National Academies Press: OpenBook

Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging (2021)

Chapter: Chapter 2 - Research Approach

« Previous: Chapter 1 - Background
Page 7
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 7
Page 8
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 8
Page 9
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 9
Page 10
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 10
Page 11
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 11
Page 12
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 12
Page 13
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 13
Page 14
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 14
Page 15
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 15
Page 16
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 16
Page 17
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 17
Page 18
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 18
Page 19
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 19
Page 20
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 20
Page 21
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 21
Page 22
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 22
Page 23
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 23
Page 24
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 24
Page 25
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 25
Page 26
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 26
Page 27
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 27
Page 28
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 28
Page 29
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 29
Page 30
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 30
Page 31
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 31
Page 32
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 32
Page 33
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 33
Page 34
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 34
Page 35
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 35
Page 36
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 36
Page 37
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 37
Page 38
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 38
Page 39
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 39
Page 40
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 40
Page 41
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 41
Page 42
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 42
Page 43
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 43
Page 44
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 44
Page 45
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 45
Page 46
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 46
Page 47
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 47
Page 48
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 48
Page 49
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 49
Page 50
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 50
Page 51
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 51
Page 52
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 52
Page 53
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 53
Page 54
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 54
Page 55
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 55
Page 56
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 56
Page 57
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 57
Page 58
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 58
Page 59
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 59
Page 60
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 60
Page 61
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 61
Page 62
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 62
Page 63
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 63
Page 64
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 64
Page 65
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 65
Page 66
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 66
Page 67
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 67
Page 68
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 68
Page 69
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 69
Page 70
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 70
Page 71
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 71
Page 72
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 72
Page 73
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 73
Page 74
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 74
Page 75
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 75
Page 76
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 76
Page 77
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 77
Page 78
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 78
Page 79
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 79
Page 80
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 80
Page 81
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 81
Page 82
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 82
Page 83
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 83
Page 84
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 84
Page 85
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 85
Page 86
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 86
Page 87
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 87
Page 88
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 88
Page 89
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 89
Page 90
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 90
Page 91
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 91
Page 92
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 92
Page 93
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 93
Page 94
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 94
Page 95
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 95
Page 96
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 96
Page 97
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 97
Page 98
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 98
Page 99
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 99
Page 100
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 100
Page 101
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 101
Page 102
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 102
Page 103
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 103
Page 104
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 104
Page 105
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 105
Page 106
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 106
Page 107
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 107
Page 108
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 108
Page 109
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 109
Page 110
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 110
Page 111
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 111
Page 112
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 112
Page 113
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 113
Page 114
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 114
Page 115
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 115
Page 116
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 116
Page 117
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 117
Page 118
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 118
Page 119
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 119
Page 120
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 120
Page 121
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 121
Page 122
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 122
Page 123
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 123
Page 124
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 124
Page 125
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 125
Page 126
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 126
Page 127
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 127
Page 128
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 128
Page 129
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 129
Page 130
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 130
Page 131
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 131
Page 132
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 132
Page 133
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 133
Page 134
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 134
Page 135
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 135
Page 136
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 136
Page 137
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 137
Page 138
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 138
Page 139
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 139
Page 140
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 140
Page 141
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 141
Page 142
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 142
Page 143
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 143
Page 144
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 144
Page 145
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 145
Page 146
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 146
Page 147
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 147
Page 148
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 148
Page 149
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 149
Page 150
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 150
Page 151
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 151
Page 152
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 152
Page 153
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 153
Page 154
Suggested Citation:"Chapter 2 - Research Approach." National Academies of Sciences, Engineering, and Medicine. 2021. Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging. Washington, DC: The National Academies Press. doi: 10.17226/26089.
×
Page 154

Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

7   Introduction The research was conducted in two sequential phases. The first phase included a critical evaluation of AASHTO T 240 and AASHTO R 28 and other laboratory conditioning procedures that have been proposed as alternates, identification and selection of improved laboratory conditioning procedures for further development in NCHRP Project 09-61, and a detailed design of experiments to further develop the selected laboratory conditioning procedures and relate them to the aging that occurs during construction and the service life of the pavement. In the second phase, the experiments were executed; the data were analyzed, and materials were prepared to assist with implementing the improved procedures. The primary products of NCHRP Project 09-61 were recommended revisions to AASHTO standard practices and methods of tests associated with laboratory conditioning of asphalt binders. These recom- mended revisions were supported by commentaries summarizing the findings from the research that support the recommended revision. Considerations for Laboratory Conditioning Quantity of Conditioned Binder for Grading An important factor that must be considered in evaluating binder conditioning methods is the amount of conditioned binder needed for performance grading. Table 1 summarizes the amount of conditioned binder needed for performance grade verification using AASHTO M 320 and AASHTO M 332 as well as the amount of conditioned binder needed for continuous grading using AASHTO M 320. Table 1 shows that low-temperature testing using the bending beam rheometer (BBR) and the direct tension (DT) test dictates the quantity of binder that must be conditioned through both the short- and long-term procedures. Each BBR beam requires approximately 16 g of binder or 32 g per test temperature when two beams are tested as required by AASHTO R 29, Standard Practice for Grading or Verifying the Performance Grade of an Asphalt Binder. Each DT specimen requires approximately 8 g of binder and AASHTO T 314, Standard Method of Test for Determining the Fracture Properties of Asphalt Binder in Direct Tension (DT), requires six specimens to be tested per temperature. Dynamic shear rheometer (DSR) testing requires only about 1 g of binder per test, and the same specimen can be tested at multiple temperatures to obtain pass/fail properties. From Table 1, the amount of conditioned binder needed for current specification testing ranges from approximately 34 g for AASHTO M 320 Table 1 and AASHTO M 332 verification without direct tension to approximately 162 g for continuous grading using AASHTO M 320 Table 2. Continuous low-temperature grading based on AASHTO R 49, Standard Practice for Determining the Low-Temperature Performance Grade (PG) of Asphalt Binder, requires C H A P T E R   2 Research Approach

8 Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging BBR and DT testing at two temperatures. DT testing is not commonly performed in practice, and the 2018 AASHTO re:source removed the test from the Performance Graded Binder proficiency sample testing. Therefore, the evaluation presented here is based on the testing required for AASHTO M 320 Table 1 and AASHTO M 332 without DT, which reduces the range of conditioned binder need for current grading to 34 g for verification and 66 g for continuous grading. The 4 mm parallel plate DSR testing has the potential to reduce the amount of binder needed for performance grading to only a few grams. Recent research performed at the Western Research Institute introduced the possibility of replacing BBR testing with low temperature, 4 mm parallel plate DSR testing and associated analyses (Farrar et al. 2015). Although 4 mm DSR testing has Test Approx. Mass per Specimen (with Waste) g M 320 Table 1 and M 332 Verification Without Direct Tension M 320 Table 1 Continuous Grading Without Direct Tension M 320 Table 1 and M 332 Verification with Direct Tension M 320 Table 2 Verification with Direct Tension M 320 Table 2 Continuous Grading with Direct Tension Number Mass,g Number Number Number Number RTFOT DSR 1 1 1 1 1 1 1 1 1 1 1 PAV DSR 1 1 1 1 1 1 1 1 1 1 1 PAV BBR 16 2 32 4 64 2 32 4 64 4 64 PAV Direct Tension 8 0 0 0 0 6 48 6 48 12 96 Total NA NA 34 NA 66 NA 82 NA 114 NA 162 Mass, g Mass, g Mass, g Mass, g Table 1. Summary of the quantity of binder needed for performance grading. Environmental Zone Subgrade Type State Years of Data Comment Dry, Freeze Coarse MT 20.1 Chip seal year 9 UT 18.0 Fog seal year 1 Fine WA 19.9 Chip seal year 9 Active SD 20.4 Swelling soil Dry, Non-Freeze Coarse CA 16.1 NM 17.5 Fine Active Wet, Freeze Coarse NY 13.8 WI 16.9 Fine MO 15.3 Active OH 14.6 Frost susceptible soil Wet, Non-Freeze Coarse MS 17.3 NC 17.3 NJ 6.4 Fine AR 17.0 Active TX 19.0 Swelling soil, chip seal year 15 In alphabetical order: AR = Arkansas, CA = California, MO = Missouri, MS = Mississippi, MT = Montana, NC = North Carolina, NJ = New Jersey, NM = New Mexico, NY = New York, OH = Ohio, SD = South Dakota, TX = Texas, UT = Utah, WA = Washington, WI = Wisconsin Table 2. Summary of SPS-8 sites.

Research Approach 9   proven acceptable for research, several issues must be addressed before it is considered viable for specification testing. These include: 1. Determining appropriate methods for temperature control during low-temperature DSR testing. Results to date from equipment using air or liquid nitrogen cooling differ signifi- cantly from equipment using Peltier systems (Reinke 2017). 2. Standardizing procedures for determining the compliance correction for the DSR. Proper compliance correction is critical for low-temperature DSR testing. 3. Standardizing procedures for sample loading, trimming, and testing. 4. Standardizing the data analysis. 5. Establishing appropriate test control to reduce testing error through ruggedness testing. 6. Conducting inter-laboratory testing to determine within and between laboratory testing errors. Without a funded, coordinated national effort, it will be several years before the issues above are resolved. Additionally, 4 mm DSR testing requires research-grade rheometers that are not available in most agency and producer laboratories. Therefore, the evaluation presented in this report assumes that low-temperature performance grading will be done with the BBR. In summary, the performance grading procedures commonly used in practice require approximately 34 g of conditioned binder for verification and 66 g of conditioned binder for continuous grading. The evaluation presented in this report is based on these quantities of conditioned binder. Target In-Service Pavement Age There is growing consensus that AASHTO R 28 is not severe enough, leading to the question: what pavement age should the improved long-term conditioning procedure target? Long-term conditioning should target the time in service where changes in binder properties lead to the onset of distress in the pavement. The Long-Term Pavement Performance (LTPP) Specific Pavement Studies Number 8 (SPS-8), Study of Environmental Effects in the Absence of Heavy Loads, provides data from a number of pavement sections that can be used to estimate when asphalt binders in typical pavements age to the point that non–load associated distresses begin to develop. The SPS-8 pavements were new pavements constructed on roadways with limited truck traffic. For flexible pavements, the SPS-8 experimental plan included two pavement structures: 4 in of asphalt concrete over 8 in of aggregate base and 7 in of asphalt concrete over 12 in of aggregate base. The experimental plan called for the two pavement structures to be constructed on three different subgrade types: fine grained, coarse grained, and active (active soil was defined as frost susceptible or swelling) in each of the four LTPP climatic zones. This resulted in an experimental design having 12 sites with 24 flexible pavement sections. The sections included in the SPS-8 experiment varied from this design as shown in Table 2. For each environmental zone, there were multiple sites constructed on coarse-grained soils, while no sites were con- structed on fine and active soils for the dry, non-freeze environmental zone. Additionally, monitoring of the New Jersey site lasted only 6.4 years, and four sections received maintenance treatments during the monitoring period. The LTPP distress monitoring included documenting the extent of rutting and various types of cracking, including fatigue, longitudinal, and transverse. The SPS-8 sections were surveyed every 1 to 2 years. Of the various measurements of cracking, transverse cracking is the best measurement of distress caused by environmental exposure. Table  3 summarizes the first

10 Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging appearance of sustained transverse cracking for both the thin and thick pavement sections at each site. Sustained transverse cracking means transverse cracking that is growing based on subsequent survey data. For some sections, small amounts of transverse cracking were reported earlier than shown in Table 3, but subsequent surveys registered no transverse cracking, suggesting the cracks healed or were the result of survey errors. Table  3 also includes averages for selected groupings. In these calculations, an age of 17.3 years was used for the North Carolina sections, which did not exhibit sustained transverse cracking. These averages show the following interesting trends: 1. Pavements constructed on swelling soils (South Dakota and Texas) exhibited transverse cracks extremely early. The age to sustained transverse cracking is similar for pavements constructed on coarse-grained soils and non-swelling fine grain soils. Since the early trans- verse cracking in the South Dakota and Texas sections appears to be caused by the effects of swelling soil, these sections were eliminated from subsequent analyses. 2. Pavements that received maintenance treatments before transverse cracks appeared (Montana, Utah, and Washington) exhibited transverse cracking later than pavements that did not receive maintenance treatments. Since the maintenance treatments likely alter the binder aging, the Montana, Utah, and Washington sections were also eliminated from subsequent analyses. 3. The age to first sustained transverse cracking is similar for the two thicknesses of asphalt concrete. This provides supporting evidence that the aged properties of the binder near the surface govern distress caused by environmental effects. Environmental Zone Subgrade Type State Asphalt Thickness 4 in 7 in Comment Dry, Freeze Coarse MT 13.0 13.0 Chip seal year 9 UT 16.7 16.7 Fog seal year 1 Fine WA 18.6 15.6 Chip seal year 9 Active SD 4.1 3.4 Dry, Non-Freeze Coarse CA 13.1 13.1 NM 9.4 15.8 Fine Active Wet, Freeze Coarse NY 7.8 8.9 WI 7.7 12.4 Fine MO 13.4 11.2 Active OH 11.8 6.4 Wet, Non-Freeze Coarse MS 14.1 10.8 NC * * None to 17.3 years NJ * * None to 6.4 years Fine AR 9.5 13.0 Active TX 5.4 5.4 Chip seal year 15 Average Coarse 12.9 14.1 17.3 years used for NC Fine 13.3 11.6 Swelling 4.8 4.4 Average After Removing Swelling Soil Dry, Freeze 16.1 15.1 Dry, Non-Freeze 11.3 14.5 Wet, Freeze 10.2 9.7 Wet, Non-Freeze 13.6 13.7 17.3 years used for NC No Maintenance Treatment 11.6 12.1 Maintenance Treatment 16.1 15.1 Table 3. Age to first sustained transverse cracking.

Research Approach 11   4. The age to first sustained transverse cracking appears to be shorter by approximately 3 years for pavements constructed in the wet, freeze environment. This may be caused by the freezing and thawing of the subgrade. Figure 1 shows the evolution of cracking calculated as the average for the 4 in and 7 in asphalt sections at each site, eliminating the sections on swelling soil and the sections that received maintenance treatments before transverse cracks appeared. Figure 1 and Table 3 suggest that the improved long-term aging procedure should target in-service aging of 10 to 12  years, which is longer than the 4 to 8  years that have been associated with the current PAV procedure (Anderson et al. 1994). The target pavement age from this analysis was used to evaluate long-term conditioning procedures and to calibrate the improved long-term conditioning procedure. Methods for Judging the Equivalency of Laboratory Conditioning and Field Aging The equivalency of various short- and long-term binder conditioning procedures to field- aged binders was evaluated using various rheological and chemical tests. Tests were selected based on their successful use in completed research studies to quantify short- and long-term aging of asphalt binders. The sections that follow document the methods that were used. Short-Term Conditioned Binder The primary tool used to assess the alternative laboratory short-term conditioning proce- dures was the AASHTO M 320 high pavement temperature specification parameter G*/sind. This parameter was measured on laboratory-conditioned binder and binder recovered from 0 5 10 15 20 25 30 35 40 45 0 2 4 6 8 10 12 14 16 18 20 Le ng th o f T ra ns ve rs e Cr ac ks , m Pavement Age, yrs AR CA MS NM NC MO NY OH WI Figure 1. Average transverse cracking in LTPP SPS-8 sections with sections on swelling soil and sections that received maintenance treatment removed.

12 Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging oven-conditioned loose mixtures using AASHTO T 315, Standard Method of Test for Deter- mining the Rheological Properties of Asphalt Binder Using a Dynamic Shear Rheometer (DSR), with a strain level of 10 percent as specified for short-term conditioned binder. The binder from short-term conditioned loose mixtures was recovered per ASTM D7906, Standard Practice for Recovery of Asphalt from Solution Using Toluene and the Rotary Evaporator. Long-Term Conditioned Binder Rheological Properties The rheological index and crossover frequency from the Christensen-Anderson master curve model were the rheological properties used to compare residue after long-term condi- tioning to field aging. The Christensen-Anderson master curve describes the shear modulus and phase angle of the binder over a wide range of temperatures and loading rates (Christensen and Anderson 1992). As discussed in greater detail below, these parameters vary in a consistent manner with increased aging for a specific binder. Another consideration in the selection of these rheological parameters was that they are obtained from DSR frequency sweep measure- ments, which require only a small amount of binder. The experiments conducted in NCHRP Project 09-61 make extensive use of binder recovered from thin slices of pavement cores. Rheological Index and Crossover Frequency. The Christensen-Anderson master curve model was formulated in a manner that the model parameters have specific physical meaning (Christensen and Anderson 1992). Equation 1 presents the Christensen-Andersen model for the frequency dependency of the binder complex shear modulus, and Equation 2 presents the Christensen-Anderson model for the frequency dependency of the phase angle. * 1 (1) log 2 log 2 G Gg c r R R ( )ω = + ω ω           − r c R ( )δ ω = + ω ω           90 1 (2)log 2 where G*(w) = complex shear modulus [pascal (Pa)], d(w) = phase angle, degrees, Gg = glassy modulus assumed equal to 1 × 109 (Pa), wr = reduced frequency at the reference temperature [radians per second (rad/sec)], wc = crossover frequency at the reference temperature (rad/sec), and R = rheological index. The Christensen-Anderson model uses two equations to represent the temperature depen- dency of an asphalt binder. Above the defining temperature (Td), the Williams-Landel-Ferry equation, Equation 3, is used. Below the defining temperature, the Arrhenius function, Equa- tion 4, is used. a T T T T T d d ( ) ( ) = − − + − log 19 92 (3) a T T Td ( ) = −    log 13016.07 1 1 (4)

Research Approach 13   where a(T) = shift factor, T = temperature (°K), and Td = defining temperature (°K). Figure  2 is a schematic of a fitted master curve and the nomenclature used with the Christensen-Anderson model. The glassy modulus, Gg, is the limiting modulus reached at high frequencies and low temperatures. For asphalt binders, a value of 1 GPa is usually assumed. The viscous asymptote is the line that the master curve approaches at low frequencies and is an indicator of the steady-state viscosity of the binder. The crossover frequency, wc, is the frequency where the phase angle is 45 degrees and is typically close to the point where the viscous asymp- tote intersects the glassy modulus. The crossover frequency is an indicator of the hardness of the binder: the lower the crossover frequency, the harder the asphalt binder. The rheological index, R, is the difference between the log of the glassy modulus and the log of the complex modulus at the crossover frequency. It is an indicator of the rheological type. As the value of the rheological index increases, the master curve becomes flatter, indicating a more gradual transition from elastic response (glassy modulus) to steady-state flow (viscous asymptote). Finally, the defining temperature, Td, is the temperature where there is a transition in the temperature dependency of the asphalt binder’s response. The defining temperature is related to, but not the same as, the glass transition temperature. As the defining temperature decreases, the temperature depen- dency above the defining temperature will decrease. Non-linear least-squares optimization was used to fit the three parameters of the Christensen- Anderson master curve model: wc, R, and Td. This was done using the Solver function in Excel. Equations 1, 3, and 4 were combined to provide the predicted value of G* as a function of the temperature and frequency used in the frequency sweep tests. The objective function for Figure 2. Typical binder master curve with the Christensen-Anderson model parameters.

14 Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging Figure 3. Effect of field aging and laboratory conditioning on binder master curves. the optimization was the sum of the squares of the difference between this predicted value of G* and the measured value of G*. The Solver function in Excel was used to minimize the objective function. The Christensen-Anderson master curve model is most accurate for conditions near the crossover frequency. Therefore, the frequency sweep testing was performed at three tempera- tures spaced 6°C apart that bracketed a phase angle of 45° at the frequency of 10 rad/sec. The DSR testing followed the procedure in AASHTO T 315 for a long-term conditioned binder using a strain level of 0.1 percent. The frequency range for the testing was 0.1 to 100 rad/sec. Effect of Aging on Christensen-Anderson Master Curve Parameters. During the SHRP, Anderson et al. demonstrated the utility of using master curves to track binder aging (Anderson et al. 1994). Figure 3 shows how the master curve of a binder changes with field aging and laboratory conditioning. With increased aging or conditioning, the master curves shift to the left and become flatter. This results in a decrease in the crossover frequency and an increase in the rheological index. For the example shown in Figure 3, PAV conditioning produces a master curve that is similar to the binder recovered from the field section. More recently, the Western Research Institute developed Christensen-Anderson master curves for: (1) binders recovered from field-aged pavements, (2) the original binder used in the pavements, and (3) the original binder after RTFOT and PAV conditioning. Recovered binders were obtained from various depths in the field cores ranging from 6.5 mm to 61 mm (Boysen and Schabron 2015). Figure 4 shows the changes in the rheological index and crossover fre- quency with increased aging and conditioning for a section from the pavement testing facility. This figure includes data for: 1. Original binder, 2. RTFOT-conditioned binder,

Research Approach 15   3. PAV-conditioned binder, 4. Recovered binder from different depths for 96 months of field aging in McLean, Virginia, and 5. Recovered binder data for the 6.5 mm depth after 96 months of field aging plus 8 months of accelerated aging. The accelerated aging was accomplished by using the heaters on the Accelerated Loading Facility (ALF) to continuously heat the pavement to a surface temperature of 70°C. Figure 4 shows a unique relationship between the rheological index and crossover frequency (binder hardness) with increasing exposure. When comparing laboratory conditioning to field aging, the laboratory conditioning should follow the same path for changes in the rheological index and crossover frequency as the field aging and move far enough along the path for the targeted time in service and depth. The laboratory conditioning in Figure 4 is along the same path as the field aging; however, the PAV is far short of reproducing the aging that occurs near the pavement surface after 96 months of exposure to the climate in McLean, Virginia. It is close to the field aging that occurred in this pavement after 96 months at a depth of about 50 mm. Note that the relationship between the rheological index and crossover frequency shown in Figure 4 is binder specific. Different binders will have different values for the rheological index and crossover frequency and their evolution; time in service will also be different. Chemical Properties Changes in the chemistry of the asphalt binder resulting from laboratory conditioning and field aging were measured using Fourier Transform Infrared Spectroscopy (FTIR). FTIR spectra have been used to determine the amount of oxygen in asphalt. They are a reliable surrogate test for determining the oxygen content in laboratory- and field-aged samples when using the absorbance intensity of carbonyl (C=O) and sulfoxide (S=O). As discussed in greater detail below, the intensity of carbonyl and sulfoxide varies in a consistent manner with increased aging for a specific binder. FTIR Analysis. Liquid cell FTIR spectra were obtained from carbon disulfide (CS2) and carbon tetrachloride (CCl4) solutions at a concentration of 50 mg binder per mL of solution. The two different solvents have different interfering regions in the spectra, so by using both solvents more of the asphalt spectra can be observed. The solutions were loaded into a 1 mm, path-length sodium chloride (NaCl) cell (International Crystal Laboratories). Spectra were Figure 4. Comparison of the rheological index and crossover frequency for laboratory-conditioned and field-aged PG 70-22 binder from the pavement testing facility (data from Boysen and Schabron 2015).

16 Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging recorded using a Perkin Elmer Frontier FTIR. Mid-infrared spectra were collected from 4,000 cm–1 to 600 cm–1 for 32 scans at a resolution of 4 cm–1. All samples were background- subtracted using the same cell loaded with the pure solvent. The background spectra were acquired before running the samples. FTIR spectra are very sensitive and well suited to track oxidative aging by monitoring the increase in the carbonyl (C=O) band near 1,700 cm–1 and the sulfoxide (S=O) band near 1,034 cm–1 as shown in Figure 5. This figure includes overlays of FTIR spectra for the binder used in section AZ1-1 of the Asphalt Research Consortium (ARC) validation site in Arizona that was laboratory conditioned in the PAV using different operating parameters. More severe conditions in the PAV produced increased C=O and S=O peaks. Research at the Western Research Institute has shown that the increase in the sum of the carbonyl and sulfoxide absor- bance (C=O + S=O) correlates with the total amount of elemental oxygen in the asphalt during oxidation. Figure 6 shows example relationships for several binders from the FHWA Pavement Testing Facility (Boysen and Schabron 2015). Effect of Aging on FTIR Spectra. Figure 7 demonstrates how carbonyl plus sulfoxide absorbance can be used to compare laboratory conditioning to field aging for a specific binder. The data in Figure 7 and Figure 4 are from the same pavement at the pavement testing facility. Figure 7 shows a unique relationship between carbonyl plus sulfoxide absorbance and cross- over frequency (binder hardness) with increasing exposure. When comparing laboratory conditioning to field aging, it is important that laboratory conditioning produce similar changes in chemistry as field aging. Like the rheological data in Figure 4, the carbonyl plus sulfoxide absorbance in Figure 7 shows the laboratory conditioning follows the same path as the field aging; however, PAV conditioning does not condition the binder enough to represent the aging that occurs near the pavement surface after 96 months of exposure. Binder Recovery This project made extensive use of binder recovered from 0.5 in thick slices of pavement cores. The extraction and recovery procedure that was used was developed at the Western Research Institute to minimize changes in the binder caused by the extraction and recovery process (Boysen and Schabron 2015). A core slice is broken into smaller pieces. Extractions were done at room temperature by mixing the pieces with a mixture of reagent-grade toluene and reagent-grade 95 percent ethanol in volumetric proportions of 85 percent toluene to 15 percent 65085010501250145016501850 Wavelength (cm-1) Original RTFOT 20 Hr PAV 40 Hr PAV C S 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 Ab so rb an ce (A U ) Figure 5. Example FTIR spectra showing carbonyl (C=O) and sulfoxide (S=O) peaks.

Research Approach 17   0.0000 0.2000 0.4000 0.6000 0.8000 1.0000 1.2000 1.4000 1.6000 0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 C= O + S =O A bs or ba nc e Oxygen, wt % PG 70-22 Air Blown SBS Modified Crumb Rubber Terpolymer Fibers Figure 6. Correlation between FTIR (C=O + S=O) absorbance values and weight percent oxygen (data from Boysen and Schabron 2015). 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 0.0001 0.001 0.01 0.1 1 10 100 1000 10000 [C =O + S= O ] A bs or ba nc e Crossover Frequency, rad/sec Lab Aged 96 Month Ambient 96 Month Ambient + 8 Month Accelerated O rig in al RT FO TP AV 6. 5 m m 19 m m 48 .5 m m 61 .5 m m 6. 5 m m Increasing Exposure Figure 7. Comparison of carbonyl plus sulfoxide absorbance and crossover frequency for laboratory-conditioned and field-aged PG 70-22 binder from the pavement testing facility (data from Boysen and Schabron 2015). ethanol. Extractions were repeated until the extract solution was near clear. The extract portions from repeat extractions were each centrifuged at 1,170 RCF (2,200 rpm with 21.6 cm radius) for 30 minutes, and the supernatant liquids were then combined to provide a sample extract solution. The solution was rotary evaporated at 70°C to provide about 150 mL of concentrated extract solution. Final solvent removal consisted of 3 hours in the hood under an argon tent followed by 4 hours in a vacuum oven (60°C, 380 mm Hg vacuum, argon bleed). The recovered

18 Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging binder samples were analyzed by liquid (CS2) FTIR to ensure complete solvent removal. Toluene removal was deemed complete when there was a negligible amount or no peak evident at 692 cm–1. Evaluation of Short-Term Conditioning Procedures Shortcomings of AASHTO T 240 In AASHTO T 240, short-term aging is simulated by exposing a moving film of asphalt binder to a stream of air at 163°C, near the upper range of typical construction temperatures. The premise for the procedure is that a container rotating in a carousel and containing a small amount of binder will produce a uniform and continuously renewed film of binder, thereby facilitating both oxidation and volatilization of the binder. While this premise is generally valid for unmodified binders conditioned at traditional HMA mixing temperatures, it may not be valid for many modified binders and temperatures associated with warm mix asphalt (WMA). The primary weakness of AASHTO T 240 is the inability of some asphalt binders to form a uniform film that is continuously renewed during the test. The uniformity of the film and how well it is renewed is viscosity dependent. Stiffer binders do not tend to “roll” in the container, and therefore the amount of aging depends on the binder’s viscosity; aging tends to decrease as the viscosity of the binder increases. Other deficiencies of AASHTO T 240 include: 1. Binders modified with elastomeric polymers tend to crawl out of the bottle due to the Weissenberg effect and the stream of air that tends to push the binder to the front of the bottle. 2. Less conditioning occurs when the test is conducted at higher elevations due to the pressure dependency of the asphalt oxidation reactions. 3. The shape of the bottle makes recovery of the binder and cleaning difficult. 4. The procedure uses a single temperature of 163°C, so the properties of binder used with WMA processes cannot be evaluated. 5. The RTFOT oven design and temperature control are outdated and could be improved. Several alternative short-term procedures have been proposed. These include the Modified German Rotating Flask (Robertson et al. 2001), the Rotating Cylinder Aging Test (Verhasselt 2003), the Ageing Profile Test (Hill et al. 2009), the Universal Simple Aging Test (Farrar et al. 2014), and the Stirred Airflow Test (Glover et al. 2001). Brief descriptions of these methods are presented below, followed by an assessment of these procedures as improvements to AASHTO T 240. Alternative Short-Term Conditioning Procedures Modified German Rotating Flask (MGRF) The procedure for the MGRF was developed by the Western Research Institute by modify- ing the procedure in German Standard DIN 52 016, Testing the Thermal Stability of Bitumen in a Rotating Flask, commonly referred to as the German Rotating Flask (GRF). The GRF was developed as a less expensive alternative to the RTFOT. It uses a rotary evaporator like that used ASTM D 5404, Standard Practice for Recovery of Asphalt from Solution Using the Rotary Evaporator, and AASHTO T 319, Standard Method of Test for Quantitative Extraction and Recovery of Asphalt Binder from Asphalt Mixtures, for recovery of binders after solvent extraction. The primary improvements to the GRF procedure incorporated in the MGRF are twofold: the sample size increased from 100 g to 200 g, and the binder is conditioned in a Morton flask (flask with indentations to promote agitation) to improve the uniformity of the binder film. Figure 8 presents a schematic of the MGRF. Operating parameters for the MGRF

Research Approach 19   are summarized in Table 4. The MGRF takes 2.5 times longer for short-term conditioning compared to AASHTO T 240. Neat and modified binders conditioned in the MGRF have rheological properties similar to those obtained by conditioning the binder in the RTFOT following AASHTO T 240 (Robertson et al. 2001, Ramaiah and D’Angelo 2003, Anderson and Bonaquist 2012). Rotating Cylinder Aging Test The Rotating Cylinder Aging Test (RCAT) was developed in Belgium primarily to simulate long-term aging (Verhasselt 2003). The procedure is standardized as European Standard EN 15323, Accelerated Long-Term Ageing/Conditioning by the Rotating Cylinder Method (RCAT). The RCAT conditions binder in an oven with a mechanism to rotate a cylinder containing binder and a solid steel roller. The operating principle for this device is the roller produces a film of binder that is renewed during rotation. Figure 9 is a schematic of the cylinder and roller. Although the device was developed to simulate long-term aging, it can be used for short-term conditioning using temperature and airflow rates that are the same as the AASHTO T 240. Table 5 summarizes the operating conditions for the RCAT. Using the commercially produced version of the device, both short- and long-term conditioning can be performed with the same equipment. A 500 g sample of binder is first short-term conditioned. Upon completion of the short-term conditioning, a portion of the sample is removed for physical property measurements. Figure 8. Schematic of the MGRF. Parameter Condition Sample Size 200 g Temperature 165°C Airflow 2 l/min Rotational Speed 20 rpm Heat-Up Time 10 min with no airflow Conditioning Time 200 min under airflow Table 4. Recommended operating parameters for the MGRF.

20 Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging The remainder of the sample is then long-term conditioned. A unique concept included in the long-term conditioning procedure is samples are removed at various times to allow character- ization of aging kinetics. Using the operating conditions listed in Table 5, the RCAT reasonably reproduces the short-term conditioning that occurs in AASHTO T 240 and the long-term conditioning that occurs in the AASHTO R 28 (Verhasselt 2003). Ageing Profile Test The Ageing Profile Test was developed in the United Kingdom (UK) to address issues associated with using AASHTO T 240 and AASHTO R 28 when conditioning polymer-modified binders (Hill et al. 2009). The procedure uses the RTFOT oven with the modified containers and screw mixers shown in Figure 10 to perform both short- and long-term conditioning of binders (Manual of Contract Documents 2019). The modifications include an aluminum container having a removable lid and coated with polytetrafluoroethylene (PTFE) as well as screws inserted into the containers to mix the binder during conditioning. Additionally, the amount of binder per container is reduced from 35 g to 19 g. The screws are designed to move binder into the container as it is rotated, but they also potentially improve the exposure of the binder to air during the test. The redesigned containers provide closer dimensional tolerances than are possible with glass containers, and the aluminum exterior provides structural stability that cannot be achieved with PTFE containers. The Ageing Profile Test combines short- and Figure 9. Rotating Cylinder Aging Test (RCAT). Test Parameter Condition Short-Term Sample Size 500 g Temperature 163 C Speed 1 rpm Airflow Air at 4 l/min Conditioning Time 235 min Long-Term Sample Size Varies Temperature 90°C Speed 1 rpm Airflow Oxygen at 4.5 l/hr Conditioning Time 140 hr ° Table 5. Operating parameters for the RCAT.

Research Approach 21   long-term conditioning into one procedure. First, the binder is short-term conditioned at 163°C for 45 minutes, and two containers are removed for characterizing the properties of the short-term conditioned residue. The temperature is then reduced to 135°C, and two containers are removed after 4 hours, 8 hours, and 22 hours to develop the long-term aging profile for the binder. Comparisons of rheological properties for a limited number of binders indicate that 8 hours of conditioning at 135°C is approximately equivalent to 65 hours of PAV conditioning at 85°C. Table 6 summarizes the operating parameters for the Ageing Profile Test. Universal Simple Aging Test The Universal Simple Aging Test (USAT) is a thin-film test designed for both short- and long-term conditioning and is intended to be used with 4 mm DSR testing (Farrar et al. 2014). The test conditions a thin film of binder that is only 0.3 mm thick compared to 3.18 mm for the PAV. Figure 11 is a schematic of the plates used to prepare the binder film. The plates were designed to fit in the PAV rack. Each slot in the plate in Figure 11 contains 1 g of binder. Short-term aging is simulated by conditioning the film in a forced draft oven for 50 minutes at 150°C for HMA or 130°C for WMA. Long-term aging is simulated by further conditioning the binder for 8 hours in a PAV at the AASHTO R 28 temperature and pressure conditions. Table 7 summarizes the operating parameters for the USAT. Stirred Airflow Test AASHTO T 240, the MGRF, the RCAT, the Ageing Profile Test, and the USAT are based on conditioning thin films. The Stirred Airflow Test (SAFT), developed at Texas A&M University, a. PTFE-coated aluminum container. b. Screws for mixing binder during the test. Figure 10. Modifications to AASHTO T 240 for the UK Ageing Profile Test. Test Parameter Condition Short-Term Sample Size 152 g (8 containers at 19 g each) Temperature 163°C Speed 15 rpm Airflow Air at 4 l/min Conditioning Time 45 min Long-Term Sample Size 114 g (6 containers at 19 g each) Temperature 135°C Speed 15 rpm Airflow Air at 4 l/min Conditioning Time 4 hr, 8 hr, and 22 hr Table 6. Operating parameters for the UK Ageing Profile Test.

22 Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging is somewhat different and utilizes air blowing to condition binders (Glover et al. 2001). Labora- tory conditioning using thin films and air blowing are similar mechanistically if the air bubbles remain small and well dispersed. Smaller air bubbles increase the ratio of surface area to volume of the asphalt, and the reaction becomes less diffusion-limited and behaves more like a thin-film reaction. Figure 12 is a schematic of the SAFT. In the SAFT, air from a nozzle submerged in the binder is dispersed by an impeller mounted to an external motor. The SAFT also includes a simple condenser to trap volatiles emitted by the binder during conditioning. Through trial and error, operating parameters for the device that approximate the conditioning that occurs in the RTFOT were determined. Table 8 summarizes the operating parameters for the SAFT. Two sets of operating parameters are given in Table 8 because the initial commercial version of the SAFT differed from the schematic shown in Figure 12. The initial commercial version used an oven to heat the vessel rather than a heating mantle, thereby eliminating the high temperatures Figure 11. USAT specimen plate. Test Parameter Condition Short-Term Sample Size 3 g per USAT plate Temperature 150°C for HMA 130°C for WMA Conditioning Time 45 min Long-Term Sample Size 3 g per USAT plate Temperature 100°C Pressure 2.1 MPa Conditioning Time 8 hr Table 7. Operating parameters for the USAT.

Research Approach 23   at the vessel wall that occur when using a heating mantle. This resulted in somewhat longer conditioning times (Anderson and Bonaquist 2012). Comparisons of rheological properties for six neat binders and five polymer-modified binders conditioned in the SAFT and the RTFOT showed the SAFT-conditioned binders had lower stiffness for higher viscosity neat binders and polymer-modified binders (Anderson and Bonaquist 2012). This may be the result of (1) the distribution of air bubbles in the SAFT being viscosity dependent, and/or (2) the Weissenberg occurs resulting in binder coalescing along the impeller, producing poorly dispersed air bubbles. Assessment of Short-Term Conditioning Procedures An engineering assessment of the various short-term conditioning procedures as replace- ments for AASHTO T 240 was conducted. The criteria used in the assessment and the results are summarized in Table 9. A discussion of several of the criteria is presented below. 1. Quantity of Binder and Number of Binders per Run. It is critical that the short-term procedure yields enough binder for performance grading, and this was one of the criteria Figure 12. Schematic of the SAFT. Parameter Texas A&M Prototype Initial Commercial Sample Size 250 g 250 g Temperature 163°C 163°C Airflow 2,000 ml/min 2,000 ml/min Stirrer Speed 700 rpm 1,000 rpm Heat-Up Time 15 min under nitrogen 15 min under nitrogen Conditioning Time 30 min under air 50 min under air Table 8. Proposed operating parameters for the SAFT.

24 Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging used to select AASHTO T 240 during the SHRP. All short-term procedures, except the USAT, meet this criterion. As discussed above, the developers of the USAT envisioned replacing current BBR testing with 4 mm DSR testing. Since 4 mm DSR testing requires research- grade equipment, and there currently is not a coordinated national effort to standardize the equipment and procedures, it is not appropriate to base the work in this project on the future use of 4 mm DSR testing for routine performance grading. Although most of the procedures in Table 9 provide more than enough binder for performance grading, AASHTO T 240 and the Aging Profile Test offer the ability to short-term condition more than one binder per run. This is a significant consideration for laboratory throughput. 2. Viscosity Effect on Film Uniformity, Diffusion Effect, and Binder Creep. Although minimizing the viscosity effect, the diffusion effect, and the creep of binder from the containers are stated as reasons for developing the various alternatives to AASHTO T 240, only the USAT, where a uniform thin film is manually prepared, fulfills these criteria. Data supporting the development of the other alternative procedures generally include a com- parison showing how well rheological properties from the alternative procedures match AASHTO T 240 for both neat and modified binders. This suggests two possibilities: (1) the alternative method does not improve film uniformity and reduce the diffusion effect over a wide range of binder consistencies, or (2) the alternative method does improve film unifor- mity over a wide range of binder consistencies, but stiffer binders actually age less in mixtures during short-term conditioning compared to softer binders. Except for the SAFT, visual evidence supports the ability of the alternative procedures to minimize binder creep from the containers. Visual evidence is not available for the SAFT, which has an impeller rotating in a closed container. There are many videos available online demonstrating the Weissenberg effect by rotating a shaft in an elastic liquid. The Weissenberg effect may occur in the SAFT, but it is not visible. Consideration Short-Term Conditioning Procedure MGRF RCAT Aging Profile Test USAT SAFT Sufficient Quantity of Binder for Current Specification Testing Yes Yes Yes No Yes Number of Binders per Run 1 1 Up to eight if only short-term is done Multiple 1 Reduces Viscosity Effect on Film Uniformity Maybe Maybe Maybe Yes Maybe Reduces Diffusion Effect Maybe Maybe Maybe Yes Maybe Reduces Creep of Binder from Container Yes Yes Yes Yes Unknown Suitable for Ground-Tire-Rubber-Modified Binders Yes Unknown Unknown No Unknown HMA and WMA Temperatures Possible Possible Yes Yes Possible Measure of Binder Volatility Mass change None Mass change Mass change Collection system Improved Residue Recovery Yes Yes Yes Yes Yes Improved Test Control Possible Yes Unknown With oven and procedure modifications Yes Possible Reduces Elevation Effect Possible Possible Possible Possible Possible AASHTO Standard Test No No Modification of T 240 No No Equipment Commercially Available Yes Yes Most Most No Equipment Cost Modest High Low Low High Additional Technician Training Cost Modest Modest Low Low Modest Table 9. Summary of assessment of available short-term procedures.

Research Approach 25   3. Suitability for Use with GTR-Modified Binders and WMA. There is renewed interest in using GTR-modified binders and adapting the performance grading system for specifying these materials. GTR-modified binders with high rubber content and large particles are particularly problematic when conditioned according to AASHTO T 240. Interestingly, one of the early studies using the GRF that was conducted in Florida to address modified binders that could not be properly conditioned by AASHTO T 240 included crumb-rubber- modified-binders that were successfully conditioned (Sirin et al. 1998). The primary concern with the procedures that use rods or screws to improve the film uniformity is whether these elements grind or break down the rubber particles during laboratory conditioning. The film in the USAT is thinner than most coarse rubber particles, making this procedure unaccept- able for these materials. Issues associated with the uniformity of the film and diffusion in AASHTO T 240 are exacerbated by the low temperatures for WMA. Procedures for WMA were specifically addressed in the development of the Ageing Profile Test and the USAT. It should be possible to address WMA with the other alternate procedures. 4. Measure of Binder Volatility. It is important to include a measure of binder volatility in the improved short-term procedure, particularly considering some of the new products and blends that are being considered. All alternate procedures, except the RCAT, provide some measure of binder volatility. The original simple condenser included in the SAFT was improved in NCHRP Project 09-36 to include polymer beads that adsorb different compo- nents (Anderson and Bonaquist 2012). This approach can also be applied to the MGRF. 5. Improved Residue Recovery and Improved Test Control. The shape of the bottles used in AASHTO T 240 makes residue recovery tedious, particularly for modified binders, and produces a significant dwell time between the recovery from the first and last container. All alternate procedures provide improved residue recovery procedures. Additionally, temperature control in the RTFOT oven is challenging and between laboratory variability, can perhaps be improved by redesigning the oven, particularly the airflow and the position of the heating elements. Such modifications are beyond the scope of this project but could be done for AASHTO T 240 and the Ageing Profile Test. Temperature control for the MGRF is provided by an oil bath, while the RCAT and USAT use modern oven designs that could be applied to the SAFT and the RTFOT. 6. Elevation Effect. Research conducted in NCHRP Project 20-07 Task 400 and by others has shown that there is an elevation effect in the properties of RTFOT residue when the conditioning is conducted at different elevations. The effect is large enough to result in different conclusions being drawn concerning the acceptability of the same binder when tested at different elevations (Advanced Asphalt Technologies 2018, Wang 2013, Velasquez et al. 2013). The elevation effect is due to the difference in atmospheric pressure with eleva- tion and the pressure dependency of the oxidation reactions in asphalt. Short-term aging at a constant absolute pressure is the best approach to minimize this effect; potentially, it can be applied to all alternate procedures, although it will require significant equipment redesign. 7. Standardization, Commercialization, and Costs. These are important considerations when implementing new equipment. Although there are draft standards or international standards for all alternate procedures none have been adopted by AASHTO or ASTM. The containers and mixing elements used in the Ageing Profile Test can be added as a revision to AASHTO T 240. All other methods would require adopting a new standard method of test. The equipment for most of the alternate methods is commercially available; the exceptions are the containers and screw mixers for the Ageing Profile Test, the USAT aging plates, and the SAFT. A commercial version of the SAFT was developed, but it is not currently being marketed. The MGRF, RCAT, and SAFT would require replacing current equipment, with the costs for the RCAT and SAFT being higher than those of the MGRF. The cost of the containers and mixing screws from the Ageing Profile Test and the USAT plates are relatively low. Finally, training costs for all the alternate procedures are low to moderate.

26 Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging Based on the engineering assessments presented above, two procedures were selected for further evaluation as possible improvements to AASHTO T 240: modifying AASHTO T 240 to use containers and mixing screws similar to those used in the UK Ageing Profile Test; and static, thin-film test similar to the USAT, but using a thicker film to provide enough conditioned binder for performance grading. The following sections present additional details for these two selections. Screw Mixers for the RTFOT Published data for the UK Ageing Profile Test show the mixing screws stop polymer-modified binders from crawling out of the container at temperatures as low as 135°C (Hill et al. 2009). However, it is not clear that the mixing screws, as designed, reduce the viscosity effect where stiffer binders tend to age less in AASHTO T 240. Data for the same binders conditioned in the standard RTFOT and the Ageing Profile Test are compared in Figure 13 and Figure 14. These figures show similar shear modulus and phase angles for binder conditioned in the RTFOT and the Ageing Profile Test for a wide range of binder stiffnesses. These comparisons suggest that the Ageing Profile Test mixing screws are not effective at improving the exposure to air for binders with a wide range of consistencies. If the mixing screws reduce the viscosity effect and improve the uniformity of the film during conditioning, then the Ageing Profile Test would be expected to produce higher moduli and lower phase angles compared to the standard RTFOT for the stiffer binders. This is clearly not the case in Figure 13 and Figure 14. The mixing screws used in the UK Ageing Profile Test are a standard screw design that the developers confirm was not optimized to minimize the viscosity effect on film uniformity and diffusion (Anderson 2016). Figure 15 shows one possible design that will eliminate binder creeping from the bottle and may improve uniformity of the film. In this design, the screw sections move material toward the center where the solid section produces a uniform film. Unfortunately, it is not possible to evaluate how well various designs will work with binders Figure 13. Comparison of shear modulus for binder conditioned in the RTFOT and the UK Ageing Profile Test (data from Hill et al. 2009).

Research Approach 27   of varying consistencies through analysis or simulation. The short-term selection experiment presented later in this chapter includes evaluation of both the original UK design and the design shown in Figure 15 to determine whether a continuous screw or a screw with a solid section to form a film is better. Static, Thin-Film Test The primary deficiency with the USAT identified by the engineering assessment is each USAT plate yields only 3 g of binder because the developers of the USAT envisioned simultaneous adoption of the USAT and 4 mm DSR testing for low-temperature grading (Farrar et al. 2014). The USAT is a static, thin-film conditioning procedure similar to AASHTO T 179, Standard Method of Test for Effects of Heat and Air on Asphalt Materials (Thin-Film Oven Test), except the film thickness is reduced from 3.18 mm to 0.3 mm to reduce the effect of diffusion. This results in more rapid short-term conditioning in the USAT (45 minutes) compared to AASHTO T 179 (5 hours). As discussed in the assessment of long-term conditioning procedures presented later, the use of thinner films for conditioning in the PAV is a promising option for improving long- term conditioning. The use of the same film thickness for short- and long-term conditioning offers the potential to simplify laboratory conditioning. A very reasonable approach is to short- term condition the binder in the pans used for the long-term procedure in a low-pressure PAV Figure 14. Comparison of phase angle for binder conditioned in the RTFOT and the UK Ageing Profile Test (data from Hill et al. 2009). Figure 15. Conceptual design of screw mixer for the RTFOT.

28 Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging at 163°C and 110 kPa (kilopascal) absolute pressure. Using a low-pressure PAV with absolute pressure control for short-term conditioning would eliminate the laboratory elevation effect that is known to be an issue with AASHTO T 240. The short-term conditioned residue from one pan would be used for testing the short-term conditioned binder. The remaining pans would be cooled, weighed for mass change determination, and then transferred to a standard PAV for long-term conditioning. This approach has advantages: 1. It removes the viscosity dependent film renewal, laboratory elevation effect, and binder recovery issues associated with the AASHTO T 240, 2. It eliminates binder leakage when conditioning heavily modified binder, and 3. The higher temperature for the short-term conditioning will improve the uniformity of the thin film for the subsequent long-term conditioning. The primary unknown for this approach is the film thickness that provides appropriate short- and long-term conditioning. A preliminary investigation of the effect of film thickness on short-term conditioning was conducted during the evaluation phase of NCHRP Project 09-61 to obtain an estimate of the film thickness in a static, thin-film test yielding approximately the same short-term conditioned residue properties as AASHTO T 240. SHRP binders AAC-1 and AAF-1 were conditioned in PAV pans in a forced draft oven at 163°C for 85 minutes. Film thicknesses of 0.8 mm, 1.08 mm, 1.59 mm, and 3.18 mm were used. SHRP binders AAC-1 and AAF-1 were used because they exhibited different sensitivities to thickness in a PAV film thickness experiment conducted during SHRP (Anderson et al. 1994). The response measured in this experiment was the AASHTO M 320 high pavement temperature parameter G*/sind at the grade temperature for each binder. The results are summarized in Table 10 and shown graphically in Figure 16 for AAC-1 and Figure 17 for AAF-1. These data and comparisons show that a film thickness of approximately 0.8 mm to 0.9 mm provides approximately the same short-term conditioning as AASHTO T 240. Evaluation of Long-Term Conditioning Procedures Shortcomings of AASHTO R 28 The PAV procedure, AASHTO R 28, was refined during SHRP (Anderson et al. 1994) and adopted in AASHTO M 320 and AASHTO M 332 as the best compromise for simulating the aging that occurs during the first 4 to 8 years of pavement service life. In AASHTO R 28, asphalt binder aging is accelerated using a combination of elevated temperature and elevated air pressure. In the current procedure, a relatively thick film of binder, 3.18 mm that has been conditioned following AASHTO T 240 is treated in pans in an oven at temperatures of either 90°C, 100°C, or 110°C under 2.1 MPa air pressure for 20 hours. The temperature that is Binder Temp, °C G*/Sinδ, kPa Original RTFOT 0.80 mm 1.08 mm 1.59 mm 3.18 mm AAC-1 52 5.06 58 1.550 3.28 3.23 2.90 2.29 1.91 64 0.705 1.51 1.46 1.30 1.03 AAF-1 58 4.25 3.49 64 1.190 3.01 3.36 2.42 1.85 1.63 70 0.566 1.33 1.46 1.08 Table 10. Effect of film thickness on short-term conditioned G*/sinc.

Research Approach 29   1.00 1.50 2.00 2.50 3.00 3.50 0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 G* /s in δ at 5 8° C, k Pa Film Thickness, mm AAC-1 Thin Film Conditioned RTFOT Original Figure 16. Short-term conditioning film thickness results for SHRP Binder AAC-1. 1.00 1.50 2.00 2.50 3.00 3.50 0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 G* /s in δ at 6 4° C, k Pa Thickness, mm AAF-1 Thin Film Conditioned RTFOT Original Figure 17. Short-term conditioning film thickness results for SHRP Binder AAF-1.

30 Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging used depends on the environmental conditions where the binder will be used, with the lower temperature used in the extreme northern United States, the middle temperature used in most of the United States, and the highest temperature used in desert climates. Practitioners have raised several concerns with AASHTO R 28, the most important being: 1. The service life simulated by the procedure is not well defined and appears to be binder dependent. The conditioning in the PAV may simulate 5 years of life in service for one binder, and 10 or more years for another binder (Anderson et al. 1994). 2. For the standard test duration, the amount of conditioning is not sufficiently severe to simulate long-term aging that occurs near the pavement surface (Hanson et al. 2009). PAV conditioning for 40 hours is now being used by researchers to better simulate longer service life (Hanson et al. 2009) and to better discriminate differences between binders, especially those containing recycled engine oil bottoms (REOB) and recycled asphalt shingles (RAS) (Reinke et al. 2016). 3. It is not clear if the conditioning temperatures included in the test procedure cover the range of environmental temperatures encountered in the United States (Houston et al. 2005). Alternative Long-Term Conditioning Procedures Available Approaches Although there are several concerns with AASHTO R 28, only a few alternative procedures have been proposed. These include the RCAT, the Ageing Profile Test, and the USAT described earlier in the section on short-term conditioning. Additionally, researchers have suggested increasing the time (Erskine et al. 2012, Reinke et al. 2016), decreasing the film thickness, (Erskine et al. 2012), or increasing the temperature (Houston et al. 2005) in the PAV to better simulate in-service aging. Interestingly, developers of alternative long-term aging procedures generally compare their procedures to AASHTO R 28 rather than actual field aging. Research in Canada that included 40-hour PAV and reduced film thickness 20-hour PAV included comparisons of rheological properties of laboratory-conditioned binder to field-aged binder. Those com- parisons suggest the standard 3.18 mm film thickness conditioned for 40 hours and thinner 0.8 mm film thickness conditioned for 20 hours represent approximately 8 to 10 years in service (Erskine et al. 2012). Figure 18 compares the low-temperature continuous grade data for 11 binders conditioned for 40 hours using the standard film thickness (3.18 mm) and 20 hours using one-quarter of the standard film thickness (0.8 mm). The data are scattered around the line of equality indicating these two conditions are approximately equivalent. In NCHRP Project 09-23, temperatures as a function of climate and time in service for standard 20-hour PAV conditioning were developed using the global aging system (Mirza and Witczak 1995) and limited recovered binder testing (Houston et al. 2005). These recommended temperatures are reproduced in Table 11 and suggest the need for somewhat higher temperatures than currently used in AASHTO R 28 if the conditioning time of 20 hours is maintained. Feasibility of Using Resonant Acoustic, Sonic, and Ultrasonic Mixing to Accelerate Oxidation Acceleration A novel approach for improving long-term laboratory conditioning that was included in the evaluation phase was to reduce the diffusion effect in the PAV by increasing the exposure of the film to air through acoustic, sonic, or ultrasonic mixing. The sections below discuss the feasibility of these methods. Resonant Acoustic Mixing. Resonant acoustic mixing introduces acoustic energy into liquids, slurries, powders, and pastes. An oscillating mechanical driver creates motion in a

Research Approach 31   mechanical system comprised of engineered plates, eccentric weights, and springs. This energy is then transferred to the material to be mixed. Low-frequency, high-intensity acoustic energy is used to create a uniform shear field throughout the entire mixing vessel. The result is rapid fluidization and dispersion of material in the container. Figure 19 is a schematic of the resonant acoustic mixing process. A representative from an engineering firm demonstrated the resonate acoustic mixing system to the research team. Mixing can be done in a heated vessel, which can also be pressurized and then evacuated to remove air bubbles after conditioning. In the initial demonstration, -34 -32 -30 -28 -26 -24 -22 -34 -32 -30 -28 -26 -24 -22 20 H ou r, 0. 8 m m T hi ck ne ss P AV L ow T em pe ra tu re Co nti nu ou s G ra de , ° C 40 hour, 3.18 mm Thickness PAV Low Temperature Continuous Grade, °C Figure 18. Comparison of PAV film thickness and aging time (data from Erskine et al. 2012). Location Mean Annual Air Temperature (MAAT) °C Recommended PAV Aging Temperature, °C 5 yrs 10 yrs 15 yrs 20 yrs Barrow, AK −11.0 85 85 90 90 Fargo, ND 5.9 95 95 100 100 Billings, MT 8.7 95 100 105 105 Chicago, IL 11.6 95 100 105 105 Washington, DC 12.9 95 100 105 110 San Francisco, CA 13.8 100 105 105 110 Oklahoma City, OK 15.9 100 105 110 110 Dallas, TX 19.3 100 110 110 115 Las Vegas, NV 20.5 100 110 115 115 Phoenix, AZ 23.6 105 110 115 120 Not previously defined, in alphabetical order: AK = Alaska, AZ = Arizona, DC = District of Columbia, IL = Illinois, ND = North Dakota, NV = Nevada, OK = Oklahoma Table 11. Recommended PAV aging temperatures from NCHRP Project 09-23 (data from Houston et al. 2005).

32 Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging PAV-conditioned binder was mixed in an unheated container at 100°C. This demonstration showed that the resonant acoustic mixer can entrain air and mix a typical PAV-conditioned binder. The firm agreed to perform additional trials. Binder samples were sent to condition at 100°C in a heated mixing vessel at atmospheric pressure for 4, 8, and 20 hours. The plan was to compare the rheological properties of these samples to the same binder conditioned in accordance with AASHTO R 28. After additional trials, it was concluded that resonant acoustic mixing would not accelerate conditioning in the PAV. Sonic Mixing. The feasibility of using sonic mixing was evaluated at the Western Research Institute using a standard sonic cleaning bath. Asphalt ARC BI-0002 was poured into tins to form 1 mm thick samples. The thickness was sufficient to be impacted by the diffusion of oxygen into the sample. Control samples were heated in an oven at 90°C for 24, 48, and 96 hours to determine the amount of oxidation at these times without mixing. Other samples were placed in a sonic bath that was filled with mineral oil and equipped with coiled tubing attached to a heater–chiller bath. The tins with asphalt were clamped into place within the sonic bath (50–60 Hz); sonic mixing was introduced, and the temperature was maintained at 60°C or 90°C for 24 and 48 hours. Upon introducing sonic mixing, a standing wave was formed in the asphalt as shown in Figure 20. FTIR data were collected on the samples to determine the amount of oxidation that had occurred in the samples. The samples were heated and stirred before sampling. Oxidative aging due to oxygen uptake can be clearly monitored by observing the carbonyl (C=O) and sulfoxide (S=O) bands in the FTIR spectra. Figure 21 shows an overlay of the 24- and 48-hour conditioned samples. Despite being able to produce a standing wave in the asphalt sample, sonic mixing did not have a significant impact on the oxidation of the samples. It was assumed that the areas of thinner and thicker asphalt generated by the standing wave could produce areas of significantly decreased diffusion and increase the rate of oxidation. It was not investigated to see if moving the sample in the sonic bath would change the rate of oxidation. Ultrasonic Mixing. Ultrasonic agitation occurs at frequencies near 20,000 Hz, and it can introduce a significant amount of energy into the system to increase the rate of chemical reactions (Suslick 1990). An ultrasonic horn, operating at 20 kHz and fitted with a microtip suitable for up to 10 mL of sample, was used to induce oxidation in asphalt. The asphalt sample had ultrasound induced in it while it was inside an aluminum block with a cell cut into it. The cell was approximately 1 in long by 1 in wide and 1.5 in tall. The aluminum block was preheated, and an asphalt sample approximately 1 in deep was poured into the cell. The ultrasonic tip was Figure 19. Schematic of resonant acoustic mixing.

Research Approach 33   Figure 20. Photograph of asphalt in a sonic bath (50–60 Hz) with a standing wave produced from sonic mixing at 60çC. Figure 21. Solution FTIR spectra for ARC BI-0002 unheated control and samples heated at 90çC using an oven and sonic mixing (50–60 Hz) for 24 and 48 hours.

34 Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging inserted 0.5 in into the asphalt, and the power and cycling of the ultrasound were adjusted to maintain a constant temperature that was measured using a thermocouple inserted into the asphalt. It was found for the ARC BI-0002 asphalt that cavitation (bubbles forming) and convection mixing would not occur until the asphalt was warmed to about 120°C [cavitation due to ultrasound usually requires a viscosity < 5,000 centipoise (cP)]. Two short-term experi- ments were performed so that the temperature was maintained at about 120°C and 140°C for two hours. FTIR data from these experiments showed that there was negligible oxidation in the sample. Since there was no observable oxidation after 2 hours of ultrasonic mixing, the experiment was repeated for 8 hours; the FTIR spectra are shown in Figure 22. After 8 hours of ultrasonic mixing, there was still no observable oxidation. Ultrasonic mixing introduces cavitation, which can cause localized areas with temperatures of up to 5,000°C and pressures up to 500 atmospheres (atm) over the course of a few micro- seconds (Suslick 1990). These conditions can be used to accelerate chemical reactions; however, in asphalt, if the conditions are not carefully controlled, chemical reactions can take place that do not occur at normal in-service pavement conditions. Without adequate control, ultrasound can also cause various modifiers to be decomposed. Because of the various technical difficulties associated with controlling ultrasonic-induced asphalt oxidation, the technology is not consid- ered feasible at this time. Assessment of Long-Term Conditioning Procedures An engineering assessment of the available long-term conditioning procedures as replace- ments for AASHTO R 28 was conducted. Resonant acoustic, sonic, and ultrasonic mixing were not included because these technologies were not considered feasible based on the feasibility evaluation described above. The criteria used in the assessment and the results are summarized in Table 12. A discussion of several of the criteria is presented below. 1. Quantity of Binder and Number of Binders per Run. It is critical that the long-term procedure yields enough binder for performance grading. All long-term procedures, except Figure 22. Solution FTIR spectra for ARC BI-0002 unheated control and samples heated at 120çC using an ultrasonic horn (20,000 Hz) for 8 hours.

Research Approach 35   the USAT, meet this criterion. As discussed above, the developers of the USAT envisioned replacing current BBR testing with 4 mm DSR testing. Since 4 mm DSR testing requires research-grade equipment, and there currently is not a coordinated national effort to stan- dardize the equipment and procedures, it is not appropriate to base the work in this project on the future use of 4 mm DSR testing for routine performance grading. Like the RTFOT, the PAV offers the advantage that multiple binders can be conditioned during one run of the equipment. AASHTO R 28 has the provision for 10 pans holding 50.0 g of binder. With the current pan and rack design and using one-quarter of the film thickness, only one binder can be conditioned per run; however, the pan and rack can be redesigned to increase the yield of the PAV at a reduced film thickness. 2. Time to Complete Conditioning. The time to complete the conditioning also affects the laboratory throughput. The 20-hour conditioning time in AASHTO R 28 is convenient for a single shift operation and allows the PAV equipment to be cycled with new binders each day if necessary. The 40-hour conditioning that has been recommended is not convenient for laboratory throughput. The RCAT and USAT conditioning times of 140 and 8 hours respec- tively produce aging approximately equivalent to AASHTO R 28. These conditioning times will be longer if it is desired to simulate greater in-service aging. 3. Temperature, Atmosphere, and Pressure. Although increasing the temperature is a very efficient way to accelerated laboratory conditioning, research on the aging of asphalt binders has shown that the oxidation reactions change when the temperature exceeds about 110°C (Peterson 2009). The 135°C used in the long-term portion of the Ageing Profile Test and suggested for long-term mix aging (Braham et al. 2009) exceeds this threshold temperature as do some of the temperatures recommended from the NCHRP Project 09-23 research. The use of pure oxygen or oxygen-enriched air, particularly under pressure, with hydrocarbon materials produces a significant laboratory safety issue. In the RCAT, conditioning is done at atmospheric pressure, while the USAT and AASHTO R 28 use a relatively high pressure of 2.1 MPa (300 psi). The long duration of the long-term conditioning in the RCAT is a function of the use of atmospheric pressure and 90°C temperature. Models of the pressure dependency Consideration Long-Term Aging Procedure RCAT Aging Profile Test USAT Extended Time PAV Thinner Film PAV Increased Temperature PAV Sufficient Quantity of Binder for Current Specification Testing Yes Yes No Yes Yes Yes Number of Binders per Run 1 1 Multiple 3 Depends on film thickness and pan/rack design 3 Time to Complete Conditioning, hrs 140 8 8 Up to 40 20 20 Temperature, °C 90 135 100 90–110 90–110 Up to 120 Atmosphere Oxygen Air Air Air Air Air Pressure, MPa Atmospheric Atmospheric 2.1 MPa 2.1 MPa 2.1 MPa 2.1 MPa Reduces Diffusion Effect Maybe Maybe Yes No Yes Maybe Correlated to Field Aging Indirectly Indirectly Indirectly Limited Limited Limited AASHTO or ASTM Standard Test Method No Modification of T 240 No Modification of R 28 Modification of R 28 Modification of R 28 Equipment Commercially Available Yes Most No Yes Except pans Yes Equipment Cost High Low Low Low Low Low Additional Technician Training Cost Modest Low Low Modest Low Low Table 12. Summary of assessment of available long-term procedures.

36 Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging of the asphalt oxidation reactions indicate it is a power function of pressure, with an exponent between about 0.2 and 0.6 (Liu et al. 1996). For an exponent between 0.2 and 0.6, 2.1 MPa pressure is on the relatively flat portion of the curve. Doubling the pressure from 2.1 MPa to 4.2 MPa would change the reaction by between 8 percent and 15 percent while significantly impacting the design of the vessel. 4. Reduces Diffusion Effect. The main criticism of AASHTO R 28 is the reaction in the relatively thick 3.18 mm film is diffusion controlled. The RCAT, Ageing Profile Test, and the USAT attempt to reduce the diffusion effect by producing thinner films. This is accomplished in the USAT, but it is not clear how well the long-term portion of the RCAT and Ageing Profile Test produce uniform thin films. Increasing temperature in the PAV will alter the diffusion rate; however, the better approach is to reduce the film thickness. As discussed earlier, limited data collected in Canada suggest conditioning using one-quarter of the current film thickness is similar to doubling the conditioning time at the current film thickness. 5. Correlated to Field Aging. Only limited correlations to field aging were completed during SHRP (Anderson et al. 1994, Lytton et al. 1993) and after (Houston et al. 2005, Hanson et al. 2009, and Erskine et al. 2012). The consensus of these limited correlations is that AASHTO R 28 approximates 5 to 8 years of in-service aging. The RCAT, Ageing Profile Test, and USAT have only been indirectly correlated to field aging by comparing results from these alternative tests to the results from AASHTO R 28. 6. AASHTO or ASTM Standard Test Method, Commercialization, and Costs. These are important considerations when implementing new equipment. Although there are inter- national standards for the RCAT and Ageing Profile Test, neither has been adopted by AASHTO or ASTM. The USAT, and changes in duration, film thickness, or temperature in the PAV could be added as revisions to AASHTO R 28. Most of the equipment for the alter- nate long-term methods is commercially available, the exceptions being the containers and screw mixers for the Ageing Profile Test and the USAT aging plates. The RCAT would require replacing current equipment at a relatively high cost. The costs associated with the additional or modified equipment for the other alternates are low. Finally, training costs for all alternate procedures are low to moderate. Based on the engineering assessment presented above, a thinner film thickness for the PAV was selected for further evaluation and calibration. A preliminary film thickness experiment was conducted to estimate the film thickness that would provide long-term conditioning approximately equivalent to 40 hours of PAV conditioning at the standard 3.18 mm film thickness. SHRP binders AAC-1 and AAF-1 were conditioned for 20 and 40 hours in the PAV at 100°C using 2.1 MPa pressure. Film thicknesses of 0.8 mm, 1.08 mm, 1.59 mm, and 3.18 mm were used. SHRP binders AAC-1 and AAF-1 were used because they exhibited different sensitivities to thickness in a PAV film thickness experiment conducted during SHRP (Anderson et al. 1994). The responses measured in this experiment were: (1) creep stiffness and m-value (the rate at which the asphalt binder stiffness changes over time) for BBR tests conducted at −6°C for AAC-1 and 0°C for AAF-1, and (2) DSR frequency sweep data at 10°C, 22°C, and 34°C. The frequency sweep data were used to calculate the crossover frequency and rheological index for the Christensen-Anderson model and the Glover-Rowe parameter (Rowe 2011). The results are presented in Table 13. Figure 23 and Figure 24 compare the BBR data from AAC-1 and AAF-1, respectively. In these figures, the stiffness increases and the m-value decreases with increasing conditioning time and decreasing film thickness. These figures show that using a film thickness of 0.8 mm to 0.9 mm produces similar low-temperature properties after 20 hours of PAV conditioning as 40 hours of PAV conditioning using the standard film thickness of 3.18 mm. Similar results are shown in Figure 25 and Figure 26 for the rheological index and the crossover frequency. In this case, the crossover frequency decreases, and the rheological index increases with increasing conditioning time and decreasing film

Binder Time, hrs Thickness, mm Rheological Index (R) Crossover Frequency (ω c), Hz Glover-Rowe Parameter, kPa Creep Stiffness, MPa m- value AAC-1 20 0.80 2.501 5.33 175.4 90.3 0.362 1.08 2.334 12.46 107.8 90.5 0.366 1.59 2.290 14.85 101.2 86.7 0.368 3.18 2.168 38.21 53.3 80.8 0.387 40 0.80 2.785 0.69 399.4 102.0 0.327 1.08 2.681 1.00 361.6 101.0 0.333 1.59 2.676 1.20 332.1 98.8 0.338 3.18 2.493 4.85 171.8 90.0 0.357 AAF-1 20 0.80 2.337 1.151 629.1 80.6 0.371 1.08 2.319 1.795 459.7 74.7 0.382 1.59 2.243 2.972 362.5 68.0 0.391 3.18 2.038 12.190 162.1 57.5 0.430 40 0.80 2.660 0.017 1952.5 108.0 0.323 1.08 2.566 0.127 1779.5 97.1 0.333 1.59 2.567 0.209 1279.4 94.2 0.343 3.18 2.289 1.821 489.8 78.4 0.380 Table 13. Effect of film thickness and conditioning time on long-term conditioned residue properties. 0.20 0.25 0.30 0.35 0.40 0.45 1.90 1.92 1.94 1.96 1.98 2.00 2.02 m -v al ue Log Stiffness, MPa AAC-1, 20 hrs AAC-1, 40 hrs 0. 80 m m 3. 18 m m 1. 59 m m 1. 08 m m 3. 18 m m 1. 59 m m 0. 80 m m 1. 08 m m Figure 23. BBR stiffness and m-values at 0çC for AAC-1 for various long-term conditioning times and film thicknesses. 0.20 0.25 0.30 0.35 0.40 0.45 1.70 1.75 1.80 1.85 1.90 1.95 2.00 2.05 m -v al ue Log Stiffness, MPa AAF-1, 20 hrs AAF-1, 40 hrs 0. 80 m m 3. 18 m m 1. 59 m m 1. 08 m m 3. 18 m m 1. 59 m m 0. 80 m m 1. 08 m m Figure 24. BBR stiffness and m-values at -6çC for AAF-1 for various long-term conditioning times and film thicknesses.

38 Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging thickness. Finally, referring to Table 13, the Glover-Rowe parameter increases with increasing conditioning time and decreasing film thickness. Again, similar Glover-Rowe parameters are obtained for 20-hour PAV conditioning using a film thickness of 0.8 mm to 0.9 mm and 40-hour PAV using the standard film thickness of 3.18 mm. The results from this preliminary experi- ment confirm that a reasonable way to obtain greater long-term conditioning is to reduce the film thickness in the PAV. PAV Pan and Rack Design for Thinner Films Using thinner films in the PAV raises two concerns. The first is whether enough binder for continuous grading can be conditioned in the PAV in a single run. The second is whether current tolerances are acceptable when conditioning thinner films. The sections that follow present analyses of these two issues. Figure 25. Rheological index and crossover frequency AAC-1 for various long-term conditioning times and film thicknesses. Figure 26. Rheological index and crossover frequency AAF-1 for various long-term conditioning times and film thicknesses.

Research Approach 39   Yield for Thinner Film AASHTO R 28 uses 10 pans that are 140 mm in diameter. Each pan receives 50.0 g of binder, yielding a total of 500 g of conditioned binder. The film thickness for AASHTO R 28 is approxi- mately 3.18 mm. Table 14 summarizes the yield for a single run of the PAV for various film thicknesses assuming 1.5 g of binder is lost in the pan after scraping. For continuous grading, 65 g of PAV-conditioned binder are needed; therefore, the minimum film thickness for the current pan and rack design is 0.6 mm or about 9.5 g of binder per pan. The pans used in the PAV are not optimized for the vessel. Figure 27 is a schematic of two commercially available PAVs showing the dimensions for the vessel, the current pans and rack, Film Thickness, mm Mass of Binder per Pan, g Yield for 10 Pans and 1.5 g per Pan Binder Loss, g Assumed % Recovery 3.2 50.0 485 97.0 3.0 47.1 456 96.8 2.5 39.3 378 96.2 2.0 31.4 299 95.2 1.5 23.6 221 93.6 1.0 15.7 142 90.4 0.8 12.6 111 88.1 0.7 11.0 95 86.4 0.6 9.4 79 84.0 0.5 7.9 64 81.0 0.4 6.3 48 76.2 0.3 4.7 32 68.1 Table 14. Current PAV yield for various film thicknesses. Figure 27. Schematic of two commercial PAVs.

40 Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging and the maximum diameter pans and maximum rack height based on the wall clearances provided in AASHTO R 28. This figure shows that larger diameter pans and a taller rack can be used in the PAV to increase the yield of conditioned binder. If larger pans are used, they would have to be designed to reduce the potential for warping, which is a significant issue with the current 140 mm diameter pans that are only 0.6 mm thick (Anderson et al. 2016). Figure 28 presents one option for a revised pan design. Its base is 3.18 mm thick with a total height of 6.36 mm; concentric rings minimize the potential for warping. The rings are spaced such that an equal amount of binder placed in each ring produces the same film thickness. The total area for conditioning binder is 190 cm2, a 23 percent increase over the area of the current 140 mm pans. The pans in Figure 28 are shorter than the 9.5 mm pans currently specified in AASHTO R 28. Keeping a 12.5 mm spacing between the pans and using all available height, 12 pans can be stacked in the PAV from Vendor 1, while 22 can be stacked in the PAV from Vendor 2. Table 15 presents the yield for one run of the Vendor 1 and Vendor 2 PAVs at Figure 28. Example of PAV pan design for thinner film thicknesses. Film Thickness, mm Mass of Binder per Pan, g Assumed % Recovery Yield for 12 Pans, g Yield for 22 Pans, g 1.0 19.4 90.4 210 386 0.9 17.4 89.4 187 342 0.8 15.5 88.1 164 300 0.7 13.6 86.4 141 259 0.6 11.6 84.0 117 214 0.5 9.7 81.0 94 173 0.4 7.8 76.2 71 131 0.3 5.8 68.1 47 87 Table 15. PAV yield for revised pan and rack for various film thicknesses.

Research Approach 41   various film thicknesses using the revised pan design and the percent recovery from Table 14. With Vendor 1’s PAV using the modified rack and pan, two binders for continuous grading can be conditioned during one PAV run down to a film thickness of 0.6 mm. For the taller PAV from Vendor 2 with the modified rack and pan, four binders for continuous grading can be conditioned during one run of the PAV down to a film thickness of 0.7 mm. Tolerances for Thinner Films The data from the preliminary film thickness experiment was used to estimate film thickness tolerances for short- and long-term conditioning. The procedure that was used is illustrated using the data shown in Figure 29 for short-term conditioning. Figure 29 shows measured G*/sind data as a function of film thickness for the four film thicknesses included in the exper- iment. It also shows equations fitted through the data for film thicknesses from 0.8 mm to 1.6 mm, the region of interest for thinner film short- and long-term conditioning procedures. The slope of the curves in Figure 29 at any point is the sensitivity of the measured specifica- tion property to changes in film thickness. Clearly, the sensitivity increases with decreasing film thickness, and binder AAF-1 is more sensitive than binder AAC-1. The slope at any point is the derivative of the curves shown in Figure 29 with respect to film thickness. Table 16 summarizes the slopes over the range of film thicknesses from 0.8 mm to 1.6 mm and normal- izes them to the measured G*/sind. An estimate of the allowable film thickness variation is also included in Table 16. This was calculated by dividing the within-laboratory coefficient of variation of 3.2 percent tabulated in AASHTO T 315 for short-term conditioned residue by the normalized slope. The allowable film thickness variation was then converted to an allow- able mass variation by multiplying the allowable film thickness variation by the pan area and the density of the binder. This estimate is also included in Table 16 for the pan shown in Figure 28, which has two areas of 95 cm2 and assumes a specific gravity of the binder of 1.02. y = 2.6982x-0.854 R² = 0.985 y = 2.9273x-0.501 R² = 0.9785 1.00 1.50 2.00 2.50 3.00 3.50 0.00 0.50 1.00 1.50 2.00 2.50 3.00 3.50 G* /s in δ at G ra de Te m pe ra tu re , k Pa Film Thickness, mm AAC-1 AAF-1 Figure 29. Effect of film thickness on properties of short-term conditioned residue.

42 Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging Based on Table 16, a tolerance of about 0.2 g seems appropriate for film thicknesses between 0.8 mm and 1.0 mm. This tolerance is somewhat less than the 0.5 g specified in AASHTO R 28, AASHTO T 179, and AASHTO T 240 but easily obtainable in binder testing laboratories. The tolerance on the levelness of the pan is more difficult to evaluate. For linear changes in G*/sind with film thickness, the critical parameter to control is the average film thickness, which is controlled by the mass of binder added to the pan. Referring to Figure 29, the linear assumption seems reasonable over about ±0.1 mm from the average thickness; therefore, it is reasonable to base the levelness tolerance on what is practical to achieve in the laboratory. Table 17 summarizes the accuracy of various levels and the resulting difference in film thickness for a 170 mm diameter pan, which is the largest pan that can be placed in the PAV. AASHTO R 28 specifies a maximum difference in film thickness of 0.5 mm over the 140 mm pan diameter. For AASHTO R 28 a quality bull’s-eye level can be used to ensure compliance. For a 140 mm diameter pan set with a quality bull’s-eye level, the difference in film thickness across the pan will be less than 0.14 mm. For film thickness of 1 mm or less, it is probably best to set the equipment using a precision machinist level having an accuracy of 0.00042 mm/mm. This will result in a difference in film thickness of 0.068 mm across a 170 mm pan, well within the range Binder Film Thickness, mm Slope, kPa/mm G*/sinδ, kPa Normalized Slope, %/mm Allowable Variation Film Thickness, mm Mass, g AAC-1 0.8 −2.050 3.27 −62.6 0.05 0.50 0.9 −1.718 3.09 −55.7 0.06 0.56 1.0 −1.466 2.93 −50.1 0.06 0.62 1.1 −1.271 2.79 −45.5 0.07 0.68 1.2 −1.115 2.67 −41.8 0.08 0.74 1.3 −0.989 2.57 −38.5 0.08 0.80 1.4 −0.885 2.47 −35.8 0.09 0.87 1.5 −0.798 2.39 −33.4 0.10 0.93 1.6 −0.724 2.31 −31.3 0.10 0.99 AAF-1 0.8 −3.485 3.26 −106.8 0.03 0.29 0.9 −2.801 2.95 −94.9 0.03 0.33 1.0 −2.304 2.70 −85.4 0.04 0.36 1.1 −1.931 2.49 −77.6 0.04 0.40 1.2 −1.643 2.31 −71.2 0.04 0.44 1.3 −1.416 2.16 −65.7 0.05 0.47 1.4 −1.235 2.02 −61.0 0.05 0.51 1.5 −1.086 1.91 −56.9 0.06 0.54 1.6 −0.964 1.81 −53.4 0.06 0.58 Table 16. Calculation of film thickness tolerance for short-term conditioning. Type of Level Accuracy Difference in Thickness Across 170 mm Diameter Pan, mmmm/mm degree Bull’s-Eye Level 0.00100 0.057 0.1700 Precision Machinist Level 0.00042 0.023 0.0680 High-Precision Machinist Level 0.00004 0.002 0.0068 Table 17. Accuracy of various levels and difference in film thickness for 170 mm diameter pan.

Research Approach 43   where the change in G*/sind is linear with thickness. Neither AASHTO T 179 nor the draft AASHTO standard for the USAT test specifies a tolerance for levelness. The USAT method specifies using a 10 in level to level the racks in the oven. Standard levels have an accuracy similar to the bull’s-eye level in Table 17, resulting in a difference in film thickness up to about 0.1 mm over the 101.6 mm length of the USAT plate. AASHTO T 240 specifies a levelness tolerance for the RTFOT carriage of ±1.0 degree. This tolerance provides up to a 2.44 mm difference in film thickness over the 140 mm-long RTFOT bottle. The analysis presented above was repeated for the properties of long-term conditioned residue, including G* • sind at 25°C, 10 rad/sec, creep stiffness, and m-value. The results are summarized in Table 18 for a film thickness of 0.8 mm. Table 18 shows that testing on the short-term residue is most sensitive to film thickness variations. Since the tolerances outlined above for the short-term residue can be achieved without major modification to equipment, they are reasonable for both short- and long-term conditioning using film thickness as low as 0.8 mm. Selection of Procedures for Further Development The evaluation presented above concluded that there are two options for an improved short- term laboratory aging procedure for asphalt binders: (1) adopt modifications to AASHTO T 240 similar to those implemented in the UK Ageing Profile Test or (2) adopt thin-film condi- tioning similar to the USAT, but use a thicker film to provide more binder for current test methods. The major unanswered question with both that required further investigation through experimentation is whether they treat neat and modified binders of various consistencies equally. Published data for both procedures are limited to comparisons with AASHTO T 240, the assumed flawed procedure that motivated the development of the alternate procedures. The improved long-term aging procedure recommended from the evaluation presented above is to reduce the film thickness in AASHTO R 28 to increase the amount of aging simu- lated by the test. The preliminary film thickness experiment conducted during the evaluation phase showed that 20 hours of conditioning at a film thickness of 0.8 mm to 0.9 mm produces rheological properties for long-term conditioned residue that are similar to those obtained from 40 hours of conditioning using the standard film thickness of 3.18 mm. The major unanswered question for long-term conditioning that required further investigation through experimenta- tion is what time in service is simulated by various PAV operating conditions for pavements constructed in different climates. Property Binder Slope Property at 0.8 mm Film Thickness %/mm Allowable Variation Film Thickness, mm Mass, g G*/sinδ AAC-1 −2.050 kPa/mm 3.27 kPa −62.6 0.05 0.50 AAF-1 −3.485 kPa/mm 3.26 kPa −106.8 0.03 0.29 G·sinδ AAC-1 −526.4 kPa/mm 3143.0 kPa −16.8 0.29 2.83 AAF-1 −1705.1 kPa/mm 7156.0 kPa −16.4 0.30 2.90 Creep Stiffness AAC-1 −9.84 MPa/mm 91.5 MPa −10.8 0.23 2.25 AAF-1 −24.43 MPa/mm 80.4 MPa −30.4 0.08 0.80 m- value AAC-1 0.010400 /mm 0.362 2.9 0.35 3.37 AAF-1 0.024100 /mm 0.373 6.5 0.15 1.50 Table 18. Summary of film thickness tolerance calculations for properties measured on short- and long-term conditioned residue.

44 Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging Short-Term Conditioning Selection Experiment Objective The evaluation of short-term conditioning procedures identified two options for improv- ing short-term laboratory conditioning of asphalt binders: (1) adopt the container and mixing screw modifications to AASHTO T 240 that were implemented in the UK Ageing Profile Test (Hill et al. 2009), or (2) adopt thin-film conditioning similar to the USAT (Farrar et al. 2014) but using a thicker film to provide more binder for current specification tests. During outreach to industry, a third option was identified: (3) modify AASHTO T 240 to include a heating step where binder in the containers is heated to the conditioning temperature before hand rotating the container to coat it just before loading the container in the Rolling Thin Film Oven (RTFO). The objective of this short-term conditioning selection experiment was to select one of these candidates for calibration during NCHRP Project 09-61. The primary concerns with AASHTO T 240 are: (1) binders of different consistencies could be treated differently because the film thickness and its renewal vary with the consistency of the binder, and (2) some heavily modified binders and GTR-modified binders crawl out of the bottle during conditioning. The candidate approaches address these concerns in different ways. The mixing screws used in the UK Ageing Profile Test are intended to move binder into the container to keep modified binders from crawling out of the container and to improve exposure to air over a range of binder consistencies. In a thin-film test like the USAT, a film of specified thickness is prepared, and it is conditioned in a pan in an oven without movement. Adding a heating step to AASHTO T 240 helps produce and maintain a more uniform film in the RTFOT container. Data in the literature for the UK mixing screw and the USAT show they can reproduce the conditioning that occurs in AASHTO T 240, the assumed flawed test that they are intended to replace. However, no data compare the conditioning from these pro- cedures and AASHTO T 240 modified with a heating step to the aging that occurs in asphalt mixtures during construction. The short-term conditioning selection experiment described below was designed to compare conditioning from the candidate approaches to AASHTO T 240 and oven-conditioned loose mix. The loose mix conditioning procedures that were used were those recommended in NCHRP Project 09-52, which were based on a field evaluation of numerous factors affecting the short-term aging characteristics of asphalt mixtures (Newcomb et al. 2015). AASHTO T 240 was included in the experiment to gather reference data for the current standard procedure; however, the final selection of the improved short-term procedure was based on comparisons to oven-conditioned loose mix. Experimental Design The short-term conditioning selection experiment was designed as a paired difference experiment where a set of binders with a wide range of consistencies were subjected to different short-term conditioning procedures. Paired difference experiments are commonly used to compare the effectiveness of different treatments. The experiments subject the same materials to different treatments and compare the results (Ott 1977). For this application, several binders were short-term conditioned using six procedures: 1. AASHTO T 240, 2. Modified AASHTO T 240 with UK mixing screw, 3. Modified AASHTO T 240 with NCHRP Project 09-61 mixing screw, 4. Static 0.8 mm film, 5. Modified AASHTO T 240 with heating step, and 6. Loose mix conditioned per NCHRP Project 09-52 recommendations.

Research Approach 45   The AASHTO M 320 parameter, G*/sind, was measured on residue from the five binder conditioning procedures and recovered binder from the loose mix conditioning. If two procedures treat the binders the same, differences in G*/sind between the two procedures— averaged over all binders tested—will not be significantly different from zero. The difference for an individual binder may be positive or negative, but the average difference over several binders should be zero. A t-test is used to assess the statistical significance of the average difference, as summarized below. Null hypothesis: µA − µB = 0 Alternative hypothesis: µA − µB > 0 or µA − µB < 0 (as appropriate) Test statistic: t d s n d =     Rejection region: Reject the null hypothesis and accept the alternative hypothesis if t > tα for n − 1 degrees of freedom. where µA = population mean for short-term conditioning treatment A, µB = population mean for short-term conditioning treatment B, t = test statistic, tα = critical value of the test statistic, d– = average of the differences in G*/sind between treatment A and treatment B, sd = standard deviation of the G*/sind differences, and n = number of binders compared. A relatively modest experiment using two neat binders, two polymer-modified binders, and a GTR-modified binder was initially planned and executed. This experiment included conditioning to simulate HMA and WMA production. Due to unexpected results from the initial experiment, the experiment was expanded to include more binders, the heating step modification of AASHTO T 240, and replication of loose mix conditioning and binder recovery. Table 19 presents the final design of the experiment for HMA conditions. Table 20 presents the final design of the experiment for WMA conditions. The GTR-modified binder was removed from the final experiment because this binder could not be properly recovered, and the basis for the final selection of an improved procedure was equivalency with loose mix conditioning. Conditioning Procedures The following procedures were used for the binder and loose mix conditioning in the short- term conditioning selection experiment. AASHTO T 240 This conditioning was conducted in accordance with AASHTO T 240, including mass change measurements. After pouring each sample, the container was rotated to precoat the container. UK and NCHRP Project 09-61 Mixing Screws These procedures generally followed the short-term conditioning specified for the UK Ageing Profile Test (Hill et al. 2009) with the following two exceptions: (1) the mass of binder

46 Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging was 35 g as specified in AASHTO T 240 rather than the 19 g specified in the UK Ageing Profile Test, and (2) the duration of the conditioning was 85 minutes rather than the 45 minutes speci- fied the UK Ageing Profile Test. Both procedures used the PTFE-coated aluminum containers from the UK Ageing Profile Test, shown in Figure 30. Figure 31 shows a photograph of the UK mixing screw that is designed to move binder back into the container as the RTFO carousel rotates. Figure 32 shows a photograph of the NCHRP Project 09-61 mixing screw that is designed to move the binder to the middle of the container where the solid section of the mixer produced a uniform film thickness. The primary difference between AASHTO T 240 conditioning and conditioning using the UK and NCHRP Project 09-61 mixers is the containers Binder Binder Conditioning 85 Minutes at 163°C Recovered from Short-Term Oven- Aged Loose Mix 2 Hours at 135°C AASHTO T 240 UK Mixing Screw NCHRP Project 09-61 Mixing Screw Static 0.8 mm Film AASHTO T 240 Hot Rep 1 Rep 2 Rep 3 Neat PG 52-34 X X X X X X X X Terpolymer PG 64-34 X X X X X X X X Neat PG 64-22 X X X X X X X X SBS PG 76-22 X X X X X X X X SBS PG 64-34 X X X X X X X X SBS PG 76-28 X X X X X X X X PG 64-22 with 3% Latex X X X X X X X X SBS PG 88-22 X X X X X X X X SBS = styrene-butadiene-styrene Table 19. Final design of the short-term selection experiment for HMA mix conditions. Binder Binder Conditioning 85 Minutes at 163°C Recovered from Short-Term Oven- Aged Loose Mix 2 Hours at 116°C AASHTO T 240 UK Mixing Screw NCHRP Project 09-61 Mixing Screw Static 0.8 mm Film AASHTO T 240 Hot Rep 1 Rep 2 Rep 3 Neat PG 52-34 X X X X X X X X Terpolymer PG X X X X X X X X Neat PG 64-22 X X X X X X X X SBS PG 76-22 X X X X X X X X SBS PG 64-34 X X X X X X X X SBS PG 76-28 X X X X X X X X PG 64-22 with 3% Latex X X X X X X X X SBS PG 88-22 X X X X X X X X Table 20. Final design of the short-term selection experiment for WMA conditions.

Research Approach 47   Figure 30. Prototype PTFE-coated aluminum RTFO container and lid. Figure 31. Stainless steel UK mixer. Figure 32. Stainless steel NCHRP Project 09-61 mixer.

48 Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging and the mixing screws were preheated in an oven set at 163°C for 90 minutes before weighing the binder into the container. After weighing the binder into the container, the container was rotated to precoat it. The container was not cooled before loading it into the RTFO carousel. The mass change measurements were, therefore, made on containers without cooling them to room temperature. Static Thin Film In this procedure, 12.5 g of binder was weighed into PAV pans and conditioned in a forced draft oven for 85 minutes. The pans were placed in a PAV rack that was leveled with a precision level. Leveling the rack was very difficult because the shelves and the bottom of the oven tended to warp when heated. AASHTO T 240 Modified with a Heating Step For this modification of AASHTO T 240, the binder was poured into containers, and the containers were cooled on a rack for 60 minutes. After 60 minutes, the initial mass for the mass change measurements was made, and the containers were transferred to an oven set to the conditioning temperature and heated in the vertical position for 15 minutes. When transferring the heated containers to the RTFO, the container was rolled once to coat it. The remainder of the conditioning and second mass change measurement was as specified in AASHTO T 240. Loose Mix Oven Conditioning For this procedure, asphalt concrete mixtures were prepared as per AASHTO R 35, Standard Practice for Superpave Volumetric Design for Asphalt Mixtures, then conditioned for 2 hours in a forced draft oven following the procedures recommended in NCHRP Project 09-52, that is, 2 hours at 135°C for HMA and 2 hours at 116°C for WMA. The binder from the mixtures was then extracted and recovered in accordance with ASTM D7906. The recovery procedure was not successful for the GTR-modified binder mixture. Particles of GTR were evident in the aggregate after extraction. Materials Eight binders were included in the short-term conditioning selection experiment. Table 21 presents viscosity and DSR properties for the original binders. The viscosity measurements Binder AASHTO T 315 Properties AASHTO T 316 Viscosity, Pa·s Test Temp, °C G*, kPa Phase Angle, ° G*/sinδ, kPa 135°C 163°C Neat PG 52-34 52 1.39 85.8 1.39 0.22 0.07 Terpolymer PG 64-34 64 1.04 67.6 1.12 0.73 0.21 Neat PG 64-22 64 1.65 86.9 1.65 0.43 0.13 SBS PG 76-22 76 1.14 75.5 1.18 1.11 0.30 GTR-Modified 82 1.17 79.5 1.19 NT* 0.99 SBS PG 64-34 64 1.99 63.4 2.23 1.21 0.37 SBS PG 76-28 76 1.46 57.9 1.72 2.36 0.54 PG 64-22 with 3% Latex 76 1.09 79.8 1.10 3.06 0.74 SBS PG 88-22 88 1.42 51.5 1.81 7.65 1.64 * GTR-modified binder not conditioned at 135°C NT = Not testable Table 21. Properties of original binders.

Research Approach 49   were made at the temperatures used in the binder conditioning procedures to simulate WMA and HMA conditions. The basis for selection of an improved short-term binder conditioning procedure was the properties of binder recovered from loose mixtures that were short-term conditioned following the recommendations from NCHRP Project 09-52, which are 2 hours in a forced draft oven at 135°C for HMA and 2 hours in a forced draft oven at 116°C for WMA. Properties of the mixture used for the short-term oven conditioning are presented in Table 22. The coarse aggregates in this mixture were diabase from Northern Virginia (VA). The fine aggregates were a combination of diabase and natural sand. The natural sand was 10 percent of the aggregate blend. Results and Analysis Relative Aging of Binder Conditioning Procedures AASHTO M 320 high-temperature rheological properties (after short-term conditioning with AASHTO T 240), and the various candidate procedures are presented in Table 23. Table 23 presents laboratory conditioning at 163°C to simulate HMA production; Table 24 presents laboratory conditioning at 135°C to simulate WMA production. These data were analyzed graphically and statistically to determine if any of the candidate procedures conditioned the binders differently than AASHTO T 240 and if so, the nature of the difference. Recall, one of the concerns with AASHTO T 240 is that binders of different consistencies could be treated differently because the film thickness and its renewal vary with the consistency of the binder. Figures 33 through 36 present plots of the AASHTO T 240 Relative Aging Index, defined by Equation 5, as a function of the viscosity of the binder at the conditioning temperature. Data are presented for both conditioning temperatures. The same scale is used in all figures. The Design Gyrations 75 Aggregate Sources Coarse VA Diabase Fine VA Diabase and Natural Sand Gradation, % Passing Sieve Size, mm 12.500 100 9.500 94 4.750 54 2.360 38 1.180 28 0.600 21 0.300 12 0.150 8 0.075 5.2 Aggregate Properties FAA 46.1 CAA 100/99 Flat & Elongated 2.7 Sand Equivalent 59.4 Binder Content, weight % 5.5 Effective Binder Content, weight % 4.8 Air Voids, volume % 3.9 Voids in Mineral Aggregate, volume % 15.9 Voids Filled with Asphalt, % 75.5 Dust to Effective Asphalt Ratio 1.1 Table 22. Properties of asphalt mixture used in the short-term selection experiment.

Binder Test Temp, °C AASHTO T 240 UK Mixing Screw NCHRP Project 09-61Mixing Screw Static 0.8 mm Film AASHTO T 240 Hot G* kPa δ ° G*/sinδ kPa G* kPa δ ° G*/sinδ kPa G* kPa δ ° G*/sinδ kPa G* kPa δ ° G*/sinδ kPa G* kPa δ ° G*/sinδ kPa Neat PG 52-34 52 3.26 81.6 3.30 3.40 81.4 3.44 3.50 81.2 3.54 3.38 81.7 3.42 3.53 81.2 3.57 Terpolymer PG 64-34 64 2.37 61.5 2.70 2.40 61.4 2.73 2.29 61.7 2.60 2.29 62.3 2.59 2.45 61.2 2.79 Neat PG 64-22 64 3.75 83.6 3.77 3.90 83.4 3.93 3.93 83.4 3.96 3.78 83.6 3.80 3.99 83.1 4.02 SBS PG 76-22 76 2.54 68.2 2.74 2.67 67.4 2.89 2.59 67.8 2.80 2.14 70.8 2.27 2.57 68.0 2.77 SBS PG 64-34 64 3.84 61.3 4.38 4.08 60.9 4.67 3.94 61.1 4.50 3.39 61.8 3.85 4.09 60.9 4.68 SBS PG 76-28 76 2.75 54.9 3.36 2.85 54.8 3.49 2.77 54.9 3.39 2.28 56.5 2.73 2.99 54.9 3.65 GTR- Modified 82 2.65 68.5 2.85 3.30 66.0 3.61 2.99 66.1 3.27 2.18 72.2 2.29 NT NT NT PG 64-22 with 3% Latex 76 2.12 76.8 2.18 2.62 77.0 2.69 2.20 77.7 2.26 2.04 79.2 2.07 2.12 77.1 2.17 SBS PG 88-22 88 2.13 49.4 2.81 3.85 45.2 5.43 2.86 47.2 3.90 2.47 49.5 3.11 2.30 49.5 3.02 Table 23. Laboratory-conditioned binder rheological properties for HMA conditions.

Binder Test Temp, °C AASHTO T 240 UK Mixing Screw NCHRP Project 09-61Mixing Screw Static 0.8 mm Film AASHTO T 240 Hot G* kPa δ ° G*/sinδ kPa G* kPa δ ° G*/sinδ kPa G* kPa δ ° G*/sinδ kPa G* kPa δ ° G*/sinδ kPa G* kPa δ ° G*/sinδ kPa Neat PG 52-34 52 2.23 83.7 2.25 2.44 83.2 2.46 2.44 83.3 2.46 2.03 84.3 2.04 2.13 83.9 2.14 Terpolymer PG 64-34 64 1.42 64.7 1.57 1.72 63.5 1.92 1.63 63.8 1.82 1.41 64.7 1.56 1.40 66.1 1.54 Neat PG 64-22 64 2.59 85.3 2.60 2.78 85.0 2.80 2.70 85.1 2.71 2.32 85.8 2.32 2.76 85.0 2.77 SBS PG 76-22 76 1.57 73.1 1.64 2.00 70.9 2.12 1.96 71.3 2.07 1.64 73.2 1.71 1.71 72.9 1.79 SBS PG 64-34 64 2.66 62.4 3.00 3.34 61.2 3.81 3.12 61.6 3.54 2.59 62.4 2.93 2.94 62.0 3.33 SBS PG 76-28 76 1.86 56.4 2.24 2.31 55.4 2.80 2.25 55.4 2.73 1.79 56.6 2.14 2.01 56.2 2.41 PG 64-22 with 3% Latex 76 1.47 78.1 1.50 2.10 76.0 2.16 1.80 76.2 1.86 1.42 78.4 1.45 1.62 77.9 1.66 SBS PG 88-22 88 1.76 50.1 2.29 3.33 46.3 4.61 2.36 48.3 3.16 1.85 50.1 2.41 1.90 49.7 2.49 Table 24. Laboratory-conditioned binder rheological properties for WMA conditions.

52 Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging Figure 33. AASHTO T 240 Relative Aging Index for the UK mixing screw. Figure 34. AASHTO T 240 Relative Aging Index for the NCHRP Project 09-61 mixing screw.

Research Approach 53   Figure 35. AASHTO T 240 Relative Aging Index for AASHTO T 240 with a heating step. Figure 36. AASHTO T 240 Relative Aging Index for the static 0.8 mm film.

54 Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging AASHTO T 240 Relative Aging Index is greater than 1.0 when the laboratory conditioning procedure ages the binder more than AASHTO T 240; it is less than 1.0 when the laboratory conditioning procedure ages the binder less than AASHTO T 240. * * * * * * (5)240 240 240 240 RAI AI AI G sin G sin G sin G sin G sin G sin T i T i original T original i T ( ) ( ) ( ) ( ) ( ) ( ) = = δ δ δ δ             = δ δ where RAIT240 = AASHTO T 240 Relative Aging Index, (G*/sind)i = AASHTO M 320 specification parameter G*/sind for candidate i, (G*/sind)T240 = AASHTO M 320 specification parameter G*/sind for AASHTO T 240, and (G*/sind)Original = original binder AASHTO M 320 specification parameter G*/sind. Figures 33 and 34 show increases in the AASHTO T 240 Relative Aging Index with increasing viscosity for the two mixing screw procedures. The larger values for the UK mixing screw compared to the NCHRP Project 09-61 mixing screw indicates the UK mixing screw design provides more effective mixing of the binder with air for highly viscous binders. When the viscosity is about 0.55 Pa•s or less, both mixing screws have little effect, increasing the aging index only about 4 percent above that for AASHTO T 240. This increase is probably the result of using preheated containers and binder in the mixing screw procedures. Figure 35 shows the AASHTO T 240 Relative Aging Index when a heating step was added to AASHTO T 240. The heating step increased the aging index approximately 6 percent above that for AASHTO T 240. The increase does not appear to be a function of the viscosity of the binder at the conditioning temperature. Similarly, Figure 36 shows the AASHTO T 240 Relative Aging Index for the static 0.8 mm film does not appear to be dependent on the viscosity of the binder at the conditioning temperature. On average, the static 0.8 mm film resulted in about a 5 percent decrease in the aging index compared to AASHTO T 240. Linear regression analyses were performed to confirm the significance of the trends shown in Figures 33 through 36. The results are summarized in Table 25. The regression analyses confirm the viscosity effect on the AASHTO T 240 Relative Aging Index for the UK and NCHRP Project 09-61 mixing screw procedures when the viscosity of the binder at the condi- tioning temperature is greater than 0.55 Pa•s. The analyses also confirm that the viscosity of the binder at the conditioning temperature does not affect the AASHTO T 240 Relative Aging Index for AASHTO T 240 modified with a heating step and the static 0.8 mm film. Method Viscosity Range, Pa‧s Observations Slope RAI T240/log(viscosity Pa‧s) p-value Slope UK Mixing Screw 0.07 to 0.54 8 0.5495 Not Significant 0.55 to 7.65 9 0.0329 0.664 NCHRP Project 09-61 Mixing Screw 0.07 to 0.54 8 0.3315 Not Significant 0.55 to 7.65 9 0.0284 0.240 AASHTO T 240 with Heating Step 0.07 to 7.65 16 0.1646 Not Significant Static 0.8 mm Film 0.07 to 7.65 17 0.3910 Not Significant Table 25. Summary of regression analysis of the AASHTO T 240 Relative Aging Index.

Research Approach 55   Comparison of Binder Conditioning Procedures to Loose Mix Conditioning The analysis presented in the preceding section showed the mixing screw procedures are effective at increasing the exposure of more viscous binders to air during short-term condi- tioning in the RTFOT. The analysis also showed the UK mixing screw design to be more effective than the NCHRP Project 09-61 design. Finally, the analysis showed that heating the binder and the containers to the conditioning temperature before short-term conditioning increased the aging index by approximately 6 percent compared to AASHTO T 240. This section compares G*/sind values measured on residue from the various short-term binder condition- ing procedures to G*/sind values measured on binder recovered from short-term conditioned loose mixtures. The mixtures were short-term conditioned in a forced draft oven following the procedures recommended in NCHRP Project 09-52 2 hours at 135°C to simulate HMA production, and 2 hours at 116°C to simulate WMA production (Newcomb et al. 2015). Binders from the short-term conditioned loose mixtures were extracted and recovered following ASTM D7906. Since loose mix conditioning and binder recovery are likely to be more variable than conditioning asphalt binder, three independent mixture samples were conditioned, recovered, and tested. Tables 26 and 27 present the results of the rheological testing for recovered binder simulating HMA and WMA production, respectively. Various graphical and statistical analyses were performed as discussed below to compare the short-term binder conditioning procedures to short-term conditioning of loose mixtures. Correlation of Aging Index Between Short-Term Binder and Short-Term Mix Conditioning. Figures 37 through 41 compare aging indices from the various short-term laboratory binder conditioning procedures with aging indices from binder recovered from short-term conditioned loose mixtures. These figures include the data, simulated HMA and WMA aging, and the line of equality. The error bars are 90 percent confidence intervals for the recovered binder aging index. For conditioning that simulates HMA, the data for all binder conditioning procedures, except the static 0.8 mm film, generally clusters around the line of equality; however, the spread of the data around the line of equality is quite large. As shown by the error bars, some of this spread is due to variability in the mixture conditioning and recovery process, but a substantial portion is due to differences in how binders age during mixture conditioning compared to binder conditioning. The static 0.8 mm film data for HMA conditioning generally plots below the line of equality. For conditioning that simulates WMA, the data for AASHTO T 240 (Figure 37), AASHTO T 240 with heating (Figure 38), and the static 0.8 mm film (Figure 41) generally plot below the line of equality, indicating these procedures age the binder less than mixture conditioning. The WMA data for the two mixing screw procedures (Figures 39 and 40) generally cluster around the line of equality. Considering both HMA and WMA conditioning, it appears that the NCHRP Project 09-61 mixing screw best reproduces the aging that occurs in oven- conditioned loose mixes. Binder Viscosity Effect. One of the concerns with AASHTO T 240 is binders having different consistencies could be treated differently because the film thickness and its renewal vary with the consistency of the binder. To compare binder and mixture conditioning, a loose mix relative aging index was calculated by dividing the aging index from the laboratory binder procedures for each binder by the corresponding aging index from loose mix conditioning. A loose mix relative aging index greater than 1 indicates that the laboratory binder procedure ages the binder more than loose mix conditioning, while a loose mix relative aging index less than 1 indicates the binder procedure ages the binder less than loose mix conditioning. Figures 42 through 46 present plots of the loose mix relative aging index as a function of the viscosity of the binder at the conditioning temperature for the various laboratory binder condi- tioning procedures. The same scale is used in each of these figures to compare the various procedures. The loose mix relative aging indices generally range from about 0.7 to 1.3. The UK mixing screw procedure produced values approaching 2.0 for the PG 82-22 binder.

Binder Test Temp, °C Replicate 1 Replicate 2 Replicate 3 Average Standard Deviation G* kPa δ ° G*/sinδ kPa G* kPa δ ° G*/sinδ kPa G* kPa δ ° G*/sinδ kPa G* kPa δ ° G*/sinδ kPa G* kPa δ ° G*/sinδ kPa Neat PG 52-34 52 3.24 83.2 3.26 3.36 83.3 3.38 3.15 83.5 3.17 3.33 83.3 3.27 0.08 0.15 0.11 Terpolymer PG 64-34 64 1.95 65.3 2.14 2.02 64.7 2.24 2.13 64.6 2.36 2.07 64.9 2.25 0.15 0.38 0.11 Neat PG 64-22 64 4.51 82.5 4.47 5.06 82.0 5.11 4.45 82.5 4.49 4.89 82.3 4.69 0.36 0.29 0.36 SBS PG 76-22 76 2.41 70.0 2.56 2.23 70.1 2.37 2.39 70.0 2.54 2.34 70.0 2.49 0.10 0.06 0.10 SBS PG 64-34 64 4.52 61.2 5.16 4.03 62.1 4.56 4.05 61.7 4.60 4.37 61.7 4.77 0.36 0.45 0.34 SBS PG 76-28 76 2.26 57.9 2.67 2.17 58.5 2.54 2.32 58.4 2.72 2.32 58.3 2.64 0.21 0.32 0.09 PG 64-22 with 3% Latex 76 2.65 71.7 2.79 2.71 77.4 2.78 2.80 76.9 2.87 2.71 75.3 2.81 0.08 3.16 0.05 SBS PG 88-22 88 2.33 52.5 2.93 2.48 51.8 3.15 2.88 51.5 3.43 2.65 51.9 3.17 0.42 0.51 0.25 Table 26. Recovered rheological properties for HMA conditions. Binder Test Temp, °C Replicate 1 Replicate 2 Replicate 3 Average Standard Deviation G* kPa δ ° G*/sinδ kPa G* kPa δ ° G*/sinδ kPa G* kPa δ ° G*/sinδ kPa G* kPa δ ° G*/sinδ kPa G* kPa δ ° G*/sinδ kPa Neat PG 52-34 52 2.33 83.4 2.35 2.61 84.1 2.63 2.38 84.6 2.39 2.52 84.0 2.46 0.17 0.60 0.15 Terpolymer PG 64-34 64 1.64 66.5 1.79 1.68 65.9 1.84 1.82 65.7 2.00 1.72 66.0 1.88 0.10 0.42 0.11 Neat PG 64-22 64 3.17 84.0 3.18 3.55 83.5 3.57 3.62 83.4 3.65 3.43 83.6 3.47 0.23 0.32 0.25 SBS PG 76-22 76 1.92 71.5 2.02 1.93 70.7 2.04 1.90 71.1 2.00 1.96 71.1 2.02 0.06 0.40 0.02 SBS PG 64-34 64 3.38 61.8 3.84 3.32 62.0 3.76 3.48 61.6 3.96 3.49 61.8 3.85 0.25 0.20 0.10 SBS PG 76-28 76 2.26 57.9 2.67 2.17 58.5 2.54 2.32 58.4 2.72 2.32 58.3 2.64 0.21 0.32 0.09 PG 64-22 with 3% Latex 76 1.80 78.0 1.84 2.11 73.4 2.20 2.16 67.3 2.35 2.04 72.9 2.13 0.22 5.37 0.26 SBS PG 88-22 88 1.56 53.4 1.94 2.08 51.4 2.66 2.06 51.5 2.64 2.10 52.1 2.41 0.56 1.13 0.41 Table 27. Recovered rheological properties for WMA conditions.

Research Approach 57   1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.6 2.8 3.0 3.2 3.4 1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.6 2.8 3.0 3.2 3.4 AA SH TO T 2 40 A gi ng In de x Recovered Binder Aging Index Simulated HMA Aging Simulated WMA Aging Figure 37. Comparison of aging indices from AASHTO T 240 conditioning and loose mix conditioning. 1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.6 2.8 3.0 3.2 3.4 1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.6 2.8 3.0 3.2 3.4 AA SH TO T 2 40 W ith H ea tin g Ag in g In de x Recovered Binder Aging Index Simulated HMA Aging Simulated WMA Aging Figure 38. Comparison of aging indices from AASHTO T 240 with heating conditioning and loose mix conditioning.

58 Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging 1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.6 2.8 3.0 3.2 3.4 1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.6 2.8 3.0 3.2 3.4 U K M ix in g Sc re w A gi ng In de x Recovered Binder Aging Index Simulated HMA Aging Simulated WMA Aging Figure 39. Comparison of aging indices from UK mixing screw conditioning and loose mix conditioning. Figure 40. Comparison of aging indices from NCHRP Project 09-61 mixing screw conditioning and loose mix conditioning.

Research Approach 59   1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.6 2.8 3.0 3.2 3.4 1.0 1.2 1.4 1.6 1.8 2.0 2.2 2.4 2.6 2.8 3.0 3.2 3.4 St ati c 0. 8 m m F ilm A gi ng In de x Recovered Binder Aging Index Simulated HMA Aging Simulated WMA Aging Figure 41. Comparison of aging indices from static 0.8 mm film conditioning and loose mix conditioning. Figure 42. Loose mix relative aging index for AASHTO T 240.

60 Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging Figure 43. Loose mix relative aging index for AASHTO T 240 with heating. Figure 44. Loose mix relative aging index for the UK mixing screw.

Research Approach 61   Figure 45. Loose mix relative aging index for the NCHRP Project 09-61 mixing screw. Figure 46. Loose mix relative aging index for the static 0.8 mm film.

62 Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging Regression analyses were conducted to determine if the loose mix relative aging index values for any of the laboratory conditioning procedures were affected by the viscosity of the binder at the conditioning temperature. The analysis was conducted separately using data for simulated HMA aging and simulated WMA aging. The results, summarized in Table 28, show the relative loose mix aging index is not affected by the viscosity of the binder at the conditioning temperature, except for simulated WMA aging using the UK mixing screw procedure; p-values less than 0.05 were considered statistically significant. Paired Difference Analysis. The final analysis that was conducted was a paired difference analysis of the aging index values between loose mix conditioning and each of the binder conditioning procedures. The analysis was conducted separately for simulated HMA and WMA conditioning. The results of the paired difference analysis are presented in Table 29 Procedure Data Set Observations Log Viscosity Slope p-value Conclusion AASHTO T 240 HMA 8 0.7271 Not statistically significant WMA 8 0.9198 Not statistically significant AASHTO T 240 with Heating Step HMA 8 0.7084 Not statistically significant WMA 8 0.1859 Not statistically significant UK Mixing Screw HMA 8 0.1015 Not statistically significant WMA 8 0.0476 Statistically significant increase NCHRP Project 09-61 Mixing Screw HMA 8 0.6391 Not statistically significant WMA 8 0.1473 Not statistically significant Static 0.8 mm Film HMA 8 0.6150 Not statistically significant WMA 8 0.3814 Not statistically significant Table 28. Summary of statistical analysis of the effect of viscosity on the loose mix relative aging index. Binder/Statistic Test Temp °C Aging Index Difference AASHTO T 240 AASHTO T 240 with Heating UK Mixing Screw NCHRP Project 09-61 Mixing Screw Static 0.8 mm Film Neat PG 52-34 52 0.02 0.22 0.12 0.19 0.11 Neat PG 64-22 64 −0.56 −0.41 −0.46 −0.44 −0.54 Terpolymer PG 64-34 64 0.50 0.58 0.53 0.41 0.40 SBS PG 76-22 76 0.21 0.24 0.34 0.26 −0.19 SBS PG 64-34 64 −0.18 −0.04 −0.05 −0.12 −0.41 SBS PG 76-28 76 0.42 0.59 0.49 0.43 0.05 PG 64-22 with 3% Latex 76 −0.58 −0.58 −0.11 −0.50 −0.68 SBS PG 88-22 88 −0.20 −0.08 1.25 0.40 −0.03 Average Difference −0.04 0.06 0.26 0.08 −0.16 Standard Deviation of Differences 0.41 0.43 0.52 0.39 0.36 t-statistic −0.31 0.42 1.44 0.58 −1.26 Critical t-statistic, 0.05 Confidence Level 1.89 1.89 1.89 1.89 1.89 Conclusion No significant difference No significant difference No significant difference No significant difference No significant difference Table 29. Summary of paired difference analysis for simulated HMA aging.

Research Approach 63   for simulated HMA aging and Table 30 for simulated WMA aging. The analysis of the simu- lated HMA aging data concluded that there is not a significant difference between loose mix conditioning and any of the laboratory binder conditioning procedures. The average aging index difference for conditioning to simulate HMA ranged from −0.16 for static 0.8 mm film to 0.26 for the UK mixing screw. This range is from −7.5 percent to 12.3 percent of the average aging index of 2.15 from the short-term conditioned loose mixtures. The analysis of the simulated WMA aging concluded that AASHTO T 240, AASHTO T 240 with heating, and the static 0.8 mm film resulted in significantly lower aging indices compared to loose mix conditioning, while there was not a significant difference for the two mixing screw pro- cedures compared to loose mix conditioning. The significantly lower average aging index values for AASHTO T 240, AASHTO T 240 with heating, and the static 0.8 mm film ranged from −20.3 to −13.3 percent of the average aging index of 1.71 from the short-term mixtures. The conditioned binder aging index values that were not significantly different from the mix conditioned values ranged from −2.3 percent to 8.1 percent of the average aging index for the short-term mixtures. Mass Change and Binder Leakage Table 31 summarizes mass change measurements and the conditions resulting in binder leakage during conditioning. The data in Table 31 show binder was often lost during condi- tioning with the mixing screws, which disagrees with the reported performance of mixing screws in the UK Ageing Profile Test (Hill et al. 2009). The likely reason binder was lost during this study is because 35 g of binder were used in this study compared to 19 g specified in the UK Ageing Profile Test. Adding the mixing screws to the container increases the height of the binder in the container, making it easier for binder to be lost through the opening in the container. For AASHTO T 240, the only binder leakage that occurred was when conditioning the PG 88-22 binder at 135°C. For AASHTO T 240 with heating, binder was lost when condi- tioning the PG 64-22 with 3 percent latex at 163°C. No binder was lost when conditioning with the static 0.8 mm film. Binder/Statistic Test Temp, °C Aging Index Difference AASHTO T 240 AASHTO T 240 with Heating UK Mixing Screw NCHRP Project 09-61 Mixing Screw Static 0.8 mm Film Neat PG 52-34 52 −0.15 −0.23 0.00 0.00 −0.30 Neat PG 64-22 64 −0.53 −0.42 −0.40 −0.46 −0.69 Terpolymer PG 64-34 64 −0.20 −0.22 0.12 0.03 −0.21 SBS PG 76-22 76 −0.32 −0.19 0.08 0.04 −0.26 SBS PG 64-34 64 −0.38 −0.23 −0.02 −0.14 −0.41 SBS PG 76-28 76 −0.23 −0.14 0.09 0.05 −0.29 PG 64-22 with 3% Latex 76 −0.57 −0.43 0.03 -0.25 −0.62 SBS PG 88-22 88 −0.07 0.04 1.21 0.41 0.00 Average Difference −0.31 −0.23 0.14 −0.04 −0.35 Standard Deviation of Differences 0.18 0.15 0.47 0.26 0.22 t-statistic −4.85 −4.26 0.85 −0.43 −4.41 Critical t-statistic, 0.05 Confidence Level 1.89 1.89 1.89 1.89 1.89 Conclusion < Loose Mix < Loose Mix No significant difference No significant difference < Loose Mix Table 30. Summary of paired difference analysis for simulated WMA conditions.

64 Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging Figures 47 through 50 compare mass change measurements from the various procedures to the mass change measured in AASHTO T 240. These figures show that although there is general agreement between AASHTO T 240 and the candidate short-term binder condition- ing procedures, changes to specification limits may be required. Mass change measurements are difficult for the mixing screw procedures because the measurements are made using hot containers. The procedures could be modified to make the mass change measurements after cooling the containers to room temperature; however, this will increase the time required to complete the short-term conditioning. PAV Operating Parameters Experiment Objective The primary improvement to AASHTO R 28 identified by the evaluation of long-term conditioning procedures was to increase the level of aging simulated by the PAV procedure. The improvement aimed to better match the properties of binders near the surface of the pavement after approximately 10 years in service. Four options were considered to increase the level of aging simulated by the PAV procedure: 1. Increasing the partial pressure of oxygen by switching to pure oxygen or oxygen-enriched air in the PAV and/or increasing the operating pressure, 2. Increasing the conditioning temperature, 3. Increasing the conditioning time, or 4. Decreasing the film thickness. Temp, °C Binder Viscosity, Pa·s Mass Change, % AASHTO T 240 AASHTO T 240 with Heating Step UK Mixing Screw NCHRP Project 09-61 Mixing Screw Static 0.8 mm Film 163 Neat PG 52-34 0.07 −0.497 −0.615 −0.181 −0.144 −0.547 Terpolymer PG 64-34 0.13 −0.420 −0.985 −0.517 −0.420 −0.896 Neat PG 64-22 0.21 −0.055 −0.206 −0.069 −0.081 −0.208 SBS PG 76-22 0.30 −0.020 −0.118 0.009 0.012 −0.125 SBS PG 64-34 0.37 −0.485 −0.672 leak leak −0.874 SBS PG 76-28 0.54 −0.148 −0.214 leak leak −0.250 PG 64-22 with 3 %-Latex 0.74 −0.004 leak leak leak −0.004 GTR-Modified 0.99 −0.265 NT −0.268 leak −0.223 SBS PG 88-22 1.64 −0.288 −0.397 leak leak −0.413 135 Neat PG 52-34 0.22 −0.666 −0.610 −0.540 −0.525 −0.433 Terpolymer PG 64-34 0.43 −1.072 −0.828 −0.881 −0.873 −0.770 Neat PG 64-22 0.73 -0.174 −0.038 −0.204 0.016 −0.032 SBS PG 76-22 1.11 −0.086 −0.012 −0.087 −0.129 −0.014 SBS PG 64-34 1.21 −0.502 −0.161 leak leak −0.172 SBS PG 76-28 2.36 −0.203 −0.037 leak leak −0.042 PG 64-22 with 3 %-Latex 3.06 −0.014 −0.035 leak leak −0.030 SBS PG 88-22 7.65 leak −0.131 leak leak −0.121 Table 31. Mass change and binder leakage.

Research Approach 65   Figure 47. Comparison of mass change from AASHTO T 240 and AASHTO T 240 with heating step. -1.2 -1.0 -0.8 -0.6 -0.4 -0.2 0.0 0.2 -1.2 -1.0 -0.8 -0.6 -0.4 -0.2 0.0 0.2 U K M ix in g Sc re w M as s C ha ng e, % AASHTO T 240 Mass Change, % Conditioned at 163 ˚C Conditioned at 135 ˚C Figure 48. Comparison of mass change from AASHTO T 240 and UK mixing screw procedure.

66 Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging Figure 49. Comparison of mass change from AASHTO T 240 and NCHRP Project 09-61 mixing screw procedure. Figure 50. Comparison of mass change from AASHTO T 240 and static 0.8 mm film.

Research Approach 67   In the first option, the use of pure oxygen or oxygen-enriched air was rejected based on laboratory safety concerns. In the presence of pure oxygen or oxygen-enriched air under high pressure, hydrocarbon materials can ignite or explode (Air Products 2014). Increasing the pressure in the PAV was also rejected. Models of the pressure dependency of the asphalt oxidation reactions indicate it is a power function of pressure with an exponent between about 0.2 and 0.6 (Liu et al. 1996). For an exponent between 0.2 and 0.6, 2.1 MPa pressure is on the relatively flat portion of the curve. Doubling the pressure from 2.1 MPa to 4.2 MPa would change the reaction by between 8 and 15 percent, while significantly impacting the design of the vessel. The remaining options—increasing the conditioning temperature, increasing conditioning time, or decreasing the film thickness—were considered viable. Research completed in NCHRP Project 09-23 recommended increasing the conditioning temperature to as high as 120°C to simulate longer in-service aging in hot climates (Houston et al. 2005). Regarding timing, many researchers have increased the conditioning time to 40 hours based on the near-surface aging study completed for the Airfield Asphalt Pavement Technology Program (Hanson et al. 2009). For film thickness, researchers in Canada showed significant increases in simulated aging by using one-quarter of the AASHTO R 28 film thickness (Erskine et al. 2012). The preliminary film thickness study conducted during the evaluation phase of this project concluded that in 20 hours of conditioning, a film thickness of 0.8 mm to 0.9 mm produces residue approximately equivalent to that obtained with 40 hours of PAV conditioning using the standard film thickness of 3.18 mm. However, there are limited data relating 40-hour PAV conditioning to in-service pavement aging. The objective of the PAV operating parameters experiment was to determine the range of PAV operating parameters (conditioning temperature, conditioning time, and film thickness) needed to simulate near-surface, in-service aging of approximately 10  years for warm and cool climates. The experiment was designed to compare rheological and chemical properties of binder conditioned in the PAV using different operating conditions to those from binder recovered from the ARC validation sections in Arizona and Minnesota after approximately 4 to 5 years and 9 to 11 years in service. The experiment included comparisons for eight field sections (four at each site), each constructed using a different asphalt binder. One binder from the Minnesota sections is polymer modified, and one is modified with REOB. One binder from the Arizona sections is an air-blown binder. Experimental Design The PAV operating parameters experiment was a classic response surface experiment (NIST/SEMATECH 2017) with conditioning temperature, conditioning time, and film thick- ness as the controlled factors. The objective of a response surface experiment is to define the shape of the response surface in a well-defined region. The response surface can then be used to determine the combination of factors that produces long-term conditioned residue with properties approximating those for binder recovered from the in-service pavements. Since response surface experiments are often used in process improvement, efficient designs that allow for interaction and non-linear effects have been developed (NIST/SEMATECH 2017). The composite face-centered design used in the PAV operating parameters experiment allows the general quadratic model given in Equation 6 to be fit using data obtained from 15 combi- nations of the three controlled factors. ˆ (6)0 1 1 2 2 3 3 12 1 2 13 1 3 23 2 3 11 12 22 22 33 32y b b x b x b x b x x b x x b x x b x b x b x= + + + + + + + + +

68 Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging where ŷ = predicted response, xi = value of response i, and bij = model coefficients. Figure  51 illustrates the face-centered composite design. This design includes testing: (1) the corners of the domain representing the extreme value combinations of the factors; (2) the midpoints of each face representing high and low values for each of the three factors, keeping the other factors at their respective mid-level; and (3) the center point of the domain. The center point tests are replicated to determine the testing error. The key to designing response surface experiments is selecting appropriate ranges for each of the factors. Figure 52 compares rheological properties of AASHTO R 28 conditioned residue to the properties of binder recovered from the ARC validation section AZ1-1 after 4 years and 9 years in service for various depths. These data were collected in an earlier study completed for the FHWA (Boysen and Schabron 2015). Clearly, the aging simulated by AASHTO R 28 is not severe enough to represent near-surface aging after 9 years in service; therefore, substantially more aging was targeted in the experiment by including higher temperatures, longer condi- tioning times, and thinner films, as discussed below. Experience from the preliminary film thickness experiment completed during the evaluation phase of this project indicated that the thinnest film that can be formed in a PAV pan using reasonable care is 0.8 mm, which corresponds to a mass of 12.5 g. The preliminary film thick- ness experiment also showed that this film thickness provides conditioning in 20 hours that is approximately the same as that in 40 hours of conditioning at the standard film thickness that uses 50.0 g of binder. Based on these limits, binder masses of 12.5 g, 18.8 g, and 25.0 g were selected for the PAV operating parameters experiment. AASHTO R 28 includes conditioning temperatures of 90°C, 100°C, and 110°C depending on the high-temperature grade or climate where the binder will be used. Research completed Figure 51. Illustration of face-centered composite design.

Research Approach 69   in NCHRP Project 09-23 recommended that higher temperatures may be needed for hotter climates (Houston et al. 2005). Considering this recommendation and the data in Figure 52 that shows the aging simulated by AASHTO R 28 is much less severe than needed to simulate 10 years in service, temperatures of 100°C, 110°C, and 120°C were selected for the PAV operating parameters experiment. The conditioning time determines the turnover cycle of the conditioning equipment. The 20-hour conditioning time in AASHTO R 28 allows the equipment to be turned over every day. Recently, a conditioning time of 40 hours has been used in several research studies. The turn- over cycle for 40-hour conditioning is 2 days. Conditioning times of 20, 30, and 40 hours were selected for the PAV operating parameters experiment. Table 32 presents the conditions used for each run of the experiment. Note that six runs were conducted at the center point to estimate the error associated with conditioning and testing. Rheological and chemical responses were measured for residue conditioned in each run. The rheological responses were the change in the Christensen-Anderson master curve parameters from RTFOT residue. The chemical responses were the change in the carbonyl absorbance and the change in the sulfoxide absorbance from the RTFOT residue. Details of the test methods were presented earlier. The binders from the ARC Arizona and Minnesota validation sites were used in the experi- ment. Original binders and field cores representing two in-service aging times, 4 to 5 years and 9 to 11 years, were available for the experiment. Table 33 summarizes climatic data for the Arizona and Minnesota sites taken from the Modern-Era Retrospective analysis for Research and Application (MERRA) temperature database in LTPPBind Online (FHWA n.d.). The climate for these sections covers much of the range of climate in the United States. Four different binders were used in the pavements at each site. Table 34 summarizes the binders and cores used in each of the validation sections. Rheological and chemical properties of recovered binder were measured for three slices from cores from each test section. Figure 53 shows the slices, which include two slices near the surface of the pavement: (1) 0.0 to 0.5 in, and (2) 0.5 to 1.0 in, and a third slice deeper in the pavement. The depth of the third slice was selected based on the paving at each site to be away from layer interfaces where tack coat may have been used. The properties of the recovered binders were collected by the Western Research 1.0 1.5 2.0 2.5 3.0 3.5 4.0 4.5 5.0 0.00001 0.0001 0.001 0.01 0.1 1 10 Rh eo lo gi ca l I nd ex Crossover Frequency, rad/sec PAV Loose Mix 48 Month Aged 108 Month Aged 6. 5 m m 6. 5 m m 19 m m 19 m m 10 8 m m 10 8 m m Figure 52. Comparison of rheological properties of binder recovered from Section AZ1-1 and AASHTO R 28 conditioned binder.

70 Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging Run Conditioning Temperature, °C Mass of Binder, g Conditioning Time, hr 1 100 12.5 20 2 120 12.5 20 3 100 25.0 20 4 120 25.0 20 5 100 12.5 40 6 120 12.5 40 7 100 25.0 40 8 120 25.0 40 9 100 18.8 30 10 120 18.8 30 11 110 12.5 30 12 110 25.0 30 13 110 18.8 20 14 110 18.8 40 15 110 18.8 30 16 110 18.8 30 17 110 18.8 30 18 110 18.8 30 19 110 18.8 30 20 110 18.8 30 Table 32. Runs for three-factor, face-centered composite design. Site Zone MAAT,°C Low Air, °C Standard Deviation Low Air, °C 7-Day Average High Air, °C Standard Deviation 7-Day Average High Air, °C Annual Degree Days > 10 °C AZ Dry, no freeze 15.4 −5.09 1.69 39.98 1.39 5600.6 MN Wet, freeze 7.3 −31.06 5.09 31.09 3.05 2488.9 Table 33. Summary of climatic data for ARC validation sites in Arizona and Minnesota. Site Construction Date Binder Grade ARC Section Source Modification Core Age, yrs AZ November 2011 PG 76-16 AZ1-1 WTI/WTS Blend Air blown 4 & 9 PG 76-16 AZ1-2 Venezuelan Blend N/A PG 76-16 AZ1-3 Rocky Mountain Blend N/A PG 76-16 AZ1-4 Canadian Blend N/A MN August 2006 PG 58-34 MN1-2 Canadian Blend Terpolymer 5 & 11 PG 58-28 MN1-3 Canadian Blend N/A PG 58-28 MN1-4 Middle East Blend REOB PG 58-28 MN1-5 Venezuelan Blend N/A Table 34. Summary of binders and cores for ARC validation sites in Arizona and Minnesota.

Research Approach 71   Institute for 4-year and 9-year cores from the Arizona sections (Boysen and Schabron 2015). The properties for the 5-year and 11-year cores for the Minnesota sections were collected during this project. Results and Analysis Varying PAV Conditions Tables 35 through 42 present the Christensen-Anderson master curve parameters and the carbonyl and sulfoxide absorbances for the binders from the Arizona and Minnesota ARC validation sites for various levels of conditioning. This includes: (1) original binder, (2) RTFOT conditioned, (3) standard PAV conditioned, (4) 40-hr PAV conditioned, and (5) the combinations of temperature, mass, and conditioning time from the face-centered composite design. The tables also show the change in the master curve parameters and change in the carbonyl and sulfoxide absorbances from the RTFOT condition. For some of the 120°C PAV conditions, the conditioned binder was so stiff that a proper DSR specimen could not be formed and tested. a. Minnesota ARC Sections. Middle column, depth, is in inches. b. Arizona ARC Sections. Middle column, depth, is in inches. Built 2006 Cores from 2011 and 2017 (years 5 and 11) Minnesota Sections S2 S3 19 mm mix 3.5 4.0 4.5 5.0 5.5 2.5 3.0 0 S1 12.5 mm mix 19 mm mix 0.5 1.0 1.5 2.0 4.5 5.0 0 19 mm mix S1 0.5 S2 1.0 1.5 2.0 2.5 19 mm mix 3.0 S3 3.5 4.0 Cores from 2005 and 2010 (years 4 and 9) Arizona Sections Built 2001 Figure 53. Schematic of Minnesota and Arizona core slices.

72 Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging Condition/ PAV Run PAV Operating Parameters CA Master Curve 20°C Reference Temperature CCl4 FTIR Change from RTFOT Temp, °C Mass, g Time, hrs Td, °C R log ωc C=O S=O Td, °C R log ωc C=O S=O Original NA NA NA −6.3 2.29 1.54 0.2473 0.2735 NA NA NA NA NA RTFOT NA NA NA −5.3 2.55 0.60 0.2895 0.2862 0.00 0.00 0.00 0.0000 0.0000 PAV 20 hr 100 50 20 −4.2 2.96 −0.69 0.3840 0.3399 1.02 0.41 −1.29 0.2031 0.0537 PAV 40 hr 100 50 40 2.2 3.14 −1.39 0.4523 0.3558 7.49 0.59 −1.99 0.2873 0.0696 1 100 12.5 20 5.2 3.11 −1.43 0.4238 0.3560 10.50 0.56 −2.03 0.2590 0.0698 2 120 12.5 20 19.2 3.49 −4.28 0.5950 0.3973 24.49 0.94 −4.88 0.4715 0.1111 3 100 25.0 20 4.9 3.06 −1.23 0.4029 0.3533 10.19 0.51 −1.83 0.2354 0.0671 4 120 25.0 20 11.4 3.46 −3.28 0.5710 0.3788 16.64 0.91 −3.88 0.4290 0.0926 5 100 12.5 40 7.7 3.29 −2.18 0.4886 0.3660 12.93 0.74 −2.78 0.3338 0.0798 6 120 12.5 40 29.0 3.86 −7.13 0.7767 0.4113 34.23 1.31 −7.73 0.6672 0.1251 7 100 25.0 40 7.6 3.20 −1.93 0.4953 0.3697 12.85 0.65 −2.53 0.3442 0.0835 8 120 25.0 40 20.6 3.95 −6.15 0.7200 0.4202 25.91 1.40 −6.75 0.6194 0.1340 9 100 18.8 30 7.4 3.05 −1.51 0.4586 0.3660 12.67 0.49 −2.11 0.3038 0.0798 10 120 18.8 30 19.7 3.67 −5.25 0.6738 0.4147 24.96 1.12 −5.85 0.5673 0.1285 11 110 12.5 30 11.4 3.26 −2.67 0.5424 0.3875 16.71 0.71 −3.27 0.4091 0.1013 12 110 25.0 30 9.0 3.46 −2.87 0.5242 0.3952 14.25 0.91 −3.47 0.3886 0.1090 13 110 18.8 20 7.9 3.17 −1.91 0.4911 0.3648 13.19 0.61 −2.51 0.3351 0.0786 14 110 18.8 40 11.7 3.38 −3.16 0.5889 0.3935 16.96 0.83 −3.75 0.4616 0.1073 15 110 18.8 30 8.5 3.40 −2.81 0.5269 0.3829 13.77 0.85 −3.41 0.3890 0.0967 16 110 18.8 30 10.6 3.39 −2.85 0.5358 0.3808 15.87 0.84 −3.45 0.3958 0.0946 17 110 18.8 30 7.2 3.41 −2.80 0.5432 0.3893 12.43 0.86 −3.40 0.4117 0.1031 18 110 18.8 30 8.2 3.45 −2.86 0.5257 0.3848 13.51 0.90 −3.46 0.3897 0.0986 19 110 18.8 30 7.2 3.43 −2.74 0.5310 0.3841 12.48 0.87 −3.34 0.3943 0.0979 20 110 18.8 30 6.6 3.38 −2.69 0.5289 0.3831 11.88 0.82 −3.29 0.3912 0.0969 CA = Christensen-Anderson CCl4 = Carbon tetrachloride NA = not applicable Table 35. Master curve and FTIR data for AZ1-1.

Research Approach 73   Table 36. Master curve and FTIR data for AZ1-2. Condition/P AV Run PAV Operating Parameters CA Master Curve 20°C Reference Temperature CCl4 FTIR Change from RTFOT Temp, °C Mass, g Time, hrs Td, °C R log ωc C=O S=O Td, °C R log ωc C=O S=O Original NA NA NA −11.1 1.89 2.38 0.1120 0.2916 NA NA NA NA NA RTFOT NA NA NA −8.0 2.12 1.61 0.1312 0.3074 0.0 0.00 0.00 0.0000 0.0000 PAV 20 hr 100 50 20 −0.8 2.47 0.39 0.2137 0.4153 7.2 0.35 −1.22 0.0825 0.1079 PAV 40 hr 100 50 40 −0.3 2.66 −0.34 0.2758 0.4703 7.8 0.54 −1.95 0.1446 0.1629 1 100 12.5 20 −5.6 2.49 0.01 0.2683 0.4929 2.4 0.37 −1.61 0.1371 0.1855 2 120 12.5 20 NT NT NT NT NT NT NT NT NT NT 3 100 25.0 20 −6.11 2.48 0.12 0.2500 0.4850 1.9 0.36 −1.50 0.1188 0.1776 4 120 25.0 20 14.1 3.75 −4.93 0.4774 0.5645 22.1 1.64 −6.55 0.3462 0.2571 5 100 12.5 40 6.8 3.07 −2.16 0.3927 0.5826 14.8 0.95 −3.78 0.2615 0.2752 6 120 12.5 40 NT NT NT NT NT NT NT NT NT NT 7 100 25.0 40 5.4 3.03 −1.92 0.3658 0.5674 13.4 0.91 −3.53 0.2346 0.2600 8 120 25.0 40 NT NT NT NT NT NT NT NT NT NT 9 100 18.8 30 1.6 2.77 −0.95 0.3243 0.5431 9.6 0.65 −2.56 0.1931 0.2357 10 120 18.8 30 NT NT NT NT NT NT NT NT NT NT 11 110 12.5 30 16.9 3.90 −5.74 0.5396 0.5886 24.9 1.78 −7.35 0.4084 0.2812 12 110 25.0 30 17.0 3.88 −5.07 0.4657 0.6053 25.0 1.76 −6.69 0.3345 0.2979 13 110 18.8 20 3.4 3.20 −2.00 0.3782 0.5529 11.4 1.08 −3.62 0.2470 0.2455 14 110 18.8 40 NT NT NT NT NT NT NT NT NT NT 15 110 18.8 30 11.8 3.68 −4.36 0.4846 0.5692 19.8 1.56 −5.97 0.3534 0.2618 16 110 18.8 30 8.6 3.68 −4.04 0.4930 0.6086 16.6 1.56 −5.66 0.3618 0.3012 17 110 18.8 30 7.4 3.76 −4.03 0.4775 0.6150 15.4 1.64 −5.65 0.3463 0.3076 18 110 18.8 30 12.3 3.43 −3.66 0.4673 0.6069 20.4 1.31 −5.28 0.3361 0.2995 19 110 18.8 30 12.6 3.58 −4.22 0.4849 0.5868 20.6 1.46 −5.83 0.3537 0.2794 20 110 18.8 30 18.1 3.81 −5.40 0.4864 0.5566 26.1 1.69 −7.01 0.3552 0.2492 NA = not applicable NT = not testable

74 Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging Table 37. Master curve and FTIR data for AZ1-3. Condition/ PAV Run PAV Operating Parameters CA Master Curve 20°C Reference Temperature CCl4 FTIR Change from RTFOT Temp, °C Mass, g Time, hrs Td, °C R log ωc C=O S=O Td, °C R log ωc C=O S=O Original NA NA NA −11.6 1.63 2.49 0.1152 0.3218 NA NA NA NA NA RTFOT NA NA NA −5.8 1.89 1.61 0.1420 0.3494 0.0 0.00 0.00 0.0000 0.0000 PAV 20 hr 100 50 20 −1.6 2.19 0.34 0.2297 0.4395 4.1 0.30 −1.27 0.0877 0.0901 PAV 40 hr 100 50 40 3.0 2.35 −0.25 0.2979 0.4988 8.7 0.46 −1.86 0.1559 0.1494 1 100 12.5 20 4.7 2.44 −0.65 0.3167 0.5540 10.5 0.54 −2.26 0.1747 0.2046 2 120 12.5 20 18.2 3.26 −5.17 0.5987 0.6273 24.0 1.36 −6.78 0.4567 0.2779 3 100 25.0 20 5.0 2.37 −0.43 0.2956 0.5327 10.7 0.48 −2.03 0.1536 0.1833 4 120 25.0 20 15.6 3.02 −3.44 0.4720 0.5946 21.2 1.12 −5.05 0.3300 0.2452 5 100 12.5 40 6.5 2.66 −1.69 0.4299 0.5682 12.2 0.76 −3.29 0.2879 0.2188 6 120 12.5 40 NT NT NT NT NT NT NT NT NT NT 7 100 25.0 40 8.9 2.65 −1.82 0.4205 0.6013 14.7 0.76 −3.42 0.2785 0.2519 8 120 25.0 40 35.1 3.39 −5.89 0.6861 0.6316 40.8 1.50 −7.50 0.5441 0.2822 9 100 18.8 30 3.6 2.64 −1.25 0.3381 0.5576 9.4 0.75 −2.85 0.1961 0.2082 10 120 18.8 30 31.5 3.43 −7.27 0.6551 0.5751 37.3 1.54 −8.87 0.5131 0.2257 11 110 12.5 30 13.1 2.93 −3.28 0.5250 0.6488 18.8 1.04 −4.88 0.3830 0.2994 12 110 25.0 30 9.9 3.04 −2.61 0.4540 0.6257 15.7 1.15 −4.22 0.3120 0.2763 13 110 18.8 20 6.3 2.64 −1.63 0.4108 0.6057 12.0 0.75 −3.23 0.2688 0.2563 14 110 18.8 40 13.6 3.07 −3.94 0.5645 0.6310 19.3 1.18 −5.55 0.4225 0.2816 15 110 18.8 30 12.3 2.88 −2.94 0.4916 0.6345 18.1 0.99 −4.54 0.3496 0.2851 16 110 18.8 30 12.5 2.98 −2.96 0.4822 0.6102 18.2 1.09 −4.57 0.3402 0.2608 17 110 18.8 30 12.5 2.96 −3.06 0.5013 0.6166 18.3 1.06 −4.67 0.3593 0.2672 18 110 18.8 30 11.0 2.87 −2.79 0.5037 0.6180 16.8 0.98 −4.39 0.3617 0.2686 19 110 18.8 30 10.2 2.89 −2.78 0.4906 0.6064 16.0 1.00 −4.39 0.3486 0.2570 20 110 18.8 30 15.0 2.90 −3.25 0.4894 0.6241 20.7 1.01 −4.85 0.3474 0.2747 NA = not applicable NT = not testable

Research Approach 75   Condition/ PAV Run PAV Operating Parameters CA Master Curve 20°C Reference Temperature CCl4 FTIR Change from RTFOT Temp, °C Mass, g Time, hrs Td, °C R log ωc C=O S=O Td, °C R log ωc C=O S=O Original NA NA NA −6.5 1.73 2.07 0.1271 0.3167 NA NA NA NA NA RTFOT NA NA NA −4.9 1.90 1.31 0.1486 0.3317 0.0 0.00 0.00 0.0000 0.0000 PAV 20 hr 100 50 20 −2.3 2.20 0.20 0.2383 0.4404 2.6 0.30 −1.11 0.0897 0.1087 PAV 40 hr 100 50 40 2.0 2.34 −0.60 0.3186 0.4903 6.9 0.44 −1.91 0.1700 0.1586 1 100 12.5 20 7.8 2.34 −1.00 0.3412 0.5456 12.6 0.44 −2.31 0.1926 0.2139 2 120 12.5 20 NT NT NT 0.6886 0.5469 NT NT NT 0.5400 0.2152 3 100 25.0 20 4.5 2.34 −0.59 0.3063 0.5223 9.4 0.44 −1.90 0.1577 0.1906 4 120 25.0 20 14.2 2.83 −3.17 0.5220 0.5828 19.1 0.93 −4.48 0.3734 0.2511 5 100 12.5 40 7.5 2.58 −1.90 0.4599 0.5768 12.4 0.68 −3.21 0.3113 0.2451 6 120 12.5 40 NT NT NT NT NT NT NT NT NT NT 7 100 25.0 40 6.7 2.55 −1.76 0.4430 0.5716 11.6 0.65 −3.07 0.2944 0.2399 8 120 25.0 40 31.2 3.20 −6.98 0.7184 0.5725 36.1 1.30 −8.29 0.5698 0.2408 9 100 18.8 30 4.8 2.61 −1.59 0.3891 0.5771 9.7 0.71 −2.90 0.2405 0.2454 10 120 18.8 30 30.1 3.15 −7.00 0.7657 0.6234 35.0 1.25 −8.31 0.6171 0.2917 11 110 12.5 30 13.7 3.02 −4.04 0.5994 0.6069 18.6 1.12 −5.35 0.4508 0.2752 12 110 25.0 30 11.9 2.84 −2.91 0.4984 0.5942 16.8 0.94 −4.22 0.3498 0.2625 13 110 18.8 20 7.9 2.60 −1.98 0.4447 0.5884 12.8 0.70 −3.29 0.2961 0.2567 14 110 18.8 40 18.2 2.98 −4.74 0.6449 0.5952 23.1 1.08 −6.05 0.4963 0.2635 15 110 18.8 30 13.2 2.77 −3.21 0.5399 0.5698 18.1 0.87 −4.52 0.3913 0.2381 16 110 18.8 30 13.2 2.77 −3.17 0.5540 0.6054 18.1 0.87 −4.48 0.4054 0.2737 17 110 18.8 30 13.2 2.77 −3.23 0.5475 0.6018 18.1 0.87 −4.54 0.3989 0.2701 18 110 18.8 30 14.2 2.79 −3.27 0.5500 0.6051 19.1 0.89 −4.58 0.4014 0.2734 19 110 18.8 30 14.6 2.90 −3.44 0.5503 0.5925 19.5 0.99 −4.75 0.4017 0.2608 20 110 18.8 30 13.2 2.82 −3.29 0.5525 0.5980 18.1 0.92 −4.60 0.4039 0.2663 NA = not applicable NT = not testable Table 38. Master curve and FTIR data for AZ1-4.

76 Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging Condition/ PAV Run PAV Operating Parameters CA Master Curve 20°C Reference Temperature CCl4 FTIR Change from RTFOT Temp, °C Mass, g Time, hrs Td, °C R log ωc C=O S=O Td, °C R log ωc C=O S=O Original NA NA NA −18.3 2.02 3.78 0.1272 0.3249 NA NA NA NA NA RTFOT NA NA NA −14.3 2.26 2.93 0.1435 0.3330 0.0 0.00 0.00 0.0000 0.0000 PAV 20 hr 100 50.0 20 −8.8 2.79 1.20 0.2351 0.4097 5.5 0.53 −1.73 0.0916 0.0767 PAV 40 hr 100 50.0 40 0.4 3.38 −0.66 0.2978 0.4681 14.7 1.12 −3.59 0.1543 0.1351 1 100 12.5 20 0.4 3.28 −0.61 0.2892 0.4989 14.7 1.03 −3.54 0.1457 0.1659 2 120 12.5 20 25.0 4.45 −8.11 0.5244 0.5720 39.3 2.20 −11.04 0.3809 0.2390 3 100 25.0 20 −3.8 3.23 −0.24 0.2811 0.4939 10.5 0.97 −3.17 0.1376 0.1609 4 120 25.0 20 37.3 4.53 −8.04 0.4565 0.5636 51.6 2.27 −10.97 0.3130 0.2306 5 100 12.5 40 10.4 3.88 −3.18 0.3822 0.5527 24.7 1.62 −6.11 0.2387 0.2197 6 120 12.5 40 NT NT NT NT NT NT NT NT NT NT 7 100 25.0 40 12.6 3.72 −3.06 0.3839 0.5335 27.0 1.46 −5.98 0.2404 0.2005 8 120 25.0 40 NT NT NT NT NT NT NT NT NT NT 9 100 18.8 30 2.6 3.70 −1.88 0.3517 0.5185 17.0 1.45 −4.81 0.2082 0.1855 10 120 18.8 30 NT NT NT NT NT NT NT NT NT NT 11 110 12.5 30 19.3 4.13 −5.50 0.4665 0.5710 33.6 1.87 −8.42 0.3230 0.2380 12 110 25.0 30 19.4 4.02 −4.91 0.4297 0.5617 33.7 1.76 −7.84 0.2862 0.2287 13 110 18.8 20 8.8 3.92 −3.23 0.3734 0.5653 23.1 1.66 −6.16 0.2299 0.2323 14 110 18.8 40 30.0 4.50 −7.92 0.4970 0.5604 44.3 2.24 −10.84 0.3535 0.2274 15 110 18.8 30 20.2 4.18 −5.47 0.4593 0.5738 34.5 1.92 −8.40 0.3158 0.2408 16 110 18.8 30 22.2 4.20 −5.79 0.4511 0.5757 36.5 1.94 −8.72 0.3076 0.2427 17 110 18.8 30 23.3 4.18 −5.94 0.4514 0.5753 37.7 1.92 −8.87 0.3079 0.2423 18 110 18.8 30 23.3 4.12 −5.55 0.4604 0.5612 37.7 1.87 −8.48 0.3169 0.2282 19 110 18.8 30 SL SL SL SL SL SL SL SL SL SL 20 110 18.8 30 SL SL SL SL SL SL SL SL SL SL NA = not applicable NT = not testable SL = sample lost Table 39. Master curve and FTIR data for MN1-2.

Research Approach 77   Table 40. Master curve and FTIR data for MN1-3. Condition/ PAV Run PAV Operating Parameters CA Master Curve 20°C Reference Temperature CCl4 FTIR Change from RTFOT Temp, °C Mass, g Time, hrs Td, °C R log ωc C=O S=O Td, °C R log ωc C=O S=O Original NA NA NA −14.1 1.62 3.81 0.1167 0.2913 NA NA NA NA NA RTFOT NA NA NA −11.4 1.83 3.00 0.1378 0.3156 0.0 0.00 0.00 0.0000 0.0000 PAV 20 hr 100 50 20 −7.8 2.26 1.68 0.2350 0.4160 3.6 0.43 −1.32 0.0972 0.1004 PAV 40 hr 100 50 40 −5.9 2.60 0.57 0.3030 0.4692 5.5 0.77 −2.43 0.1652 0.1536 1 100 12.5 20 −5.4 2.66 0.39 0.3147 0.5138 6.0 0.83 −2.61 0.1769 0.1982 2 120 12.5 20 34.9 4.10 −8.23 0.6207 0.5314 46.4 2.27 −11.24 0.4829 0.2158 3 100 25.0 20 −5.5 2.55 0.70 0.2702 0.4796 6.0 0.72 −2.30 0.1324 0.1640 4 120 25.0 20 17.3 3.53 −4.05 0.4950 0.5234 28.7 1.70 −7.05 0.3572 0.2078 5 100 12.5 40 5.4 3.03 −1.33 0.4116 0.5343 16.9 1.20 −4.34 0.2738 0.2187 6 120 12.5 40 NT NT NT NT NT NT NT NT NT NT 7 100 25.0 40 2.1 3.10 −1.29 0.4071 0.5394 13.5 1.27 −4.29 0.2693 0.2238 8 120 25.0 40 26.1 4.05 −8.41 0.7172 0.5650 37.5 2.22 −11.41 0.5794 0.2494 9 100 18.8 30 −0.2 2.82 −0.37 0.3483 0.5397 11.2 0.99 −3.38 0.2105 0.2241 10 120 18.8 30 33.4 4.20 −9.12 0.6897 0.5431 44.8 2.37 −12.12 0.5519 0.2275 11 110 12.5 30 11.8 3.48 −3.74 0.5377 0.5705 23.2 1.65 −6.75 0.3999 0.2549 12 110 25.0 30 6.2 3.40 −2.56 0.4762 0.5273 17.6 1.57 −5.57 0.3384 0.2117 13 110 18.8 20 7.5 3.10 −1.64 0.4026 0.5613 18.9 1.27 −4.64 0.2648 0.2457 14 110 18.8 40 27.4 3.71 −6.23 0.5860 0.5983 38.9 1.87 −9.24 0.4482 0.2827 15 110 18.8 30 13.5 3.41 −3.43 0.5049 0.5613 24.9 1.58 −6.43 0.3671 0.2457 16 110 18.8 30 11.1 3.40 −3.16 0.4999 0.5548 22.5 1.57 −6.17 0.3621 0.2392 17 110 18.8 30 11.8 3.34 −3.27 0.5006 0.5708 23.2 1.51 −6.27 0.3628 0.2552 18 110 18.8 30 12.9 3.37 −3.32 0.4980 0.5735 24.3 1.53 −6.32 0.3602 0.2579 19 110 18.8 30 13.1 3.27 −3.06 0.5055 0.5782 24.5 1.44 −6.07 0.3677 0.2626 20 110 18.8 30 13.1 3.23 −3.23 0.5007 0.5741 24.5 1.40 −6.23 0.3629 0.2585 NA = not applicable NT = not testable

78 Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging Table 41. Master curve and FTIR data for MN1-4. Condition/ PAV Run PAV Operating Parameters CA Master Curve 20°C Reference Temperature CCl4 FTIR Change from RTFOT Temp, °C Mass, g Time, hrs Td, °C R log ωc C=O S=O Td, °C R log ωc C=O S=O Original NA NA NA −13.5 2.02 3.45 0.1406 0.3257 NA NA NA NA NA RTFOT NA NA NA −11.4 2.31 2.55 0.1734 0.3391 0.0 0.00 0.00 0.0000 0.0000 PAV 20 hr 100 50 20 −7.2 2.74 1.12 0.2791 0.4201 4.2 0.43 −1.42 0.1057 0.0810 PAV 40 hr 100 50 40 −5.3 2.96 0.24 0.3428 0.4650 6.1 0.65 −2.31 0.1694 0.1259 1 100 12.5 20 −6.1 2.97 0.27 0.3241 0.4800 5.3 0.67 −2.27 0.1507 0.1409 2 120 12.5 20 19.6 3.84 −5.01 0.5752 0.5698 31.0 1.54 −7.56 0.4018 0.2307 3 100 25.0 20 −7.1 2.88 0.52 0.3085 0.4722 4.3 0.57 −2.03 0.1351 0.1331 4 120 25.0 20 11.8 3.64 −3.11 0.4973 0.5540 23.2 1.33 −5.66 0.3239 0.2149 5 100 12.5 40 −0.9 3.17 −0.64 0.4053 0.5132 10.6 0.86 −3.19 0.2319 0.1741 6 120 12.5 40 NT NT NT NT NT NT NT NT NT NT 7 100 25.0 40 2.3 3.13 −0.79 0.4058 0.5190 13.7 0.82 −3.34 0.2324 0.1799 8 120 25.0 40 30.4 3.83 −6.81 0.6539 0.5450 41.8 1.52 −9.36 0.4805 0.2059 9 100 18.8 30 −0.1 3.02 −0.07 0.3624 0.4877 11.3 0.71 −2.62 0.1890 0.1486 10 120 18.8 30 18.9 3.66 −4.88 0.6279 0.5827 30.4 1.35 −7.43 0.4545 0.2436 11 110 12.5 30 13.7 3.48 −2.87 0.4853 0.5672 25.1 1.18 −5.42 0.3119 0.2281 12 110 25.0 30 4.7 3.44 −1.86 0.4577 0.5449 16.1 1.13 −4.40 0.2843 0.2058 13 110 18.8 20 0.4 3.24 −1.04 0.3972 0.5281 11.8 0.93 −3.59 0.2238 0.1890 14 110 18.8 40 13.0 3.65 −3.54 0.5213 0.5731 24.4 1.34 −6.09 0.3479 0.2340 15 110 18.8 30 5.1 3.26 −1.52 0.4675 0.5335 16.5 0.96 −4.07 0.2941 0.1944 16 110 18.8 30 7.4 3.41 −2.23 0.4684 0.5500 18.8 1.10 −4.78 0.2950 0.2109 17 110 18.8 30 5.2 3.35 −1.89 0.4725 0.5409 16.6 1.04 −4.43 0.2991 0.2018 18 110 18.8 30 5.0 3.43 −1.97 0.4649 0.5479 16.4 1.13 −4.51 0.2915 0.2088 19 110 18.8 30 5.0 3.48 −2.18 0.4685 0.5495 16.4 1.17 −4.73 0.2951 0.2104 20 110 18.8 30 5.7 3.55 −2.39 0.4721 0.5507 17.1 1.24 −4.94 0.2987 0.2116 NA = not applicable NT = not testable

Research Approach 79   Condition/ PAV Run PAV Operating Parameters CA Master Curve 20°C Reference Temperature CCl4 FTIR Change from RTFOT Temp, °C Mass, g Time, hrs Td, °C R log ωc C=O S=O Td, °C R log ωc C=O S=O Original NA NA NA −15.1 1.44 4.16 0.2232 0.3003 NA NA NA NA NA RTFOT NA NA NA −12.4 1.59 3.63 0.2496 0.3099 0.0 0.00 0.00 0.0000 0.0000 PAV 20 hr 100 50 20 −10.3 1.86 2.69 0.3338 0.4244 2.1 0.27 −0.94 0.0842 0.1145 PAV 40 hr 100 50 40 −7.8 2.14 1.94 0.4120 0.4538 4.5 0.54 −1.70 0.1624 0.1439 1 100 12.5 20 −7.3 2.09 1.94 0.3984 0.4704 5.1 0.50 −1.70 0.1488 0.1605 2 120 12.5 20 17.9 3.64 −5.04 0.8773 0.5448 30.3 2.05 −8.68 0.6277 0.2349 3 100 25.0 20 −8.6 2.00 2.21 0.3719 0.4594 3.8 0.41 −1.42 0.1223 0.1495 4 120 25.0 20 0.0 2.98 −1.06 0.6495 0.5530 12.4 1.38 −4.69 0.3999 0.2431 5 100 12.5 40 −4.4 2.45 0.73 0.5303 0.5124 7.9 0.86 −2.91 0.2807 0.2025 6 120 12.5 40 NT NT NT NT NT NT NT NT NT NT 7 100 25.0 40 −6.2 2.40 0.82 0.5151 0.5276 6.2 0.81 −2.81 0.2655 0.2177 8 120 25.0 40 16.4 3.80 −5.12 0.8805 0.5593 28.7 2.21 −8.76 0.6309 0.2494 9 100 18.8 30 −7.1 2.22 1.49 0.4436 0.5154 5.3 0.63 −2.15 0.1940 0.2055 10 120 18.8 30 23.5 3.72 −6.01 0.8911 0.5449 35.8 2.13 −9.65 0.6415 0.2350 11 110 12.5 30 4.0 3.06 −1.53 0.7371 0.5258 16.4 1.47 −5.16 0.4875 0.2159 12 110 25.0 30 −0.5 2.76 −0.32 0.6062 0.5460 11.9 1.17 −3.95 0.3566 0.2361 13 110 18.8 20 −6.4 2.43 0.82 0.5282 0.5246 5.9 0.83 −2.82 0.2786 0.2147 14 110 18.8 40 6.4 3.25 −2.43 0.7755 0.5713 18.8 1.66 −6.06 0.5259 0.2614 15 110 18.8 30 −1.5 2.80 −0.70 0.6587 0.5609 10.8 1.21 −4.33 0.4091 0.2510 16 110 18.8 30 −1.3 2.90 −0.61 0.6630 0.5360 11.1 1.31 −4.25 0.4134 0.2261 17 110 18.8 30 −0.6 2.89 −0.87 0.6710 0.5514 11.8 1.30 −4.50 0.4214 0.2415 18 110 18.8 30 −0.9 2.87 −0.69 0.6583 0.5387 11.4 1.28 −4.32 0.4087 0.2288 19 110 18.8 30 −0.3 2.88 −0.82 0.6759 0.5265 12.1 1.28 −4.46 0.4263 0.2166 20 110 18.8 30 −0.9 2.93 −0.95 0.6548 0.5240 11.4 1.34 −4.58 0.4052 0.2141 NA = not applicable NT = not testable Table 42. Master curve and FTIR data for MN1-5.

80 Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging Various graphs were constructed using data in Tables 35 through 42 to map the changes in rheology and chemistry of each binder with increasing severity of PAV conditioning. The most informative graphs were plots showing the change in the rheological index (Figures  54 and 55), change in the crossover frequency as a function of change in the carbonyl absorbance, and change in the crossover frequency as a function of change in the sulfoxide absorbance from the RTFOT condition (Figures 56 through 59). The same scale is used in the figures for ease of comparison. All graphs are coded by PAV operating temperature because this factor has the greatest effect on the measured data. Within a temperature, the severity of PAV conditioning increases with thinner films and longer conditioning times. The standard PAV conditioning (50.0 g at 100°C for 20 hours) and 50.0 g at 100°C for 40 hours are labeled “STD PAV” and “40 hr PAV.” These graphs show that for each binder, increasing severity of PAV conditioning produces rational and predict- able rheological and chemical changes. There are no abrupt changes evident for any of the binders. The data for more severely conditioned binder, however, is somewhat more vari- able. Figures 54 and 55 show the change in the rheological index as a function of the change in the crossover frequency for the Arizona and Minnesota binders, respectively. Increasing severity of PAV conditioning is movement to the left and up in these figures. For a specific binder, as the binder becomes harder, crossover frequency decreases, and the rheological index increases, resulting in a flatter master curve. Figures 56 and 57 show the change in the crossover frequency as a function of the change in the carbonyl absorbance for the Arizona and Minnesota binders, respectively. Increasing severity of PAV conditioning is movement to the right and down in these figures. For a given binder, as the carbonyl absorbance increases, the binder becomes harder as indicated by a decrease in the crossover frequency. Figures 58 and 59 show similar plots using the change in crossover frequency as a function of the change in the sulfoxide absorbance as the measure of the change in chemistry. Comparing these figures to those for the carbonyl absorbance show changes in carbonyl absorbance are better related to the changes in rheology. The change in rheological index versus change in crossover frequency to the center point of the experiment, 18.8 g at 110°C for 30 hours, provides a measure of the relative sensitivity of each binder to increased severity of conditioning. Figure 60 shows the change in the rheological index versus the change in the crossover frequency to the center point of the experiment. MN1-2 is the most sensitive to 30 hours of PAV conditioning at 110°C using a mass of 18.8 g, while AZ1-1 is the least sensitive to this level of conditioning. AZ1-3, AZ1-4, MN1-4, and MN1-5 have similar sensitivity to conditioning and are less sensitive than AZ1-2 and MN1-3, which also have similar sensitivity to PAV conditioning. Equation 6 was used to develop response surfaces for the change in crossover frequency, the change in carbonyl absorbance, and the change in sulfoxide absorbance from the RTFOT condition. The strong correlation between the crossover frequency and the rheological index for each binder shown in Figure 54 and Figure 55, indicates that only one of these parameters is needed as an indicator of the change in the rheological properties. The crossover frequency was selected because it has a larger range compared to the rheological index. The response surfaces were fit using stepwise linear regression after applying appropriate transformations to the factor variables (for example, temperature squared, temperature times thickness, etc.). Vari- ables having p-values less than 0.05 were considered significant and included in the models. The response surface models are summarized in Table 43 for the change in the crossover frequency, Table 44 for the change in carbonyl absorbance, and Table 45 for the change in sulfoxide absorbance. These tables include the model coefficients that were found to be statistically significant and the explained variance for the models.

a. Binder AZ1-1. c. Binder AZ1-3. b. Binder AZ1-2. d. Binder AZ1-4. Figure 54. Change in the rheological index as a function of change in crossover frequency for Arizona binders.

a. Binder MN1-2. c. Binder MN1-4. b. Binder MN1-3. d. Binder MN1-5. Figure 55. Change in the rheological index as a function of change in crossover frequency for Minnesota binders.

a. Binder AZ1-1. c. Binder AZ1-3. b. Binder AZ1-2. d. Binder AZ1-4. Figure 56. Change in crossover frequency as a function of change in carbonyl absorbance for Arizona binders.

a. Binder MN1-2. c. Binder MN1-4. b. Binder MN1-3. d. Binder MN1-5. Figure 57. Change in crossover frequency as a function of change in carbonyl absorbance for Minnesota binders.

a. Binder AZ1-1. c. Binder AZ1-3. b. Binder AZ1-2. d. Binder AZ1-4. Figure 58. Change in crossover frequency as a function of change in sulfoxide absorbance for Arizona binders.

a. Binder MN1-2. c. Binder MN1-4. b. Binder MN1-3. d. Binder MN1-5. Figure 59. Change in crossover frequency as a function of change in sulfoxide absorbance for Minnesota binders.

Research Approach 87   0.00 0.50 1.00 1.50 2.00 2.50 -10.00 -9.00 -8.00 -7.00 -6.00 -5.00 -4.00 -3.00 -2.00 -1.00 0.00 ΔR Δlogωc AZ1-1 AZ1-2 AZ1-3 AZ1-4 MN1-2 MN1-3 MN1-4 MN1-5 Figure 60. Path length of change in rheology to the center point of the PAV operating parameters experiment (18.8 g, 110çC, 30 hrs). Binder Intercept Temp, °C Mass, g Time, hr Temp2, °C Mass2, g Time2 Temp, °C x Mass, g Time, hr x Mass, g Temp, °C x Time R2 AZ1-1 −78.14194 1.46927 −0.30044 0.50216 −0.00702 0.00306 −0.00533 98.8 AZ1-2 −80.58201 1.52488 −0.37838 1.40548 −0.00716 0.00393 −0.01511 93.0 AZ1-3 −57.39055 1.27047 −0.88928 0.37877 −0.00705 0.00885 −0.00442 95.5 AZ1-4 −61.42471 1.33470 −1.16930 0.65404 −0.00733 0.01181 −0.00709 99.1 MN1-2 −111.74651 2.17303 0.06417 0.86208 −0.01067 −0.00997 −0.00038 99.4 MN1-3 −82.62631 1.94919 −1.72682 0.59401 −0.01124 0.01723 −0.00700 98.0 MN1-4 −65.89802 1.36973 −0.62555 0.52082 −0.00722 0.00632 −0.00579 97.7 MN1-5 −96.95516 2.11394 −1.61016 0.67732 −0.01151 0.01609 −0.00747 98.8 Table 43. Crossover frequency response surface models. Binder Intercept Temp, °C Mass, g Time, hr Temp2, °C Mass2, g Time2 Temp, °C x Mass, g Time, hr x Mass, g Temp, °C x Time R2 AZ1-1 2.60784 −0.05221 0.01282 −0.01800 0.00027 −0.00013 0.00022 99.3 AZ1-2 −0.95791 0.00823 −0.00331 −0.00025 0.00021 97.9 AZ1-3 −0.64964 0.04359 −0.01734 0.00007 −0.00045 0.00023 99.3 AZ1-4 −1.12499 0.05972 0.00011 −0.00016 −0.00062 0.00017 99.0 MN1-2 1.05393 −0.03117 0.02570 0.00020 −0.00019 −0.00026 0.00017 99.8 MN1-3 −0.67017 0.04067 −0.02083 0.00007 −0.00042 0.00027 99.5 MN1-4 1.91606 −0.04621 0.02747 −0.00679 0.00027 −0.00009 −0.00028 0.00017 99.8 MN1-5 −3.01002 0.03017 0.08485 −0.01826 −0.00087 0.00026 99.4 Table 44. Carbonyl absorbance response surface models.

88 Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging Binder Intercept Temp, °C Mass, g Time, hr Temp2, °C Mass2, g Time2 Temp, °C x Mass, g Time, hr x Mass, g Temp, °C x Time R 2 AZ1-1 −4.34739 0.075706 0.02178 −0.00031 −0.00003 −0.00019 96.4 AZ1-2 −2.89778 0.051930 0.014874 −0.00025 −0.00044 0.00003 87.9 AZ1-3 −2.82837 0.048085 0.00423 −0.00022 84.6 AZ1-4 −0.04491 0.01497 −0.00022 −0.00023 0.00007 68.8 MN1-2 −4.34739 0.075706 0.02178 −0.00031 −0.00003 −0.00019 96.4 MN1-3 −3.00861 0.055169 0.014874 −0.00025 −0.00043 0.00003 87.9 MN1-4 −2.82837 0.048085 0.012742 −0.00019 −0.00001 −0.00011 91.6 MN1-5 −2.42692 0.041219 0.014168 −0.00016 −0.00011 85.0 Table 45. Sulfoxide absorbance response surface models. The crossover frequency response surface models for the various binders are compared in Figure 61 for the Arizona binders and Figure 62 for the Minnesota binders. Figure 63 and Figure  64 present the comparisons of the carbonyl absorbance response surface models and Figure 65 and Figure 66 show the comparisons of the sulfoxide absorbance response surface models. These comparisons present the changes from the RTFOT condition as a function of temperature, with the data points coded by binder mass and conditioning time. The same scale is used in all figures to make the comparisons easier. The graphs are truncated at the temperature where the change in the logarithm of the crossover frequency reached the maximum that could be measured for the binders: −8 for the Arizona binders and −12 for the Minnesota binders. Recall, some binders when conditioned at 120°C were so stiff that proper DSR test specimens could not be formed. The crossover frequency response surfaces show large differences in the sensitivity of the binders to the various PAV operating parameters. All binders are most sensitive to changes in the PAV conditioning temperature, with binders AZ1-2 (Figure 61b) and MN1-2 (Figure 61a) being the most sensitive. The increasing spread in the data with increasing temperature indicates the sensitivity to changes in conditioning time and binder mass increases with increasing temperature. Binder AZ1-2 (Figure 61b) is also quite sensitive to changes in conditioning time, as shown by the large spread for different conditioning times. The effect of binder mass is relatively small for this binder. The other binders show similar sensitivity to changes in condi- tioning time and binder mass. The carbonyl absorbance response surface models are compared in Figure 63 for the Arizona binders and Figure 64 for the Minnesota binders. The carbonyl absorbance is most sensitive to changes in PAV conditioning temperature for all binders. Binders AZ1-4 (Figure 63d) and MN1-5 (Figure 64d) are most sensitive to conditioning temperature changes. All binders, except AZ1-2 (Figure 63b), exhibit increased sensitivity to thickness with increasing temperature. Binders AZ1-1 and AZ1-2 (Figures 63a and 63b respectively) are more sensitive to change in conditioning time compared to binder mass, as shown by the distinct grouping for different conditioning times. The sulfoxide absorbance response surfaces are compared in Figure 65 for the Arizona binders and Figure 66 for the Minnesota binders. The sulfoxide absorbance is most sensitive to changes in PAV conditioning temperature for all binders. Binders AZ1-3 (Figure 65c), MN1-2 (Figure 66a), and MN1-3 (Figure 66b) have peak sulfoxide absorbances in the range of PAV conditioning temperatures from 100°C to 115°C. The sulfoxide absorbance is relatively insensitive to binder mass. Binders AZ1-1 (Figure 65a), AZ1-2 (Figure 65b), and MN1-5 (Figure 66d) exhibit no change in sulfoxide absorbance over the range of 12.5 g to 25.0 g in a standard PAV pan.

a. Binder AZ1-1. c. Binder AZ1-3. b. Binder AZ1-2. d. Binder AZ1-4. Figure 61. Comparison of the crossover frequency response surface models for the Arizona binders.

a. Binder MN1-2. c. Binder MN1-4. b. Binder MN1-3. d. Binder MN1-5. Figure 62. Comparison of the crossover frequency response surface models for the Minnesota binders.

a. Binder AZ1-1. c. Binder AZ1-3. b. Binder AZ1-2. d. Binder AZ1-4. Figure 63. Comparison of the carbonyl absorbance response surface models for the Arizona binders.

a. Binder MN1-2. c. Binder MN1-4. b. Binder MN1-3. d. Binder MN1-5. Figure 64. Comparison of the carbonyl absorbance response surface models for the Minnesota binders.

a. Binder AZ1-1. c. Binder AZ1-3. b. Binder AZ1-2. d. Binder AZ1-4. Figure 65. Comparison of the sulfoxide absorbance response surface models for the Arizona binders.

a. Binder MN1-2. c. Binder MN1-4. b. Binder MN1-3. d. Binder MN1-5. Figure 66. Comparison of the sulfoxide absorbance response surface models for the Minnesota binders.

Research Approach 95   Field Cores The Christensen-Anderson master curve parameters and the carbonyl absorbance for binder recovered from the field cores are presented in Table 46 for the cores from the Arizona ARC validation sections and Table 47 for the cores from the Minnesota ARC validation sections. These tables include the measured data for slices from three depths and the change in the master curve parameters and the carbonyl absorbance from the RTFOT condition. Plots of the change in the crossover frequency and change in the carbonyl absorbance from RTFOT-conditioned binder as a function of depth are shown in Figure 67 for the Arizona sections and Figure 68 for the Minnesota sections. The same scale is used in all plots to facilitate the comparisons. In these plots, increased aging is indicated by decreases in the crossover frequency and increases in the carbonyl absorbance. The data generally exhibit logical trends of increased aging near the surface of the pavement that attenuates with depth, and the aging is greater for the longer time in service. Aging appears to decrease more rapidly with depth for the Arizona sections compared to the Minnesota sections. For the Arizona sections, the rheological and chemical properties at 0.75 in indicate only slightly more aging compared to the properties at 3.25 in (Figure 67); for the Minnesota sections, the rheological and chemical properties at 0.75 in are significantly more aged compared to the properties at 3.75 in (Figure 68). These plots also indicate the precision of the testing of recovered binders. For example, the 4-year crossover frequency for AZ1-2 indicates slightly greater aging at 3.25 in compared to 0.75 in (Figure 67b). The difference is approximately 20 percent of change in the crossover frequency for the 0.75 in depth. Another Location CA Master Curve 20°C Reference Temperature CCl4 FTIR Change from RTFOT Section Age, yrs Depth, in Td, °C log ωc R C=O S=O Td, °C log ωc R C=O S=O AZ1-1 4 0.25 1.7 −1.58 2.94 0.5023 0.4202 7.0 −2.17 0.39 0.2128 0.1340 9 0.25 1.5 −2.40 3.34 0.5303 0.4398 6.8 −3.00 0.78 0.2408 0.1536 4 0.75 4.8 −1.34 2.96 0.4426 0.3985 10.0 −1.94 0.41 0.1531 0.1123 9 0.75 3.7 −1.89 3.18 0.4729 0.4147 9.0 −2.49 0.63 0.1834 0.1285 4 3.25 −4.3 −0.87 2.92 0.4797 0.3642 1.0 −1.47 0.37 0.1902 0.0780 9 3.25 3.2 −1.06 2.82 0.5024 0.3704 8.5 −1.66 0.27 0.2129 0.0842 AZ1-2 4 0.25 −2.6 0.05 2.39 0.3129 0.5634 5.5 −1.57 0.27 0.1817 0.2560 9 0.25 4.8 −1.12 2.76 0.3549 0.6383 12.8 −2.73 0.64 0.2237 0.3309 4 0.75 −9.4 0.78 2.20 0.2615 0.5263 −1.4 −0.83 0.08 0.1303 0.2189 9 0.75 1.5 −0.18 2.53 0.3046 0.5889 9.5 −1.80 0.41 0.1734 0.2815 4 3.25 −9.0 0.53 2.31 0.2809 0.5538 −0.9 −1.08 0.19 0.1497 0.2464 9 3.25 −0.4 0.06 2.53 0.3068 0.5831 7.6 −1.56 0.41 0.1756 0.2757 AZ1-3 4 0.25 0.6 −0.51 2.25 0.3881 0.5781 6.4 −2.12 0.36 0.2461 0.2287 9 0.25 6.0 −1.70 2.54 0.4614 0.6340 11.7 −3.31 0.64 0.3194 0.2846 4 0.75 −1.1 0.36 2.02 0.3309 0.5322 4.7 −1.25 0.13 0.1889 0.1828 9 0.75 −0.3 −0.57 2.26 0.4075 0.6142 5.5 −2.17 0.37 0.2655 0.2648 4 3.25 −1.6 0.70 1.98 0.3602 0.5561 4.2 −0.91 0.09 0.2182 0.2067 9 3.25 0.3 0.00 2.16 0.4277 0.6090 6.1 −1.60 0.27 0.2857 0.2596 AZ1-4 4 0.25 6.2 −1.17 2.43 0.4122 0.6023 11.1 −2.47 0.53 0.2636 0.2706 9 0.25 12.4 −2.18 2.37 0.4682 0.6074 17.3 −3.49 0.47 0.3196 0.2757 4 0.75 −0.2 −0.24 2.21 0.3611 0.5704 4.7 −1.55 0.31 0.2125 0.2387 9 0.75 1.6 −0.68 2.32 0.4127 0.6105 6.5 −1.99 0.42 0.2641 0.2788 4 3.25 −3.2 −0.24 2.21 0.4032 0.5956 1.7 −1.54 0.31 0.2546 0.2639 9 3.25 0.6 −0.51 2.30 0.4279 0.6058 5.5 −1.81 0.40 0.2793 0.2741 Table 46. Master curve and FTIR data for Arizona field cores.

96 Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging example is all Minnesota sections have greater carbonyl absorbance at a depth of 3.75 in for the 5-year cores compared to the 11-year cores (Figure 68), which is not rational and likely due to variability in the data. The difference is approximately 25 percent of the average change in carbonyl plus absorbance measured for 3.75 in. There are three sources of variability that could account for the difference of 20 to 25 percent: the extraction and recovery process, the testing, and data analysis. Comparison of PAV Conditioning and Field Aging The selection of appropriate PAV operating parameters was based on comparing rheological and chemical properties for various PAV operating conditions to the properties of the field cores. The rheological properties are compared in Figure 69 for the Arizona sections and Figure 70 for the Minnesota sections. These figures were prepared by adding the field core data (field-aged data) to the plots presented earlier showing the rheological index as a function of crossover frequency for the various PAV operating conditions. The field core data are coded by time in service and labeled with the depth at the midpoint of the 0.5 in thick slice. The same scale is used in all plots to facilitate the comparisons. Except for section MN1-2, the section with the terpolymer- modified binder, there is reasonable agreement between the lab and field data. The field data plot close to the regression line formed from the data for the various PAV conditions. The field data generally exhibit decreasing crossover frequency and increasing rheological index with increasing exposure. Increasing exposure occurs closer to the surface and for longer time in service. Location CA Master Curve 20°C Reference Temperature CCl4 FTIR Change from RTFOT Section Age, yrs Depth, in Td, °C log ωc R C=O S=O Td, °C log ωc R C=O S=O MN1-2 5 0.25 −9.3 1.03 2.45 0.4407 0.5585 5.1 −1.90 0.20 0.2972 0.2255 11 0.25 −8.0 0.39 2.50 0.4511 0.5678 6.4 −2.54 0.25 0.3076 0.2348 5 0.75 −11.0 1.65 2.13 0.4391 0.5463 3.4 −1.28 −0.12 0.2956 0.2133 11 0.75 −4.2 0.45 2.08 0.3417 0.4792 10.1 −2.48 −0.17 0.1982 0.1462 5 3.75 −13.0 2.35 2.23 0.3343 0.4309 1.4 −0.58 −0.03 0.1908 0.0979 11 3.75 −15.2 2.76 2.13 0.2786 0.4390 −0.9 −0.16 −0.12 0.1351 0.1060 MN1-3 5 0.25 −7.5 0.77 2.35 0.3674 0.5221 3.9 −2.23 0.52 0.2296 0.2065 11 0.25 −5.8 0.02 2.42 0.4079 0.5511 5.6 −2.98 0.59 0.2701 0.2355 5 0.75 −8.0 1.31 2.28 0.3258 0.5088 3.4 −1.69 0.45 0.1880 0.1932 11 0.75 −7.5 0.84 2.30 0.3857 0.5561 3.9 −2.16 0.47 0.2479 0.2405 5 3.75 −10.1 2.02 2.08 0.2746 0.4376 1.3 −0.99 0.25 0.1368 0.1220 11 3.75 −12.0 2.42 2.00 0.2206 0.3819 −0.5 −0.59 0.17 0.0828 0.0663 MN1-4 5 0.25 −7.1 0.25 2.76 0.3939 0.5272 4.4 −2.30 0.45 0.2205 0.1881 11 0.25 3.1 −0.72 2.90 0.4372 0.5643 14.6 −3.27 0.59 0.2638 0.2252 5 0.75 −8.1 0.81 2.69 0.3640 0.5318 3.4 −1.74 0.39 0.1906 0.1927 11 0.75 −8.5 0.12 2.85 0.4162 0.5573 2.9 −2.43 0.54 0.2428 0.2182 5 3.75 −9.8 1.83 2.43 0.2679 0.4675 1.6 −0.72 0.13 0.0945 0.1284 11 3.75 −12.7 2.05 2.38 0.2517 0.4238 −1.3 −0.50 0.07 0.0783 0.0847 MN1-5 5 0.25 −7.8 1.46 2.04 0.4407 0.5585 4.5 −2.17 0.45 0.1911 0.2846 11 0.25 −8.5 1.46 2.10 0.4511 0.5678 3.8 −2.18 0.50 0.2015 0.2939 5 0.75 −10.9 2.71 1.42 0.4391 0.5463 1.4 −0.93 −0.17 0.1895 0.2724 11 0.75 −10.5 2.32 1.87 0.3417 0.4792 1.8 −1.32 0.27 0.0921 0.2053 5 3.75 −13.3 2.96 1.73 0.3343 0.4309 −1.0 −0.67 0.14 0.0847 0.1570 11 3.75 −13.5 3.00 1.74 0.2786 0.4390 −1.1 −0.63 0.15 0.0290 0.1651 Table 47. Master curve and FTIR data for Minnesota field cores.

a. Binder AZ1-1. c. Binder AZ1-3. b. Binder AZ1-2. d. Binder AZ1-4. Figure 67. Change in crossover frequency and carbonyl absorbance for Arizona field cores.

a. Binder MN1-2. c. Binder MN1-4. b. Binder MN1-3. d. Binder MN1-5. Figure 68. Change in crossover frequency and carbonyl absorbance for Minnesota field cores.

a. Binder AZ1-1. c. Binder AZ1-3. b. Binder AZ1-2. d. Binder AZ1-4. Figure 69. Comparison of PAV-conditioned and field-aged change in the rheological index as a function of crossover frequency for Arizona binders.

a. Binder MN1-2. c. Binder MN1-4. b. Binder MN1-3. d. Binder MN1-5. Figure 70. Comparison of PAV-conditioned and field-aged change in the rheological index as a function of crossover frequency for Minnesota binders.

Research Approach 101   The negative change in the rheological index (field-aged rheological index less than RTFOT rheological index) for some of the MN1-2 data suggests that all the polymer may not have been extracted and recovered. The rheological index of polymer-modified binders is generally higher than neat binders. Analysis of the FTIR spectra was conducted to determine if the recovered binders contained terpolymer. Figure 71 compares FTIR spectra for three of the recovered binder samples to FTIR spectra from the PAV-conditioned samples at approximately the same carbonyl absorbance around 1700 cm–1. The peak around 1732 cm–1 corresponds to the carbonyl of the ester in the terpolymer. Although both the recovered field binders and the PAV- conditioned binders exhibit a peak around 1732 cm–1, the peak is always lower for the field binders at approximately the same oxidation level, indicating a lower polymer content. The crossover frequency as a function of the carbonyl absorbance from the recovered field binders (field-aged binders) is compared to PAV binders, seen in Figure 72 for the Arizona sections and Figure 73 for the Minnesota sections. These figures were prepared by adding the recovered field core data to the plots presented earlier showing the crossover frequency as a function of the carbonyl absorbance for the various PAV operating conditions. The field core data are coded by years in service and labeled with the depth at the midpoint of the 0.5 in thick slice. The same scale is used in all plots to facilitate the comparisons. The trends in both the field-aged and PAV-conditioned data are similar; increasing exposure results in an increase in the carbonyl absorbance and a decrease in the crossover frequency. However, there is a somewhat poorer agreement between the field and PAV data in these plots than the plots containing only rheological data. Binders AZ1-1 and MN1-5 show reasonable agreement, but for the other binders, the change in the crossover frequency is smaller for the field-aged binders compared to the PAV-conditioned binders at the same change in the carbonyl absor- bance. The comparisons in Figure 72 and Figure 73 show that accelerated aging in the PAV will not always simultaneously match the change in rheology and the change in the chemistry of field-aged binders. Figure 71. Comparison of the carbonyl of the ester in the terpolymer (}1732 cm–1) for recovered field binder with PAV-conditioned binder at approximately equal oxidation (}1700 cm–1).

a. Binder AZ1-1. c. Binder AZ1-3. b. Binder AZ1-2. d. Binder AZ1-4. Figure 72. Comparison of PAV-conditioned and field-aged change in crossover frequency as a function of carbonyl absorbance for Arizona binders.

a. Binder MN1-2. c. Binder MN1-4. b. Binder MN1-3. d. Binder MN1-5. Figure 73. Comparison of PAV-conditioned and field-aged change in crossover frequency as a function of carbonyl absorbance for Minnesota binders.

104 Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging Optimization The response surface models were used in an optimization to select the PAV operating param- eters that best matched the field data for each validation site. The objective function for the opti- mization was to minimize the sum of the square of the relative error between PAV-conditioned binder and the field-aged binder, as defined by Equation 7. SSRE PAV Field Field ∑= − ×   100 (7) 2 where SSRE = sum of relative errors squared, PAV = PAV-conditioned property for the appropriate response surface, and Field = field-aged property. The optimization was performed for the slice representing the top 0.5 in in the pavement and the slice representing 0.5 in to 1.0 in in the pavement using the Solver function in Excel. The lowest slices were not included in the optimization because the recovered field binders from these slices were generally aged less than the PAV conditions used to fit the response surfaces. Binder MN1-2 was not included in the optimization due to the concerns about incomplete extraction and recovery of the polymer discussed earlier. The optimization was first conducted allowing conditioning time, binder mass, and condi- tioning temperature to vary within the range of the experiment. The optimization was performed for four criteria: 1. Minimizing the relative error for the change in the logarithm of the cross over frequency, 2. Minimizing the relative error for the change in the carbonyl absorbance, 3. Minimizing the relative error for the change in the sulfoxide absorbance, and 4. Minimizing the combined relative error for the carbonyl absorbance and the crossover frequency. The results of this optimization are summarized in Table 48. The table shows the operating parameters that provided the best fit for each site and each criterion, the average error for the change in the logarithm of the crossover frequency, and the change in the carbonyl absorbance and the sulfoxide absorbance. The best match for the sulfoxide absorbance requires more severe aging conditions compared to the best match for the logarithm of the crossover frequency and the carbonyl absorbance. To approximate the aging of the top 0.5 in of the pavement will require higher temperatures and thinner films than currently used in the PAV. The PAV data at the optimum conditions based on minimizing the combined relative error for the crossover frequency and the carbonyl absorbance are compared to the field data in Figure 74. As shown by the regression equations in the graphs, these operating parameters provide a relatively unbiased estimate of field aging based on the change in the logarithm of the crossover frequency. These operating parameters, however, produce conditioning binders with lower carbonyl absorbance compared to the field binders. Considering the data in Table 48, a constrained optimization was conducted using a mass of 12.5 g and a conditioning time of 20 hours to determine the optimum conditioning tempera- tures for the current PAV conditioning time using the thinnest film tested. The optimum PAV conditioning temperatures for the Arizona and Minnesota sites for the top 0.5 in slice and the 0.5 in to 1.0 in slice that minimize the combined relative error for the carbonyl absorbance and the crossover frequency are summarized in Table 49 along with the average relative errors. Using 12.5 g and 20 hours of conditioning, the optimum temperatures for the 0.5 in to 1.0 in slice are similar to the current AASHTO R 28 temperature of 100°C. The optimum temperatures for

Research Approach 105   Site Slice, in Criteria PAV Operating Parameters Average Relative Error, % Temperature, °C Mass, g Time, hr ∆log ωc ∆(C=O) ∆(S=O) AZ 0–0.5 ωc 106.7 12.5 20 −4 −14 −25 C=O 110.7 17.0 20 14 −2 −23 S=O 120.0 12.5 38 254 112 −7 ωc & C=O 107.5 12.5 20 1 −10 −24 0.5–1.0 ωc 100.0 18.0 22 −2 −28 −31 C=O 105.0 16.0 20 28 −3 −21 S=O 120.0 12.8 29 339 142 1 ωc & C=O 100.0 12.5 24 7 −16 −26 MN 0–0.5 ωc 100.8 12.5 27 7 −16 −26 C=O 100.0 25.0 36 15 -6 −21 S=O 118.5 25.0 37 219 118 −1 ωc & C=O 100.0 12.5 31 2 −11 −21 0.5–1.0 ωc 100.0 12.5 20 3 −10 −21 C=O 100.0 12.5 20 3 −10 −21 S=O 106.0 13.6 20 76 66 −2 ωc & C=O 100.0 12.5 20 3 −10 −21 Table 48. Summary of optimization of PAV operating parameters 9–10 years of in-service aging. a. Crossover frequency. Figure 74. Comparison of optimized PAV-conditioned and field-aged properties. (continued on next page)

106 Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging b. Carbonyl absorbance. Figure 74. (Continued). the top slice, however, are higher: 107.5°C for the Arizona site and 103.9°C for the Minnesota site. The PAV data at these optimum conditions are compared to the field data in Figure 75. These operating parameters also provide a relatively unbiased estimate of field aging based on the change in the logarithm of the crossover frequency; however, they produce conditioned binders with lower carbonyl absorbance compared to the field binders. There is little difference in the spread of the data when the mass of the binder is constrained to 12.5 g, and the condition- ing time is constrained to 20 hours. Long-Term Conditioning Calibration Experiment Objective The objective of the long-term conditioning calibration experiment was to determine appro- priate temperatures for 12.5 g, 20-hr, 2.1 MPa PAV conditioning for various climates. The basic approach consisted of comparing rheological and chemical properties of laboratory-conditioned Site Slice, in Criteria PAV Operating Parameters Average Relative Error, % Temperature, °C Mass, g Time, hr ∆log ωc ∆(C=O) ∆(S=O) AZ 0–0.5 C=O & ωc 107.5 12.5 20 1 −10 −24 AZ 0.5–1.0 C=O & ωc 102.1 12.5 20 7 −18 −26 MN 0–0.5 C=O & ωc 103.9 12.5 20 −2 −1 −21 MN 0.5–1.0 C=O & ωc 100.0 12.5 20 3 −10 −21 Table 49. Summary of constrained optimization of PAV operating parameters.

Research Approach 107   a. Crossover frequency. b. Carbonyl absorbance. Figure 75. Comparison of constrained, optimized PAV-conditioned and field-aged properties.

108 Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging binders to those for binder recovered from cores from in-service pavements. The rheological properties used were the change in the Christensen-Anderson master curve parameters from the (RTFOT) residue. The chemical properties were the change in the carbonyl and sulfoxide absorbances from the RTFOT residue. Experimental Design Effect of Storage Aging The pavements used in the long-term conditioning calibration experiment were pavements from the LTPP program that had original binder stored in the LTPP Materials Reference Library (MRL) and cores that were taken 7 to 16 years after construction. The cores were stored in the MRL for 12 to 16 years, leading to concern that the cores may have aged significantly during stor- age. Before conducting the long-term calibration experiment, a storage aging experiment was conducted using 12 in diameter cores from two of the pavements. This experiment investigated whether an aging gradient was present, moving from the edge of the core to the middle. Blocks 2.4 in by 2.4 in by 1.0 in thick were cut along the diameter of the core as shown in Figure 76. The blocks were cut into 0.5 in thick slices. The binder in each slice was recovered, and DSR fre- quency sweep tests were performed to determine Christensen-Anderson master curve param- eters. FTIR testing was conducted to determine carbonyl and sulfoxide absorbance. Details of the test methods were described earlier. A gradient from the edge to the middle in the slices is an indication that aging occurred during storage. An approximate magnitude of the aging during storage was determined by comparing the gradient to the change in properties measured on the original binder after RTFOT conditioning. This testing was performed on two cores with the properties summarized in Table 50. The results are summarized in Table  51 for the Mississippi core and Table  52 for the New York core. Figure 77 and Figure 78 show plots of carbonyl absorbance, carbonyl plus 12 in 2.4 in 2.4 in Figure 76. Schematic of testing to determine whether significant aging occurred during storage.

Research Approach 109   State Age Storage Time, yrs MAAT, °C Air Voids, % VMA, % VBE, % Binder Grade MS 11.9 11.7 17.0 3.7 15.5 11.8 AC-30 NY 13.2 11.9 8.6 13.3 23.7 10.4 AC-10 Table 50. Cores for studying storage aging. Depth, in Offset from Center, in FTIR Absorbance Master Curve Parameters C=O S=O C=O + S=O Td C° log ωc R 0.25 4.8 0.4302 0.6188 1.0490 −5.03 0.12 2.32 0.25 2.4 0.4390 0.6227 1.0617 −4.63 0.00 2.34 0.25 0.0 0.4339 0.5849 1.0188 −5.37 −0.01 2.35 0.25 2.4 0.4512 0.6247 1.0759 −5.69 0.07 2.35 0.25 4.8 0.4590 0.6325 1.0915 −4.39 −0.29 2.40 0.75 4.8 0.3058 0.5650 0.8708 −7.59 1.28 1.99 0.75 2.4 0.3577 0.5763 0.9340 −5.73 0.94 2.06 0.75 0.0 0.3560 0.5718 0.9278 −6.37 1.21 2.04 0.75 2.4 0.2618 0.5553 0.8171 −8.07 1.63 1.91 0.75 4.8 0.3132 0.5660 0.8792 −5.91 1.20 2.02 1.25 4.8 0.3418 0.5716 0.9134 −6.71 1.36 1.95 1.25 2.4 0.3460 0.5696 0.9156 −6.81 1.31 1.95 1.25 0.0 0.3503 0.5677 0.9180 −7.51 1.19 2.04 1.25 2.4 0.3410 0.5673 0.9083 −7.75 1.33 1.98 1.25 4.8 0.2776 0.5567 0.8343 −8.01 1.60 1.92 Table 51. FTIR and master curve data from the storage aging study for the Mississippi section. Depth, in Offset from Center, in FTIR Absorbance Master Curve Parameters C=O S=O C=O + S=O Td C° log ωc R 0.25 4.8 0.4467 0.6319 1.0786 5.14 −1.01 2.14 0.25 2.4 0.4610 0.6379 1.0989 2.98 −0.59 2.22 0.25 0.0 0.4599 0.6370 1.0969 −4.30 −0.17 2.34 0.25 2.4 0.4943 0.6586 1.1529 −3.83 −0.30 2.34 0.25 4.8 0.5275 0.6713 1.1988 −5.14 −0.13 2.34 0.75 4.8 0.4377 0.6029 1.0406 −5.13 0.08 2.18 0.75 2.4 0.4098 0.5960 1.0058 −4.76 0.20 2.19 0.75 0.0 0.4023 0.5928 0.9951 −4.95 0.60 2.17 0.75 2.4 0.4395 0.6073 1.0468 −5.48 0.51 2.20 0.75 4.8 0.4561 0.6152 1.0713 −5.19 0.32 2.27 1.25 4.8 0.3903 0.5816 0.9719 −7.05 0.70 2.11 1.25 2.4 0.3617 0.5732 0.9349 −5.55 0.78 2.14 1.25 0.0 0.3538 0.5744 0.9282 −5.88 1.06 2.06 1.25 2.4 0.3698 0.5784 0.9482 −5.55 0.78 2.14 1.25 4.8 0.4117 0.5996 1.0113 −6.05 0.73 2.17 Table 52. FTIR and master curve data from the storage aging study for the New York section.

a. Carbonyl absorbance. b. Carbonyl plus sulfoxide absorbance. c. Crossover frequency. d. Rheological index. 1.50 1.60 1.70 1.80 1.90 2.00 2.10 2.20 2.30 2.40 2.50 0.0 1.0 2.0 3.0 4.0 5.0 R Offset From Center, in 0.25 in Depth 0.75 in Depth 1.25 in Depth Linear (0.25 in Depth) Linear (0.75 in Depth) Linear (1.25 in Depth) Figure 77. Depth and offset effects for Mississippi core.

a. Carbonyl absorbance. b. Carbonyl plus sulfoxide absorbance. c. Crossover frequency. d. Rheological index. Figure 78. Depth and offset effects for New York core.

112 Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging sulfoxide absorbance, crossover frequency, and rheological index as a function of offset from the center of the core for the Mississippi and New York cores, respectively. These figures show somewhat different results. There is not a clear effect of offset from the center for the Mississippi section. The New York section, however, shows a clear offset effect for the 0.25 in and 0.75 in data. Both cores were stored for approximately the same length of time. A harder binder, AC-30, was used in the Mississippi section compared to an AC-10 for the New York section. Addition- ally, the air void content for the New York section was high at 13.3 percent. The air void content of the Mississippi section was 3.7 percent. Linear regression analyses were conducted to confirm the trends shown in Figure 78 and Figure 79. Dummy variables were included in the analysis to permit the intercept and slopes to vary with depth. The best fit models with statistically significant variables are summarized in Table 53. The best fit models have intercepts that change with depth; however, the offset slope does not vary with depth. For the Mississippi core, the offset slope is zero, indicating no offsetting effect. For the New York core, there is a significant offset effect for the carbonyl absorbance, the carbonyl plus sulfoxide absorbance, and the crossover frequency. This indicates that some aging occurs during storage for cores with higher air void contents. Table 54 compares the change in the property across the New York core to the change in the property from the RTFOT condition for the binder used in the core. This analysis shows the gradient accounts for approximately 0 to 20 percent of the property change for high air void content cores. Therefore, some additional aging likely occurred during storage in the higher air void content cores. To minimize this effect, the center portion of the core was used in the long-term conditioning calibration experiment. Figure 79. Geographical distribution of LTPP sections used in the long-term conditioning calibration experiment.

Research Approach 113   Property Core Depth, in Intercept Offset Slope Offset from Center, in Explained Variance C=O Absorbance MS 0.25 0.4427 0 0.810.75 0.3189 0 1.25 0.3313 0 NY 0.25 0.4536 0.0084 0.850.75 0.4048 0.0084 1.25 0.3532 0.0084 C=O + S=O Absorbance MS 0.25 1.0593 0 0.840.75 1.3357 0 1.25 0.8978 0 NY 0.25 1.0915 0.0117 0.870.75 0.9982 0.0117 1.25 0.9252 0.0117 Log ωc MS 0.25 −0.02 0 0.930.75 1.25 0 1.25 1.36 0 NY 0.25 −0.22 −0.075 0.890.75 0.56 −0.075 1.25 1.02 −0.075 R MS 0.25 2.35 0 0.950.75 2.00 0 1.25 1.96 0 NY 0.25 2.28 0 0.540.75 2.21 0 1.25 2.12 0 Table 53. Summary of statistical analysis from the storage aging study. Property Depth, in Aged Value RTFOT Value Delta Aging Gradient, in Delta Gradient % of Change from RTFOT C=O 0.25 0.4536 0.1279 0.3257 0.0084 0.0403 12.4 0.75 0.4048 0.1279 0.2769 0.0084 0.0403 14.6 1.25 0.3532 0.1279 0.2253 0.0084 0.0403 17.9 C=O + S=O 0.25 1.0915 0.4284 0.6631 0.0117 0.0562 8.5 0.75 0.9982 0.4284 0.5698 0.0117 0.0562 9.9 1.25 0.9252 0.4284 0.4968 0.0117 0.0562 11.3 Log ωc 0.25 −0.22 2.77 −2.99 −0.075 −0.3600 12.0 0.75 0.56 2.77 −2.21 −0.075 −0.3600 16.3 1.25 1.02 2.77 −1.75 −0.075 −0.3600 20.6 R 0.25 2.28 1.72 0.56 0 0 0 0.75 2.21 1.72 0.51 0 0 0 1.25 2.12 1.72 0.40 0 0 0 Table 54. Aging gradient effect in the New York core. Main Calibration Experiment Design Table 55 list the pavements that were used in the long-term conditioning calibration experi- ment. These pavements were in 23 U.S. states and two Canadian provinces and cover a wide range of climates. The mean annual air temperatures during the period from construction to coring ranged from approximately 1.7°C to 20.0°C. Figure 79 shows the geographical distribu- tion of the pavements that were used.

114 Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging The selected pavements include new, full-depth construction and overlays of existing asphalt concrete and Portland cement concrete pavements. The thickness of the asphalt concrete placed during the LTPP construction ranged from 3.7 in to 10.5 in. Recovered binder properties were measured for three 0.5 in thick slices: 0 to 0.5 in; 0.5 in to 1.0 in; and 1.5 in to 2.0 in. These depths were selected based on the depth effect in the Christensen modification of the Witczak global aging system model (Advanced Asphalt Technologies LLC 2004). This depth effect is shown in Figure 80. At the midpoint of the slices, the relative aging is 1.0 for the top slice, 0.35 for the second slice, and 0.16 for the fourth slice. The three recovered binders at each site were subjected to DSR frequency sweep testing to determine Christensen-Anderson master curve parameters and FTIR testing to determine the carbonyl and sulfoxide absorbances. The cores from the MRL were 10 in to 14 in in diameter. Based on the findings of the storage aging experiment, the recov- ered binder testing was conducted on 4-in-diameter cores removed from the middle of the larger core. Bulk specific gravity (AASHTO T 166), maximum specific gravity (AASHTO T 209), asphalt content (AASHTO T 164), and extracted aggregate gradations (AASHTO T 30) were measured on the remaining ring portion. In a few cases where the LTPP aggregate specific gravities appeared to be in error, aggregate specific gravity measurements (AASHTO T 84 and AASHTO T 85) were made on the extracted aggregate. No. LTPP ID State/Province Age, yrs Region MMAT, °C Binder Type Storage Time, yrs 1 010102 AL 12.4 Wet, no freeze 17.9 AC-20 14.0 2 050804 AR 7.3 Wet, no freeze 17.8 Unknown 14.4 3 060603 CA 13.5 Wet, no freeze 9.1 AR-4000 13.8 4 170603 IL 15.9 Wet, freeze 11.1 Unknown 13.3 5 180603 IN 12.9 Wet, freeze 10.0 Unknown 16.3 6 190108 IA 14.5 Wet, freeze 11.2 Unknown 12.3 7 230504 ME 9.1 Wet, freeze 6.1 Unknown 15.0 8 240507 MD 13.2 Wet, no freeze 11.4 AC-20 14.2 9 270504 MN 14.7 Wet, freeze 4.2 Pen 85/100 14.2 10 280805 MS 11.9 Wet, no freeze 17.0 AC-30 11.7 11 290507 MO 7.6 Wet, freeze 14.8 Unknown 13.3 12 300806 MT 9.3 Dry, freeze 1.7 Pen 85/100 12.2 13 340507 NJ 13.6 Wet, freeze 11.9 AC-20 13.4 14 350802 NM 9.4 Dry, no freeze 17.2 AC-10 13.3 15 360802 NY 13.2 Wet, freeze 8.6 Unknown 11.9 16 370802 NC 8.7 Wet, no freeze 16.8 Unknown 12.9 17 400603 OK 14.1 Wet, no freeze 15.4 AC-20 12.9 18 420608 PA 12.7 Wet, freeze 8.5 AC-20 14.2 19 460804 SD 14.2 Dry, freeze 7.7 Pen 120/150 12.0 20 480802 TX 11.0 Wet, no freeze 20.0 AC-20 12.1 21 550806 WI 7.7 Wet, freeze 5.8 Unknown 14.0 22 810504 AB 15.7 Dry, freeze 3.2 Pen 200/300 13.1 23 830504 MB 14.7 Wet, freeze 2.9 Pen 150/200 12.9 24 06A806 CA 7.8 Dry, no freeze 18.1 AR 4000 12.2 25 48A504 TX 15.9 Wet, no freeze 18.9 AC-10 with latex 12.0 Not previously defined, in alphabetical order: AL = Alabama, IN = Indiana, IA = Iowa, ME = Maine, MD = Maryland, PA = Pennsylvania, AB = Alberta, Canada, MB = Manitoba, Canada Table 55. Summary of LTPP sections to be used in the long-term conditioning calibration experiment.

Research Approach 115   To determine appropriate PAV conditioning temperatures for each pavement, the original binder samples were first RTFOT-conditioned. The RTFOT-conditioned residue was then further conditioned in the PAV using 12.5 g in standard PAV pans, 20 hours conditioning time, and 2.1 MPa air pressure. Three temperatures were used: 85°C, 100°C, and 115°C. Rheological and chemical properties of the original binder, RTFOT residue, and the residue from the three PAV conditioning temperatures were measured. DSR frequency sweep testing was conducted to determine the Christensen-Anderson master curve, and FTIR testing was conducted to determine carbonyl and sulfoxide absorbance. The master curve, carbonyl absorbance, and sulfoxide absorbance data were analyzed to determine equivalent PAV conditioning tempera- tures for each slice for each pavement. Statistical analysis of the equivalent PAV temperatures was performed to develop a model of the equivalent PAV conditioning temperature as a function of pavement age, climate, volumetric properties, depth, and binder properties. This model was then used to recommend PAV conditioning temperatures for AASHTO M 320 and AASHTO M 332 grading. Results and Analysis Long-Term Conditioning Calibration Database Workbook An Excel workbook for the long-term conditioning calibration experiment was assembled to organize the data for statistical analysis. The workbook has 11 spreadsheets. The main elements of these worksheets are discussed below. Location Spreadsheet. Table 56 shows the information contained in the location spread- sheet. Most of these data were extracted from the LTPP InfoPave database (FHWA 2017). The LTPP Section ID can be used to access additional information for each pavement section. The elevation data were obtained from Google Earth for the longitude and latitude from LTPP InfoPave (Google Earth 2020). Two of the sections, Montana and Manitoba, received chip seals before the cores were extracted. The time in service for these sections was calculated from the Figure 80. Depth effect from the Christensen modification of the Witczak global aging system model.

State/ Prov- ince LTPP ID Route Construction Date Chip Seal Date Coring Date Time in Service, yrs Time in Storage, yrs Latitude Longitude Elevation, ft AL 010102 US-280 westbound 03/01/1993 07/21/2005 12.4 14.0 32.63570 −85.29572 680 AR2 050804 US-65 southbound 12/01/1997 03/01/2005 7.3 14.4 34.19870 −91.96281 210 CA 060603 I-5 northbound 05/06/1992 10/20/2005 13.5 13.8 41.35826 −122.35965 3917 IL 170603 I-57 northbound 05/24/1990 04/28/2006 15.9 13.3 39.94639 −88.30302 690 IN 180603 US-31 northbound 06/11/1990 04/30/2003 12.9 16.3 41.18426 −86.24781 866 IA 190108 US-61 southbound 11/01/1992 04/27/2007 14.5 12.3 40.68581 −91.25078 662 ME 230504 I-95 northbound 06/15/1995 08/05/2004 9.1 15.0 45.06010 −68.68832 140 MD 240507 US-15 northbound 03/31/1992 06/03/2005 13.2 14.2 39.30141 −77.52040 379 MN 270504 US-2 westbound 09/15/1990 06/10/2005 14.7 14.2 47.52810 −95.19382 1469 MS 280805 State-315 northbound 01/01/1996 11/13/2007 11.9 11.7 34.44519 −89.87261 276 MO 290507 US-65 northbound 08/17/1998 03/31/2006 7.6 13.3 36.51644 −93.22952 1343 MT 300806 State-273 northbound 06/01/1994 09/01/2003 06/01/2007 9.3 12.2 46.14162 −112.89166 5091 NJ 340507 I-195 westbound 07/20/1992 03/05/2006 13.6 13.4 40.17780 −74.51911 123 NM 350802 I-10 frontage eastbound 11/01/1996 03/31/2006 9.4 13.3 32.19354 −108.29852 4573 NY 360802 State-947A eastbound 08/01/1994 09/27/2007 13.2 11.9 43.35552 −77.92370 253 NC 370802 1245 northbound 12/10/1997 08/25/2006 8.7 12.9 34.80880 −77.66349 70 OK 400603 I-35 southbound 08/08/1992 08/29/2006 14.1 12.9 36.71348 −97.34595 1021 PA 420608 I-80 westbound 09/03/1992 05/06/2005 12.7 14.2 40.98949 −77.81956 1267 SD 460804 State-1804 southbound 06/01/1993 08/01/2007 14.2 12.0 45.92797 −100.40881 1684 TX 480802 2223 eastbound 07/01/1996 06/27/2007 11.0 12.1 30.77167 −96.40077 386 WI 550806 State-29 eastbound 11/30/1997 08/12/2005 7.7 14.0 44.88162 −89.31584 1310 ALB 810504 Provincial-16 westbound 09/30/1990 06/16/2006 15.7 13.1 53.58277 −116.03169 2887 MAN 830504 Provincial-1 westbound 09/08/1989 06/01/2004 09/15/2006 14.7 12.9 49.65974 −96.29205 956 CA2 06A806 Sycamore St. northbound 09/01/1999 06/05/2007 7.8 12.2 37.41774 −120.76171 120 TX2 48A504 US-175 southbound 10/16/1991 08/21/2007 15.9 12.0 32.61340 −96.40476 455 Table 56. Long-term conditioning calibration database location spreadsheet.

Research Approach 117   construction date to the date the chip seal was placed. The chip seal was removed by diamond sawing and not included in the testing. Layer Thickness Spreadsheet. The pavements used in the long-term conditioning cali- bration experiment included new construction and overlays of asphalt and Portland cement concrete pavements. Table 57 shows the information contained in the layer thickness spread- sheet. This spreadsheet lists the type of construction and shows the thickness of the asphalt layers placed during the construction represented by the original binder samples from the MRL. The thickness of the new asphalt ranged from 3.7 in to 10.5 in. Climate Spreadsheet. Various climatic data are available for the LTPP sections through LTPP InfoPave. Table 58 and Table 59 present the climatic data used in the long-term calibra- tion experiment. These data were extracted from two locations in LTPP InfoPave using the National Aeronautics and Space Administration’s (NASA) MERRA climatic data. The annual precipitation, annual average air temperature, and annual average daily solar radiation data shown in Table 59 were extracted from the climate section of LTPP InfoPave and averaged over the time between construction and coring, or for the Montana and Manitoba sections, between construction and the application of chip seal treatments. The degree days, coldest air temperature, highest 7-day average air temperature, and 50 and 98 percent reliability pavement temperatures were extracted from the LTPPBind Online tool using the latitude and longitude State/ Province LTPP Section Construction Type New Asphalt Layer Thickness, in Surface 2 3 4 AL 010102 New construction 1.4 2.8 AR2 050804 New construction 1.6 5.7 CA 060603 Overlay of PCC 4.8 IL 170603 Overlay of PCC 1.5 2.2 IN 180603 Overlay of PCC 0.7 3.0 IA 190108 New construction 2.0 4.1 4.5 ME 230504 Overlay of AC 2.0 3.7 MD 240507 Overlay of AC 2.0 4.7 MN 270504 Overlay of AC 2.0 3.3 MS 280805 New construction 2.0 2.0 MO 290507 Overlay of AC 2.0 4.5 MT 300806 New construction 6.9 NJ 340507 Overlay of AC 2.0 2.9 2.6 NM 350802 New construction 7.0 NY 360802 New construction 0.9 2.1 4.6 NC 370802 New construction 1.8 2.4 2.7 OK 400603 Overlay of PCC 4.0 PA 420608 Overlay of PCC 2.0 2.6 3.0 1.1 SD 460804 New construction 7.1 TX 480802 New construction 2.5 4.0 WI 550806 New construction 2.0 5.0 ALB 810504 Overlay of AC 4.8 MAN 830504 Overlay of AC 5.6 CA2 06A806 New construction 6.8 TX2 48A504 Overlay of AC 2.1 2.9 Note: PCC = Portland cement concrete. Table 57. Long-term conditioning calibration database pavement section spreadsheet.

118 Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging of the section listed in Table 56. The pavement temperatures were calculated for the surface of the pavement. Volumetric Property Spreadsheets. The database for the long-term conditioning calibra- tion experiment included: • Spreadsheet with aggregate and volumetric properties for the surface layer, and • Spreadsheet with aggregate and volumetric properties for the layer below the surface layer for six sections where the surface layer was less than 2 in thick. Aggregate specific gravity and absorption data were extracted from LTPP InfoPave. The properties that were measured included: 1. Bulk specific gravity, AASHTO T 166, 2. Maximum specific gravity, AASHTO T 209, 3. Asphalt content, AASHTO T 164, 4. Extracted aggregate gradation, AASHTO T 30, and 5. Aggregate specific gravity and absorption, AASHTO T 84 and AASHTO T 85, for sections where the aggregate specific gravity data in LTPP InfoPave appeared to be in error (IA, MD, and NC). State/ Province LTPP Section Avg. Annual Precipitation, mm Avg. Annual Air Temperature, °C Avg. Daily Shortwave Radiation, W/m2 AL 010102 1350.2 17.9 152915.8 AR2 050804 1408.1 17.8 151776.2 CA 060603 832.8 9.1 152712.7 IL 170603 1102.8 11.1 137973.0 IN 180603 1095.4 10.0 133629.3 IA 190108 1022.1 11.2 138028.7 ME 230504 1232.4 6.1 122946.2 MD 240507 1259.7 11.4 139732.9 MN 270504 813.0 4.2 124069.0 MS 280805 1395.4 17.0 151218.5 MO 290507 1015.7 14.8 150737.1 MT 300806 770.3 1.7 138636.8 NJ 340507 1325.1 11.9 136766.8 NM 350802 229.1 17.2 184250.5 NY 360802 1021.6 8.6 127251.8 NC 370802 1461.5 16.8 146039.8 OK 400603 809.9 15.4 155134.3 PA 420608 1351.9 8.5 130301.3 SD 460804 555.9 7.7 133909.5 TX 480802 1188.8 20.0 155431.0 WI 550806 1001.5 5.8 128559.0 ALB 810504 723.8 3.2 106073.6 MAN 830504 712.2 2.9 118510.4 CA2 06A806 696.4 18.1 165409.2 TX2 48A504 1069.9 18.9 154238.9 Table 58. Long-term conditioning calibration database climatic spreadsheet, LTPP InfoPave data.

Research Approach 119   Table 59. Long-term conditioning calibration database climatic spreadsheet, LTPPBind Online data. State/ Prov- ince LTPP Section Avg. Annual Degree Days Over 10°C April to September Temperature, °C Coldest Air Highest 7-Day Avg High Air 50% Reliability, High Pavement 98% Reliability, High Pavement 50% Reliability, Low Pavement 98% Reliability, Low Pavement AL 010102 3739.5 −17.2 37.6 61.1 65.4 −9.4 −15.4 AR2 050804 3867.9 −22.9 40.1 62.9 67.2 −13.9 −21.0 CA 060603 2350.5 −21.9 35.3 57.3 61.2 −15.5 −22.0 IL 170603 2778.3 −34.8 33.0 55.7 60.1 −24.3 −32.8 IN 180603 2561.0 −37.3 31.7 54.3 58.6 −26.4 −35.6 IA 190108 2865.9 −35.3 35.0 57.2 62.0 −24.8 −32.9 ME 230504 1847.8 −42.4 28.5 50.7 54.6 −31.4 −39.7 MD 240507 2827.0 −31.0 34.0 56.7 60.9 −21.4 −29.1 MN 270504 1995.9 −45.4 31.0 52.1 56.5 −34.5 −41.9 MS 280805 3594.9 −22.5 37.8 60.9 65.3 −13.8 −21.2 MO 290507 3302.5 −24.4 37.5 60.2 64.6 −15.7 −22.5 MT 300806 1088.5 −40.5 27.0 49.1 53.4 −30.4 −39.3 NJ 340507 2457.3 −22.4 31.0 54.0 58.1 −15.3 −21.9 NM 350802 3880.8 −13.5 38.9 62.3 66.2 −6.6 −11.4 NY 360802 1185.9 −19.2 24.8 48.0 52.0 −14.2 −20.0 NC 370802 3334.8 −20.0 34.8 58.3 62.4 −12.1 −18.9 OK 400603 3643.9 −23.4 39.5 61.9 66.2 −15.0 −21.4 PA 420608 2283.6 −29.4 31.1 53.9 58.2 −20.7 −28.5 SD 460804 2611.5 −37.3 35.2 56.0 60.3 −28.1 −36.1 TX 480802 4057.4 −16.5 39.8 63.2 67.4 −8.5 −14.3 WI 550806 1873.0 −42.7 28.3 50.5 54.8 −31.7 −39.3 ALB 810504 1356.0 −45.3 28.0 47.8 51.9 −36.9 −46.3 MAN 830504 1984.0 −45.5 31.4 51.9 56.1 −35.4 −42.1 CA2 06A806 4060.1 −6.1 42.7 64.4 68.2 −2.8 −7.8 TX2 48A504 3934.3 −17.4 39.7 62.9 67.1 −9.6 −15.4 Surface layer volumetric properties for aggregate and mixture properties are presented in Table 60 and Table 61, respectively. Table 62 and Table 63 present second layer volumetric properties—aggregate and mixture properties—for the layer below the surface layer for the six sections where the surface layer was less than 2 in thick. The cores used in the long-term calibration experiment were taken between the wheel paths where limited post-construction compaction occurred. The air void contents range from less than 1 percent to over 13 percent. Some of the mixtures have low effective binder contents. Keep in mind that the pavements used in the long-term calibration experiment were constructed before the adoption of AASHTO M 323 for designing asphalt mixtures and the widespread implementation of volumetric controls during construction. Recovered Binder Property Spreadsheets. Rheological and chemical properties of binder recovered from slices at three depths in each of the pavements from the long-term calibration experiment are presented in Table 64 and Table 65. The test and analysis methods used were described earlier. Table 64 presents the Christensen-Anderson master curve parameters and goodness of fit statistics using a reference temperature of 25°C. The master curve fitting was done using G* as the criterion; therefore, the explained variance for G* is somewhat better than that of the phase angle. Table 65 presents the carbonyl, sulfoxide, and carbonyl plus sulfoxide absorbance from the FTIR spectra.

120 Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging State/ Prov- ince LTPP Section Aggregate Bulk Specific Gravity Aggregate Water Absorption, % Aggregate Gradation, % Passing Sieve Size in mm, % Coarse Fine Com- bined Coarse Fine 25.0 19.0 12.5 9.5 4.75 2.36 1.18 0.60 0.30 0.15 0.075 AL 010102 2.625 2.728 2.673 0.48 0.22 100.0 98.4 90.4 80.0 60.1 46.5 35.1 25.8 17.1 10.5 6.0 AR2 050804 2.610 2.603 2.607 0.53 0.43 100.0 100.0 93.0 85.3 65.0 47.0 34.3 25.8 17.7 11.4 7.1 CA 060603 2.709 2.696 2.703 0.33 0.54 100.0 100.0 91.3 80.5 59.5 45.2 30.7 21.0 14.7 10.7 8.0 IL 170603 2.660 2.650 2.656 0.97 1.67 100.0 100.0 100.0 99.2 62.3 40.1 28.1 20.5 12.8 8.9 7.5 IN 180603 2.410 2.637 2.519 2.37 1.17 100.0 100.0 100.0 95.8 60.6 48.2 37.9 27.7 12.8 5.7 4.1 IA 190108 2.492 2.624 2.539 3.72 0.97 100.0 99.5 82.3 73.8 52.9 35.9 27.3 19.9 9.9 6.8 6.1 ME 230504 2.680 2.677 2.679 0.60 0.60 100.0 100.0 92.2 81.3 49.9 35.8 24.8 17.2 12.4 9.7 7.7 MD 240507 2.662 2.676 2.666 1.09 0.52 100.0 100.0 97.8 88.9 48.9 31.9 23.2 17.6 11.8 8.8 7.4 MN 270504 2.683 2.643 2.661 0.70 0.70 100.0 100.0 91.6 82.3 68.1 56.9 43.8 27.9 12.3 6.4 4.3 MS 280805 2.550 2.633 2.585 0.87 0.60 100.0 100.0 100.0 94.4 59.9 41.9 29.6 21.0 12.1 8.9 7.0 MO 290507 2.654 2.636 2.647 0.62 0.94 100.0 100.0 98.6 88.4 60.8 39.5 25.5 18.2 13.1 9.5 7.3 MT 300806 2.683 2.653 2.674 0.63 0.83 100.0 100.0 80.4 64.5 39.7 29.6 23.5 19.5 15.9 12.2 8.4 NJ 340507 2.910 2.827 2.873 0.60 1.03 100.0 97.0 85.1 76.4 55.7 45.0 36.1 28.7 18.4 10.3 6.3 NM 350802 2.223 2.443 2.313 4.10 1.97 100.0 97.7 82.0 72.7 52.8 40.9 29.2 20.7 14.2 10.1 7.3 NY 360802 2.637 2.600 2.619 0.47 1.07 100.0 100.0 100.0 99.4 83.0 49.1 32.3 23.7 17.5 12.8 9.7 NC 370802 2.415 2.637 2.549 3.72 0.60 100.0 100.0 100.0 94.9 75.8 60.2 50.6 44.8 34.4 9.8 5.9 OK 400603 2.590 2.573 2.584 1.75 0.80 100.0 100.0 96.6 91.9 61.1 35.5 23.9 17.2 11.7 8.6 7.2 PA 420608 2.663 2.737 2.697 1.17 1.00 100.0 100.0 100.0 99.4 64.3 46.2 28.4 18.4 12.6 9.5 7.2 SD 460804 2.633 2.580 2.608 0.60 1.37 100.0 100.0 92.5 81.6 60.6 47.3 35.8 25.3 13.0 7.0 4.8 TX 480802 2.513 2.540 2.524 0.87 1.40 100.0 100.0 99.7 92.1 65.1 41.7 31.0 25.2 18.2 11.6 7.0 WI 550806 2.630 2.633 2.632 0.67 0.90 100.0 100.0 97.4 94.3 77.2 54.8 39.0 27.1 13.8 8.1 6.0 ALB 810504 2.620 2.587 2.605 0.67 1.37 100.0 100.0 91.4 80.0 59.5 43.5 34.2 28.2 19.3 11.9 7.7 MAN 830504 2.603 2.623 2.614 1.13 1.13 100.0 100.0 93.4 83.5 64.3 52.8 42.7 32.8 22.1 10.3 6.0 CA2 06A806 2.650 2.613 2.638 0.70 0.69 100.0 96.0 75.5 65.8 43.2 31.9 25.6 21.3 14.7 8.7 5.3 TX2 48A504 2.827 2.620 2.735 0.57 0.93 100.0 100.0 87.1 77.6 54.4 44.5 36.9 31.3 24.9 11.8 6.0 Table 60. Long-term conditioning calibration database, surface layer volumetric properties spreadsheet, aggregate properties. State/ Prov- ince LTPP Section Specific Gravity Asphalt Content, wt % Voids in Mineral Aggregate, vol % Voids Filled With Asphalt, vol % Air Voids, vol % Effective Binder Content, vol %Bulk Maximum Effective AL 010102 2.384 2.465 2.692 5.71 15.9 79.3 3.3 12.6 AR2 050804 2.286 2.436 2.630 5.12 16.8 63.3 6.2 10.6 CA 060603 2.303 2.543 2.710 4.03 18.2 48.2 9.4 8.8 IL 170603 2.286 2.505 2.701 4.82 18.1 51.6 8.7 9.3 IN 180603 2.300 2.477 2.720 5.97 14.2 49.5 7.1 7.0 IA 190108 2.292 2.495 2.661 4.21 13.5 39.9 8.1 5.4 ME 230504 2.357 2.475 2.685 5.29 16.7 71.4 4.8 11.9 MD 240507 2.416 2.507 2.727 5.32 14.2 74.5 3.6 10.6 MN 270504 2.257 2.454 2.665 5.42 19.8 59.4 8.0 11.7 MS 280805 2.311 2.399 2.600 5.49 15.5 76.3 3.7 11.8 MO 290507 2.359 2.495 2.682 4.67 15.0 63.8 5.5 9.6 MT 300806 2.400 2.532 2.708 4.26 14.1 63.0 5.2 8.9 Table 61. Long-term conditioning calibration database, surface layer volumetric properties spreadsheet, mixture properties.

Research Approach 121   LTPP Section Aggregate Bulk Specific Gravity Aggregate Water Absorption, % Aggregate Gradation, % Passing Sieve Size in mm, % Coarse Fine Combined Coarse Fine 25.0 19.0 12.5 9.5 4.75 2.36 1.18 0.60 0.30 0.15 0.075 AL 010102 2.815 2.668 2.776 0.33 0.48 100.0 88.4 67.1 53.5 31.8 26.2 23.0 18.4 11.9 7.6 5.4 AR2 050804 2.607 2.600 2.604 0.50 0.47 97.9 85.5 70.2 61.7 45.5 34.1 25.9 20.2 14.3 9.3 5.8 CA 060603 IL 170603 2.610 2.593 2.605 1.07 1.40 100.0 98.7 89.2 82.2 50.2 31.5 21.7 16.1 11.8 8.6 7.2 IN 180603 2.430 2.640 2.501 2.43 1.37 100.0 98.9 81.4 70.2 45.8 34.0 27.0 20.3 11.0 6.1 4.2 IA 190108 ME 230504 MD 240507 MN 270504 MS 280805 MO 290507 MT 300806 NJ 340507 NM 350802 NY 360802 2.657 2.610 2.639 0.40 1.07 100.0 96.4 91.3 84.8 59.5 38.9 27.8 21.0 15.1 10.2 7.4 NC 370802 2.415 2.587 2.481 3.72 1.20 100.0 97.3 79.7 68.0 47.0 38.4 32.1 26.4 17.3 7.1 4.0 OK 400603 PA 420608 SD 460804 TX 480802 WI 550806 ALB 810504 MAN 830504 CA2 06A806 TX2 48A504 State/ Prov- ince Table 62. Long-term conditioning calibration database, second layer volumetric properties spreadsheet, aggregate properties. State/ Prov- ince LTPP Section Specific Gravity Asphalt Content, wt % Voids in Mineral Aggregate, vol % Voids Filled With Asphalt, vol % Air Voids, vol % Effective Binder Content, vol %Bulk Maximum Effective NJ 340507 2.596 2.656 2.876 4.63 13.8 83.6 2.3 11.6 NM 350802 2.038 2.252 2.469 6.89 18.0 47.1 9.5 8.5 NY 360802 2.118 2.443 2.664 5.71 23.7 44.0 13.3 10.4 NC 370802 2.153 2.392 2.627 6.34 20.9 52.1 10.0 10.9 OK 400603 2.287 2.454 2.619 4.37 15.4 55.7 6.8 8.6 PA 420608 2.459 2.512 2.751 5.70 14.0 85.0 2.1 11.9 SD 460804 2.319 2.443 2.659 5.59 16.1 68.4 5.1 11.0 TX 480802 2.291 2.450 2.631 4.76 13.6 52.2 6.5 7.1 WI 550806 2.294 2.441 2.678 6.06 18.1 66.8 6.0 12.1 ALB 810504 2.360 2.467 2.636 4.40 13.4 67.7 4.3 9.1 MAN 830504 2.330 2.454 2.658 5.27 15.6 67.5 5.1 10.5 CA2 06A806 2.289 2.510 2.676 4.13 16.8 47.7 8.8 8.0 TX2 48A504 2.524 2.546 2.765 5.10 12.4 93.0 0.9 11.5 Table 61. (Continued).

122 Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging State/ Prov- ince LTPP Section Specific Gravity Asphalt Content, wt % Voids in Mineral Aggregate, vol % Voids Filled with Asphalt, vol % Air Voids, vol % Effective Binder Content, vol %Bulk Maximum Effective AL 010102 2.542 2.614 2.795 4.05 12.1 77.3 2.8 9.4 AR2 050804 2.341 2.465 2.628 4.27 14.0 63.9 5.0 8.9 CA 060603 IL 170603 2.306 2.491 2.679 4.72 15.6 52.5 7.4 8.2 IN 180603 2.148 2.492 2.681 4.73 18.2 24.1 13.8 4.4 IA 190108 ME 230504 MD 240507 MN 270504 MS 280805 MO 290507 MT 300806 NJ 340507 NM 350802 NY 360802 2.294 2.532 2.733 4.81 17.2 45.5 9.4 7.8 NC 370802 2.115 2.389 2.570 5.06 19.1 39.8 11.5 7.6 OK 400603 PA 420608 SD 460804 TX 480802 WI 550806 ALB 810504 MAN 830504 CA2 06A806 TX2 48A504 Table 63. Long-term conditioning calibration database, second layer volumetric properties spreadsheet, mixture properties. Lab Conditioned Binder Master Curve Spreadsheets. Rheological and chemical proper- ties were measured for the original binder and residue from various laboratory conditioning. The laboratory conditioning included: • RTFOT, • Standard PAV (50.0 g, 20-hr, 2.1 MPa at 100°C or 90°C depending on climate), • 12.5 g, 20-hr, 2.1 MPa, 85°C, • 12.5 g, 20-hr, 2.1 MPa, 100°C, and • 12.5 g, 20-hr, 2.1 MPa, 115°C. The Christensen-Anderson master curve parameters are presented in Table 66 for the original binder, RTFOT-conditioned residue, and standard PAV-conditioned residue. The Christensen- Anderson master curve parameters for 12.5 g, 20-hr PAV conditioning at the three temperatures are presented in Table 67. The carbonyl, sulfoxide, and carbonyl plus sulfoxide absorbance from the FTIR spectra are shown in Table 68 for the original binder, RTFOT-conditioned residue, and standard PAV-conditioned residue, and Table 69 for 12.5 g, 20-hr PAV conditioning at three temperatures. Equivalent 12.5 g, 20-hr PAV Conditioning Temperatures Rheological-Based Properties. One of the issues identified during the PAV operating parameters experiment was the Christensen-Anderson master curve parameters matched the laboratory-conditioned master curve parameters at different temperatures. It was not clear which temperature yielded the best agreement for the master curves, so an average was used.

State/ Prov- ince LTPP Section 0.0 to 0.5 in Slice 0.5 in to 1.0 in Slice 1.5 in to 2.0 in Slice Td °C ωc rad/sec R G* r2 δ r 2 Td °C ωc rad/sec R G* r2 δ r 2 Td °C ωc rad/sec R G* r2 δ r 2 AL 010102 −0.39 0.20 2.97 0.998 0.946 −8.15 3.27 2.50 0.997 0.985 −11.01 9.73 2.42 0.998 0.986 AR2 050804 1.41 1.25 2.19 0.997 0.943 −7.77 2.32 2.20 0.995 0.965 -8.76 3.38 2.20 0.998 0.967 CA 060603 5.71 0.26 1.98 0.998 0.986 5.63 0.50 1.95 0.998 0.986 5.09 0.62 1.98 0.998 0.983 IL 170603 4.35 0.21 2.37 0.997 0.954 5.78 0.29 2.15 0.997 0.964 5.83 0.53 2.10 0.998 0.967 IN 180603 1.62 1.10 2.22 0.998 0.974 1.15 8.75 2.49 0.996 0.960 −6.80 2.11 2.21 0.995 0.972 IA 190108 1.95 1.21 2.49 0.997 0.933 3.19 2.44 2.31 0.998 0.946 3.25 2.13 2.32 0.997 0.946 ME 230504 −12.73 84.59 2.10 0.998 0.985 −13.07 184.66 1.99 0.998 0.980 −13.34 228.89 2.00 0.998 0.979 MD 240507 −11.34 3.02 2.35 0.998 0.993 −10.66 10.64 2.21 0.998 0.993 −11.20 17.65 2.23 0.998 0.989 MN 270504 −7.86 3.30 2.37 0.996 0.964 -6.61 6.81 2.23 0.995 0.973 -6.60 5.82 2.24 0.996 0.972 MS 280805 1.37 0.69 2.27 0.998 0.977 −7.57 7.06 2.03 0.999 0.992 −8.50 11.82 1.94 0.998 0.987 MO 290507 1.12 0.37 2.32 0.998 0.980 2.13 1.16 2.28 0.999 0.982 −7.49 2.02 2.28 0.998 0.987 MT 300806 −7.27 19.64 2.10 0.996 0.967 −9.60 222.33 1.87 0.997 0.978 −6.84 32.10 2.07 0.996 0.968 NJ 340507 −3.04 3.23 2.53 0.998 0.951 −7.12 3.22 2.12 0.998 0.992 −6.92 3.56 2.06 0.998 0.990 NM 350802 −7.71 2.15 2.72 0.996 0.959 −8.09 5.49 2.64 0.996 0.960 −9.48 10.33 2.54 0.996 0.955 NY 360802 4.24 0.08 2.11 0.999 0.995 2.15 1.16 2.04 0.998 0.979 −8.79 5.85 2.05 0.997 0.979 NC 370802 1.43 0.26 2.63 0.999 0.980 −1.42 0.79 2.56 0.999 0.978 -0.55 1.11 2.50 0.999 0.980 OK 400603 10.71 0.02 2.56 0.997 0.906 3.95 0.17 2.34 0.996 0.940 3.42 1.28 2.22 0.997 0.943 PA 420608 −11.81 32.28 2.21 0.998 0.988 −10.27 37.19 2.11 0.998 0.989 −11.10 35.16 2.11 0.998 0.991 SD 460804 1.67 0.63 2.40 0.998 0.958 3.90 0.22 2.48 0.998 0.947 −6.28 1.87 2.38 0.996 0.972 TX 480802 4.87 0.13 2.33 0.996 0.923 5.51 0.64 2.20 0.996 0.920 5.04 0.86 2.17 0.995 0.919 WI 550806 −11.96 16.80 2.22 0.997 0.989 −12.41 48.07 2.10 0.998 0.983 −13.14 80.91 2.05 0.998 0.981 ALB 810504 −9.18 52.75 1.80 0.999 0.993 −12.56 233.51 1.70 0.999 0.995 −11.05 219.74 1.67 0.999 0.996 MAN 830504 −11.96 16.80 2.22 0.999 0.987 −12.41 48.07 2.10 0.999 0.991 −13.14 80.91 2.05 0.999 0.992 CA2 06A806 6.23 0.72 1.69 0.997 0.972 6.41 1.55 1.57 0.997 0.972 7.81 1.01 1.54 0.997 0.974 TX2 48A504 −12.90 27.47 2.49 0.996 0.961 −8.75 256.43 1.89 0.997 0.981 −8.00 91.23 1.94 0.997 0.974 Table 64. Long-term conditioning calibration database: recovered binder master curve spreadsheet, 25çC reference temperature.

124 Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging State/ Province LTPP Section 0 to 0.5 in Slice 0.5 in to 1.0 in Slice 1.5 in to 2.0 in Slice C=O S=O C=O + S=O C=O S=O C=O + S=O C=O S=O C=O + S=O AL 010102 0.4806 0.5503 1.0309 0.3546 0.5366 0.8912 0.2770 0.4405 0.7175 AR2 050804 0.5608 0.5244 1.0852 0.5034 0.5334 1.0368 0.4871 0.5416 1.0287 CA 060603 0.7462 0.8479 1.5941 0.6936 0.8049 1.4985 0.6722 0.6289 1.3011 IL 170603 0.5000 0.6636 1.1636 0.4457 0.6844 1.1301 0.4169 0.6315 1.0484 IN 180603 0.5066 0.5029 1.0095 0.4656 0.5240 0.9896 0.4273 0.5674 0.9947 IA 190108 0.5108 0.5270 1.0378 0.4690 0.5191 0.9881 0.4725 0.5167 0.9892 ME 230504 0.2977 0.4373 0.7350 0.2599 0.4449 0.7048 0.2539 0.4238 0.6777 MD 240507 0.4257 0.5603 0.9860 0.3347 0.5087 0.8434 0.3161 0.5058 0.8219 MN 270504 0.4140 0.5422 0.9562 0.3660 0.5264 0.8924 0.3773 0.5543 0.9316 MS 280805 0.4834 0.5913 1.0747 0.3793 0.6028 0.9821 0.3887 0.5128 0.9015 MO 290507 0.4996 0.5716 1.0712 0.4641 0.5539 1.0180 0.4673 0.5206 0.9879 MT 300806 0.3467 0.5223 0.8690 0.2139 0.4099 0.6238 0.3111 0.5253 0.8364 NJ 340507 0.6777 0.5661 1.2438 0.5862 0.6280 1.2142 0.5788 0.5846 1.1634 NM 350802 0.5111 0.3958 0.9069 0.4958 0.3983 0.8941 0.4774 0.3909 0.8683 NY 360802 0.5529 0.5937 1.1466 0.4601 0.5672 1.0273 0.4002 0.4651 0.8653 NC 370802 0.4123 0.5490 0.9613 0.3727 0.5069 0.8796 0.3745 0.4402 0.8147 OK 400603 0.7283 0.4449 1.1732 0.5547 0.4162 0.9709 0.4577 0.4058 0.8635 PA 420608 0.2690 0.4639 0.7329 0.2405 0.4563 0.6968 0.2435 0.4623 0.7058 SD 460804 0.5678 0.5103 1.0781 0.6473 0.5146 1.1619 0.5322 0.5203 1.0525 TX 480802 0.5232 0.4459 0.9691 0.4893 0.4462 0.9355 0.4786 0.4485 0.9271 WI 550806 0.4641 0.5054 0.9695 0.4139 0.5081 0.9220 0.3988 0.5049 0.9037 ALB 810504 0.4403 0.6700 1.1103 0.3796 0.6451 1.0247 0.3701 0.6323 1.0024 MAN 830504 0.4243 0.5536 0.9779 0.3907 0.5844 0.9751 0.3747 0.5871 0.9618 CA2 06A806 0.9293 0.5551 1.4844 0.8427 0.5424 1.3851 0.8802 0.5638 1.4440 TX2 48A504 0.3810 0.4320 0.8130 0.2040 0.3506 0.5546 0.2621 0.4170 0.6791 Table 65. Long-term conditioning calibration database: recovered binder FTIR spreadsheet. State/ Province LTPP Section Original RTFOT STD PAV Td, °C ωc, rad/sec R Td °C ωc, rad/sec R Temp, °C Td, °C ωc, rad/sec R AL 010102 −10.08 1438.68 1.92 −7.55 128.28 2.21 100 −5.20 5.42 2.61 AR2 050804 −8.12 2011.47 1.44 −6.65 401.53 1.67 100 −5.19 64.61 1.89 CA 060603 −9.43 6410.42 1.41 −7.09 1322.32 1.61 100 −3.94 99.63 1.88 IL 170603 −8.70 1913.37 1.50 −6.06 328.19 1.75 100 −3.71 15.48 2.13 IN 180603 −11.36 2893.13 1.58 −8.50 482.91 1.79 100 −5.71 15.76 2.22 IA 190108 −9.58 4638.85 1.55 −7.67 999.32 1.75 100 −4.83 29.80 2.23 ME 230504 −12.58 7049.62 1.64 −9.54 953.38 1.88 100 −7.41 79.32 2.25 MD 240507 −12.87 1951.14 1.83 −10.66 247.04 2.12 100 −6.88 9.02 2.51 MN 270504 −8.86 3506.82 1.78 −6.94 864.78 1.95 100 −4.03 21.26 2.40 MS 280805 −9.10 1840.46 1.49 −7.41 347.09 1.72 100 −5.56 32.99 2.01 MO 290507 −9.41 2496.77 1.39 −7.58 662.20 1.58 100 −5.86 57.28 1.87 MT 300806 −9.82 4865.83 1.52 −7.38 947.75 1.78 100 −4.41 48.97 2.15 NJ 340507 −12.12 4519.95 1.37 −9.89 1187.47 1.57 100 −5.79 104.90 1.83 NM 350802 −8.54 2294.59 1.88 −6.90 505.85 2.06 100 −5.54 31.29 2.50 Table 66. Long-term conditioning calibration database: original binder, residue from RTFOT, and standard PAV-conditioned residue master curve spreadsheet, 25çC reference temperature.

Research Approach 125   State/ Province LTPP Section 85°C 100°C 115°C Td, °C ωc, rad/sec R Td, °C ωc, rad/sec R Td, °C ωc, rad/sec R AL 010102 −5.12 8.24 2.54 −3.88 0.65 2.83 17.64 0.000020 3.65 AR2 050804 −4.94 71.61 1.88 −4.32 17.60 2.08 3.01 0.276191 2.58 CA 060603 −3.66 114.65 1.86 −1.64 6.25 2.15 14.14 0.000869 3.05 IL 170603 −3.28 14.48 2.11 −1.62 0.94 2.47 21.58 0.000006 3.30 IN 180603 −5.58 19.09 2.16 3.42 0.69 2.51 20.84 0.000009 3.33 IA 190108 −4.62 38.12 2.20 −2.79 3.18 2.50 15.18 0.000600 3.22 ME 230504 −7.00 113.27 2.17 −5.62 17.03 2.38 6.37 0.013702 3.20 MD 240507 −6.66 15.17 2.43 −4.92 1.15 2.71 21.02 0.000002 3.67 MN 270504 −3.40 35.84 2.32 −1.00 0.71 2.82 21.05 0.000042 3.51 MS 280805 −5.05 35.11 1.97 −3.79 4.06 2.24 14.74 0.000799 3.02 MO 290507 −5.04 58.96 1.85 −4.62 8.60 2.07 19.25 0.000965 2.66 MT 300806 −4.59 57.03 2.09 −3.06 4.36 2.42 12.60 0.001722 3.09 NJ 340507 −6.66 127.44 1.80 −4.82 15.99 2.06 14.02 0.000414 3.04 NM 350802 −6.08 55.97 2.40 −5.01 7.28 2.69 3.30 0.230757 3.18 NY 360802 −4.51 27.62 2.09 −2.97 1.25 2.50 21.55 0.000029 3.14 NC 370802 −5.72 6.12 2.49 2.17 0.24 2.90 23.09 0.000003 3.61 OK 400603 −2.64 4.81 2.52 5.27 0.64 2.69 14.06 0.001629 3.17 PA 420608 −6.90 25.64 2.34 −5.17 2.16 2.57 12.07 0.000361 3.42 SD 460804 −7.29 182.21 2.09 −5.65 17.35 2.38 11.08 0.002335 3.19 TX 480802 −1.77 5.84 2.16 7.94 0.74 2.30 14.66 0.000705 2.99 WI 550806 −6.86 71.97 2.11 −5.54 5.86 2.43 17.56 0.000089 3.28 ALB 810504 −8.36 541.51 1.70 −5.95 68.92 1.91 6.78 0.025790 2.71 MAN 830504 −10.26 188.81 2.09 −7.77 10.28 2.44 18.77 0.000013 3.58 CA2 06A806 1.94 73.99 1.39 2.25 11.58 1.55 7.33 0.255286 2.01 TX2 48A504 −3.64 47.03 2.14 −3.00 5.36 2.42 5.72 0.025447 3.05 Table 67. Long-term conditioning calibration database: 12.5 g, 20-hr PAV conditioning master curve spreadsheet, 25çC reference temperature. State/ Province LTPP Section Original RTFOT STD PAV Td, °C ωc, rad/sec R Td °C ωc, rad/sec R Temp, °C Td, °C ωc, rad/sec R NY 360802 −9.74 2744.08 1.51 −8.49 594.03 1.72 90 −4.87 72.21 1.98 NC 370802 −11.07 1359.11 1.85 −8.57 106.29 2.18 100 −6.19 4.32 2.56 OK 400603 −6.47 1418.90 1.74 −5.04 166.06 2.04 100 −2.59 3.93 2.54 PA 420608 −11.77 1665.93 1.77 −9.61 298.75 2.02 100 −7.24 18.52 2.36 SD 460804 −13.38 9098.73 1.59 −10.65 2039.77 1.81 100 −7.34 111.72 2.18 TX 480802 −4.48 1446.61 1.49 −3.09 214.30 1.70 100 −1.93 8.46 2.09 WI 550806 −13.47 8301.17 1.54 −10.48 1278.58 1.80 100 -6.63 55.95 2.20 ALB 810504 −17.66 31826.57 1.29 −13.51 5590.54 1.47 90 −10.54 1236.01 1.64 MAN 830504 −18.05 18481.28 1.50 −15.04 3404.23 1.76 100 −10.53 132.39 2.17 CA2 06A806 −1.81 2645.32 1.07 −0.89 832.75 1.18 100 1.10 104.40 1.36 TX2 48A504 −8.49 3444.04 1.64 −6.20 604.72 1.83 100 −4.11 37.67 2.18 Table 66. (Continued).

126 Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging A different approach was used for the data from the calibration experiment. First, the master curve parameters for the laboratory-conditioned binder were plotted as a function of temper- ature, as shown in Figure 81 for the Missouri binder. The bilinear relationships shown in Figure 81 were used in a constrained optimization to interpolate and, in a few cases, extrapolate to determine the appropriate master curve parameter as a function of temperature. The con- strained optimization used the measured data for the binder recovered from the field cores and the laboratory relationships for the master curve parameters to determine the temperature that provided a Christensen-Anderson master curve that best fit the measured data for the recovered binder. Subjective optimization was performed by varying the temperature to optimize the explained variance for the G* and phase angle master curves. Figure 81 compares the fitted master curves for the binder recovered from the top slice from the Missouri section. Goodness of fit statistics for all sections are summarized in Table 70, Table 71, and Table 72 for the three slices. As expected, the constrained fitting using the master curve temperature relationships from the 12.5 g, 20-hr PAV conditioning was poorer than the unconstrained fitting. How- ever, in most cases, the laboratory-constrained fitting provided reasonable master curves as shown in Figure 82. Conversely, for a few of the sections (California, New York, Oklahoma, and South Dakota), the fitting was poorer as shown in Figure 83. This indicates that PAV State/ Province LTPP Section Original RTFOT STD PAV C=O S=O C=O + S=O C=O S=O C=O + S=O Temp, °C C=O S=O C=O + S=O AL 010102 0.0953 0.2725 0.3678 0.1200 0.2878 0.4078 100 0.2216 0.3779 0.5995 AR2 050804 0.0831 0.2606 0.3437 0.1035 0.2772 0.3807 100 0.2112 0.3709 0.5821 CA 060603 0.1776 0.2792 0.4568 0.2162 0.2920 0.5082 100 0.3248 0.4330 0.7578 IL 170603 0.1148 0.2886 0.4034 0.1347 0.3117 0.4464 100 0.2297 0.4169 0.6466 IN 180603 0.5714 0.5871 1.1585 0.1543 0.3089 0.4632 100 0.2613 0.4117 0.6730 IA 190108 0.1201 0.2727 0.3928 0.1471 0.2906 0.4377 100 0.2575 0.3903 0.6478 ME 230504 0.0898 0.2661 0.3559 0.1121 0.2881 0.4002 100 0.2182 0.3768 0.5950 MD 240507 0.1027 0.2741 0.3768 0.1284 0.2991 0.4275 100 0.2272 0.4103 0.6375 MN 270504 0.1083 0.2662 0.3745 0.1323 0.2990 0.4313 100 0.2431 0.3874 0.6305 MS 280805 0.1289 0.2669 0.3958 0.1572 0.2903 0.4475 100 0.2566 0.3864 0.6430 MO 290507 0.2006 0.2789 0.4795 0.2336 0.2911 0.5247 100 0.3288 0.3895 0.7183 MT 300806 0.0901 0.2705 0.3606 0.1179 0.2969 0.4148 100 0.2363 0.3907 0.6270 NJ 340507 0.1984 0.2856 0.4840 0.2347 0.2978 0.5325 100 0.3310 0.4001 0.7311 NM 350802 0.2139 0.2405 0.4544 0.2449 0.2563 0.5012 100 0.3697 0.3083 0.6780 NY 360802 0.1072 0.2860 0.3932 0.1279 0.3005 0.4284 90 0.1945 0.3829 0.5774 NC 370802 0.1059 0.2669 0.3728 0.1388 0.2838 0.4226 100 0.2454 0.3815 0.6269 OK 400603 0.1017 0.2522 0.3539 0.1343 0.2701 0.4044 100 0.2686 0.3503 0.6189 PA 420608 0.1054 0.2833 0.3887 0.1300 0.2967 0.4267 100 0.2311 0.3818 0.6129 SD 460804 0.0981 0.2843 0.3824 0.1286 0.3058 0.4344 100 0.2408 0.4181 0.6589 TX 480802 0.1027 0.2886 0.3913 0.1302 0.3016 0.4318 100 0.2507 0.3856 0.6363 WI 550806 0.1039 0.3046 0.4085 0.1314 0.3126 0.4440 100 0.2418 0.4271 0.6689 ALB 810504 0.2388 0.2941 0.5329 0.2566 0.3302 0.5868 90 0.2879 0.4142 0.7021 MAN 830504 0.1454 0.3179 0.4633 0.1676 0.3336 0.5012 100 0.2593 0.4225 0.6818 CA2 06A806 0.2063 0.2976 0.5039 0.2548 0.2966 0.5514 100 0.3950 0.4247 0.8197 TX2 48A504 0.1067 0.2948 0.4015 0.1331 0.3073 0.4404 100 0.2506 0.3975 0.6481 Table 68. Long-term conditioning calibration database: original binder, residue from RTFOT, and standard PAV-conditioned residue FTIR spreadsheet.

Research Approach 127   Table 69. Long-term conditioning calibration database 12.5 g, 20-hr PAV conditioning FTIR spreadsheet. State/ Province LTPP Section 85°C 100°C 115°C C=O S=O C=O + S=O C=O S=O C=O + S=O C=O S=O C=O + S=O AL 010102 0.1898 0.4008 0.5906 0.2770 0.4600 0.7370 0.5220 0.5470 1.0690 AR2 050804 0.1823 0.3843 0.5666 0.2696 0.4274 0.6970 0.4681 0.4615 0.9296 CA 060603 0.3147 0.4576 0.7723 0.4097 0.5711 0.9808 0.6729 0.6118 1.2847 IL 170603 0.2149 0.4533 0.6682 0.3081 0.5159 0.8240 0.5714 0.5871 1.1585 IN 180603 0.2382 0.4396 0.6778 0.3353 0.5291 0.8644 0.6071 0.5733 1.1804 IA 190108 0.2291 0.4083 0.6374 0.3132 0.4533 0.7665 0.5628 0.5074 1.0702 ME 230504 0.1794 0.4008 0.5802 0.2569 0.4481 0.7050 0.4832 0.5161 0.9993 MD 240507 0.1946 0.4155 0.6101 0.2805 0.4939 0.7744 0.5928 0.5673 1.1601 MN 270504 0.2061 0.4216 0.6277 0.3008 0.5096 0.8104 0.5191 0.4917 1.0108 MS 280805 0.2249 0.4097 0.6346 0.3250 0.4693 0.7943 0.6551 0.5414 1.1965 MO 290507 0.3031 0.4133 0.7164 0.3957 0.4885 0.8842 0.7226 0.5681 1.2907 MT 300806 0.2078 0.4188 0.6266 0.2929 0.4888 0.7817 0.5390 0.5325 1.0715 NJ 340507 0.3041 0.4166 0.7207 0.3979 0.5011 0.8990 0.7841 0.5654 1.3495 NM 350802 0.3440 0.3138 0.6578 0.4073 0.3356 0.7429 0.5364 0.3505 0.8869 NY 360802 0.2102 0.4329 0.6431 0.3007 0.5174 0.8181 0.5557 0.5498 1.1055 NC 370802 0.2162 0.3929 0.6091 0.3156 0.4542 0.7698 0.6101 0.5087 1.1188 OK 400603 0.2443 0.3671 0.6114 0.3526 0.3825 0.7351 0.5438 0.4023 0.9461 PA 420608 0.2020 0.4018 0.6038 0.2839 0.4728 0.7567 0.5202 0.5418 1.0620 SD 460804 0.2067 0.4221 0.6288 0.2910 0.4967 0.7877 0.5382 0.5445 1.0827 TX 480802 0.2456 0.4284 0.6740 0.3563 0.4526 0.8089 0.5790 0.4844 1.0634 WI 550806 0.2148 0.4391 0.6539 0.3032 0.5176 0.8208 0.5726 0.5772 1.1498 ALB 810504 0.2930 0.4949 0.7879 0.3805 0.6085 0.9890 0.6167 0.7283 1.3450 MAN 830504 0.2344 0.4615 0.6959 0.3328 0.5390 0.8718 0.5729 0.6204 1.1933 CA2 06A806 0.4161 0.4751 0.8912 0.5654 0.5156 1.0810 0.8417 0.5638 1.4055 TX2 48A504 0.2377 0.4157 0.6534 0.3366 0.4517 0.7883 0.5363 0.4839 1.0202 conditioning provides a poorer simulation of the aging that occurs with the binders used in these three sections. Chemical-Based Properties. Calculating equivalent PAV temperatures using the FTIR absorbance data was less complicated. Bilinear relationships for the carbonyl, sulfoxide, and carbonyl plus sulfoxide absorbance were developed from the laboratory-conditioned data as shown in Figure 84 for the Missouri binder. These were then used to interpolate, and in some cases, extrapolate to determine the equivalent PAV conditioning temperatures. For several binders, the sulfoxide absorbance for the recovered binder was significantly different than that for the laboratory-conditioned binder, yielding either extremely high or extremely low equiva- lent PAV temperatures for sulfoxide absorbance. The equivalent PAV temperatures calculated from the FTIR spectra are summarized in Table 73 for the three slices. Figure 85 compares the equivalent PAV temperatures from the carbonyl absorbance to the equivalent PAV temperatures from the sulfoxide absorbance. The figure shows there is poor agreement between the equiva- lent PAV temperatures based on these two absorbances. Figure 86 compares the equivalent PAV temperatures from the carbonyl absorbance to the equivalent PAV temperatures from the carbonyl plus sulfoxide absorbance. This figure shows good agreement between the two equivalent PAV temperatures, indicating the dominant effect of carbonyl on the equivalent PAV temperature.

128 Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging a. Defining temperature. b. Crossover frequency. c. Rheological index. Figure 81. Plots of Christensen-Anderson master curve parameters as a function of conditioning temperature for 12.5 g, 20-hr PAV conditioning for the Missouri section.

Research Approach 129   LTPP Section Equivalent PAV Temperature °C Explained Variance Unconstrained Optimization Laboratory Temperature Relationships, Constrained Optimization G* Phase Angle G* Phase Angle AL 101.2 0.998 0.946 0.998 0.950 AR2 108.0 0.997 0.943 0.909 0.909 CA 107.9 0.998 0.986 0.817 0.817 IL 101.7 0.997 0.954 0.921 0.913 IN 99.7 0.998 0.974 0.930 0.930 IA 101.3 0.997 0.933 0.979 0.909 ME 87.6 0.998 0.985 0.985 0.985 MD 99.6 0.998 0.993 0.941 0.941 MN 96.8 0.996 0.964 0.925 0.925 MS 102.7 0.998 0.977 0.961 0.961 MO 105.7 0.998 0.980 0.989 0.984 MT 90.1 0.996 0.967 0.963 0.963 NJ 101.4 0.998 0.951 0.920 0.920 NM 105.8 0.996 0.959 0.960 0.960 NY 106.9 0.999 0.995 0.806 0.806 NC 100.7 0.999 0.980 0.933 0.933 OK 110.3 0.997 0.906 0.812 0.812 PA 84.2 0.998 0.988 0.983 0.983 SD 105.0 0.998 0.958 0.888 0.888 TX 103.8 0.996 0.923 0.917 0.917 WI 101.0 0.997 0.989 0.984 0.983 ALB 100.7 0.999 0.993 0.982 0.982 MAN 98.7 0.999 0.987 0.971 0.971 CA2 110.6 0.997 0.972 0.943 0.943 TX2 82.3 0.996 0.961 0.897 0.897 Table 70. Equivalent 12.5 g, 20-hr PAV temperatures, and goodness for fit statistics from master curve data for the 0 to 0.5 in slice. LTPP Section Equivalent PAV Temperature, °C Explained Variance Unconstrained Optimization Laboratory Temperature Relationships, Constrained Optimization G* Phase Angle G* Phase Angle AL 93.1 0.997 0.985 0.980 0.980 AR2 107.6 0.995 0.965 0.968 0.968 CA 106.1 0.998 0.986 0.841 0.841 IL 102.2 0.997 0.964 0.859 0.859 IN 97.1 0.995 0.972 0.958 0.958 IA 99.0 0.998 0.946 0.926 0.893 ME 81.2 0.998 0.980 0.979 0.979 MD 90.8 0.998 0.993 0.938 0.938 MN 94.4 0.995 0.973 0.908 0.908 MS 98.3 0.999 0.992 0.980 0.980 MO 103.4 0.999 0.982 0.982 0.982 MT 85.2 0.996 0.968 0.961 0.954 Table 71. Equivalent 12.5 g, 20-hr PAV temperatures, and goodness for fit statistics from master curve data for the 0.5 in to 1.0 in slice. (continued on next page)

130 Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging LTPP Section Equivalent PAV Temperature, °C Explained Variance Unconstrained Optimization Laboratory Temperature Relationships, Constrained Optimization G* Phase Angle G* Phase Angle AL 84.6 0.998 0.986 0.982 0.982 AR2 105.1 0.998 0.967 0.973 0.973 CA 105.1 0.998 0.983 0.868 0.868 IL 101.2 0.998 0.967 0.852 0.852 IN 86.7 0.996 0.960 0.890 0.891 IA 100.2 0.997 0.946 0.930 0.892 ME 78.8 0.998 0.979 0.984 0.984 MD 86.1 0.998 0.989 0.965 0.965 MN 95.5 0.996 0.972 0.901 0.901 MS 94.6 0.998 0.987 0.970 0.970 MO 101.5 0.998 0.987 0.983 0.983 MT 75.9 0.997 0.978 0.991 0.984 NJ 102.6 0.998 0.990 0.981 0.981 NM 95.6 0.996 0.955 0.971 0.962 NY 95.7 0.997 0.979 0.934 0.934 NC 94.1 0.999 0.980 0.950 0.950 OK 100.0 0.997 0.943 0.843 0.843 PA 85.1 0.998 0.991 0.963 0.963 SD 104.3 0.998 0.991 0.939 0.939 TX 100.0 0.996 0.972 0.972 0.910 WI 92.4 0.998 0.981 0.980 0.980 ALB 91.1 0.999 0.996 0.992 0.993 MAN 90.0 0.999 0.992 0.981 0.981 CA2 109.7 0.997 0.974 0.907 0.907 TX2 72.8 0.997 0.981 0.993 0.989 Table 72. Equivalent 12.5 g, 20-hr PAV temperatures, and goodness for fit statistics from master curve data for the 1.5 in to 2.0 in slice. LTPP Section Equivalent PAV Temperature, °C Explained Variance Unconstrained Optimization Laboratory Temperature Relationships, Constrained Optimization G* Phase Angle G* Phase Angle NY 101.0 0.998 0.979 0.839 0.839 NC 96.3 0.999 0.978 0.950 0.950 OK 107.5 0.996 0.940 0.811 0.811 PA 84.3 0.998 0.989 0.963 0.963 SD 108.0 0.998 0.947 0.889 0.889 TX 100.1 0.996 0.920 0.959 0.914 WI 94.0 0.998 0.983 0.984 0.984 ALB 91.2 0.999 0.995 0.992 0.992 MAN 93.0 0.999 0.991 0.975 0.975 CA2 106.9 0.997 0.972 0.933 0.933 TX2 79.2 0.997 0.974 0.975 0.975 NJ 102.6 0.998 0.992 0.989 0.988 NM 100.4 0.996 0.960 0.978 0.966 Table 71. (Continued).

Research Approach 131   a. Constrained optimization using laboratory-conditioned temperature relationships for Td, ωc, and R. b. Unconstrained optimization. Figure 82. Comparison of best-fit master curve from constrained laboratory-conditioned binder temperature relationships and best-fit master curve from unconstrained optimization for the top slice from the Missouri section. a. Constrained optimization using laboratory-conditioned temperature relationships for Td, ωc, and R. b. Unconstrained optimization. Figure 83. Comparison of best-fit master curve from constrained laboratory-conditioned binder temperature relationships and best-fit master curve from unconstrained optimization for the top slice from the California section.

132 Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging Figure 84. Plots of FTIR absorbances as a function of conditioning temperature for 12.5 g, 20-hr PAV conditioning for the Missouri section. Equivalent PAV Conditioning Temperature, °C LTPP Section 0 to 0.5 in Slice 0.5 in to 1.0 in Slice 1.5 in to 2.0 in Slice C=O S=O C=O + S=O C=O S=O C=O + S=O C=O S=O C=O + S=O AL 112.5 115.6 113.3 104.8 113.2 107.0 100.0 95.1 98.0 AR2 122.0 142.7 125.0 117.7 146.6 121.9 116.4 150.2 121.4 CA 119.2 202.0 130.3 116.2 186.2 125.6 115.0 121.3 115.8 IL 110.9 131.1 115.2 107.8 135.5 113.7 106.2 124.4 110.1 IN 109.5 95.6 106.9 107.2 99.1 105.9 105.1 113.0 106.2 IA 111.9 120.4 113.4 109.4 118.2 110.9 109.6 117.6 111.0 ME 102.7 96.6 101.5 100.2 99.0 100.0 99.4 92.3 96.7 MD 107.0 113.6 108.2 102.6 103.0 102.7 101.7 102.4 101.8 MN 107.8 72.7 110.9 104.5 85.9 106.1 105.3 62.5 109.1 MS 107.2 125.4 110.5 102.5 127.8 107.0 102.9 109.0 104.0 MO 104.8 115.7 106.9 103.1 112.3 104.9 103.3 106.0 103.8 MT 103.3 111.5 104.5 101.1 112.5 102.8 86.1 83.1 84.7 NJ 110.9 115.2 111.5 107.3 129.6 110.5 107.0 119.5 108.8 NM 112.1 160.6 117.1 110.3 163.1 115.8 108.1 155.7 113.1 NY 114.8 135.3 117.1 109.4 123.1 110.9 105.9 90.7 102.5 NC 104.9 126.1 108.2 102.9 114.5 104.7 103.0 96.6 101.9 OK 129.5 147.3 131.1 115.9 125.5 116.8 108.2 117.7 109.1 PA 97.3 98.1 97.7 92.1 96.5 94.1 92.6 97.8 95.0 SD 116.8 104.3 114.8 121.6 105.6 119.0 114.6 107.4 113.5 TX 111.2 95.8 109.4 109.0 96.0 107.5 108.2 97.5 107.0 WI 109.0 97.7 106.8 106.2 98.2 104.6 105.3 97.6 103.8 ALB 103.8 107.7 105.1 99.8 104.6 101.5 98.2 103.0 100.6 MAN 105.7 102.7 105.0 103.6 108.4 104.8 102.6 108.9 104.2 CA2 119.8 112.3 118.6 115.1 108.3 114.1 117.1 115.0 116.8 TX2 103.3 91.8 101.6 88.7 85.5 87.6 79.9 57.9 74.0 Table 73. Equivalent 12.5 g, 20-hr PAV temperatures from FTIR spectra.

Research Approach 133   Figure 85. Comparison of equivalent PAV temperatures based on carbonyl absorbance and sulfoxide absorbance. Figure 86. Comparison of equivalent PAV temperature based on carbonyl absorbance and carbonyl plus sulfoxide absorbance. Comparison of Rheological- and Chemical-Based Equivalent 12.5 g, 20-hr PAV Temper- atures. Figure 87 compares equivalent PAV temperatures from rheology to the various equivalent PAV temperatures from the FTIR absorbance data. These figures show similar trends between the equivalent PAV temperatures from rheology and those from carbonyl absorbance and carbonyl plus sulfoxide absorbance. Although the trends are similar, the equivalent PAV temperatures to match changes in carbonyl are approximately 10°C higher than the equivalent PAV temperatures to match changes in rheology. Equivalent PAV temper- atures based on changes in sulfoxide absorbance are not strongly related to the equivalent PAV temperatures based on changes in rheology. Graphical Analysis of Rheological-Based Equivalent 12.5 g, 20-hr PAV Temperatures Since the equivalent PAV temperatures based on changes in rheological properties and those based on changes in carbonyl and carbonyl plus sulfoxide exhibited similar trends, a detailed analysis of the equivalent PAV temperatures was performed only on the equivalent PAV

134 Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging a. Carbonyl to rheology. b. Sulfoxide to rheology. c. Carbonyl plus sulfoxide to rheology. Figure 87. Comparison of equivalent 12.5 g, 20-hr PAV conditioning temperatures from rheology and FTIR spectra.

Research Approach 135   temperature based on changes in rheological properties. The first step in the detailed analysis was to perform a graphical analysis to determine the factors affecting the equivalent PAV temperatures. This section presents a summary of these analyses. Depth. The equivalent PAV temperatures generally decrease with depth; however, the change with depth varies between sections. The decrease in the equivalent PAV temperature for the 1.5 in to 2.0 in slice compared to the 0 to 0.5 in slices varies from less than 1 percent to as high as 16 percent. Figure 88 shows the range-of-the-depth effect averaged over sections exhibiting similar behavior. The depth effect does not appear to be related to the other param- eters discussed below. Pavement Age. Figure 89 shows the effect of pavement age on the equivalent PAV temper- atures. Data for only the 0 to 0.5 in slice are shown. Pavement age alone is not a good indicator of the equivalent PAV temperature. The observation that age is not a good predictor of the equivalent PAV temperature suggests there is an interaction of several factors. Climate. The climatic factors extracted from the LTPP InfoPave database for the sections used in the PAV calibration experiment included: 1. Mean annual air temperature, 2. Annual degree days over 10°C, 3. 50 and 98 percent reliability high pavement temperature, 4. 50 and 98 percent reliability low pavement temperature, 5. Average daily shortwave radiation, 6. Average annual precipitation. The extracted data were for the period between the construction of the section and extraction of the cores used in the laboratory testing. Figure 90 compares plots of the equivalent PAV Figure 88. Effect of depth on equivalent PAV temperature.

136 Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging temperature for the 0 to 0.5 in slice as a function of various measures of temperature. The mean annual air temperature, the 98 percent reliability high pavement temperature, and the average of the 98 percent reliability high and low pavement temperatures appear to be better indicators of the effect of temperature than does cumulative degree days. The data for the Texas-2 and Pennsylvania sections plot below the other data. As discussed in the next section, these pave- ments have lower air voids compared to the other pavements. Plots of the equivalent PAV temperature for the 0 to 0.5 in slice as a function of shortwave radiation and precipitation are shown in Figure 91 and Figure 92, respectively. The shortwave radiation effect is similar to the air and pavement temperature effects shown in Figure 90. The equivalent PAV temperature is not strongly related to the average annual precipitation. Volumetric Factors. The pavement cores used in the PAV calibration experiment were taken between the wheel paths. Bulk specific gravity, maximum specific gravity, asphalt content, and gradation were measured on sections of the cores not used for extraction. Bulk specific grav- ity data for the aggregates used in the mixtures were extracted from the LTPP InfoPave database. The following volumetric factors were calculated from the data: 1. Air void content, 2. Voids in the mineral aggregate, 3. Effective volume of binder, and 4. Voids filled with asphalt. Figure 93 compares plots of the equivalent PAV temperature for the 0 to 0.5 in slice as a function of various volumetric properties. The equivalent PAV temperatures are most strongly related to the air void content and the voids filled with asphalt. Note the low air void content and high voids filled with asphalt for the Texas-2 and Pennsylvania sections, which have equivalent PAV temperatures lower than expected based on the climate. Figure 89. Effect of age on equivalent PAV temperature.

a. Mean annual air temperature. b. Cumulative degree days. c. Average of 98 percent reliability high and low temperature. d. Ninety-eight percent reliability high pavement temperature. Figure 90. Effect of temperature on equivalent PAV temperature.

138 Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging Figure 91. Effect of shortwave radiation on equivalent PAV temperature. Figure 92. Effect of precipitation on equivalent PAV temperature.

a. Air void content. b. Voids in the mineral aggregate. c. Effective volume of binder. d. Voids filled with asphalt. Figure 93. Effect of volumetric composition on equivalent PAV temperature.

140 Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging Figure 94 compares plots of the equivalent PAV temperature for the 0 to 0.5 in slice as a func- tion of various gradation parameters and apparent film thickness. The equivalent PAV tempera- ture is not strongly related to any of the gradation parameters. Rational Statistical Model of Rheological-Based Equivalent 12.5 g, 20-hr PAV Temperatures The primary conclusion drawn from the graphical analysis was the equivalent PAV temperature is affected by depth, climate, and how well the pavement was compacted, but the equivalent PAV temperature is not strongly related to any single factor. A rational model for the equivalent PAV temperature was, therefore, formed to analyze the data. This model assumes the equivalent PAV temperature is a function of the exposure of the binder and the sensitivity of changes in the rheological properties of the binder to aging temperature. The exposure depends on (1) the time in service, (2) temperature at the project site, (3) air void content of the pavement, and (4) the depth in the pavement. The sensitivity of the rheological properties of the binder to aging temperature was estimated by fitting the change in the rheo- logical index from the RTFOT condition to PAV conditioning at 85°C, 100°C, and 115°C to an Arrhenius equation. R Ae BTPAV∆ = − (8) where DR = change in the rheological index from RTFOT to PAV condition, TPAV = PAV conditioning temperature in °K, and A and B are fitting parameters. The parameter B is an indicator of the sensitivity of the binder rheological properties to aging temperature. Table 74 presents the Arrhenius parameters for the binders in the pavement sec- tions used in the PAV calibration experiment. The last column provides the relative sensitivity of the binder rheology to aging temperature. The New Jersey, California, Manitoba, and Alberta binders are more sensitive to aging temperature, while the Oklahoma, Texas, New Mexico, and Iowa binders are less sensitive to aging temperature. The form of the rational model is similar to the modification to the Witczak global aging model developed by Christensen in NCHRP Project 01-42 (Advanced Asphalt Technologies LLC 2004). An exposure index is calculated from Equation 9. Then, knowing the exposure index, the equivalent PAV temperature for 12.5 g, 20-hr, 2.1 MPa conditioning is calculated from Equation 10. 1 1 0.25 (9) 1 273.15 1 273.15 VTM 3.86 0.37 EI t e e d B T= +              ( ) − + −        − where EI = exposure index, t = time in service, yrs, B = binder specific Arrhenius coefficient, T = average of the 98% reliability maximum and minimum pavement temperatures, °C, e = 2.71828 (Euler’s number), VTM = in-place air void content, %, and d = depth from the pavement surface, in.

a. Percent passing 4.75 mm sieve. b. Percent passing 0.075 mm sieve. c. Sieve size for 50 percent passing. d. Apparent film thickness. Figure 94. Effect of gradation on equivalent PAV temperature.

142 Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging T EIPAV ( )= 72.75 (10)0.096 where TPAV is equivalent 12.5 g, 20-hr, 2.1 MPa PAV conditioning temperature. The first term inside the braces in Equation 9 is an Arrhenius function that accounts for the aging temperature at the project location. It includes the binder-specific coefficient, B, determined from the change in the rheological index as a function of PAV conditioning temperature. The second term inside the braces accounts for the effect of air voids. It is a sigmoidal function centered at an air void content of 3.86 percent. Finally, the third term inside the braces accounts for the depth effect. The best fit model for the depth effect was a power model. Attempts to relate the exponent in the depth effect to other properties did not signifi- cantly improve the model. The constants in Equation 9 and Equation 10 were determined by fitting these equations to the data from the PAV calibration experiment. Non-linear optimization was used to fit the data. Table 75 presents the data used in the optimization. Figure 95 compares the measured and predicted equivalent PAV temperature values at the three depths. The overall explained variance for the model is 0.65 with a standard error of estimate of 5.2 degrees. The model is reasonably accurate considering the nature of the data. The accuracy of the model decreases with depth into the pavement. For the 0 to 0.5 in slice, the explained variance is 0.69, and the standard error of estimate is 4.1°C. For the 0.5 in to 1.0 in slice, the explained variance is 0.64, and the standard error of estimate is 5.1°C. Finally, for the 1.5 in to 2.0 in slice, the explained variance is only 0.53, and the standard error of estimate is 6.5°C. The poorer fit with depth may be due to air void LTPP Section A B B/Baverage AL 18.02 6871.5 1.02 AR2 17.28 6764.0 1.01 CA 21.14 8086.0 1.20 IL 17.82 6754.1 1.01 IN 17.19 6513.9 0.97 IA 14.52 5498.9 0.82 ME 18.30 7028.6 1.05 MD 19.30 7346.4 1.09 MN 17.74 6699.2 1.00 MS 19.68 7554.2 1.12 MO 16.43 6366.5 0.95 MT 17.48 6680.9 0.99 NJ 22.59 8644.4 1.29 NM 14.30 5510.0 0.82 NY 16.29 6181.5 0.92 NC 18.51 7041.2 1.05 OK 10.40 4006.6 0.60 PA 17.71 6775.0 1.01 SD 19.33 7392.3 1.10 TX 12.27 4702.0 0.70 WI 19.04 7254.0 1.08 ALB 20.27 7812.0 1.16 MAN 20.93 7914.4 1.18 CA2 15.94 6277.0 0.93 TX2 16.33 6268.8 0.93 Table 74. Arrhenius parameters for binders used in the PAV calibration experiment.

Research Approach 143   gradients or other changes with depth. The lowest slice of four of the pavements and the second and lowest slice for two of the pavements used mixes that were different than the surface mix and had different air void contents. When the air voids in these mixes were less than the air voids for the surface mix, the lower air void content was used in the fitting. The relative importance of the predictor variables in the model was assessed through a sensi- tivity analysis. The sensitivity analysis was conducted for each predictor variable by varying the predictor variable over the range of the data from Table 75 while holding the other predictor variables at the median value. Table 76 summarizes the ranges and median values used in the sensitivity analysis. The results of the sensitivity analysis are presented in Figure 96 through Figure 100. Figure 96 shows the effect of pavement temperature on the equivalent 12.5 g, 20-hr, 2.1 MPa PAV temperatures. The effect is nearly linear, increasing at the rate of 8°C per 10°C change in the average of the 98 percent reliability high and low pavement temperatures. This results in approximately a 20°C change in the equivalent PAV temperature for the range of climates in LTPP Section T Avg 98% High and Low Temp, °C t Time in Service, yrs VTM Air Void Content, % B Binder, Arrhenius Exponent Equivalent PAV Temperature, °C 0.25 in 0.75 in 1.75 in AL 24.98 12.4 3.29 6871.5 101.2 93.1 NA AL 24.98 12.5 2.80 6871.5 NA NA 84.6 AR2 23.12 7.3 6.16 6764.0 108.0 107.6 NA AR2 23.12 7.3 5.00 6764.0 NA NA 105.1 CA 19.61 13.5 9.44 8086.0 107.9 106.1 105.1 IL 13.67 15.9 8.74 6754.1 101.7 102.2 NA IL 13.67 15.9 7.40 6754.1 NA NA 101.2 IN 11.50 12.9 7.15 6513.9 99.7 97.1 85.9 IA 14.56 14.5 8.14 5498.9 101.3 99.0 100.2 ME 7.45 9.1 4.77 7028.6 87.6 81.2 78.8 MD 15.89 13.2 3.63 7346.4 99.6 90.8 86.1 MN 7.30 14.7 8.03 6699.2 96.8 94.4 95.5 MS 22.07 11.9 3.67 7554.2 102.7 98.3 94.6 MO 21.04 7.6 5.45 6366.5 105.7 103.4 101.5 MT 7.04 9.3 5.33 6680.9 90.1 85.2 75.9 NJ 18.10 13.6 2.26 8644.4 101.4 102.6 102.6 NM 27.39 9.4 9.50 5510.0 105.8 100.4 95.6 NY 15.98 13.2 13.30 6181.5 106.9 NA NA NY 15.98 13.2 9.40 6181.5 NA 101.0 95.7 NC 21.77 8.7 10.00 7041.2 100.7 96.3 94.1 OK 22.41 14.1 6.81 4006.6 110.3 107.5 100 PA 14.84 12.7 2.11 6775.0 84.2 84.3 85.1 SD 12.10 14.2 5.08 7392.3 105.0 108.0 104.3 TX 26.54 11.0 6.49 4702.0 103.8 100.1 100.0 WI 7.75 7.7 6.02 7254.0 101.0 94.0 92.4 ALB 2.78 15.7 4.35 7812.0 100.7 91.2 91.1 MAN 7.02 14.7 5.05 7914.4 98.7 93.0 90.0 CA2 30.20 7.8 8.80 6277.0 110.6 109.7 106.5 TX2 25.86 15.9 0.86 6268.8 82.3 79.2 72.8 Table 75. Data used to develop a rational statistical model of equivalent 12.5 g, 20-hr, 2.1 MPa PAV conditioning.

144 Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging a. 0 to 0.5 in slice. b. 0.5 to 1.0 in slice. c. 1.5 to 2.0 in slice. Figure 95. Predicted versus measured equivalent PAV temperatures, including depth effect.

Research Approach 145   Property T Avg 98% High and Low Temp, °C t Time in Service, yrs VTM Air Void Content, % B Binder, Arrhenius Exponent d Depth, in Minimum 2.78 7.25 0.86 8644.38 0.25 Median 15.98 12.89 6.02 6763.98 0.75 Maximum 30.20 15.94 13.30 4006.62 1.75 Table 76. Data range and values used in sensitivity analysis. Figure 96. Sensitivity of 12.5 g, 20-hr, 2.1 MPa equivalent PAV temperature model to pavement temperature. the United States, which is in agreement with the temperatures used in AASHTO M 320 and AASHTO M 332, which vary from 90°C to 110°C. Figure 97 shows the effect of pavement age on the equivalent 12.5 g, 20-hr, 2.1 MPa PAV temperatures. The effect is nearly linear, increasing at the rate of approximately 0.9°C per year. For the binders tested, standard PAV conditioning at 100°C is approximately equal to 12.5 g, 20-hr, 2.1 MPa PAV conditioning at 85°C. Figure 97 is consistent with the observation that standard PAV conditioning underestimates the aging that occurs in service near the surface of the pavement. Figure 98 shows the effect of air void content on the equivalent 12.5 g, 20-hr, 2.1 MPa PAV temperatures. This figure shows the equivalent PAV temperature is a maximum for air voids above approximately 6.5 percent. As air voids decrease, the equivalent PAV temperature decreases significantly. The large effect of air voids was evident in the raw equivalent PAV temperature data.

146 Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging Figure 97. Sensitivity of 12.5 g, 20-h, 2.1 MPa equivalent PAV temperature model to pavement age (yr). Figure 98. Sensitivity of 12.5 g, 20-hr, 2.1 MPa equivalent PAV temperature model to air void content.

Research Approach 147   Figure 99. Sensitivity of 12.5 g, 20-hr, 2.1 MPa equivalent PAV temperature model to binder Arrhenius exponent. Figure 100. Sensitivity of 12.5 g, 20-hr, 2.1 MPa equivalent PAV temperature model to depth.

148 Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging The Texas-2 section with air void content of 0.86 percent had the lowest equivalent PAV temper- ature but was in one of the hottest environments. The Pennsylvania section with air voids of 2.86 percent also had lower equivalent PAV temperatures than expected for the climate at this site. Figure 98 also indicates that current compaction specifications targeting 7 percent to 8 percent air voids do not significantly reduce the aging of binders in pavements. Figure 99 shows the effect of the binder Arrhenius exponent on the equivalent 12.5 g, 20-hr, 2.1 MPa PAV temperatures. Recall that this parameter is a measure of the temperature sensitivity of the binder aging. Binders with higher Arrhenius exponent age more rapidly as the temperature increases. Over the range of sensitivities for the binders from the LTPP sections, the effect is approximately one-half of that for pavement temperature. Finally, Figure 100 shows the effect of depth on the equivalent 12.5 g, 20-hr, 2.1 MPa PAV temperatures. This figure shows the model is least sensitive to depth, decreasing about 8°C per inch near the pavement surface and decreasing to about 3°C per inch deeper in the pavement. 12.5 g, 20-hr, 2.1 MPa PAV Conditioning Temperatures for Performance Grading Considering the difference in the equivalent 12.5 g, 20-hr, 2.1 MPa PAV temperatures for the 0 to 0.5 in slice and the 0.5 in to 1.0 in slice of about 5°C and the standard error of the model of about 5°C, the model was also fit to the average of the equivalent PAV temperature for the top two slices. This model, which represents the aging in the top inch of the pavement, is given in by Equation 11 and Equation 12. The explained variance for this model is 0.69 with a standard error of estimate of 4.3°C. Figure 101 compares the measured and predicted equivalent PAV temperatures. Figure 101. Predicted versus measured equivalent PAV temperatures for top 1.0 in of pavement.

Research Approach 149   1 1 (11) 1 273.15 1 273.15 3.85 EI t e e B T VTM = +          ( ) − + −        − T EIPAV ( )= 72.02 (12)0.093 The model given by Equation 11 and Equation 12 was used to determine 12.5 g, 20-hr, 2.1 MPa PAV conditioning temperatures for AASHTO M 320 and AASHTO M 332 grading. Based on LTPPBind 3.1, the average of the 98 percent reliability high and low pavement temperatures in the United States and Canada ranges from −6°C to 33°C in 3°C increments (Pavement Systems LLC 2005). Note that adjustments for traffic speed and traffic volume should not be applied to these temperatures. Based on typical compaction specifications and Figure 98 showing the sensitivity of the equivalent PAV temperatures to air voids, an air void content of 7.0 percent was selected. The median binder Arrhenius exponent of 6764 was used. The results are shown in Table 77. This table shows the equivalent 12.5 g, 20-hr, 2.1 MPa PAV conditioning temperature calculated from Equation 11 and Equation 12, the recommended 12.5 g, 20-hr, 2.1 MPa PAV conditioning temperature, the percentage of the weather stations from LTPPBind 3.1 for each recommended conditioning temperature, and the corresponding binder grades based on climate without adjustment for traffic speed or traffic volume. The three lowest and the highest average 98 percent reliability high and low pavement temperatures in Table 77 are outside the range of the data used in the calibration experiment. The current range of PAV temperatures in AASHTO M 320 and AASHTO M 332, from 90°C to 110°C covers 98 percent of the weather stations from LTPPBind 3.1 for the United States and Canada. A PAV temperature of 100°C covers 41 percent of the weather stations. Average 98% Reliability High and Low Pavement Temperature Calculated PAV Temperature, °C Recommended Temperature, °C % of LTPPBind 3.1 Stations PG Grade Based on Environment −61 84.4 85 1 PG 40-52, PG 46-52, PG 40-46 −31 86.6 01 88.9 90 4 PG 52-52, PG 46-46, PG 40-40, PG 46-40, PG 52-46, PG 40-343 91.1 6 93.4 95 20 PG 58-46, PG 52-40, PG 46-34, PG 40-28, PG 58-40, PG 52-34, PG 46-28, PG 40-229 95.7 12 97.9 100 41 PG 64-40, PG 58-34, PG 52-28, PG 46-22, PG 40-16, PG 64-34, PG 58-28, PG 52-22, PG 46-16, PG 40-10, PG 64-28, PG 58-22, PG 52-16, PG 46-10 15 100.2 18 102.5 21 104.8 105 20 PG 70-28, PG 64-22, PG 58-16, PG 52-10, PG 70-22, PG 64-16, PG 58-1024 107.1 27 109.3 110 13 PG 70-16, PG 64-10, PG 70-10 30 111.6 331 115.0 115 1 PG 76-10 1 Outside range of data used in calibration. Table 77. Recommended 12.5 g, 20-hr PAV conditioning temperatures for performance grading.

150 Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging Sensitivity Experiment Objective The objectives of the sensitivity experiment were to: (1) confirm the practicality of using 12.5 g, 20-hr, 2.1 MPa long-term conditioning for a range of binders, and (2) determine the approximate magnitude of changes to intermediate and low-temperature performance grading resulting from the increased aging simulated by 12.5 g, 20-hr, 2.1 MPa long-term conditioning. Both are important implementation considerations. Experimental Design The sensitivity experiment was designed to compare intermediate and low-temperature performance grading for binders’ long-term conditioning in accordance with AASHTO R 28 and 12.5 g, 20-hr, 2.1 MPa PAV conditioning. Ten binders that were included in NCHRP Project 09-59 and NCHRP Project 09-60 were included in the experiment. Table  78 lists the binders that were included in the sensitivity experiment. This table shows the AASHTO M 320 performance grade of each binder, the type of modification, the average of the 98 percent reliability high and low pavement temperature at the project site if the binder was used in a field project, the average of the 98 percent reliability high and low pavement temperature based on the environmental grade where the binder may be used if it was not used in a specific project, and the conditioning temperature for 12.5 g, 20-hr, 2.1 MPa PAV conditioning based on Table 77. Each binder was first RTFOT conditioned; then, the RTFOT residue was conditioned at the temperatures shown. Intermediate- and low-temperature grading was then conducted on the PAV residue. During conditioning, it was observed that the PG 76 binders did not fully coat the pan, which led to additional work to define a film formation step at a higher temperature under nitrogen. That work estab- lished a temperature of 135°C for 30 minutes under industrial nitrogen as the conditions for the film formation step. DSR testing of RTFOT residue before and after the film formation step confirmed that the rheological properties were not significantly affected by the film formation step. Only the conditioning of the PG 76 binders in Table 78 included the film formation step. Binder Project AASHTO M 320 Grade Modification Average 98% Reliability High and Low PAV Conditioning Temperature, °C Project Site, °C Environmental Grade, °C AAK-1 SHRP 64-22 None NA 21 105 AAM-1 SHRP 64-16 None NA 24 105 AZ1-1 ARC 76-16 Air Blown 28.5 NA 110 AZ1-3 ARC 76-16 None 28.5 NA 110 MN1-2 ARC 58-34 Terpolymer 8.8 NA 95 MN1-4 ARC 58-28 None 8.8 NA 95 MN1-5 ARC 58-28 None 8.8 NA 95 ME3 NCHRP 09-60 76-28 Unknown 14.3 NA 100 AC1928 NCHRP 09-59 76-34 SBS NA 12 100 AC1879 NCHRP 09-59 76-22 SBS/PPA NA 21 105 Table 78. Binders and conditioning temperatures used in sensitivity experiment.

Research Approach 151   Results and Analysis Test results are presented in Table 79 for the AASHTO M 320 and AASHTO M 332 inter- mediate temperature parameter G* • sind and Table 80 for the low temperature creep stiffness and m-value. Intermediate data are shown for two PAV conditions: • Standard AASHTO R 28 conditioning at 100°C using 50.0 g, conditioned for 20 hours at 2.1 MPa air pressure, and • 12.5 g, 20-hr, 2.1 MPa conditioning at the temperatures from NCHRP Project 09-61. Low-temperature data are shown for these PAV conditions as well as 50 g, 40-hr, 2.1 MPa conditioning at 100°C. The 50.0 g PAV-conditioned data were collected in NCHRP Projects 09-59 and 09-60. Continuous intermediate- and low-temperature data are summarized in Table  81 and Table 82, respectively. Note that different conditioning temperatures were used. All 50.0 g PAV conditioning used 100°C, while 95°C, 100°C, and 105°C were used for the 12.5 g conditioning depending on the climate where the binder was used or would likely be used. For 12.5 g, 20-hr, 2.1 MPa conditioning at the temperature determined in this project, the intermediate con tinuous grade temperature increased approximately 1°C to 10°C above that for AASHTO T 28 conditioning. Similarly, for the 12.5 g, 20-hr, 2.1 MPa conditioning at the temperature determined in this project, the low continuous grade temperature increased approximately 0.5°C to 13°C above that for AASHTO T 28 conditioning. The largest increase occurred when the 12.5 g conditioning was conducted at 105°C, while the smallest increase occurred when the 12.5 g conditioning was conducted at 95°C. All 50.0 g conditioning was performed at 100°C. Note that binders ME3 and AC 1928 were conditioned at 100°C using 40-hr at 50.0 g and 20-hr at 12.5 g. Low temperature, continuous grade temperatures, and DTc values are within 0.6°C under these two accelerated aging conditions. The continuous grade and DTc data in Table 82 show 12.5 g, 20-hr, 2.1 MPa conditioning at the temperatures determined in this project may be more or less severe compared to 50.0 g, 40-hr, 2.1 MPa conditioning depending on the temperature used. When conditioned at higher temperatures, 12.5 g, 20-hr conditioning is more severe. When conditioned at the same temper- ature, the 12.5 g, 20-hr conditioning is approximately equivalent to 50.0 g, 40-hr conditioning. When conditioned at lower temperatures, the 12.5 g, 20-hr conditioning is less severe.

Conditioning G*·sinδ, kPa at Binder Mass, g Time, hr Temp, °C 4°C 7°C 10°C 13°C 16°C 19°C 22°C 25°C 28°C 31°C 34°C 37°C 40°C AAK 50 20 100 6360 4300 12.5 20 105 5500 4010 AAM 50 20 100 5000 3530 12.5 20 105 5970 4410 AZ1-1 50 20 100 5010 3690 12.5 20 110 5330 4160 AZ1-3 50 20 100 5030 3550 12.5 20 110 5540 4270 MN1-2 50 20 100 5510 3710 12.5 20 95 5120 3660 MN1-4 50 20 100 5700 4090 12.5 20 95 5160 3770 MN1-5 50 20 100 7740 4980 12.5 20 95 5990 3950 ME3 50 20 100 5100 3540 12.5 20 100 5170 3720 AC1928 50 20 100 6450 4720 12.5 20 100 5950 4510 AC1879 50 20 100 6730 4630 12.5 20 105 5110 3850 Table 79. AASHTO M 320 and AASHTO M 332 intermediate temperature data for the sensitivity experiment binders.

Conditioning BBR Stiffness, MPa at BBR m-value, at Binder Mass, g Time, hr Temp, °C 6°C 0°C −6°C −12°C −18°C −24°C − 30°C 6°C 0°C −6°C −12°C −18°C −24°C −30°C AAK 50 20 100 150 342 0.376 0.313 50 40 100 192 392 0.327 0.273 12.5 20 105 139 259 0.323 0.292 AAM 50 20 100 181 357 0.300 0.256 50 40 100 88.7 199 0.32 0.272 12.5 20 105 104 197 0.302 0.277 AZ1-1 50 20 100 80.6 157 0.321 0.292 50 40 100 44.6 89.4 0.345 0.297 12.5 20 110 54.6 102 0.305 0.277 AZ1-3 50 20 100 157 321 0.341 0.284 50 40 100 197 366 0.304 0.26 12.5 20 110 78.8 147 0.321 0.283 MN1-2 50 20 100 279 557 0.318 0.266 50 40 100 153 310 0.323 0.288 12.5 20 95 150 328 0.327 0.298 MN1-4 50 20 100 113 244 0.331 0.289 50 40 100 65.0 137 0.338 0.299 12.5 20 95 127 253 0.319 0.281 MN1-5 50 20 100 246 598 0.359 0.269 50 40 100 271 606 0.336 0.262 12.5 20 95 261 619 0.342 0.275 ME3 50 20 100 194 407 0.329 0.280 50 40 100 116 240 0.345 0.296 12.5 20 100 233 536 0.301 0.259 AC1928 50 20 100 141 281 0.302 0.285 50 40 100 44.1 83.1 0.304 0.288 12.5 20 100 48.4 88.7 0.306 0.287 AC1879 50 20 100 181 397 0.340 0.278 50 40 100 230 387 0.303 0.263 12.5 20 105 136 270 0.317 0.271 Table 80. AASHTO M 320 and AASHTO M 332 low-temperature data for the sensitivity experiment binders.

154 Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging Binder Continuous Grade 50 g, 20 hr 12.5 g, 20 hr AAK-1 20.8 28.9 AAM-1 19.0 20.1 AZ1-1 22.0 25.8 AZ1-3 28.1 38.2 MN1-2 10.7 13.2 MN1-4 11.2 13.3 MN1-5 16.0 17.3 ME3 16.2 19.3 AC1928 6.4 8.9 AC1879 21.4 28.2 Table 81. Intermediate temperature continuous grade data for the sensitivity experiment binders. Binder Continuous Grade ∆T c 50 g, 20 hr 50 g, 40 hr 12.5 g, 20 hr 50 g, 20 hr 50 g, 40 hr 12.5 g, 20 hr AAK-1 −27.0 −24.9 −20.4 2.4 −0.9 −3.0 AAM-1 −22.0 −18.4 −16.5 −4.5 −6.6 −9.5 AZ1-1 −20.3 −15.6 −11.0 −7.5 −10.8 −15.4 AZ1-3 −20.2 −16.5 −7.2 −1.2 −3.6 −9.7 MN1-2 −34.6 −31.9 −33.3 1.4 −1.8 0.3 MN1-4 −32.3 −27.8 −30.9 −3.3 −6.5 −4.6 MN1-5 −29.3 −28.8 −29.0 2.4 1.9 2.6 ME3 −31.4 −27.5 −28.1 −0.1 −2.3 −1.7 AC1928 −34.7 −23.5 −23.9 −5.9 −16.7 −16.2 AC1879 −25.7 −22.4 −18.1 −0.2 −2.7 −4.8 Table 82. Low-temperature continuous grade data for the sensitivity experiment binders.

Next: Chapter 3 - Findings and Application »
Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging Get This Book
×
MyNAP members save 10% online.
Login or Register to save!
Download Free PDF

Asphalt binders experience aging that occurs in two distinct stages under quite different conditions: (1) short-term during construction (plant mixing, storage, placement, and compaction) and (2) long-term during the service life of the pavement.

The TRB National Cooperative Highway Research Program's NCHRP Research Report 967: Asphalt Binder Aging Methods to Accurately Reflect Mixture Aging documents research conducted to improve laboratory binder conditioning methods to accurately simulate the short-term and long-term aging of asphalt binders, and to calibrate the improved procedures to the aging that occurs during mixture production, transport, and placement as well as during the service life of the pavement structure.

  1. ×

    Welcome to OpenBook!

    You're looking at OpenBook, NAP.edu's online reading room since 1999. Based on feedback from you, our users, we've made some improvements that make it easier than ever to read thousands of publications on our website.

    Do you want to take a quick tour of the OpenBook's features?

    No Thanks Take a Tour »
  2. ×

    Show this book's table of contents, where you can jump to any chapter by name.

    « Back Next »
  3. ×

    ...or use these buttons to go back to the previous chapter or skip to the next one.

    « Back Next »
  4. ×

    Jump up to the previous page or down to the next one. Also, you can type in a page number and press Enter to go directly to that page in the book.

    « Back Next »
  5. ×

    To search the entire text of this book, type in your search term here and press Enter.

    « Back Next »
  6. ×

    Share a link to this book page on your preferred social network or via email.

    « Back Next »
  7. ×

    View our suggested citation for this chapter.

    « Back Next »
  8. ×

    Ready to take your reading offline? Click here to buy this book in print or download it as a free PDF, if available.

    « Back Next »
Stay Connected!