Keywords

1 Introduction

Because of the great both economic and environmental benefits, the use of Reclaimed Asphalt (RA) has gained popularity. However, the stiff binder released by RA aggregates, especially when modified RA is considered, restricts its maximum amount allowed in Hot Mix Asphalt (HMA) due to the risk of mixtures with a reduced workability and more prone to some pavement distresses [1]. In this sense, Warm Mix Asphalt (WMA) technology can represent an effective technical solution for avoiding the downsides of high RA content. Indeed, the reduction of production temperatures by 20–40 °C induces lower bitumen aging, thus mitigating the stiffening effect of the aged bitumen from RA and not penalizing the workability of the mixture [2]. Despite this benefit, additional concerns are related to the presence of Polymer Modified Bitumens (PMB) especially if both virgin and RA bitumens are polymer modified. Indeed, possible interactions between virgin and aged polymer chains as well as WMA additives may modify the rheological behavior of binder phase compromising the performance of the mixtures [3,4,5].

Given this background, the objective of this research is to evaluate the feasibility of using percentages of modified RA up to 25%, in WMA mixtures prepared with a modified bitumen. A laboratory mechanical characterization, also carried out on a conventional recycled HMA, allowed the potentialities of WMA technology combined to recycling techniques to be evaluated.

2 Experimental Program

Virgin (limestone, basalt, calcareous filler) and modified RA aggregates were used to produce the aggregate blends. The bitumen content of RA (Dmax = 5 mm) was 8.1% by aggregate weight and the recovered bitumen had a penetration of 21 dmm and a softening point of 81 °C. An SBS polymer modified bitumen classified as PmB 45/80-70 was selected as the binder. A chemical WMA additive available on the market, having an alkyl-amide polyamines-based composition and characterized by a density at 25 °C of 0.94 g/cm3, a flash point higher than 150 °C and a viscosity at 25 °C of 150 cP, was chosen.

Three dense-graded bituminous mixtures for wearing course were investigated: two WMAs and one HMA mixture as reference. The grading curves of the three asphalt mixtures were different and designed according to the Italian technical specifications for road construction [6]. The mixtures were coded as:

  • WMA-20-3.7, containing 20% RA, 3.7% added binder and 5.1% total binder;

  • WMA-25-3.7, containing 25% RA, 3.7% added binder and 5.5% total binder;

  • HMA-15-4.6, containing 15% RA, 4.6% added binder and 5.6% total binder.

The WMA mixtures were mixed and compacted at 130 °C and 120 °C respectively, whereas the HMA mixture were mixed and compacted at 170 °C and 160 °C, respectively.

For each mixture a series of cylindrical specimens (D = 100 mm) were prepared by means of a gyratory compactor to the same target air voids content (equal to 5%). Specimens were then subjected to a mechanical characterization in terms of Indirect Tensile Strength (ITS), water sensitivity (Indirect Tensile Strength Ratio—ITSR), stiffness (Dynamic Modulus |E*| and Indirect Tensile Stiffness Modulus—ITSM), fatigue (Indirect Tensile Fatigue Test—ITFT) and rutting resistance.

3 Results and Discussions

3.1 Indirect Tensile Strength (ITS) Test and Water Sensitivity

ITS test was performed at 25 °C, according to EN 12697-23. Six replicates were performed for each mixture, three in dry condition and three in wet condition, where the conditioning was performed according to EN 12697-12 (Method A). The Indirect Tensile Strength Ratio (ITSR) value was calculated as ratio between the ITS values in wet and dry conditions. Figure 1 shows the average ITS and ITSR values for all the investigated mixtures. In both dry and wet conditions, the ITS values of WMA mixtures are higher than the reference HMA and satisfied the local technical specification requirements (0.95 < ITS < 1.70). However, WMAs show a slightly lower moisture resistance, especially for higher RA contents.

Fig. 1
figure 1

ITS and ITSR results

Full size image

3.2 Stiffness

Dynamic complex modulus was measured to evaluate the stiffness of mixtures, at 4 temperatures (from 5 to 50 °C) and 6 frequencies (from 0.1 to 20 Hz) by means of Asphalt Mixture Performance Tester (AMPT). Two replicates were performed for each mixture.

The Huet–Sayegh (HS) model [7] was used to fit the frequency dependency of the dynamic modulus |E*| and the Williams-Landel-Ferry (WLF) equation [8] was used to model the shift factor trend with temperature. The dynamic modulus master curves of all mixtures at Tref = 20 °C are shown in Fig. 2a and the parameters of HS (Eg, Ee, k, h, δ, t) and WLF (C1 and C2) models are listed in Table 1. It can be observed that both WMA mixtures showed lower dynamic moduli with respect to the reference HMA, over the entire range of loading frequency. This is mainly imputable to the less oxidation effects that WMA mixtures experienced during the lower temperature production process. Therefore, compared to the reference HMA mixture, the WMAs showed a significantly reduced stiffening effect mainly due to aging. Moreover, both WMA mixtures showed a similar trend for the master curves, highlighting the prevalent role of aging rather than RA content on the development of stiffness properties.

Table 1 HS and WLF model parameters
Full size table

Stiffness properties were also measured by the ITSM at 20 °C, according to EN 12697-26. Figure 2b reports the comparison between the ITSM values and the dynamic moduli identified in the master curves at the same temperature (20 °C) and frequency (2 Hz–comparable with the rise time of 124 ms) of the ITSM tests. Figure 2b shows that ITSM results are consistent with the dynamic modulus data obtained through uniaxial compression test, confirming lower stiffness of both WMA mixtures compared to the reference mixture.

Fig. 2
figure 2

Dynamic modulus master curves of mixtures at Tref = 20 °C (a), ITSM versus |E*| @ 2 Hz at 20 °C (b)

Full size image

3.3 Rutting Performance

Rutting resistance of mixtures was determined by performing cyclic compression tests at 60 °C, according to EN 12697-25. Two replicates were performed for each mixture. A cyclic axial stress with an amplitude of 0.3 MPa and a steady confining pressure of 0.2 MPa were adopted during the entire test (10,000 load cycles). The cumulative axial strain of the specimen was recorded and the creep curve was built.

Table 2 The rutting parameter results
Full size table

The creep rate fc (i.e. slope of the least square linear fit of the quasi linear part of the creep curve) and the permanent deformation after 1000 load cycles ε1000 (i.e. deformation calculated from the least square power fit of the quasi linear part at 1000th cycle) were used to characterize the rutting potential of the mixtures. Specifically, the lower the value of these parameters, the higher the rutting resistance of the mixture.

The average values of the rutting parameters of WMA mixtures, shown in Table 2, are remarkably lower than those of the reference mixture, indicating that WMA mixtures exhibited higher rutting resistance. These results highlight that despite the mitigated stiffness due to the reduced production temperature, the WMA mixtures would not experience any permanent deformation accumulation risk in the early period of their service life.

3.4 Fatigue Performance

ITFT was carried out at 20 °C according to the British standard BS DD ABF (1997). During the test, the number of loading cycles was recorded until the complete failure of the specimen was obtained.

Three initial strain levels (100, 125 and 150 με) were selected for each mixture and the corresponding stresses to be applied were calculated by using the ITSM values (Fig. 2b) measured before the fatigue tests. Two specimens were tested for each strain level. The ITFT results are shown in Fig. 3, where the initial horizontal strain \({\upvarepsilon }_{\mathrm{h},\mathrm{ initial}}\) is plotted as a function of the number of loading cycles to failure Nf. Test data were analyzed according to a power law regression and the fatigue curve (i.e. \({\upvarepsilon }_{{\text{h}}} = {\text{a}} \cdot {\text{N}}_{{\text{f}}}^{{ - {\text{b}}}}\)) was reported for each mixture. The obtained high R2 values (> 0.95) for all mixtures indicate good reliability to predict fatigue life by using this fitting model. From the analysis of fatigue curves, it can be noted that, at high and medium strain levels, WMA mixtures can withstand higher numbers of loading cycles than the reference one, highlighting a better fatigue behavior for WMA mixtures; whereas, at low strain levels, comparable behaviour is expected between WMAs and HMA. These results are consistent with the stiffness analysis, which suggests that the higher stiffness of the reference mixture could result in a more brittle tendency. Moreover, the comparison between WMA mixtures shows that a higher RA content results in a slightly higher fatigue resistance, especially at higher strain level.

Fig. 3
figure 3

Fatigue curve of mixtures at 20 °C

Full size image

4 Conclusions

This paper focused on the performance analysis of dense graded recycled WMA mixtures for wearing courses produced with different modified RA amounts (20 and 25%) and an SBS modified bitumen. The mixtures were analyzed in terms of strength and stiffness properties, moisture susceptibility, rutting and fatigue behavior by testing gyratory compacted specimens.

Compared to a conventional recycled HMA, the investigated WMA mixtures showed:

  • higher strength properties in both dry and wet conditions but a higher moisture susceptibility that tends to increase with the increase in RA content;

  • lower stiffness properties and higher fatigue resistance. This finding highlights that lower production temperatures (i.e. WMA technologies) reduce short-term aging effects leading to stiffness mitigation as well as improved fatigue performance. The increase in RA content did not affect the stiffness of WMAs but allowed the achievement of higher fatigue resistance;

  • significantly higher rutting resistance despite lower stiffness properties. Besides, no RA content effect on the rutting behavior was observed.

Overall, the outcomes of this experimental investigation strongly supported the practical use of percentages of modified RA, up to 25%, in WMA mixtures prepared with a modified bitumen.