an oven at 105
o
C for 24 hours and then
allowed to cool for 24 hours at 20
o
C. The FBA that passed 5 mm sieve (hereafter
FBA sand) was used to replace natural sand. The Lytag used was with a size of 8
mm and was supplied by Finlay Concrete Products, Northern Ireland, U.K. It was
also oven dried at 40
o
C for 24 hours and cooled to 20
o
C before casting. Table 1
reports the chemical compositions of OPC, FA, FBA and Lytag. The specific gravity
and 1-hour water absorption of basalt, natural sand, FBA sand and Lytag are reported
in Table 2. Fig. 1 presents the particle size distribution of basalt, natural sand, FBA
sand and Lytag.
2.2 Mix proportions
Two control mixes containing OPC, basalt and natural sand were designed for a 28-
day compressive strength of 20 N/mm
2
 (Series M) and 40 N/mm
2
 (Series H)
respectively, for a slump in the range of 30-60 mm. For each control mix, 30% of
OPC, 100% of natural sand, and 100% of basalt were then replaced with FA, FBA,
and Lytag, respectively. The binder content (OPC or OPC + FA) was kept the same
as that of the control mix for each series when the natural sand and basalt were
replaced with FBA and Lytag, respectively.   Yun Bai, Ratiyah Ibrahim, and P.A. Muhammed Basheer 79
Table 1: Chemical composition of cement, PFA, FBA, and Lytag
Oxide
composition (%)
OPC FA FBA Lytag
SiO2 20.6 59.01 61.78 53.19
Al2O3 5.7 22.8 17.8 26.3
Fe2O3 2.9 8.8 6.97 10.26
CaO 63.6 2.38 3.19 2.02
MgO 1.8 1.39 1.34 1.45
Na2O 0.12 0.74 0.95 0.96
K2O 0.75 2.8 2 3.99
SO3 3.2 0.27 0.79 -
Cl 0.01 0.01 - -
LOI 1.5 6.7 3.61 4
Table 2: Property of aggregates
Property Basalt Natural sand FBA sand Lytag
Specific gravity (S.S.D.) 2.91 2.66 1.58 1.52
1-hour water absorption (%) 1.1 1.1 32.2 12.31
0
2 0
4 0
6 0
8 0
10 0
0 .0 1  0 .1  1  10  10 0
Nom i n a l   ape rtu re   s i z e   o f  te s t  s i e ve   (m m )
Cumula tive percentage
passing (%)
FBA Sand Natural Sand Lytag Basalt
Fig. 1: Particle distribution of FBA sand, natural sand, Lytag, and basalt
For each series, three different mixes were studied. Mix 1:
100%OPC+100%NS+100%CA (control). Mix 2: 100%OPC+100%FBA+100%LG.
Mix 3: 70%OPC+30%FA+100%FBA+100%LG. The water content (and therefore
W/C) of Mix 2 and Mix 3 was adjusted by carrying out trials so that the workability
measured in terms of slump was in the range of 30-60mm. The volume ratio between
the fine aggregate and the coarse aggregate for each test series was kept the same as 80  International Workshop on Sustainable Development and Concrete Technology
that obtained for the respective control mix. The resulting mix proportions, which
were used in this investigation, are reported in Table 3.
Table 3: Mix proportions (kg/m
3
) and properties of fresh concretes
M ix  No  W/C  Ceme nt  FA
Fre e
Wate r
Sand FBA Basalt Lytag
Me asure d
Slump
(mm)
Me asure d
Air Content
(%)
M1 0.65 330 - 215 820 - 1040 - 27 2
M2 0.4 330 - 132 - 552 - 616 51 5
M3 0.32 231 99 106 - 562 - 627 43 5
H1 0.47 460 - 215 715 - 1025 - 50 1.2
H2 0.32 460 - 147 - 477 - 602 30 5
H3 0.29 322 138 133 - 473 - 599 34 5
2.3  Batching and mixing
For each mix, the required quantities of the constituents were batched by weight. The
water required for 1-hour absorption by the aggregates (basalt, natural sand, FBA
sand and Lytag) was added to the mix water in addition to the free water shown in
Table 3. Different mixing procedures were used for NWC and LWC, which are
described below.
Mixing procedure for NWC (control): The manufacturing of NWC was carried out
based on reference 8. Approximately half the basalt, all the natural sand and the
remaining basalt were added, in this order, evenly into the pan. The aggregates were
mixed for 30 seconds. The mixing was continued and about half the mixing water
(i.e. free water as shown in Table 3 plus that required for 1 hour water absorption)
was added during the next 15 seconds. After mixing for a total of 3 minutes, the
mixer was stopped and the contents were left covered for 15 minutes.  The cement
was then added evenly over the aggregate. The mixer was started and the mixing was
continued for 30 seconds. The mixer was then stopped and any material adhering to
the mixer blades were cleaned off into the pan. Without delay, the mixing was
recommenced and the remaining mixing water was added over the next 30 seconds.
The mixing was continued for 3 minutes after all the materials were added.
Mixing procedure for LWC: The procedure given in the Lytag Information
Document [9] was used to modify the manufacturing procedure for the LWC. About
half the mixing water (free water as shown in Table 3 plus that required for 1 hour
water absorption) was added. Then all the Lytag and all the FBA were added in this
order, evenly into the pan [9] and mixed for 3 minutes. The mixer was stopped and
left covered for 15 minutes. Thereafter, the procedure was the same as that for the
NWC. Yun Bai, Ratiyah Ibrahim, and P.A. Muhammed Basheer 81
2.4  Specimen preparation and curing
For each mix, nine 100-mm size cubes were cast to determine the compressive
strength at the age of 3, 7, and 28 days. At 28 days, the same cubes used for
compressive strength were also used to  test the density at saturated-surface dried
(SSD) condition. Three 250x250x110-mm slabs were cast to investigate pull-off
tensile strength, permeation properties and porosity of the concrete at the age of 28
days.
All specimens were cast in two layers and compacted on a vibrating table until air
bubbles appearing on the surface stopped.  They were left in the mould in the
laboratory at 20(±1)
o
C for one day and then removed from the moulds. After that,
they were cured in water at 20(±1)
o
C for two days, and then wrapped in polythene
sheet and left in the laboratory at 20(±1)
o
C until they were tested. (The three-day
specimens were tested immediately after removing from the water bath, instead of
wrapping in polythene sheet.)
2.5  Details of tests
For each mix, the air content and workability (in terms of slump) of the fresh
concrete were measured. The air content was measured by following a procedure
given in BS 1881: part 106: 1983 [10]. The slump test was carried out in accordance
with BS 1881: Part 102: 1983 [11].
At the age of 3, 7, and 28 days, the compressive strength was determined by crushing
three 100-mm cubes in accordance with BS 1881:Part 116: 1983 [12] and an average
of the three values was obtained. Prior to the compressive strength test at the age of
28 days, the cubes were used to test the SSD density by following BS 1881: Part 114:
1983 [13].
At the age of 28 days, the slabs were dried at 40(±1)
o
C and 22(±1)% Relative
Humidity (RH) in a drying cabinet for two weeks and then cooled to room
temperature 20(±1)
o
C for one day. The air permeability and water absorption
(sorptivity) were tested on three slabs per mix by using the “Autoclam Permeability
System” [14] on the mould finished face and average values of both the air
permeability and the sorptivity were calculated. The surface tensile strength was
measured by carrying out the Limpet pull-off test [15] at two locations on the mould
finished surfaces of the three slabs immediately after the permeation test. All the six
results were averaged and reported. After the pull-off test, one Φ75-mm core was
taken from each of the slab and the water absorption test was carried out by
following BS 1881: Part 122: 1983 [16]. The porosity of the concretes was then
calculated based on the volume of the voids occupied by the absorbed water.  82  International Workshop on Sustainable Development and Concrete Technology
3. Results and Discussion
3.1  Properties of fresh concrete
Fig. 2 shows the free water content for different mixes. It can be seen that when the
FBA sand and Lytag were used to replace natural sand and basalt, respectively, the
water demand of the concrete decreased. This is attributable to the spherical/round
particle shape of both FBA sand and Lytag [4, 17], which, compared to the angular
particles of sand and basalt, have a “ball-bearing effect” and thus reduce the water
demand of the fresh concrete. When 30% of the OPC was replaced with FA in both
series, there was also a water reduction compared to the mix 2. This is again
attributable to the “ball-bearing effect” of the FA particles. Therefore, it can be seen
from the above results that when FA, FBA and Lytag were used to manufacture
lightweight concrete, the water demand of the concrete decreased.
0
50
100
150
200
250
300
M H
Series
Free Water (kg/m3)
100%OPC+100%NS+100%CA
100%OPC+100%FBA+100%LG
70%OPC+30%FA+100%FBA+100%LG
Fig. 2: Free water content of NWC and LWC
3.2  Density
Table 4 reports the density of hardened  concrete at saturated-surface dried (SSD)
condition measured at 28 days. It can be seen that when natural sand and basalt were
replaced with FBA sand and Lytag respectively, there was a significant reduction in
the density of hardened concrete for both series. This suggests that the low density of
both FBA sand and Lytag is beneficial to produce LWC. When FA was used in mix
3 to replace 30% of the OPC, the density was further reduced. This, again, is due to
the lower density of FA compared to that of OPC. Thus, it can be concluded that the
low density of FA, FBA and Lytag is a benefit for manufacturing lightweight
structural concretes. In the current study, the SSD density in the range of 1560-1960
kg/m
3
was achieved.
  Yun Bai, Ratiyah Ibrahim, and P.A. Muhammed Basheer 83
Table 4: Density (kg/m
3
) of hardened concrete at 28 days (SSD)
M1 M2 M3 H1 H2 H3
1977 1725 1559 2471 1952 1819
3.3  Compressive strength
Fig. 3 presents the compressive strength of both series at 3, 7, and 28 days. Fig. 4
illustrates the relationship between the 28-day compressive strength and the SSD
density. In Fig. 5, the contribution of different mix to the compressive strength is
compared in terms of the specific strength, i.e., ratio of strength to relative density.
From Fig. 3 it can be seen that when the FBA sand and Lytag were used to replace
natural sand and basalt respectively, different effects can be observed for series M
and H. In series H, the compressive strength decreased from H1 to H3 at all the ages.
However, in series M, this trend was visible only for the 3-day results. At the age of
7 and 28 days, there was an increase in strength when the NS and CA were replaced
with FBA and Lytag, respectively.
As indicated in Fig. 4, except for one data point corresponding to mix M1, there is a
linear relationship between the density and the compressive strength, i.e. the
compressive strength is directly proportional to the SSD density of hardened
concrete. This indicates that the lightweight was achieved at the cost of reduction in
the compressive strength. Nevertheless, it is still possible to manufacture LWC with
SSD density in the range of 1560-1960 kg/m
3
 and 28-day compressive strength in the
range of 20-40 N/mm
2
.
 
0
20
40
60
3 7 28
Age (Days)
Comp. Strength (N/mm2)
M1 M2 M3
(a) Series M
0
20
40
60
3 7 28
Age (Days)
Comp. Strength (N/mm2)
H1 H2 H3
(b) Series H
Fig. 3: Compressive strength of NWC and LWC 84  International Workshop on Sustainable Development and Concrete Technology
20
30
40
50
60
1400 1600 1800 2000 2200 2400 2600
Density (kg/m
3
)
Comp. Strength (N/mm2)
M1 M2 M3 H1 H2 H3
Fig. 4: Compressive strength vs. density
0
5
10
15
20
25
M H
Series
Specific Strength
100%OPC+100%NS+100%CA
100%OPC+100%FBA+100%LG
70%OPC+30%FA+100%FBA+100%LG
Fig. 5: Specific strength of NWC & LWC
Fig. 5 indicates that the specific strength for M2 and M3 are higher than M1, which
suggests that for the same weight of concrete, LWC provided marginally higher
compressive strength than NWC. For series H, the specific strength for H2 and H3
are lower than H1. However, the difference was small. Therefore, it can be
concluded that FA, FBA and Lytag can be favorably used to manufacture medium
strength LWC. In the case of high strength concrete, these replacements would result
in decrease in compressive strength of the concrete.  
3.4  Pull-off surface tensile strength
Fig. 6 presents the results of the pull-off test. It can be seen that, for Series M, the
surface tensile strength of M2 and M3 are higher than that of M1, and that of M2 is
equal to that of M3. However, for Series H, the pull-off tensile strength of H2 and
H3 are lower than that of H1, and the value for H3 is also lower than that for H2.
Thus, in terms of the pull-off surface tensile strength, FA, FBA and Lytag have a
beneficial effect on medium strength LWC, but a slightly adverse effect on highstrength LWC. Yun Bai, Ratiyah Ibrahim, and P.A. Muhammed Basheer 85
0
2
4
6
M H
Series
Surf. Tens. Strength
(N/mm2)
100%OPC+100%NS+100%CA
100%OPC+100%FBA+100%LG
70%OPC+30%FA+100%FBA+100%LG
Fig. 6: Surface tensile strength of
NWC and LWC
3.5  Permeability
The near surface permeation property
was evaluated by using the “Autoclam
Permeability System.” Figs. 7 and 8
show the air permeability and
sorptivity results, respectively.
It can be seen that when FBA sand and
Lytag were used to replace natural
sand and basalt to manufacture LWC,
the air permeability dramatically
increased. From Fig. 2, it can be seen
that, due to the water reduction effect
of FBA and Lytag, the free water of
mix 2 for both
series is lower than mix 1. Since the binder content was the same for all the mixes in
each series, the decreased free water content would result in a decreased free waterbinder ratio, which should have decreased the air permeability [18]. In addition,
although the particles of Lytag are quite porous [17], they have no effects on the air
permeability of LWC [19]. Thus, the increased air permeability should be
attributable to the porous particles of the FBA sand [5].  However, the air
permeability indices of Mix 3 for both series are lower than mix 2, but still higher
than mix 1.  The decrease of the air permeability indices of mix 3 can be considered
to be due to the physical filling effect and pozzolanic reaction of FA, leading to the
densification of the pore structure. This reveals that the increased air permeability
caused by the porous FBA particles can partly be compensated by the FA. However,
since the slabs were only 28 days old, the pozzolanic reaction has not fully
developed. Thus, a long-term study is required in order to investigate any possible
further beneficial effect of FA on the LWC.
0.00
0.40
0.80
1.20
1.60
M H
Series
Air Perm. Index
(ln(bar)/min)
100%OPC+100%NS+100%CA
100%OPC+100%FBA+100%LG
70%OPC+30%FA+100%FBA+100%LG
Fig. 7: Air permeability of NWC & LWC
0.00
2.00
4.00
6.00
M H
Series
Sorpt. Index    
(M3*10-7/min0.5)
100%OPC+100%NS+100%CA
100%OPC+100%FBA+100%LG
70%OPC+30%FA+100%FBA+100%LG
Fig. 8: Sorptivity of NWC & LWC 86  International Workshop on Sustainable Development and Concrete Technology
The sorptivity result in Fig. 8 indicates that when natural sand and basalt were
replaced with the FBA sand and Lytag, the sorptivity indices for both series were
higher than the corresponding control (mix 1). This is mainly attributable to the
porous FBA and Lytag particles. However, when FA was used in mix 3 to replace
30% of the OPC, the sorptivity did not decrease as it did in air permeability. Thus,
the FA has no beneficial effect on reducing the sorptivity of LWC at 28 days. On the
contrary, the sorptivity increased. Again a long-term study is required to investigate
any further beneficial effect.
0.00
4.00
8.00
12.00
M H
Series
Porosity (%)
100%OPC+100%NS+100%CA
100%OPC+100%FBA+100%LG
70%OPC+30%FA+100%FBA+100%LG
Fig. 9: Porosity of NWC & LWC
3.6  Porosity
The porosity result is reported in Fig. 9.
The trend was similar to that of air
permeability in Fig. 7, i.e., when natural
sand and basalt were replaced by the FBA
sand and Lytag respectively, the porosity
of LWC increased. However, when FA
was used in mix 3 to replace cement, the
porosity was lower than mix 2, but still
higher than mix 1 for both series. The
result further confirms that whereas FBA
sand and Lytag increase the porosity of
LWC, the FA would partly compensate
the detrimental effect caused by FBA sand
and Lytag on the porosity and air
permeability.
4. Conclusions
•  By using FA, FBA and Lytag, it is possible to manufacture lightweight
concrete with density in the range of 1560-1960 kg/m
3
.
•  In terms of contribution to the compressive strength by per unit weight of
concrete, FA, FBA, and Lytag can be beneficially used to manufacture
medium strength concrete.
•  LWC incorporating FBA and Lytag resulted in an increase in the
permeability; by replacing 30% of OPC with FA, the permeability of LWC
could be improved.
•  In order to manufacture durable LWC, measures should be taken to further
improve the permeation property.
Acknowledgments
The FBA for this research was supplied by Connexpo (N.I.) Ltd and, the Lytag was
supplied by Finlay Concrete Products (N.I.). The facilities provided by the School of Yun Bai, Ratiyah Ibrahim, and P.A. Muhammed Basheer 87
Civil Engineering at Queen’s University, Belfast, are gratefully acknowledged. Mr.
Yun Bai is in receipt of the Overseas Research Students Award at Queen’s and is on
leave of absence from Ningxia Communication Department, China. Ms. Ratiyah
Ibrahim was funded by the Government of Brunei.