Low Temperature Crack Resistance of Stone Mastic Asphalt Affected by Its Nominal Maximum Size and Asphalt Binders

: To study the effect of nominal maximum particle size (NMS) of stone mastic asphalt (SMA) gradation and a mixed modiﬁed asphalt binder on the low temperature crack resistance of SMA, SMA asphalt mixtures with three different NMS, two styrene-butadiene-styrene (SBS) and crumb rubber modiﬁer (CRM) hybrid-modiﬁed asphalt binders were designed.Thermal strength restrained specimen test (TSRST) and bending beam test (BBT) were performed on the mixtures. The freezing-fracture temperature and strength from TSRST, bending tensile strain and stiffness from BBT at low temperature were evaluated. Finally, an intrinsic relationship was established between these properties obtained from the two tests. The results showed that the following: (1) the SMA-5, the best of all the three NMS, has a freezing-fracture temperature of − 31.5 ◦ C, which is 18.9% and 48.6% lower than those of SMA-10 and SMA-13, respectively, while its freezing fracture strength is 6.15 MPa, which is 95.2% and 243.6% higher than those of SMA-10 and SMA-13, respectively; (2) the bending failure strain of SMA-5 is 4649 µε , which is higher than those of SMA-10 and SMA-13 by 11.3% and 21.9%, respectively; (3) increased CRM dose in the SBS-modiﬁed SMA improves the crack resistance at low temperature; (4) the bending failure strain of the SMA has the best correlation with the freezing-fracture temperature, with a correlation coefﬁcient of about 0.8.


Introduction
Polymer-modified stone mastic asphalt (SMA) is widely used in highway pavements with high traffic volume in China to improve the overall surface layer performance at high temperature.The main reason is that SMA has good rutting assistance by the stone-tostone aggregate skeleton structure for loading and transferring, in addition to the better performance of the modified binders over unmodified binders.With the increase in traffic volume and the appearance of extreme low temperature climate, more and more cracks appear in asphalt pavement.These cracks could be caused by repeated loading fatigue and low temperature shrinkage.Cracking at low temperature, as one of the cracks asphalt pavements suffer, destroys not only the continuity of the pavement but also the integrity of the pavement structure.It seriously affects the normal service life of asphalt pavement [1,2].The modified SMA has been approved as one of the asphalt mixtures that performed well in crack resistance.
There are many factors influencing the low temperature crack resistance of asphalt mixtures, including the gradation and maximum nominal particle size of the mixture, the grade of asphalt binder, volumetric properties of the mixture, etc. [3,4].Research exist on how modified asphalt binders and gradation of aggregates affect the cracking resistance of asphalt mixtures.In addition, there are many tests and index to evaluate the crack resistance of asphalt mixtures.Wang found that by using a modified asphalt binder of 4% rock asphalt with Pen 90 base asphalt, the asphalt mixtures reached a bending strain of 2403 µε, meeting the design requirements of low temperature in cold regions in winter (≥2300 µε) [5].Ji et al. found that the failure strain of the SBS/CRM hybrid-modified SMA-13 asphalt mixture of increased by about 14% compared with that of the SBS-modified one at −10 • C [6].Le et al. showed that the low temperature rheological properties of asphalt could be significantly improved by using a Sasobit modifier with a dosage of 3.0% [7].Fu et al. selected bis (2-ethylhexyl) adipate (DOA) as a plasticizer to modify base asphalt, and the results showed that when the amount of plasticizer was increased to 3.0%, the failure strain of modified asphalt increased by about 66% compared with base asphalt [8].The failure strain increased by about 66% compared with the base asphalt.Chen et al. adopted the improved Superpave design method to design a rubber asphalt anti-crack layer asphalt mixture, which could achieve a better function of absorbing and reflecting cracks [9].Wan studied the crack resistance of a rubber-asphalt mixture with three gradation types of SMA-13, AC-13 and Novachip (ultra-thin wearing course) Type-B; the results showed that the splitting and bending strength of SMA-13 with a smaller void at a low temperature were greater than that of the Novachip Type-B rubber asphalt mixture with a larger void [10].Wang et al. analyzed the anti-reflection crack performance of AC-10 and CAM-10 asphalt mixtures of strength absorbing layers with different void ratios through the four-point bending test and overlay test.He found that the CAM-10 asphalt mixture with a target void ratio of 2% had the best anti-cracking performance at a low temperature [11].Marasteanu et al. proposed the superiority of the lowtemperature performance test of asphalt based on prefabricated notch three-point bending fracture mechanics [12,13].Yan et al. applied the disc tensile test (DCT) to conduct a fracture test of the asphalt mixture at −12 • C, and selected the fracture strain capacity limit as the evaluation index [14].
There are many research projects on the gradation, the NMS of gradation, and the shape of aggregates on the rutting resistance [15][16][17][18][19].However, there is little research on the effect of gradation NMS on the low temperature crack resistance of the SMA mixtures.Therefore, in order to investigate the influence of NMS and modified asphalt binder on the low temperature crack resistance of SMA, six SBS and CRM hybrid-modified SMA mixtures with three nominal maximum sizes were designed.The freezing-fracture temperature and strength of different asphalt mixtures were measured by TSRST, and the failure strain and stiffness of different asphalt mixtures were measured by BBT at low temperature.The low temperature cracking resistance of SMA was analyzed and evaluated by four test indexes of the above two test methods.Finally, the Grey relational analysis method was used to establish the correlation between the properties of low temperature shrinkage and bending loading resistance.

Materials
SBS-modified asphalt used in this study was taken from Suzhou Sanchuang Road Engineering Co., Ltd., Jiangsu Province.The main properties are shown in Table 1.The base asphalt was Pen 70 petroleum asphalt cement from Zhejiang Weike, and the three indexes and dynamic viscosity of the asphalt binder were in line with JTG F40-2004 "Technical Specifications for Construction of Highway Asphalt Pavement".The dosage of SBS used for the modification was 5%, and the main properties are shown in Table 2.A crumb rubber modifier (CRM) was added into the mixing drum using the dry mixing method to prepare all the hybrid SBS/CRM-modified asphalt mixtures.The CRM content added was 10% and 20%.The size of CRM used was 80 mesh, and the properties of the CRM are shown in Table 3. Tables 4 and 5 showed the properties of mineral filler and the physical properties of lignin fiber.The coarse and fine aggregates are both basalt, which were divided into the following four groups: 10-15 mm, 5-10 mm, 3-5 mm and 0-3 mm.One can refer to Table 6 for their technical properties.Three asphalt mixtures of SMA-13, SMA-10 and SMA-5 were designed by following the JTG F40-2004 Technical Specification for Construction of Highway Asphalt Pavement.The pass rate and combined gradation curves are presented in Figure 1   Three asphalt mixtures of SMA-13, SMA-10 and SMA-5 were designed by following the JTG F40-2004 Technical Specification for Construction of Highway Asphalt Pavement.The pass rate and combined gradation curves are presented in Figure 1 and Table 7.The Marshall test method was used to determine the optimum asphalt content of all the mixtures.The volumetric properties of the asphalt mixtures are shown in Table 8.

Methods
For TSRST, the asphalt mixture specimen was obtained by cutting a slab that was compacted by the rolling method.A UTM universal testing machine was used for TSRST, as shown in Figure 2. Prism samples with sizes of 50 mm × 50 mm × 250 mm were tested at an initial temperature of 5 • C and a cooling rate of 15 The Marshall test method was used to determine the optimum asphalt content of all the mixtures.The volumetric properties of the asphalt mixtures are shown in Table 8.

Methods
For TSRST, the asphalt mixture specimen was obtained by cutting a slab that was compacted by the rolling method.A UTM universal testing machine was used for TSRST, as shown in Figure 2. Prism samples with sizes of 50 mm × 50 mm × 250 mm were tested at an initial temperature of 5 °C and a cooling rate of 15 °C/h.The mean values of three parallel specimens were obtained for analysis.In the bending beam test, a bending beam creep tester was used, as shown in Figure 3.The loading rate was 50 mm/min.The beam size was 200 mm × 30 mm × 35 mm, and the test temperatures were −10 °C, −5 °C and 0 °C.Three parallels were tested and the results were averaged.

Thermal Strength Restrained Specimen Test
The results were presented in six groups, as shown in Table 9 and Figure 4. Th

Thermal Strength Restrained Specimen Test
The results were presented in six groups, as shown in Table 9 and Figure 4.The groups of 1, 2 and 3 were designated to the three gradations of SMA-13, SMA-10 and SMA-5 with a fixed asphalt binder modified with 5% SBS + 10% CRM, respectively.The groups of 4, 5 and 6 were designated to SMA-13, SMA-10 and SMA-5 with a fixed asphalt binder modified with 5% SBS + 20% CRM, respectively.Three parallel trials were conducted for each group.

Thermal Strength Restrained Specimen Test
The results were presented in six groups, as shown in Table 9 and Figure 4.The groups of 1, 2 and 3 were designated to the three gradations of SMA-13, SMA-10 and SMA-5 with a fixed asphalt binder modified with 5% SBS + 10% CRM, respectively.The groups of 4, 5 and 6 were designated to SMA-13, SMA-10 and SMA-5 with a fixed asphalt binder modified with 5% SBS + 20% CRM, respectively.Three parallel trials were conducted for each group.Freezing fracture strength/MPa  Generally, the freezing fracture temperature ranged from −17 • C to −33 • C, and the freezing fracture strength ranged from 1.5 to 6.5 KN for the mixtures discussed.The effect of NMS and asphalt binders on the results are discussed in the following section.
(1) The freezing fracture temperatures of groups 1, 2 and 3 decreased from −20.4 • C to −23.8 • C and −29.2 • C with the decreased NMS from 13, 10 to 5 mm, respectively.The freezing fracture temperatures of SMA-10 and SMA-5 decreased by 16.7% and 43.1% as compared with that of SMA-13, respectively.Similarly, the freezing fracture temperatures of groups 4, 5 and 6 became lower with the decreased NMS.Therefore, when the amount of CRM in the mixed modified binder was constant, the decrease in NMS can reduce the freeze-crack temperature of SMA.This may be caused by a higher OAC for a smaller NMS, as shown in Table 8.In addition, with the same gradation, increasing the content of CRM can slightly reduce the freezing fracture temperatures of SMA, because the higher content of CRM increased the viscosity of the mixed modified asphalt binder.
(2) The freezing fracture strength of groups 1, 2 and 3 increased from 1.6 MPa of SMA-13, 2.47 MPa of SMA-10 to 5.0 MPa of SMA-5.An increasing rate of 54.4% and 212.5% was obtained for SMA-13 and SMA-10, as compared with SMA-5, respectively.This trend was true for groups 4, 5 and 6, which achieved a higher increase rate in the strength for the same decrease in the NMS.In simple terms, the freezing fracture strength was increased by the deceased NMS when the amount of CRM in the mixed modified binder was constant.This may be again caused by a higher OAC for a smaller NMS, as shown in Table 8.Furthermore, addition of a higher CRM content, resulting in a viscous asphalt binder, will result in the higher freezing fracture strength of SMA.

Bending Beam Test
Table 10 shows the bending beam test results of six different types of asphalt mixtures at different temperatures.Figure 5 shows that the NMS has an obvious influence on the bending failure strain of hybrid-modified SMA when the test temperature and CRM content are constant.In addition, the bending failure strain decreased generally as the test temperature increased and the dose of CRM also decreased.
. Sci. 2022, 12, x FOR PEER REVIEW 8 of Figure 5 shows that the NMS has an obvious influence on the bending failure stra of hybrid-modified SMA when the test temperature and CRM content are constant.addition, the bending failure strain decreased generally as the test temperature increas and the dose of CRM also decreased.Taking the case of the modified binder of 5% SBS + 20% CRM as an example, t bending fracture strains of SMA-5, SMA-10 and SMA-13 at 0 °C were 4669 με, 4177 με a 3813 με, respectively.The bending fracture strain decreased by 10.2% and 18.0% for SM 10 and SMA-15, compared with the value of SMA-5.For the case of the modified bind of 5% SBS + 10% CRM, when the CRM content was reduced, the bending fracture strain SMA-5 is 11.3% and 21.9% higher than those of SMA-13 and SMA-10.The trend was general true for the other test temperatures and doses of CRM.
Figure 6 showed the effect of NMS on the stiffness modulus of the hybrid-modifi SMA.The stiffness increased generally as the test temperature increased and also the do of CRM.At the same temperature, the stiffness of SMA-10 was greater than that of t Taking the case of the modified binder of 5% SBS + 20% CRM as an example, the bending fracture strains of SMA-5, SMA-10 and SMA-13 at 0 • C were 4669 µε, 4177 µε and 3813 µε, respectively.The bending fracture strain decreased by 10.2% and 18.0% for SMA-10 and SMA-15, compared with the value of SMA-5.For the case of the modified binder of 5% SBS + 10% CRM, when the CRM content was reduced, the bending fracture strain of SMA-5 is 11.3% and 21.9% higher than those of SMA-13 and SMA-10.The trend was in general true for the other test temperatures and doses of CRM.
Figure 6 showed the effect of NMS on the stiffness modulus of the hybrid-modified SMA.The stiffness increased generally as the test temperature increased and also the dose of CRM.At the same temperature, the stiffness of SMA-10 was greater than that of the other two NMS.The trend was true for both doses of CRM.It must be noted that the tread of the stiffness changed with the NMS in a different way from that of the freezing fracture strain.
of 5% SBS + 10% CRM, when the CRM content was reduced, the bending fracture strain SMA-5 is 11.3% and 21.9% higher than those of SMA-13 and SMA-10.The trend was general true for the other test temperatures and doses of CRM.
Figure 6 showed the effect of NMS on the stiffness modulus of the hybrid-modifi SMA.The stiffness increased generally as the test temperature increased and also the do of CRM.At the same temperature, the stiffness of SMA-10 was greater than that of t other two NMS.The trend was true for both doses of CRM.It must be noted that the tre of the stiffness changed with the NMS in a different way from that of the freezing fractu strain.The relationship between bending stiffness of the SMA and test temperature w studied.Taking the case of the SMA-13 with hybrid-modified 5%SBS + 20%CRM as example, when the test temperature increased from −10 °C to −5 °C and 0 °C, the bendi stiffness decreased from 2966 MPa, 2471 MPa to 2337 MPa.A decreasing rate of 16.7% a The relationship between bending stiffness of the SMA and test temperature was studied.Taking the case of the SMA-13 with hybrid-modified 5%SBS + 20%CRM as an example, when the test temperature increased from −10 • C to −5 • C and 0 • C, the bending stiffness decreased from 2966 MPa, 2471 MPa to 2337 MPa.A decreasing rate of 16.7% and 21.2% was observed for temperatures of −5 • C and 0 • C, respectively, as compared with that of −10 • C. The higher the temperature, the smaller the stiffness and the better the crack resistance at low temperature.These trends are basically similar to the existing findings of asphalt mixtures.

Grey Relational Analysis
Grey relational analysis was used to study the internal relationship between the properties of the load-caused cracking resistance and temperature shrinkage cracking of the SBS/CRM-modified asphalt mixtures under low temperature.As one of the Grey system analysis methods, the Grey relational analysis method measures the correlation degree between factors according to the similarity or dissimilarity of the development situation among factors (Grey relational degree) [20].
As the constrained test piece temperature strength test is one of the best evaluation methods for evaluating the low temperature performance of SMA in the engineering field [21], the freezing fracture temperature and strength of the SMA were taken as reference series in this study.The bending beam test indexes for evaluating vehicle load cracking were taken as comparison series.Grey relational degree analysis was carried out on the bending fracture strain (X1~X3) and stiffness (X4~X6) of SMA at different test temperatures.The results are shown in Table 11 and Figure 7.It can be observed from Figure 7 that the correlation degree between freezing fracture temperature and low temperature bending beam test in the temperature-strength test of constrained specimens is the best.Both are above 0.7, which was greater than the correlation coefficient of freezing fracture strength, indicating that freezing fracture temperature It can be observed from Figure 7 that the correlation degree between freezing fracture temperature and low temperature bending beam test in the temperature-strength test of constrained specimens is the best.Both are above 0.7, which was greater than the correlation coefficient of freezing fracture strength, indicating that freezing fracture temperature can better characterize the crack resistance of SMA under a low temperature load.Secondly, the correlation degree between bending fracture strain and freezing fracture temperature at three different temperatures was about 0.8.The correlation degree of the stiffness at each temperature was lower than that of the bending fracture strain, indicating that bending fracture strain was more suitable than stiffness to characterize the low temperature performance of the SMA with hybrid modification.In addition, the variation trend of the correlation degree between freezing fracture strength and low temperature bending beam test was basically similar to that of the freezing fracture temperature.The correlation degree between freezing fracture strength and bending stiffness at 0 • C is poor, only 0.634.It is consistent with the analysis results of stiffness above.

Conclusions
In this paper, the influence of different NMS and different CRM contents on the cracking resistance of SBS/CRM composite-modified asphalt SMA mixture was studied.The cracking resistance at low temperature of six different SMA mixtures was evaluated using the bending beam test and thermal strength restrained specimen test.The relationship between the properties obtained from BBT and TSRST was established.
The experimental results show that reducing NMS and increasing CRM content can significantly improve the low temperature crack resistance of the SMA mixture.NMS of SMA has a significant effect on the freezing fracture temperature and strength of the mixtures.The freezing fracture temperature was lowered by deceased NMS, regardless of the dose of CRM for the hybrid-modified binders.Addition of a higher CRM content, resulting in a viscous asphalt binder, will produce a lower freezing fracture temperature and higher strength.NMS of SMA has a significant effect on the bending fracture strain but has no obvious effect on the stiffness.The bending failure strain was lowered by deceased NMS, regardless of the dose of CRM for the hybrid-modified binders, while NMS of 10 mm meant the SMA could reach its highest stiffness.Addition of a higher CRM content, resulting in a viscous asphalt binder, will produce a higher strain and stiffness.The correlation degree between bending fracture strain and temperature was very good, with the coefficient of correlation between 0.77 and 0.81, while the coefficient of correlation degree between freeze fracture stiffness and strength decreased slightly, between 0.63 and 0.67.It is suggested that freezing fracture temperature and strength may be more suitable for characterizing crack resistance of asphalt mixture at low temperature.Addition of CRM in the SBS-modified binders and aggregates during the dry process made the SMA behavior better regarding the cracking resistance for both the freezing temperature and strength

Figure 2 . 1 Figure 3 .
Figure 2. Test setup using UTM.In the bending beam test, a bending beam creep tester was used, as shown in Figure 3.The loading rate was 50 mm/min.The beam size was 200 mm × 30 mm × 35 mm, and the test temperatures were −10 • C, −5 • C and 0 • C. Three parallels were tested and the results were averaged.Appl.Sci.2022, 12, x FOR PEER REVIEW 6 of 1

Figure 3 .
Figure 3. Test setup for low temperature bending beam creep.

Figure 3 .
Figure 3. Test setup for low temperature bending beam creep.

Figure 4 .
Figure 4. Freezing fracture temperature and freezing fracture strength of different asphalt mixtures.

Figure 4 .
Figure 4. Freezing fracture temperature and freezing fracture strength of different asphalt mixtures.

Figure 7 .
Figure 7. Correlation coefficient between different properties at low temperature.

Figure 7 .
Figure 7. Correlation coefficient between different properties at low temperature.

Table 2 .
Properties of SBS.

Table 3 .
Properties of CRM.

Table 4 .
Properties of mineral filler.

Table 5 .
Physical properties of lignin fiber.

Table 6 .
Properties of aggregate.

Table 6 .
Properties of aggregate.

Table 7 .
The sieve passing rate of three kinds of aggregates.

Percentage of Mass Passing of Different Particle Sizes/%
Figure 1.Combined gradation curves.

Table 7 .
The sieve passing rate of three kinds of aggregates.
• C/h.The mean values of three parallel specimens were obtained for analysis.

Table 9 .
Test results of asphalt mixture restraint specimens.

Table 9 .
Test results of asphalt mixture restraint specimens.

Table 10 .
Bending beam test results of asphalt mixtures.

Table 11 .
SBS/CRM-modified SMA performance index at low temperature.