2.1 Definition and types of emulsifiers

An emulsifier refers to a substance that mixes two types of immiscible substances uniformly to make them stable. An emulsifier for asphalt serves to mix the asphalt binder, which is an oil component, and water. Emulsifiers are classified into oil-in-water (hereinafter O/W), water-in-oil (W/O) and water-in-oil-in-water (W/O/W) complexes according to the dispersion type ^{(Bakry et al. 2016)}. Fig. 1 shows a diagram for each type of emulsifier dispersion ^{(James 2006)}.

Fig. 1 Illustration of the emulsifier system (Bakry et al. 2016)

O/W takes the form of water, but when dispersed, it takes the form of oil. W/O takes the form of oil, but when dispersed, it takes the form of water. The dispersion type of every emulsifier currently used in the market is one of these three types, and the maker does not disclose which type is used. Emulsifiers can be classified into cationic, anionic, and nonionic according to the charge of the hydrophilic head. Table 1 shows the structure of emulsifiers by ionic type ^{(Evans and Wennerström 1999; James 2006; Khan et al. 2018)}. Cationic emulsifiers contain nitrogen in their head to generate cations in water, while anionic emulsifiers contain oxygen in their head to generate anions in water. Nonionic emulsifiers are electrically neutral because their heads are not charged, so they have free affinity with water and are less affected by electrolytes and pH. Generally, cationic and anionic emulsifiers, which enable asphalt mass production and are inexpensive, are used, but nonionic emulsifiers have been developed and used to secure solubility and stability without being affected by the charge and pH of asphalt and aggregates ^{(Lin 1968; Lin 1979; Takamura et al. 1979; Shinoda et al. 1980; Gullapalli and Sheth 1999)}.

2.2 Cold mix emulsified asphalt curing mechanism

Cold mix emulsified asphalt concrete exhibits its performance and strength through the mechanisms of breaking, flocculation, settlement, coalescence, and cures of the emulsified asphalt binder. Fig. 2 shows the curing mechanism of emulsified asphalt concrete in this process ^{(Im and Kim 2014; Khan et al. 2016)}. When the compaction of emulsified asphalt concrete starts, breaking begins when contacting with the flocculator on the surface of the aggregate by compaction energy. The breaking rate can be controlled with the type and content of the emulsifier. Cationic emulsifiers begin to be chemically broken when they come into contact with flocculators, while anionic emulsifiers are not chemically broken even when they come into contact with flocculators, but breaking occurs when water evaporates ^{(James 2006)}. In addition, the breaking of emulsified asphalt can be improved if the absorption rate of the aggregate is high or the pores of the particle size are large so that water is quickly removed ^{(Khan et al. 2016)}. When breaking of the emulsified asphalt begins, flocculation occurs, in which the asphalt particles in the emulsified asphalt aggregate together. In general, small asphalt particles surround large asphalt particles, and particles with increased weight are settled and coalesce into the aggregate. As shown in Fig. 2, if flocculation is deflected and early coalescence occurs in a state without sufficient breaking, the strength and durability of the mixture decrease because the adhesion to the aggregate decreases. Therefore, in order to increase the breaking rate and to reduce the flocculation and sedimentation speed, a stabilizer should be added, or the asphalt content should be adjusted. According to the curing speed of emulsified asphalts, the latter are divided into rapid setting, medium setting, and slow setting. This study used a slow setting type. The reason is that RAP has a larger contact angle with emulsified asphalt than virgin aggregate and a small contact area. Therefore, this was used to demonstrate the physical performance of emulsified asphalt concrete by cures in a state in which sufficient contact is made ^{(Wang et al. 2020; Konaté et al. 2021)}.

Fig. 2 Schematic illustrations of mechanism of emulsified asphalt(Im and Kim 2014;Khan et al. 2016)

3.1 Emulsified asphalt binder

In this study, 4 types of emulsifiers were used to compare and analyze the effect of types of emulsifiers on the physical performance of cold recycle emulsified asphalt concrete. All of the emulsifiers used were of a slow setting type, and two types of cationic emulsifiers (A and B), one anionic emulsifier (C), and one nonionic emulsifier (D) were prepared. Table 2 summarizes the properties of each emulsifier used. Different characteristics are shown depending on the type of emulsifier, but the color, viscosity measurement temperature, and number of amine groups present inside also differ. Therefore, they are expected to show different characteristics when manufactured with cold mix asphalt binder. The asphalt binder used to manufacture the emulsified asphalt binder was PG 64-22 (penetration level 60-80) grade. Table 3 summarizes the physical properties of the asphalt binder. The emulsified asphalt binder was prepared by first mixing water (H2O) heated to 60 °C, hydrochloric acid (HCl), and an emulsifier at the appropriate production ratio suggested by the supplier, and then stirring with the asphalt binder in a colloid mill. The colloid mill was used by DenimoTech’s BLM-P in Denmark, and agitation was performed at 160 °C for 90±10 seconds. Table 4 shows the physical properties of the 4 types of emulsified asphalt binders.

3.2 Aggregate

3.2.1 Recycled asphalt concrete aggregate and virgin aggregate

This study used recycled asphalt concrete collected from the ○○ road pavement overlay construction site in Seoul. Fig. 3 shows the RAP production process.

The asphalt binder was recovered by the method of KS F 2354 ^{(KATS 2018a)} and a penetration test of the asphalt binder was carried out. After obtaining the particle size with a sieving test in RAP, we adjusted the virgin aggregate and the mixing ratio according to the target particle size, and carried out the mixing design. Fig. 4 shows a picture of the RAP. After collection, physical and sieving tests were performed following natural drying in sunlight for more than 96 hours. As physical properties of coarse aggregate, density and abrasion ratio (KS F 2503, ^{KATS 2019)}, stability (KS F 2507, ^{KATS 2017a)} and elongated particle contents (KS F 2575, ^{KATS 2018b)} were evaluated. For the physical properties of fine aggregate, density and water absorption (KS F 2504, ^{KATS 2020)} grade shape test (KS F 2384, ^{KATS 2003)}, and passing finer \#200 (KS F 2511, ^{KATS 2017b)} were evaluated. The particle size analysis was performed by carrying out a sieving test at least 10 times after classifying the samples using the quartering method.

Fig. 3 Production process of cold recycle asphalt concrete mixture

Fig. 4 RAP aggregate

3.2.3 Physical properties of aggregate and particle size analysis results

Table 6 summarizes the physical properties of the RAP and new aggregates. The quality criteria for coarse and fine aggregates of the Ministry of Land, Infrastructure and Transport’s guidelines for asphalt concrete pavement construction (2017.04) were applied and all criteria were satisfied. Fig. 5 shows the sieving test results for the RAP. The BB-1CR particle size, which is the basic particle size standard of the Ministry of Land, Infrastructure and Transport’s guidelines for asphalt concrete pavement construction (2017.04), was used as the target particle size. The results of several sieving tests showed that coarse aggregate of 10 mm or more exceeded the upper limit, and fine aggregate of 5 mm or less also exceeded the lower limit.

Table 6 Physical properties of virgin aggregate and RAP

		Criteria	Aggregate
			Virgin	RAP
Coarse aggregate	Density (g/㎤)	≥2.5	2.623	2.539
	Absorption (%)	≤3.0	1.8	2.1
	Wear rate (%)	≤40 (for BB)	31.2	37.9
	Stability (%)	≤12	6.4	12
	Elongated particles (%)	≤10	5	8
Fine aggregate	Density (g/㎤)	-	2.642	2.564
	Absorption (%)	-	1.3	1.9
	Void content (%)	≥45	95	46.4
Passing finer #200 (%)		-		2.56

Fig. 5 Particle size distribution curve of RAP

3.3 Cement

Since cold recycle emulsified asphalt concrete mixture contains moisture, cement is generally used for rapid discharge ^{(Terrell and Wang 1971; Saadoon et al. 2018)}. A small amount of cement is used (about 1-2 %) to improve the curing rate, not to develop strength development ^{(Saadoon et al. 2017)}. If the amount of cement is increased by 2 % or more to manufacture cold mix emulsified asphalt, the strength development is changed to cement, so it is not called cold mix asphalt concrete, but cold mix stabilization treatment. Therefore, cold mix stabilization treatment is not classified as cold mix emulsified asphalt concrete. Table 7 shows the physical properties of cement. Type 1 Portland cement from S company was used.

3.4 Preparation of mix design

3.4.1 Particle size adjustment

Fig. 6 shows the synthetic particle size used in this study. The BB-1CR particle size is the base particle size of asphalt pavement, and the synthetic particle size was set according to the median of the upper and lower limits. Based on the results of the particle size adjustment, the ratio of virgin aggregate to RAP aggregate, 65:35, was suitable and this ratio was used.

Fig. 6 Particle size distribution used for mix design

3.4.2 Optimal water contents and coating ratio test

Since water is used in cold recycle emulsified asphalt concrete mixture, unlike in the general hot mix asphalt mixture design, it is necessary to determine the optimal water contents so that the emulsified asphalt binder is properly coated on the aggregate surface. This is because if the amount of moisture before mixing is not appropriate, the emulsified asphalt is combined with the powder of the aggregate and the coating is not even, resulting in deterioration of the asphalt pavement performance ^{(Ling et al. 2013)}. The optimal water contents was determined with the coating ratio test. Put 1 batch (about 1,200 g) of dry aggregate in a mixer, apply water as much as the dry weight thinly with a spray and mix until fully aborbed. After adding the emulsified asphalt binder and mixing with a mixer for 90 seconds±10 seconds, dry in the air using a fan until the prepared mixture becomes dry on the surface. The coating rate was calculated as a percentage by dividing the weight of the asphalt-coated aggregate and the total weight before mixing through visual inspection. The test is repeated while increasing the water by 1 % based on the dry aggregate weight. In this study, the coating rate test was performed by controlling the moisture to 2 % to 7 % of the aggregate weight and adjusting the content of the emulsified asphalt binder to 3 % to 7 %. The test results are presented in Figs. 7~10. The results showed that the coating ratio increased and then decreased as the moisture increased. It was also found that as the content of the emulsified asphalt binder increased, the coating ratio increased up to near the optimal water contents, and there was no constant trend thereafter. In addition, the test results showed that the coating ratio in the content of the emulsified asphalt binder was highest in the vicinity of the optimal water contents, except for B. According to the type of emulsified asphalt binder and the water contents of the aggregate, different coating ratios were observed. Therefore, they were found to have a close relationship with each other and to affect the asphalt coating ratio. In this study, the highest coating rate was adopted and the water content was determined as 4 % for emulsified asphalt A and 2 % for emulsified asphalt B, C, and D.

Fig. 7 Coating ratio of emulsified asphalt A

Fig. 8 Coating ratio of emulsified asphalt B

Fig. 9 Coating ratio of emulsified asphalt C

Fig. 10 Coating ratio of emulsified asphalt D

3.5 Test methods and criteria

In order to compare and evaluate the performances of the cold recycle emulsified asphalt concrete mixtures prepared with different emulsifiers, Marshall stability and flow tests, which are indexes of the strength of asphalt concrete, and air void and indirect tensile strength (ITS) tests, which are indicators of the crack resistance, were carried out according to the relevant KS regulations. The indirect tensile strength ratio (TSR), an index of water resistance, was tested by immersing the water-treated specimens in a constant temperature water bath at 25 °C for 24 hours without freezing and thawing. A DS test, which is an index of plastic deformation, was conducted at 60 °C with a wheel tracking tester. Table 9 shows the items and specification criteria for the tests performed.

4.1 Marshall stability and flow value test results

Fig. 11 compares the Marshall stability and flow values of the different specimens. All specimens satisfied the specification criteria, but in the case of cationic emulsified asphalt concrete specimens, the Marshall stability and flow value increased as the amine content increased. On the other hand, among the four types of emulsifiers, DC0, a sample using a nonionic emulsifier, showed the highest Marshall stability and flow value.

Fig. 11 Changes in Marshall stability according to the type of emulsified asphalt

Fig. 12 Comparison of marshall stability of specimens AC for cement contents

Fig. 13 Comparison of marshall stability of specimens BC for cement contents

Fig. 14 Comparison of marshall stability of specimens CC for cement contents

Fig. 15 Comparison of marshall stability of specimens DC for cement contents

Figs. 12~15 show the changes in the Marshall stability and flow value of each sample according to the different amounts of cement added. Regardless of the type of emulsified asphalt concrete, the addition of cement recorded a slight increase in Marshall stability. An increase in the Marshall stability of cold recycle emulsified asphalt concrete mixture according to the addition of cement was also found in previous studies ^{(Niazi and Jalili 2009; Liu et al. 2020)}. On the other hand, the change in flow value of each specimen was insignificant and there was no regularity, so no clear relationship between the cement contents and the flow value was evidenced.

4.2 Air void measurement results

Table 10 shows the air void for each specimen. The air void was found to be 11 % to 12 % in all specimens, indicating that the change in the type of emulsifier was not related to the air void.

4.3 ITS and TSR test results

Fig. 16 shows the ITS and TSR test results for each specimen type. The AC0 and BC0 specimens did not meet the current criteria of 0.4 MPa, and DC0 using a nonionic emulsifier was 0.44 MPa, showing the highest ITS among the four types of specimens. All the TSRs were found to be 0.7 or higher, but for specimens A and B, the ITS was below the specification criteria. Therefore, it can be said that the TSR evaluation was meaningless.

Figs. 17~20 show the changes in the ITS and TSR of each sample according to the cement contents. The ITS increased by adding cement, but the increase was insignificant. The D specimens using nonionic emulsifiers had the highest ITS, and the ITS measured in air was 0.4 MPa or more, which meets the current criteria, but wet specimens were significantly affected by moisture. The ITS of specimen C using anionic emulsifier was found to be improved compared to that of specimens using cationic emulsifier.

Fig. 16 Changes in ITS & TSR to the type of emulsified asphalt

Fig. 17 Comparison of ITS & TSR of specimens AC for cement contents

Fig. 18 Comparison of ITS & TSR of specimens BC for cement contents

Fig. 19 Comparison of ITS & TSR of specimens CC for cement contents

Fig. 20 Comparison of ITS & TSR of specimens DC for cement contents

4.4 DS test results

Fig. 21 compares the DS values of the specimens using different emulsifier types. The DS of the DC0 specimen was the highest, but the difference was insignificant. Figs. 22~25 show the DS test results for each specimen according to the cement contents. In all specimens, it was found that the DS increased when the cement contents were increased, regardless of the type of emulsifier. This result is similar to that of previous research ^{(Li et al. 2019)}.

Fig. 21 Changes in dynamic stability to the types of emulsified asphalt concrete

Fig. 22 Comparison of dynamic stability of specimens AC for cement contents

Fig. 23 Comparison of dynamic stability of specimens BC for cement contents

Fig. 24 Comparison of dynamic stability of specimens CC for cement contents

Fig. 25 Comparison of dynamic stability of specimens DC for cement contents