Modified Bitumen

Laboratory Investigation of Polymer Modified Bitumens

Ulf lsacsson and Xiaohu Lu

Roya! Institute ofTechnology, Division of Highway Engineering, S100 44 Stockholm, Sweden

ABSTRACT

This paper deals with laboratory characterization of various pure and polymer modified bitumens, in which five bitumens from four sources and six polymers of different types were used. The effects of base bitumen and polymer type and content on compatibility, storage stability, rheology and aging were investigated. The results indicated that compatibility and storage stability of polymer modified binders were largely dependent on polymer content and were influenced by the charncteristics of base bitumens and polymers.

Polymer modification increased the elastic responses and dynamic moduli of bitumens at intermediate and high temperatures, and influenced complex and stiffness moduli at low temperatures. It also reduced the temperature susceptibili ty and the glass transition and limiting stiffness temperatures of birumens. The degree of the improvement generally increased with polymer content, but varied with bitumen sourceJgrade and polymer type.

Artificial aging may cause oxidation of bitumen and degradation of polymer, and consequently, change the microstructure and rheological properties of the modified binders. These changes were largely influenced by polymer nature and content. Due to temperature and frequency dependence, aging index is not a suitable parameter for characterization of aging susceptibility of polymer modified binders.

The effectiveness of polymer modification as reflected by different testing parameters may vary considerably. The usefulness of empirical test methods (e.g. penetration, softening point and Fraass breaking point) for characterizing polymer modified bilwnens is discussed. To predict significant benefit from polymer modification in practice, it is necessary to use appropriate and fundamental pararnetm .

Keywords: polymer modified birumens, compatibility, storage stability, temperature susceptibility, rheology, low-temperature creep, aging

INTRODUCTION

The behaviour of bitumens in service is governed by their initial engineering properties as well as by the mechanical and environmental conditions to whieh they are subjected. To enable pavements to accommodate increasing traffic intensity and axle loads in varying climatic environments, high quali ty bilwnens are required. Special binders are also needed for other applications, such as bridges, runways and slurry seals. These examples suggest a potential area for bitumen modification.

The use of polymers as bi tumen modifim is not a new phenomenon. As early as 1823, an English cork manufacturer was granted a patent for a binder containing natural rubber. After the Second World War, synthetic polymer began to compete with natural rubber as an additive in road bitumen. Over the years, an increased interest in bitumen modification using different types of synlhel ic polymers has been obselVed in many countries (1-5).

Despite numerous investigations in this area, polymer modified bitumens have not yet been characterized unambiguously, because of the complex nature of bitumen and the complex interaction between bitumen and polymer (1 , 6, 7).

This paper presents a laboratory investigation of a number of pure and polymer modified bitumens. Five bitumens from four sources and six polymers of different types were used. The main objective of the work is to study fundamental properties of the binders, which include rheology, temperature susceptibility, compatibility, storage stability, low-temperature creep responses and aging characteristics. The effects of bitumen source/grade and polymer type and content on these properties are evaluated. Relationships between parameters of dynamic mechanical analysis, creep test and conventional methods are also examined.

MATERIALS AND TEST METHODS

Materials Five base bitumens, denoted A, B, C, 0 and E, were used in this study. They were

obtained from Venezuela (Laguna), Mexico, Saudi Arabia and Russia, respectively. The physical properties of the bitumens are given in Table I.

Table 1 Physical Properties of Ibe Bitumens Used in This Study

Bitumen code A 8 C 0 E Gn'" 885 BISo BISo BI80 BI80

Soo~, Venezuela Venezuela Mexico _.

Russia Penetration 5C, dmm 84 113 18. 149 '81 Softening point C "J 46.0 ]S.S 41.0 42.8 44.3 Viscosity (al1]5C, mm2fs (e) 3S4 21. 208 236 172 fraus breakin~ point, C to/ 11 16 17 16 22 p,

-1.00 -0.89 -0.09 -0.10 1.29 PVN -0.60 -0.53 -0.57 -0.59 -0.85 6' (e)

'" '" . . ,. ''' ' ASTM 0 5, ASl M 0 3, ASTM 021 0, IP 80, Penetration mdex (8), Penetration

vi5(:osity number (9).

The polymers investigated are styrene-butadiene-styrene (SBS), styrene-ethylene-butylene-styrene (SEBS), ethylene vinyl acetate (EVA) and ethylene butyl acrylate (EBA) copolymers. The SSS polyme~ used are powdered ](raton D-1I01 and Kraton D-1184, supplied by the Shell Chemical Company. Kraton 0-1101 is a linear SBS polymer containing 31 percent styrene, and Kralon-1184 a branched polymer containing 30 percent styrene. The SEBS polymer is Kralon G 1650 (Shell), which contains 29 percent styrene. The two EVA copolymers used are Elvax 260 (EVA I) and EJvax 420 (EVA2), supplied by DuPont. Melt indices (MI) of Elvax 260 and Elvax 420 are 6 and ISS, and their vinyl acetate contents are 2S and IS percent, respectively. The EBA polymer was produced by Neste and supplied by Nynas Petroiewn.

Polymer modified bitwnens (totally 36) were prepared using a low shear mixer (RW 20 DZM-P4, Janke & Kunkel) at ISOC and a speed of 125 rpm. The polymer contents used were ], 6 and 9 percent by weight of blend. In preparation, 600 g of the bitumen was heated to fluid condition and poured into a 2000 ml spherical flask . Upon reaching 175C, a preweighed amount of polymer was added to the bitumen. Mixing was then continued at ISDe for two hour.;. After completion, the polymer-bitumen was removed from the flask and divided into

l

small containers. The blend was cooled to room temperature, sealed with aluminium foil and stored for further tesling. The process for preparing the modified binders exposed the bitumen to high temperature and air for an cxtended lime, which led to hardening of the bitumen. For the accurate evaluation of polymer effects, the base bitumens were also subjected to the same treatment as the polymer-bitumen blends.

ConvtntionQ/ Bindu Tuts The standard methods used include softening point (ASTM D 36), penetration (ASTM

D 5), kinematic viscosity (ASTM D 2170) and Fraass breaking point (IP 80).

Dynllmic Mechanical Analysis (DMA) DMA with frequency sweeps (0. 1 to 100 radls) and temperature sweeps (-30 10 13Sq

was carried out using a Rheometries rheometer (RDA II). Parallel plates, gap 1.5 mm for l 8 mm and I mm for CD 25 nun, were used in a temperature range of )0 to SOC and 40 to 135C. respectively. In temperature sweeps, when reducing the temperature, the nonnal force is kept at zero by slightly reducing the gap. The gap was calibrated at the starling temperature. The effect of the thennal expansion of the instrument is deducted during the increase of the temperature. A sinusoidal strain was then applied by an actuator. The actual strain and torque were measured and input to a computer for calculating various viscoelastic parameters such as storage modulus (G,), loss modulus (G~), complex modulus (G') and phase angle (0).

Creep Tests Using a Bending Beam Rheometer (BBR) Creep tests were carried out at four different temperatures (-35, 25, -15 and -IOC)

using a bending beam rheometer (BBR), Cannon Instrument Company. In tests, the binder was heated to fluid condition, poured into the mold and then allowed to cool at room temperature for about 90 minutes. The sample was cooled to approximately -SC for I minute and demolded. After demolding, the sample beam (125 mm long, 12.5 mm wide and 6.25 mm thick) was submerged in a constant temperature bath and kept at each test temperature (starting at -35C) for 30 minutes. A constant load of 100 g was then applied to the rectangular beam of the binder which was supported at both ends by stainless sleel half-rounds (102 mm apart), and the deflection of center point was measured continuously. Creep stiffness (5') and creep rate (m) of the binders were detennined at several loading times ranging from 8 to 240 seconds.

Thin Layer Chromatography with Flame Ionization Detector (TLC-FID) In TLCFID, the sample to be analyzed is dissolved in a solvent and SPOiled at one end

of a chromarod (quartz rod coated with a thin layer of sintered silica or alumina). The rods are then developed with suitable solvents, after which the solvents are removed by heating. The rods are scanned at a chosen speed through a hydrogen flame, and the separated fractions successively vaporizedipyrolYled. The FlO signals from each fraction are amplified and recorded as separate peaks. In this study, 8 commercial equipment, an latroscan MK5 analyzer (latron Laboratories Inc., Tokyo, Japan), was used. 2 percent (w/v) solutions of bitumens were prepared in dichloromethane, and [ IJI sample solution spotted on chrnmarods using a spotter. The separation of bitumen into four generic fractions (saturates, aromatics, resins and asphaltenes) was perfonned by a three-stage development using n-heptane, toluene and dichloromethanelmethanol (9515 by volume), respectively.

Gel Permeation Chromatography (GPC) [n GPC, molecules that differ in size are separated by passing the sample through a

stationary phase consisting of porous cross-linked polymeric get. The pores of the gel exclude

,

molecules larger than a certain critical size, while smaller molecules can permeate the gel slruCtun~ by diffusion. The sample components art eluted in order of decreasing size or molecular weight. The GPC systems used were three ultra-styrage! columns arranged in the order of pore size (100, 500 and 500A) with a refractive index detector (Water 410). In the analysis, 5 percent binder solutions were prepared in tetrahydrofuran (THF). A broad molecular weight polystyrene standard was used 10 calibrate the instrument.

Fluorescence Microscopy Fluorescence microscopy was used to study morphology (compatibility) of polymer

modified binders. The modified binder was illuminated using a blue light for excitation and the fluorescent yellow light re-

,

The weight average (M ... ) and number average (M.) molecular weights as well as polydispmity (MJM.) of the base bitumens were detennined by means of OPC. Figure I gives examples of bitumen GPe chromatograms. A small difference was observed between the ope data of bitumens C and 0, which correlates well with the results obtained using TLC-FID. Compared to other bitumens, bitumen E contains more of the larger molecular species; the fraction of molecules with M ... over 2500 glmole is 18 percent. This bi tumen also shows the widest molecular weight distribution (polydispersity 2.49) of the five bitumens.

Table 2 Chemieal Charaderistits of Bue Bitumens

Bitumen Code A B C 0 E Saturates % 11.3 13.0 10.4 10.9 16.9 Aromatics % 54.7 56.4 64.8 63.6 51.1 Resins % 18.7 17.3 III 14.8 18.3 Asphahenes (% J 5.4 III 11.1 10.7 J]J Molecular weight, M. 650 600 750 140 720 Molccular weight. M. ])00 ])00 1400 1300 1800 Polydispmiry (MJM.) 1.1l6 2.08 1.84 1.80 2.49 Fraction with Mw >2.Sxl0l (%) IJ 11 11 II 18

Conventional Hindu ruts The test resul ts obtained using conventional methods are presented in Table 3. As

expected. polymer modification caused an increase in binder consistency (decrease in penetration and increase in kinematic viscosity and softening point). The consistency changes may increase with polymer content, but vary with bitumen source/grade and polymer type. For a given polymer content, the bi tumens were observed to be less susceptible to EVA addition compared to the other polymers. However, in some cases, changes in the consistency as measured by those conventional parameters were not consistent.

Polymer modification also reduced Fraass breaking point of the binders. The reduction was comparably small (a few degree). For most of the modified binders, Fraass breaking point did not decrease proportionally with increasing polymer content. Since the Fraass test applies a high strain rate to binder samples, the actual differeoce in the fracture properties of the base and modified bitumens may be reduced by this test. This means that Fraass breaking point probably cannot accurately relate polymer perfonnance to low temperature properties of binders.

Tahle 3 Binder Test Re.sults Obtained Using Conventional Methods

Binder Viscosity Penelrati()l'l Softening Fraass PI PVN @mc. @25C. poin!, C breaking mm1/s dmm I ooin!, C

Bitumen A 370 77 47.3 ." -0.86 -0.64

A + 3Y, linear SBS 821 64 52.1 ." -0.1 1 0.32 A + W. linear SBS 2020 49 71.5 12 3.25 1.23

A + 9% linear SSS 4780 37 84.0 ." 4.08 2.00 A + 6!-', branched SSS 3830 50 82.3 ." 459 2. 12 A + 6Y. SEBS 2740 . 0 71.8 16 2.55 1.38

A + 6Y. EVAI 1860 54 59.5 1' 1.12 1.24 A "'6% EVA2 1080 54 63.5 12 1.90 0.49 A +6% EBA 2020 44 74.8 Il 3.25 1.09 Bitumen B 224 165 39.5

." -1.l0 .Q.55 B + 3% linear SBS 506 III 55J 16 2.70 0.40 B + 6Y, linear SBS 1060 .. 17.2 17 5.61 1.10 B+9Y. linearSBS 3350 70 85.1 19 5.95 2.44 B + 6Y, branched SSS 2390 84 82.5 18 6.14 2.24 B+ J!/tSE8S 577 III 58.0 19 lOS 0.48 8 + 6Y, SE8S 1630 61 66.8 18 2.84 L22 B +9Y. SEBS 3950 46 78 .S ll 3.90 2.05 8+3%EVA I 469 141 42.2

." .(l.61 0.48 8 +6%EVAI 1140 97 53.4 20 1.48 IJ' B+ 9%EVAI 2970 76 59.8 19 2.15 2.40 B+3% EVA2 342 129 46.8

." 0.64 -0.16 B+6%EVA2 602 96 57.4 20 2.39 0.35 B+9%EVA2 1070 77 64.4

." 3.10 0.9) B+)%EBA 472 121 51.6 20 1.80 0.28 8 +6%EBA 1180 82 70.2 18 4.33 1.16 8 +9'hIEBA 3040 61 77.8

." 4.58 2.10 Bitumen C 237 151 41.5 16 .{l.44 -0.51 C + )% linear SSS 461 120 51.2 18 1.60 0.23 C + 6% linear SBS ])50 .. 77.2 20 5.61 1.45 C + 9% linear SBS 3450 67 85J 20 5.81 2.42 C + 6% branched SBS 3890 89 85.4

." 6.12 3.06 C+6%SEBS 1490 67 70.5

." 3.74 1.22 C+6%EVAI 940 100 57.7

." 2.60 1.09 C+6%EVA2 584 .. 59.8 21 2.62 0.20 C + 6Y. ESA 1080 n 7J.l 20 4.06 0.85 Bitumen D 250 110 410 1 6 -0.6 1 -0.66 D + 6% linear SBS 1220 80 75.2 19 5.00 1.17 D + 6% branched SBS 4360 77 84.0 19 6.12 2.98 Bimmen E 187 166 44.0 ll 0.72 -0.85 E+ 6% linear SBS 1640 70 75.2 26 4.58 1.4\ E + 6Y. branched SSS 3220 69 81.5 25 5.52 2.37 .

. The but bitumens have been subJet!ed to the same treatment IS the polymer bitumen blends .

Compatibility and Storage Sillbilily Fluorescence Microscopy Studies Due to the chemical dissimilarity (e.g. differences in

molecular weight, polarity and structure) between bitumen and polymer, most of the modified binders obtained by mechanical mixing are thermodynamically immiscible. In describing such systems, the term ~compati bility" is often used. Compatibility can be defined as the slate of dispersion between two dissimilar components. The property was studied using fluorescence microscopy by characterizing the nature of the continuous phase and the fineness of the dispersion of the discontinuous phase. At room temperature and a magnification of 250, the modified binders may show either a continuous bi tumen phase with dispersed polymer particles or a continuous polymer phase with dispersed birumen globules, or two interlocked continuous phases. as illustrated in Figure 2.

iI'

..

. ' . ' .. '" . M ... ~. J

I . o

8itwnen 8 + 3% SEaS BilUmen B + 9% SEBS

~tr . , ' . . .

Bitumen B + 3% EVA] Bitumen B + 9'10 EVA]

Figure 2 Fluorescence Photomicrognilphs or Polymer Modified Bitumens

The morphology (compatibility) is the result of the mutual effect of polymer and bitumen, and consequently is influenced by bitumen composition and polymer nature and content. In general, at a low polymer content, the small polymer spheres swollen by bitumen compatible fractions (e.g. aromatic oils) are spread homogeneously in a continuous bitumen phase. Compared with 3 percent elastomer (SBS and SEBS) modified binders, those containing thermoplastics (EVA and EBA) of the same polymer conlent revealed much finer dispersion of the polymer. Belter dispersion of the polymers was also observed for the modified binders prepared from the bitumens with higher aromatic and lower asphaJtenc contents. By increasing polymer content, a continuous polymer phase may be obtained (Figure 2). The minimum pcrcentagl! of polymer to ensure the fonnation of its continuous phasc depends to a great extent on the base bitumen and the polymer itself. In most cases,

appearance of a continuous polymer structure was observed to begin a\ a polymer content of about 6 percent.

The morphology should also be influenced by testing temperature, This means that the morphology of polymer modified binders observed at ambient temperature does not necessarily reflect thai at high temperatures.

Phase Separalion Polymer modified binders may separate inlo JXllymcNich and asphaltenerich pbases during hot storage. The phase separation has been CQnfirmcd by tluorescence micrographs and Fourier Transform Infrared Spectroscopy (FTJR) of the modified binders before and after lube storage al 180C for three days (\ 0). The two phases differ widely in their rheological properties and chemical constitutions. In this study, the logarithm of the ratio of complex modulus (DMA al 25C) of the asphaltenerich phase to that of the JXI1ymer-rich phase is defined as the separation index, I" and is used to evaluate the slorage stability of the modified binders (Figure 3),

\.ooE+09 r----------------,

, l.ooE+08

l.ooE+07

--Bilumen A + 6% lOB'" bc:fln stonge - . _. Top sroion iIfIa storage Bottom sectiOll after $Iorage

. . .

-. .. . . . .

. ...

e LOOE+06 I~~~ .~ .. ~.q~._.; .. ~ .. -.~. ::::::l I, ~ l.ooE+05 - . -... _.- " _.

~ _.-' LooE+il4 U

l.ooE+{I3

I.OOE+{I2 I, " Jog/O(bonom section),IG' {top section)]

I.00E+{I1 L_~~=_~~~~~~~....J l.ooE-01 l.ooE+OO I.(IOE+OI l.ooE+{I2

Figure J Complex Modulus as a Function of Frequency for a EBA Modified Bitumen berore and after Storage atl80C ror Three Days

The investigation indicates that the storage stability decreases with increasing polymer content. For the modified binders with a low content of polymer (3 percent), almost no phase separation (I, close to 0) was observed. The storage stability is also influenced by characteristics of the base bitumen and the polymer. Higher content of aromatics and/or lower content of asphaltenes are observed to be favourable with regard to storage stability, as illustrated in Figure 4. Among all the modified binders prepared, the modified binders with EBA display the lowest storage stability. For the elastomer (SBS and SEBS) modified binders, those containing branched SBS show the lowest storage stability, For the EVA modified binders, better storage stability is observed for those containing the polymer with lower VA content and higher Mr.

The storage stability wa~ also assessed by softening point. Unfortunately, the conventional evaluation did not show any correlation with DMA method. This implies that different conclusions could be drawn from the storage stability if using different evaluation methods.

9

Since many polymer modified bitumens display a tendency towards phase separation, some agitation is recommended during the storage of hoI blends. However, such a measure should not be taken wilhoullimil and at the expense of the oxidative degradation of polymers.

2' ,-------~~~-. O''IIINIrSES

,.

, o!

. n~SBS

' . o

00 ,

. L~ __ ~~ __ ~____' H O 50.0 SSO 600 6S 0 700

2' ,-----__________ ,

2.'

; ! U ! 5 LO

" . ,

O' nlftMS8S " ''''_sas

o

o ,

o

.. '---~-----------' LO.O ,,. ... , ..

CMI .. L" ",.",,"II .. n 1% . , ",

Figure 4 Relationship between the SeparttioD Index (at 1 radls) or SBS Modifi ed Binder! lod Bitumen Composi tion

Comparison oj Storage Stability and CompatibiUry In general, storage stability of polymer modified binders may be associated with the compatibility between bilwnen and polymer. A quaJitalive comparison between separation index and fluorescence photomicrograph indicates that, at a given polymer conlenl, the modified binders showing a finer polymer dispersion generally exhibit less phase separation during static hot storage. In this case, the separation index (storage stability) can be regarded as a measurement of the compatibility of polymer modified binders. However, high phase separation was observed in modified binders with high polymer content showing a continuous polymer phase in their microstructures; the modified binders with such desirable microstructures cannot be judged as incompatible systems. Obviously, the compatibility of polymer modified binders defined in different ways or evaluated by different methods is not necessarily consistent. This also suggests that using only one method for determination of the degree of compatibility of polymer modified binders could be misleading.

DynIJmic Rhe%gic4/ Propertiu Dynamic rheological properties refer to responses of a material to periodically varying

strains or stresses (11). ln DMA, the ratio of the peak: stress to the peak strain is defined as the complex modulus (G ), which is a measure of the overall resistance to deformation of a material. lbe in-phase and out-Qf-phase components of G are defined as the storage modulus (G' ) and the loss modulus (G-), respectively. Storage modulus is proportional to the stress in-phase with the strain and provides information on the elastic responses of a material, while loss modulus is proportional to the stress out-of-phase with the strain and is associated with viscous effects. The phase difference between the stress and strain in an oscillatory deformation is defined as phase angle (li). This parameter is a measure of the viscoelastic character of the material. A purely viscous liquid and an ideal elastic solid demonstrate li of 90" and 0, respectively. The viscoelastic parameters of bitumens are functions of temperature and frequency, which may be modified by the addition of polymers.

Phase Angle and Viscoelastic Properties Phase angle as a function of temperature was detennined at I radfs over a temperature range from 30 to 135C. Examples are gh'en in Figure 5. It is evident that polymer modification increases the elastic response (decreased phase angle) of bitumens. The effectiveness of polymer modification is largely dependent on temperature, and is innuenced by bitumen source/grade and polymer type and content. As the temperature exceeds about 2OC, the polymers begin significantly to improve the elasticity of the modified binders. This can be attributed to the viscosity of the mobile bitumen components, which is low enough to allow the elastic network of the polymer to innuence the mechanical properties of the modified binders. With increasing temperature, the phase angle may pass through a maximum and then through a minimum, or show a smoothly increasing function of temperature, depending on base bitumen and polymer nature and content (Figure 5). The phase angle maximum and minimum correspond to the transi tion and plateau regions of storage modulus, respectively, which may be indicative of a continuoll5 IX'lymer network.

I ......

I ...... --,

I."'" --...... ~- ~

__ c .. _ _

'.01( ... --' !" I --........ -I . '" j"

I I . '" " ,-~ ,. --

--< ... _-__ c . .. __ l.,M --, -_ .... _-

,--~ . , " ,. , . ~ . , " ,. ON ,.

'-"' '-'"

' .oK ... l.oK"

I . "" 1 c ... .. ! " i UK'" uK'" , ,

, uK.f) ,

i ' .... M I , .... ,

.... _ .... -__ c ... _

I ...... ..

__ c ... ~ __ c .. ",-

,-, , ~ .. , "

,. ON ,. ~ .. , " '"

, . T ...... __ ra T ..... ~ .... ra

Figure 5 Storage Modulus and Phase Angle as a Function ofTemperatun at 1 radls

In this investigation, the contribution of storage modulus (G') to complex modulus (G-) is arbitrarily considered to be negligible as a > 75~, since in this case, G' is less than 27% of G", and JG 1 '" JGt G-' '" G- . The temperature above which the phase angle is greater than 75 is denoted T11' , and used to quantitatively compare the clastic response of the binders. II was observed thai polymer modification increased T15' of the binders. and the improvement may increase with polymer content. The improvement is also dependent on the base bitumen and polymer Iype. The innuence of the polymers tested may be ranked as branched SBS, SEBS/linear sas. EBA, EV A2 and EVA I, branched sas being the most

II

effective. This temperature is observed 10 statistically correlate with the softening point, as illustrated in Figw-e 6. For all the binders studied, the correlation coefficient obtaioed is 0.95. However, al the softening point, all the base bitumens are purely viscous, as indicated by phase angles of 90, while the modified binders still display considerable elasticity. This observation suggests that the softening point is related to dilTerent properties when applied to pun: and polymer modified bitumens.

JO " "

., 70 80 90 100 Soflnln~ Point [OC1

Figure 6 T 7S. as a FUDdioD of Softening Point

Polymer modification also influences the frequency dependence of phase angle, as illustrated in Figure 7. The improved viscoelastic properties of polymer modified binders can be observed over a wide frequency range. Although the frequency dependence of phase angle is similar for the base bitumens, significant differences can be observed among the modified binders. This may result from different degrees of molecular interactions between base bitumens and polymers.

,

. " io , ..

L.

,-

-_. -_ .. ... -

~_I .. ... -_ .......

._10" , oK ... ._."oM!

,

,

"

--. .. _c ... .. ~-< .... .... ... _ c ... .... ~_c ... _

Figure 7 Effect of Polymer Modification on .'requeney Dependenee of Phase Angle at 60C

Dynamic Moduli and High Temperature Properties As indicated in Figure 5, polymer modified binders exhibit four regions in the plots of storage modulus versus temperature: glassy, transition, plateau, and terminal or flow, similar to those for polymer materials. These regions can be associated qual itatively with different kinds of molecular responses (12). Different degrees of prominence may be observed in the four regions, depending on the chemical and physical nature (e.g. molecular weight, microstructure and network density) of the binders. Of these regions, the plateau corresponds to the phase angle minimum and may indicate the presence of polymer networks (cross-links and entanglements) and/or polymer crystallinity in the modified binders. For the base bitumens and modified binders with a low content of polymer, the plots display only three regions but not the plateau. In this case, the transition goes directly from the glassy region to the terminal region, and correspondingly, no phase angle maximum and minimum occur.

For optimal performance at elevated temperatures, the selected polymer should be able to form a continuous elastic network when dissolved/dispersed in the bitumen. This is determined by several factors such as the chemical and physical properties of the polymer and the bitumen, polymer content and the thermal/mechanical history of the blend. For example, for the modified binders containing a low content of polymer (e.g. 3 percent) and the modified binders with bitumen E, no apparent phase angle minima are observed (Figure 5). Correspondingly, the plateau region of these modified binders is not easily recognized, indicating that there is no, or a very weak, polymer network. Compared with thermoplastics (EVA and EBA), elastomers (SBS and SEBS) may result in a more pronounced plateau region. Due to the increased elastic response (increased storage modulus and decreased phase angle) and increased complex modulus, the SHRP rutting parameter, Gt/sinS, is improved. The magnitude of the increase is dependent on the bitumen source/grade and polymer content and type. It is also highly dependent on testing conditions (e.g. frequency and temperature), as ilhlslrated in Figure 8.

I.fllt.., r ---- - ----,

I.t_t!

-_. __ A ........ _ _ "I~[WI.,

... _ ..... f'/>J __ A ... ..

, __ " 1.101;'"

f,..,._,I" .... ) UOE" ,

I . ~IO n~---------

~ u N:>ti T UtE ...

~ I.M'" , . [ '" 1 [ '"

1 . ['" 1.1IK41 L~~ _ _ ~~~ __ ~

M .H 1. " " " ,. 1H '" T .. ,.. ... "1'Cl

Figure 8 C/sinS as a Fundion of Frequency (at 60C) and Temperature (It 1 rad/s)

Since in European countries, the contribution of the bitumen to the susceptibility of permanent deformation is usually assessed by testing the softening point, a comparison has been made between the SHRP rulling parameter and softening point (13). It was observed that no significant linear relationship existed between those two parameters if all the binders tested were analyzed together. However, such linear relationships could be established if modified bindern produced from one and the same bitumen were examined, as indicated in Figure 9. The correlation coefficients of the three regression lines are in the range 0.84 to 0.95.

15 r---------------------,

10

5

35 55

65

75

,

Sobnln" Poin t reI

R'" O.U

R" o.as

R" O.93

85 95 105

Figure 9 Relationship between G"sino (at 10 radls and 60C) and Softening Point

13

BLACK Diagram Polymer modification of bitumen rheology is also identified in Figure 10, where BLACK diagrams (complex modulus as a function of phase angle) are displayed. These diagrams were generated with temperature sweeps (from )0 to 135C) al 1 radfs. Evidenliy, al a low polymer conlent (3 percent in this study), the effect of polymer modification appears insignificant, and the behaviour of the modified binders remains close to thaI of the base bitumens. After modification with a sufficiently high polymer content (t?: 6 percent), the binders change fundamentally in their rheological behaviour, as indicated by a substantial decrease in phase angle (8 substantial increase in the elastic response) with decreasing complex modulus. In these cases, the rheological response of the modified binders is mainly imposed by the viscoelastic properties of the polymers, and those of the bitumens waning considerably. These observations imply that the rheological properties of polymer modified binders are governed by their continuous phase.

,~~

,-~

,--,

, r-

ltllK>M

.i UII[~ i t .. ... {I.." At ... ,f--

'f--,

1.l1li;"

1.l1li; ...

,-

, ~ 1:- ' k '\ ~ := I -

I. " M " ~ - ..... 1 ...

,.II1II:'"

-~ f .. _oM i ' .. U H l: UIK ....

{ '.111:'" ~ 1.lII:olII

1.11('" '.ME

,,,&4,

iR -_. ... _ ..... \1 ... __ ... "tv ...

-_ .. "' ......

,. " " M " ~ -......

Figure 10 Complex Modulus as a Function of Ph an Angle

"

The rheological changes are also influenced by the characteristics of the bitumens and the polymers. For the modified binders with SEBS or SSS, complex modulus deviales from that of the base bitumens a\ phase angles as low as lit' (corresponding to binder glassy slale al low temperatures). However. the curves of the modified binders with EVA and EBA are almost the same as those of the base bitumens at points where 0 5 500 and Gt ~ 10 MPa. These differences suggest that SEBS and sas may improve bitumen rheology over a wide temperature range, while EVA and EBA show their effect mainly at high temperatures.

Low-Tempuafure Propertits For a flexible pavement, one of the failure modes is low-temperature cracking. This

occurs when the Ulenna] stress induced at low temperatures exceeds the tensile strength of the asphalt pavement. Low-temperature cracking can be a serious problem in cold areas. To reduce the risk of crack.ing, the binder should have good flexibility (low stiffness and complex modulus and high ability of stress relaxation) at the lowest pavement temperature. In this paper, the influence of polymer modification on the low.temperature properties of bitumens is studied using DMA and BBR.

DMA CharaclerizaliQn The DMA observations indicate that, in low temperature regions (below OC), all the modified binders show a lower complex modulus titan the corresponding base bitumens. However, the glassy modulus G, of the binders is close to I GPa and insignificantly influenced by polymer modification. This value of glassy modulus reflects the rigidity of the carbon hydrogen bonds as the bitwnens reach their minimum thennodynamic equilibrium voiwne {i4}.

BtllalllU Codt

A C 0

-j

iJ JO ..

-15 " : 20

;; 25

..

< -30 ~ -35

.bo.sc ~il"",." 9 6% Ii,",,,, sas o 6% br,"ciltd SBS -40 O~6%SEBS Q '6%EVAI 1I'!I' 6%EVA2

1iI ~6%EBA -45

Figurt II Erreet of Polymer Type on tbe DMA Glau Transition Temperature

In quantitatively characterizing the low temperature properties of bitumen, one of the parameters often used is the glass transi tion temperature, T,. In this study, detennination ofT, of the binders was initially perfonned using differential scanning calorimetry (OSC). Unfortunately, relevant results were not obtained. Most of the binders displayed a wide transi tion region in DSC, and T, was difficult to recognize. The glass transit ion temperature was then determined from peak loss modulus at I radls using DMA. In general, the modified hinders demonstrate lower T, compared to the base bitumens. However, the effect of 3

"

percent polymer on binder T I is very small, If 6 or 9 percent polymer is added, the range of reduction in T, is between 10 and 20C for the elastomer (SSS and SEBS) modified binders, but only a few degree for those containing lhennoplaslics (EVA and EBA), ils illustrated in Figure 11. It should be noted that values of the DMA TI art frequency dependent and will also be affected by heating rate at the temperature sweep. To allow meaningful comparisons, careful control of lest conditions is necessary.

Low-Tempera/ure Creep Response In the SHRP binder speeification, a limited creep stiffuess (S) and logarithmic creep rate (m-value, which is related to binder stress relaxation ability) have been used as performancebased criteria at low temperatures (15). The measurements of low-temperatUIC creep responses are conducted at JOe above the minimum pavement design temperature. In this study, the SHRP bending beam rheometer was employed to detennine binder low-temperature creep responses at six different loading times (8, 15, 30, 60, 120 and 240 sec) and four different temperatures (-35, -25, -15 and -IOC). The stiffiless of the binder is calculated from the dimensions of the beam, the applied load and the measured deflection. Typical examples of creep response are shown in Figure 12, in which the influence of polymer modification is clearly illustrated .

r:-c:---,--------, ...,..._. i 1'"

~ ...

A .. II

..... _ .. "' ...

__ 1 .... - _---- ---_ ......... -

L_~_~~~~_---' 51 100 151 180 150

LoM,..T I-!

",,--------, ...,..._. --_ ..... ... --_ ...... --_ .. ".-

Figure 12 Creep Responses or the Base Ind SEOS Modified Bitumen 0 at -2SC

Table 4 indicates that polymer modified binders generally display a lower creep stiffness than the corresponding base bitumen, especially at temperatures lower than 15e. The improvement increases with polymer content, but varies with the base bitumen and polymer type. At 35C and for a given polymer content, the modified binders containing elastomers (SSS and SEBS) show a higher reduction in creep stiffness than those with thermoplastics (EVA and EBA). However, al the other three low temperatures, varying changes are observed in the creep stiffness. Some modified binders (e.g. bi tumens A and B modified with SEBS) even show an increased creep stiffness at 15C. Moreover, the usc of a softer bitumen seems favourable for reducing creep stiffness (cf. B85 and BI80). As a result of the reduced creep stiffness, the modified binders exhibit a lower limiting temperature (by definition, the temperature at 300 MPa creep stiffiless) than the base bitumens. However, degree of the improvement is comparatively low. As regards the effect of polymer modification on binder stress retaxation ability, both increased and decreased mvalues are observed, depending on the testing temperature and loading time. Therefore, no definite conclusions could be drawn on this mailer.

Table 4 Cmp Stiffness and m-value of the Binders at a Loading Time of 60 sec anti Dirrerent Temperatures

Binder Cree Stiffness M Pa m-va[ue -[SC -2SC -35C -ISC -2SC -35C

Bitumen A 185 789 1590 0.45 0.26 0.09 A + W.linear sas 152 680 1320 0.43 0.24 0.12 A+6t.linearSBS I2J 482 884 0.40 0.23 0.12 A+9%linearSaS 161 550 682 OJ5 0.2 1 0.11 A + 60/. branched SBS 146 484 "'7 0.38 0.20 0.13 A+6%SEBS 195 684 1140 0.38 0.20 0.11 A+6%EVAI III 624 1560 0.47 0.30 0.13 A+6%EVA2 180 790 2000 0.42 0.25 0.12 A+6%EBA 16 1 656 1620 0.41 0.26 0.13 Bitumen B 54 487 1560 0.61 0.35 0.15 B + 3% linear SBS 63 400 1080 0.54 0.32 0.13 B + 6% linear sas 55 371 921 0.52 OJO 0.15 B + 9010 linear SBS

" J72 664 0.45 0.29 0.16

B + 6% branclted SSS 56 304 m 0.49 OJI 0.17 B + 3% SEllS 63 42J 1IS0 0.54 0.33 0.1 5 B + 6';' SEBS 80 329 774 0.41 0.28 0.[9 B + 9";' SEaS 60 230 575 0.38 0.25 0.21 S+3%EVAI 40 350 1360 0.64 0.41 0.17 B+6%EVAI 36 326 1200 0.6[ 0.42 0.19 S +9%EVAI 31 302 1210 0.58 0.46 0.20 B+3%EVA2 50 3" 1190 0.56 0.40 0.16 B+6%EVA2 48 366 1240 0.54 0.38 0.18 B+9%EVA2 27 265 1140 0.59 0.42 0.19 B+3%EBA 45 J83 1250 0.58 0.37 0.18 B+6%ESA J8 298 1160 0.52 OJ8 0.19 B+9%EBA 27 234 1020 0.54 0.42 0.21 Bitumen C 99 4J7 931 0.49 0.32 0.19 C + 3% linear sas !OS J7I 812 0.45 0.29 0.20 C + 6% linear sas 95 295 491 0.42 0.27 0.18 C + 9% linear SBS 91 257 628 0.40 0.29 0.19 C + 6% branched sas 81 328 686 0.43 0.29 0.20 C+6%SEBS 97 JJJ 765 0.42 0.28 0.1 8 C+6%EVAI 89 J57 958 0.47 OJ! 0.19 C+6%EVA2 94 411 1040 0.44 0.32 0.19 C+6%EBA 60 167 875 0.35 0.16 0.20 Bitumen D 144 481

'" 0.46 0.29 0.16

D + 6% linear SSS 117 l77 796 0.42 0.27 0.17 0+ 6% branched sas 117 354 "4 0.38 0.26 0.18 Bitumen E 24 151 516 0.49 0.38 0.30 E + 6% linear sas 40 176 366 0.39 0.35 0.27 E + 6% branched SBS 42 185 446 0.34 0.30 0.26

11

CorreIa/ions between Different Low-Tempera/ure Parameters Relationships between different low-temperature parameters were examined. It was observed thai creep stiffness may statistically (risk level 5 percent) correlate with complex modulus at low temperatures. AI 25C, the correlation coefficient R is 0.80 for the creep stiffness at 8 sec and the tDmplex modulus at 1 radls. In addition, for all the binders tested, statistically significant relationships exist between the limiting stiffness temperature and Fraass breaking point (R = 0.86), and between the limiting stiffness temperature and DMA T, (R .. 0.62). The [ow conelation coefficient is due to the modified binders containing II high content (6 and 9 percent) of elastomers, which show significant differences in the degree of modification with respect to those parameters.

Tempua(ure SlISctptibility Temperature susceptibility of the binders is one of the important parameters

detennining asphalt perfonnance in road service. Asphalt mixtures containing the binders with lower temperature susceptibility should be more resistant to cracking and rutting at low and high temperatures, respectively. Temperature susceptibility is usually defined as the change in binder properties as a function of temperature. Since binder properties may be characterized by means of various parameters, different approaches have been proposed to evaluate temperature susceptibility. Traditionally, temperature susceptibility is quantified through consistency measurements made at two different temperatures. Two classical methods are the Penetrat ion Index, PI, and Penetration Viscosity Nwnber, PYN. Lower values of PI and PVN indicate higher temperature susceptibility. The results obtained (Table 3) show that the classical parameters are improved by polymer modification. For all the binders, PI may statistically correlate with PVN (R = O.S3). However, the temperature susceptibility of binders may be ranked differently by PI and PVN. As mentioned earier, tests of penetration and softening point are empirical and, in principle, only valid for unmodified bitwnens. Consequently, conventional temperature susceptibility parameters such as PI and llYN are probably not relevant indicators in the case of polymer modified binders.

In DMA, the temperature susceptibility of binders may be evaluated by measurements of various viscous and elastic parameters at different temperatures and frequencies. Observations described earier indicate that polymers modify the rheology of bitwnen. Through polymer network and/or crystallinity, the modified binders become more elastic and less temperature susceptible over the nonnal service temperature range compared 10 the base bitumens (Figure S). AI high temperaltU"es, however, the polymer modified binders are likely to cxhibit temperature susceptibility that is the same as, or slightly higher than, thaI of the base bitumens. This is a result of the weakening and dissociation of the polystyrene domains (SSS and SEBS) or the melting of Ihe crystalline portions (packed polymer segments of EVA and EBA). To perfonn an evaluation similar 10 viscosity temperature susceptibility (VTS) (16), the double logarithm of complex modulus, complex viscosity (,, ' .. O'/ro) and dynamic viscosi ty (,, ' = GM/Il) was plotted against the logarithm of temperature in K for both unmodified and modified binders. Unfortunately, these plots were not straight lines over a wide temperature range, and the slope (temperature susceptibility) changed with the temperature range being considered. This means that temperature susceptibility cannot be evaluated as a single-valued parameter such as VTS.

Aging Properties Aging of the binders was perfonned using the Thin Film Ovcn Test (TFOT, ASTM D

1754) and the Rolling Thin Film Oven Test (RTFOT, ASTM D 2872), respectively. Aluminium pans (TFOD and glass containers (RTFOD were heated to 160C before loading the sample. The sample holders were kept in an oven at 160C for about IS minutes (RTFOT

"

glass containers in horizontal position) and then aged according 10 the standardized procedures: 163C and 75 minutes for RITOr and, 163C and 5 hours for TFOT, The aged binders were evaluated by measuring their rheological properties.

Figure I) reveals the viscoelastic changes oflhe binders after aging. AI the frequency studied ( I radls), aging increases the complex modulus of the base bitumens. However, the innuence of aging on polymer modified binders is temperature dependent. For the base bitumens and modified binders wi th 3 percent SSS, as well as those containing thermoplastics, the elastic response was slightly increased by aging, as indicated by a small decrease in phase angle in a temperature range of 0 10 about 7OC. In contrast, when aging is performed on the modified binders with 6 or 9 percent SBS and SEBS, large increases in phase angle were observed at high temperatures, and consequently, the depth of the phase: angle minimum is reduced, which would imply polymer network damage.

, -~ , ....

_ .1fOT -_, -_ .. ,.. .... .

, ... f!

f 'lIl* ; '1Il>f)

J'II(I('I)O t ,lIlotl J UltEoO:I ,00."

-_ ......... -.... "101 _ _ A

-- , .... _--_ ..... ... -

" I-111 ," ' 0 _I1IOT --,

-_ .. "'- '"

" -_ ......... ,..

.... "'" -_A , a~ -_ ......... '"

-_ .. "'--,- 0 0

,. ' 0

'" '-'"

, -~ , ..

..... ,...". __ 0 _ _ 0''' __

,-~

f '- * t "Oli>f) 1 ' 00 ...

.... lTfOT-- _ _ O"~_I&I

!. \ " i. ' .>f) J U II!>f) r "

' ... ' "

, .. ,-

0 .... 10 .. ' . .. .. , .

,~'" '-'"

Figure 13 Effect of RTFOT Aging on Tempel'lllture DependeD~e of Complu Modulus and Pbase Angle

In addi tion, the frequency dependence of the rheological parameters is innucnc.ed by aging process (Figure 14). For the base bitumens, aging increases complex modulus and decreases phase angle over a frequency range of 0.1 to 100 rad/s, indicating the aged bitumens become more elastic. For the modified binders, the curves may cross, both for complex modulus and phase angle, and changes in elasticity (phase angle) are dependent on the frequency considered.

The rheological changes are consistent with the degradation of polymers and oxidation of bi tumens (17). An increasing content of funct ional groups may change bitumen polari ty and molecular association. As a resull, compatibility (microstructure) between thc polymer

19

and bitumen is changed, as shown in Figure 15. On the other hand, breakdown of polymer reduces the number of large polymer molecules, and consequently, the effectiveness of the polymer in significantly modifying bitumen rheology is reduced. Obviously, the rheological changes in aged modified binders are dependent on a combined effect of bitumen oxidation and polymer degradation, which varies with types of bitumen and polymer, as well as polymer content.

'*~ . i ,OOl'" ~ "1 lODE'"

i ", I , t:~ , , 0' , _____ a_.""" I IIlIl ...

_ _ ..... tr .....

0 -. . _.---. _ .... -

,-- ,-, ' *~ 'OK.o, '*4 f_~""1

'*4 r-:--:-;-~~~~~~~l. . .

~100Jt: .... I . Ioot:'" j,00l~ i A 100!",

. ........... -...... " .,

"1 ", .' , 0'

---_ ............ ""'" -_ .... p ..... .....

_ ... . _ . .... ... "' ... ""', 10 .-_ ....... ., ...... ,'" ,~~ L~ _______ -"

,_ ,lOEoOO ,RoO, I E'" f_I'~

'*~ . ;: , ....

~ .. ~ 1 1.0lIIE>fJ j

"' I, toIioOl .' , ! ~---.. ~ .... -.... ., JI.'" -_ .. "''' ..........

-. ._ ..... _ ........ ._ ..... _ ........

,*4

,-, ' *~ 'Q'.", , .~ f_l--t

'.4 r~~~~~~~~~-,' ......-:.

., , .,

--~-.. " ... -."'" -_ .. "' ....... ..... -.... . _ ......... -."'"

.' ,

.,

. _ .. "' ...... ,"'"

" ,~~ L ____ ~~ ___ ....J , OQE.II I lOX" I 111(>01

''''''''''''o4IIl Figure 14 Complex ModulU5lnd Phase Angle It 60C as I Funttion or Frequenty Cor Base and Polymer Modified Bitumen B berore Ind after RTFOT

Bitumen B + 6% EaA bcfoo RTFOT Bitumen a + 6% EBA after RTfOT

Figure IS Errut of RTFOT Aging on the Microstrutlure of Polymer Modified Binden

T raditionlllly, the aging sensitivity of bitumens is evaluated by means of aging index, which is defined as the ratio of a physical/rheological parameter of the aged binder to that of the original binder. For conventional binders, high values of the parameter indicate a high degree of bitwnen aging (hardening). However, aging index of polymer modified binders is largely influenced by the evaluation temperature and frequency, as illustrated in Figure 16. The varying indices are caused by degradation of the JXllymcr and oxidation of the bitumen, which are to some extent compensatory. Evidently, aging index is not an accurate parameter in the case of polymer modified binders, and it does nol seem possible 10 apply a certain value of aging index in characterizing their aging properties.

.

.

~ 2.0 .."

r ... ,,/ " ._I.! . , ....

: 1.0 .. ~ ... -. ~ Q.5

' ..

0.' L~~~~~_~~~..J ~ ~ G 10 ~ ~ ~ ]00 ]M ]~

Tom"" .. t." for Mouori"1 AItiDJ: ladn t"tl

, 0

-_. -_ .. , .. ,-, . -_ .............

'V ,

OJ 1.00&-411 1.00+1

Figure 16 Temperature and Frequency Dependence of RTFOT Agiog Iodex

CONCLUSIONS

----

I.OOE+Gl

Polymer modified bitumens exhibit a two-phase microstructure. Under the influence of gravitational fields, a phase separation can OCCut. The phase separation may be determined by hot storage test. The compatibility and storage stability of polymer modified binders depend greatly on polymer content and are also influenced by the chemical composition of base bitumens and the nature of polymers. By mixing a small amount (3 percent in this study) of polymer, the modified binders show dispersed polymer particles in the continuous bitumeo matrix. At an increased polymer content, a continuous polymer phase may be observed. The minimum percentage of the polymer to ensute the formation of its continuous phase depends on the base bitumen as well as the polymer itself. Regardless of the nature of the two phases, the storage stability of modified binders decreases with increasing polymer content. Among all the modified binders prepared, those containing EBA display the lowest storage stability. For the elastomer (SBS and SEBS) modified binders, those containing branched SSS are the most unstable. In the case of EVA modified binders, better storage stability is observed for those containing the polymer with a lower VA content and a higher MI. Better compatibility and higher storage stability may also be achieved by using a base bi tumen with a higher content of aromatics and/or a lower content of asphaltenes.

A Qualitative comparison between separation index and fluorescence micrograph indicates that, at a given polymer content, the modified binders showing finer polymer dispersion in micrographs generally exhibit less phase separation during hot storage. Such a comparison cannot be made among modified binders with differeot polymer contents. This means that the compatibility defined by different methods is not necessarily consistent, and the degree of compatibility cannot be determined using only one method. Moreover,

"

compatibility and storage stability may influence the rheological properties of modified binders; unfortunately, no definite relationships between tbese parameters were established in this study. The complexity in such relationships increases when modified binders with different kinds of polymer art compared.

The rheological properties of bitumens are improved by polymer modification. The improved viscoclastic properties may be identified by various parameters obtained using dynamic mechanical analysis, such as phase angle and storage modulus. Four regions (glassy, transition, plateau, and tenninal or flow) are recognized in plOlS of storage modulus vrnus temperature for polymer modified binders of sufficiently high polymer content (6 percent in this study). In phase anglcltemperature diagrams, phase angle maximum and minimum are observed. On the other hand, for base bitumens and modified binders with a low content of polymer (3 percent in !his study), !he transition goes straight from the glassy 10 the flow regions with no region of rubberlike behaviour. The degree of the modification with respect to the increased elastic response is greatly affected by the characteristics of base bitumens and polymers. For a given base bitumen, the influence of the six polymers tested may be ranked as branched SBS, SEBS/linear SBS, EBA, EVA2 and EVAI, branched SBS being the most effettive. The improved dynamic viscoelastic properties should be favourable for resistance to rutting al high temperatures.

Polymer modification improves the low.temperature properties of bitumens, as indicated by decreases in creep stiffness, Fraass breaking point and glass transition and limiting stiffness temperatures. The relative improvement varies with the base bitumen, polymer type and polymer content. Compared with thennoplastics (EVA and EBA), the elastomers (SBS and SEBS) seem more effettive in improving bitumen low-temperature parameters (except for the limiting stiffness temperature). For a given polymer, the BI80 bitumens may result in a relatively higher reduction in creep stiffness than the B8S bitumen. The influences of polymer modification may also vary with testing conditions (temperature and loading time). The improved low temperature parameters should be favourable with respect to asphalt cracking perfonnance.

The RTFOT and TFOT aging causes oxidation of bitumen and degradation of polymer, and consequently, change the microstructure and rheological properties of the modified binders. These changes largely depend on the characteristics of the polymer used and its content. For the modified binders with thennoplastics (EVA and EBA) as well as those with a low content of elastomers (SBS and SEBS), aging increases complex modulus and elastic response (decreased phase angle) and reduces temperature susceptibility. On the other hand, for the modified binders with sufficiently high content of elastomers (;2: 6 percent in Ihis study), the influence of aging on those parameters is strongly dependent on temperature and frequency.

Aging index may be applied for evaluating !he base bitwnens and certain modified binders with no significantly chemical changes in the polymer during aging, but it is not suitable if the aging process causes a considerable polymer degradation. In this case, values of aging index are largely influenced by !he temperature and frequency dependence of polymer effect

ACKNOWLEDGEMENTS

The authors wish to thank Jonas Ekblad, Britt Wideman and Clarissa Villalobos for their laboratory assistance. The authors also gratefully acknowledge the Swedish Transpon and Communications Research Board (KFB) for the financial support of this study, and Nynas Petroleum for supplying a number ofbase bitumens and polymers.

REFERENCES

D.N. lillie, tW. Bulton, R.M. White, EK Ensley, Y. Kim and SJ. Atuned, "Investigation of asphalt additives", Report No FHWA/RD-871OO1. U.S. Department ofTransportation. Federal Highway Administration, 1987.

2 a.N. King, H.W. King, O. Harders, W. Arand and P.-P. Planche, "lnflucnceofasphalt grade and polymer concentration on the low temperature performance of polymer modified asphalt", J. Assoc. of Asphalt Paving Technologists, Vol. 62, 1993, pp. ]-22.

] H I. Collins, M.G. Bouldin, R. Gelles. and A. Berker, "Improved performance of paving asphalls by polymer modification", J. Assoc. of Asphalt Paving Technologists. Vol. 60, 199I,pp.43-79.

4 n . Goodrich, "Asphaltic binder rheology, asphalt concrete rheology and asphalt concrete mix properties", J. Assoc. of Asphalt Paving Technologists, Vol. 60, pp. 1991 , pp. 80- 120.

5 S. Piazza, A. Arcozzi and C. Verga, "Modified bitumens containing thennoplastic polymers~, Rubber Chemistry and Technology, Vol. 53, 1980, pp. 994- 1005.

6 G. Van Gooswil!igen and W.C. Vonk, "The role of bitumen in blends with thermoplastic rubbers for roofing applications", Proceedings of the Vl lntemational Conference "Roofing and Waterproofing, The International Waterproofing Association, London, 1986, pp. 45-52.

7 B. Brule, Y. Brion and A. Tanguy, "Paving asphalt polymer blends: relationships between composition, structure and properties~ , Pmc. Assoc. of Asphalt Paving Technologists, Vol. 57, 1988, pp. 41-64.

8 D. Whiteoak, "The Shell Bitumen Handbook", Shell Bitumen UK, 1991, p.74. 9 N.W. Mcleod, "A 4-year survey of low temperature transverse pavement cracking on

three Ontario test roads", Proc. Assoc. of Asphalt Paving Technologists, Vol. 41 , 1972, pp. 424-493.

10 X. Lu, U. lsacsson and 1. Ekblad, "Phase separation of SBS polymer modified bitumens", Journal of Materials in Civil Engineering, ASCE. 1998, in press.

II O. Kramer and J.D. Ferry, "Dynamic mechanical properties", Science and Technology of Rubber, edited by F. R. Eirich, Academic Press, New York, 1978.

12 J.D. Ferry, Viscoelastic Properties of Polymers", Wiley, New York, 1980. 13 G.N. King, H.W. King, O. Harders, P. Chavenot and J.-P. Planche, "Influence of

asphalt grade and polymer concentration on the high temperature performance of polymer modified asphalt". J. Assoc. of Asphalt Paving Technologists, Vol. 61, 1992, pp. 29-66.

14 H.U. Bahia and DA Anderson, "The new proposed rheological properties of asphalt binders: why are they required and how do they compare to conventional properties", Physical Properties of Asphalt Cement Binders, ASTM STP 124\, John C. Hardin, Ed., American Society for Testing and Materials, Philadelphia, 1995, pp. 1-27.

15 DA Anderson and T.W. Kennedy, "Development of SHRP binder specification", J. Assoc. of Asphalt Paving Technologists, Vol. 62, 1993, pp. 481-507.

16 V.P. Puzinauskas, Properties of asphalt cements-, Pmc. Assoc. of Asphalt Paving TechnologislS, Vol. 48, 1979, pp. 646-710.

17 X. Lu. and U. lsacsson, -Chemical and rheological evaluation of ageing properties of SBS polymer modified bitumens-, Fuel, Vol. 77, 1998, pp. 961-972.