Road Materials and Pavement Design

ISSN: (Print) (Online) Journal homepage: https://www.tandfonline.com/loi/trmp20

Behaviour of binder blends: experimental results

and modelling from LVE properties of pure binder,
RAP binder and rejuvenator

Andrei Forton, Salvatore Mangiafico, Cedric Sauzéat, Hervé Di Benedetto &

Paul Marc

To cite this article: Andrei Forton, Salvatore Mangiafico, Cedric Sauzéat, Hervé Di Benedetto
& Paul Marc (2021): Behaviour of binder blends: experimental results and modelling from LVE
properties of pure binder, RAP binder and rejuvenator, Road Materials and Pavement Design, DOI:
10.1080/14680629.2021.1905699

To link to this article: https://doi.org/10.1080/14680629.2021.1905699

Published online: 30 Mar 2021.

Submit your article to this journal

Article views: 12

View related articles

View Crossmark data

Full Terms & Conditions of access and use can be found at

https://www.tandfonline.com/action/journalInformation?journalCode=trmp20
ROAD MATERIALS AND PAVEMENT DESIGN
https://doi.org/10.1080/14680629.2021.1905699

Behaviour of binder blends: experimental results and modelling

from LVE properties of pure binder, RAP binder and rejuvenator
Andrei Fortona,b , Salvatore Mangiaficoa , Cedric Sauzéata , Hervé Di Benedettoa and
Paul Marcb
a Université de Lyon/ENTPE, LTDS (UMR CNRS 5513) Vaulx-en-Velin Cedex, France; b Department of Overland
Communication Ways, Foundation and Cadastral Survey, University Politehnica Timisoara Timisoara, Romania

ABSTRACT ARTICLE HISTORY

The first objective of this study is to highlight the effect of the rejuvenator Received 28 July 2020
on the linear viscoelastic (LVE) properties of various blends of a fresh 50/70 Accepted 12 March 2021
binder and a reclaimed asphalt pavement (RAP) binder, with a rejuvena- KEYWORDS
tor (a mixture of vegetal oils). A total of 17 binders (including pure fresh RAP binder; rejuvenator;
and RAP binders) produced as blends of two base binders and a rejuvena- binder blending; LVE;
tor in various proportions were tested. Dynamic shear rheometer tests were 2S2P1D model; estimations
performed, and the experimental results were modelled using the 2S2P1D
model. The second main objective was to present and analyse two different
approaches for estimation of 2S2P1D parameters and temperature shift fac-
tors of binder blends, based on reduced experimental data (only data for the
base constituents). Good correlation was found between the direct deter-
mination and the estimated results. However, a more accurate estimation
was performed with the second approach.

Introduction
The reuse of reclaimed asphalt pavement (RAP) in the production of new bituminous mixtures has
become over the years a common strategy with a double benefit: producing more economical and
eco-friendlier road materials (Moon et al., 2014; Tapsoba et al., 2014; Hong & Prozzi, 2018; Mazzoni
et al., 2018).
Besides these two major benefits, it was shown that the use of RAP materials within bituminous
mixtures can provide also a beneficial impact on the mechanical behaviour of the final mixture (Visin-
tine et al., 2013; Manke et al., 2019; Kaseer et al., 2019, among others). However, when high amounts of
RAP (more than 20%) are used in the production of a new bituminous mixture, an increase of stiffness
was observed due to the aged RAP binder (Li et al., 2008; Walther et al., 2019).
By investigating the interaction between a virgin binder and a RAP binder, Noferini et al. (2017)
showed that the rheological properties and the viscosity of the binder blends are little influenced when
less than 10% of RAP material is used in the asphalt mixtures. Therefore, the effect of the RAP binder
is significant on the properties of binder blends when more than 10% RAP material are used in the
production of new asphalt mixture. In this case, recycling agents or rejuvenators are frequently used
in order to increase the amounts of RAP and to assure a good workability and satisfactory mechanical
properties of the final mix.
The potential of different recycling agents to rejuvenate the aged RAP binder was highlighted in
various studies by evaluating the rheological or conventional properties, of various binder blends and

CONTACT Andrei Forton andrei.forton@student.upt.ro

© 2021 Informa UK Limited, trading as Taylor & Francis Group

2 A. FORTON ET AL.

also by investigating the effect of rejuvenators on the thermomechanical performances of bituminous

mixtures (produced with RAP material). Mogawer et al. (2013) showed that rejuvenators can mitigate
the stiffness of the rejuvenated binder blends and the cracking performances of the corresponding
bituminous mixtures were improved. However, when the rejuvenator is not used in a proper dosage,
it can lead to an adversely impact on the moisture susceptibility and the rutting performances of
mixtures.
Zaumanis et al. (2014) used the Superpave PG tests in order to determine the optimum dosage of six
rejuvenators and highlighted that a proper amount of rejuvenators can reduce the performance grade
of a rejuvenated binder blend to the level of a fresh binder, can ensure excellent rutting resistance,
good cracking performances and can provide a longer fatigue life of 100% recycled asphalt mixtures.
Therefore, the type and the optimum dosage of rejuvenators play an important role in the behaviour
of binder blends and bituminous mixtures produced with RAP binder/material.
Nowadays many products of vegetal origin are used as rejuvenators. Mangiafico et al. (2017)
showed that the addition of a recycling agent of vegetal origin into recycled bituminous mixtures
can reduce the norm of complex modulus and when more than 40% of RAP material were used, an
improvement on the fatigue performances was observed.
In two previous studies, Forton et al. (2020, 2021) showed the potential of a rejuvenator of veg-
etal origin to contabalance the effect of an RAP binder, in terms of mechanical characteristics, and
finally to provide some rejuvenated blends that may have a similar penteration grade, ring and ball
temperature, Fraass temperature and similar steady shear viscosity as the one of the considered fresh
binder.
Menapace et al. (2018) and Oldham et al. (2018) cocluded that the reduction of stiffness should
not be used as a single indicator of the rejuvenation effect of a recycling agent. A true rejuvenator
should lead to a restoration of both stiffness and morphology of the final binder blends. Therefore,
the selection of a rejuvenator and its proper dosage based on only one direction could not meet all
the specifications of a fresh virgin binder.
Many theoretical/analogical linear viscoelastic (LVE) models were proposed over the years in order
to perform a mathematical approximation of the real behaviour of the tested materials. Olard and Di
Benedetto (2003) developed a unique rheological model 2S2P1D model (abbreviation of the combi-
nations of two springs, two parabolic creep elements and one dashpot) that can be used in order to
model and to predict the LVE behaviour of both binders and bituminous mixtures over a wide range
of temperatures and frequencies.
The characterization of the LVE behaviour of bituminous materials can be very useful in order to
highlight the effects of the addition of an RAP material and a rejuvenator.
Mangiafico et al. (2014) proposed an estimation method for the 2S2P1D constants and temperature
shift factors for binder blends produced by mixing one type of RAP binder and different types of fresh
binders. A good correlation between the original and the estimated values of 2S2P1D constants was
obtained.
This study has two main objectives. Firstly, to investigate the effect of a rejuvenator on the
LVE properties of various blends of a fresh binder (50/70 penetration grade) and an RAP-extracted
binder, with and without rejuvenator (a mixture of vegetal oils). A total of 17 binders were pro-
duced, including pure fresh and RAP binders and their rheological behaviour was characterised in
terms of complex shear modulus by means of dynamic shear rheometer (DSR) tests. The rheological
properties of all binders were described by using the 2S2P1D model developed at ENTPE labora-
tory (Olard & Di Benedetto, 2003; Di Benedetto et al., 2004) which was chosen due to its ability to
satisfactorily describe the LVE properties of binders. The second main objective of this work was
to present and analyse two different approaches for estimation of 2S2P1D parameters and temper-
ature shift factors of binder blends, based on reduced experimental data (data only for the base
constituents). Finding a proper estimation method could allow to determine 2S2P1D parameters
and temperature shift factors of binder blends which do not need to be prepared and tested in the
laboratory.
 ROAD MATERIALS AND PAVEMENT DESIGN 3

Figure 1. Proportion of the three base components in tested binders: pure binder (50/70), RAP binder (RAP) and rejuvenator (Rej).

Table 1. Properties of the two base binders and the density of the rejuvenator.
Properties Fresh binder 50/70 RAP binder Rejuvenator
Density at 25°C (kg/m3 ) 1016 1060 930
Ring and ball temperature (°C) 48.60 79.00 –
Fraass temperature (°C) −14 −2 –
Penetration (0.1 mm) 54 11 –

Materials and methods

Materials
Several blends were produced by mixing three base materials: a pure fresh binder (a straight-run 50/70
penetration grade binder), an RAP binder (extracted from a well-known RAP material according to the
EN 12697-1, with a 4.0% binder content measured trough extraction and recovery according to EN
12697-3) and a rejuvenator (a mixture of vegetal oils). All materials are commonly available in Romania.
The experimental plan includes a total of 17 binders (Figure 1): pure fresh and RAP binders,
3blends of RAP binder and the rejuvenator and 12 blends of fresh and RAP binders with and without
rejuvenator.
The rejuvenator contents used for blends were 0%, 5%, 10% and 15% by mass of the RAP binder,
corresponding to 0.0%, 0.2%, 0.4% and 0.6% by mass of RAP material.
Blending proportions between these three base components were calculated in order to reproduce
real ratios within corresponding bituminous mixtures (with a 5.6% total binder content) containing
25%, 50%, 75% and 100% RAP material. The detailed proportions for all tested binders are shown in
Figure 1.
The European conventional properties of the same binder blends were investigated in Forton
et al. (2020). Therefore, Table 1 summarises the conventional properties (penetration, softening point,
Fraass breaking point) of the two base binders together with the density of the rejuvenator.
Blends were named after the penetration grade of the fresh binder, the amount of RAP material
used to produce the corresponding bituminous mixture and the rejuvenator dosage by mass of RAP
binder.
The same binder blends were used in a previous article (Forton et al., 2020, 2021) where the con-
ventional properties and the steady shear viscosity of these materials were investigated. Also, in these
4 A. FORTON ET AL.

previous studies, the production process of all binder blends was well described. However, a brief
presentation of this process is described in the following lines.
The studied binder blends were produced following the same procedure which was chosen in order
to reproduce in the laboratory the actual production process of bituminous mixtures containing RAP
material and rejuvenator. Therefore, all blends were produced in two steps:

− Step 1: the 50/70 fresh binder was heated in a thermostatic oven at 160°C (chosen with respect
to the Romanian technical specifications for the production of bituminous mixtures) for 30 min.
Then the rejuvenator at ambient temperature was added to the virgin binder. These preliminary
blends were mixed for 5 min and then again heated for 20 min at the same temperature;
− Step 2: the RAP binder, which was previously heated at the same temperature for the same time
as the fresh binder, was added to the preliminary blends (previous step) and these blends were
mixed for another 5 min in order to produce a homogenous final binder blends.

Rheological method
For all binders, complex modulus tests were performed by using a DSR apparatus with a 25 mm diame-
ter plate-plate configuration (1 mm gap), at 7 temperatures (from 25°C to 85°C with a 10°C increment)
and at 10 frequencies (from 0.1 to 10 Hz). All tests were performed in strain control mode with a 5%
targeted shear strain amplitude. However, the torque limit of the instrument was reached for strain
amplitudes lower than the 5% targeted value during some tests at low temperatures and high fre-
quencies (Forton et al., 2020). Therefore, these tests were performed at variable strain amplitudes. All
data points obtained at strain amplitudes lower than the targeted 5% shear strain amplitude are plot-
ted with crosses (×) and were not considered in the rest of the analysis. It must be mentioned that at
temperatures higher than 45°C, the tests were carried out for sure in the LVE domain. Since the LVE limit
was not explicitly studied in this work, maybe at low temperature ( < 35°C), the tests were performed
at the limit of the LVE domain. A wider temperature and frequency ranges should be considered in a
future work and tests should be performed by using both 25 mm/8 mm plate-plate configurations.
The time–temperature superposition principle was applied to obtain temperature shift factors aT
(Di Benedetto et al., 2007; Nguyen et al., 2013; Mangiafico et al., 2014, 2019a). Firstly, the shift factors
were obtained free and then the Williams–Landel–Ferry (WLF) equation was used (data shown in the
next section) to approximate these values, in order to verify if the free fitting is close to the WLF fitting
(data shown in Figure 4(b)). Master curves were derived from experimental results at the reference
temperature Tref = 55 ◦ C according to Equations (1) and (2):
|G∗ (T, f )| = |G∗ (Tref , aT · f )| (1)
ϕ(T, f ) = ϕ(Tref , aT · f ) (2)
where T and f are, respectively, the test temperature and frequency and Tref is the reference temper-
ature. Values of aT shift factors for all tested binders are reported in Table A1 (Appendix).

Rheological modelling method

Experimental results were analysed using 2S2P1D model (Di Benedetto et al., 2004, 2007; Olard & Di
Benedetto, 2003; Mangiafico et al., 2014, 2017, 2019a), developed at ENTPE, which consists of two elas-
tic (spring) elements, two parabolic creep elements and one viscous (dashpot) element. Each element
is associated with a corresponding parameter, as in Equation (3).
G∞ − G 0
G∗ (iωτ ) = G0 + −k
(3)
1 + δ(iωτ ) + (iωτ )−h + (iωβτ )−1
where ω is the pulsation; τ is the characteristic time, a function of temperature and accounting for
time–temperature superposition principle as in Equation (4); G∞ (glassy modulus) is the value of the
 ROAD MATERIALS AND PAVEMENT DESIGN 5

Figure 2. (a) Schematic representation of 2S2P1D model; (b) 2S2P1D parameters meaning on a Cole–Cole curve of a bitumen.

complex shear modulus when ωτ → ∞; G0 (static modulus) is the value of the complex shear modulus
when ωτ → 0, generally equal to 0 in the case of pure binders; δ, k and h are dimensionless constants
and β is a dimensionless parameter related to Newtonian viscosity η, Equation (5).
τ
aT = (4)
τ0

where τ0 is the characteristic time at reference temperature Tref and aT is the shift factor, varying with
temperature.
η = (G∞ − G0 )βτ (5)
Moreover, Figure 2 shows the schematic representation of the 2S2P1D model regarding the con-
stitutive elements and the associated parameters (Figure 2(a)) and also, the meaning of each 2S2P1D
parameter on a general Cole–Cole curve of a binder (Figure 2(b)).
Values of aT can be approximated by the WLF Equation (6) (Di Benedetto et al., 2007; Ferry, 1980;
Nguyen et al., 2013; Mangiafico et al., 2014):

−C1 (T − Tref )
log(aT ) = (6)
C2 + (T − Tref )

where C1 and C2 are empirical constants depending on the material and Tref .
All experimental results were modelled using the 2S2P1D model and the WLF equation: master
curves at Tref = 55 ◦ C, Black and WLF curves for all the tested binders.

Results
In order to investigate the influence of RAP binder and rejuvenator content on the LVE behaviour of the
blends, master curves of norm |G∗ | and phase angle ϕ of complex shear modulus at a reference temper-
ature of 55°C and Black diagrams were built for all blends. In this work, for the phase angle the symbol
ϕ was used instead δ which is usually used for the phase angle obtained from DSR measurements due
to the fact that the symbol δ was used for a 2S2P1D parameter.
As an example, master curves of |G∗ | and ϕ (Figure 3(a)), temperature shift factors vs. temperature
(Figure 3(b)) and Black diagrams (Figure 3(c)) of blends of RAP binder and rejuvenator are shown in
Figure 3.
As expected, master curves, Black curves and aT shift factors of the blends produced with rejuvena-
tor fall between corresponding master curves of fresh 50/70 binder and RAP binder. On the other side,
the increase of the rejuvenator content into the blends corresponds to master curves, Black curves and
shift factors progressively approaching those of the fresh binder.
From our results, the blends produced with fresh and RAP binders and 10% Rej by mass of the
RAP binder presents close properties to those of the fresh 50/70 binder, independently of the RAP
6 A. FORTON ET AL.

Figure 3. Complex shear modulus test results for the blends produced with RAP binder and rejuvenator: (a) master curves of the
norm of complex shear modulus and phase angle at a reference temperature of 55°C; (b) temperature shift factors; (c) Black diagram.

binder content (master curves and aT shift factors are significantly close to those of the fresh 50/70
binder). However, these blends produced with 10% Rej are stiffer and they present much lower phase
angle values compared to the fresh binder used in this study. This trend indicates that this rejuvena-
tor is capable to counterbalance the effect of the hard-aged RAP binder into the blends, in term of
mechanical characteristics.
Moreover, the blends produced with fresh and RAP binders and 15% Rej by mass of the RAP
binder show a behaviour which may correspond to a ‘softer binder’ than the fresh binder used in
this study. In Figure 3, it can be observed that the blend RAP + 15% Rej presents a behaviour that
corresponds much more with the data obtained for the 50/70 binder. This trend can be explained
by the fact that the ‘optimal dosage’ of the rejuvenator is probably closer to 15% Rej by mass of
the RAP binder (corresponds to 0.6% rejuvenator by mass of the RAP material for the correspond-
ing bituminous mixture) in order to obtain a blend who presents a similar behaviour as a fresh 50/70
penetration grade binder. By increasing the rejuvenator content above this ‘optimal percentage’
softer binders than the 50/70 penetration grade binder can be obtained. However, in order to deter-
mine the optimal dosage of a rejuvenator, this process must not be performed in only one direction
due to the fact that this dosage may not meet all the specifications of a fresh virgin binder. There-
fore, DSR tests must be combined with other tests at low temperature and also with a chemical
investigation of rejuvenated binder blends in order to determine the proper optimal dosage of the
rejuvenator.
 ROAD MATERIALS AND PAVEMENT DESIGN 7

Figure 4. 2S2P1D model and WLF equation fitting of DSR complex shear modulus test results for the fresh 50/70 and RAP binders:
(a) master curves of phase angle and norm of the complex shear modulus (Tref = 55◦ C); (b) WLF curves; (c) Black curves.

Taking into consideration that all binder blends were produced by mixing two base binders (fresh
and RAP binders) and a rejuvenator, first the experimental results obtained for fresh and RAP binders
were modelled (Figure 3).
All binders were tested at intermediate and high temperatures, and by plotting the obtained data
in a Cole–Cole representation, it was observed that the experimental data fit only on the first part of
the curve. By fitting the 2S2P1D model on the data obtained for the two base binders, it was observed
that the glassy modulus G∞ and the parameter k which are characterising the ‘final part’ of the 2S2P1D
model Cole–Cole plot do not show an important variation. A similar comment can be done also for
the shape parameter δ. Therefore, in this work, we imposed and assumed that G0 , G∞ , kandδ have the
same constant values for all binders. It must be mentioned that probably this assumption could not
be valid if a larger temperature range will be considered (DSR tests performed at lower temperatures).
Therefore, further investigation must be done.
For this reason, G0 , G∞ , k and δ were imposed constant for all tested binders. The values of these
constant parameters were established according to the 2S2P1D fitting model of fresh 50/70 and RAP
binders (G0 = 0 MPa, G∞ = 980 MPa, k = 0.25 and δ = 4.3). Figure 4(a–c) shows the 2S2P1D model
and WLF equation fitting for the base binders (pure fresh 50/70 and RAP binders). The experimental
results obtained for all binders were modelled using the same proposed approach.
Values of 2S2P1D constants (h, β and τ ) for all tested blends are reported in Table 2 and in Figure
5(a–c) as a function of RAP binder content.
As expected, h, β and τ show a remarkable tendency with the increase of RAP binder and rejuve-
nator contents in the final blends, as shown in Figure 5. Similar observation can be done also for the
8 A. FORTON ET AL.

Figure 5. 2S2P1D constants and shift factors for all tested binders: (a) h tendency with RAP binder content; (b) β tendency with
RAP binder content; (c) τ tendency with RAP binder content; (d) aT at 65°C for Tref = 55◦ C tendency with RAP binder content.

shift factors tendency with the increase of both RAP and rejuvenator amounts (Figure 5(d) shows an
example regarding the aT shift factors obtained for all tested binders corresponding to the 65°C for
Tref = 55◦ C). Linear regressions were performed for values of h, log β, log τ and log aT of blends having
the same rejuvenator content (0%, 5%, 10% and 15%).
As it can be observed in Figure 5(a), constant h shows a linear relationship with the RAP binder
content. For the blends produced without rejuvenator, the increase of RAP binder content corre-
sponds to a decrease of constant h values (tending towards values of RAP binder). On the other side,
the increase of rejuvenator content corresponds to an increase of h values. For the binder blends
produced with 10% Rej, a lower value of the R2 was obtained (0.813), value that is not meaningful
since the h values of all these blends remain approximately constant independently of the RAP binder
content.
A similar remark is valid for constants β, τ and aT temperature shift factors (at 65°C for Tref = 55◦ C)
in a logarithmic scale (Figure 5(b–d)). Values of β and τ are increasing with the increase of RAP binder
content and on the contrary, they are decreasing with the increase of rejuvenator content. An inverse
tendency can be observed for the aT shift factors at 65°C for Tref = 55◦ C.
 ROAD MATERIALS AND PAVEMENT DESIGN 9

Table 2. 2S2P1D constants for all tested binders.

G0 G∞ k δ h β τ
Binders (MPa) (MPa) (−) (−) (−) (−) (s)
50/70 0.633 124 4.60 × 10−9
RAP 0.596 630 8.70 × 10−8
RAP + 5% Rej 0.614 360 1.99 × 10−8
RAP + 10% Rej 0.634 198 6.53 × 10−9
RAP + 15% Rej 0.645 135 2.65 × 10−9
50/70 + 25% RAP 0.625 166 7.90 × 10−9
50/70 + 25% RAP + 5% Rej 0.629 158 6.42 × 10−9
50/70 + 25% RAP + 10% Rej 0.633 132 4.60 × 10−9

0.25

4.30
980
50/70 + 25% RAP + 15% Rej 0 0.636 120 3.91 × 10−9
50/70 + 50% RAP 0.620 235 1.22 × 10−8
50/70 + 50% RAP + 5% Rej 0.627 195 7.20 × 10−9
50/70 + 50% RAP + 10% Rej 0.632 145 5.00 × 10−9
50/70 + 50% RAP + 15% Rej 0.637 124 3.50 × 10−9
50/70 + 75% RAP 0.614 310 2.28 × 10−8
50/70 + 75% RAP + 5% Rej 0.623 238 1.08 × 10−8
50/70 + 75% RAP + 10% Rej 0.631 166 5.75 × 10−9
50/70 + 75% RAP + 15% Rej 0.638 125 3.53 × 10−9

For simplicity in the following section the values of 2S2P1D constants (h, β and τ ) and values of aT
shift factors determined for all the blends by fitting the 2S2P1D model on the experimental results, as
described so far, are referred to as ‘direct determination’.

Estimation of 2S2P1D constants and temperature shift factors

In accordance with the observed tendencies, two different approaches were proposed in order to
estimate 2S2P1D constants (h, β and τ ) and temperature shift factors (aT ) of all binder blends.
Both estimation approaches are based on the same hypothesis, which supposes that the data
obtained for the base materials (5 binders for the first approach; 2 binders and equivalent values for
the rejuvenator in the second approach) are modelled by using the 2S2P1D model and the obtained
2S2P1D parameters and shift factors are then used as input values in order to estimated same param-
eters for several other binder blends (12 binder blends in the first approach and 15 binder blends in
the second approach). The estimated values of the parameters and the values obtained from direct
determination (data reported in Table 2) were then plotted for the considered binder blends in order
to verify the correlation between them. More details for each estimation approach are given in the fol-
lowing subsection. It must be specified that similar estimations were performed for the conventional
parameters (Forton et al., 2020) and for the steady shear viscosity (Forton et al., 2021) obtain for the
same binder blends.

First estimation approach

The first approach supposes that values of 2S2P1D constants (h, β and τ ) and temperature shift factors
(aT ) for fresh binder and for the blends of RAP and rejuvenator (RAP + Rej) are known (results for 5
binders considered as input values).
Mangiafico et al. (2014) proposed an estimation method for the 2S2P1D constants and temperature
shift factors for binder blends produced by mixing one type of RAP binder and different types of fresh
binders (using a linear relation in order to estimate constants h and β and a logarithmic relation in
order to estimate parameter τ and aT temperature shift factors).
Due to the fact that constant β is related to the Newtonian viscosity (Equation (5)), in this study, the
estimation of this constant was performed by using a logarithmic relation.
Equation (7) was used to estimate constant h, while Equation (8) (which is identical to Equation (7)
in log–log scale) was used in order to estimate constants β and τ and temperature shift factors aT of
10 A. FORTON ET AL.

Figure 6. Experimental values of h, β, τ and aT at 65°C for Tref = 55◦ C and equivalent values of the rejuvenator as a function of
the rejuvenator content: re-calculation of R2 by imposing the equivalent values of the rejuvenator as an intercept point for all linear
regressions.

12 binder blends.

hblendapp.1 = a · h50/70 + (1 − a) · hRAP+Rej (7)

log Xblendapp.1 = a · log X50/70 + (1 − a) · log XRAP+Rej (8)

Second Estimation approach

The second approach requires only values of relative concentrations, 2S2P1D constants (h, β and τ )
and temperature shift factors (aT ) of the base components: fresh binder, RAP binder and rejuvenator
(equivalent values). Equations (9) and (10) were considered.
With this approach, estimated values of the 15 blends (hblendapp.2 and Xblendapp.2 for Equations (9)
and (10), respectively) were calculated from experimental values obtained for fresh binder (h50/70 and
X50/70 for Equations (9) and (10), respectively) and RAP binder (hRAP and XRAP for Equations (9) and
(10), respectively) and from equivalent values for the rejuvenator (hRej and XRej for Equations (9) and
(10), respectively) as a function of the relative mass concentrations of fresh binder, RAP binder and
rejuvenator, respectively a, b and c (a + b + c = 1).

hblendapp.2 = a · h50/70 + b · hRAP + c · hRej (9)

log Xblendapp.2 = a · log X50/70 + b · log XRAP + c · log XRej (10)
 ROAD MATERIALS AND PAVEMENT DESIGN 11

Figure 7. Plots of estimated vs. experimental values: (a) constant h; (b) constant β; (c) constant τ forTref = 55◦ C; (d) aT temperature
shift factors for Tref = 55◦ C.

Equivalent values for the rejuvenator (hRej and XRej ) were obtained by optimising the correlation
between estimated values of blends (hblendapp.2 and Xblendapp.2 ) and experimental results (Table 2) by
maximising the R2 values using the least square method. It must be mentioned that these equivalent
values of the rejuvenator were used only in the context of the blending rule, as they are not intended
to reflect actual properties of the rejuvenator.
Estimated values of 2S2P1D constants (h, β and τ ) and temperature shift factors (aT ) of the 15
binder blends and the equivalent values for the rejuvenator are reported in Table A2 (Appendix).
Figure 6 was plotted in order to highlight the differences between the R2 values obtained for the
free linear regression performed on the experimental results and linear regressions performed by
imposing as intercept the equivalent values for the rejuvenator. Therefore, the experimental results
(direct determination) of constants h (Figure 6(a)), β (Figure 6(b)), τ (Figure 6(c)) and of temperature
shift factors aT at 65°C (Tref = 55 ◦ C, Figure 6(d)) were plotted as a function of the rejuvenator con-
tent, together with the equivalent values determined for the rejuvenator. As it can be observed, linear
regressions can be performed with excellent approximation, for all series of blends (25% RAP, 50%
RAP, 75% RAP and 100% RAP), by imposing as intercept point the equivalent values determined for
12 A. FORTON ET AL.

Figure 8. Comparison between test results, curves built using experimental and estimated 2S2P1D constants for the 50/70 + 75%
RAP + 5% Rej blend: (a) master curves of phase angle and norm of the complex shear modulus (Tref = 55◦ C); (b) Black curve.

Figure 9. (a) Error between |G∗ |measured and |G∗ |2S2P1D (from direct determination, from the first and second estimation of 2S2P1D
constants) for the blend 50/70 + 75% RAP + 5% Rej; (b) Error between ϕmeasured and ϕ2S2P1D (from direct determination, from the
first and second estimation of 2S2P1D constants) for the blend 50/70 + 75% RAP + 5% Rej.
 ROAD MATERIALS AND PAVEMENT DESIGN 13

the rejuvenator (hRej and XRej ). The difference between the R2 values obtained for the free linear regres-
sions and the forced linear regressions were insignificant (influence only on the 3rd decimal). It can be
noticed that R2 values are satisfactory (always higher than 0.958).

Analysis of results
Figure 7 shows the correlation plots of estimated (by applying both approaches) vs. experimental
results for constants h (Figure 7(a)), β (Figure 7(b)), τ (Figure 7(c)) and temperature shift factors, aT
(Figure 7(d)).
For both correlations (first and second estimations vs. direct determination results) coefficients of
determination (R2 ) were calculated with respect to the equality line. Values for the base constituents
were not taken into account: for the first estimation approach, the direct determination results for five
binders and for the second estimation approach, the direct determination results for two binders and
the equivalent values for the rejuvenator. As it can be observed in Figure 7, greater values of R2 were
found with the second approach, except for constant τ .
All estimated values reported in Tables A2, A3 and A4 (Appendix) were used to simulate the LVE
behaviour of all blends by re-implementing them in the 2S2P1D model. The resulting master curves
of phase angle and norm of the complex shear modulus and Black curves were compared with those
built using the experimental data. As an example, Figure 8 shows the curves for the blend 50/70 + 75%
RAP + 5% Rej and as it can be observed that master curves and Black curves obtained with both
approaches are almost superposed to the experimental ones.
For all the blends, it was observed that the estimated LVE behaviour is satisfactorily close to the
2S2P1D curves based on the experimental data. These results are showing that these two approaches
used to estimate 2S2P1D constants and temperature shift factors can be considered valid for the tested
binders.
As an example, for the blend 50/70 + 75% RAP + 5% Rej, the errors between the measured norm
of complex shear modulus |G∗ | and phase angle ϕ and the calculated |G∗ | and ϕ from the 2S2P1D
constants (original direct determination and estimated values) are shown in Figure 9.
The errors were calculated for all binder blends for the results obtained at temperatures from 55°C
to 85°C (temperatures for which the measurements were performed at a constant 5% shear stain
amplitude) over the whole frequency range (0.1–10 Hz).
Acceptable errors, in accordance with the findings of Mangiafico et al. (2019b) were generally
observed over the whole range of frequencies and temperatures for both each binder blend.

Conclusions
A fresh 50/70 penetration grade binder, anRAP-extracted binder and a rejuvenator (mixture of veg-
etal oils) were mixed in different amounts, producing 17 different materials (including the two base
binders, three blends of RAP binder with the rejuvenator and 12 blends of fresh and RAP binder with
and without the rejuvenator).
It was observed that the blends produced with 15% Rej presents a behaviour that corresponds
much more with the data obtained for the 50/70 binder. This trend can be explained by the fact that
the ‘optimal dosage’ of the rejuvenator is probably closer to 15% Rej by mass of the RAP binder (cor-
responds to 0.6% rejuvenator by mass of the RAP material for the corresponding bituminous mixture).
However, in order to determine the optimal dosage of a rejuvenator, this process must not be per-
formed on only one direction due to the fact that this dosage may not meet all the specifications of a
fresh virgin binder.
The experimental data were analysed using the 2S2P1D model in which four parameters (G0 , G∞ , k
and δ) were constant for all tested binders and only three constants (h, β and τ ) were variable. 2S2P1D
constants and the temperature shift factors of all binders were then compared according to their RAP
binder content. Constants h, β and τ and shift factors aT show some remarkable tendencies with
14 A. FORTON ET AL.

the increase of RAP binder and rejuvenator contents in the final blends. In particular, parameter h
shows a linear relationship with the RAP binder content and on the other side constants β, τ and aT
temperature shift factors follow a linear evolution in semi-log axes.
Based on these observed tendencies, two different estimation approaches were proposed in
order to estimate h, β, τ and aT for all the blends from the direct determination values for the
base constituents. It was showed that greater values of R2 were found with the second approach,
except for parameter τ . Master curves of phase angle and norm of the complex shear modulus
and Black curves were recalculated using the estimated values and errors between |G∗ |measured and
|G∗ |2S2P1D(original)/est.1/est.2 and ϕmeasured and ϕ2S2P1D(original)/est.1/est.2 were determined for all binder
blends, showing satisfactory correlation with the direct original fitting for each blend. The advantage
of the second approach is that less input data is needed (only data obtained and modelled for the three
base materials) and it can be used in order to estimate 2S2P1D parameters and temperature shift fac-
tors for other combinations of RAP binder/rejuvenator. Further investigation is still needed in order
to validate these methods. Different other binder blends produced with other types of binders and
rejuvenators should be tested over a wider range of temperatures and frequencies in order to verify
the validity of the estimation method. Finding a proper estimation method could allow to determine
2S2P1D parameters and temperature shift factors of binder blends which do not need to be prepared
and tested in the laboratory.

Acknowledgments
The authors received no specific funding for this work.

Disclosure statement
No potential conflict of interest was reported by the author(s).

References
Di Benedetto, H., Neifar, M., Sauzéat, C., & Olard, F. (2007). Three-dimensional thermo-viscoplastic behaviour of bituminous
materials: The DBN model. Road Materials and Pavement Design, 8(2), 285–316. https://doi.org/10.1080/14680629.2007.
9690076
Di Benedetto, H., Olard, F., Sauzéat, C., & Delaporte, B. (2004). Linear viscoelastic behavior of bituminous mate-
rials: From binders to mixes. Road Materials and Pavement Design, 5(sup1), 163–202. [Special Issue:EATA 2004].
https://doi.org/10.1080/14680629.2004.9689992
Ferry, J. D. (1980). Viscoelastic properties of polymers (3rd ed.). Wiley.
Forton, A., Mangiafico, S., Sauzéat, C., Di Benedetto, H., & Marc, P. (2020). Properties of blends of fresh and RAP
binders with rejuvenator: Experimental and estimated results. Construction and Building Materials, 236, 117555.
https://doi.org/10.1016/j.conbuildmat.2019.117555
Forton, A., Mangiafico, S., Sauzéat, C., Di Benedetto, H., & Marc, P. (2021). Steady shear viscosity of blends of fresh and
RAP binders with rejuvenator: Experimental and estimated results. Construction and Building Materials, 269, 121236.
https://doi.org/10.1016/j.conbuildmat.2020.121236
Hong, F., & Prozzi, J. A. (2018). Evaluation of recycled asphalt pavement using economic, environmental, and energy
metrics based on long-term pavement performance sections. Road Materials and Pavement Design, 19(8), 1816–1831.
https://doi.org/10.1080/14680629.2017.1348306
Kaseer, F., Bajaj, A., Martin, A. E., Arambula-Mercado, E., & Hajj, E. (2019). Strategies for producing asphalt mixtures with
high RAP content. Journal of Materials in Civil Engineering, 31(11), 05019002. https://doi.org/10.1061/(ASCE)MT.1943-
5533.0002910
Li, X., Marasteanu, M. O., Williams, R. C., & Clyne, T. R. (2008). Effect of RAP (proportion and type) and binder grade on the
properties of asphalt mixtures. Transportation Research Record: Journal of the Transportation Research Board, 2051(1),
90–97. https://doi.org/10.3141/2051-11
Mangiafico, S., Di Benedetto, H., Sauzéat, C., Olard, F., Pouget, S., & Planque, L. (2014). New method to obtain viscoelas-
tic properties of Bitumen blends from pure and reclaimed asphalt pavement binder constituents. Road Materials and
Pavement Design, 15(2), 312–329. https://doi.org/10.1080/14680629.2013.870639
Mangiafico, S., Sauzéat, C., & Di Benedetto, H. (2019a). Comparison of different blending combinations of virgin and RAP-
extracted binder: Rheological simulations and statistical analysis. Construction and Building Materials, 197, 454–463.
https://doi.org/10.1016/j.conbuildmat.2018.11.217
 ROAD MATERIALS AND PAVEMENT DESIGN 15

Mangiafico, S., Sauzéat, C., & Di Benedetto, H. (2019b). 2S2P1D model calibration error from user panel for one Bitumen and
one bituminous mixture. Advances in Materials Science and Engineering, 2019, 1–16. https://doi.org/10.1016/j.conbui-
ldmat.2018.11.217. https://doi.org/10.1155/2019/6547025
Mangiafico, S., Sauzéat, C., Di Benedetto, H., Pouget, S., Olard, F., & Planque, L. (2017). Complex modulus and fatigue
performances of bituminous mixtures with reclaimed asphalt pavement and a recycling agent of vegetable origin.
Road Materials and Pavement Design, 18(2), 315–330. https://doi.org/10.1080/14680629.2016.1213509
Manke, N. D., Williams, R. C., Sotoodeh-Nia, Z., Cochran, E. W., Porot, L., Chailleux, E., Pouget, S., Olard, F., Barco Carrion, A.
J. D., Planche, J.-P., & Lo Presti, D. (2019). Performance of a sustainable asphalt mix incorporating high RAP content and
novel bio-derived binder. Road Materials and Pavement Design, 23, 1. https://doi.org/10.1080/14680629.2019.1643769
Mazzoni, G., Bocci, E., & Canestrari, F. (2018). Influence of rejuvenators on bitumen ageing in hot recycled asphalt mixtures.
Journal of Traffic and Transportation Engineering (English Edition), 5(3), 157–168. https://doi.org/10.1016/j.jtte.2018.01.
001
Menapace, I., Garcia Cucalon, L., Kaseer, F., Arámbula-Mercado, E., Epps Martin, A., Masad, E., & King, G. (2018). Effect of
recycling agents in recycled asphalt binders observed with microstructural and rheological tests. Construction and
Building Materials, 158, 61–74. https://doi.org/10.1016/j.conbuildmat.2017.10.017
Mogawer, W. S., Booshehrian, A., Vahidi, S., & Austerman, A. J. (2013). Evaluating the effect of rejuvenators on the degree
of blending and performance of high RAP. RAS, and RAP/RAS Mixtures. Road Materials and Pavement Design, 14(sup2),
193–213. https://doi.org/10.1080/14680629.2013.812836
Moon, K. H., Falchetto, A. C., Marasteanu, M., & Turos, M. (2014). Using recycled asphalt materials as an alternative material
source in asphalt pavements. KSCE Journal of Civil Engineering, 18(1), 149. https://doi.org/10.1007/s12205-014-0211-1
Nguyen, Q. T., Di Benedetto, H., Sauzéat, C., & Tapsoba, N. (2013). Time temperature superposition principle for
bituminous mixes in the linear and nonlinear domains. Journal of Materials in Civil Engineering, 25(9), 1181–1188.
https://doi.org/10.1080/19648189.2009.9693176 https://doi.org/10.1061/(ASCE)MT.1943-5533.0000658
Noferini, L., Simone, A., Sangiorgi, C., & Mazzotta, F. (2017). Investigation on performances of asphalt mixtures made with
reclaimed asphalt pavement: effects of interaction between virgin and RAP bitumen. International Journal of Pavement
Research and Technology, 10(4), 322–332. https://doi.org/10.1016/j.ijprt.2017.03.011
Olard, F., & Di Benedetto, H. (2003). General “2S2P1D” model and relation between the linear viscoelastic behaviors of bitu-
minous binders and mixes. Road Materials and Pavement Design, 4(2), 185–224. https://doi.org/10.1080/14680629.2003.
9689946
Oldham, D., Hung, A., Parast, M. M., & Fini, E. H. (2018). Investigating Bitumen rejuvenation mechanisms using
a coupled rheometry-morphology characterization approach. Construction and Building Materials, 159, 37–45.
https://doi.org/10.1016/j.conbuildmat.2017.10.113
Tapsoba, N., Sauzéat, C., Di Benedetto, H., Baaj, H., & Ech, M. (2014). Behaviour of asphalt mixtures contain-
ing reclaimed asphalt pavement and asphalt shingle. Road Materials and Pavement Design, 15(2), 330–347.
https://doi.org/10.1080/14680629.2013.871091
Visintine, B., Khosla, N. P., & Tayebali, A. (2013). Effects of higher percentage of recycled asphalt pavement on pavement
performance. Road Materials and Pavement Design, 14(2), 432–437. https://doi.org/10.1080/14680629.2013.779310
Walther, A., Büchler, S., Cannone Falchetto, A., Wang, D., Riccardi, C., & Wistuba, M. P. (2019). Experimental investigation
on asphalt mixtures prepared with reclaimed asphalt pavement and rejuvenators based on the BTSV method. Road
Materials and Pavement Design, 20(7), 1695–1708. https://doi.org/10.1080/14680629.2019.1594053
Zaumanis, M., Mallick, R. B., Poulikakos, L., & Frank, R. (2014). Influence of six rejuvenators on the performance properties
of reclaimed asphalt pavement (RAP) binder and 100% recycled asphalt mixtures. Construction and Building Materials,
71, 538–550. https://doi.org/10.1016/j.conbuildmat.2014.08.073
16 A. FORTON ET AL.

Appendix

Table A1. aT temperature shift factors (Tref = 55◦ C) for all tested binders.
aT (−)
Binders 25°C 35°C 45°C 55°C 65°C 75°C 85°C
50/70 3.96 × 102 41.00 5.30 1 2.36 × 10−1 7.40 × 10−2 2.62 × 10−2
RAP 1.41 × 103 88.00 7.90 1 1.60 × 10−1 3.30 × 10−2 7.70 × 10−3
RAP + 5% Rej 5.82 × 102 52.00 6.00 1 2.10 × 10−1 5.30 × 10−2 1.62 × 10−2
RAP + 10% Rej 3.00 × 102 33.00 4.90 1 2.53 × 10−1 7.70 × 10−2 2.70 × 10−2
RAP + 15% Rej 1.82 × 102 22.00 4.21 1 2.97 × 10−1 1.02 × 10−1 4.12 × 10−2
50/70 + 25% RAP 5.59 × 102 48.00 5.70 1 2.20 × 10−1 6.04 × 10−2 1.97 × 10−2
50/70 + 25% RAP + 5% Rej 4.88 × 102 43.50 5.50 1 2.31 × 10−1 6.49 × 10−2 2.20 × 10−2
50/70 + 25% RAP + 10% Rej 4.27 × 102 39.80 5.30 1 2.40 × 10−1 7.00 × 10−2 2.44 × 10−2
50/70 + 25% RAP + 15% Rej 3.76 × 102 36.90 5.08 1 2.49 × 10−1 7.47 × 10−2 2.70 × 10−2
50/70 + 50% RAP 6.37 × 102 52.50 5.98 1 2.08 × 10−1 5.30 × 10−2 1.68 × 10−2
50/70 + 50% RAP + 5% Rej 4.89 × 102 44.03 5.53 1 2.25 × 10−1 6.37 × 10−2 2.07 × 10−2
50/70 + 50% RAP + 10% Rej 3.78 × 102 37.06 5.11 1 2.43 × 10−1 7.33 × 10−2 2.55 × 10−2
50/70 + 50% RAP + 15% Rej 2.96 × 102 31.46 4.74 1 2.61 × 10−1 8.37 × 10−2 3.10 × 10−2
50/70 + 75% RAP 8.75 × 102 63.80 6.70 1 1.92 × 10−1 4.62 × 10−2 1.30 × 10−2
50/70 + 75% RAP + 5% Rej 5.85 × 102 48.77 5.92 1 2.16 × 10−1 5.75 × 10−2 1.79 × 10−2
50/70 + 75% RAP + 10% Rej 3.99 × 102 37.85 5.28 1 2.42 × 10−1 7.07 × 10−2 2.43 × 10−2
50/70 + 75% RAP + 15% Rej 2.78 × 102 29.72 4.72 1 2.69 × 10−1 8.61 × 10−2 3.24 × 10−2

Table A2. Estimated values of 2S2P1D constants for all binders.

1st estimation 2nd estimation
Materials h (−) β (−) τ (s) h (−) β (−) τ (s)
50/70 N∗ N∗ N∗ N∗ N∗ N∗
RAP N∗ N∗ N∗ N∗ N∗ N∗
Rej equivalent – – – ∗∗0.985 ∗∗3.52 × 10−3 ∗∗7.01 × 10−20
RAP + 5% Rej N∗ N∗ N∗ 0.614 354 2.31 × 10−8
RAP + 10% Rej N∗ N∗ N∗ 0.631 210 6.92 × 10−9
RAP + 15% Rej N∗ N∗ N∗ 0.647 130 2.30 × 10−9
50/70 + 25% RAP 0.626 166 7.78 × 10−9 0.626 166 7.78 × 10−9
50/70 + 25% RAP + 5% Rej 0.629 151 6.04 × 10−9 0.630 151 6.22 × 10−9
50/70 + 25% RAP + 10% Rej 0.633 136 4.92 × 10−9 0.633 137 4.97 × 10−9
50/70 + 25% RAP + 15% Rej 0.636 126 4.12 × 10−9 0.636 125 4.00 × 10−9
50/70 + 50% RAP 0.620 222 1.31 × 10−8 0.620 222 1.31 × 10−8
50/70 + 50% RAP + 5% Rej 0.626 184 7.89 × 10−9 0.626 183 8.35 × 10−9
50/70 + 50% RAP + 10% Rej 0.633 148 5.25 × 10−9 0.632 151 5.37 × 10−9
50/70 + 50% RAP + 15% Rej 0.637 128 3.71 × 10−9 0.638 126 3.52 × 10−9
50/70 + 75% RAP 0.614 296 2.22 × 10−8 0.613 296 2.22 × 10−8
50/70 + 75% RAP + 5% Rej 0.622 222 1.03 × 10−8 0.623 220 1.11 × 10−8
50/70 + 75% RAP + 10% Rej 0.634 161 5.60 × 10−9 0.632 167 5.79 × 10−9
50/70 + 75% RAP + 15% Rej 0.640 130 3.36 × 10−9 0.641 127 3.10 × 10−9
Notes: N∗ – no estimation performed (experimental results). ∗∗ – equivalent values for the rejuvenator obtained by optimising the
correlation between estimated and experimental results.
 ROAD MATERIALS AND PAVEMENT DESIGN 17

Table A3. Estimated values of temperature shift factors aT (Tref = 55◦ C) by applying the 1st approach for all binders analysed in
this approach.
1st estimation
Binders aT 25°C aT 35°C aT 45°C aT 65°C aT 75°C aT 85°C
50/70 + 25% RAP 4.97 × 102 46.99 5.69 2.20 × 10−1 6.41 × 10−2 2.11 × 10−2
50/70 + 25% RAP + 5% Rej 4.25 × 102 42.85 5.42 2.31 × 10−1 6.96 × 10−2 2.40 × 10−2
50/70 + 25% RAP + 10% Rej 3.75 × 102 39.32 5.22 2.39 × 10−1 7.46 × 10−2 2.64 × 10−2
50/70 + 25% RAP + 15% Rej 3.39 × 102 36.20 5.06 2.47 × 10−1 7.89 × 10−2 2.87 × 10−2
50/70 + 50% RAP 6.23 × 102 53.86 6.11 2.05 × 10−1 5.55 × 10−2 1.69 × 10−2
50/70 + 50% RAP + 5% Rej 4.56 × 102 44.75 5.55 2.26 × 10−1 6.54 × 10−2 2.19 × 10−2
50/70 + 50% RAP + 10% Rej 3.56 × 102 37.76 5.14 2.42 × 10−1 7.51 × 10−2 2.65 × 10−2
50/70 + 50% RAP + 15% Rej 2.92 × 102 32.17 4.85 2.58 × 10−1 8.39 × 10−2 3.13 × 10−2
50/70 + 75% RAP 7.82 × 102 61.73 6.56 1.92 × 10−1 4.80 × 10−2 1.36 × 10−2
50/70 + 75% RAP + 5% Rej 4.89 × 102 46.70 5.67 2.21 × 10−1 6.16 × 10−2 2.01 × 10−2
50/70 + 75% RAP + 10% Rej 3.39 × 102 36.31 5.07 2.45 × 10−1 7.57 × 10−2 2.66 × 10−2
50/70 + 75% RAP + 15% Rej 2.54 × 102 28.75 4.65 2.69 × 10−1 8.89 × 10−2 3.39 × 10−2

Table A4. Estimated values of temperature shift factors aT (Tref = 55◦ C) by applying the 2nd approach for all binders analysed in
this approach.
2nd estimation
Materials aT 25°C aT 35°C aT 45°C aT 65°C aT 75°C aT 85°C
Rej equivalent ∗1.67 × 10−4 ∗2.13 × 10−3 ∗5.36 × 10−2 ∗20.60 ∗1.83 × 102 ∗2.83 × 103
RAP + 5% Rej 6.60 × 102 53.06 6.23 2.02 × 10−1 4.97 × 10−2 1.42 × 10−2
RAP + 10% Rej 3.31 × 102 33.49 5.02 2.49 × 10−1 7.23 × 10−2 2.47 × 10−2
RAP + 15% Rej 1.76 × 102 22.01 4.12 3.01 × 10−1 1.02 × 10−1 4.09 × 10−2
50/70 + 25% RAP 4.97 × 102 46.99 5.69 2.20 × 10−1 6.41 × 10−2 2.11 × 10−2
50/70 + 25% RAP + 5% Rej 4.36 × 102 43.03 5.46 2.29 × 10−1 6.87 × 10−2 2.34 × 10−2
50/70 + 25% RAP + 10% Rej 3.82 × 102 39.41 5.24 2.38 × 10−1 7.37 × 10−2 2.59 × 10−2
50/70 + 25% RAP + 15% Rej 3.37 × 102 36.20 5.04 2.48 × 10−1 7.89 × 10−2 2.87 × 10−2
50/70 + 50% RAP 6.23 × 102 53.86 6.11 2.05 × 10−1 5.55 × 10−2 1.69 × 10−2
50/70 + 50% RAP + 5% Rej 4.78 × 102 45.10 5.63 2.23 × 10−1 6.39 × 10−2 2.09 × 10−2
50/70 + 50% RAP + 10% Rej 3.70 × 102 37.96 5.19 2.41 × 10−1 7.33 × 10−2 2.56 × 10−2
50/70 + 50% RAP + 15% Rej 2.89 × 102 32.18 4.80 2.60 × 10−1 8.37 × 10−2 3.12 × 10−2
50/70 + 75% RAP 7.82 × 102 61.73 6.56 1.92 × 10−1 4.80 × 10−2 1.36 × 10−2
50/70 + 75% RAP + 5% Rej 5.24 × 102 47.21 5.79 2.17 × 10−1 5.95 × 10−2 1.87 × 10−2
50/70 + 75% RAP + 10% Rej 3.58 × 102 36.63 5.14 2.43 × 10−1 7.30 × 10−2 2.53 × 10−2
50/70 + 75% RAP + 15% Rej 2.49 × 102 28.74 4.59 2.71 × 10−1 8.87 × 10−2 3.38 × 10−2
Notes: ∗ – equivalent values for the rejuvenator obtained by optimising the correlation between estimated and experimental results.