Investigation of effect of bitumen chemical composition, polymer and warm mix asphalt additives on the properties of bitumen by using response surface method

Bitumen, a binder material used in asphalt applications, is one of the important components that affects the service-life of asphalt. Bitumen chemistry, and amounts of Saturate, Aromatic, Resin, Asphaltene (SARA) fractions within the bitumen, affect the properties of both base and modified bitumen. The desired properties of bitumen change with varying content of bitumen and purpose of application. To improve the properties and the performance while using in the asphalt mixtures usually the polymer addition is utilized. While thermoplastic elastomers are mostly preferred, plastomers such as polypropylene, ethylene vinyl acetate, and natural / synthetic rubber can also be used for the modification of bitumen. The addition of materials like Styrene-Butadien-Styrene (SBS) copolymer can simultaneously develop the durability and flexibility at high temperatures by containing the both soft and hard segments in its structure. In addition, hot mix asphalt additives can be used to prepare bitumen-aggregate mixtures at lower temperatures in order to improve the application conditions. For this purpose, organic or chemical additives included into the bitumen, as well as processes with the foaming techniques can be used for lowering the mixing and layout temperatures. Influence of all these additives on the properties of bitumen changes with the type and amount of additive and bitumen chemistry. An additive applied to improve the some properties of bitumen may adversely affect another feature of the bitumen and consequently asphalt. Especially when the addition of these additives together and variety in the bitumen chemistry are taken into consideration, the desired properties may not be obtained in the final product. Therefore, understanding the simultaneous effect of the variation of bitumen composition, amount of polymer and warm mix additive on the bitumen properties is aimed in this study. Experimental statistical methods were used for this purpose and the effects of independent parameters on the response functions individually and simultaneously were investigated by using Response Surface Method. The composition of bitumen and the amount of polymer and warm mix asphalt additives were defined as independent variables in this study. Dynamic shear rheometer (DSR), Bending beam rheometer (BBR), Elastic recovery and Viscosity were determined as response functions of this system. The effects of these independent variables on the response functions were investigated by the full factorial analysis initially and then the response surface method. In the experimental study, SARA analysis was performed on the bitumens which are in the same penetration grade and obtained from Izmit, İzmir and Batman refineries. Furthermore, Fourier Transform Infrared Spectroscopy (FTIR), Simulated Distillation Gas Chromatography (GC-SimDis), Differential Scanning Calorimetry (DSC) and Thermal Gravimetric Analysis (TGA) were performed for detailed characterization of the bitumens. Styrene-Butadiene-Styrene (SBS) tri-block copolymer and a commercial paraffin wax were used as warm mix asphalt additives for the modifications of the bitumens. FTIR, DSC and TGA analysis were carried out to determine whether these additives can be used in bitumen blends. When the results of characterization tests and chemistry of bitumen were considered together, it can be seen that the same penetration grade bitumens may have different chemical properties and when the performance grade was determined according to TS EN 14023 standardization it may be also found different. This result shows that it cannot be achieved same performance grade by applying the same modifications to the same penetration grade bitumens which have different chemical compositions. Moreover the difference of SARA fractions in the three bitumens could not be seen in the FTIR analysis. The reason for the lower initial boiling point observed in GC-Simdis and the low temperatures at which loss of mass starts for Batman bitumen in TGA analysis are due to the light fractions of the Batman bitumen rather than the amount of saturated content. On the other hand, the high amount of aromatic and saturated structures in the bitumen decreased the glass transition temperature of Izmit bitumen. For the characterization analysis of the additives, it has been proved that the properties of the additives are the same as those of the commercially available common ones having without extra features. The Gaestel index (GI), one of the independent parameters, was used to define the bitumen chemistry in a single variable and examine the effect of bitumen chemistry on the response functions. The GI values of İzmit, İzmir and Batman bitumens were calculated by using the determined SARA fractions as 0.487, 0.622 and 0.988, respectively. In the experimental design matrix, 3 discrete values (0.487, 0.622, 0.988) were defined for the GI as discrete numeric variable, and 2-5%, 1-3% upper and lower limit values for the SBS and Wax, respectively, which were defined as continuous variables. Within these limits, bitumen modifications were made according to the experimental matrix designed by JMP® software and DSR, Viscosity, Elastic recovery and BBR tests which were determined as response functions of system, were carried out. Model equations were constructed with response functions of independent variables by using the data obtained from the measurements and the response surface method. Three-dimensional response surfaces and two-dimensional interaction curves are plotted by using the model data that indicates the effect of each parameter on the properties of bitumen. For the individual analysis, all independent variables were found as significant in different degrees for DSR and BBR response functions. The increasing amount of SBS promotes the value of DSR and creep stiffness exponentially. However this increasing trend affects the creep stiffness negatively since it has the upper limit in the standard. Furthermore the bitumen composition which was defined as GI in the analysis, was an effective term on the all response functions especially for elastic recovery, viscosity and DSR properties which improve rapidly with the increasing value of GI. The higher amount of asphaltene and saturates or lower amount of resins and aromatics enhance the higher value of GI. On the other hand, the amount of wax were determined as insignificant on the viscosity and elastic recovery properties of bitumen despite the fact that the wax is known as an organic additive for viscosity decreasing in the asphalt mixtures. This may be the result of the fact that the viscosity decreasing effect of Wax is very small, comparing with the increasing effect of the SBS and GI. Finally, the optimum values of variables were determined by performing the simultaneous analysis of independent variables with the four response functions together within the desired performance limits. The optimization analysis showed that all variables have effects on the response functions in different degrees. When the result of simultaneous analysis of response functions were compared to the individual analysis, there were terms added or eliminated from the model expression. Especially the GI*GI term which is important for the effect of bitumen composition on the response function was found as insignificant for DSR, Viscosity and Elastic recovery responses in simultaneous analysis. The reason for this was shown as the given limitation for the maximization of DSR and Elastic recovery and the minimization of viscosity. The optimum variable values within the given response limits are GI: 0.988, SBS: 3.2% and Wax: 3%. Tests within the scope of EN 14023 standard and asphalt tests were carried out on modified bitumen prepared in optimum values to investigate the improvement on the performance on both bitumen and asphalt properties. The performance grade of neat Batman bitumen was found to be PG 70-22 while it was determined as PG 82-16 after the modification at optimum conditions. As a result of the modification, while the high temperature performance of bitumen improves as two performance grade, the low temperature performance was affected negatively. The reason maybe that all the modified bitumen samples has the result on BBR as -16 since the addition of SBS increases the creep stiffness and although the amount of SBS is minimum (%2), the performance criterion was not achieved on -22. As a result of the Marshall asphalt design for determining the asphalt properties, aggregate-bitumen blends had 4.8% of the unmodified bitumen, whereas it was necessary to use 6.2% of the modified bitumen. Asphalt samples prepared by using the amounts of bitumen found in design and tested on the wheel tracking test machine. The amount of rutting were found as 13% for unmodified bitumen, while in the modified bitumen this value was improved by 50% to 6.7%. Thus, the DSR, viscosity, elastic recovery and creep stiffness properties of the bitumen have been successfully modelled based on the bitumen content, amount of polymer and warm mix asphalt additives and optimum mixing ratios have been successfully determined by using the "Response Surface Method". The widespread use of experimental design methods will provide a scientific and realistic basis for the preparation of asphalt mixture formulations instead of personal experience as well as the prevention of waste of time.