Decalcification resistance of alkali-activated slag

Journal of Hazardous Materials 233–234 (2012) 112–121

Contents lists available at SciVerse ScienceDirect

Journal of Hazardous Materials journal homepage: www.elsevier.com/locate/jhazmat

Decalcification resistance of alkali-activated slag ´ Nataˇsa Marjanovic, ´ Violeta Nikolic´ Miroslav M. Komljenovic´ ∗ , Zvezdana Baˇscˇ arevic, Institute for Multidisciplinary Research, University of Belgrade, Kneza Viˇseslava 1, 11030 Belgrade, Serbia

h i g h l i g h t s     

The effects of decalcification on properties of alkali-activated slag were studied. Decalcification was performed by concentrated NH4 NO3 solution (accelerated test). Portland-slag cement (CEM II/A-S 42.5 N) was used as a benchmark material. Decalcification led to strength decrease and noticeable structural changes. Alkali-activated slag showed significantly higher resistance to decalcification.

a r t i c l e

i n f o

Article history: Received 5 November 2011 Received in revised form 6 June 2012 Accepted 29 June 2012 Available online 6 July 2012 Keywords: Alkali-activated slag Decalcification Accelerated leaching

a b s t r a c t This paper analyses the effects of decalcification in concentrated 6 M NH4 NO3 solution on mechanical and microstructural properties of alkali-activated slag (AAS). Portland-slag cement (CEM II/A-S 42.5 N) was used as a benchmark material. Decalcification process led to a decrease in strength, both in AAS and in CEM II, and this effect was more pronounced in CEM II. The decrease in strength was explicitly related to the decrease in Ca/Si atomic ratio of C–S–H gel. A very low ratio of Ca/Si ∼0.3 in AAS was the consequence of coexistence of C–S–H(I) gel and silica gel. During decalcification of AAS almost complete leaching of sodium and tetrahedral aluminum from C–S–H(I) gel also took place. AAS showed significantly higher resistance to decalcification in relation to the benchmark CEM II due to the absence of portlandite, high level of polymerization of silicate chains, low level of aluminum for silicon substitution in the structure of C–S–H(I), and the formation of protective layer of polymerized silica gel during decalcification process. In stabilization/solidification processes alkali-activated slag represents a more promising solution than Portland-slag cement due to significantly higher resistance to decalcification. © 2012 Elsevier B.V. All rights reserved.

1. Introduction Stabilization/solidification (S/S) is accepted as a wellestablished technique for the treatment of hazardous and radioactive wastes prior to re-use or final disposal [1–4]. The degree of effectiveness of S/S products is defined basically by mechanical and structural stability. Cement-based materials are most commonly used due to their low cost, significant durability and simple processing techniques [5–10]. The main purpose of the leaching studies of radioactive or other hazardous wastes that are incorporated in blocks of suitable embedding materials is to assess their potential hazard to the environment, when these blocks come into contact with water during long-term storage or disposal [7].

∗ Corresponding author. Tel.: +381 11 20 85 048; fax: +381 11 30 55 289. E-mail addresses: [email protected], [email protected] ´ [email protected] (Z. Baˇscˇ arevic), ´ [email protected] (M.M. Komljenovic), ´ [email protected] (V. Nikolic). ´ (N. Marjanovic), 0304-3894/$ – see front matter © 2012 Elsevier B.V. All rights reserved. http://dx.doi.org/10.1016/j.jhazmat.2012.06.063

An important and widespread process of degradation of cementbased materials is decalcification, which is the consequence of calcium ions migration from solidified cement matrix into surrounding aggressive medium [11]. Since decalcification is a very slow reaction, this is not the degradation process frequently seen in cement-based structures. Decalcification of cement-based materials has been identified as an important issue for the long-term radioactive waste disposal in deep geological formations [12–24]. As some radionuclides show stability over a longer period of time (period of half-decay of several hundred years), decalcification of cement-based materials affected by water represents the worst case scenario, which is to be considered seriously when designing containers for radioactive waste disposal. Decalcification also affects concrete structures which have been in contact with pure or acidic waters for a longer period of time: dams, tunnels, pools, water pipes, etc. [25]. Most often, decalcification has been analyzed in ordinary Portland cement (OPC or CEM I) [12–17,26–34]. Different methods have been developed to study the process of decalcification. Water was frequently used as the aggressive medium: pure and mineralized [33,34], distilled [27], as well as deionized [11,14,15,29,31].

M.M. Komljenovi´c et al. / Journal of Hazardous Materials 233–234 (2012) 112–121

A whole palette of accelerated tests was developed in order to examine the process of decalcification in real time. In some cases organic acids were used [30], while in others gradient of electric potential [14]. Sulfate salts, such as (NH4 )2 SO4 [26,28,35] and Na2 SO4 [35], were also used. Chemical, mineral and mechanical similarity of leaching by water and concentrated solution of ammonium nitrate was established by Carde and Franc¸ois [13]. Decalcification by highly concentrated 6 M NH4 NO3 solution offers some key advantages compared to leaching by water. The rate of leaching is increased by two orders of magnitude and additionally, the leaching reaction in 6 M NH4 NO3 solution is almost entirely a pure decalcification, i.e. there is a small loss of silicon, even at very low ratios of Ca/Si = 0.3 [11]. Therefore, instead of deionized water, most tests are carried out with strongly acidified solutions, such as ammonium nitrate [11–13,16,17,19,24,26,28,32,35–37]. Although calcium leaching is a well-analyzed phenomenon, a solution that would lead to its prevention has yet to be found, notwithstanding serious efforts to minimize its harmful effect. As portlandite is the weakest link in the cement system, it is clear that cementitious materials, which contain the lowest amount of portlandite, show the strongest resistance to the process of leaching. In order to slow down the process of calcium leaching, along with the ordinary Portland cement, different authors have used various additives: silica fume [12,14,31,32], blast furnace slag [14], fly ash [31], and calcium-carbonate [27]. Along with these materials, the effect of the process of decalcification on properties of other types of cement was also examined: CEM II [24,34,37], CEM III [30,34], CEM V [34], white Portland cement (WPC) [11,35], sulfate resistant Portland cement (SRPC) [28,29,35], as well as low-heat Portland cement and high early-strength Portland cement [15]. The process of decalcification of hydrated clinker minerals was also investigated: tricalcium silicate (C3 S) [11,27,35,36], tricalcium aluminate (C3 A) and tetracalcium aluminofferite (C4 AF) [27], as well as C–S–H [36]. In S/S processes alkali-activated cements have been recognized as a more promising option than Portland cement, due to lower leachability of contaminants from alkali-activated cement stabilized hazardous and radioactive wastes [4,38–46]. Alkali-activated cement consists of an alkaline activator and cementitious material, such as blast furnace slag, coal fly ash, phosphorus slag, steel slag, red mud, and metakaolin, or a combination of two or more of them. However, to the best of our knowledge and according to available literature, there has been no research on the effects of decalcification process on the properties of alkaliactivated slag. This paper investigates the effects of decalcification process on mechanical and microstructural properties of alkali-activated slag (AAS).

2. Experiment design The design of experimental research is based on two main states of examined materials: (a) Non-degraded material – initial, i.e. reference state (cured in a humid chamber), and (b) Chemically degraded material – asymptotic final state (exposed to accelerated leaching in concentrated NH4 NO3 solution). These two main states can be considered as asymptotic physical states of cementitious materials, especially those used in concrete structures exposed to long term impact of pure or mineralized water [17].

113

Table 1 Chemical composition (%, m/m) and physical characteristics of GBFS and CEM II. Composition/characteristics investigated LOI at 1000 ◦ C SiO2 Al2 O3 Fe2 O3 CaO MgO SO3 S MnO Na2 O K2 O Sum Density (kg/m3 ) Specific surface area (Blaine) (m2 /kg)

GBFS 2.13 37.50 7.27 0.73 38.48 10.86 0.39 1.51 0.30 0.54 0.26 99.97 2890 390

CEM II 2.80 21.70 4.62 2.98 62.53 1.56 2.44 0.00 0.22 0.45 0.60 99.90 3020 380

Decalcification was performed by concentrated 6 M NH4 NO3 solution in a period of 90 days. NH4 NO3 solution was completely renewed every 30 days. Kinetics of decalcification process was investigated based on the changes in mortar compressive strength. Microstructural changes were investigated on the paste by X-ray diffraction and SEM/EDS analysis. Numerous differences in experimental conditions used to examine Portland cement systems leave little space for comparison of decalcification process of AAS and the data from literature referring to Portland cement. Consequently, in these experiments, Portlandslag cement (CEM II/A-S 42.5 N) was used as a benchmark material and the decalcification process was investigated under equal conditions for both systems, thus making this comparison credible. 2.1. Materials The following materials were used: • Granulated blast furnace slag (GBFS) – “U.S. Steel”, Serbia. • Sodium silicate–water glass (Na2 O·nSiO2 ) – “Galenika - Magmasil”, Serbia • Sodium hydroxide (98% NaOH) – “Zorka-Pharm”, Serbia. • Portland-slag cement (CEM II/A-S 42.5 N) – “Titan”, Serbia • Ammonium nitrate (99% NH4 NO3 ) – “Superlab”, Serbia. Sodium silicate was used as an alkaline activator. Starting sodium silicate modulus n = SiO2 /Na2 O (mass ratio) was 2.97 (9.43% Na2 O, 28.0% SiO2 ). In order to reduce undesirable shrinking and avoid fast setting of AAS, low sodium silicate modulus (n = 0.6) was used in all experiments, while Na2 O concentration was 4% in respect to the slag mass [47]. Sodium silicate modulus was adjusted by adding NaOH. Chemical composition and physical characteristics of GBFS and CEM II are given in Table 1. GBFS was ground so that its specific surface area (Blaine) was approximately 400 m2 /kg. 2.2. Sample preparation AAS paste was prepared by adding activator to water and then mixing it with ground slag. Water/binder ratio was 0.25 (water represents the total amount of water in the system, including water from the activator, while binder represents the total slag mass and solid part of the activator). Water/cement ratio was 0.25. Sample dimensions were 25 mm × 25 mm × 30 mm. Sample labels are given in Table 2. Mortar prisms (40 mm × 40 mm × 160 mm) were prepared according to Serbian standard SRPS EN 196-1 (2008), which is in compliance with European EN 196-1 standard. Cement/sand

114

M.M. Komljenovi´c et al. / Journal of Hazardous Materials 233–234 (2012) 112–121

Table 2 Sample labels. Time (days)

28 28 + 30 28 + 60 28 + 90

Table 3 pH of 6 M NH4 NO3 solution. Reference samples

Degraded samples – after leaching (6 M NH4 NO3 )

AAS

CEM II

AAS

CEM II

REF0z REF1z REF2z REF3z

REF0c REF1c REF2c REF3c

– Iz IIz IIIz

– Ic IIc IIIc

and AAS/sand mass ratios were 1:3. Water/cement ratio was 0.5 and water/binder 0.43. It is well known that water/cement, i.e. water/binder ratio has a significant effect on the degradation kinetics [29]. In AAS case water/binder ratio of 0.5 could not be used due to the bleeding. In order to make these two systems comparable, the AAS water/binder ratio was determined providing the same rheological properties (flow table test) as benchmark CEM II. Mortar prisms were kept in the mold for the first 24 h, in an airconditioned room (20 ± 2 ◦ C; 50 ± 5% relative humidity). After being taken out of the mold the prisms were cured in a humid chamber (20 ± 2 ◦ C; 90 ± 5% rel. hum.) up to 28 days. Water curing of AAS is not a desirable option, as it would lead to premature leaching and unavoidable loss of strength [48]. 2.3. Decalcification (leaching) resistance pH value of NH4 NO3 solution before and after leaching was measured by pH-meter (pHTESTR30, “EUTECH”, Holland). Complete replacement of NH4 NO3 solution was performed every 30 days over a period of 90 days. In order to test resistance to leaching, after 28 days of curing in a humid chamber, a series of three mortar prisms were placed in plastic containers with NH4 NO3 (solution/solid mass ratio was 4) at 20 ± 2 ◦ C. Mortar prisms were completely immersed in the solution allowing all surfaces to be in contact with the aggressive medium. Mortar strength was tested according to SRPS EN 196-1 (2008) standard. Strength testing was performed by hydraulic press “SEIDNER” after 30, 60 and 90 days of exposure to NH4 NO3 . X-ray diffraction (XRD) analysis was performed on pastes by PHILLIPS PW 1710 powder diffractometer. After being cured (in a humid chamber or NH4 NO3 solution) for a required period of time, the reaction was stopped by grinding the paste samples in isopropanol (1/2 h). The samples were then rinsed with acetone and dried to the constant mass at 50 ◦ C. Microstructure analysis was performed on the fractured surface of samples by scanning electron microscope (SEM, TESCAN VEGA TS 5130 MM). X-ray microanalysis (EDS) was performed by INCA PentaFET-x3-Oxford Instruments. Prior to analysis, all samples were Au-Pd coated. SEM/EDS analysis was performed on an average distance of 2–3 mm from the sample surface exposed to the effects of environment. The results of EDS analysis represent the average values of 20–30 individual EDS analysis for each examined sample.

Sample labels

AAS CEM II

pH before leaching

4.95

pH after leaching 30 days

60 days

90 days

8.45 8.83

8.12 8.67

8.00 8.36

3.2. Mortar compressive strength Exposure of mortar prisms to 6 M NH4 NO3 solution led to a decrease in strength, both in AAS and CEM II (Table 4). After 30 days of leaching compressive strength of CEM II mortar prisms drastically decreased, i.e. relative (degraded/reference) strength was 0.41. After 60 days of leaching CEM II relative strength was 0.19. Over the same period, AAS mortar strength had a significantly lower decrease (relative strength was 0.94 after 30 days and 0.91 after 60 days of leaching). After 90 days of leaching compressive strength of CEM II was practically unchanged. This might mean that CEM II strength has reached its minimum. Over the same period, strength of AAS had an additional 6% of decrease, i.e. relative strength was 0.85. The data shown clearly indicate that AAS showed significantly higher resistance to leaching compared to the CEM II benchmark samples. 3.3. X-ray diffraction analysis The starting slag sample contained amorphous phase along with some crystalline phases: melilite (Ca,Na)2 (Al,Mg,Fe2+ )(Si,Al)2 O7 and merwinite Ca3 Mg(SiO4 )2 (Fig. 1). The main product of alkali-activation was poorly crystallized calcium silicate hydrate C–S–H(I), imperfect version of 1.4 nm tobermorite. It is well-known that C–S–H(I) can accommodate a substantial concentration of defects such as the omission of bridging tetrahedra, or variations in the contents of interlayer Ca and of protons attached to Si–O− [36]. These changes allow variations in Ca/Si ratio, ranging from 0.67 to 1.5 The presence of secondary calcite, formed as a result of carbonation during curing, was also detected. XRD analysis of degraded

3. Results and discussion 3.1. pH of NH4 NO3 solution Portlandite decomposes when pH of the pore solution decreases below 12 [49]. As pH decreases, C–S–H decalcifies, and when pH drops below 9, C–S–H starts to decompose. During the whole experiment, pH value of NH4 NO3 solution was below 9 (Table 3), i.e. the conditions were appropriate for dissolution of both portlandite and C–S–H.

Fig. 1. Diffractograms of AAS reference and degraded samples.

M.M. Komljenovi´c et al. / Journal of Hazardous Materials 233–234 (2012) 112–121

115

Table 4 Mortar compressive strength of AAS and CEM II samples. Time (days)

Compressive strength (N/mm2 ) Reference samples

28 28 + 30 28 + 60 28 + 90 a

Relative strength (degraded/reference samples) Degraded samples (6 M NH4 NO3 )

AAS

CEM II

AAS

CEM II

AAS

CEM II

48.7 49.5 50.4 51.5

52.6 55.9 59.2 61.2

48.7 46.7 46.1 43.9

52.6 23.0 11.5 11.0

1.00a 0.94 0.91 0.85

1.00a 0.41 0.19 0.18

Initial strength – before leaching (coefficient 1.00).

AAS indicated the presence of melilite and merwinite, crystalline phases originating from the starting slag, and C–S–H(I). Possible crystalline products of the reaction with NH4 NO3 in the observed period were not noticed, but the intensity of peaks corresponding to merwinite significantly decreased. This practically means that NH4 NO3 solution caused intensive merwinite dissolution. The starting CEM II sample contained main clinker minerals: C3 S, C2 S, C3 A, and C4 AF (Fig. 2). The main products of hydration were C–S–H(I) and portlandite (calcium hydroxide). The presence of secondary calcite was also detected. XRD analysis of degraded CEM II indicated the presence of C–S–H(I), portlandite, and secondary calcite. The intensity of peaks corresponding to portlandite significantly decreased in the observed period, which means that NH4 NO3 solution causes intensive portlandite dissolution. C4 AF is the most stable non-hydrated phase in cement paste exposed to acidic aggressive medium [30,50].

distributed inclusions in a glassy phase [52]. The decrease of peak intensities corresponding to merwinite in AAS degraded samples was observed by XRD analysis (Fig. 1). This means that partial solubility of unreacted slag grains, particularly merwinite, is caused by the effects of NH4 NO3 . Fig. 5a (BSE mode) shows the microstructure of CEM II reference sample with unreacted slag or clinker grains incorporated in the matrix of reaction products – C–S–H(I) and clearly visible IP of reaction. The microstructure of CEM II reference sample, consisting of reaction products of different shapes, is shown in Fig. 5b Prominent porosity of CEM II degraded samples, caused by the effects of NH4 NO3 , is shown in Fig. 6. Microstructural changes of CEM II due to leaching were not explored in detail, considering that, as mentioned in Section 1, this system has already been widely analyzed. 3.5. X-ray microanalysis (EDS)

3.4. Scanning electron microscopy (SEM) SEM analysis of AAS reference samples showed unreacted slag grains incorporated in the matrix of reaction products – C–S–H(I) (Fig. 3). The differentiated border zone between unreacted slag grains and surrounding matrix, or the so-called inner products (IP) of reaction, was also visible (Fig. 3a; BSE mode). IP are probably responsible for possible slowdown in reaction rate, while a significant part of slag remains unreacted [51]. Foil-like morphology of C–S–H(I) is shown in Fig. 3b. SEM analysis of AAS degraded samples showed unreacted slag grains incorporated in the matrix of reaction products – C–S–H(I), as well as IP of reaction (Fig. 4a; BSE mode). Fig. 4b (BSE mode) shows partial solubility of slag grains. This phenomenon was not detected within AAS reference samples. Slag crystalline phases (melilite and merwinite), of different sizes and forms, can be found as randomly

Fig. 2. Diffractograms of CEM II reference and degraded samples.

3.5.1. Average content of main elements and their ratios During curing of AAS reference samples in a humid chamber (from 28 to 28 + 90 days), the average content of main elements present and their ratios were variable (Table 5). Ca/Si atomic ratio decreased from value 0.84 (after 28 days) to 0.77 (after 28 + 90 days). These values are in accordance with literature data for a similar system [53]. A gradual decrease in Ca/Si ratio during time indicated an increase of length of silicate chains in the C–S–H(I) structure [36,52]. Humid chamber conditions might also have led to partial decalcification of C–S–H(I) due to carbonation [54]. Al/Si atomic ratio ranging from 0.14 (after 28 days) to 0.25 (after 28 + 90 days), indicated possible increase in the level of aluminum substitution in C–S–H(I) [51,55]. On the other hand, aluminum substitution in C–S–H(I), which is more favorable from the energy aspect, leads to the increase in average length of silicate chains [56], which is in accordance with a decrease in Ca/Si ratio. While exposing AAS samples to NH4 NO3 for a period of 28 to 28 + 90 days, the average content of main elements present in AAS was also variable (Table 5). Ca/Si atomic ratio intensively decreased from value 0.84 (after 28 days) to 0.33 (after 28 + 90 days). The decrease in Ca/Si ratio during leaching, i.e. decalcification of C–S–H(I) gel, occurred intensively in a period 28 + 60 days, decreasing from value 0.84 to 0.31. After that, the Ca/Si ratio practically did not change, i.e. a plateau of Ca/Si ∼0.3 was reached. Preservation of silicon in the leached solid phase prevented complete dissolution of solid C–S–H(I) gel [11,21,31]. According to Taylor’s model [52], structure of C–S–H(I) with low Ca/Si = 0.66 ratio shows similarities with 1.4 nm tobermorite, but with the presence of certain defects. C–S–H(I) comprises of tetrahedral, very long silicate layers (chains) linked to octahedral layer of calcium: Te–Oc–Te (Te – tetrahedral, Oc – octahedral) structure. It is presumed that the silicate chains are infinite. Silicate chains are composed of dimers which are linked to bridging tetrahedra, balanced by a proton (H+ ). At the atomic ratio Ca/Si < 0.66 there are

116

M.M. Komljenovi´c et al. / Journal of Hazardous Materials 233–234 (2012) 112–121

Fig. 3. Microstructure of AAS reference samples: (a) REF1z; (b) REF2z.

Fig. 4. Microstructure of AAS degraded samples: (a) Iz; (b) IIz.

Fig. 5. Microstructure of CEM II reference samples: (a) REF0c; (b) REF2c.

M.M. Komljenovi´c et al. / Journal of Hazardous Materials 233–234 (2012) 112–121

117

Fig. 6. Microstructure of CEM II degraded samples: (a) Ic; (b) IIc.

Table 5 Average content of AAS main elements (atomic %) and their ratios. Sample labels

Time (days)

Si

Ca

Al

Mg

Na

REF0z REF1z REF2z REF3z

28 28 + 30 28 + 60 28 + 90

14.06 14.53 10.39 10.84

11.99 11.60 8.46 8.22

2.01 2.38 2.07 2.29

2.54 3.41 3.22 4.34

5.28 9.84 10.85 10.99

0.84 0.83 0.80 0.77

0.14 0.19 0.22 0.25

Iz IIz IIIz

28 + 30 28 + 60 28 + 90

15.06 21.85 23.34

9.40 5.99 7.19

2.56 2.62 3.28

4.61 4.96 6.33

0.26 0.15 0.11

0.65 0.31 0.33

0.18 0.13 0.15

no more calcium ions in the interlayer space, so C–S–H(I) starts to decompose [11,57]. As a consequence of the decalcification process the structure becomes more heterogeneous, and the composition of C–S–H(I) gel becomes richer in silicon (Fig. 7). Thus, a very low ratio of Ca/Si ∼0.3 can be explained by coexistence of C–S–H(I) with low Ca/Si ratio and polymerized silica gel [11,36,51]. The mechanism of C–S–H(I) decomposition due to leaching (decalcification) is quite similar to the mechanism of decalcification of C–S–H(I) due to carbonation [11,54,58]. Al/Si atomic ratio ranges in a very narrow interval from 0.14 (after 28 days) to 0.15 (after 28 + 90 days). Compared to the reference samples, which in the same period showed an increase in

Fig. 7. Silicon and calcium content of AAS reference and degraded samples after 28 + 60 days.

Ca/Si

Al/Si

Al/Si ratio, a slight change in this ratio during leaching was probably related to aluminum leaching [59], i.e. to a lower degree of its substitution in the C–S–H(I) structure. Sodium can be partially adsorbed by C–S–H, while incorporation of sodium in the C–S–H structure is quite limited [60]. As sodium is present in the pore system in the form of free ions [51,60], it was leached almost completely already after the first 30 days (Table 5). 3.5.2. Mg/Si ratio versus Al/Si in AAS samples Concentration of calcium is uniform only in non-degraded sample, as well as in asymptotically decalcified sample [16]. Taking into consideration that calcium leaching from solid paste causes its inhomogeneous distribution, it appears that it is better for compositional analysis to be normalized in relation to silicon content than in relation to the calcium content [51]. Fig. 8 shows the results of compositional analysis. Magnesium was clearly always present with C–S–H(I) gel (Fig. 8), and it is unlikely that the substitution of Mg in C–S–H(I) occurs. Linear dependency of Mg/Si ratio with Al/Si ratio of AAS reference samples indicated the presence of a phase similar to hydrotalcite [52,53]. Since XRD analysis did not confirm the presence of such a crystalline phase, it can be assumed that it is a finely dispersed (Mg-Al hydroxide) gel within C–S–H(I), with organized hydrotalcite structure on a micro- or nanometer level [53,61–63]. In this case, the slope of the plot on the Mg/Si-Al/Si diagram represents the Mg/Al atomic ratio in the hydrotalcite gel (Fig. 8). An average Mg/Al atomic ratio in the hydrotalcite gel was almost constant and amounted to 2.38 (Fig. 8a), which is in accordance with the results of other authors, who reported Mg/Al atomic ratio of 2.3 [51,62]. A nominally similar form of hydrotalcite occurs regardless of the system chemistry, i.e. the environment in which the slag hydrates: activated slag, non-activated, Portland-slag

118

M.M. Komljenovi´c et al. / Journal of Hazardous Materials 233–234 (2012) 112–121

Fig. 8. Mg/Si ratio versus Al/Si in AAS samples: (a) reference; (b) degraded.

cement or slag in the presence of calcium hydroxide [62]. Excessive aluminum that has not been incorporated into hydrotalcite gel, is incorporated in C–S–H (I), substituting silicon predominantly in the position of bridging tetrahedra [53,55,61,64]. Consequently, when ratio Mg/Si = 0, positive X-axis interception (Al/Si = 0.067, Fig. 8a) represents the relative amount of aluminum incorporated into C–S–H(I). Positive X-axis interception also indicates the fact that there is no substitution of Mg in the C–S–H(I) nanostructure [62]. On the other hand, aluminum substitution for silicon in the C–S–H(I) structure leads to the increase of the silicate chain length, as well as the increase in the level of polymerization [56,57,65]. If we assume that the average length (the number of tetrahedra) of the silicate chains is 8–11 [33,36,54], then the value Al/Si = 0.067 indicates that approximately 17–20% of bridging silicon tetrahedra were substituted by aluminum. When testing resistance to leaching, the average Mg/Al atomic ratio was 1.89 (Fig. 8b), with positive X-axis interception almost equal to zero, i.e. practically there was no more aluminum incorporated into C–S–H. Absence of the bridging tetrahedra in silicate chains led to depolymerization of C–S–H, i.e. to breaking of long silicate chains into fragments that contained (3n − 1) tetrahedra [36,56]. Detected decrease in Mg/Al atomic ratio can be the consequence of one of the following processes: (1) Magnesium leaching from hydrotalcite gel and/or (2) Aluminum leaching from C–S–H gel; when released aluminum can be incorporated into polymerized silica gel, which results in forming of aluminosilicate gel [31]. During the leaching process calcium left the C–S–H interlayer space, leaving the silicon not balanced by calcium anymore, but by hydrogen (Ca2+ was replaced by H+ ). Calcium leaching led to the formation of silica gel and coexistence with C–S–H gel. Due to the breaking of silicate chains, aluminum was no longer in the equilibrium position of bridging tetrahedra, so it was released from the C–S–H gel and then leached. A similar occurrence of aluminum release from C–S–H gel was noted in AAS during the process of carbonation [54]. The dissolution of the multi-oxide system, such as the C–S–H gel, in acidic medium, occurs through a series of reactions of metalproton exchange, until its structure disintegrates completely. For instance, the dissolution of basaltic glass in acidic medium undergoes the following stadiums of exchange: (1) Alkali metal H; (2) Ca H; (3) Mg H; (4) Al H; (5) Breaking of Si O bonds, which most likely involves absorption of H2 O rather than Si–H exchange [59]. Based on the results of SEM/EDS analysis, it was confirmed that during the process of C–S–H gel leaching a substitution of

sodium, calcium and aluminum with proton occurred, while the Mg–H exchange was not clearly confirmed. A likely reason might be the fact that Mg was not incorporated in the structure of C–S–H gel, but formed a separate Mg-Al (hydrotalcite) gel. Although the method of calcium leaching by concentrated solution of NH4 NO3 is used to study the structure of C–S–H of different Ca/Si ratios [36], it is obvious that additional research should be conducted, in order to clarify in more detail the destruction of C–S–H structure due to leaching. High atomic ratio Ca/Si ∼2.6, in CEM II reference samples (Table 6), can be explained by the coexistence of C–S–H gel and calcium hydroxide. After 60 days of leaching, Ca/Si atomic ratio rapidly decreased to the value Ca/Si ∼1.4. Although in CEM II reference samples the content of sodium was very low, it is, as is the case in AAS, almost completely leached within the first 30 days. The Al/Si ratio did not change significantly during leaching. 3.6. Correlation of strength and microstructure (EDS analysis) Correlation of compressive strength of reference and degraded samples and Ca/Si atomic ratio is presented in Fig. 9. In AAS (Fig. 9a) and CEM II reference samples (Fig. 9b) the increase in strength (up to 28 + 90 days) was related to the minor changes in the Ca/Si atomic ratio. The decrease in strength of degraded samples was related to the decrease in Ca/Si atomic ratio. A large decrease in strength of CEM II (>80%, Fig. 9b) was related to the large decrease in Ca/Si atomic ratio. The dissolution of portlandite is what usually primarily occurs, while the decalcification of C–S–H gel occurs only if portlandite is inaccessible or locally depleted [11]. During leaching, portlandite dissolves completely and creates higher porosity [32]. Dissolution of portlandite due to NH4 NO3 (Fig. 2), had a destructive effect on mechanical characteristics of CEM II mortar. After almost complete leaching of portlandite (after 28 + 30 days, Ca/Si ∼2.0), a further decrease in strength occurred due to the dissolution of C–S–H, until it reached minimal, or the so-called residual strength [12], with ratio Ca/Si ∼1.4 (after 28 + 60 days). As it was mentioned in Section 3.5, the preservation of silica in the solid phase prevented complete dissolution of solid C–S–H gel. In the AAS case, portlandite was not detected, only C–S–H(I) and crystalline phases of melilite and merwinite, originating from slag (Fig. 1). When the appropriate value of pH was reached, C–S–H gel started to dissolve and then precipitated again, but with a lower Ca/Si ratio, leading to a progressive homogenization of cement paste. Dissolution of C–S–H, as well as merwinite (Figs. 1 and 4b), had a smaller effect on the mechanical characteristics of AAS.

M.M. Komljenovi´c et al. / Journal of Hazardous Materials 233–234 (2012) 112–121

119

Table 6 Average content of CEM II main elements (atomic %) and their ratios. Sample labels

Time (days)

Si

Ca

Al

Mg

Na

Ca/Si

Al/Si

REF0c REF1c REF2c

28 28 + 30 28 + 60

8.89 9.65 8.16

21.58 25.46 20.59

1.07 1.73 1.32

0.54 0.37 0.65

0.78 0.28 0.19

2.60 2.63 2.58

0.13 0.21 0.17

Ic IIc

28 + 30 28 + 60

11.07 15.86

19.29 16.89

2.11 2.80

0.62 0.65

0.08 0.07

2.00 1.39

0.20 0.22

Fig. 9. Correlation of strength and Ca/Si ratio: (a) AAS; (b) CEM II.

Consequently, the dissolution of portlandite and decalcification of C–S–H with high Ca/Si ratio in solid Portland cement paste left behind a very porous, corroded layer, while the absence of portlandite and decalcification of C–S–H with low Ca/Si ratio in AAS resulted in thick protective layer of silica gel [49]. On the other hand, the results of SEM/EDS analysis indicated that the chemistry of C–S–H phase showed more complexity than suggested by XRD analysis. High level of cross linking meant that it was relatively difficult to remove Si from C–S–H by chemical aggression, while aluminum substitution for silicon in a tetrahedral network represented potential weakness due to instability of tetrahedral aluminum [66]. Finally, it can be summed up that AAS had a very pronounced resistance to decalcification, due to: the absence of portlandite, high level of polymerization of silicate network (low Ca/Si atomic ratio ∼0.8), a relatively low level of aluminum substitution (up to 20% of bridging tetrahedra), as well as the formation of protective layer of polymerized silica gel.

4. Conclusions This paper investigates the effects of decalcification process on mechanical and microstructural properties of alkali-activated slag (AAS). The blast furnace slag was activated by the solution of sodium silicate. Portland-slag cement (CEM II/A-S 42.5 N) was used as a benchmark material. Decalcification was performed under the effect of concentrated 6 M NH4 NO3 solution and the testing lasted for 90 days. During decalcification process the decrease in strength occurred, which was explicitly related to the structural changes. In CEM II portlandite dissolution occurred at first, followed by the dissolution of C–S–H gel, which was closely connected with a significant decrease in Ca/Si atomic ratio. In AAS, the presence of portlandite was not detected, only C–S–H(I) was found, as well as crystalline phases of melilite and

merwinite from the slag. Decalcification process of AAS occurred dominantly in C–S–H(I) gel. Although a decrease in Ca/Si atomic ratio occurred, the decrease in strength was significantly lower compared to the benchmark sample CEM II. A very low Ca/Si ratio was the consequence of coexistence of C–S–H(I) gel with low Ca/Si ratio and silica gel. Decalcification also caused almost total leaching of sodium and tetrahedral aluminum from C–S–H(I) gel. It is possible that subsequent incorporation of aluminum in polymerized silica gel occurred. During AAS decalcification there were simultaneous minor structural changes in Mg-Al (hydrotalcite) gel, along with partial dissolution of merwinite in unreacted slag grains. Partial dissolution of unreacted slag grains represents the proof of phase separation in the blast furnace slag. A very pronounced AAS resistance to decalcification process can be explained by the following reasons: (1) Absence of portlandite. (2) High level of polymerization of silicate chains in C–S–H(I) gel structure. (3) Low level of aluminum substitution for silicon in C–S–H(I) gel structure. (4) Formation of protective layer of polymerized silica gel during decalcification process. In stabilization/solidification processes alkali-activated slag represents a more promising solution than Portland-slag cement due to significantly higher resistance to decalcification. Acknowledgements This work was carried out within the project TR34026. The authors are grateful to the Ministry of Education and Science of the Republic of Serbia for financial support. We thank Dr. Aleksandra Rosic´ (Faculty of Mining and Geology, Belgrade University) for XRD analysis.

120

M.M. Komljenovi´c et al. / Journal of Hazardous Materials 233–234 (2012) 112–121

References [1] F.P. Glasser, Fundamental aspects of cement solidification and stabilisation, J. Hazard. Mater. 52 (1997) 151–170. [2] J. Li, J. Wang, Advances in cement solidification technology for waste radioactive ion exchange resins: a review, J. Hazard. Mater. B135 (2006) 443–448. [3] R. Malviya, R. Chaudhary, Factors affecting hazardous waste solidification/stabilization: a review, J. Hazard. Mater. B137 (2006) 267–276. [4] C. Shi, A. Fernandez-Jimenez, Stabilization/solidification of hazardous and radioactive wastes with alkali-activated cements, J. Hazard. Mater. B137 (2006) 1656–1663. [5] X.C. Qiao, C.S. Poon, C.R. Cheeseman, Transfer mechanisms of contaminants in cement-based stabilized/solidified wastes, J. Hazard. Mater. B129 (2006) 290–296. [6] R. Malviya, R. Chaudhary, Leaching behavior and immobilization of heavy metals in solidified/stabilized products, J. Hazard. Mater. B137 (2006) 207–217. [7] K.G. Papadokostaki, A. Savidou, Study of leaching mechanisms of caesium ions incorporated in ordinary Portland cement, J. Hazard. Mater. 171 (2009) 1024–1031. [8] A. Antemira, C.D. Hills, P.J. Carey, K.H. Gardner, E.R. Bates, A.K. Crumbie, Longterm performance of aged waste forms treated by stabilization/solidification, J. Hazard. Mater. 181 (2010) 65–73. [9] M.I. Ojovan, G.A. Varlackova, Z.I. Golubeva, O.N. Burlaka, Long-term field and laboratory leaching tests of cemented radioactive wastes, J. Hazard. Mater. 187 (2011) 296–302. [10] S.B. Eskandera, S.M. Abdel Azizb, H. El-Didamonyc, M.I. Sayed, Immobilization of low and intermediate level of organic radioactive wastes in cement matrices, J. Hazard. Mater. 190 (2011) 969–979. [11] J.J. Chen, J.J. Thomas, H.M. Jennings, Decalcification shrinkage of cement paste, Cem. Concr. Res. 36 (2006) 801–809. [12] C. Carde, R. Franc¸ois, J.M. Torrenti, Leaching of both calcium hydroxide and C–S–H from cement paste: modeling the mechanical behavior, Cem. Concr. Res. 26 (1996) 1257–1268. [13] C. Carde, R. Franc¸ois, Effect of the leaching of calcium hydroxide from cement paste on mechanical and physical properties, Cem. Concr. Res. 27 (1997) 539–550. [14] H. Saito, A. Deguchi, Leaching tests on different mortars using accelerated electrochemical method, Cem. Concr. Res. 30 (2000) 1815–1825. [15] J.S. Ryu, N. Otsuki, H. Minagawa, Long-term forecast of Ca leaching from mortar and associated degeneration, Cem. Concr. Res. 32 (2002) 1539–1544. [16] F.H. Heukamp, F.-J. Ulm, J.T. Germaine, Poroplastic properties of calciumleached cement-based materials, Cem. Concr. Res. 33 (2003) 1155–1173. [17] G. Constantinides, F.-J. Ulm, The effect of two types of C–S–H on the elasticity of cement-based materials: results from nanoindentation and micromechanical modeling, Cem. Concr. Res. 34 (2004) 67–80. [18] F.-J. Ulm, F. Heukamp, J. Germaine, Residual design strength of cementbased materials for nuclear waste storage systems, Nucl. Eng. Des. 211 (2002) 51–60. [19] C. Carde, G. Escadeillas, R. Francois, Use of ammonium nitrate solution to simulate and accelerate the leaching of cemente pastes due to deionized water, Mag. Concr. Res. 49 (1997) 295–301. [20] P. Faucon, F. Adenot, J.F. Jacquinot, J.C. Petit, R. Cabrillac, M. Jorda, Longterm behaviour of cement pastes used for nuclear waste disposal: review of physico-chemical mechanisms of water degradation, Cem. Concr. Res. 28 (1998) 847–857. [21] A.W. Harris, M.C. Manning, W.M. Tearle, C.J. Tweed, Testing of models of the dissolution of cements-leaching of synthetic C–S–H gels, Cem. Concr. Res. 32 (2002) 731–746. [22] C. Gallé, H. Peycelon, P. Le Bescop, Effect of an accelerated chemical degradation on water permeability and pore structure of cement based materials, Adv. Cem. Res. 16 (2004) 105–114. [23] K. Yokozeki, K. Watanabe, N. Sakata, N. Otsuki, Modeling of leaching from cementitious materials used in underground environment, Appl. Clay Sci. 26 (2004) 293–308. [24] F. Agostini, Z. Lafhaj, F. Skoczylas, H. Loodsveldt, Experimental study of accelerated leaching on hollow cylinders of mortar, Cem. Concr. Res. 37 (2007) 71–78. [25] F.P. Glasser, J. Marchand, E. Samson, Durability of concrete – degradation phenomena involving detrimental chemical reactions, Cem. Concr. Res. 38 (2008) 226–246. [26] U. Schneider, S.-W. Chen, The chemomechanical effect and the mechanochemical effect on high-performance concrete subjected to stress corrosion, Cem. Concr. Res. 28 (1998) 509–522. [27] S. Catinaud, J.J. Beaudoin, J. Marchand, Influence of limestone addition on calcium leaching mechanisms in cement-based materials, Cem. Concr. Res. 30 (2000) 1961–1968. [28] R. Jauberthie, F. Rendell, Physicochemical study of the alteration surface of concrete exposed to ammonium salts, Cem. Concr. Res. 33 (2003) 85–91. [29] Y. Maltais, E. Samson, J. Marchand, Predicting the durability of Portland cement systems in aggressive environments – laboratory validation, Cem. Concr. Res. 34 (2004) 1579–1589. [30] A. Bertron, J. Duchesne, G. Escadeillas, Accelerated tests of hardened cement pastes alteration by organic acids: analysis of the pH effect, Cem. Concr. Res. 35 (2005) 155–166.

[31] A. Hidalgo, S. Petit, C. Domingo, C. Alonso, C. Andrade, Microstructural characterization of leaching effects in cement pastes due to neutralisation of their alkaline nature – Part I: Portland cement pastes, Cem. Concr. Res. 37 (2007) 63–70. [32] J.J. Gaitero, I. Campillo, A. Guerrero, Reduction of the calcium leaching rate of cement paste by addition of silica nanoparticles, Cem. Concr. Res. 38 (2008) 1112–1118. [33] M. Moranville, S. Kamali, E. Guillon, Physicochemical equilibria of cementbased materials in aggressive environments – experiment and modeling, Cem. Concr. Res. 34 (2004) 1569–1578. [34] S. Kamali, M. Moranville, S. Leclercq, Material and environmental parameter effects on the leaching of cement pastes: experiments and modelling, Cem. Concr. Res. 38 (2008) 575–585. [35] K.-C. Werner, Y. Chen, I. Odler, Investigations on stress corrosion of hardened cement pastes, Cem. Concr. Res. 30 (2000) 1443–1451. [36] J.J. Chen, J.J. Thomas, H.F.W. Taylor, H.M. Jennings, Solubility and structure of calcium silicate hydrate, Cem. Concr. Res. 34 (2004) 1499–1519. [37] N. Burlion, D. Bernard, D. Chen, X-ray microtomography: application to microstructure analysis of a cementitious material during leaching process, Cem. Concr. Res. 36 (2006) 346–357. [38] C.Y. Rha, S.K. Kang, C.E. Kim, Investigation of the stability of hardened slag paste for the stabilization/solidification of wastes containing heavy metal ions, J. Hazard. Mater. B73 (2000) 255–267. [39] Q. Guangren, L. Yuxiang, Y. Facheng, S. Rongming, Improvement of metakaolin on radioactive Sr and Cs immobilization of alkali-activated slag matrix, J. Hazard. Mater. B92 (2002) 289–300. [40] G. Qian, D.D. Sun, J.H. Tay, Immobilization of mercury and zinc in an alkaliactivated slag matrix, J. Hazard. Mater. B101 (2003) 65–77. [41] Z. Yunsheng, S. Wei, C. Qianli, C. Lin, Synthesis and heavy metal immobilization behaviors of slag based geopolymer, J. Hazard. Mater. 143 (2007) 206–213. [42] M. Izquierdo, X. Querol, J. Davidovits, D. Antenucci, H. Nugteren, C. FernándezPereira, Coal fly ash-slag-based geopolymers: microstructure and metal leaching, J. Hazard. Mater. 166 (2009) 561–566. [43] M. Izquierdo, X. Querol, C. Phillipart, D. Antenucci, M. Towler, The role of open and closed curing conditions on the leaching properties of fly ash-slag-based geopolymers, J. Hazard. Mater. 176 (2010) 623–628. [44] M. Komljenovic, Z. Bascarevic, V. Bradic, Mechanical and microstructural properties of alkali-activated fly ash geopolymers, J. Hazard. Mater. 181 (2010) 35–42. [45] Y. Luna Galiano, C. Fernández Pereira, J. Vale, Stabilization/solidification of a municipal solid waste incineration residue using fly ash-based geopolymers, J. Hazard. Mater. 185 (2011) 373–381. [46] R. Shirley, L. Black, Alkali activated solidification/stabilisation of air pollution control residues and co-fired pulverised fuel ash, J. Hazard. Mater. 194 (2011) 232–242. ´ Effects of dosage and modulus of water glass [47] D. Kriˇzan, B. Zˇ ivanovic, on early hydration of alkali-slag cements, Cem. Concr. Res. 32 (2002) 1181–1188. [48] F. Sajedi, H.A. Razak, The effect of chemical activators on early strength of ordinary Portland cement-slag mortars, Constr. Build. Mater. 24 (2010) 1944–1951. [49] C. Shi, J.A. Stegemann, Acid corrosion resistance of different cementing materials, Cem. Concr. Res. 30 (2000) 803–808. [50] A. Bertron, G. Escadeillas, J. Duchesne, Cement pastes alteration by liquid manure organic acids: chemical and mineralogical characterization, Cem. Concr. Res. 34 (2004) 1823–1835. [51] A.R. Brough, A. Atkinson, Sodium silicate-based, alkali-activated slag mortars. Part I. Strength, hydration and microstructure, Cem. Concr. Res. 32 (2002) 865–879. [52] H.F.W. Taylor, Cement Chemistry, Academic Press Ltd, 1992. [53] M. Ben Haha, G. Le Saout, F. Winnefeld, B. Lothenbach, Influence of activator type on hydration kinetics, hydrate assemblage and microstructural development of alkali activated blast-furnace slags, Cem. Concr. Res. 41 (2011) 301–310. [54] M. Palacios, F. Puertas, Effect of carbonation on alkali-activated slag paste, J. Am. Ceram. Soc. 89 (2006) 3211–3221. [55] J.E. Oh, S.M. Clark, P.J.M. Monteiro, Does the Al substitution in C–S–H(I) change its mechanical property? Cem. Concr. Res. 41 (2011) 102–106. [56] H. Manzano, J.S. Dolado, A. Ayuela, Aluminum incorporation to dreierketten silicate chains, J. Phys. Chem. B 113 (2009) 2832–2839. [57] P. Faucon, A. Delagrave, J.C. Petit, C. Richet, J.M. Marchand, H. Zanni, Aluminum incorporation in calcium silicate hydrates (C–S–H) depending on their Ca/Si ratio, J. Phys. Chem. B 103 (1999) 7796–7802. [58] F. Matsushita, Y. Aono, S. Shibata, Calcium silicate structure and carbonation shrinkage of a tobermorite-based material, Cem. Concr. Res. 34 (2004) 1251–1257. [59] E.H. Oelkers, General kinetic description of multioxide silicate mineral and glass dissolution, Geochim. Cosmochim. Acta 65 (2001) 3703–3719. [60] S.D. Wang, The role of sodium during the hydration of alkali-activated slag, Adv. Cem. Res. 12 (2000) 65–69. [61] I.G. Richardson, A.R. Brough, G.W. Groves, C.M. Dobson, The characterisation of hardened alkali-activated blast-furnace slag pastes and the nature of the calcium silicate hydrate (C–S–H), Cem. Concr. Res. 24 (1994) 813–829. [62] J.M. Richardson, J.J. Biernacki, Stoichiometry of slag hydration with calcium hydroxide, J. Am. Ceram. Soc. 85 (2002) 947–953.

M.M. Komljenovi´c et al. / Journal of Hazardous Materials 233–234 (2012) 112–121 [63] I.G. Richardson, G.W. Groves, Microstructure and microanalysis of hardened cement pastes involving ground granulated blast-furnace slag, J. Mater. Sci. 27 (1992) 6204–6212. [64] A. Fernandez-Jimenez, F. Puertas, I. Sobrados, J. Sanz, Structure of calcium silicate hydrates formed in alkaline-activated slag: influence of the type of alkaline activator, J. Am. Ceram. Soc. 86 (2003) 1389–1394.

121

[65] J. Schneider, M.A. Cincotto, H. Panepucci, 29 Si and 27 Al highresolution NMR characterization of calcium silicate hydrate phases in activated blast-furnace slag pastes, Cem. Concr. Res. 31 (2001) 993–1001. [66] H. Xu, J.L. Provis, J.S.J. van Deventer, P.V. Krivenko, Characterization of aged slag concretes, ACI Mater. J. 105 (2008) 131–139.