Powder Technology 254 (2014) 314–323

Contents lists available at ScienceDirect

Powder Technology
journal homepage: www.elsevier.com/locate/powtec

Influence of natural pozzolan, silica fume and limestone fine on strength,

acid resistance and microstructure of mortar
Y. Senhadji a,⁎, G. Escadeillas b, M. Mouli c, H. Khelafi c, Benosman d
a
Civil Engineering Department, University of MASCARA, Bp 736 route de Mamounia, Mascara 29000, Algeria
b
Université de Toulouse, UPS, INSA, LMDC, 135, Avenue de Rangueil, 31 077 Toulouse Cedex 4, France
c
Department of Civil Engineering, Laboratory of Materials, ENPO, 31000 Oran, Algeria
d
Faculty of Science, Laboratory of Polymer Chemistry, University of Oran, 31000 Oran, Algeria

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

Article history: Based on an ongoing experimental programme, this research focuses on the effect of various supplementary
Received 6 November 2013 cementitious materials (SCMs) (natural pozzolan (NP)/silica fume (SF)/limestone fine (LF) at various substitution
Received in revised form 3 January 2014 levels) on the microstructure and mechano-chemical resistance of blended mortar. The paper primarily considers
Accepted 12 January 2014
the characteristics of these materials, including their strength and the effects of aggressive chemical environments,
Available online 21 January 2014
by using sulphuric acid and nitric acid. The porosity and pore size distribution of the mortars are also examined using
Keywords:
mercury intrusion porosimetry (MIP).
Blended cements The microstructural changes in pastes caused by SCMs and the acid attack of the solution are analysed and related
Pozzolan powder to the phase composition found by X-ray diffraction (XRD). Microstructural investigations, such as scanning
Supplementary cementitious materials electron microscopy (SEM) and energy dispersive X-ray spectroscopy (EDS), were also used to support the
Acid attack explanation for these mechanisms.
Microstructure The results, according to ASTM C267, showed that the addition of natural pozzolan or limestone fine would
Calcium hydroxide improve the acid resistance of mortar, but at different rates depending on the proportion of SCMs. On the
other hand, mortars with silica fume are severely damaged in the sulphuric acid environment.
© 2014 Elsevier B.V. All rights reserved.

1. Introduction matrix, thus improving resistance to acid attack [11,12]. In a word,

Supplementary cementitious materials can be either self-cementing
or pozzolanic, or both, depending on their chemical composition [1].
They can be obtained as industrial by-products such as silica fume, fly
ash, and ground granulated blast-furnace slag or from agricultural wastes
such as rice husk ash [1–3]. Furthermore, several SCMs, such as volcanic
tuffs, pumice, calcined clay and others, can be obtained naturally [2,3]
and SCMs such as high reactivity metakaolin can be manufactured indus-
trially [4]. Some SCMs contain silica in amorphous form, which would
react with calcium hydroxide (C–H) to form more cementitious calcium
silicate hydrate (C–S–H, C_CaO; S_SiO2 and H_H2O) and contribute
to the concrete strength. These materials include volcanic ash, diatoma-
ceous earths, condensed silica fume, pulverised fly ash, natural pozzolan
and rice husk ash [5].
Today, the incorporation of SCMs has gained in importance because
it improves the durability of concrete [6–8], leads to a significant reduc-
tion in CO2 emissions, and reduces the unit cost of concrete production
[9]. The use of SCM can lead to densification of the mortar [10] and
cannot only contribute to the decrease or disappearance of calcium
hydroxide but also modify the chemical composition of the cement
⁎ Corresponding author. Tel.: +213 771 655 016; fax: +213 45 80 41 64.
E-mail address: senhadjidz@yahoo.fr (Y. Senhadji).
SCMs contribute greatly to the production of economic and durable con-
crete structures. Moreover, according to Hidalgo et al. [13], the capabil-
ity of pozzolanic materials to enhance the strength of concrete is more
closely associated with physical than chemical effects.
From the literature review, it is clear that, in the context of durability
and especially of resistance to acid attacks, there is debate over the
incorporation of mineral admixtures in concrete. First, the addition
of mineral admixtures may have little influence on the chemical
resistance. Second, the lower water requirement of concrete containing
mineral admixtures (except silica fume) coupled with pore refinement
through the formation of additional C–S–H may lead to a denser micro-
structure of the cement matrix, thus reducing the rate of diffusion of
aggressive solutions [8,14]. Third, the pozzolanic reaction consumes
calcium hydroxide (C–H) from the cement paste and thereby reduces
the amount of C–H available per unit area to react with the acidic
solution, which may result in faster movement of the corrosion front.
For instance, Durning and Hicks [15] and Mehta [16] reported that the
incorporation of silica fume increased the resistance of concrete to 1%
sulphuric acid attack through reduced C–H content and lower perme-
ability. Conversely, Monteny et al. [17] reported a negative effect of silica
fume incorporation in concrete specimens exposed to 0.5% sulphuric
acid. They stated that a refined pore structure with higher capillary
suction would cause deeper penetration of acidic solutions into concrete
and increase the surface area in contact with acid [17].

0032-5910/$ – see front matter © 2014 Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.powtec.2014.01.046
 Y. Senhadji et al. / Powder Technology 254 (2014) 314–323 315

Other research [7,11] has shown that the action of acid attack is de- Table 1
pendent on the type of acid present. In the case of attack by hydrochloric Chemical and physical properties of cement and supplementary materials.

and nitric acids, the calcium salt formed is soluble in water, leading to Oxides in clinker (%) OPC Natural Silica Limestone
higher mass loss. On other hand, in the case of sulphuric acid attack, Pozzolan fume
the calcium salt formed is not soluble in water. It becomes deposited SiO2 21.35 47.21 88.10 7.89
in voids and causes internal stresses that lead to disruption and loss of Al2O3 4.59 18.85 5.39 2.58
strength of the matrix. Fe2O3 5.52 9.99 0.84 1.13
CaO 63.89 10.84 1.49 45.45
In a series of chemical tests with a range of SCC mixtures, Bassuoni and
MgO 1.37 4.38 1.50 1.72
Nehdi [6] reported that the resistance to sulphuric acid attack of SO3 2.72 0.50 0.06 0.95
self-consolidating concrete (SCC) incorporating combinations of SCMs K2O 0.41 0.20 0.08 0.21
was moderately improved by addition of an organic corrosion Na2O 0.13 0.81 0.52 0.00
inhibitor (OCI), while sulphate resistant Portland cement (SRPC) and Cl – – – –
LOI 2.47 3.91 1.70 42.48
air-entrainment offered little improvement in strong and very strong
Calcite content (CaCO3) – – – 86.98
sulphuric acid solutions. Mineralogical compounds in clinker (Bogue calculation)
In Algeria, most of the cement is blended with additions such as C2S, % 25.69
limestone and natural pozzolan. Natural pozzolan is currently used in C3S, % 47.15
C3A, % 2.84
cement manufacture by at least eight of the fourteen Algerian cement
C4AF, % 16.70
plants while two cement plants use limestone. These cement plants
usually add between 15 and 20% of natural pozzolan and 10% of Physical properties
limestone filler as cement replacement by weight. Apart from the inter- Setting time (min.) Initial set 105
nal quality control testing for conformity to standards requirements, no Final set 225
Specific gravity [kg/m3] 3.10 2.62 2.24 2.70
detailed investigation has been made to evaluate the effect of the
Fineness, [m2/kg] (EN196-6) 325 430 23,000a 456
interaction between limestone and natural pozzolan additions on the a
According to the data sheet.
properties of cement mortar and concrete [18].
The present study is aimed at evaluating the response to sulphuric
and nitric acid of local SCM mortars prepared with ordinary Portland cement and NP are given in Table 1. The Algerian natural pozzolan
cement and natural-pozzolan/silica fume/limestone at various replace- shows a strong acidic character, having a (SiO2 + Al2O3) content ranging
ment levels. The chemical degradation caused by dilute solutions of around 57% of the total.
nitric and sulphuric acids is intended to simulate the deterioration of
mortars exposed to the aggressive environments. A specific method
standardised in ASTM C267 [19] was a major tool used in the current 2.1.3. Silica fume (SF)
study to evaluate resistance to acidic chemical environments. Condensed silica fume has been classified as a pozzolan by the
The effects of SCMs and exposure to these media on these cements American Concrete Institute (ACI) Committee 226 on cementitious
were investigated by using mercury intrusion porosimetry tests and materials [22]. ASTM C 618 [21] recognises three classes of pozzolan:
by the visual and chemical information provided by scanning electron N, F, and C. Silica fume is somewhat closer to the N class. However,
microscopy (SEM) and X-ray diffraction (XRD). Hence, the objective of particular provision is needed to classify silica fume appropriately.
the present investigation is to address some of the uncertainties The superfineness of SF (specific surface area 15,000–30,000 m2/kg
concerning the acid attack of mortar containing mineral admixtures. compared to 200–500 m2/kg for cement) and its high silica dioxide
content enhance its pozzolanic action. The silica fume used in this
2. Materials and methods experimental programme was a commercial product called Sikacrete
HD from Sika Algeria. Its chemical composition and selected properties
2.1. Constituent materials are shown in Table 1.

The physical and chemical properties of the various materials used
in the present work are highlighted below. The choice of local
Algerian materials was based on their abundant availability and their
moderate cost.
2.1.1. Cement
The cement used was a local ordinary Portland cement type (CEM I
42.5 R), manufactured by the Zahana Cement Company located in the
west of Algeria. The cement factory conforms to the Algerian standard
NA 442 [20] (which is mainly based on the European Standard EN
197-1). The physical properties of clinker, and its mineralogical and
chemical composition data are presented in Table 1.
2.1.2. Natural pozzolan (NP)
The natural pozzolans used as mineral admixtures in Portland cement
must meet certain chemical and physical requirements. For instance,
ASTM C 618 [21] Class N admixtures must have a minimum content of
70% in SiO2 + Al2O3 + Fe2O3, whereas Algerian natural pozzolan 76%.
Although, the importance of the content (SiO2 + Al2O3) is emphasised
by the fact that the active vitreous phases generally are richer in silica
and alumina content. NP used in this study was obtained from natural de-
posits in the west of Algeria. The pozzolan was ground in a laboratory mill
to a specific surface of 350 m2/kg. The chemical composition of the
2.1.4. Limestone fine (LF)
Crushed limestone is the main source of aggregates used in concrete.
However, its production is associated with high percentages of fines
that can make these aggregates unacceptable in the concrete mix
design. As a result, over 20% of such products cannot be used and
become hazardous to the environment. Many studies have shown that
the main effects of limestone filler are of a physical nature. It leads to
better packing of the cement granular skeleton and a large dispersion
of cement grains [23]. The crushed limestone used here was obtained
from limestone quarries in Oran City. It was ground to a specific surface
area of 456 m2/kg, with a maximum nominal size of 80 μm. The basic
properties of crushed limestone are shown in Table 1.
The granulometric study of the three mineral additions was carried
out using CILAS 1090 apparatus, which analyses particle sizes from
0.02 to 500 μm by laser diffraction. The particle size distributions are
shown in Fig. 1.
2.1.5. Sand
Local, well graded, crushed sand from a quarry at Kristel (Oran) was
used. Its physical proprieties were as follows: relative specific density,
2650 kg/m3, water absorption, 1.5%, bulk density, 1500 kg/m3 and
maximum grain size, 3 mm.
316 Y. Senhadji et al. / Powder Technology 254 (2014) 314–323

100 are the average of three tests. Because the durability of cementitious
NP materials is as important as their strength, the performance of SCM
SF mortars in aggressive media such as sulphuric and nitric acids was stud-
80 LF
LS
ied and compared with that of ordinary Portland cement. The mortar
Percentage passing (%)

was cast in cubic steel moulds 50 ∗ 50 ∗ 50 mm3. The mortar specimens

were cured in water at 20 ± 3 °C for 28 days before being subjected to
60 acid attack.
The relative acid attack was determined in accordance with ASTM
C267 [19]. Three specimens of each mortar mix were immersed either
40 in 5% sulphuric acid (H2SO4) or in 5% Nitric acid (HNO3) for a period
of 8 weeks. The other three specimens were cured in water. The volume
of acid solution was taken as four times the volume of the immersed
20 specimens to ensure that all surfaces of the samples were in contact
with the solution. It was refreshed every week. Specimens were
extracted from the solution weekly, rinsed three times with tap water
0 to remove loose reaction products, blotted with a paper towel and left
0,1 1 10 100 to dry at 20 ± 3 °C and 50% RH for 30 min before weighing and optical
Particule diameter (micron) microscope inspection. For each specimen, the cumulative mass loss at
each week was calculated using Eq (1):
Fig. 1. Granulometry of the SCMs.

2.1.6. Admixture
A superplasticiser (SP) provided by the company GRANITEX
(Algeria) was used. This superplasticiser, MEDAPLAST-SP40, was incor-
porated in all mixes and the content was adjusted slightly for each mix
to provide efficient dispersion of the cementitious particles and to
maintain similar consistency. The solid content, pH and specific gravity
of the SP are 40%, 8.2 and 1.07 respectively.
2.2. Specimen preparation, curing and casting
To find the effect of SCMs on Portland cement, a total of 10 mortar
mixtures were prepared with a constant binder/sand/water proportion
of (1/3/0.5) and superplasticiser was added to improve workability. The
binder consisted of cement with NP or SF or LF. The detailed mix propor-
tions of the mortars are shown in Table 2. All replacements were made
by mass of cement. To increase the homogeneity, the binders (OPC, with
NP, SF or LF) were first mixed together in the mixer for 2 min, and then
the water and sand were added. After that, the mixes were poured into
steel moulds and compacted. The specimens were surface-smoothed
and covered with plastic film. All specimens were removed from the
moulds 1 day after casting. Thereafter, they were cured in water at
room temperature (20 ± 3 °C).
The flowability of the mortars was determined by flow table tests
according to American standard ASTM C 230 [24]. The mortar sample
spread diameter, measured before and after 15 strokes (1 stroke per
second), represented the consistency of a mortar. The flow values
were between 97% and 113%.
For each mix, prisms 50 ∗ 50 ∗ 50 mm3 were tested to determine
compressive strengths, at 2, 7, 28, 90 and 360 days. The results reported
W 1 −W 2
Weight loss ð% Þ ¼  100 ð1Þ
W1
Where W1 and W2 are the weight of the specimen before immersion
and cleaned after immersion respectively.
2.3. Microstructural Investigations
Mercury intrusion porosimetry (MIP) with a contact angle of 130°
and maximum pressure of about 450 MPa was used to investigate the
hardened mortar samples with an age of 360 days or older. Capillary
pores between 0.0025 μm and 100 μm were measured. X-ray diffraction
(XRD), scanning electron microscopy (SEM), and energy-dispersive
X-ray spectroscopy (EDS) were performed in order to explore the
links between the macroscopic and microstructural characteristics of
mortar with and without mineral additions. The XRD analyses were s
carried out using a Siemens D-5000 X-ray diffractometer with Kα cobalt
anticathode (λ = 1.789 Å). The 2-Theta values ranged from 5° to 30°
and were recorded with a 0.04° step and an acquisition time of 2 s per
step. For investigations by SEM and EDXA, a JEOL—model JSM 6380LV
equipped with a Falcon Energy EDXA System 60SEM was used.
3. Results and discussion
3.1. The compressive strength
3.1.1. Effect of natural pozzolan
The influence of natural pozzolan on the compressive strengths of
the Portland cement-NP mortars is shown in Table 3. NP induced reduc-
tions in the compressive strengths of mortar at all levels of replacement
at 2, 7, 28 and 90 days with the exception that 20% NP replacement for

Table 2 Table 3
Basic properties of the materials studied. Compressive strengths of mortars.

Sample code Cement (%) Mineral admixture (%) W/C SP Mortars Compressive strength [MPa]
(%)
NP SF LF 2 Days 7 Days 28 Days 90 Days 360 Days

OPC 100 – – 1 0.5 1 OPC 22.9 32.3 45.5 58.8 59.9

NP15 85 15 – 1.4 0.5 1.4 NP15 16.9 27.4 39.6 53.4 55.9
NP20 80 20 – 1.6 0.5 1.6 NP20 11.4 27.9 41.8 52.5 61.5
NP25 75 25 – 1.7 0.5 1.7 NP25 13.0 27.1 37.8 48.6 56.5
SF5 95 – 5 1.3 0.5 1.3 SF05 18.3 34.0 46.9 59.9 62.9
SF7.5 92.5 – 7.5 1.4 0.5 1.4 SF7.5 18.9 32.9 49.9 60.4 63.2
SF10 90 – 10 1.4 0.5 1.6 SF10 19.2 32.0 47.2 63.6 65.8
LS5 95 – – 1.6 0.5 1.6 LF05 19.2 37.1 38.9 42.6 48.1
LS10 90 – – 1.7 0.5 1.7 LF10 17.6 33.6 42.8 44.9 46.8
LS15 85 – – 1.8 0.5 1.8 LF15 16.5 28.4 38.5 39.3 44.4
 Y. Senhadji et al. / Powder Technology 254 (2014) 314–323
OPC increased compressive strength a little at 360 days. However,
reductions were 9, 11 and 17% for 15, 20 and 25% NP replacement of
OPC, respectively, at 90 days. Reductions at early ages due to NP
replacement increased with increasing NP level and decreased with
curing time.
This is consistent with previous studies, which have shown that NP
does not contribute notably to the early strength development of
cementitious systems [18]. This could be explained by the slowness of
the pozzolanic reactions between the glassy particles in the NP and
the C–H. The pozzolanic reaction with C–H occurs slowly because C–H
has first to be produced as a result of the hydration of Portland cement.
Kaid et al. [25] studied the pozzolanic activity of the NP used here.
Algerian volcanic pozzolan consumed 12 g of C–H per 100 g of pozzolan
at 20 °C in 28 days and 22 g in 360 days. However, because the contin-
uation of the pozzolanic reaction and the formation of secondary C–S–H
and C–A–S–H enhanced the structure of the hydrated cement paste and
the paste–aggregate interface (ITZs, Interfacial Transition Zone) and
decreased the capillary porosity of the mortar, a greater degree of hy-
dration was achieved, resulting in strengths at 360 days of age that
were comparable to those of ordinary Portland cement specimens.
This evidence can also be correlated to the pore size distribution results.
3.1.2. Effect of silica fume
The results showed that SF replacement at the levels of 5, 7.5 and
10% also decreased the compressive strength of mortars at 2 and
7 days. Results of numerous studies have indicated that SF slows the
rate of hardening and reduces the early compressive strength of con-
crete [26]. After 28, 90 and 360 days, addition of silica fume produced
the highest increase in strength when compared to the other SCMs
(natural pozzolan and limestone). For example, with the addition of
5%, 7.5%, and 10% of SF, the compressive strength increased by 5%, 6%,
and 10%, respectively at the mortar age of 360 days. Addition of 10% of
silica fume by mass was the optimum, which produced the highest
increase in strength.
The increase in strength of silica fume concrete could be attributed
to the improvement in the bond between the hydrated cement matrix
and the aggregate. This is due to the conversion of calcium hydroxide,
which tends to form on the surface of aggregate particles to calcium
silicate hydrate [9].
3.1.3. Effect of limestone filler
The results showed that the compressive strength decreased with
increasing LF content at all ages. In Table 3, it can be observed that the
compressive strength of control OPC mortar achieved values of 22.9,
32.3, 45.5, and 59.9 MPa at the ages of 2, 7, 28 and 360 days respectively.
The compressive strength of mortars (LF5, LF10, LF15) containing
limestone filler ranged from 19.2 to 16.5 MPa at the age of two days,
37.1 to 28.4 MPa at the age of seven days, 38.9 to 38.5 MPa at the age
of 28 days, and 48.1 to 44.4 MPa at the age of 360 days. Numerous
studies have indicated that LF slows the rate of hardening and reduces
the early compressive strength of concrete and mortar [27,28].
The findings of the present study are in contrast with those reported
by Nehdi et al. [29]who shown that limestone filler does not only
physically participate in the hydration process by providing additional
surface area for nucleation and growth of hydration products, limestone
filler also has chemically reactivity [29]. Sellevold et al. [30] found that
the presence of nucleation sites accelerates the hydration of the cement.
Bonavetti et al. [31] also showed that 10–20% limestone replacement
at various water to cement ratio (0.25–0.50), the degree of cement paste
hydration was highly more rapid during the first 7 days in the higher
W/C ratio with limestone addition [31].
3.2. Mercury intrusion porosimetry
Porosity is the characteristic property that is influenced by all the
parameters related to mixture proportions and, consequently, it acts
317
as a detailed record of these influences. A major factor that significantly
affects both total porosity and pore size distribution of a cement matrix
is the partial replacement of Portland cement by mineral admixtures,
although the nature of the filler seems to have an influence [32,33].
The cumulative pore volume at 360 days versus pore radius for four
mortar mixtures made with different binder types is shown in Fig. 2. The
porosities of the samples after 360 days of maturation decreased to 47%,
14% and 18% of the value of OPC mortar, for samples replaced with
NP20, SF7.5 and LF10 respectively. The total pore volume of mortar
made with Portland cement during the 12 months was higher than
that of mortar made with binary binders. The mean pore size of pure
cement composite was 60 nm, while those of NP20, SF7.5 and LF10
mortar composites were 16, 36 and 31 nm, respectively. It can be seen
that the pozzolan reduced the mean pore size of the products.
The investigation of porosity and pore size distribution of the mortar
mixtures, OPC, NP20, SF7.5 and LF10, selected for this study demon-
strated that Portland cement mortar mixtures were more porous than
mortar containing mineral admixtures. In fact, LF10 and NP20 mortar
had a large proportion of pores smaller than 5 nm compared to Portland
cement mortar, even the volume of the pore size population between
2.5 and 125 nm was more noticeable for NP20 mortar than for SF7.5
and OPC mortars. In addition, the average pore radius of OPC concrete
was higher than that of binder with mineral admixtures. The real effect
of SCMs is to refine pore systems over time and, therefore, to increase
the total volume of pores smaller than 50 nm. It has been shown that
the largest pore population of mortar with 7.5% of silica fume has a
pore radius range between 25 and 50 nm. Furthermore, experimental
results for concrete mixtures with different binder types reveal that
using mineral admixtures, especially pozzolanic materials, in concrete
is a more effective way of minimizing total porosity and refining pore
structure than extending the curing period [34].
Najimi et al. [1] found that natural zeolite (natural pozzolan)
contributed significantly to reduce porosity of concrete due to the filling
effect of ultrafine natural zeolite as well as its pozzolanic reaction.
3.3. Acid attack
Figs. 3 and 4 show the results of weight change versus time for
mortar and composite specimens exposed to 5% HNO3 or 5% H2SO4
solutions for 1, 14, 28 and 56 days. There is a permanent percentage
loss in weight with time for the mortar and composites immersed in
the different acid solutions.
3.3.1. HNO3 attack
Fig. 3 shows that, during 56 days' exposure to 5% nitric acid solution,
all cement–NP composites corroded more slowly in acid solution than
the control specimen of plain cement mortar. For the composites
containing 15%, 20% and 25% of pozzolan, the loss in weight was lower
than that of the corresponding OPC mortar by 13%, 22% and 15% respec-
tively. For specimens containing SF, there was significant change in the
weight loss. The losses in weight were reduced by as much as 25%, 23%
and 30% with the addition of 5%, 7.5% and 10% silica fume respectively.
Similar conclusions were drawn by Safwan and Abou-Zeid [26] from
durability studies of a concrete made with Portland cement and 40% of
SF. The chemical resistance to nitric acid attack of SF and NP mortars
could be attributed to the following: (a) pozzolanic materials reduce
permeability [35], thereby preventing the ingress of water and transport
of alkali and hydroxyl ions; (b) pozzolanic materials increase strength
and stiffness, resulting in better resistance to cracking and less expan-
sion; (c) replacing a portion of cement with a less-alkaline pozzolanic
material decreases the total amount of alkali present; and (d) pozzolans
react with calcium hydroxide to form calcium silicate hydrate with
a low CaO/SiO2 ratio. Formation of C–S–H depletes C–H and the low
C/S ratio enables the entrapment of alkalis, both of which reduce
the amount of hydroxyl ions available to participate in the chemical
attack [10].
318 Y. Senhadji et al. / Powder Technology 254 (2014) 314–323

50
-3 -3 -3 Total porosity
(2.5-5)10 (5-12.5)10 (12.5-25)10
45 -3 -3 -3 OPC 9.66%
(25-50)10 (50-62.5)10 (62.5-125)10
NP20 5.07%
0.125-0.25 0.25-0.5 0.5-2.5 2.5-5
SF7.5 8.32%
40 5-12.5 12.5-25 25-50 50-100 LF10
LS10 7.95%

35

Relative volume (%)

30

25

20

15

10

OPC NP20 SF7.5 LF10

Pores radius ranges (Micron)

Fig. 2. Differential pore size distribution and total porosity of the samples OPC, NP20, SF7.5 and LF10 after 360 days.

The LF mortars showed similar reduction in loss of weight to that of
the NP mortars, but with slightly lower values at all tested ages.
According to the results shown in Fig. 3, after 56 days of immersion,
the specimens of mortars with 5, 10 and 15% LF had 12, 13 and 12.5%
mass losses respectively, while the OPC reference mortar had the
greatest mass loss, of 14.6%.
3.3.2. H2SO4 attack
In the case of sulphuric acid illustrated in Fig. 4, the comparison of
loss in weight between mortar containing NP and control mortar
shows that all batches of pozzolanic mortar had lower mass loss than
the control at the ages of 1, 14, 28 and 56 days. The decrease of loss in
weight of the mortar due to addition of NP can be attributed to the
chemical properties and pozzolanic activity of the NP.
As explained above, SF significantly improved the compressive
strength and reduced the porosity of concrete. Accordingly, it can be
argued that the mechanisms that improved compressive strength and
reduced porosity were not successful in improving the resistance of
concrete to attack by sulphuric acid at relatively high concentration. In
other words, converting C–H into calcium silicate hydrate and refining
the porosity of concrete did not improve its resistance to severe
sulphuric acid attack. The findings of the present study are in agreement
with those reported by Torii and Kawamura [36], who found that partial
replacement of OPC by SF and fly ash could not effectively prevent the
acid deterioration of concrete specimens immersed in H2SO4.
Although natural pozzolan and silica fume have comparable effects
on the physical properties of hardened concrete, the chemical effect of
NP on the composition of the hydrated cement paste differs from that
of SF. This is due to the creation of calcium aluminate hydrates
(C2ASH8, C4AH13, C3AH6) in the NP-modified mortar. It is believed that
calcium aluminate hydrates may be more chemically stable in highly
acidic environments than calcium silicate hydrates. This is indirectly

16
OPC
14
NP15
12 NP20
Weight loss (%)

10 NP25

SF5
8
SF7.5
6 SF10

4 LF05

LF10
2
LF15
0
1 Day 14 Days 28 Days 56 Days
Different mortar mixtures immersed in 5% HNO3

Fig. 3. Weight loss of specimens under nitric acid exposure.

 Y. Senhadji et al. / Powder Technology 254 (2014) 314–323 319

14
OPC
12
NP15

NP20
10

Weight loss (%)

NP25
8
SF5

SF7.5
6
SF10
4
LF05

LF10
2
LF015
0
1 Day 14 Days 28 Days 56 Days
Different mortar mixtures immersed in 5% H2SO4

Fig. 4. Weight loss of specimens under sulfuric acid exposure.

supported by findings in the literature that high alumina cements and
calcium aluminate cements improve the resistance of concrete to
sulphuric acid attack, possibly due to the high aluminate contents of
these cements [6].
From Fig. 4, it can be seen that mortars with 5, 10 and 15% LF that are
immersed in H2SO4 are more resistant to weight loss. Specimens with
the highest percentage of limestone (15% CaCO3) had the minimum
weight loss of 6.3%. The better resistance of cements with limestone
additions to acid attacks (Figs. 3 and 4) can be attributed to three impor-
tant factors: (i) LF was much finer than OPC. It filled the micropores in
cement mortar and the ability of mortar to resist nitric and sulphuric
acids attack was improved by the reduced permeability and porosity
(discussed above—Fig. 2), (ii) the presence of a high calcium carbonate
(CaCO3) content increased the capacity of limestone mortars to
consume more aggressive acid, and (iii) the decreased proportion of
cement reduced the portlandite (C–H) content.
However, by comparing the aggressiveness of the sulphuric acid
medium and the nitric acid medium, we conclude that, for the same
concentration, the overall degree of attack tended to be more severe
in nitric solution than sulphuric solution. This was probably due to the
fact that, in the case of nitric acid attack, the calcium salt (calcium ni-
trate) formed is very soluble in water (266.0 g/100 g water at 20 °C),
leading to higher mass loss. On other hand, in the case of sulphuric
acid attack, the calcium salt (gypsum) formed has very low solubility
in water (0.22 g/100 g water at 20 °C) [37]. Rendell and Jauberthie
[14] reported that, when cement mortar was exposed to sulphuric
OPC
1.2
LF15 1 NP15
0.8
0.6 Compressive Strength
LF10 0.4 NP20 (28 Days)
0.2
Weigth Loss (HNO3)
0 the white-coloured part of the layer consisted mostly of SiO2.aq and
after 56 days
LF5 NP25 Weigth Loss (H2SO4)
after 56 days
SF10 SF5
SF7.5
Fig. 5. Relation between compressive strength and chemical resistance.
acid, a dense layer of gypsum was formed, which was capable of
retarding the deterioration process by acting as a surface sealing layer.
3.4. Correlation between compressive resistance and acids resistance
Fig. 5 illustrates the relationship between compressive strength after
28 days and mass loss after 56 days in acidic solution. In fact, there is not
a direct relationship between the compressive strength and the mass
loss for the mortar specimens immersed in HNO3 solutions. In the
same way, the external state of degradation of the samples after
exposure to the sulphuric acid attack was not a good indicator of their
compressive strength. In fact, there was a divergence between the loss
of mass and the compressive strength.
The trend observed for chemical resistance to HNO3 and H2SO4
solutions is affected by multiple factors and not only by compressive
strength. The following examples are noteworthy: 1) the concentration
and pH of the aggressive solution, 2) the mobility of the solution,
3) the w/c of the binders, 4) the type of cement, and 5) the mineral
admixtures, among others.
3.5. Visual inspection
An optical microscope was used to observe the deteriorated
corroded surface of specimens at regular intervals. Photographs of cor-
roded specimen surfaces taken after 56 days of exposure are presented
in Fig. 6. For specimens in nitric and sulphuric acid respectively, the
deterioration of the surface was seen to increase with time, though
the extent of deterioration among the ten series of samples could not
be easily differentiated through visual inspection. A layer of corrosion
products with cracks and pores visible under an optical microscope
was observed to form on the surface of immersed specimens early in
the first day of their immersion in the solution of nitric acid. The colour
of the corroded cementitious material changed. It was white in the
external zone adjacent to the surface of the corroded layer; underneath,
there was a thin rust-brown zone (Fig. 6(A)). Pavlik [37] reported that
the rust-brown colour related to the presence of ferric oxide in the
deeper zone. The thickness of the layer of corrosive products increased
progressively over time. This resulted in a lower diffusion resistance of
the corroded layer and a greater rate of corrosion caused by nitric acid
was observed. The difference between the corrosion rates of control
mortar and blended cement mortars in nitric acid was noticeable
(change in volume). After 56 days, we found that blended mortars
with SCMs behaved better than Portland cement mortar.
320 Y. Senhadji et al. / Powder Technology 254 (2014) 314–323

E
E
G

Relative
OPC

NP15

NP20

NP25
0
(A) Specimen after 8weeks in Nitric acid 5 10 15 20 25 30
2 THETA-Scale

Fig. 7. XRD patterns of OPC, NP15, NP20 and NP25 after 360 days of immersion in tap
water.

(Figs. 7 and 8) and indicated the effectiveness of SF and NP as pozzolanic

(B) Specimen after 8 weeks in Sulphuric acid
Fig. 6. Deterioration of specimens after 8 weeks of immersion in HNO3 and H2SO4
solutions.
Sulphuric acid exposure caused extensive formation of gypsum in
the surface regions of the specimens, as could be seen from subsequent
X-ray analysis. Visually, we clearly saw the degraded state of the differ-
ent samples after 7 weeks of immersion in 5% H2SO4 solutions
(Fig. 6(B)). Mortar specimens with silica fume exposed to sulphuric
acid indicated that the surface underwent significant deterioration
since it became highly porous and rich in weak crystalline structures
(calcium sulphate). However, the difference between the corrosion
rates of NP and LF mortar specimens in sulphuric acid was relatively
very small. The cores of the samples were not affected.
3.6. Mineralogical analysis of hardened mortar
additives, being consistent with previous reports [15,38]. Durning and
Hicks [15] reported that, in a mixture of 30% SF and 70% PC, C–H
disappeared entirely.
Mehta and Monteiro [2] found that C–H also decreased when SF and
FA (fly ashes) were used together. As shown in sections (3–1 and 3–2)
the diffractogram test results were in good agreement with those of the
compressive strength and MIP tests. X-ray analysis patterns of lime-
stone mortars did not show any change compared to control mortar.
In order to identify the deterioration products of the acid attacks
after 56 days of exposure in 5% HNO3 and 5% H2SO4 solutions, samples
from the surface layer were analysed by X-ray diffraction. Figs. 9 and
10 show the powder XRD patterns of OPC, NP20, SF7.5 and LF10 mortars
immersed in nitric and sulphuric acids respectively. It can be seen from
these diagrams that, after 56 days, ettringite and quartz remained
persistent in mortars with SCMs. We also note that portlandite
was completely decomposed. The superposition of XRD patterns
(Fig. 9—after HNO3 attack) (Fig. 10—after H2SO4 attack) shows that
there is disappearance of crystalline hydrated phases and appearance
of new ones. They confirm the appearance of a large quantity of gypsum
P
E
E G
OPC
Relative

3.6.1. X-ray diffraction

The hydration products were mineralogically determined by X-ray SF5
diffraction. The results are given in Figs. 7 and 8. From the X-ray analysis
patterns of the ten hydrated mortars, the expected crystalline hydration
products were clearly visible and it was determined that the mortars SF7.5
had hydrated, giving cement hydration products such as portlandite
(P), calcium silicate hydrate and ettringite (E) in all samples. Ettringite
SF10
was apparent in all composite cement systems. This front of ettringite
0
was therefore composed of the primary ettringite formed when the 5 10 15 20 25 30
cement set. Not surprisingly, the amount of calcium hydroxide present 2 THETA-Scale
in the SMCs systems was seen to vary between the NP15–NP25 and
SF5–SF10 mortars. This was more obvious in the NP25 and SF10 mortars Fig. 8. XRD patterns of OPC, SF5, SF7.5 and SF10 after 360 days of immersion in tap water.
 Y. Senhadji et al. / Powder Technology 254 (2014) 314–323 321

P
E
E CN
E OPC
Relative

NP20

SF7.5

LF10
0
5 10 15 20 25 30
2 THETA-Scale

Fig. 9. XRD patterns of OPC, NP20, SF7.5 and LF10 mortars after 56 days of immersion in 5%
HNO3.

(G) on the specimens that were exposed to the sulphuric acid, and a
trace of calcium nitrate (Ca(NO3)2) for the specimens that were exposed
to nitric acid. The small quantity of calcium nitrate is due to its high
solubility in water. The washing of each specimen after the period of
immersion removed almost all of this salt from the surface.

3.6.2. Scanning electron microscopy and EDS

Scanning electron microscopy (SEM) and EDS were also carried out
to identify the products formed through attack by solutions of sulphuric
acid and nitric acid and the results are exhibited in Figs. 11 and 12, re-
spectively. Fig. 11(A) and (B) presents the SEM photos of OPC and
NP20 samples exposed to a 5% solution of sulphuric acid. In addition,
an EDS spectrum was obtained for the sample at a selected spot.
Fig. 11(A), taken around a crack on the surface of the core sample
shows deposition of a large number of needle-like crystals (gypsum),
smaller than 10 μm in diameter and 20–50 μm in length, covering the
mortar surface. In the pores or surface cracks, there are many
club-shaped gypsum crystals as shown in Fig. 11(B). The formation of

G
Fig. 11. Elemental analysis obtained for OPC (A) and NP20 (B) samples and SEM image of
the external surface obtained after 8 weeks of immersion in 5% sulphuric acid.

E G
E P gypsum leads to the cracking, spalling, and decomposition of mortar
E
Relative

OPC
exposed to a sulphuric acid environment. EDS spectra of these needles
showed major elements such as calcium (Ca), oxygen (O) and sulphur
(S), confirming the identification of gypsum formation. The images
NP20 and elemental analysis obtained by SEM corroborated some of the
results discussed above.
In the case of HNO3 attack, the damaged OPC mortar was selected as
SF7.5 an example to observe the microstructure of deterioration and the SEM
images are shown in Fig. 12. No portlandite crystals exist in the acid-
attacked mortar. It is clear that exposure to a solution of nitric acid gen-
erates many more pores and micro-cracks in the mortar specimens.
LF10
0 4. Conclusions
5 10 15 20 25 30
2 THETA-Scale
This study investigated the effect of local SCMs (especially natural
Fig. 10. XRD patterns of OPC, NP20, SF7.5 and LF10 mortars after 56 days of immersion in pozzolan) on the resistance of mortar to sulphuric and nitric acid attacks
5% H2SO4. and on its compressive strength, X-ray analysis microstructures and
322 Y. Senhadji et al. / Powder Technology 254 (2014) 314–323
Fig. 12. SEM images obtained for the external porous texture of OPC sample after being
attacked by the 5% nitric acid solution, A (x250) and B (x45).
porosity. Based on the test data and analyses of the results presented in
this paper, the following conclusions can be drawn.
• From the test results it is deduced that the additions of silica fume and
natural pozzolan generally enhance the mechanical behaviour of
concrete at late curing ages by refinement of microstructure and
transforming C–H into C–S–H or C–A–S–H through the pozzolanic
reaction.
• In contrast, natural pozzolan reduces the compressive strengths of
mortar at all levels of replacement at early ages compared to the
control mortar. Limestone mortars have lower compressive strength
than the Portland cement mortar and mortar with pozzolanic additions.
• Results of loss in weight after acidic attacks indicate generally superior
performance of natural pozzolan and limestone mortars compared to
the respective control mortar and mortar that incorporates silica-
fume, especially after exposure to sulphuric acid.
• There are fundamental differences in the pore system characteristics of
Portland cement and blended cement mortars. Mortar with Portland
cement generally has a higher total porosity than that with blended ce-
ment. Within the hydration process, the reduction of larger volume
pores is more marked for mortar containing mineral admixtures than
for mortar with Portland cement. Furthermore, this reduction is more
significant when the admixtures are pozzolanic. It was also found that
the incorporation of mineral admixtures considerably reduced the aver-
age pore radius and increased the number of smaller pores.
• Using instrumental analysis such as XRD and SEM, the formation of
gypsum is clearly observed to be the main cause of the damage to the
cement matrix in the case of sulphuric attack. This was also confirmed
through visual observation.
Acknowledgments
The financial support of the Ministry of Higher Education and
Scientific Research of Algeria, under the grant Accord C.N.E.P.R.U, Refs.
J0401920110016 and J045520120005, is gratefully acknowledged. The
authors greatly appreciated the technical support from the LMDC
Laboratory of INSA Toulouse.
References
[1] M. Najimi, J. Sobhani, B. Ahmadi, M. Shekarchi, An experimental study on durability
properties of concrete containing zeolite as a highly reactive natural pozzolan,
Constr. Build. Mater. 35 (2012) 1023–1033.
[2] P.K. Mehta, P.J.M. Monteiro, Concrete: Microstructure, Properties, and Materials,
Second edition, McGraw-Hill Companies, New York, USA, 1993, p. 548.
[3] A.M. Neville, Properties of Concrete, Fourth edition John Wiley & Sons, Inc., New
York, USA, 1996. (pp. 844).
[4] A.A.A. Hassan, M. Lachemi, K.M.A. Khandaker Hossain, Effect of metakaolin and silica
fume on the durability of self-consolidating concrete, Cem. Concr. Compos. 24 (3)
(2012) 801–807.
[5] R. Mejia de Gutiérrez, Effect of supplementary cementing materials on the concrete
corrosion control, Rev. Metal Madrid Vol. Extr. (2003) 250–255.
[6] M.T. Bassuoni, M.L. Nehdi, Resistance of self-consolidating concrete to sulfuric acid
attack with consecutive pH reduction, Cem. Concr. Res. 37 (2007) 1070–1084.
[7] S. Goyal, M. Kumar, D.S. Sidhu, B. Bhattacharjee, Resistance of mineral admixture
concrete to acid attack, J. Adv. Concr. Technol. 7 (2) (2009) 273–283.
[8] Y. Senhadji, G. Escadeillas, H. Khelafi, M. Mouli, A.S. Benosman, Evaluation of natural
pozzolan for use as supplementary cementitious material, Eur. J. Environ. Civ. Eng.
16 (1) (2012) 77–96.
[9] E. Hewayde, M.L. Nehdi, E. Allouche, G. Nakhla, Using concrete admixtures for
sulphuric acid resistance, Proc. Inst. Civ. Eng. Constr. Mater. 160 (CM1) (2007)
25–35.
[10] K. Baltakys, R. Jauberthie, R. Siauciunas, R. Kaminskas, Influence of modification of
SiO2 on the formation of calcium silicate hydrate, Mater. Sci. Pol. 25 (3) (2007)
663–670.
[11] A. Bertron, J. Duchesne, G. Escadeillas, Accelerated tests of hardened cement pastes
alteration by organic acids: analysis of the pH effect, Cem. Concr. Res. 35 (2005)
155–166.
[12] O. Oueslati, J. Duchesne, The effect of SCMs and curing time on resistance of mortars
subjected to organic acids, Cem. Concr. Res. 42 (2012) 205–214.
[13] A. Hidalgo, C. Andrade, C. Alonso, Role of alkaline reserve in the acidic resistance of
cement pastes, Workshop on the Role of Calcium Hydroxide in concrete, The
American Ceramic Society, Anna Maria Island, Florida, 2001. 93–112 (Nov. 1–3).
[14] F. Rendell, R. Jauberthie, The deterioration of mortar in sulphate environments,
Constr. Build. Mater. 13 (6) (1999) 321–327.
[15] T.G. Durning, M.C. Hicks, Using microsilica to increase concrete's resistance to
aggressive chemicals, Concr. Int. 13 (3) (1991) 42–48.
[16] P.K. Mehta, Studies on chemical resistance of low water/cement ratio cements, Cem.
Concr. Res. 15 (6) (1985) 969–978.
[17] J. Monteny, N. De Belie, L. Taerwe, Resistance of different types of concrete mixtures
to sulfuric acid, Mater. Struct. 36 (258) (2003) 242–249.
[18] M. Ghrici, S. Kenai, M. Mansour-Said, A. El-Hadj Kadri, Some engineering properties
of concrete containing natural pozzolana and silica fume, J. Asian Archit. Build. Eng.
5 (2) (2006) 349–354.
[19] ASTM C267, Standard specification for chemical resistance of mortars, grouts, and
monolithic surfacings and polymer concretes 1997, Annual Book of American
Society for Testing Materials, Philadelphia, V 4.05, 1997.
[20] NA 442, Norme Algérienne, [Algerian Standards] Liants Hydrauliques—Ciments
Courants, Composition, Spécification, IANOR, Algiers, 20032003.
[21] ASTM C 618–01, Standard specification for coal ash and raw or calcined natural
pozzolan for use as mineral admixture in concrete, Annual book of ASTM, Concrete
and Aggregates, vol. 04.022002. (2002).
[22] ACI Committee 226, Silica fume in concrete, American Concrete Institute, Detroit
USA, ACI Mater. J. 84 (2) (1987) 158–166.
[23] Z. Guemmadi, M. Resheidat, H. Chabil, B. Toumi, A Better Use of Limestone Filler in
Concrete, Proceedings of International Conference on Construction and Building
Technology ICCBT2008. 307–318.
[24] ASTM C 230, Standard Specifications for Flow Table for Use in Tests of Hydraulic
Cement, the American Society of Testing and Materials, vol. 04.012008.
[25] N. Kaid, M. Cyr, S. Julien, H. Khelafi, Durability of concrete containing a natural poz-
zolan as defined by a performance-based approach, Constr. Build. Mater. 23 (2009)
3457–3467.
[26] A.K. Safwan, M.N. Abou-Zeid, Characteristics of silica-fume concrete, J. Mater. Civ.
Eng. (ASCE) 6 (3) (1994) 357–375.
[27] S.T. Lee, R.D. Hooton, H.S. Jung, D.H. Park, C.S. Choi, Effect of limestone filler on the
deterioration of mortars and pastes exposed to sulfate solutions at ambient
temperature, Cem. Concr. Res. 38 (2008) 68–76.
 Y. Senhadji et al. / Powder Technology 254 (2014) 314–323
[28] P. Turgut, Masonry composite material made of limestone powder and fly ash,
Powder Technol. 204 (2010) 42–47.
[29] M. Nehdi, S. Mindess, P.C. Aitcin, Rheology of high-performance concrete: effect of
ultra fine particles, Cem. Concr. Res. 28 (5) (1998) 687–697.
[30] E.J. Sellevold, D.H. Bager, E. Klitgaard-Jensen, T. Knudsen, “Silica Fume-Cement
Pastes: Hydration and Pore Structure”, Condensed Silica Fume in Concrete, Institutt
for Bygningsmateriallære, Norges Tekniske Høgskole, Universitetet i Trondheim,
Trondheim, Norway, 1982. 19–50 BML 82.610, February.
[31] V. Bonavetti, H. Donza, V. Rahhal, E. Irassar, Influence of initial curing on the proper-
ties of concrete containing limestone blended cement, Cem. Concr. Res. 30 (5)
(2000) 703–708.
[32] P. Diederich, M. Mouret, A. de Ryck, F. Ponchon, G. Escadeillas, The nature of
limestone filler and self-consolidating feasibility—relationships between physical,
chemical and mineralogical properties of fillers and the flow at different states,
from powder to cement-based suspension, Powder Technol. 218 (2012) 90–101.
323
[33] I.B. Celik, The effects of particle size distribution and surface area upon cement
strength development, Powder Technol. 188 (3) (2009) 272–276.
[34] J.F. Martirena, R.L. Day, B. Middendorf, M. Gehrke, L. Martínez, J.M. Dopico,
Lime-pozzolan binder as a very fine mineral admixture in concrete, International
symposium on ultra high performance concrete2004. (Published by M. Schmidt, E
Fehling, C Geisenhanslüke (3) 1–12).
[35] S. Shi Hui, W.Xu. Bi, C. Zhou Xiao, Influence of mineral admixtures on compressive
strength, gas permeability and carbonation of high performance concrete, Constr.
Build. Mater. 23 (2009) 1980–1985.
[36] K. Torii, M. Kawamura, Effects of fly ash and silica fume on the resistance of mortar
to sulphuric acid and sulphate attack, Cem. Concr. Res. 24 (2) (1994) 361–370.
[37] V. Pavlik, Effect of carbonates on the corrosion rate of cement mortars in nitric acid,
Cem. Concr. Res. 30 (2000) 481–489.
[38] A.K.H. Kwan, J.J. Chen, Adding fly ash microsphere to improve packing density,
flowability and strength of cement paste, Powder Technol. 234 (2012) 19–25.