Effect of ageing on bitumen chemistry and rheology

Ž .Construction and Building Materials 16 2002 15�22 Effect of ageing on bitumen chemistry and rheology Xiaohu Lu�, Ulf Isacsson Di�ision of Highway Engineering, Royal Institute of Technology, SE-100 44 Stockholm, Sweden Received 15 October 2000; received in revised form 16 July 2001; accepted 25 October 2001 Abstract Effect of ageing on bitumen chemistry and rheology was studied. Seven bitumens were aged according to the thin film oven test Ž . Ž .TFOT and the rolling thin film oven test RTFOT . The binders were characterised using infrared spectroscopy, chromatogra- phy and dynamic mechanical analysis. Statistical correlation between different chemical parameters, as well as between chemical and rheological parameters, was examined. The relationship between TFOT and RTFOT was also investigated. It was observed that ageing influenced bitumen chemistry and rheology significantly. However, chemical and rheological changes were generally not consistent, and consequently, ageing susceptibility of bitumens may be ranked differently when different evaluation methods are used. Regardless of the type of the parameters measured, a strong correlation was observed between TFOT and RTFOT, and the two ageing procedures show similar severity. � 2002 Elsevier Science Ltd. All rights reserved. Keywords: Bitumens; Ageing; Chemical analysis; Rheological characterisation 1. Introduction Bitumen ageing is one of the principal factors caus- ing the deterioration of asphalt pavements. Important ageing related modes of failure are traffic and ther- mally induced cracking, and ravelling. In bitumen age- ing, two types of mechanisms are involved. The main ageing mechanism is an irreversible one, characterised by chemical changes of the binder, which in turn has an impact on the rheological properties. The processes contributing to this type of ageing include oxidation � � � �1�3 , loss of volatile components 4,5 and exudation Žmigration of oily components from the bitumen into . � �the aggregate 6 . The second mechanism is a re- � �versible process called physical hardening 7�9 . Physi- cal hardening may be attributed to molecular structur- Žing, i.e. the reorganisation of bitumen molecules or .bitumen microstructures to approach an optimum � Corresponding author. Tel.: �46-8-7908703; fax: �46-8-108124. Ž .E-mail address: [email protected] X. Lu . thermodynamic state under a specific set of conditions � �1,10 . Bitumen ageing occurs during the mixing and con- struction process as well as during long-term service in the road. The circumstances at different ageing stages vary considerably. The factors affecting bitumen ageing include characteristics of the bitumen and its content in the mix, nature of aggregates and particle size dis- tribution, void content of the mix, production related factors, temperature and time. All these factors oper- ate at the same time, making the process of bitumen ageing very complex. As summarised in a state-of-the- � �art report 11 , a number of laboratory methods have been used in the quantitative determination of bitumen ageing at various stages of the production process as well as in service. The simulation of field ageing entails increasing the temperature, decreasing bitumen film thickness, increasing oxygen pressure, or using combi- nations of these factors. Kinetics of bitumen ageing � �may vary with test conditions 3,12 . In this paper, two standardised methods, the thin Ž .film oven test TFOT, ASTM D 1754 and the rolling 0950-0618�02�$ - see front matter � 2002 Elsevier Science Ltd. All rights reserved. Ž .PII: S 0 9 5 0 - 0 6 1 8 0 1 0 0 0 3 3 - 2 ( )X. Lu, U. Isacsson �Construction and Building Materials 16 2002 15�2216 Ž .thin film oven test RTFOT, ASTM D 2872 , were used to age bitumens. The effect of ageing on bitumen chemistry and rheology was investigated using infrared spectroscopy, chromatography and dynamic mechanical analysis. Correlation between chemical and rheological changes was examined. 2. Experimental 2.1. Bitumens Seven bitumens, denoted B1 to B7, were used in this study. The penetration grades of B1, B2 and B7 were 60, 85 and 370, respectively, and the other bitumens were penetration 180. The crude source of bitumens B1, B2, B3 and B7 was Venezuela, and the sources of B4, B5 and B6 were Mexico, Saudi Arabia and Russia, respectively. 2.2. Ageing procedures Ageing of bitumens was performed using TFOT Ž . Ž .ASTM D 1754 and RTFOT ASTM D 2872 . Stan- dardised conditions, i.e. 163 �C and 75 min for RTFOT, and 163 �C and 5 h for TFOT, were used. The aged bitumens were evaluated by measuring their rheologi- cal properties and chemical characteristics. ( )2.3. Fourier transform infrared FTIR spectroscopy ŽA FTIR spectrometer, Infinity 60AR Mattson, reso- �1 .lution 0.125 cm , was used to determine the functio- nal characteristics of bitumens before and after ageing. Five percent by weight solutions of bitumens were Ž .prepared in carbon disulfide. Blank solvent and sam- ple scans were performed using circular sealed cells Ž .ZnSe windows and 1 mm thickness . All spectra were obtained by 32 scans with 5% iris and 4 cm�1 resolu- tion in wavenumbers ranging from 1900 to 500 cm�1. 2.4. Thin-layer chromatography with flame ionization ( )detection TLC�FID Ž .In the TLC�FID, 2% w�v solutions of bitumens were prepared in dichloromethane, and 1 �l sample solution spotted on chromarods using a spotter. The separation of bitumen into four generic fractions Ž .saturates, aromatics, resins and asphaltenes was per- formed by a three-stage development using n-heptane, Table 1 Effect of ageing on bitumen composition Ž .Bitumens IR absorbance Generic fractions % Carbonyl compounds Sulfoxides Saturates Aromatics Resins Asphaltenes B1-unaged 3.68 0.71 10.3 54.9 23.5 11.3 B1-TFOT 4.52 0.84 10.7 51.2 27.6 10.5 B1-RTFOT 4.34 0.79 8.8 53.4 26.0 11.4 B2-unaged 3.55 0.73 10.7 56.7 21.2 11.4 B2-TFOT 4.91 0.90 10.4 48.2 28.5 12.9 B2-RTFOT 4.88 0.91 9.8 49.7 27.2 13.3 B3-unaged 3.86 0.68 14.2 62.5 13.2 10.1 B3-TFOT 5.53 0.90 13.0 47.6 27.8 11.6 B3-RTFOT 5.45 0.92 12.0 49.8 25.6 12.6 B4-unaged 0.49 0.65 9.3 62.7 19.1 8.9 B4-TFOT 1.40 1.21 10.0 52.5 29.2 8.3 B4-RTFOT 1.49 1.00 9.0 60.0 22.7 9.2 B5-unaged 0.49 0.62 9.0 64.2 19.3 7.5 B5-TFOT 1.24 1.05 8.9 53.0 30.2 7.9 B5-RTFOT 1.31 1.13 10.2 53.2 29.6 7.0 B6-unaged 4.97 0.67 17.1 54.6 20.9 7.4 B6-TFOT 5.94 1.14 17.3 44.9 29.8 8.0 B6-RTFOT 5.72 0.91 17.1 43.8 30.1 9.0 B7-unaged 4.91 0.76 15.2 54.1 22.6 8.1 B7-TFOT 5.45 0.83 13.8 50.6 25.7 9.9 B7-RTFOT 5.45 0.81 13.3 52.2 24.2 10.3 ( )X. Lu, U. Isacsson �Construction and Building Materials 16 2002 15�22 17 Žtoluene and dichloromethane�methanol 95�5 by .volume , respectively. The fractions were determined Žby means of Iatroscan MK-5 analyzer Iatron Labora- .tories Inc., Tokyo, Japan . 2.5. High performance gel permeation chromatography ( )HP-GPC The HP-GPC system used was Waters 515 HPLC pump equipped with Waters 410 differential refrac- tometer. Three ultra-styragel columns were arranged in ˚the order of pore size of 100, 500 and 500 A. The system temperature was 35 �C. In the analysis, 5% by weight solutions of bitumens were prepared in tetrahy- Ž .drofuran THF and 50 �l sample solution was injected into the column. The flow rate of the THF mobile phase was 1 ml�min. To calibrate the instrument, a series of polystyrene standards were used. ( )2.6. Dynamic mechanical analysis DMA ŽDMA was carried out using a rheometer RDA II, .Rheometrics . Temperature sweeps with 2 �C incre- Ž .ments from �30 to 135 �C were performed at 1 rad�s and frequency sweeps were applied over the range of 0.1�100 rad�s at 25 and 60 �C. Parallel plates with � 8 mm�1.5 mm gap and � 25 mm�1 mm gap were used in the temperature ranges of �30 to 50 �C and 40 to 135 �C, respectively. The strains used varied with tem- perature to ensure tests were within the linear vis- coelastic region. 3. Results and discussion 3.1. Functional groups As mentioned earlier, oxidation is the most impor- tant ageing mechanism. It can be verified and quantita- tively measured by functional group analysis using FTIR. A typical infrared spectrogram is shown in Fig. 1. The absorbance bands at 1705 cm�1 are due to the ŽC�O stretch in carbonyl compounds e.g. ketones, .carboxylic acids and anhydrides , while those at 1030 cm�1 are due to the S�O stretch in sulfoxides. Conse- quently, the peak areas at the two wavenumbers may be considered as concentration measures of carbonyl compounds and sulfoxides, respectively. As indicated in Table 1, carbonyl compounds and sulfoxides are formed in the process of TFOT and RTFOT and the degree of the oxidative changes is dependent on the bitumen. For most of the bitumens tested, there is a small difference between TFOT and RTFOT with regards to formation of the functional groups. The mechanism of bitumen oxidation is very com- plex. It could be that oxidation of methylene and Fig. 1. Effect of ageing on bitumen FTIR sepctrogram. degradation of unsaturated chains and�or naphthenic rings of benzene systems lead to ketones and carboxylic acids, respectively, and oxidation of thio-ethers to sul- foxides. In addition, aromatization and chain scission may occur during oxidative ageing, which do not result � �in oxygen incorporation in the bitumen 1 . The func- tionalities formed should introduce an increase in the overall polarity of the bitumen, which in turn will influence bitumen rheology. 3.2. Generic fractions The effect of ageing on the chemical composition of bitumens was also studied using TLC�FID. Using this method, four generic fractions, namely saturates, aro- matics, resins and asphaltenes, were determined. As indicated in Fig. 2 and Table 1, ageing decreases aro- matics and at the same time increases the content of resins and asphaltenes. However, the content of satu- rates changes slightly due to their inert nature to oxygen. Since the fractionation of bitumens is mainly based on molecular polarity, the compositional changes should imply transformation of different fractions, i.e. aromatics� resins� asphaltenes. Similar to FTIR observation, a small difference is observed between TFOT and RTFOT as measured by changes in bitumen Ž .generic fractions Table 1 . 3.3. GPC parameters In HP-GPC, the sample components are eluted in order of decreasing molecular weight. An example of HP-GPC profiles is given in Fig. 3, and accordingly, two GPC fractions, Fraction-I and Fraction-II, are evalu- ated. For unaged bitumens, the elution times of the two fractions are 15.5�19.5 and 19.5�29.0 min, respec- Ž .tively; the transition point of elution time 19.5 min shifts approximately 0.5 min after bitumen ageing. The GPC results are summarised in Table 2. In determining ( )X. Lu, U. Isacsson �Construction and Building Materials 16 2002 15�2218 Fig. 2. Effect of ageing on bitumen TLC-FID chromatogram. molecular weights, a series of polystyrene standards were used. The contents of fractions were determined by area normalization of chromatograms. As can be seen, during ageing, the content of Frac- Ž .tion-I large molecules is increased at the expense Ž .oxidation of Fraction-II. However, for most of the Ž .bitumens, the weight average M and number aver-w Ž .age M molecular weight of the two fractions changen slightly after ageing. On the other hand, increases in Ž .molecular weight and polydispersity M �M arew n observed for the whole bitumen system when subjected Fig. 3. Effect of ageing on bitumen GPC chromatogram. to ageing; this is due to the content changes of Frac- tions-I and -II. Using GPC analysis, it can be demon- strated that association of smaller molecules with higher polarity may contribute to the high molecular weight � �fraction of the bitumen 13,14 . This means that the increased content of Fraction-I may also be indicative of the formation of highly polar functional groups during ageing. Table 2 also indicates that TFOT and RTFOT result Table 2 Effect of ageing on bitumen GPC parameters Bitumens M M M �M M -I M -II M -I M -II I-% II-%n w w n n n w w B1-unaged 699 3970 5.69 10 600 449 12 900 1280 17.1 82.9 B1-TFOT 753 4760 6.32 9100 425 12 600 1080 28.8 71.2 B1-RTFOT 776 5070 6.53 9500 424 13 300 1100 28.6 71.4 B2-unaged 840 4170 4.96 11 900 685 14 800 1580 19.3 80.7 B2-TFOT 890 5280 5.94 10 400 640 14 400 1360 29.9 70.1 B2-RTFOT 792 4980 6.29 9130 569 13 200 1190 31.8 68.2 B3-unaged 670 3320 4.95 10 000 547 12 300 1340 18.0 82.0 B3-TFOT 742 4430 5.97 8940 526 12 200 1170 29.4 70.6 B3-RTFOT 753 4580 6.08 8810 526 12 200 1140 30.9 69.1 B4-unaged 817 2860 3.50 11 380 729 12 900 1630 10.9 89.1 B4-TFOT 868 3410 3.93 11 000 737 12 800 1590 16.2 83.8 B4-RTFOT 864 3460 4.01 10 600 716 12 700 1520 17.3 82.7 B5-unaged 750 3040 4.05 11 200 660 13 000 1580 12.7 87.3 B5-TFOT 800 3580 4.47 10 800 667 12 900 1550 17.8 82.2 B5-RTFOT 798 3640 4.56 10 900 662 13 000 1550 18.2 81.8 aB6-unaged 726 6040 8.31 8780 508 12 100 1130 32.1 67.9 B6-TFOT 759 6840 9.01 8610 505 15 100 884 35.9 64.1 B6-RTFOT 745 6660 8.95 8320 498 11 700 1080 35.9 64.1 B7-unaged 537 3270 6.09 10 700 546 13 600 1420 20.9 79.1 B7-TFOT 580 4370 7.54 9430 546 12 800 1270 30.3 69.7 B7-RTFOT 583 4580 7.86 9210 540 12 900 1220 33.0 67.0 a Fraction-I of unaged and aged bitumen B6 consists of two narrow peaks. ( )X. Lu, U. Isacsson �Construction and Building Materials 16 2002 15�22 19 Fig. 4. Correlation between TFOT and RTFOT as evaluated by GPC. Ž .in similar changes in GPC parameters Fig. 4 , and ageing sensitivity is dependent on the bitumens. The results are consistent with those obtained using FTIR and TLC�FID. 3.4. Rheological measurements Bitumen is a viscoelastic material, which displays either elastic or viscous behaviour, depending on tem- perature and time of loading. At sufficiently low tem- peratures and�or short loading times, bitumen behaves essentially as an elastic solid. As temperature increases and�or loading time increases, the viscous property of bitumen becomes more obvious. At sufficiently high temperatures and�or long loading times, bitumen is essentially a Newtonian liquid, and can be described by a shear rate independent viscosity value. The tempera- ture and time dependence of bitumen rheology may be changed by the process of ageing, as illustrated in Figs. 5 and 6, as well as in Table 3. The results show that, with ageing, the complex Ž � . Ž .modulus G , storage modulus G� and the loss mod- Ž . Ž .ulus G� increase and the phase angle � decreases. These indicate that ageing makes the mechanical properties of the bitumen more solid-like. The magni- tude of the changes is dependent on bitumen type and evaluation conditions. For example, at 25 �C and 10 Ž .rad�s, the complex modulus may increase by 1 B6 to Ž . Ž .105% B7 and phase angle may decrease by 3 B6 to Ž .10% B1 after ageing. On the other hand, ageing has Ž � .minimal effect on dynamic moduli G , G� and G� at Ž .low temperatures �0 �C . A significant decrease in phase angle is generally observed at temperatures of Ž .0�70 �C Fig. 5 . Since phase angle is a measure of the Žratio between loss modulus and storage modulus tan � .�G��G � , the decreased phase angle implies that ageing leads to a greater increase in storage modulus than in loss modulus. However, at high temperatures Ž .�40 �C , the increase in complex modulus is mainly Ž .attributed to the increased viscous component G� . Fig. 5 also indicates that at a temperature range of 0�50 �C the dynamic moduli become less susceptible to temperature after ageing. In the plot of phase angle versus temperature, three DMA characteristic temperatures, T , T and T , can15 45 75 be defined. T is the temperature at which phase angle45 Table 3 Complex modulus and phase angle at 10 rad�s and two different temperatures Ž . Ž .Bitumens Complex modulus Pa Phase angle � 25 �C 60 �C 25 �C 60 �C B1-unaged 8.39E5 2.77E3 69.7 85.4 B1-TFOT 1.53E6 4.97E3 62.4 82.3 B1-RTFOT 1.37E6 5.25E3 62.7 81.7 B2-unaged 5.65E5 1.92E3 72.7 86.2 B2-TFOT 8.47E5 2.99E3 66.3 84.0 B2-RTFOT 8.97E5 3.82E3 65.1 82.9 B3-unaged 1.29E5 6.86E2 77.3 88.0 B3-TFOT 2.50E5 1.10E3 71.4 85.8 B3-RTFOT 2.49E5 1.43E3 70.6 85.7 B4-unaged 1.62E5 6.43E2 75.6 88.5 B4-TFOT 2.52E5 8.13E2 70.5 87.3 B4-RTFOT 2.62E5 1.02E3 70.5 87.1 B5-unaged 2.22E5 7.75E2 75.0 88.2 B5-TFOT 3.79E5 1.77E3 69.7 87.3 B5-RTFOT 3.70E5 1.76E3 69.6 86.7 B6-unaged 1.66E5 4.75E2 61.1 85.1 B6-TFOT 1.87E5 7.51E2 57.6 82.7 B6-RTFOT 1.67E5 1.12E3 59.3 82.3 B7-unaged 4.37E4 2.25E2 78.5 89.1 B7-TFOT 8.98E4 4.60E2 75.2 87.8 B7-RTFOT 9.02E4 4.35E2 75.3 87.9 ( )X. Lu, U. Isacsson �Construction and Building Materials 16 2002 15�2220 Fig. 5. Dynamic modulus and phase angle as functions of tempera- ture at 1 rad�s for bitumen B1 before and after TFOT. Žequals 45 � i.e. loss modulus and storage modulus are .equal while T and T represent the temperatures at15 75 which phase angles equal 15 and 75 �, respectively. The Ž .elastic contribution G� to bitumen complex modulus Ž � .G would be dominant at temperatures lower than T , since in this case G� is less than 27% of G� and15 � 2 2'G � G� �G� G�. On the other hand, bitumen is essentially fluid as the temperature exceeds T , be-75 Fig. 6. Complex modulus and phase angle as functions of frequency at 25 �C for bitumen B1 before and after TFOT. Table 4 Effect of ageing on DMA characteristic temperatures Ž . Ž . Ž .Bitumens T �C T �C T �C15 45 75 B1-unaged �17 1 28 B1-TFOT �15 4 38 B1-RTFOT �16 4 39 B2-unaged �18 �2 23 B2-TFOT �17 0 34 B2-RTFOT �18 0 34 B3-unaged �23 �9 14 B3-TFOT �22 �6 24 B3-RTFOT �22 �6 28 B4-unaged �24 �4 18 B4-TFOT �24 �2 26 B4-RTFOT �23 �1 26 B5-unaged �22 �2 18 B5-TFOT �22 1 27 B5-RTFOT �22 1 28 aB6-unaged � 4 34 B6-TFOT � 6 39 B6-RTFOT � 6 40 B7-unaged �26 �13 7 B7-TFOT �25 �11 18 B7-RTFOT �25 �11 19 a Lower than �30 �C. cause of the negligible contribution of G� to G�. The interval between T and T can be considered as a75 15 temperature range of binder viscoelastic behaviour. Table 4 shows the effect of ageing on T , T and T .15 45 75 The temperatures were obtained by using temperature sweeps at a frequency of 1 rad�s. As expected, T and45 T are largely dependent on the bitumen and increase75 with ageing. Differences in T and T between the45 75 unaged bitumens are as large as 17 and 41, respec- tively, and ageing may increase T and T by 3 and 1045 75 �C, respectively. T also depends on source�type of15 Žthe bitumens and is slightly influenced by ageing Table . Ž .4 . Statistical investigation level of significance 0.01 indicated that T , T and T relate to the weight15 45 75 Ž .average molecular weight M . The correlation coef-w ficients are 0.62, 0.55 and 0.76, respectively. DMA measurements also show a good relationship between TFOT and RTFOT. Examples of correlation are given in Fig. 7. For the bitumens studied, TFOT and RTFOT show similar severity. 3.5. Ageing index Ageing susceptibility of bitumens may be evaluated ( )X. Lu, U. Isacsson �Construction and Building Materials 16 2002 15�22 21 Fig. 7. DMA comparison of TFOT and RTFOT aged bitumens. by means of an ageing index, which is defined as the ratio of a chemical or physical parameter of the aged binder to that of the original binder. Ageing indices obtained using chemical and rheological measurements are shown in Table 5. Relationships between different ageing indices have been examined. It was found that Ž .good statistical correlation R�0.93 exists between G� and M ratios, however, no correlation or veryw weak correlation exists between other ageing indices. This observation suggests that the ageing susceptibility of bitumens would be ranked differently when a dif- ferent ageing indices are used. It also indicates that the chemical and rheological changes of bitumens during the process of ageing are not necessarily consistent. 4. Conclusions Ageing influences bitumen chemistry and rheology significantly. Chemical changes include the formation of carbonyl compounds and sulfoxides, transformation of generic fractions, and increases in amount of large Ž .molecules or molecular association , molecular weight and polydispersity. As a result of the chemical changes, the mechanical properties of aged bitumens become more solid-like, as indicated by increased complex modulus and decreased phase angle. However, the chemical and rheological changes are generally not consistent. Consequently, ageing susceptibility of bitu- mens may be ranked differently when different evalua- tion methods are used. Using either chemical analyses or rheological measurements, a strong correlation is observed between TFOT and RTFOT, and the two methods show similar severity. Table 5 Ageing indices obtained using chemical and rheological measurements Bitumens B1 B2 B3 B4 B5 B6 B7 aTFOT ageing index Carbonyls ratio 1.23 1.38 1.43 2.86 2.53 1.20 1.10 Sulfoxides ratio 1.18 1.23 1.32 1.86 1.69 1.70 1.09 Resins ratio 1.17 1.35 2.11 1.53 1.56 1.43 1.14 M ratio 1.20 1.27 1.33 1.19 1.18 1.13 1.34w GPC Fraction-I ratio 1.68 1.55 1.63 1.49 1.40 1.07 1.45 � Ž .G ratio 25 �C, 10 rad�s 1.82 1.50 1.94 1.56 1.71 1.13 2.05 � Ž .G ratio 60 �C, 10 rad�s 1.79 1.56 1.60 1.26 2.28 1.58 2.04 aRTFOT ageing index Carbonyls ratio 1.18 1.37 1.41 3.04 2.67 1.15 1.10 Sulfoxides ratio 1.11 1.25 1.35 1.54 1.82 1.36 1.07 Resins ratio 1.11 1.28 1.93 1.19 1.53 1.44 1.07 M ratio 1.28 1.20 1.38 1.21 1.20 1.10 1.40w GPC Fraction-I ratio 1.67 1.65 1.72 1.59 1.43 1.08 1.58 � Ž .G ratio 25 �C, 10 rad�s 1.63 1.59 1.93 1.62 1.67 1.01 2.06 � Ž .G ratio 60 �C, 10 rad�s 1.90 1.99 2.08 1.59 2.27 1.58 1.93 aAgeing index is defined as the ratio of a chemical or rheological parameter after and before ageing. ( )X. Lu, U. Isacsson �Construction and Building Materials 16 2002 15�2222 Acknowledgements The authors would like to thank Jonas Ekblad, Britt Wideman and Clarissa Villalobos for their laboratory assistance during the testing programme. References � �1 Branthaver JF, Petersen JC, Robertson RE, Duvall JJ, Kim SS, Harnsberger PM, Mill T, Ensley EK, Barbour FA, Schabron JF. Binder Characterization and Evaluation, vol. 2: Chemistry, SHRP-A-368. Washington, DC: National Research Council, 1993. � �2 Tuffour YA, Ishai I. The diffusion model and asphalt age- hardening. Proc. Assoc. Asphalt Paving Technol. 1990;59:73�92. � �3 Petersen JC, Branthaver JF, Robertson RE, Harnsberger PM, Duvall JJ, Ensley EK. Effects of physicochemical factors on asphalt oxidation kinetics. Transport. Res. Record 1993; 1391:1�10. � �4 Traxler RN. Relation between asphalt composition and harden- ing by volatilization and oxidation. Proc. Assoc. Asphalt Paving Technol. 1961;30:359�377. � �5 van Gooswilligen G, De Bats FTh, Harrison T. Quality of paving grade bitumen � a practical approach in terms of functional tests, Proceedings of 4th Eurobitume Symposium, vol. I, Madrid, October 1989, pp. 290�297. � �6 Curtis CW, Ensley K, Epps J. Fundamental Properties of Asphalt�Aggregate Interactions Including Adhesion and Ab- sorption, SHRP-A-341. Washington, DC: National Research Council, 1993. � �7 Brown AB, Sparks JW, Smith FM. Steric hardening of asphalts. Proc. Assoc. Asphalt Paving Technol. 1957;26:486�494. � �8 Traxler RN, Coombs CE. Development of internal structure in asphalts with time. Proc. Am. Soc. Testing Mater. Ž .1937;37 II :549�557. � �9 Bahia HU, Anderson DA. Glass transition behaviour and phys- ical hardening of asphalt binders. J. Assoc. Asphalt Paving Technol. 1993;62:93�129. � �10 Petersen JC. Chemical composition of asphalt as related to asphalt durability: state of the art. Transport. Res. Record 1984;999:13�30. � �11 Johansson LS, Lu X, Isacsson U. Ageing of Road Bitumens � State of the Art, TRITA-IP FR 98-36. Sweden: Royal Institute of Technology, 1998. � �12 Verhasselt AF, Choquet FS. Comparing field and laboratory ageing of bitumens on a kinetic basis. Transport. Res. Record 1993;1391:30�38. � �13 Hattingh MM. The fractionation of asphalt. Proc. Assoc. As- phalt Paving Technol. 1984;53:197�215. � �14 Dukatz Jr EL, Anderson DA, Rosenberger JL. Relationship between asphalt flow properties and asphalt composition. Proc. Assoc. Asphalt Paving Technol. 1984;53:160�185.