Influence of low-temperature physical hardening on stiffness and tensile strength of asphalt concrete and stone mastic asphalt

Construction and Building Materials 61 (2014) 191–199

Contents lists available at ScienceDirect

Construction and Building Materials journal homepage: www.elsevier.com/locate/conbuildmat

Influence of low-temperature physical hardening on stiffness and tensile strength of asphalt concrete and stone mastic asphalt Józef Judycki ⇑ Department of Civil and Environmental Engineering, Gdansk University of Technology, Narutowicza Street 11/12, 80-952 Gdansk, Poland

h i g h l i g h t s  Stiffness of asphalt mixes increased after isothermal storage at low temperature.  No changes of tensile strength after isothermal storage were noted.  Physical hardening of asphalt concrete and stone mastic asphalt was different.  Two measures of physical hardening of asphalt mixes were proposed.

a r t i c l e

i n f o

Article history: Received 21 November 2013 Received in revised form 27 February 2014 Accepted 9 March 2014 Available online 28 March 2014 Keywords: Physical hardening Asphalt concrete Stone mastic asphalt Stiffness modulus Indirect tensile strength Low-temperature cracking

a b s t r a c t This paper presents laboratory testing of stiffness modulus and indirect tensile strength of three asphalt concrete (AC) and three stone mastic asphalt (SMA) mixes after isothermal storage at temperature of 20 °C, at different time intervals up to 16 days. The tests under repeated dynamic loading showed physical hardening of all tested mixes which was manifested by an evident increase of their stiffness moduli after isothermal storage. Contrary to expectations, the tests on all mixes did not show any evident changes of the indirect tensile strengths which after isothermal storage remained almost constant, within normal scatter range. The differences in the physical hardening of asphalt concrete AC and stone mastic asphalt SMA were found. At the beginning of isothermal storage at 20 °C the physical hardening was slower for the SMA. After 5–16 days of storage the SMA showed greater physical hardening than the asphalt concrete AC. Two measures which allowed the quantification of intensity of the physical hardening were introduced. Ó 2014 Elsevier Ltd. All rights reserved.

1. Introduction The physical hardening of asphalt binder is the term denoting the process of increase of stiffness of asphalt binder during extended storage at isothermal conditions. The process was first investigated in 1936 by Traxler and Schweyer [1] who studied different asphalt binders and observed substantial increase in their viscosities with time of storage at room temperature. They called this phenomenon ‘‘age hardening’’. In 1937 Traxler and Coombs [2] explained the process of increase of viscosity of binders at isothermal conditions by the sol–gel transition. In 1950s Brown et al. [3] investigated several asphalt binders after isothermal storage and found reversible increase of their stiffness. They called the process ‘‘steric hardening’’ and explained it by the collapse of the molecular free volume at low temperature, isothermal sol–gel ⇑ Tel.: +48 601 22 81 90. E-mail address: [email protected] http://dx.doi.org/10.1016/j.conbuildmat.2014.03.011 0950-0618/Ó 2014 Elsevier Ltd. All rights reserved.

transition and wax crystallization. The term ‘‘physical hardening’’ of asphalt binders was first used by Blokker and Van Hoorn [4] in 1959 who explained the process by wax crystallization and asphaltene aggregation. Struik [5] in 1970s tested the same process for several materials including polymers and some asphalt binders and called it the ‘‘physical aging’’ to distinguish it from the ‘‘chemical aging’’ which involves chemical reactions. In 1990 Pechenyi and Kuznetsov [6] explained the reversible hardening of asphalt binders during isothermal storage by the formation of partially ordered structures in the process of crystallization of asphalt material. The physical hardening of asphalt binders at low temperatures, became the subject of intense research during last two decades, when it was noted that the process may have a significant impact on low-temperature behavior of asphalt pavements. It is known that physical hardening is related to the chemical composition and the source of asphalt binder base. Shrinkage and collapse of free molecular volume have been quoted as its prime cause [7,8].

192

J. Judycki / Construction and Building Materials 61 (2014) 191–199

The molecular structuring and aggregation of asphaltenes and possibly crystallization of wax have been identified as additional factors responsible for the physical hardening [9–12]. The process of physical hardening of asphalt binders has been well documented by many researchers during the last two decades [7–10,13–15]. So far, much less attention has been paid to physical hardening of asphalt mixes. The question whether physical hardening occurs in asphalt pavement at cold temperatures and what is its importance in development of low-temperature cracking has been seriously disputed. Some researchers claim that the physical hardening in asphalt mixes can be cancelled by stress relaxation and does not affect low-temperature cracking of asphalt pavements [16,17]. Others [10–12,18–20] are of the opinion that the physical hardening is a very important factor and the asphalt mixes containing binders showing significant physical hardening at low temperatures are more susceptible to low-temperature cracking, what has been proven in field tests in Canada [11,20,21]. The dispute whether physical hardening occurs in asphalt mixes is caused by confounded results of laboratory tests, which were not always well designed to investigate this phenomenon. In most cases the thermal stress restrained specimen tests (TSRST) were used to test the physical hardening of asphalt mixes [14,22,23]. However, this method did not show a clear effect of physical hardening in asphalt mixes, which was the main basis to question the importance of physical hardening. A disadvantage of the TSRST test, when applied to the physical hardening, is that it does not measure directly the mix stiffness but rather the thermal stresses induced in a restrained asphalt specimen. The thermal stresses are in a very complex way related to several factors, such as elastic and viscous properties of the mix, its stress relaxation potential, thermal contraction properties of the mix, as well as specimen conditioning, rate of cooling and the thermal history. A better insight into physical hardening was obtained with a recently developed asphalt thermal cracking analyzer (ATCA) presented by Baglieri et al. [18], Bahia et al. [19] and Tabatabaee et al. [24]. The results from ATCA tests allowed the verification of the thermal stress model which included physical hardening [25]. There are very few published reports from testing of unrestrained asphalt concrete specimens, in which the stiffness modulus of asphalt mix was directly measured after isothermal storage at low temperatures [14.26]. In one of them [14], asphalt concrete specimens were tested in uniaxial tension–compression after storage at a low temperature and physical hardening was very evident. In another research [26], very small beams (6.25  12.5  102 mm) cut from compacted asphalt concrete were tested in bending beam rheometer (BBR) according to AASHTO T313-06. In this case, the results were very much scattered, a part of the specimens showed physical hardening but a significant part did not. Not only stiffness but also tensile strength is extremely relevant to the development of low temperature cracks in asphalt pavements. In the available literature on physical hardening there is no evidence that tensile strength of asphalt mixes was tested after extended storage at low temperatures. It was reported by El Hussein et al. [27], that micro-cracking developed in asphalt concrete samples during storage at low temperature, due to large differences between temperature contraction coefficients of asphalt binder and aggregate, which might lower tensile strength. Therefore, the tensile strength of asphalt mixes after prolonged storage at low temperatures is worth investigating.

2. Objectives The main objective of this study was to determine effects of isothermal storage at low temperature on the stiffness and tensile strength of two types of asphalt mixes, namely asphalt concrete

AC and stone mastic asphalt SMA. Stiffness moduli and tensile strength of asphalt mixes were measured in indirect tensile test, after isothermal storage of specimens at 20 °C, at different time intervals up to 16 days. The temperature of 20 °C was selected for testing as it was a typical low temperature in Poland during severe winters. The second objective was to identify the influence of the mix type (asphalt concrete AC and stone mastic asphalt SMA), type of asphalt binder used, either neat and polymer modified, on the low-temperature physical hardening. The SMA is becoming the most popular wearing course on newly constructed main roads in Europe. The most research on physical hardening was carried out in the USA and Canada on asphalt concrete mixes and properties of SMA are less known. Therefore, part of this research was devoted to the low-temperature physical hardening of the SMA. 3. Material tested In total, six asphalt mixes designed for wearing courses for medium and heavy traffic were tested for physical hardening. These included three asphalt concrete (AC) and three stone mastic asphalt (SMA) mixes. In order to obtain more realistic data the tested mixes were collected from asphalt plants and brought to the laboratory for testing. The mixes were produced by three different contractors and differed in mix composition, asphalt binders and aggregates used. The composition of tested asphalt mixes and types of aggregates used are presented in Table 1. The mixes tested for physical hardening in this study represented all typical materials which have been used in Poland for wearing courses, including two types of mixes (AC and SMA), different asphalt binders (neat and polymer modified), two methods of polymer modifications (in refinery and in asphalt plant), different mineralogical types of aggregates, mix gradings and compositions. Testing of mixes sampled from asphalt plants had the advantage over laboratory produced mixes that real technological aging and normal scatter of mix composition were included. This gave the opportunity to assess the physical hardening in the realistic materials used for road construction. Liquid adhesive agent (fatty amines) was added to all mixes to improve their water and frost resistance. In case of SMA the anti-dripping stabilizer in form of the cellulose fibers was also added. Properties of asphalt binders are presented in Table 2. Three types of asphalt binders were used: neat, modified in refinery with styrene–butadiene–styrene (SBS) copolymer and modified with linear styrene–butadiene (SB) copolymer added in asphalt plant. All the binders were produced in the same refinery from the Ural crude oil. The binders contained low amount of paraffin. The harder grade 35/50 was produced with air blowing. The grade 50/70 was produced alternatively by compounding or by air blowing. The method of production of the base binder used for modification with the SBS polymer is unknown to the author. Asphalt binders were not tested in this study, but the data on their properties were provided by the refinery.

4. Testing of physical hardening of asphalt mixes 4.1. Preparation of specimens Loose asphalt mixes collected from the asphalt plants were reheated, remixed and compacted in a gyratory compactor, to produce cylindrical specimens 100 mm in diameter and approximately 63 mm in height. It was expected that level of compaction and voids content in the compacted mix may affect physical hardening. Therefore, a great effort was undertaken to produce homogenous specimens. Greater number of specimens than required for testing of physical hardening was produced. Specimens whose voids differed more than 0.5% from the average value were rejected from further testing. It was expected that besides the voids content also initial stiffness of asphalt mix might affect the process of physical hardening, in this way that stiffer specimens might harden during isothermal storage in different way than softer ones. To check homogeneity of mix stiffness all the specimens approved for testing of physical hardening were, before the proper test, examined in indirect tension modulus testing under repeated loading at +20 °C, in the Nottingham Asphalt Tester (NAT), according to the European Standard PN-EN 12697-26 at 0.12 s time of loading pulse, frequency of 20 cycles per minute and vertical pressure of 220 kPa. Despite the

193

J. Judycki / Construction and Building Materials 61 (2014) 191–199 Table 1 Properties of tested asphalt mixes designed for wearing courses. Properties

Asphalt concrete

Stone mastic asphalt

AC1

AC2

AC3

SMA1

SMA2

SMA3

A

B

C

C

A

C

0/12.8 50/70 Neat

0/16 35/50 Neat

Type of coarse aggregate (P2 mm)

Crushed granite

0/11 80C SBS modified in refinery Crushed granite

Mixture of crushed granite and gravel Limestone

Crushed gravel

Crushed granite

Limestone

Limestone

Limestone

0/11 35/50 Neat Crushed granite Crushed granite Limestone

0/11 50B SBS modified in refinery Crushed granite

Type of sand (0.075/2 mm)

Mixture of crushed basalt and granite Crushed granite

0/12.8 50/70 SB modified in asphalt plant Crushed gravel

5.5 0.4

5.2 0.4

5.7 0.4

6.5 0.4

6.1 0.4

6.3 0.4

None

None

None

0.4

0.4

0.4

61.0 31.0 8.0 13.5

68.4 24.0 7.6 12.8

61.6 31.0 7.4 13.1

77.5 11.6 10.9 17.4

78.1 11.3 10.6 16.7

77.5 11.6 10.9 17.2

2.614 28 2.2

2.446 31 2.3

2.412 23 2.4

2.438 35 4.0

2.441 25 3.6

0.19

0.18

0.23

0.23

0.22

9.0 7,712

7.8 4,291

9.6 2,939

5.9 4,065

6.1 3,301

564

477

267

402

417

7.3

11.1

9.1

9.9

12.6

General data Asphalt plant and contractor designation Nominal size of aggregate, mm Type of asphalt binder

Type of filler (<0.074 mm) Mix composition Asphalt binder content, % by wt. Adhesive agent content, % by wt. of asphalt binder Stabilizer (cellulose fibers), % by wt. of mix Fraction retained on 2 mm, % Fraction 0.074/2 mm, % Fraction passing 0.074 mm, % Content of asphalt mastic (sum of asphalt binder and filler), %

Properties of specimen selected for testing of physical hardening Specific gravity, G/cm3 2.456 Number of selected specimens, n 24 Voids content, Average, 1.7 % St. dev., 0.20 % C.V., % 11.9 Stiffness modulus, at Average, 5,914 +20 °C, t = 0.12 s MPa 482 St. dev., MPa C.V., % 8.2

Crushed granite Limestone

St. dev. – standard deviation, C.V. – coefficient of variation.

same way of compaction and very similar voids content the stiffness moduli at +20 °C varied (see Table 1). To minimize the likely influence of the initial stiffness of specimens on evaluation of physical hardening, and to avoid misleading results, a special procedure was used to form a series of nominally uniform specimens. One series for testing the physical hardening after a given isothermal storage period consisted of four nominally uniform specimens, with the exception of mix AC3 for which five uniform specimens were tested in a series. All the specimens produced from a given asphalt mix were divided into four (or five in the case of AC3) groups with very close values of the stiffness modulus S at +20 °C; S1 > S2 > S3 > S4. Each such group consisted of six specimens. For each test series of four specimens, the first specimen was taken randomly from the stiffness group S1, the second from S2, the third from S3 and the fourth from S4. In this way each one of the four specimens in a given test series came from a different group of initial mix stiffness thereby averaging the effect of variable initial stiffness of specimens on the physical hardening. 4.2. Procedure of testing physical hardening Two properties of asphalt mixes were tested after a given period of isothermal storage: stiffness modulus (S) and indirect tensile strength (ITS). The specimens were put into a freezer preset at temperature 20 °C. Three temperature sensors were installed: one in a control asphalt mix specimen, in a drilled hole filled afterwards with asphalt binder. Second sensor was installed on the side wall of a control specimen. A third sensor registered temperature inside

the freezer. The temperature of the specimens was kept at a constant level of 20 °C with ±0.5 °C accuracy with rare maximum deviation up to ±1 °C. The moment when the temperature of control specimen reached 20 °C was treated as the starting point of isothermal storage. The periods of isothermal storage were equal to 1 h, 3 h, 24 h (1 day), 72 h (3 days) and 384 h (16 days). After a given isothermal storage period a specimen was taken from the freezer and put in the Nottingham Asphalt Tester (NAT) for stiffness modulus testing at 20 °C in indirect tension, according to PN-EN 12697-26. Load was applied across the vertical diameter of the specimen through two curved loading strips. The time of loading pulse was 0.12 s with load repetitions every 3 s, what resulted in the frequency of 20 cycles per minute. The test was run under the controlled stress mode with constant horizontal stress pulse 460 kPa for asphalt concrete (AC) and 440 kPa for stone mastic asphalt (SMA). The specimens were tested in two perpendicular planes and results were averaged. The Poisson’s ratio assumed for calculations of stiffness modulus was equal to 0.15 at 20 °C (after [28]). Testing of stiffness modulus of asphalt mixes at 20 °C in the Nottingham Asphalt Tester has been a common practice at the Road Laboratory of Gdansk Technical University and no any problems in measurements of diametric displacements were encountered at such temperature. The repeatability of results was very good, what can be seen in Table 3. The testing device was inside the storage chamber preset to 20 °C. It was extremely important to run the test in a quick and reliable manner to minimize the heating up of the specimen when it was for a short time in a room temperature. Excessive heating up

194

J. Judycki / Construction and Building Materials 61 (2014) 191–199

Table 2 Properties of asphalt binders. Properties

a

Designation of asphalt binder 35/50

50/70

50B

80C

Type of asphalt binder Penetration at 25 °C, 0.1 mm Softening point, °C Fraass breaking temperature, °C Paraffin content, % Viscosity at 60 °C, Pa s Elastic recovery, %

Neat 45 54 13 1.3 671 –a

Neat 64 48 15 1.3 294 –a

SBS modified 58 58 14 –a –a 87

SBS highly modified 60 75 22 –a –a 99

After RTFOT Decrease of penetration, % Increase of softening point, °C Elastic recovery, %

32 6 –a

33 6 –a

25 3 81

10 1 95

‘‘–’’ Not tested.

  jAVG  X max j jAVG  X min j  100% RS ¼ max ; AVG AVG

of a specimen would cause recovery of the physical hardening. However, the temperature measurements on the control specimen indicated that the increase of temperature during the whole process of stiffness modulus testing did not exceed 0.5 °C and rarely 1 °C, so it can be safely assumed that during the testing the effect of physical hardening was not disrupted. After testing of stiffness modulus the specimen was put to the second freezer preset at 20 °C and kept there for 30 min before testing for indirect tensile strength. The indirect tensile strength test was performed according to PN-EN-12697-23 with constant loading rate 20 mm/min.

ð1Þ

where AVG, Xmax, Xmin – average, maximum and minimum results obtained in a series. As usual, the scatter of indirect tensile strength (ITS) was greater than the scatter of stiffness modulus (S). In both cases the scatters were in a normal range observed in asphalt mix testing. The relative scatter of results RS was selected in this study to evaluate the dispersion of results instead of the coefficient of variation because number of specimens tested in a uniform series was small and equal to 4 or 5. Due to the currently ongoing discussion whether the physical hardening occurs in the asphalt mixes or not and the divergent results presented earlier in [14,22,23,26] it was decided to present in Tables 3 and 4 all the results with the relative scatter characterizing their dispersion in order to prove that the physical hardening do occur in the tested asphalt mixes. As an example, Fig. 1 presents typical changes of stiffness modulus (S) and indirect tensile strength (ITS) in relation to duration of isothermal storage for asphalt concrete AC1 and stone mastic asphalt SMA1. The scatters of results (maximum and minimum

5. Test results Four nominally uniform specimens for the AC1, AC2, AC3, SMA1 and SMA2 mixes and five for the SMA3 mix were tested in a series at a given isothermal storage periods. Tables 3 and 4 present the average values of stiffness modulus (S) and indirect tensile strength (ITS) tested at 20 °C after six periods of isothermal storage at 20 °C. Tables 3 and 4 present also the relative scatter RS of results calculated as:

Table 3 Stiffness modulus S of asphalt mixes at 20 °C and relative scatter of results RS. Isothermal storage time at 20 °C, (h)

Designation of asphalt mix AC1

1 3 24 72 120 384

AC2

AC3

SMA1

SMA2

SMA3

S (MPa)

RS (%)

S (MPa)

RS (%)

S (MPa)

RS (%)

S (MPa)

RS (%)

S (MPa)

RS (%)

S (MPa)

RS (%)

24,092 25,454 26,376 26,936 26,974 27,323

8.5 4.5 2.3 2.7 3.8 3.0

26,223 27,581 27,894 28,488 29,123 29,045

4.6 6.6 6.8 1.7 3.2 3.5

25,002 25,773 26,619 27,122 27,730 28,050

4.1 4.5 1.1 1.0 1.5 1.9

21,003 21,760 22,188 23,114 23,800 24,528

12.0 5.1 5.0 5.8 5.5 5.4

22,557 22,946 23,974 24,300 25,073 25,521

6.0 5.4 3.7 4.5 1.2 3.4

23,920 24,657 25,760 26,868 27,977 27,967

5.3 7.1 2.2 3.2 2.7 1.2

Table 4 Indirect tensile strength ITS of asphalt mixes at 20 °C and relative scatter of results RS. Isothermal storage time at 20 °C, (h)

Designation of asphalt mix AC1

1 3 24 72 120 384 a

AC2

AC3

SMA1

SMA2

SMA3

ITS (MPa)

RS (%)

ITS (MPa)

RS (%)

ITS (MPa)

RS (%)

ITS (MPa)

RS (%)

ITS (MPa)

RS (%)

ITS (MPa)

RS (%)

4.37 –a 4.30 4.56 4.38 4.62

5.4 –a 20.2 17.0 22.9 18.6

3.65 3.92 3.51 4.27 3.68 3.66

10.1 17.6 15.5 12.5 15.6 19.8

4.23 4.61 4.48 3.17 3.92 4.03

11.7 8.7 6.9 12.6 15.1 16.4

4.11 3.72 3.87 3.67 3.63 3.58

16.3 9.9 2.9 21.1 13.2 4.8

2.88 3.59 3.45 3.06 2.96 3.08

13.6 30.9 14.7 23.5 12.5 9.9

3.18 3.42 3.64 3.49 3.59 3.54

17.6 8.3 12.2 7.5 6.8 16.6

‘‘–‘‘ Rejected results with the highest deviations from the average value.

195

J. Judycki / Construction and Building Materials 61 (2014) 191–199

recorded values of moduli and strengths) are indicated in Fig. 1. The same character of changes as shown in Fig. 1 was observed for all the tested mixes, both for asphalt concrete AC and for SMA. Fig. 2 presents stiffness modulus and indirect tensile strength in relation to isothermal storage time for all the tested mixes in semi-logarithmic scale. The test results indicated clearly that the stiffness modulus of all tested mixes increased with time of isothermal storage at 20 °C, whereas the indirect tensile strength did not change with the time of isothermal storage. In case of indirect tensile strength the scatter of registered strength values was greater than the effect of isothermal storage time.

6. Analysis of results 6.1. Measures of intensity of physical hardening Two measures of intensity of physical hardening of asphalt mixes were introduced: Stiffness Growth Rate SGR and Physical Hardening Ratio PHR. These measures allow the assessment of the sensitivity of various asphalt mixes to physical hardening. The relation between stiffness moduli and logarithm of isothermal storage time is approximately linear, as shown in Fig. 2, and can be expressed as:

SðtÞ ¼ Sð1Þ þ SGR  log t

ð2Þ

where S(t) – stiffness modulus of asphalt mix after isothermal storage time t, from t = 1 h to t = 384 h, in MPa, S(1) – stiffness modulus of asphalt mix after isothermal storage time t = 1 h, approximated from the test data with Eq. (2), in MPa, log t – logarithm of time t dðsÞ given in hours, SGR – Stiffness Growth Rate, SGR ¼ dðlog , which reptÞ resents the increase of stiffness modulus in MPa, after each decade of time of isothermal storage expressed in hours. For each subsequent decade 1–10 h, 10–100 h and 100–1000 h the stiffness increases by SGR value, expressed in MPa.

Table 5 presents the values of Stiffness Growth Rate SGR, S(1) and the determination coefficients R2 of the relation (2) for all the tested mixes. The determination coefficients were in a range from 0.90 to 0.99, what indicates a quite strong relationship. The S(1) values, given in Table 5, were calculated from the regression Eq. (2) and therefore slightly differ from the values of stiffness moduli obtained from measurements at isothermal storage time t = 1 h given in Table 3. The second measure of physical hardening is the Physical Hardening Ratio (PHR) which is defined as follows:

PHR ¼ SðtÞ=Sð1Þ

ð3Þ

where S(t) – stiffness modulus of asphalt mix after t hours of isothermal storage, MPa, S(1) – stiffness modulus of asphalt mix at the beginning of isothermal storage at t = 1 h, MPa, t – time of isothermal storage, h. Table 6 presents the calculated values of Physical Hardening Ratio on the basis of test results presented in Table 3. Fig. 3 presents the Physical Hardening Ratio in relation to isothermal storage time for asphalt concrete AC and stone mastic asphalt SMA. The following empirical formula was selected to fit the experimental data:

PHR ¼ 1 þ a1  log t þ a2  ðlog tÞ

2

ð4Þ

where PHR – Physical Hardening Ratio, PHR P 1.0, a1, a2– experimental constants, a1 P 0; t – time of isothermal storage, in hours; t P 1 h, Table 7 presents parameters of Eq. (4) and coefficients of determination R2. The proposed formula fits the experimental data very well (R2 = 0.911–0.980) what is illustrated in Fig. 4 for the AC1 and SMA1 mixes. It is worth noting that the sign of the parameter a2 is different for the AC and SMA mixes. The sign ‘‘–’’ of a2 is characteristic for AC mixes and indicates that the curve PHR (log t) is convex what is caused by the fast increase of stiffness modulus S of AC mixes during first hours of isothermal storage and very small, or no increase, for long time of storage. The sign ‘‘+’’ of a2 is characteristic for SMA mixes and indicates that the curve PHR (log t) is concave

Fig. 1. Typical changes of the stiffness modulus (S) and indirect tensile strength (ITS) tested at 20 °C in relation to the isothermal storage time at 20 °C for asphalt concrete (AC1) and stone mastic asphalt (SMA1).

196

J. Judycki / Construction and Building Materials 61 (2014) 191–199

Fig. 2. The stiffness modulus and indirect tensile strength at 20 °C in relation to isothermal storage time for all the tested mixes.

Table 5 Stiffness Growth Rate SGR and approximated S(1) at 20 °C. Mix type

Type of asphalt binder

S(1) (MPa)

dðsÞ SGR ¼ dðlog tÞ

R2

AC1 AC2 AC3 SMA1 SMA2 SMA3

50/70 Neat 35/50 Neat 50/70 Modified with SB in asphalt plant 80C Modified with SBS in refinery 35/50 Neat 50B Modified with SBS in refinery

24,527 26,601 25,079 20,915 22,443 23,839

1,193 1,044 1,172 1,301 1,159 1,685

0.94 0.90 0.99 0.95 0.98 0.97

Table 6 Physical Hardening Ratio PHR at 20 °C. Isothermal storage time at 20 °C, h

AC1 50/70 neat

AC2 35/50 neat

AC3 50/70 modified in asphalt plant

SMA1 80C SBS modified in refinery

SMA2 35/50 neat

SMA3 50B SBS modified in refinery

1 3 24 72 120 384

1.000 1.057 1.095 1.118 1.120 1.134

1.000 1.052 1.064 1.086 1.111 1.108

1.000 1.031 1.065 1.085 1.109 1.122

1.000 1.036 1.056 1.101 1.133 1.168

1.000 1.017 1.063 1.077 1.112 1.131

1.000 1.031 1.077 1.123 1.170 1.169

what is caused by the slower increase of stiffness modulus S of SMA mixes during first hours of isothermal storage and continuous steady increase for longer time of storage. It can be clearly seen in Fig. 4 that the sign of the parameter a2 indicates the nature of growth of stiffness modulus with time of isothermal storage. 6.2. Effect of physical hardening on stiffness modulus of asphalt mixes Asphalt binders used for asphalt mixes were not tested in this study thus the physical hardening potential of the binders was unknown. Therefore the analysis presented below is based on the data from the mix testing.

Isothermal storage at 20 °C caused the increase of stiffness moduli measured at 20 °C, for all the tested mixes, both for AC and SMA. It can be clearly seen in Fig. 1 and Table 3 that the stiffness modulus increases with the duration of isothermal storage time, initially at a high rate until 24 h and with slowing rate after 24 h of storage. Between 120 h and 384 h (5 and 16 days) of isothermal storage the increase of stiffness modulus is very small. The increase of stiffness of the tested mixes measured at 20 °C after 384 h (16 days) of isothermal storage at 20 °C was in a range from 11% to 17% (Table 6). The results obtained in this research show good agreement with the data presented earlier by Soenen et al. [14] for dense graded asphalt concrete. In [14] tests were

J. Judycki / Construction and Building Materials 61 (2014) 191–199

197

Fig. 3. Physical Hardening Ratio at 20 °C for asphalt concrete (AC) and stone mastic asphalt (SMA).

Table 7 Parameters of the formula for Physical Hardening Ratio PHR. Parameters

AC1

AC2

AC3

SMA1

SMA2

SMA3

a1 a2 R2

0.0981 0.0185 0.980

0.0721 0.0117 0.911

0.0522 0.0018 0.985

0.0336 0.0123 0.970

0.0352 0.0064 0.983

0.0649 0.0022 0.957

performed at 15 °C and 25 °C, and after 24 h of isothermal storage the increase of mix stiffness was in a range (5–12%). In this research the tests were performed at 20 °C and after 24 h the increase was in a similar range of 6–10% (Table 6). Two factors affecting physical hardening of asphalt mixes will be discussed further: the effect of the type of the mix (AC and SMA) and the effect of the type of asphalt binder (neat and polymer modified). The process of physical hardening of SMA and AC mixes is different. The rate of physical hardening during first hours of isothermal storage is faster for AC than for SMA. However, after

384 h the Physical Hardening Ratio becomes greater for SMA (Table 6). Fig. 5 shows that 60% of maximum hardening recorded after 384 h in AC occurs after 10–40 h, while in SMA after 40–70 h. In addition, AC is stiffer at low temperature than SMA and its indirect tensile strength is slightly greater (Fig. 2). Some dissimilarity in physical hardening and in low-temperature properties between AC and SMA are caused by the differences in their composition, grading and voids content. In SMA, crushed aggregate forms a continuous matrix in which voids are filled with asphaltic mastic (mixture of asphalt binder and filler) while AC is gradually graded. SMA contains much more asphalt binder and asphaltic mastic (the sum of binder and filler). It is worth to have a close look at Table 5 at the Stiffness Growth Ratio of mixes containing polymer modified and neat binders. The greatest Stiffness Growth Rate (SGR) during isothermal storage was recorded for the SMA3, for which SGR = 1685, what means that the stiffness S increased by 1685 MPa for each subsequent decade of time in hours. The mix with the second greatest SGR = 1301 was the SMA1. It is characteristic that the

Fig. 4. Comparison of calculated (lines) and measured (points) Physical Hardening Ratios for the AC1 and SMA1 mixes (a) PHR vs. time in normal scale and (b) PHR vs. time in logarithmic scale.

Fig. 5. Relative hardening of AC and SMA in relation to storage time.

198

J. Judycki / Construction and Building Materials 61 (2014) 191–199

SMA3 and the SMA1 contained the binders modified with SBS copolymer in the same refinery. All other mixes had a less SGRvalue in a range from 1044 to 1172. Three of them (AC1, AC2, SMA2) contained unmodified neat binder and the AC3 contained 50/70 binder modified with SB copolymer in asphalt plant. It is interesting that the AC1 mix with neat unmodified binder 50/70 showed almost the same physical hardening (SGR = 1193) as the AC3 mix with the same 50/70 binder which was modified in asphalt plant with the SB copolymer (SGR = 1172). These results suggest that in the case of the AC3 mix the addition of the SB copolymer to the binder in asphalt plant did not affect the physical hardening of the mix. Other low-temperature properties of these mixes (AC1 and AC3) only slightly differed. The described results indicate that the susceptibility to physical hardening of mixes containing polymer modified binders is mostly dependent on the properties of the base binders which were used for modification. However, this finding requires further more systematic investigations. 6.3. Effect of isothermal storage at 20 °C on tensile strength Contrary to expectations, the indirect tensile strengths of all the tested mixes were not affected by isothermal storage and were the same, allowing for normal deviations from the average values, during 16 days of isothermal storage at 20 °C (Table 4 and Figs. 1 and 2). It was reported in 1989 by El Hussein et al. [27], that microcracking occurred in asphalt samples during storage at low temperature due to the large differences in thermal contraction coefficients between asphalt binder and aggregate. Such microcracking would have caused some decrease in the tensile strength of the mixes. However, it has been proven in this study that there was no reduction in the indirect tensile strength during isothermal storage at 20 °C up to 16 days for all the tested mixes, both AC and SMA. The phenomenon suggested by El Hussein et al. [27] that low temperature induces micro-cracking in asphalt film covering grains of aggregate, and in consequence causes decrease in tensile strength of the asphalt mix, has not been proven in this research. It might be possible that such micro-cracking phenomenon can occur in asphalt mixes at much more extreme low temperatures than 20 °C but it requires further testing.

7. Conclusions 1. The laboratory tests have shown evident physical hardening of all the tested mixes (asphalt concrete AC and stone mastic asphalt SMA) manifested by an increase of their stiffness moduli tested under repeated dynamic loading after isothermal storage at low temperature. The maximum recorded increase of stiffness was from 11% to 17% after 120 h of storage at 20 °C. 2. Two measures of physical hardening: Stiffness Growth Rate (SGR) and Physical Hardening Ratio (PHR) were introduced to differentiate the susceptibility of different asphalt mixes to physical hardening. 3. It was found, that the process of physical hardening of stone mastic asphalt SMA and asphalt concrete AC was different. The initial growth of stiffness during storage at 20 °C was slower for the SMA. However, after extended storage at 20 °C during 5–16 days the Physical Hardening Ratio became greater for the SMA as compared with the AC mixes. In general, the stiffness and indirect tensile strength of the SMA was lower than that of the AC mixes. The dissimilarities in physical hardening and in low-temperature properties between the AC and SMA mixes were caused by the differences in their composition, grading, and the contents

of asphalt binders, asphaltic mastic (the sum of asphalt binder and filler) and voids. Considering the growing applications of the SMA for wearing courses of asphalt pavements the low-temperature behavior of these mixes requires further studies. 4. The tested mixes showed different behavior after lowtemperature storage what indicated that the mix type, its composition and asphalt binders used are important. The results indicated that the susceptibility to physical hardening of mixes containing polymer modified binders is primarily dependent on the properties of base binder used for polymer modification. These aspects require further studies. 5. The evident increase of stiffness of asphalt mixes during isothermal storage at low temperature which has been confirmed in this study should be considered in the analysis of the development of thermal stresses and low-temperature cracking in asphalt pavements. 6. Not only stiffness but also tensile strength of asphalt mixes is extremely relevant to the development of low temperature cracks in asphalt pavements. It has been proven in this research that the indirect tensile strength at 20 °C for all six tested asphalt mixes (AC and SMA) remained almost constant during isothermal storage at 20 °C for 16 days, allowing some normal scatter of results.

References [1] Traxler RN, Schweyer HE. Increase in viscosity of asphalts with time. Proc Am Soc Test Mater 1936;36(part II):544–50. [2] Traxler RN, Coombs CE. Development of internal structure in asphalt with time. Proc Am Soc Test Mater 1937;37(Part II):549–55. [3] Brown AB, Sparks JW, Smith FM. Steric hardening of asphalts. Proc Assoc Asphalt Paving Technol 1957;26:486–94. [4] Blokker PC, Van Hoorn H., Durability of bitumen in theory and practice. In: Proc fifth world petroleum congress, section VI, paper 27, June 1–5. 1959. p. 417– 32. [5] Struik LCE. Physical ageing in amorphous polymers and other materials. Amsterdam, The Netherlands: Elsevier; 1978. [6] Pechenyi BG, Kuznetsov OI. Formation of equilibrium structures in bitumens. Trans Khimiya i Technologiya Topiv i Masel 1990;7:32–4. [7] Anderson DA, Christensen DW, Bahia HU, Dongre R, Sharma MG, Antle CE. Binder characterization and evaluation. In: Physical characterization, vol. 3. Publication SHRP-A-369, Strategic Highway Research Program, National Research Council, Washington D.C., 1994. [8] Bahia HU, Anderson DA. Physical hardening of paving grade asphalt as related to compositional characteristics. In: 204th American Chemical Society National Meeting, Division of Fuel Chemistry, vol. 37, no. 4, preprints of papers, Washington D.C., August, 1992. [9] Anderson DA, Morasteanu MO. Physical hardening of asphalt binders relative to their glass transition temperature. Transport Res Rec 1999;1661:27–34. [10] Hesp SAM, Iliuta S, Shirokoff JW. Reversible aging in asphalt binders. Energy Fuel 2007;21:1112–21. [11] Hesp SAM, Soleimani A, Subramani S, Philips T, Smith D, Marks P, et al. Asphalt pavement cracking: analysis of extraordinary life cycle variability in Eastern and Northeastern Ontario. Int J Pavement Eng 2009;10(3):209–27. [12] Togunde OP, Hesp SAM. Physical hardening in asphalt mixtures. Int J Pavement Res Technol 2012;5(1):46–53. [13] Lu X, Isacsson U. Laboratory study on the low temperature physical hardening of conventional and polymer modified bitumens. Constr Build Mater 2000;14:79–88. [14] Soenen H, Ekblad J, Lu X, Redelius P. Isothermal hardening in bitumen and in asphalt mixes. In: Proceedings of the 3rd Euroasphalt and Eurobitume congress, vol. 2, Vienna, May, 2004, p. 1364–75. [15] Tabatabaee H, Mangiafico S, Velasquez R, Bahia HU. Investigation of low temperature cracking in asphalt pavements. National pooled fund study – phase II, task 2, subtask on physical hardening. University of WisconsinMadison, October, 2010. [16] Shenoy A. Stress relaxation can perturb and prevent physical hardening in a constrained binder at low temperatures. Int J Road Mater Pavement Des 2002;3(1):87–94. [17] Dongre R. Effect of physical hardening on stress relaxation behaviour of asphalt pavements. In: Proceedings of Euroasphalt & Eurobitume congress, Barcelona, 2000. p. 220–7. [18] Baglieri O, Dalmazzo D, Barazia M, Tabatabaee HA, Bahia HU. Influence of physical hardening on low-temperature properties of bitumen and asphalt mixtures. Procedia – Soc Behav Sci 2012;53:504–13.

J. Judycki / Construction and Building Materials 61 (2014) 191–199 [19] Bahia HU, Tabatabaee H, Valasquez R. Importance of bitumen physical hardening for thermal stresses buildup and relaxation in asphalt. In: Proceedings of the 5th Euroasphalt and Eurobitume Congress, Istanbul, June, 2012. [20] Hesp SAM, Genin SN, Scafe D, Shurvell HF, Subramani S. Five years performance review of a Northern Ontario pavement trial: validation of Ontario’s double-edge-notched tension (DENT) and extended bending beam rheometer (BBR) test methods. Proc Can Tech Asphalt Assoc 2009;54:99–126. [21] Zhao MO, Hesp SAM. Performance grading of the Lamont, Alberta C-SHRP pavement trial binders. Int J Pavement Eng 2006;7(3):199–211. [22] Lu X, Isacsson U, Ekblad J. Influence of polymer modification on low temperature behaviour of bituminous binders and mixtures. In: 6th RILEM symposium, performance testing and evaluation of bituminous materials, Zurich, 2003. p. 435–41. [23] Romero P, Youtcheff J, Stuart K. Low-temperature physical hardening of hotmix asphalt. Transport Res Rec 1999;1661:22–6.

199

[24] Tabatabaee HA, Velasquez R, Arshadi A, Bahia HU. Investigation of low temperature cracking in asphalt pavements. In: National pooled fund study – phase II, task 5 – Modeling of asphalt mixtures contraction and expansion due to thermal cycling, University of Wisconsin-Madison, January, 2012. [25] Tabatabaee HA, Velasquez R, Bahia HU. Modeling thermal stress in asphalt mixtures undergoing glass transition and physical hardening. Transport Res Rec 2012;2296:106–14. [26] Falchetto AC, Marasteanu MO. Physical hardening: from binders to mixtures. Int J Roads Airports 2011;1(1):18–34. [27] El Hussein HM, Kim KW, Ponniah J. Asphalt concrete damage associated with extreme low temperatures. J Mater Civil Eng 1989;10(4):269–74. [28] Kim YR, Seo Y, King M, Momen M. Dynamic modulus testing of asphalt concrete in indirect tension mode. J Transport Res Board, Natl Res Counc, Washington, D.C. 2004;1891:163–73.