The materials studied under cyclic loading are composed of a mixture of asphalt concrete AC50/70 and the coarse aggregates SACC and pebbles. The latter is what is currently largely used in the asphalt mixtures of pavements, the exploitation of which is responsible for the deterioration of the Amazon rivers beds. In this study, pebbles are used only as a comparison to SACC, which is studied as an alternative for the serious environmental problems created by the exploitation of pebbles.

1) Aggregates

The SACC was obtained from typical clay soil of Amazon-Brazil. The SACC was characterized according to the Brazilian Association of Technical Standards (ABNT) using NBR-7181 for the grain size, NBR-6459 for the liquid limit and NBR-7180 for the plastic limit. The calcination potentiality of the raw clay to form aggregates was obtained following the instructions of the Federal National Department of Transport Infrastructure (DNIT) of Brazil (ME222, ME223 and ME225). To produce aggregates from raw clay, massive bricks, with dimensions of 60 mm × 110 mm × 200 mm and two longitudinal holes, were crushed into grain-sized portions.

The two coarse materials were evaluated according to the instructions of the American Society for Testing and Materials (ASTM). The grain size and the Los Angeles abrasion value follow ASTM-C136 and ASTM- C131, respectively. The apparent specific gravity (Gsa), bulk specific gravity (Gsb), bulk specific gravity in the condition saturated surface dry (Gsbssd) and absorption were obtained via adherence to ASTM-C127. The coarse materials were also characterized in relation to the Rodded Unit Weight (Wur) and Loose Unit Weight (Wul) following the American Association of State Highway and Officials (AASHTO).

Three fine aggregates were used to constitute the asphalt mixtures: two sands, obtained from the local supplier, with different grain sizes, and the powder resulting from the procedure of brick grinding. They were characterized by ASTM-C136 (grain size), ASTM-C128 (Gsa, Gsbssd and absorption) and Rodded Unit Weight (AASHTO).

The grain-size distributions for the natural soil (NSOIL), which was used to produce the sintered aggregate of calcined clay, are shown in Figure 1: the sands MAO and COA, contained in AC 50/70 mixtures with SACC and pebbles, respectively; the artificial sand, obtained from powder of SACC (SACC-P); and the coarse aggregates SACC and pebbles. The consistency indexes of NSOIL are: LL = 48%, PL = 22% and PI = 26%. From these results, NSOIL is classified as belonging to the groups A7-6 and CL, according to the AASHTO Classification System and Unified Classification System, respectively.

The mineralogy of the natural soil used to obtain the sintered aggregate consists of kaolinite, goethite, hematite, illite and quartz. The characterizations of the coarse aggregate and the sand are shown in Table 1 and Table 2, respectively.

2) Binder and filler

The petroleum asphalt cement AC 50/70 was used as the binder, and rheological analysis, performed according to the Superior Performance Asphalt Pavements, classified it as PG64 + 2. The Portland cement, the filler in the composition of the asphalt mixtures, exhibits 3.15/cm³ of apparent specific gravity and a 0.075 mm sieve passing percentage of 100% according to DNIT-ME085 and DNIT-ME367, respectively.

The mechanical behavior of asphalt mixtures depends largely on temperature, varying from elastic to nonlinear viscoelastic, as well as on rate of application of the load or strain. These characteristics are derived primarily from the asphalt binder. In this study, we perform the Dynamic Modulus |E^*| under both constant tension and constant strain regimes. This propriety was obtained for two types of AC 50/70 mixtures: with SACC, as an alternative coarse material, and with pebbles, which is a conventional coarse aggregate used in civil engineering and taken as the reference material for SACC.

1) Samples

Following the Superpave system for the granulometric compositions of asphalt mixtures, the Nominal Maximum Sizes (NMS) adopted was 19.00 mm. The percentages of mineral components of the mixtures were defined according to the Bailey method [21] .

In Table 3, we present the composition of the concrete asphalt mixture with SACC and pebbles. The AC 50/70 mixture with SACC consists of 74.25% SACC, 14.03% SACC sand (SACC-P), 10.32% Coari sand (COA) and 1.40% Portland cement. The AC 50/70 mixture with pebbles consists of 67.20% pebbles, 29.7% Manaus sand (MAO) and 3.1% Portland cement.

The air voids (Va) and voids filled with asphalt (VFA) were determined by compacting the mixtures according to the Marshall method of dosage. Based on these parameters, and following the method of DNIT, the content of the composite was determined; the volumetric parameters are shown in Table 4. Aiming to simulate the

short-termaging effect, before compaction, the mixtures remained in an oven at 10˚C above the temperature of compaction for two hours. Based on the guidelines of DNIT, its preparation and compaction temperatures were determined.

To perform the dynamic module test, the samples were prepared with 100 mm diameters and 200 mm heights. The samples were subjected to pulses of sinusoidal loads, under strain control mode (by applying axial tension- compression loading) and under tension control mode, at temperatures of −5˚C, 10˚C, 25˚C and 35˚C and load application frequencies of 0.01, 0.05, 0.10,0.50, 1.00, 5.00, 10.00 and 25.00 Hz. To prevent sample damage, the axial deformation did not exceed the value of 75 µε. The tests were performed using a Universal Testing Machine UTM-14, made by IPC Co. Ltd. of Australia. The deformations in the samples were monitored with micros LVDTs (Linear Variable Differential Transducers). Such devices were bonded at four diametrically opposed points of the samples at distances of 25 mm to150 mm from each face.

The Dynamic Modulus (E^∗) of AC 50/70 mixtures with SACC and pebbles was obtained under both the strain control mode and tension control mode. In Figure 2(a), we present |E^*| at constant tension regime for AC 50/70 mixed with SACC for three different temperatures, T = 40˚C, T = 10˚C and T = −5˚C, as a function of the frequency, ω. In Figure 2(b), we show |E^*| as a function of the reduced frequency ω_R = a_Tω, where a_T is a temperature shift factor [22] .

Taking a reference temperature T₀ in a double logarithmic scale, log(α_T) is the amount of shift necessary to form a unique curve from different temperatures, known as a master curve. In the present study, the reference temperature T₀ is taken as 10˚C. The results present in Figure 2(b) are compared with the master curve obtained by Zheng et al. [23] , analyzing asphalt mixtures over various frequency and temperature ranges. According to these authors, |E^*| of the mixtures as a function of the reduced frequency is represented by the function

where is the complex modulus in the limit ω_R→∞, ω_c is a frequency location parameter, and k and me are

dimensionless shape parameters. The parameters ω_c, k and m_e are estimated using a minimum squared error fitting procedure. This function presents a spectrum that increases monotonically from zero, at ω_R = 0, until its saturation at as ω_R→∞. The width of the relaxation spectrum is characterized by a shape index defined as R = (m_e/k)log2. The higher this index, the more gradual the transition from the elastic to the viscous behavior is. The solid lines in Figure 2(b) represent the quality of the fit to Equation (1) for each sample.

In Figure 3(a), we present the values of as a function of ω_R under both constant tension and constant strain regimes, and we observe good agreement between the two methods. In Figure 3(b), we present the temperature shift factor a_T as a function of T obtained under constant tension (C. Tension) and constant strain (C. Strain) regimes, with both coinciding and approaching a straight line.

The comparison between the Dynamic Modulus of the mixtures of AC 50/70 with SACC and with pebbles is shown in Figure 4, where we present, in (a), |E^*| as a function of the reduced frequency ω_R and, in (b), |E^*| as a function of the frequency ω, obtained under constant stress. In (a), the black squares and the red circles represent the AC 50/70 mixture with pebbles and the AC 50/70 mixture with SACC, respectively. From these results, we observe that for low reduced frequency ω_R, the mixture with SACC is stronger than the mixture with pebbles, and, for high ω_R, the inverse occurs and the mixture with pebbles is stronger than the mixture with SACC. For a given frequency, a small reduced frequency corresponds to a high temperature environment. In (b), we verify that for T = −5˚C and T = 10˚C, the mixture with pebbles is stronger than the mixture with SACC. When increasing the temperature and for low frequency, the mixture with SACC becomes stronger than that with pebbles. For T = 40˚C, for example, the mixture with SACC is stronger than the mixture with pebbles for frequencies lower than 1.0 Hz. In tropical zones, the temperature of asphalt pavement can rise to 70˚C. Our results shows

that, in such zones and where sedimentary soil is more plentiful, with the lack of granular material, the sintered aggregate of calcined clay (SACC) can be a good alternative to natural granular aggregate for concrete asphalt mixtures.