Click for next page ( 25


The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 24
24 As shown, all three tests generally meet the requirements SAFT may be overwhelming volatile loss, especially given the established for an improved short-term binder aging test. Each low amounts of volatiles that have been collected by previous has been shown to be relatively simple, applicable to both researchers using this test. Whether adequate mixing of air and neat and modified binders, and to reasonably reproduce the binder will occur at the lower temperatures needed for extend- level of aging that occurs in the RTFOT. Procedures for extend- ing the test to long-term aging is an issue that was addressed ing the RCAT and the MGRF to long-term aging at reason- experimentally in NCHRP Project 9-36. able temperatures and without pressure have been developed, Another important consideration for an improved short- but these will require modification and further development term aging procedure is the ability to quantify the amount of to adapt them to specification testing in the United States. volatile material lost from the binder. Only the SAFT offers a Although a long-term aging procedure has not been devel- method for directly measuring the volatile compounds lost oped for the SAFT, it appears that the device is extendable to during the aging process. It includes a condenser for collect- long-term aging, provided sufficient dispersion of air can ing the volatile compounds from the exhaust air from the be accomplished in stiff modified binders at temperatures vessel. The volatile compounds trapped by the condenser can around 90C to 100C. This may require redesign of the stir- be weighed to determine the amount of material lost and can ring mechanism to adequately disperse air in highly viscous be analyzed to determine their composition. Although this media. The flexibility to design this portion of the device is a approach requires additional development to reduce the level distinct advantage for extending the test to long-term aging. of variability reported in the initial development studies, it is better than determining mass change by weighing the filled 3.2.3.2 Promising Tests vessels before and after aging as is done with the RTFOT and MGRF. The change in mass after short-term aging is the net The RCAT and MGRF are conceptually similar and also result of a loss in mass due to the loss of volatile compounds very similar to the RTFOT. Both involve exposing a relatively and an increase in mass due to oxidation. It appeared that the thin film of binder to air or oxygen. The binder film is renewed MGRF could be modified to collect volatile compounds using by rotating the vessel containing the binder. The RCAT vessel an approach similar to that used in the SAFT. and the MGRF both include elements to mix the binder using gravity. The RCAT includes an internal shaft mechanism that keeps the binder evenly dispersed in the vessel. In the MGRF, 3.3 Selection Study the indentations in the Morton flask mix the binder during the 3.3.1 Introduction short-term test. In the long-term test proposed in Germany, steel balls are added to a smooth flask to provide mixing. The The primary finding from the review of the literature and MGRF is, by far, the more appealing of these two devices from research in progress was that the MGRF and SAFT are both the perspective of cost, equipment availability, and testing promising approaches for an improved short-term aging pro- time for long-term aging. The test can be assembled from off- cedure to be used in the United States with AASHTO M320. the-shelf components for a reasonable cost. The RCAT is The equipment required for both tests is relatively inexpensive, more expensive (approximately $18,000) and available from and they are easy to perform, applicable to both neat and mod- only a single supplier. Because the long-term RCAT requires ified binders, and reasonably reproduce the level of aging that 1 week to complete, and the equipment for the MGRF is more occurs in the RTFOT. The major issue unresolved through the readily available at a lower cost, the RCAT was not considered literature review and review of research in progress was which for further development in NCHRP Project 9-36. of these two tests was best suited for extension to long-term The SAFT uses a different approach in which air is dispersed aging. Only limited data was available on a long-term version of in the binder using a stirring mechanism. It is essentially a the MGRF, and no long-term aging data was available for the small, laboratory-scale air blowing still. A thin film is not pro- SAFT. The selection study was designed to investigate the feasi- duced in this device. Instead, air from a nozzle submerged in bility of extending these tests to long-term aging. Since cost, the binder is dispersed in the binder by an impeller attached complexity, and ability to simulate the RTFOT were judged to to an external motor. It appears that for short-term aging, this be similar for the two tests, the extendibility to long-term aging approach is much more efficient at mixing air with the binder became an important factor in selecting the short-term test because the duration of the test is approximately one-half to method for further development in Project 9-36. one-quarter of that for the tests involving thin films with sim- The selection study was conducted in two parts. The first ilar temperatures and airflow rates. It is not known whether part of the study was an assessment of various modifications the more rapid aging observed in the SAFT is the result of that could easily be made to the MGRF and SAFT to produce more rapid volatile removal or more rapid oxidation. Of con- prototype long-term versions of these tests. The goal in this cern is the possibility that more vigorous oxidation in the effort was to obtain approximately the same degree of aging

OCR for page 24
25 that occurs in the PAV. The second part of the selection study i - RTFOT was a formal experiment designed to address whether the RA = 100% (1) degree of aging in the prototype long-term versions of the PAV - RTFOT tests is affected by the large differences in viscosities for neat and modified binders at the selected long-term aging temper- where ature of 100C. This section presents key findings from the RA = relative long-term aging selection study. The complete selection study report is included i = dynamic viscosity for configuration i, measured as Appendix B. at 60C, 0.1 rad/s RTFOT = dynamic viscosity for RTFOT aged, measured at 60C, 0.1 rad/s 3.3.2 Development of Initial Prototype PAV = dynamic viscosity for PAV aged, measured at Long-Term Aging Tests 60C, 0.1 rad/s 3.3.2.1 MGRF The investigation of various alternatives for a long-term Attempts to develop a prototype long-term version of the version of the MGRF procedure started with the MGRF, MGRF that approximates the aging produced in the PAV which uses a 2,000-mL Morton flask. The alternatives that focused on methods to enhance mixing and to create a film were investigated included a 2,000-mL round flask with the that is continuously renewed within the flask. This was accom- addition of steel balls and rollers to enhance mixing and for- plished by adding various mixers and scrapers and varying the mation of a film and the use of scrapers to create and renew rotational speed of the flask. For all of this testing, a tempera- the film. As shown in Figure 3-9, the Morton flask was mar- ture of 100C, an airflow rate of 36 L/h, and an aging time of ginally successful for long-term aging for the PG 58-28 binder, 48 hours were used. Table 3-7 summarizes the chronological but was not successful for aging the PG 82-22 binder. The use order of the various configurations that were attempted. Fig- of a round flask with steel balls, as used in Germany, increased ure 3-9 compares the degree of aging achieved with each con- the aging of the PG 82-22 slightly while the use of rollers that figuration relative to the aging obtained with the PAV. conformed to the shape of the flask did not. The simple scrap- Schematic diagrams of selected configurations are shown in ers designed to fit in a round flask appear to remove much of Figures 3-10 through 3-13. The measure of the degree of aging the viscosity effect, resulting in similar aging of the PG 58-28 shown in Figure 3-9 is defined by Equation 1. The relative and PG 82-22 binders, but the degree of aging after 48 hours aging according to this equation is simply the change in vis- is only one-third of that obtained in the PAV. cosity above RTFOT aging caused by the prototype long-term test divided by the increase in viscosity that occurs during PAV 3.3.2.2 SAFT aging. For all equipment configurations and binders, relative aging is reported based on the dynamic viscosity measured at Attempts to develop a prototype long-term version of the 60C and 0.1 rad/s. SAFT that approximates the aging produced in the PAV Table 3-7. Summary of long-term rotating flask configurations tested. Number Flask Mixer Speed Figure Observations 1 Morton None 4 rpm Not 1. Adequate film for PG 58-22 binder. shown 2. Does not produce a moving film for PG 82-22 binder. 3. Low relative aging for both binders. 2 Morton 3 Steel Balls 1 rpm Figure 1. Not used with PG 58-28 binder. 3-10 2. Does not produce a moving film for PG 82-22 binder. 3. Low relative aging for PG 82-22 binder 3 Round 1 Football- 1 rpm Not 1. Not used with PG 58-28 binder. Shaped shown 2. Does not produce a moving film for PG 82-22 binder. Roller 3. Low relative aging for PG 82-22 4 Round 2 Football- 1 rpm Figure 1. Not used with PG 58-28 binder. Shaped 3-11 2. Does not produce a moving film for PG 82-22 binder. Rollers 3. Low relative aging for PG 82-22. 5 Round Single 1 rpm Figure 1. Does not produce a film for PG 58-28 binder. Scraper 3-12 2. Generated a renewed film for PG 82-22, but film thickness increased with aging time. 3. Low relative aging for both binders. 6 Round Double 1 rpm Figure 1. Not used with PG 58-28 binder. Scraper 3-13 2. Generated a renewed film for PG 82-22. Film thickness relatively constant with aging time. 3. Low relative aging for PG 82-22 binders.

OCR for page 24
26 Less Aged Than PAV More Aged Than PAV Round + Double Scraper 37% 27% Round + Scraper 32% Round + 2 Rollers PG 58-28 7% PG 82-22 Morton + 3 Balls 15% 55% Morton 12% 0% 20% 40% 60% 80% 100% 120% 140% 60 C DYNAMIC VISCOSITY INCREASE, PERCENT OF PAV Figure 3-9. Relative aging from Equation 1 for various long-term rotating flask configurations. Figure 3-10. Schematic of 2,000-mL Morton flask Figure 3-11. Schematic of 2,000-mL round flask with with three steel balls. two football-shaped rollers.

OCR for page 24
27 Clamp to hold scraper stationary Figure 3-12. Schematic of 2,000-mL round flask with single scraper. focused on the design of an impeller that could efficiently mix same used in Figure 3-9 and defined in Equation 1. The relative air with the binders over a wide range of viscosities. The design aging according to this equation is simply the change in viscos- proceeded from the impeller used in the short-term version of ity above RTFOT aging caused by the prototype long-term test the test, which is very efficient at mixing air with low viscosity divided by the increase in viscosity that occurs during PAV binders, to a helix impeller which is efficient at mixing highly aging. The dynamic viscosity was measured at 60C, 0.1 rad/s. viscous fluids, and finally to a helix/turbine impeller which The original impeller worked well with the PG 58-28 binder, combines the benefits of both. For all of this testing, a temper- but it did not provide adequate mixing of the PG 82-22 binder. ature of 100C and an airflow rate of 36 L/h were used. When this impeller is used with extremely viscous materials, Table 3-8 summarizes the chronological order of the various the entire mass of material spins with the impeller. The helix configurations that were attempted. Figure 3-14 compares the impeller, which is frequently used to mix very viscous and degree of aging achieved with each configuration relative to the particulate-filled fluids, worked well with the PG 82-22 aging achieved with the PAV. Schematic diagrams of selected binder, but apparently did not disperse air as efficiently in configurations are shown in Figures 3-15 through 3-17. The the less viscous PG 58-28 binder. The helix/turbine impeller, measure of the degree of aging shown in Figure 3-14 is the which includes a helix to move the binder vertically in the Clamp to hold scraper stationary Figure 3-13. Schematic of 2,000-mL round flask with double scraper.

OCR for page 24
28 Table 3-8. Summary of long-term SAFT configurations tested. Number Impeller Speed Schematic Observations 1 Original 700 Figure 3-15 1. Good mixing of PG 58-28 binder. rpm 2. Could not stir PG 82-22 binder. 3. High relative aging for PG 58-28. 2 Helix 220 Figure 3-16 1. Good mixing of both PG 58-28 and PG 82-22 binders. rpm 2. Better mixing of PG 82-22 binder occurs at lower speeds. 3. High relative aging of PG 82-22 binder. 4. Moderate relative aging of PG 58-28 binder. 3 Helix/ 350 Figure 3-17 1. Good mixing of PG 58-28, PG 82-22, and PG 76-22 binders. 4-Bladed rpm 2. Aging at 48 hrs exceeds PAV conditions for the PG 58-28 and Turbine PG 82-22. 3. Less difference in aging between PG 58-28 and PG 82-22 than observed with helix. 4 Helix/ 350 Not shown 1. Good mixing of PG 58-28, PG 82-22, and PG 76-22 binders. 8-Bladed rpm 2. PAV aging obtained at approximately 40 hours. Turbine vessel, and a turbine to mix air with the binder, resulted in the at 60C, 0.1 rad/s. The degree of aging appears to increase with best performance over the range of binders investigated. At increasing binder stiffness, which is counterintuitive. 48 hours, the degree of aging obtained in the PG 58-28 and The testing described earlier found that it is possible to PG 82-22 binders exceeded that obtained in the PAV. extend the SAFT to a long-term aging test. The following sec- The last iteration of the impeller design was a helix/turbine tion discusses the experiment on viscosity effects that was impeller with eight turbine blades. With this impeller, PAV conducted to quantify the significance of the differences aging conditions were reached after approximately 40 hours. between the aging of the PG 58-28 and the PG 82-22 binder Figure 3-18 shows the degree of aging obtained with this con- shown in Figure 3-18. figuration for the three binders included in the selection study. The measure of the degree of aging shown in Figure 3-18 is the 3.3.3 Viscosity Effects Experiment same used in Figures 3-9 and 3-14 and defined in Equation 1. The relative aging according to this equation is simply the Only the final iteration (a helix/turbine impeller with eight change in viscosity above RTFOT aging caused by the proto- turbine blades) of the long-term version of the SAFT was sub- type long-term test divided by the increase in viscosity that jected to the experiment on viscosity effects. Table 3-9 sum- occurs during PAV aging. The dynamic viscosity was measured marizes the testing conditions for the long-term SAFT. Less Aged Than PAV More Aged Than PAV 57% Helix/Turbine 24 hrs 63% 71% Helix/Turbine 48 hrs 125% 136% PG 76-22 LDPE Helix 24 hrs 21% 40% PG 58-28 Neat PG 82-22 SBS Helix 48 hrs 65% 95% Original 48 hrs 85% 0% 20% 40% 60% 80% 100% 120% 140% 60 C DYNAMIC VISCOSITY INCREASE, PERCENT OF PAV Figure 3-14. Relative aging from Equation 1 for various long-term SAFT configurations.

OCR for page 24
29 Temp. Probe 600ml/min. Probe to 600ml/min. Probe to Air Temp. Air Temp. Control Control 220 rpm 700rpm Heating Heating Mantle Mantle Figure 3-15. Schematic of long-term SAFT with Figure 3-16. Schematic of long-term SAFT with original impeller. helix impeller. The viscosity effects experiment was designed to investi- gate the effect of viscosity on the degree of aging that occurs in the long-term SAFT. This was accomplished by aging split View A 3-D samples of RTFOT-aged PG 58-28 and PG 82-22 binders in the PAV and the long-term SAFT, and comparing rheological properties at high, intermediate, and low pavement tempera- tures. The following properties were measured for the RTFOT, PAV, and long-term SAFT: Shear modulus and phase angle from a DSR frequency 600ml/min. sweep at 60C using frequencies from 0.1 to 100.0 Hz. Air Clamp Shear modulus and phase angle from a DSR frequency sweep at 25C using frequencies from 0.1 to 100.0 Hz. Creep stiffness and m-value at 60 sec from BBR tests con- Bath 350 rpm ducted at -12C. Wax Three independent samples were aged in the long-term SAFT and PAV and tested as outlined above. Regression analysis A was used to compare the rheological properties between the PAV and the long-term SAFT. Details of the statistical analysis are presented in Appendix B on the project webpage. The sta- tistical analysis found that there was a statistically significant difference in aging between the PAV and the long-term SAFT and that the difference was binder dependent. Table 3-10 summarizes the key differences and compares them to the Figure 3-17. Schematic of long-term SAFT with single operator coefficient of variation for the DSR and BBR helix/turbine impeller.

OCR for page 24
30 Less Aged Than PAV More Aged Than PAV 58-28 NEAT 0.80 76-22 LDPE 0.93 82-22 SBS 1.15 0% 20% 40% 60% 80% 100% 120% 140% 60 C DYNAMIC VISCOSITY INCREASE, PERCENT OF PAV Figure 3-18. Relative aging from Equation 1 for final iteration of the long-term SAFT (helix/eight-bladed turbine, 200 rpm, 36 L/h airflow, 100C, 40 hours). tests. The bias of the long-term SAFT relative to the PAV is ence was most evident in the high pavement temperature approximately twice the AASHTO single-operator precision; tests, but also occurred in the intermediate and low pave- therefore, these biases have engineering significance as well ment temperature tests that are used in AASHTO M320. statistical significance. Probably more important than the Differences between the aging produced by the long-term finding that the aging was different between the PAV and SAFT and the PAV appear to be temperature dependent. the long-term SAFT was the fact that the two binders aged The differences were greater at the upper grading tempera- differently. The PG 82-22 binder aged more in the long- ture than at the lower grading temperature. This implies term SAFT than in the PAV, while the PG 58-28 binder aged that the two aging procedures produce materials that are more in the PAV than in the long-term SAFT. This differ- different rheologically. There are two possible explanations for the binder effect. First, the helix/turbine impeller and its rotational speed may Table 3-9. Testing conditions for not be properly optimized for lower viscosity binders. Second, the long-term SAFT. the air dispersion mechanism in the long-term SAFT may Condition Value age polymers more than, or in a different way than, the high- Sample Size 250 g pressure aging occurring in the PAV. Additional testing of Aging Temperature 100C neat and modified binders, both having a wide range of con- Impeller Type Helix + 8-Bladed Turbine Impeller Speed 350 rpm sistency, is needed to determine the cause of this effect and to Airflow Rate 36 L/h further improve the long-term SAFT. This additional testing Aging Time 40 hours was beyond the scope of NCHRP Project 9-36. Table 3-10. Comparison of long-term SAFT bias with DSR and BBR precision. Long-Term SAFT Bias Relative to AASHTO Single Property PAV, % Operator Coefficient PG 58-28 PG 82-22 of Variation, % G* at 60 C -10 +14 7.9 G* at 25 C -6 +13 7.9 S at -12 C +3 +7 3.2 m at -12 C -3 +3 1.4