Characterization of the Cracking Resistance Gradient of Bitumen Emulsion-Based Cold In-Place Recycling Mixtures over Curing by Semi-Circular Bending Test

Abstract

To better reveal the performance development of bitumen emulsion-based cold in-place recycling (BE-CIR) mixture over curing, a semi-sealed laboratory curing method was proposed in this research to simulate the in situ moisture evaporation process and cracking resistance of the BE-CIR specimen at different depths during a curing time of 28 days, which was also investigated by the semi-circular bending (SCB) test. The influencing factors of cement content (1.5% to 2.5%), initial moisture content (3.5% to 4.5%), curing temperature (25 °C to 45 °C) and relative humidity were investigated, and the significance of different factors affecting the performance development was also analyzed. The results indicate significant variations in cracking performance parameters at different depths, with the top part exhibiting notably higher tensile strength and fracture energy compared to the bottom part, and a gradient index (GI) is proposed to describe the difference. Cement content affected early tensile strength and fracture energy, while the initial moisture content affected the development rate of the performance. The influence of curing temperature was extensive, and as the temperature increased beyond 40 °C, the strength of the effect decreased. High humidity during the early stage of curing inhibited the strength formation and development of fracture energy. The performance development of the BE-CIR mixture is more significantly influenced by the moisture migration process, which is governed by curing temperature and relative humidity, as opposed to the cement content and initial moisture content.

1. Introduction

In recent years, with the continuing attention given by society to environmental issues, low-energy and low-emission technologies for asphalt pavements have also received rapid development. Bitumen emulsion-based cold in-place recycling (BE-CIR) as a cost-effective and environment-friendly maintenance technology has gained worldwide application [1,2,3,4]. The composition of the BE-CIR mixture includes emulsified bitumen, virgin aggregates, recycled asphalt pavement (RAP), additional water, and various inorganic additives (e.g., Portland cement) [5]. Due to the low viscosity of bitumen emulsion, BE-CIR mixture can be mixed, paved and compacted at an ambient temperature, as opposed to the traditional hot-mix asphalt (HMA) [6,7]. The adding of Portland cement can improve the mechanical properties of mixtures by promoting the dispersion of the bitumen, regulating the breaking of the emulsion and accelerating the curing time [8,9,10]. The additional water is an essential component, since it contributes to homogeneous distribution of bitumen droplets and good encapsulation of aggregates [11,12,13].

Because of the wide variation in the component materials, the mechanical performance of BE-CIR mixtures is a combination of the properties of the bitumen and the cement [14]. Bitumen is a viscoelastic binder with low stiffness, while cement is a hydraulic binder with brittle hydration products and high compressive strength and stiffness. Increasing the cement content in this composite leads to variations in the microstructure of the co-binder, and consequently increases the stiffness and decreases the temperature dependence of the mechanical properties [15,16,17]. Despite such variability, a common feature of BE-CIR mixtures’ mechanical performance is their evolutionary behavior with curing time [18,19,20]. During the curing process, the water content is reduced to equilibrium with the surroundings due to evaporation, and the mixture evolves from the short-term ‘‘fresh” state to the long-term ‘‘hardened” state under the interaction of different physical and chemical mechanisms, including emulsion breaking and setting and cement hydration.

The BE-CIR mixtures have been investigated in a number of studies, showing good rutting resistance and relatively high modulus, while its crack resistance and fatigue durability are relatively weak compared to HMA [21,22,23,24,25,26]. Given that the fracture characteristics of the BE-CIR layer directly affect the service life of the pavement structure, it is important to study the growth behavior of the fracture properties in order to facilitate the enhancement of the cracking resistance of the mixes. Factors influencing the mechanical properties of BE-CIR include not only material composition and compaction, which are similar to those of HMA, but also the curing process [27,28,29]. Saadoon et al. [30,31] studied the performance of CIR mixtures with different cements and emulsified bitumen. The Marshall stability was found to have a good relation with the water loss, which is mainly influenced by the bitumen emulsion, but not with the type and content of cement. Kim [32] confirmed that the moisture content and length of the curing period have significant effects on the properties of the CIR mixtures made with emulsion asphalt. Effects of different curing conditions in terms of temperature (varied 15 °C–60 °C) and relative humidity (varied 6%–95%) were studied [33,34,35]. It was demonstrated that higher curing temperatures can lead to an increase in strength, and that higher humidity at an early stage can improve the hydration of cement [27,36]. However, the extent of influence between material design parameters and curing conditions has not been comprehensively compared.

In addition, from the perspective of field performance, a unique gradient distribution of the in situ moisture content during curing in BE-CIR pavement was identified by monitoring the humidity at different depths using humidity sensors [37]. The tensile strength of field cores determined using semi-circular bending tests (SCB) also exhibited gradient features which are higher in the top half and lower in the bottom. However, the reasons for the gradient features of field cores include moisture evaporation and vehicle loading pressure. It is essential to identify the reason for the gradient feature and investigate the forming process. Additionally, another work has evaluated the effects of field-water evaporation and heat-transfer conditions in the laboratory [38]. A semi-sealed laboratory curing condition was promoted as being more consistent with field performance.

To the authors’ best knowledge, the BE-CIR mixtures are mostly viewed as an integrated whole when evaluating their mechanical properties, which may display a great difference from examples in the field. Unlike the hot-mix asphalt, the strength formation of BE-CIR is time-variable due to the coupling effects of cement hydration, emulsion breaking and water evaporation. However, the observed depth variation mechanism and its corresponding influencing factors have not been investigated. Therefore, it is essential to identify the reason for the gradient feature and investigate the forming process. This paper intends to reach a better understanding of the gradient developmental behavior of crack resistance in BE-CIR mixtures under different influencing factors.

2. Scope and Objectives

The crack resistance associated with different depths in BE-CIR mixtures was evaluated by an SCB test under semi-sealed laboratory curing conditions. The main objectives are listed as follows:

• Investigate the gradient characteristics of the cracking resistance of BE-CIR mixture affected by water evaporation;
• Characterize the influence of factors including mixture design parameters and curing conditions on the gradient developmental behavior of fracture properties;
• Evaluate the influence of various factors on the fracture properties and gradient distribution.

3. Materials and Experimental Design

3.1. Materials

The RAP materials were collected from a highway pavement maintenance project using the BE-CIR technique in the Jiangsu Province of China. As shown in Figure 1, the gradation of RAP material from this project complied with the medium-gradation limitations specified by the Jiangsu Province agency [39]. The type of emulsified bitumen was cationic slow-setting, and the cement was Portland Ordinary Cement P.O. 42.5; the properties of these materials are listed in

Table 1. Properties of emulsified bitumen and P.O. 42.5 cement.

Properties of the Emulsified Bitumen		Value
Residue content (%)		64
Penetration (25 °C; 0.1 mm)		59.1
Soft point (°C)		59.9
Ductility (15 °C; cm)		110
Sieve residue (1.18 mm; %)		0.03
Storage stability (5 d, 25 °C; %)		1.47
Storage stability (1 d, 25 °C; %)		0.42
Properties of Cement
Setting time (min)	Initial set	3.14
	Final set	4.62
Flexural strength (MPa)	3 d	4.9
	28 d	7.2
Compressive strength (MPa)	3 d	24.1
	28 d	51.4

, and are in accordance with the Chinese standards for cold recycling [40]. In the BE-CIR project, 100% of RAP materials were recycled and no addition aggregates were added in the field. So, similarly, the mixture samples were prepared with a 100% reuse of RAP.

3.2. Specimen Preparation and Curing Conditions

Firstly, the RAP was dried in an oven at 60 °C to a constant mass and then cooled naturally to an ambient temperature in preparation for mixing. The mixing was performed in three steps: firstly, the RAP and cement were mixed, then water was added to the mix, and finally the emulsified bitumen was added. The mixing time for each step was 60 s. This procedure enabled the aggregates and fillers to be wet before coming into contact with the binder so that the adsorption capacity of the emulsified bitumen could be reduced, which favored the homogeneity of the mix. Moreover, the Chinese specification stipulates that the mixing time should not be longer than 60 s after the addition of asphalt emulsion, in order to avoid the breaking of asphalt emulsion. After mixing, the specimens of 150 mm in diameter and 105 ± 2 mm in height were compacted by the Superpave Gyratory Compaction (SGC); approximately 3800 g of the mixture was used.

The accelerated curing condition usually employed in most studies involves placing the specimen in the oven unsealed at an ambient temperature. However, these methods ignore the boundary condition of water evaporation under field conditions, and the moisture migration may contribute to the inconsistent strength-growth behavior.

Based on a previous study, a semi-sealed laboratory curing condition was designed to simulate the in situ curing condition, and proved to be more consistent with field cores [38]. This method employed a waterproof fabric made of Teflon material wrapped around the side and bottom surfaces of the specimen to ensure that moisture within the mixture could only evaporate from the upper surface, as shown in Figure 2. The environmental chamber was then used for curing at different combinations of temperature and relative humidity.

3.3. Testing Program

In this paper, a total of four influencing factors were investigated, including the two mixture-design factors of cement content and initial moisture content and the two curing environmental factors of temperature and relative humidity. The orthogonal experimental design method was used; a total of 10 different BE-CIR mixtures were designed for testing, and the influencing factors corresponding to each BE-CIR were as shown in

Table 2. Design of different BE-CIR.

Mix ID	Cement Content (%)	Initial Moisture Content (%)	Curing Temperature (°C)	Relative Humidity of Curing (%)
BE-CIR1	1.5	4	35	30%
BE-CIR 2	2
BE-CIR 3	2.5
BE-CIR 4	2	3.5
BE-CIR 5		4.5
BE-CIR 6		4	25
BE-CIR 7			30
BE-CIR 8			40
BE-CIR 9			45
BE-CIR 10			35	90% (first 5 days)

. The asphalt emulsion content of all BE-CIR was set at 3.3% according to the mix design used during the field project. In addition, a relatively high cement content of 2% was adopted in the previous project to address the lack of high temperature performance and weak early strength. Therefore, in the orthogonal experimental design, a cement content of 2%, an initial moisture content of 4% and the average curing temperature of 35 °C used in the field project were centered. Three cement contents of 1.5%, 2% and 2.5%, and three initial moisture contents of 3.5%, 4% and 4.5% were studied, respectively. The initial moisture content consists of the water in the emulsified bitumen and the extra water. All BE-CIR had the same emulsified bitumen content of 3.3%, and the moisture in the emulsified bitumen was 1.188% according to the residue content. The different initial moisture content was controlled by the extra moisture. Since the BE-CIR pavement is not suitable for construction in environments with low temperature, the range of the curing temperature was set to extend from 25 °C to 45 °C, and every 5 °C was set for different groups. In terms of the relative humidity of curing, except for a constant relative humidity of 30%, this paper considers one situation in which relative humidity was higher in the early 5 days of curing (90%), which represents rain in this curing period.

In order to investigate the growth of crack resistance over curing time, for each BE-CIR, 12 parallel specimens were prepared for the following tests at different curing days. Four different curing durations of 3, 7, 14 and 28 days were selected, respectively, and there were three parallel specimens for test for each curing day. At each curing day set for a test, the mass of each sample was firstly measured to calculate the water-loss rate by comparing the result with the initial mass after compaction. The wrapped waterproofing was then removed and the specimens cut for an SCB test according to the process described in Figure 3. The sample size for the SCB test was a semicircle with a diameter of 150 mm and a thickness of 50 mm. Since this paper investigated the differences in performance development in the depth direction, four semicircular samples obtained from one specimen should be divided into two groups, top and bottom, respectively. The top part represented the portion containing a surface in direct contact with air during curing, and the bottom part was the remaining portion. Therefore, during each curing interval, three parallel specimens were utilized to obtain a total of 12 semicircular samples—six from the top portion and six from the bottom portion.

3.4. SCB Test

A UTM-25 universal testing system was used to perform the SCB tests in an environmental chamber according to AASHTO TP 124 [41]. The span between supporting rollers was set to 0.8 times the specimen diameter, which was 120 mm, as shown in Figure 4a,b. Before the test, specimens were placed into the environmental chamber for 4 h at the test temperature of 25 °C. The loading rate was set to 50 mm/min, and the applied load and displacement of the load actuator could be obtained. According to the load–displacement curve, as shown in Figure 4c, two indices, tensile strength (${\sigma }_t$) and fracture energy ($G_f$), could be calculated by Equation (1) and Equation (2), respectively [42]. Apparently, these two indicators represent different perspectives on the cracking resistance of a mixture. The tensile strength denotes the ultimate load that the sample can withstand; the greater the strength, the stiffer the sample and the more difficult it is to crack. The fracture energy represents the flexibility of the sample, and the greater the fracture energy, the better the ability of the sample to work with cracks.

$${\sigma }_t=\frac{4.8F_{max}}{BD}$$

(1)

$$G_f=\frac{W_f}{Are{a}_{lig}}$$

(2)

where ${\sigma }_t$ = SCB tensile strength (MPa); $F_{max}$ = peak load (kN); $B$ = specimen thickness (mm); $D$ = specimen diameter (mm); $G_f$ = fracture energy (J/m²); $W_f$ = fracture work (J), $\int Pdu$, where P = applied load (kN), u = vertical displacement (m); $Are{a}_{lig}$ = ligament area (m²), $BD/2$.

In addition, the gradient index (GI) is defined to characterize the difference in cracking resistance in the depth direction. The gradient indices of tensile strength and fracture energy were calculated by Equation (3) and Equation (4), respectively.

$$\mathrm{GI}-{\sigma }_t=\frac{{\sigma }_t^{top}-{\sigma }_t^{bottom}}{{\sigma }_t^{bottom}}\times 100$$

(3)

$$\mathrm{GI}-G_f=\frac{G_f^{top}-G_f^{bottom}}{G_f^{bottom}}\times 100$$

(4)

where $\mathrm{GI}-{\sigma }_t$ = the gradient index of tensile strength; $\mathrm{GI}-G_f$ = the gradient index of fracture energy; ${\sigma }_t^{top},{\sigma }_t^{bottom}$ = tensile strength of top or bottom layer; $G_f^{top},G_f^{bottom}$ = fracture energy of top or bottom layer.

4. Results

4.1. Water Loss

The results for water loss for different BE-CIRs over curing time are shown in Figure 5. Overall, there was a trend of rapid increase followed by gradual stabilization of water loss for all samples except BE-CIR10. The rapid increase was mainly observed within 0–7 days of curing, and the variation was very small after a curing of 14 days.

According to Figure 5a, representing BE-CIRs with different levels of cement content, the water loss at 1.5% cement content was significantly smaller than with the other two cement contents, and the difference appeared mainly in the period of 0–7 days curing, while it became smaller after curing for 14 days. This phenomenon arises from the direct relationship between cement content and the quantity of free water combined with the cement, leading to an elevated rate of water loss. However, when the cement content was increased from 2% to 2.5%, the difference in water loss was minor and only slightly increased, from 0.78% to 0.83% for BE-CIR with 2.5% cement content at curing 1 d. The potential explanation for this occurrence could be that when the cement content surpassed 2%, an excess of cement particles caused uneven bonding with the bitumen film. As a result, some cement particles failed to make contact with the free water, thereby impeding further increases in water loss. The results for different initial water contents are shown in Figure 5b, which indicates that the water loss of the three mixtures was relatively similar to each other at curing 1 d, at approximately 1%, and with the increase of the curing, the water-loss rates of the three mixtures progressively exhibited the law of increasing according to the increased initial water content. Nevertheless, by using the initial moisture content to back-calculate the residual moisture content, it can be found that the residual moisture content was almost the same, at 0.8%, for different mixtures after curing for 28 days. This indicates that the initial moisture content could only affect the early rate of water loss but not the final residual moisture content.

Figure 5c illustrates the variation of water loss of BE-CIRs over time under different curing temperatures. It can be noticed that the water-loss rates for different days increased with the increase of the curing temperature. There was a significant increase in the water loss for every 5 °C increase in the interval from 25 °C to 40 °C, while the water loss at 40 °C and 45 °C were much closer to each other, which suggests that when the curing temperature was greater than 40 °C, the rate of water evaporation inside the BE-CIR mixtures might have reached its peak. This indicates that the curing temperature significantly affects the water migration and evaporation process of the mixture. In addition, comparing the water-loss curves at 40 °C and 25 °C, the water loss of the BE-CIR had reached 3% at 7 d when the curing temperature was 40 °C, and became stable after 14 d, while the water loss at a curing temperature of 25 °C was about 0.7% lower than that at 40 °C after curing 28 days. It can be assumed that there was still some residual water inside the BE-CIR mix at lower conditioning temperatures, and thus it might still take 2–3 months for the water to be completely evaporated.

The variation in water loss for two different humidity conditions is presented in Figure 5d. The water loss of BE-CIR10 was negative under the high-humidity condition for the first 5 curing days, which implies that the mixture could absorb a certain amount of additional water, and there exists an extreme value for this water absorption, which is about 0.5%. When the ambient relative humidity decreased to normal humidity, the water loss started to increase, and the rate of this increment was basically the same as that of BE-CIR4. This suggests that the occurrence of rainy conditions at the early stage of the curing will delay the water dissipation process accordingly.

4.2. Performance Parameters from SCB and Gradient Distribution

4.2.1. Effect of Cement Content

Figure 6 illustrates the SCB indices and GI over curing of different cement contents. It can be observed that both tensile strength and fracture energy continuously become larger along with the curing. However, the incremental rates of the two indices were different, and the tensile strength was constantly increasing in different curing periods, while the fracture energy increased significantly in curing 3–7 d, and the growth rate decreased in the subsequent period. This is explained by the fact that the tensile strength could more significantly be influenced by the hydration of the cement, which continues throughout the 28 days of curing. On the other hand, the fracture energy could be significantly influenced by the emulsified bitumen. During the curing period of 3–7 days, there was a rapid dissipation of water, causing the emulsified bitumen to rupture. Consequently, the adhesion between the bitumen film and the aggregate increased rapidly, resulting in a swift escalation of the fracture energy.

Considering different cement contents, at 3 days of curing, the higher the cement content, the higher the tensile strength ${\sigma }_t$, especially when the cement content increased from 2% to 2.5% and the tensile strength of the top part increased from 0.7 MPa to 1.00 MPa, with an increment of 42.8%; the increment for the bottom part was slightly smaller but also went up, rising to 28%, indicating that an increase in the cement content can improve the early strength of the BE-CIR mixture. However, as the curing proceeded, the tensile strength of the BE-CIR with 2.5% cement content grew more slowly and was already smaller than those of the other two cement content levels at 7 days of curing. After curing 14 days, the tensile strength of the BE-CIR with 2.0% cement content was the highest, followed by that of 1.5%, and that of the 2.5% was the lowest. This indicates that there is an optimum cement content for the tensile strength of BE-CIR mixes in the long term. When the cement content is higher, it will lead to a decrease, in the long term, of tensile strength, although it can improve the early strength. The fracture energy was different from tensile strength to some extent. At 3 days of curing, the mixture with the lower cement content similarly had the lower fracture energy, while after 7 days, the higher the cement content the lower the fracture energy. The underlying cause is attributed to the fact that as the cement content increases, the mixture becomes stiffer, leading to reduced resistance against bending and tensile deformation. Consequently, this results in a decrease in fracture energy.

Further, comparing the performance of the top and bottom parts, the tensile strength and fracture energy of the top part was always greater than those of the bottom part, regardless of days of curing. This phenomenon was mainly due to the restriction of unidirectional evaporation of water, which can lead to an uneven distribution of residual water within the mixture; different water contents could lead to different degrees of cement hydration, forming different microscopic morphologies and adhesive capacities of the cement-emulsified asphalt mortar. As the curing proceeded, the gradient index (GI) of both ${\sigma }_t$ and $G_f$ kept decreasing, which also meant that the difference in cracking performance between the top and bottom parts kept decreasing. For a GI of ${\sigma }_t$, the higher the cement content, the greater the GI. The GI-${\sigma }_t$ of BE-CIR with 2.5% cement content could even rise beyond 50% at 3 d of curing. Regarding a GI of $G_f$, the mixture with 2.5% cement content had the highest GI at 3 d of curing, while the mixture with 1.5% cement content was the highest after curing for 28 d. This also demonstrates that the higher the cement content, the more significant the influence on the inhomogeneous distribution of the residual water is at the early stage of curing. Meanwhile, cement plays a weaker role in the fracture-energy growth behavior, compared to its role in tensile strength.

4.2.2. Effect of Initial Moisture Content

Results for the SCB Indices and GIs of different initial moisture content are illustrated in Figure 7. The development of both ${\sigma }_t$ and $G_f$ for BE-CIRs with different initial water contents were almost similar. The tensile strengths of the different mixtures were almost the same at 3 and 28 days of curing. And only during 7 to 14 days of curing there was a (small) increase in ${\sigma }_t$ with an increase in initial moisture content, which may have been due to a more rapid evaporation of water in this period. However, there were relatively large differences in $G_f$ between three BE-CIRs. The mixture with the largest initial moisture content, of 4.5%, had the largest $G_f$, indicating that the rapid migration of moisture during the curing period could contribute to fracture resistance development.

From the perspective of GI, the GI of ${\sigma }_t$ decreased continuously with curing time, whereas the GI of $G_f$ increased from 3 to 7 days, reached a maximum at 7 days of curing, and then decreased until the 28-day mark. The maximum of GI-$G_f$ could be over 80%. And the GI-$G_f$ could drop to 20% after curing 28 days. In addition, the greater the initial water content, the greater the GI of both ${\sigma }_t$ and $G_f$. The rationale behind this is straightforward: a higher initial water content can cause a less uniform distribution of residual water in the depth direction.

4.2.3. Effects of Curing Temperature

Figure 8 shows the SCB Indices and GIs of BE-CIRs with different curing temperatures. At 3 days of curing, there was a large difference between the ${\sigma }_t$ and $G_f$ of BE-CIR mixtures at different curing temperatures. The ${\sigma }_t$ of the BE-CIR at the lowest temperature of 25 °C was only 0.44 MPa for the top part and 0.40 MPa for the bottom part, while it reached 1.14 MPa and 0.92 MPa for the top and bottom parts, respectively, at a curing temperature of 45 °C, which were both more than twice as much as those at 25 °C. Similarly, the $G_f$ of the BE-CIR at 45 °C was more than three times that of the mix at 25 °C. This demonstrates that there was a significant impact made by the curing temperature on the early development of anti-cracking properties. After curing for 7 days, the ${\sigma }_t$ of BE-CIR with a curing temperature of 40 °C was already closer to that of the mixture at 45 °C, with a difference of only 0.13 MPa in the top layer, while the difference with the mixture cured at 35 °C was 0.42 MPa. The results for $G_f$ presented larger differences between curing temperatures from 25 °C to 35 °C, with smaller differences between those from 35 °C to 45 °C. Combined with the results of the water loss, when the curing temperature rose to more than 40 °C, the water loss was basically the same at 7 days of curing, showing similar results with the tensile strength, indicating that the variation of the moisture inside the mixture generated quite a significant influence. Moreover, after 28 days of curing, the ${\sigma }_t$ and $G_f$ of BE-CIRs still increased with the increase of the curing temperature, and the fracture energy was significantly less at curing temperatures of 25 and 30 °C.

Considering the differences between top and bottom parts, the effects of the curing temperature were obvious. The GIs for the higher curing temperatures of 40 and 45 °C varied relatively closely over time and corresponded to smaller values, reaching only 30% during the first 7 days of the curing and dropping to less than 10% after 28 days. The other three curing temperatures had similar GI-${\sigma }_t$, at 15%, after curing for 28 days, while the GI-$G_f$ of the lowest curing temperature (25 °C) presented a much larger value, at 60%, than the other two values, which reached about 30%. It is known that different curing temperatures will result in different rates of moisture evaporation, which leads to different residual moisture contents as well as a gradient distribution, ultimately contributing to an inconsistent development of properties.

4.2.4. Effects of Relative Humidity

For easy expression, the constant relative humidity and early high relative humidity are replaced by 30% and 90% in this section; the results are presented in Figure 9. Due to the mixtures of 90% curing at a high relative humidity during the first 5 days, and the fact that water did not dissipate, or even the possible existence of water absorption, both the ${\sigma }_t$ and the $G_f$ of the top and bottom parts were basically less than those for the mixtures curing in a constant low relative humidity. The difference of performance between the two categories of mixtures was larger in the top parts than in bottom parts. In addition, the GI of ${\sigma }_t$ and $G_f$ was close to 0 or even negative at 3 and 7 days. This is further evidence that evaporation of moisture is directly responsible for the gradient in anti-cracking properties of the BE-CIR mixture. The presence of rain in the early stages of curing could lead to a delay in moisture migration, and consequently to a delay in the occurrence of performance development.

4.3. Multifactor Analysis of Variance

The multi-factor analysis of variance (MANOVA) test was conducted by MATLAB to investigate the significance of the four factors. A variable is significant when the p-value is less than the level of significance (0.05). Results of MANOVA for both ${\sigma }_t$ and $G_f$ at different curing times are listed in

Table 3. MANOVA of factors for properties at different curing times.

p-Value	Top Layer				Bottom Layer
	3 d	7 d	14 d	28 d	3 d	7 d	14 d	28 d
Factors	${\sigma }_t$
Cement content	0.0018	0.8175	0.8047	0.7591	0.1292	0.3635	0.6669	0.4106
Initial moisture content	0.6732	0.2672	0.1855	0.8778	0.6322	0.7351	0.7579	0.7195
Curing temperature	<0.0001	<0.0001	0.0007	0.0001	0.0005	<0.0001	0.001	0.0038
Relative humidity	0.0211	0.0042	0.2022	0.0726	0.9652	0.6468	0.1225	0.0787
Factors	$G_f$
Cement content	0.5065	0.4467	0.0083	0.0013	0.2497	0.091	0.0544	0.0228
Initial moisture content	0.8663	0.5907	0.3563	0.1447	0.6329	0.6269	0.5379	0.6985
Curing temperature	<0.0001	0.011	0.0046	0.0014	0.0001	0.001	0.0019	0.002
Relative humidity	0.2641	0.045	0.6027	0.4465	0.0047	0.799	0.8142	0.3328

. According to the p-value, the value for cement content was only significant in the ${\sigma }_t$ of the top part at 3 days of curing, the top part at day 14, and the bottom part at 28 days of curing. This indicates that the cement content will primarily exert a significant impact on both the early strength and the long-term ability to resist deformation and function in the presence of cracks. The initial moisture content was not significant for properties at any curing time. The effect of curing temperature at each curing time was significant, with a p-value of less than 0.01, while the p-value of relative humidity was only smaller than 0.05 for day 3 or day 7, which was mainly due to the humidity mode used in this paper. As humidity conditions represent the weather during curing, the results prove that the presence of rain during the curing process has a significant influence on the anti-cracking properties. Therefore, compared to moisture content, the migration or evaporation process of water has a greater influence on the formation of the strength and long-term performance.

In addition, MANOVA of factors for the gradient index was also employed, as shown in

Table 4. MANOVA of factors for the gradient index.

p-Value	GI
	3 d	7 d	14 d	28 d
Factors	${\sigma }_t$
Cement content	0.1199	0.0236	0.6743	0.2873
Initial moisture content	0.2592	0.0558	0.5349	0.5541
Curing temperature	0.6127	0.0022	0.0534	0.0578
Relative humidity	0.008	<0.0001	0.2465	0.0918
Factors	$G_f$
Cement content	0.7288	0.0962	0.029	0.001
Initial moisture content	0.1566	0.8608	0.6378	0.0839
Curing temperature	0.7818	0.8263	0.0210	0.0095
Relative humidity	0.071	0.0007	0.1634	0.6983

. The most significant of the four factors is the curing temperature. However, comparing the different days of curing, the significance of curing temperature on each performance level is mainly evident after 7 days of curing. For the other three factors, relative humidity had a significant effect on the GI at 7 days due to the particular curing condition, while the cement content only manifested significance at ${\sigma }_t$ after 7 days, which was due to the effects of the different cement contents on the early strength. Therefore, the more significant factors influencing the gradient characteristics presented by BE-CIR mixtures are the curing temperature and relative humidity.

5. Discussion of the Influencing Factors

The curing process of BE-CIR mixture is usually accompanied by a continuous decrease in water content due to evaporation and an increase in strength due to emulsion breaking and cement hydration [31,34,43,44]. The depth and time variations of BE-CIR pavement are affected by both the material component and environmental conditions. In this study, the two material-component factors of cement content and initial moisture content as well as the two environmental factors of curing temperature and relative humidity were examined. The addition of cement can improve the early strength of BE-CIR, but the improvement effect is diminished when the cement content exceeds 2%. Furthermore, it was observed that the depth fluctuation of BE-CIR became more pronounced with an increase in cement content, suggesting that the hydration extent was influenced by the gradient evaporation process. As for the initial moisture content, it can influence the rate of water dissipation significantly, but it has little effect on the final residual moisture content trapped in the BE-CIR. The initial moisture content may mainly affect the time and depth variation of BE-CIR pavement during a curing process of around 2 weeks. Compared with the material component, the environmental factors showed more extensive effects on the performance variation of the BE-CIR. Especially after curing 28 days, when the difference of ${\sigma }_t$ between 25 °C and 45 °C could be 0.4 MPa, while it was less than 0.1 MPa between different cement contents and moisture contents. As the temperature surpassed 40 °C, the acceleration impact became less pronounced. Moreover, during the initial stage of curing, high humidity hindered water dissipation from the mixture, and the mixture could even have absorbed approximately 0.5% more water from the air. Excess water also impeded cement hydration, suspending the development of crack-resistance properties [17,35]. Overall, the growth in the cracking resistance of BE-CIRs is extremely dependent on the moisture environment they are subjected to. Consequently, understanding and possibly controlling moisture levels throughout the construction and curing phases is crucial for ensuring the longevity of BE-CIR pavements.

6. Conclusions

This paper evaluated the cracking resistance of different parts of BE-CIR mixtures by SCB tests based on a semi-sealed laboratory curing condition simulating in situ moisture evaporation. The influence of four factors, comprising cement content, initial moisture content, curing temperature and relative humidity, were investigated. Based on the analysis, the following conclusions can be drawn:

(1)

The performance growth characteristics of BE-CIR mixes at various depths exhibited notable differences, evident through the top part’s higher tensile strength and fracture energy compared to the bottom part. This occurrence can be attributed to the non-uniform distribution of residual water within the mixture due to moisture evaporation. Consequently, this results in distinct microscopic morphologies and adhesive capabilities of the cement-emulsified asphalt mortar, ultimately leading to differences in crack resistance.

(2)

Increasing cement content can only improve the early tensile strength of the BE-CIR. Meanwhile, BE-CIR with higher cement content become stiffer and consequently less deformable, resulting in lower fracture energy. The initial moisture content mainly influenced the rate of water dissipation, but not the final residual moisture content. In addition, higher initial moisture content contributed to a more significant gradient characterization of the fracture properties due to the increase in residual water.

(3)

Elevating the temperature resulted in a considerable acceleration of water loss, leading to enhanced fracture properties and a reduction in the gradient index. In addition, during the initial stages of curing, high humidity prevented water dissipation from the mixture, which resulted in hindering the development of fracture resistance.

(4)

The findings indicated that the moisture migration process was governed by curing temperature and relative humidity, which were more significant than cement content and initial moisture content to the formation of cracking resistance performance. As for the gradient index of the fracture properties, the effect of curing temperature was the most significant, particularly after a curing period of 7 days.

In summary, it could be concluded that BE-CIR mixture subjected to different moisture environments will exhibit different cracking resistance developmental behaviors. Therefore, future research efforts should investigate the change of moisture in the BE-CIR layer directly, and figure out the impact of moisture on the microstructure and micromechanics of cement-emulsified asphalt mortar during curing to better understand the gradient development of performance.