E ﬀ ect of LDH on the dissolution and adsorption behaviors of sulfate in Portland cement early hydration process

: In recent years, it has been widely recognized that the incorporation of Mg – Al – LDH into cement - based materials can improve the salt corrosion resistance of cement - based materials. The reason for the improvement comes from the anion adsorption capacity of Mg – Al – LDH. It was con ﬁ rmed that the addition of Mg – Al – LDH would accelerate the setting and hardening of cement paste. With the increase in the Mg – Al - LDH content, the initial setting time of cement slurry with di ﬀ erent gypsum con tents will decrease by 10 – 50% and the viscosity of the cement slurry will increase by 100 – 200%. Depending on di ﬀ erentgypsumcontents,thedegreeofcementhydration varied. This article also found that the gypsum in the cement has a negative e ﬀ ect on the resistance to salt erosion, which was brought about by the Mg – Al - LDH adsorption capacity.


Introduction
With the continuous development of the field of civil engineering, the research on the durability of concrete has gradually drawn more and more attention, especially revolving around the durability research of sulfate corrosion. With the infiltration and corrosion of − SO 4 2 , Afm in concrete will be transformed into Aft, resulting in the expansion and cracking of concrete. How to improve the sulfate corrosion resistance of concrete has become one hot research direction in the field of concrete materials. Layered double hydroxide (LDH) is a relatively new material, which has a wide range of applications in the fields of environment, chemistry, biology, and energy. Its general structural formula is [M 2+ 1−x M 3+ x (OH) 2 ] x+ (A n− ) x/n ·mH 2 O [1] and its most common form in nature is Mg-Al-LDH. In recent years, Mg-Al-LDH has been incorporated into concrete to study its effect on the improvement of concrete resistance to chloride corrosion and sulfate corrosion, due to LDH's unique interlayer anion exchangeability in a solution environment ( [2][3][4]). There are many ways to synthesize Mg-Al-LDH, ranging over coprecipitation method, sol-gel method, hydrothermal synthesis method, ion exchange method, structure reconstruction method, mechanochemical method, etc. The particle size, crystallinity, and adsorption performance of LDH of different synthesis methods will also be different [5]. LDH has the function of structural reconstruction after high-temperature calcination. LDH after high temperature calcination is called LDO [6].
Later, Mg-Al-LDH was incorporated into cement-based materials as an anionic adsorbent. Studies have confirmed that the incorporation of Mg-Al-LDH into cement-based materials can enhance the ability of cement-based materials to resist sulfate corrosion and carbonate salt corrosion. In the pore solution simulating alkali-activated slag cement, Mg-Al-LDHs were bound to chloride ions through surface adsorption and ion exchange, which accounted for 90 and 10% of the total bound ions, respectively [7]. Shui et al. [8] found that the incorporation of Mg-Al-LDH can improve the chloride ion-binding capacity of cement paste, and theoretically proved that the incorporation of Mg-Al-LDH can delay the corrosion of steel bars by chloride salts in concrete. Yang et al. [9] confirmed that the incorporation of Mg-Al-LDHs reduced the diffusion rate of chloride ions in the mortar, and there is no obvious negative effect on the mechanical strength of the mortar. By testing the chloride equilibrium isotherm, corrosion potential, and polarization resistance of steel, Xu et al. [10] confirmed that the incorporation of LDH in a saturated calcium hydroxide solution environment can inhibit the corrosion of steel. Tatematsu and Saski [11] added LDH to prepare the concrete specimen, and proved that the long-term corrosion inhibition effect of LDH on steel is effective. Qu et al. [12] also confirmed the effect of LDH particle size and dosage on the chloride ion penetration resistance of cement-based materials through experiments. In addition to slowing down the corrosion of steel bars in concrete, LDH can improve the sulfate and carbonate resistance of cement-based materials [13][14][15][16].
While LDH improves the corrosion resistance of cementbased materials to inorganic salts, it also affects other properties of cement-based materials. Many scholars have found through experiments that the compressive strength of cement-based materials will decrease with the addition of the content of Mg-Al-LDH [4,14,17]. These findings support that Mg-Al-LDH does not participate in the hydration reaction in cement-based materials, which will create new weak interfacial regions. However, the addition of Ca-Al-LDH and LDO can improve the compressive strength of cement-based materials [12,18,19]. On the one hand, Ca-Al-LDH promotes the hydration reaction, and on the other hand, the high level of water absorption of Ca-Al-LDH and LDO reduce the water-cement ratio. Most of the research shows that the addition of LDH can significantly reduce the setting time of cementbased materials [13,18,20], with the increase of LDH content, the setting time is shortened further. Reasons behind this would vary depending on the type of LDH. Nanomaterials can shorten the setting time of cementbased materials and Ca-Al-LDH will form seeds in the cement paste to shorten the setting time. The reason for which Mg-Al-LDH shortens the setting time mainly depends on its interlayer anion. The type of interlayer anion determines whether Mg-Al-LDH will have water absorption and inhibitory effect on gypsum.
Mg-Al-LDH has a great influence on the early hydration process of cement paste because of the water absorption properties and the adsorption capacity of Mg-Al-LDH. The improvement of the corrosion resistance of cement paste by Mg-Al-LDH is mainly due to its anion adsorption capacity. In theory, the − SO 4 2 of gypsum in cement plays a negative role in the corrosion resistance of Mg-Al-LDH. The proposed study focuses on the effect of two different anion-based Mg-Al-LDH on the early hydration process of cement paste and explores the interaction between Mg-Al-LDH and gypsum in cement during cement hydration. The method used for synthesizing Mg-Al-LDH in this article was the urea synthesis method.
In this experiment, according to the mixing ratio, 31.73 g Mg(NO 3 ) 2 ·6H 2 O, 23.24 g Al(NO 3 ) 3 ·6H 2 O, 50 g CO(NH 2 ) 2 , and 10.7 g NH 4 NO 3 were poured into beakers, then added deionized water to each beaker, and stirred and dissolved to 50 mL. The four solutions were mixed and stirred at 40°C under nitrogen atmosphere for 4 h. Then let the mixture react at 90°C for 40 h. The obtained sample was fully washed with deionized water to pH 7 and then washed again to ensure that the residual urea in the sample is fully washed at 25°C. The washed filter cake was put into an oven to dry and grinded it to obtain Mg-Al-− NO 3 -LDH.

Adsorption kinetic experiments
The 0.2 g Mg-Al-LDH was added to 50 mL of the 5 mmol·L −1 Na 2 SO 4 solution and put it into a shaking bed for shaking at a shaking speed of 240 rpm. After shaking at different times, the centrifuge tube was allowed to stand and 3 mL of the supernatant was taken. The supernatant was filtered with a 0.22 µm membrane filter, and the ion concentration of − SO 4 2 in the filtrate was measured by an ion chromatography. Each of the above experiments was repeated three times and the mean value is given as the result. The obtained results were fitted with pseudo-first-order kinetic model (1) and pseudo-second-order kinetic model (2), respectively [21]. (1) where k 1 (min −1 ) and k 2 [g·(mg·min) −1 ] are the adsorption rate constants, respectively, and q e and q t (mg·g −1 ) represent the adsorption capacity at equilibrium and time t, respectively.

Adsorption isotherm experiments
The 0.2 g Mg-Al-LDH was added to 50 mL of Na 2 SO 4 solution of different concentrations and put it into a shaking bed for shaking at a shaking speed of 240 rpm. After shaking for 3 h, the centrifuge tube was allowed to stand and 3 mL of the supernatant was taken. The supernatant was filtered with a 0.22 µm membrane filter, and the ion concentration of − SO 4 2 in the filtrate was measured by an ion chromatography. Each of the above experiments was repeated three times and the mean value is given as the result. The obtained results were fitted with Langmuir isotherm model (3) and Freundlich isotherm model (4), respectively [21].
where K L (L·mg −1 ) and K f (mg (1−n) L n ·g −1 ) are the Langmuir adsorption constant and the Freundlich affinity coefficient, respectively; Q (mg·g −1 ) is the maximum adsorption capacity; C e (mg·L −1 ) is the concentration of adsorbate at equilibrium; and n is the Freundlich linearity constant.

Cement paste specimens
To compare and evaluate the interaction between Mg-Al-LDH and gypsum in cement paste, cement paste specimens were prepared to test for setting time, early hydration process, − SO 4 2 dissolution, and compressive strength. The water-cement ratio of the cement paste specimen is 0.35 and the mixing ratio of gypsum and Mg-Al-LDH is shown in Table 1. The experiment used cement clinker without gypsum and the composition of cement clinker is shown in Table 2.
For each group of cement paste specimens, six 20 mm × 20 mm × 20 mm cubes for the compressive strength test and the early hydration process test were prepared. The standard curing age of the test piece was 3 days.

Macro performance 2.4.1 Viscosity of early hydration of cement
The viscosity of cement paste was measured by a DV-11+ viscometer from Brookfield, USA. The rotational viscometer used a No. 63 rotor, which was a cross-shaped rotor. The viscosity test program started from 15 rad·min −1 of rotation speed, increased by 15 rad·min −1 each time, took and maintained each rotation speed for 15 s, took 1 data value per second, increased logarithmically to 150 rad·min −1 , and recorded the viscosity value.

Heat of hydration
The heat of hydration was measured using the American TAM Air eight-channel isothermal calorimeter. Each group of samples was tested with external stirring. The heat of hydration test time was 30 h. The ambient temperature was set to 20 ± 1°C, and the reference sample was distilled water with the same specific heat.

Microstructures
The XRD diffractometer is produced by the Netherlands Panalytical Company. The copper target was used for the test. The test voltage was 40 kV and the scanning rate was 0.024°·s −1 . Finally, the obtained XRD patterns were analyzed by using Jade6 software to compare with the card database.

Thermogravimetric and differential scanning calorimetry (TG-DSC)
The TG-DSC test was performed by the SDT650 synchronous thermal analyzer produced by TA Instruments in the United States. The experimental heating rate was 10°C·min −1 .

X-ray fluorescence (XRF) surface scanning analysis
The X-ray fluorescence (XRF) surface scanning analysis was tested by Bruker-M4 Plus micro-area X-ray fluorescence spectrometer of Shanghai Boyue Instruments. The test sample was a 3 cm block with a flat surface. The test method can detect the distribution of elements on the surface of the sample. , which proved that the synthesized products were pure and correct. . The correlation coefficients (R 2 ) of Langmuir and Freundlich models were 0.9999 and 0.9778, respectively. It is evident that, with the increasing initial sulfate concentration, the sulfate adsorption amount was increased dramatically and then reached equilibrium gradually. These isotherms were best fitted by the Langmuir model with higher R 2 values than the Freundlich model. A fundamental assumption of the Langmuir adsorption isotherm model is monolayer adsorption. This is consistent with the adsorption mechanism of Mg-Al-− NO 3 -LDH. Thus, the isotherm adsorption behavior supports that the synthesized product is Mg-Al-− NO 3 -LDH.    close to the experimental result. This could explain that the adsorption rate is controlled by the chemisorption mechanism, including chemical reaction, electron gain or loss, or electron sharing. This result indicated that Mg-Al-− NO 3 -LDH adsorbed − SO 4 2 mainly through anion substitution.  Figure 4(b) shows the relationship between the content of Mg-Al-− NO 3 -LDH and gypsum on the setting time of cement paste. In this experiment, the cement paste without gypsum achieved rapid setting. The cement paste reached the initial setting state when it was transferred from the mixing pot to the setting time test mold. From Figure 4(a), when LDH was not added, the initial setting time of the cement paste is significantly prolonged when the amount of gypsum in the cement increased from 0 to 3%. However, when the gypsum content increased from 3 to 4%, the initial setting time of cement did not change significantly, which proved that the retarding effect of gypsum reached the maximum effect.

Cement paste setting time
According to Figure 4(a), the initial setting time of the cement paste decreased when increasing the content of Mg-Al-− NO 3 -LDH, which also showed that the retardation effect of gypsum was inhibited. However, given the same Mg-Al-− NO 3 -LDH content, the initial setting time of the cement paste with 4% gypsum content was longer than that of the cement paste with 3% gypsum content. This phenomenon shed light on that the addition of Mg-Al-− NO 3 -LDH would play a role in accelerating the cement paste setting, in the cement-gypsum retarding system, while an appropriate increase in the amount of gypsum would also inhibit the accelerating effect of Mg-Al-−

Heat of hydration
The first descending phase of the hydration exothermic curve describes the dissolved exothermic release of cement particles, during which C 3 A reacts with − SO 4 2 from gypsum to produce a large amount of Aft. The first hydration exothermic peak represents the rapid formation of CSH and CH, and the cement slurry reaches the final setting state. The second hydration exothermic peak represents the reduction of − SO 4 2 concentration in hydration products, and Aft is transformed into Afm. Figure 5(a-c) shows the early hydration exothermic curves of cement paste with different Mg-Al-LDH contents when the gypsum content is 2, 3, and 4%, respectively. The peak corresponding time of each curve is shown in Table 3. The results of the hydration exotherm were analyzed together with the results of setting time. In the cement-gypsum retardation system, the addition of Mg-Al-− NO 3 -LDH inhibits the retardation of gypsum. In the absence of Mg-Al-− NO 3 -LDH, the content of 3 and 4% gypsum has the same retardation effect on cement. Nevertheless, the setting time of the cement paste with 4% gypsum content is significantly longer than that of the cement paste with 3% gypsum content, when Mg-Al-− NO 3 -LDH was added. had no obvious effect on the setting time and early hydration heat of the cement paste. It shows that the effect of Mg-Al-LDH on the hydration reaction of the cement-gypsum system is mainly due to the ion adsorption performance brought by its interlayer anion species.  Figure 7 provides the time-dependent change curve of the viscosity of the cement slurry with different ratios. Each curve in Figure 6 has an upward and downward curve. The upward curve represents the test process when the rotational speed increases, and the downward curve represents the test process when the rotational speed decreases. Both the upward and downward curves show that the higher the rotational speed, the higher the viscosity.   changes in S4L3 and S4L5 were different from those of S4L0. Although the viscosity of all three increases with the increase of time, their increasing trends were different. Figure 7 clearly shows that the viscosity increase trend of S4L0 increases with time, but the viscosity increase trend of S4L3 and S4L5 tends to be flat as time increases. The incorporation of Mg-Al-− NO 3 -LDH made the viscosity increase in the cement paste largely occurred in the early stage, and the viscosity growth trend slowed down after 40 min. The reason might be that the incorporation of Mg-Al-− NO 3 -LDH can help the cement paste to establish a hydration structure with greater thixotropy, and the flocculation phenomenon of the cement paste was enhanced. This also explained why the S4L5-3 curve has a lower viscosity value at 150 rad·min −1 . Because of the flocculation effect of S4L5 cement paste, part of the water was wrapped by the flocculation structure. After a period of hydration, the flocculation structure was broken by the rotor rotation, and this part of the water was released and the viscosity of the system was reduced. However, a hydration structure network with high viscosity had been initially formed inside the cement paste system. When the rotation speed was reduced, the rotation speed was not enough to break the slurry structure again. Therefore, the viscosity of the S4L5-3 curve increased rapidly when it dropped to 15 rad·min −1 , and the viscosity value of the S4L5-3 exceeded the other three curve values of the S4L5.

TG-DSC results of the effect of Mg-Al-LDH on cement hydration
From the results of the cement setting time, hydration heat, and viscosity tests, it can be concluded that Mg-Al-− NO 3 -LDH had an inhibitory effect on the retardation of gypsum in the cement system and shortened the time for the hydration product AFt to be converted into AFm in the cement paste. However, we need to further analyze the effect of Mg-Al-− NO 3 -LDH incorporation on the early hydration of cement The TG-DSC results of the cement system with 4% gypsum content are shown in Figure 8. According to the calculation and analysis in Figure 8, under the condition of removing the influence of Mg-Al-− NO 3 -LDH content on the TG-DSC results, in the cement paste of 4% gypsum system, when the content of Mg-Al-− NO 3 -LDH increases, the content of calcium hydroxide in the cement paste after hydration for 1 day decreases, and the degree of early hydration is lower. According to the previous test conclusion, the incorporation of Mg-Al-− NO 3 -LDH will promote the hardening of cement paste. In the meantime, the incorporation of Mg-Al-− NO 3 -LDH will reduce the degree of early cement hydration reaction, according to the TG-DSC analysis. The results showed that the hydration ranking was: S4L0 > S4L3 > S4L5.
The viscosity value test gave the reason for this phenomenon, owing to the fact that Mg-Al-− NO 3 -LDH would absorb a part of the water in the aggregated cement, and the amount of water used by the mixing water for cement hydration decreased, resulting in an increase in the viscosity of the system, resulting in incomplete hydration of cement particles and reduction in the hydration process. The answers to why LDH promoted the coagulation and hardening of cement paste were that, on the one hand, it adsorbed part of the water and reduced the water-cement ratio and, on the other hand, the adsorption of − SO 4 2 dissolved in gypsum in the cement paste system.
The incorporation of Mg-Al-− NO 3 -LDH would reduce the setting time of the 4% gypsum system cement paste,  while it reduces the early hydration degree. The more Mg-Al-− NO 3 -LDH was incorporated, the greater the effect.
3.6 XRD results of the effect of Mg-Al-LDH on cement hydration Figure 9 shows the XRD results of the 3-day-old cement paste samples. From Figure 10, with the increase in the Mg-Al-− NO 3 -LDH content, the diffraction peak of Mg-Al-− NO 3 -LDH became more obvious; however, the content of CH was gradually decreasing. With the increase in the Mg-Al-− NO 3 -LDH dosage, the peak intensity of C 3 S is higher. This showed that the hydration ranking was S4L0 > S4L1 > S4L3 > S4L5 > S4L7.

Early compressive strength results of cement paste specimens
The 3-day compressive strength results of the cement paste specimens are shown in Table 4. Regardless of the content of gypsum, the early compressive strength of the   -LDH did not participate in the cement hydration reaction but created an additional interfacial weakness in the specimen. It can be seen that under the condition of 4% gypsum content, the compressive strength of the specimen decreased with the increase in the Mg-Al-− NO 3 -LDH content, this phenomenon was consistent with the previous experimental results. According to the previous experiments, we know that the reason for why Mg-Al-− NO 3 -LDH reduced the degree of hydration reaction was that it would adsorb free water used to participate in the hydration reaction. Because, at 4% gypsum content, the cement paste mixed with Mg-Al-− NO 3 -LDH also could maintain a longer setting time. However, under the condition of 2 and 3% gypsum content, the addition of Mg-Al-− NO 3 -LDH would greatly reduce the setting and hardening time of the cement paste so that the free water content in the cement paste would decrease rapidly. With the rapid reduction in free water content in the hydration reaction system, Mg-Al-− NO 3 -LDH would release the adsorbed free water. Therefore, in the early hydration reaction, the unhydrated cement particles had enough water for the hydration reaction, which reduced the early negative influence of Mg-Al-− NO 3 -LDH on the hydration reaction. When Mg-Al-− NO 3 -LDH rapidly released free water, it played a role in promoting the coagulation of the cement paste so that the early hydration degree of the paste sample would be improved. This was the reason for the phenomenons of S2L5 > S2L3 > S2L1 > S2L0 and S3L7 > S3L5. The experimental results confirmed that the effect of Mg-Al-− NO 3 -LDH on the hydration of cement paste was related to the gypsum content in the cement paste.

Conclusion
This research investigated the combined effect of Mg-Al-LDH and − SO 4 2 from gypsum on the early hydration of cement in the cement paste by testing the setting time, hydration heat, viscosity of cement paste with different ratios, and XRD, TG-DSC, compressive strength, and XRF surface scan of cement paste specimens with different ratios. Based on the presented results, the following conclusions can be drawn: 1. In the gypsum-cement system, the addition of Mg-Al-− NO 3 -LDH will shorten the setting time of the cement paste and advance the time of the hydration exothermic peak, which will make the hydration product AFt more easily converted into AFm. has no obvious effect on the setting time, viscosity, and hydration of cement. But it will exist as a defect and reduce the compressive strength of cement specimens. 6. In future studies, the effect of Mg-Al-LDH on the erosion resistance of sulfoaluminate cement remains to be investigated.
Funding information: This research was funded by the "National Natural Science Foundation of China" No. 52178196.