Construction and Building Materials 186 (2018) 1072–1081 Contents lists available at ScienceDirect Construction and Building Materials journal homepage: www.elsevier.com/locate/conbuildmat Advances in understanding and analyzing the anti-diffusion behavior in complete carbonation zone of MSWI bottom ash-based alkali-activated concrete Guodong Huang a,b,c, Yongsheng Ji a,c,⇑, Linglei Zhang a,c, Jun Li a,c, Zhihui Hou a,c a b c State Key Laboratory for Geomechanics and Deep Underground Engineering, China University of Mining and Technology, Xuzhou 21116, Jiangsu, China School of Civil Engineering and Construction, Anhui University of Science and Technology, Huainan 232001, Anhui, China School of Mechanics and Civil Engineering, China University of Mining and Technology, Xuzhou 21116, Jiangsu, China h i g h l i g h t s BA displayed a unique anti-diffusion behavior under natural conditions. The anti-diffusion behavior was influenced by amount of sodium silicate used. Anti-diffusion effect of BA significantly reinforced its carbonation resistance. a r t i c l e i n f o Article history: Received 17 January 2018 Received in revised form 7 July 2018 Accepted 7 August 2018 Keywords: Carbonation Municipal solid waste incineration bottom ash Alkali-activated Anti-diffusion Concrete a b s t r a c t This study investigated the carbonation resistance of municipal solid waste incinerator (MSWI) bottom ash alkali-activated concrete (BA) and Portland cement concrete (CC) through the accelerated carbonation test. The results show that BA has lower resistance against carbonation compared to CC under the conditions of the accelerated carbonation experiment. However, BA displayed a unique anti-diffusion behavior under natural conditions which significantly reinforced its carbonation resistance. The antidiffusion behavior was closely related to the amount of liquid sodium silicate in the activator. Moreover, the pH value in complete carbonation zone of BA was much higher compared to that of CC. The main reason is that liquid sodium silicate turned into magadiite in BA and then the magadiite turned back to liquid sodium silicate to release a large amount of OH under low concentrations of CO2. This resulted in a rapid increase in pH value of the pore fluid and generation of the anti-diffusion behavior. Ó 2018 Published by Elsevier Ltd. 1. Introduction Alkali-activated materials are inorganic nonmetallic compounds which are prepared by the action of strong alkali on industrial solid wastes consisting of aluminosilicates . Compared to conventional Portland cement materials, alkali-activated materials have good mechanical performance, high temperature resistance, resistance against acids, alkali, and salts, and low permeability. Moreover, the production of alkali-activated materials does not require high temperature calcination or sintering processes unlike the production of Portland cement. Alkali-activated materials can also undergo complete polymerization at room temperature ⇑ Corresponding author at: State Key Laboratory for Geomechanics and Deep Underground Engineering, China University of Mining and Technology, Xuzhou 21116, Jiangsu, China. E-mail address: email@example.com (Y. Ji). https://doi.org/10.1016/j.conbuildmat.2018.08.038 0950-0618/Ó 2018 Published by Elsevier Ltd. [2–4]. Furthermore, alkali-activated materials have lower energy consumption, almost no pollution, and are easy to recycle and reuse [4–6]. Therefore, they are environmentally friendly green materials which may be used as a building material to replace Portland cement in the future [7–10]. However, studies on the durability of alkali-activated materials have found that the carbonation resistance of alkali-activated concrete is weaker than that of Portland cement concrete, particularly under the conditions of accelerated carbonation test. Laboratory tests of alkali-activated concrete have shown relatively high carbonation rates and this problem is more serious as the concentration of CO2 increases . The current research indicates that due to a certain concentration of CO2 in the atmosphere, it enters into reinforced Portland cement concrete via diffusion and dissolves in the pore liquid, lowering its pH value. The steady decrease in pH value of the pore fluid will cause the deactivated film of reinforcing bar to become unstable and even damaged, resulting in accelerated 1073 G. Huang et al. / Construction and Building Materials 186 (2018) 1072–1081 corrosion of the rebar and then a reduction in bearing capacity . Therefore, rebar corrosion due to carbonation is a common problem for the ordinary Portland cement (OPC) system, particularly in warm and relatively humid environments. Moreover, alkali-activated materials have higher carbonation rates than Portland cement concrete, which makes them more prone to rebar corrosion issues due to carbonation. Several studies have investigated the carbonation of alkaliactivated slag (AAS) concrete and other related alkali-activated materials . These studies indicate that alkali-activated concrete exhibits a higher carbonation rate than Portland cement concrete at the same compressive strength and that the carbonation mechanism for the two materials is different . Moreover, some studies have shown that the higher carbonation rate of alkali-activated materials results in the degradation of its mechanical properties, although there is some controversy about this [15–17]. However, other studies have found results that are contrary to the reports of carbonation problems in alkali-activated mortars and concretes, as mentioned above . These studies have experimentally demonstrated that alkali-activated concretes removed from accelerated carbonation test after extended periods do not display carbonation issues . This indicates that the outcome of the accelerated test is not an accurate predictor of in-service performance of alkali-activated materials and the carbonation resistance of alkali-activated materials in natural carbonation is equivalent to that of Portland cement concrete . The researchers pointed out that the change in pore solution equilibria causes the formation of sodium bicarbonates during accelerated carbonation, compared to the hydrous sodium carbonates formed during natural carbonation . This shifts the carbonation mechanism to favor more rapid reaction progress, and results in a higher apparent degree of acceleration (compared to natural conditions) than in Portland cements . Shi et al. reported that the natural carbonation rates for alkali silicate-activated slag concretes are lower than 1 mm/year, while laboratory studies have shown that carbonation depths in alkaliactivated concretes range from 13 mm to 25 mm after 240 h of exposure to 7% CO2 . Xu et al. studied aged concretes (up to 35 years) based on slag activated with carbonates or carbonate/ hydroxide blends, and observed an acceptable carbonation rate for all samples (0.03–0.5 mm/year), verifying the high stability of these AAS binders [22,23]. Bernal et al. believe that accelerated carbonation testing is unduly aggressive towards alkali-activated concretes, and test results must be cautiously interpreted [14,15,19,20]. Therefore, there is a need for reasonable evaluation of carbonation resistance for predicting the service life of alkaliactivated concrete under carbonating conditions. The main factors affecting carbon resistance in alkali-activated materials need to be analyzed by studying the carbonation process and the change in the product before and after carbonation. This study focuses on evaluating the effect of accelerated carbonation on alkali-activated concrete produced mainly by MSWI bottom ash, and the improvement in carbonation resistance due to anti-diffusion behavior in carbonation zone. The anti-diffusion behavior refers to the automatic alkaline recovery in the complete carbonation zone of alkali-activated concrete when removed from the accelerated carbonation chamber. By macroscopically studying the anti-diffusion behavior in carbonation zone and by microstructural analysis (Fourier Transform Infrared Spectroscopy (FT-IR) and X-ray diffraction (XRD)) of the products before and after carbonation, we can elaborate on the mechanism of anti-diffusion behavior occurring in alkali-activated materials and perform a reasonable evaluation of their carbonation resistance. 2. Experimental materials and methods 2.1. Materials (1) MSWI bottom ash The MSWI bottom ash used in this experiment was supplied by GCL-Poly (Xuzhou, China) Renewable Energy Power Generation Co., Ltd. The municipal solid waste was first separated via magnetic separation and then calcined at 800 °C. The bottom ash was washed at the recycle station and its chemical composition is shown in Table 1. In this experiment, the bottom ash was ground into a fine powder with surface area and specific gravity of 400 m2/kg and 2.47, respectively. The particles were mainly spherical with a median particle size of 47 mm and 28% was retained on a 45 lm sieve. (2) Granulated blast furnace slag (GBFS) and slaked lime The GBFS used in the experiment is S95 granulated blast furnace slag powder, which conforms to the Chinese standard GB/T 18046-2008 . The surface area and specific gravity of GBFS are 416 m2/kg and 2.89, respectively, and 35% was retained on a 45 lm sieve. Its chemical composition is also shown in Table 1. Industrial grade slaked lime was used. Its purity is 95%, specific gravity is 2.24, and surface area is 400 m2/kg. (3) Activator Sodium hydroxide (NaOH) and sodium silicate solution (Na2SiO3) with 9.65% Na2O, 25.22% SiO2, and 65.13% H2O were used as the activators. The sodium silicate solution has a modulus (molar ratio of SiO2 and Na2O in liquid sodium silicate) of 2.6. A 4.8 M NaOH solution (NaOH dosage and water content are given in Table 2) was prepared and this solution was used to obtain each set of concrete samples. The NaOH solution was allowed to stand for 24 h before use. (4) Others The Portland cement used for the experiment was the PO 42.5 type, which conforms to the Chinese standard GB 175-2007 . Commercially available limestone with a maximum size of 30 mm and a specific gravity of 2.67 in saturated surface Table 1 Chemical composition of raw material (%). Raw material SiO2 CaO Al2O3 Fe2O3 MgO K2O Na2O Zn, Pb, Ba, Sr, Cu Loss GBFS Cement MSWI bottom ash 31.35 26.55 43.82 34.65 62.9 24.44 18.65 7.77 14.18 0.57 3.62 6.18 9.31 2.68 3.26 – 1.5 2.52 – 0.31 2.24 – – 0.61 0.7 3.2 1.62 Table 2 Mix proportion of the concrete kg/m3. C-1 C-2 C-3 C-4 cement MSWI bottom ash GBFS Ca(OH)2 water sodium silicate solution NaOH sand gravel compressive strength (MPa 28 d) 330 0 0 0 0 165 165 165 0 132 132 132 0 33 33 33 165 92.4 133.3 0 0 88 0 254 0 17.6 31.7 0 651 651 651 651 1056 1056 1056 1056 45.2 49.6 46.7 18.5 Note: For example (C-2), Dissolve 17.6 kg of NaOH into 92.4 kg of water, which the molar solubility of NaOH solution is 4.8 M. G. Huang et al. / Construction and Building Materials 186 (2018) 1072–1081 dry conditions was used as the coarse aggregate. For the fine aggregate, river sand with a specific gravity of 2.56 and a fineness modulus of 2.8 was used. Tap water was used for the test. 2.2. Methods 2.2.1. Mix proportion of concrete Table 2 shows the mix proportion of the CC and BA. The groups from C-1 to C-2 represent CC and BA, used for comparing the carbonation resistance of the two kinds of concrete. The groups from C-2 to C-4 represent BA mixed with different kinds of alkali-activators, which were used to determine the main factor causing the anti-diffusion behavior in carbonation zone. 2.2.2. Casting and curing of concrete samples The CC and BA samples were mixed in a temperature controlled room at 20 ± 2 °C. Before preparing BA, the bottom ash was mixed with NaOH solution to form a slurry. This slurry was then aged for 4 h as a defoaming pre-treatment, which can eliminate the effect of foaming and expansion on the performance of BA when it comes into contact with alkali. The casting procedure was started by mixing the slurry, GBFS and slaked lime for 1 min, and then sand and sodium silicate solution were added and stirred for 30 s. Finally, coarse aggregates were added to the mixture and stirred for 1 min. The casting procedure of CC was started by mixing the Portland cement, sand and water for 1 min, then the coarse aggregates were added and stirred for 1.5 min. The casted sample dimensions were 100 mm 100 mm 400 mm and the liquid-solid ratio of BA and CC was 0.5, where the liquid was the total mass of NaOH solution and sodium silicate solution. All samples were placed in a curing room after casting and demolded after 24 h. The curing was performed in a 20 ± 2 °C curing room (>95% room humidity) until testing. 2.2.3. Accelerated carbonation test An accelerated carbonation test was performed according to the ‘‘standard for test methods of long-term performance and durability of ordinary concrete” (GB/ T50082-2009) . The environmental parameters of the carbonation chamber were: CO2 mass fraction of 20%, temperature of 20 ± 2 °C, and relative humidity of 70 ± 5%. Four days before the start of the test, the samples were dried at 60 °C for 48 h and placed in a natural environment for 2 d. All surfaces were sealed with molten paraffin, except for one. Treated samples were placed into a carbonation box and carbonation was accelerated to specific durations (14 d, 28 d, and 60 d). To measure carbonation depth, the cured samples were removed from the box and cut into samples with a thickness of about 60 mm each. The remaining powder on the section was removed, and then 1% phenolphthalein indicator was sprayed to determine the depth of the fully carbonized zone based on the indicator color. After 30 s, according to the original calibration points, the full carbonation depth of every measuring point was measured using vernier calipers. The remaining part of the sample was sealed with molten paraffin, and then placed into the carbonation chamber to continue carbonation. 2.2.4. Sample collection of carbonized powder and determination of pH value The samples from C-1 to C-4 were taken starting from the outer to inner regions in the cross section of the concrete and the sample range was the surface to test center of concrete (0–30 mm). The samples were taken every 5 mm and the stones were removed leaving behind only the mortar. Then, the mortar was ground into powder and poured into a 25-ml beaker. A small amount of distilled water was added (water to solid ratio = 1:3) and mixed to form a sticky paste. The pH value was measured using an acidometer and the value was read after 5 min. 2.2.5. Microstructural analysis of BA samples (1) XRD analysis The paste samples of CC and BA (C-1 and C-2), which were cured for 28 days, were analyzed using an X-ray diffractometer (XRD, Bruker, D8 Advance, Germany) with Cu K a 1 radiation and 2 h scanning range of 5°–70°. (2) FT-IR analysis The paste samples of CC and BA (C-1 and C-2), which were cured for 28 days, were analyzed by Fourier Transform Infrared Spectroscopy (FT-IR, Bruker, VERTEX 80v, Germany). The wavenumber range was from 400 to 4000 cm1. 3. Results and discussion properties and the compressive strength is up to 45.2 MPa, 49.6 MPa, and 46.7 MPa respectively, which they have the same level of compressive strength. However, the compressive strength of sample C-4 is only 18.5 MPa, which is obviously lower than other samples and this will seriously affected the resistance of carbonation. The carbonation depth curves of CC and BA (error bars represent one standard deviation) are shown in Fig. 1. The 14 d, 28 d, and 60 d carbonation depths of CC specimen (C-1) were 6 mm, 12 mm, and 14 mm, respectively, while the corresponding carbonation depths of BA specimen (C-2) were 16 mm, 20 mm, and 26 mm. The results indicate that BA is far less resistant against carbonation than CC at high carbon concentrations. When only NaOH was used as an activator, the carbonation resistance of BA specimen (C-3) showed a slight improvement and the 14 d, 28 d, and 60 d carbonation depths reduced to 14 mm, 18 mm, and 22 mm, respectively. When only sodium silicate solution was used as an activator, the carbonation resistance of BA specimen (C-4) showed a significant decrease and the 14 d, 28 d, and 60 d carbonation depths increased to 17 mm, 21 mm, and 27 mm, respectively. 3.2. The anti-diffusion behavior in carbonation zone of BA After the carbonation of CC specimen (C-1) (Fig. 2, a-1, a-2, a-3), the uncarbonized area showed red color and the fully carbonized area showed no color after spraying phenolphthalein indicator, and the uncarbonized and fully carbonized areas were clearly separated by a red line. As time went on, the dividing line (red line) was still obvious and remained at the same position. Moreover, the colors in uncarbonized area and fully carbonized area also remained constant without any further changes. The carbonation behavior of BA was completely different from that of CC. An anti-diffusion behavior was observed in the complete carbonation zone of BA specimen (C-2) (Fig. 2). It can be seen from the 14 d (b-1), 28 d (c-1) and 60 d (d-1)) carbonation test results of BA specimen (C-2) that the uncarbonized area showed a reddish color (below the red line), while the fully carbonized area showed no color (above the red line) after the phenolphthalein indicator was sprayed. This same behavior was observed for CC too. However, the difference was that, about ten minutes later, the color in fully carbonized area rapidly changed from colorless Carbonation depth (mm) 1074 C-1 C-2 C-3 C-4 20 10 0 0 3.1. Mechanical properties and carbonation depth of CC and BA Table 2 shows the compressive strength of the CC and BA after 28 days of curing. Samples C-1 to C-3 all have good mechanical 15 30 45 60 Time (day) Fig. 1. Carbonation depth curves of CC and BA; Error bars represent one standard deviation. G. Huang et al. / Construction and Building Materials 186 (2018) 1072–1081 1075 Fig. 2. Anti-diffusion behavior of BA samples; (a-1, a-2, a-3,) 14 d, 28 d, and 60 d carbonation depth of sample (C-1); (b-1, c-1,d-1) 14 d, 28 d, and 60 d carbonation depth of sample (C-2); (b-2, c-2, d-2) in the anti-diffusion process of sample (C-2); (b-3, c-3, d-3) complete anti-diffusion behavior of sample (C-2). to light red (above the red line) and the color in the uncarbonized area (below the red line) changed from red to crimson (Fig. 2, b-2, c-2 and d-2). Moreover, the dividing line between the uncarbonized area and fully carbonized area began to push back quickly (red line). As the anti-diffusion behavior continued, the light red color in fully carbonized area constantly deepened and the crimson color in uncarbonized area continued to expand. This behavior lasted for about three minutes. Ultimately, the anti-diffusion behavior was completed after about fifteen minutes. It can be seen from Fig. 2 (b-3, c-3, and d-3) that the fully carbonized area completely changed from light red to crimson and the color was consistent with the uncarbonized area. Moreover, the crimson color 1076 G. Huang et al. / Construction and Building Materials 186 (2018) 1072–1081 completely covered the uncarbonized area and the dividing line between carbonized and uncarbonized areas also disappeared completely. 3.3. Influence of alkali-activator on anti-diffusion in carbonation zone The anti-diffusion behavior of carbonized concrete is shown in Table 3. It can be seen from Table 3 that the CC specimen (C-1) and the BA specimen (C-3), which used only NaOH as the activator, did not display the anti-diffusion behavior. However, the BA specimen (C-2) which was activated by sodium silicate solution and NaOH as well as the BA specimen (C-4) which was activated only by sodium silicate solution both showed the anti-diffusion behavior in carbonation zone. By comparing the compositions of BA specimens (C-2, C-3, and C-4), it can be inferred that the occurrence of the anti-diffusion behavior in BA is closely related to the sodium silicate solution. Moreover, by comparing the start time and duration of anti-diffusion behavior in BA specimens (C-2 and C-4), it can be seen that with the increase in content of sodium silicate solution, the start time was significantly advanced and the anti-diffusion behavior was stronger. This suggests that the antidiffusion reaction rate and the degree of anti-diffusion are obviously enhanced with the increase in content of sodium silicate solution. Fig. 3. Distribution of pH value in CC and BA after 60d carbonation; Error bars represent one standard deviation. 3.4. Determination of pH value in alkali-activated concrete The pH value distribution of carbonized concrete (error bars represent one standard deviation) is shown in Fig. 3. It can be seen that the pH value of CC specimen (C-1) remained at 8.5 in the range of 0–10 mm and this region was a fully carbonized region with a relatively sufficient carbonation reaction. Then, the pH value increased from 8.5 to 13.0 in the depth from 10 to 15 mm and this region was a pH change zone of the carbonation reaction. Finally, the pH value remained at 13 from 15 mm depth to inside and this region was a stable uncarbonized zone. Similar to the pH distribution trend for CC specimen (C-1), the pH value of BA specimen (C-3) without anti-diffusion characteristics remained at 8.5 in the range of 0–20 mm. The pH value increased from 8.5 to 13.6 in the depth from 20 to 25 mm and then remained at 13.6 in the depth from 25 mm to inside. The pH value distribution of BA specimen (C-2) was completely different from that of the CC specimen (C-1). The pH value of BA specimen (C-2) remained at 11.3 in the fully carbonized region from 0 to 25 mm, even after carbonation. This is significantly higher than the pH value in the fully carbonized area of CC specimen (C-1), suggesting that the carbonation reverse diffusion can significantly improve the pH value in fully carbonized zone. Then, the pH value increased from 11.3 to 13.9 in the depth from 25 to 30 mm and remained at 13.9 in the depth from 30 mm to the interior of the uncarbonized zone, which is higher compared to the CC specimen (C-1) in the same zone. The pH distribution trend of BA specimen (C-4) is similar to that of BA specimen (C-2) in the fully carbonized region and pH change zone. Only in the uncarbonized zone, the pH value of BA specimen (C-4) from 30 mm to the interior remained at 12.4, which is much lower than that of BA specimen (C-2). This is because NaOH was not used as the activator. The above results indicate that even in the fully carbonized region of BA specimens (C-2 and C-4), the pH can still be maintained at a higher value. This is a very exciting result and it indicates that even if the BA specimen has been completely carbonized, the passivation film of internal steel is still stable, which greatly reduces the corrosion rate of the steel bar and improves its durability. 3.5. Microanalysis 3.5.1. XRD analysis Fig. 4 shows the XRD patterns of the CC specimen (C-1) after 60 days of carbonation. It can be seen from Fig. 4(a) that the uncarbonized zone of CC specimen (C-1) had a very complex mineral composition and its components included crystals such as quartz (SiO2; PDF #01-070-2536), gehlenite (2CaO∙Al2O3∙SiO2; PDF #00035-0755), calcite (CaCO3; PDF #00-047-1743), tobermorite (Ca5(Si6O16)(OH2); PDF #01-089-6458), hillebrandite (Ca2(SiO3) (OH2); PDF #00-042-0538), dicalcium silicate hydrate (Ca2SiO43H2O; PDF #00-029-0374), and portlandite (Ca(OH2); PDF #00-044-1481) as well as partially amorphous silicates and aluminosilicates with amorphous chemical compositions . The mineral composition of the complete carbonation zone in CC specimen (C-1) changed significantly compared to the uncalcined zone. Fig. 4(b) shows that the characteristic peak of portlandite decreased dramatically after carbonation and the peak of calcite increased markedly. This result indicates that carbonation promoted the dissolution of portlandite (Ca(OH2), releasing a large amount of OH involved in the carbonation reaction, and generated a large amount of calcite. Moreover, the characteristic peaks of gehlenite, tobermorite, hillebrandite and dicalcium silicate Table 3 Anti-diffusion behavior of carbonation. C-1 C-2 C-3 C-4 Anti-diffusion behavior start time end time NaOH Ca(OH)2 water sodium silicate solution no yes no yes – 10 min – 5 min – 12 min – 6 min 0 17.6 31.7 0 0 33 33 33 165 92.4 133.3 0 0 88 0 165 G. Huang et al. / Construction and Building Materials 186 (2018) 1072–1081 (a) XRD pattern of uncarbonized zone (a) 1077 XRD pattern of uncarbonized zone (b) XRD pattern of complete carbonation zone (b) XRD pattern of complete carbonation zone Fig. 5. XRD pattern of BA specimen (C-2). Fig. 4. XRD pattern of CC specimen (C-1). hydrate all decreased to different extents, which indicated that these minerals are all involved in carbonation in varying degrees and play a certain role in resisting carbonation . Fig. 5 shows the XRD patterns of BA specimen (C-2) after 60 days of carbonation. It can be seen from Fig. 5(a) that the uncarbonized zone in BA specimen (C-2) had a significant peak of magadiite that was formed by self-coagulated liquid sodium silicate which did not participate in the polymerization reaction . At the same time, the presence of portlandite was due to the addition of Ca(OH)2 during preparation of the BA [29,30]. The characteristic peaks of quartz, calcite, gehlenite and other crystals also appeared in the XRD pattern of CC specimen (C-1). 1078 G. Huang et al. / Construction and Building Materials 186 (2018) 1072–1081 The complete carbonation zone in BA specimen (C-2) showed a certain degree of change compared to the uncalcined zone. From Fig. 5(b), it can be seen that the characteristic peak of portlandite reduced significantly due to carbonation. More importantly, the characteristic peak of magadiite decreased dramatically compared to the uncalcined zone. This indicates that magadiite turns back to liquid sodium silicate when the pH value in pore fluid and the concentration of CO2 decrease to a certain level, and then it releases a large amount of OH to increase the pH value in pore fluid [19,20]. 3.5.2. FT-IR analysis Fig. 6 shows the FT-IR spectra of the CC specimen (C-1). It can be seen from Fig. 6(a) that the uncarbonized zone in CC specimen (C1) exhibited infrared vibration modes at 954 cm1 corresponding to the stretching vibration of Si-O-Si bonds. This specific frequency is characteristic of silicon tetrahedra (SiO4) in the chain structure of calcium silicate hydrate (C-S-H). The shoulders at 1424 cm1 are related to O-C-O bonds and this specific frequency is characteristic of calcite. The mode at 1643 cm1 is assigned to the asymmetric stretching vibration of H-OH bonds, and this specific frequency is characteristic of water. The band at 3460 cm1 is associated with bending vibrations of –OH bonds, and is related to bound Ca (OH)2 [27,28]. These results are consistent with XRD results of the uncarbonized zone in CC specimen (C-1). Fig. 6(b) shows the FT-IR spectrum of complete carbonation zone in CC specimen (C-1). The characteristic peak of –OH bond at 3445 cm1 significantly decreased while the characteristic peaks of O-C-O bond at 875 cm1 and 1413 cm1 significantly increased, indicating that the carbonation consumed a large amount of Ca (OH)2 and caused the formation of a large amount of calcium carbonate, which is consistent with the XRD results. Fig. 7 shows the FT-IR spectra of the BA specimen (C-2). It can be seen from Fig. 7(a) that the uncarbonized zone in BA specimen (C-2) exhibited infrared vibration modes at 1033 cm1 associated with bending vibrations of O-Si-O bonds, related to bound magadiite. A characteristic –OH peak appeared at 3459 cm1 due to –OH mainly from the sodium silicate solution and the addition of a small amount of Ca(OH)2. The characteristic peaks of quartz, calcite, and other crystals also appeared in the spectrum of BA specimen (C-2), which is consistent with the XRD results [14,15,20]. Fig. 7(b) shows the FT-IR spectrum of the complete carbonation zone in BA specimen (C-2). It can be seen that the characteristic peaks of O-C-O bonds at 874 cm1 and 1453 cm1 increased significantly, indicating that Ca(OH)2 was consumed and a large amount of calcite was produced. More importantly, the characteristic peak of –OH at 3459 cm1 showed no obvious reduction after carbonation, while the characteristic peaks of O-Si-O bonds at 1033 cm1, related to bound magadiite, decreased significantly. This result indicates that magadiite turned from solid back to liquid and released a large amount of OH. Thus, the characteristic peaks of magadiite significantly decreased and there was no change in the characteristic peak of –OH. This is consistent with the XRD results [19,20,31]. (a) FT-IR pattern of uncarbonized zone 3.6. Discussion 3.6.1. Reaction and carbonation mechanism of CC and BA During the hydration process of CC (Fig. 8(a)), a large amount of Ca(OH)2 was formed by hydration reaction of 3CaOSiO2 as well as 2CaOSiO2 [27,28]. The apparent characteristic peaks of C-S-H and portlandite in Fig. 4(a) and the obvious characteristic peaks of Si-OSi and –OH bonds in Fig. 6(a) confirmed the aforementioned mechanism. As the carbonation reaction progressed (Fig. 8(b)), the OH in the pore solution of CC was continuously consumed. However, the crystalline Ca(OH)2 continued to dissolve, supplementing the (b) FT-IR pattern of complete carbonation zone Fig. 6. FT-IR pattern of CC specimen (C-1). consumption of OH. In addition, the Ca(OH)2 crystals formed by the hydration of Portland cement accounted for 20% to 30% of the total mass in binding material, and the pH value of CC remained at a high level until the solid-state Ca(OH)2 was depleted. 1079 G. Huang et al. / Construction and Building Materials 186 (2018) 1072–1081 (a) FT-IR pattern of uncarbonized zone to BA since it contains many substances that can participate in carbonation [7,10,28], e.g. Ca(OH)2. The polymerization process of the BA specimen (C-2) is shown in Fig. 8(c) (reactions (1)–(3)). Due to the strongly alkaline environment provided by the activator, active calcium as well as active aluminum in the slag (GBFS, MSWI bottom ash) dissolved rapidly and polymerized with the active silicon provided by the liquid sodium silicate, thus producing C-S-H gel and C-A-S-H gel [7,32,33]. Furthermore, insufficient reaction led to excessive crystallization of sodium silicate in the strong alkaline environment, which then precipitated into the BA. The apparent characteristic peaks of C-S-H, C-A-S-H, and magadiite in Fig. 5(a) and the obvious characteristic peaks of the Si-O-Si and O-Si-O bonds in Fig. 7(a) confirmed the aforementioned mechanism. The carbonation process of the BA specimen (C-2) is shown in Fig. 8(d) (reaction (4)). During the carbonation reaction, large amounts of OH in the pore liquid were consumed, resulting in a weakly alkaline environment . Such an environment promoted the dissolution of portlandite as well as magadiite and supplemented the consumed OH in the pore liquid. However, during the preparation of BA, the majority of portlandite and sodium silicate were consumed in the polymerization reaction (reactions (3) and (5)); therefore, the remaining content decreased . Moreover, the anti-diffusion effect produced by magadiite was applicable only at low concentration of CO2, ultimately leading to poorer resistance against carbonation in the accelerated carbonation test. The significantly decreased characteristic peaks for portlandite and magadiite in Figs. 5(b) and 7(b) confirmed the described mechanism. In summary, during the polymerization process of the BA specimen (C-2), the portlandite mineral could not be generated to resist carbonation. Also, the polymerization process consumed large amounts of Ca(OH)2 and liquid sodium silicate which were added during the preparation of BA. Therefore, the carbonation resistance of BA was much weaker than that of CC [36,37,38]. Ca2þ þ SiO3 2 þ nH2 O ! C S H # ðstrong alkali environmentÞ 3þ Ca2þ þ Al ð1Þ þ SiO3 2 þ nH2 O ! C A S H # ðstrong alkali environmentÞ Naþ þ nSiO3 2 þ nH2 O ! Na2 O nSiO2 nH2 O # ðSurplusÞ ð2Þ ð3Þ Na2 O nSiO2 nH2 O Naþ þ nOH þ nSiO3 2 Fig. 7. FT-IR pattern of BA specimen (C-2). The significant decrease in the characteristic peak of portlandite in Fig. 4(b) and the obvious decrease in the characteristic peak of –OH bond in Fig. 6(b) confirmed the aforementioned mechanism. In summary, CC has good resistance against carbonation compared ðlowCO2 concentrationandweak - alkali environmentÞ ð4Þ CaðOHÞ2 ðsÞ þ SiO3 2 þ nH2 O ! C S H ðIn the solid phaseÞ ð5Þ 3.6.2. The mechanism of anti-diffusion in carbonation zone The carbonation process of BA involves the penetration of CO2 and blocking by OH (Fig. 9). Specifically, CO2 dissolves into the pore fluid and continues to push inward, while OH released from the dissolution of magadiite as well as a small amount of Ca(OH)2 blocks the entry of CO2. The advancing speed of the carbonized surface (pH change zone in Fig. 9) is determined by the concentrations and diffusion rates of CO2 and OH on both sides. In the accelerated carbonation test, the concentration of CO2 is set at 20%±3% and the high concentration of CO2 gas rapidly intrudes into the surface of BA. The diffusion rate of CO2 (VCO2) is very high and the consumed CO2 can be quickly replenished by the high concentration of CO2. However, the amount of substances in BA that can resist carbonation is very limited. Furthermore, the conversion of magadiite into sodium silicate solution to release OH is strongly inhibited in high concentrations of CO2. The small amount of residual Ca(OH)2 in BA is unable to completely consume 1080 G. Huang et al. / Construction and Building Materials 186 (2018) 1072–1081 Fig. 8. The process of reaction and carbonation in CC and BA; (a) The reaction process of CC specimen (C-1); (b) The carbonation process of CC specimen (C-1); (c) The reaction process of BA specimen (C-2); (d) The carbonation process of BA specimen (C-2). Y pH PH change zone CO2 13.9 CO2 - OH After anti-diffusion Concrete surface Complete carbonation zone Uncarbonized zone 8.5 Before anti-diffusion 0 11.3 25 30 X Depth (mm) Fig. 9. The mechanism of anti-diffusion in BA specimen (C-2). progress of the carbonized surface (pH change zone in Fig. 9) in BA is very rapid and the pH value in complete carbonation zone is very low, as shown by the blue line in Fig. 9. The concentration of CO2 dropped from 20% to 0.03% when BA was removed from the carbonation box to the natural environment. The decrease in CO2 concentration led to the reduction in diffusion rate of CO2. Also, the magadiite rapidly decomposed and converted into liquid sodium silicate to release a large amount of OH. As a result, the diffusion rate of CO2 was much lower than the compensation rate of OH (VOH- VCO2). This conclusion was confirmed by the decrease in the characteristic peak of magadiite in Figs. 5(b) and 7(b). Due to the anti-diffusion effect, the pH value in the pore fluid of BA increased rapidly and the carbonized surface (pH change zone in Fig. 9) pushed back quickly. Ultimately, the completely carbonized region turned from colorless to red in Fig. 2 and the pH value in completely carbonized region increased from 8.5 to 11.3 (red line in Fig. 9) after anti-diffusion effect. Therefore, the anti-diffusion behavior can significantly enhance the carbonation resistance of BA and also alleviate the carbonization process under natural conditions. As a result, the rebar corrosion resistance and durability of BA reinforced concrete can both be improved. 4. Conclusions the CO2 in the pores in time, which causes the diffusion rate of CO2 to be much faster than the compensation rate of OH (VCO2 V OH). This conclusion was confirmed by the decrease in the characteristic peak of magadiite in Figs. 5(a) and 7(a). Therefore, the inward Accelerated carbonation testing of CC and BA binders shows that the carbonation resistance of BA is much lower than that of CC, due to the absence of an anti-carbonization component such G. Huang et al. / Construction and Building Materials 186 (2018) 1072–1081 as Ca(OH)2. BA displays a unique anti-diffusion behavior in the complete carbonation zone. However, the anti-diffusion behavior effect of BA is clearly reflected only at low concentrations of CO2 (0.03%) and this effect is inhibited under the experimental conditions of the accelerated carbonation test (CO2 concentration of 20%). The anti-diffusion behavior causes a significant rise in pH value of pore fluid in the complete carbonation zone of BA, which markedly improves the anti-carbonation properties of BA. The corresponding pH value for BA is significantly higher than that for CC. The rapid decomposition of magadiite releasing a large amount of OH is the main reason for the anti-diffusion behavior in BA specimens. With the increase in content of liquid sodium silicate, the anti-diffusion behavior of BA in the carbonation zone is more evident. The findings of this work can help in the further development of corrosion resistant building materials. Conflict of interest The authors declare that they have no conflict of interest. Acknowledgement The authors are grateful to the financial support from the Central Universities (2018XKQYMS02). The research works belongs to one part of the projects which are financially supported by FRFCU. References  John L. Provis, Angel Palomo, Caijun Shi, Advances in understanding alkaliactivated materials, Cem. Concr. Res. 78 (2015) 110–125.  X. Gao, Q.L. Yu, H.J.H. Brouwers, Reaction kinetics, gel character and strength of ambient temperature cured alkali activated slag–fly ash blends, Constr. Build. Mater. 80 (2015) 105–115.  N.K. Lee, H.K. Lee, Reactivity and reaction products of alkali-activated, fly ash/ slag paste, Constr. Build. Mater. 81 (2015) 303–312.  Jian Zhang, Caijun Shi, Zu.hua. Zhang, Ou. Zhihua, Durability of alkali-activated materials in aggressive environments: a review on recent studies, Constr. Build. Mater. 152 (2017) 598–613.  P. Tang, M.V.A. Florea, P. Spiesz, H.J.H. Brouwers, Characteristics and application potential of municipal solid waste incineration (MSWI) bottom ashes from two waste-to-energy plants, Constr. Build. Mater. 83 (2015) 77–94.  Nabottom Ashjyoti Saikia, Gilles Mertens, Koenraad Van Bottom Ashlen, Jan Elsen, Tom Van Gerven, Carlo Vandecasteele, Pre-treatment of municipal solid waste incineration (MSWI) bottom ash for utilisation in cement mortar, Constr. Build. Mater. 96 (2015) 76–85.  Chao Li, Henghu Sun, Longtu Li, A review: the comparison between alkaliactivated slag (Si+Ca) and metakaolin (Si+Al) cements, Cem. Concr. Res. 40 (2010) 1341–1349.  Hu. Mingyu, Zh.u. Xiaomin, Long Fumei, Alkali-activated fly ash-based geopolymers with zeolite or bentonite as additives, Cem. Concr. Compos. 31 (2009) 762–768.  F. Puertas, B. González-Fonteboa, I. González-Taboada, M.M. Alonso, M. TorresCarrasco, G. Rojo, F. Martínez-Abella, Alkali-activated slag concrete: fresh and hardened phenomenon, Cem. Concr. Compos. 85 (2018) 22–31.  C. Shi, A. Fernández-Jiménez, A. Palomo, New cements for the 21st century: the pursuit of an alternative to Portland cement, Cem. Concr. Res. 41 (2011) 750– 763.  Raphaëlle Pouhet, Martin Cyr, Carbonation in the pore solution of metakaolinbased geopolymer, Cem. Concr. Res. 88 (2016) 227–235.  F. Puertas, M. Palacios, T. Vázquez, Carbonation process of alkali-activated slag mortars, J. Mater. Sci. 41 (2006) 3071–3082.  M. Palacios, F. Puertas, Effect of carbonation on alkali-activated slag paste, J. Am. Ceram. Soc. 89 (2006) 3211–3221. 1081  Susan A. Bernal, John L. Provis, David G. Brice, Adam Kilcullen, Peter Duxson, Jannie S.J. van Deventer, Accelerated carbonation testing of alkali-activated binders significantly underestimates service life: The role of pore solution chemistry, Cem. Concr. Res. 42 (2012) 1317–1326.  Susan A. Bernal, Ruby Mejía de Gutierrez, John L. Provis, Volker Rose, Effect of silicate modulus and metakaolin incorporation on the carbonation of alkali silicate-activated slags, Cem. Concr. Res. 40 (2010) 898–907.  S.A. Bernal, E. Rodríguez, R. Mejía de Gutiérrez, V. Rose, F. Puertas, S. Delvasto, Carbonation phenomenon of mortar produced by alkali-activation of a granulated blast furnace slag, Proceedings of 23rd International Conference on Solid Waste Technology and Management, Widener University, Philadelphia, PA, 2008.  B. Johannesson, P. Utgenannt, Microstructural changes caused by carbonation of cement mortar, Cem. Concr. Res. 31 (2001) 925–931.  C. Shi, P.V. Krivenko, D.M. Roy, Alkali-Activated Cements and Concretes, Taylor & Francis, Abingdon, UK, 2006.  Susan A. Bernal, John L. Provis, Brant Walkley, Rackel San Nicolas, John D. Gehman, David G. Brice, Adam R. Kilcullen, Peter Duxson, Jannie S.J. van Deventer, Gel nanostructure in alkali-activated binders based on slag and fly ash, and effects of accelerated carbonation, Cem. Concr. Res. 53 (2013) 127– 144.  S.A. Bernal, J.L. Provis, R. Mejia de Gutierrez, J.S.J. van Deventer, Accelerated carbonation testing of alkali-activated slag/metakaolin blended concretes: effect of exposure conditions, Mater. Struct. 48 (2015) 1–17.  M. Criado, A. Palomo, A. Fernández-Jiménez, Alkali activation of fly ashes. Part 1: effect of curing conditions on the carbonation of the reaction products, Fuel 84 (2005) 2048–2054.  Susan A. Bernal, Effect of the activator dose on the compressive strength and accelerated carbonation resistance of alkali silicate-activated slag/metakaolin blended materials, Constr. Build. Mater. 98 (2015) 217–226.  Hailong Ye, Aleksandra Radlin’ska, Carbonation-induced volume change in alkali-activated slag, Constr. Build. Mater. 144 (2017) 635–644.  GB/T18046-2008, Ground granulated blast furnace slag used for cement and concrete, 2008.  GB175-2007, Common Portland Cement, 2007.  GB/T50082-2009, Standard for test methods of long-term performance and durability of ordinary concrete, 2009.  Merah Ahmed, Krobba Benharzallah, Effect of the carbonatation and the type of cement (CEM I, CEM II) on the ductility and the compressive strength of concrete, Constr. Build. Mater. 148 (2017) 874–886.  Yu Min, Bao Hao, Ye Jianqiao, Chi Yin, The effect of random porosity field on supercritical carbonation of cement-based materials, Constr. Build. Mater. 146 (2017) 144–155.  Ehsan Ul Haq, Sanosh Kunjalukkal Padmanabhan, Antonio Licciulli, In-situ carbonation of alkali activated fly ash geopolymer, Constr. Build. Mater. 66 (2014) 781–786.  Jingliang Dong, Lijiu Wang, Tingting Zhang, Study on the strength development, hydration process and carbonation process of NaOH-activated Pisha Sandstone, Constr. Build. Mater. 66 (2014) 154–162.  Kirubajiny Pasupathy, Marita Berndt, Arnaud Castel, Jay Sanjayan, Rajeev Pathmanathan, Carbonation of a blended slag-fly ash geopolymer concrete in field conditions after 8 years, Constr. Build. Mater. 125 (2016) 661–669.  Tzen-Chin Lee, Wei-Jer Wang, Ping-Yu Shih, Kae-Long Lin, Enhancement in early strengths of slag-cement mortars by adjusting basicity of the slag prepared from fly-ash of MSWI, Cem. Concr. Res. 39 (2009) 651–658.  K. Dombrowski, A. Buchwald, M. Weil, The influence of calcium content on the structure and thermal performance of fly ash bottom ashsed geopolymers, J. Mater. Sci. 42 (2007) 3033–3043.  Raphael Bucher, Paco Diederich, Gilles Escadeillas, Martin Cyr, Service life of metakaolin-based concrete exposed to carbonation Comparison with blended cement containing fly ash, blast furnace slag and limestone filler, Cem. Concr. Res. 99 (2017) 18–29.  M.S.H. Khan, A. Noushini, A. Castel, Carbonation of a low-calcium fly ash geopolymer concrete, Mag. Concr. Res. 69 (2017) 24–34.  Maria Chiara Bignozzi, Stefania Manzi, Maria Elia Natali, William D.A. Rickard, Arie van Riessen, Room temperature alkali activation of fly ash: the effect of Na2O/SiO2 ratio, Constr. Build. Mater. 69 (2014) 262–270.  A. Younsi, Ph. Turcry, A. Aït-Mokhtar, S. Staquet, Accelerated carbonation of concrete with high content of mineral additions: effect of interactions between hydration and drying, Cem. Concr. Res. 43 (2013) 25–33.  Kiacher Behfarnia, Majid Rostami, An assessment on parameters affecting the carbonation of alkali-activated slag concrete, J. Cleaner Prod. 157 (2017) 1–9.