Comparison was made between blended binders and the extracted and recovered (E&R) binder based upon results of rheological tests including the PG designations, rutting performance, fatigue life and characterization. They determine the set of blending conditions, if any, that results in a blended binder with the most rheological affinity to the E&R binder. Subsequently, the assessment of the rejuvenating efficiency is more realistic. The rheological tests’ results along with the comparisons are presented within this section.
4.1. Performance grade (PG)
Since the E&R asphalt binder was treated as short term (ST) or rolling thin film oven (RTFO) aged binder, the high temperature (HT) PG was determined only based upon G*/sin (δ) ≥ 2.2 kPa (RTFO criterion). As shown in Fig. 5(a), the HTe according to the unaged blended binder (G*/sin (δ) ≥ 1 kPa), namely HTe,unaged, decreases with blending severity increases; i.e. intense blending (IB) results in the lowest HTe,unaged while low blending (LB) gives the highest. In contrast, the trend of the HTe based upon RTFO-aged blended binder, namely HTe,st−aged, reverses and, with blending severity increases, there is increase in HTe,ST−aged; i.e. intense blending (IB) results in the highest HTe,unaged while low blending (LB) gives the lowest. Table 1 reveals that, with blending condition severity increase, the difference between HTe,unaged and HTe,ST−aged, namely ΔHTe, decreases; i.e. IB has the lowest and ΔHTe and LB has the highest. Rejuvenation using bio-rejuvenator is an intermolecular interactions between asphaltene dimers and amide groups in bio-rejuvenators (Zadshir et al., 2018; Behnood, 2019). Time, temperature and rotational speed are the effective factors of the interaction’s progress and completion (Behnood, 2019). ΔHTe can be an indicator showing the completion of the rejuvenation process for a set of blending conditions. It can be concluded that, since the blending conditions of LB including time and rotational speed are not long and fast enough, respectively, to complete the interaction of rejuvenation process, it continued during RTFO aging and, after RTFO aging, there is a rejuvenated binder that is evaluated based upon ST-aged criterion. As a result, ΔHTe is large; HTe,,unaged is high; HTe,ST−aged is very low; and the HT PG of LB is, subsequently, quite lower than IB and MB.
However, finding the blended asphalt binder with the most similarity to E&R binder is more important as it deals with reality of the blend inside the mixture. From Fig. 5(a), it is evident that medium blending (MB) has approximately the same G*/sin (δ) vs temperature curve as the E&R. Consequently, both MB and E&R have the same HT PG designation.
With respect to low temperatures criteria (stiffness and m-value from BBR test), all asphalt binders behave approximately the same (Fig. 5(b)). It indicates that, regardless of blending conditions, PAV aging results in almost the same thermal cracking resistance for all rejuvenated asphalt binders. This could mean that the PAV process eradicates the rejuvenation impact in all binders equally. Yet, it does not mean that all asphalt binders have equal aging potential. It is evident from Table 1 that the E&R has the lowest aging potential because, with respect to absolute value, its ΔTc is the lowest. ΔTc is the difference between stiffness- and m-value-based effective LT (LTe) and is an indicator of aging potential (Elkashef et al., 2017). Most aged asphalt binders have stiffness-based LTe smaller than the m-value-based LTe (Elkashef et al., 2017). So, LT PG is generally controlled by m-value and ΔTc is generally negative. It can be inferred that higher value of ΔTc shows higher ability to relax stresses and less aging potential. E&R is less prone to aging because the major aging probably takes place during extraction and recovery. Consequently, the E&R has already done its aging and it is less prone to further aging.
After determining the LT and HT PG, there is an intermediate temperature control (G*.sin (δ) ≤ 5000 kPa). As shown in Fig. 5 (c), all asphalt binders securely pass the criterion. The E&R’s temperature vs G*.sin (δ) curve is the same as the IB’s curve. While, with respect to performance grading, MB mimics ST-aged E&R and IB mimics PAV-aged E&R, LB manifests a very different behavior compared to E&R binder. Thus, IB or MB is recommended for PG determination as both yield the same PG as E&R. Also, since HT PG under any blending condition is controlled by RTFO-aged criteria, performing PG testing on the unaged blended binders can be considered optional.
Table 1
| IB | MB | LB | E&R | RAP binder |
HTe corresponding to unaged (HTe,unaged) (˚C) | 80.1 | 82.5 | 93.6 | NA | 96.6 |
HTe corresponding to ST-aged (HTe,ST−aged) (˚C) | 81.6 | 79.1 | 69.1 | 79.8 | 89.1 |
ΔHTe=HTe,unaged –HTe,ST−aged | -1.5 | 3.4 | 24.5 | NA | 7.5 |
LTe (˚C) | -22.17 | -24.11 | -22.29 | -23.94 | -16.2 |
PG | 76 − 22 | 76 − 22 | 64 − 22 | 76 − 22 | 88 − 16 |
ΔTc (˚C) | -4.61 | -3.57 | -3.2 | -0.07 | -6.2 |
4.2. Multiple stress creep and recovery (MSCR)
Based on PG designation, IB and MB show the closest behavior to E&R. But, due to the shortcomings of the PG grading system in addressing traffic impact on rutting potential, MSCR was adopted in PG+ (T Harman et al., 2011). Based upon non-recoverable creep compliance (Jnr) at 3.2 kPa stress level and the percent difference between Jnr at 3.2 kPa and 0.1 kPa stress levels, all asphalt binders have standard traffic designation letter “S”. From Fig. 6, it can be seen that the Jnr values of the E&R at 3.2 kPa stress level fall between the Jnr values of the IB and MB. The E&R has greater percent difference between Jnrs compared to the blended binders. Still, it can be inferred that IB, MB and E&R are in the same PG category, namely PG 76S-22, whereas LB is PG 64S-22. It corroborates the similarity of the IB and MB with the E&R, and the discrepancy between the LB and the E&R.
4.3. Linear amplitude sweep (LAS)
Fatigue analysis is essential for asphalt binders and mixtures containing very high RAP since fatigue performance is the primary concern of these types of asphalt mixtures (Sabouri et al., 2014). It is common to perform tension/compression (T/C) fatigue test at 10 ˚C in Canada for asphalt mixture fatigue performance (Perraton et al., 2010). But, as given Table 2, performing LAS at 10 ˚C gave G*s at the frequency of 10 Hz which are higher than the limit recommended by Teymourpour and Bahia (2014) (G*s greater than 60 MPa). Therefore, as recommended by Safaei and Castetorena (2016), LAS was performed at the average climatic PG temperature minus 4 ˚C as recommended by Safaei and Castetorena (2016). For PG 76S-22, it is 23 ˚C. The same was applied for LB (PG 64S-22) because of comparison purposes. LAS was performed on two replicates for each binder.
Figure 7(a) and Fig. 7(b) show the curves derived from LAS testing for fatigue life and damage growth, respectively. It can be observed that the binders have more or less the same damage characteristic curve, whereas fatigue life curves vary. If fatigue life is considered, there is differences between the extracted and recovered (E&R) binder and the blended ones; i.e. IB, MB and LB. The E&R shows higher fatigue life. However, the blended binders have intertwined fatigue lives over the strain levels (Fig. 7(b)). Taking the shear strain level of 5% as the one most representing the mixture (Tran et al., 2012), it can be observed that the fatigue lives of the binders blended under different blending conditions are close.
Figure 7 shows that damage and fatigue life have different results with regard to the fatigue performance of the binders. Damage is an intrinsic property of a material that is not impacted by the shear strain level. Here, fatigue life is an estimation based upon damage characteristics curves.
Considering the damage resistance (Fig. 7 (a)), LB has a very close integrity parameter (C) vs damage intensity curve to E&R’s one. However, all the binders have almost the same damage endurance. Thus, regardless the fatigue life, all binder blends can represent E&R with respect to fatigue performance. It can be inferred that the blending conditions do not maintain much of their influence on the binders after PAV aging. Perhaps, PAV totally releases the rejuvenation impacts on all binders and leaves them in a same non-rejuvenated state. Further studies are required on it.
Table 2
Binder | G* @ 10 ˚C and 10 Hz (MPa) |
IB | 79.3 |
MB | 67.0 |
LB | 74.3 |
E&R | 88.0 |
4.4. Linearity and complex shear modulus
The characterization of binders was conducted through linearity and complex shear modulus testing on the RTFO-aged binders. So, the results are limited to the short term (ST) aged state of the binders. LB is the softest asphalt binder with the highest LVE limit at both intermediate and high temperature (Fig. 8). Except the Cole-Cole curve, E&R curves always fall between IB and MB (Figs. 8 to 12). This complies with the results obtained from DSR testing on the ST-aged binder; i.e. G*/sin (δ) and Jnr. Cole-Cole curves indicate that the LB and MB have roughly the same curve, which is between E&R and IB.
As given in Table 3, parameters such as “h” and “k” of the 2S2P1D model are roughly the same for all binder types. “h” along with the static modulus (E00 = 0) shape the very left part of the Cole-Cole curves. This part represents the high temperatures behavior of binders. Figure 9 proves that all asphalt binders, regardless the blending conditions, behave more or less the same at high temperatures.
As the temperature decreases, the IB, E&R, MB and LB curves tend to diverge. This is due to the parameter “δ” and glassy modulus (E0). Comparing the Cole-Cole curves of the MB and LB indicates that the influence of the E0 is far greater than “δ”. Also, because of the constant value for the parameter “k”, there are almost constant distances between curves within the post-peak domain. Therefore, the origin of the difference among the Cole-Cole curves is the glassy modulus.
The complex shear modulus master curves at the reference temperature of 16 ˚C confirms that IB is the stiffest and LB is the softest asphalt binder type (Fig. 11). The results of the phase angle master curves comply with complex shear modulus one. As shown in Fig. 12, the difference between the curves can be as large as 10 ˚. Phase angle tends toward 90 ˚ and 0 ˚ with the frequency tending to infinitive low and high asymptote, respectively.
Table 3
| E00 (kPa) | E0 (kPa) | K | H | Δ | τE (s) | β |
E&R | 0 | 670000 | 0.30 | 0.62 | 3.50 | 0.00030 | 500 |
IB | 0 | 610000 | 0.30 | 0.62 | 4.00 | 0.00080 | 500 |
MB | 0 | 650000 | 0.29 | 0.62 | 3.50 | 0.00015 | 350 |
LB | 0 | 650000 | 0.29 | 0.63 | 3.20 | 0.00010 | 250 |