Resistance of alkali-activated binders to organic acid attack: Assessment of evaluation criteria and damage mechanisms

Construction and Building Materials 151 (2017) 405–413

Contents lists available at ScienceDirect

Construction and Building Materials journal homepage: www.elsevier.com/locate/conbuildmat

Resistance of alkali-activated binders to organic acid attack: Assessment of evaluation criteria and damage mechanisms Andreas Koenig ⇑, Annemarie Herrmann, Steffen Overmann, Frank Dehn Institute of Mineralogy, Crystallography and Materials Science, Leipzig University, Scharnhorststrasse 20, 04275 Leipzig, Germany

h i g h l i g h t s
Resistance of low-Ca AABs under organic acid attack was proved.
Strength rise during acid attack is linked to new phase formation near the surface.
Organic acid resistance of AABs decreases with increasing CaO content of the binder.
CaO-limit of 10 m.-% seems appropriate to differentiate AABs regarding durability.

a r t i c l e

i n f o

Article history: Received 5 January 2017 Received in revised form 16 June 2017 Accepted 18 June 2017 Available online 26 June 2017 Keywords: Alkali-activated binder Geopolymer binder Organic acid attack Degradation Micro X-ray tomography

a b s t r a c t This paper summarizes the results of investigations into the material resistance of mortars and concretes based on alkali-activated binders (AABs) under organic acid attack (acetic, propionic, lactic acid with pH  3). The development of the residual strength, degradation depth and mass loss over time were recorded after six, twelve and 18 weeks of storage. Additionally, the damage mechanisms were examined by different imaging techniques, mainly micro X-ray computer tomography (mXCT) and microscopy, including spatially resolved porosity as well as phase analyses. The degree of degradation, especially in the pre-damaged zone, decreases with a decreasing Ca-content of the binder. No damage zones, an increased residual strength and a densification (detected as a reduction in porosity) at the acidexposed fringe zone was observed for so-called Geopolymer binders (low-Ca AABs) the longer the exposure to acid was. A lower degree of degradation was detected for high-Ca AABs compared to cementitious reference mortars/concretes. Ó 2017 Elsevier Ltd. All rights reserved.

1. Introduction The acid resistance of cementitious materials is influenced by the impermeability of the concrete matrix and the resistance of the strength forming phases. Apart from technological concrete optimizations, such as a decreasing w/c-ratio [1] or an optimized grading curve [2], the best potential for improvement with respect to acid resistance is to generate more stable phases. The use of Supplementary materials such as fly ash or particularly silica fume leads to a reduction of the acid-soluble Ca(OH)2 in Portland cement-based binders, and at the same time to the formation of C-S-H-phases with lower C/S ratios [3–6]. In contrast to fly ash and silica fume, the addition of metakaolin results in the formation of C-A-S-H-phases with an enhanced acid resistance compared to ordinary C-S-H-phases [3]. In order to reduce the amount of

⇑ Corresponding author. E-mail address: [email protected] (A. Koenig). http://dx.doi.org/10.1016/j.conbuildmat.2017.06.117 0950-0618/Ó 2017 Elsevier Ltd. All rights reserved.

acid-soluble Ca(OH)2 and thereby to generally improve the acid resistance, a reduction of the clinker content in cements is decisive. Accordingly, the aforementioned pozzolans are often combined to form cementitious composites [7,8]. Nevertheless, the advances in cementitious materials regarding the durability are limited due to the thermodynamic instability of hydrate phases for pH-values below nine [9,10]. The resistance of cementitious materials is reduced in many applications with challenging exposure conditions, e.g. biogenic sulfuric or organic acid attack [6,11]. Initially, alternative cements such as high alumina cements [12,13] or sulfate-activated granulated blast furnace slag [13,14] were used to improve the acid resistance of concretes. More recently, alkali-activated binders became a promising alternative for highly durable concretes due to their high resistance within aggressive environments, which was already demonstrated for sulphate [15], sulfuric acid [16–18], hydrochloric acid [19], nitric acid [18] or acetic acid [17].

406

A. Koenig et al. / Construction and Building Materials 151 (2017) 405–413

Next to mineral acids, organic acids can be of particular importance for sewage or agricultural concrete applications. Even if the acidity of organic acids like acetic, propionic, lactic or butyric acids is lower than mineral acids, concrete attack isn’t less aggressive (i.e. due to buffer effects and the solubility of the organic salts). A more detailed description of the damage mechanism of weak acids is published in [20]. This paper presents a detailed experimental program to characterize the damage mechanisms of mortars and concretes based on four alkali-activated binders exposed to organic acids. An organic acid mix of acetic, propionic and lactic acid with a pH of 3 was chosen to represent typical exposure conditions for highly-stressed agricultural constructions, i.e. silage clamps [11]. The experimental results obtained on alkali-activated mortars with different CaO contents are compared with reference mortars based on ordinary Portland cement (OPC) as well as on OPC partially enhanced by fly ash (OPC + FA). Finally, recommendations based on the findings that were obtained are formulated for the future evaluation of acid resistance. 2. Significance of work The broad spectrum of investigated binder compositions and assessment criteria herein, will thereby enable a sufficient and comparative assessment of different AAB systems as well as the used evaluation criteria and is at the same time highlighting the novelty of the paper. 3. Materials and methods 3.1. Cementitious materials and testing conditions 3.1.1. Experimental program This paper assesses the acid resistance of four different types of alkali-activated binder, an alkali-activated granulated blast furnace slag (AAS), an alkali-activated low-calcium fly ash (AAFA), a high-Ca multi-component AAB consisting of fly ash and slag (80-20, 50-50) as well as two reference mixes (OPC, OPC + FA) in terms of the development of mass loss, residual compressive strength and degradation depth. Moreover, the damage mechanisms of the binder systems which are quite different from one another (OPC, AAS, AAFA) were comparatively evaluated by Xray diffraction (XRD), micro X-ray computer tomography (mXCT) and mercury intrusion porosimetry (MIP).

3.1.2. Compositions of raw materials and mix designs A low-calcium fly ash (FA) according to EN 450-1:2012-10, granulated blast furnace slag (S) and two ordinary Portland cements (CEM I 32.5 R, CEM I 52.5 R) according to EN 197-1:2011-11 were used. The chemical composition, amorphous percentage and loss of ignition of the raw materials used were determined by Xray fluorescence (XRF), X-ray diffraction analyses (XRD) and gravimetric analysis. The results are shown in Table 1. Alkali-activated and multi-component mortars as well as concretes were activated by the addition of liquid sodium hydroxide (19 M) and sodium silicate with a SiO2 /Na2O molar ratio of 3.4 and solid content of 34.5 m.-%. The specific dosage was adjusted to the chemical composition of the raw materials (Table 2). Quarzitic gravel (q = 2.63 kg/dm3) with a maximum size of 16 mm and locally available river sand were used as aggregates for the concrete mix. The mortars contained CEN sand according to EN 196-1:2005-05. The workability was adjusted in both cases by the addition of a polycarboxylate ether superplasticizer (PCE-SP) for OPC and by the addition of Lignosulphonate-superplasticizer (LSF-SP) for AAS, 80-20 and 50-50, to ensure good compatability and prevent segregation.

3.1.3. Casting Mortars were produced in accordance to EN 196-1:2005-05 with a binder to sand ratio of 1:3, mixed in a TESTING Blum & Feuerherdt mortar mixer of the type 1.0205, consolidated via shock table and cast in molds of the size 40  40  160 mm3. In addition, concrete prisms (100  100  400 mm3) and concrete cubes with an edge length of 150 mm were cast. Dry and liquid constituents were mixed in a compulsory mixer for 3 min. Chemical additions were then added and mixing continued for a further 60 s. before the concrete mix was placed in steel molds and consolidated by vibration. Alkali-activated fly ash concretes were cast in plastic molds to avoid any loss of surface quality caused by a high bonding between the concretes and the steel molds.

3.1.4. Curing The curing conditions depend on the type of binder, in particular their Cacontent and reaction mechanisms. Therefore, only alkali-activated fly ash (AAFA) mortars and concretes were cured at 75 °C for the first 48 h. The heat-cured specimens as well as all the other specimens stored at standard climate conditions (20 °C, 65% RH) were removed from the mold after 2 days. The OPC-based specimens were then stored under water until testing, whereas the AABs were wrapped in foil and stored in standard climate conditions.

3.1.5. Experimental set up All samples were stored in water until saturation (at least three days) before the acid tests. This procedure ensures a diffusion-controlled process during the acid attack. A total of 54 mortar specimens, six concrete prisms and six concrete plates (100  100  40 mm3, sawn from concrete cubes) were immersed in the organic acid for at least 28 days in accordance with an in-house testing procedure, presented in [20]. The test equipment and the testing procedure developed here enable a constant stress level and reproducible results. The accelerated test set up is primarily designed to compare the performance of different kinds of mortar and/or concrete. A reliable, probabilistic model for service life predictions is currently not available since the complexity of field conditions is not predictable [21]. The liquid test medium was a homogenous mix of 1.5% acetic acid, 0.5% propionic acid and 3.0% lactic acid with a pH-value of 3. The acid mix is based on own investigations of agricultural constructions such as silage clamps [11]. The acid test solution had a volume of 320 dm3 and a temperature of 25–28 °C. The samples had a total surface area of 246 dm2. The pH value was regulated by an automatic titration unit filled with organic acid mix while the acid solution was homogenized by a mixer with a pump system. A flow rate of between 0.05 and 0.30 m/s prevailed in the acid container. The acid test solution was replaced every four weeks. The CaOcontent of the continuously stirred test medium was determined by the complete replacement of the test solution at different times (6 wk.: 1345.0 mg/L, 12 wk.: 626.7 mg/L, 18 wk.:1237.0 mg/L). Furthermore, additional 48 mortar specimens and six concrete prisms were stored in a water bath as a reference for the corresponding removal date. The damage potential was assessed after six, twelve and 18 weeks of immersion. Three mortar prisms were removed from the acid tank and two prisms were removed from water bath each time. The residual compressive strength was determined on four mortar cubes (two exposed and two unexposed references), which were sawn from the prisms. The third exposed specimen contained carbon rods (Ø 3 mm) and was used for image-generating analysis of the degradation depth. In contrast, all the concrete specimens were removed at the same time. After 18 weeks of immersion, one concrete prism was removed from the acid tank and one from the water tank to determine the residual compressive strength. The concrete plates which were also removed after 18 weeks of immersion were used to examine the depth of concrete degradation.

3.2. Mass loss To ensure a suitable statistical basis, all mortar and concrete prisms that were exposed to acid and water were weighed using a digital scale with an accuracy of ±0.1 g; the results were recorded before exposure as well as at the specific removal date. For this purpose, the mortar and concrete prisms were roughly dried on a saturated surface. With a progressive exposure time, the number of samples gradually

Table 1 Chemical composition by XRF, X-ray amorphous percentage by XRD and Loss On Ignition (LOI) of the raw materials (%). Raw material

Portland cement (CEM I 32.5 R) Portland cement (CEM I 52.5 R) Fly Ash (FA) Blast furnace slag (S)

Chemical composition SiO2

Al2O3

Fe2O3

CaO

MgO

SO3

K2O

Na2O

20.10 19.97 52.85 35.35

4.82 5.56 25.57 11.18

3.20 3.20 9.41 0.57

63.50 63.59 3.26 40.33

3.46 2.08 1.74 6.83

1.27 3.74 0.48 2.69

0.99 1.13 2.42 0.68

0.21 0.27 1.44 0.18

LOI

X-ray amorphous quantity

2.22 1.55 1.55 –

4.0 2.1 63.2 98.8

407

A. Koenig et al. / Construction and Building Materials 151 (2017) 405–413 Table 2 concrete mix designs (kg/m3). Mixture

OPC OPC-FA AAS 50–50 80–20 AAFA *

Powder

Activator solution

CEM I 32.5 R

CEM I 52.5 R

FA

450 – – – – –

– 347 – – – –

– 103 – 225 360 450

S

– – 450 225 90 –

w/b ratio*

Sodium hydroxide (19 M)

Sodium silicate

Water

– – 33 45 45 90

– – 67 112.5 112.5 45

180 175 129 84 84 55

0.4 0.39 0.45 0.42 0.42 0.3

Superplasticizer

Aggregate

PCE

Sand 0– 2 mm

Quartzitic gravel 2– 8 mm

8– 16 mm

678 675 644 633 624 675

507 504 481 473 466 504

509 506 483 475 468 506

1.25 2.25 – – – –

LSF

– – 13.5 13.5 13.5 –

w/b = water/binder ratio of the sum total of water and the binder (sum total of powder).

decreased from nine to three mortar prisms per mix design. The recorded data for the concrete specimens cannot be considered as statistically representative since only one specimen was tested. 3.3. Residual compressive strength The compressive strength was determined as an average value of four cubes with a size of 40  40  40 mm3 for mortars and three cubes with a size of 100  100  100 mm3 for concretes. Mortar prisms or concrete prisms were cut into cubes not later than three days after removal using a wet saw. Afterwards, the cut surfaces were polished and used as a testing surface to determine the residual strength of the undamaged core rather than to characterize the residual strength of the damaged edge zone. Residual compressive strength tests were performed seven days after the specific removal time as well as after complete drying of the specimens. The failure load of mortar specimens was determined based on EN 196-1:2005-05 with a servo hydraulic compression and bending test device of the type RT 200/10-1s from Testing Blum & Feuerherdt GmbH. The concretes were tested based on EN 123901:2012-12 with a servo hydraulic compression testing device of the type 102/3000, walter & bai ag. It is assumed that an undamaged specimen core with unaffected strength will fully bear the applied compressive load while a damaged zone cannot bear any load at all. An equivalent degradation depth (Eq. (1)) based on the resulting residual strength and considering geometrical changes was calculated according to [20]:

DDcal ¼

sffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi# " 1 F a  1000  a 2 fw

exposed surfaces on opposite sides (Fig. 1c). Directly after polishing (4 weeks after removal), microscopy of the samples was examined. Therefore, the endings of the carbon rods mark the original level of the undamaged surface, which is important for the determination of the depth of erosion. A minimum of eight high-resolution images with increased color intensity were recorded for each polished specimen surface with a digital reflected-light microscope (Olympus SZX 10). The high number of images is related to the maximum aggregates size. According to Koenig and Dehn [20] increasing aggregate sizes will increase degradation depth and thus necessitate increasing measurement lengths. Subsequently, the single pictures were assembled and the degradation depth was determined at a measuring distance with a minimum length of 70 mm using the image edition software analysis docu (Olympus Soft Imaging Solutions). The average and the standard deviation were calculated based on 100 measuring points. A more detailed description for RLM analysis on degraded specimens is presented in [20].

3.5. Methods for microstructural investigations on OPC, AAS and AAFA 3.5.1. Introduction Due to the strong similarities between OPC and OPC + FA, AAS, 50-50 and 80-20 as well as AAFA, the microstructural investigations are partially limited to representatives of each group, namely OPC (main phases: C-S-H + CH), Ca-rich AAB (main phases: C-S-H + C-A-S-H) and AAB with a low Ca-content (main phases: N-A-S-H), so-called Geopolymer binders [22].

ð1Þ

where: DDcal = Calculated depth of degradation in mm. a = Edge length of the reference cube in mm. Fa = Failure load of the acid-exposed specimen in N. fw = Compressive strength of the samples exposed to water in N/mm2. On the basis of the minimum and maximum compressive strength the deviations were calculated for each single calculated depth of degradation. 3.4. Reflected-light microscopy (RLM) Every sample contains two parallel carbon rods, which were placed in the middle of the sample (Fig. 1a) and leveled to the sample surface by polishing. At the specific removal time, the microscopy samples were embedded in resin within 14 days after removal. Subsequently the samples were cut along the midline marked by two carbon rods to provide a cross section (Fig. 1b) with two acid-

3.5.2. Micro X-ray computer tomography (mXCT) Micro X-ray computer tomography (mXCT) was used to characterize the spatially-resolved porosity and for the comparative determination of the degradation depth on exposed mortar samples. The grey-value-specific zoning of the matrix results from an alteration of the material-specific absorption coefficient, this being affected by density and chemical composition. Thus, bright areas indicate dense and dark areas less dense local materials. The mXCT analysis was performed with an Xray power of 26.3 Watts (beam energy 180 kV and flux 180 mA), a copper filter at the focal spot and a step size of 0.2/360° (1.800 positions). The voxel edge length was maintained at 25.6 mm (V = 16.777 mm3) as an indicator for maximum resolution. After six weeks of immersion, subsamples with acid-exposed surfaces were cut from the center of the mortar prisms (1.5  2  4 mm3). The spatially total porosity was analyzed perpendicular to the sample surface by ‘‘ImageJ 1.47v” (National Institutes of Health) and based on image stacks (2D). Contrary, the pore distribution was determined on the rendered dataset (3D) of the same region of interest (ROI) by ‘‘VGStudio Max 2.0” (Volume Graphics GmbH). Because of the limited resolution of 16.777 mm3 the calculated values from rendered dataset can only represent pores with a volume of >135 mm3 (16.777 mm3, 8 voxels) or rather an equivalent pore

Fig. 1. Preparation of microscopy samples: (a) reduction of specimen length, (b) cutting the resin embedded sample along the midline marked by two carbon rods (dotted line), (c) microscopy analyses of polished samples.

408

A. Koenig et al. / Construction and Building Materials 151 (2017) 405–413

radius of 25.6 mm. The small difference between results obtained by ImageJ (spatially total porosity) and VGStudio (sum of pore-volume) can be attributed to the minimum voxel number.

3.5.3. X-Ray diffraction (XRD) The quality of the mineral phases was determined using a Seifert-FPM XRD 7 Xray diffractometer (XRD) operating in Bragg-Brentano geometry with a Cu-anode and a graphite monochromator (40 kV/40 mA (kKa¯ = 1.5418 Å). Data was collected over a range of 5–70° with a step size of 0.03° and a counting period of 6 s. The powdered and dried material samples with a particle size <0.125 mm (acid-exposed mortars extracted from the undamaged core (C) as well as from the acid-exposed (E) fringe, see Fig. 4) were analyzed with the addition of an internal standard (inert Fluorite, F). The qualitative analysis was carried out using the ‘‘DIFFRAC.EVA” software (Bruker Corporation, USA). The comparison of diffractograms of the undamaged core and the acid-exposed samples allows conclusions about the stability of strength giving reaction products. An quantification was not performed because of the high percentage of amorphous phases before (C-S-H-, C-A-S-H-, N-A-S-Hphases) and after the acid attack.

3.5.4. Mercury intrusion porosimetry (MIP) The mercury intrusion porosimetry (MIP) was used to determine the pore distribution in undamaged mortars in addition to mXCT-analyses. MIP was obtained for OPC, AAS and AAFA mortars of a comparable age to the specimens analysed by mXCT. Mortar specimens were dried at 105 °C until a constant mass and subsequently analysed with ‘‘Autopore 9229” from ‘‘Micromeritics”. Based on the pressure range of the equipment, pore radii between 3.5 nm and 23 lm could be determined.

4. Results 4.1. Mass loss The mortar and concrete specimens lost mass with an increasing exposure time (Fig. 2). The rate of mass loss is significantly higher within the first six weeks than between the first (week six) and the last removal date (week 18). A correlation between mass loss and CaO-content of the binder could be observed. The low-Ca AAB mortar/concrete (AAFA) had the lowest mass loss, whereas the highest mass loss was recorded for OPC mortars/concretes. In principal, the mortar samples showed a higher mass loss than concrete samples. Specimens stored in water showed a slight mass gain over time (<0.8 M.-% after 18 weeks). 4.2. Residual compressive strength At the beginning of storage, the OPC-based and high-Ca AABs could be classified in a comparable strength class of C35/45 or C40/50 and therefore met the requirements of exposure class XA3 as defined by EN 206. The AAFA mortar and concretes achieved a lower strength and thus do not fulfill the normative requirements of this exposure class.

Fig. 2. Time-dependent mass change for acid-exposed and water-cured mortar and concrete specimens as a function of the CaO-content of the binder.

An increase in strength up until twelve weeks for the reference specimens, followed by a slight decrease in strength up to 18 weeks for water storage, were obtained for all types of binder (Table 3). A continuous decrease in strength over time was detected for specimens immersed in the organic acid solution, except for AAFA. Fig. 3 illustrates the residual compressive strength calculated as the quotient of the strength obtained with acidexposed and water cured specimens at the specific exposure time. The evaluation based on relative strength was chosen to exclude the effect of a strength class. Similar to the mass loss, the rate of strength is more distinct in the first six weeks of immersion. Unlike, the strength development beyond six weeks of exposure follows a linear behavior. In contrast to all the other binders, acid-exposed AAFA specimens display a steadily increase in strength over time, which can be observed in acid-exposed and water-cured mortars as well as in concretes. The results of this study prove that the residual strength decreases with an increasing Ca-content of the binder. The residual concrete compressive strength is generally higher than the corresponding mortar strength. 4.3. Reflected light microscopy (RLM): depth of degradation Different degradation zones can be identified on polished, acidexposed samples based on color differences (Fig. 4b). Within the cross sections of OPC, OPC + FA, AAS and 50-50, the depth of erosion (DE) was followed by a depth of reaction (DR) including a dark and a red-yellow zone. The dark color in the near-surface area is related to the specimen preparation since an increase in porosity caused by the acid attack facilitates the penetration of colored resin during embedding. In multi-component AABs, an additional brighter zone below the red-yellow zoning was observed, whose intensity increased with the amount of slag. A small bright zone was detected in the AAFA samples. The depth of degradation (DD) reflects the sum of the depth of erosion (DE) and depth of reaction (DR) and arises with increasing exposure time (Fig. 4a). 4.4. Microstructural investigations 4.4.1. Porosity measurements by mXCT and MIP The fundamental differences between the tested binders also cause substantial differences in microstructural-related properties. Mercury intrusion porosimetry (MIP) analyses (Table 4) show a uniform pore size distribution in OPC with a total porosity (TP) of 5.1%. Additionally, an uneven pore size distribution for AABs occurred in a pore size range between 1 and 5.000 mm for AAS (TP 4.7%) and was concentrated at pore sizes over 2.000 mm or below 0.1 mm for AAFA (TP 7.9%). The mXCT-analysis showed a change in density and porosity inside the acid-exposed areas of OPC, AAS or AAFA. The total porosity of the undamaged core is in conformity with MIP (Table 4) results and shows comparable results. A change in density of the quartzitic aggregates was not observed. Unlike AAFA mortars, the density of OPC and AAS mortars decreased with a simultaneous increase in porosity (total porosity, pore size and pore connectivity). Only AAFA mortars displayed a reduction of the total porosity and pore sizes, as concluded in Table 4. The depth of degradation recorded using mXCT correlates well with the degradation depth determined by RLM (Fig. 4). 4.4.2. Phase analyses by XRD In OPC specimens (Fig. 5 left), ettringite [E, Ca6Al2[(OH)12| (SO4)3]26 H2O)] and portlandite [P, (Ca(OH)2)] were dissolved completely by the acid attack. At the same time, the X-ray amorphous percentage as well as the quartz content [Q, (SiO2)] increased and rock-forming minerals such as muscovite [MS,

409

A. Koenig et al. / Construction and Building Materials 151 (2017) 405–413 Table 3 Compressive strength (MPa) of acid-exposed and water-cured mortar prisms after zero, six, twelve and 18 weeks of immersion, based on different binder systems. Exposure Time

0 wk.

6 wk.

12 wk.

18 wk.

OPC

Water Organic acid mix

51.9 –

51.1 31.4

61.3 26.9

61.8 24.0

OPC + FA

Water Organic acid mix

73.6 –

73.7 57.1

82 53.1

78.4 33.9

AAS

Water Organic acid mix

58.7 –

59.8 50.6

64.9 45.8

59.6 34.8

50-50

Water Organic acid mix

79.6 –

86.5 76.33

93.2 65.1

87.7 57.1

80-20

Water Organic acid mix

51.2 –

59.9 51.1

63.8 42.7

60 37.6

AAFA

Water Organic acid mix

21.5 –

29.4 25.2

26.8 33.6

26 31.5

The X-ray pattern of AAFA mortars, shown in Fig. 5 (right), mostly consists of rock-forming minerals and fly ash minerals like mullite (ML). An increase in intensity can be obtained for the acidexposed zones when compared to the undamaged core of AAFAs, similar to AAS mortars. A clear signal was also observed in the range of 27.3–27.9 2H.

5. Discussion 5.1. High-Ca binders: OPC, OPC+FA

Fig. 3. Compressive strength of acid-exposed mortars and concretes with a varying CaO-content relative to the compressive strength obtained from reference specimens stored under water at the corresponding age.

Fig. 4. Depth of degradation as a function of CaO-content and time (left); detected zones exemplarily shown on an acid-exposed AAS mortar (right).

(KAl2[(OH,F)2|AlSi3O10)], orthoclase [O, (K[AlSi3O8)] as well as chamosite [C, (Fe2+.Mg,Fe3+)5Al[(OH,O)8|AlSi3O10)] decreased in the acid-exposed areas. In AAS mortars, only mineral phases of the aggregates used such as quartz (Q), orthoclase (O), muscovite (MS) as well as chamosite (C), and an X-ray amorphous percentage were identified. All of the mentioned mineral phases are acid-insoluble and were also detected in the acid-exposed surface, even if with an increased intensity (Fig. 5, center). A signal which was obtained between 27.3 and 27.9 2H can probably be assigned to anorthite (CaAl2Si2O8, plagioclase feldspar), which has similar reflexes and atoms like as free ions in the test solution.

5.1.1. OPC The XRD-analyses confirmed that the hardened cement paste (i.e. C-S-H-phases, ettringite, portlandite) of Ca-rich binders based on common cements hydrolyses quickly under acid attack. Portlandite, the mineral phase of hydrated cement with the highest solubility, is initially dissolved with a decreasing pH-value (12.2 after [23] – 12.5 after [9]). According to technical literature, the stability of ettringite, the second mineral phase obtained in Ca-rich binders, is restricted by a lower pH-limit of 10.5 [24–26]. A similar limiting value was defined for C-S-H phases [27], even though this value must be lower with decreasing C/S ratios for thermodynamic reasons [10,21,23,28]. The lack of ettringite (XRD) and the significant reduction of the Ca-concentration [4,29] in the acid-exposed surface indicate that the C-S-H phases of the hydrated cement were degraded completely, causing a high mass loss and high degradation depths (Fig. 2 and Fig. 4). An increase in porosity could be proven by mXCT-analysis for the remaining X-ray amorphous (Fig. 5 left) and Si-rich matrix [20] of the exposed surface. This leads to the conclusion that the acid-degraded cross-section is unable to assume any load-bearing function (Fig. 6). This assumption is confirmed by the significant loss of strength over time (Fig. 3). The high porosity is caused by the leaching of the hydrated Portland cement phases but also by the erosion of small aggregates. That is the reason why the amount of rock-forming minerals like muscovite and orthoclase as well as the microstructure-characterizing mineral chamosite decreased. Another reason could be a mobilization of the alkalis in feldspars under acidic conditions like it is well known under alkaline conditions [30]. Unlikely to [2,3,31,32], which could show that the degradation depth (DD) of normal cement is reduced significantly through the substitution of Portland cement by fly ash. Due to the reduction of the Ca(OH)2 content from around 21 to 19 wt-% (according to [33,34] obtained for comparable hardened cement specimens with an age of 56 days), the DDs of OPC and OPC + FA are similar in this study. This can be attributed to the low reactivity and low contents of the fly ash.

410

A. Koenig et al. / Construction and Building Materials 151 (2017) 405–413

Table 4 Conclusion of image-generating methods performed on OPC, AAS and AAFA mortar. OPC 6 wk.

AAS

AAFA

18 wk.

6 wk.

18 wk.

6 wk.

18 wk.

6616 mm

3419 mm

5097 mm

–

–

RLM-imagesa

Depth of degradation 3679 mm mXCT- Results (after 6 weeks of immersion) mXCT cross-sections

Depth of degradation

a

106 mm

–

Un-damaged core

Acid-exposed

Un-damaged core

Acidexposed

Un-damaged core

51.3 ± 2.6%

6.2 ± 0.8%

30.4 ± 3.8%

5.5 ± 1.1%

4.4 ± 1.7%

7.3 ± 1.4%

54.5 0.0 0.0 0.0 0.1

0.0 0.9 1.3 0.4 0.1

31.2 0.0 0.0 0.0 0.1

0.0 0.2 0.9 0.5 0.2

0.0 0.0 1.3 1.3 0.9

0.0 0.5 1.9 1.2 0.7

MIP- Results on undamaged specimens with comparable age Pore distribution in mL/g 10–23 mm 0.01 1–10 mm 0.15 0.1–1 mm 0.28 10–100 nm 0.96 3–10 nm 0.52

0.00 0.13 0.22 0.61 0.41

Total porosityb Pore distribution in % 100–10 mm3 10–1 mm3 1–0.1 mm3 0.1–0.01 mm3 <0.01 mm3

b

129 mm Acid-exposed

0.00 0.20 0.58 2.09 0.85

The red lines mark the DR, blue line symbolizes the midline of the DR. The DE could not be visualized due the low thickness. Based on a voxel size of 25.6 mm (resolution) and a constant volume of degraded (28.2  15.4  10.3 mm3) and undamaged (4.1  15.1  18.0 mm3) area.

Fig. 5. X-ray patterns of 18-week, acid-exposed mortar samples based on different binders for comparison of the mineralogical composition of the undegraded core (C) and the acid-exposed surface (E), Keylist of minerals: Anorthite (A), chamosite (C), ettringite (E), fluorite (F), muscovite (MS) mullite (ML), orthoklas (O), quartz (Q).

In addition to the pozzolanic activity the constant w/b-ratio has the consequence that the water to cement ratio increases from 0.40 (OPC) to 0.50 (OPC + FA) and therefore the capillary porosity was increased (see Table 2). This might neutralise the effect of the reduced Ca(OH)2 content [2] and the reduced CaO/SiO2 of the new C-S-H phases [35–37]. The higher strengths before the started acid tests could result from the higher strength class of the OPC (32.5 compare to 52.5) and not from the pozzolanic activity of the fly ash [33,34]. The lower mass loss of OPC + FA is assigned to the inert fly ash components as well as the pozzolanic activity of the glassy fly ash components. According to the authors investigations, this fly ash

[38] contains 11.6 Vol.-% (34.9 w.-%) of inert components, namely mullite, quartz and maghemite with a density of 3.59 g/cm3 (webmineral.com). The differences in portlandite contents between OPC (22% by cement mass) and OPC + FA (17% by cement mass) amount to max. 5 w.-% according [33,34] after 18 weeks. As a result of the consumed Ca(OH)2, only a small part of the SiO2 in fly ash glass had reacted. For example: After 18 weeks the volume in the depth of reaction (DR) of one mortal prism (40  40  160 mm3) can be calculated based on a DR of approx. 5.0 mm (see Fig. 4) with approx. 0.113 dm3 (40  HYPERLINK "SPS:refid::"[160  10]  2 + HYPERLINK "SPS:refid::"[40  10]  HYPERLINK "SPS:refid::"[160  10]

A. Koenig et al. / Construction and Building Materials 151 (2017) 405–413

411

of erosion (Fig. 4). Furthermore, microscope images showed that the degree of degradation is lower in the damaged area than in pure alkali-activated slag binders. The acid resistance increased with an increasing fly ash content and decreasing slag content. A clear delimitation between degraded and undamaged areas becomes more difficult using image generating analysis for binders with increasing fly ash contents. The residual strength has to be determined alongside the mass loss and degradation depth for a quantification of the degree of degradation.

5.3. Low-Ca AAB – AAFA

Fig. 6. Measured average degradation depth vs. degradation depth calculated from the average residual strength after six, twelve and 18 weeks (three points).

+ 40  40  5 mm3) in comparison to the prism volume of 0.256 dm3 (40  40  160 mm3). The mass of the inert parts in the fly ash in the DR volume amounted approx. 2.0 g (=0.35 w.-%) in one prism. Concurrently, the reduced portlandite content (5% by cement mass) in this DR volume resulted approx. 2.3 g (0.40 w.-%). The comparison between the determined difference in the mass loss of approx. 2.5 w.-% (see Fig. 2) and the calculated summed value of 0.75 w.-% indicate, that maybe additional mode of actions could be involved like for example the glass parts in FA, which could not react because of the decrease pH-value in the DR and the low pozzolanic activity. The enhanced residual compressive strength despite the increase in the depth of degradation implies that the compressive strength in the degraded area is considerably higher than for pure OPC. The discussed points like the formation of more stable C-S-Hphases instead of portlandite as well as the non-reacted parts in the FA, but also the higher-strength cement grades could be the reasons for the higher compressive strength in DR.

5.2. Ca-rich AAB: AAS and AAS + FA (50-50 & 80-20) In contrast to ordinary Portland cements, the strength-forming phases of alkali-activated binders based on granulated blast furnace slag with comparatively high Al-contents consist of calcium alumina silicate hydrates (C-A-S-H), such as gehlenite hydrates (C2ASH8) with a C/S ratio 1.5 [36,39,40] or calcium aluminate hydrates e.g. hydrogarnet (C3AH6) [39] along with C-S-H-phases with a low C/S ratio [41,42]. Thus, the characteristically low Cacontents lead to the improvement of the chemical resistance to acid attack from a thermodynamic point of view [7,13,43]. At the same time, the Si- and Al-rich, strength-forming phases result in an enhanced residual strength (Fig. 3) and a reduced mass loss (Fig. 2) compared to hydrated Portland cement phases. Due to a low mass loss and strength decrease (Fig. 2 and Fig. 3) over the exposure time, it is assumed that this zone is less or not damaged. In fact, it is suggested that the granulated blast furnace slag is colored by metal sulfides that are oxidized next to colorless sulfate compounds by the admission of air [39] or low alkalinity [44]. This color change is considerably higher than the pH limit of 9.5 for passivation and is therefore not a real damage indicator. Compared to Ca-rich AAS, a slower mass loss (Fig. 2), higher residual strength (Fig. 3), and lower depth of degradation (Fig. 4) were observed for high-Ca AABs (50-50 & 80-20) consisting of fly ash and granulated blast furnace slag of different proportions. The sample geometry is constant due to a non-observable depth

The heat-cured AAFA binder does not undergo erosion but shows a brighter area near the exposed surface (<500 mm/18 wk.). The residual strength of mortars and concretes increases by around 20% up to twelve weeks of immersion and stagnates in week 18. The mXCT analyses on exposed surfaces of opposite sides showed an increase in the total porosity of the fringe compared to the undamaged core with a low number of pores and sizes ranging between 0.01 and 100 mm3. A clear and new signal was obtained at 27.7 2H for the surface-near area in an XRD analysis. The XRD signal could be assigned to anorthite (CaAl2Si2O8, plagioclase feldspar). Anorthite is a well-known mineral, which is typically formed under high temperatures between 627 and 727 °C [45] but can also be formed under lower temperature (200 °C) and an applied pressure between 6 and 68 MPa. The formation of anorthite necessitates CaO, reactive Al2O3 and SiO2 which can be supplied by i.e. slag [46], clay-limestone mixes [45,47], brown coal fly ash [48] or high alumina cements (HAC) [49]. It is assumed that alkalis, characterized by high mobility, are leached out from the pore solution as well as from the crystal structure of the N-A-S-H phases during acid exposure in the near-surface area of AAFA [50]. Simultaneously, Ca-ions are present in the acid solution due to the leaching process in the Ca-rich samples. Similar to phosphoric acidbased Geopolymers [51–53], the fly ash particles near the exposed surface can be activated by the acidic solution containing high amounts of dissociated Ca2+ and Al3+. The low peak width at half-height of the anorthite reflex and the filled pores (mXCT) indicate that the new crystals were precipitate out of the pore solution. Experimental studies about the dissolution [54] and forming process [55] of anorthite by acids under normal conditions could serve as proof of this theory. A phase transformation from an amorphous to a crystal structure, based on the substitution between Si and Al [56] as well as between alkalis and alkalis earth for the chargebalance [57] is highly improbable. Up to now, activation through alternative acids such as organic acids is unknown in both the construction and the ceramics industry. Only Bakharev [17] detected bright reaction products of near-surface sodium hydroxide- and sodium silicate-activated, low-Ca fly ashes exposed to acetic acid. A radiographic allocation linked to an increase in strength was not determined by [17]. The residual strength characteristics are decisive for a future quantification of the degradation of low-Ca binders. A characterization of degradation depth based on optical analyses alone is not sufficient as single method. The low compressive strength of the AAFA concrete before the acid attack (mortar strength of 21.5 N/mm2) complicated their use as a structure material for highly-stressed agricultural constructions. However, the very high compressive strength, which could be demonstrated herein for more component alkaliactivated binders based on fly ash and slag (80-20) show, that the addition of low amounts of slag can improve the strength significantly. Another possibility for structural use of alkali-activated binders could be the combination of two different concretes cast fresh in fresh to combine advantageous properties (mechanical and acid resistance).

412

A. Koenig et al. / Construction and Building Materials 151 (2017) 405–413

5.4. Evaluation of the load-bearing capacity of the damaged cross section Fig. 6 shows the microscopically-obtained degradation depth (DD) as a function of the calculated degradation depth (DDcal) in accordance to [20] and Eq. (1) and allows drawing conclusions about the load-bearing capacity (LBC) of the degraded area. Assuming that a degraded area is damaged (in terms of the dissolution of Ca-based, strength-forming phases and an increase in porosity) the damaged part of the cross-section is unable to bear any load after acid exposure. This theoretical assumption is shown in Fig. 6 as a dashed line. The distances between the linear functions and the single measuring points are the result of the variation in the accelerated stress by the organic acids, the inhomogeneity of the samples and the measurement uncertainty (compressive strength, light microscopy). The sign of the slope is an indicator for an increasing () or decreasing (+) residual strength. Test data that lies below the dashed line, as in the case of OPC, means that the LBC of the damaged area is zero and the optically measured DD may underestimate the real depth of performance degradation. Test data above of the dashed line (e.g. OPC + FA, AAS, 50-50 and 80-20) indicates that the cross-section of the damaged fringe still contributes to the LBC slightly. The deviation from the theoretical assumption increased with a decreasing CaOcontent for AABs. Only the test data for AAFA contradict the theoretical assumption. AAFA, a nearly Ca-free AAB, displays an extreme behavior since both the test data and the curve gradient have shifted to the negative scale. The negative scale is the result of the increasing compressive strength based on the new mineral phase (Fig. 3) as well as the decreasing porosity (Table 4). These results show that the acid resistance as representative of the concrete durability properties of AABs is largely affected by the Ca-content of the binder. Where AAS and Ca-rich AABs showed enhanced, but still cement-like durability properties (see Figs. 2 and 3), low-Ca AAB showed a significant increase in chemical resistance. Following the proposal of [38], and in accordance with the experimental results obtained herein (Fig. 6), a CaO limiting value of <10 w.-%. appears appropriate for the classification of AAB durability properties. 6. Conclusions The type of the binder used for mortars and concretes largely influences the material resistance, in particular the acid resistance of the strength-forming texture as well as the damage mechanisms, as a consequence of an organic acid exposure. The most relevant findings of the current study are summarized in the following: 1. Mortars and concretes based on alkali-activated binders possess an improved resistance to organic acids compared to Portland cement, with and without the substitution of siliceous fly ash. 2. The acid resistance of the strength-forming phases was linked to the binder composition. Therefore, the resistance increases with a decreasing CaO-content: 3. C-S-H (Portland cement) C-A-S-H (Ca-rich single or multi component AAB) N-A-S-H (low-Ca AAB and Geopolymer). A CaO limiting value of 10 m.-% seems to be appropriate to distinguish concrete durability properties of alkali-activated low-Ca and high-Ca binders. Additional tests for AABs with CaO contents of between 3.6 and 10.9 w.-% are necessary for further limitation. 4. The decrease in the Ca-content of alkali-activated binders was accompanied by a reduction of the mass loss and an improvement of the residual compressive strength. The reduction of

mass loss largely depends on the low Ca-leaching potential of the reaction products of AABs causing low erosion and low degradation depths. 5. The low-Ca AAB based on an alkali-activated, low-Ca fly ash (AAFA) exhibits the lowest mass loss and degradation depth despite having the lowest strength class and highest porosity. Furthermore, in low-Ca AAB mortars as well as in low-Ca AAB concretes, an increase in strength during ongoing exposure could be proven. It is assumed that this increase in strength is probably related to a densification of the exposed surface and the formation of anorthite. 6. In the near-surface area of the low-Ca AAB sample (AAFA), Ca2+ in the acid solution was integrated in minerals like anorthite. The porosity therefore decreased and the strength increased. The verifiable, subsequent incorporation of Ca2+ may be a promising alternative and a binder-specific curing method for low-Ca AABs in terms of strength enhancement and permeability reduction. 7. The relationship between the microscopically-obtained degradation depth and the prediction model based on residual compressive strength leads to the conclusion that the lower the CaO content, the lower the degradation depth. For a reliable assessment of the degradation depth or the material resistance, it is recommended that validation criteria be extended and that optically obtained degradation depths are linked to mass loss and residual strength.

Acknowledgements The authors gratefully acknowledge the technical support of the Leipzig Institute for Materials Research and Testing (MFPA Leipzig GmbH) represented by Mr. Karsten Hellmich. References [1] V. Pavlík, Corrosion of hardened cement paste by acetic and nitric acids Part III: influence of water/cement ratio, Cem. Concr. Res. 26 (3) (1996) 475–490. [2] R. Hüttl, B. Hillemeier, High performance concrete: an example of acid resistance, BFT Int. 1 (2000) 52–60. [3] D. Roy, P. Arjunan, M. Silsbee, Effect of silica fume, metakaolin, and lowcalcium fly ash on chemical resistance of concrete, Cem. Concr. Res. 31 (12) (2001) 1809–1813. [4] A. Bertron, G. Escadeillas, J. Duchesne, Cement pastes alteration by liquid manure organic acids: chemical and mineralogical characterization, Cem. Concr. Res. 34 (10) (2004) 1823–1835. [5] V. Pavlík, S. Uncˇík, The rate of corrosion of hardened cement pastes and mortars with additive of silica fume in acids, Cem. Concr. Res. 27 (11) (1997) 1731–1745. [6] N. de Belie, M. Debruyckere, D. van Nieuwenburg, B. de Blaere, Attack of concrete floors in pig houses by feed acids: Influence of fly ash addition and cement-bound surface layers, J. Agric. Eng. Res. 68 (2) (1997) 101–108. [7] N. de Belie, H.J. Verselder, B. de Blaere, D. van Nieuwenburg, R. Verschoore, Influence of the cement type on the resistance of concrete to feed acids, Cem. Concr. Res. 26 (11) (1996) 1717–1725. [8] E. Gruyaert, P. van den Heede, M. Maes, N. de Belie, Investigation of the influence of blast-furnace slag on the resistance of concrete against organic acid or sulphate attack by means of accelerated degradation tests, Cem. Concr. Res. 42 (1) (2012) 173–185. [9] J. Kiekbusch, Acid Attack on Cement Based Materials (in German: Säureangriff auf zementgebundene Materialien), Shaker, Aachen, 2007. [10] Babushkin VI, Thermodynamics of silicates [Place of publication not identified]: Springer-Verlag, Berlin An, 2012. [11] A. Koenig, F. Dehn, Biogenic acid attack on concretes in biogas plants, Biosyst. Eng. 147 (2016) 226–237. [12] A. Allahverdi, F. Skvara, Acidic corrosion of hydrated cement based materials: part 2 kinetics of the phenomenon and mathematical models, Ceram. Silik. 44 (4) (2000) 152–160. [13] C. Shi, J. Stegemann, Acid corrosion resistance of different cementing materials, Cem. Concr. Res. 30 (5) (2000) 803–808. [14] A.M. Rashad, Y. Bai, P. Basheer, N.C. Collier, N.B. Milestone, Chemical and mechanical stability of sodium sulfate activated slag after exposure to elevated temperature, Cem. Concr. Res. 42 (2) (2012) 333–343.

A. Koenig et al. / Construction and Building Materials 151 (2017) 405–413 [15] D. Bondar, C.J. Lynsdale, N.B. Milestone, N. Hassani, Sulfate resistance of alkali activated pozzolans, Int. J. Concr. Struct. Mater. 9 (2) (2015) 145–158. [16] N.K. Lee, H.K. Lee, Influence of the slag content on the chloride and sulfuric acid resistances of alkali-activated fly ash/slag paste, Cem. Concr. Compos. 72 (2016) 168–179. [17] T. Bakharev, Resistance of geopolymer materials to acid attack, Cem. Concr. Res. 35 (4) (2005) 658–670. [18] R.R. Lloyd, J.L. Provis, Jannie S.J. van Deventer, Acid resistance of inorganic polymer binders. 1. Corrosion rate, Mater. Struct. 45 (1–2) (2012) 1–14. [19] J. Temuujin, A. Minjigmaa, M. Lee, N. Chen-Tan, A. van Riessen, Characterisation of class F fly ash geopolymer pastes immersed in acid and alkaline solutions, Cem. Concr. Compos. 33 (10) (2011) 1086–1091. [20] A. Koenig, F. Dehn, Main considerations for the determination and evaluation of the acid resistance of cementitious materials, Mater, Struct, 2015. [21] König, A., 2013. Biogenic acid attack on concrete in biogas plants. Damage mechanism and development potential (Dissertation). [22] A. Herrmann, A. König, F. Dehn, Structural concrete based on alkali-activated binders: terminology, reaction mechanisms, mix design, and performance, Struct. Concr. (2017). [23] Scrivener, K.L., Young. J.F., 1997. Mechanisms of chemical degradation of cement-based systems. In: Proceedings of the Material Research Society’s Symposium on Mechanisms of Chemical Degradation of Cement-based Systems, Boston, USA, 27-30 November 1995, London, New York, E & FN Spon. [24] D. Damidot, F.P. Glasser, Thermodynamic investigation of the CaO Al2O3 CaSO4 H2O system at 25 °C and the influence of Na2O, Cem. Concr. Res. 23 (1) (1993) 221–238. [25] A. Gabrisová, J. Havlica, S. Sahu, Stability of calcium sulphoaluminate hydrates in water solutions with various pH values, Cem. Concr. Res. 21 (6) (1991) 1023–1027. [26] C.J. Warren, E.J. Reardon, The solubility of ettringite at 25 °C, Cem. Concr. Res. 24 (8) (1994) 1515–1524. [27] R.E. Beddoe, H.W. Dorner, Modelling acid attack on concrete: part I. The essential mechanisms, Cem. Concr. Res. 35 (12) (2005) 2333–2339. [28] Grandia, F., Galindez. J.-M., Molinero. J., Arcos. D., 2010. Evaluation of low-pH cement degradation in tunnel plugs and bottom plate systems in the frame of SR-Site. Stockholm. [29] A. Koenig, F. Dehn, Acid resistance of ultra high-performance concrete (UHPC), in: K. Sobolev, S.P. Shah (Eds.), Nanotechnology in Construction, Springer International Publishing, Cham, 2015, pp. 317–323. [30] J. van Aardt, S. Visser, Calcium hydroxide attack on feldspars and clays: possible relevance to cement-aggregate reactions, Cem. Concr. Res. 7 (6) (1977) 643–648. [31] K. Torii, M. Kawamura, Effects of fly ash and silica fume on the resistance of mortar to sulfuric acid and sulfate attack, Cem. Concr. Res. 24 (2) (1994) 361– 370. [32] Z.-T. Chang, X.-J. Song, R. Munn, M. Marosszeky, Using limestone aggregates and different cements for enhancing resistance of concrete to sulphuric acid attack, Cem. Concr. Res. 35 (8) (2005) 1486–1494. [33] B.K. Marsh, R.L. Day, Pozzolanic and cementitious reactions of fly ash in blended cement pastes, Cem. Concr. Res. 18 (2) (1988) 301–310. [34] B. Meng, Limitations for the applicability of pozzolans in concrete, FraunhoferIRB-Verlag, Stuttgart, 1998. [35] I.G. Richardson, The nature of the hydration products in hardened cement pastes, Cem. Concr. Compos. 22 (2000) 97–113. [36] B. Lothenbach, K. Scrivener, R.D. Hooton, Supplementary cementitious materials, Cem. Concr. Res. 41 (12) (2011) 1244–1256.

413

[37] A.V. Girão, I.G. Richardson, R. Taylor, R. Brydson, Composition, morphology and nanostructure of C–S–H in 70% white Portland cement–30% fly ash blends hydrated at 55 °C, Cem. Concr. Res. 40 (9) (2010) 1350–1359. [38] A. Herrmann, A. König, F. Dehn, Proposal for the classification of alkaliactivated binders and geopolymer binders, Cem. Int. 03 (2015) 62–69. [39] Stark, J., Wicht, B., 2000. In German: Zement und Kalk: Der Baustoff als Werkstoff; mit 90 Tabellen. Basel [u.a.]: Birkhäuser. [40] I.G. Richardson, G.W. Groves, Microstructure and microanalysis of hardened cement pastes involving ground granulated blast-furnace slag, J. Mater. Sci. 27 (22) (1992) 6204–6212. [41] W. Kunther, B. Lothenbach, J. Skibsted, Influence of the Ca/Si ratio of the C–S–H phase on the interaction with sulfate ions and its impact on the ettringite crystallization pressure, Cem. Concr. Res. 69 (2015) 37–49. [42] J.M. Richardson, J.J. Biernacki, P.E. Stutzman, D.P. Bentz, Stoichiometry of slag hydration with calcium hydroxide, J. Am/ Ceram. Soc. 85 (4) (2002) 947–953. [43] O. Oueslati, J. Duchesne, The effect of SCMs on the corrosion of rebar embedded in mortars subjected to an acetic acid attack, Cem. Concr. Res. 42 (2) (2012) 467–475. [44] A. Roy, Sulfur speciation in granulated blast furnace slag: An X-ray absorption spectroscopic investigation, Cem. Concr. Res. 39 (8) (2009) 659–663. [45] H. El-Didamony, K. Khalil, M. El-Attar, Physicochemical characteristics of fired clay-limestone mixes, Cem. Concr. Res. 30 (1) (2000) 7–11. [46] J. Bai, A. Chaipanich, J. Kinuthia, M. O’Farrell, B. Sabir, S. Wild, et al., Compressive strength and hydration of wastepaper sludge ash–ground granulated blastfurnace slag blended pastes, Cem. Concr. Res. 33 (8) (2003) 1189–1202. [47] C.-W. Wu, C.-J. Sun, S.-H. Gau, C.-L. Hong, C.-G. Chen, Mechanochemically induced synthesis of anorthite in MSWI fly ash with kaolin, J. Hazard. Mater. 244–245 (2013) 412–420. [48] M. Enders, The CaO distribution to mineral phases in a high calcium fly ash from Eastern Germany, Cem. Concr. Res. 26 (2) (1996) 243–251. [49] D.M. Roy, E.L. White, C.A. Langton, K.G. Zimmerman, Hydrated calcium aluminosilicate cements for hydrothermal bonding, Cem. Concr. Res. 8 (4) (1978) 509–511. [50] R.R. Lloyd, J.L. Provis, Jannie S.J. van Deventer, Pore solution composition and alkali diffusion in inorganic polymer cement, Cem. Concr. Res. 40 (9) (2010) 1386–1392. [51] L.-P. Liu, X.-M. Cui, Y. He, S.-D. Liu, S.-Y. Gong, The phase evolution of phosphoric acid-based geopolymers at elevated temperatures, Mat. Lett. 66 (1) (2012) 10–12. [52] C.-M. Guo, K.-T. Wang, M.-Y. Liu, X.-H. Li, X.-M. Cui, Preparation and characterization of acid-based geopolymer using metakaolin and disused polishing liquid, Ceram. Int. 42 (7) (2016) 9287–9291. [53] D.S. Perera, J.V. Hanna, J. Davis, M.G. Blackford, B.A. Latella, Y. Sasaki, et al., Relative strengths of phosphoric acid-reacted and alkali-reacted metakaolin materials, J. Mater. Sci. 43 (19) (2008) 6562–6566. [54] E.H. Oelkers, J. Schott, Experimental study of anorthite dissolution and the relative mechanism of feldspar hydrolysis, Geochim. Cosmochim. Acta 59 (24) (1995) 5039–5053. [55] F. Steinhäußer, K. Hradil, S. Schwarz, W. Artner, M. Stöger-Pollach, A. SteigerThirsfeld, et al., Wet chemical porosification of LTCC in phosphoric acid: Anorthite forming tapes, J. Eur. Ceram. Soc. 35 (15) (2015) 4181–4188. [56] W. Loewenstein, The distribution of aluminum in the tetrahedra of silicates and aluminates, Am. Mineral. 39 (1954) 92–96. [57] C.K. Yip, G.C. Lukey, J.L. Provis, Jannie S.J. van Deventer, Effect of calcium silicate sources on geopolymerisation, Cem. Concr. Res. 38 (4) (2008) 554–564.