Construction and Building Materials 79 (2015) 73–82

Contents lists available at ScienceDirect

Construction and Building Materials

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

Durability performance and engineering properties of shale and volcanic

ashes concretes
Mohammed Seddik Meddah ⇑
Sultan Qaboos University, Department of Civil & Architectural Engineering, Faculty of Engineering, P.O. Box: 33, P.C: 123, Al-Khod, Muscat, Oman

h i g h l i g h t s

 Using natural pozzolana and volcanic ash in concrete production may reduce Portland cement consumption.
 At low to moderate replacement level, these two pozzolanas provide satisfactory performance.
 Overall, SA pozzolana seems to perform better than NP.

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

Article history: Designing a sustainable concrete has become a vital requirement for today’s concrete. Introducing vari-
Received 4 August 2014 ous cementing and pozzolanic materials in concrete in replacement of Portland cement seems to be the
Received in revised form 18 December 2014 appropriate way to lower the environmental impact of concrete industry.
Accepted 4 January 2015
This paper reports results on concrete performance produced with two types of pozzolanas including
Available online 21 January 2015
natural pozzolana (NP) obtained from volcanic ash and shale ash (SA) used, in various proportions ranged
from 10–45%, as a partial substitute of Portland cement (PC). Concrete mixtures were designed with a
Keywords:
wide range of water-to-cementitious ratios (w/c) ranged from 0.79 to 0.45. The key mechanical proper-
Carbonation
Chloride diffusion
ties and durability performance of binary blended cement concretes were investigated.
Compressive strength Using both NP and SA has resulted in a strength loss; while SA seems to perform better than NP in
Drying shrinkage terms of strength development and durability performance. The blended cement concretes with 10–
Freeze–thaw 15% pozzolanas was found to have a good resistance to carbonation and chloride ions ingress and are
Natural pozzolana freeze–thaw durable. However, with replacement level higher than 20%, the durability factor, chloride
Portland cement ions and carbonation resistance drop down.
Setting time Overall, the results indicate that the mechanical and durability performances of binary blended
Shale ash
cements with NP and SA are strongly linked to their intrinsic characteristics including chemical compo-
Volcanic ash
sition, fineness, particle size distribution and potential reactivity.
Ó 2015 Elsevier Ltd. All rights reserved.

1. Introduction efforts to reduce the CO2 emission by minimising clinker produc-

Portland cement industry is one of the heaviest contributors to
the global CO2 emission with an estimated contribution of approx-
imately 5–7% of the global CO2 emissions [1]. The annual cement
production has passed from 2.6 billion tonnes in 2007 [2] to 4 bil-
lion tonnes in 2013 [3] and a rapid increase is expected in the com-
ing years. As the pressure to reduce CO2 emissions is rising under
the new environmental regulations, research is being directed
towards increasing the replacement levels of clinker and optimis-
ing different combinations of supplementary cementitious materi-
als (SCMs). Thus, the cement industry is now putting considerable
⇑ Tel.: +968 24142672.
E-mail address: seddikm@squ.edu.om
tion and using blended cements, in which part of the clinker is
replaced by SCMs or fillers. In fact, SCMs whether natural pozzo-
lans (volcanic tuff, zeolite and diatomite) or thermally treated clays
(metakaolin) or by-products industrial wastes such as silica fume,
fly ash, slag have been used for many years to develop Portland-
composite cements not only for environmental considerations
and sustainable development issues but also to enhance concrete
strength and durability performance [4].
Indeed, environmental issues are one of the major concerns but
also cost reduction, energy saving, the use of conventional raw
material or industrial by-products as well as concrete performance
improvement, are among the main objectives targeted by the use
of different mineral admixtures in concrete production. The bene-
fits of natural pozzolana (NP) and shale ash (SA) as well as other

http://dx.doi.org/10.1016/j.conbuildmat.2015.01.020
0950-0618/Ó 2015 Elsevier Ltd. All rights reserved.
74 M. Seddik Meddah / Construction and Building Materials 79 (2015) 73–82

SCMs used as a partial replacement of Portland cement (PC) are
now well established. In 2006, around 300 million tonnes of vari-
ous natural pozzolanas were estimated available while only 50%
are used [5].
The BS EN 197-1 (2000) identifies type II cements (CEMII/A-LL
32,5/42,5) that may contain various materials as main constitu-
ents, in percentages ranging from 6% to 35%. Natural pozzolana,
fly ash, slag and limestone are the main materials that are permit-
ted by the EN 197-1.
Natural pozzolanas such as diatomite have been widely used as
a partial substitute of Portland cement in many applications
because it contains reactive SiO2 and due to their beneficial effects
including reduction in heat evolution, decreased permeability and
increased resistance to chemical attack [6,7]. However, NP are
often associated with a reduction in early-ages strength [8–10]
but an improved hydration and strength at medium and later ages
when moist-cured for longer time [11]. Adding 10% raw diatomite
to PC resulted in a comparable strength with respect to the refer-
ence concrete while beyond this limit, compressive strength
decreases due to the high water demand induced by the increased
fineness. The strength development of diatomite concrete was
reported to increase depending on the increased diatomite content
and reactive silica content [12].
Recent results revealed that ternary system containing 20%
limestone (LS) filler and 30% NP has led to improved early and
long-term compressive and flexural strengths and enhanced dura-
bility against sulphate, acid and chloride ions ingress [11].
The use of shale ash is, however, limited and shown less success
and efficiency compared to the other types of pozzolanas. Several
authors [13,14] have reported that the use of burnt shale in cement
pastes increased the water demand and prolonged the setting time.
The burnt shale ash has a low pozzolanic activity which results in a
lower strength development of concrete/mortar compared to con-
trol mixture as reported by some authors [13,15]. Meanwhile,
Baum et al. [16] have found that shale ash concrete exhibited less
shrinkage than PC concrete while Bentur and Grinberg [14] have
reported that shale ash increased drying shrinkage of concrete.
In this paper, the effect of volcanic ash and shale ash on con-
crete performance is investigated. Attention is particularly directed
to the key mechanical and durability performances of concrete
specimens made with these two types of SCMs as a partial substi-
tute of Portland cement. The results presented herein are part of an
extensive research project aims to develop concrete made with
various Portland-composite cements.
3. Experimental work
3.1. Materials
Portland cement CEM I 42.5 N conforming to EN 197-1:2000
was used in all control mixtures while various proportions of NP
and SA were added as a partial substitution of PC to obtain binary
binders. These binary binders were formulated varying the
replacement (by weight) of PC by NP from 0% to 45% and the
replacement of SA from 0% to 40%. The Natural Pozzolana used is
a volcanic dust with a rough surface texture while the burnt shale
ash has a high inherent fineness. The chemical and mineralogical
compositions along with physical properties of PC and the two
pozzolanic admixtures (NP and SA) used are given in Table 1;
and their particle size distribution are shown in Fig. 1. Both NP
and SA used could, relatively, be considered as high siliceous.
Two size fractions of 5–10 mm and 10–20 mm of crushed gran-
ite were used as coarse aggregate (CA) while river sand 0–5 mm
was used as fine aggregate. To obtain a nominal targeted slump
value of 75 ± 5 mm, a superplasticizer conforms to BS 5075: Part
3 with a proper content (expressed by % of weight of cement)
was introduced while an air-entraining agent conforming to BS
5075: Part 2 was used for the purpose of freeze–thaw resistance
of air-entrained concrete.
3.2. Details of mixtures, concrete mixing and specimens
Both PC and binary concretes were designed with a wide range
of water to cementitious materials (w/c) of 0.79, 0.65, 0.60, 0.52
and 0.45. The cement content was varying from 235 to 410 kg/
m3. The total coarse aggregate content was kept constant
(1200 kg/m3) while the fine aggregate content was slightly
adjusted to maintain the yield. Table 2 provides the mix propor-
tions of both PC and binary cements concretes investigated. All
concrete mixtures were produced in a planetary concrete mixer
and were appropriately labelled as given in Table 2. After mixing,
slump test was carried out before concrete was cast in the moulds.
For all concretes types, three samples of both Portland cement
and binary cement concretes containing natural pozzolana and
shale ash, referred to as PC, NP and SA respectively, are examined.
Blended mixtures NP and SA label are followed by a number desig-
nating the replacement level of PC by NP and SA. Concrete speci-
mens were cast in metal moulds (cubes, cylinders and prisms) in

Table 1
2. Research significance
Chemical composition and physical properties of cement and pozzolanas used.

Today’s concrete industry is facing multiple challenges, espe- Constituents PC NP SA

cially ensuring adequate mechanical and durability performance Chemical composition, %

with low environmental impact. There is no doubt that the use SiO2 21.4 58.1 56.1
Al2O3 4.7 13.1 23.6
of various pozzolans and SCMs in concrete production have consid-
Fe2O3 2.7 4.4 8.2
erable beneficial effects not only on mechanical and durability per- CaO 65.2 9.2 2.3
formance of concrete but also economical and environmental MgO 1.0 2.6 1.9
advantages. Since many decades, Silica fume (SF), fly ash (FA), SO3 2.9 – 0.27
and slag are the most popular cementing materials that are widely TiO2 – 0.6 1.2
Total alkali, Na2Oe 0.55 1.053 2.023
used in concrete. Whereas, NP and SA have been first used in LOI 0.9 6.4 3.2
cement-based materials since several decades before the develop-
Bogue composition, %
ment and effective use of SF, FA and slag, their use as cementing C3S 67.3 – –
materials is still limited and only limited data on their overall per- C2S 10.6 – –
formances are yet available. C3A 7.9 – –
This research highlights the mechanical and durability perfor- C4AF 8.2 – –
mance of binary concrete designed with various proportions of Physical properties
NP and SA. Contrary to what has been established by some authors, Density, kg/m3 3140 2480 2670
Fineness, m2/kg 381 453 1590
SA has a great potential of use as a cementing material and per-
Residue on 45 lm, % 6.2 32.3 21.7
forms much better than NP.
 M. Seddik Meddah / Construction and Building Materials 79 (2015) 73–82 75

100 accordance with BS EN 12390-3. The flexural strength test was

determined under three points bending loading on a prismatic
90 specimen of 100  100  500 mm while modulus of elasticity
was determined on a 150  300 mm cylinder specimens at 28 days
80
of age in accordance with EN 12390-5 and BS EN 1352:1997,
Percentage passing (%)

70 respectively. Specimens were wet cured at 20 ± 2 °C from demoul-

ding at 24 h of age till testing time. All the results reported are the
60 average of three tested specimens. It should be noted that because
of the considerable number of mixtures investigated, flexural
50
PC strength and some durability testing were conducted only on some
40 selected mixtures.
NP
30
SA 3.3.1. Shrinkage
20 Drying shrinkage was measured on 75  75  300 mm prisms
equipped with stainless steel studs at either end of the specimen
10 45 μm in accordance with BS 1920-8:2009. The prisms were demoulded
24 h after casting and then stored in a temperature and humidity
0
1 10 100 1000 controlled room (20 ± 2 °C and 55 ± 5% RH). The length change
readings were periodically recorded for a period of 90 days using
Sieve size (µm) a strain frame.
Fig. 1. Particle size distribution of Portland cement and the two pozzolanas used.
3.3.2. Freezing–thawing (F–T)
Prismatic specimens of 75  75  300 mm were used to assess
three layers and compacted with a plate vibrator as specified by BS concrete resistance to rapid freezing-thawing according to ASTM
1881: Part 108: 1983. Afterwards, all concrete specimens were C666 (2003) standard test method. Specimens were cured in water
covered with plastic sheet and kept in their moulds in a laboratory at 20 ± 2 °C prior to their exposure to freezing and thawing cycles
environment for the first 24 h. Specimens were then demoulded and then subjected to temperature cycling from +4 to 18 °C. The
and wet-cured at 20 ± 2 °C for the first 28 days and subsequently, duration of the cycles is 2–5 h. The freezing portion of the cycle is
specific curing regimes were applied depending on the durability accomplished by air cooling (similar to air conditioning) and the
test. thawing portion by submersion in water.

3.3. Testing methods 3.3.3. Carbonation

Accelerated carbonation test was carried out on cubes of
Compression test was carried out on a concrete cubic specimen 100 mm wet-cured during the first 28 days. The specimens were
of 100 mm at different ages of 2, 7, 28, 60, 180 and 365 days in then pre-conditioned by drying in laboratory for 14 days. After-

Table 2
Mixture composition for 1 m3 of PCNP and PCSA concretes.

Constituents materials, kg/m3

w/c (design strength, MPa) PC content SA,% SP in PCSA,% NP,% SP in PCNP,% Aggregates Water
0/5 Sand 5/10 mm 10/20 mm
0.79 (25) 235 – 0.30 – 0.30 700 400 800 185
10 0.43 15 0.30
20 0.69 – –
30 1.03 30 0.50
40 1.46 45 0.77
0.65 (35) 285 – 0.24 – 0.24 730 400 800 185
10 0.38 15 0.24
20 0.62 – –
30 1.05 30 0.48
40 1.43 45 0.72
0.60 (40) 310 – 0.13 – 0.13 710 400 800 185
10 0.35 15 0.13
20 0.69 – –
30 0.95 30 0.35
40 1.30 45 0.52
0.52 (50) 355 – 0.11 – 0.11 670 400 800 185
10 0.39 15 0.11
20 0.68 – –
30 1.26 30 0.33
40 1.43 45 0.66
0.45 (60) 410 – 0.21 – 0.21 625 400 800 185
10 0.42 15 0.21
20 0.71 – –
30 1.26 30 0.50
40 1.60 45 0.76
76 M. Seddik Meddah / Construction and Building Materials 79 (2015) 73–82
wards, specimens were sealed, from all sides, using a bituminous
coating paint except the top side of the specimen which was
exposed, in a carbonation tank, to a CO2 enriched atmosphere con-
taining 4% of CO2 at 20 ± 2 °C and 55 ± 5% of RH. The depth of car-
bonation in the tested concrete was measured by applying a
phenolphthalein colour indicator spray on a freshly broken piece
of the specimens after 2, 4, 8, 12 and 20 weeks of exposure. This
turns non-carbonated concrete pink, and remains colourless in car-
bonated concrete. The depth of the uncoloured zone of the con-
crete (the carbonated layer) from the edges of broken piece is
measured at 3 points and the mean value is reported as the carbon-
ation depth.
3.3.4. Chloride ions penetration
Chloride ions ingress was measured across slice of concrete
specimens measuring 100 mm Ø and 25 mm thick. These slices
were cut from a concrete cylinder 150  300 mm wet-cured for
14 days. The tested slices were located in a standard two compart-
ments diffusion test cells. The inner part of the cell was filled with
calcium hydroxide solution and the cell was partially immersed in
a chloride tank, thereby exposing the outer face of concrete to the
chloride solution. The electrochemical chloride ingress test method
used called potential difference, where a potential difference of
7.5 V is passed across the specimen. The chloride transport rate
is evaluated from frequent chloride analysis of the liquid in the
two chambers. The Fick’s first law was used to calculate the chlo-
ride diffusion coefficient (Eq. (1)) and the coefficients were derived
from steady-state diffusion experiments.
@c
Q ¼ Dc
dx
where Q is the mass transport rate one square-metre of concrete
(mol/m2/s), Dc is the diffusion coefficient (m2/s), c is the total chlo-
ride concentration in the pore water (mol/m3), t is the time, and x is
the distance from the exposed surface (m).
3.3.5. Initial surface absorption (ISA)
This test consists of the measurement of the rate at which water
flows into the capillary pore network of concrete through a known
surface area. Estimation of the volume flow is obtained by mea-
surement of the length of flow along a capillary of known dimen-
sion. The initial surface absorption of the investigated mixes was
determined on 150 mm cubes after the concrete was water-cured
for 28 days as per BS 1881-208 Part 5. The specimens were oven-
dried at 105 °C to constant weight prior to the test and left to cool
to the laboratory temperature (17–20 °C) in a desiccator for a per-
iod of 24 h. The contact area is defined by a plastic cell sealed onto
the concrete surface and should not be less than 5000 mm2. Water
is introduced into the cell via a connecting point and maintained at
a head of 200 mm using a filter funnel. A second connection point
to the cap leads to a horizontal capillary tube. The connection to
the reservoir is closed and the absorption is measured by observing
the movement of the end of the water line in the capillary tube
with an affixed scale at 10 min. The ISA measurements were taken
at 10 min after the initiation of the test and the measurements are
referred to as ISA-10 at 10 min. More details about the testing and
setup used could be found in earlier publication by the author [17].
4. Test results and discussion
4.1. Fresh properties
As abovementioned, the amount of superplasticizer used was
adjusted to obtain a nominal slump value of 75 ± 5 mm for all con-
crete mixtures investigated. Regardless of the blended cement type
used, decreasing the w/c of concrete has, generally, resulted in an
increase of the amount of superplasticizer required to reach the
targeted slump value. As could be seen in Table 2, the use of both
NP and, especially SA has resulted in a significant increase of the SP
content to reach the target slump value. In fact, the major factor
affecting the SP demand in such a cementing system is mainly
the physical properties of the pozzolana used including shape,
the particle size distribution, the specific surface area, the amount
of ultrafine particles in the mixture and its chemical composition,
particularly the amount of unburned carbon. As for the two pozzol-
anas investigated herein, SA has significantly increased the amount
of SP required (double even more for some mixtures) compared to
NP. This increase of SP demand is mainly attributed to the high
specific surface area and considerable amount of ultrafine particles
in SA used which requires high content of SP for a better lubrica-
tion and dispersion of SA particles. In terms of workability, it seems
that the use of blended cement with NP is more advantageous than
the SA.
On the other hand, all blended cements tested showed longer
initial setting time compared to Portland cement as could be seen
in Table 3. The delay in setting time was, to some extent, more pro-
nounced for SA than for NP. The PC mix had the shortest setting
time with 105 min while NP45 had the longest initial setting time
with 198 min, followed by the SA30 with 188 min. Obviously, the
delay in setting time for mixes with NP and SA compared to the
PC mix is attributed to their low early-age reactivity compared to
clinker. In fact, setting time is intimately related to the hydration
kinetics of the cementing system which is also linked to the reac-
tivity of the pozzolana used and hence, to its chemical composi-
tion. Using NP and SA would lower the clinker-free lime and C3A
ð1Þ content in the mixtures and results in a delayed setting time.
4.2. Mechanical properties
Tables 4 and 5 show the compressive strength results of the
tested concrete mixtures. Obviously, for both control and blended
cement concretes, increasing the curing time and decreasing the w/
c of concrete has resulted in a substantial increase in the compres-
sive strength. Regardless of the w/c and concrete’s age, the use of
various amounts of NP and SA has led to a reduction in the com-
pressive strength. This reduction in the compressive strength of
blended cement concretes is proportional to the amount of pozzo-
lana added and was more pronounced when high content (over
30%) of pozzolanas was embedded in the mixture. For the same
replacement level of 30%, it was found that the use of NP has led
to a larger reduction in the compressive strength compared to
the SA.
The use of SA in concrete mixes seems to perform better than
mixes with NP and has less negative effect on the compressive
strength. Meanwhile, the use of 10% of SA has an insignificant
effect on the compressive strength while beyond this limit, impor-
tant strength loss was recorded.
Table 3
Consistency and initial setting time of PC and blended cements
used.
Mixture code Initial setting time, min
PC 105
NP15 138
NP30 147
NP45 198
SA10 149
SA20 167
SA30 188
SA40 179
 M. Seddik Meddah / Construction and Building Materials 79 (2015) 73–82 77

Table 4
Compressive strength of the PCNP concretes.

PCNP Mixture code NP, % Compressive strength, MPa

Age, days
1 2 7 28 60 180 365
w/c = 0.79 PC 0 7.5 12.0 18.5 27.5 31.0 34.0 35.0
235 kg/m3 NP15 15 5.5 9.5 14.5 22.0 25.5 28.5 30.0
NP30 30 4.0 7.0 11.0 17.0 19.5 24.5 25.5
NP45 45 3.0 4.5 8.0 13.0 15.5 20.5 21.0
w/c = 0.65 PC 0 10.5 17.5 26.5 37.0 41.0 44.5 46.0
285 kg/m3 NP15 15 8.0 14.0 21.0 29.5 33.5 38.5 40.0
NP30 30 6.0 10.5 17.5 25.0 28.5 33.0 34.5
NP45 45 4.0 7.0 12.0 18.0 21.0 26.5 27.5
w/c = 0.60 PC 0 12.0 20.0 30.5 41.0 45.0 49.0 50.5
310 kg/m3 NP15 15 9.5 16.0 24.5 33.0 37.0 43.0 44.5
NP30 30 7.0 12.0 20.5 28.5 32.0 37.5 38.5
NP45 45 5.0 8.5 15.0 21.5 25.5 31.5 32.5
w/c = 0.52 PC 0 16.0 26.0 38.5 50.0 54.0 58.5 61.0
355 kg/m3 NP15 15 13.0 21.0 32.0 41.5 46.0 51.5 53.5
NP30 30 9.5 16.0 27.0 37.0 40.0 46.5 48.0
NP45 45 7.0 11.5 20.0 28.5 32.5 40.0 42.0
w/c = 0.45 PC 0 20.5 32.0 47.5 59.0 63.5 68.0 71.0
410 kg/m3 NP15 15 16.5 27.0 40.0 51.5 56.0 60.0 63.0
NP30 30 13.0 20.5 34.0 45.0 49.0 54.5 57.5
NP45 45 9.5 15.0 26.0 37.0 41.5 50.0 52.5

Table 5
Compressive strength of the PCSA concretes.

PCSA Mixture code SA, % Compressive strength, MPa

Age, days
1 2 7 28 60 180 365
w/c = 0.79 PC 0 7.5 12.0 18.5 27.5 31.0 34.0 35.0
235 kg/m3 SA10 10 7.0 11.5 18.0 26.5 30.0 33.0 34.0
SA20 20 6.0 9.5 15.0 23.0 26.5 29.5 30.5
SA30 30 4.5 7.5 11.5 19.5 23.0 25.5 26.5
SA40 40 3.5 5.5 8.0 16.0 19.5 22.0 22.5
w/c = 0.65 PC 0 10.5 17.5 26.5 37.0 41.0 44.5 46.0
285 kg/m3 SA10 10 10.0 16.5 25.5 36.0 40.0 44.0 45.5
SA20 20 8.0 14.5 22.0 32.0 36.0 40.0 41.0
SA30 30 6.5 11.5 17.5 27.0 31.0 35.0 36.0
SA40 40 5.0 8.5 13.0 23.0 27.0 29.5 30.5
w/c = 0.60 PC 0 12.0 20.0 30.5 41.0 45.0 49.0 50.5
310 kg/m3 SA10 10 11.5 19.0 29.5 40.0 44.0 48.5 50.0
SA20 20 9.5 16.5 25.5 35.5 40.0 44.0 45.5
SA30 30 8.0 13.0 20.5 30.5 34.5 38.5 39.5
SA40 40 6.0 10.0 16.0 26.5 31.0 34.0 35.0
w/c = 0.52 PC 0 16.0 26.0 38.5 50.0 54.0 58.5 61.0
355 kg/m3 SA10 10 15.0 25.0 37.5 49.0 53.0 57.5 60.0
SA20 20 13.0 22.0 33.0 44.5 49.5 54.0 56.5
SA30 30 10.0 17.5 27.0 38.5 43.5 48.5 50.5
SA40 40 8.0 13.5 21.0 33.5 39.5 42.5 44.5
w/c = 0.45 PC 0 20.5 32.0 47.5 59.0 63.5 68.0 71.0
410 kg/m3 SA10 10 19.5 30.5 46.0 57.5 62.5 67.0 70.5
SA20 20 17.0 27.5 41.5 53.5 59.0 64.0 66.5
SA30 30 14.0 22.5 34.5 47.0 53.0 58.5 61.0
SA40 40 11.0 17.5 27.5 41.0 48.0 52.0 54.5

For all the PCNP mixtures, the strength loss was ranging from
11% to 62% while for the PCSA mixtures, the strength loss was
ranging from around 1% to 57% depending on the w/c, concrete’s
age and the pozzolana content. The obtained results indicate that
to limit the reduction effect on the compressive strength, it is rec-
ommended to use a maximum of 20% SA and 15% NP or lower.
Available results [18,19] showed that pre-mixed blended lime-
stone or NP-cements with low PC replacement level up to 15%
could slightly improve compressive strength, or at least led to com-
parable strength to the PC concrete. On the other hand, López and
Castro [20] have found that mixtures with 33% NP have led to
around 27% compressive strength reduction which is relatively in
agreement with the present results.
In fact, it is well recognised that when blended cements are pro-
duced by intergrinding clinker together with the cementitious
materials (genuine blended cement), the resulted performance sur-
pass the one mixed separately in the concrete mixer.
Obviously, mechanical performance of any cementing system is
strongly linked to its reactivity which in turn linked to the chemi-
cal composition and intrinsic characteristics of each cementitious/
pozzolana material. The chemical composition given in Table 1
showed that NP has relatively high LOI of 6.4% which could be con-
78 M. Seddik Meddah / Construction and Building Materials 79 (2015) 73–82

w/c particles are, the higher the reactivity will be. The results shown
28 days PCNP in Tables 4 and 5 and Figs. 2 and 3 indicate that SA was relatively
0.79
70
0.65 more reactive than NP used. In fact, Fig. 1 shows also that NP par-
ticles are slightly coarser than their corresponding PC and SA.
Compressive strength, MPa

60 0.60
One major distinction between the two pozzolanas materials
50 0.52
used (NP and SA) to be noted is their kinetic of reaction. Using
0.45 NP has resulted in an increasing rate of strength loss up to 7 days
40
followed by a decreasing rate of the strength loss onwards. Never-
30 theless, a decreasing trend of the strength loss over time was
observed when the SA was used. It could be speculated that NP
20
has a predominant negative effect at early ages (up to 7 days)
10 which turns into a positive effect at later-ages while SA has a
decreasing rate of the strength loss with extending curing time.
0 Indeed, the addition of pozzolanic materials is expected to
0 10 20 30 40 50 improve the interface properties by a pozzolanic reaction and a fill-
NP content, % ing effect and hence, improving the strength and impermeability of
Fig. 2. Compressive strength development of Portland and natural pozzolana the concrete. However, this improvement depends on the pozzola-
cement concretes. nic materials parameters including reactivity, oxide composition
and the amount of pozzolana used which could affect its pozzola-
nicity and hence, determine its pozzolanic activity index (ASTM
w/c C311). The relative strength, which represents the relation of
28 days strength of a given mix at a given age to the strength of the refer-
70 PCSA 0.79
ence mix, may indicate the degree of pozzolanicity and the overall
0.65
Compressive strength, MPa

60 contribution of the cementitious materials in strength gain/loss of

0.60
the blended cement concrete.
50 0.52 The relative strength of the mixtures designed with NP and SA
0.45 was calculated at 7 and 28 days and the results are presented in
40
Table 6. At both 7 and 28 days and for both mixture types (PCNP
30 and PCSA), the relative strength ratio was always under the unit
1. It should be mentioned that for the five w/c mixes investigated,
20
the relative strength of the SA10 mix was very close to the unit 1
10 which indicates a high pozzolanicity of SA when used at low con-
tent. The relative strength results presented in Table 6 indicate, for
0 both mixture types, a strength loss at 7 and 28 days compared to
0 10 20 30 40 50 the control mix. This strength loss is strongly dependent on the
SA content, % proportion of pozzolanas used. As speculated earlier, the SA con-
Fig. 3. Compressive strength development of Portland and shale ash cement tent incorporated in partial replacement of Portland cement should
concretes. not exceed a maximum of 20% while the use of NP, however,
should be limited to around 15% to reach a strength value nearly
sidered as higher than the limit of around 6% set by BS 3892-1993 comparable to that of Portland cement concrete. This indicates that
and ASTM C618. In addition, reactivity of cementing materials is both NP and SA could generate an effective pozzolanicity when
also linked to its fineness and particle size distribution. Finer the incorporated at a moderate replacement level ranging from 15%

Table 6
Relative compressive strength of the binary concrete mixtures.

SA mixtures Relative strength NP mixtures Relative strength

7 days 28 days 7 days 28 days
w/c = 0.79 SA10 0.97 0.96 NP15 0.78 0.80
SA20 0.81 0.84 – – –
SA30 0.77 0.85 NP30 0.60 0.62
SA40 0.43 0.58 NP45 0.43 0.47
w/c = 0.65 SA10 0.96 0.97 NP15 0.79 0.80
SA20 0.83 0.87 – – –
SA30 0.66 0.73 NP30 0.66 0.68
SA40 0.49 0.62 NP45 0.45 0.49
w/c = 0.60 SA10 0.97 0.98 NP15 0.80 0.81
SA20 0.84 0.87 – – –
SA30 0.67 0.74 NP30 0.67 0.70
SA40 0.52 0.65 NP45 0.49 0.52
w/c = 0.52 SA10 0.97 0.98 NP15 0.83 0.83
SA20 0.86 0.89 – – –
SA30 0.70 0.77 NP30 0.70 0.74
SA40 0.55 0.67 NP45 0.52 0.57
w/c = 0.45 SA10 0.97 0.96 NP15 0.84 0.87
SA20 0.87 0.91 – – –
SA30 0.73 0.80 NP30 0.72 0.76
SA40 0.58 0.70 NP45 0.55 0.63
 M. Seddik Meddah / Construction and Building Materials 79 (2015) 73–82 79

Table 7 ucts), strength and modulus of elasticity development are affected

The 28-day compressive and flexural strengths, and modulus of elasticity of concrete
by mineralogical features of the clinker, pozzolanic reactions, fine-
mixtures investigated.
ness, reactive SiO2 ratio and water demand of the cement mixtures.
Mixture code Cube compressive Flexural Modulus of Further investigation regarding CH and C-S-H is needed to investi-
(w/c = 0.60) strength , MPa strength, MPa elasticity, GPa
gate the pozzolanic reaction between the NP/SA and portlandite
PC 41.0 5.0 30.0 and to establish the efficiency of these Mineral admixtures as poz-
NP15 33.0 (19.5) – 27.5 (8)
zolanic materials.
NP30 28.5 (30.5) – 25.0 (16.7)
NP45 21.5 (47.6) 3.5 (30) 22.5 (25)
SA10 40.0 (2.4) – 29.5 (1.7) 4.3. Durability performance
SA20 35.5 (13.4) – 27.5 (8)
SA30 30.5 (25.6) 4.0 (20) 24.0 (20)
4.3.1. Drying shrinkage
SA40 26.5 (35.4) – 21.5 (28)
The ultimate magnitude of drying shrinkage strains measured
Note: ()% reduction compared to the control mixture. after 90 days of exposure to dry-air is summarised in Table 8. It
could be seen that NP has slightly reduced the total magnitude of
shrinkage by an average of 10–17% while SA has, generally,
Table 8
increased the drying shrinkage magnitude by approximately 12%.
Drying shrinkage strains and durability factor of concretes investigated. As discussed above, the high fineness of SA and its chemical reac-
tivity might contribute to the increase in shrinkage magnitude
Mixture code (w/ Drying shrinkage strains, ASTM durability factor,
c = 0.52) (microstrains) (fresh air content = 5%)
observed for the PCSA mixtures. Obviously, the high fineness of
NP, and especially SA would increase the surface area of the
PC 680 96.0 (>80%) Reliable
cement paste and its ultimate magnitude of shrinkage strains.
NP15 645 83.3 (>80%) Reliable
NP30 607 62.5 (>60%) Just above The finer and reactive is the pozzolanic material, the larger will
the limit be the shrinkage strains magnitude generated as also indicated in
NP45 575 41.7 (<60%) Very poor other sources [22,23].
SA10 690 95.8 (>80%) Reliable
SA20 710 85.4 (>80%) Reliable
SA30 740 70.8 (>70%) Acceptable 4.3.2. Initial surface absorption (ISA)
SA40 767 58.3 (<60%) Poor ISA test is designed to evaluate the porosity of concrete and
cement based-materials. Fig. 4 provides, for both PC and blended
cement concretes, the test results of the ISA-10 to water versus
to 20%. The results in Table 6 showed also that SA is always more the pozzolanas content. The results indicate that increasing the
reactive than NP, particularly for the same replacement level of NP or SA content in concrete leads to an increase of the ISA-10.
30%. The increase of ISA-10 seems to be more pronounced in case of
Likewise the compressive strength, the use of NP and SA has NP compared to SA. At all concrete’s ages and for the same replace-
resulted in a reduction in both the flexural strength and modulus ment level of 30%, the PCSA showed a lower ISA-10 than the PCNP.
of elasticity as could be seen in Table 7. The reduction in modulus On the other hand, for both the PC and blended cement concretes,
of elasticity was proportional to the replacement level and was and regardless of the type and content of pozzolana used, the ISA-
ranging between 5% and 33%. In fact, while modulus of elasticity 10 decreases when increasing the curing time as could be seen in
is intimately linked to the aggregate bulk rather than the cement Figs. 5 and 6. Indeed, due to the increase of the curing time up to
paste, it is obvious that modulus of elasticity is also affected by 6 months, pozzolanic reaction will progress and results in the for-
the three phases of concrete material: aggregate bulk, cement mation of additional hydrate products, less porosity and high com-
paste and the interfacial transition phase [21]. The results obtained pactness, disconnected pore system and hence, lower ISA-10
seem to indicate that the blended cement pastes are weaker than values [24,17]. The lowest value of ISA-10 of 25  102 mm2/s,
their corresponding PC-cement pastes which in turn led to the obtained after 6 months of curing, was recorded for the PCSA10
reduction in the modulus of elasticity. Moreover, since the strength mix while the highest value of 76  102 mm2/s was recorded for
of cement paste is a function of the hydrated part (hydrate prod- the PCNP45 mix. Definitely, higher the pozzolana content, higher
the ISA-10 values will be.

120 NP-0.79 28 days

28d
NP-0.60 80 NP
100 NP-0.45 70
SA-0.79
ISA-10, m/m2/sec 10 -2

ISA-10, m/m2/sec 10 -2

SA-0.60 60
80
SA-0.45 50
60 40

30
40
20
20 PC NP15
10
NP30 NP45
0 0
0 10 20 30 40 50 0 50 100 150 200
Pozzolanas content, % Age, days

Fig. 4. Effect of natural pozzolana and shale ash contents on the ISA. Fig. 5. ISA versus time for natural pozzolana concretes.
80 M. Seddik Meddah / Construction and Building Materials 79 (2015) 73–82

80 indicator of its resistance to carbonation than the type of cement

SA
used. Reducing the w/c of concrete results in a strength improve-
70 ment and hence, better concrete resistance to carbonation. On
60 the other hand, since no strength improvement was recorded by
ISA-10, m/m2/sec 10-2

using either NP or SA; similarly, no enhancement of the resistance

50 to carbonation was found as a result of the presence of NP or SA.
40
4.3.4. Chloride ingress resistance
30 Chloride diffusion coefficients (D) for both PC mixes and mixes
made with various amounts of NP and SA are determined based on
20
Fick’s second law and the results are presented in Figs. 8 and 9.
10 PC SA10 SA20 Regardless of the type and content of pozzolanas used, the results
SA30 SA40 indicate that increasing the w/c of concrete has led to significant
0
increase in the chloride diffusion coefficient. Moreover, except a
0 50 100 150 200
slight decrease of the chloride diffusion coefficient for mixes with
Age, days
10–20% SA, increasing the pozzolana content in the mixture has
Fig. 6. ISA versus time for shale ash concretes. also resulted in an increase of the chloride diffusion coefficient.
For the PC mixes, it was found that the chloride diffusion coeffi-
cient ranged from 1.1  1010 m2/s to 2.15  1010 m2/s depending
on the w/c. However, for mixes containing NP and SA, the chloride
45 SA-0.65 20 weeks exposure
diffusion coefficient varies considerably depending on both the w/c
40 SA-0.52 of concrete and NP/SA content. Results presented in Figs. 8 and 9
Carbonation depth, mm

NP-0.65 showed that for mixes containing various contents of cementing

35
NP-0.52
30
25
20
15
10
5
0 ticles. Higher is the reactivity of pozzolana used, larger is the
0 10 20 30
Pozzolanas content, %
Fig. 7. Concrete carbonation depth versus NP and SA contents after 20 weeks
exposure.
4.3.3. Carbonation
As well known, carbonation is a process that results in a reduc-
tion of concrete pH to below 9 and could lead to the initiation of
steel reinforcement corrosion.
Decreasing the w/c from 0.65 to 0.52 has resulted in a signifi-
cant reduction in the carbonation depth while increasing the
amount of pozzolanas embedded in the mix has led to an increase
of the carbonation depth. It could be seen in Fig. 7 that using up to
10% of either NP or SA has no significant effect on the carbonation
depth of the binary concretes. However, increasing the pozzolanas
content beyond this limit has led to sharp and substantial increase
of the carbonation depth. At the same replacement level (30%) of
PC by the pozzolanas and for both w/c ratios, the depth of carbon-
ation recorded when using NP was higher than in case of SA. As
discussed above, SA seems to be quite more reactive and finer par-
ticles than NP. This relative reactivity and particle size of SA may
result in a formation of additional hydrate products in the system
(a secondary C-S-H) which contributes to fill the capillary pores,
refine the coarser capillary pores and disconnect the pore network.
Also the filling capacity of the powder could reduce the permeabil-
ity and penetration of CO2 into the treated PCSA/PCNP concretes.
For the mixes designed with a w/c of 0.65, a maximum carbon-
ation depth of 42 mm was recorded when embedding 45% NP;
however, the maximum carbonation depth was reduced to a half
(22 mm) for the same mixtures when reducing their w/c to 0.52.
In fact, it is well recognised that concrete strength is the best
material up to 25–30%, the chloride diffusion coefficient could
reach an average value of 3.3  1010 m2/s depending on the w/c
of the mixture. However, for a high content of pozzolana over
40%, the chloride diffusion coefficient may reach a maximum value
of 5.8  1010 m2/s with an increasing gap when over 30% of
pozzolanas (NP or SA) is used (Figs. 8 and 9).
Indeed, this suggests that the differences between the transport
coefficients of the PC and PCNP or PCSA are mainly due to the dif-
ferent pore structure developed in the presence of pozzolana par-
40 50
amount of secondary C–S–H formed and hence, lower the pore vol-
ume and finer the pore structure will be. Therefore, this reduces
the fraction of interconnected pores which in turn reduces the per-
meability and diffusivity within the blended cement matrix. As
could be seen in Figs. 8 and 9, the actual trends suggest that using
up to 15% of NP, no significant effect on the chloride diffusion coef-
ficients was observed while using up to 20% SA had reduced the
chloride diffusion coefficients.
Overall, the use of pozzolana such as NP and SA could physically
improve the pore structure of the cement matrix, particle packing
by filling the voids between the clinker particles as well as being a
reactive material and consequently, decreases chloride diffusion
and penetration of other harmful materials. Nevertheless, perme-
ability to chloride ions is intimately linked to the mixture design
parameters and microstructure of the resulted matrix, in particular
the capillary pore network characteristics. The larger the volume of
coarser and connected pores is, the higher the permeability and
chloride diffusion coefficient will be.
4.3.5. Resistance to freezing–thawing
The durability factor DF is commonly used to quantify the
freeze and thaw resistance of concrete. It is expressed by the rela-
tion in Eq. (2).
PN
DF ¼ ð2Þ
M
where P is the relative dynamic modulus of elasticity at N cycles (%);
N is the number of cycles at which P reaches the specified minimum
value for discontinuing the test or specified number of cycles at
which the exposure is to be terminated, whichever is less; and M
is the specified number of cycles at which the exposure is to be
terminated.
 M. Seddik Meddah / Construction and Building Materials 79 (2015) 73–82 81

45 air content of 5%, the amount of pozzolanic materials used seems

PCNP to be the major factor affecting the DF of air-entrained concretes.
Coefficient of diffusion, ×10-11 m 2/sec

40 NP45
Obviously, freezing-thawing durability factor increases as
35 strength increases and permeability decreases. The results
obtained revealed that permeability to chloride increases and com-
30 NP30 pressive strength decreases when increasing the replacement level
of PC by NP or SA above 15%. Similarly, the DF seems to drop down
25
NP15 when the replacement level of PC by NP is higher than 15% while
20 mixes with SA showed an adequate DF of 85.4 even at 20% replace-
PC ment level. It could be speculated that the combination PCSA has
15 provided an appropriate particle packing and strength develop-
10 ment which contributed in the satisfactory DF obtained with this
mixes. Moreover, the use of SA seems to produce more secondary
5 C–S–H gel and denser than the PCNP mixes, which indeed refine
the capillary pores and enhances durability performance of the
0
PCSA concrete.
0.40 0.45 0.50 0.55 0.60 0.65 0.70
w/c
5. Conclusions and recommendations
Fig. 8. Coefficient of diffusion of concrete made with various NP contents at
different w/c.
Two types of pozzolanas (volcanic and shale ashes) were exam-
ined in terms of mechanical and durability performances when
used in binary concretes. Based on the test results reported herein,
45 PCSA PC SA20
the following concluding remarks can be made:
Coefficient of diffusion, ×10-11 m 2/sec

40 The type and composition of blended cement seems to have a

35
30
25
20
15
10
5 durability performance. The compressive strength of both NP and
0 SA blended cements decreases with increasing PC replacement
0.40 0.45 0.50 0.55
w/c
Fig. 9. Coefficient of diffusion of concrete made with various SA contents at
different w/c.
The durability factor (DF) of both the PC and NP/SA mixes are
given in Table 8. It could be seen that except the NP45 mix, all
the other control and blended mixes have developed a good freez-
ing–thawing resistance with a durability factor of around 60% or
above. More importantly, mixtures with a pozzolana content up
to 15% have always exhibited a durability factor over 80%; whereas
beyond 15% replacement level, blended mixtures showed lower
freezing–thawing resistance with a DF ranging between 80% and
60%. In fact, ASTM C494 (1999) [25] standard requires a minimum
DF of 80% for satisfactory performance while Aïtcin [26] has sug-
gested that the DF should be above 60% after 300 cycles for con-
crete to be freeze thaw resistant. Other researchers [27,28] have
suggested an acceptable DF values ranging from 70% to 75%.
It is well known that freezing–thawing resistance, which is usu-
ally assessed and characterised by the DF, is governed by several
parameters such as strength and permeability but it is strongly
linked to the air content and the air system parameters including
total volume of air, the spacing factor and specific surface area of
air bubbles. These parameters are in turn affected by the mix
design parameters, particularly w/c, type and content of pozzol-
anas used and water reducing agent. The results indicate that, for
the mixes investigated herein (w/c = 0.52) with a constant fresh
SA40
significant influence on both the fresh and hardened properties
of concrete. The use of NP and SA has decreased the normal consis-
tency of blended cements due to their high specific surface area
SA30 compared to PC. Meanwhile, for the same slump value, the use of
both NP and SA has increased the superplasticizer demand com-
pared to PC, whereas the NP-cement requires lower superplasticiz-
SA10 er content than its corresponding SA-cement.
Even though that NP and SA particles are relatively reactive
materials, the results suggest a maximum replacement level ran-
ged between 15% and 20% to limit the strength loss and drop in
0.60 0.65 0.70 level at all tested w/c and ages.
Experimental results showed that air-entrained concrete mixes
designed with a w/c of 0.52 and up to 20% of pozzolanas (NP or SA)
would be expected to provide a good concrete freeze–thaw resis-
tance with a durability factor higher than 80%. Obviously, increas-
ing the w/c would reduce the resistance of both PC and blended
cement concretes to chloride ions diffusion, freeze–thawing and
carbonation.
Results showed that up to 20% replacement level, the NP and SA
blended cement concretes have exhibited a low initial surface
absorption, and good resistance to carbonation and chloride
ingress.
The performance and potential of each cementing system PCNP
or PCSA is intimately linked to the pozzolanicity of NP and SA
which is also depending on the reactive silica and/or alumina of
these pozzolanic materials used in the system.
References
[1] Damtoft JS, Lukasik J, Herfort D, Sorrentino D, Gartner EM. Sustainable
development and climate change initiatives. Cem Concr Res
2008;38(2):115–27.
[2] United States Geological Survey. USGS Mineral Program Cement Report; 2008.
p. 44–45.
[3] United States Geological Survey. USGS Mineral Program Cement Report; 2014.
pp. 38–39.
[4] Suzuki M, Meddah MS, Sato R. Use of porous ceramic waste aggregates for
internal curing of high-performance concrete. Cem Concr Res
2009;39(5):373–81.
[5] ECRA European Cement Research Academy. Development of state of the art-
techniques in cement manufacturing, (CSI/ ECRA Technology papers). Geneva;
4 June 2009.
82 M. Seddik Meddah / Construction and Building Materials 79 (2015) 73–82
[6] Rodriguez-Camacho RE. Using Natural Pozzolans to Improve the Sulfate
Resistance of Cement Mortars. In: Malhotra VM, editor. International
Conference, Bangkok, Thailand. ACI SP 178–53; 1998. p. 1021–1040.
[7] Yılmaz Bülent, Ediz Nezahat. The Use of Raw and Calcined Diatomite in
Cement Production. Cem Concr Compos 2008;30(3):202–11.
[8] Mehta PK. Studies on blended Portland cements containing Santorin earth.
Cem Concr Res 1981;11(4):507–18.
[9] Massazza F. Pozzolanic Cements. Cem Concr Compos 1993;15(4):185–214.
[10] Tagnit-Hamou A, Petrov N, Luke K. Properties of concrete containing
diatomaceous earth. ACI Mater J 2003;100(1):73–8.
[11] Ghrici M, Kenai S, Said-Mansour M. Mechanical properties and durability of
mortar and concrete containing natural pozzolana and limestone blended
cements. Cem Concr Compos 2007;29(7):542–9.
[12] Stamatakis MG, Fragoulis D, Csirik G, Bedelean I, Pedersen S. The influence of
biogenic micro-silica-rich rocks on the properties of blended cements. Cem
Concr Compos 2003;25(2):177–84.
[13] Smadi M, Yeginobali A, Khedaywi T. Potential uses of Jordanian spent oil burnt
shale as a cementive material. Mag Concr Res 1989;41(148):183–90.
[14] Bentur A, Grinberg T. Hydration and properties of pastes and mortars made
with oil-burnt shale. Am Ceram Soc Bull 1982;61(12):1296–300.
[15] Kheddaywi T, Yeginobali A, Smadi M, Cabrera J. Pozzolanic activity of
Jordanian oil shale ash. Cem Concr Res 1990;20(6):843–52.
[16] Baum H, Soroka I, Bentur A. Properties and structure of oil-shale ash pastes I:
composition and physical features. Cem Concr Res 1985;15(2):303–14.
[17] Limbachiya MC, Meddah MS, Ouchagour Y. Use of recycled concrete aggregate
in fly-ash concrete. Constr Build Mater 2012;27(1):439–49.
[18] Güneyisi E, Gesoglu M, Özturan T, Mermerdas K, Özbay E. Properties of
mortars with natural pozzolana and limestone-based blended cements. ACI
Mater J 2011;108(5):493–500.
[19] Voglis N, Kakali G, Chaniotakis E, Tsivilis S. Portland-limestone cements. Their
properties and hydration compared to those of other composite cements. Cem
Concr Compos 2005;27(2):191–6.
[20] López M, Castro JT. Effect of natural pozzolans on porosity and pore
connectivity of concrete with time. Rev Ing Constr 2010;25(3):419–31.
[21] Meddah MS, Zitouni S, Belaâbes S. Effect of size and content of coarse
aggregate on the compressive strength of concrete. Constr Build Mater
2010;24(4):505–12.
[22] Meddah MS, Tagnit-Hamou A. Evaluation of rate of deformation for early-age
concrete shrinkage analysis and time zero determination. ASCE J Mater Civ Eng
2011;23(7):1076–86.
[23] Meddah MS, Tagnit-Hamou A. Pore structure of concrete with mineral
admixtures and its effect on self-desiccation shrinkage. ACI Mater J
2009;106(3):241–50.
[24] Meddah MS, Lmbachiya MC, Dhir RK. Potential use of binary and composite
limestone-cements in concrete. Constr Build Mater 2014;58:193–205.
[25] ASTM C494/C 494M-99. Standard specification for chemical admixtures for
concrete, ASTM Committee C09. West Conshohocken, PA, USA: ASTM
Standards; 1999.
[26] Aïtcin P-C. Binders for durable and sustainable concrete. Taylor & Francis;
2008.
[27] Zia P, Ahmad S, Leming M, Schemmel J, Elliott R. Mechanical behaviour of high
performance concrete: very early strength of concrete, Report SHRP-C-363, V3,
1993, Strategic Highway Research Program. Washington DC: National
Research Council; 1993. p. 117.
[28] Richardson DN. Quick test for durability factor estimation. Missouri University
of Science and Technology & Missouri Department of Transportation; 2009. p.
113.