National Academies Press: OpenBook

Acceptance Tests for Surface Characteristics of Steel Strands in Prestressed Concrete (2008)

Chapter: Appendix B - Evaluation of Mechanical and Chemical Test Methods

« Previous: Appendix A - Review of Strand Bond Literature
Page 39
Suggested Citation:"Appendix B - Evaluation of Mechanical and Chemical Test Methods." National Academies of Sciences, Engineering, and Medicine. 2008. Acceptance Tests for Surface Characteristics of Steel Strands in Prestressed Concrete. Washington, DC: The National Academies Press. doi: 10.17226/14206.
×
Page 39
Page 40
Suggested Citation:"Appendix B - Evaluation of Mechanical and Chemical Test Methods." National Academies of Sciences, Engineering, and Medicine. 2008. Acceptance Tests for Surface Characteristics of Steel Strands in Prestressed Concrete. Washington, DC: The National Academies Press. doi: 10.17226/14206.
×
Page 40
Page 41
Suggested Citation:"Appendix B - Evaluation of Mechanical and Chemical Test Methods." National Academies of Sciences, Engineering, and Medicine. 2008. Acceptance Tests for Surface Characteristics of Steel Strands in Prestressed Concrete. Washington, DC: The National Academies Press. doi: 10.17226/14206.
×
Page 41
Page 42
Suggested Citation:"Appendix B - Evaluation of Mechanical and Chemical Test Methods." National Academies of Sciences, Engineering, and Medicine. 2008. Acceptance Tests for Surface Characteristics of Steel Strands in Prestressed Concrete. Washington, DC: The National Academies Press. doi: 10.17226/14206.
×
Page 42
Page 43
Suggested Citation:"Appendix B - Evaluation of Mechanical and Chemical Test Methods." National Academies of Sciences, Engineering, and Medicine. 2008. Acceptance Tests for Surface Characteristics of Steel Strands in Prestressed Concrete. Washington, DC: The National Academies Press. doi: 10.17226/14206.
×
Page 43
Page 44
Suggested Citation:"Appendix B - Evaluation of Mechanical and Chemical Test Methods." National Academies of Sciences, Engineering, and Medicine. 2008. Acceptance Tests for Surface Characteristics of Steel Strands in Prestressed Concrete. Washington, DC: The National Academies Press. doi: 10.17226/14206.
×
Page 44
Page 45
Suggested Citation:"Appendix B - Evaluation of Mechanical and Chemical Test Methods." National Academies of Sciences, Engineering, and Medicine. 2008. Acceptance Tests for Surface Characteristics of Steel Strands in Prestressed Concrete. Washington, DC: The National Academies Press. doi: 10.17226/14206.
×
Page 45
Page 46
Suggested Citation:"Appendix B - Evaluation of Mechanical and Chemical Test Methods." National Academies of Sciences, Engineering, and Medicine. 2008. Acceptance Tests for Surface Characteristics of Steel Strands in Prestressed Concrete. Washington, DC: The National Academies Press. doi: 10.17226/14206.
×
Page 46
Page 47
Suggested Citation:"Appendix B - Evaluation of Mechanical and Chemical Test Methods." National Academies of Sciences, Engineering, and Medicine. 2008. Acceptance Tests for Surface Characteristics of Steel Strands in Prestressed Concrete. Washington, DC: The National Academies Press. doi: 10.17226/14206.
×
Page 47
Page 48
Suggested Citation:"Appendix B - Evaluation of Mechanical and Chemical Test Methods." National Academies of Sciences, Engineering, and Medicine. 2008. Acceptance Tests for Surface Characteristics of Steel Strands in Prestressed Concrete. Washington, DC: The National Academies Press. doi: 10.17226/14206.
×
Page 48
Page 49
Suggested Citation:"Appendix B - Evaluation of Mechanical and Chemical Test Methods." National Academies of Sciences, Engineering, and Medicine. 2008. Acceptance Tests for Surface Characteristics of Steel Strands in Prestressed Concrete. Washington, DC: The National Academies Press. doi: 10.17226/14206.
×
Page 49
Page 50
Suggested Citation:"Appendix B - Evaluation of Mechanical and Chemical Test Methods." National Academies of Sciences, Engineering, and Medicine. 2008. Acceptance Tests for Surface Characteristics of Steel Strands in Prestressed Concrete. Washington, DC: The National Academies Press. doi: 10.17226/14206.
×
Page 50
Page 51
Suggested Citation:"Appendix B - Evaluation of Mechanical and Chemical Test Methods." National Academies of Sciences, Engineering, and Medicine. 2008. Acceptance Tests for Surface Characteristics of Steel Strands in Prestressed Concrete. Washington, DC: The National Academies Press. doi: 10.17226/14206.
×
Page 51
Page 52
Suggested Citation:"Appendix B - Evaluation of Mechanical and Chemical Test Methods." National Academies of Sciences, Engineering, and Medicine. 2008. Acceptance Tests for Surface Characteristics of Steel Strands in Prestressed Concrete. Washington, DC: The National Academies Press. doi: 10.17226/14206.
×
Page 52
Page 53
Suggested Citation:"Appendix B - Evaluation of Mechanical and Chemical Test Methods." National Academies of Sciences, Engineering, and Medicine. 2008. Acceptance Tests for Surface Characteristics of Steel Strands in Prestressed Concrete. Washington, DC: The National Academies Press. doi: 10.17226/14206.
×
Page 53
Page 54
Suggested Citation:"Appendix B - Evaluation of Mechanical and Chemical Test Methods." National Academies of Sciences, Engineering, and Medicine. 2008. Acceptance Tests for Surface Characteristics of Steel Strands in Prestressed Concrete. Washington, DC: The National Academies Press. doi: 10.17226/14206.
×
Page 54
Page 55
Suggested Citation:"Appendix B - Evaluation of Mechanical and Chemical Test Methods." National Academies of Sciences, Engineering, and Medicine. 2008. Acceptance Tests for Surface Characteristics of Steel Strands in Prestressed Concrete. Washington, DC: The National Academies Press. doi: 10.17226/14206.
×
Page 55
Page 56
Suggested Citation:"Appendix B - Evaluation of Mechanical and Chemical Test Methods." National Academies of Sciences, Engineering, and Medicine. 2008. Acceptance Tests for Surface Characteristics of Steel Strands in Prestressed Concrete. Washington, DC: The National Academies Press. doi: 10.17226/14206.
×
Page 56
Page 57
Suggested Citation:"Appendix B - Evaluation of Mechanical and Chemical Test Methods." National Academies of Sciences, Engineering, and Medicine. 2008. Acceptance Tests for Surface Characteristics of Steel Strands in Prestressed Concrete. Washington, DC: The National Academies Press. doi: 10.17226/14206.
×
Page 57
Page 58
Suggested Citation:"Appendix B - Evaluation of Mechanical and Chemical Test Methods." National Academies of Sciences, Engineering, and Medicine. 2008. Acceptance Tests for Surface Characteristics of Steel Strands in Prestressed Concrete. Washington, DC: The National Academies Press. doi: 10.17226/14206.
×
Page 58
Page 59
Suggested Citation:"Appendix B - Evaluation of Mechanical and Chemical Test Methods." National Academies of Sciences, Engineering, and Medicine. 2008. Acceptance Tests for Surface Characteristics of Steel Strands in Prestressed Concrete. Washington, DC: The National Academies Press. doi: 10.17226/14206.
×
Page 59
Page 60
Suggested Citation:"Appendix B - Evaluation of Mechanical and Chemical Test Methods." National Academies of Sciences, Engineering, and Medicine. 2008. Acceptance Tests for Surface Characteristics of Steel Strands in Prestressed Concrete. Washington, DC: The National Academies Press. doi: 10.17226/14206.
×
Page 60
Page 61
Suggested Citation:"Appendix B - Evaluation of Mechanical and Chemical Test Methods." National Academies of Sciences, Engineering, and Medicine. 2008. Acceptance Tests for Surface Characteristics of Steel Strands in Prestressed Concrete. Washington, DC: The National Academies Press. doi: 10.17226/14206.
×
Page 61
Page 62
Suggested Citation:"Appendix B - Evaluation of Mechanical and Chemical Test Methods." National Academies of Sciences, Engineering, and Medicine. 2008. Acceptance Tests for Surface Characteristics of Steel Strands in Prestressed Concrete. Washington, DC: The National Academies Press. doi: 10.17226/14206.
×
Page 62
Page 63
Suggested Citation:"Appendix B - Evaluation of Mechanical and Chemical Test Methods." National Academies of Sciences, Engineering, and Medicine. 2008. Acceptance Tests for Surface Characteristics of Steel Strands in Prestressed Concrete. Washington, DC: The National Academies Press. doi: 10.17226/14206.
×
Page 63
Page 64
Suggested Citation:"Appendix B - Evaluation of Mechanical and Chemical Test Methods." National Academies of Sciences, Engineering, and Medicine. 2008. Acceptance Tests for Surface Characteristics of Steel Strands in Prestressed Concrete. Washington, DC: The National Academies Press. doi: 10.17226/14206.
×
Page 64
Page 65
Suggested Citation:"Appendix B - Evaluation of Mechanical and Chemical Test Methods." National Academies of Sciences, Engineering, and Medicine. 2008. Acceptance Tests for Surface Characteristics of Steel Strands in Prestressed Concrete. Washington, DC: The National Academies Press. doi: 10.17226/14206.
×
Page 65
Page 66
Suggested Citation:"Appendix B - Evaluation of Mechanical and Chemical Test Methods." National Academies of Sciences, Engineering, and Medicine. 2008. Acceptance Tests for Surface Characteristics of Steel Strands in Prestressed Concrete. Washington, DC: The National Academies Press. doi: 10.17226/14206.
×
Page 66
Page 67
Suggested Citation:"Appendix B - Evaluation of Mechanical and Chemical Test Methods." National Academies of Sciences, Engineering, and Medicine. 2008. Acceptance Tests for Surface Characteristics of Steel Strands in Prestressed Concrete. Washington, DC: The National Academies Press. doi: 10.17226/14206.
×
Page 67
Page 68
Suggested Citation:"Appendix B - Evaluation of Mechanical and Chemical Test Methods." National Academies of Sciences, Engineering, and Medicine. 2008. Acceptance Tests for Surface Characteristics of Steel Strands in Prestressed Concrete. Washington, DC: The National Academies Press. doi: 10.17226/14206.
×
Page 68
Page 69
Suggested Citation:"Appendix B - Evaluation of Mechanical and Chemical Test Methods." National Academies of Sciences, Engineering, and Medicine. 2008. Acceptance Tests for Surface Characteristics of Steel Strands in Prestressed Concrete. Washington, DC: The National Academies Press. doi: 10.17226/14206.
×
Page 69
Page 70
Suggested Citation:"Appendix B - Evaluation of Mechanical and Chemical Test Methods." National Academies of Sciences, Engineering, and Medicine. 2008. Acceptance Tests for Surface Characteristics of Steel Strands in Prestressed Concrete. Washington, DC: The National Academies Press. doi: 10.17226/14206.
×
Page 70
Page 71
Suggested Citation:"Appendix B - Evaluation of Mechanical and Chemical Test Methods." National Academies of Sciences, Engineering, and Medicine. 2008. Acceptance Tests for Surface Characteristics of Steel Strands in Prestressed Concrete. Washington, DC: The National Academies Press. doi: 10.17226/14206.
×
Page 71
Page 72
Suggested Citation:"Appendix B - Evaluation of Mechanical and Chemical Test Methods." National Academies of Sciences, Engineering, and Medicine. 2008. Acceptance Tests for Surface Characteristics of Steel Strands in Prestressed Concrete. Washington, DC: The National Academies Press. doi: 10.17226/14206.
×
Page 72
Page 73
Suggested Citation:"Appendix B - Evaluation of Mechanical and Chemical Test Methods." National Academies of Sciences, Engineering, and Medicine. 2008. Acceptance Tests for Surface Characteristics of Steel Strands in Prestressed Concrete. Washington, DC: The National Academies Press. doi: 10.17226/14206.
×
Page 73
Page 74
Suggested Citation:"Appendix B - Evaluation of Mechanical and Chemical Test Methods." National Academies of Sciences, Engineering, and Medicine. 2008. Acceptance Tests for Surface Characteristics of Steel Strands in Prestressed Concrete. Washington, DC: The National Academies Press. doi: 10.17226/14206.
×
Page 74
Page 75
Suggested Citation:"Appendix B - Evaluation of Mechanical and Chemical Test Methods." National Academies of Sciences, Engineering, and Medicine. 2008. Acceptance Tests for Surface Characteristics of Steel Strands in Prestressed Concrete. Washington, DC: The National Academies Press. doi: 10.17226/14206.
×
Page 75
Page 76
Suggested Citation:"Appendix B - Evaluation of Mechanical and Chemical Test Methods." National Academies of Sciences, Engineering, and Medicine. 2008. Acceptance Tests for Surface Characteristics of Steel Strands in Prestressed Concrete. Washington, DC: The National Academies Press. doi: 10.17226/14206.
×
Page 76
Page 77
Suggested Citation:"Appendix B - Evaluation of Mechanical and Chemical Test Methods." National Academies of Sciences, Engineering, and Medicine. 2008. Acceptance Tests for Surface Characteristics of Steel Strands in Prestressed Concrete. Washington, DC: The National Academies Press. doi: 10.17226/14206.
×
Page 77
Page 78
Suggested Citation:"Appendix B - Evaluation of Mechanical and Chemical Test Methods." National Academies of Sciences, Engineering, and Medicine. 2008. Acceptance Tests for Surface Characteristics of Steel Strands in Prestressed Concrete. Washington, DC: The National Academies Press. doi: 10.17226/14206.
×
Page 78
Page 79
Suggested Citation:"Appendix B - Evaluation of Mechanical and Chemical Test Methods." National Academies of Sciences, Engineering, and Medicine. 2008. Acceptance Tests for Surface Characteristics of Steel Strands in Prestressed Concrete. Washington, DC: The National Academies Press. doi: 10.17226/14206.
×
Page 79
Page 80
Suggested Citation:"Appendix B - Evaluation of Mechanical and Chemical Test Methods." National Academies of Sciences, Engineering, and Medicine. 2008. Acceptance Tests for Surface Characteristics of Steel Strands in Prestressed Concrete. Washington, DC: The National Academies Press. doi: 10.17226/14206.
×
Page 80
Page 81
Suggested Citation:"Appendix B - Evaluation of Mechanical and Chemical Test Methods." National Academies of Sciences, Engineering, and Medicine. 2008. Acceptance Tests for Surface Characteristics of Steel Strands in Prestressed Concrete. Washington, DC: The National Academies Press. doi: 10.17226/14206.
×
Page 81
Page 82
Suggested Citation:"Appendix B - Evaluation of Mechanical and Chemical Test Methods." National Academies of Sciences, Engineering, and Medicine. 2008. Acceptance Tests for Surface Characteristics of Steel Strands in Prestressed Concrete. Washington, DC: The National Academies Press. doi: 10.17226/14206.
×
Page 82
Page 83
Suggested Citation:"Appendix B - Evaluation of Mechanical and Chemical Test Methods." National Academies of Sciences, Engineering, and Medicine. 2008. Acceptance Tests for Surface Characteristics of Steel Strands in Prestressed Concrete. Washington, DC: The National Academies Press. doi: 10.17226/14206.
×
Page 83
Page 84
Suggested Citation:"Appendix B - Evaluation of Mechanical and Chemical Test Methods." National Academies of Sciences, Engineering, and Medicine. 2008. Acceptance Tests for Surface Characteristics of Steel Strands in Prestressed Concrete. Washington, DC: The National Academies Press. doi: 10.17226/14206.
×
Page 84
Page 85
Suggested Citation:"Appendix B - Evaluation of Mechanical and Chemical Test Methods." National Academies of Sciences, Engineering, and Medicine. 2008. Acceptance Tests for Surface Characteristics of Steel Strands in Prestressed Concrete. Washington, DC: The National Academies Press. doi: 10.17226/14206.
×
Page 85
Page 86
Suggested Citation:"Appendix B - Evaluation of Mechanical and Chemical Test Methods." National Academies of Sciences, Engineering, and Medicine. 2008. Acceptance Tests for Surface Characteristics of Steel Strands in Prestressed Concrete. Washington, DC: The National Academies Press. doi: 10.17226/14206.
×
Page 86
Page 87
Suggested Citation:"Appendix B - Evaluation of Mechanical and Chemical Test Methods." National Academies of Sciences, Engineering, and Medicine. 2008. Acceptance Tests for Surface Characteristics of Steel Strands in Prestressed Concrete. Washington, DC: The National Academies Press. doi: 10.17226/14206.
×
Page 87
Page 88
Suggested Citation:"Appendix B - Evaluation of Mechanical and Chemical Test Methods." National Academies of Sciences, Engineering, and Medicine. 2008. Acceptance Tests for Surface Characteristics of Steel Strands in Prestressed Concrete. Washington, DC: The National Academies Press. doi: 10.17226/14206.
×
Page 88
Page 89
Suggested Citation:"Appendix B - Evaluation of Mechanical and Chemical Test Methods." National Academies of Sciences, Engineering, and Medicine. 2008. Acceptance Tests for Surface Characteristics of Steel Strands in Prestressed Concrete. Washington, DC: The National Academies Press. doi: 10.17226/14206.
×
Page 89
Page 90
Suggested Citation:"Appendix B - Evaluation of Mechanical and Chemical Test Methods." National Academies of Sciences, Engineering, and Medicine. 2008. Acceptance Tests for Surface Characteristics of Steel Strands in Prestressed Concrete. Washington, DC: The National Academies Press. doi: 10.17226/14206.
×
Page 90
Page 91
Suggested Citation:"Appendix B - Evaluation of Mechanical and Chemical Test Methods." National Academies of Sciences, Engineering, and Medicine. 2008. Acceptance Tests for Surface Characteristics of Steel Strands in Prestressed Concrete. Washington, DC: The National Academies Press. doi: 10.17226/14206.
×
Page 91
Page 92
Suggested Citation:"Appendix B - Evaluation of Mechanical and Chemical Test Methods." National Academies of Sciences, Engineering, and Medicine. 2008. Acceptance Tests for Surface Characteristics of Steel Strands in Prestressed Concrete. Washington, DC: The National Academies Press. doi: 10.17226/14206.
×
Page 92
Page 93
Suggested Citation:"Appendix B - Evaluation of Mechanical and Chemical Test Methods." National Academies of Sciences, Engineering, and Medicine. 2008. Acceptance Tests for Surface Characteristics of Steel Strands in Prestressed Concrete. Washington, DC: The National Academies Press. doi: 10.17226/14206.
×
Page 93
Page 94
Suggested Citation:"Appendix B - Evaluation of Mechanical and Chemical Test Methods." National Academies of Sciences, Engineering, and Medicine. 2008. Acceptance Tests for Surface Characteristics of Steel Strands in Prestressed Concrete. Washington, DC: The National Academies Press. doi: 10.17226/14206.
×
Page 94
Page 95
Suggested Citation:"Appendix B - Evaluation of Mechanical and Chemical Test Methods." National Academies of Sciences, Engineering, and Medicine. 2008. Acceptance Tests for Surface Characteristics of Steel Strands in Prestressed Concrete. Washington, DC: The National Academies Press. doi: 10.17226/14206.
×
Page 95
Page 96
Suggested Citation:"Appendix B - Evaluation of Mechanical and Chemical Test Methods." National Academies of Sciences, Engineering, and Medicine. 2008. Acceptance Tests for Surface Characteristics of Steel Strands in Prestressed Concrete. Washington, DC: The National Academies Press. doi: 10.17226/14206.
×
Page 96
Page 97
Suggested Citation:"Appendix B - Evaluation of Mechanical and Chemical Test Methods." National Academies of Sciences, Engineering, and Medicine. 2008. Acceptance Tests for Surface Characteristics of Steel Strands in Prestressed Concrete. Washington, DC: The National Academies Press. doi: 10.17226/14206.
×
Page 97
Page 98
Suggested Citation:"Appendix B - Evaluation of Mechanical and Chemical Test Methods." National Academies of Sciences, Engineering, and Medicine. 2008. Acceptance Tests for Surface Characteristics of Steel Strands in Prestressed Concrete. Washington, DC: The National Academies Press. doi: 10.17226/14206.
×
Page 98
Page 99
Suggested Citation:"Appendix B - Evaluation of Mechanical and Chemical Test Methods." National Academies of Sciences, Engineering, and Medicine. 2008. Acceptance Tests for Surface Characteristics of Steel Strands in Prestressed Concrete. Washington, DC: The National Academies Press. doi: 10.17226/14206.
×
Page 99
Page 100
Suggested Citation:"Appendix B - Evaluation of Mechanical and Chemical Test Methods." National Academies of Sciences, Engineering, and Medicine. 2008. Acceptance Tests for Surface Characteristics of Steel Strands in Prestressed Concrete. Washington, DC: The National Academies Press. doi: 10.17226/14206.
×
Page 100
Page 101
Suggested Citation:"Appendix B - Evaluation of Mechanical and Chemical Test Methods." National Academies of Sciences, Engineering, and Medicine. 2008. Acceptance Tests for Surface Characteristics of Steel Strands in Prestressed Concrete. Washington, DC: The National Academies Press. doi: 10.17226/14206.
×
Page 101
Page 102
Suggested Citation:"Appendix B - Evaluation of Mechanical and Chemical Test Methods." National Academies of Sciences, Engineering, and Medicine. 2008. Acceptance Tests for Surface Characteristics of Steel Strands in Prestressed Concrete. Washington, DC: The National Academies Press. doi: 10.17226/14206.
×
Page 102
Page 103
Suggested Citation:"Appendix B - Evaluation of Mechanical and Chemical Test Methods." National Academies of Sciences, Engineering, and Medicine. 2008. Acceptance Tests for Surface Characteristics of Steel Strands in Prestressed Concrete. Washington, DC: The National Academies Press. doi: 10.17226/14206.
×
Page 103
Page 104
Suggested Citation:"Appendix B - Evaluation of Mechanical and Chemical Test Methods." National Academies of Sciences, Engineering, and Medicine. 2008. Acceptance Tests for Surface Characteristics of Steel Strands in Prestressed Concrete. Washington, DC: The National Academies Press. doi: 10.17226/14206.
×
Page 104
Page 105
Suggested Citation:"Appendix B - Evaluation of Mechanical and Chemical Test Methods." National Academies of Sciences, Engineering, and Medicine. 2008. Acceptance Tests for Surface Characteristics of Steel Strands in Prestressed Concrete. Washington, DC: The National Academies Press. doi: 10.17226/14206.
×
Page 105
Page 106
Suggested Citation:"Appendix B - Evaluation of Mechanical and Chemical Test Methods." National Academies of Sciences, Engineering, and Medicine. 2008. Acceptance Tests for Surface Characteristics of Steel Strands in Prestressed Concrete. Washington, DC: The National Academies Press. doi: 10.17226/14206.
×
Page 106
Page 107
Suggested Citation:"Appendix B - Evaluation of Mechanical and Chemical Test Methods." National Academies of Sciences, Engineering, and Medicine. 2008. Acceptance Tests for Surface Characteristics of Steel Strands in Prestressed Concrete. Washington, DC: The National Academies Press. doi: 10.17226/14206.
×
Page 107
Page 108
Suggested Citation:"Appendix B - Evaluation of Mechanical and Chemical Test Methods." National Academies of Sciences, Engineering, and Medicine. 2008. Acceptance Tests for Surface Characteristics of Steel Strands in Prestressed Concrete. Washington, DC: The National Academies Press. doi: 10.17226/14206.
×
Page 108
Page 109
Suggested Citation:"Appendix B - Evaluation of Mechanical and Chemical Test Methods." National Academies of Sciences, Engineering, and Medicine. 2008. Acceptance Tests for Surface Characteristics of Steel Strands in Prestressed Concrete. Washington, DC: The National Academies Press. doi: 10.17226/14206.
×
Page 109
Page 110
Suggested Citation:"Appendix B - Evaluation of Mechanical and Chemical Test Methods." National Academies of Sciences, Engineering, and Medicine. 2008. Acceptance Tests for Surface Characteristics of Steel Strands in Prestressed Concrete. Washington, DC: The National Academies Press. doi: 10.17226/14206.
×
Page 110
Page 111
Suggested Citation:"Appendix B - Evaluation of Mechanical and Chemical Test Methods." National Academies of Sciences, Engineering, and Medicine. 2008. Acceptance Tests for Surface Characteristics of Steel Strands in Prestressed Concrete. Washington, DC: The National Academies Press. doi: 10.17226/14206.
×
Page 111
Page 112
Suggested Citation:"Appendix B - Evaluation of Mechanical and Chemical Test Methods." National Academies of Sciences, Engineering, and Medicine. 2008. Acceptance Tests for Surface Characteristics of Steel Strands in Prestressed Concrete. Washington, DC: The National Academies Press. doi: 10.17226/14206.
×
Page 112
Page 113
Suggested Citation:"Appendix B - Evaluation of Mechanical and Chemical Test Methods." National Academies of Sciences, Engineering, and Medicine. 2008. Acceptance Tests for Surface Characteristics of Steel Strands in Prestressed Concrete. Washington, DC: The National Academies Press. doi: 10.17226/14206.
×
Page 113
Page 114
Suggested Citation:"Appendix B - Evaluation of Mechanical and Chemical Test Methods." National Academies of Sciences, Engineering, and Medicine. 2008. Acceptance Tests for Surface Characteristics of Steel Strands in Prestressed Concrete. Washington, DC: The National Academies Press. doi: 10.17226/14206.
×
Page 114
Page 115
Suggested Citation:"Appendix B - Evaluation of Mechanical and Chemical Test Methods." National Academies of Sciences, Engineering, and Medicine. 2008. Acceptance Tests for Surface Characteristics of Steel Strands in Prestressed Concrete. Washington, DC: The National Academies Press. doi: 10.17226/14206.
×
Page 115
Page 116
Suggested Citation:"Appendix B - Evaluation of Mechanical and Chemical Test Methods." National Academies of Sciences, Engineering, and Medicine. 2008. Acceptance Tests for Surface Characteristics of Steel Strands in Prestressed Concrete. Washington, DC: The National Academies Press. doi: 10.17226/14206.
×
Page 116
Page 117
Suggested Citation:"Appendix B - Evaluation of Mechanical and Chemical Test Methods." National Academies of Sciences, Engineering, and Medicine. 2008. Acceptance Tests for Surface Characteristics of Steel Strands in Prestressed Concrete. Washington, DC: The National Academies Press. doi: 10.17226/14206.
×
Page 117
Page 118
Suggested Citation:"Appendix B - Evaluation of Mechanical and Chemical Test Methods." National Academies of Sciences, Engineering, and Medicine. 2008. Acceptance Tests for Surface Characteristics of Steel Strands in Prestressed Concrete. Washington, DC: The National Academies Press. doi: 10.17226/14206.
×
Page 118
Page 119
Suggested Citation:"Appendix B - Evaluation of Mechanical and Chemical Test Methods." National Academies of Sciences, Engineering, and Medicine. 2008. Acceptance Tests for Surface Characteristics of Steel Strands in Prestressed Concrete. Washington, DC: The National Academies Press. doi: 10.17226/14206.
×
Page 119
Page 120
Suggested Citation:"Appendix B - Evaluation of Mechanical and Chemical Test Methods." National Academies of Sciences, Engineering, and Medicine. 2008. Acceptance Tests for Surface Characteristics of Steel Strands in Prestressed Concrete. Washington, DC: The National Academies Press. doi: 10.17226/14206.
×
Page 120
Page 121
Suggested Citation:"Appendix B - Evaluation of Mechanical and Chemical Test Methods." National Academies of Sciences, Engineering, and Medicine. 2008. Acceptance Tests for Surface Characteristics of Steel Strands in Prestressed Concrete. Washington, DC: The National Academies Press. doi: 10.17226/14206.
×
Page 121
Page 122
Suggested Citation:"Appendix B - Evaluation of Mechanical and Chemical Test Methods." National Academies of Sciences, Engineering, and Medicine. 2008. Acceptance Tests for Surface Characteristics of Steel Strands in Prestressed Concrete. Washington, DC: The National Academies Press. doi: 10.17226/14206.
×
Page 122

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.

39 A P P E N D I X B Evaluation of Mechanical and Chemical Test Methods

CONTENTS Introduction ................................................................................................................................................................................... 41 Overview of Test Program ..................................................................................................................................................... 41 Strand Samples ....................................................................................................................................................................... 42 Mechanical Test Methods and Results .......................................................................................................................................... 44 Transfer Length Testing......................................................................................................................................................... 44 Pull-Out Testing..................................................................................................................................................................... 55 Surface and Chemical Test Methods and Results ......................................................................................................................... 69 Contact Angle Measurement................................................................................................................................................. 70 Examination under Ultraviolet Light.................................................................................................................................... 73 Testing pH .............................................................................................................................................................................. 73 Loss on Ignition...................................................................................................................................................................... 77 Loss in Hot Alkali Bath .......................................................................................................................................................... 78 Change in Corrosion Potential.............................................................................................................................................. 79 Surface Roughness ................................................................................................................................................................. 79 Corrosion Rate ....................................................................................................................................................................... 80 Organic Residue Extraction................................................................................................................................................... 81 Atomic Absorption and Colorimetric Analyses ................................................................................................................... 83 Evaluation of Test Methods ........................................................................................................................................................... 84 Variability in Test Methods ................................................................................................................................................... 84 Evaluation of Mechanical Test Methods............................................................................................................................... 84 Evaluation of Surface and Chemical Test Methods.............................................................................................................. 86 Summary of Test Method Correlation................................................................................................................................ 103 Precision Testing .................................................................................................................................................................. 108 Development of Thresholds......................................................................................................................................................... 110 Regression and Prediction Intervals.................................................................................................................................... 111 Thresholds Based on Regression with Single Predictor ..................................................................................................... 117 Thresholds Based on Regression with Multiple Predictors................................................................................................ 119 Summary....................................................................................................................................................................................... 121 References ..................................................................................................................................................................................... 122 40

Introduction A number of test methods were proposed for use as part of a quality control (QC) program to evaluate bond of pre- stressed concrete strand. These were classified as performance- based (i.e., mechanical) tests and surface and chemical tests. The mechanical methods included: • Pull out from concrete, • Pull out from Portland cement mortar, • Pull out from gypsum plaster-based mortar. The surface and chemical methods included: • Contact angle, • Examination under UV light, • pH, • Loss on ignition, • Loss in alkali bath, • Change in corrosion potential, • Corrosion rate, • Surface roughness, • Organic residue extraction/Fourier transform infrared (FTIR) spectroscopical analysis, • Elemental analysis, – AA, – Visible light spectroscopy. These tests, as well as transfer length tests, have been per- formed as part of this research program, which is reviewed in detail in this appendix. The purpose of this program was to determine if any of these proposed tests are applicable for use in a QC program. The QC tests have been divided into two categories, depending on the complexity and time required to conduct the tests: Level I and Level II QC tests. A summary of the tests methods, their QC Level, and the test objectives are given in Tables 2 and 3 of the Project Report. Overview of Test Program To evaluate these test methods, several rounds of exper- imentation were conducted: screening, correlation, and precision. Screening Testing The first round of experiments consisted of “screening” experiments. The objective for the screening experimentation was to eliminate those tests that are not helpful for predicting bond performance. Thus, the first step in this round of test- ing was to estimate the correlation between each surface or chemical test and bond performance; the second step was to identify those methods where some degree of correlation was indicated. For each source of strand, bond performance was measured in terms of pull-out stresses, transfer lengths, or both. Accordingly, for the purpose of this analysis, the bond performance was treated as the independent variable. The best experimental design for estimating a correlation is to place the design points as far apart as possible in terms of the inde- pendent variable. Thus, the optimal statistical design is to run each test on strands that show a range of bonding performance. For the screening experiments high, medium, and low bond- ing sources were desired. However, efforts to obtain a very low bonding strand were not successful. Although reports of low- bonding-strand incidents continue to surface in the precast concrete industry, “unused” samples of such strand remained elusive. Therefore, the screening tests on new strands were run on what are essentially high bond and intermediate bond strands. Some strand from a project in India with apparently very poor bond properties was tested. However, the surface chemicals on this strand did not appear to be the same as for North American manufactured strand, and therefore, the test results could not be directly compared to the rest of the test program. Correlation Testing The second round of experiments was performed for confir- mation and calibration purposes. This round involved running additional tests using those methods that showed promise in the screening experiments. These selected tests were conducted on five new strand samples. This complete data set was then used to assess the correlation between the QC tests and bond performance, and to determine if the tests were able to accu- rately identify good and bad strand. It was also used as a basis for discussing pass/fail criteria for acceptable bond performance. Precision Testing A third round of testing was conducted to determine the precision, that is, repeatability, of those methods showing good correlation with bond strength. This was intended to form the basis of precision statements to be included in pub- lished test methods. Basis for Evaluation—Transfer Length and Pull-Out Tests Transfer length is the most reliable and realistic measure of bond performance. During the screening testing, the evalua- tion of correlations between the pull-out testing and the bond performance were based on performance as measured with transfer length tests conducted on the same source of strands. 41

Pull-out testing was conducted as part of the screening studies using three materials as the test matrix: a concrete, a Portland cement mortar, and a gypsum plaster mortar. Based on comparisons with transfer length tests conducted in this study and described in this document, the concrete pull-out test showed the best correlation with bond quality. The surface and chemical test methods were evaluated in the Screening Round based on the results of pull-out tests from concrete, again on strand samples from the same source. However, the evaluation of correlation of test results to bond in the Correlation Round of testing was based on results from a mortar pull-out test program associated with NCHRP Project 12-60 Transfer, Development, and Splice Length for Strand/Reinforcement in High-Strength Concrete. The Principal Investigator from this project supplied the strand samples for this portion of the study. No concrete pull-out testing was conducted in the Correlation Round of the experimental program. Strand Samples To assess the effectiveness of the mechanical and surface chemistry-based testing procedures, it was essential that samples representing the range of possible performance be evaluated. Since neither precasters nor strand suppliers were enthusiastic about associating themselves with poor-bonding strand, obtaining samples of strand from the lower end of the performance spectrum was difficult. The strand sources included in testing for this program are listed in Table B-1. This table also includes the first slip or 0.1-in. slip stress measured in concrete pull-out tests or in mortar pull-out tests, depending on what was available. Each pull-out stress is the average of the pull-out stresses from at least six individual pieces of strand. The bond stresses are cal- culated based on the actual surface area and the embedment length of the strand. These strand sources fall into three groupings: historic, re- cently manufactured, and OSU (Oklahoma State University). Historic Strand—This study initially identified samples of strand for testing from prior tests conducted at Kansas State University (KSU) by Bob Peterman and at Stresscon by Don Logan that cover a wide range of pull-out behavior. These are referred to as “historic” strand and were manufactured between 1997 and 2004. Figure B-1 is a plot of first slip stress or stress at 0.1-in. end slip versus maximum stress from the data avail- able from historic concrete pull-out tests. When suggested minimum pull-out loads for acceptable bonding performance (suggested by Don Logan of Stresscon, based on a limited number of flexural beam tests conducted in the mid-1990s and his engineering judgment) are con- verted to bond stresses, they are 425 and 955 psi for the first slip and maximum stresses, respectively. These thresholds have been reproduced in Figure B-1. Recently manufactured strand—Figure B-1 also shows the concrete pull-out performance of recently manufactured 42 Strand Geometry Mortar Pull-Out Testing Concrete Pull-Out Testing (LBPT) Strand Source ID Size (in.) Measured Diameter (in.) Pitch (in.) Lay (Handed- ness) Location Date 0.1-in. Slip Stress (psi) Location Date 0.1-in. Slip Stress (psi) Historic Strand KSU-F 1/2 Special 0.524 7 5/8 Left -- -- -- KSU Mar 2004 241 KSU-H 1/2 Special 0.523 7 1/2 Left -- -- -- KSU Mar 2004 209 SC-F 1/2 0.503 8 Left -- -- -- SC May 1997 223 SC-H 1/2 Special 0.530 7 1/4 Left -- -- -- SC Nov 2002 472 SC-IS 1/2 0.501 7 Left -- -- -- SC Mar 2003 682 101 6/10 0.601 8 1/2 Left -- -- -- SC Oct 2004 241 Recently Manufactured Strand 102 1/2 0.501 7 1/2 Left KSU Jun 2005 315 KSU Jun 2005 441 103 1/2 0.503 8 Left KSU Jun 2005 397 KSU Jun 2005 944 151 1/2 Special 0.517 7 1/2 Left KSU Jun 2005 273 KSU Jun 2005 541 153 6/10 0.588 9 Right -- -- -- KSU/SC Jun/Aug 2006 142/406 OSU Strand 349 1/2 0.505 8 3/4 Left OSU Jun 2004 156 -- -- -- 548 1/2 0.500 7 5/8 Left OSU Jan-Feb 2004 623 -- -- -- 697 1/2 0.503 7 1/4 Left OSU May 2004 606 -- -- -- 717 1/2 0.500 8 Left OSU Feb 2004 206 -- -- -- 478 * 1/2 0.499 7 5/8 Left OSU May-June 2004 409 -- -- -- 960 * 1/2 0.500 7 1/2 Left OSU May-June 2004 409 -- -- -- * Samples designated 478 and 960 were from same source. Table B-1. Strand sources.

samples identified during this project. These recently manu- factured samples were obtained in large quantities for the purpose of this research and were used in the screening exper- iments. Notice that the greatest variation in the recently manufactured strand is not in terms of maximum stress but in terms of the early bond stress. The recently manufactured strand sources (102, 103, and 151) were selected because initial testing indicated that they represented a range of first slip pull-out performance. None of these strands had significantly low maximum load pull-out performance. The bond stress at 0.1-in. slip of Source 102, measured in concrete pull-out testing performed as part of this project, is slightly above Logan’s 425 psi threshold. Of the 31 sources pre- sented on this plot, 13 are to the left of Source 102. Only one of these (D-5/96) was available in enough quantity to enable additional testing, and the condition of this strand is variable. As stated, the three recently manufactured sources of strand obtained for testing were identified as 102, 103, and 151. Source 103 is the strand used by Stresscon in their ordinary production of precast/prestressed concrete. It is a source with a proven record of good bond from pull-out tests, flexure beam tests, transfer length tests, and end slips observed in hollow core precast/prestressed concrete members. Source 102 is from a lot of strand delivered to a Midwest precast concrete producer at the same time as another lot from the same strand manufacturer who experienced excessive slip in hollow core products produced by the precaster. After this precaster experienced the slip problems, he contracted with Peterman of KSU to perform concrete pull-out tests. Tests were run on this strand with the results shown in Figure B-2. Later, it was re- ported that this same strand manufacturer was routinely run- ning Portland cement mortar pull-out tests. Source 151 is the same strand manufacturer as 102, but from a precaster in the eastern United States who had not reported any strand slip problems. As received, the spool of strand from the east coast precaster had small areas of rust on the outer windings of the spool. These outer windings were removed, and the tested strand was taken from an inner winding on the same spool. The 102 and 103 strands are both 0.5-in. diameter. The 151 strand is 0.52-in. diameter. One additional recently manufactured strand was received after the completion of the Screening Round of evaluation. This strand was supplied from India and was numbered 153. This strand was tested in portions of the correlation program, but was not included in the correlation analysis, since it was learned that the manufacturing processes used in its production were markedly different from those used to produce strand in the United States. OSU strand—The sample sources used for the Correlation Round of testing were selected by Bruce Russell of OSU. These sources of strand had been tested in work performed by Russell for the NCHRP Project No. 12-60 Transfer, De- velopment, and Splice Length for Strand/Reinforcement in High-Strength Concrete, the Oklahoma Department of Trans- portation, and NASPA, also known as the Committee of the American Wire Products Association. Two of the six strand sources provided by Russell were actually the same strand source, a fact that was not known before the testing was com- pleted. This was intended to test the repeatability of the surface and chemical test methods. Transfer length and mortar pull- out test results were provided in tabular form by Russell after the chemical and surface testing had been completed. These 43 A BC D E F G H J S1 S2 S3 S4 S5 hollowcore H-11/02 U -11/02 F-5/97 IT-11/02 TW-5/96 D-5/96 NC-05/01 IS-3/03 101-A-0.6 102-A-0.5 103-B-0.5102-A-0.5 103-B-0.5 151-Z-0.5 200 400 600 800 1000 1200 200 300 400 500 600 700 800 900 1000 Bond Stress at First Slip or 0.1-in. M ax S tre ss (p si) Historical results at first slip NCHRP tests at 0.1-in. slip Max Threshold 1st Slip Threshold Figure B-1. Correlation between maximum stress and first observed or 0.1-in. slip stress for historic and recently manufactured strand.

tables are reproduced in Table B-2 and Table B-3. Transfer length testing had been conducted on only two sources, while mortar pull-out testing had been conducted on all five sources. Although Russell did not provide a description of his test methodology, it is believed that the data were obtained as described in the recently published Master of Science thesis, “Assessing the Bond Quality of Prestressing Strands Using NASP Bond Test” (Chandran 2006). Russell did not supply the strand tested in Project 12-60 with the lowest bond. Mechanical Test Methods and Results The original project scope included the development of a performance-based test method for use in evaluating strand bond. The Screening Round of the experimental program in- cluded mechanical testing of strand sources using transfer length, pull-out from concrete (large concrete block pull-out test [LBPT]), pull out from Portland cement mortar, and pull out from gypsum plaster-based mortar tests. Each test is pre- sented separately in this section. The number of samples of each strand source tested by each mechanical test method is shown in Table B-4 and B-5. Transfer Length Testing Transfer length testing was conducted in three rounds, with its procedure changing slightly for each subsequent round of testing. During the Trial Round, two 4 in. × 4 in. × 16 ft-prisms were cast for strand designations 102 and 151. The size of the prisms was based on two considerations: (1) the cross-sectional area of the prism was sized to obtain a concrete compressive stress of approximately 2000 psi after release of prestress; (2) its length was designed to obtain four transfer lengths per specimen—two from the initial release of prestress and two from saw-cutting the specimen at its mid-point. The prism had to be long enough to assure that the transfer lengths from the four ends would not overlap. Figure B-3 shows a schematic of the transfer length speci- men, as well as the concrete mixture proportions used for the Trial Round and for Round 1. The cement was a Type III with a blaine fineness of 564 m2/kg. The coarse aggregate was a siliceous gravel meeting ASTM C33 #67 gradation require- ments, and the fine aggregate had a fineness modulus of 2.89. The mixture proportions were determined essentially accord- ing to the procedure given in ACI 211.1-91, Standard Practice for Selecting Proportions for Normal, Heavyweight, and Mass Concrete. Figure B-4 schematically shows the transfer length specimen mounted in its stressing frame. Figure B-5 shows one such frame ready for casting, with prestressing strand tensioned in the form. After the prisms were cast and the formwork was stripped, Whittemore points were attached to both sides of the prisms as shown in Figure B-6. The strain in the beam relative to the position along the beam was measured based on readings taken with a Whittemore gauge (Figure B-7). This device measures the change in location of Whittemore 44 Figure B-2. Large concrete block pull-out behavior of strands tested at KSU that demonstrated excessive slip in a hollow-core product. In this figure, behavior is compared against that of a well-bonding control.

Table B-2. NASPA (mortar) pull-out and transfer length test data accompanying strand from OSU.

46 Pullout Bond Stress Contact Angle pH Strand Source ID Transfer Length Concrete Mortar Hydrocal As- Rec’d After Ca(OH)2 After Ignition pH Meter Duotest pH-Fix 7.5-9.5 Historic Strand KSU-F -- 6 -- -- 1 1 -- -- -- -- KSU-H -- 6 -- -- 2 2 -- -- -- -- SC-F -- 6 -- -- 3 3 -- -- -- -- SC-H -- 6 -- -- 3 3 -- -- -- -- SC-IS -- 6 -- -- 3 3 -- -- -- -- 101 -- 6 -- -- 3 3 -- -- -- -- Recently Manufactured Strand 102 6 6 -- 6 3 3 -- 2 2 2 103 6 6 -- 6 3 3 -- 2 2 2 151 6 6 -- 6 3 3 -- 2 2 2 153 -- 6 -- -- -- 3 -- -- 6 -- OSU Strand Samples 349 -- -- 6 -- -- 3 3 -- 6 -- 548 -- -- 18 -- -- 3 3 -- 6 -- 697 -- -- 11 -- -- 3 3 -- 6 -- 717 -- -- 6 -- -- 3 3 -- 6 -- 478 * -- -- 12 -- -- 3 3 -- 6 -- 960 * -- -- 12 -- -- 3 3 -- 6 -- * Samples designated 478 and 960 were from same source. Table B-3. Transfer length and beam test data accompanying strand from OSU. Table B-4. Number of samples tested for each mechanical and surface and chemical test method.

points to 1/10,000th of an inch. The Whittemore points are cylindrical buttons that contain a central indentation into which styli on the gauge are inserted. After the initial readings were recorded, the strand was re- leased gently, and surface strains were measured. These two trial tests resulted in transfer lengths of 20 to 28 in. for each source. It was determined that the concrete strength may have been too high and the release method too gradual to provide a true measure of transfer length. In Round 1, two prisms were cast for each of the three sources of recently manufactured strand (102, 103, and 151); a total of six stressing frames similar to those needed for the Trial Round were used. After casting, in preparation for measuring the strand end slip, a small notch was made at the exposed ends of the strand. The distances from this small cut to the end of the prism were measured with calipers, in digi- tal photos, and with a steel scale (see Figure B-8). This posi- tion was later re-measured with all three of these methods after release and at later times to determine the end slip of the strand. Initially, the release was achieved by quickly cutting the strand with an acetylene torch similar to what is commonly done in precast concrete plants. However, this resulted in splitting and fracturing at the end region of the concrete for three specimens (Figure B-9). For the remaining three specimens, release was achieved by gradually heating up the strand until it failed in tension due to its reduced tensile strength at high temperatures. Even using this method, one of the three remaining prisms split at its end. The strength of the concrete at the time of prestress transfer was 4060 psi. The strength of the concrete at the time of center was 4810 psi. The strain versus position was determined after the strand was released. The prisms were then cut in half (Figure B-10), and the strain in the prisms was measured along with the end slip (measured at the far ends of the prisms) and strand suck-in (measured at the saw-cut ends). To examine the effect of time on transfer length, end slip, and suck-in, these measurements were repeated 28 days, 6 months, and 22 months after release. Strand suck-in measurements were abandoned after the 28-day data collection because of their inefficacy. During the initial saw-cutting process, when the saw initially nicked the strand, the individual wires fractured sequentially, leaving an uneven surface from which accurate measurement of the strand displacement proved dif- ficult (see Figure B-11). A typical plot of strain versus position on the prisms is given in Figure B-12. This plot shows several sets of strain 47 LAB Change in Corrosion Potential Corrosion Rate AA and Color. Analysis Strand Source ID LOI Method 1 Method 2 As- Rec’d After Ca(OH)2 After Ignition Surf. Rough. As- Rec’d After Ca(OH)2 After Ignition Org. Ext. Res. Water Acid Historic Strand KSU-F 1 2 -- -- -- -- -- 1 -- -- 2 2A 2A KSU-H 2 2 -- -- -- -- -- 2 -- -- 2 2A 2A SC-F 2 2 -- -- -- -- -- - -- -- 3 3A 3A SC-H 2 2 -- -- -- -- -- 2 -- -- 3 3A 3A SC-IS 2 2 -- -- -- -- -- 2 -- -- 3 2A 2A 101 2 2 -- 2 -- -- -- 2 -- -- 3 3AB 3A Recently Manufactured Strand 102 2 2 2 3 1 -- 6 2 1 -- 3 3AB 3A 103 2 2 2 3 1 -- 6 2 1 -- 3 3AB 3A 151 2 2 2 3 1 -- 6 2 1 -- 3 3AB 3A 153 -- -- -- 3 -- 3 12 -- 3 3 3 3 3 OSU Strand Samples 349 3 -- -- 4 -- 3 12 -- 3 3 3 3 3 548 3 -- -- 3 -- 3 12 -- 3 3 3 3 3 697 3 -- -- 3 -- 3 12 -- 3 3 3 3 3 717 3 -- -- 3 -- 3 12 -- 3 3 3 3 3 478 * 3 -- -- 3 -- 3 12 -- 3 3 3 3 3 960 * 3 -- -- 3 -- 3 12 -- 3 3 3 3 3 * Samples designated 478 and 960 were from same source. A Acid wash performed on strand subsequent to water wash. B Both warm- and hot-water washes performed. Table B-5. Number of samples tested for each mechanical and surface and chemical test method.

data collected over a series of months as detailed above. Additionally, this plot shows lines drawn at 100% of the plateau strain for each measurement. The transfer length measured in this test is defined as the intersection of the strain on this plateau and the linear extension of the slope from the transfer region. Because of the problems experienced with this first set of transfer length specimens, a second set was fabricated and tested (Round 2). In Round 2, the use of saw cutting to meas- ure strand suck-in was omitted due to the previously men- tioned ineffectiveness of this method. End slip measurements for this round of testing were continued using the scale and photos. End slips were also measured using a depth gauge. As a final change, in Round 2, the cross-sectional dimensions of the prisms were increased to prevent splitting on rapid re- lease of prestress. Rapid release was not abandoned, as was done by necessity during Round 1, because releasing the strand in this way realistically imitates what is done at a typical precaster’s yard. To prevent end-splitting, specimen cross-sections were increased to 4.5 in. × 4.5 in., which re- duced the compressive stress on the concrete to 1500 psi. The Round 2 specimen and stressing frame are presented schematically in Figure B-13 and Figure B-14, respectively. Figure B-13 also lists the concrete mixture proportions for Round 2. The strength of the concrete at the time of prestress transfer was 4380 psi. Because the center saw cut was eliminated, it was possible to cast two 8-ft long specimens end-to-end within the fabri- cation frame. A space was left between the specimens using a stiff standoff, exposing a portion of the strand. The strand in this segment between the two specimens was torch cut after the initial release (see Figure B-15). 48 Figure B-3. Transfer length specimen used in the Trial Test and Round 1 Test.

Two additional issues were experienced in this second round: 1. While making the notch in the wire for the end slip meas- urement on two of the prisms, the individual wire fractured and some loss of prestress occurred. 2. It was not initially known that one of the strands was 0.52-in. diameter (1/2 in. special) instead of 0.5-in. diameter. Thus, this larger strand had a slightly lower bond stress at transfer. To account for these issues and enable comparisons of stress transfer behavior between the tested strand sources, the average bond stress over the transfer length was calculated based on the measured transfer length and strand tensile stress. This approach eliminates complications from strands of vary- ing sizes and varying initial stress conditions. The average bond stress, Ut, is calculated as (Eq. 1) where fse is the effective prestress after transfer, Aps is the cross-sectional area of the strand, Cp is the circumferential perimeter of the strand (4/3 π db) and Lt is the transfer length. Average bond stress is thus dependent both on effective pre- stress as well as transfer length for a given strand geometry. For the specimens from the two rounds of testing, a transfer U f A C L t se ps p t = 49 Figure B-4. Transfer length stressing frame and specimen orientation, Trial Test and Round 1 Test.

length of 30 in. translated to an average bond stress of 375 to 425 psi. The approach used to evaluate the average bond stress is the same as that used to derive the formula for predicting transfer length in ACI 318 Section 12.9, which is (Eq. 2)L f d t se b = 3 where Lt = transfer length in inches, fse = the effective stress in the strand after the losses, and db = nominal diameter of the strand in inches. The “3” in the denominator is based on an assumed average bond stress of 400 psi and the prevailing strand geometry in the 1950s when the formula was devel- oped (Tabatabai and Dickson 1993). For the purpose of this calculation, fse was taken as the difference between the stresses in the strand before release and the elastic losses only. The elastic loss was determined based on the strain measured im- mediately after release in the central region of the test prism over which the strain is approximately constant, assuming no relaxation losses in the strand. Results of Transfer Length Testing Table B-6 and Table B-7 show the average bond stress over transfer length measured for Rounds 1 and 2, respectively. The combined results of Rounds 1 and 2 are presented in Table B-8 and graphically in Figure B-16. Table B-9 presents coefficients of variation for these results. The combined Round 1 and 2 results are averages of six transfer length measurements from the two rounds from each source: two from Round 1 and four from Round 2. Other data collected during Round 1 were left out of this inter-round analysis because of the procedural differences between the two rounds relating to prestress release. The data omitted (grayed in Table B-6) represented bond stresses obtained from the end of the prisms at which the strands were released and those from the prisms saw-cut ends. The end slip was measured by three methods: using a scale, using calipers/depth gauge, and comparing digital photos taken before and after release. Figure B-17 presents an 50 Figure B-5. Tensioned strand in form used for casting transfer length specimen. Figure B-6. Specimens with Whittemore buttons attached. Figure B-7. Whittemore gauge readings used to calculate strain.

average of the end slips measured with a scale for both rounds of testing excluding data omitted for inter-round bond stress analysis. Previous researchers (Rose and Russell 1997) demon- strated that transfer length is proportional to end slip. Therefore, since transfer length is in the denominator of the expression given as Eq. 2, it would be expected that average bond stress over transfer length is inversely proportional to end slip. This has been confirmed in this study as shown in Figure B-18, which depicts the correlation between bond stress and the inverse of end slip for measurements made using a scale, photos, and a set of calipers. Each point on this figure represents the average of several measurements made on a single strand source by a given method. The methods of end slip measurement that were judged best 51 Figure B-8. Measurement of strand slippage from dead end—Strand 103B. Figure B-9. Fractured end of prism after rapid release of strand. Figure B-10. Cutting specimen in half.

were the use of a scale or calipers, because they correlated best with the bond stress calculated from surface strain measurements. Discussion of Transfer Length Testing The average bond stress over the transfer length decreased over time. The rate of this decrease was typically largest dur- ing the first month after transfer and then was roughly con- stant during the next two time spans (28 days to 6 months and 6 months to 18 to 22 months). Source 102, initially the poorest bonding strand, displayed less change in bond stress over time. While the average bond stress decreased, the end slips increased. Strand from the poorest-bonding source (102) had the highest end slips; the best-bonding strand had the lowest end slips. While a correlation exists between end slip and transfer length or bond over transfer length measurements, it is not without some limitations. Examining the end slip measurements for Source 102 over time (Figure B-17) can reveal the limits of the end slip-average bond stress over transfer length correlation. The measured average end slip decreased between 28-day and 6-month measurements; this 52 Figure B-11. Strand appearance after mid-point saw cut of transfer length prism. 0 200 400 600 800 1000 1200 1400 1600 1800 2000 0 20 40 60 80 100 120 140 160 180 Position (in.) St ra in (u E) After Release Strain Plateau After Release After Cut Strain Plateau (S) - After Cut Strain Plateau (N) - After Cut 28 Days After Release Strain Plateau (S) - 28 days Strain Plateau (N) - 28 days 6 Mos. After Release Strain Plateau (S) - 6 mos. Strain Plateau (N) - 6 mos. 22 Mos. After Release Strain Plateau (S) - 22 mos. Strain Plateau (N) - 22 mos. Figure B-12. Strains measured on transfer length beam containing strand from Source 103.

apparent trend is likely related to the measurement technique and no increase was noted in the bond stress for these data points. The largest increase in end slip for all sources occurred between the last two measurements, while the largest de- crease in bond stress typically occurred between the first two. Due to the imperfect correlation between end slip and transfer length bond stress, caution should be used when extrapolating transfer length stress or transfer length itself from end slip. In part, this is because end slip is a difficult property to measure and a certain amount of inaccuracy can be expected in this single measurement. The transfer length and the aver- age bond stress over the transfer length are calculated based on numerous discrete measurements. Multiple methods for measuring end slip were employed in this study; all were difficult to correlate with transfer length bond stress. It was initially thought that using digital photos might decrease error, but instead, this method proved less effective than the alternatives. Due to the large amount of time required to perform the photo analysis and the lack of improvement in the results, it seems that measuring end slip directly, either with a scale or calipers, was a better method. Measurements made using a depth gauge were ineffective, just as were attempted measurements of strand suck-in at cut beam ends for Round 1. Recent research (Peterman 2007) has indicated that greater and more variable transfer lengths may be experi- enced by strand located within 8 in. of an as-cast top surface. This may indicate that the specimen size used in this study may have been too shallow. In the transfer length specimens, the strands are about 2 in. below the as-cast top surface and would therefore be expected to suffer from this effect. Furthermore, research has indicated a strong correlation between concrete strength and bond properties (Stocker and 53 Figure B-13. Transfer length specimen used in the Round 2 Test.

Sozen 1971, Lane 1998, Mitchell et al. 1993). Thus, in an effort to differentiate the innate bond properties of a strand from the results of transfer length tests, a concrete strength correction factor should be employed to interpret transfer length data. Transfer lengths determined by side strains measured using mechanical strain gages (Whittemore or DEMEC gages) are considered less reliable as the concrete ages. Creep of the con- crete results in a non-linear relationship between the strain in the concrete and the strain in the steel. Instead, recent research suggests that the slip of the strand may be a better indicator of transfer length over a long period of time. This measurement has proven to be a better long-term indicator of transfer length and is easier to perform than the mechanical strain measurement. Transfer Length Testing for Correlation Round of Evaluation As mentioned, the sampled sources used for the Correlation Round of testing were selected by Bruce Russell of OSU. Transfer length data were provided accompanying two of the sources (see Table B-2 and Table B-3) and it is assumed the testing protocol is as reported by Chandran (2006). Russell reported transfer length data calculated from measured end slips and measured directly from surface strain measurements 54 Figure B-14. Transfer length stressing frame and specimen orientation Round 2 Test.

(DEMEC). These data were provided for Sources 697 and 717. The transfer lengths reported for Source 717 based on surface strain data were about 25 in. However, based on end slip meas- urements, the reported transfer lengths for Source 717 were 38 to 75 in. Agreement between the measurement methods was better for Source 697, with surface-strain based transfer lengths ranging from 24 to 27 in. and end slip based transfer lengths ranging from 19 to 24 in. Because of the large unex- plained variation between the results of the two transfer length measurement methods, and because there were only two data points available, the transfer length data were not considered in the correlation analysis conducted to evaluate surface and chemical test methods. Lastly, for the same two sources, Russell provided results from flexural beam tests conducted to evaluate development length, which similarly were not analyzed. Pull-Out Testing A key component of this project as originally defined was the development of a procedure for qualifying the bond char- acteristics of prestressing strand using some sort of pull-out test of untensioned strand. Pull-out tests to determine the bond characteristics of steel in concrete have been used for over 100 years. Pull-out tests have been used to judge the bond quality of prestressing strand for almost 50 years. Refinements to historic pull-out test methods continue to develop. These refinements deal not so much with the basic methodology, but rather with the small variations in testing procedures. At the current time, two types of pull-out tests are most common and appear to be the most viable. The first method involves pulling untensioned strand out of a block of concrete. The second method involves pulling untensioned strand out of a steel cylinder filled with mortar. 55 Figure B-15. Strands between specimens after torch cut. Average Bond Stress over Transfer Length (psi) Strand Sample ID Beam End Initial After Cut 28 Days 6 Months 22 Months S-S 437 423 380 364 328 S-N -- 596 558 477 397 N-S -- 525 486 437 380 102-A-0.5-1A N-N 375 354 336 345 341 S-S 397 436 422 385 364 S-N -- 569 569 524 468 N-S -- 513 476 476 416 102-A-0.5-2A N-N 344 364 369 359 331 S-S 422 436 429 390 357 S-N -- 627 547 485 459 N-S -- 612 485 436 402 103-B-0.5-1A N-N 485 443 436 415 378 S-S 514 475 450 428 428 S-N -- 694 626 571 494 N-S -- 626 546 514 459 103-B-0.5-2A N-N 597 514 475 467 428 S-S 371 341 319 332 323 S-N -- 586 413 355 345 N-S -- 360 341 336 327 151-Z-0.5-1C N-N 332 336 319 315 315 S-S 349 359 354 334 309 S-N -- 399 369 330 302 N-S -- 434 427 419 359 151-Z-0.5-2C N-N 330 309 302 309 258 Bond stress values shaded in gray were measured at central cut or at the end of specimen influenced by release method and were ignored during computation of inter-round averages. Table B-6. Average bond stress over transfer length from Round 1.

In its current form, the concrete pull-out test resembles a method developed by Moustafa (1974). The method was primarily developed to judge the capacity of strand to be used as lifting loops to handle product during shipping and erection. The test developed by Moustafa was modified by Logan (1997) to judge the bond quality of strand in preten- sioned applications. Further developments of the method have occurred and are the basis for the testing reported herein. In its current form, the mortar pull-out test method resem- bles a method originally developed for the Post Tensioning Institute in 1994 (Hyett et al., 1994, Post-Tensioning Institute 1996). The method was primarily developed to judge the bond quality of prestressing strand used in rock anchors. The method became the basis of ASTM A981-97(2002), Standard Test Method for Evaluating Bond Strength for 15.2 mm (0.6 in.) Diameter Prestressing Steel Strand, Grade 270, Uncoated, Used in Prestressed Ground Anchors. Later, this method was modi- fied by Russell and Paulsgrove (1999) for NASPA and became known as the NASPA test. The NASPA test has been modi- fied slightly by this research project to make it less sensitive to the test apparatus. One of the goals of this project was to try to eliminate variables by using standardized and universally available embedment media referred to in the NCHRP Project 10-62 Request for Proposal as a “surrogate homogeneous material.” Accordingly, a third type of pull-out test was attempted: pulling untensioned strand out of a steel pipe filled with a modified gypsum plaster (Hydrocal). The three types of pull-out tests were performed on the three sources of strand at KSU in March and May of 2005. In summary, these are: (1) LBPTs, (2) ordinary mortar cylin- der pull-out tests, and (3) Hydrocal mortar cylinder pull-out tests. Each of these test procedures and the results are described in detail below. Large Concrete Block Pull-Out Test In the LBPT, six prestressing strands are embedded in a 2 ft × 2 ft × 32 in. block of concrete. (Note—the length of the block can vary, depending on the number of strands being tested simultaneously.) The bottom 4 in. and top 2 in. of the strand are debonded, leaving an 18-in. long bond length. The layout of the test specimen is presented in Figure B-19, and the test frame is shown in Figure B-20. The concrete mix and other test parameters are as follows: Concrete mix (quantities are per cubic yard): 660 lbs Type III cement 1,100 lbs sand 1,900 lbs-3/4 in. top size coarse aggregate with average Mohs hardness > 6.0 26 oz-water reducer 292 lbs water Required properties: Required strength at time of test = 3500 to 5900 psi Required test age = 1 day with heat cure = 2 days with ambient temperature cure Measured properties in this test series: Age at test = 2 days Slump: 1 3/4 in. 56 Average Bond Stress over Transfer Length (psi) Strand Sample ID Beam End After Release 28 Days 6 Months 18 Months S-S -- -- -- -- S-N -- -- -- -- N-S 401 322 310 311 102-A-0.5-8A N-N 334 265 250 273 S-S 667 442 404 370 S-N 400 302 278 271 N-S 538 442 389 333 103-B-0.5-5A N-N 621 466 431 324 S-S 375 318 304 277 S-N 386 335 312 270 N-S 430 374 361 291 151-Z-0.5-5C N-N 410 358 334 302 Table B-7. Average bond stress over transfer length from Round 2. Average Bond Stress over Transfer Length (psi) Strand Source ID After Cut or Release 28 Days 6 Months 18-22 Months 102 363 323 316 314 103 531 427 397 351 151 375 334 323 286 Round 1 measurements made at released ends (S-S) or saw-cut ends (S-N or N-S) were ignored in computation of averages. Table B-8. Average bond stress by strand source as a function of time (Rounds 1 and 2).

Compressive strength at 1 day = 5300 psi Compressive strength at time of testing (2 days): 5300 psi Modulus of elasticity at 2 days: 3680 ksi (89% of ACI- 318 8.5.1) Splitting tensile strength at 2 days: 487 psi (6.7 × ) Test parameters: Number of strands per block: 6 from each source Block dimensions: 2 ft × 2 ft × 32 in. Strand Embedment: 18 in. (top 2 in. and bottom 4 in. debonded in block, 24 in. − 2 in. − 4 in. = 18 in.) Load rate: 20 kips/min Items to record during test: Time, date, load, ram travel, end slip, first movement as observed from live end. Monitor all items continuously during test up to maxi- mum load. ′fc 0 5. A photo of the LBPT setup is presented in Figure B-21. The concrete pull-out specimens were cast in the standard (upright) position on blocking to allow the center wires of the strands to pass through the bottom of the form (see Figure B-22 through Figure B-24). The center wire (king wire) of the strand was extended through the bottom of the specimen to be connected to a linear variable displacement transducer (LVDT) so that free-end slip could be measured during test- ing (see Figure B-25). To test, the block was rotated to the horizontal position, so that the LVDT that measures strand slip could be installed (see Figure B-21). The load apparatus was mounted to the side surface of the block. The particular load frame used at KSU employs a yoke so that the strand is being pulled about 8 in. from the surface of the concrete (see Figure B-26). The strand was loaded at a rate of 20 kips/min. The time, load, ram movement, and strand end slip were all recorded continuously during the test (see Figure B-27). The test was terminated just after the test sample reached the maximum load. The average bond stress at 0.1-in. slip for each source tested (calculated based on the nominal surface area of the embedded section of the strands) from the concrete pull-out testing is given in Table B-10. Mortar Pull-Out Test In the mortar pull-out test, a prestressing strand is embed- ded in 5-in. diameter by 18-in. long steel cylinders filled with 57 250 300 350 400 450 500 550 After Cut or Release 28 Days 6 Months 18-22 Months Time of Measurement A ve ra ge B on d St re ss (p si) 102 103 151 Figure B-16. Average bond stress over transfer length by strand source as a function of time (Rounds 1 and 2). Coefficient of Variation for Average Bond Stresses over Transfer Length (psi) Strand Source ID After Cut or Release 28 Days 6 Months 18-22 Months 102 8% 13% 15% 10% 103 19% 15% 16% 15% 151 12% 8% 7% 7% Round 1 measurements made at released ends (S-S) or saw-cut ends (S-N or N-S) were ignored in computation of coefficients of variation. Table B-9. Coefficient of variation for average bond stress by strand source (Rounds 1 and 2).

58 (a) (b) 0.08 0.10 0.12 0.14 0.16 0.18 0.20 0.22 After Cut or Release 28 Days 6 Months 18-22 Months Time of Measurement En d Sl ip (in .) 102 103 151 0.08 0.10 0.12 0.14 0.16 0.18 0.20 0.22 After Cut or Release 28 Days 6 Months 18-22 Months Time of Measurement En d Sl ip (in .) 102 103 151 Figure B-17. End slip as a function of time (Rounds 1 and 2) measured with a scale.

59 R2 = 0.3322 R2 = 0.9024 R2 = 0.7033 200.0 250.0 300.0 350.0 400.0 450.0 500.0 550.0 4.0 5.0 6.0 7.0 8.0 9.0 10.0 11.0 12.0 1 / End Slip (in.-1) A ve ra ge B on d St re ss o ve r T ra ns fe r L en gt h (p si) Measure w/ scale Measure w/ calipers Measure w/ photos Figure B-18. Correlation between average bond stress over the transfer length and the inverse of end slip as measured with three different methods. Figure B-19. Pull-out test in concrete specimen dimensions and details.

mortar (Portland cement, sand, and water). The top 2 in. of the embedded portion of the strand is debonded, leaving a 16-in. bond length. Six cylinders are tested for each source of strand. This test is a slight modification to the NASPA test re- ported by Russell and Paulsgrove (1999). The NASPA test uses displacement-rate control (the rate of movement of the hy- draulic ram is specified to be 0.1 in./min in the NASPA test) while the test described herein uses load-rate control with a load rate of 5 kips/min, which was found to be similar to the rates used under displacement-rate control (according to a conversation held with Russell). Also, the test described herein continues up to maximum load while the NASPA test stops at an end slip of 0.1 in. The mortar pull-out test was conducted with load rate con- trol instead of displacement-rate control because it is believed that the load rate is an important test parameter that will influence the test results. For instance, in 2002, tests were 60 Figure B-20. Test setup for pull-out test in concrete.

performed at KSU using displacement-rate control conform- ing exactly to the NASPA procedure at that time. The strands tested were from the same reel as strands tested at the Uni- versity of Oklahoma (OU). However, the pull-out forces measured at KSU were much higher than measured at OU. Part of the reason for this difference is because OU used a flexible load frame (see Figure B-28), while KSU used a stiff load frame (see Figure B-29). Consequently, while both tests were conducted at similar displacement rates, the rate of load (about 7 to 11 kips/min) at KSU was much higher than at OU (reported to be about 5 kips/min). It is a well known phenomenon that higher load rates can result in higher load values in structural materials testing. For a test to be univer- sally applicable, the test should not be load frame dependant. Accordingly, the test was changed from a displacement-rate controlled test to a load-rate controlled test. In a load- rate controlled test, the design of the test frame is unrelated to the test result. The load rate selected was similar to the rate reportedly achieved at OU, 5 kips/min. (At a presentation to the PCI Prestressing Steel Committee in April 2004, Don Pellow presented some additional information from Russell, indicating that the load rate achieved at OU was actually closer to 7 kips/min.) The specimen configuration is shown in Figure B-30, and the test frame is shown in Figure B-31. 61 Figure B-21. Test set-up. Concrete block is turned on its side, loading apparatus placed at one side (live end) and measurement of end slip instrumentation mounted at the other side (dead end). Figure B-22. View of large block forms during concrete pour. The student is internally vibrating the concrete. Figure B-23. View of inside of form. Figure B-24. Large blocks immediately after casting. Concrete is cured at lab room temperature.

The mortar mix and other parameters used in the test are as follows: Mortar mix: 1 part Type III cement by weight 2 parts ASTM C33 sand by weight 0.45 parts water by weight Required properties: Required strength at time of test = 3500 to 5000 psi Age at test: 22 to 26 h Measured properties: 1 day compressive strength: 3700 psi +/− 200 psi Test parameters: Number of strands per cylinder = 1 Number of cylinders per test = 6 Cylinder dimensions: steel tube 5-in. diameter by 1/8-in. wall thickness × 18-in. long with welded steel base plate with center hole to pass strand through Strand embedment: 16 in. (top 2 in. of specimen unbonded) 62 Figure B-25. Close-up view of LVDT used to measure end slip. Note single king wire of strand extends through the concrete. Figure B-26. Loaded end (live end) of strand. Figure B-27. Data acquisition system. Average Pull-Out Stress at 0.1-in. Slip (psi) Strand Source ID Concrete Portland Cement Mortar Gypsum Plaster Mortar Historic Strand A KSU-F 241 -- -- KSU-H 209 -- -- SC-F 223 -- -- SC-H 472 -- -- SC-IS 682 -- -- 101 241 -- -- Recently Manufactured Strand 102 441 315 588 103 944 397 621 151 541 273 619 153 142/406B -- -- OSU Strand 349 -- 156 -- 548 623 697 -- 606 -- 717 -- 206 -- 478 * -- 409 -- 960 * -- 409 -- * Samples designated 478 and 960 were from same source. A All Historic Strand concrete pull-out values represent average pull-out stress at first slip rather than at 0.1-in. slip. B Sample from India was tested at Stresscon and KSU and found to be highly variable. Table B-10. Pull-out test results from concrete, Portland cement mortar, and gypsum plaster mortar.

Load rate: 5 kip/min (note: the NASPA test uses a live end displacement rate of 0.1 in./min) Items to record: Load, ram travel, end slip, first visual movement. Monitor continuously during test. Measure mortar cube strength at beginning and end of test period. Photographs of the mortar pull-out test specimens and test- ing procedure are given in Figure B-32 through Figure B-38. The strand displacement rates achieved during the critical portion of the mortar pull-out tests (around the 0.1-in. end slip) are close to 0.1 in./min (ipm), which is the rate specified by the NASPA test, as presented in Figure B-39. The average bond stresses (calculated based on the nomi- nal surface area of the embedded section of the strands) from the Portland cement mortar pull-out testing are given in Table B-10. Hydrocal Pull-Out Test A range of Hydrocal-based mixtures were evaluated as possible surrogate materials for use in a pull-out test. 63 Figure B-28. NASPA test specimen at OU. Note flexible frame. Figure B-29. NASPA test specimen at KSU. Note stiff frame.

The final mixture used in this testing contained Hydrocal White (a material similar to plaster of paris made by United States Gypsum), Ottawa graded sand (ASTM C778), cal- cium hydroxide flakes, USG retarder for lime-based plas- ters, and water. This formulation of Hydrocal is almost pure plaster of paris and was chosen because it is produced at only one manufacturing facility from consistent raw materials. Also, like cement, it is a calcium compound (plaster of paris is hemihydrated calcium sulfate). Calcium hydroxide flakes were added to simulate the alkalinity of concrete, and the Hydrocal was combined with sand to limit the heat production generated during the rapid plaster hydration. Aside from water and sand contents, the strength of this Hydrocal mixture is influenced by the mixing procedure (longer, more vigorous mixing results in greater strength) and moisture content after curing (specimens that were dried resulted in higher strengths). Two-inch cubes could be pro- duced that demonstrated compressive strengths of more than 8500 psi at 48 h, if oven-dried for 24 h after demolding. How- ever, that same mixture produced strengths of only 3900 psi if kept moist which is the condition of the Hydrocal mortar within the steel pipe. Because it would take over 7 days to fully dry the plaster, the specimens were tested in a wet condition. In this test, a prestressing strand is embedded 12 in. in a 3-in. diameter steel cylinder filled with mortar (gypsum/lime plaster, sand, and water). Six cylinders are tested for each source of strand. The test is conducted similarly to the mortar pullout test. The layout of the test is presented in Figure B-40, and the test frame is shown in Figure B-41. The specimen materials, configuration, and testing procedures are shown in Figure B-42 through Figure B-50. The Hydrocal mix and other parameters used in the test are as follows: Hydrocal mortar mix: 1 part Hydrocal White by weight 0.75 parts Ottawa graded sand (ASTM C778) by weight 0.05 parts Ca(OH)2 flakes by weight 0.0002 parts USG retarder for lime-based plasters 0.35 parts water by weight 64 Figure B-30. Pull-out test in mortar specimen.

Required properties: Required strength at time of test = 3000 to 4000 psi Age at test: 22 to 26 h Measured properties: 1 day compressive strength: 3540 +/− 300 psi Test parameters: Number of strands per cylinder = 1 Number of cylinders per test = 6 Cylinder dimensions: standard steel pipe 3-in. diameter by 12-in. long. Welded steel base plate with center hole. Strand embedment: 12 in. Load rate: 5 kips/min Items to record: Load, ram travel, end slip. Monitor con- tinuously during test. Measure mortar cube strength at beginning and end of test period. The average bond stresses (calculated based on the nom- inal surface areas of the embedded sections of the strands) from the Hydrocal mortar pull-out testing are given in Table B-10. Pull-Out Test Results The tests performed in the Screening Round resulted in six data sets for each of the three recently manufactured strand sources and for each of the three pullout test meth- ods for a total of 54 data sets. In each data set is recorded time, load, ram travel, and end slip. For the concrete pull- out tests, there is also trigger data corresponding to the point where slip at the loaded end was first noticed. This “observed first slip” is important to maintain because it re- lates back to historic pullout data recorded by Logan (1997) and others. To enable comparisons of the shape of the stress-slip curves for each strand source, average curves were calculated using a computer program written in Visual Basic by WJE. The data logging equipment used during testing records data at set 65 Figure B-31. Test setup for pull-out test in mortar. Figure B-32. Mortar cylinder.

intervals of time. This results in slightly different load incre- ments for each data set. Accordingly, it is necessary to generate equivalent load increments in each data set so that average slips can be calculated for each increment of load. A load in- crement of 100 lbs was chosen. The program interpolates the end slips, average, and standard deviation at each increment of load. The average bond stress is plotted against the end slip in the figures shown below with the standard deviation for end slip at that load increment included in a horizontal error bar on each side of each data point. The averaged bond stress versus slip curves appear in Fig- ure B-51 to Figure B-53 for all three pull-out test methods. The previously described NASPA test uses a pull-out crite- ria of 0.1-in. end slip as a way to compare different sources of strand. Based on these test results, this appears to be a good criterion for evaluating the concrete pull-out test as well. This target value replaces the concept of first observable slip in the old Moustafa test because the first observed slip is subject to greater operator error. The “first observable slip” and “0.1-in. slip” are measured in different ways, yet were found to be close in value to one another. The bond stresses measured at the first observed slip and the 0.1-in. end slip measured in a range of concrete 66 Figure B-33. Top view of empty mortar cylinder. Figure B-34. Mortar in mixer. Figure B-35. Mortar cubes. Figure B-36. Vibrating mortar in specimen.

pull-out tests are plotted together in Figure B-54. It comes as no surprise that the first observed slip at the live end of the strand occurs at varying values of measured end slip at the dead end of the strand. However, this plot, which also in- cludes a line of equality, shows that these values are generally similar. This is to be expected since the stress at first observed slip and at 0.1-in end slip both occur early in the pull-out test. This similarity is significant because the stress at 0.1-in end slip was not recorded during the historic concrete pull-out tests conducted on the many historic samples that were in- cluded in the screening evaluation of the chemical and surface test methods reported in the next section, Surface and Chem- ical Test Methods and Results. Therefore, in discussing the correlation between bond performance and the chemical test 67 Figure B-37. Specimen in test frame. Figure B-38. Specimens ready to test. 0 1 2 3 4 5 6 7 0 50 100 150 200 250 300 Time (sec) ra m tr av el (in .) 103B 151Z 500 lbs 800 lbs 0.1 " slip 0.1 " slipRate = 0.09 ipm Rate = 0.12 ipm Figure B-39. Ram travel versus time.

results, both types of bond characterization are used together. The average bond stresses, calculated based on the nominal surface area of the embedded section of the strands, for con- crete pull-out tests conducted by Logan at Stresscon are given in Table B-10. Mortar Pull-Out Testing for Correlation Round of Evaluation As mentioned, the sample sources used for the Correlation Round of testing were selected by Bruce Russell of OSU. Mortar pull-out data were provided for each of the sources (see Table B-2), and the testing protocol is as reported by Chandran (2006). The reported NASPA pull-out forces rep- resent the average load at 0.1-in. slip for multiple (5 to 12) specimens, all tested on the same day with the same batch of mortar. Per the protocol outlined in the thesis, the mortar pull-out force was measured on strands embedded in 5-in. diameter by 18-in. long cylinders (with 16 in. of strand in di- rect contact with mortar). These mortar pull-out tests were conducted under displacement-rate control, with an addi- tional criterion for load rate. This is in contrast to mortar pull-out tests conducted by this project, which were per- formed under load-rate control. However, the displacement rate for this project’s tests ranged from 0.09 to 0.12 in./min, which is close to the rate of 0.1 in./min specified by the NASPA method. For comparison with mortar pull-out test results for the Screening Round of testing, the loads at 0.1-in. slip provided from OSU were converted to average bond stresses at 0.1-in. slip. This was done by dividing the given load by the surface area of the strand in contact with mortar (33.51 in.2 for 0.5-in. diameter strand). These bond stresses are given in Table B-10. Standard deviations for the OSU bond stresses were not provided along with the data, but were computed based on the full set of data presented in Chandran (2006) for 68 Figure B-40. Pull-out test for the Hydrocal mortar specimen.

all but one source (Source 717). The standard deviation was computed considering all specimens’ load data rather than breaking these data into subsets based on batch. Since indi- vidual test data were not available for Source 717, the standard deviation plotted with this source was the standard deviation for Batch 8N, which included 12 specimens. Discussion of Mortar and Large Concrete Block Pull-Out Testing The relationship between pull-out strength in concrete and pull-out strength in mortar is not well established. However, there is at least one data set for comparison. Pull-out strengths in mortar were measured by Russell in the Round II NASPA tests (Russell and Paulsgrove 1999). These test re- sults were found to correlate quite well with the first observed slip in the large concrete block tests performed on the same strand at Stresscon (unpublished data collected previously by the research team) as shown in Figure B-55. The data suggest that the first observed slip threshold for the large concrete block pull-out test (LBPT) should be 1.5 times that of the NASPA test. Thus, the NASPA test criteria for 1/2-in. strand of 10,500 lbs would correspond to a first observed slip of 15.8 kips in the LBPT. This value is close to the 16-kip pull-out threshold recommended by Logan. Surface and Chemical Test Methods and Results In this section, the experimental study conducted to evaluate the surface and chemical test methods is presented. The test procedure and results are covered separately for each method. The number of samples of each strand source tested by each 69 Figure B-41. Test setup for pull-out test in Hydrocal mortar. Figure B-42. Empty cylinders and mix ingredients. Figure B-43. End plates welded to pipe bases.

surface and chemical test method is shown in Table B-4 and B-5. Contact Angle Measurement The contact angle is a measure of surface tension (wet- ability). It was anticipated that the presence of drawing lubricants would affect this property. Measurements were taken with the strand: (1) in an as-received condition, (2) after immersing the strand sample in a saturated calcium hydrox- ide [Ca(OH)2] solution, and (3) after an ignition process. The calcium hydroxide exposure (also called a lime dip) will convert sodium soaps (e.g., sodium stearates) to insoluble calcium salts. For example, water-soluble sodium stearate (a soap or wetting agent) is converted to a film of insoluble calcium stearate (a wax-like, water-repellent that increases the surface-energy of the strand). This conversion reaction was chosen to simulate the reaction of concrete with surface residues of soaps and is intended to produce a condition where the effect of similar calcium stearate compounds on the contact angle are compared, even if the original residue did not result from a calcium stearate-based lubricant. The ignition process was performed on samples to volatilize organic compounds expected to be present in the drawing lubricants. The contact angle is measured on the projected shadow of a small drop of distilled water that has been applied to the strand surface with a syringe as pictured in Figure B-56. 70 Figure B-44. Mixture in bucket and mixing attachment to drill. Figure B-45. Filling the cylinders. Figure B-46. Vibrating the Hydrocal mixture.

The process of making a measurement with this method is described as follows: • The strand sample was laid on the stage with the section of wire to be tested facing upward. • The projection lens was adjusted so that silhouettes of both the wire surface and the syringe needle were in focus. • The helical wire surface was adjusted to achieve a level silhouette on the projection screen under the syringe needle that was adjusted so that it was directly over the high point of the wire surface. • A distilled water droplet, with a diameter of seven units on the scale imprinted on the projection screen, was formed using a microdroplet syringe. • The cylinder stage was raised until the wire surface touched the water droplet and then lowered until the water droplet released from the syringe needle. • The cylinder stage and projection screen were adjusted to align the left-hand side of water droplet silhouette with the origin of the scale on screen. • The dial of the protractor component of the instrument was adjusted so that the indicator line intersected the apex of the water droplet silhouette. • The contact angle, which is twice the angle measured from the apex of the drop, was read from the protractor scale (the units are already doubled on the projection screen scale shown in Figure B-56) and recorded. Six individual contact angle readings were taken per strand sample in each of the three conditions that were tested. Read- ings of a strand sample (one reading per outer wire) were taken in its as-received condition. Readings of a strand sam- ple were also taken after immersing the strand sample in a sat- urated calcium hydroxide [Ca(OH)2] solution to stimulate the environment surrounding the strand when it is in con- crete. The immersion time in the calcium hydroxide solution was 10 min and, before testing, the strand was rinsed with water and dried in the following manner: (1) immersion in de-ionized water for 5 min, (2) drying by allowing vertical strand to drip while exposing to a hot air stream from a heat gun, and (3) setting the strand on a clean surface and cooling to room temperature. Finally, readings were taken after an ig- nition process similar to that discussed for the Weight Loss on Ignition (LOI) test: Pieces of strand approximately 9-in. long were dried for 4 h at 110°C, allowed to cool in a desicca- tor for at least 12 h, ignited for 30 min at 415°C, allowed to cool in a desiccator for at least 12 h and then tested. The contact angle values measured are reported in Table B-11. This gives the average contact angle for each source. When possible, contact angle readings were performed on three samples of a strand source. Contact angle testing after the calcium hydroxide exposure was also included in the Correlation Round. The ignition process to remove organic residue was only used in the 71 Figure B-47. Making mortar cubes. Figure B-48. Completed specimens ready for test.

72 Figure B-49. Specimen in test frame. Figure B-50. Close up of dead end showing LVDT attachment. 0 200 400 600 800 1000 1200 -0.5-0.45-0.4-0.35-0.3-0.25-0.2-0.15-0.1-0.050 End slip (in.) B on d st re ss (p si) 102A 103B 151Z Legend observation of first slip Figure B-51. Bond stress versus end slip from large concrete block pull-out test.

Correlation Round as a frame of reference representing strand without any residue. Examination under Ultraviolet Light The examination under ultraviolet (UV) light was conducted using a range of light sources having different wavelengths. The most promising had a 366-nm wavelength. The exami- nations revealed detectable fluorescence in three of the strand samples examined. In the first and third, SC-F and 102, the fluorescence was found to be diffuse over the surface of the strand and appeared to coincide with rust product. In the second, SC-IS, the florescence was confined to the interstitial valleys between the wires that make up the strand and appeared as localized speckles (Figure B-57). Observations of the fluo- rescence are given in Table B-12. This test method was not included in the Correlation Round. Testing pH Testing of the pH of the surface was attempted with each of the strand sources to see if measurement of the alkalinity of a solution generated by placing drops of water on the residue could be linked to bond. Testing of the pH of the 73 0 100 200 300 400 500 600 -1.2-1.05-0.9-0.75-0.6-0.45-0.3-0.150 End Slip (in.) B on d st re ss (p si) 102A 103B 151Z 0 100 200 300 400 500 600 700 -1.5-1.35-1.2-1.05-0.9-0.75-0.6-0.45-0.3-0.150 End Slip (in.) B on d St re ss (p si) 102A 103B 151Z Figure B-52. Bond stress versus end slip from mortar pull-out test. Figure B-53. Bond stress versus end slip from Hydrocal pull-out test.

74 Comparison of bond stress at 1st obs. slip vs. 0.1-in. end slip 800 700 600 500 B on d St re ss a t 1 st o bs er ve d sl ip (p si) Bond stress at 0.1-in. end slip (psi) 400 300 200 100 0 0 100 200 300 400 500 600 700 800 900 1000 0 0 5 10 15 20 St re ss co n fir st o bs s lip (k ips ) 25 30 35 y = 1.4681x R2 = 0.9023 5 10 OU 0.1-in. slip force (kips) 15 20 25 Figure B-54. Correlation between bond stress at first slip and bond stress at 0.1-in. slip. Figure B-55. Comparison of pull-out capacities in mortar versus concrete.

surface was conducted using a pH meter, indicator papers, and indicator solutions. When using universal indicator paper (Tridicator by Fil- chem, Inc.), droplets of de-ionized water were placed on the strand surface and allowed to dissolve water-soluble surface residue for 30 sec before a piece of indicator paper was touched to the surface. Droplets of water were applied to the outer peak surface of an individual wire in two loca- tions and to the interstitial valley in two locations on each piece of strand tested. A photo of the droplets and the universal indicator paper, after the color change, is given in Figure B-58. Testing also included experimentation with indicator solutions that change color in a pH range deemed suitable for these tests (7 to 10). These indicators (and the pH range over which they change color) included phenolphthalein (8.2 to 10), curcumin (7.4 to 8.6), thymol blue (8 to 9.6), and cresol red (7 to 8.8). These solutions were found to be inappropriate for testing strand samples since the low surface tension of the solutions caused them to slide over the strand surface. A photo showing the indicator solution colors and the type of color change observed is given in Figure B-59. Additional pH testing was performed to expand on the res- olution of the measurement technique, since this property showed some promise as an identifier of poor bond. This ad- ditional testing was done using a hand-held pH meter (Exstik manufactured by Extech) and using specific pH indicator papers (manufactured by Macherey-Nagel), both of which measure pH at a greater resolution than was done with the universal indicator paper. The pH of the aqueous solution produced by applying a de-ionized water droplet to the crevice between two adjacent wires of a strand sample and allowing that droplet to dissolve surface residues for a set period of time was measured. The Exstik pH meter is pictured in Figure B-60. This instru- ment was calibrated using pH 7 and pH 10 buffer solutions before each use and reads pH to two decimal places. This meter required that 2 drops of water be placed next to each other on a sample crevice. The reading was taken by placing the meter against the elongated water drop, causing the probe surface to wet sufficiently to wet both ports on the probe tip for each reading. The meter was held against the surface for 10 to 20 sec 75 (a) (b) Figure B-56. Contact angle meter (a) and close-up of drop projection (b). Average Contact Angle (°) Strand Source ID As-Received After Ca(OH)2 After Ignition Historic Strand KSU-F 80 106 -- KSU-H 91 103 -- SC-F -- -- -- SC-H 87 80 -- SC-IS 83 79 -- 101 55 94 -- Recently Manufactured Strand 102 60 87 -- 103 71 79 -- 151 107 98 -- 153 -- 75 -- OSU Strand 349 -- 87 11 548 -- 79 15 697 -- 68 8 717 -- 94 12 478 * -- 73 7 960 * -- 76 9 * Samples designated 478 and 960 were actually from same source. Table B-11. Contact angle.

to allow the meter to reach an equilibrium reading. Sometimes the meter readings seemed fairly stable, but other readings tended to drift significantly (usually upward) the longer the instrument was held in place. All four types of high-resolution pH indicator paper from Macherey-Nagel (Part Nos. 90211, 90305, 92160, and 92170) shown in Figure B-61 were tested to measure the pH of the strand surface. The pH indicator papers that appeared most suitable for this application based on the ease of use and the applicable range were the Duotest and pH-Fix 7.5 to 9.5 indi- cators, and the data presented were collected using these indicators. These read with 0.3 and 0.2 pH unit resolutions, respectively. To measure pH with pH indicator papers, droplets of de-ionized water were placed in crevices and allowed to dwell for 30 sec. The paper edge was then dipped into the water droplet until it made contact with the bottom of the crevice, and it was withdrawn. The Duotest paper was read immediately, but the pH-Fix paper was allowed to stand for 1 min for complete color development before reading. Ten pH measurements were conducted using each of the three high-resolution methods alternately along the length of two strand samples. Each measurement site was separated by approximately 4 in. so the individual test locations for each method were spaced about 1-ft apart. Droplets of water were allowed to stand for 5 min before testing. The averages of the pH readings for each sample are given in Table B-13. The pH of the de-ionized water, according to the pH meter was between 5.5 and 6.2. The average of the pH measured on samples from the same source appears repeatable for Sources 102 and 151, with both sources resulting in a high pH. However, the pH on one of the samples from Source 103 was high (9.02), while the other sample was low (7.62). Nevertheless, 76 (a) (b) Figure B-57. Surfaces of strand examined under ultraviolet light. Sample SC-F is on the left (a), and sample SC-IS is on the right (b). Strand Source ID Fluorescence Historic Strand KSU-F None KSU-H None SC-F Inconsistent, dull glow, concentrated on one side of strand (rust) SC-H None SC-IS Speckles in crevices 101 None Recently Manufactured Strand 102 Dull glow in crevices (rust) 103 None 151 None Table B-12. UV fluorescence. Figure B-58. Surface of strand during pH testing (the bottom of the universal indicator paper, as shown, was touched to the water and has changed color).

there is general agreement between the types of measurement, albeit with an apparent offset separating the readings taken with the different methods. Despite the higher resolution of measurement, the pH meter resulted in wider scatter and had a greater standard deviation over the 10 individual readings. The Duotest paper is the easiest to read and resulted in little scatter. The results from each of the three methods are given in Table B-14. The Duotest procedure was included in the Correlation Round. Loss on Ignition The weight LOI represents the weight of compounds that can be volatilized or burned off the strand surface at high tem- perature. This property was measured with the expectation that the weight lost may consist mainly of organic residues such as drawing lubricants. The LOI was measured using the following procedure: Pieces of strand approximately 9-in. long were dried for 4 h at 110°C, allowed to cool in a desiccator for at least 12 h, weighed three times to 0.1-mg precision then ignited for 30 min at 415°C, allowed to cool in a desiccator for at least 12 h, and then weighed again three times on the same balance. The LOI reported is the change in average weight divided by the surface area of the strand. The average LOI for the strands tested are given in Table B-15. This test method was included in the Correlation Round. 77 Figure B-59. Indicator solutions (from left to right, Thymol Blue, Cresol Red, Curcumin, Phenolphthalein) and strips of indicator-treated paper that had been applied to strand surface. Figure B-60. Exstik hand-held pH meter. Figure B-61. High-resolution pH indicator papers. The Duotest (left) and pH-Fix 7.5–9.5 (3rd from left) were preferred. Table B-13. pH measurements with universal indicator. Average pH reading Strand Source ID Interstitial Valley Peak Historic Strand KSU-F 9.0 8.0 KSU-H 8.8 7.3 SC-H 7.0 7.0 SC-IS 7.0 7.0 Recently Manufactured Strand 102 9.0 8.2 103 7.0 7.0 151 7.7 7.0

Loss in Hot Alkali Bath The weight loss after hot alkali bath (LAB) represents the weight of compounds that can be washed off the strand surface in a hot sodium hydroxide solution. As with the LOI test, this property was measured with the expecta- tion that the weight lost may consist mainly of drawing lubricants. The LAB was measured using the following procedures: Pieces of strand approximately 9-in. long were dried for 4 h at 110°C, allowed to cool in a desiccator for at least 12 h, weighed three times to 0.1-mg precision then cleaned accord- ing to one of the procedures detailed below, dried in an oven (110°C) for 4 h, allowed to cool in a desiccator for at least 12 h, and then weighed again three times on the same scale. The LAB reported is the change in average weight divided by the surface area of the strand. The first cleaning procedure (Method 1) used a hot sodium hydroxide solution in an ultrasonic cleaning tank. The sodium hydroxide solution was prepared by slowly adding 500 gm of NaOH pellets to 1,500 mL of de-ionized water (originally about 22°C). (The dissolution of NaOH pellets in water was an exothermic reaction, causing the solution temperature to rise significantly. The temperature of the solution after about 15 min of stirring was 77°C.) The stainless steel tub of the ultrasonic cleaning tank was conditioned by rinsing twice with 80°C water. The hot rinse water was poured out before adding NaOH solution that had been heated to 80°C on a hot plate. The strand samples were disassembled into the indi- vidual wires, and the king (center) wire was not included in the testing. Two sets of wires were placed in the tank (each set resting on one half of the tank bottom), and the ultrasonic cleaner was run for 10 min. Each wire was removed individ- ually and thoroughly wiped with a paper towel. After the first wiping procedure, each wire was rinsed with de-ionized water and thoroughly wiped again with a paper towel. A second round of tests was initiated to see if the original pro- cedure, which removed a large amount of material, was overly aggressive. In the modified tests (Method 2), the strands were tested intact and either the wiping process or the ultrasonic ag- itation was eliminated. Neither of these modifications appeared to have a significant effect on the results of the test. The average LAB values for the strands tested are given in Table B-16. This test method was not included in the Correlation Round. 78 Table B-14. pH measurements with pH meter and with Duotest and pH-fix indicators. Strand Source ID Average LOI (mg/cm2) Historic Strand KSU-F 0.102 KSU-H 0.080 SC-F -- SC-H 0.036 SC-IS 0.012 101-A-0.5 0.017 Recently Manufactured Strand 102 0.051 103 -0.021 151 0.002 153 0.375 OSU Strand 349 0.139 548 0.059 697 0.036 717 0.086 478 * 0.041 960 * 0.045 * Samples designated 478 and 960 were actually from same source. Table B-15. Average weight loss on ignition (LOI). Strand Source ID Average LAB - Original Procedure (mg/cm3) Average LAB - Modified Procedure (mg/cm3) Historic Strand KSU-F 0.288 - KSF-H 0.242 - SC-H 0.493 - SC-IS 0.360 - Recently Manufactured Strand 101* 0.195 0.197 A 102* 0.368 0.391 B 103* 0.416 0.375 B 151* 0.185 0.167 B A - No wipe B - No ultrasonic agitation * - Strand tested intact Table B-16. Average weight loss after alkali bath (LAB). Average pH Strand Source ID pH Meter Duotest pH-Fix 7.5-9.5 Recently Manufactured Strand 102 9.36 8.7 9.0 103 8.32 7.4 8.0 151 9.13 8.3 8.4 153 -- 7.3 -- OSU Strand 349 -- 7.2 -- 548 -- 7.1 -- 697 -- 7.0 -- 717 -- 7.5 -- 478 * -- 7.1 -- 960 * -- 7.2 -- * Samples designated 478 and 960 were actually from same source.

Change in Corrosion Potential Past studies of the corrosion resistance of prestressing strand in concrete have suggested that strand with a coating of residue does not corrode as readily as a clean strand. To assess the potential for corrosion, the strand samples were placed in a solution of de-ionized water, and the corrosion potential measured with a reference cell (saturated calomel reference electrode) was monitored versus time. This corrosion poten- tial is determined by the amount of ferrous ions in solution surrounding the sample, and a greater drop in this potential is indicative of a greater tendency to corrode. Four strand samples were tested at a time using a multi- plexer that allowed readings of the potential to be made using a single reference cell. Figure B-62 is a photo of the test setup. While data were collected continuously for 6 h (Figure B-63), an exposure time of 6 h was arbitrarily chosen as measure- ment cutoff points. Measurements were taken with the strand in as-received condition, after immersing the strand sample in a saturated calcium hydroxide [Ca(OH)2] solution, and after an ignition process. The immersion time in the calcium hydroxide solution was 10 min and, before testing, the strand was rinsed with water and dried in the following manner: (1) immersion in de-ionized water for 5 min, (2) drying by allowing vertical strand to drip while exposing to a hot air stream from a heat gun, and (3) setting the strand on a clean surface and cooling to room temperature. For the ignition samples, readings were taken after an ignition process similar to that discussed for the LOI test: For this process, pieces of strand approximately 9-in. long were dried for 4 h at 110°C, allowed to cool in a desicca- tor for at least 12 h, ignited for 30 min at 415°C, allowed to cool in a desiccator for at least 12 h, and then tested. The average changes in corrosion potential from the values initially measured are reported in Table B-17. This test, conducted on samples in as-received condition, was included in the Correlation Round. The ignition cleaning regime to remove organic residues was used only in the Correlation Round as a frame of reference representing strand without any residue. Surface Roughness Microscopic examinations of sectioned portions of wire taken from strand have indicated that an observable difference in the surface roughness of the good- and poor-bonding strand sources exists. Based on images captured using a scanning electron microscope, the depth of the roughened surface fea- tures is typically 3 μm (0.0001 in.) or less. Trials with a portable profilometer suitable for a QC setting (Figure B-64) were conducted to determine whether these physical measurements could accurately represent the surface roughness and to in- vestigate the correlation with bond performance. This system works by measuring the deflection of a diamond probe, with a tip radius of 2-μm (shown in Figure B-64), as it is dragged 2 mm across the surface of the sample. The surface profile can be plotted, but it is most commonly interpreted using one of a range of possible parameters cal- culated based on the profile. The three parameters deemed to be most applicable to this project are Ra, the average rough- ness; Rz, the average roughness depth; and Pc, the peak count. The definitions of the first two parameters are illustrated in Figure B-65. The peak count is defined by the instrument manufacturer as “the number of profile features whose peaks and following valleys project beyond the upper and lower cut-off thresholds respectively.” For our evaluation, the upper and lower thresholds were located a distance 0.13 μm above and below the mean depth of the profile. The Ra, Rz, and Pc were measured at least twice per wire on each of the six exterior wires on one piece of strand. The average Ra, Rz, and Pc, for each of the tested strand sources is given in Table B-18. This test method was included in the Correlation Round. 79 Figure B-62. Setup for monitoring corrosion potential versus time. Saturated calomel reference electrode is in center of figure.

current near to the corrosion potential, is inversely propor- tional to the corrosion current, which quantifies the rate at which the electrochemical corrosion reaction is occurring. This relationship is governed by the Stern-Geary equation: icorr = B/Rp, where icorr is the corrosion current density (current per unit area), and B is constant for the metal being testing. The value of B is usually assumed to be 26 mV for steel that is actively corroding, which is the assumption made here. The potential of the sample is measured relative to a refer- ence electrode, in this case a silver-silver chloride half-cell, which is also submerged in the solution. The solution used in these tests was 0.1% Cl as NaCl by weight. The corrosion rate was measured by varying the corrosion potential ±15 mV from the corrosion potential after the strand sample was 80 -0.4 -0.35 -0.3 -0.25 -0.2 -0.15 -0.1 -0.05 0 0 30 60 90 120 150 180 210 240 270 300 330 360 Time (min.) Ch an ge in C or ro si on P ot en tia l (V ) 101-A-0.6-4B 101-A-0.6-5B 102-A-0.5-1D 102-A-0.5-2E 103-B-0.5-1D 103-B-0.5-2E 151-Z-0.5-1F 151-Z-0.5-2G Figure B-63. Change in corrosion potential versus time for recently manufactured strand. Change in Corrosion Potential (V) As-Received After Ca(OH)2 After Ignition Strand Source ID 1 h 6 h 1 h 6 h 1 h 6 h Historic Strand 101 -0.189 -0.349 -- -- -- -- Recently Manufactured Strand 102 -0.16 -0.267 -0.188 -0.315 -- -- 103 -0.077 -0.172 -0.059 -0.156 -- -- 151 -0.16 -0.322 -0.212 -0.377 -- -- 153 -0.203 -0.272 -- -- -0.137 -0.268 OSU Strand 349 -0.228 -0.289 -- -- -0.074 -0.163 548 -0.064 -0.080 -- -- -0.136 -0.222 697 -0.090 -0.154 -- -- -0.142 -0.237 717 -0.170 -0.241 -- -- -0.196 -0.314 478 * -0.216 -0.272 -- -- -0.080 -0.186 960 * -0.154 -0.211 -- -- -0.064 -0.140 * Samples designated 478 and 960 were actually from same source. Table B-17. Average change in corrosion potential. Corrosion Rate To further explore the interaction between strand bond and corrosion, the instantaneous rate of corrosion of samples of strand in a salt solution was measured with a polarization resistance technique. The polarization resistance technique measures the corrosion current, which quantifies the rate at which the electrochemical corrosion reaction is occurring. This is a much faster test than the test for change in corrosion potential, but requires specialized equipment (a potentiostat). The polarization resistance technique measures the shift in potential of a metal sample from a stable corrosion potential due to an external current. The polarization resistance, Rp, that is, the ratio of the change in potential to the applied

exposed to the chloride solution for 5 min. The exposure length of the samples was 10 cm, and the submerged ends of the samples were coated with epoxy resin so that the cut sur- face of the sample would not participate in the measurement. Figure B-66 is a photo of the test setup. Measurements were taken with the strand in an as-received condition, after immersing the strand sample in a saturated calcium hydroxide [Ca(OH)2] solution and after an ignition process. The immersion time in the calcium hydroxide solu- tion was 10 min and, before testing, the strand was rinsed with water and dried in the following manner: (1) immersion in de-ionized water for 5 min, (2) drying by allowing vertical strand to drip while exposing to a hot air stream from a heat gun, and (3) setting the strand on a clean surface and cooling to room temperature. For the ignition samples, readings were taken after an ignition process similar to that discussed for the LOI test reported above. For this process, pieces of strand approximately 9-in. long were dried for 4 h at 110°C, allowed to cool in a desiccator for at least 12 h, ignited for 30 min at 415°C, allowed to cool in a desiccator for at least 12 h, and then tested. The ignition cleaning process was used in the Correlation Round as a frame of reference to represent strand without any organic residue. The average corrosion rates (the corrosion current per unit area) are given in Table B-19. The corrosion rate test was conducted on samples in the as-received condition and after calcium hydroxide exposure in the Correlation Round. Organic Residue Extraction The tests for identification and quantification of organic drawing-compound residues were based on solvent extraction procedures, together with gravimetric and FTIR analyses. Essentially, the amount of material extracted from a defined length of strand was determined by weighing the extraction residue on an analytical balance. The material(s) in the 81 2 µm probe tip (a) (b) Figure B-64. Profilometer (a) and close up of pickup probe (b). Figure B-65. Definition of surface roughness parameters (from Operating Instructions for Perthometer M4P-150). Strand Source ID Ra (µm) Rz (µm) PC (µm) Recently Manufactured Strand 102 0.66 3.85 183 103 0.40 2.53 210 151 0.54 3.13 199 153 0.41 2.45 244 OSU Strand 349 0.56 3.33 200 548 0.26 1.65 212 697 0.81 4.72 181 717 1.20 5.83 110 478 * 0.42 2.52 212 960 * 0.50 3.21 203 * Samples designated 478 and 960 were actually from same source. Table B-18. Average roughness parameters.

extraction residue were then identified by FTIR analysis of the residue. The FTIR spectrum obtained is like a fingerprint of the material. The extraction procedure used is a modification of a proce- dure found in ASTM C114 for organic materials in cement. Multiple extractions were used to differentiate between various forms of drawing-compound residue. For example, sodium stearate is soluble in warm or hot water, but calcium stearate and stearic acid are insoluble in water. The strand was first washed with warm or hot water to remove water-soluble materials, such as sodium stearate. Then, the strand was exposed to hydrochloric acid and chloroform to extract water- insoluble residues such as calcium stearate and stearic acid. In the Screening Round of evaluation, the extraction method was done with two successive procedures over a 12-in.-long strand segment: (1) a water-wash procedure, and (2) an acid/ solvent-wash procedure. During the water-wash procedure, the strand was washed with either warm (initially 120°F) or hot (initially 212°F) water. The wash solution was acidified, and then the chloroform-soluble organic components were extracted from this acidified solution with chloroform. In the subsequent acid/solvent-wash procedure, the same strand seg- ment was washed with a hydrochloric acid solution (10% HCl), followed by a chloroform-wash. The organic compounds in the aqueous acid wash were then extracted with the chloro- form wash solvent. After both of these wash procedures, the chloroform and acidified solutions were separated in a sepa- ratory funnel. (The acidified solution was saved for the ele- mental analysis discussed below.) The chloroform solvent with dissolved organic components was evaporated, and the residue after evaporation was weighed. The weight per unit area was then calculated. The average residue weights generated for three samples from each source by the water-wash procedure and the acid/ solvent-wash procedures are given in Table B-20, along with the “total,” which is the sum of the two concentrations. The extraction was performed using water at a higher temperature in the second half of the test program since the warm-water wash removed little organic residue on any of the strand. It had been expected that the water wash would differentiate sodium stearate lubricants that are water-soluble from calcium stearate lubricants, which are not. With the exception of Source 101, minimal amounts of water-soluble residue suggest that sodium stearate had not been applied, or had been removed by rain or some other water exposure (washing process), or is less water- soluble than originally thought. The hot water was intended to dissolve more of the sodium stearate, if present. At the conclusion of the Screening Round, it was observed that the water temperature had little effect in most cases, but that the residue concentrations measured with the warm-water method seemed to generally correlate better with bond tests. Therefore, only a warm-water wash was used in the Correlation 82 Figure B-66. Setup for corrosion rate testing. Average Icorr (µA/cm2)Strand Source ID As-Received After Ca(OH)2 After Ignition Historic Strand KSU-F 32.8 -- -- KSU-H 56.1 -- -- SC-F -- -- -- SC-H 87 -- -- SC-IS 83 -- -- 101 55 -- -- Recently Manufactured Strand 102 32.7 52.8 -- 103 9.4 10.2 -- 151 23 44.0 -- 153 -- 2.9 19.3 OSU Strand 349 3.1 3.1 13.3 548 3.9 2.0 12.3 697 76.8 46.4 29.4 717 30.5 12.9 14.3 478 * 10.7 10.1 14.8 960 * 10.8 8.7 10.5 * Samples designated 478 and 960 were actually from same source. Table B-19. Average corrosion rate.

Round. In addition, to minimize the effort spent on perform- ing the time-consuming chloroform organic extraction, the wash solutions from the warm-water and acid/chloroform washes were combined, and a single separation was performed. Therefore, only one FTIR scan and residue weight determina- tion was made per piece of strand. However, this is considered essentially equivalent to the combination of sequential warm- water and acid/chloroform washes performed in the Screening Round. Since quantifying the water-soluble materials was still of interest, because it might provide insight into possible clean- ing methods, separate warm-water washes were performed on additional pieces of strand. This wash solution was acidified and saved for elemental analysis. The organic residue concen- trations from the single separation of the combined warm- water and acid/chloroform washes for the OSU strand sources is also included in Table B-20. The FTIR spectra of the residues indicate that a salt of a fatty acid, such as stearic acid, was present in the extractions. A summary of the spectra interpretations, based on multiple spectra for each source, is listed in Table B-21. Atomic Absorption and Colorimetric Analyses To identify the chemical composition of residual inor- ganic components of pretreatment chemicals and drawing compounds, chemical analyses of the acidified water-extract and acid-solvent-extract solutions, which had been separated from the chloroform, were performed. Either zinc phosphate or borax (sodium borate) is often applied to the wire before the drawing process begins to help drawing lubricants stick to the surface of the rod stock. Most common drawing lubricants are expected to include stearate salts, particularly sodium and calcium stearates. The elemental concentrations 83 Average Residue (mg/cm2) - Warm water Average Residue (mg/cm2) - Hot Water Strand Source ID Warm- Water Wash Acid/Solvent Wash Total Hot-Water Wash Acid/Solvent Wash Total Average Residue (mg/cm2) - Total A Historic Strand KSU-F 0.006 0.137 0.143 -- -- -- -- KSU-H 0.006 0.112 0.118 -- -- -- -- SC-F 0.005 0.102 0.107 -- -- -- -- SC-H 0.002 0.040 0.042 -- -- -- -- SC-IS 0.001 0.016 0.016 -- -- -- -- 101 0.006 0.05 0.055 0.033 0.034 0.067 -- Recently Manufactured Strand 102 0.005 0.064 0.069 0.005 0.062 0.067 -- 103 0.001 0.01 0.012 0.002 0.023 0.025 -- 151 0.005 0.034 0.038 0.017 0.024 0.04 -- 153 -- -- -- -- -- -- 0.186 OSU Strand 349 -- -- -- -- -- -- 0.022 548 -- -- -- -- -- -- 0.047 697 -- -- -- -- -- -- 0.019 717 -- -- -- -- -- -- 0.117 478 * -- -- -- -- -- -- 0.033 960 * -- -- -- -- -- -- 0.035 * Samples designated 478 and 960 were from same source. A All OSU strand tested for total extraction residue with a combined warm-water and acid/solvent wash. Table B-20. Organic residue extraction concentrations. Strand Source FTIR Spectra Interpretation* Historic Strand KSU-F stearic acid KSU-H stearic acid SC-F stearic acid SC-H stearic acid, acrylic SC-H stearic acid SC-IS fatty acid, trace styrene 101 stearic acid Recently Manufactured Strand 102 stearic acid 103 fatty acid 151 stearic acid 153 fatty acid OSU Strand 349 fatty acid, resin 478 stearic acid 548 fatty acid 697 fatty acid, resin 717 stearic acid 960 stearic acid *Stearate salts are converted to stearic acid by the organic residue extraction process. Table B-21. Interpretation of FTIR spectra.

of sodium, potassium, calcium, and zinc were determined by atomic absorption spectroscopy. The solutions were also scanned for detectable quantities of aluminum during the Screening Round. Colorimetric analyses of the wash solu- tions for boron and phosphate ( as total phosphate) were performed using visible light spectroscopy. Atomic Absorption Studies Dilutions were made with de-ionized water as needed to produce final solutions with concentrations in the linear ranges of the desired elements under study. Lanthanum chlo- ride, 0.1% by volume, was added at the rate of 5 mL/100 mL of solution to prevent ionization of the desired elements. Standard solutions of sodium, potassium, calcium, and zinc were made using purchased NIST traceable 1000 mg/L atomic absorption standards. These standards were diluted to pro- duce linear standards for the concentration ranges desired. The atomic absorption spectrometer was calibrated using the prepared standards. Each diluted solution was run twice with recalibration of the spectrometer for each run to determine the concentration of the selected element. The values were averaged, and the concentration of each element in milligrams per centimeter of strand was calculated. To detect the pres- ence but not exact quantities of aluminum, the diluted wash solutions were checked for the presence of aluminum without instrument calibration. The average concentrations measured with atomic absorp- tion spectroscopy taken as the average of three samples from each source are given in Table B-22 and Table B-23. The quantity of aluminum was below detectable levels. Visible Light Spectroscopy Portions of the filtered wash solutions from the water-wash and acid/chloroform-wash procedures were also analyzed for boron using sulfuric-carminic acid with visible light spec- troscopy. An aliquot of the wash solution was mixed with the sulfuric-carminic acid solution and allowed to stand for 25 min. The absorbance of the solution was read using a Hach visible light spectrometer set at 605 nm. Boron standards made from boric acid were used to prepare calibration curves. Examples of the appearance of the solutions are given in Figure B-67. The concentration of boron in milligrams per liter (mg/L) of wash solution was determined from the calibration curves obtained. The average boron concentration of three samples from each source is given in Table B-22 and Table B-23. During the Correlation Round, portions of the filtered acid extraction solutions were analyzed for phosphate using molybdic acid reaction and visible light spectroscopy. The orthophosphate in the solution reacts with the molybdate in an acidic medium, producing a phosphomolybdate complex PO43− that is blue in color. The color of the solution is proportional to the concentration of phosphate. An aliquot of the extrac- tion solution was mixed with the molybdic acid solution and allowed to stand for 2 min. The absorbance of the solution was read using a Hach visible light spectrometer set at 890 nm. Based on this value, the concentration of phosphate in the extraction solution was determined and this was converted to a strand surface concentration. The results of phosphate analyses can be found in Table B-23. Evaluation of Test Methods This evaluation was conducted in three evaluation rounds: screening, correlation, and precision. The evaluation of the Screening and Correlation Rounds are presented together, even though correlation testing was not conducted on all the test methods attempted. The precision testing is discussed separately. Variability in Test Methods The goal of this project was the identification of test methods that could accurately measure or predict the bond quality of prestressing strand in concrete. To determine whether a pro- posed method under consideration was able to achieve this goal, the results obtained with that test method were compared with other test results deemed to be a reliable measure of bond quality. It should be noted that all test methods have a certain level of uncertainty and that poor correlation will result if there is variability or inaccuracy in either the method being evaluated or the test result against which it is compared. Each data point on the plots presented in this section is the average of a series of measurements using each technique. Horizontal and vertical error bars on the plots display one standard deviation in each direction from the data point. Assuming a normal distribution, the range covered by these bars includes 68% of the measured data. Evaluation of Mechanical Test Methods Before assessing the correlation of the surface and chemical test methods to bond performance, the correlation between the mechanical test methods was evaluated. This was possible only for the recently manufactured strand, since they were the only available sources that could be obtained in sufficient length to support transfer length testing. The transfer length test, measured based on strain profiles, is considered the most accurate measure of bond performance. This is because this method most closely simulates the in-service conditions in which prestressing strand is used, evaluating the bond per- formance of strand in concrete in a stressed condition. To allow for comparisons to be made across samples with different 84

Sodium (mg/cm2) Potassium (mg/cm2) Calcium (mg/cm2) Zinc (mg/cm2) Aluminum (mg/cm2) Boron (mg/cm2)Strand Source ID Water Acid Total Water Acid Total Water Acid Total Water Acid Total Water Acid Total Water Acid Total Historic Strand KSU-F 0.227 0.033 0.260 0.061 0.014 0.075 0.005 0.073 0.078 0.002 0.004 0.006 N.D. N.D. N.D. 0.075 0.024 0.099 KSU-H 0.174 0.031 0.205 0.049 0.012 0.061 0.007 0.065 0.072 0.002 0.004 0.006 N.D. N.D. N.D. 0.061 0.024 0.085 SC-F 0.013 0.080 0.093 0.004 0.008 0.012 0.076 0.705 0.781 0.002 0.876 0.878 N.D. N.D. N.D. 0.002 0.008 0.010 SC-H 0.018 0.130 0.148 0.005 0.023 0.028 0.004 0.072 0.076 0.002 0.874 0.876 N.D. N.D. N.D. 0.004 0.013 0.017 SC-IS 0.009 0.188 0.197 0.002 0.025 0.027 0.005 0.023 0.028 0.002 1.407 1.409 N.D. N.D. N.D. 0.002 0.009 0.011 101 (warm) 0.140 0.030 0.170 0.034 0.008 0.042 0.002 0.014 0.016 0.002 0.002 0.004 N.D. N.D. N.D. 0.090 0.009 0.099 101 (hot) 0.154 0.026 0.180 0.035 0.008 0.043 0.002 0.013 0.015 0.002 0.002 0.004 N.D. N.D. N.D. 0.096 0.007 0.103 Recently Manufactured Strand 102 (warm) 0.150 0.037 0.187 0.019 0.008 0.027 0.013 0.269 0.282 0.002 0.002 0.004 N.D. N.D. N.D. 0.033 0.018 0.051 102 (hot) 0.174 0.018 0.192 0.029 0.004 0.033 0.019 0.272 0.291 0.002 0.002 0.004 N.D. N.D. N.D. 0.040 0.008 0.048 103 (warm) 0.014 0.098 0.112 0.002 0.013 0.015 0.005 0.019 0.024 0.002 0.894 0.896 N.D. N.D. N.D. 0.006 0.013 0.019 103 (hot) 0.015 0.101 0.116 0.002 0.014 0.016 0.006 0.017 0.023 0.003 0.937 0.940 N.D. N.D. N.D. 0.006 0.013 0.019 151 (warm) 0.063 0.020 0.083 0.005 0.002 0.007 0.002 0.027 0.029 0.002 0.002 0.004 N.D. N.D. N.D. 0.042 0.021 0.063 151 (hot) 0.072 0.019 0.091 0.004 0.003 0.007 0.005 0.024 0.029 0.003 0.002 0.005 N.D. N.D. N.D. 0.053 0.021 0.074 N.D. = Not detected Table B-22. Elemental analysis—average for each source—results from Screening Round testing.

strand diameters and to directly account for the stress trans- ferred to the concrete, the average bond stress over the transfer length was calculated in the manner discussed in the previous section on Mechanical Test Methods and Results. As reported, the bond stresses of the recently manufac- tured strands were measured at 0.1-in. slip during pull-out testing in concrete, mortar, and Hydrocal. This concrete bond stress is compared to the average bond stress over the transfer length for each strand source in Figure B-68. A nearly linear relationship exists between the concrete bond stress and the bond stress over transfer length and the ranking of the three sources in terms of bond quality is consistent. In Figure B-69, the mortar bond stress has a fair correlation with the mortar bond stress at 0.1-in. slip. Figure B-70 compares the Hydrocal bond stress to the bond stress over the transfer length. A weak relationship between Hydrocal pull-out stress and average bond stress over transfer length is visible, but the correlation is less clear than with the concrete or the mortar, because of the large variability in the pull-out performance. All three methods differentiate between the higher bond control strand (Source 103) and the marginal bond strands (Sources 102 and 151). The concrete pull-out test is more dis- criminating than the other two test methods since the high bond strand had a pull-out bond stress almost twice the value of the marginal bond strands. In the mortar cylinder test, the high bond strand only had 1.3 times the pull-out value of the marginal bond strand. The Hydrocal mortar pull-out test is the least discriminating, with the high bond strand bond stress of only 1.2 times the marginal bond strand’s at 0.1-in. slip. As can be seen in Figure B-53, at 0.5-in. slip, one of the marginal bond strand sources carried an even higher stress in the Hydrocal than the high bond strand, rendering this test method unworthy of further consideration. The comparatively smaller range in bond stresses meas- ured using the mortar pull-out test was expected, based on the literature. However, more discrimination was expected than was recorded in this test series. During the NASPA tests performed at KSU in 2002, the tested strands had bond stresses at 0.1-in. end slip ranging from 167 to 715 psi. The range in the most recent tests reported herein ranged from 273 to 397 psi, admittedly on a different series of strands. One primary dif- ference between the two test programs is the mortar strength. In the 2002 tests, the mortar strengths were about 5200 psi. In the current series of tests, the mortar strength was about 3700 psi. Another difference is the use of load control versus displacement-rate control. In the tests run in 2002, load rates as high as 11 kips/min were achieved for the high bond strand compared to the 5 kips/min used in the current series. No additional mechanical testing was conducted in the Correlation Round of the evaluation. Evaluation of Surface and Chemical Test Methods To compare the surface and chemical testing to bond be- havior, the results of these tests were plotted against pull-out 86 Sodium (mg/cm2) Potassium (mg/cm2) Calcium (mg/cm2) Zinc (mg/cm2) Boron (mg/cm2) Phosphate (mg/cm2)Strand Source ID Warm Water Acid Total Warm Water Acid Total Warm Water Acid Total Warm Water Acid Total Warm Water Acid Total Warm Water Acid Total Recently Manufactured Strand 153 0.030 0.086 0.115 0.009 0.024 0.033 0.016 0.579 0.595 0.003 1.209 1.212 0.002 0.009 0.011 -- -- 1.467 OSU Strand 349 0.133 0.048 0.181 0.003 0.004 0.006 0.030 0.436 0.466 0.004 0.960 0.964 0.002 0.006 0.008 -- -- 1.229 548 0.016 0.148 0.164 0.004 0.039 0.043 0.018 0.075 0.093 0.004 1.395 1.399 0.002 0.005 0.007 -- -- 1.605 697 0.017 0.048 0.065 0.003 0.003 0.005 0.044 0.385 0.429 0.004 0.946 0.950 0.002 0.007 0.009 -- -- 1.206 717 0.019 0.111 0.130 0.002 0.002 0.004 0.016 0.294 0.310 0.004 0.890 0.894 0.003 0.011 0.014 -- -- 1.080 478 * 0.029 0.135 0.165 0.009 0.056 0.065 0.004 0.052 0.056 0.004 0.960 0.964 0.002 0.009 0.011 -- -- 0.902 960 * 0.026 0.130 0.156 0.009 0.053 0.061 0.003 0.048 0.051 0.004 0.968 0.972 0.003 0.011 0.014 -- -- 0.933 * Samples designated 478 and 960 were actually from same source. Table B-23. Elemental analysis—average for each source—results from Correlation Round testing. Figure B-67. Boron analyses: lighter sample is a blank while the darker sample on the right is a sample taken from strand and contains boron.

test behavior measured in the LBPTs in the Screening Round of our evaluations and in the mortar pull-out tests in the Cor- relation Round of our evaluations. Additional comparisons were made between the results of the surface and chemical test methods and the transfer length test results for the recently manufactured strand. Since the transfer length was not meas- ured on the historic strand and the available transfer length results for the samples provided by OSU were deemed unre- liable, the plots presented here show only correlation with pull-out behavior. As discussed above, in all of the historic strand sources, first slip measurements were recorded by visual inspection. For recently manufactured strand (strands from Sources 102, 103, and 151), the stress used for plotting is taken as the stress when end slip was 0.1 in. Contact Angle The contact angle measured on the as-received strand did not correlate well with average bond stress over the transfer length or the bond stress at first or 0.1-in. slip during concrete pull-out testing. This was probably because the lubricant con- tained mixed forms of stearate. The contact angle on strand after ignition process also did not correlate well with the bond 87 102 103 151 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 0.25 0.3 0.35 0.4 0.45 0.5 0.55 0.6 0.65 0.7 Average Bond Stress over Transfer Length (ksi) A ve ra ge B on d St re ss a t 0 .1 -in . s lip (k si) 151 103 102 0.20 0.25 0.30 0.35 0.40 0.45 0.25 0.3 0.35 0.4 0.45 0.5 0.55 0.6 0.65 0.7 Average Bond Stress over Transfer Length (ksi) A ve ra ge B on d St re ss a t 0 .1 -in . s lip (k si) Figure B-68. Concrete pull-out bond stress versus bond stress over transfer length (recently manufactured strand). Figure B-69. Mortar pull-out bond stress versus bond stress over transfer length (recently manufactured strand).

stress at 0.1-in. slip during mortar pull-out testing. In fact, nearly all ignited sources tested gave similar results. Figure B-71 is a plot of the contact angle after the strand was dipped in a saturated calcium hydroxide solution versus stress at 0.1-in. slip in concrete. The contact angle after expo- sure to calcium hydroxide correlated well. The contact angle was lower with greater bond stress. This makes conceptual sense since strand that is more water-repellant will produce a higher contact angle and apparently is less likely to bond well. Measured contact angles below approximately 85° after the calcium hydroxide exposure were indicative of good bond. Because of this good correlation, this test was included in the correlation testing. A comparison of contact angle after expo- sure to calcium hydroxide and 0.1-in. slip stress in mortar is given in Figure B-72. These results from the Correlation Round confirmed the relationship between and contact angle after calcium hydroxide exposure and bond. Examination under Ultraviolet Light Table B-24 shows the bond stress at first or 0.1-in. slip dur- ing concrete pull-out testing and the bond stress over the 88 151 103 102 0.50 0.55 0.60 0.65 0.70 0.75 0.25 0.30 0.35 0.40 0.45 0.50 0.55 0.60 0.65 0.70 Average Bond Stress over Transfer Length (ksi) A ve ra ge B on d St re ss a t 0 .1 -in . s lip (k si) KSU-F KSU-H SC-H SC-IS 101 151 103 102 60 70 80 90 100 110 120 0.0 0.2 0.4 0.6 0.8 0.9 Average Concrete Bond Stress (ksi) Co nt ac t A ng le (° ) a fte r L im e D ip Contact angle - first slip Contact angle - 0.1-in. slip Figure B-70. Hydrocal pull-out bond stress versus bond stress over transfer length (recently manufactured strand). Figure B-71. Correlation between contact angle on strand after a dip in calcium hydroxide solution and bond stress in concrete at first or 0.1-in. slip (historic and recently manufactured strand).

transfer length measured for each of the recently manufac- tured strand sources and the observed fluorescence under UV light. No correlation was found for either property. It is pos- sible that this approach could be useful if inert, fluorescing tracer compounds were intentionally added to the drawing lubricant materials during their manufacture. However, since the commonly available drawing materials do not contain such materials, this method proved ineffective and was not included in the Correlation Round of evaluation. Testing pH Initial testing with the lower resolution, universal Tridi- cator paper on both the historic and recently manufactured strand indicated a limited correlation with bond. This corre- lation was strongest when the droplets were applied to the crevice between wires on the strand surface. Therefore, additional testing of the crevice area was conducted on re- cently manufactured strand (no additional historic strand was available) with three different high-resolution measure- ment techniques: pH meter, Duotest paper, and pH-Fix paper. A higher pH corresponded to a lower average bond stress for all methods. The quality of the correlations differed depending on the measurement method, with the Duotest paper showing the best correlation. The relationship between pH measured with this method and the 0.1-in. slip stress during concrete pull-out testing is plotted in Figure B-73. As bond stress increased, the pH measured with each of these 89 349 548 697 717 478 960 102 103 151 60 65 70 75 80 85 90 95 100 105 110 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 Average Mortar Bond Stress at 0.1-in. Slip (ksi) Co nt ac t A ng le (° ) a fte r L im e D ip OSU Strand Recently-Manufactured Strand Figure B-72. Correlation between contact angle on strand after a dip in calcium hydroxide solution and bond stress in mortar at first or 0.1-in. slip (OSU strand). Strand Source ID Fluorescence Bond Stress at First Slip or 0.1-in. Slip (psi) Average Bond Stress over Transfer Length (psi) Historic Strand KSU-F None 241 - KSU-H None 209 - SC-F Inconsistent dull glow, concentrated on one side of strand (rust) 223 - SC-H None 472 - SC-IS Speckles in crevices 682 - 101 None 241 - Recently Manufactured Strand 102 Dull glow in crevices (rust) 441 363 103 None 944 531 151 None 541 375 Table B-24. UV fluorescence and mechanical properties.

methods decreased for all three sources of recently man- ufactured strand. The pH testing conducted in the Screening Round was deemed initially successful at finding a correlation with bond. Of the methods used to measure pH, the Duotest paper was most effective, since it was easiest to use and produced the least scatter. Therefore, correlation testing using the pH test was performed with the Duotest indicator paper. Comparison of pH and 0.1-in. slip stress in mortar for the recently manu- factured and OSU strands is given in Figure B-74. When all the measurements are considered, the relation- ship between poor bonding performance and pH is un- clear. A lower pH (<8) does not guarantee satisfactory bonding performance. However, the pH test results appear to be influenced by the presence of borax pretreatments, which were apparently used on a number of the historic and recently manufactured strands, based on the elemental analysis of the residue extracts. Analysis of the samples from OSU indicated that none of these strands were pre- treated with borax. Therefore, distinguishing between these 90 102 151 103 6.5 7.0 7.5 8.0 8.5 9.0 9.5 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 Average Concrete Bond Stress (ksi) pH Figure B-73. Correlation between the Duotest paper pH reading and the bond stress in concrete at 0.1-in. slip (recently manufactured strand). 960 478 717 697 548349 151 103 102 6.5 7 7.5 8 8.5 9 9.5 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 Average Mortar Bond Stress at 0.1-in. Slip (ksi) pH OSU Strand Recently-Manufactured Strand Figure B-74. Correlation between Duotest Paper pH reading and bond stress in mortar at 0.1-in. slip (recently manufactured and OSU strand).

latter sources using the pH test method is apparently not as useful. Loss on Ignition The weight LOI versus concrete pull-out bond stress at first or 0.1-in. slip is plotted in Figure B-75. This shows that, as the average bond stress increased, the weight loss on ignition de- creased. The samples of the strand from Source 103 did not lose weight, but in fact gained weight due to oxidation of the steel. Since good correlation was observed, this test was included in the Correlation Round of testing. The results of this testing are compared to the mortar pull-out bond stress at 0.1-in. slip in Figure B-76. The correlation here is less clear, even among the recently manufactured strand, which showed better corre- lation when compared to concrete pull-out results. However, high weight losses above 0.7 mg/cm2 did appear to be consis- tent with low bond. While this test appears to have moderate effectiveness at predicting mortar pull-out stresses, it is easy to conduct and is therefore recommended as part of a larger 91 SC-IS SC-H KSU-H KSU-F 151 103 102 -0.03 -0.01 0.01 0.03 0.05 0.07 0.09 0.11 0.0 0.2 0.4 0.6 0.8 1.0 Average Concrete Bond Stress (ksi) W ei gh t L os s on Ig ni tio n (m g/c m2 ) Loss on Ignition - first slip Loss on Ignition - 0.1-in. slip Figure B-75. Correlation between weight loss on ignition and bond stress in concrete at first or 0.1-in. slip (historic and recently manufactured strands). 960 478 717 697 548 349 151 103 102 -0.05 0.0 0.05 0.1 0.15 0.2 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 Average Mortar Bond Stress at 0.1-in. Slip (ksi) W ei gh t L os s on Ig ni tio n (m g/c m2 ) OSU Strand Recently-Manufactured Strand Figure B-76. Correlation between weight loss on ignition and bond stress in mortar at 0.1-in. slip (recently manufactured and OSU strands).

QC program. Since many materials that are not related to bond may be ignited during this test, the use of this method without other companion testing is not recommended. Loss in Hot Alkali Bath No trends were apparent between the weight loss after soak- ing the strand in a hot alkali (sodium hydroxide solution) bath and average bond stress over transfer length or pull-out bond stress at first or 0.1-in. slip using either of the attempted methods. The weight loss in alkali bath and bond stress in concrete at first or 0.1-in. slip using Method 1 (the more ag- gressive cleaning procedure) is shown in Figure B-77. Based on this lack of correlation, this test method was not included in the Correlation Round of evaluation. Change in Corrosion Potential When compared to concrete pull-out bond stresses at first or 0.1-in. slip, a greater drop in corrosion potential correlated with lower bond stress in the strands as-received (Figure B-78). This correlation was also noted after exposure to calcium 92 KSU-F KSU-H SC-H SC-IS 151 103 10 2 -0.1 0.2 0.4 0.6 0.8 1.0 0.0 0.2 0.4 0.6 0.8 1.0 Average Concrete Bond Stress (ksi) W ei gh t L os s in A lk al i B at h (m g/c m2 ) LAB - first slip LAB - 0.1-in. slip Figure B-77. Correlation between weight loss in alkali bath and bond stressing concrete at first or 0.1-in. slip using Method 1 (historic and recently manufactured strands). 151 101 102 103 -0.40 -0.35 -0.30 -0.25 -0.20 -0.15 -0.10 -0.05 0.00 0.0 0.2 0.4 0.6 0.8 1.0 1.2 Average Concrete Bond Stress (ksi) Ch an ge in C or ro si on P ot en tia l (V ) as -R ec ei ve d Figure B-78. Correlation between corrosion potential change and bond stress in concrete at 0.1-in. slip (recently manufactured strand in as-received condition).

hydroxide solution, but less pronounced. Similar trends were observed based on the transfer length data. Note that a greater corrosion potential drop implies a greater readiness to corrode. Based on this apparent correlation, a 6-h test to measure the change in corrosion potential of strand in the as-received condition was included in the Correlation Round. The results of that testing are shown in Figure B-79. Testing after ignition was not helpful for evaluating bond performance. Since a greater drop in corrosion potential means that the strand corroded more easily, this was contrary to what was expected. Instead, it had been anticipated that a strand with some surface residue would be more likely to resist corrosion. This expectation was based on field observations of flash rusting on clean steel and based on work by Perenchio et al. (1989). The metallurgical examination of the strand surface may explain why the corrosion performance is contrary to expectations. The half-cell tests provide indications of active corrosion, but not where the corrosion is occurring. It is pos- sible that the relatively rough surfaces found in low bond strand provide small corrosion potential sites at the “peaks” of the surface profile, compared to the “valleys,” where most of the lubricant presumably resides. Thus, the corrosion activity measured may be occurring in small localized areas. The greater surface area per unit length of the rougher strand when considered on a microscopic scale may also support more corrosion. Surface Roughness Increased surface roughness as measured using the param- eter Ra corresponded to a lower bond stress in concrete at 0.1-in. slip, as shown in Figure B-80. This was consistent with the visual interpretation of SEM micrographs discussed in Appendix D. Similar trends were observed based on the transfer length data. As a result, this test was included in the Correlation Round of evaluation. The results of that testing are shown in Figure B-81. These data show a slight correlation but, as can be seen from the standard deviations in these fig- ures, there is significant variation in the individual readings. The measurements are relatively quick and easy to take, but this test method does not appear to be effective enough to dif- ferentiate the mortar pull-out bond quality of the strand sources. Corrosion Rate As seen in Figure B-82, corrosion rate correlated well with the concrete bond stress at first or 0.1-in. slip. Note that a higher corrosion rate implies a greater readiness to corrode. For each test of as-received strand and strand after exposure to calcium hydroxide solution, higher bond stress corresponded to lower corrosion rates. Similar trends were observed based on the transfer length data. Because of this good correlation, this test was included in the Correlation Round of evaluation. In that round, the corrosion rate did not correlate well in any of the three conditions (as-received, after calcium hydroxide solution exposure, and after ignition) tested. The results of the as-received tests are plotted versus mortar pull-out test results in Figure B-83. The trend that corrosion occurs more readily on strands with poorer bond properties, at least when concrete pull-out testing is considered, is similar to that observed during the 93 -0.400 -0.350 -0.300 -0.250 -0.200 -0.150 -0.100 -0.050 0.000 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 Average Mortar Bond Stress at 0.1-in. Slip (ksi) Ch an ge in C or ro si on P ot en tia l (V ) as -R ec ei ve d 349 548 697 717 478 960 102 103 151 OSU Strand Recently-Manufactured Strand Figure B-79. Correlation between corrosion potential change and bond stress in mortar at 0.1-in. slip (recently manufactured and OSU strand in as-received condition).

change in corrosion potential measurements, and is also contrary to prior expectations. Possible explanation for the unexpected result is described in the section entitled Change in Corrosion Potential. Organic Residue Extraction For the sources tested as part of this study, as extraction residue decreased, bond stress increased. The relationship is evident when the “total” extraction residue, that is the combined residue concentration from both the water and acid/chloroform wash solutions, is considered. As shown in Figure B-84, the “total” extraction residue correlates well with the concrete pull-out bond stress at first or 0.1-in. slip for both the hot- and warm-water wash procedures. A similar strong relationship exists between the extracted organic residue and the average bond stress over the transfer length. Because of this good correlation, this test was included in the Correlation Round of evaluation. In that round the mor- tar pull-out bond stress at 0.1-in. slip correlated less well with 94 102 103 151 0.00 0.20 0.40 0.60 0.80 1.00 1.20 0.4 0.5 0.6 0.7 0.8 0.9 1.0 1.1 Average Concrete Bond Stress (ksi) Su rfa ce R ou gh ne ss , R a (µ m) 349 548 697 717 478 960 102 103 151 0.00 0.20 0.40 0.60 0.80 1.00 1.20 1.40 1.60 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 Average Mortar Bond Stress at 0.1-in. Slip (ksi) Su rfa ce R ou gh ne ss , R a (µ m) OSU Strand Recently-Manufactured Strand Figure B-80. Correlation between surface roughness parameter Ra and bond stress in concrete at 0.1-in. slip (recently manufactured strand). Figure B-81. Correlation between surface roughness parameter Ra and bond stress in mortar at 0.1-in. slip (recently manufactured and OSU strand in as-received condition).

total residue (Figure B-85). However, samples with a high concentration of organic residue (greater than 0.05 mg/cm2) had a bond stress at 0.1-in. slip in mortar lower than 0.4 ksi. Elemental Analyses The elemental analysis was conducted on the extraction solutions in both the screening and correlation testing. This testing was performed to provide information about compo- nents in pretreatments and lubricants used to manufacture strand and the effect these components may have on bond performance. Sodium—The surface concentration of sodium in the water wash per unit area of strand versus the concrete pull- out bond stress at first or 0.1-in. slip is illustrated in Fig- ure B-86. The sodium concentration from the water-wash solutions and the combined water and acid/solvent wash concentrations showed a similar relationship to bond stresses: a high concentration of sodium correlates to a low concrete 95 151 103 102 101 SC-IS SC-H KSU-H KSU-F 151 103 102 -20 -10 0 10 20 30 40 50 60 70 0.0 0.2 0.4 0.6 0.8 1.0 Average Concrete Bond Stress (ksi) Co rr os io n Ra te (µ A/ cm 2 ) Corrosion rate Corrosion rate - after lime dip 151 103 102 349 548 697 717 478 960 -20 0 20 40 60 80 100 120 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 Average Mortar Bond Stress at 0.1-in. Slip (ksi) Co rr os io n Ra te (µ A/ cm 2 ) Recently-Manufactured Strand OSU Strand Figure B-82. Correlation between corrosion rate as-received and after calcium hydroxide (lime) solution exposure and concrete bond stress at first or 0.1-in. slip in concrete (historic and recently manufactured strand). Figure B-83. Correlation between corrosion rate as-received and bond stress in mortar at 0.1-in. slip (recently manufactured and OSU strand).

bond stress, but that a low concentration of sodium does not guarantee a high concrete bond stress. The surface concen- tration of sodium from the acid/chloroform wash following the warm-water wash did not correlate well with bond. The results from a comparison of water-soluble sodium with mortar pull-out test results in the Correlation Round are similar to that compared to concrete pull out, as shown in Figure B-87. Calcium—Figure B-88 shows the calcium surface concen- tration per unit area of strand from the acid/chloroform wash. Nearly all calcium removed from these strands was removed by the acid/chloroform wash. The initial water washes were ineffective at removing compounds containing this element, and the plot of the total calcium concentration mimics that of the acid/chloroform wash. In the Correlation Round, the total calcium surface concentration showed little correlation to bond in mortar (Figure B-89). Potassium—Figure B-90 shows the relationship between the potassium surface concentration from the water wash and concrete pull-out bond stress, and Figure B-91 illustrates the 96 101 15 1 103 102 101 SC-IS SC-H SC- F KSU- H KSU-F 151 103 102 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.14 0.16 0.0 0.2 0.4 0.6 0.8 1.0 Average Concrete Bond Stress (ksi) Ex tr ac tio n Re si du e (m g/c m2 ) Hot water/acid-chl. washes - first slip Hot water/acid-chl. washes - 0.1-in. slip Warm water/acid-chl. washes - first slip Warm water/acid-chl. washes - 0.1-in. slip 960 478 717 697 548 349 151 103 102 0.000 0.020 0.040 0.060 0.080 0.100 0.120 0.140 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 Average Mortar Bond Stress at 0.1-in. Slip (ksi) Ex tr ac tio n Re si du e (m g/c m2 ) OSU Strand - Warm water/acid-chl. washes Recently-Manufactured Strand - Warm water/acid-chl. washes Figure B-84. Correlation between the “total” combined residues (water plus acid/chloroform washes) and bond stress at first or 0.1-in. slip in concrete (historic and recently manufactured strand). Figure B-85. Correlation between the “total” combined residues (water plus acid/chloroform washes) and bond stress at first or 0.1-in. slip in mortar (recently manufactured and OSU strand).

relationship between the potassium surface concentration from the acid/chloroform wash and the concrete bond stress. As can be seen, for most sources, the majority of the potassium was removed by the water wash procedure. However, for a number of sources, particularly SC-H and SC-IS, potassium compounds of apparently low water-solubility were present, and these compounds required the acid/chloroform wash procedure to remove them. Since the amount of potassium removed by the water wash was generally greater than by the acid/chloroform wash in the Screening Round, it appeared that a high level of water-soluble potassium corresponds to a low concrete bond stress, but at a low level of potassium, the bond stress may be high or low. This relationship was less clear in the Correlation Round, as shown in Figure B-92, which shows the potassium surface concentration from the water wash and mortar pull-out bond stress. Zinc—Zinc was found in non-negligible concentrations only in the acid/chloroform wash. Figure B-93 shows the total zinc concentration versus the concrete bond stress at first 97 101 SC-IS SC-H SC-F KSU-H KSU-F 151 103 102 0.00 0.05 0.10 0.15 0.20 0.25 0.0 0.2 0.4 0.6 0.8 1.0 Average Concrete Bond Stress (ksi) M as s So di um p er A re a (m g/c m2 ) Warm water wash - first slip Warm water wash - 0.1-in. slip 349 548 697 717 478 960 102 103 151 0.000 0.020 0.040 0.060 0.080 0.100 0.120 0.140 0.160 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 Average Mortar Bond Stress at 0.1-in. Slip (ksi) M as s So di um p er A re a (m g/c m2 ) OSU Strand - Warm water wash Recently-Manufactured Strand - Warm water wash Figure B-86. Correlation between sodium concentration from warm-water wash and bond stress at first or 0.1-in. slip in concrete (historic and recently manufactured strand). Figure B-87. Correlation between sodium concentration from warm-water wash and bond stress at 0.1-in. slip in mortar (recently manufactured and OSU strand).

or 0.1-in. slip. A high concentration of extracted zinc was associated with both high and low bond stresses and did not correlate well with bond quality. This was confirmed in the Correlation Round (Figure B-94). Boron—Unlike the zinc compounds, the boron compounds in the strand surface residue were largely water-soluble. Figure B-95 shows the total concentration of boron com- pared to the concrete pull-out bond stress. Low boron con- tents appear to correlate with higher average bond stress over the transfer length. Following the trend observed in much of the elemental analysis, a high boron level signifies a low bond stress, but a low boron level does not necessarily guarantee a high bond stress. This was confirmed in the Correlation Round (Figure B-96). Phosphate—Not surprisingly, like the concentration of zinc with which it probably was applied to the strand in a pre- treatment process, the phosphate concentration did not cor- relate well with bond (Figure B-97). High concentrations of 98 SC- F KSU-H KSU- F SC-H SC-IS 101 10 2 103 151 0.0 0.2 0.4 0.6 0.8 1.0 1.2 0.0 0.2 0.4 0.6 0.8 1.0 Average Concrete Bond Stress (ksi) M as s Ca lc iu m p er A re a (m g/c m2 ) Acid/chl. wash after warm water - first slip Acid/chl. wash after warm water - 0.1-in. slip 960 478 717 697 548 349 151 103 102 0.000 0.100 0.200 0.300 0.400 0.500 0.600 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 Average Mortar Bond Stress at 0.1-in. Slip (ksi) M as s Ca lc iu m p er A re a (m g/c m2 ) OSU Strand - Warm water/acid-chl. washes Recently-Manufactured Strand - Warm water/acid-chl. washes Figure B-88. Correlation between calcium concentration from acid/chloroform wash after warm-water wash and bond stress at first or 0.1-in. slip in concrete (historic and recently manufactured strand). Figure B-89. Correlation between total calcium concentration from warm-water and acid/chloroform washes and bond stress at 0.1-in. slip in mortar (recently manufactured and OSU strand).

extracted phosphate were associated with both high and low bond stresses. Interpretation of Elemental Analyses Relative to Manufacturing Processes The atomic absorption and colorimetric analyses done on wash solutions from strands from various sources measured varied levels of sodium, calcium, potassium, zinc, boron, and phosphate. The presence and concentration of these elements is largely governed by the specific pretreatment process and wire-drawing lubricants used in the manufacturing of each specific strand source, though other sources of these elements may be possible. The pretreatments commonly used in strand production are zinc phosphate and sodium borate, also known as borax. Lubricants typically consist primarily of either sodium stearate, calcium stearate, or some other fatty acid. Table B-25 and Table B-26 show each strand source ranked in 99 KSU-F KSU-H SC-F SC-H SC-I S 101 15 1 10 3 10 2 0.00 0.01 0.02 0.03 0.04 0.05 0.06 0.07 0.08 0.0 0.2 0.4 0.6 0.8 1.0 Average Concrete Bond Stress (ksi) M as s Po ta ss iu m p er A re a (m g/c m2 ) Wa rm water wash - first slip Wa rm water wash - 0.1-in. slip KSU-F KS U- H SC-F SC-H SC-IS 101 102 103 151 0.000 0.005 0.010 0.015 0.020 0.025 0.030 0.035 0.040 0.0 0.2 0.4 0.6 0.8 1.0 Average Concrete Bond Stress (ksi) M as s Po ta ss iu m p er A re a (m g/c m2 ) Acid/chl. wash after warm water - first slip Acid/chl. wash after warm water - 0.1-in. slip Figure B-90. Correlation between potassium concentration from the warm-water wash and bond stress at first or 0.1-in. slip in concrete (historic and recently manufactured strand). Figure B-91. Correlation between potassium concentration from acid/chloroform wash after warm-water wash and bond stress in concrete at first or 0.1-in. slip (historic and recently manufactured strand).

ascending order of concrete or mortar pull-out bond per- formance together with the prominent elements removed dur- ing the extraction process, and the presumed pretreatment and lubricants used based on the prominent elements. Knowledge about the types of pretreatments and drawing lubricants can be an important part of the interpretation of the QC test results. These results of the elemental analyses suggest that the wires in the sampled strands were pretreated using one of two methods during the manufacturing process. Those strands that carried high amounts of boron typically carried low amounts of zinc and phosphate. As expected, the lubricants appeared to be either calcium or sodium/potassium stearates but a number of strands showed evidence of both. This may result from different types of lubricants being used in separate dies in the drawing process. There is a greater amount of sodium, potassium, and calcium than would be expected from the stearate compounds alone (based on the concentration of the organic residues extracted). Therefore, other sources of these elements are contributing to the values measured here. Possible sources include chemicals 100 349 548 697717 478 960 102 103 151 0.000 0.010 0.020 0.030 0.040 0.050 0.060 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 Average Mortar Bond Stress at 0.1-in. Slip (ksi) M as s Po ta ss iu m p er A re a (m g/c m2 ) OSU Strand - Warm water/acid-chl. washes Recently-Manufactured Strand - Warm water/acid-chl. washes 101 SC-IS SC-H SC-F KSU-H KSU-F 15 1 103 102 -0.1 0.1 0.3 0.5 0.7 0.9 1.1 1.3 1.5 0.0 0.2 0.4 0.6 0.8 1.0 Average Concrete Bond Stress (ksi) M as s Zi nc p er A re a (m g/c m2 ) Total warm water and acid/chl. wash - first slip Total warm water and acid/chl. wash - 0.1-in. slip Figure B-92. Correlation between total potassium concentration from warm-water wash and bond stress at 0.1-in. slip in mortar (recently manufactured and OSU strand). Figure B-93. Correlation between zinc concentration from combined warm-water and acid/chloroform washes and bond stress at first or 0.1-in. slip in concrete (historic and recently manufactured strand).

in the pretreatment processes described above; detergents used for cleaning; or fillers, such as lime, used in some drawing lubricants. Combined Elemental Analysis For any single element or organic residue quantified (as discussed above), good bond performance was observed only with strand sources on which low concentrations were found. However, poor bond performance was found with strands with both high and low concentrations of these single entities. This may be explained by suggesting that high concentrations of any of these element or organic residues has the potential to produce poor bond, and that while a poor bonding strand may have a low concentration of one single element, it is likely to be high in at least one other. To investigate this idea, analysis was done to compare bond performance associated with various combinations of these elements. The objective 101 960 478717 697 548 349 151 103 102 0.000 0.200 0.400 0.600 0.800 1.000 1.200 1.400 1.600 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 Average Mortar Bond Stress at 0.1-in. Slip (ksi) M as s Zi nc p er A re a (m g/c m2 ) OSU Strand - Warm water/acid-chl. washes Recently-Manufactured Strand - Warm water/acid-chl. washes 10 1 SC-IS SC-H SC-F KSU-H KSU- F 151 103 102 0.00 0.02 0.04 0.06 0.08 0.10 0.12 0.0 0.2 0.4 0.6 0.8 1.0 Average Concrete Bond Stress (ksi) M as s B or on p er A re a (m g/c m2 ) Total warm water and acid/chl. wash - first slip Total warm water and acid/chl. wash - 0.1-in. slip Figure B-94. Correlation between zinc concentration from combined warm-water and acid/chloroform washes and bond stress at 0.1-in. slip in mortar (recently manufactured and OSU strand). Figure B-95. Correlation between concentration from boron in combined warm-water and acid/chloroform washes and bond stress at first or 0.1-in. slip in concrete (historic and recently manufactured strand).

of this combination approach was to determine which ele- ments had the most significant effect on bond and to identify an indicator or score that might be used to predict poor bond. This score, which has been called the combined index, was defined so that it would increase if any of the included elements were found on the strand. This combined index was defined as the average of the scaled residue concentrations for certain combinations of elements after the concentrations were scaled from 0 to 1, to account for differences in the magnitude of the measured concentra- tions. This was done for the total concentrations, that is, the sum of the materials removed using water and acid/chloroform washes in the Screening Round or the combined wash solu- tions in the Correlation Round, measured with warm water as the first step in the wash procedure. The scaling was performed by dividing the concentrations measured for each source by 102 960 478 717 697 548 349 151 103 102 0.000 0.010 0.020 0.030 0.040 0.050 0.060 0.070 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 Average Mortar Bond Stress at 0.1-in. Slip (ksi) M as s B or on p er A re a (m g/c m2 ) OSU Strand - Warm water/acid-chl. washes Recently-Manufactured Strand - Warm water/acid-chl. washes 960 478 717 697 548 349 0.800 0.900 1.000 1.100 1.200 1.300 1.400 1.500 1.600 1.700 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 Average Mortar Bond Stress at 0.1-in. Slip (ksi) M as s Ph os ph at e pe r A re a (m g/c m2 ) OSU Strand - Warm water/acid-chl. washes Figure B-96. Correlation between boron concentration from combined warm-water and acid/chloroform washes and bond stress at 0.1-in. slip in mortar (recently manufactured and OSU strand). Figure B-97. Correlation between phosphate concentration from combined warm water and acid/chloroform washes and bond stress at 0.1-in. slip in mortar (recently manufactured and OSU strand).

the maximum measured for that compound in each round. In this way, the maximum concentration was assigned a value of 1, and lower values were distributed between 0 and 1. The idea behind the averaging was that if any of the concentrations were high, it would be reflected in the index for that source. Table B-27 shows the scaled concentrations for sodium, potassium, calcium, zinc, boron, and the organic residue for those sources for which concrete pull-out testing results were available. The relationship between bond and combined in- dices based on many element and organic residue combina- tions were explored, from individual pairs of elements to combinations of all the scaled concentrations. The combined index based on boron, calcium, and extracted organic residue was found to have the highest R2 value (R2 = 0.78) when plotted against concrete pull-out bond stress, as seen in Figure B-98. This figure illustrates a good correlation between poor bond stress and these three normalized values present together. The final column in Table B-27 is the combined index of boron, calcium, and organic residue used to generate this plot. A similar analysis was performed for those samples for which mortar testing was available and Table B-28 show the scaled concentrations for sodium, potassium, calcium, zinc, boron, and the organic residue. Once again, the combined index based on boron, calcium, and organic residue was found to have the highest R2 value (as shown in Figure B-99), but this time the correlation was much less (R2 = 0.28). Summary of Test Method Correlation A large experimental program has been conducted to sup- port the evaluation of various proposed test methods for strand bond that were intended for use as part of a QC program. These were classified as performance-based (i.e., mechanical) tests and surface and chemical tests. The correlation between bond and the methods that fall under each of these classifica- tions were evaluated differently based on the strand sources that were collected for testing and the data quantifying bond performance that were available. The evaluation of correla- tions between the various pull-out testing methods and the bond were based on bond performance as measured with transfer length tests. The correlations between the surface and chemical test methods were evaluated based on the results of 103 Strand Prominent Elements Presumed Pretreatment Presumed Lubricants Concrete Pull- Out Bond Stress (ksi) 153 Ca, Zn, P zinc phosphate calcium salt offatty acid - KSU-H Na, K, B borax Na/K stearate 0.209 SC-F Ca, Zn zinc phosphate calcium stearate 0.223 101 Na, K, B borax Na/K stearate 0.241 KSU-F Na, K, B borax Na/K stearate 0.241 102 Na, Ca, B borax Na/Ca stearate 0.441 SC-H Na, K, Zn zinc phosphate sodium stearate 0.472 151 Na, B borax calcium stearate 0.541 SC-IS Na, Zn zinc phosphate sodium salt of fatty acid 0.682 103 Na, Zn zinc phosphate sodium salt of fatty acid 0.944 Table B-25. Compounds likely used in manufacture of each source—historic and recently manufactured strands. Strand Prominent Elements Presumed Pretreatment Presumed Lubricants Mortar Pull- Out Bond Stress (ksi) 349 Ca, Zn, P zinc phosphate calcium salt of fatty acid and resin 0.156 717 Na, Ca, Zn, P zinc phosphate Na/Ca stearate 0.206 478 Na, K, Zn, P zinc phosphate Na/K stearate 0.409 960 Na, K, Zn, P zinc phosphate Na/K stearate 0.409 697 Ca, Zn, P zinc phosphate calcium salt of fatty acid and resin 0.606 548 Na, K, Zn, P zinc phosphate Na/K salt of fatty acid 0.623 Table B-26. Compounds likely used in manufacture of each source—OSU strands.

104 Source Zinc Potassium Sodium Calcium Boron Organic Residue Combined Index for B, Ca, and Organic Residue Historic Strand KSU-F 0.004 1.000 1.000 0.100 1.000 1.000 0.700 KSU-H 0.004 0.798 0.790 0.092 0.856 0.829 0.592 SC-F 0.623 0.158 0.356 1.000 0.101 0.749 0.617 SC-H 0.620 0.372 0.568 0.097 0.168 0.295 0.187 SC-IS 1.000 0.361 0.757 0.035 0.113 0.114 0.087 101 0.002 0.567 0.651 0.020 0.989 0.388 0.466 Recently Manufactured Strand 102 0.003 0.357 0.738 0.362 0.515 0.487 0.455 103 0.636 0.199 0.430 0.030 0.182 0.081 0.098 151 0.003 0.088 0.308 0.035 0.604 0.259 0.300 Table B-27. Scaled concentrations and combined index for comparison with concrete pull out. KSU-F KSU-H SC-F SC-H SC-IS 101 (warm) 102 (warm) 103 (warm) 151 (warm) y = -0.9622x + 0.8181 R2 = 0.7813 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 0.9 1.0 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 Combined Index, B, Ca, Organic Residue Co nc re te P ul l O ut B on d St re ss (k si) Figure B-98. Comparison between concrete pull-out bond stress at first or 0.1-in. slip and the combined index of boron, calcium, and organic residue. Source Zinc Potassium Sodium Calcium Boron Organic Residue Combined Index for B, Ca, and Organic Residue Recently Manufactured Strand 102 0.003 0.483 1.000 0.648 0.853 0.594 0.698 103 0.642 0.270 0.583 0.055 0.302 0.098 0.152 151 0.003 0.119 0.417 0.064 1.000 0.316 0.460 OSU Strand 349 0.688 0.067 0.251 1.000 0.104 0.186 0.430 548 1.000 0.697 0.769 0.172 0.083 0.403 0.220 697 0.678 0.045 0.248 0.882 0.114 0.160 0.386 717 0.638 0.034 0.577 0.675 0.177 1.000 0.617 478 0.688 1.000 0.704 0.119 0.146 0.283 0.183 960 0.694 0.944 0.678 0.109 0.177 0.297 0.194 Table B-28. Scaled concentrations and combined index for comparison with mortar pull out.

pull-out tests from concrete in the Screening Round and based on pull-out tests from mortar in the Correlation Round. The correlations with bond are discussed below for each classification. To provide a quantitative measure of the goodness-of-fit to aid in the evaluation of these methods, a linear regression has been performed, and the coefficient of determination (R2) de- termined for the relationship between each proposed test method and the bond quality measure. No physical basis for a linear relationship between these measures of bond is known; however, the linear relationship was assumed as the simplest model relating the parameters. The coefficient of determination (R2) is a measure of the adequacy of a regression model; that is, it describes the amount of variability in the data explained by the regression model. The closer the R2 is to 1.0, the more com- pletely the model describes the relationship between the test method results and the basis for evaluation. These coefficients of determination are presented in Table B-29 and Table B-30. To further evaluate the validity of these methods, the significance of the linear models developed based on these data was evaluated by the calculation of P-values for the coefficients (slope) from the linear models. The coefficient from the linear model is judged to be significant when there is a sufficiently high confidence that it is not equal to zero. If this is true, the relationship represented by the model is sta- tistically significant and the results of the surface tests are meaningful in the prediction of the pull-out test. A 95% con- fidence level is commonly used to evaluate significance. The level of confidence of significance on the coefficient is given by (1 − P-value) × 100%, so a P-value < 0.05 implies that the con- fidence interval does not include zero with higher than 95% confidence. The P-value was determined for the relationship between selected test methods and the bond quality measures and these are presented in Table B-31. This concept of P-value is revisited in Regression and Prediction Intervals in the next section, Development of Thresholds. Correlation with Bond—Mechanical Test Methods The concrete pull-out test results correlated better with bond quality than the other pull-out test methods that were evalu- ated, based on comparisons with transfer length tests con- ducted on three strand sources. However, pull-out testing from mortar also showed promise as a means to evaluate bond, and the existing correlation is deemed sufficient to justify further study. This limited program cannot be considered a definitive evaluation of these methods. Nevertheless, this conclusion is contrary to that of other studies of pull-out test methods, in- cluding that sponsored by NASPA which have concluded that the mortar pull-out test is superior at assessing bond per- formance (Russell and Paulsgrove 1999, Russell 2001, Russell 105 960 478 717 697548 349 151 (warm) 103 (warm) 102 (warm) y = -0.4259x + 0.5351 R2 = 0.2762 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.0 0.1 0.2 0.3 0.4 0.5 0.6 0.7 0.8 Combined Index, B, Ca, Organic Residue M or ta r P ul l O ut B on d St re ss (k si) Figure B-99. Comparison between mortar pull-out bond stress at 0.1-in. slip and the combined index of boron, calcium, and organic residue. Test Method QC Level Coefficient of Determination (R2) from Regression with Average Bond Stress over Transfer Length Concrete Pull Out II 0.98 Mortar Pull Out II 0.85 Hydrocal Mortar Pull Out II 0.36 Table B-29. Coefficient of determination (R2) from linear regression with average bond stress over transfer length.

106 Coefficient of Determination (R2) from Regression with Mechanical Test Test Method QC Level Concrete Pull Out (0.1-in. and First Slip) Mortar Pull Out (0.1-in. Slip) As-Received I 0.04 0.35 After Ca(OH)2 Dip I 0.61 0.57 After Ca(OH)2 Dip— Stearate Only† I 0.44 0.84 Contact Angle (°) After Ignition I N.A. 0.00 pH I 0.97 0.18 Loss on Ignition I 0.86 0.16 Loss on Alkali Bath I 0.17 0.76* As-Received I 0.72 0.68 After Ca(OH)2 Dip I 0.80 1.00* Change in Corrosion Potential after 6 h After Ignition I N.A. 0.00 Surface Roughness, Ra I 0.93 0.16 As-Received II 0.67 0.09 After Ca(OH)2 Dip II 1.00 0.00 Corrosion Rate After Ignition II N.A. 0.18 Total II 0.81 0.12 Organic Residue Extraction Total—Stearate only† II 0.88 0.63 Warm Water II 0.34 0.31 Sodium Total II 0.12 0.02 Warm Water II 0.39 0.00 Potassium Total II 0.28 0.14 Warm Water II 0.17 0.06 Calcium Total II 0.22 0.07 Warm Water II 0.21 0.05 Zinc Total II 0.25 0.21 Warm Water II 0.30 0.10 Boron Total II 0.28 0.11 Phosphate Total II N.A. 0.17 Combined Index for B, Ca, & Org. Res. Scaled for combination II 0.78 0.28 R2 values presented in bold are for those methods recommended for use in a QC program. * Test method not included in Correlation Round—regression based on three sources. † Only those sources identified as containing primarily stearate-based compounds by FTIR analysis are considered. Table B-30. Coefficient of determination (R2) from linear regression with concrete and mortar pull out at 0.1-in. and first slip. P-value from Regression with Mechanical Test Test Method QC Level Concrete Pull Out (0.1-in. and First Slip) Mortar Pull Out (0.1-in. Slip) After Ca(OH)2 Dip I 0.039 0.019 Contact Angle (°) After Ca(OH)2 Dip - Stearate Only† I 0.262 0.029 Loss on Ignition I 0.003 0.285 Change in Corrosion Potential after 6 h As-Received I 0.356 0.006 Total II 0.002 0.353 Organic Residue Extraction Total—Stearate only† II 0.006 0.110 † Only those sources identified as containing primarily stearate-based compounds by FTIR analysis are considered. Table B-31. P-value from linear regression with concrete and mortar pull out at 0.1-in. and first slip.

2006). This correlation was not explored further in the Correlation Round of testing. Correlation with Bond—Surface and Chemical Test Methods At the initiation of this study, the surface and chemical methods were divided into Level I and II QC tests, based on the required effort and complexity of each test. These correla- tions are discussed separately here, since the level of correlation required to justify the use of each test method is different for each QC level. As can be seen in Table B-30, a number of the surface and chemical test methods that showed good cor- relation with concrete pull-out test results did not correlate as well with the mortar pull-out test results. This may be in- dicative of the inadequacy of the surface and chemical methods, but may also be related to inaccuracies in the pull-out test methods. For the contact angle and organic residue extraction meas- urement methods, coefficients of determination have been calculated using only data from sources identified in the FTIR analyses as carrying only stearate-based lubricants. This was done to eliminate the potentially confounding influences of non-stearate-based lubricants. Analyzing these data in this manner has a practical motivation, since such models could be useful in a production setting where the lubricant in use is known to be only stearate-based. Finally, the significance of the relationships between the pull-out test results and the surface and chemical method re- sults given in terms of the P-values for the coefficient are con- sidered for those methods that showed promise based on the coefficients of determination or other factors. Level I QC tests—The objective of the Level I QC test methods is to quickly and easily determine if strand prop- erties that have been correlated with questionable bond are present. The minimum correlation required for these tests to be useful is somewhat lower than for the Level II QC tests. • Contact Angle—Contact angle correlated with bond only after the strand sample was subjected to exposure to a sat- urated calcium hydroxide solution. This correlation is higher for those sources judged to carry only stearate- based lubricants, when performance based on mortar pull- out testing is considered. Both P-values calculated when comparing this test against mortar and concrete pull-out testing are less than 0.05, suggesting that there is greater than 95% confidence that the relationships between the pull-out tests and this test method are significant. It is rec- ommended that this method be included as part of a future QC program. • Examination under UV light—A limited quantity of fluo- rescing material was observed, and no correlation to bond was found. This method should be abandoned. • Testing pH—The pH test was successful in finding a cor- relation with bond as measured by concrete-based pull-out testing on a limited data set, but it was unsuccessful at finding a similar correlation based on mortar pull-out test results. It also appeared that this test was only effective to differentiate strands produced with a borax pretreatment. Therefore, this method may be applicable only for strand produced with borax pretreatments. More study is needed before a recommendation regarding the adaptation of this method can be made. • Loss on Ignition—A good correlation was found between the ignition loss and bond performance measured in con- crete pull-out tests. The P-value calculated when comparing this test against concrete pull-out testing is less than 0.01, suggesting that there is greater than 99% confidence that this test method is significant at predicting concrete pull out. This correlation or significance was not found based on mortar pull-out test results. This is one of the easiest tests to perform and is recommended for future testing, though not as a sole measure of bond performance. • Loss in Alkali Bath—Multiple cleaning procedures using sodium hydroxide solutions were attempted, but no cor- relation was observed between the weight loss and bond in concrete. While a higher correlation was found with the mortar test, this higher correlation is based only on three sources. Interestingly, this is the only method suggested by the Wire Association International manual for measuring surface residues on wire. It is recommended that this test be abandoned. • Change in Corrosion Potential—Although the mechanism behind the observed trend is somewhat uncertain, the drop in corrosion potential showed a good correlation with bond in both the Screening and Correlation Rounds of evaluation. The P-value calculated when comparing this test against mortar pull-out testing is less than 0.01. When compared to concrete, the coefficient of determination is similarly high, but the P-value is also high (0.356), sug- gesting that this relationship is not significant based on a 95% confidence threshold. This high P-value is likely in- fluenced by the fact that this analysis is based on only three samples. A nearly perfect correlation (R2 = 1.00) with mortar pull out was measured in the Screening Round for this test when performed on samples after exposure to a saturated calcium hydroxide solution. However, this higher correlation is only based on three sources, and it was judged that this additional effort is not worthwhile. Therefore, it is recommended that this method, conducted on strands in as-received condition, be included as part of a future QC program. 107

• Surface Roughness—The surface roughness parameter, Ra, correlated well with bond in concrete based on only three sources, but not with bond in mortar. Since an increased roughness was associated with poor bond, it appears that correlation to bond is not a direct effect, but is related to the tendency of the wire surfaces to retain residue. The pro- filometer used to measure this property is convenient for use in a QC setting, but does not appear to be sensitive enough to measure the roughness at the scale needed, nor does it test a sufficiently large surface of the strand for the test result to be representative of a property that can be tied to bond performance. Therefore, it is recommended this method be abandoned in its current form. However, we believe that this test method could be viable if performed using a different methodology. Unfortunately, project funds did not permit further development of this test method. Level II QC tests—The objective of Level II QC testing is to provide a more conclusive prediction of bond perform- ance than possible with the Level I QC tests. These tests re- quire either more advanced methods or more complicated equipment. The minimum correlation required for these tests is higher than for the Level I QC tests. • Corrosion Rate—A strong correlation was measured between corrosion rate and pull-out bond in concrete. However, the correlation between corrosion rate and pull- out bond in mortar was relatively weak. Given this lack of consistent correlation, the uncertainty about the mecha- nisms involved in establishing the initial good correlation, and the complexity and equipment-dependent nature of this test, it is not recommended for inclusion in a future QC program. • Organic Residue Extraction—The concentration of the organic residue correlated well with the bond performance in concrete, but only moderately with bond in mortar. This test is time-consuming to perform, but gives the best direct measure of the type and quantity of drawing lubricants left on the strand surface during the manufacturing process. Of all the methods proposed, this method evaluates the property of the strand tied most directly to bond quality. Therefore, it is recommended that this method be included as part of a future QC program. FTIR spectroscopy should continue to be performed on the organic residues to ensure that residues being evaluated are consistent. This is necessary because the effect of residues with different chemistries is unlikely to be proportionally similar (e.g., a stearate-based lubricant residue will likely effect bond differently than a non-stearate-based lubricant residue). FTIR analyses will also identify contamination of the samples from other organic materials, such as oils, greases, or form release agents. The correlation between mortar pull-out stress and residue concentration was much higher when those sources carrying only stearate-based lubricants were included in the correlation analysis. Nevertheless, the P-values calcu- lated when comparing this test against mortar pull-out testing considering all samples was less than 0.01, provid- ing further support to the recommendation to use this test method. • Elemental Analysis—Atomic absorption and visual light spectroscopy were used to determine the surface concen- tration of various elements. The concentrations of sodium and boron showed signs of a correlation with the mechan- ical properties measured in pull-out tests in both concrete and mortar. The concentrations of zinc, however, did not. The elemental analysis gives some insight into the type of pretreatment and lubricant in use and was useful for the purposes of this study. The combined index based on boron, calcium, and organic residue, combines the elemen- tal analyses with the organic residue extraction procedures, and showed good correlation to pull-out bond in concrete. This correlation was not found with mortar pull-out results. Given the cost and equipment-dependent nature of the atomic absorption testing, it is not recommended for in- clusion in a future QC program. In summary, the methods that are recommended for inclu- sion in future QC programs are as follows: • Weight Loss on Ignition (LOI) of Steel Strand, • Determination of the Surface Tension of Steel Strand by Contact Angle Measurement, • Change in Corrosion Potential of Steel Strand, and • Organic Residue Extraction with FTIR Analysis. Precision Testing The recommended QC methods have been written in AASHTO/ASTM standard method format in Appendix C, where they are titled: 1. Test Method for the Determination of the Surface Tension of Steel Strand by Contact Angle Measurement, 2. Test Method for Weight Loss on Ignition (LOI) of Steel Strand, 3. Test Method for Change in Corrosion Potential of Steel Strand, and 4. Test Method for Identification and Quantification of Residue on Steel Strand by Extraction, Gravimetric, and Spectroscopical Analyses. Testing was conducted to provide the basis for a precision statement accompanying the proposed test methods devel- oped for identifying strand bond performance. 108

To determine the precision (i.e., the repeatability) of the methods, the selected tests were repeated up to six times on samples of strand obtained from the same source. This testing was conducted on a single source identified as a middle range performer in that particular test during the correlation testing. The results of this testing are presented below according to ASTM practice. A determination of bias in the testing methods is not possible at this time since a known reference sample can- not be selected in a universally acceptable manner. The sources of strand that were used for this testing are given in Table B-32. These sources were chosen because they produced middle-of-the-range results obtained dur- ing the original testing programs. The results for the four test methods are given in Table B-33 through Table B-36. The precision and bias statements to be added to the stan- dard test methods are in the following discussion. Test Method for Weight Loss on Ignition (LOI) of Strand Single-operator precision—The single-operator standard deviation was found to be 0.014 mg/cm2∗. Therefore, results of two properly conducted tests by the same operator on the same source are not expected to differ from each other by more than 0.041 mg/cm2∗. (Numbers followed by an asterisk in this and the following sections represent, respectively, the (1s) and (d2s) limits as described in ASTM C670, Standard Practice for Preparing Precision and Bias Statements for Test Methods for Construction Materials [ASTM 2003].) Bias—Since there is no accepted reference material suitable for determining the bias in this test method, no statement on bias is made. Test Method for Contact Angle Measurement of a Water Droplet on a Strand Surface Single-operator precision—The single-operator standard deviation was found to be 4°F∗. Therefore, results of two prop- erly conducted tests by the same operator on the same source are not expected to differ from each other by more than 10°F∗. Bias—Since there is no accepted reference material suit- able for determining the bias in this test method, no state- ment on bias is made. Test Method for Change in Corrosion Potential of Strand Single-operator precision—The single-operator standard deviation was found to be 0.047 V∗. Therefore, results of two 109 Test Method Strand Source Organic Residue Extraction 102 Contact Angle 102 Change in Corrosion Potential 103 Weight Loss on Ignition 717 Sample Set (three pcs. of strand) Average Concentration (mg/cm2) 1 0.057 2 0.061 3 0.072 4 0.055 5 0.088 6 0.078 Average for 6 repeats 0.069 Standard Deviation for 6 repeats 0.013 Sample Set (three pcs. of strand) Average Weight LOI (mg/cm2) 1 0.086 2 0.079 3 0.102 4 0.086 5 0.079 6 0.115 Average for 6 repeats 0.091 Standard Deviation for 6 Repeats 0.014 Table B-32. Strand source for precision testing. Table B-33. Precision test results for organic residue extraction. Table B-34. Precision test results for weight loss on ignition. Sample Set (three pcs. of strand) Average Contact Angle after Lime Dip (°) 1 68 2 75 3 71 4 71 5 77 6 73 Average for 6 repeats 73 Standard Deviation for 6 Repeats 4 Table B-35. Precision test results for contact angle.

properly conducted tests by the same operator on the same source are not expected to differ from each other by more than 0.133 V∗. Bias—Since there is no accepted reference material suit- able for determining the bias in this test method, no state- ment on bias is made. Test Method for Identification and Quantification of Strand Surface Residue by Extraction, Gravimetric, and Spectroscopical Analyses Single-operator precision—The single-operator standard deviation was found to be 0.013 mg/cm2∗. Therefore, results of two properly conducted tests by the same operator on the same source are not expected to differ from each other by more than 0.037 mg/cm2∗. Bias—Since there is no accepted reference material suitable for determining the bias in this test method, no statement on bias is made. In Table B-37, the results of the precision testing program are compared to those obtained during the screening and cor- relation testing program. Two tests, namely the Test Method for Change in Corrosion Potential of Strand and the Test Method for Contact Angle Measurement of a Water Droplet on a Strand Surface, were conducted on the same sources of strand that were tested as part of the Screening Round of evaluation in 2004. The results of the Precision testing program obtained in 2007 gave average results that differed from the earlier result by an amount greater than the acceptable single-operator ranges listed in the precision statements above. This is explained by the fact that these methods are heavily dependent on the physical condition of the strand surface, and the surface of these strand samples had undergone some corrosion. The results of the other two test methods are less sensitive to such corrosion and gave results that were more consistent with previous testing efforts. Development of Thresholds For the recommended surface and chemical test methods to be useful in a QC setting, thresholds for acceptable bond behavior are needed. The usefulness of acceptance/rejection thresholds for the surface and chemical test results is depend- ent on the precise correlation of these results with minimum acceptable bond strengths established by physical test methods. The validity of these thresholds is also dependent on the valid- ity of the physical test methods (such as pull-out tests) used to measure bond performance. At the direction of the NCHRP supervisory panel, the transfer length testing originally planned for this test program as a basis for developing thresholds for the surface and chemical test results was not conducted. Instead, the basis available for developing thresholds for the chemical and surface test methods are the acceptance limits proposed by Russell and adopted by NASPA. The bond strength thresholds proposed by Russell are stated in terms of the force at 0.1-in. slip measured by the NASPA mortar pull-out test procedure. They are based on a set of 110 Test Method Strand Source Screening or Correlation Testing Result Precision Testing Result* Reported Single Operator Precision (d2s) Organic Residue Extraction (mg/cm2) 102 0.069 0.069 0.037 Contact Angle (°) 102 87 73 10 Change in Corrosion Potential (V) 103 -0.167 -0.334 0.133 Weight Loss on Ignition (mg/cm2) 717 0.086 0.091 0.014 *Average of all Precision testing results. Table B-37. Results of precision testing compared to correlation testing. Sample Set (three pcs. of strand) Change in Corrosion Potential (V) After 6 h 1 -0.402 2 -0.358 3 -0.317 4 -0.280 5 -0.313 Average for 5 repeats -0.334 Standard Deviation for 5 Repeats 0.047 Table B-36. Precision test results for change in corrosion potential.

development length tests conducted in parallel with the de- velopment of the NASPA Strand Bond Test (Russell 2001, Russell 2006). The thresholds were derived using development length tests on four strand sources (referred to as the NASPA Round III study [Russell 2001]), and they are defined in terms of acceptance criteria for the average force at 0.1-in. slip from six pull outs with a lower criterion for any single measurement of the six pull outs. The Round III report proposed thresholds of 7300 and 5500 lbs, for the minimum permissible average and single test result, respectively for 1/2-in. diameter strand (Russell 2001). These minimum thresholds have since been increased to 10500 and 9000 lbs, but without additional test- ing (Russell 2006). For 0.6-in. diameter strand, the suggested thresholds are 12600 and 10800 lbs for the minimum permis- sible average and single test result, respectively (Russell 2006). No threshold has been suggested for other sizes of strand. Despite the somewhat limited scope of the development process used to establish these NASPA test thresholds, the threshold determination effort for the surface and chemical testing conducted in this study was performed assuming that these thresholds were well-defined lower bounds. As has been done throughout this study, the thresholds were converted to bond stresses calculated as the force divided by the nominal surface area (the nominal perimeter of the strand multiplied by the embedment length) to support comparisons among all of the tested strands. When converted to a bond stress, the minimum threshold on the average of six tests of 10,500 lbs is equal to 0.313 ksi. This value was used as the basis for the threshold analysis. Note that since rigorous mechanical testing was not per- formed on many of the strand sources used to develop the correlations between performance and these chemical and surface tests, the scientific basis for such thresholds is less than would be desirable. As a result, conservative thresholds will be proposed. Regression and Prediction Intervals While some analysis has already been made using linear re- gression methods, a background on regression techniques is presented here as a basis for the discussion of prediction in- tervals, which are less commonly used, but which are needed for the development of thresholds. In this analysis, the pull- out performance is treated as the dependant variable and is plotted on the y-axis, while the variables (i.e., test method re- sults) on which the predictions are based are plotted on the x-axis. Regression with a Single Predictor Regression is a standard statistical technique for modeling the relationship between two or more variables. In the case of the mortar pull-out stress of the strand, the goal is to predict the mortar pull-out stress as a function of the proposed QC tests or chemical analyses. Since these tests can be done more eas- ily than the actual pull-out test, it would be helpful to predict the pull-out stress by running a few simple tests and then using a prediction function to estimate the pull-out stress for that strand. The model for the response (the pull-out stress in this case) consists of two major parts: the prediction formula and the error that captures the variability in the response (the pull- out stress in this case). The response is usually denoted by y and the predictor variable is usually denoted by x. The most common prediction formula is a simple linear model with an unknown slope, β1, and intercept, β0. The model for the re- sponse is then: (Eq. 3) where ε is the error term, which is typically assumed to be a normal distributed random variable with mean zero and an unknown variance, σ2. The idea is that the average response (e.g., pull-out stress) is linearly related to the value of the predictor variable (e.g., the change in corrosion potential as- received); however, due to random variation (for example, variation in the testing and measurement system), the actual pull-out stress for any given strand sample may be somewhat higher or lower than the predicted average pull-out stress based on that predictor (e.g., corrosion potential). Figure B-100 shows the example of mortar pull-out stress versus the change in corrosion potential as-received. In the plot, the linear trend is apparent, but individual strand sam- ples often deviate substantially from the line due to random error. The standard deviation of the observations from the line is estimated by the value S shown in the box in the upper right-hand corner of the plot. S2 is the estimator of the variance of ε, the normally distributed random variable used to model the random error. Although the true values of β0 and β1 are unknown, the regression provides the estimates that produce the smallest possible value of S, in this case these estimates are and where the “hat” notation indi- cates that these are estimators of β0 and β1. The other impor- tant measure provided in the plot, R2, is a measure of how much of the variability of the data is explained by the linear model. This value ranges between 0 and 1, where zero means that the slope is flat and thus no linear relationship exists be- tween x and y. An R2 of 100% indicates that the data all fall exactly on the single line generated by the regression. In this case, the R2 value is 68.2%, which indicates that there is clearly some relationship between slip stress and corrosion; however, there is also a substantial amount of error. Figure B-101 shows the standard regression output from Excel’s Data Analysis Tool Pack. The Regression Statistics βˆ1 = 1.743,βˆ0 = 0.7659 y x= + +β β ε0 1 111

Table provides Standard Error (S), R2, and R2 adjusted. The multiple R is just the square root of R2 and is sometimes called the correlation. The ANOVA (Analysis of Variance) analysis performed by Excel includes a statistical test to determine if the amount of variability explained by the fitted model is significantly more than would be expected from a model fitted to random data from a normal distribution. All the entries in the ANOVA table are intermediate values for calculating the “Significance F” value. The “Significance F” value shows how likely it would be that random, normally distributed data would fit a linear model as well as this model fits these data. Here, the proba- bility is very low (0.0061) indicating that it is very unlikely that these data are just random data and much more likely that the data are actually following the linear model. Finally, the bottom table provides the estimates for the coefficients ( and ). The P-values in the bottom table are the result of statistical tests that test to see if the true intercept, β0, and the true slope, β1, are equal to zero. The low P-values indicate that neither the slope nor the in- tercept is likely to be zero. For the intercept, this test is not very interesting, but if the slope is zero, that would indicate no linear relationship between the pull-out stress and the change in corrosion potential as-received. Notice that the P-value for the slope (in the row labeled “Change in Corr. Pot. As Received”) is exactly the same as the “Significance F” value above. For the case of a single predictor variable, these two tests are exactly equivalent. In the next section where multiple predictors are discussed, these tests will no longer be equivalent. βˆ1 = 1.7432βˆ0 = 0.7659 112 y = 1.7425x + 0.7659 R2 = 0.6823 0.000 0.100 0.200 0.300 0.400 0.500 0.600 0.700 -0.4 -0.3 -0.3 -0.2 -0.2 -0.1 -0.1 0.0 Change in Corrosion Potential (V) as-Received A ve ra ge M or ta r B on d St re ss a t 0 .1 -in . S lip (ks i) S=0.0971930 Figure B-100. Fitted line plot for mortar pull-out stress versus the change in corrosion potential as-received. SUMMARY OUTPUT Regression Statistics Multiple R 0.826022976 R Square 0.682313957 Adjusted R Square 0.636930237 Standard Error 0.097193005 Observations 9 ANOVA df SS MS F Significance F Regression 0.1420221 0.142022 15.03433 0.00607324 Residual 7 0.066125 0.009446 Total 8 0.208147 Coefficients Standard Error t Stat P-value Lower 95% Upper 95% Intercept 0.765888672 0.105371 7.268469 0.000167 0.516724919 1.015052 Change in Corr. Pot. As Received 1.74252891 0.449405 3.877413 0.006073 0.679854919 2.805203 Figure B-101. Excel output for regression with single predictor.

The last two entries in each row of the bottom table are 95% confidence intervals for the intercept and slope. These are ranges that 95% of the time will cover the true values β0 and β1. The most useful piece of information here is similar to the conclusions made earlier and that is that with 95% confi- dence β1, the slope is greater than zero—again indicating a linear relationship between pull-out stress and change in cor- rosion potential as-received. The main reason for developing this regression is to allow for prediction of the pull-out stress by measuring the change in corrosion potential. However, entering the measured change in corrosion potential into the prediction formula gives the average estimated pull-out stress and does not account for variation that is bound to occur in the test results or uncertainty in the regression model. This vari- ation is evidenced by the fact that all the points used to de- velop the regression did not fall on the line, that is the R2 value was not 100%. Instead, what is needed to interpret and prac- tically apply a change in corrosion potential test result is the computation of a lower bound on the interval that, with 90% confidence, includes the pull-out stress for a strand sample with that change in potential test result. This type of interval is known as a one-sided prediction interval and is a standard part of regression theory and practice. A two-sided prediction interval is used when both an upper and a lower bound are required. The one-sided prediction interval will be focused on here. The prediction interval concept is a necessary part of the development of acceptance/rejection thresholds for the rec- ommended QC tests, since, to conservatively verify that a ˆ ˆ ˆy x= +β β0 1 specified pull-out bond stress is likely to be achieved, the threshold on the QC test must be chosen as the value where the prediction interval lower bound is equal to the pull-out stress threshold. The model gives an estimate of the average pull-out stress if the pull-out test was actually conducted repeatedly on the same source of strand. For a given measurement of the predictor, half of the actual pull-out test results would be expected to fall above this average and half would fall below. The distribution of individual pull-out obser- vations about that average pull-out stress is the basis for the prediction interval, which is calculated based on the variabil- ity in the data used for the regression. This concept is demonstrated graphically in Figure B-102, which shows the prediction interval lower bound plotted along with the regression line, and data for the mortar pull out plotted versus the change in corrosion potential. If a spec- ified threshold on mortar pull out is defined as 0.313 ksi, the threshold on the corrosion potential is the value where the pull-out threshold and the curve delineating the lower bound of the prediction interval intersect, shown by the red lines in the plot. In this case, the threshold would be approximately −0.175 V. Unfortunately, Excel does not provide prediction intervals as a part of its standard output. However, a formula is provided in (Eq. 4) that allows for the calculation of a 90% prediction interval lower bound for pull-out stress of a new strand for the predictor. In the example, this is a prediction interval for pull- out stress with a measured value of change in corrosion po- tential as received. Of course, this prediction interval assumes that the model in (Eq. 3) is the correct model. ˆ ˆ ˆy x= +β β0 1 113 Change in Corr. Pot. (V) y = 1.7413x + 0.7656 R2 = 0.6818 0.000 0.100 0.200 0.300 0.400 0.500 0.600 0.700 -0.35 -0.3 -0.25 -0.2 -0.15 -0.1 -0.05 0 M or ta r P ul l-o ut 0 .1 -in S lip S tre ss (k si) Change in Corr. Pot. (V) Prediction Interval Lower Bound Threshold on Change in Corr. Pot. (V) Specified Pull- Out Stress Threshold Change in Corr Pot. Where Pull-Out Threshold Intersects Prediction Interval Figure B-102. Threshold determination using the prediction interval.

(Eq. 4) Each of the variables in (Eq. 4) is explained in the following table. To help with the calculations below, the vector contains Prediction Interval (x x t n p 0 0 1 0 90 ) ˆ ˆ . ,= + − −β β S n x x Sxx 1 1 0 2+ + −( ) all the values of the predictor variable that were used in the re- gression (the prime indicates the transpose of the vector, so xD is defined as a column vector), x D = − − − − −( . , . , . , . , . ,0 289 0 080 0 154 0 241 0 272 . , . . , . ) − − − − ′ 0 211 0 267 0 172 0 322 . 114 Explanation of Variables in Eq. 4 Variable Description Numerical Value in This Example The estimate of the intercept. 0.7658 The estimate of the slope. 1.7425 S The standard deviation of the observations from the 0.097193 fitted line. x0 The value of x for which y is to be predicted. Any value of corrosion potential for which a prediction is needed. For example, x0 = −0.1. The average of all the x values used in the regression. Sxx This is a measure of the spread in the x values in the 0.04683 regression. It is defined as n The total number of observations in the regression. 9 p The total number of β’s in the model (β0, β1) 2 t.90,n−p The 90th percentile of the t-distribution with n-p 1.42 degrees of freedom. See the table below for approximate for n-p = 7 values. t-distribution for other confidence levels are available in most standard statistics textbooks and using Excel’s TINV() function. S nxxx = ′ −x xD D 2 x = −0 228.x βˆ1 βˆ0 n-p t.90,n−p 1 3.08 2 1.89 3 1.64 4 1.53 5 1.48 6–7 1.42 8–9 1.39 10–13 1.36 14–27 1.33 >27 1.30 Approximate Values of the 90th Percentile of the t-Distribution

Following (Eq. 4), the 90% prediction interval for a new strand sample that has a change in corrosion potential as-received of x0 = −0.1 V, would be This means that for a measured change in corrosion potential as-received of −0.100 V, the mortar pull-out stress at 0.1-in. slip would be expected, with 90% confidence, to be greater than 0.425 ksi. For the purpose of determining thresholds on other surface and chemical tests, the prediction interval over the range of the measured responses is needed. This is calculated by vary- ing x0 in Eq. 4. Regression with Multiple Predictors Often the response of interest depends on more than one predictor variable. In that case, additional terms are added to the regression model as shown in Eq. 5 for k predictors. (Eq. 5) When regression is performed with multiple predictors, the regression output is very similar to the regression output for a single predictor. Again, the R2 and R2 adjusted can be calculated and should be interpreted as measures of how well the model fits the data. The difference between the two statistics is that, since adding predictors makes the model more flexible and thus better able to fit the data, the R2 adjusted measure includes a penalty for additional predictors in the model. Thus, the goal of multiple regression is to find a model that has a relatively high R2 value with as few predictors as possible, and maximizing R2 adjusted accomplishes this goal. However, if too many predictors are put into the model, there will be very few degrees of freedom (n-p) for estimating the error variance (S). Again the goal is to have low values of the F-significance and low P-values for each of the slope estimates (i.e., β coefficients). An Excel regression output for a multiple regression is shown in Figure B-103, and it can be seen that this regression is a better model for the data than the single predictor regression shown above, as measured by R2 and R2 adjusted. However, two of the slope estimates (Weight LOI and Contact Angle After Dip) have P-values that are rel- atively large. Notice that the 95% confidence intervals of the slopes for these two factors contain zero; this indicates that the data are not conclusively supporting that these slopes are different than zero. Thus, the relationship between these y x x xk k= + + + ⋅ ⋅ ⋅ + +β β β β ε0 1 1 2 2 = + − − + + 0 7658 1 7425 0 1 1 42 0 097193 1 1 9 . . ( . ) . ( . ) ( . ( . )) . . − − − = 0 1 0 228 0 04683 0 425 2 measures and pull-out stress is not conclusive at the chosen confidence level. If this multiple regression model were to be used for pre- diction of the pull-out stress, a prediction interval is still needed to determine thresholds. Determining the predic- tion interval for models based on multiple predictors is possible; however, it is somewhat more complicated and involves the matrix manipulations demonstrated by the fol- lowing example. For this example, the data used to develop the model by regression are as follows: Change in Mortar Pull- Weight Contact Corrosion Out Stress Loss on Angle after Potential at 0.1-in. Ignition Lime Dip As-Received Slip (ksi) x1 x2 x3 y 0.139 87 −0.289 0.156 0.059 79 −0.08 0.623 0.036 68 −0.154 0.606 0.086 94 −0.241 0.206 0.041 73 −0.272 0.409 0.045 76 −0.211 0.409 0.051 87 −0.267 0.315 −0.021 79 −0.172 0.397 0.002 98 −0.322 0.273 Let the matrix X be defined as a column of ones (repre- senting the coefficient to be multiplied by the intercept, β0) and then a column of levels for each of the other predictors in each observation. The 90% prediction interval is then given by: (Eq. 6) where each of the variables in (Eq. 6) is explained in the fol- lowing table. Prediction Interval (x x xk1 2, , . . . , ) ˆ = +β0 ˆ ˆ ˆ ,. ,β β β1 1 2 2 90x x x t Sk k n p+ ⋅ ⋅ ⋅ − ′ ′− −x X X x( ) 1 X = 1 1 0 139 0 059 87 79 0 289 0 080 1 0 036 68 0 15 . . . . . . − − − 4 1 0 086 94 0 241 1 1 1 1 1 0 041 0 045 0 051 0 021 0 . . . . . . − − . . . . . .002 73 76 87 79 98 0 272 0 211 0 267 0 172 0 3 − − − − − 22 ⎡ ⎣ ⎢⎢⎢⎢⎢⎢⎢⎢⎢ ⎤ ⎦ ⎥⎥⎥⎥⎥⎥⎥⎥⎥ 115

The 90% prediction interval lower bound for is 0.252 ksi. This means that for a strand source with measured LOI of 0.04 mg/cm2, a contact angle after lime dip of 78°, and a change in corrosion potential as-received of −0.250 V, the mortar pull-out stress at 0.1-in. slip would be expected, with 90% confidence, to be greater than 0.252 ksi. x 0 = − ⎛ ⎝ ⎜⎜⎜ ⎞ ⎠ ⎟⎟⎟ 1 0 04 78 0 250 . . Unfortunately, the prediction intervals based on regression with multiple predictors cannot be plotted in two dimensions as was done in Figure B-103, for the single predictor example. The predicted pull-out stress (y in Eq. 5) is not uniquely de- termined by a single combination of predictors (x1, x2, . . .), but can be found based on numerous combinations. However, the prediction interval for the pull-out stress will be different depending on the specific combination of predictors that is used. That means that when multiple regression is used to improve the predictive ability of the model, a single 116 Explanation of Variables in Eq. 6 Variable Description Numerical Value in This Example The estimate of the intercept. 0.7658 The estimate of the slope for predictor variable i. 1.7425 X A vector of levels for each predictor at which y is to For example, be predicted (including a 1 for the intercept). n The total number of observations in the regression. 9 k The number of predictors in the model. 3 p = k + 1 The total number of β’s in the model. 4 t.90,n−p The 90th percentile of the t-distribution with n-p 1.48 for n-p = 5 degrees of freedom. See the table above for approximate values. t-distribution for other confidence levels are available in most standard statistics textbooks and using Excel’s TINV() function. x 0 = − ⎛ ⎝ ⎜⎜⎜ ⎞ ⎠ ⎟⎟⎟ 1 0 04 78 0 250 . . βˆi βˆ0 SUMMARY OUTPUT Regression Statistics Multiple R 0.922308764 R Square 0.850653457 Adjusted R Square 0.761045531 Standard Error 0.078849246 Observations 9 ANOVA df SS MS F Significance F Regression 0.1770613 0.05902 9.49306 0.016594131 Residual 0.0310865 0.006217 Total 8 0.208147 Coefficients Standard Error t Stat P-value Lower 95% Upper 95% Intercept 1.202523363 0.243392 4.940688 0.00432 0.576864667 1.828182 Weight Loss on Ignition -0.84478002 0.617114 -1.368919 0.229318 -2.431123313 0.741563 Contact Angle After Lime Dip -0.00632965 0.003492 -1.812644 0.129631 -0.015305971 0.002647 Change in Corr. Pot. As Received 1.179495734 0.450947 2.6156 0.047349 0.020300812 2.338691 Figure B-103. Excel output for regression with multiple predictors.

threshold cannot be defined. Instead, for a specific set of pre- dictors, a new prediction interval must be calculated based on the set of data used to develop a regression model. The lower bound of the newly calculated prediction interval must then be compared with the specified pull-out threshold. For the example calculation performed above, the lower bound is 0.252 ksi. This is lower than the specified pull-out threshold of 0.313 ksi, so it cannot be predicted that this source of strand will exceed that pull-out stress 90% of the time (the de- fined confidence level). The framework for completing this cal- culation and comparison is given in the Microsoft Excel spreadsheet developed in this study. Selection of Confidence Level For the threshold determinations performed based on the data collected in this study, the confidence level was taken as 90%. This means that for a given surface and chemical test result, 10% of the pull-out results would be expected to fall below that prediction interval. This confidence level is lower than the 95% confidence interval that is most commonly used as the basis for probabilistic design in structural engineering analysis. However, using a confidence level as high as 95% makes determination of the thresholds for the surface and chemical tests very conservative. Thresholds Based on Regression with Single Predictor The test methods that were recommended for inclusion are: • Weight LOI, • Contact Angle Measurement after Lime Dip, • Change in Corrosion Potential, • Organic Residue Extraction. The efforts made to define thresholds for each of these methods were based on single variable linear regressions and are described individually below. The results are summarized in Table B-38. Weight Loss on Ignition (LOI)—The prediction interval for LOI with a one-sided confidence level of 90% is shown in Figure B-104. As can be seen in this figure, the prediction in- terval does not exceed 0.313 ksi anywhere over the range of test results observed in this study. For that reason, no threshold can be determined. Contact Angle Measurement after Lime Dip—The pre- diction interval for Contact Angle with a one-sided confidence level of 90% is shown in Figure B-105. As can be seen in this figure, this prediction interval exceeds 0.313 ksi when the 117 Predictor Constant ( x 0 ) Coefficient ( ) Coefficient of Determination ( R 2 ) Weight Loss on Ignition (m g/cm 2 ) 0.445 -1.403 0.16 Contact Angle after Li me Dip (°) 1.393 -0.012 0.57 Change in Corrosion Potential After 6 h(V)—As-Received 0.766 1.741 0.68 Extracted Organic Residue (m g/cm 2 ) 0.453 -1.752 0.12 Extracted Organic Residue (m g/cm 2 )—Stearate only 0.436 -1.943 0.63 Table B-38. Regression coefficients for single-predictor models. Loss on Ignition (mg/cm2) y = -1.4031x + 0.4454 R2 = 0.1595 -0.100 0.000 0.100 0.200 0.300 0.400 0.500 0.600 0.700 -0.04 -0.02 0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 0.16 M or ta r P ul l-O ut 0 .1 -in . S lip S tre ss (k si) Loss on Ignition (mg/cm2) Prediction Interval Lower Bound Threshold on Loss on Ignition (mg/cm2) Figure B-104. Prediction interval (confidence level  90%) for Weight Loss on Ignition. Threshold not possible.

contact angle is less than 73°. Therefore, based on these data and the NASPA defined threshold on mortar pull-out stress at 0.1-in. slip, a Contact Measurement after Lime Dip of 73° or lower is recommended to give a good (90%) confidence of adequate bond. This test must be run on recently manufac- tured strand with no surface weathering or rust (i.e., bright strand). Change in Corrosion Potential—The prediction interval for Change in Corrosion Potential with a one-sided confi- dence level of 90% is shown in Figure B-106. As can be seen in this figure, this prediction interval exceeds 0.313 ksi when the change in the corrosion potential is less negative than −0.175 V. Therefore, based on these data and the NASPA defined threshold on mortar pull out 0.1-in. slip stress, a Change in Corrosion Potential of −0.175 V or more (less negative) is recommended to give a good confidence of ade- quate bond. Organic Residue Extraction—The prediction interval for organic residue extraction with a one-sided confidence level of 90% is shown in Figure B-107. As can be seen in this figure, the prediction interval does not exceed 0.313 ksi anywhere over the range of test results observed in this study. For that reason, no threshold can be determined. A similar analysis was attempted considering only those sources with organic residue that the FTIR analyses in- dicated was primarily stearate. This was done to eliminate potentially confounding influences of non-stearate-based lubricants. The prediction interval for this stearate residue 118 Contact Angle after Lime (°) y = -0.0123x + 1.3929 R2 = 0.5701 -0.100 0.000 0.100 0.200 0.300 0.400 0.500 0.600 0.700 0 20 40 60 80 100 120 M or ta r P ul l-O ut 0 .1 -in . S lip S tre ss (k si) Contact Angle After Lime (°) Prediction Interval Lower Bound Threshold on Contact Angle After Lime (°) Figure B-105. Prediction interval (confidence level  90%) for Contact Angle after Lime Dip. Change in Corr. Pot. (V) y = 1.7413x + 0.7656 R 2 = 0.6818 0.000 0.100 0.200 0.300 0.400 0.500 0.600 0.700 -0.35 -0.3 -0.25 -0.2 -0.15 -0.1 -0.05 0 M or ta r P ul l-O ut 0 .1 -in . S lip S tre ss (k si) Change in Corr. Pot. (V) Prediction Interval Lower Bound Threshold on Change in Corr. Pot. (V) Figure B-106. Prediction interval (confidence level  90%) for Change in Corrosion Potential.

with a one-sided confidence level of 90% is shown in Fig- ure B-108. As can be seen in this figure, the Coefficient of Determination is higher, but the prediction interval still does not exceed 0.313 ksi anywhere over the range of test results observed in this study, and no threshold can be determined. Thresholds Based on Regression with Multiple Predictors An attempt was also made to look for linear combinations of multiple results obtained from all the evaluated methods that would correlate with bond performance. While numerous combinations were examined, the three combinations that showed the best correlation, based on the adjusted coefficient of determination (R2 adj.), were: • Contact Angle Measurement after Lime Dip & Change in Corrosion Potential, • Contact Angle Measurement after Lime Dip & Organic Residue Extraction (100% stearate only), • Weight Loss on Ignition (LOI) & Contact Angle Measure- ment after Lime Dip & Change in Corrosion Potential. Note that for multiple-predictor regression, a larger num- ber of variables will increase the R2. Therefore, the adjusted R2 statistic, which accounts for the number of degrees of freedom in the data set, was calculated as a means to compensate for 119 Organic Residue (mg/cm2) y = -1.7515x + 0.453 R2 = 0.1241 0.000 0.100 0.200 0.300 0.400 0.500 0.600 0.700 0 0.02 0.04 0.06 0.08 0.1 0.12 0.14 M or ta r P ul l-O ut 0 .1 -in . S lip S tre ss (k si) Organic Residue (mg/cm2) Prediction Interval Lower Bound Threshold on Organic Residue (mg/cm2) Extracted Organic Residue (mg/cm2) y = -1.9426x + 0.4355 R2 = 0.6282 0.000 0.050 0.100 0.150 0.200 0.250 0.300 0.350 0.400 0.450 0.000 0.020 0.040 0.060 0.080 0.100 0.120 0.140 M or ta r P ul l-O ut 0 .1 -in . S lip S tre ss (k si) Prediction Interval Lower Bound Threshold on Extracted Organic Residue (mg/cm2) Organic Residue (mg/cm2) Figure B-107. Prediction interval (confidence level  90%) for Organic Residue. Threshold not possible. Figure B-108. Prediction interval (confidence level  90%) for Organic Residue when FTIR analysis indicates organic residue is primarily stearate. Threshold not possible.

this potentially misleading effect. The regression coefficients for these three models and the R2 adj. are given in Table B-39 to Table B-41. The R2 adj. values for these combinations were high and equal to 0.73, 0.98 and 0.76, respectively. The R2 adj. for each of these combinations was higher than the R2 for the single-predictor regression models. The regression that indicated that the last combination of predictors listed above (Contact Angle Measurement after Lime Dip & Organic Residue Extraction) was a good predic- tor of bond was performed based only on those strand sources that the FTIR analysis of the organic residue identi- fied as being stearate only. This limited the number of data points used to develop the regression model to five, but was done as a means of eliminating potentially confounding in- fluences of non-stearate-based lubricants on the results obtained by the contact angle and organic residue extraction measurement methods. Given the high level of correlation with the multiple regression approach, this model may be particularly useful in a production setting where the lubricant in use is known. The prediction interval cannot be shown in a two- dimensional plot as was done with the single variable models. This is because there are multiple combinations of variables that can combine to give the same output. For this reason, a separate prediction interval must be calculated for each combination of variables. A Microsoft Excel-based spread- sheet has been developed for this purpose. To give a sense of how these multiple regression models might be used, tables have been prepared showing the predicted pull out, the lower bound on the prediction interval and the result of a compar- ison with the specified mortar pull-out threshold of 0.313 ksi, for each of the three multiple regression models. These are shown in Table B-42 to Table B-44. For example, Table B-42 was developed for the model based on Contact Angle Measurement after Lime Dip & Change in Corrosion Potential. The first row of this table shows the results of these two individual tests obtained for Source 349. Based on the regression model, the predicted mortar pull-out stress at 0.1-in. slip is 0.264 ksi for these two results. The lower bound on the prediction interval for that combination of the two test results must be calculated 120 Predictor Coefficient Constant 1.209 Contact Angle after Lime Dip (°) -0.007 Change in Corrosion Potential After 6 h (V)—As-Received 1.233 Adjusted Coefficient of Determination (R2 adj.) 0.727 Predictor Coefficient Constant 0.864 Contact Angle after Lime Dip (°) -0.006 Extracted Organic Residue (mg/cm2) -1.093 Adjusted Coefficient of Determination (R2 adj.) 0.976 Predictor Coefficient Constant 1.203 Weight Loss on Ignition (mg/cm2) -0.846 Contact Angle after Lime Dip (°) -0.006 Change in Corrosion Potential After 6 h (V)—As-Received 1.178 Adjusted Coefficient of Determination (R2 adj.) 0.761 Table B-39. Regression coefficients for model based on Contact Angle Measurement after Lime Dip & Change in Corrosion Potential. Table B-40. Regression coefficients for model based on Contact Angle Measurement after Lime Dip & Organic Residue Extraction (100% stearate only). Table B-41. Regression coefficients for model based on Weight Loss on Ignition (LOI), Contact Angle Measurement after Lime Dip & Change in Corrosion Potential. Mortar Pull Out 0.1-in Slip Stress (ksi) Strand Source ID Contact Angle after Lime Dip (°) Change in Corrosion Potential (V) Experimentally Determined in Pull-Out Test Value Predicted by Regression for QC Results Lower Bound of Prediction Interval Pass/Fail* Based on Prediction Interval from QC Tests Pass/Fail* Based on Pull-Out Test Result 349 87 -0.289 0.156 0.264 0.131 Fails Fails 548 79 -0.080 0.623 0.576 0.420 Passes Passes 697 68 -0.154 0.606 0.559 0.417 Passes Passes 717 94 -0.241 0.206 0.276 0.136 Fails Fails 478 73 -0.272 0.409 0.38 0.232 Fails Passes 960 76 -0.211 0.409 0.435 0.303 Fails Passes 102 87 -0.266 0.315 0.291 0.161 Fails Passes 103 79 -0.172 0.397 0.463 0.331 Passes Passes 151 98 -0.322 0.273 0.149 0.003 Fails Fails * Threshold for passing is 0.313 ksi. Table B-42. Evaluation of prediction interval for model based on Contact Angle Measurement after Lime Dip & Change in Corrosion Potential.

specifically using those values and is 0.131 ksi. Since 0.131 ksi is less than the mortar pull-out threshold of 0.313 ksi, this source fails to meet the minimum required combined Con- tact Angle Measurement after Lime Dip & Change in Corro- sion Potential performance. For Source 548, the prediction interval calculated for the specific combination of test results measured for that source is 0.420 ksi, and this source “passes” since this value exceeds the threshold. Summary An experimental program was conducted to evaluate a num- ber of test methods proposed for use as part of a QC program to evaluate bond of strand. This included limited mechanical testing (pull-out testing from concrete, Portland cement mor- tar, and gypsum plaster-based mortar) and extensive surface and chemical testing (Contact Angle, Examination Under UV light, pH, LOI, Loss in Alkali Bath, Change in Corrosion Potential, Corrosion Rate, Surface Roughness, Organic Residue Extraction/FTIR Analysis, and Elemental Analysis). These tests, as well as transfer length tests, have been conducted on a range of strand sources to help establish correlations between the proposed QC tests methods and bond quality. To evaluate these test methods, several rounds of evaluation were conducted: screening, correlation, and precision testing. The objective for the Screening Round was to eliminate those tests that do not show promise for predicting bond performance. The Correlation Round included those methods that showed promise in the screening experiments and was conducted to confirm that the selected QC tests correlated with bond performance over a larger sample set and were able to accurately identify good and bad strand. The Precision Testing was intended to form the base for precision statements to be included in the published test methods. The Screening Round was conducted using strand collected by the project team over the 5 years before this project began. Obtaining additional strand with a range of bonding qualities for the purposes of the Correlation Round directly from strand suppliers proved difficult. Sections of strand were provided by Bruce Russell of OSU, who had previously conducted mortar pull-out tests on these sources of strand in work performed for the NCHRP Project No. 12-60 Transfer, Development, 121 Mortar Pull Out 0.1-in Slip Stress (ksi) Strand Source ID Contact Angle after Lime Dip (°) Extracted Organic Residue (mg/cm2) Experimentally Determined in Pull-Out Test Value Predicted by Regression for QC Results Lower Bound of Prediction Interval Pass/Fail* Based on Prediction Interval from QC Tests Pass/Fail* Based on Pull Out Test Result 717 94 0.117 0.206 0.211 0.176 Fails Fails 478 73 0.033 0.409 0.42 0.388 Passes Passes 960 76 0.035 0.409 0.401 0.371 Passes Passes 102 87 0.069 0.315 0.303 0.274 Fails Passes 151 98 0.037 0.273 0.276 0.24 Fails Fails * Threshold for passing is 0.313 ksi. Mortar Pull Out 0.1-in Slip Stress (ksi) Strand Source ID Weight Loss on Ignition (mg/cm2) Contact Angle after Lime Dip (°) Change in Corrosion Potential (V) Experimentally Determined in Pull-Out Test Value Predicted by Regression for QC Results Lower Bound of Prediction Interval Pass/Fail* Based on Prediction Interval from QC Tests Pass/Fail* Based on Pull Out Test Result 349 0.139 87 -0.289 0.156 0.193 0.044 Fails Fails 548 0.059 79 -0.08 0.623 0.558 0.407 Passes Passes 697 0.036 68 -0.154 0.606 0.56 0.424 Passes Passes 717 0.086 94 -0.241 0.206 0.25 0.114 Fails Fails 478 0.041 73 -0.272 0.409 0.385 0.243 Fails Passes 960 0.045 76 -0.211 0.409 0.435 0.309 Fails Passes 102 0.051 87 -0.267 0.315 0.294 0.169 Fails Passes 103 -0.021 79 -0.172 0.397 0.517 0.378 Passes Passes 151 0.002 98 -0.322 0.273 0.2 0.049 Fails Fails * Threshold for passing is 0.313 ksi. Table B-43. Evaluation of prediction interval for model based on Contact Angle Measurement after Lime Dip & Organic Residue Extraction (100% stearate only). Table B-44. Evaluation of prediction interval for model based on weight loss on ignition, Contact Angle Measurement after Lime Dip & Change in Corrosion Potential.

and Splice Length for Strand/Reinforcement in High-Strength Concrete; the Oklahoma Department of Transportation; and NASPA, also known as the Committee of the American Wire Products Association [AWPA]). Although pull-out testing from concrete appears to corre- late best with transfer length, the most reliable and realistic measure of bond performance, the Correlation Round of this test program had to be based on available mortar pull-out results provided by Russell from the NCHRP 12-60 program. The four test methods that showed the best correlation with bond and that are recommended for inclusion in future QC programs are: 1. Weight Loss on Ignition (LOI) of Strand (QC-I), 2. Contact Angle Measurement of a Water Droplet on a Strand Surface (QC-I), 3. Change in Corrosion Potential of Strand (QC-I), and 4. Organic Residue Extraction with FTIR Analysis (QC-II). The QC tests have been divided into two categories, de- pending on the complexity and time required to conduct the tests: Level I (QC-I) and Level II (QC-II) tests. The QC level is shown in the list above. Thresholds for these QC tests have been developed where possible based on prediction intervals for the regression calculated from the available data and a minimum criterion on the mortar pull-out stress adopted by NASPA. Regression with multiple predictors has also been performed to see if results of these methods can be combined to better predict bond. The three combinations that showed the best correlation, based on the adjusted coefficient of determination (R2 adj.), were: 1. Weight Loss on Ignition (LOI) & Contact Angle Measure- ment after Lime & Change in Corrosion Potential, 2. Contact Angle Measurement after Lime & Change in Cor- rosion Potential, and 3. Contact Angle Measurement After Lime & Organic Residue Extraction (100% stearate only). The adjusted coefficients of determination for each of these combinations were higher than the coefficients of determina- tion for the single-predictor regression models. Thresholds for multiple-predictor regressions cannot be determined using the same procedure used for single-predictor regressions. Instead, the lower bound on the prediction interval must be calculated for each combination of test results. A computational tool in the form of a Microsoft Excel spreadsheet has been developed for this purpose. References Chandran, K. (2006). Assessing the Bond Quality of Prestressing Strands Using NASP Bond Test, Master of Science Thesis, Oklahoma State University, Stillwater, OK. Hyett, A.J., Dube, S., and Bawden, W.F. (1994, November). “Laboratory Bond Strength Testing of 0.6" 7-Wire Strand from 7 Different Manufacturers.” Final Report. Department of Mining Engineering, Queen’s University, Kingston, Ontario. Lane, S.N. (1998, December). “A New Development Length Equation for Pretensioned Strands in Bridge Beams and Piles,” Report No. FHWA- RD-98-116, Federal Highway Administration, Structures Div., McLean, VA. Logan, D.R. (1997 March-April). “Acceptance Criteria for Bond Qual- ity of Strand for Pretensioned Prestressed Concrete Applications,” PCI Journal, pp. 52–90. Mitchell. D., Cook. W. D., Khan. A. A., and Tham, T. (1993, May-June). “Influence of High Strength Concrete on Transfer and Develop- ment Length of Pretensioned Strand,” PCI Journal, Vol. 38. No. 3., pp. 52–66. Moustafa, S. (1974). “Pull-Out Strength of Strand Lifting Loops.” Tech- nical Bulletin 74-B5, Tacoma, WA: Concrete Technology Associates. Perenchio, W.F., Fraczek, J., and Pfeiffer, D.W. (1989). Corrosion Protection of Prestressing Systems in Concrete Bridges. NCHRP Report No. 313, Transportation Research Board, National Research Council, Washington, D.C. Peterman, R.J. (2007, May-June). “The Effects of As-Cast Depth and Concrete Fluidity on Strand Bond,” PCI Journal, pp. 72–101. Post-Tensioning Institute. (1996). Recommendations for Prestressed Rock and Soil Anchors, 3rd Ed., Phoenix, AZ. Rose, D.R., and Russell, B.W. (1997, July/August). “Investigation of Stan- dardized Tests to Measure the Bond Performance of Prestressing Strand,” PCI Journal, Vol. 42, No. 4, pp. 56–80. Russell, B.W., and Paulsgrove, G.A. (1999). “NASP Strand Bond Testing Round Two—Assessing Repeatability and Reproducibility of the Moustafa Test, the PTI Bond Test and the NASP Bond Test.” Final Report 99-04. The University of Oklahoma, Fears Structural Engi- neering Laboratory, Norman, OK. Russell, B.W. (2001). “Final Report—NASP Round III Strand Bond Testing,” Okalahoma State University, Stillwater, OK. Russell, B.W. (2006, June). “Final Report—NASP Round IV Strand Bond Testing” Okalahoma State University, Stillwater, OK. Stocker, M.F., and Sozen, M.A. (1971). “Investigation of Prestressed Reinforced Concrete for Highway Bridges, Part V: Bond Character- istics of Prestressing Strand,” Engineering Experiment Station 503, The University of Illinois, Urbana-Champaign, IL. Tabatabai, H., and Dickson, T.J. (1993, Nov-Dec). “The History of the Pretensioned Strand Development Length Equation,” PCI Journal, Vol. 38, No. 6, pp. 64–75. 122

Next: Appendix C - Specifications for Standard Surface Test Methods »
Acceptance Tests for Surface Characteristics of Steel Strands in Prestressed Concrete Get This Book
×
MyNAP members save 10% online.
Login or Register to save!
Download Free PDF

TRB's National Cooperative Highway Research Program (NCHRP) Report 621: Acceptance Tests for Surface Characteristics of Steel Strands in Prestressed Concrete explores tests to identify and measure residues on the surface of steel pre-stressing strands and to establish thresholds for residue types found to affect the strength of the strand's bond to concrete.

  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!