Construction and Building Materials 197 (2019) 681–692

Contents lists available at ScienceDirect

Construction and Building Materials

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

Sulphuric acid exposure of conventional concrete and alkali-activated

concrete: Assessment of test methodologies
Lei Gu, Terry Bennett, Phillip Visintin ⇑
School of Civil, Environmental and Mining Engineering, The University of Adelaide, SA 5005, Australia

h i g h l i g h t s

Compares the efficacy of indicators and measures of sulphuric acid degradation.

Considers impact of sulphuric acid degradation on OPC and alkali activated concrete.

Test methodology ensures consistent boundary conditions.

Results show no single universal indicator of resistance to sulphuric acid.

The same indicators cannot be applied to OPC and alkali activated concrete.

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

Article history: Biogenic sulphuric acid corrosion can lead to significant degradation of Ordinary Portland Cement Concrete
Received 13 July 2018 (OPCC) and cause significant damage to wastewater infrastructure. Recent studies have suggested that
Received in revised form 19 November 2018 alkali-activated concrete (AAC) has improved durability compared to OPCC in a range of aggressive environ-
Accepted 22 November 2018
ments. In this study, the resistance of both OPCC and class F fly ash based AAC to sulphuric acid is investi-
Available online 30 November 2018
gated in order to develop a simple laboratory test method, capable of indicating the susceptibility of both
materials to biogenic acid corrosion. Tests that provide an indication of the susceptibility of cementitious
Keywords:
materials to corrosion (change of mass, length, compressive strength, and cross-section) are undertaken
Concrete degradation
Sulphuric acid corrosion
under exposure to 1% (pH = 1.0) and 3% (pH = 0.52) sulphuric acid concentration for 495 and 112 days
Microstructure respectively, to establish the ability to accelerate testing. Further, characterization tests are conducted to
Alkali-activated concrete investigate the change of specimens at the microstructural level. The indicators of susceptibility to corro-
Assessment of test methodologies sion are discussed based on their potential advantages and limitations as well as application to both OPCC
and Class F fly ash based AAC. The results show that although the alkali-activated matrix partially retains
strength after sulphuric acid attack, the coarse aggregates in the alkali-activated concrete are more prone
to attack than those in an OPC matrix. It is also shown that using a higher concentration sulphuric acid accel-
erates the degradation of specimens sufficiently to perform an indicator test within 90 days.
Ó 2018 Elsevier Ltd. All rights reserved.

1. Introduction The mechanism of biogenic sulphuric acid corrosion is complex,

and its replication in a laboratory environment requires sophisti-
One of the most rapid cases of concrete degradation is that cated experimental apparatus and bacterial growth. Although pos-
which occurs in sewage collection systems. In these systems the sible in a multidisciplinary university research environment [5],
combination of bacteria, moisture, and oxygen in the atmosphere the equipment and skills required are not readily available in a typ-
results in the production of biogenic sulphuric acid [1,2]. It has ical civil engineering materials testing laboratory. As an alterna-
been estimated that the annual maintenance costs associated with tive, Aiken et al. [6] suggest the direct exposure to silage effluent
biogenic corrosion are over €450 million in Germany and £85 mil- to avoid sophisticated equipment. While providing a direct indica-
lion in the UK; while in the United States it is expected that US$390 tion of performance, using this approach it may be difficult to pro-
billion dollars will be spent on the repair of wastewater infrastruc- duce a standardised testing regime that can be performed at
ture over the period from 2002 to 2022 [3,4]. different laboratories. As an alternative to the direct simulation
of the wastewater environment, previous studies have instead
⇑ Corresponding author. focused on the exposure of cementitious materials to chemical
E-mail address: [email protected] (P. Visintin).
sulphuric acid, in this way it is possible to expose materials to

https://doi.org/10.1016/j.conbuildmat.2018.11.166
0950-0618/Ó 2018 Elsevier Ltd. All rights reserved.
682 L. Gu et al. / Construction and Building Materials 197 (2019) 681–692

pH typical of wastewater environments but without the need for 2.1. Sample preparation
generating it through bacterial action [5,7–9].
As the primary focus of this study is to evaluate test methodologies, as opposed
In addition to simplification of the exposure regime, a simple, to direct comparisons of different mixes, the materials to be investigated have been
repeatable methodology is desirable. Candidate approaches, such chosen to be to be susceptible to attack, that is conventional concretes are manufac-
as length change, mass loss, corrosion depth, acid consumption tured using OPC rather than Sulphate Resisting Cement, and with crushed lime-
and reduction in alkalinity provide a means of indicating the dura- stone and river sand as coarse and fine aggregate respectively.
The complete chemical composition of the OPC is provided in Table 1, along
bility performance of Ordinary Portland Cement Concrete (OPCC),
with that of the fly ash used to manufacture the AAC. The activator used in the man-
but may not translate well to other cementitious materials. For ufacture of the AAC consisted of laboratory grade sodium silicate (Na2SiO3) and
example, the standard test methodology for acid resistance (ASTM 14 M sodium hydroxide (NaOH) which was pre-mixed in a ratio of Na2SiO3: NaOH
C 267 [10]) only covers silicate concrete and hydraulic materials, of 1.5:1. The sodium silicate solution used in activator had a SiO2 to Na2O ratio of
and is not specifically applicable to other concretes. 3.22 and was comprised of 62.5% water by the total weight. The total Na2O content
amounted to 6.5% of the binder weight.
Recently, alkali-activated concrete (AAC) has been suggested as The mix proportions of the (OPCC and the AAC are given in Table 2, in which the
an alternative to OPCC, with potential for enhanced durability AA concrete mix was designed according to Albitar et al. [14]. The water to cement
characteristics in civil engineering applications [11,12]. This ratios of OPCC_30 and OPCC_50 were 0.64 and 0.47, respectively and were designed
improved durability has been mainly attributed to the different according to Neville [15]. All the aggregates were used in saturated surface dry
condition.
chemical composition of the matrix material. For example, Izzat
After casting, the OPCC specimens were cured in their moulds at 20 ± 2 °C for
et al. [13] found that class C fly ash based geopolymer mortar 24 h before demoulding, after which they were cured in a fog room at a constant
had lower mass loss and compressive strength reduction when temperature of 22 °C for 28 days. The AAC specimens were cured in moulds at
samples were immersed in 3% sulphuric acid solution for 120 days 20 ± 2 °C for 48 h before demoulding, specimens were then covered using double
when compared to the OPC mortar. Thokchom et al. [11] also con- layer vacuum bags and cured in the fog room at 22 °C for 28 days. As a control
the 28 day compressive strength of all mixes was determined from compressive
cluded that class F fly ash based geopolymer mortar had good tests conducted on standard 100 mm  200 mm cylinders according to ASTM C39
resistance against 10% sulphuric acid for 18 weeks. [16] and was found to be 36.4 MPa, 57.8 MPa and 22.2 MPa for mixes OPCC_30,
The aim of this programme of work is to assess the applicability OPCC_50 and AAC respectively.
of a simplified exposure regime and commonly applied durability
indicators to vastly different binders. Further, it is intended that 2.2. Test methodology
a test methodology be developed that can capture degradation in
For testing, the specimens were laid on a plastic grid and submerged in a bath of
a short period of time. While both OPCC and AAC are considered, 1% (w/w %) sulphuric acid solution. In order to manage the pH change, the sulphuric
direct comparison of the performance of OPCC and AAC is not acid solution to specimen ratio was chosen as 4.5 (suggested by ASTM C1012 [17] in
the intended focus as this would be highly dependent on the mate- sulphate attack test). The pH level was maintained at 1.0 ± 0.05, monitored by a
rial design. portable digital pH meter, and adjusted twice a week by the addition of 35% (w/
w %) sulphuric acid. The choice of 1% sulphuric acid was based on simulating a
This paper provides the details of the experimental program, pH that can exist in wastewater environments where severe degradation is
including material information, test procedures and instrumenta- observed [4,7]. The immediate exposure to 1% sulphuric acid is however accelerated
tion. The results of the experimental program are then presented, in comparison to field like conditions where the process of bacterial growth can
and the different measurement techniques are discussed in refer- take several years to reach this pH level [18].
Throughout the test period the solution was continuously circulated using
ence to their potential advantages and limitations. A detailed dis-
pumps to avoid the development of strong ion gradients, to assist with this a min-
cussion on the results is then presented, where the different imum spacing of 5 cm between specimens was maintained (Fig. 1(a)). To eliminate
measurement techniques are assessed based on their ability to rep- the influence of evaporation, lids was used to cover the test tubs. Fig. 1(b) shows the
resent changes in compressive strength. pH value change of OPC and AA concrete specimens during the entire immersion
period from which it can be seen that the pH monitoring resulted in small varia-
tions in pH throughout the testing period.
To provide an indication of sulphuric acid degradation, acid consumption, mass
loss, length change, pH change, corrosion depth and microstructural change were
2. Materials and methodologies all observed for a period of 495 days. As a direct measure of the impact of degrada-
tion the change in compressive strength is also quantified. The methodology for
Specimens are manufactured to provide an indication of sulphuric acid degra- each test is summarised as follows:
dation in terms of acid consumption, mass loss, length change, pH change, corrosion
depth and microstructural change. As a direct measure of the impact of degradation
Coarse aggregate samples were immersed in 1% sulphuric acid solution and EDX
the change in compressive strength is also quantified. analysis performed at 14 and 495 days to determine the chemical composition
In this section, information on the materials used to manufacture the concrete, of the reaction product.
details of test specimens and the test methodology employed are presented.

Table 1
Chemical composition of binders.

Oxide (wt%) SiO2 Al2O3 Fe2O3 CaO MgO Na2O K2O SO3 TiO2 P2O5 Mn2O3 LOI*
Cement 19.89 4.69 3.16 63.35 1.94 0.22 0.40 2.77 0.27 0.03 0.07 3.55
Fly ash 54.24 28.94 4.01 3.58 1.80 0.60 1.46 0.25 1.21 0.27 0.06 3.12

LOI: Loss on ignition.

Table 2
Mix proportion of OPC and AA concrete.

Mix ID Cement Fly ash Limestone Sand Water Superplasticizer Activator

(kg/m3) (kg/m3) (kg/m3) (kg/m3) (kg/m3) (kg/m3) (kg/m3)
CC_30 280 0 1110 740 180 0 0
CC_50 360 0 1110 740 171 1 0
AAC 0 424.8 1176 576 16.8 48 158.4
 L. Gu et al. / Construction and Building Materials 197 (2019) 681–692 683

Fig. 1. (a) Specimen arrangement and test setup; (b) pH value of OPC and AA concrete specimens under 1% sulphuric acid attack.

The microstructural change of OPCC and AAC specimens before and after sul-
The compressive strength of cylindrical specimens (100 mm in diameter and
phuric acid attack (at days 42, 142 and 495) was examined by scanning electron 200 mm in height) was determined according to ASTM C39 [16]. To ensure
microscopy (SEM) and energy dispersive x-ray (EDX) using a Phillips XL 30 the loading faces were parallel to the loading platens the ends of the specimens
instrument. For the analysis, samples were taken from the surface (4  5 mm that had been damaged by exposure to sulphuric acid were capped with a gyp-
depth) at the mid-height of a standard (100  200 mm) cylindrical specimen. sum paste (3 replicates were tested at ages 42, 142, 260, 375 and 495 days).
Samples were coated with platinum prior to SEM analysis, and EDX was per-
formed at an accelerating voltage of 15 kV.

The consumption of acid solution was recorded twice weekly, when the pH of 3. Test results of concretes under 1% sulphuric acid attack
the sulphuric acid solution was measured and adjusted accordingly.

Change in surface alkalinity was determined using a 1% Phenolphthalein solu- In this section, the change in physical, mechanical and
tion to identify the proportion of the cross section with a pH greater than 8.2. microstructural properties of the OPCC and ACC samples exposed
This solution was applied to a cylindrical disk of diameter 100 mm and thick-
to 1% sulphuric acid are presented.
ness 10 mm that was cut from the mid-height of a standard 100 mm  200 mm
cylinder. The change in alkalinity was measured at 42, 140, 260, 375, 495 days.
The depth of area with a pH of greater than 8.2 was determined by measuring 3.1. Visual appearance
the radius of unaffected area at nine locations with a digital microscope.

The change in mass after periods of exposure was measured using prism spec-
imens 75 mm by 75 mm in cross section and 285 mm in length. The mass of a Fig. 2 shows the visual appearance of OPCC and AAC specimens
single specimen was periodically measured (at days 9, 23, 51, 142, 260, 375, and after 495 days immersion in 1% sulphuric acid. In Fig. 2(a) it can be
495). In order to remove the influence of surface moisture, according to De Belie seen that the OPCC_30 specimen showed a mild degree of corro-
et al. [2], prior to weighing, the specimen was dried at 20 ± 2 °C for 48 h. sion, characterised by a small degree of expansion and spalling of

The length change of a prismatic specimen (75  75  285 mm) was deter-
mined by measuring the length of a single specimen, per concrete, between
cement paste and aggregates. The OPCC_50 specimen in Fig. 2(b)
the DEMEC gauges. According to the procedure outlined by De Belie et al. showed severe corrosion, characterised by significant expansion
[19], tests were conducted at ages 42, 142, 260, 375 and 495 days after first dry- and large areas in which the cement matrix and aggregates in
ing specimens at 20 ± 2 °C for 4 h. the surface spalled, leading to a very porous surface structure.

Fig. 2. Specimens after immersion in 1% sulphuric acid for 495 days: (a) OPCC_30; (b) OPCC_50; (c) AAC.
684 L. Gu et al. / Construction and Building Materials 197 (2019) 681–692

The increase in surface damage with compressive strength is also (a) showed a dense amorphous structure before immersion with
observed by Aiken et al. [20]. The AAC specimen Fig. 2(c) showed the typical calcium silica hydrate (C-S-H) chemical composition
the most significant corrosion with a large amount of spalling of expected from OPC [19] (Table 3). The rod like structure in area
the mortar and coarse aggregates, leading to an irregular shape B of Fig. 5(a) was identified as ettringite which formed as a result
and smaller size of the specimen when compared to that before of the high aluminium and sulphur concentrations identified in
immersion. Table 3. The sulphur content observed in the sample before immer-
sion is present due to the addition of gypsum during cement man-
3.2. Influence of sulphuric acid on coarse aggregate ufacture [21].
After 495 days of immersion in sulphuric acid, the OPC paste in
Fig. 3 presents the visual appearance of limestone aggregate Fig. 5(b) exhibited monoclinic crystals which were reasoned to be
before and after 1% sulphuric acid exposure. It can be seen that gypsum (CaSO42H2O), based on the atomic ratio of Ca:S:O
apart from the formation of a white leachate, the limestone aggre- obtained from the EDX analysis (see Table 3). Also of note in
gate after 14 days immersion in sulphuric acid in Fig. 3(b) showed Fig. 5(b) is that some amorphous structures are attached to the
no significant difference to that before immersion (Fig. 3(a)). Fig. 4 gypsum crystals, these can be inferred to be silica gels leached
shows the results of an XRD analysis of the white leachate which from the C-S-H during the sulphuric acid attack, as identified by
can be concluded to be predominantly gypsum. high silicon and oxygen concentrations detected by EDX. The
Although it is hard to distinguish the source of gypsum in OPCC leaching of silica gels from C-S-H under sulfuric acid attach is also
because of the presence of C–H and C-S-H in the matrix, the addi- observed by previous study [22]. The presence of a large amount of
tional gypsum formed on the surface of limestone aggregate would expansive gypsum crystals in the matrix resulted in the porous
grow together with the gypsum formed from cement paste. structure shown in Fig. 5(b).
Fig. 3(c) shows the severe degradation of limestone after The AA paste in area D in Fig. 5(c) also showed a dense amor-
495 days immersion in 1% sulphuric acid. This severe degradation phous structure before immersion with a typical sodium alumi-
can be attributed to the continued leaching of calcium which nosilicate hydrate (N-A-S-H) chemical composition [23] shown in
resulted in a very weak internal structure due to the high CaO con- in Table 3. The AA paste after 495 days immersion in area E in
tents of aggregate (43–46%). Fig. 5(d) showed a similar structure in morphology from that in
Fig. 5(c) before immersion. However, from the Table 3, the surface
3.3. Microstructural performance element analysis showed the atomic ratio of Si-to-Na increased
from 7.9 in area D to 15.2 in area E, indicating that the sodium ions,
3.3.1. Surface morphology and element change of OPC and AAC matrix which act as charge compensation cations of the aluminosilicate
Fig. 5 shows the SEM images of the OPCC and AAC concrete framework, were leached out and exchanged by hydrogen or
samples before and after 495 days sulphuric acid immersion, while hydronium ions from the solution. The atomic ratio of Si-to-Al
Table 3 shows the results of surface element analysis by EDX at the increased from 2.0 in area D to 3.1 in area E, indicating that the
areas marked A to E in Fig. 5. The cement matrix in area A in Fig. 5 tetrahedral Al from the aluminosilicate framework is dissolved

(a) Before immersion (b) 14 days immersion

(c) 495 days immersion

Fig. 3. Limestone aggregate before and after 1% sulphuric acid exposure.
 L. Gu et al. / Construction and Building Materials 197 (2019) 681–692 685

G tion, the OPCC samples show the formation of gypsum crystals

G - gypsum
(Fig. 6(b)). The development of gypsum crystals in the ITZ causes
G
the debonding of aggregates from the specimens shown in Fig. 2
G (a) and (b) and leads to significant decrease in the strength of the
G G specimen.
The ITZ with fine aggregates of an AAC sample after immersion
(Fig. 6(d)) showed a similar structure in morphology from that
before immersion in Fig. 6(c) suggesting AA paste remained in
good form under sulphuric acid attack. A similar observation has
also been made in previous studies [23].

0 10 20 30 40 50 60 70 80
2 (degrees) 3.3.3. Change of coarse aggregate interfacial transition zone in AAC
The SEM image in Fig. 7 show the presence of gypsum in the
Fig. 4. XRD pattern of the white leachate from limestone aggregate after 1%
sulphuric acid exposure.
vicinity of the ITZ surrounding the coarse aggregate in an AAC sam-
ple, as identified by the chemical composition obtained from EDX
analysis (Table 4). This expansive mechanism caused the signifi-
by breaking the Si-O-Al bonds. A similar observation has also been cant spalling of aggregates as seen in Fig. 2(c).
made in previous studies [4,24]. In area E, 2.1% sulphur was Unlike the OPCC degradation under sulphuric acid attack, the
detected, indicating some phases of sulphate generated in the AA degradation process in AAC does not generate any product that
paste after sulphuric acid attack. Whereas the 1.1% calcium content can act as a short-term protective barrier [4]. Therefore, when sul-
detected in area D decreased to zero in area E, indicating that the phuric acid penetrates through the AA fly ash matrix, the aggre-
small amount of calcium in the AA paste totally dissolved in the gates are directly expose to the acidic environment. This allows
sulphuric acid. The change of potassium and iron content in Table 3 for a more rapid dissolution of the calcium compounds from the
might be useful to further studies if not used explicitly in this coarse aggregate (limestone). These calcium ions subsequently
study. formed gypsum when the sulphate ions diffused in the matrix,
resulting in a porous and soft structure near the ITZ, leading to
the debonding of aggregates from the specimens shown in Fig. 2
3.3.2. Change of fine aggregate interfacial transition zone (c) and significant decrease in the strength of the specimen. Hence
Fig. 6 shows the change in interfacial transition zone (ITZ) it can be suggested from this experiment that investigations con-
between the paste and fine aggregates of OPCC and AAC samples. cerning the mechanisms affecting durability must focus not only
It can be seen from Fig. 6(a) and (c) that the ITZ of both unexposed on the performance (strength, porosity, diffusivity, etc.) of the
OPCC and AAC samples indicate a firm bond between the matrix paste but must also consider the influence of the aggregate and
and aggregate. After 495 days of exposure to sulphuric acid solu- the ITZ.

Fig. 5. SEM images and surface element analysis of OPC and AAC matrix before and after 495 days immersion.
686 L. Gu et al. / Construction and Building Materials 197 (2019) 681–692

Table 3
Surface element analysis in the marked area in Fig. 5 by EDX.

Element Area A Area B Area C Area D Area E

wt% at% wt% at% wt% at% wt% at% wt% at%
Oxygen 40.7 60.8 39.0 59.5 56.4 74.4 41.6 55.7 40.2 54.4
Sodium – – – – – – 3.6 3.3 2.2 2.0
Aluminium 3.0 2.7 4.8 4.3 – – 16.4 13.0 12.0 9.6
Silicon 14.0 11.9 1.3 1.2 – – 34.0 26.0 39.4 30.4
Sulphur – – 9.9 7.6 19.6 12.9 – – 3.2 2.1
Potassium 2.7 1.6 – – – – – – 1.4 0.8
Calcium 35.6 21.2 45.0 27.5 24.0 12.7 2.1 1.1 – –
Iron 4.0 1.7 – – – – 2.3 0.9 1.74 0.7

wt%: Weight percentage; at%: Atom percentage.

Fig. 6. Change of fine aggregate interfacial transition zones.

3.4. Consumption of sulphuric acid

Acid consumption per specimen provides a direct measure of

the sulphuric acid consumed, indicating the quantity of the chem-
ical reactions during the corrosion process [4]. Therefore, acid con-
sumption might be viewed as a direct chemical indicator of the
performance of the cementitious material when exposed to an
acidic environment.
Fig. 8(a) presents the consumption of sulphuric acid per speci-
men for all material types. The acid consumption of OPCC_30 spec-
imens increased slowly to 0.35 kg during the initial 260 days and
subsequently stabilised. The acid consumption of OPCC_50 speci-
mens increased rapidly to 0.43 kg during the initial 142 days, and
also subsequently increased at a reduced rate to 0.78 kg after
142 days. From these results, it is shown that OPCC specimens with
higher cement content (i.e. 380 kg/m3 in OPCC_50) have higher
acid consumption than those with lower cement content (i.e.
280 kg/m3 in OPCC_30). The more rapid consumption of acid can

Fig. 7. Coarse aggregate interfacial transition zone after 495 days immersion.
 L. Gu et al. / Construction and Building Materials 197 (2019) 681–692 687

Table 4
Surface element analysis in the marked area in Fig. 7 by EDX.

Element Oxygen Aluminium Silicon Sulphur Calcium

Area A wt% 53.2 1.3 4.3 16.7 24.4
at% 71.4 1.0 3.3 11.2 13.1

wt%: Weight percentage; at%: Atom percentage.

Fig. 8. Acid consumption of OPC and AA concrete specimens under 1% sulphuric Fig. 9. Depth of lost alkalinity of OPC and AA concrete specimens under 1%
acid attack. sulphuric acid attack.

be used to explain the apparent higher degradation for specimens 3.6. Mass change
with a higher cement content in Fig. 2.
The acid consumption of AAC specimens remained at 0 during Fig. 10 presents the mass loss of concrete specimens relative to
the first 14 days of exposure before increasing significantly. This the initial weight of the specimen prior to immersion, with three
phenomenon was attributed to the formation of barrier of carbon measurements taken at each time. It can be observed that the mass
dioxide on the surface of AAC specimens discussed in Gu et al. loss of OPCC_30 specimens increased to around 2% during the ini-
[25]. After this initial period acid consumption increased at a rela- tial 142 days and subsequently stabilised. The mass loss of
tively constant rate. Among all the tests, the AAC specimens con- OPCC_50 specimens increased almost linearly to around 5.5% dur-
sumed much more sulphuric acid than the OPCC specimens.

3.5. Depth of the area lost alkalinity

Using Phenolphthalein, the depth into the specimens at which

the pH of the concrete has reduced to less than 8.2 was measured
from the average of nine measurements on a sample. The results
from four unique samples, per concrete type, are shown in Fig. 9
at each exposure time.
Examining the results in Figs. 8 and 9 together, it can be seen
that the more sulphuric acid consumed by specimens, the larger
depth of the area lost alkalinity developed. Furthermore, the
OPCC_30 and OPCC_50 specimens showed a much smaller depth
of reduced alkalinity (less than 9 mm) than that of AAC (greater
than 40 mm) in Fig. 9. Lloyd et al. [26] also observed a large depth
of reduced alkalinity (over 10 mm after 84 days immersion) when
testing the resistance of class F fly ash based AAC against 1% sul-
phuric acid.
The AAC fly ash matrix has lower initial pH environment in its
pore fluid (i.e. 10.5–12.6 [27]) than that in hydrated OPCC matrix
(i.e. 13–14 [28]). The depth of the area lost alkalinity can only be
an indicator to represent the degradation when comparing the
same material and is unlikely to be a reasonable indicator for com- Fig. 10. Mass loss of OPC and AA concrete specimens under 1% sulphuric acid
paring the performance of OPCC and AAC. attack.
688 L. Gu et al. / Construction and Building Materials 197 (2019) 681–692

ing the entire test period. The mass loss of AAC specimens both OPCC_30 and ACC a strength gain occurs due to the continued
increased slowly during the initial 51 days and subsequently hydration of the OPCC and AAC fly ash matrix as discussed in Gu
rapidly increased to over 40% after 495 days. Ariffin et al. [29] et al. [25]. It can be observed that the compressive strength
observed a larger mass loss in OPC than that in AAC after OPCC_30 and OPCC_50 specimens lost almost 40% and 55% during
18 months immersion when testing the resistance of OPC and the entire test period when compared to that before immersion.
AAC against 2% sulphuric acid. The contradictory results in this The compressive strength of AAC specimens increased slowly dur-
study were mainly due to the spalling off of large pieces of concrete ing the initial 42 days and subsequently lost to around 50% initial
caused by the de-bonding of limestone coarse aggregate used in strength after 495 days. Ariffin et al. [29] observed a larger com-
the AAC mix. The difference in mass loss between the OPCC and pressive strength loss in OPC than that in AAC after during the
AAC specimens is an indicator of the degradation level of the spec- immersion period. The compressive strength change is highly
imens. That is, as shown in Fig. 2, OPCC_30 and OPCC_50 speci- mix dependent between the studies, especially when using the
mens only suffered from minor spalling compared to AAC which coarse aggregate susceptible to attack.
suffered a significant loss of aggregate due to damage to the ITZ.
Thus, mass loss may be a reliable indicator only when minor degra- 4. Discussion
dation occurs, and would be significantly affected by de-bonding
and spalling off of coarse aggregates and large pieces of mortar In order to identify a test methodology for rapidly assessing the
when severe degradation occurs. performance of cementitious materials subjected to sulphuric acid
Care should be taken in directly comparing the performance of attack, the results of the long term exposure tests presented in Sec-
different mixes when using this measure, for example the mass tion 3 (exposure to 1% sulphuric acid solution for 495 days) are
loss can be related to the aggregate content and therefore mass lost compared to those obtained from identical concrete specimens
in the form of spalling as is shown in Fig. 2. subjected to 3% (w/w %) sulphuric acid exposure for a period of
112 days.
3.7. Length change The tests performed at 1% sulphuric acid concentration (pH 1),
are representative of the lowest pH expected to be encountered
Fig. 11 presents the change in length of the concrete specimens in field conditions where sulphur oxidising bacteria are present
relative to their initial length. As shown in Fig. 9, the sulphuric acid and have developed over time [18]. These tests required a period
only penetrates 5–9 mm from the surface in OPC concrete speci- of at least one year to observe a noticeable reduction in compres-
mens whereas the acid penetrates around 40 mm in AA concrete sive strength. To further accelerate the test methodology, to per-
specimens. Thus, the expansion products in OPCC specimens are form a test within a reasonable amount of time, specimens
only generated near to the surface whereas, a large proportion of prepared at the same time were immersed in 3% (pH 0.52) sul-
the AAC specimens are affected by sulphuric acid. Sata et al. [30] phuric acid concentration. The results of these tests are docu-
observed a larger length change in OPC than that in bottom ash mented in [25] and are here compared with the current results.
based AAC after 12 months immersion when testing the resistance The cement content of OPCC has significant influence on the
of OPC and AAC against 3% sulphuric acid. The contradictory results sulphuric acid resistance of concrete based on the test results pre-
in this study were mainly due to expansive gypsum generated sented in Section 3. It is shown that specimens with higher cement
from the limestone coarse aggregate used in the AAC mix. content (i.e. 380 kg/m3 in OPCC_50) have more severe degradation
in visual appearance, higher acid consumption, larger depth of lost
alkalinity, larger mass loss, larger expansion in length, and larger
3.8. Compressive strength
compressive strength loss than those with lower cement content
(i.e. 280 kg/m3 in OPCC_30). A similar observation has also been
Fig. 12 presents the compressive strength of the concrete spec-
made by Aiken et al. [20]. The morphology and surface element
imens at each time period. It can be noted that at early ages for

Fig. 11. Length change of OPC and AA concrete specimens under 1% sulphuric acid Fig. 12. Compressive strength change of OPC and AA concrete specimens under 1%
attack. sulphuric acid attack.
 L. Gu et al. / Construction and Building Materials 197 (2019) 681–692 689

change of OPCC_30 and OPCC_50 before and after immersion The depth of lost alkalinity shown in Fig. 14(b) indicates very
detected by SEM and EDX was similar for both materials, indicating similar acid penetration depths for OPCC under the differing expo-
no difference in the reaction process for each material. Further, sure regimes. The higher acid consumption in 3% concentration
higher strength concrete usually has lower porosity than that of (shown in Fig. 14(a)), however suggests that the degradation
lower strength concrete [15,21], indicating that the difference in mechanism is altered with pH as the higher ion gradient would
strength reduction can be explained in terms of initial porosity. It be expected to give rise to greater penetration depth. It is possible
is likely that during the sulphuric acid degradation process, the that the different concentration of sulphuric acid affects the devel-
C–H and C-S-H in the OPC paste is dissolved, generating expansive opment of gypsum crystals and the development of a protective
gypsum and amorphous silica as discussed in Section 3. In this pro- barrier. This protective barrier could be either chemical via a local
cess, the higher cement content in concretes with a higher initial reduction of pH [2] or physical in the form of a blocking of the pore
strength, would lead to the production of more expansive gypsum, space.
leading to higher porosity in the degraded zone of specimens. This In contrast, the depth of penetration in AAC continues to
in turn would allow easier penetration of the acid in concretes with increase with time suggesting that the reaction products in AAC
a higher cement content. Further study on the performance (poros- have a much smaller effect in creating an impediment to diffusion.
ity, permeability, diffusivity) in the front of acid attack of concrete This was also observed in numerous studies of acid resistance of
should be conducted to support the hypothesis. AA pastes and mortars [11,20,26], an in depth description of the
The visual appearance of typical specimens under 3% sulphuric differences observed in AA and OPC pastes and mortars, particu-
acid attack at the end of their immersion periods are shown in larly with respect to the role of the Ca/Si ratio of the virgin material
Fig. 13. It can be observed by comparing Fig. 2(a) with Fig. 13(a) in creating expansive reaction products, is elucidated by Aiken
that the OPCC specimens immersed in the 3% concentration dis- et al. [20].
played greater degradation on its surface than the specimen The mass loss of OPCC in 1% concentration decelerates after
exposed to the lower concentration (1%) in Fig. 3(b), albeit for a 142 days exposure, whereas in 3% concentration is appears to
greater period of time. Thus indicating that there is pH dependence accelerate after 42 days, shown in Fig. 14(c). This difference in
on the degradation mechanism. mass loss can be attributed to the different levels of degradation
Fig. 13(b) shows representative samples of AAC under 3% sul- observed on the surface of the 3% exposure specimens as shown
phuric acid attack at the end of its immersion period. Comparing in Fig. 13(a) compared to that seen in the 1% specimens (Fig. 2
Fig. 2(c) with Fig. 13(b) it can be observed that both specimens (a)). The mass loss, in both exposure regimes, for the AAC speci-
showed more significant degradation and had a large amount of mens is accelerating with time, and is reflective of the large
loose material on their surface than the OPC samples. This is likely amount of spalling of materials seen in Figs. 13(b) and Fig. 2(c).
due to the greater depth of acid penetration, and the resulting A noticeable difference in length change is observed between
higher degree of damage to the ITZ observed in the AAC. the two concrete types in Fig. 14(d). This is not unexpected, as
In Fig. 14 the mean indicator measurements and compressive measurements were performed with a DEMEC gauge at the centre
strength development with immersion time are plotted for both of the specimens and therefore this reflects the depth of penetra-
concrete types and immersion regimes. In Fig. 14(a) the marked tion of the acid.
difference seen between acid consumption levels for the different The substantial mass loss recorded in these particular AAC spec-
concretes at lower acid concentration is not repeated when sub- imens and increased length change further demonstrates that
jected to higher concentration. whilst the resistance of a paste or mortar might be indicative of
the potential for a material to be resistant to acid attack, the com-
posite material as a whole must be considered. The limestone used
in this study might have contributed to the formation of a benefi-
cial diffusion barrier (chemical or physical) in OPCC, but deleteri-
ous expansion in AAC.
In Fig. 14(e) the change in compressive strength over time of
both the OPCC and AAC concretes under both exposure regimes
are shown. These results indicate that the total exposure period
required to observe the same reduction in strength, for both mate-
rials, is substantially reduced (from 400 days to 110 days).
The similar reductions in compressive strengths are used to
assess the reliability of the other measurements to be utilised as
indicators. To facilitate this we choose to normalise mass loss
and loss of alkalinity. The normalised depth of alkalinity jdAlk j is
calculated as

dpH
jdAlk j ¼
D

where dpH is the depth of lost alkalinity and D is the original diam-
eter of the specimen. The normalised mass loss jMj is calculated as

Mt
jM j ¼
M0

where M t is the mass at an immersion time t and are shown in

Fig. 15.
Fig. 13. The visual appearance of OPCC and AAC under 3% sulphuric acid attack for From the inspection of the sensitivity of the measurements to
112 days: (a) CC_30 3%; (b) AAC 3% changes in compressive strength, it can be concluded that:
690 L. Gu et al. / Construction and Building Materials 197 (2019) 681–692

Fig. 14. The comparison of properties under 3% sulphuric acid attack and normalised properties under 1% sulphuric acid attack: (a) acid consumption (b) depth of area lost
alkalinity; (c) mass loss; (d) length change; (e) compressive strength change.
 L. Gu et al. / Construction and Building Materials 197 (2019) 681–692 691

(a) (b)

(c) (d)
Fig. 15. The properties of 405-day OPCC and AAC under 1% sulphuric acid attack: (a) Acid consumption per sample; (b) Normalised depth of area lost alkalinity; (c)
Normalised mass loss; (d) Length change.

Acid consumption per specimen, shown in Fig. 15(a) shows a 5. Conclusions

poor correlation with the change in compressive strength and
is inconsistent between material types. An OPC based concrete and Alkali activated concrete were

The retained alkalinity in Fig. 15(b) and mass loss in Fig. 15(c) immersed in 1% sulphuric acid solutions for 495 days. The results
could be employed as indicators for OPC based materials as of microstructural analysis, strength and a range of candidate indi-
the change in strength reduction corresponds to the same cator measurements are presented. This was performed in order to
changes in these measurements, however neither measure identify a simple and rapid test procedure to identify the suscepti-
show consistent variation for the AAC specimens; bility of a variety of cementitious materials.

Although used as a common indicator of durability for other A test methodology has been developed such that tightly con-
chemical exposure regimes, for example sulphate attack (ASTM trolled ‘reservoir’ boundary conditions (i.e relatively constant pH
C1012 [17]), the length change of the specimens in Fig. 15(d), with time) can be achieved through the use of sufficiently large
for all material types, shows no association with the change in immersion tanks, employment of pumps within the reservoir to
compressive strength. reduce the potential for ion gradients and regular observation

None of the measurements appear to give a reliable indicator of and correction of the acid concentration levels.
the loss of compressive strength for AAC and therefore none can The results presented here are compared with tests on identical
be used as means for comparing the performance of different concretes (prepared in the same batches) but subjected to 3% sul-
cementitious materials. phuric acid concentrations. From a combination of the microstruc-
tural analyses, recorded acid consumption per specimen and the
The reduction of compressive strengths shown in Fig. 14(e), depth of acid penetration (as indicated through measuring the
does however suggest that a rapid test of sulphuric acid resistance depth of lost alkalinity) it was found that:
could be performed using compressive strength measurements,
with complimentary testing of other performance measures, such
The mechanism of acid degradation was different between the
as depth of penetration, within a reasonable time frame of circa two material types.
100 days.
692 L. Gu et al. / Construction and Building Materials 197 (2019) 681–692

The response of OPCC specimens was dominated by the devel- [4] C. Grengg, F. Mittermayr, N. Ukrainczyk, G. Koraimann, S. Kienesberger, M.
Dietzel, Advances in concrete materials for sewer systems affected by
opment of a barrier of expansive products which impeded the
microbial induced concrete corrosion: a review, Water Res. 134 (2018) 341–
diffusion process. The observed strength reduction resulted 352.
from changes in the microstructure of the material. [5] J. Monteny, E. Vincke, A. Beeldens, N. De Belie, L. Taerwe, D. Van Gemert, W.

Continued penetration of sulphuric acid was observed in AAC Verstraete, Chemical, microbiological, and in situ test methods for biogenic
sulfuric acid corrosion of concrete, Cem. Concr. Res. 30 (4) (2000) 623–634.
specimens. The meso-structural mechanism of spalling due to [6] T.A. Aiken, W. Sha, J. Kwasny, M.N. Soutsos, Resistance of geopolymer and
expansion of the coarse aggregates dominated the reduction Portland cement based systems to silage effluent attack, Cem. Concr. Res. 92
in strength. (2017) 56–65.
[7] M. Alexander, A. Bertron, N. De Belie, Performance of Cement-Based Materials
in Aggressive Aqueous Environments, Springer, Dordrecht, 2013.
The evaluation of the measurements taken reveal that there [8] C.W. Fourie, Acid Resistance of Sewer Pipe Concrete Doctoral dissertation,
was no single universal indicator of resistance to sulphuric acid University of Cape Town, 2007.
[9] J.P. Frost, D.J. Armstrong, Concrete and silage effluent-a cyclical exposure
that could be applied as a simple test technique for both materials, method for accelerated corrosion testing, Farm Building Progr. 116 (1994) 27–
for the following reasons: 30.
[10] ASTM, C. 267-01: Standard Test Methods for chemical resistance of mortars,
grouts and monolithic surfacing and polymer concretes, ASTM International,

Visual appearance is too subjective and can be highly influenced West Conshohocken, USA, 2012.
by the occurrence of spalling. [11] S. Thokchom, P. Ghosh, S. Ghosh, Resistance of fly ash based geopolymer

The consumption of acid, was recorded as a potential indicator mortars in sulfuric acid, ARPN J. Eng. Appl. Sci. 4 (1) (2009) 65–70.
[12] P. Chindaprasirt, U. Rattanasak, Improvement of durability of cement pipe
of total chemical activity, but found to be inconsistent with
with high calcium fly ash geopolymer covering, Constr. Build. Mater. 112
changes in compressive strength. Thus showing that the mech- (2016) 956–961.
anism of degradation is a complex interplay between ion trans- [13] A.M. Izzat, A.M.M. Al Bakri, H. Kamarudin, L.M. Moga, G.C.M. Ruzaidi, M.T.M.
port and chemical activity. Faheem, A.V. Sandu, Microstructural analysis of geopolymer and ordinary
Portland cement mortar exposed to sulfuric acid, Mater. Plast. 50 (3) (2013)

The depth of lost alkalinity is highly dependent on the initial 171–174.
alkalinity and therefore cannot be used as a universal indicator [14] M. Albitar, M.M. Ali, P. Visintin, M. Drechsler, Effect of granulated lead smelter
across concrete types. slag on strength of fly ash-based geopolymer concrete, Constr. Build. Mater. 83
(2015) 128–135.

Due the occurrence of spalling of large aggregates, mass change [15] A.M. Neville, Properties of Concrete, Longman House, Burnt Mill, Harlow,
can be highly variable, making it a poor indicator when applied England, 1995.
to concretes. [16] ASTM, C. 39: Standard A. Standard Test Methods for Compressive Strength of
Cylindrical Concrete Specimens, ASTM International, West Conshohocken,

The use of length change as an indicator of degradation has USA, 2018.
been utilised successfully for materials where the generation [17] ASTM, C. 1012: Standard Test Method for Length Change of Hydraulic-Cement
of expansive reaction products is consistent. However, the Mortars Exposed to a Sulfate Solution, ASTM International, West
Conshohocken, USA, 2018.
results obtained suggest sensitivity to the pH of the reservoir [18] R.L. Islander, J.S. Devinny, F. Mansfeld, A. Postyn, H. Shih, Microbial ecology of
and material type. crown corrosion in sewers, J. Environ. Eng. 117 (6) (1991) 751–770.
[19] N. De Belie, J. Monteny, L. Taerwe, Apparatus for accelerated degradation
testing of concrete specimens, Mater. Struct. 35 (7) (2002) 427–433.
Equivalent reductions in compressive strength were observed
[20] T.A. Aiken, J. Kwasny, W. Sha, M.N. Soutsos, Effect of slag content and activator
at approximately 100 days in 3% solution to that of 400 days of dosage on the resistance of fly ash geopolymer binders to sulfuric acid attack,
1% solution. Measureable changes in the mass loss, acid consump- Cem. Concr. Res. (2018).
tion, length change and loss of alkalinity at 100 days in 3% solution, [21] H.F.W. Taylor, Cement Chemistry, second ed., Thomas Telford, London, 1997.
[22] C. Shi, P.V. Krivenko, D.M. Roy, Alkali-Activated Cements and Concretes, Taylor
therefore demonstrates that a test procedure to indicate the sus- & Francis, Abingdon, UK, 2006.
ceptibility of concretes can be performed within a reasonable time [23] J.L. Provis, J.S.J. Van Deventer, Geopolymers: Structures, Processing, Properties
frame for standard testing. and Industrial Applications, Woodhead, Cambridge, 2009.
[24] T. Bakharev, Resistance of geopolymer materials to acid attack, Cem. Concr.
Res. 35 (4) (2005) 658–670.
Conflict of interest [25] L. Gu, P. Visintin, T. Bennett, Evaluation of accelerated degradation test
methods for cementitious composites subject to sulfuric acid attack;
application to conventional and alkali-activated concretes, Cem. Concr.
The authors declared that there is no conflict of interest. Compos. 87 (2018) 187–204.
[26] R.R. Lloyd, J.L. Provis, J.S.J. van Deventer, Acid resistance of inorganic polymer
binders. 1. Corrosion rate, Mater. Struct. 45 (2012) 1–14.
References
[27] M. Izquierdo, X. Querol, J. Davidovits, D. Antenucci, H. Nugteren, C.
FernándezPereira, Coal fly ash-slag-based geopolymers: microstructure and
[1] H. Yuan, P. Dangla, P. Chatellier, T. Chaussadent, Degradation modeling metal leaching, J. Hazard. Mater. 166 (1) (2009) 561–566.
ofconcrete submitted to biogenic acid attack, Cem. Concr. Res. 70 (2015) 29– [28] C.L. Page, K.W.J. Treadaway, Aspects of the electrochemistry of steel in
38. concrete, Nature 297 (1982) 109–115.
[2] N. De Belie, J. Monteny, A. Beeldens, E. Vincke, D. van Gemert, W. Verstraete, [29] M.A.M. Ariffin, M.A.R. Bhutta, M.W. Hussin, M.M. Tahir, N. Aziah, Sulfuric acid
Experimental research and prediction of the effect of chemical and biogenic resistance of blended ash geopolymer concrete, Constr. Build. Mater. 43 (2013)
sulfuric acid on different types of commercially produced concrete sewer 80–86.
pipes, Cem. Concr. Res. 34 (12) (2004) 2223–2236. [30] V. Sata, A. Sathonsaowaphak, P. Chindaprasirt, Resistance of lignite bottom ash
[3] G.D. Ma, A. Gutiérrez-Padilla, S. Bielefeldt, M. Ovtchinnikov, J. Hernandez, geopolymer mortar to sulfate and sulfuric acid attack, Cem. Concr. Compos. 34
Silverstein, Biogenic sulfuric acid attack on different types of commercially (5) (2012) 700–708.
produced concrete sewer pipes, Cem. Concr. Res. 40 (2) (2010) 293–301.