- Email: [email protected]
Chemical and rheological investigation of high-cured crumb rubber-modified asphalt
Construction and Building Materials 123 (2016) 847–854
Contents lists available at ScienceDirect
Construction and Building Materials journal homepage: www.elsevier.com/locate/conbuildmat
Chemical and rheological investigation of high-cured crumb rubber-modified asphalt Naipeng Tang, Weidong Huang ⇑, Feipeng Xiao The Key Laboratory of Road and Traffic Engineering, Ministry of Education, Tongji University, Shanghai 201804, PR China
h i g h l i g h t s Active Polymer Index was defined to characterize the curing process of CRM Asphalt. Interaction temperature contributed more to the degradation than interaction time. There was linear relationship between chemical and rheological properties. API derived from ATR-FTIR test is a good indicator of degradation of crumb rubber.
a r t i c l e
i n f o
Article history: Received 28 April 2016 Received in revised form 18 June 2016 Accepted 26 July 2016
Keywords: High-cured CRM asphalt Degradation ATR-FTIR Active Polymer Index Elastic recovery
a b s t r a c t In the production of Terminal Blend (TB) rubberized asphalt, degradation of crumb rubber in asphalt matrix can help improve the storage stability and workability of crumb rubber modified asphalt. However, degradation process of crumb rubber modified (CRM) binders was not clear from molecular components to rheological properties. This study investigated chemical and rheological perspectives of the degradation process of CRM asphalt under high interaction temperature (220–280 °C) and extended curing time (2–8 h), namely high-cured CRM asphalt. Attenuated Total Reflection (ATR) Fourier Transform Infrared (FT-IR) Spectroscopy was used to characterize the degradation behaviors of CRM binders at varying curing temperatures and times. Active Polymer Index (API) was defined as the ratio between band areas at 978–918 cm1 and 850–785 cm1 of FTIR tests to characterize the released active polymer from vulcanized crumb rubber. Besides, dynamic shear rheometer (DSR) was used to perform elastic recovery (ER) test, dynamic oscillatory test and multiple stress creep recovery (MSCR) test. Test results indicated that the critical degradation temperature was 260 °C in this limited study. Interaction temperature contributed more to the degradation than interaction time. Pearson correlation analysis indicated that API derived from ATR-FTIR test was a good indicator of degradation of CRM binders. Ó 2016 Elsevier Ltd. All rights reserved.
1. Introduction Crumb rubber has been used in pavement industry over 50 years. Traditional wet-process technique, such as Asphalt Rubber, has shown disadvantages in storage stability and workability [1]. Attempts have been made to improve the storage stability or decrease the viscosity of crumb rubber modified (CRM) asphalt through various physical and chemical methods. Additives like vegetable oil, warm mix asphalt additives (Rediset and Evotherm) and bio-binder were found to effectively decrease the viscosity of CRM asphalt [2,3]. Activated or treated crumb rubber by using chemicals like furfural (C5H4O2) and polymeric compatibilizer were proved to improve the storage stability of CRMA significantly ⇑ Corresponding author. E-mail address: [email protected] (W. Huang). http://dx.doi.org/10.1016/j.conbuildmat.2016.07.131 0950-0618/Ó 2016 Elsevier Ltd. All rights reserved.
[4,5]. Besides, storage stability of CRMA can also be improved by using high shear method or using finer powdered rubber (200 mesh) [6,7]. Most of the above studies were concerned with the partial degradation of crumb rubber in asphalt matrix. In recent years, Terminal Blend (TB) rubberized asphalt technology arose in America has shown its potential in greatly improving storage stability of rubber/asphalt composite because crumb rubber is fully digested into the asphalt, while maintaining considerable performance in pavement practice. Full scale projects and heavy vehicle simulator testing in California indicated that fatigue performance of TB rubberized asphalt is better than that of asphalt rubber and base binder [8]. The FHWA’s Accelerated Loading Facility (ALF) testing results indicated that Texas TB rubberized asphalt offer better rutting resistance than Arizona wet process asphalt rubber when the latter provide better fatigue performance [9]. It was reported that the production of TB rubberized asphalt can be
848
N. Tang et al. / Construction and Building Materials 123 (2016) 847–854
achieved through degradation of crumb rubber in asphalt matrix using heat and shear, meanwhile polymers might be added to the degraded rubber/asphalt composite to meet the performance grade criteria [10]. It has been confirmed by various research that degradation of CRM binders can affect rheological properties significantly [11–13]. Meanwhile, molecular weight distribution and molecular structure present remarkable change from micro perspective [12,14]. It would be better to understand the degradation mechanism through qualitative or quantitative analysis of molecular weight distribution and molecular structure in combination with rheological investigations. The current research on molecular weight distribution of CRM binders can’t fully characterize the rubber/asphalt composite as some dissoluble rubber components were removed by syringe filter before Gel Permeation Chromatography (GPC) test [15]. As a rapid test method, Fourier Transform Infrared Spectroscopy (FT-IR) gains popularity and quantitative analysis using FT-IR has been successfully applied to polymer modified asphalt [16]. However, FT-IR was mostly used to qualitatively detect the released rubber components into asphalt [14,15]. The study of degradation process of crumb rubber in asphalt matrix is limited. The objective of this study is to investigate degradation process of high-cured CRM binders in terms of their chemical and rheological properties. Attenuated Total Reflection (ATR) Fourier Transform Infrared (FT-IR) Spectroscopy is used to explore the main element components of CRM binders at varying reacting temperatures and times. Active Polymer Index (API) is employed to characterize the released active polymer content from the vulcanized crumb rubber. Dynamic Shear Rheometer (DSR) is used to perform elastic recovery (ER) test, dynamic oscillatory test and multiple stress creep recovery (MSCR) test in this study.
2. Materials and methods 2.1. Materials One PG 64-22 base binder provided by ESSO Asphalt Company and one minus 30 mesh crumb rubber provided by China Rubber Resource Regeneration were used to prepare high-cured CRM binders. A 20% crumb rubber (by the weight of total binder) was mixed with base binder at four temperatures of 220 °C, 240 °C, 260 °C and 280 °C for 2, 4, 6 and 8 h. The blending speed was 400 rpm. The preparation of samples was reported by Abdelrahman [11], Billiter [12], Zanzotto [13] and Flanigan [17]. Besides, crumb rubber modified asphalt produced by adding 18% crumb rubber to base binder, blending for 0.5 h at 190 °C was served as a control binder in the aftermentioned FT-IR analysis. The detailed test samples are presented in Table 1. For example, for binder blended at 220 for 2 h, the produced binder is marked as E22.2 in Table 1. As storage stability is very important for engineering use, the separation test were performed for each sample through keeping an aluminium tube containing 50 ± 0.5 g binder at 163 °C
Table 1 Description of the test samples. 80% base binder + 20% crumb rubber Blending temperature
220 °C 240 °C 260 °C 280 °C
Blending time 2h
4h
6h
8h
E22.2 E24.2 E26.2 E28.2
E22.4 E24.4 E26.4 E28.4
E22.6 E24.6 E26.6 E28.6
E22.8 E24.8 E26.8 E28.8
Table 2 Storage stability test results. Blending temperature
Softening point
Blending time
°C
2h
4h
6h
8h
220 °C
Up Bottom Difference
67.3 68 0.7
58.3 61.9 3.6
69.6 68.2 1.4
58.9 55.7 3.2
240 °C
Up Bottom Difference
53.4 57.6 4.2
47.4 50.3 2.9
54.7 54.2 0.5
51.3 49.2 2.1
260 °C
Up Bottom Difference
49.1 52.3 3.2
48.2 49.9 1.7
44.8 46.8 2
47.3 48.2 0.9
280 °C
Up Bottom Difference
49.1 50.8 1.7
48.5 49 0.5
46.1 45.5 0.6
47.5 46.8 0.7
for 48 h according to ASTM D7173. The top and bottom parts of the tube were collected for softening point test (ASTM D36) and the softening point difference between the top and bottom is used as an index to evaluate the storage stability. In China, if the difference is less than 2.5 °C, the binder is considered to have good storage stability. The storage stability test results are presented in Table 2. On the whole, increasing interaction temperature and extending curing time will help to improve the storage stability. 2.2. Attenuated Total Reflection (ATR) Fourier Transform Infrared (FTIR) Spectroscopy The infrared spectra values were collected using a Bruker TENSOR FT-IR spectrometer equipped with a reflection diamond ATR accessory. About 1 g asphalt binder was put on the surface of ATR diamond and fixed using the steel loader. The asphalt binder completely contacted the diamond homogenously. Thirty-two scans were averaged within the wavenumber range of 4000– 600 cm1 for each sample and then these average spectrum values were obtained. In this study, the peak at 965 cm1 attributed to CAH out of plane bending of trans-alkene showed a significant change during the interaction process of high-cured CRM asphalt. As FTIR peak intensities are sensitive to the concentration of components and the thickness of the sample [14], it would be better to normalize the band areas between 978 and 918 cm1 to eliminate the effect of sample thickness on FTIR results. In this study, the peak at 810 cm1 attributed to C-H out of plane bending of aromatics in base binder was found to vary little in terms of band areas and it has been successfully used as a reference peak in the quantitative analysis of SBS modified asphalt [18]. In this study, band areas between 978 and 918 cm1 (AR965) were divided by that of 850–785 cm1 (AR810) and the ratio is defined as Active Polymer Index (Eq. (1)).
Active Polymer Index ðAPIÞ ¼ AR965 =AR810
ð1Þ
The AR965 and AR810 were calculated in Thermo Scientific OMNICTM software automatically. The integration scopes were from 978 to 918 cm1 for 965 cm1 and 850 to 785 cm1 for 810 cm1 (Fig. 1). Band areas values rather than peak absorbance were used in this study because band areas varied little within three replicates of each sample [19]. The carbonyl index was also calculated to investigate the aging status of high cured CRM binders during the preparation. The calculation method was similar to that of API and the integration scopes for carbonyl were from 1678 to 1725 cm1. The base binder
849
N. Tang et al. / Construction and Building Materials 123 (2016) 847–854 0.30
peak strain
0.28 810
3
2 unrecovered strain
Strain
0.26
0.24
1
965
Absorbance
recovered strain
4
0.22
0 0.20 1000
950
900
850
800
750
0
700
2
4
6
8
10
Time (s)
Wavenumber (cm -1) Fig. 1. Illustration of band areas between 978–918 cm1 and 850–785 cm1.
Fig. 2. Typical creep and recovery cycle in MSCR test.
under Rolling Thin Film Oven (RTFO) short term aging and Pressure Aging Vessel (PAV) long term aging were used as control binders.
modulus curve were then used to build phase angle master curve. The newly developed double-logistic (DL) mathematical function by Asgharzadeh and Tabatabaee [24] was taken for the modelling of phase angle master curve. The reference temperature of 25 °C was chosen in the production of complex modulus and phase angle master curve [25].
2.3. Elastic Recovery test based on dynamic shear rheometer Elastic recovery (ER) was used to investigate the elastic properties of high-cured CRM asphalt. The test procedure was originally developed by Clopotel [20] and shown as follows: (1) Rolling Thin Film Oven (RTFO) aged sample is tested on the 8 mm plate in TA AR1500ex DSR. (2) The sample is stored at 25 °C for 20 min before test. (3) A constant strain rate of 0.023 s1 is applied for 2 min in a peak hold mode of flow procedure provided by TA’s instrument control software. (4) A constant zero shear stress is applied for a period of 1 h in a peak hold mode of flow procedure provided by TA’s instrument control software. (5) Calculate the elastic recovery at the end of relaxation using Eq. (2).
ER-DSR ¼
recovered strain at the end of the relaxation step strain at the end of loading step 100% ð2Þ
2.5. Multiple stress creep recovery test The MSCR test was also performed on DSR AR1500ex to obtain percent recovery of high-cured CRM binders at 64 °C under 0.1 kPa and 3.2 kPa creep stress according to ASTM D7405 [26]. The MSCR test was operated in rotational mode using 1 s creep load followed by 9 s recovery for each cycle. Ten creep and recovery cycles were run at 0.1 kPa creep stress followed by ten at 3.2 kPa creep stress. Fig. 2 presents the typical creep and recovery cycle. For each cycle, the percent recovery (R) is calculated using the following equation:
R¼
ep eu 100% ep
ð3Þ
where ep represents the peak strain, eu represents the unrecovered strain, r is the stress level. Then, the average percent recovery of 10 cycles at 0.1 kPa and 3.2 kPa are calculated and expressed as R0.1, R3.2 respectively. Besides, stress sensitivity parameter, Rdiff is calculated using the following equation:
Rdiff ¼ ðR0:1 R3:2Þ=ðR0:1Þ 100
ð4Þ
2.4. Dynamic oscillatory test Dynamic shear oscillatory test was performed on TA Dynamic Shear Rheometer (DSR) AR1500ex to obtain phase angle (d) of high-cured CRM binders. The test was performed at 64 °C and 10 rad/s using 25 mm parallel plate geometry according to ASTM D7175 [21]. Two replicates were tested for each binder and the average value was reported. All tested samples were Rolling Thin Film Oven (RTFO) aged according to ASTM D2872 [22]. Besides, frequency sweeps between 0.1 and 30 Hz at temperatures between 5 and 75 °C. The 8 mm diameter, 2 mm gap, parallel plate testing geometry was used for the tests between 5 and 25 °C and the 25 mm diameter, 1 mm gap was used from 35 to 75 °C. The strain amplitude for the frequency sweep tests was within the linear viscoelastic (LVE) response of the binder. The sigmoidal model developed in National Cooperative Highway Research Program (NCHRP) Project A-37A was taken to construct complex modulus master curve [23]. The shift factors for the construction of complex
3. Results and discussion 3.1. ATR-FTIR test results and analysis Fig. 3 presents the ATR spectrum of base binder, crumb rubber modified asphalt (CRMA, 190 °C, 0.5 h) and high-cured CRM asphalt. CRMA was produced by adding 18% crumb rubber to base binder, blending for 0.5 h at 190 °C, served as a control binder. In the interest of brevity, high-cured CRM binders interacted at four different temperatures for 4 h are presented here. As presented in Fig. 3, compared to base binder, CRMA showed increase at 1020 cm1, 873 cm1 and 670 cm1, which are attributed to CAOAS asymmetrical stretching in organosulfate (R1OASO2AOR2), CAH out of plane bending in 1,2,4-substituted aromatics and SAC respectively. However, when it comes to high-cured CRM binders, peak at 670 cm1 disappeared and peak at 1020 cm1 and
N. Tang et al. / Construction and Building Materials 123 (2016) 847–854
Base binder °C °C °C °C °C 670
965
Absorbance
873
1020
850
Table 3 Carbonyl index calculation results. Carbonyl index Blending temperature 220 °C 240 °C 260 °C 280 °C
Blending time 2h
4h
6h
8h
0.00115 0.00095 0.00014 0.00004
0.00062 0.00033 0.00023 0.00037
0.00026 0.00029 0.00040 0.00061
0.00057 0.00054 0.00049 0.00063
Base binder RTFO
Base binder PAV
0.00044
0.00456
binder underwent RTFO short term aging. The aging caused by the high curing conditions was far from the PAV aging effect. 1100
1000
900
800
700
3.2. ER-DSR Test Results and analysis
Wavenumber (cm -1)
873 cm1 decreased sharply. Therefore, one should note that not only SAS and SAC but also organosulfate and 1,2,4-substituted aromatics were broken during the high-cure process. Meanwhile, it should be noted that there was new peak formed at 965 cm1 which is attributed to CAH out of plane bending of trans-alkene. The new peak was due to the break of SAC bond and reconstruction of C@C thereby [27]. Besides, with the increase of interaction temperature, the newly formed peak was gradually intensified, which was quantified in this study (Fig. 4). At presented in Fig. 4, only small amount of C@C were gradually reconstructed with the increase of interaction time at 220 °C and 240 °C. This can be explained as only small part of S-C were broken at 220 °C and 240 °C because the heat energy was not high enough to break strong SAC bonds although the interaction time was extended to 8 h. When the interaction temperature was further elevated to 260 °C, SAC were gradually broken, especially for the first 4 h. When it comes to 280 °C, SAC was severely broken. To investigate the aging effect during the preparation of highcured CRM binders, the carbonyl index was calculated and the results are presented in Table 3. The base binder under RTFO short term aging and PAV long term aging were used as control binders to see whether high curing methods had caused serious oxidative aging. It can be seen that there is no clear relationship between carbonyl index and interaction temperature and time. In most cases, the aging status of CRM binders is similar to that of base
Fig. 5 illustrates the elastic recovery curve of high-cured CRM asphalt prepared at different temperatures for 2 h. It is clear that recovery behavior is weakening with the increase of interaction temperature. In Fig. 6, when the interaction temperature was elevated to 260 °C, the elastic recovery decreased drastically for the first 2 h. Since the elastic loss was due to the degradation of crumb rubber, it can be deduced that 260 °C was the critical degradation 3
°C
°C
°C
°C
1
0 0
10
20
30
40
50
60
Time (min) Fig. 5. Elastic recovery curve of high-cured CRM asphalt.
100
°C
°C
°C
°C
°C
90
0.35
80
Elastic Recovery (%)
0.30
Active Polymer Index
°C
2
0.40 °C
°C
Strain
Fig. 3. ATR spectrums of base binder, CRMA and high-cured CRM asphalt. ⁄ Absorbance of each sample is vertical shifted to make a comparison.
0.25 0.20 0.15 0.10
70 60 50 40 30 20
0.05 0.00
10
2
4
6
8
Interaction Time (h) Fig. 4. Plot of Active Polymer Index (API) of high-cured CRM asphalt.
0
2
4
6
8
Interaction Time (h) Fig. 6. Elastic recovery results for high-cured CRM asphalt.
851
N. Tang et al. / Construction and Building Materials 123 (2016) 847–854
asphalt cement with crumb rubber [11,28]. In Fig. 7, phase angle of high-cured CRM asphalt binders tested at 64 °C are presented. It is obvious that higher temperature and longer interaction time resulted in higher phase angle. It can be easily explained by the gradual loss of cross-linked network of crumb rubber in asphalt matrix. Compared to elastic recovery, phase angle was much more sensitive to the degradation of molecular structure within binders cured at 260 °C. Similar to elastic recovery, phase angle of binders cured at 280 °C varied little and the values are much closer to that of base binder (87.7), indicating that crumb rubber has reached the limit of degradation, which is consistent with FT-IR analysis. It is interesting that phase angle of binder cured at 220 °C, 8 h is quite close to that of binder cured at 240 °C, 2 h. Meanwhile, phase angle of binder cured at 240 °C, 8 h is close to that of binder cured at 260 °C, 2 h. Similar phenomenon can be observed between binder cured at 260 °C, 8 h and binder cured at 280 °C, 2 h. However, for the binders cured at 280 °C for the first 4 h, the phase angle behavior and API behavior is not consistent. It is assumed that 280 °C severely destroyed S-C as well as some carbon chain in rubber structure. The increase of API at 280 °C from 2 h to 4 h might be due to the broken of carbon chain of rubber structure (Polybutadiene phase) and the C@C came back from the polymerization reaction in the production of rubber. However, the phase angle has reached considerable high value (84.87) at 280 °C, 2 h, further broken of carbon chain in rubber structure will cause minor effect to elastic behavior. To make a more detailed comparison between CRM binders cured at different temperatures, the rheological master curves for base binder and CRM binders with different curing temperatures at 2 h are plotted in Fig. 8. The difference between binders cured at 220 °C, 2 h and 240 °C, 2 h showed little difference with respect to the complex modulus and phase angle over a wide frequency domain. In addition, the two binders cured at 220 °C, 2 h and 240 °C, 2 h were much stiffer than base binder and phase angle values were much lower than that of base binder at lower frequency domain, which means the rubber particles were contributing to the elastic behavior of modified binders at lower frequency domain (correspond to high temperature). However, when the interaction temperature were elevated to 260 and 280 °C, both the complex modulus and phase angle values had changed greatly, which is consistent with the FTIR, ER analysis results. Again, due to the serious degradation of crumb rubber in the asphalt matrix, the phase
temperature in this study, which was also confirmed by FTIR analysis. Besides, one should note that interaction time showed differential effect on elastic recovery properties of binders prepared at different temperatures. For 220 °C, elastic recovery dropt slowly with the increase of interaction time, while for 240 °C, elastic recovery decreased much more prominently. When it comes to 260 °C, there is no significant change of elastic recovery with extended interaction time. Similar phenomenon can be observed for 280 °C. This might be caused by the change of molecular chain at different interaction temperature. However, in this study, ER results of binders cured at 260 °C were less sensitive to the molecular change, which might be caused by the ER test temperature. When it comes to 280 °C, SAC were severely broken. In combination of FT-IR results, it can be concluded that 4 h was sufficient enough to break the cross-link bonds and additional interaction time will not cause significant change of elastic recovery. 3.3. Oscillatory test results and analysis Phase angle indicates the elastic component of asphalt binder. Lower phase angle means that binder is much more elastic. Phase angle has been successfully used to explain the interaction of
°C
90
°C
°C
°C
°C
80
70
60
50
2
4
6
8
Interaction Time (h) Fig. 7. Plot of phase angle against interaction time.
90
9
80
8
70
7
60
6
50
5
40
4
30
3
20
2
10
1
Log Complex Modulus (Pa)
Base Binder
Phase Angle (°)
40
220°C 2h 240°C 2h 260°C 2h 280°C 2h Base Binder |G*| 220°C 2h |G*| 240°C 2h |G*|
0
-5
-4
-3
-2
-1
0
1
2
3
4
Log Reduced Frequency (Hz) Fig. 8. Plot of rheological master curves at 25 °C.
5
0
260°C 2h |G*| 280°C 2h |G*|
852
N. Tang et al. / Construction and Building Materials 123 (2016) 847–854
°C
100
°C
°C
3.4. MSCR test results and analysis
°C
°C
90 80 70 60 50 40 30 20 10 2
4
6
8
Interaction Time (h) Fig. 9. Plot of MSCR percent recovery against interaction time, 0.1 kPa, 64 °C.
100
°C
°C
°C
°C
°C
80
60
40
20
0 2
4
6
8
Interaction Time (h) Fig. 10. Plot of MSCR percent recovery against interaction time, 3.2 kPa, 64 °C.
120 °C
°C
°C
°C
The percent recovery from MSCR test has been confirmed to distinguish the cross-linked polymer modified binders from noncross-linked ones [29,30]. Therefore, the percent recovery can be used to characterize the degradation process of crumb rubber in asphalt matrix in this study. Fig. 9 presents the percent recovery results tested at 0.1 kPa, 64 °C. For 220 °C interaction temperature, R0.1 decreased slowly with the increase of interaction time, which is similar to ER results. For 240 °C interaction temperature, R0.1 maintained considerable value for the first 6 h, while decreased below 80% at 8 h. When it comes to 260 °C, there was considerable change of R0.1, which demonstrates that R0.1 was much more sensitive to the molecular change when the interaction temperature was elevated to 260 °C. When the interaction temperature was further elevated to 280 °C, significant difference can be observed between 260 °C and 280 °C. However, there was no significant difference between interaction times. Again, crumb rubber has reached the limit of degradation at 280 °C. Besides, the equivalent relation between interaction temperature and time was also observed by comparing R0.1 from MSCR test. Fig. 10 presents the percent recovery results tested at 3.2 kPa, 64 °C. R3.2 showed totally different behavior compared to R0.1. R3.2 was obviously lower than R0.1 for the same binder. This can be explained by the stress sensitivity of modified binder. When the stress was elevated to 3.2 kPa, the degraded crumb rubber in asphalt matrix carried more of the load. Therefore, it was easier for the rubber molecular chain to start slipping. As the vulcanized structure of crumb rubber were gradually destroyed with the increase of interaction temperature and time, the binder behaved more in non-linear manner at 3.2 kPa. Fig. 11 clearly shows the difference between R0.1 and R3.2. When the interaction temperature is not less than 240 °C, the percent recovery difference are above 75%, which indicates that 3.2 kPa creep stress is much stricter than 0.1 kPa in evaluating the creep recovery ability of high-cured CRM asphalt. If one compares the FT-IR analysis results and MSCR percent recovery results, it will be found that R0.1 behaves more like Active Polymer Index than R3.2 does in this limited study. Therefore, 0.1 kPa is recommended for use in MSCR test to evaluate the degradation process of high-cured CRM asphalt. 3.5. Statistical analysis
100
Rdiff (%)
80
60
40
20
2
4
6
8
Interaction Time (h) Fig. 11. Plot of Rdiff against interaction time, 64 °C.
angle plateau region jumped to a high level and the modulus dropt drastically. It was noticed that the complex modulus of binders cured at 260 and 280 °C, 2 h were even lower than base binder at lower frequency domain although the phase angle values were still lower than that of base binder.
3.5.1. Effect of interaction temperature and time From above test results and analysis, it was found that both interaction temperature and time affected the degradation process. However, it seemed that interaction temperature played a more important role than interaction time based on the existed results. To confirm this, regression analysis between the test results and interaction temperature and time were performed using statistical software SPSS. The regression analysis results are presented in Table 4. The linear model was used in this study and R2 of regression functions were above 0.8. Especially, ER-DSR and phase angle showed strong linear correlation with interaction temperature and time since the regression R2 were 0.986 and 0.978 respectively. In Table 4, unstandardized coefficients (in the regression functions) and standardized coefficients for interaction temperature and time are presented. The latter ones ignore the independent variable’s scale of units, which makes comparisons easy. In this study, even though interaction temperature had a small unstandardized coefficient compared to interaction time, interaction temperature contributed more to the test results (API, ER-DSR, Phase Angle, R0.1 and R3.2) because it had a larger standardized coefficient. When it comes to the significance of independent variables,
853
N. Tang et al. / Construction and Building Materials 123 (2016) 847–854 Table 4 Regression analysis results. Dependent Variable
Regression function
R2
API ER-DSR Phase Angle R0.1 R3.2
z = 0.012x + 0.003y 0.629 z = 1.552x 1.027y + 307.701 z = 1.349x + 0.473y 52.434 z = 3.862x 1.424y + 435.40 z = 2.098x 0.759y + 218.14
0.849 0.986 0.978 0.882 0.850
Standardized coefficients
p-Value (t-test method)
x
y
x
y
constants
0.346 0.148 0.271 0.246 0.246
0.868 0.982 0.951 0.906 0.889
0.004 0.001 <0.001 0.023 0.039
<0.001 <0.001 <0.001 <0.001 <0.001
<0.001 <0.001 <0.001 <0.001 <0.001
x represents interaction time. y represents interaction temperature. z represents dependent variable.
Table 5 Pearson Correlation Analysis Results. API
Phase Angle
ER-DSR
R0.1
R3.2
API
Pearson correlation Sig. (2-tailed) N
1 – 16
0.933** <0.001 16
0.895** <0.001 16
0.970** <0.001 16
0.804** <0.001 16
Phase angle
Pearson correlation Sig. (2-tailed) N
0.933** <0.001 16
1 – 16
0.976** <0.001 16
0.945** <0.001 16
0.935** <0.001 16
ER-DSR
Pearson correlation Sig. (2-tailed) N
0.895** <0.001 16
0.976** <0.001 16
1 – 16
0.917** <0.001 16
0.900** <0.001 16
R0.1
Pearson correlation Sig. (2-tailed) N
0.970** <0.001 16
0.945** <0.001 16
0.917** <0.001 16
1 – 16
0.794** <0.001 16
R3.2
Pearson correlation Sig. (2-tailed) N
0.804** <0.001 16
0.935** <0.001 16
0.900** <0.001 16
0.794** <0.001 16
1 – 16
– Data is unavailable. Correlation is significant at the 0.01 level (2-tailed).
**
p-values derived from t-test method are also presented in Table 4. Commonly, if p-value is less than 0.05, the independent variable is considered to be significant to the regression model. Obviously, both interaction time and temperature contribute significantly to API, ER-DSR, Phase Angle, R0.1 and R3.2. However, interaction temperature contributes more than interaction time as interaction temperature has larger standardized coefficients than interaction time.
3.5.2. Correlation between API, ER-DSR, phase angle, R0.1 and R3.2 As API is derived from chemical analysis, it is necessary to validate its reliability in detecting the degradation process of highcured CRM asphalt. Pearson correlation analysis was performed between API, ER-DSR, phase angle, R0.1 and R3.2. The analysis results are presented in Table 5. On the whole, these parameters correlate well to each other. API derived from ATR-FTIR spectrum is strongly correlated to phase angle, ER-DSR and R0.1, which indicates that API is capable of detecting the degradation process of high-cured CRM asphalt. Of all these parameters, phase angle shows the best correlation to other parameters. As discussed in MSCR test results and analysis, R3.2 is not recommended for use in detecting the degradation of crumb rubber in asphalt matrix. Correlation analysis confirmed this as R3.2 didn’t show advantage in correlating to other parameters. In summary, API is a good indicator of degradation of crumb rubber in asphalt matrix. Meanwhile, DSR based Elastic Recovery, phase angle from oscillatory test and percent recovery at 0.1 kPa from MSCR test are good indicators of degradation of high-cured CRM asphalt.
4. Conclusions and recommendations The degradation process of high-cured CRM asphalt was investigated through ATR-FTIR quantitative analysis and various rheological test methods. The following conclusions can be drawn from the test results and data analysis: (1) The critical degradation temperature is 260 °C in this study as ATR-FTIR test and various rheological tests indicate. When it comes to 280 °C, degradation limit is achieved. (2) In this limited study, dynamic oscillatory test and MSCR test indicate that effect of 20 °C increment of interaction temperature, curing for 2 h was similar to that of 6 h increment of interaction time. Regression analysis indicates that interaction temperature contributes more to the degradation than interaction time. (3) Active Polymer Index (API) derived from ATR-FTIR test is a good indicator of degradation of crumb rubber. API is strongly correlated to phase angle, elastic recovery and percent recovery at 0.1 kPa. (4) DSR based elastic recovery, phase angle from oscillatory test and percent recovery at 0.1 kPa from MSCR test are good indicators of degradation of high-cured CRM asphalt. Percent recovery at 3.2 kPa is not recommended for use in detecting the degradation of crumb rubber in asphalt matrix. (5) Based on the rheological and chemical test results, recommendations are as follows. The production temperature of TB rubberized should not be above 260 °C to avoid significant elastic loss of CRM binders. FTIR
854
N. Tang et al. / Construction and Building Materials 123 (2016) 847–854
test is recommended as a fast and effect method to detect the degradation status in the production of TB rubberized asphalt. Acknowledgement The authors gratefully acknowledge the financial supports by the National Natural Science Foundation of China under grant number 51478351. References [1] F.J. Navarro, P. Partal, F. Martínez-Boza, C. Gallegos, Thermo-rheological behaviour and storage stability of ground tire rubber-modified bitumens, Fuel 83 (2004) 2041–2049. [2] S. Biro, A. Geiger, J. Kay, Asphalt rubber with temporarily decreased viscosity, in: Asphalt Rubber 2009 Conference, Rubber Pavements Association, Nanjing, 2009, pp. 461–474. [3] B. Amadou, H. Shahrzad, F. ElhamH, Investigating effect of amine based additives on asphalt rubber rheological properties, Raleigh, in: Y.R. Kim (Ed.), International Society for Asphalt Pavements, CRC Press, North Carolina, USA, 2014, pp. 921–931. [4] K.M. Shatanawi, S. Biro, A. Geiger, S.N. Amirkhanian, Effects of furfural activated crumb rubber on the properties of rubberized asphalt, Constr. Build. Mater. 28 (2012) 96–103. [5] G. Cheng, B. Shen, J. Zhang, A study on the performance and storage stability of crumb rubber-modified asphalts, Petrol Sci. Technol. 29 (2011) 192–200. [6] W. Daly, I. Negulescu, Characterization of asphalt cements modified with crumb rubber from discarded tires, Transp. Res. Rec. (1997) 37–44. [7] D.L. Presti, G. Airey, Tyre rubber-modified bitumens development: the effect of varying processing conditions, Road Mater. Pav. Des. 14 (2013) 888–900. [8] R.G. Hicks, D. Cheng, T. Duffy, Evaluation of Terminal Blend Rubberized Asphalt in Paving Applications, California Pavement Preservation Center, Chico, California, 2010. [9] X. Qi, A. Shenoy, G. Al-Khateeb, T. Arnold, N. Gibson, J. Youtcheff, T. Harman, Laboratory characterization and fullscale accelerated performance testing of crumb rubber asphalts and other modified asphalt systems, Asphalt Rubber Conference, Palm Springs, USA, 2006. [10] A.R.G. The University Of Wisconsin-Madison, The Effects of digesting Crumb Rubber in Modified Binders, The University of Wisconsin-Madison Asphalt Research Group in collaboration with the Recycled Materials Resource Center, 2011. [11] M. Abdelrahman, S. Carpenter, Mechanism of interaction of asphalt cement with crumb rubber modifier, Transp. Res. Rec. (1999) 106–113.
[12] T. Billiter, R. Davison, C. Glover, J. Bullin, Production of asphalt-rubber binders by high-cure conditions, Transp. Res. Rec. (1997) 50–56. [13] L. Zanzotto, G. Kennepohl, Development of rubber and asphalt binders by depolymerization and devulcanization of scrap tires in asphalt, Transp. Res. Rec. (1996) 51–58. [14] A. Ghavibazoo, M. Abdelrahman, M. Ragab, Mechanism of crumb rubber modifier dissolution into asphalt matrix and its effect on final physical properties of crumb rubber-modified binder, Transp. Res. Rec. (2013) 92–101. [15] M. Ragab, M. Abdelrahman, A. Ghavibazoo, Performance enhancement of crumb rubber-modified asphalts through control of the developed internal network structure, Transp. Res. Rec. (2013) 96–104. [16] J. Masson, L. Pelletier, P. Collins, Rapid FTIR method for quantification of styrene-butadiene type copolymers in bitumen, J. Appl. Polym. Sci. 79 (2001) 1034–1041. [17] T.P. Flanigan, One-stage abrasive absorption process for producing tire rubber modified asphalt cement systems and products thereof, Google Patents 1996. [18] D. Sun, L. Zhang, Quantitative determination of SBS content in SBS modified asphalt, J. Build. Mater. 16 (2013) 36. [19] I. Yut, A. Zofka, Attenuated total reflection (ATR) fourier transform infrared (FT-IR) spectroscopy of oxidized polymer-modified bitumens, Appl. Spectrosc. 65 (2011) 765–770. [20] C.S. Clopotel, H.U. Bahia, Importance of elastic recovery in the DSR for binders and mastics, Eng. J. 16 (2012) 99–106. [21] ASTM, Standard Test Method for Determining the Rheological Properties of Asphalt Binder Using a Dynamic Shear Rheometer D7175-08, 2008. West Conshohocken, PA. [22] ASTM, Standard Test Method for Effect of Heat and Air on a Moving Film of Asphalt (Rolling Thin-Film Oven Test), D2872–122012. [23] N.I.M. Yusoff, M.T. Shaw, G.D. Airey, Modelling the linear viscoelastic rheological properties of bituminous binders, Constr. Build. Mater. 25 (2011) 2171–2189. [24] S.M. Asgharzadeh, N. Tabatabaee, K. Naderi, M. Partl, An empirical model for modified bituminous binder master curves, Mater. Struct. 46 (2013) 1459– 1471. [25] G.D. Airey, Rheological properties of styrene butadiene styrene polymer modified road bitumens, Fuel 82 (2003) 1709–1719. [26] ASTM, Standard Test Method for Multiple Stress Creep and Recovery (MSCR) of Asphalt Binder Using a Dynamic Shear Rheometer D7405-10a, 2010. West Conshohocken, PA. [27] B. Adhikari, D. De, S. Maiti, Reclamation and recycling of waste rubber, Prog. Polym. Sci. 25 (2000) 909–948. [28] M. Abdelrahman, Controlling performance of crumb rubber-modified binders through addition of polymer modifiers, Transp. Res. Rec. (2006) 64–70. [29] J. D’Angelo, R. Dongré, Practical use of multiple stress creep and recovery test, Transp. Res. Rec. 2126 (2009) 73–82. [30] W. Huang, N. Tang, Characterizing SBS modified asphalt with sulfur using multiple stress creep recovery test, Constr. Build. Mater. 93 (2015) 514–521.
Contents lists available at ScienceDirect
Construction and Building Materials journal homepage: www.elsevier.com/locate/conbuildmat
Chemical and rheological investigation of high-cured crumb rubber-modified asphalt Naipeng Tang, Weidong Huang ⇑, Feipeng Xiao The Key Laboratory of Road and Traffic Engineering, Ministry of Education, Tongji University, Shanghai 201804, PR China
h i g h l i g h t s Active Polymer Index was defined to characterize the curing process of CRM Asphalt. Interaction temperature contributed more to the degradation than interaction time. There was linear relationship between chemical and rheological properties. API derived from ATR-FTIR test is a good indicator of degradation of crumb rubber.
a r t i c l e
i n f o
Article history: Received 28 April 2016 Received in revised form 18 June 2016 Accepted 26 July 2016
Keywords: High-cured CRM asphalt Degradation ATR-FTIR Active Polymer Index Elastic recovery
a b s t r a c t In the production of Terminal Blend (TB) rubberized asphalt, degradation of crumb rubber in asphalt matrix can help improve the storage stability and workability of crumb rubber modified asphalt. However, degradation process of crumb rubber modified (CRM) binders was not clear from molecular components to rheological properties. This study investigated chemical and rheological perspectives of the degradation process of CRM asphalt under high interaction temperature (220–280 °C) and extended curing time (2–8 h), namely high-cured CRM asphalt. Attenuated Total Reflection (ATR) Fourier Transform Infrared (FT-IR) Spectroscopy was used to characterize the degradation behaviors of CRM binders at varying curing temperatures and times. Active Polymer Index (API) was defined as the ratio between band areas at 978–918 cm1 and 850–785 cm1 of FTIR tests to characterize the released active polymer from vulcanized crumb rubber. Besides, dynamic shear rheometer (DSR) was used to perform elastic recovery (ER) test, dynamic oscillatory test and multiple stress creep recovery (MSCR) test. Test results indicated that the critical degradation temperature was 260 °C in this limited study. Interaction temperature contributed more to the degradation than interaction time. Pearson correlation analysis indicated that API derived from ATR-FTIR test was a good indicator of degradation of CRM binders. Ó 2016 Elsevier Ltd. All rights reserved.
1. Introduction Crumb rubber has been used in pavement industry over 50 years. Traditional wet-process technique, such as Asphalt Rubber, has shown disadvantages in storage stability and workability [1]. Attempts have been made to improve the storage stability or decrease the viscosity of crumb rubber modified (CRM) asphalt through various physical and chemical methods. Additives like vegetable oil, warm mix asphalt additives (Rediset and Evotherm) and bio-binder were found to effectively decrease the viscosity of CRM asphalt [2,3]. Activated or treated crumb rubber by using chemicals like furfural (C5H4O2) and polymeric compatibilizer were proved to improve the storage stability of CRMA significantly ⇑ Corresponding author. E-mail address: [email protected] (W. Huang). http://dx.doi.org/10.1016/j.conbuildmat.2016.07.131 0950-0618/Ó 2016 Elsevier Ltd. All rights reserved.
[4,5]. Besides, storage stability of CRMA can also be improved by using high shear method or using finer powdered rubber (200 mesh) [6,7]. Most of the above studies were concerned with the partial degradation of crumb rubber in asphalt matrix. In recent years, Terminal Blend (TB) rubberized asphalt technology arose in America has shown its potential in greatly improving storage stability of rubber/asphalt composite because crumb rubber is fully digested into the asphalt, while maintaining considerable performance in pavement practice. Full scale projects and heavy vehicle simulator testing in California indicated that fatigue performance of TB rubberized asphalt is better than that of asphalt rubber and base binder [8]. The FHWA’s Accelerated Loading Facility (ALF) testing results indicated that Texas TB rubberized asphalt offer better rutting resistance than Arizona wet process asphalt rubber when the latter provide better fatigue performance [9]. It was reported that the production of TB rubberized asphalt can be
848
N. Tang et al. / Construction and Building Materials 123 (2016) 847–854
achieved through degradation of crumb rubber in asphalt matrix using heat and shear, meanwhile polymers might be added to the degraded rubber/asphalt composite to meet the performance grade criteria [10]. It has been confirmed by various research that degradation of CRM binders can affect rheological properties significantly [11–13]. Meanwhile, molecular weight distribution and molecular structure present remarkable change from micro perspective [12,14]. It would be better to understand the degradation mechanism through qualitative or quantitative analysis of molecular weight distribution and molecular structure in combination with rheological investigations. The current research on molecular weight distribution of CRM binders can’t fully characterize the rubber/asphalt composite as some dissoluble rubber components were removed by syringe filter before Gel Permeation Chromatography (GPC) test [15]. As a rapid test method, Fourier Transform Infrared Spectroscopy (FT-IR) gains popularity and quantitative analysis using FT-IR has been successfully applied to polymer modified asphalt [16]. However, FT-IR was mostly used to qualitatively detect the released rubber components into asphalt [14,15]. The study of degradation process of crumb rubber in asphalt matrix is limited. The objective of this study is to investigate degradation process of high-cured CRM binders in terms of their chemical and rheological properties. Attenuated Total Reflection (ATR) Fourier Transform Infrared (FT-IR) Spectroscopy is used to explore the main element components of CRM binders at varying reacting temperatures and times. Active Polymer Index (API) is employed to characterize the released active polymer content from the vulcanized crumb rubber. Dynamic Shear Rheometer (DSR) is used to perform elastic recovery (ER) test, dynamic oscillatory test and multiple stress creep recovery (MSCR) test in this study.
2. Materials and methods 2.1. Materials One PG 64-22 base binder provided by ESSO Asphalt Company and one minus 30 mesh crumb rubber provided by China Rubber Resource Regeneration were used to prepare high-cured CRM binders. A 20% crumb rubber (by the weight of total binder) was mixed with base binder at four temperatures of 220 °C, 240 °C, 260 °C and 280 °C for 2, 4, 6 and 8 h. The blending speed was 400 rpm. The preparation of samples was reported by Abdelrahman [11], Billiter [12], Zanzotto [13] and Flanigan [17]. Besides, crumb rubber modified asphalt produced by adding 18% crumb rubber to base binder, blending for 0.5 h at 190 °C was served as a control binder in the aftermentioned FT-IR analysis. The detailed test samples are presented in Table 1. For example, for binder blended at 220 for 2 h, the produced binder is marked as E22.2 in Table 1. As storage stability is very important for engineering use, the separation test were performed for each sample through keeping an aluminium tube containing 50 ± 0.5 g binder at 163 °C
Table 1 Description of the test samples. 80% base binder + 20% crumb rubber Blending temperature
220 °C 240 °C 260 °C 280 °C
Blending time 2h
4h
6h
8h
E22.2 E24.2 E26.2 E28.2
E22.4 E24.4 E26.4 E28.4
E22.6 E24.6 E26.6 E28.6
E22.8 E24.8 E26.8 E28.8
Table 2 Storage stability test results. Blending temperature
Softening point
Blending time
°C
2h
4h
6h
8h
220 °C
Up Bottom Difference
67.3 68 0.7
58.3 61.9 3.6
69.6 68.2 1.4
58.9 55.7 3.2
240 °C
Up Bottom Difference
53.4 57.6 4.2
47.4 50.3 2.9
54.7 54.2 0.5
51.3 49.2 2.1
260 °C
Up Bottom Difference
49.1 52.3 3.2
48.2 49.9 1.7
44.8 46.8 2
47.3 48.2 0.9
280 °C
Up Bottom Difference
49.1 50.8 1.7
48.5 49 0.5
46.1 45.5 0.6
47.5 46.8 0.7
for 48 h according to ASTM D7173. The top and bottom parts of the tube were collected for softening point test (ASTM D36) and the softening point difference between the top and bottom is used as an index to evaluate the storage stability. In China, if the difference is less than 2.5 °C, the binder is considered to have good storage stability. The storage stability test results are presented in Table 2. On the whole, increasing interaction temperature and extending curing time will help to improve the storage stability. 2.2. Attenuated Total Reflection (ATR) Fourier Transform Infrared (FTIR) Spectroscopy The infrared spectra values were collected using a Bruker TENSOR FT-IR spectrometer equipped with a reflection diamond ATR accessory. About 1 g asphalt binder was put on the surface of ATR diamond and fixed using the steel loader. The asphalt binder completely contacted the diamond homogenously. Thirty-two scans were averaged within the wavenumber range of 4000– 600 cm1 for each sample and then these average spectrum values were obtained. In this study, the peak at 965 cm1 attributed to CAH out of plane bending of trans-alkene showed a significant change during the interaction process of high-cured CRM asphalt. As FTIR peak intensities are sensitive to the concentration of components and the thickness of the sample [14], it would be better to normalize the band areas between 978 and 918 cm1 to eliminate the effect of sample thickness on FTIR results. In this study, the peak at 810 cm1 attributed to C-H out of plane bending of aromatics in base binder was found to vary little in terms of band areas and it has been successfully used as a reference peak in the quantitative analysis of SBS modified asphalt [18]. In this study, band areas between 978 and 918 cm1 (AR965) were divided by that of 850–785 cm1 (AR810) and the ratio is defined as Active Polymer Index (Eq. (1)).
Active Polymer Index ðAPIÞ ¼ AR965 =AR810
ð1Þ
The AR965 and AR810 were calculated in Thermo Scientific OMNICTM software automatically. The integration scopes were from 978 to 918 cm1 for 965 cm1 and 850 to 785 cm1 for 810 cm1 (Fig. 1). Band areas values rather than peak absorbance were used in this study because band areas varied little within three replicates of each sample [19]. The carbonyl index was also calculated to investigate the aging status of high cured CRM binders during the preparation. The calculation method was similar to that of API and the integration scopes for carbonyl were from 1678 to 1725 cm1. The base binder
849
N. Tang et al. / Construction and Building Materials 123 (2016) 847–854 0.30
peak strain
0.28 810
3
2 unrecovered strain
Strain
0.26
0.24
1
965
Absorbance
recovered strain
4
0.22
0 0.20 1000
950
900
850
800
750
0
700
2
4
6
8
10
Time (s)
Wavenumber (cm -1) Fig. 1. Illustration of band areas between 978–918 cm1 and 850–785 cm1.
Fig. 2. Typical creep and recovery cycle in MSCR test.
under Rolling Thin Film Oven (RTFO) short term aging and Pressure Aging Vessel (PAV) long term aging were used as control binders.
modulus curve were then used to build phase angle master curve. The newly developed double-logistic (DL) mathematical function by Asgharzadeh and Tabatabaee [24] was taken for the modelling of phase angle master curve. The reference temperature of 25 °C was chosen in the production of complex modulus and phase angle master curve [25].
2.3. Elastic Recovery test based on dynamic shear rheometer Elastic recovery (ER) was used to investigate the elastic properties of high-cured CRM asphalt. The test procedure was originally developed by Clopotel [20] and shown as follows: (1) Rolling Thin Film Oven (RTFO) aged sample is tested on the 8 mm plate in TA AR1500ex DSR. (2) The sample is stored at 25 °C for 20 min before test. (3) A constant strain rate of 0.023 s1 is applied for 2 min in a peak hold mode of flow procedure provided by TA’s instrument control software. (4) A constant zero shear stress is applied for a period of 1 h in a peak hold mode of flow procedure provided by TA’s instrument control software. (5) Calculate the elastic recovery at the end of relaxation using Eq. (2).
ER-DSR ¼
recovered strain at the end of the relaxation step strain at the end of loading step 100% ð2Þ
2.5. Multiple stress creep recovery test The MSCR test was also performed on DSR AR1500ex to obtain percent recovery of high-cured CRM binders at 64 °C under 0.1 kPa and 3.2 kPa creep stress according to ASTM D7405 [26]. The MSCR test was operated in rotational mode using 1 s creep load followed by 9 s recovery for each cycle. Ten creep and recovery cycles were run at 0.1 kPa creep stress followed by ten at 3.2 kPa creep stress. Fig. 2 presents the typical creep and recovery cycle. For each cycle, the percent recovery (R) is calculated using the following equation:
R¼
ep eu 100% ep
ð3Þ
where ep represents the peak strain, eu represents the unrecovered strain, r is the stress level. Then, the average percent recovery of 10 cycles at 0.1 kPa and 3.2 kPa are calculated and expressed as R0.1, R3.2 respectively. Besides, stress sensitivity parameter, Rdiff is calculated using the following equation:
Rdiff ¼ ðR0:1 R3:2Þ=ðR0:1Þ 100
ð4Þ
2.4. Dynamic oscillatory test Dynamic shear oscillatory test was performed on TA Dynamic Shear Rheometer (DSR) AR1500ex to obtain phase angle (d) of high-cured CRM binders. The test was performed at 64 °C and 10 rad/s using 25 mm parallel plate geometry according to ASTM D7175 [21]. Two replicates were tested for each binder and the average value was reported. All tested samples were Rolling Thin Film Oven (RTFO) aged according to ASTM D2872 [22]. Besides, frequency sweeps between 0.1 and 30 Hz at temperatures between 5 and 75 °C. The 8 mm diameter, 2 mm gap, parallel plate testing geometry was used for the tests between 5 and 25 °C and the 25 mm diameter, 1 mm gap was used from 35 to 75 °C. The strain amplitude for the frequency sweep tests was within the linear viscoelastic (LVE) response of the binder. The sigmoidal model developed in National Cooperative Highway Research Program (NCHRP) Project A-37A was taken to construct complex modulus master curve [23]. The shift factors for the construction of complex
3. Results and discussion 3.1. ATR-FTIR test results and analysis Fig. 3 presents the ATR spectrum of base binder, crumb rubber modified asphalt (CRMA, 190 °C, 0.5 h) and high-cured CRM asphalt. CRMA was produced by adding 18% crumb rubber to base binder, blending for 0.5 h at 190 °C, served as a control binder. In the interest of brevity, high-cured CRM binders interacted at four different temperatures for 4 h are presented here. As presented in Fig. 3, compared to base binder, CRMA showed increase at 1020 cm1, 873 cm1 and 670 cm1, which are attributed to CAOAS asymmetrical stretching in organosulfate (R1OASO2AOR2), CAH out of plane bending in 1,2,4-substituted aromatics and SAC respectively. However, when it comes to high-cured CRM binders, peak at 670 cm1 disappeared and peak at 1020 cm1 and
N. Tang et al. / Construction and Building Materials 123 (2016) 847–854
Base binder °C °C °C °C °C 670
965
Absorbance
873
1020
850
Table 3 Carbonyl index calculation results. Carbonyl index Blending temperature 220 °C 240 °C 260 °C 280 °C
Blending time 2h
4h
6h
8h
0.00115 0.00095 0.00014 0.00004
0.00062 0.00033 0.00023 0.00037
0.00026 0.00029 0.00040 0.00061
0.00057 0.00054 0.00049 0.00063
Base binder RTFO
Base binder PAV
0.00044
0.00456
binder underwent RTFO short term aging. The aging caused by the high curing conditions was far from the PAV aging effect. 1100
1000
900
800
700
3.2. ER-DSR Test Results and analysis
Wavenumber (cm -1)
873 cm1 decreased sharply. Therefore, one should note that not only SAS and SAC but also organosulfate and 1,2,4-substituted aromatics were broken during the high-cure process. Meanwhile, it should be noted that there was new peak formed at 965 cm1 which is attributed to CAH out of plane bending of trans-alkene. The new peak was due to the break of SAC bond and reconstruction of C@C thereby [27]. Besides, with the increase of interaction temperature, the newly formed peak was gradually intensified, which was quantified in this study (Fig. 4). At presented in Fig. 4, only small amount of C@C were gradually reconstructed with the increase of interaction time at 220 °C and 240 °C. This can be explained as only small part of S-C were broken at 220 °C and 240 °C because the heat energy was not high enough to break strong SAC bonds although the interaction time was extended to 8 h. When the interaction temperature was further elevated to 260 °C, SAC were gradually broken, especially for the first 4 h. When it comes to 280 °C, SAC was severely broken. To investigate the aging effect during the preparation of highcured CRM binders, the carbonyl index was calculated and the results are presented in Table 3. The base binder under RTFO short term aging and PAV long term aging were used as control binders to see whether high curing methods had caused serious oxidative aging. It can be seen that there is no clear relationship between carbonyl index and interaction temperature and time. In most cases, the aging status of CRM binders is similar to that of base
Fig. 5 illustrates the elastic recovery curve of high-cured CRM asphalt prepared at different temperatures for 2 h. It is clear that recovery behavior is weakening with the increase of interaction temperature. In Fig. 6, when the interaction temperature was elevated to 260 °C, the elastic recovery decreased drastically for the first 2 h. Since the elastic loss was due to the degradation of crumb rubber, it can be deduced that 260 °C was the critical degradation 3
°C
°C
°C
°C
1
0 0
10
20
30
40
50
60
Time (min) Fig. 5. Elastic recovery curve of high-cured CRM asphalt.
100
°C
°C
°C
°C
°C
90
0.35
80
Elastic Recovery (%)
0.30
Active Polymer Index
°C
2
0.40 °C
°C
Strain
Fig. 3. ATR spectrums of base binder, CRMA and high-cured CRM asphalt. ⁄ Absorbance of each sample is vertical shifted to make a comparison.
0.25 0.20 0.15 0.10
70 60 50 40 30 20
0.05 0.00
10
2
4
6
8
Interaction Time (h) Fig. 4. Plot of Active Polymer Index (API) of high-cured CRM asphalt.
0
2
4
6
8
Interaction Time (h) Fig. 6. Elastic recovery results for high-cured CRM asphalt.
851
N. Tang et al. / Construction and Building Materials 123 (2016) 847–854
asphalt cement with crumb rubber [11,28]. In Fig. 7, phase angle of high-cured CRM asphalt binders tested at 64 °C are presented. It is obvious that higher temperature and longer interaction time resulted in higher phase angle. It can be easily explained by the gradual loss of cross-linked network of crumb rubber in asphalt matrix. Compared to elastic recovery, phase angle was much more sensitive to the degradation of molecular structure within binders cured at 260 °C. Similar to elastic recovery, phase angle of binders cured at 280 °C varied little and the values are much closer to that of base binder (87.7), indicating that crumb rubber has reached the limit of degradation, which is consistent with FT-IR analysis. It is interesting that phase angle of binder cured at 220 °C, 8 h is quite close to that of binder cured at 240 °C, 2 h. Meanwhile, phase angle of binder cured at 240 °C, 8 h is close to that of binder cured at 260 °C, 2 h. Similar phenomenon can be observed between binder cured at 260 °C, 8 h and binder cured at 280 °C, 2 h. However, for the binders cured at 280 °C for the first 4 h, the phase angle behavior and API behavior is not consistent. It is assumed that 280 °C severely destroyed S-C as well as some carbon chain in rubber structure. The increase of API at 280 °C from 2 h to 4 h might be due to the broken of carbon chain of rubber structure (Polybutadiene phase) and the C@C came back from the polymerization reaction in the production of rubber. However, the phase angle has reached considerable high value (84.87) at 280 °C, 2 h, further broken of carbon chain in rubber structure will cause minor effect to elastic behavior. To make a more detailed comparison between CRM binders cured at different temperatures, the rheological master curves for base binder and CRM binders with different curing temperatures at 2 h are plotted in Fig. 8. The difference between binders cured at 220 °C, 2 h and 240 °C, 2 h showed little difference with respect to the complex modulus and phase angle over a wide frequency domain. In addition, the two binders cured at 220 °C, 2 h and 240 °C, 2 h were much stiffer than base binder and phase angle values were much lower than that of base binder at lower frequency domain, which means the rubber particles were contributing to the elastic behavior of modified binders at lower frequency domain (correspond to high temperature). However, when the interaction temperature were elevated to 260 and 280 °C, both the complex modulus and phase angle values had changed greatly, which is consistent with the FTIR, ER analysis results. Again, due to the serious degradation of crumb rubber in the asphalt matrix, the phase
temperature in this study, which was also confirmed by FTIR analysis. Besides, one should note that interaction time showed differential effect on elastic recovery properties of binders prepared at different temperatures. For 220 °C, elastic recovery dropt slowly with the increase of interaction time, while for 240 °C, elastic recovery decreased much more prominently. When it comes to 260 °C, there is no significant change of elastic recovery with extended interaction time. Similar phenomenon can be observed for 280 °C. This might be caused by the change of molecular chain at different interaction temperature. However, in this study, ER results of binders cured at 260 °C were less sensitive to the molecular change, which might be caused by the ER test temperature. When it comes to 280 °C, SAC were severely broken. In combination of FT-IR results, it can be concluded that 4 h was sufficient enough to break the cross-link bonds and additional interaction time will not cause significant change of elastic recovery. 3.3. Oscillatory test results and analysis Phase angle indicates the elastic component of asphalt binder. Lower phase angle means that binder is much more elastic. Phase angle has been successfully used to explain the interaction of
°C
90
°C
°C
°C
°C
80
70
60
50
2
4
6
8
Interaction Time (h) Fig. 7. Plot of phase angle against interaction time.
90
9
80
8
70
7
60
6
50
5
40
4
30
3
20
2
10
1
Log Complex Modulus (Pa)
Base Binder
Phase Angle (°)
40
220°C 2h 240°C 2h 260°C 2h 280°C 2h Base Binder |G*| 220°C 2h |G*| 240°C 2h |G*|
0
-5
-4
-3
-2
-1
0
1
2
3
4
Log Reduced Frequency (Hz) Fig. 8. Plot of rheological master curves at 25 °C.
5
0
260°C 2h |G*| 280°C 2h |G*|
852
N. Tang et al. / Construction and Building Materials 123 (2016) 847–854
°C
100
°C
°C
3.4. MSCR test results and analysis
°C
°C
90 80 70 60 50 40 30 20 10 2
4
6
8
Interaction Time (h) Fig. 9. Plot of MSCR percent recovery against interaction time, 0.1 kPa, 64 °C.
100
°C
°C
°C
°C
°C
80
60
40
20
0 2
4
6
8
Interaction Time (h) Fig. 10. Plot of MSCR percent recovery against interaction time, 3.2 kPa, 64 °C.
120 °C
°C
°C
°C
The percent recovery from MSCR test has been confirmed to distinguish the cross-linked polymer modified binders from noncross-linked ones [29,30]. Therefore, the percent recovery can be used to characterize the degradation process of crumb rubber in asphalt matrix in this study. Fig. 9 presents the percent recovery results tested at 0.1 kPa, 64 °C. For 220 °C interaction temperature, R0.1 decreased slowly with the increase of interaction time, which is similar to ER results. For 240 °C interaction temperature, R0.1 maintained considerable value for the first 6 h, while decreased below 80% at 8 h. When it comes to 260 °C, there was considerable change of R0.1, which demonstrates that R0.1 was much more sensitive to the molecular change when the interaction temperature was elevated to 260 °C. When the interaction temperature was further elevated to 280 °C, significant difference can be observed between 260 °C and 280 °C. However, there was no significant difference between interaction times. Again, crumb rubber has reached the limit of degradation at 280 °C. Besides, the equivalent relation between interaction temperature and time was also observed by comparing R0.1 from MSCR test. Fig. 10 presents the percent recovery results tested at 3.2 kPa, 64 °C. R3.2 showed totally different behavior compared to R0.1. R3.2 was obviously lower than R0.1 for the same binder. This can be explained by the stress sensitivity of modified binder. When the stress was elevated to 3.2 kPa, the degraded crumb rubber in asphalt matrix carried more of the load. Therefore, it was easier for the rubber molecular chain to start slipping. As the vulcanized structure of crumb rubber were gradually destroyed with the increase of interaction temperature and time, the binder behaved more in non-linear manner at 3.2 kPa. Fig. 11 clearly shows the difference between R0.1 and R3.2. When the interaction temperature is not less than 240 °C, the percent recovery difference are above 75%, which indicates that 3.2 kPa creep stress is much stricter than 0.1 kPa in evaluating the creep recovery ability of high-cured CRM asphalt. If one compares the FT-IR analysis results and MSCR percent recovery results, it will be found that R0.1 behaves more like Active Polymer Index than R3.2 does in this limited study. Therefore, 0.1 kPa is recommended for use in MSCR test to evaluate the degradation process of high-cured CRM asphalt. 3.5. Statistical analysis
100
Rdiff (%)
80
60
40
20
2
4
6
8
Interaction Time (h) Fig. 11. Plot of Rdiff against interaction time, 64 °C.
angle plateau region jumped to a high level and the modulus dropt drastically. It was noticed that the complex modulus of binders cured at 260 and 280 °C, 2 h were even lower than base binder at lower frequency domain although the phase angle values were still lower than that of base binder.
3.5.1. Effect of interaction temperature and time From above test results and analysis, it was found that both interaction temperature and time affected the degradation process. However, it seemed that interaction temperature played a more important role than interaction time based on the existed results. To confirm this, regression analysis between the test results and interaction temperature and time were performed using statistical software SPSS. The regression analysis results are presented in Table 4. The linear model was used in this study and R2 of regression functions were above 0.8. Especially, ER-DSR and phase angle showed strong linear correlation with interaction temperature and time since the regression R2 were 0.986 and 0.978 respectively. In Table 4, unstandardized coefficients (in the regression functions) and standardized coefficients for interaction temperature and time are presented. The latter ones ignore the independent variable’s scale of units, which makes comparisons easy. In this study, even though interaction temperature had a small unstandardized coefficient compared to interaction time, interaction temperature contributed more to the test results (API, ER-DSR, Phase Angle, R0.1 and R3.2) because it had a larger standardized coefficient. When it comes to the significance of independent variables,
853
N. Tang et al. / Construction and Building Materials 123 (2016) 847–854 Table 4 Regression analysis results. Dependent Variable
Regression function
R2
API ER-DSR Phase Angle R0.1 R3.2
z = 0.012x + 0.003y 0.629 z = 1.552x 1.027y + 307.701 z = 1.349x + 0.473y 52.434 z = 3.862x 1.424y + 435.40 z = 2.098x 0.759y + 218.14
0.849 0.986 0.978 0.882 0.850
Standardized coefficients
p-Value (t-test method)
x
y
x
y
constants
0.346 0.148 0.271 0.246 0.246
0.868 0.982 0.951 0.906 0.889
0.004 0.001 <0.001 0.023 0.039
<0.001 <0.001 <0.001 <0.001 <0.001
<0.001 <0.001 <0.001 <0.001 <0.001
x represents interaction time. y represents interaction temperature. z represents dependent variable.
Table 5 Pearson Correlation Analysis Results. API
Phase Angle
ER-DSR
R0.1
R3.2
API
Pearson correlation Sig. (2-tailed) N
1 – 16
0.933** <0.001 16
0.895** <0.001 16
0.970** <0.001 16
0.804** <0.001 16
Phase angle
Pearson correlation Sig. (2-tailed) N
0.933** <0.001 16
1 – 16
0.976** <0.001 16
0.945** <0.001 16
0.935** <0.001 16
ER-DSR
Pearson correlation Sig. (2-tailed) N
0.895** <0.001 16
0.976** <0.001 16
1 – 16
0.917** <0.001 16
0.900** <0.001 16
R0.1
Pearson correlation Sig. (2-tailed) N
0.970** <0.001 16
0.945** <0.001 16
0.917** <0.001 16
1 – 16
0.794** <0.001 16
R3.2
Pearson correlation Sig. (2-tailed) N
0.804** <0.001 16
0.935** <0.001 16
0.900** <0.001 16
0.794** <0.001 16
1 – 16
– Data is unavailable. Correlation is significant at the 0.01 level (2-tailed).
**
p-values derived from t-test method are also presented in Table 4. Commonly, if p-value is less than 0.05, the independent variable is considered to be significant to the regression model. Obviously, both interaction time and temperature contribute significantly to API, ER-DSR, Phase Angle, R0.1 and R3.2. However, interaction temperature contributes more than interaction time as interaction temperature has larger standardized coefficients than interaction time.
3.5.2. Correlation between API, ER-DSR, phase angle, R0.1 and R3.2 As API is derived from chemical analysis, it is necessary to validate its reliability in detecting the degradation process of highcured CRM asphalt. Pearson correlation analysis was performed between API, ER-DSR, phase angle, R0.1 and R3.2. The analysis results are presented in Table 5. On the whole, these parameters correlate well to each other. API derived from ATR-FTIR spectrum is strongly correlated to phase angle, ER-DSR and R0.1, which indicates that API is capable of detecting the degradation process of high-cured CRM asphalt. Of all these parameters, phase angle shows the best correlation to other parameters. As discussed in MSCR test results and analysis, R3.2 is not recommended for use in detecting the degradation of crumb rubber in asphalt matrix. Correlation analysis confirmed this as R3.2 didn’t show advantage in correlating to other parameters. In summary, API is a good indicator of degradation of crumb rubber in asphalt matrix. Meanwhile, DSR based Elastic Recovery, phase angle from oscillatory test and percent recovery at 0.1 kPa from MSCR test are good indicators of degradation of high-cured CRM asphalt.
4. Conclusions and recommendations The degradation process of high-cured CRM asphalt was investigated through ATR-FTIR quantitative analysis and various rheological test methods. The following conclusions can be drawn from the test results and data analysis: (1) The critical degradation temperature is 260 °C in this study as ATR-FTIR test and various rheological tests indicate. When it comes to 280 °C, degradation limit is achieved. (2) In this limited study, dynamic oscillatory test and MSCR test indicate that effect of 20 °C increment of interaction temperature, curing for 2 h was similar to that of 6 h increment of interaction time. Regression analysis indicates that interaction temperature contributes more to the degradation than interaction time. (3) Active Polymer Index (API) derived from ATR-FTIR test is a good indicator of degradation of crumb rubber. API is strongly correlated to phase angle, elastic recovery and percent recovery at 0.1 kPa. (4) DSR based elastic recovery, phase angle from oscillatory test and percent recovery at 0.1 kPa from MSCR test are good indicators of degradation of high-cured CRM asphalt. Percent recovery at 3.2 kPa is not recommended for use in detecting the degradation of crumb rubber in asphalt matrix. (5) Based on the rheological and chemical test results, recommendations are as follows. The production temperature of TB rubberized should not be above 260 °C to avoid significant elastic loss of CRM binders. FTIR
854
N. Tang et al. / Construction and Building Materials 123 (2016) 847–854
test is recommended as a fast and effect method to detect the degradation status in the production of TB rubberized asphalt. Acknowledgement The authors gratefully acknowledge the financial supports by the National Natural Science Foundation of China under grant number 51478351. References [1] F.J. Navarro, P. Partal, F. Martínez-Boza, C. Gallegos, Thermo-rheological behaviour and storage stability of ground tire rubber-modified bitumens, Fuel 83 (2004) 2041–2049. [2] S. Biro, A. Geiger, J. Kay, Asphalt rubber with temporarily decreased viscosity, in: Asphalt Rubber 2009 Conference, Rubber Pavements Association, Nanjing, 2009, pp. 461–474. [3] B. Amadou, H. Shahrzad, F. ElhamH, Investigating effect of amine based additives on asphalt rubber rheological properties, Raleigh, in: Y.R. Kim (Ed.), International Society for Asphalt Pavements, CRC Press, North Carolina, USA, 2014, pp. 921–931. [4] K.M. Shatanawi, S. Biro, A. Geiger, S.N. Amirkhanian, Effects of furfural activated crumb rubber on the properties of rubberized asphalt, Constr. Build. Mater. 28 (2012) 96–103. [5] G. Cheng, B. Shen, J. Zhang, A study on the performance and storage stability of crumb rubber-modified asphalts, Petrol Sci. Technol. 29 (2011) 192–200. [6] W. Daly, I. Negulescu, Characterization of asphalt cements modified with crumb rubber from discarded tires, Transp. Res. Rec. (1997) 37–44. [7] D.L. Presti, G. Airey, Tyre rubber-modified bitumens development: the effect of varying processing conditions, Road Mater. Pav. Des. 14 (2013) 888–900. [8] R.G. Hicks, D. Cheng, T. Duffy, Evaluation of Terminal Blend Rubberized Asphalt in Paving Applications, California Pavement Preservation Center, Chico, California, 2010. [9] X. Qi, A. Shenoy, G. Al-Khateeb, T. Arnold, N. Gibson, J. Youtcheff, T. Harman, Laboratory characterization and fullscale accelerated performance testing of crumb rubber asphalts and other modified asphalt systems, Asphalt Rubber Conference, Palm Springs, USA, 2006. [10] A.R.G. The University Of Wisconsin-Madison, The Effects of digesting Crumb Rubber in Modified Binders, The University of Wisconsin-Madison Asphalt Research Group in collaboration with the Recycled Materials Resource Center, 2011. [11] M. Abdelrahman, S. Carpenter, Mechanism of interaction of asphalt cement with crumb rubber modifier, Transp. Res. Rec. (1999) 106–113.
[12] T. Billiter, R. Davison, C. Glover, J. Bullin, Production of asphalt-rubber binders by high-cure conditions, Transp. Res. Rec. (1997) 50–56. [13] L. Zanzotto, G. Kennepohl, Development of rubber and asphalt binders by depolymerization and devulcanization of scrap tires in asphalt, Transp. Res. Rec. (1996) 51–58. [14] A. Ghavibazoo, M. Abdelrahman, M. Ragab, Mechanism of crumb rubber modifier dissolution into asphalt matrix and its effect on final physical properties of crumb rubber-modified binder, Transp. Res. Rec. (2013) 92–101. [15] M. Ragab, M. Abdelrahman, A. Ghavibazoo, Performance enhancement of crumb rubber-modified asphalts through control of the developed internal network structure, Transp. Res. Rec. (2013) 96–104. [16] J. Masson, L. Pelletier, P. Collins, Rapid FTIR method for quantification of styrene-butadiene type copolymers in bitumen, J. Appl. Polym. Sci. 79 (2001) 1034–1041. [17] T.P. Flanigan, One-stage abrasive absorption process for producing tire rubber modified asphalt cement systems and products thereof, Google Patents 1996. [18] D. Sun, L. Zhang, Quantitative determination of SBS content in SBS modified asphalt, J. Build. Mater. 16 (2013) 36. [19] I. Yut, A. Zofka, Attenuated total reflection (ATR) fourier transform infrared (FT-IR) spectroscopy of oxidized polymer-modified bitumens, Appl. Spectrosc. 65 (2011) 765–770. [20] C.S. Clopotel, H.U. Bahia, Importance of elastic recovery in the DSR for binders and mastics, Eng. J. 16 (2012) 99–106. [21] ASTM, Standard Test Method for Determining the Rheological Properties of Asphalt Binder Using a Dynamic Shear Rheometer D7175-08, 2008. West Conshohocken, PA. [22] ASTM, Standard Test Method for Effect of Heat and Air on a Moving Film of Asphalt (Rolling Thin-Film Oven Test), D2872–122012. [23] N.I.M. Yusoff, M.T. Shaw, G.D. Airey, Modelling the linear viscoelastic rheological properties of bituminous binders, Constr. Build. Mater. 25 (2011) 2171–2189. [24] S.M. Asgharzadeh, N. Tabatabaee, K. Naderi, M. Partl, An empirical model for modified bituminous binder master curves, Mater. Struct. 46 (2013) 1459– 1471. [25] G.D. Airey, Rheological properties of styrene butadiene styrene polymer modified road bitumens, Fuel 82 (2003) 1709–1719. [26] ASTM, Standard Test Method for Multiple Stress Creep and Recovery (MSCR) of Asphalt Binder Using a Dynamic Shear Rheometer D7405-10a, 2010. West Conshohocken, PA. [27] B. Adhikari, D. De, S. Maiti, Reclamation and recycling of waste rubber, Prog. Polym. Sci. 25 (2000) 909–948. [28] M. Abdelrahman, Controlling performance of crumb rubber-modified binders through addition of polymer modifiers, Transp. Res. Rec. (2006) 64–70. [29] J. D’Angelo, R. Dongré, Practical use of multiple stress creep and recovery test, Transp. Res. Rec. 2126 (2009) 73–82. [30] W. Huang, N. Tang, Characterizing SBS modified asphalt with sulfur using multiple stress creep recovery test, Constr. Build. Mater. 93 (2015) 514–521.