Effects of curing on emulsion cold mix asphalts and their extracted binder

10 This paper focuses on the physicochemical changes that happen in cold mix asphalts during curing, and more specifically, while and after transitioning to different simulated seasons. Several tests were carried out in order to better grasp the influence of the weather (temperature and humidity) on the curing of such materials. The mechanical behaviour of the mix was 15 assessed using oedometer tests. The physicochemical evolutions of extracted binders, such as oxidation and rheology, were evaluated. The results show stiffening of the mix and ageing of the binder linked to a higher temperature and a lower humidity. A low temperature and high moisture seem to slow down these evolutions. However the binder behaviour does not explain the whole mix behaviour as the kinetics between them are not always similar. Thus other 20 mechanisms are yet to be found and taken into account to fully understand cold mix asphalts behaviour.


Introduction
Cold mix asphalts are a composite material obtained by mixing a bitumen emulsion (bitumen 5 droplets dispersed in an aqueous phase) with aggregates. At the final stage aggregates are linked together by the organic binder after the emulsion break. The material stiffens with time while water drains off in a phase called curing.
These materials are mostly used for pavement (road repair and structure of low traffic roads).
This type of material could be an answer to the ecological issue. Contrary to hot mix asphalts 10 which are manufactured at 160 °C, cold mix asphalts process requires emulsified bitumen and cold and wet rocks, saving energy to heat it up and preventing fumes generation. Nevertheless, cold mix mechanical behaviour during curing is relatively unknown and final performances are difficult to predict. The presence of water and its removal are one of the main differences compared to hot mix asphalt and they greatly influence the evolution of such a material. 15 Furthermore it is known that the curing of emulsion cold mix asphalt is highly dependent on external parameters such as the climatic conditions [1,2] or the composition [3]. Apart from increase of stiffness, the curing also influences other aspects of the material as stiffening and oxidation (ageing) of the binder, as is has already been studied for hot and cold bituminous mixes [4,5,6,7]. 20 To improve knowledge on emulsion cold mix asphalts and thus facilitate their use, an assessment of their behaviour during curing and a better understanding of their curing mechanisms are needed. The ambition of this paper is to give an idea of the different aspects of curing on emulsion cold mix asphalts and understand the influence of the implementation 3 season. These issues were assessed by comparing the mix mechanical evolution with the physicochemical characteristics of the binder with time. This comparison will give information about the different physical and chemical phenomena involved in the curing process. More precisely, cold mix asphalt samples were tested by oedometer test all along curing and their binder was extracted and studied at different curing times, with IR, rheology and calorimetry. 5 The rheological evolution of the binder allows a direct understanding of the influence of the binder on the whole mix. A change in glass transition measured with calorimetry could account for binder weakening during curing. The oxidation (quantified with IR) and increase of modulus of the binder would indicate its ageing.

Materials
The emulsion cold mix asphalt samples were made with a 70/100 bitumen emulsion at 65/35 bitumen-water ratio and virgin aggregates. The residual binder content was 4.55% and the total water content 6.50%. The material was compacted at 10% void content in cylindrical molds 15 (120 mm diameter, 60 mm height).

Conditioning procedures
All the samples were cured in a BIA CL1-30 climatic chamber at the following parameters: -Three samples were cured at 35 °C and 20 % RH during two months followed by 10 °C 20 and 80 %RH for two months. This group is called group A.
-Two samples went through two months at 10 °C and 80 %RH and then two months at 35 °C and 20 %RH. This group is called group B.

4
The 35 °C and 20 %RH curing parameters have been numerously applied in the literature [6,7,8,10] in particular when used for 15 days to simulate two or three summers of curing onsite. In this study, they are used to simulate warm season. The 10 °C -80 %RH parameters are chosen to simulate a cold season.
The samples were then removed from the chamber at specific times and separated into two 5 testing groups: a batch of samples was tested with oedometric testing while the binder was extracted from samples of the other batch all along the curing process.
2.3 Test procedure 2.3.1 Oedometric tests 10 The mechanical behaviour of the samples was evaluated with a Schenck Prüfrahmen Typ press to measure the oedometric moduli at different curing times according to Lambert's method [9]. This technique uses cyclic compression of a specimen that is set in a cylindrical mold (120 mm diameter) to calculate the stiffness of a non-cohesive material. As the sample is contained in the mold, only vertical displacements occur. The determination of the modulus is calculated by 15 dividing the stress amplitude with the strain amplitude from the stress-strain cycles obtained with oedometer testing.
Along with the oedometer tests, these samples were regularly weighed. Assuming that when the weight was minimal the samples were dry, the water content versus time was deduced with the equation 100% × (msample(t) -msample(min))/(msample(t=0) -msample(min)). The binder was extracted from the samples using an Asphalt Analysator and tetrachloroethylene as solvant. Prior to the extraction process, those samples were lyophilized with an Alpha 1-2 plate LDplus (Bioblock-Christ) freeze dryer in order to remove water, without using drying and heating steps to prevent further curing and oxidation. The collected binder was then centrifuged and distilled to remove the sand from the mix and the 5 perchloroethylene from the extraction. After this procedure, the rheological behaviour, the glass transition and the oxidation of the extracted binders were studied.

Complex modulus measurement
The rheology measurements were done using an Anton Paar MCR102 DSR. Two plate-10 plate sample holders were used: an 8 mm diameter one for a range of temperature from 20 to -10 °C (5000 Pa imposed stress) and a 25 mm diameter one for a range from 20 to 80 °C (500 Pa imposed stress). Both ranges of temperatures were tested between 0.1 and 100 Hz.

DSC
15 Two different samples were tested in calorimetry for each extracted binder with a Mettler Toledo DSC3 DSC, the data was retrieved during an increase of temperature at 10 °C/min. The data given in the results section of this paper are the mean values between the two tests.

20
Finally, the FTIR tests consisted in measuring the oxidation levels SO and CO of the binders by transmission on thin film, on five samples for each extracted binder. The device used was a Perkin Elmer Spectrum. 6

3.1
Evolution of the water content and samples geometry

Rheological behaviour of the mixture
The stiffness modulus of the samples was assessed with an oedometer device [9]. Usually, other types of mechanical tests (indirect tensile or two-point bending for example) are used for the 5 characterization of asphalt mixes as they are already very stiff at an early age. Asphalt cold mixes take time to cure and become more cohesive to be tested with these devices, they are very weak at young age. Therefore the oedometer test appears to be a suitable testing technique to measure their mechanical behaviour.   Moreover, these evolutions can be compared to the water content in the samples ( fig. 6).
Actually, in a general way, the modulus is lower when the water content is higher, and vice 10 versa, when the water content is low, the modulus is higher. Besides the important stiffening at the beginning of curing for group A happens at the same time as the fast water loss of the samples ( fig.1-a). In addition, the sudden increase of stiffness at the very beginning of the warm season for group B (around 70 days) can be correlated to the sharp decrease of water content at the same time ( fig.1-b). But the presence of water does not explain the whole behaviour of the 15 material as it continues to evolve even when the water content is stable [10].    Thus a cold period appears to hinder the stiffening of a binder, in cold mix asphalts.  The series of the first 4 points (0, 14, 29 and 59 curing days at 10 °C -80 %RH) on figure   11 show an even more significant trend to the one seen on figure 10 for the same simulated season: during the 2 months at 10 °C and 80 %RH, group A has lost half the stiffness that B has, and B has gained in phase angle where A has not. This difference in kinetics between A 20 and B during cold conditions may be related to their initial states : group A has already been through a warmer period which may have started physicochemical processes that B has not seen yet, which would explain the inertia of softening seen on figure 10 between points 4 and 5.
Then, after the transition to 35 °C -20 %RH for group B (point 5), the binder gets even stiffer and elastic than its initial state, which implies a certain reversibility in the mechanisms which occur during a cold period.
It is important to keep in mind that here the curing is only of 4 months in a climatic chamber this is why the data changes are very low. What is interesting is to extrapolate those results to longer curing times, where the changes would be far more significant, but the trends would 5 probably remain the same. Infra-red oxidation levels were measured on the extracted binders of the samples from groups A and B. Figure 13 shows the carbonyl (figure 13-a) and sulfoxide (figure 13-b) 15 oxidation indexes of these specimens at different curing times.  The coefficients are presented in Table 1 and 2. These coefficients can help to understand at what amount the stiffness increased : if the coefficient is close to 1, then the 15 modulus did not evolve significantly ; if the coefficient is lower than 0, the modulus decreased with time ; finally if the coefficient is higher than 1 then the stiffness improved. Table 1 represents the coefficients between the final and the initial shear modulus and Table   2 the same coefficients on the oedometric modulus, for each group and each modality, tested at 15 °C and 10 Hz. The 15 °C data for the extracted binder were extrapolated from the curves of 20 |G*| versus test temperature to allow an easy comparison between the mix and the binder data.  However, the results concerning the asphalt mix (Table 2) are mildly different as the total (« Total curing process ») and the 10 °C -80 %RH coefficient of group B are a bit higher than the data from group A. More precisely, during the first modality (10 °C -80 %RH) the stiffness of group B increases as much as the stiffness of group A (around 2 for both groups) whereas throughout the second modality, the coefficients of group B are a bit higher than the one of 5 group A. This shows that different phenomena are at stake depending on the curing parameters:

Glass temperature and oxidation levels
when the first modality is a cold one (10 °C -80 %RH), not every curing physicochemical processes are triggered, which results in a continuation of stiffening after transition to the warm modality (35 °C -20 %RH). On the contrary when the first modality is a warm one, the majority of the curing processes are induced, that is why the stiffening coefficients of the cold season as 10 second modality are close to 1.
Finally the coefficients for each modality are reasonably close between the binder and the mix except for the first modality of group B: the mix coefficient reach 2 whereas the binder coefficient only 0.6. This indicates that the binder evolution only is not enough to justify the mix evolution: other phenomena come into play. A hypothesis could deal with the bitumen- 15 aggregate contacts quality or the void geometry under the influence of water.
Concerning this part, it is important to keep in mind that these coefficients may also slightly differ with the curing time chosen for the calculations. For instance the very initial state (curing time = 0) of the mixes are not known as the oedometer tests have begun at 3 days of curing, this may explain the proximity in coefficients between both groups for the first modality ; if the 20 initial time was 0, the coefficients of A would be higher than for B. Moreover there is no oedometer data between 49 and 71 days for group A so its coefficients may be fairly underestimated.

Conclusion
This paper has the purpose of bringing some new data concerning emulsion cold mix asphalts, giving leads and distinguishing the main phenomena influencing curing, and showing that it is often interesting to compare different scales of the studied material to understand its behaviour 5 as much as possible.
The main results are : (1) The thickness of the samples varies with the water content which can influence the stiffness, (2) Regardless of the simulated season sequence inside the climatic chamber, based on The binder evolution does not entirely explain the mix evolution. Other physicochemical processes are to be taken into account to completely fathom this type 20 of materials, such as the influence of water or the void geometry.
These materials appear to be an attractive way towards a more reasoned use of bitumen as less bitumen is required to implement them. Contrary to hot mix asphalts, their behaviour progresses with time and more research could be done to improve their optimal performances.