Effect of MgO, Mg-Al-NO3 LDH and calcined LDH-CO3 on chloride resistance of alkali activated fly ash and slag blends

The mechanism of layered double hydroxides (LDHs) formation with the addition of magnesium oxide (MgO), the LDHs reconstruction of calcined layered double hydroxides (CLDHs) and the effect of externally added synthetic LDHs in alkali-activated materials (AAMs) system are the important factors influencing the chloride resistance of AAMs. The objective of this research is to investigate the different effects of MgO, nitrate intercalated LDH (Mg-Al-NO3 LDH) and calcined natural LDH-CO3 on the alkali activated fly ash and slag blends (AAFS) in terms of chloride resistance. Mg-Al-NO3LDH was successfully synthesized with the co-precipitation method. CLDH was also obtained by thermal treating natural Mg-AlCO3 LDH. The reaction heat flow and reaction products of the AAFS were tested by isothermal calorimetry and X-ray diffraction (XRD), respectively. The microstructure of the AAFS was characterized by the nitrogen sorption tests (BET) and mercury intrusion porosimetry (MIP). The compressive strength of AAFS was also tested. The non-steady-state migration (NSSM) was applied to determine the chloride resistance of AAFS. The Ordinary Portland Cement (OPC) samples were also studied for comparative purpose. The results of the calorimeter test showed that the MgO, Mg-Al-NO3 LDH, and CLDH delayed the time to reach the reaction peak (TRRP) at 3.9%, 11.2%, and 9.2%, respectively. The MIP results illustrated that mesopores and micropores of AAFS were refined by the three admixtures. The NSSM results indicated that CLDH can significantly improve the chloride resistance of AAFS binder comparing to the samples containing MgO and Mg-Al-NO3 LDH. The different influences of MgO, Mg-Al-NO3 LDH and CLDH were compared in this study. 2020 The Author(s). Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/).


Introduction
Numerous military and civilian concrete structures exposed to the marine environment face the risk of chloride corrosion [1]. Chloride penetration into reinforced concrete can cause local destruction of the passive layer, leading to localized corrosion.
Consequently, the usability and load-bearing capacity decreased, even the failure of the structure could occur. Facing a large amount of chloride ingress problems, approximately 3.4% of the global GDP ($2.5 Trillion USD) is spent each year to prevent, mitigate, and repair the infrastructure damage due to chloride-induced corrosion [2]. As we can see from Fig. 1, remarkable chloride ingress problems are universal worldwide [3]. The corrosion problems of offshore concrete structures lead to an urgent demand for the research about the chloride resistance of the alkali activated materials (AAMs). The effect of the alkaline environment in AAMs on https://doi.org/10.1016/j.conbuildmat.2020.118865 0950-0618/Ó 2020 The Author(s). Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/). passivate steel bars may be very important in improving the durability of the reinforced concrete structure. Meanwhile, the use of by-product in AAMs also meet the demand for sustainable development worldwide. To solve these issues, many methods were developed to prevent chloride penetration. Table 1 illustrates the investigations by diverse researches subjected to chloride resistance.
Recently, Layered double hydroxides (LDHs) are widely investigated to absorb undesired anions in aqueous solution [18]. LDHs are one of the primary reaction products in alkali activated slag. The skeleton of LDHs consists of the divalent metal ion and trivalent metal ion, for example, Mg 2+ and Al 3+ [19]. The interlayer of LDHs has the capability of binding the anions in AAMs, for instance, chloride ions and sulfate ions. The interlayer anion can exchange with other anions and the affinity order of anion is characterized by Costa [20] as: Thus, the amount of Mg-Al-NO -3 -LDH influences the chloride resistance in the silicate matrix, because the NO -3 can exchange the Cl in the chloride environment. Xu et al. [21] prepared the Mg-Al-NO -3 -LDH and Mg-Al-NO -2 -LDH by the co-precipitation method, and both Mg-Al-LDHs exhibited the excellent chloride removal and corrosion inhibition capacities. Qu et al. [22] reported that the finer particle size of the Ca-Al-NO -3 -LDH enhanced the Fig. 1. Cost of corrosion related to GDP in different countries [3]. 2019 [17] chloride resistance of the cement matrix, which is attributed to both the physical barrier effect of the increased tortuosity and chemical binding capacity of the LDHs. Regardless of the references mentioned above, the synthesis pH of LDH, flow rates, temperature, pressure environment partly contribute to the property of the final product [23]. However, few literature reports the externally added synthetic LDHs in AAMs pertaining to chloride resistance [2]. Thus, there are still largely unknown areas in terms of LDHs applied in AAMs matrix to promote the chloride resistance. The influence of the synthetic and externally added LDHs in AAMs needs to be deeply investigated. Magnesium oxide (MgO) is introduced in the AAMs to deal with the chloride ingress problem, because the MgO content in the silicate system influences the formation of LDHs in the AAMs matrices, thus improving chloride resistance [8]. High contents of MgO (higher than 8 wt%) increase the formation of the hydrotalcitelike phase and decrease the C-S-H phase. Consequently, the chemical binding capacity increases by more LDHs content but the physical binding capacity decreases by the less C-S-H gel [24]. With the addition of MgO in AAMs, it is hard to control the content of LDHs formation, because the reaction is affected by many parameters, e.g. the two metal ions molar ratio [25]. Therefore, further investigations are needed for a desirable control of MgO content in AAMs to increase the chloride binding capacity.
The calcined layered double hydroxides (CLDHs) can be calcined from the natural hydrotalcite (Mg-Al-CO 3 LDH) and rebuild the layered double structure in the high alkaline system [26]. Sato et al. [27] investigated the effect of CLDH on the cement concrete structure, the results of the rapid chloride penetration test showed that the CLDH promoted the chloride binding capacity due to the reconstruction of a LDH-like phase in the cementitious matrix (Fig. 2). The utilization of the CLDH in the sodium carbonate activated slag exhibited a higher chloride resistance than the reference samples without CLDH incorporation [28]. CLDH rebuilds the LDH-like phase to absorb the chloride ion and increases the chemical chloride binding capacity. Therefore, CLDH enhances the physical absorption of chloride ions. CLDHs has the potential to act as an effective chloride adsorbent in AAMs matrix. The CLDH content promotes the LDH-like phase formation, thereby affecting the chloride binding capacity of AAMs. As mentioned above, there were studies about applying CLDHs in sodium carbonate activated slag, but very limited researches about CLDHs applied in alkali activated slag and fly ash. Consequently, investigations on chloride adsorption of CLDHs are needed to understand the influence of the CLDHs dosage in alkali activated slag and fly ash.
The MgO and CLDH influence the LDH formation in AAMs, and externally added LDHs can play the role of filler and chemical absorbent in AAMs, thereby affecting the chloride resistance of AAMs. Therefore, the objective of this research is to investigate the different effects of MgO, nitrate intercalated LDHs and calcined LDH-CO 3 on chloride resistance of the AAFS binder. The reaction heat flow, reaction products, pore structures, mechanical properties and chloride resistance of alkali activated fly ash and slag incorporated with MgO, LDHs, and CLDHs are investigated. Furthermore, the mechanism to form the LDH in AAFS affected by MgO and CLDH is analyzed. The Ordinary Portland Cement (OPC) based samples are also used as the comparison to the AAMs samples. The respective contributions to enhance the chloride resistance of the three admixtures are evaluated.

Starting materials
Class F fly ash (FA), ground granulated blast-furnace slag (GGBS) and Enci Portland Cement CEM Ⅰ 52.5 R were utilized in this research. The chemical compositions were determined by X-ray fluorescence (XRF), as shown in Table 2.
The MgO utilized in this study was Magnesia 291 (provided by the MAGNESIA GERMANY) with a purity of 96%. The reactivity of the MgO was tested by the reaction time with acetic acid following [29]. The reaction time was 19.86 s, which was defined as a medium active magnesia. The specific surface area of the MgO was 51.39 m 2 /g and Fig. 3 showed the crystalline pattern of MgO.

Calcination
Step 2 precipitate was filtered, thoroughly washed, and dried at 75°C for 24 h. The set-up of the Mg-Al-NO 3 LDHs synthesis is shown in Fig. 4.
CLDH was calcined from the natural hydrotalcite (LDH-CO 3 ) with a heating rate of 5°C /min and kept at 550°C for 3 h. Then the material was cooled naturally in the furnace to 105°C before it was moved to a sealed centrifuge tube and kept in a desiccator under vacuum. The characteristic peaks of the LDH and CLDH are presented in Fig. 5, showing the crystalline patterns of the LDH and CLDH. The feature peaks of LDH are situated at 13.6°, while the feature peaks of the CLDH were situated at 41.2°, 51.3°, and 75.6°. Fig. 5d illustrates that the CLDH shows the nonhomogeneity state, which proves that the double layered structure collapsed to the non-rule layered structure.

Sample preparation
MgO, Mg-Al-NO 3 LDH, and CLDH replaced 2 wt% and 4 wt% of the alkali activated binder, which are described as modified samples. GGBS / FA mass ratio is 7:3 (see the sample ID in Table 3).
The activator used in this research was a mixture of analytical grade sodium hydroxide pellets (99% purity) and sodium silicate solution, and additional water is added to the solution to meet the requirement of liquid/solid (L/S) ratio. The SiO 2 / Na 2 O molar ratio was 1.5. The composition of the sodium silicate is shown in Table 4. For the mixtures shown in Table 3, the first letter P and M represent paste sample and mortar sample, respectively. A0 represents AAMs without additional admixture, while M represents MgO, L represents LDH and C represents CLDH, and the number represents the mass content equals to binder percentage.
The starting raw materials were mixed with the designed proportion in the concrete mixer following the preparation procedure suggested by . Paste samples were cast in the steel molds (40 mm Â 40 mm Â 40 mm) followed by the vibration. All the mortar samples were cast in the Ø 100 mm Â 200 mm cylinder mold followed by the vibration. The samples were all cured in the ambient temperature for intended period, e.g. 3, 7 or 28 days.

X-ray diffractometry
X-ray diffractometry (XRD) was performed by using a Bruker D4 Phaser instrument with Co-Ka radiation (40 kV, 30 mA). The powdered specimens were measured with a step size of 0.05°from 10°t o 90°2h, and the counting time was 1 s/step.

Isothermal calorimetry
The heat flow of the samples by different precursors were measured by an isothermal calorimeter set at 20°C (TAM AIR Calorimetry). It was noteworthy that the initial 4-6 min after mixing could not be measured due to the sample preparation procedure and the initial 0.5-1 h of the heat flow data could be inaccurate because of the instability of apparatus disturbed by the loading process. The results were normalized by the mass of the solid, excluding the water amount.

Nitrogen physisorption test
The nitrogen sorption test was conducted by TriStar Ⅱ 3020, Micromeritics. The pore size distribution was determined by the Barett, Jonyer, and Halenda (BJH) method [31] from the desorption branch. Before the test, paste samples were ground to powder and the samples were immersed in the 2-propanol for 24 h to stop the   Table 3 Formulations of the pastes and mortars using different additions.

Mercury intrusion porosimetry
The samples for the mercury intrusion porosity (MIP) analysis were selected from mortars, with the dimensions ranging of 3-5 mm. The selected samples were immersed in the 2-propanol for 24 h to halt the hydration process and then put in the vacuum drying oven for 3 days in order to remove all the pore water. The MIP tests were conducted by using a Micromeritics Autopore 9600 Mercury Porosimeter.

Compressive strength
After 28 days of curing, the strength of the samples was determined according to . The samples were centered on the platens of the machine. Subsequently, the load was increased smoothly at the rate of 2400 N/s over the entire load application until the fracture load.

Non-steady-state migration experiments
The Ø 100 Â 200 mm cylindrical mortars were cast and cut into 50 ± 2 mm thick slice. The cylindrical samples were put in the vacuum container for vacuum treatment for 3 h with the pressure in the range of 10-50 mbar (1-5 kPa), with the vacuum pump still running, then the samples were immersed in the saturated Ca (OH) 2 solution for 18 ± 2 h. Subsequently, the non-steady-state migration (NSSM) experiments were carried out according to the NT Build 492 [32].
After the NSSM experiments, the samples were split along the diameter of cylindrical surface into two part to expose the inner faces. Subsequently, the AgNO 3 solution was sprayed on the split surface, AgCl (white precipitation) was formed on the surface to reveal the chloride penetration depths. The non-steady-state migration coefficient was calculated by [32]: where: D nssm : non-steady-state migration coefficient, Â10 À12 m 2 /s; U: the absolute value of the applied voltage, V; T: the average value of the initial and final temperatures in the anolyte solution,°C; L: the thickness of the specimen, mm; the average value of the penetration depths, mm; t: test duration, hour.
The scheme of the NSSM test set up is shown in Fig. 6.
The primary differences between OPC pastes and AAFS pastes are obvious. The normalized hydration heat peak of the AAFS binder is 75.90% smaller than that of the OPC binder, comparing to the P-OPC and the P-A0 (Fig. 8). It is clear from Fig. 8 that increasing the MgO, LDHs and CLDHs content slightly prolong the TRRP.
Due to the high MgO content in the P-M2 and P-M4 (higher than 8 wt%), the high MgO systems can preferentially form hydrotalcite by reacting Al with Mg [24], and the hydration heat shows the prolonged TRRP. As shown in Fig. 7b and Fig. 8, the P-M2 and P-M4 extend the TRRP at 3.2% and 3.9% comparing to the P-A0, respectively. Firstly, MgO dissolve in the solution and release Mg 2þ ; Then, the free Mg 2þ react with the OHin the alkaline environment forming the Mg(OH) 2 precipitation and decreased slightly the pH value. Consequently, the MgO addition increases the TRRP and delays the hydration process slightly. Meanwhile, the increased MgO content has a slight effect on TRRP in the high MgO (higher than 8 wt%) and Al 2 O 3 (higher than 14 wt%) blended system [35].
The addition of LDHs in the geopolymer pastes shows a slight prolonged TRRP and decreased heat peak. The P-L2 and P-L4 exhibit the 7.2% and 11.2% delay comparing to the P-A0 reference, respectively. Firstly, the nitrate ion in the LDH releases some NO -3 , and the OHin the AAFS matrix partly exchange with NO -3 , resulting in the pH value decrease of the AAFS matrix. Thus, it delays the TRRP. Secondly, the agglomeration of LDH materials can also extend the TRRP in AAFS matrix. Therefore, the incorporation of LDH exhibit a decrease in TRRP and reaction heat peak.
Ke et al. [28] utilized the CLDHs to expedite the reaction kinetics of the different Na 2 CO 3 activated blast furnace slags. The CLDHs was associated with the removal of dissolved CO 2-3 from the fresh cement, yielding a rise in the pH value, and the potential seeding effects. While the CLDHs in sodium silicate and sodium hydroxide activated fly ash and slag system demonstrate the opposite effect, the CLDHs react withOH -, forming the LDH-OH and yielding a decrease in the pH value, consequently prolonging the TRRP. Therefore, the TRRP prolongs 0.81 h (P-C2) and 1.19 h (P-C4) comparing to the reference group P-A0 (shown in Fig. 8), which extends the TRRP 6.2% and 9.2%, respectively.

Reaction products
The XRD patterns of the samples show that the primary hydration products of AAFS binder are C-S-H, hydrotalcite-like phase, and Mullite. Fig. 9a) and b) show the crystalline phases in all the AAFS samples exhibit the same reaction products at 3 curing days and 28 curing days, respectively. Comparing Fig. 9 a) and b), the 3-day crystalline products of AAFS are practically consistent with the 28-day crystalline products. The primary AAFS crystalline products are formed in the initial 3 days. The observed results are consistent with the results presented in section 3.1, that the primary reaction heat flows and products are formed in the early stage. However, the OPC samples show the peak changes between 3 days and 28 days, which illustrates that the hydration still goes on after 3 days.  The main characteristic peak of hydrotalcite is identified at 13.6°from Fig. 9a) and b). However, Ke et al. [36] reported the main peak of the XRD pattern for a similar hydrotalcite-like phases after filtration was centered at 11.6°. The shift in the hydrotalcitelike peak can be attributed to the partial carbonation of the LDH.
Periclase is identified in specimens prepared with MgO addition (P-M2 and P-M4), which indicated that the MgO is not fully reacted with slag to form LDH. However, according to the previous study [8], the content of LDH formed from MgO in AAFS binder has the potential to grow over time, due to the solid state reaction of MgO in alkali activated cement.
The XRD patterns of P-L2, P-L4, P-C2, and P-C4 demonstrate almost the same intensity of LDHs. The characteristic peak of the CLDH phase is not found either in 3-day or 28-day XRD patterns, while the peak of the hydrotalcite-like phase appears, proving the CLDH completes the reconstruction process in AAFS to form the LDH.
Combining Fig. 9a) and b), the MgO and CLDH participate in the LDH formation in the AAFS matrix. Meanwhile, the externally added LDH can remain in the AAFS matrix. The three admixtures show the same results in the AAFS matrix in terms of the crystalline products.

Pore structure
As seen in Fig. 10a, the OPC pastes obtain the most mesopores (20-50 nm) and macropores (50-100 nm) and the lowest micropores. Meanwhile, the AAFS pastes are also the highest in mesopores and the lowest in micropores and macropores, as shown in Fig. 10b. The amount of the mesopores are nearly the same for all the AAFS pastes, which are quite different from the mesopores in OPC pastes. OPC pastes show a larger pore size than those of AAFS pastes. Fig. 10b shows the pore size distribution of AAFS pastes incorporating 2 wt% and 4 wt% MgO, LDH and CLDH. It shows the porosity slightly increases from the reference sample P-A0 to samples containing 2 wt% and 4 wt% MgO and CLDH. Furthermore, the porosity decreases with the LDH addition. Therefore, the pore size distribution is slightly improved by the LDH addition and the microstructure is refined.
The pore size distribution results of mortars from the MIP measurement are shown in Fig. 11. It is obvious that the pore size dis-tribution of OPC mortars mainly situates between the 10-100 nm (mesopores and macropores), which are consistent with the BJH method results shown in Fig. 10a.
The AAFS mortars are the highest in mesopores and the lowest in macropores. As described in Fig. 11b, the addition of MgO (M-M2 and M-M4) increases the total volume of the macropores comparing to the reference M-A0. However increasing the content of MgO can reduce the mesopores of AAFS mortars, which forms a small amount of the Mg(OH) 2 to fill the pores and corresponds well with the reference [37,38]. Fig. 12 and Fig. 13 show the cumulative porosity and the total porosity of AAFS and OPC samples. The total porosity of OPC mortars (20.01%) is much higher than the M-A0 mortars (3.969%).
The porosity of AAFS decreases with the increasing MgO dosage, which is generally in agreement with the BJH method result (Fig. 10 b). The LDH admixture in the AAFS samples also shows a slight influence on the porosity, but the 2 wt% and 4 wt% replacement of the fly ash and slag exhibit little difference in terms of the porosity (4.64% and 4.57%, respectively). However, comparing the porosity of P-A0, the incorporation of 2 wt% and 4 wt% CLDH (M-C2 and M-C4) has a positive correlation on total porosity, increasing up to 4.35% and 4.87%, respectively.
In general, the incorporation of MgO, LDH, and CLDH have slight effects on total porosity, while the unreacted MgO functions as the inert filler in the AAFS matrix. The LDH and CLDH admixtures have a slight effect on the microstructures. The AAFS mortars exhibit almost the same porosity, ranging from 3.96% to 4.93%.
Combining the nitrogen sorption results and MIP measurement results, the pore size of AAFS mortars mainly range from 3 to 10 nm (mesopores), while the pore size of OPC mortars mainly widely situate from 10 to 100 nm (both mesopores and macropores) The total porosity of the AAFS mortars are from 3.96% to 4.93%, while the OPC mortars is around 20.01%.

Compressive strength
Fig. 14 shows the compressive strength of the AAFS and OPC pastes at the curing age of 3, 7 and 28 days. Fig. 15 shows the relative strength of blended pastes at 3, 7 and 28 days compared to the AAFS (P-A0). The increasing trend of compressive strength shows that the strength of AAFS samples may continue to increase after the 28-day curing time, due to the incorporation of fly ash with a slow pozzolanic reactivity. These observed trends are related to the XRD results shown in section 3.1, and also in agreement with [39,40].
The compressive strength increases with the increasing content of the MgO, and this agrees with previous researches [41][42][43]. P-M2 and P-M4 exhibit higher compressive strength than other samples, up to 96.75 MPa and 103.58 MPa at 28 days curing age, respectively. The increase in relative strength for increased MgO content is attributed to the total content of MgO. Hwang et al. [43] reported that the content of MgO at 12 wt% (including the MgO inside binder material) in alkali activated slag showed the highest compressive strength. As the MgO dosage increases in P-M2 and P-M4, the compressive strength increases. The observed results agree with the mesopores distribution shown in Section 3.3, and the increased MgO content decreases the macropores and porosity that leads to the higher compressive strength.
Increasing the dosage of LDH in AAFS shows a slight decrease in compressive strength at 28 days, compared to the reference sample (P-A0). According to [44], the strength of the concrete samples gradually decreased with the increasing content of LDH. Meanwhile, the macropores of LDH samples increased with the rising    dosage of LDH in Section 3.3, the obtained compressive strength is consistent with these results.
The compressive strength development of CLDH addition in AAFS exhibits the same trend with LDH addition samples, namely the CLDH addition in AAFS slightly declines the compressive strength. Furthermore, P-C2 and P-C4 show the same compressive strength level with P-L2 and P-L4, at the 28 days curing age, respectively. The observed results are in agreement with [45], who reported that the CLDH content higher than 1 wt% resulted in the reduction of compressive strength.

Chloride migration
The results of the NSSM experiments are shown in Table 5 and Fig. 16. The OPC reference illustrates the highest coefficient D nssm among all mortar specimens. Meanwhile, the AAFS reference (M-A0) also shows the highest coefficient D nssm in all AAFS mortars, excluding the samples with 2 wt% LDH replacement (M-L2). Fig. 17 shows the relationship between replacing materials content and chloride migration coefficients. With the increasing replacement content, the MgO, LDH and CLDH mortars show the decreasing trend of the coefficient. In general, the incorporation of MgO, LDH, and CLDH increases the chloride resistance of AAFS mortars.
As one can see from Fig. 16 and Fig. 17, the MgO admixture decreases the chloride migration coefficient. The chloride binding capacity is due to (1) the LDH phase in the AAFS [8,46,47] and (2) pore size distribution [48][49][50]. On the one hand, the XRD pattern (in Section 3.1) shows that the hydrotalcite phase is formed in the AAFS samples, which can increase the chloride binding capacity. The MgO is a significant resource to form the LDH, and the chloride migration coefficient proves that the MgO addition enhances the hydrotalcite formation and then promotes the chloride resistance of AAFS [8]. The results in Table 5 illustrates that the coefficient D nssm of M-A0, M-M2, and M-M4 are 5.42, 4.68 and 4.55 Â 10 À12 m 2 /s, respectively. The chloride migration coefficient declines by 13.7% (sample M-M2) and 16.1% (sample M-M4), compared to the reference sample M-A0. On the other hand, the pore size distributions have a large effect on the chloride resistance of the AAMs samples with supplementary cementitious materials (SCMs), especially the mesopores and macropores affect significantly the permeability of concrete. The relationship between microstructure and permeability is investigated by many researchers [17,51]. As we can conclude from Section 3.3 that the increasing MgO content leads to the lower mesopores and macropores, thus the observed results are consistent with the NSSM test. Fig. 16 and Fig. 17 also show the nitrate intercalated LDH incorporation in the AAFS matrix in terms of the chloride resistance. Comparing M-L2 and M-L4 to M-A0, the chloride migration coefficient of M-L2 is higher than the reference M-A0 at 7.2%, while the coefficient of M-L4 is lower than the reference sample at 17.2%. The obtained results indicate that the nitrate intercalated LDH has a remarkable increase in the chloride binding capacity of AAFS with    the increasing content of LDH. The slight increase in the M-L2 coefficient is possibly due to the high applied voltage during the NSSM test. Spiesz et al. [52] reported that the coefficient D nssm was nearly 10% larger for samples under the 60 V than the samples under the lower voltages (35 V and 47.5 V). Generally, the dosage of the nitrate intercalated LDH has a direct influence on the chloride resistance of the AAFS matrix, Fig. 17 shows that the chloride resistance of samples is sensitively related to the increasing LDH dosage.
The CLDH incorporation has the highest contribution to chloride binding capacity in the AAFS matrix as shown in Fig. 16 and Fig. 17. The coefficient of M-C2 and M-C4 is lower than M-A0 at 24.2% and 46.9%, respectively. Obviously, the CLDH illustrates the higher chloride resistance than the MgO and LDH addition. This phenomenon may be attributed to the filler effect of LDH formation from CLDH on the AAFS mortars, and the microstructures also prove that the mesopores and macropores decrease as the CLDH content increased. Meanwhile, the mesopores and macropores results are consistent with the observed chloride penetrations in the AAFS matrix. The CLDH addition in AAFS has the lowest mesopores and macropores, thus CLDH contributes to the highest chloride resistance in all the AAFS samples in this study.
The porosity of the samples shows relatively close value (see section 3.3), thus the chloride binding capacity of AAFS samples and OPC samples is predominated by the chemical absorption. The secondary reaction product of AAMs, LDH, induces the main influence. The LDH reconstruction of CLDH has high effectiveness, and LDH addition directly affects the chloride binding capacity. While the MgO has relatively complex process to form LDH in AAMs. Thus, the CLDH admixture shows the highest chloride binding capacity, LDH admixture has moderate chloride resistance, and MgO admixture has the lowest effectiveness in chloride resistance.

Conclusions
In this paper, the reaction heat flow, reaction products, microstructure, compressive strength and chloride resistance of alkali activated fly ash and slag blends with the addition of MgO, LDH and CLDH are studied. The different effects of the MgO, Mg-Al-NO 3 LDH and CLDH incorporation are evaluated. The three admixtures in AAFS are proved to enhance the mechanical property, chloride resistance, and microstructure. According to the experimental results, the following conclusions can be drawn: The MgO, LDH and CLDH admixtures delay the time to reach the reaction peak (TRRP) in the AAMs matrix up to 3.9%, 11.2%, and 9.2%, respectively. The MgO admixture first forms Mg(OH) 2 precipitation, while the Mg-Al-NO 3 LDH and CLDH absorb the OH À . Therefore, all three admixtures slightly decrease the pH value in the AAFS matrix, and prolong the time to the reaction peak (TTRP). MgO promotes the formation of the in-situ growth of LDH phase. The CLDH completes the LDH-reconstruction process in AAFS, during the LDH-reconstruction process, the pH decreases with the consumption of OHby CLDH. Meanwhile, the externally added Mg-Al-NO 3 LDH remains in the AAFS matrix. The LDHs phase influenced by the three admixtures shows the partial carbonation in the AAFS matrix, and the carbonation results in the shift of LDH-NO 3 characteristic peak from 11.6°to 13.6°. The microstructures of AAFS are refined by the inner filler effect with unreacted MgO. The MgO shows the highest refinement on mesopores and macropores. While the Mg-Al-NO 3 LDH and CLDH admixtures have a limited effect on the microstructure. The increased MgO and CLDH content result in the higher compressive strength in AAFS, which is related to the mesopores refinement. While the LDH admixture shows limited influence on the compressive strength. MgO, Mg-Al-NO 3 LDH , and CLDH improved the chloride resistance of the AAFS matrix. The CLDH shows the highest chloride binding capacity and enhancement.
CRediT authorship contribution statement

Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.