Evaluation of a rejuvenator as an additive in asphalt Valoración de un rejuvenecedor como aditivo en el asfalto

Ecuadorian asphalt has the particularity of being prone to premature aging; therefore, the asphalt pavements have insufficient durability. The objective of this experimental work is to assess the effect that a modification with the rejuvenating agent Sylvaroad RP-1000 produces on the properties of AC 20 asphalt from the Esmeraldas Refinery, as well as evaluating its impact on the quality of a typical asphalt mix, without reclaimed asphalt pavement, prepared with modified and unmodified asphalt. The asphalt used is also classified as PG 64-22 and when mixed with 2.5% by weight of the rejuvenator, the grade changed to PG 58-28. Additional studies were carried out using the Δ Tc, Glover-Rowe parameters and the corresponding transition temperatures, showing improvements in the results. At that point, a dense mix with 5.5% asphalt was designed. Stiffness modulus, cyclic compression and fatigue modulus tests were applied with the NAT equipment to the specimens. When using modified asphalt, the fatigue behavior improved remarkably. Additionally, an aging process in oven at 85 ° C was applied to both mixtures, measuring the stiffness modulus and fatigue at 8 days, which demonstrated an adequate behavior only in the mixture made with rejuvenator. en los resultados. Luego se diseñó una mezcla densa con 5.5% de asfalto; a las briquetas confeccionadas se le aplicaron pruebas de módulo de rigidez, compresión cíclica y fatiga con el equipo NAT. Al emplear asfalto modificado el comportamiento a la fatiga mejoró notablemente. Adicionalmente, se aplicó un proceso de envejecimiento en horno a 85°C a ambas mezclas, valorándose el módulo y la fatiga a los 8 días, que demostraron un buen comportamiento solo en la mezcla con rejuvenecedor.


Introduction
For economic and environmental reasons, the use of Reclaimed Asphalt Pavement (RAP) has been increasing in many countries; however, the use of high percentages of RAP in a recycled mix will make it less workable, more difficult to compact, and more prone to cracking and weathering (Li et al., 2008), (Mogawer et al., 2012). One solution to these problems is the use of softer base asphalts or the use of recycling or rejuvenating agents. These agents can restore the rheological characteristics of the recycled binders and their mixture, with the base binder to the desired performance requirements .
It is known that flexible pavements in Ecuador present some types of deterioration within a few years of being constructed or rehabilitated. Fatigue or block cracking are the most common types of damage. This may be attributed, in part, to the fact that most of the asphalt used comes from the Esmeraldas Refinery, which has the particularity of being prone to premature aging and therefore to low durability. This can be verified by viscosity classification, where non-compliance with quality requirements in the residue of the rotary thin film test or RTFOT is frequent . Therefore, in works developed at the Road Laboratory of the Catholic University of Santiago de Guayaquil (UCSG), rejuvenating additives have been directly used with Ecuadorian asphalt cement to try to decrease the intensity of its aging, also to assess its incidence on the behavior of the mixtures (Icaza and Mera, 2018).
The objective of the present experimental work is to evaluate the effect on the properties of AC 20 asphalt, coming from the Esmeraldas Refinery, its modification with the rejuvenating agent Sylvaroad RP-1000; as well as to evaluate its incidence on the quality of a typical asphalt mix, without milled material, prepared with unmodified asphalt and modified with rejuvenator.

Classification by viscosity
The base binder used was asphalt cement from Esmeraldas, which was modified with the addition of the indicated rejuvenating agent. For the dosing of the rejuvenating agent, the first step was to ensure compliance with the RTFOT residue requirements. Subsequently, for the mixtures, it was verified that this dosage was not excessive, because if it were, it would produce a very soft binder that would have a negative impact on the rutting resistance of the mixture. Conversely, a very low dosage could help reduce the binder's brittleness but without a pronounced effect on the improvement of fatigue resistance.
The modification employed involved mixing 97.5% by weight of AC 20 base asphalt with 2.5% by weight of rejuvenator. (Table 1) shows the test results for viscosity classification, according to the Ecuadorian standard (INEN 2515(INEN , 2010. In the base asphalt, although the viscosity test on the original bitumen at 60°C, would allow the sample to be identified as AC 20. The viscosity and ductility requirements are not met in the RTFO residue tests, which is a fairly widespread problem in this asphalt and which is precisely the reason for this research. When modified with rejuvenator, the residue requirements can be met, but the viscosity of the original asphalt is then not met. This behavior should be considered when adding rejuvenators in the manner proposed in this study. (Table 2) shows the values obtained in the Superpave tests according to INEN 3030-2017. Theoretically a PG 64 -22 asphalt could satisfy the temperature ranges of all our geographical regions, obtained according to the Superpave criteria and shown in (Table 3) . However, in our roads there is a poor performance in the face of intermediate temperatures, identified with fatigue cracking. It is widely known that the control parameter for these temperatures employed by Superpave, i.e., "G*.Senδ", correlates poorly with fatigue.

Classification by performance grades
On the other hand, as pavements age they begin to exhibit cracking and aggregate spalling. Although the stresses produced by traffic increase destructions, the evolution of rheological properties of aged asphalt can be damaging enough by itself to cause block cracking due to stresses caused by thermal gradients (King et al., 2012). This criteria is considered valid for the ranges of fluctuating daily temperatures throughout the year in Ecuador, with more significant gradients in the Sierra region, where precisely these types of damage: block cracking and spalling or "peeling", are more frequent and intense. With the addition of a rejuvenator, as shown in (Table 2), a PG 58-28 is obtained, which can undoubtedly help to reduce these negative effects, with greater possibilities in the Sierra region due to the high temperatures reached.

Additional studies on binders
A few years ago, an expert from the Asphalt Institute expressed the following opinion: The next step in the evolution of asphalt technology is to set a parameter for the control of binders at intermediate temperatures, i.e., related to deteriorations associated with fatigue cracking and durability. A group of new tests and parameters have been proposed: Linear Amplitude Sweep (LAS), R-value, Glover-Rowe, ΔTc and Double Edge Notched Tension (DENT) to name a few, but asphalt scholars have yet to agree on the one they believe best relates to cracking at intermediate temperatures (Anderson, 2016a).
Therefore, it was deemed appropriate to perform some additional tests within our scope.

Parameter ΔTc
This parameter is determined by performing additional calculations with the data obtained from the beam flexural beam rheometer (BBR) test. "ΔTc" is calculated according to (Equation 1): Donde Tc,m es la temperatura crítica correspondiente al valor de la pendiente "m" (pendiente de relajamiento de Where Tc,m is the critical temperature corresponding to the value of slope "m" (stiffness relaxation slope) equal to 0.300 with loading time of 60 seconds and Tc,S is the critical temperature corresponding to the value of stiffness equal to 300 MPa obtained with loading time of 60 seconds.
Although the "ΔTc" is obtained at low temperatures, this parameter is an indicator of binder quality and durability that can correlate very well with other parameters linked to intermediate temperatures (Anderson et al., 2011). As the asphalt ages, the value of "ΔTc" increases, indicating what is considered a loss in relaxation properties. The values related to the onset of cracking and the presence of cracks are 2.5 and 5.0°C respectively. The calculated values are presented in (Table 4).
It is highlighted that in some works (Anderson, 2016b) the parameter "ΔTc" is designated as the difference between "Tc,S" and "Tc,m", which could create some confusion in the interpretation of results due to the change of sign it produces.

G'/ (ƞ'/G') = G*. ((cos δ) ² / sin δ). ω (2)
Provided that the test frequency (ω) the variables: complex modulus (G*) and phase angle (δ) are known, a damage curve can be created in Black's Diagram. The Glover -Rowe parameter is obtained from frequency swept DSR tests at 15°C and 0.005 rad/s (King et al, 2012). Values below 180 kPa suggest that no block cracking will exist, between 180 and 600 kPa that cracks are developing and above 600 kPa that cracking already exists. The results are summarized in (Table 4).

Viscoelastic transition temperature
It is known that in the master curves of an asphalt the crossover frequency represents a balance between the elastic component (G') of the complex modulus and the viscous component (G''), and a transition from a more  In the previously referred work and from correlations with the Glover -Rowe parameter, the behavior of binders remaining below the warning threshold (32°C), after 20 h of aging in PAV equipment, and below the cracking limit (45°C) after 40 h of aging in PAV, tested at 10 rad/s, is considered satisfactory. (Table 4) shows the transition temperatures obtained according to these criteria.
( Table 4) also shows that for the PAV residue in the base asphalt, both the ΔTc value and the transition temperature are in the cracking zone according to the criteria used; this situation is not reflected by the Glover-Rowe parameter. On the other hand, the addition of the rejuvenator significantly improved the behavior evaluated with these criteria.

Asphalt mix characteristics
For this research, a dense-grained mix with a maximum nominal size of 12.5 mm was used as a reference, using good quality mineral aggregates of basaltic origin. (Table 5) shows the characteristics of the aggregates and (Figure 1) shows the combined granulometry used.
A Marshall design was carried out to determine an optimum asphalt content of 5.5% (by weight). With this percentage, briquettes were made for each type of asphalt (without and with rejuvenator). Before the compaction process, the mixes were placed in an oven at 135°C for 2 hours.    (Cooper, 2002), which partially coincide with the corresponding European standards. (Vila, 2017).

Rigidity Module
Its determination was performed at the temperature of 20°C, using a controlled deformation level of 5 microns and sinusoidal waves with time interval between the beginning of the load pulse and the point at which the load is maximum, of 0.12 seconds. The results of these tests are presented in (Table 6).

Permanent deformation under cyclic compression
In this test a loading cycle consists of applying a stress for 1 second followed by 1 second of rest, with quadratic waves. The test was performed at a temperature of 40°C, with a load magnitude of 100 Kpa, measuring the vertical deformations caused by 3600 repetitions of this load. The results are shown in (Table 7).

According to the proposed requirements, the qualification of asphalt mixtures, based on the percentage of deformation in the uniaxial cyclic compression test, is:
Satisfactory mixes: deformations less than or equal to 1%, and Inadequate mixes: deformations greater than 1%. Analyzing the average values of the briquettes, it was observed that the two mixtures have values below 1% deformation, which indicates that satisfactory behavior is expected in the mixtures with respect to plastic deformation. However, the addition of a rejuvenator always tends to increase the deformations, in this case in a very small range.

Indirect tensile fatigue under controlled strain
In order to perform the fatigue test, it is necessary to previously determine the module of rigidity (Sm), using the same stress (σ) with which the aforementioned test will be performed. With the fatigue test, the number of load applications required to reach breakage, or a maximum deformation of 5 mm is obtained. The loading time is 120 milliseconds and the temperature for the study is 20°C. Considering the Poisson's coefficient (μ) with a value of 0.35, it is possible to calculate the initial tensile unit strain (ɛ) according to (Equation 3):  Table 7. Results of the uniaxial cyclic compression tests (Figure 2) represents the obtained adjustment lines. It is important to mention that according to the requirements referred to for the fatigue case, it is noticed that if the points for the analyzed mixture fall below the line corresponding to the percentile (dotted black line), the estimated behavior will be inadequate. If the points fall above the average line (dashed black line), the estimated performance will be satisfactory. Between the two lines the behavior can be considered as tolerable.

Uniaxial cyclic compression test at 40º C, % Specimen
As shown in (Figure 2), the trend line of the asphalt with rejuvenator is located above the average line, which indicates an adequate fatigue behavior according to the proposed criteria. For the base asphalt (without rejuvenator) the behavior is between tolerable and adequate.

Study of laboratory aged mixtures
From the previous analyses, it can be stated that both mixtures present acceptable behaviors, which is a very interesting result. However, it must be considered that in all the tests carried out, the mixtures have not undergone any long-term aging process. For this reason and taking into account that fatigue behavior, is one of the most important in asphalt pavements in Ecuador, it was decided to make sets of briquettes with the mixtures studied to evaluate their performance after an aging process in the laboratory.
This evaluation consisted of keeping the briquettes for 8 days in an oven at 85°C to develop long-term aging, which can be estimated at approximately more than 9 years or more than 18 years, depending on whether the zone is dry -with freeze or wet -without freeze (Bell et al., 1994). These estimates were obtained for some North American asphalt cements placed in different regions of that country, so they should be considered only as references. To consider the effect of aging with days in the kiln, stiffness modules were measured every two days on briquettes of both combinations during a 10-day period. The results obtained can be seen in (Table 8)   As shown in (Figure 3), the average stiffness modulus of the mixes corresponding to the base asphalt tends to stabilize at a value close to 7000 MPa after 8 days, while the average modulus of the mix with rejuvenator stabilizes at a value of approximately 5000 MPa after 4 days. In other words, when approaching a maximum stiffness in the mix due to aging, the use of the rejuvenator at a dosage of 2.5% by weight allows a decrease in the modulus of 2000 MPa, which represents 28.6%, which should have a positive impact on fatigue behavior.
The results of the mix with asphalt without additives, shown in (Figure 3), are lower than those obtained in a previous study (Cedeño, 2015), where for the mix used, modules close to 9000 MPa were obtained at 8 days. This difference in values should consider, among others, the quality of the mineral aggregates used.

Mix Specimen
Original Asphalt  The results of the fatigue tests performed on the two mixes under non-aging conditions and after 8 days of aging are shown in (Figure 4). As can be seen, the mix with rejuvenator has a much better fatigue behavior in both conditions, moving from a "satisfactory" to a "tolerable" position, while the mix without rejuvenator ends the 8 days in an "inadequate" state.

Conclusions
The rejuvenator used as an additive for Ecuadorian asphalt AC 20 allows meeting the corresponding requirements to the residue of the RTFO test, which is technically very convenient to counteract the usual failures that occur in our asphalt pavements. On the other hand, additional studies using the parameters ΔTc, Glover -Rowe and the transition temperature, all associated to the behavior before intermediate temperatures, show a great potential for its application. For the PAV residue in the base asphalt, both the ΔTc value and the transition temperature are in the cracking zone according to the criteria used, which allows us to identify this problem in our asphalt. The addition of the rejuvenator significantly improved the behavior evaluated with these criteria.
The results of the tests on the asphalt mixes, without aging, show that the mixes made with binder without and with rejuvenator meet the performance requirements performed with the NAT equipment. However, in the tests on specimens with long-term aging (8 days in an oven at 85°C), it was observed that the effect of such aging had a stronger impact on the mix with the base asphalt without rejuvenator, in which the stiffness modulus stabilized at approximately 7000 MPa, while the modulus obtained with rejuvenator stabilized at a value of 5000 MPa. This denotes that the addition of such a product would allow obtaining a less stiff mixture over time.
When analyzing the fatigue laws for the two mixtures, at the end of the indicated aging process, it can be concluded that the rejuvenator improves the performance of the mixture because the corresponding adjustment line is above the 15th percentile line, qualifying it as a "tolerable" mixture. On the other hand, the adjustment line for the mix with asphalt without rejuvenator is below the aforementioned limit, reaching "inadequate" levels.
In general, although the direct application of rejuvenators to asphalt cement produces a decrease in its original viscosity, it is possible to produce asphalt mixtures that perform satisfactorily in our climatic conditions. Especially the fatigue behavior, which is our major problem, is significantly improved.