National Academies Press: OpenBook
« Previous: Chapter 2 - State of the Art and Current Practice
Page 17
Suggested Citation:"Chapter 3 - Laboratory Testing Program." National Academies of Sciences, Engineering, and Medicine. 2008. Development of a Recommended Practice for Use of Controlled Low-Strength Material in Highway Construction. Washington, DC: The National Academies Press. doi: 10.17226/13900.
×
Page 17
Page 18
Suggested Citation:"Chapter 3 - Laboratory Testing Program." National Academies of Sciences, Engineering, and Medicine. 2008. Development of a Recommended Practice for Use of Controlled Low-Strength Material in Highway Construction. Washington, DC: The National Academies Press. doi: 10.17226/13900.
×
Page 18
Page 19
Suggested Citation:"Chapter 3 - Laboratory Testing Program." National Academies of Sciences, Engineering, and Medicine. 2008. Development of a Recommended Practice for Use of Controlled Low-Strength Material in Highway Construction. Washington, DC: The National Academies Press. doi: 10.17226/13900.
×
Page 19
Page 20
Suggested Citation:"Chapter 3 - Laboratory Testing Program." National Academies of Sciences, Engineering, and Medicine. 2008. Development of a Recommended Practice for Use of Controlled Low-Strength Material in Highway Construction. Washington, DC: The National Academies Press. doi: 10.17226/13900.
×
Page 20
Page 21
Suggested Citation:"Chapter 3 - Laboratory Testing Program." National Academies of Sciences, Engineering, and Medicine. 2008. Development of a Recommended Practice for Use of Controlled Low-Strength Material in Highway Construction. Washington, DC: The National Academies Press. doi: 10.17226/13900.
×
Page 21
Page 22
Suggested Citation:"Chapter 3 - Laboratory Testing Program." National Academies of Sciences, Engineering, and Medicine. 2008. Development of a Recommended Practice for Use of Controlled Low-Strength Material in Highway Construction. Washington, DC: The National Academies Press. doi: 10.17226/13900.
×
Page 22
Page 23
Suggested Citation:"Chapter 3 - Laboratory Testing Program." National Academies of Sciences, Engineering, and Medicine. 2008. Development of a Recommended Practice for Use of Controlled Low-Strength Material in Highway Construction. Washington, DC: The National Academies Press. doi: 10.17226/13900.
×
Page 23
Page 24
Suggested Citation:"Chapter 3 - Laboratory Testing Program." National Academies of Sciences, Engineering, and Medicine. 2008. Development of a Recommended Practice for Use of Controlled Low-Strength Material in Highway Construction. Washington, DC: The National Academies Press. doi: 10.17226/13900.
×
Page 24
Page 25
Suggested Citation:"Chapter 3 - Laboratory Testing Program." National Academies of Sciences, Engineering, and Medicine. 2008. Development of a Recommended Practice for Use of Controlled Low-Strength Material in Highway Construction. Washington, DC: The National Academies Press. doi: 10.17226/13900.
×
Page 25
Page 26
Suggested Citation:"Chapter 3 - Laboratory Testing Program." National Academies of Sciences, Engineering, and Medicine. 2008. Development of a Recommended Practice for Use of Controlled Low-Strength Material in Highway Construction. Washington, DC: The National Academies Press. doi: 10.17226/13900.
×
Page 26
Page 27
Suggested Citation:"Chapter 3 - Laboratory Testing Program." National Academies of Sciences, Engineering, and Medicine. 2008. Development of a Recommended Practice for Use of Controlled Low-Strength Material in Highway Construction. Washington, DC: The National Academies Press. doi: 10.17226/13900.
×
Page 27
Page 28
Suggested Citation:"Chapter 3 - Laboratory Testing Program." National Academies of Sciences, Engineering, and Medicine. 2008. Development of a Recommended Practice for Use of Controlled Low-Strength Material in Highway Construction. Washington, DC: The National Academies Press. doi: 10.17226/13900.
×
Page 28
Page 29
Suggested Citation:"Chapter 3 - Laboratory Testing Program." National Academies of Sciences, Engineering, and Medicine. 2008. Development of a Recommended Practice for Use of Controlled Low-Strength Material in Highway Construction. Washington, DC: The National Academies Press. doi: 10.17226/13900.
×
Page 29
Page 30
Suggested Citation:"Chapter 3 - Laboratory Testing Program." National Academies of Sciences, Engineering, and Medicine. 2008. Development of a Recommended Practice for Use of Controlled Low-Strength Material in Highway Construction. Washington, DC: The National Academies Press. doi: 10.17226/13900.
×
Page 30
Page 31
Suggested Citation:"Chapter 3 - Laboratory Testing Program." National Academies of Sciences, Engineering, and Medicine. 2008. Development of a Recommended Practice for Use of Controlled Low-Strength Material in Highway Construction. Washington, DC: The National Academies Press. doi: 10.17226/13900.
×
Page 31
Page 32
Suggested Citation:"Chapter 3 - Laboratory Testing Program." National Academies of Sciences, Engineering, and Medicine. 2008. Development of a Recommended Practice for Use of Controlled Low-Strength Material in Highway Construction. Washington, DC: The National Academies Press. doi: 10.17226/13900.
×
Page 32
Page 33
Suggested Citation:"Chapter 3 - Laboratory Testing Program." National Academies of Sciences, Engineering, and Medicine. 2008. Development of a Recommended Practice for Use of Controlled Low-Strength Material in Highway Construction. Washington, DC: The National Academies Press. doi: 10.17226/13900.
×
Page 33
Page 34
Suggested Citation:"Chapter 3 - Laboratory Testing Program." National Academies of Sciences, Engineering, and Medicine. 2008. Development of a Recommended Practice for Use of Controlled Low-Strength Material in Highway Construction. Washington, DC: The National Academies Press. doi: 10.17226/13900.
×
Page 34
Page 35
Suggested Citation:"Chapter 3 - Laboratory Testing Program." National Academies of Sciences, Engineering, and Medicine. 2008. Development of a Recommended Practice for Use of Controlled Low-Strength Material in Highway Construction. Washington, DC: The National Academies Press. doi: 10.17226/13900.
×
Page 35
Page 36
Suggested Citation:"Chapter 3 - Laboratory Testing Program." National Academies of Sciences, Engineering, and Medicine. 2008. Development of a Recommended Practice for Use of Controlled Low-Strength Material in Highway Construction. Washington, DC: The National Academies Press. doi: 10.17226/13900.
×
Page 36
Page 37
Suggested Citation:"Chapter 3 - Laboratory Testing Program." National Academies of Sciences, Engineering, and Medicine. 2008. Development of a Recommended Practice for Use of Controlled Low-Strength Material in Highway Construction. Washington, DC: The National Academies Press. doi: 10.17226/13900.
×
Page 37
Page 38
Suggested Citation:"Chapter 3 - Laboratory Testing Program." National Academies of Sciences, Engineering, and Medicine. 2008. Development of a Recommended Practice for Use of Controlled Low-Strength Material in Highway Construction. Washington, DC: The National Academies Press. doi: 10.17226/13900.
×
Page 38
Page 39
Suggested Citation:"Chapter 3 - Laboratory Testing Program." National Academies of Sciences, Engineering, and Medicine. 2008. Development of a Recommended Practice for Use of Controlled Low-Strength Material in Highway Construction. Washington, DC: The National Academies Press. doi: 10.17226/13900.
×
Page 39
Page 40
Suggested Citation:"Chapter 3 - Laboratory Testing Program." National Academies of Sciences, Engineering, and Medicine. 2008. Development of a Recommended Practice for Use of Controlled Low-Strength Material in Highway Construction. Washington, DC: The National Academies Press. doi: 10.17226/13900.
×
Page 40
Page 41
Suggested Citation:"Chapter 3 - Laboratory Testing Program." National Academies of Sciences, Engineering, and Medicine. 2008. Development of a Recommended Practice for Use of Controlled Low-Strength Material in Highway Construction. Washington, DC: The National Academies Press. doi: 10.17226/13900.
×
Page 41
Page 42
Suggested Citation:"Chapter 3 - Laboratory Testing Program." National Academies of Sciences, Engineering, and Medicine. 2008. Development of a Recommended Practice for Use of Controlled Low-Strength Material in Highway Construction. Washington, DC: The National Academies Press. doi: 10.17226/13900.
×
Page 42
Page 43
Suggested Citation:"Chapter 3 - Laboratory Testing Program." National Academies of Sciences, Engineering, and Medicine. 2008. Development of a Recommended Practice for Use of Controlled Low-Strength Material in Highway Construction. Washington, DC: The National Academies Press. doi: 10.17226/13900.
×
Page 43
Page 44
Suggested Citation:"Chapter 3 - Laboratory Testing Program." National Academies of Sciences, Engineering, and Medicine. 2008. Development of a Recommended Practice for Use of Controlled Low-Strength Material in Highway Construction. Washington, DC: The National Academies Press. doi: 10.17226/13900.
×
Page 44
Page 45
Suggested Citation:"Chapter 3 - Laboratory Testing Program." National Academies of Sciences, Engineering, and Medicine. 2008. Development of a Recommended Practice for Use of Controlled Low-Strength Material in Highway Construction. Washington, DC: The National Academies Press. doi: 10.17226/13900.
×
Page 45
Page 46
Suggested Citation:"Chapter 3 - Laboratory Testing Program." National Academies of Sciences, Engineering, and Medicine. 2008. Development of a Recommended Practice for Use of Controlled Low-Strength Material in Highway Construction. Washington, DC: The National Academies Press. doi: 10.17226/13900.
×
Page 46
Page 47
Suggested Citation:"Chapter 3 - Laboratory Testing Program." National Academies of Sciences, Engineering, and Medicine. 2008. Development of a Recommended Practice for Use of Controlled Low-Strength Material in Highway Construction. Washington, DC: The National Academies Press. doi: 10.17226/13900.
×
Page 47
Page 48
Suggested Citation:"Chapter 3 - Laboratory Testing Program." National Academies of Sciences, Engineering, and Medicine. 2008. Development of a Recommended Practice for Use of Controlled Low-Strength Material in Highway Construction. Washington, DC: The National Academies Press. doi: 10.17226/13900.
×
Page 48
Page 49
Suggested Citation:"Chapter 3 - Laboratory Testing Program." National Academies of Sciences, Engineering, and Medicine. 2008. Development of a Recommended Practice for Use of Controlled Low-Strength Material in Highway Construction. Washington, DC: The National Academies Press. doi: 10.17226/13900.
×
Page 49
Page 50
Suggested Citation:"Chapter 3 - Laboratory Testing Program." National Academies of Sciences, Engineering, and Medicine. 2008. Development of a Recommended Practice for Use of Controlled Low-Strength Material in Highway Construction. Washington, DC: The National Academies Press. doi: 10.17226/13900.
×
Page 50
Page 51
Suggested Citation:"Chapter 3 - Laboratory Testing Program." National Academies of Sciences, Engineering, and Medicine. 2008. Development of a Recommended Practice for Use of Controlled Low-Strength Material in Highway Construction. Washington, DC: The National Academies Press. doi: 10.17226/13900.
×
Page 51
Page 52
Suggested Citation:"Chapter 3 - Laboratory Testing Program." National Academies of Sciences, Engineering, and Medicine. 2008. Development of a Recommended Practice for Use of Controlled Low-Strength Material in Highway Construction. Washington, DC: The National Academies Press. doi: 10.17226/13900.
×
Page 52
Page 53
Suggested Citation:"Chapter 3 - Laboratory Testing Program." National Academies of Sciences, Engineering, and Medicine. 2008. Development of a Recommended Practice for Use of Controlled Low-Strength Material in Highway Construction. Washington, DC: The National Academies Press. doi: 10.17226/13900.
×
Page 53
Page 54
Suggested Citation:"Chapter 3 - Laboratory Testing Program." National Academies of Sciences, Engineering, and Medicine. 2008. Development of a Recommended Practice for Use of Controlled Low-Strength Material in Highway Construction. Washington, DC: The National Academies Press. doi: 10.17226/13900.
×
Page 54
Page 55
Suggested Citation:"Chapter 3 - Laboratory Testing Program." National Academies of Sciences, Engineering, and Medicine. 2008. Development of a Recommended Practice for Use of Controlled Low-Strength Material in Highway Construction. Washington, DC: The National Academies Press. doi: 10.17226/13900.
×
Page 55
Page 56
Suggested Citation:"Chapter 3 - Laboratory Testing Program." National Academies of Sciences, Engineering, and Medicine. 2008. Development of a Recommended Practice for Use of Controlled Low-Strength Material in Highway Construction. Washington, DC: The National Academies Press. doi: 10.17226/13900.
×
Page 56
Page 57
Suggested Citation:"Chapter 3 - Laboratory Testing Program." National Academies of Sciences, Engineering, and Medicine. 2008. Development of a Recommended Practice for Use of Controlled Low-Strength Material in Highway Construction. Washington, DC: The National Academies Press. doi: 10.17226/13900.
×
Page 57
Page 58
Suggested Citation:"Chapter 3 - Laboratory Testing Program." National Academies of Sciences, Engineering, and Medicine. 2008. Development of a Recommended Practice for Use of Controlled Low-Strength Material in Highway Construction. Washington, DC: The National Academies Press. doi: 10.17226/13900.
×
Page 58
Page 59
Suggested Citation:"Chapter 3 - Laboratory Testing Program." National Academies of Sciences, Engineering, and Medicine. 2008. Development of a Recommended Practice for Use of Controlled Low-Strength Material in Highway Construction. Washington, DC: The National Academies Press. doi: 10.17226/13900.
×
Page 59
Page 60
Suggested Citation:"Chapter 3 - Laboratory Testing Program." National Academies of Sciences, Engineering, and Medicine. 2008. Development of a Recommended Practice for Use of Controlled Low-Strength Material in Highway Construction. Washington, DC: The National Academies Press. doi: 10.17226/13900.
×
Page 60

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.

17 Introduction The key objectives of the laboratory component of this project were to identify the most important CLSM properties affecting performance in the four target appli- cations (backfill, utility bedding, void fill, and bridge approaches), to develop and recommend a suite of test methods to assess these properties, and to understand what CLSM characteristics (e.g., materials, mixture proportions) most impact performance. This chapter summarizes the key findings of the laboratory study performed under NCHRP Project 24-12(01) and is aimed at meeting the above objectives. Information on the research approach, constituent mate- rials, mixture proportions, and test methods are described in this chapter. A more comprehensive summary of this lab- oratory testing was provided in the NCHRP Project 24- 12(01) Interim Report (Folliard et al. 2001). In addition, more detailed information on the corrosion testing and service life estimation models is provided in Appendix A. The main findings of this laboratory component, coupled with the field testing program (Chapter 4), led to the devel- opment of appropriate test methods (Appendix B), recom- mended specifications (Appendix C), and recommended practices (Appendix D). Research Approach As a precursor to the laboratory program, the important (or potentially important) CLSM properties were identified that may impact performance in the four target applications. These properties are identified in Table 3.1. Based on the application-specific properties listed in Table 3.1 and combined with a synthesis of available literature, a gen- eral classification of CLSM properties was developed, whereby the various CLSM properties of interest were grouped into three categories (important, potentially important, and less important): I. Important CLSM properties • Flow • Setting time • Unconfined compressive strength • Corrosion II. Potentially important CLSM properties • Excavatability • Subsidence • Freezing and thawing • Segregation and bleeding • Triaxial shear • CBR • Resilient modulus • Water permeability • Drying shrinkage • Leaching/environmental impact III. Less important CLSM properties • Direct shear strength • Air/gas permeability • Consolidation • Thermal conductivity The general classification of the properties by relative im- portance, as shown above, was then used in developing the laboratory testing program described in this chapter, result- ing in significant efforts being placed on evaluating the “important” properties, less emphasis being placed on the “potentially important” properties and no laboratory testing centered on the “less important” properties. Information on specific materials and mixture proportions is provided next, followed by discussion on the overall testing matrix, which was developed using the classification of the CLSM properties by relative importance. Materials A range of materials, summarized in Table 3.2, was selected for inclusion in the laboratory study to ensure widespread applicability of test results. General information about the C H A P T E R 3 Laboratory Testing Program

18 portland cement and the three fly ashes (Class F, Class C, and high carbon) used is provided in this table, and more specific information about the chemical and physical prop- erties of these materials can be found in the NCHRP Project 24-12(01) Interim Report (Folliard et al. 2001). Three types of fine aggregates were used throughout this project: con- crete sand conforming to ASTM C 33, foundry sand espe- cially blended for CLSM, and bottom ash passing a No. 4 (4.75 mm) sieve. Figure 3.1 compares the gradations of the three materials. The concrete sand meets the requirements of ASTM C 33 but approaches the coarse limit of the grada- tion band. The bottom ash was found to be slightly coarser and the foundry sand slightly finer than the ASTM C 33 gra- dation limits. Mixture Proportions Based on a survey of current practice (performed as part of the original NCHRP Project 24-12), the most common types of CLSM mixtures were selected for the laboratory study. These common mixture types were further delin- eated by defining a range of typical proportions (e.g., 30 to 60 kg/m3 of portland cement). For convenience, the mix- tures selected for the laboratory study can be classified as follows: • CLSM (with fine aggregates) – Type I portland cement: 1 type, 2 levels (30 kg/m3, 60 kg/m3) – Fly ash: 3 types, 3 levels (0 kg/m3, 180 kg/m3, 360 kg/m3) – Fine aggregate: 3 types, 1 level (1500 kg/m3) – Air content: 3 levels (entrapped air only, 15% to 20% air, 25% to 30% air) (Air-entraining agents were not used for CLSM containing fly ash) • CLSM (without fine aggregates) – Type I portland cement: 1 type, 1 level (60 kg/m3) – Fly ash: 3 types, 1 level (1200 kg/m3) – Air content: 1 level (entrapped air only) • CLSM (with set accelerator) – Selected mixtures from the test matrix CLSM Application Important Properties Potentially Important Properties Backfill Flow Compressive strength Excavatability Hardening time Settlement Corrosion of metal utilities Subsidence Freeze-thaw resistance Leaching and environmental impact Utility bedding Flow Compressive strength Hardening time Corrosion of metal utilities Freeze-thaw resistance Leaching and environmental impact Thermal conductivity Void fill Flow Subsidence Settlement Unconfined compressive strength Bridge approaches Flow Compressive strength Hardening time Shear strength Resilient modulus/CBR Settlement Freeze-thaw resistance Leaching and environmental impact Table 3.1. CLSM applications and relevant properties. Materiala Description Portland cement ASTM C 150 Type I (S.G.=3.15) Fly ash ASTM C 618 Class F (CaO=1.6%, LOI = 2.9%, S.G.=2.41) ASTM C 618 Class C (CaO=26.7%, LOI = 0.37%, S.G.=2.51) High-carbon fly ash (CaO=6.0%, LOI = 14.44%, S.G.=2.09) Fine aggregate ASTM C 33 concrete sand (S.G.=2.60, Absorption=1.0%, FM = 3.0) Foundry sand (ferrous) (S.G.=2.36, Absorption=5.6%, FM = 2.14, LOI=4.5%) Bottom ashb (S.G.= 2.28, Absorption = 8.9%, FM = 2.89) Chemical admixtures Air-entraining agent (liquid, designed specifically for CLSM) Accelerating admixture (non-chloride) aMore information on these materials can be found in NCHRP Project 24-12(01) Interim Report (Folliard et al. 2001). bBottom ash is classified as a fine aggregate because of similar particle size. Table 3.2. Materials included in the laboratory program.

For each of these mixtures defined above, the types and amounts of cement, fly ash, and aggregates were selected prior to mixing (as described later), and the water content of each mixture was then adjusted to achieve a flow of 200 to 250 mm, as measured by ASTM D 6103. The 38 mixtures included in the initial laboratory study were classified and tested accord- ing to their expected ability to provide information on the fol- lowing three groups of CLSM properties (based on expected level of importance): I. Important CLSM properties (flow, setting time, uncon- fined compressive strength, and corrosion) – Measured for all 38 mixtures in initially proposed Phase I study II. Potentially important CLSM properties (excavatability, subsidence, freezing and thawing, segregation and bleed- ing, triaxial shear, CBR, resilient modulus, water perme- ability, drying shrinkage) – Measured for selected mixtures only (6) – Only “order of magnitude” values sought III. Less important CLSM properties (direct shear strength, air/gas permeability, consolidation, thermal conductiv- ity, leaching/environmental impact) – Not included in laboratory study – Literature-based and existing-practice–based cover- age only After selecting representative materials and a range of mixture proportions, as previously defined, a statistical soft- ware program (ECHIP) was used to generate the majority of the mixtures within the test matrix. This software uses ex- perimental design concepts to produce statistically signifi- cant results with a minimal number of trials. In other words, rather than producing CLSM with every possible combina- tion of material and dosage, which would not be feasible, an optimized test matrix was produced that could be used to predict test results across the entire spectrum of variables. In addition, the program can be used to statistically compare the results of one test to another or the effects of individual or combined variables on test results. The program was also designed to assess the repeatability of test results by requir- ing duplication of certain mixtures within the test matrix. Initially, two separate mixture series were generated using the statistical software: one for non–air-entrained CLSM (with fly ash) and one for air-entrained CLSM (with- out fly ash). The non–air-entrained mixtures are shown in Table 3.3 (a mixture number followed by “r”, such as 1r, denotes a mixture repeated or duplicated for statistical purposes). The air-entrained mixtures originally proposed for study were selected using the statistical software, but after diffi- culties were encountered in generating entrained air in cer- tain mixtures, the decision was made to include mixtures covering all of the selected variables. That is, two cement con- tents (30 kg/m3 and 60 kg/m3), two target air contents (15 to 20 percent and 25 to 30 percent), and two aggregate types (concrete sand and bottom ash) were used in all combina- tions to create a total of eight mixtures. From these eight mix- tures, three were selected for replicate mixtures, bringing the total number of air-entrained mixtures to eleven, as shown in Table 3.4. The mixtures shown in Table 3.5 were strategically chosen to investigate specific mixture types of interest to the research team. The mixtures represented typical CLSM paste mixtures 19 100 80 60 Pe rc en t P as sin g Concrete Sand Bottom Ash Foundry Sand ASTM C33 Limits 40 20 0 0.1 1 Sieve Size (mm) 5 Figure 3.1. Gradations of aggregates used in study.

(i.e., 5 percent cement, 95 percent fly ash) and also included mixtures containing an accelerating admixture. Lastly, this table includes non–air-entrained CLSM mixtures containing foundry sand (selected after the difficulties encountered in entraining air in mixtures containing foundry sand). After casting and testing the initially proposed mixtures (as summarized in Tables 3.3 to 3.5), additional mixtures were cast to further investigate or refine selected test meth- ods or to study selected CLSM properties in more detail. The mixtures were based in most cases on previously cast mix- tures (from the original 38 mixtures), but there were other mixtures, such as rapid-setting CLSM containing only Class C fly ash as a binder, that were included to better reflect cur- rent practice in some parts of the country. Nine sets of addi- tional mixtures were cast and will be referred to throughout this report by as mixture series A through I, as summarized in Table 3.6. Because the compressive strength of CLSM is the most common property measured (and often the only 20 Mixture Cement Content (kg/m3) Fly Ash Typeb Fly Ash Content (kg/m3) Fine Aggregate Typec Water Demand (kg/m3) Flow (cm) Total Bleeding (%) Air Content (%) Fresh Density (kg/m3) 1 30 C 180 CS 211 20.0 NA 0.9 1965 2 60 C 180 CS 206 20.0 2.45 1.0 2108 1r 30 C 180 CS 206 21.0 2.08 0.9 1974 15 30 C 360 FS 486 20.0 0.13 2.8 1741 3 60 C 360 BA 577 17.8 4.32 1.7 1754 8 60 HC 180 FS 532 24.1 1.04 3.3 1647 10 30 HC 180 BA 628 14.0 4.81 2.0 1681 9 60 F 360 FS 520 20.0 0.54 2.5 1684 5 60 F 180 BA 600 17.8 5.84 2.5 1739 12 30 C 360 BA 572 21.6 3.64 2.7 1774 4 30 F 360 CS 220 20.0 0.39 2.2 2199 7 30 F 180 FS 501 20.0 0.57 2.1 1817 3r 60 C 360 BA 541 20.0 2.58 2.1 1997 4r 30 F 360 CS 220 21.6 2.92 1.8 2211 13 60 C 360 FS 499 20.0 0.00 1.8 1902 5r 60 F 180 BA 600 16.0 7.20 1.4 1887 14 60 F 360 CS 216 21.6 1.00 1.3 2174 2r 60 C 180 CS 206 25.0 0.21 0.5 2291 11 60 HC 360 BA 573 23.0 6.42 1.7 1743 6 30 HC 360 CS 315 20.0 2.26 1.3 2103 aECHIP randomizes order of mixtures and provides for duplicates bC = Class C, HC = High Carbon, F = Class F. cCS = Concrete Sand, FS = Foundry Sand, BA = Bottom Ash. Fine aggregate content was held constant at 1500 kg/m3. Table 3.3. Non–air-entrained CLSM mixture proportions (using statistical software)a. Mixture Cement Content (kg/m3) Fly Ash Typeb Fine Aggregate Typec Water Demand (kg/m3) Flow (cm) Total Bleeding (%) Air Content (%) Fresh Density (kg/m3) 18 60 None CS 200 21.6 0.70 16.5 1836 17a 30 None BA 582 12.7 4.35 20.0 1447 16 30 None CS 295 20.0 2.33 16.0 1922 21 30 None CS 170 18.0 0.62 25.5 1789 22 60 None CS 131 20.0 0.05 26.5 1748 22r 60 None CS 136 18.0 0.43 25.5 1802 16r 30 None CS 295 19.1 2.35 15.5 1874 19a 30 None BA 492 13.0 1.08 25.0 1385 20a 60 None BA 525 13.0 3.41 18.5 1485 23 60 None BA 454 14.0 1.30 28.5 1382 20r 60 None BA 525 13.0 1.44 15.5 1511 aThese mixtures were substituted for the originally proposed mixtures because of extreme difficulty in entraining air in mixtures containing foundry sand. The originally proposed mixtures containing foundry sand were still cast, but without entrained air. bFly ash was not used for these mixtures. cFine aggregate content was held constant at 1500 kg/m3. CS = Concrete Sand, BA = Bottom Ash. Table 3.4. Air-entrained CLSM mixture proportions.

hardened property measured), the research team placed par- ticular emphasis on developing a refined test that is more re- liable and reproducible. Issues such as load rate, curing condi- tion, temperature effects, and capping methods were studied in detail. The development of an improved compressive strength test method is also critical because of the inclusion of strength in most specifications, especially as in relation to excavatability. Tables 3.7 through 3.15 show the mixture proportions, along with selected fresh properties, for the additional investi- gations. To be consistent with the initial mixtures that contain aggregates, the aggregate content was held constant at 1500 kg/m3. These tables also contain some information on fresh CLSM characteristics and in some cases data are provided on properties, such as compressive strength. The findings of these investigations are provided in more detail later in this chapter. Because of relevance to field applications and based on im- portant findings related to corrosion of ductile iron specimens embedded in the initial 38 CLSM mixtures, an expanded and detailed long-term corrosion study was performed (Phase II). In Phase II, additional CLSM mixtures were prepared and dif- ferent corrosion scenarios were evaluated. The mixture pro- portions and fresh properties for the mixtures in Phase II are shown in Table 3.16. Testing Program Overview This section provides information on the test methods used. Test methods are grouped into three categories based on the characteristics that they are intended to measure; fresh properties, hardened properties, and durability characteris- tics. Some characteristics were studied in more detail than others. A summary of the measured characteristics and used methods is provided in Table 3.17. Mixing Procedure Trial mixing was performed for the initial 38 mixtures to de- termine their approximate water demand for a target flow of 200 to 250 mm. Flow was measured following ASTM D 6103. After determining the quantities of water for the target flow, the actual mixtures were cast and test samples were prepared. For the smaller mixture volumes, a 0.028 m3 drum mixer was used. For the larger mix volumes (needed for measuring additional characteristics on the selected mixtures), a high- capacity (0.056 to 0.070 m3) laboratory mixer was used. Mixing procedures were different for non–air-entrained and air-entrained mixtures. For non–air-entrained mixtures, 21 Mixture Cement Content (kg/m3) Fly Ash Typeb Fly Ash Content (kg/m3) Fine Aggregate Typec Water Demand (kg/m3) Flow (cm) Total Bleeding (%) Air Content (%) Fresh Density (kg/m3) 25 60 HC 1200 None 853 24.0 7.38 1.3 1322 27 60 F 1200 None 486 23.0 1.28 0.7 1638 28a 60 F 180 CSd 220 20.0 1.33 1.4 2182 29a 60 None 0 FSd 373 23.0 0.28 2.6 1812 24 60 F 1200 None 486 24.0 2.25 2.8 1635 30a 30 None 0 FSd 414 20.0 0.40 2.0 1789 26a 60 None 0 CSd 136 16.5 0.00 25.5 1802 aMixtures contain accelerating admixture. bHC = High Carbon, F = Class F cCS = Concrete Sand, FS = Foundry Sand dFine aggregate content was held constant at 1500 kg/m3. Table 3.5. Additional CLSM mixtures. Mixture Series Description Number of Mixtures A Effects of load rate on compressive strength (Table 3.7) 7 B Effects of curing and air drying on compressive strength (Table 3.8) 2 C Long-term strength gain and excavatability (Table 3.9) 9 D Freeze-thaw resistance (Table 3.10) 11 E Alternative capping materials for compression cylinders (Table 3.11) 18 F Effects of drainage on compression cylinders (Table 3.12) 8 G Effects of storage conditions on compressive strength (Table 3.13) 10 H Effects of temperature and humidity on compressive strength (Table 3.14) 6 I Permeability and triaxial shear strength (Table 3.15) 6 Table 3.6. Mixture series and their descriptions.

22 Mixture Cement Content (kg/m3) Fly Ash Typea Fly Ash Content (kg/m3) Fine Aggregate Typeb Water (kg/m3) Flow (mm) Air Content (%) Density (kg/m3) A-1 60 None 0 CS 156 175 25.0 1739 A-2 60 F 360 FS 520 200 1.9 1755 A-3 60 F 1200 None 486 213 1.6 1620 A-4 30 C 180 CS 265 330 1.7 2161 A-5 30 C 180 CS 213 216 –c 2226 A-6 60 F 1200 None 501 216 –c 1635 A-7 60 None 0 CS 156 165 24.5 1740 aF = Class F, C = Class C. bCS = Concrete Sand, FS = Foundry Sand. cToo low to measure. Table 3.7. Mixture proportions for load rate study. Mixturea Cement Content (kg/m3) Fly Ash Type Fly Ash Content (kg/m3) Fine Aggregate Typeb Water (kg/m3) Flow (mm) B-1 30 Class C 180 CS 203 250 B-2 30 Class C 180 CS 189 200 aB-1 and B-2 were cast on different days and different water contents were used to obtain the desired flow. bCS = Concrete Sand. Table 3.8. Mixture proportions for cylinder curing/conditioning study. Mixture Cement(kg/m3) Sand Content (kg/m3) Fly Ash Content (kg/m3) Fly Ash Typea Water (kg/m3) Flow (mm) Air Content (%) Density (kg/m3) C-1 60 None 1195 F 485 200 Entrapped 1637 C-2 0 2000 275 C 252 229 Entrapped 2148 C-3 30 1500 0 None 112 178 28.0 1642 C-4 15 1500 180 F 177 200 Entrapped 2192 C-5 30 1500 180 F 175 200 Entrapped 2158 C-6 15 1500 180 HC 224 216 Entrapped 2095 C-7 30 1500 180 HC 224 216 Entrapped 2115 C-8 15 1500 180 C 170 206 Entrapped 2190 C-9 45 1500 0 None 103 178 25.5 1652 aF = Class F, C = Class C, HC = High Carbon. Table 3.9. Mixture proportions used for excavation boxes and companion samples. Mixture Cement(kg/m3) Fine Aggregate Typea Fly Ash Typeb Water (kg/m3) Flow (mm) Air (%) Density (kg/m3) 28-Day Strength (MPa) D-1 30 CS None 119 180 27.0 1630 0.13 D-2 30 CS F 205 200 Entrapped 2196 1.02 D-3 30 CS HC 256 229 Entrapped 2078 0.79 D-4 30 CS C 200 216 Entrapped 1980 1.47 D-7 30 FS F 425 238 Entrapped 1835 0.11 D-6 30 FS HC 481 229 Entrapped 1757 0.12 D-5 30 FS C 399 200 Entrapped 1800 0.20 D-8 30 BA F 357 200 Entrapped 1870 0.38 D-9 30 BA HC 407 200 Entrapped 1733 0.25 D-10 30 BA C 282 200 Entrapped 1896 0.53 D-11 45 CS None 96 152 30.0 1569 0.34 aCS = Concrete Sand, FS = Foundry Sand, BA = Bottom Ash. Fine aggregate content was 1500 kg/m3. bF = Class F, HC = High Carbon, C = Class C. When included, fly ash content was 180 kg/m3 Table 3.10. Mixture proportions used for freezing and thawing study.

23 Mixture Cement Content (kg/m3) Fly Ash Typea Fly Ash Content (kg/m3) Concrete Sand (kg/m3) Water (kg/m3) Flow (mm) Aird (%) Density (kg/m3) E-1 60 F 1140 None 480 200 1.4 1630 E-2 None C 180 2000 250 200 1.2 1626 E-3 30 None 0 1500 109 175 29.0 1651 E-4 15 F 180 1500 184 200 1.7 1693 E-5 30 F 180 1500 176 200 2.5 2180 E-6 15 HC 180 1500 202 200 1.8 2122 E-7 30 HC 180 1500 229 225 E 2136 E-8 15 C 180 1500 179 200 1.8 2235 E-9 30 C 180 1500 238 21 E 2176 E-10 15 F 180 1500 241 19 E 2130 E-11 60 None 0 1500 153 18 29.1 1622 E-12 60 F 1200 0 500 22 E 1602 E-13 0 C 224 1672 165 19 4.0 2179 E-14 30 None 0 1500 130 20 29.5 1539 E-15 60 None 0 1500 130 22 28.5 1539 E-16 60 F 1200 0 485 42b 1.0 1795 E-17 30 F 180 1500 175 10c 2.3 2051 E-18 60 F 180 1500 175 14c 2.5 2083 aF= Class F, C = Class C, HC = High Carbon. bDifficult to obtain adequate flow by simply adding water. cToo much water included in the mixture. dE = Entrained. Table 3.11. Mixture proportions for alternative capping materials study. Mixture Cement Content (kg/m3) Fly Ash Typea Fly Ash Content (kg/m3) Concrete Sand (kg/m3) Water (kg/m3) Flow (mm) Total Bleeding (%) Air Content (%)b Fresh Density (kg/m3) F-1 60 F 1140 None 480 200 2.30 1.4 1630 F-2 None C 180 2000 250 200 0.00 1.2 1626 F-3 30 None 0 1500 109 175 0.00 29.0 1651 F-4 15 F 180 1500 184 200 2.07 1.7 1693 F-5 30 F 180 1500 175 200 0.87 – 2170 F-6 15 HC 180 1500 224 213 2.50 – 2100 F-7 30 HC 180 1500 224 225 3.04 – 2142 F-8 15 C 180 1500 170 200 0.85 – 2218 aF = Class F, C = Class C, HC = High Carbon. b “–” = too low to measure. Table 3.12. Mixture proportions for drainage condition study. Mixture Cement(kg/m3) Fly Ash Content (kg/m3) Fly Ash Typea Fine Aggregate Typeb Water (kg/m3) Flow (mm) Air contentc (%) G-1 60 1140 F None 485 200 E G-2 0 275 C CS 252 200 E G-3 30 0 None CS 112 187 29.0 G-4 15 180 F CS 177 200 E G-5 30 180 F CS 175 200 E G-6 45 0 None CS 103 190 30.0 G-7 30 180 F FS 349 216 E G-8 30 180 C FS 352 190 E G-9 30 180 F BA 424 140 E G-10 30 180 C BA 367 152 E aF= Class F, C = Class C. bFine aggregate content was 1500 kg/m3. Only G-2 had 2000 kg/m3. CS = Concrete Sand, FS = Foundry Sand, BA = Bottom Ash. cE = Entrained air. Table 3.13. Mixture proportions for cylinder storage study.

24 Mixture Cement Content (kg/m3) Fly Ash Typea Fly Ash Content (kg/m3) Concrete Sand (kg/m3) Water (kg/m3) Flow (mm) Air Content (%) Density (kg/m3) H-1 60 F 1200 None 492 220 2.4 1631 H-2 15 F 240 1500 197 240 1.2 2191 H-3 15 C 240 1500 175 240 1.4 2212 H-4 30 C 180 1500 181 200 1.2 2163 H-5 30 F 180 1500 188 220 1.4 2210 H-6 60 None 0 1500 123 190 25.5 1603 aF= Class F, C = Class C. Table 3.14. Mixture proportions for the temperature and drying effects study. Mixture Cement(kg/m3) Fly Ash (kg/m3) Sand (kg/m3) Water (kg/m3) Flow (mm) Air Content (%) Fresh Density (kg/m3) Dry Density (kg/m3) Moisture Content (%) I-1 30 180 1500 283 200 1.3 2036 1746 16.55 I-2 60 180 1535 327 210 1.3 2077 1748 18.79 I-3 120 180 1485 335 220 0.7 2087 1759 18.61 I-4 60 0 1471 212 190 22.5 1607 1415 13.59 I-5 60 0 1181 172 210 27.0 1569 1381 13.65 I-6 60 1200 0 500 200 0.5 1529 1133 35.00 Table 3.15. Mixture proportions for triaxial shear and water permeability studies. dry materials (fine aggregate, fly ash, and cement) were first mixed with approximately half of the expected mixing water (based on trial mixing) for 3 minutes, followed by a 2-minute rest period. After the rest period, the remainder of the batched water was added, followed by 3 additional minutes of mixing. Immediately after mixing, flow measurements were taken. In most cases, because of the benefit of trial mixing, the target flow of 200 to 250 mm was obtained. If the flow was less than desired, small amounts of water were added, followed by an additional minute of mixing to obtain the target flow. For some mixtures the desired minimum flow was difficult to achieve because of tendencies for bleeding and segregation. In those cases flow values less than 200 mm were accepted. For mixing air-entrained CLSM, mixing water was held back and a relatively dry consistency (i.e., zero slump) mixture was obtained in the mixer. An AEA specifically formulated for CLSM was then added with additional water. This process was necessary because of the high potency of the AEA. If the AEA was added to an already fluid mixture, the flow would far ex- ceed the desired range and the mixture would often suffer from excessive bleeding. The researchers found obtaining both the desired flow and air content to be challenging (but generally feasible). The researchers did not focus on optimizing the mixture proportions for optimal workability (i.e., flow, bleeding, etc.); the main objective was to obtain valid and direct com- parisons of constituent material types and contents. How- ever, with high amounts of fines and/or air entrainment, selected mixtures can be modified to obtain desired work- ability levels. For example, introducing additional fly ash to CLSM containing bottom ash has been shown by the re- searchers (and others) to be an effective method of reduc- ing segregation and bleeding while maintaining the required workability. Fresh CLSM Test Methods Flow Immediately after mixing, the flow was measured following ASTM D 6103. This method, which measures the diameter of a CLSM “pancake” after a 75 × 150 mm cylinder is slowly lifted, was found to be generally easy to perform and was also quite reproducible. Air Content and Unit Weight ASTM C 231 (pressure method), which is typically used for conventional concrete, was used, with slight modification, to measure the air content and unit weight of fresh CLSM mix- tures. The only modification was that the material was placed in one layer without rodding, instead of being placed in three equal layers and then consolidated. Setting Time and Bleeding Setting and hardening of CLSM mixtures were evaluated using three methods: needle penetration (ASTM C 403), soil penetrometer (or “pocket” penetrometer), and pocket vane

shear testing. The setting and hardening of fresh CLSM sam- ples that were placed in 150 × 150 mm containers were mea- sured using a needle penetrometer and a soil penetrometer. A larger container was used for measurements using the vane shear tester. Before each measurement, the bleed water was removed and weighed. Depth of penetration for the needle penetrometer and the soil penetrometer was approximately 25 mm and 6.4 mm, re- spectively. The pocket vane shear tester only measures the shear resistance of CLSM at the upper 3 mm layer. Segregation The segregation of six selected CLSM mixtures was mea- sured quantitatively. A specially designed mold, which con- sisted of three separate cylindrical sections, was used for this purpose. Each cylindrical section had a diameter of 100 mm and a height of 75 mm. The sections were connected verti- cally to produce a sample cylinder with a diameter of 100 mm and a height of approximately 225 mm. After the samples had set, steel plates, acting as “guillotines,” were inserted at the junctions between the cylinder sections, thus yielding three separate samples (upper, middle, and lower). Each sample was then wet sieved, using the No. 4, No. 8, No. 16, No. 30, No. 50, No. 100, and No. 200 sieves. Each portion retained on these sieves was then dried in the oven at 110°C for 24 hours and weighed. Material passing the No. 200 sieve was not collected. Using the resultant gradation from each of the three sections, a “pseudo” fineness modulus (FM) was cal- culated, using the same mathematical approach as typically 25 Mixture Cement(kg/m3) Fine Aggregate Typea, b Fly ash (kg/m3) Fly Ash Typea, c Water (kg/m3) Flow (mm) Air Content (%) Density (kg/m3) A1a 63 CS 1200 F 184 209 1.5 1605 A1b 63 – 1200 F 432 203 1.3 1591 A1c 63 – 1200 F 515 200 1.0 1605 A2a – S 206 C 134 200 1.5 2177 A2b – S 206 C 200 305 0.6 2180 A3a 30 CS – – 98 178 30.0 1602 A3b 30 S – – 118 200 25.0 1695 A3c 30 S – – 112 200 29.0 1593 A4a 15 CS 180 F 190 216 1.5 2194 A4b 15 S 180 F 204 229 1.3 2169 A4c 15 S 180 F 196 216 1.5 2167 A5a 30 CS 180 F 184 203 2.0 2185 A5b 30 S 180 F 188 203 2.3 2163 A5c 30 S 180 F 170 225 1.0 2177 A6a 15 CS 180 HC 190 210 2.0 2115 A6b 15 S 180 HC 224 203 2.0 2097 A6c 15 S 180 HC 216 206 1.0 2084 A7a 30 CS 180 HC 232 203 2.3 2099 A7b 30 S 180 HC 232 203 1.3 2111 A7c 30 S 180 HC 214 206 1.8 1978 A8a 15 CS 180 C 168 216 4.8 2155 A8b 15 S 180 C 168 216 1.8 2220 A8c 15 S 180 C 174.4 200 1.5 2179 B4a 30 CS 180 C 186 216 4.8 2170 B4b 30 S 180 C 144 216 1.3 2225 B4c 30 S 180 C 184 200 1.8 2228 B6a 30 CS 180 HC 472 209 2.3 1753 B6b 30 FS 180 HC 494 203 1.8 1765 B6c 30 FS 180 HC 524 200 1.5 1750 B7a 30 FS 180 C 484 222 1.5 1795 B7b 30 FS 180 C 426 229 3.0 1848 B9a 15 BA 180 HC 324 165 1.8 1821 B9b 30 BA 180 HC 324 145 2.8 1760 B10a 30 BA 180 C 318 175 1.5 1852 B10b 30 BA 180 C 318 200 2.0 1848 a “–” indicates that this item was not used in the mixture. bFine aggregate content was kept constant 1500 kg/m3. CS = Concrete Sand, S= Sand, FS = Foundry Sand, BA = Bottom Ash. cF = Class F, C = Class C, HC = High Carbon. Table 3.16. CLSM mixtures for corrosion study and their fresh properties.

used for determining the FM of the fine aggregate. As is the case with the normal treatment of FM, two different grada- tions can yield the same FM. Therefore, the overall grada- tion was also considered when the results were analyzed, as described later. Details of this test method can be found in Appendix B. Subsidence The surface settlement of a 100 × 600 mm cylindrical CLSM sample was monitored with time. The mold was pre- pared by stacking three 100 × 200 mm molds. A PVC pipe is also a good alternative. A small device to facilitate accurate measurement of the surface height changes was developed, as described in Appendix B. Hardened CLSM Test Methods Unconfined Compressive Strength Because of the importance of the compressive strength of CLSM in specifications, design, and construction, consider- able effort was placed in developing a test method with im- proved accuracy and reliability, and upon developing this method, work was done to better understand the effects of materials and mixture proportions on CLSM strength. This section describes the basic procedures followed to test the un- confined compressive strength of the initial 38 mixtures, in- cluding methods of preparing test cylinders, curing, capping, and testing. After describing this approach, information is provided on the various modifications and improvements in- vestigated using the mixtures previously shown in Tables 3.7 through 3.15. The results of the initial compressive strength study, as well as the findings of the various follow-up studies, were used to develop and recommend an improved uncon- fined compression test for CLSM, as presented in Appendix B. Because the strength of CLSM is relatively weak (compared to concrete), careful handling of the test cylinders is necessary, especially when stripping the cylinders from the molds. There- fore, plastic cylinder molds were pre-cut down the sides (two vertical cuts from top to bottom on opposite sides of the cylin- der) and taped closed. After the CLSM was mixed, the cylin- ders were filled, while being tapped lightly on the sides to re- move large entrapped air voids. Plastic lids were then placed firmly on the cylinders, and the specimens were moved im- mediately to the moist-curing or “fog” room, which was maintained at 100 percent relative humidity (RH) and 23°C. For most of the mixtures tested, the cylinders were kept in the molds for 7 days (or less for tests performed at earlier ages) and were then stripped by simply removing the tape and re- moving the CLSM specimens from the cylinders. Conven- tional stripping tools were not used because of possible dam- age to the specimens. The cylinders were then kept outside of their molds in the fog room until testing. Some CLSM mix- tures tend to leach and soften upon long-term fog-room ex- posure. Further studies on this issue and other cylinder stor- age issues are described later in this chapter. Moist curing was selected for this study so that test results can be compared from one laboratory to another, even though CLSM is rarely, if ever, moist cured in the field. The same ar- gument can also be presented for concrete testing. That is, con- crete is rarely moist cured for more than 7 days (if at all) in field applications, but standard curing in a fog room provides a benchmark for specification and construction acceptance. Although an ASTM method currently exists for measur- ing the unconfined compressive strength of CLSM (ASTM D 4832), some modifications were made to the method for this project, as described later. Most of the compression tests were performed on a relatively low-load capacity machine 26 CLSM Characteristic Initial Study Additional Studies Flow ASTM D 6103 Soil pocket penetrometerSetting/ hardening time ASTM C 403 Pocket torvane Sample size effect Small vs. large machine Loading rate Effect of drying samples Alternative capping materials Effects of drainage Compressive strength ASTM D 4832 Curing methods, conditions CLSM vs. sand ASTM G 1 Galvanic cells ASTM G 1, modified G 109 pH ASTM G 51 Resistivity ASTM G 57 Segregation, bleeding ASTM C 940 Subsidence No Standard Triaxial shear strength USACE EM 1110-2-1906 California bearing ratio AASHTO T 193 Resilient modulus AASHTO T 292 Water permeability ASTM D 5084 Drying shrinkage No standard Excavatability No standard Splitting tensile strength Chloride diffusion ASTM C 1152 Freezing and thawing ASTM D 560 Effects on permeability Direct shear strength None Thermal conductivity None Air/gas permeability None Consolidation None Leaching None Chemical and toxicity analyses Table 3.17. Overview of laboratory testing program.

(100 kN Instron), but some testing was also performed on a larger capacity machine (1780 kN Tinius Olson) to evaluate the effects of machine capacity. When using the smaller ma- chine, displacement control was used. Additional testing was performed to examine the effects of cylinder size (75 × 100 mm, 100 × 200 mm, and 150 × 300 mm), using a constant appar- ent strain rate. For this testing, the crosshead displacement was set at 0.38 mm/min for the 100 mm high specimen, 0.51 mm/min for the 200 mm high specimen, and 0.76 mm/min for the 300 mm high specimen. The objective was to produce failure in about the same amount of time for each cylinder size for a given mixture. A floating, spherical head was used to minimize eccentricities in loading. For the larger capacity compression machine, load- controlled testing was employed, as is the case for concrete test- ing. The typical load rates used for concrete, 138 to 345 kPa/s, would fail most CLSM specimens in a matter of seconds. Thus, a lower load rate was selected (6.9 kPa/s). This lower load rate was possible on the machine for this study but may not be available for many standard concrete compression machines. Sulfur capping was used for almost all of the cylinders, ex- cept for some weaker mixtures at early ages where use of sul- fur caps was not possible. For these mixtures, neoprene pads were used. As previously mentioned, several variations were investigated for the unconfined compressive test, including cylinder size, machine capacity, capping method, load rate, and curing. A description of these investigations is presented next, and the findings are included later in this chapter. Effects of Loading Rate on Compressive Strength. ASTM D 4832 gives little guidance regarding load rate, stating only to “Apply the load at a constant rate such that the cylinder will fail in not less than 2 min.” Because of the vagueness in defining the load rate, additional testing was performed to investigate the effects of loading rate on compressive strength. Using the seven CLSM mixtures summarized in Table 3.7, the effects of displacement rate (cross-head displacement of small load- frame) on compressive strength and deformation at peak load were studied. The following loading rates (or more accurately, deflection rates) were evaluated: 0.13 mm/min, 0.25 mm/min, 0.38 mm/min, 0.51 mm/min, and 0.89 mm/min. The aim was to determine a suitable load rate range that produces repeat- able compressive strength values and can be performed in a rel- atively short time. The latter concern was because several mix- tures from early research took a relatively long time (i.e., greater than 10 to 15 minutes) to fail in compression under dis- placement control, which would not be ideal for a testing lab- oratory that must test many cylinders daily. Cylinder Curing and Conditioning. Another possible source of error and confusion in ASTM D 4832 involves the curing conditions and the treatment of cylinders before test- ing. According to ASTM D 4832, CLSM cylinders are cured in the molds (in the fog room) until the time of testing. This procedure is different from the normal concrete approach to stripping the cylinders from their molds after about the first day of curing and then curing them in the fog room. ASTM D 4832 also specifies test cylinders must undergo a drying time of 4 to 8 hours after their moist-curing period ends and before they are tested in compression. Concrete cylinders, on the other hand, are specified to remain moist until the time of testing, with no required drying time. Research was con- ducted to investigate the effects of cylinder storage (i.e., in or out of molds) and specimen conditioning or drying prior to testing. To evaluate the effects of different curing (or specimen storage) regimes, the 10 mixtures in Table 3.13 (mixtures G-1 to G-10) were cast and test cylinders were prepared. This study aimed to identify possible differences in compressive strength when four different curing conditions were used, as summarized in Table 3.18. Curing condition A, described as “normal” in the table, was the method used most throughout this project, and curing condition C was identical to the method specified in ASTM D 4832. For curing condition D, the cylinders were placed outside the laboratory and were ex- posed to the high summer temperature and dry atmosphere of Austin, Texas. All cylinders were capped with sulfur cap- ping compound and tested at a loading rate of 0.38 mm/min. The mixtures listed in Table 3.8 were used to study the ef- fects of drying time (0.5, 2, 4, and 8 hours) on strength val- ues. Cylinders for this test were cured for various ages (7, 28, and 91 days) in a fog room, and then dried for the different time periods. The cylinders were then sulfur capped and tested in a deflection-controlled machine (0.38 mm/min). Effects of Curing Temperature and Humidity on Com- pressive Strength. As already addressed, the temperature to which CLSM is exposed during its strength-gain process may be very important, especially when mixtures containing 27 Curing Condition Curing Regime A (Normal) Keep sample in mold with cap on, for 7 days in fog room. Then strip cylinder and keep cylinders in fog room until time of testing. B (Mold) Keep sample in mold, with cap on, for 7 days in fog room. Then remove cap and keep cylinder in mold in fog room until time of testing. C (Cap) Keep sample in mold, with cap on, in the fog room until time of testing. D (Outside) Keep sample in mold, with cap off, outdoors until time of testing. Table 3.18. Additional curing regimes evaluated.

certain fly ashes are used. Because CLSM is used in many dif- ferent environments in practice, the same mixture propor- tions could exhibit different strength values. This study was intended to identify factors affecting strength gain of CLSM mixtures. This study was a follow-up to earlier testing that suggested that temperature plays a major role in CLSM strength development. Three curing temperatures (10°C, 21°C, and 38°C) and six CLSM mixtures (H-1 to H-6 in Table 3.14) were selected to study the strength gain of CLSM across a range of practical construction conditions. CLSM was cast into standard cylin- der molds (75 mm × 150 mm) and moved to the appropriate temperature-controlled chamber until the date of testing. The cylinders were stored in two different manners. Half the cylin- ders from each mixture were stripped after 3 days and returned to the same chamber until the time of testing (without control over relative humidity in the chamber). This condition is des- ignated later in this report as “dry” curing. Temperature and humidity were monitored throughout the test. The other half of the specimens from a given mixture were kept inside the molds with the caps firmly placed on top until the day of test- ing (designated as “wet” curing). These cylinders were placed directly next to the cylinders that had already been stripped. Cylinders were tested for compressive strength at 7, 28, and 91 days. The moisture contents of tested specimens were measured to assess the effects of curing conditions on the moisture content (or evaporable water content) and strength of CLSM. Effects of Drainage Conditions on Compressive Strength. Unlike conventional concrete, CLSM is very rarely, if ever, cured. During the strength-gain process of CLSM, it is often continuously in contact with the surrounding soil and/or struc- ture. Different environments may significantly affect the final strength of CLSM as the water-cement ratio may be affected by the seepage of water into surrounding materials or the loss of water through evaporation of bleed water. The effects of seepage and evaporation were investigated in a study using the mixtures detailed in Table 3.12 (F-1 to F-8). This study also investigated the effects of temperature on strength gain, using the fog room as a control and ambient conditions (hot Texas summer weather) as a test condition. As described later in this report, the findings of this temper- ature effects study were quite interesting, and subsequent testing was performed using controlled-temperature envi- ronments to further elucidate the influence of temperature on CLSM strength, especially for mixtures containing high volumes of fly ash. To simulate field conditions, plastic molds were buried in loose sand and CLSM mixtures were cast directly into the molds. Before the cylinders were cast, the plastic molds were subjected to different treatments to simulate various drainage conditions. To simulate the condition of no water loss, CLSM mixtures were cast in plastic molds without holes and tight lids were placed on the cylinders (condition “cap”). To simulate the condition that only surface water evaporation is possible, mixtures were cast in plastic molds without holes and lids (condition “no cap”). To simulate moderate water seepage, mixtures were cast into molds (without caps) with seven uniformly distributed 3.6 mm di- ameter holes on the bottom (condition “bottom holes”). To simulate a more severe drainage condition, mixtures were cast into molds with holes not only on the bottom but also on the side (condition “side holes”), again without caps being placed on the top of the cylinders. There were thirty- six holes on the walls and seven holes on the bottom per mold. All drilled holes were 3.6 mm in diameter. To avoid local effects, CLSM specimens from a given mixture were randomly placed throughout the test box. Alternative Capping Materials for Compression Testing. In preliminary testing, as well as the testing of the initial 38 mixtures included in this study, sulfur capping was found to be an effective method of obtaining repeatable compressive strength data. However, for early age samples and/or for par- ticularly low strength cylinders, it may not be possible to cap cylinders with sulfur because of the risk of specimen damage. Neoprene pads were used in these cases, but because only lim- ited testing was performed, the researchers decided to signifi- cantly expand the scope of the original work to investigate a range of neoprene (or other) pads with varying properties. It is well established that higher strength concrete requires higher neoprene durometer values, and vice versa. Thus, for CLSM, softer neoprene pads (much softer than those used for concrete) were expected to be needed. Other motiva- tions for studying alternatives to sulfur capping are the po- tential health concerns over the fumes generated from sulfur capping stations and the length of time needed to cap cylin- ders with sulfur. To address these important capping-related issues, a com- prehensive investigation of alternative capping materials was launched. Included in this study were sulfur caps, gypsum (or hydrostone) caps, and neoprene pads with durometer values of 20, 40, 50, 60, and 70. In a previous study, Sauter and Crouch (2000) used soft neoprene pads made of wet-suit rub- ber to measure the compressive strengths of excavatable CLSM cylinders. This idea was extended under this project to examine soft non-neoprene rubber pads and two-layer systems. A commercially available sorbothane viscoelastic polyurethane rubber material was identified and chosen for this study. The Shore OO durometer hardness of this material is 50, which is approximately a Shore A durometer hardness of 5, according to the producer. Pads were single-layer rubber sheets with a thickness of 12.7 mm. In the early stages of this 28

study, the researchers were concerned that the polyurethane pads may be too soft to dissipate the stress concentrations. As a result, a two-layer pad system was also tested, which consisted of polyurethane-neoprene (P-N) pads. Testing of samples included glued and unglued systems. The glued sys- tem used rubber cement to bond the polyurethane and neo- prene. For comparison, neoprene pads, 13 mm thick, with a Shore A durometer hardness of 50 were also used in certain tests. Table 3.19 summarizes the variables tested in this pro- gram. Nine mixtures (E-series) were used in this study, as previously described in Table 3.11. Excavatability The excavatability of CLSM was assessed for six of the orig- inal thirty-eight CLSM mixtures to gain an “order of magni- tude feel” for the relative ease of excavating various CLSM mix- tures. CLSM was cast into 450 × 450 × 300 mm plywood boxes and allowed to harden. Attempts were made to correlate “walk- ability” with soil penetrometer values as the CLSM gained strength in the first few hours. Long-term excavatability was assessed at an age of approximately 9 months using typical hand tools, including a shovel and a pick for six selected mix- tures. The compressive strength of laboratory-cured cylinders was also measured. In addition, a relatively new instrument, the Humboldt GeoGauge, was used at the time of excavation to attempt to correlate excavatability with the stiffness of CLSM, as measured by the GeoGauge. After the initial excavation study, a more comprehensive study on long-term strength gain and excavatability was launched. Nine CLSM mixtures (C-1 to C-9 in Table 3.9) were included in the study. A field penetrometer (field version of ASTM C 403) was used to evaluate the strength gain of CLSM mixtures. The dynamic cone penetrometer (DCP) was also used to estimate the excavatability of CLSM mixtures. The DCP is a modified and simplified version of the penetrometer used by the Country Roads Board, Victoria, Australia. It is used by geotechnical engineers to obtain an index of in-situ CBR and to estimate the strength of soil as a function of depth. The test- ing consists of dropping a hammer (8 kg in weight) from a height of 575 mm, which forces a steel rod with a conical head into the CLSM or soil. The penetration depth per blow was recorded. The corresponding DCP index value was used to es- timate a soil strength value (CBR). A second approach that was used in this study to predict excavatability follows a procedure developed and used in Hamilton County, Ohio. This approach uses a removability modulus (RE), as shown in Equation 3.1a: where W = In-situ unit weight (lb/ft3) C = 28-day unconfined compressive strength (psi) When SI units are used, as required by AASHTO, the equa- tion is rewritten as shown in Equation 3.1b: where W = In-situ unit weight (kg/m3) C = 28-day unconfined compressive strength (kPa) RE W C= × × 1 5 0 5 6 0 619 10 3 1 . .. ( . )b RE W C= × × 1 5 0 5 6 104 10 3 1 . . ( . )a 29 Sulfur Cap NeopreneCap Polyurethane Cap P-NU a P-NbCLSM Mixture Curing Condition 7 d 28 d 91 d 7 d 28 d 3 d 7 d 28 d 91 d 7 d 28 d 7 d 28 d E-9 E-10 E-11 E-12 E-13 E-14 E-15 E-16 E-17 Lab E-13 E-14 E-15 E-16 E-17 E-18 Field aP-N cap unbonded. bP-N cap bonded. Table 3.19. Summary of various capping systems used to test compressive strength of CLSM.

Engineers in Hamilton County, Ohio, and the city of Cincinnati have found this methodology to be effective in lim- iting long-term strength gain and ensuring future excavatabil- ity. The research team used the same approach in calculating RE values for the C-series mixtures and comparing the results to other direct or indirect indices of CLSM strength gain. In preliminary investigations within this project, the split- ting tensile strength of CLSM was identified as a potential in- dicator of excavatability. Splitting tensile strength is also a very simple property to measure (without the need to cap the cylinders). The stress conditions of CLSM specimens under splitting tensile testing may be quite similar to stress condi- tions of CLSM mixtures under digging conditions with a shovel or backhoe. The E-series mixtures (Table 3.11) were used to evaluate various capping materials and methods, with some of the key capping-related parameters shown in Table 3.19. In addition, mixtures E1 through E8 were tested exten- sively to evaluate the effect of durometer value (neoprene pad hardness) on CLSM strength, as described later in this report. California Bearing Ratio Moist-cured specimens were tested at an age of 28 days using a slightly modified version of AASHTO T 193. The only modification was that the CLSM was placed into the molds without compaction, as is required for testing soils. After 7 days of curing, the collar of the mold was removed, and the surface of the CLSM specimen was trimmed level using a straight edge. Resilient Modulus Moist-cured specimens were tested at an age of 28 days using a slightly modified version of AASHTO T 292, with the modification relating to the deviator stresses. In trial testing, the deviator stresses listed in Table 4 of AASHTO T 292 were not found to be sufficiently high to introduce deformations. The selection of deviator stresses was based on previous re- search performed at Texas A&M University. Load condi- tioning of 41 kPa was used for the 1000 repetitions. Since the completion of the laboratory component of this project, AASHTO T 292 has been replaced by AASHTO T 307. Re- search should be conducted using this new test method in the future to ensure that it is a viable test method for evalu- ating CLSM. Water Permeability The water permeability of six CLSM mixtures, moist cured for 28 days, was measured using ASTM D 5084. A back pres- sure of 69 kPa was applied and maintained until no additional water entered the sample (approximately 30 minutes). This condition was assumed to represent saturation. Because the samples were moist cured prior to testing, the samples were es- sentially already saturated prior to sample conditioning. Thus, the requirement that the B-value (ratio of pore water pressure to confining stress) be greater than or equal to 0.95 was waived for these tests. A confining pressure of 173 kPa was applied during the tests. Triaxial Shear Strength A commonly used soil triaxial test method (USACE EM 1110-2-1906) was used for testing the shear strength of CLSM. The samples were cast in Shelby tubes (approximately 70 mm diameter) and were stripped after 7 days. Testing was performed under consolidated and drained conditions at this time, and additional specimens were tested at an age of 28 days (note that the specimens were moist cured in a fog room from the time of stripping until the time of testing). The pore water pressure was maintained at 34.5 kPa, and the confining pressures were 69, 103.5, and 172.5 kPa. The loading rate was 0.38 mm/min, the same loading rate used for most standard unconfined compression tests. The test was terminated when the residual strength was reached or the stress-strain curve became essentially flat. By curve fitting, the effective internal friction angle, φ′, and the effective cohesion, c′, were deter- mined. Various other shear strength test methods could also be used for evaluating CLSM; the method selected for this study should not be considered as the only viable approach. Drying Shrinkage No standard methods exist to measure the drying shrink- age of CLSM. A method commonly used in Germany for self- leveling floor screeds was modified and used in this study. CLSM was cast into an 87.5 × 26.3 × 1000 mm steel channel. The channel had one fixed end plate with an anchor and one movable end plate with an anchor. Before the CLSM was cast, wax paper was placed on the inside of the channel to reduce friction. CLSM was then placed in the channel forms. The amount of shrinkage was measured using a linear variable dif- ferential transformer (LVDT) that measured the displacement of the movable end plate. Shrinkage measurements were taken daily for the first week and once a week thereafter. Durability Test Methods Corrosion A comprehensive laboratory corrosion program was per- formed, with the objective to characterize the corrosion per- formance of ductile iron and galvanized steel embedded in CLSM and to identify key parameters that significantly influ- 30

ence the corrosion performance of these materials. This re- search was performed in two phases: the first phase was a smaller scale study (using the 38 initial CLSM mixtures), and the second phase was a more significant follow-up study aimed at confirming the findings from the first phase and de- veloping a thorough understanding of the corrosion of met- als in CLSM. Metallic coupons machined from ductile iron and galva- nized steel pipes were tested in two conditions: uncoupled and coupled. Figures 3.2 and 3.3 show the samples for the un- coupled and coupled conditions, respectively. In the uncoupled state, metallic coupons were embedded in 75 × 150 mm plastic cylindrical molds containing CLSM. The center of the metallic coupon was placed at the center of the cylinder, 50 mm from the top surface. Because CLSM is a low-strength material, care was taken not to damage the sam- ples after casting. Precutting the plastic molds longitudinally and taping these cuts closed prior to casting minimized the damage for the uncoupled specimens. After casting, the tape was removed and the plastic mold was separated (not re- moved) from the CLSM sample surface. Coupled samples were prepared to address the issue of metals not being completely embedded in CLSM in the field applications. For the coupled conditions, pairs of ductile iron or galvanized steel coupons were embedded in 100 × 200 mm plastic molds that were half-filled with CLSM and half with soil. In this condition, one of the metallic coupons was completely embedded in CLSM and the other coupon was completely embedded in soil and they were connected with a 10 ohm resistor at the top as shown in Figure 3.3. The metallic coupons were secured such that both were approx- imately 5 mm from the CLSM/soil interface. Six holes (4 mm diameter) were drilled at 15 mm above the bottom of each cylinder and the holes were wrapped with a filter paper that would allow the exposure solution to enter into the cylinders while preventing the soils from being washed from the molds. Control samples were similar to the uncoupled samples, but metallic coupons were completely embedded in sand. Ductile iron coupons, 13 × 24 × 4 mm in size, were machined from a 300 mm diameter commercially available ductile iron pipe (AWWA C151, Grade 60-42-10) and zinc galvanized steel coupons, 13 × 24 × 3.5 mm in size, were machined from a 300 mm diameter zinc galvanized steel culvert (uncoated thick- ness approximately 3.40 mm). All CLSM samples were cured for 28 days at 23 ± 2°C and a relative humidity greater than 98 percent. Later, samples were exposed to a 3.0 percent sodium chloride solution or distilled water. The liquid level was maintained at a level of 90 mm throughout the test program. As previously stated, the corrosion study was performed in two phases. In the first phase, a large number of CLSM mix- tures were evaluated with a low number of samples per CLSM mixture. In the second phase, a lower number of CLSM mix- tures were evaluated with a higher number of samples com- pared to the first phase. The number of samples was increased in the second phase for a better statistical analysis. In both phases, uncoupled and coupled samples were prepared and tested. In the Phase I investigation, the initial thirty-eight CLSM mixtures (thirty mixtures and eight duplicates) were evaluated to determine the influence of CLSM constituent materials and 31 Coupon CLSM or sand 100 mm 75 x 150 mm cylinder Figure 3.2. Corrosion test setup for comparing corro- sion performance of coupons in CLSM and sand (uncoupled condition). Solution level Metallic Coupons SoilCLSM Resistor Figure 3.3. Corrosion test setup for compar- ing corrosion performance of galvanic coupled coupons in CLSM and sand.

proportioning on the corrosion of metals embedded in CLSM. The mixture proportions and fresh CLSM character- istics are shown in Tables 3.3 through 3.5. In this first phase, only ductile iron coupons were evaluated. Three coupled and uncoupled samples for each of the thirty-eight CLSM mix- tures and five control samples were fabricated. All of the sam- ples were exposed to 3.0 percent sodium chloride solution for 18 months. The control samples and the soil section of cou- pled samples were filled with a sand meeting the “graded sand” requirements of ASTM C 778, “Standard Sand.” In the Phase II investigation, a total of 13 CLSM mixtures were selected and cast to evaluate the corrosion of metals em- bedded in CLSM. The mixture proportions and fresh CLSM characteristics are shown in Table 3.16. Lower case letters added to the mixture designation indicate separate batches. Ductile iron and galvanized steel coupons were evaluated for corrosion activity. A minimum of five coupled and five un- coupled samples were prepared for each of the thirteen CLSM mixtures and exposure conditions. More than 1000 samples were evaluated in the Phase II study. Half of the samples were exposed to 3.0 percent sodium chloride solution and the re- maining samples were exposed to distilled water. All samples were exposed for 26 months. Sand and clay were used to fill the soil section of each coupled sample. The sand met the “graded sand” requirements of ASTM C 778. The clay used was obtained from the National Geotechnical Experimenta- tion Site located at the Texas A&M University Riverside Cam- pus. The plastic and liquid limits of the clay were 20.9 percent and 53.7 percent, respectively, and the hydraulic conductivity coefficient was 5 × 10−4 m/year. In both phases, metallic coupons were removed from the samples at the end of the exposure period and were evalu- ated for mass loss following ASTM G 1, “Preparing, Clean- ing, and Evaluating Corrosion Test Specimens.” Ductile iron coupons were cleaned using cleaning procedure C.3.5 and galvanized steel coupons were cleaned using cleaning procedure C.9.5. In the case of the coupled samples, only the coupons embedded in the sand were evaluated for mass loss as they were determined early in this study to be the anode. The coupon embedded in the CLSM section of these samples exhibited limited corrosion, if any. Evaluation of the corrosion performance of coupons was based on the percent mass loss due to corrosion (amount of mass loss resulting from corrosion divided by the original mass of coupons). In the Phase I study, the resistivity of the CLSM and sand were evaluated using a resistivity box (or soil box) as de- scribed in ASTM G 57, “Field Measurement of Soil Resistiv- ity Using the Wenner Four-Electrode Method.” Resistivity measurements were obtained from saturated samples 182 days after casting. These samples were cast at the same time with the corrosion samples (i.e., the CLSM came from the same batch for both sample types). In the Phase II study, the resis- tivity of CLSM and soils were not measured from separately cast samples, but from each of the actual exposed uncoupled and coupled samples following ASTM G 57. In the Phase I study, two 50 × 100 mm cylinders were cast for each CLSM mixture at the same time as the corrosion samples were cast to evaluate their pH. At 182 days after cast- ing, the CLSM cylinders were removed from the curing room, and pore solution was extracted from the samples and im- mediately evaluated for pH. In the Phase II study, CLSM and distilled water solutions (1:1 by weight) were prepared from each exposed uncoupled and coupled sample to evaluate for pH. In both phases, a pH combination electrode connected to a bench top multimeter with a precision of 0.01 was used to measure the pH. In the second phase, the pH of soil sam- ples used in the coupled samples was also determined using 1:1 by weight distilled water solutions. Because only one type of clay and only one type of sand was used in the samples, only randomly selected soil samples from coupled samples exposed to the chloride and distilled water environments were collected and tested. One soil pH value was determined for each type of soil exposed to each type of environment in a coupled sample. Chloride contents were determined using a test method developed under the Strategic Highway Research Program. This method rapidly determines the chloride content in ce- mentitious materials (Cady and Gannon 1992). Freezing and Thawing ASTM D 560, a method designed to measure the freeze- thaw resistance of soil-cement mixtures, was used with one modification: thawed specimens were not brushed because of the low strength of CLSM. CLSM samples were exposed to a temperature change from −18°C (a freezer) to 23°C (the fog room) in each cycle. Samples were exposed to 12 cycles, un- less they suffered severe damage at an earlier time. Mass loss was monitored as an indicator of damage. In the initial study, six cylinders from each mixture were exposed to freeze-thaw cycles. Three of these cylinders were moist cured for 7 days and the other three were moist cured for 28 days prior to freeze-thaw cycling. Because the tests typically used for con- crete, such as the ASTM C 666, were found to be too severe in preliminary trials for CLSM, the modified soil cement method was found to be a more suitable approach. In addition to the original six mixtures, a follow-up study was conducted to specifically investigate the effects of freeze- thaw damage on CLSM permeability. Eleven mixtures (D-series in Table 3.10) were used to study the freezing and thawing effects on permeability. The specimens were 100 × 125 mm and were subjected to freezing and thawing cycles at an age of 28 days (as per the modified version of ASTM D 560 described 32

in the previous paragraph). Because some CLSM specimens may suffer significant damage from freeze-thaw cycles, shrinkwrap was used to keep the specimens intact, thereby al- lowing for subsequent measurement of water permeability. Porous stones were secured at both ends of the specimens to facilitate the permeability measurements. After completing the freezing and thawing test, water permeability (or hy- draulic conductivity) was measured using the falling-head method. For reference, specimens from each mixture that were not subjected to freeze-thaw testing were also tested for water permeability. Leaching and Environmental Impact Coal combustion products, such as fly ash, and other by- product materials, such as silica fume and slag, have been used successfully and safely for years in conventional con- crete. CLSM has proven to be especially well-suited as a con- sumer of various by-product materials; further, by-product materials that are not typically allowed in conventional con- crete, such as fly ash not meeting ASTM C 618, are routinely used in CLSM. Therefore, there has been some concern about the potential for leaching of constituents in by-product ma- terials (e.g., heavy metals, organics) from CLSM and their im- pact on the environment. To address this issue, by-product materials evaluated in this project were tested to determine their chemical composition and potential for leaching from CLSM. For each of the by-products included in the initial labo- ratory study (three fly ashes, one bottom ash, and one foundry sand), the total heavy metal concentration was de- termined following EPA Method 610, “Determination of Certain Polynuclear Aromatic Hydrocarbons in Municipal and Industrial Discharges Using Liquid-Liquid Extraction and HPLC and/or Gas Chromatography as Provided Under 40 CFR 136.1,” where nitric acid and hydrogen peroxide were used to digest the materials. The eight elements ana- lyzed were arsenic, barium, cadmium, chromium, lead, mer- cury, selenium, and silver. Because this testing determines the total amount of heavy metals, and not the leachable amount, the extraction values may be 20 times the amount that might be leached in the TCLP limits (a “rule of thumb” value). If any of the by-products tested in this project yielded values in excess of the toxicity limits (20 times the TCLP limit), TCLP was then conducted to assess the type and amount of heavy metals that may be leached from the materials. The TCLP (EPA Method 1311) is one of the most common tests performed on materials to determine their potential for leaching. The concentrations of the eight heavy metals in the extracts were compared to the TCLP limits given in EPA publication 40 CFR 261.24 as shown in Table 3.20. As described later in this chapter, the materials tested in this project were found to easily “pass” the TCLP, meaning the heavy metal concentrations were very low and not of con- cern. However, if the TCLP values exceeded any of those listed in Table 3.20, another level of testing would have been initiated, specifically the American Nuclear Society leachate test (ANS 16.1). This test is a monolith test that measures the actual leachates from CLSM containing the by-product material of interest and is a better indicator of actual leach- ing potential. Results and Discussion The main findings from the laboratory study are presented next, with emphasis on selecting or developing appropriate test methods to measure key CLSM properties and building an understanding of how specific materials, mixture proportions, etc. affect CLSM performance. Fresh Properties An important aspect of this study was the assessment of the fresh or plastic properties of CLSM, both in terms of evaluat- ing candidate test methods and determining the relationship among constituent materials, mixture proportions, and fresh CLSM properties. Tables 3.3 through 3.5 summarize some of the important parameters, including water demand (to obtain the target flow), air content, flow, unit weight, and bleeding (%) for the initial 38 mixtures. The air content, flow, unit weight, and bleeding tests were found to be effective and user- friendly. Clearly, any change in source material, mixture pro- portions, or curing regime would impact each of the relevant fresh properties of CLSM. For this project, given that the water content was adjusted for each mixture to achieve a target flow (200 to 250 mm), some interesting observations could be made about what factors most affect flow characteristics, as discussed next. 33 Table 3.20. TCLP limits of heavy metals. Element TCLP Limits(ppm) Arsenic 5.0 Barium 100.0 Cadmium 1.0 Chromium 5.0 Lead 5.0 Mercury 0.2 Selenium 1.0 Silver 5.0 Source: 40 CFR 261.24

Water Demand Throughout this study, water (and sometimes AEA) was added to CLSM mixtures so that their flow values were be- tween 200 and 250 mm; this procedure allowed interesting in- formation and trends to be gleaned from what most affects water demand. The effects of material type and quantity on the water demand of CLSM were analyzed using a statistical pro- gram (ECHIP) for the non–air-entrained mixtures; the results are shown in Figure 3.4. The Pareto graph in Figure 3.4 illustrates the statistically sig- nificant variables that affect water demand. In this graph, the ef- fect is the difference between the specified variable level and the reference variable level. The reference variable levels are con- crete sand (river), Class C fly ash, 180 kg/m3 fly ash content, and 30 kg/m3 cement content. The difference was positive for bot- tom ash, foundry sand, and high-carbon fly ash. The difference was negative for 60 kg cement, Class F fly ash, 360 kg fly ash, 60 kg cement and high-carbon fly ash, 60 kg cement and Class F fly ash, and 60 kg cement and 360 kg fly ash. The figure shows that the fine aggregate type was the most significant factor af- fecting the water demand of mixtures. The use of high-carbon fly ash also increased the water demand. There is no significant difference between the use of the Class C and Class F fly ash. In addition, analysis of variance (ANOVA) calculations identified significant variables as fly ash type, fine aggregate type, and the interactions between cement content and fly ash type. Bleeding and Segregation Bleeding and segregation affect the subsidence and the uni- formity of the placed CLSM mixtures. Using ECHIP, the ef- fects of various factors on bleeding were evaluated. The use of foundry sand was found to reduce the bleeding significantly, while the bottom ash was found to significantly increase the bleeding. This finding indicates the importance of the fine ag- gregate on bleeding of water in fresh CLSM mixtures. The fly ash type had minimal effect on the bleeding of CLSM mix- tures. Little segregation was found in the five selected CLSM mixtures. Setting and Hardening The setting/hardening behavior of CLSM is important for many applications, especially for those where early strengths are needed to satisfy construction demands (i.e., timing be- tween lifts or early opening to traffic). Test methods are needed to easily assess the setting of CLSM, both in the laboratory and in the field. This section discusses some of the important find- ings regarding the setting time of CLSM, as measured by the needle penetrometer (ASTM C 403), soil pocket penetrometer, and pocket vane shear test. When the needle penetration of CLSM is measured, a certain minimum strength of CLSM is required to obtain meaningful test results. Thus, comparing the setting time of CLSM mixtures to each other at predefined time increments is often not feasible, but rather, the timing of measurements should be a function of constituent materials and mixture proportions. Also, because a needle penetrometer penetrates deeper into mixtures than a soil penetrometer, it is less subject to bleed water effects, as discussed next. Despite these differences in penetration depth and contact angle, there was a fairly reasonable correlation between soil penetrometer and needle penetrometer values for the 38 mix- tures, as shown in Figure 3.5 (with an R2 of approximately 0.75 for all the CLSM penetration data combined). Figure 3.6 shows the relationship between the soil pocket penetrometer and the vane shear device; although a general trend exists, it is not statis- tically strong. The walkability time was assessed by preparing large CLSM boxes that were walked on at various ages. The soil penetrom- eter values were found to range from 4.32 to 7.35 kPa (average of 6.14 kPa) when CLSM mixtures were able to support the weight of an average person with about 6.4 mm indentation. More comprehensive (and realistic) data were generated in the field tests on walkability times and how they relate to penetra- tion data and other parameters (described in Chapter 4). Subsidence All of the CLSM mixtures exhibited measurable subsi- dence, with the exception of mixture 23, which had relatively poor flowability. Except for mixture 23, a reasonable correla- tion existed between subsidence and bleeding. Mixture 6 had the highest subsidence of about 2.5 percent of the placement height. Comparison of the results obtained from mixtures 26 34 0 20 40 60 80 100 120 140 160 Bottom Ash Foundry Sand High Carbon Fly Ash 60 kg & High Carbon Fly Ash 60 kg & Type F Fly Ash 60 kg Class F Fly Ash 60 kg & 360 kg Fly Ash Fly Ash Content Water Demand (kg/m3) Source: after Du et al. (2004) Figure 3.4. Pareto-effects graph for water demand of non–air-entrained CLSM mixtures.

and 22r, which were identical except the use of an accelerat- ing admixture in 26, indicated that the use of an accelerating admixture reduced bleeding and slightly reduced subsidence. The results of subsidence testing are shown in Table 3.21. Hardened CLSM Properties Unconfined Compressive Strength A great deal of emphasis was placed in assessing the uncon- fined compressive strength of CLSM. This section first summa- rizes the findings from the initial mixtures (38 in all), in which several aspects of compression testing were examined, includ- ing the effects of cylinder size, capping material, and load rate. Based on these original mixtures, some useful predictive mod- els were developed to predict the strength gain (short and long term) of CLSM. The original study led to several follow-up studies, each of which focused in more detail on issues involv- ing testing parameters. Detailed investigations on load rate, curing and conditioning of cylinders, effects of drainage on strength, and the use of alternative capping materials were per- formed. The findings of the initial broad study and the later de- tailed studies were used to refine and improve existing methods of measuring the unconfined compressive strength of CLSM. The unconfined compressive strengths of the originally pro- posed mixtures at 3, 7, 28, and 91 days are shown in Table 3.22. 35 0 0.1 0.2 0.3 0.4 0.5 0.6 0 2 4 6 8 10 12 So il Pe ne tro m et er (M Pa ) ASTM C403 Penetration (MPa) Figure 3.5. Comparison between ASTM C 403 and soil penetrometer values. 0 0.1 0.2 0.3 0.4 0.5 0 0.1 0.2 0.3 0.4 0.5 V an e Sh ea r ( M Pa ) Soil Penetrometer (MPa) Figure 3.6. Comparison between soil pocket penetration and vane shear values.

36 Table 3.21. Subsidence results of six selected CLSM mixtures. Mixture 4 Mixture 24 Time (h) Subsidence (mm) Time (h) Subsidence (mm) 1.25 3.0 3.0 5.3 2.30 3.4 5.5 6.8 3.57 3.4 6.0 6.8 4.57 3.5 5.17 3.4 Mixture 23 Mixture 6 Time (h) Subsidence (mm) Time (h) Subsidence (mm) 3 0.3 1.00 9.4 4 0.3 2.17 10.5 3.33 12.7 4.33 14.9 6.50 15.9 Mixture 22r Mixture 26 Time (h) Subsidence (mm) Time (h) Subsidence (mm) 3.17 1.2 3.67 1.5 5.17 2.2 5.67 1.5 6.33 2.2 Note: The total specimen height was 600 mm. Table 3.22. Unconfined compressive strength of original 38 CLSM mixtures. Mixture 3-day fc(MPa) C.O.V. (%) 7-day fc (MPa) C.O.V. (%) 28-day fc (MPa) C.O.V. (%) 91-day fc (MPa) C.O.V. (%) 1 0.12 8.2 0.21 6.8 1.09 4.9 1.87 2.8 2 0.29 2.4 1.76 10.1 3.69 4.0 6.26 13.5 1r 0.13 14.4 0.24 1.2 1.35 7.7 2.34 0.3 15 0.07 8.2 0.11 5.7 0.18 4.0 0.25 1.4 3 0.33 10.5 0.57 2.4 1.36 6.8 2.02 2.9 8 0.09 9.8 0.11 8.3 0.25 4.7 0.33 10.6 10 0.12 2.3 0.16 15.1 0.22 7.9 0.26 1.3 9 0.09 9.7 0.13 12.4 0.22 5.0 0.25 2.9 5 0.14 13.2 0.18 8.2 0.46 16.1 0.57 5.3 12 0.30 16.2 0.27 6.2 0.57 4.7 0.86 3.4 4 0.34 4.9 0.48 6.2 0.79 11.0 1.08 13.6 7 0.09 3.2 0.11 3.4 0.12 9.7 0.16 5.8 3r 0.46 13.1 0.58 4.4 1.49 5.8 1.97 8.3 4r 0.41 13.6 0.57 5.8 0.94 4.0 1.03 6.9 24 0.34 4.8 0.22 1.4 0.44 0.1 0.58 4.6 23a – – 0.04 6.4 0.14 9.5 0.18 7.6 18a – – 0.33 6.8 0.70 1.1 0.79 4.0 14 0.58 6.6 1.07 13.5 2.15 8.1 3.49 16.8 2r 0.42 9.3 1.58 2.9 4.90 2.4 6.87 0.3 29 0.18 0.6 0.31 0.2 0.63 1.9 0.98 6.1 30 0.09 7.1 0.14 3.1 0.26 8.6 0.28 2.8 17a – – 0.01 31.8 0.07 18.9 0.13 16.8 11 0.33 1.7 0.42 4.3 0.75 3.5 0.94 4.7 6 0.40 10.2 0.47 0.7 0.83 4.9 1.09 4.7 16a – – 0.06 11.9 0.13 12.0 0.16 8.5 21a – – 0.09 10.6 0.16 11.8 0.18 11.7 22a – – 0.43 9.0 0.73 4.3 1.01 4.8 22r 0.32 4.2 0.50 9.7 0.96 17.5 0.93 7.0 5r 0.17 12.5 0.28 10.3 0.55 10.1 0.78 16.3 26 0.43 7.1 0.76 8.2 1.14 15.8 1.53 2.6 16ra – – 0.07 9.7 0.15 23.6 0.17 8.3 13 0.28 0.5 0.35 3.3 0.74 3.2 1.12 4.4 25 0.17 4.4 0.30 5.7 0.40 30.9 0.50 9.0 19a – – 0.02 0.71 0.06 45.0 0.06 13.2 20a – – 0.04 36.4 0.21 1.0 0.29 26.0 27 0.22 4.6 0.29 1.7 0.36 3.8 0.55 6.6 20ra – – 0.04 49.9 0.15 32.8 0.24 10.4 28 0.28 3.0 0.47 0.9 0.70 1.95 0.94 0.2 aMixtures were too weak to be tested at 3 days.

The data shown in this table were for small, sulfur-capped cylinders (75 × 150 mm) tested on a smaller capacity machine at a loading rate of 0.38 mm/min, as previously described. The results were found to be repeatable, with quite low values of coefficient of variation. An interesting observation that illustrates the uniqueness of CLSM is that most mixtures show a drastic change in the load- deflection curve as the curing time is increased. Figure 3.7 illustrates this behavior for Mixture 12, which was typical of most CLSM mixtures. At early ages, CLSM acts more like a soil, with more ductile behavior, but as time progresses, CLSM begins to act more like concrete, with higher strength and lower ductility. Efforts were made in this project, based on the strength re- sults for the initial 38 mixtures, to develop predictive models for the compressive strength of CLSM. Various models and statistical approaches were considered. No single model was found to work well for the entire range of materials and mix- ture proportions; however, predictive models for subsets of the mixtures were found to be quite accurate. For instance, sepa- rate models were developed for air-entrained (both for “high” and “moderate” air contents) and non–air-entrained CLSM. The results obtained from mixtures containing bottom ash or foundry sand were not included in the data that were used to develop the predictive compressive strength model of air- entrained CLSM (up to 91 days). These data were excluded because air entrainment was found to be too difficult in mix- tures containing these aggregates. The model predicting the compressive strength of air-entrained CLSM mixtures is shown in Equation 3.2 (Du et al. 2002): where f ′c = compressive strength (MPa) a = 0.3074  ln(t) + 0.2289 b = 0.0086  ln(t) − 0.272 t = age (days) w/c = water-cement ratio The measured and predicted compressive strengths using the model shown in Equation 3.2 for air-entrained mixtures are plotted in Figure 3.8. There was very good correlation, with an R2 value of 0.97. This model was also found to be effective in predicting long-term strength gain (i.e., beyond 91 days). For example, cylinders from mixture 22r that were tested for compressive strength after 256 days had an average strength of 1.0 MPa, compared to the predicted value of 1.1 MPa. ′= ⋅ ( )f a ec b w c ( . )3 2 37 0 500 1000 1500 2000 2500 3000 0 1 2 3 4 5 Deformation (mm) Lo ad (N ) 3 days 7 days 28 days Figure 3.7. Load-deformation response of mixture 12 at 3, 7, and 28 days. 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 1.2 Pr ed ic te d St re ng th (M Pa ) Measured Strength (MPa) Figure 3.8. Measured vs. predicted strengths of air-entrained CLSM.

A predictive model was also developed for non–air- entrained CLSM mixtures. If the water-cement ratio was the only variable used to predict the compressive strength, the model yielded an R2 value of 0.8. To improve this accuracy, a model was developed that included the water-cement ratio, aggregate type, fly ash type, and the fly ash content as strength- predicting variables. A critical aspect to this approach was to assign numerical values to the non-numerical variables used in the model. Through an iterative process, the constant, k, was selected for the materials used in this investigation. Con- crete or river sand (kriver sand) was assigned a value of 1.0, foundry sand (kfoundry sand) a value of 0.2, bottom ash (kbottom ash) a value of 1.0, Class C fly ash (kC ash) a value of 2.2, Class F fly ash (kF ash) a value of 1.0, and high-carbon fly ash (kHC ash) a value of 0.75. The equation for predicting the compressive strength, S(t), is shown in Equation 3.3 (Du et al. 2002), and a comparison between predicted and actual compressive strengths is shown in Figure 3.9. where S(t) = compressive strength (MPa) t = age (days) b0(t) = 0.0007 • t2 + 0.13 • t − 0.76 b1(t) = 0.0001 • t2 + 0.013 • t + 0.42 b2(t) = 0.00008 • t2 + 0.015 • t + 0.094 b3(t) = 0.003 • t − 1.03 S t b t k kagg type b t fly ashtype b( ) = ( ) ( ) ( )• •( )0 1 2. t b t fly ashcontent b t w c k ( ) ( ) ( ) • •( ) ( ) ( . )3 4 3 3 b4(t) = 0.75 − 0.018 • t when t ≤ 30 days b4(t) = 0.22 when t > 30 days Although the models presented previously are valid for only the materials used in this project, they provide important in- sights into the strength development of CLSM mixtures. Most of the significant effects were related to the influence of water demand on compressive strength. Mixtures containing ma- terials that increased the required water content for the target spread (such as foundry sand and high-carbon fly ash) gener- ally yielded lower compressive strengths compared to mix- tures with lower water contents. The chemical reactivity of fly ash was found to be critical, because the strength of CLSM mixtures containing Class C fly ash was higher than similar mixtures containing Class F or high-carbon fly ash. Class C fly ash has a higher CaO content than Class F fly ash (and the high-carbon fly ash used in this study) and it increases the early and final strengths of the mixtures. Because the devel- oped models are valid for only the specific materials used in this study, the researchers recommend preparing and testing a series of trial mixtures to predict the strength of CLSM mix- tures containing different materials. As stated earlier, various modifications to the unconfined compression test were studied in the initial investigation, some of which were later addressed in more detailed research. The following paragraphs briefly discuss the findings from this initial investigation that focused on cylinder capping methods, cylinder size, and testing machine capacity. Table 3.23 shows a comparison between sulfur-capped cylinders and neoprene-capped cylinders for the selected six CLSM mixtures. In general, sulfur-capped cylinders yielded 38 0.0 0.2 0.4 0.6 0.8 1.0 1.2 0.0 0.2 0.3 0.5 0.7 0.8 1.0 Pr ed ic te d St re ng th (M Pa ) Measured Strength (MPa) Source: Du et al. (2002) Figure 3.9. Measured vs. predicted strengths of non– air-entrained CLSM mixtures.

higher strengths than cylinders using neoprene pads. The neoprene pads used in this initial study had a durometer value of 50, which is a typical durometer value for conven- tional concrete cylinder testing. A more comprehensive study on capping materials, including neoprene pads with signifi- cantly lower durometer values, was subsequently performed, as described later in this chapter. In a study focusing on the effects of cylinder size (75 × 100 mm vs. 150 × 300 mm), cylinder size across this range was found to have little impact on strength values. This study used deflection-controlled testing with all cylinders tested at the same effective strain rate. Thus, like conventional con- crete, different cylinder sizes of CLSM can be used provided that the length-diameter ratio remains at 2:1, that the cylin- der size is sufficiently large for the maximum aggregate size, and that the load capacity of the machine is adequate to ac- curately measure the peak load. Most of the compression testing in this project used a smaller capacity testing machine under deflection control. However, many laboratories that typically test conventional concrete are currently using larger capacity concrete compression machines under load rate control. To assess the relative difference in strength values, a limited study was performed to compare the results of a small-capacity (100 kN) machine under deflection control to a larger capacity (1780 kN) compression machine using load control. It should be noted that in order to meet the time-to-failure limits described in ASTM D 4832 (not less than 2 minutes), a load rate of 6.9 kPa/s was used for the large ma- chine. The results indicated significantly lower strength values and higher variations for the large compression machine, com- pared to the results from the smaller, deflection-controlled machine. Thus, caution should be taken when using a large- capacity (Tinius Olson, 1780 kN capacity) machine under load control. When using a large-capacity machine for testing CLSM, one must ensure the machine is properly calibrated in the lower range of load values typically encountered for CLSM. Effects of Loading Rate. Because of the general lack of guidance provided in ASTM D 4832 regarding loading rate, the researchers placed additional emphasis on assessing the ef- fects of loading rate on the compressive strength of CLSM. The loading rate is an important parameter when considering compression testing. The testing of CLSM cylinders should first of all be accurate. If strength values are found to be strongly influenced by loading rate, then a finite range must be defined and required for accurate testing. The loading rate also determines the length of time needed to test a given cylin- der. This required length of time must be sufficient to ensure accuracy (i.e., not a sudden cylinder failure) and to complete a given test in a reasonable amount of time. In a laboratory or testing facility, the time required to test cylinders may be crit- ical, especially if many tests are performed in a single day. A deflection-controlled machine, often used to test soils, was used for this investigation with the following rates of deflection: 0.13 mm/min, 0.25 mm/min, 0.38 mm/min, 0.51 mm/min, and 0.89 mm/min. The deformation at peak load was used to determine if changes in loading rate resulted in changes in modes of failure, such as a change from relatively ductile to brittle failure. The effects of load rate on compressive strength were found to vary, depending upon the type of mixture and the age of testing. Interestingly, some mixtures (e.g., mixture A-2) were relatively insensitive to load rate, whereas others were quite sensitive. No consistent trend among all mixtures suggested that compressive strength was either directly or indirectly proportional to load rate (based on the range of rates eval- uated). In general, the range of loading rates from 0.25 to 0.64 mm/min produced the most consistent results. Also, within this range of loading rates, there was little impact on the deformation at peak load, suggesting that the mode of fail- ure (e.g., ductile vs. brittle) was not greatly affected by loading rate modifications. Therefore, a range of loading of 0.25 to 0.64 mm/min is recommended (as detailed in Appendix B) based on the findings that these rates generated accurate and repeat- able strength values in a reasonable amount of time. Effects of Cylinder Curing and Conditioning. The over- all objective of this study was to determine the most efficient 39 Table 3.23. Comparison of compressive strengths with different capping materials and methods. 3 days 7 days 28 days Mixture Sulfur(C.O.V.) MPa (%) Neoprene (C.O.V.) MPa (%) Sulfur (C.O.V.) MPa (%) Neoprene (C.O.V.) MPa (%) Sulfur (C.O.V.) MPa (%) Neoprene (C.O.V.) MPa (%) 4 0.34 (4.9) 0.25 (15.1) 0.48 (6.2) 0.34 (6.0) 0.79(11.0) 0.55 (14.4) 6 0.45 (6.8) 0.25 (14.7) 0.47 (0.7) 0.36 (6.3) 0.81 (4.9) 0.61 (3.2) 23a – – 0.04 (6.42) 0.03 (12.92) 0.14 (9.52) 0.12 (3.14) 24 0.34 (4.8) 0.15 (0.3) 0.22 (1.4) 0.19 (20.5) 0.44 (0.1) 0.30 (18.8) 22r 0.32 (4.2) 0.22 (8.1) 0.50 (9.7) 0.37 (10.6) 0.93 (17.5) 0.54 (4.4) 26 0.43 (7.1) 0.29 (9.1) 0.76 (8.2) 0.42 (11.2) 1.11 (15.8) 0.73 (7.2) aMixture was too weak to be tested at 3 days.

and accurate method(s) of curing and conditioning CLSM test cylinders. CLSM samples were exposed to the four different curing conditions (A, B, C, and D), as previously described in Table 3.18. All cylinders were capped with sulfur capping compound and tested at a loading rate of 0.38 mm/min at 28 and 91 days. The curing regime recommended by ASTM D 4832 is labeled as curing condition C in this study. The compressive strengths measured at 28 and 91 days are shown in Tables 3.24 and 3.25, respectively. The remainder of this section discusses these results and also describes some of the nuances observed when testing different materials and mixture proportions. Some interesting observations were made that illustrate that the strength of CLSM is significantly affected by variations in curing conditions, and further, that these variations are a function of specific materials and mix- ture proportions. In general, there were substantial differences between the strength of CLSM cylinders stored outdoors (in hot Austin, Texas, weather) and cylinders cured in the fog room. How- ever, there was not a consistent trend for all the mixtures studied, illustrating that the effects of temperature and cur- ing conditions are sensitive to material type and proportions. Some mixtures lost strength when stored outdoors, whereas others showed significant increases in strength when stored outdoors. The largest increase in strength for mixtures stored outdoors was observed for mixtures G-2, G-8, and G-10, which con- tained Class C fly ash, where the high temperatures helped to activate the fly ash. Mixture G-2, a rapid-setting mixture with 275 kg/m3 of Class C fly ash and no portland cement, showed a 40 percent increase in strength when stored outside rather than in the fog room. Mixtures containing Class F fly ash also exhibited higher strengths for cylinders cured outdoors in a hot climate, but the differences were more pronounced for Class C fly ash. In fact, the difference between fog room–cured and outdoor-cured cylinders was as high as 250 percent for mix- tures containing Class C fly ash. Thus, using laboratory-cured cylinders to assess the field performance of CLSM containing fly ash (especially high-calcium fly ash) in hot environments must be done with caution. The effects of temperature on CLSM hydration are especially important when large amounts of fly ash are used and when the fly ash–cement ratio is high. This temperature-driven impact on strength triggered a more comprehensive study on the effects of temperature on strength gain, as described later in this chapter. For high air-content mixtures G-3 and G-6, the effects of curing methods varied with cement contents (and strength lev- els). Mixture G-3 contained 30 kg/m3 of cement and exhibited relatively low strengths. In fact, mixture G-3 suffered such a re- duction in strength when stored outdoors (compared to fog room curing) that cylinders could not be tested at an age of 28 days. At 91 days, cylinders could be tested, but the resultant strengths were significantly lower than fog room–cured cylin- ders. Mixture G-6, which contained 45 kg/m3 of cement, ex- hibited less difference in strength (comparing outdoor curing to fog room curing) than the lower cement-content mixture (G-3). For mixture G-6, there was still a reduction in strength for outdoor-stored cylinders compared to fog-room cured cylinders at 28 days, but this difference became negligible at 91 days. These findings suggest that high air-content CLSM mixtures (without fly ash) benefit more from moist curing than they do from high-temperature exposures. Overall, these findings regarding curing temperature and cylinder storage led the research team to initiate a final inves- tigation on the effects of curing temperature and humidity on compressive strength, as discussed in the next section. The effect of drying time on the compressive strength of CLSM cylinders was evaluated; the results are shown in Table 3.26. CLSM cylinders were first cured in the fog room for various time periods (7, 28, and 91 days) and then allowed to dry at room temperature for various time periods before compression testing. 40 Table 3.24. Compressive strength at 28 days using different curing conditions. Curing Condition A (Normal) Curing Condition B (Mold) Curing Condition C (Cap) Curing Condition D (Outside)Mixture Average (kPa) C.O.V (%) Average (kPa) C.O.V (%) Average (kPa) C.O.V (%) Average (kPa) C.O.V (%) G-1 559.1 3.2 365.1 2.3 344.3 6.0 536.0 5.9 G-2 246.8 16.5 267.9 4.7 269.2 14.3 380.1 5.2 G-3 89.5 24.1 93.6 24.6 59.4 6.5 –a –b G-4 247.8 11.6 167.0 16.1 164.9 4.8 –b –b G-5 893.5 1.7 877.0 3.3 991.3 4.4 1259.0 4.8 G-6 369.5 7.8 326.1 10.2 306.6 7.3 295.5 8.2 G-7 150.8 11.5 145.4 4.8 161.5 3.7 280.0 8.3 G-8 170.3 9.2 137.3 6.4 160.9 16.1 600.1 8.3 G-9 317.9 12.9 328.6 13.3 279.0 18.7 496.0 14.5 G-10 486.3 3.6 412.2 9.6 411.6 6.1 986.9 4.4 aMixture was too weak to be tested. bNot enough specimens were available for testing at this age.

The standard test for compressive strength of concrete cylinders, ASTM C 39, requires samples to be tested in a moist state. Neville (1996) pointed out that testing samples when moist would provide more reproducible results because the moisture conditions (especially near the surface) of dried samples may vary. For air-dried CLSM samples compressive strength first increased with increasing drying periods and then dropped. The largest compressive strength values were obtained from samples that were dried for 2 hours. Re- searchers observed that, similar to drying concrete, drying CLSM samples could increase the measured compressive strength values as much as 17 percent. This observation indi- cates that it is not necessary to air-dry CLSM cylinders for 4 to 8 hours before capping as required by the ASTM D 4832. This procedure eliminates any potential differences in strength due to variable moisture conditions of cylinders and increases the number of cylinders that can be tested in a given day at a testing facility. Effects of Curing Temperature and Humidity on Com- pressive Strength. This study was a follow-up to previ- ous testing that suggested that temperature plays a major role in CLSM strength development. Three curing tem- peratures (10°C, 21°C, and 38°C) and six CLSM mixtures (H-1 to H-6 in Table 3.14) were selected to study the strength gain of CLSM across a range of practical construc- tion conditions. After casting the mixtures into plastic molds, the cylinders were immediately transported to the appro- priate temperature-controlled chambers. After 3 days of storage in the chambers, half the cylinders from each mix- ture were stripped or removed from the molds and re- turned to the same chamber until the time of testing. This regime is referred to as “dry” curing in subsequent discus- sions. The other half of the cylinders from each mixture were kept inside the molds with the caps firmly in place until the day of testing (designated as “wet” curing). These cylinders were placed directly next to the cylinders that had already been stripped. The results of compression testing at 7, 28, and 91 days are shown in Table 3.27. Also included in this table are the mois- ture contents of cylinders that were just tested in compression, which were measured to assess the effects of curing conditions on the moisture content (or evaporable water content) and strength of CLSM. Mixtures containing fly ash exhibited significant strength gain at 38°C, compared to lower curing temperatures. The in- crease in strength was more pronounced for Class C fly ash, compared to Class F fly ash, mainly because of the differ- ence in their reactivity. For example, at 38°C curing temper- ature, the increase in compressive strength of Class C fly ash– containing mixture H-5 was 160 percent. However the increase in compressive strength of Class F fly ash–containing mixture H-2 was only 40 percent. The CaO content of fly ash is gen- erally the most important factor affecting the compressive 41 Table 3.25. Compressive strength at 91 days using different curing conditions. Curing Condition A (Normal) Curing Condition B (Mold) Curing Condition C (Cap) Curing Condition D (Outside)Mixture Average (kPa) C.O.V (%) Average (kPa) C.O.V (%) Average (kPa) C.O.V (%) Average (kPa) C.O.V (%) G-1 754.6 7.7 500.7 5.0 479.4 6.0 808.8 2.3 G-2 346.1 7.8 308.6 11.1 266.5 15.3 422.7 4.5 G-3 101.9 7.3 97.4 2.0 96.3 15.5 49.6 6.1 G-4 305.8 11.4 272.9 7.9 239.0 8.1 408.5 7.1 G-5 1248.8 19.2 1342.3 14.2 1244.3 7.9 –a –a G-6 378.0 13.9 367.7 3.7 366.6 7.1 378.0 13.9 G-7 175.8 10.7 179.3 4.1 199.6 7.6 271.9 8.3 G-8 218.0 2.0 210.5 8.8 201.5 12.6 943.0 2.7 G-9 408.8 6.3 347.3 14.0 353.4 24.1 501.1 13.7 G-10 785.9 17.8 688.0 6.2 617.7 5.4 1080.1 6.3 aNot enough specimens were available for testing at this age. Table 3.26. Influence of air drying on compressive strength (mixture B-2). Drying Time 0-0.5 h 2 h 4 h 8 h Age Strength (kPa) C.O.V. (%) Strength (kPa) C.O.V. (%) Strength (kPa) C.O.V. (%) Strength (kPa) C.O.V. (%) 7 days 310.1 5.8 329.2 11.3 357.5 2.7 363.2 2.0 28 days 1575.1 1.9 1536.5 7.8 1447.2 5.7 1649.8 3.3 91 days 3289.4 2.0 3226.9 4.6 3430.4 3.9 3536.1 4.6

strength of CLSM mixtures, especially at high temperatures. In general, the compressive strength values of CLSM mix- tures without fly ash were less sensitive to curing temperature than mixtures containing fly ash. Air drying of CLSM cylinders from the third day of curing generally increased their 7-day strength, compared to the samples that were kept continuously in molds for 7 days. However, the 91-day compressive strength of air-dried cylin- ders was generally lower compared to the samples that were kept in molds. At 28 days, air-dried cylinders and the samples that were kept in molds gave mixed results. This study reinforced the need to recognize that field in- stallations of CLSM may possess vastly different strengths than one might predict from laboratory-cured tests, especially when CLSM contains fly ash and is used in hot climates. As such, CLSM mixtures that are produced with locally available materials for specific field applications should be tested in field conditions. Issues such as the long-term strength gain of CLSM mixtures in the field conditions should be addressed prior to the use of CLSM mixtures. An assessment of the on-site strength of CLSM should take into account laboratory-obtained test results, but it should also take into account climatic conditions. An understanding of material reactivity is helpful in extrapolat- ing laboratory results to field performance. Effects of Drainage Conditions on Compressive Strength. To evaluate the effect of different drainage conditions on the compressive strength of CLSM a specially designed “curing box” was constructed, as described earlier in this chapter. Re- searchers evaluated four different storage conditions, which 42 Table 3.27. Effects of temperature and relative humidity on CLSM compressive strength. Age 7 days 28 days 91 days Mixture Temp. a (°C) MCb (%) CSc (kPa) C.O.V.d (%) MCb (%) CSc (kPa) C.O.V. d (%) MCb (%) CSc (kPa) C.O.V. d (%) 10 D 36.9 328.6 3.5 17.8 548.5 7.4 2.0 258.5 2.7 10 W 38.0 210.5 6.0 37.1 240.9 7.8 27.0 323.4 5.5 21 D 18.1 367.5 2.8 1.6 607.3 6.5 1.1 480.2 6.5 21 W 38.0 299.2 23.0 37.8 266.2 11.8 26.8 632.3 3.3 38 D 19.5 828.5 2.2 1.8 722.3 4.4 1.1 756.7 12.2 H-1 38 W 22.6 440.4 13.5 35.3 802.2 2.4 26.0 917.3 0.4 10 D 7.4 188.3 10.2 1.2 314.1 7.0 2.7 118.9 13.6 10 W 10.3 151.9 8.6 9.6 194.2 12.3 5.5 260.5 3.6 21 D 1.8 214.1 2.3 0.3 171.4 10.1 6.5 142.7 7.5 21 W 9.8 172.4 2.1 9.1 233.1 6.5 3.3 345.3 10.2 38 D 0.3 256.6 12.2 0.2 211.9 4.9 12.2 263.6 3.2 H-2 38 W 8.6 220.9 11.7 8.9 458.2 7.3 0.4 634.6 2.9 10 D 6.0 1384.3 1.5 1.6 2315.4 4.9 1.1 1395.4 13.0 10 W 8.7 937.7 16.2 9.1 1141.2 8.9 7.9 1367.9 3.2 21 D 1.6 1213.4 5.1 0.6 963.6 9.6 0.5 852.9 8.9 21 W 8.7 695.7 8.8 8.6 786.2 3.2 7.4 919.8 6.7 38 D 0.3 1222.5 6.8 0.0 944.5 3.8 0.4 1042.0 13.5 H-3 38 W 6.3 864.0 2.1 6.2 3844.8 7.4 2.3 3880.9 20.4 10 D 7.4 314.0 2.8 1.4 1486.2 2.8 1.1 895.8 14.8 10 W 10.0 185.9 9.4 37.7 893.2 9.6 8.1 1670.3 4.9 21 D 1.2 669.6 3.0 0.3 628.6 9.8 0.4 458.6 9.9 21 W 8.8 501.7 4.1 7.7 1570.7 7.3 3.8 3743.6 4.9 38 D 0.4 2615.0 8.5 0.9 2041.2 3.5 0.3 2060.6 5.6 H-4 38 W 5.9 2098.8 9.2 3.7 12116.8 11.1 1.4 11512.6 7.0 10 D 8.7 273.1 6.9 1.4 711.4 3.2 0.8 421.9 5.8 10 W 10.0 232.6 4.5 9.7 544.5 9.6 8.6 1362.6 6.3 21 D 1.4 420.8 3.9 0.3 411.2 5.3 0.3 330.7 12.1 21 W 10.7 316.8 6.7 10.2 815.2 2.9 8.8 1497.7 4.6 38 D 0.3 1524.7 9.1 0.2 1423.7 4.6 0.1 1339.4 11.9 H-5 38 W 7.4 1472.5 7.3 7.7 2282.0 8.6 3.5 2638.2 12.0 10 D 6.3 281.9 16.7 1.6 740.5 8.7 1.1 669.3 3.4 10 W 8.7 210.6 6.9 8.1 470.5 1.2 7 922.2 5.6 21 D 1.0 480.4 16.9 0.3 434.5 15.0 0.3 372.9 18.7 21 W 7.5 371.4 15.8 6.2 744.7 7.2 6.2 929.6 9.4 38 D 0.4 816.8 11.9 0.2 828.0 28.1 0.2 782.0 4.1 H-6 38 W 6.1 562.5 10.7 4.8 786.3 12.6 0.4 991.3 7.7 aD = cylinders stripped after 3 days, W = cylinders kept in mold until time of testing. bMC = moisture content cCS = compressive strength dC.O.V. = coefficient of variation

ranged from “normal” curing in a fog room to curing in cylin- ders that allowed seepage from the bottom and/or sides and evaporation from the top (to mimic field conditions in a trench, for example). The “curing box” was kept at a higher tempera- ture than fog room–cured cylinders, and therefore, this study also was intended to assess temperature-related effects. The main finding from this study was that the effects of water seepage to adjacent sand and loss of water by evapo- ration did not significantly impact the strength of CLSM. The study also confirmed that temperature plays a key role in many CLSM mixtures and suggested that drainage and evaporation may not be as critical as temperature-induced effects. Alternative Capping Materials for Compression Testing. The capping materials evaluated in this study included neo- prene pads, sulfur caps, and gypsum caps (or “hydro- stone”). Neoprene pads with Shore A durometer values of 20, 40, 50, 60, and 70 were evaluated. CLSM cylinders were capped and tested after 7, 28, and 91 days of curing using a load rate of 0.38 mm/min. Gypsum paste prepared for capping had a gypsum-water ratio of 0.3 and required approximately 40 minutes to harden. Because gypsum capping was a time-consuming process, it was only used for 28-day compression testing. Table 3.28 summarizes the strength data for the various capping methods and materi- als. Table 3.29 shows the corresponding coefficients of vari- ation of measured compressive strength using the various capping methods. For almost all cases, sulfur capping yielded the highest strength values for all eight mixtures tested. Also, in gen- eral, sulfur capping generated the lowest variations com- pared to the other capping methods. Lower strength cylin- ders tested with higher hardness value neoprene pads ex- hibited higher variations in the results. Compressive strength results obtained using durometer 20 neoprene pads per- formed better, especially with weaker cylinders, and exhib- ited only slightly larger variations than the results obtained using sulfur capping. ASTM D 4832 states that capping systems are acceptable when the average strength obtained is not less than 80 percent of the average strength of companion cylinders capped with sulfur capping compound. According to this criterion, only the use of gypsum and neoprene pads with a durometer value of 50 could be qualified using the ASTM C 1231 method. How- ever, the qualification method described in ASTM C 1231 is developed for concrete samples; if the ASTM C 1231 process is slightly modified to recognize the unique properties of CLSM (see Folliard et al. [2001] for more details), neoprene pads with a durometer value of 20 could be qualified as an acceptable capping material. In an additional study, samples of four CLSM mixtures were tested after 7 days of curing using sulfur capping com- pound, neoprene pads with a Shore A durometer hardness of 50, polyurethane pads, and unbonded polyurethane- neoprene pads. Figure 3.10 shows the ratios of compressive strength values obtained using different capping methods to the compressive strength values obtained using sulfur capping for similar samples. The abscissa of the plot is the mean compressive strength of the samples capped with sulfur compound. Re- sults indicated that, for compressive strength values lower 43 Table 3.28. Compressive strength results using different capping materials. MixtureCapping Material Age E-1 E-2 E-3 E-4 E-5 E-6 E-7 E-8 7 days 0.31 0.11 0.08 0.13 0.20 0.34 0.43 0.19 28 days 0.53 – 0.14 0.29 0.33 1.03 1.19 0.66 Sulfur 91 days 0.85 – 0.14 0.45 0.48 1.58 1.71 1.24 Gypsum 28 days 0.52 – 0.11 0.24 0.36 0.95 1.11 0.59 7 days 0.28 0.09 0.06 0.12 0.20 0.28 0.36 0.17 28 days 0.42 – 0.11 0.24 0.34 0.92 1.05 0.42 Neoprene Pad D70 91 days – – 0.10 0.34 0.46 1.12 1.39 0.91 7 days 0.29 0.09 0.05 0.11 0.16 0.33 0.36 0.18 28 days 0.57 – 0.12 0.26 0.27 0.66 1.02 0.42 Neoprene Pad D60 91 days – – 0.10 0.37 0.31 1.29 1.50 0.89 7 days 0.26 0.08 0.05 0.13 0.21 0.34 0.40 0.19 28 days 0.51 – 0.10 0.25 0.28 0.87 0.96 0.47 Neoprene Pad D50 91 days – – 0.10 0.35 0.32 1.23 1.44 1.08 7 days 0.27 0.07 0.04 0.13 0.22 0.33 0.34 0.21 28 days 0.51 – 0.10 0.22 0.25 0.91 1.01 0.50 Neoprene Pad D40 91 days 0.79 – 0.14 0.33 0.45 1.36 1.37 0.80 7 days 0.25 0.09 0.04 0.11 0.20 0.32 0.40 0.17 28 days 0.71 – 0.09 0.24 0.31 0.90 0.96 0.57 Neoprene Pad D20 91 days – – 0.13 0.40 0.36 1.30 1.43 1.05 “–” = Not enough specimens were available for testing at this age.

than approximately 200 kPa, the non-sulfur capping meth- ods generally underestimate the compressive strength. How- ever, for compressive strength values greater than 200 kPa, the use of non-sulfur capping methods provided results that were acceptable following the criteria given in ASTM D 4832. As noted, for different capping methods to be acceptable, the ASTM D 4832 standard requires the obtained compressive strength values to be not less than 80 percent of the corre- sponding values obtained using sulfur caps. Based on these results, the following recommendations can be made with regard to generating acceptable strength data using unbonded pads: • CLSM with compressive strength lower than 1.0 MPa should be tested using unbonded polyurethane pads (Shore OO 50, equal to Shore A durometer 5) • CLSM with compressive strength between 1.0 and 2.0 MPa should be tested using either polyurethane pads (Shore OO 50) or neoprene pads (Shore A durometer 50) • CLSMwithcompressivestrengthgreaterthan2.0MPa should be tested using neoprene pads (Shore A durometer 50) The selection of durometer 50 neoprene pads for higher strength CLSM mixtures was due to the general availability of these pads in concrete laboratories and because the pads can be 44 Table 3.29. Coefficients of variation (%) for compressive strengths using different capping materials. MixtureCapping Material Age E-1 E-2 E-3 E-4 E-5 E-6 E-7 E-8 7 days 1.6 5.8 5.0 7.1 15.7 17.6 8.4 6.9 28 days 16.3 – 14.2 5.7 20.7 13.3 4.0 10.3 Sulfur 91 days 3.6 – 15.3 10.6 10.5 6.9 5.3 8.1 Gypsum 28 days 18.0 – 7.7 6.1 5.5 5.3 7.8 14.9 7 days 5.3 12.4 20.3 5.9 3.7 12.0 9.7 9.5 28 days 45.9 – 19.3 22.0 44.5 16.1 14.9 20.8 Neoprene Pad D70 91 days – – 22.8 6.8 24.9 8.3 3.0 3.7 7 days 18.0 6.5 15.5 7.2 17.3 12.0 13.2 22.1 28 days 12.9 – 2.9 2.5 4.0 23.9 7.6 22.0 Neoprene Pad D60 91 days – – 30.4 12.9 16.4 12.5 6.6 21.4 7 days 17.5 5.7 28.7 15.9 7.4 0.7 4.6 30.0 28 days 18.1 – 15.5 5.0 20.9 15.1 6.4 6.5 Neoprene Pad D50 91 days – – 17.0 22.2 12.6 4.2 4.0 11.5 7 days 17.9 27.6 23.3 21.4 10.1 13.0 6.8 15.2 28 days 15.4 – 14.0 9.5 20.5 19.5 10.8 11.2 Neoprene Pad D40 91 days 12.7 – 2.7 4.6 27.8 6.6 14.6 9.4 7 days 16.5 1.9 20.5 15.8 10.3 7.3 9.6 16.6 28 days 7.1 – 1.1 14.5 10.1 19.2 6.8 5.9 Neoprene Pad D20 91 days – – 6.1 7.4 22.2 9.1 6.0 23.5 “–” = Not enough specimens were available for testing at this age. 0.50 0.60 0.70 0.80 0.90 1.0 1.1 0 200 400 600 800 1000 1200 Polyurethane pads Unbonded PN Bonded PN D 50 R at io o f S tre ng th M ea su re d U sin g Su lfu r C ap s Compressive Strength by Sulfur Caps (kPa) Figure 3.10. Comparison of strength values from different capping methods.

qualified in accordance with ASTM D 4832. Polyurethane pads were found to be too weak to use under high compression loads. Excavatability This section summarizes the results of tests that are directly or indirectly related to the excavatability of CLSM. Included are the initial findings from tests conducted on selected CLSM mixtures (six from the original mixture series) and the results of subsequent, more comprehensive testing on exca- vatability and related indices. The splitting tensile test was also evaluated as a potential index of excavatability, and as such, the tensile results are provided in this section. As previously described in this chapter, the excavatability of CLSM was assessed for six of the original thirty-eight mix- tures by casting CLSM into 450 × 450 × 300 mm plywood boxes. The early strength or stiffness of CLSM was assessed using a soil penetrometer, and these values were correlated with “walkability” or the time at which an average person can walk on the material. Soil penetrometer values in the range of 4.32 to 7.35 kPa were found to correlate with initial walkabil- ity. Long-term excavatability was assessed for the six CLSM mixtures at an age of approximately 9 months using typical hand tools, including a shovel and a pick. Just prior to assess- ing the excavatability, the “stiffness” of the samples was mea- sured using the GeoGauge instrument (as described earlier). Compressive strengths of laboratory-cured cylinders were also measured at the time of excavation. As shown in Table 3.30, there was no clear correlation be- tween compressive strength, excavatability, and stiffness (as measured by the GeoGauge). For example, the laboratory- cured compressive strength of mixture 23 was quite low, but the field-cured excavation box was not excavatable. Previ- ous testing has shown that laboratory-cured cylinders may not be accurate indicators of in-situ strength or stiffness, es- pecially when CLSM is exposed to higher temperatures in the field (as was the case for these samples). Also, the results suggest that compressive strength, by itself, may not be a good predictor of excavatability. Another example of lack of correlation was the fact that mixture 24 had a higher stiff- ness than mixture 22r, yet it was much easier to excavate. The findings of this initial study led the researchers to per- form more comprehensive research on excavatability, in- cluding the assessment of other test methods and indirect indices, as described next. The researchers performed a comprehensive follow-up study to the initial excavatability investigation. A wide range of CLSM mixtures (C-series) was included in the investiga- tion, and the following methods or approaches were assessed as possible indices (direct or indirect): • Unconfined compressive strength (field-cured cylinders) • Field penetrometer (field version of ASTM C 403 needle penetrometer) • DCP • CBR (estimated from DCP) • Stiffness gauge (GeoGauge) • RE • Splitting tensile strength Table 3.31 summarizes additional results from the exca- vatability study, including DCP values, stiffness values (using GeoGauge), and calculated RE values. The table also shows the compressive strength for laboratory-cured cylinders (at 28 days) and field-cured cylinders, which were cured adjacent to the excavatability boxes and tested at the time of excava- tion (240 days). The densities of the laboratory-cured and field-cured cylinders were measured before testing them in compression, and these values were used in RE calculations. The relative ease of excavation was assessed using a hand shovel. The GeoGauge was used to assess the relative stiffness of the CLSM specimens. As CLSM mixtures were quite strong (relative to soil), a thin layer of wet fine sand was placed on the surface prior to testing, as per the recommendations of the manufacturer. Three readings were taken for each mixture. 45 Table 3.30. Results of initial excavation study. Mixture Compressive Strength (MPa) Strength C.O.V. (%) Stiffnessa (MN/m) Stiffness C.O.V. (%) Relative Ease of Excavation (with shovel and/or pick)b 24 0.31 4.86 11.62 10.27 1 22r 1.01 5.49 9.76 5.40 7 6 0.92 9.66 13.20 6.72 9 4 0.70 5.24 30.72 9.10 10 26 1.61 5.91 34.57 2.09 10 23 0.12 22.31 17.15 7.51 9 aStiffness was measured using the GeoGauge device. bEach mixture was assigned an ease of excavation value from 1 to 10, where 1 is easiest (able to excavate with minimal pressure applied to shovel and/or pick) and 10 is most difficult (not able to excavate with shovel and/or pick, even under heavy pressure).

The variations were quite high for the device, with coeffi- cients of variation as high as 40 percent for some specimens. There was no clear trend between stiffness values and DCP, nor was there a clear trend between stiffness values and actual excavatability (by shovel). In general, the GeoGauge was not found to be an effective means of assessing the properties of CLSM, both because of poor reproducibility and inability to predict excavatability. The DCP index value, which indicates the penetration depth per blow, was measured for each of the excavation boxes. The minimum value for a recordable blow corresponded to a pen- etration of at least 25 mm. DCP values were found to decrease until the specimens ultimately suffered large cracks. After the large cracks appeared, the DCP values progressively decreased. Thus, the lowest index value was taken for each mixture and used in Table 3.31 because it represented the most difficult por- tion to excavate, thus providing a conservative index. The correlation between DCP index and the RE values (based on 240-day field-cured cylinders) is shown in Figure 3.11. As shown in the figure, a DCP index of 5 mm per blow correlated well with an RE value of 1.0. This correlation suggests that the DCP may be an effective, user-friendly method of assessing excavatability in the field. This approach was further investi- gated in the field testing component of this project (Chapter 4), where excavatability will be assessed not only using hand tools, but also using typical, commercial excavation equipment (i.e., backhoe). Another parameter that may potentially be used as an index for excavatability is the splitting tensile strength of CLSM. Some preliminary trials found that tensile strength may, in fact, be more suitable than compressive strength in assessing excavatability. Although splitting tensile tests were not performed on the C-series mixtures, some tests were per- formed on other mixture series. The results are provided in this section because of the potential of applying tensile data to excavatability predictions. The splitting tensile strengths of a range of CLSM mixtures (E-series) were measured, as shown in Table 3.32. A split cylin- der from mixture E-1 is shown in Figure 3.12. For the E-series CLSM mixtures, the splitting tensile strength to compressive strength ratio ranged from 9 percent to 17 percent, which is higher than those typically observed for conventional concrete. Unlike concrete, this ratio did not substantially decrease with an increase in compressive strength. Additional splitting tensile tests were performed using the E- series mixtures to assess the effects of drying on tensile strength and the tensile–compressive strength ratio. This testing was ini- tiated because drying generally has a more profound effect on tensile strength than compressive strength, at least in the case of conventional concrete. This behavior is generally attributed to the effects of microcracks. The results, shown in Table 3.33, confirmed that drying had a similar effect on CLSM, signifi- cantly lowering the tensile–compressive strength ratio. 46 Table 3.31. Results of follow-up excavatability study. MixtureTests C-1 C-2 C-3 C-4 C-5 C-6 C-7 C-8 C-9 28-day compressive strength (lab-cured) (kPa) 533 807 144 292 1027 332 1192 658 417 Density (kg/m3) 1512 1946 1724 1858 2094 2252 2143 2171 1660 240-day compressive strength (field-cured) (kPa) 1134 493 63 447 3290 362 1397 1615 199 RE (using 28-day lab- cured cylinders) 0.84 1.51 0.53 0.85 1.9 1.2 2.12 1.6 0.85 RE (using 240-day field-cured cylinders) 1.22 1.18 0.35 1.05 3.39 1.25 2.29 2.51 0.59 Modified RE (using 28-day lab- cured cylinders) 0.79 1.50 0.51 0.83 1.91 1.23 2.14 1.63 0.82 Modified RE (using 240-day field- cured cylinders) 1.16 1.17 0.34 1.03 3.42 1.29 2.32 2.55 0.57 DCP index (mm) 6.4 6.5 29 5.2 0.5 5 1.1 0.6 10 CBR (%) 37 36 7 46 100 48 100 100 22 Stiffness (using GeoGauge) (MN/m) 19.37 40.56 19.03 28.97 30.87 18.54 11.96 23.89 23.03 Relative ease of excavation (with shovel)a 3 7 1 6 9 8 8 10 4 aEach mixture was assigned an ease of excavation value ranging from 1 to 10, where 1 is easiest (able to excavate with minimal pressure applied to shovel) and 10 is most difficult (not able to excavate with shovel, even under heavy pressure).

California Bearing Ratio and Resilient Modulus CBR and resilient modulus of six CLSM mixtures were measured following modified AASHTO T 193 and T 292, re- spectively. Table 3.34 shows the measured CBR and resilient modulus values. With the exception of some mixtures that contained fly ash or high air content, observed CBR values were high, indicating that the tested mixtures would func- tion as a suitable base or subbase material. More important, the results and experience confirm that it is feasible to deter- mine CBR and resilient modulus values for CLSM using equipment commonly used to evaluate soils in typical test- ing laboratories. Water Permeability The water permeability (or hydraulic conductivity) test re- sults of six CLSM mixtures (I-series) are shown in Table 3.35. According to Bowles (1984), all of these permeability values (measured after 28 days of moist curing) were in the range of silty clays, silty or clayey fine sands, silts, clayey silts, and clays. Results indicate that water-cement ratio was an important fac- tor affecting the coefficient of permeability. Generally, per- meability decreased with decreasing water-cement ratio. In- terestingly, the high air content of mixture I-5 did not increase its permeability significantly, indicating that the entrained air bubbles were not well connected. The water permeability of the CLSM samples was easily measured using equipment commonly used to characterize soils. Additional information on the effect of freeze-thaw damages on water permeability of CLSM samples is provided in the section “Freezing and Thawing.” Triaxial Shear Strength Using the same materials and mixture proportions as the water permeability study, the triaxial shear strength of sev- eral CLSM mixtures was measured. The results, shown in 47 0.0 5.0 10.0 15.0 20.0 25.0 30.0 35.0 40.0 0 0.5 1 1.5 2 2.5 3 3.5 D CP In de x (m m pe r b low ) Modified RE (using field cured cylinders) y = 5.3212 x -1.7953 R2 = 0.928 Figure 3.11. Correlation between DCP index and RE calculated using 240-day compressive strength (field-cured cylinders). Table 3.32. Compressive and splitting tensile strengths at 7, 28, and 91 days. 7 days 28 days 91 days Mixture Average (kPa) C.O.V. (%) fst/fc (%) Average (kPa) C.O.V. (%) fst/fc (%) Average (kPa) C.O.V. (%) fst/fc (%) E-1 57.0 4.0 18.5 62.5 9.8 11.7 161.5 8.2 19.0 E-2 21.1 5.3 17.5 28.4 19.9 14.5 31.2 19.1 11.9 E-3 14.4 8.2 17.1 20.6 17.5 14.3 27.2 11.7 18.9 E-4 25.0 14.5 19.6 29.5 8.3 10.1 57.6 6.7 12.8 E-5 46.9 12.9 23.6 101.1 23.1 9.8 150.7 23.0 9.5 E-6 34.2 19.3 10.0 57.3 14.7 17.3 61.9 8.3 13.0 E-7 50.5 6.9 11.8 164.2 30.3 13.8 188.5 18.5 11.0 E-8 33.2 16.3 17.3 75.4 8.1 11.5 146.1 1.8 11.8 f'st = splitting tensile strength f'c = compressive strength

Table 3.36, confirmed the observation of other researchers (Bhat and Lovell 1996) that the strength of CLSM is com- posed of both chemical bonding and internal frictional re- sistance. For the mixtures investigated in the present study, the behavior was found to be a function of specific material and mixture proportions, and the effects changed with in- creased curing time. For mixtures I-1, I-2, and I-3, the internal friction angles and cohesion both increased with time (between 7 and 28 days). Their friction angles at 28 days were in the range of very dense granular soil, and the mixtures behaved like dense sand, with lower residual strengths than ultimate strengths. For mixture I-4, the strength development was manifested mainly as an increase in internal friction angle, whereas for mixture I-5, an 48 Figure 3.12. A cylinder from mixture E-1, before and after being tested for splitting tensile strength. Table 3.33. Effects of temperature and drying conditions on splitting tensile strength of CLSM. Mixture H-1 Mixture H-2 Condition Average (kPa) C.O.V. (%) fst/fc (%) Average (kPa) C.O.V. (%) fst/fc (%) 10°C, dry 16.7 23.1 6.5 6.0 20.2 5.1 10°C, wet 25.6 19.3 7.9 – – – 21°C, dry 23.8 8.6 4.9 8.8 39.6 6.1 21°C, wet 89.0 27.9 14.1 – – – 38°C, dry 55.0 18.8 7.3 18.9 11.4 7.2 38°C, wet 74.5 6.1 8.1 53.4 15.5 8.4 Mixture H-3 Mixture H-4 Condition Average (kPa) C.O.V. (%) fst/fc (%) Average (kPa) C.O.V. (%) fst/fc (%) 10°C, dry 100.4 14.1 7.2 65.5 33.0 7.3 10°C, wet 166.3 32.1 12.2 95.8 101.4 5.7 21°C, dry 49.2 8.5 5.8 32.2 5.0 7.0 21°C, wet 114.5 15.5 12.4 455.9 36.0 12.2 38°C, dry 87.8 7.5 8.4 158.6 6.3 7.7 38°C, wet 525.4 30.9 13.5 1791.7 12.5 15.6 Mixture H-5 Mixture H-6 Condition Average (kPa) C.O.V. (%) fst/fc (%) Average (kPa) C.O.V. (%) fst/fc (%) 10°C, dry 28.8 12.3 6.8 56.1 67.5 8.4 10°C, wet 133.0 7.5 9.8 84.5 22.8 9.2 21°C, dry 25.6 16.6 7.7 64.1 5.3 17.2 21°C, wet 127.8 8.6 8.5 142.8 7.3 15.4 38°C, dry 106.8 9.1 8.0 114.1 16.6 14.6 38°C, wet 198.2 6.0 7.5 125.3 12.5 12.6 f'st = splitting tensile strength f'c = compressive strength “–” = Not enough specimens were available for testing at this age.

increase in cohesion was the dominant factor. These mixtures (I-4 and I-5) had high air contents and exhibited behavior similar to loose or uncompacted sands. It was interesting to note that for mixture I-6, the friction angle decreased and co- hesion greatly increased with time. Drying Shrinkage As stated earlier, there are no standard methods to evalu- ate the drying shrinkage of CLSM and only limited emphasis was placed on this topic. A method developed in Germany to measure the shrinkage of conventional concrete for flooring applications was used without modification in this study to measure the shrinkage of CLSM mixtures. The results are shown in Table 3.37. The temperature during the testing pe- riod was 20°C and the relative humidity was approximately 60 to 65 percent. For mixtures without air entrainment, most of the shrinkage occurred during the first day, probably due to early bleeding and subsidence; however, the method used in this study could not detect this shrinkage. It is possible that all of the shrinkage could not be detected because CLSM did not exhibit sufficient early strength (or stiffness) to cause a detectable movement of the end anchor. More research is needed to examine drying shrinkage of CLSM and to develop a suitable test method. Because the topic of drying shrinkage was not identified as a critical issue for this project, no further research was performed on this topic. Durability Test Methods Corrosion Phase I, Uncoupled Samples. To evaluate the potential influence of resistivity, pH, fly ash type, fine aggregate type, water–cementitious materials ratio (w/cm), and cement content on the corrosion activity of ductile iron coupons embedded completely in CLSM or sand, the percent mass loss of coupons embedded in thirty different CLSM mixtures (and eight duplicated mixtures) was evaluated. The box plot showing the distribution of the percent mass loss values of the ductile iron coupons is given in Figure 3.13. Because mixtures 21 and 23 were not significantly different from other mixtures but the results obtained from them seem to be an anomaly, their data were not included in the statistical analysis. A multiple regression analysis and an analysis of variance were performed with the logarithm of percent mass loss data of the 36 CLSM samples as the response variable. Com- parison of all possible main effect models for the maximum adjusted R2 and minimum mean sum of error (MSE) indi- cates that the best model to predict mass loss of ductile iron pipe completely embedded in CLSM has three explanatory 49 Table 3.34. Resilient modulus and CBR values for selected CLSM mixtures after 28-day moist curinga. Mixture Regression Equation R 2 CBR (% ) 4 M r =3.00×10 10 (S d ) -3.3517 Sd Total Mr (kPa) (GPa) 276 199.45 345 66.23 414 52.52 0.9155 215.93 6 M r =3.00×10 6 (S d ) -2.6284 Sd Total M r (kPa) (GPa) 69 69.28 138 3.01 207 1.94 276 2.06 0.8486 175.83 23 M r =1.46×10 5 (S d ) -2.2978 S d Total Mr (kPa) (GPa) 34.5 57.61 69 4.79 103.5 3.49 138 2.36 0.9155 20.01 24 M r =1.00×10 8 (S d ) -2.8847 Sd Total Mr (kPa) (GPa) 138 134.32 207 20.63 276 11.19 345 8.47 414 4.74 0.9492 61.76 22r M r =3.06×10 2 (S d ) -0.4929 Sd Total Mr (kPa) (GPa) 207 23.33 276 17.99 345 16.6 414 15.77 483 15.25 552 13.54 0.9395 114.68 26 M r =6.57×10 5 (S d ) -1.4393 Sd Total Mr (kPa) (GPa) 414 104.44 483 89.48 690 72.93 828 33.25 0.8122 150.00 aConfining pressure 21 kPa. Sd = Deviator stress, Mr = Resilient modulus. Table 3.35. Water permeability of selected CLSM mixtures. Mixture Permeability(mm/s) I-1 2.46 × 10-3 I-2 5.33 × 10-4 I-3 1.45 × 10-4 I-4 4.20 × 10-3 I-5 6.75 × 10-3 I-6 2.89 × 10-4

50 Table 3.37. Drying shrinkage of selected CLSM mixtures. Shrinkage Strain (x 10-6)Time Mixture 4 Mixture 24 Mixture 23 Mixture 6 Mixture 22r Mixture 26 1 day 2260 2830 80 1440 90 10 2 days 2280 2850 80 1450 90 30 3 days 2280 2860 80 1450 90 50 4 days 2280 2860 80 1450 100 70 5 days 2300 2860 80 1460 100 100 6 days 2310 2880 80 1480 110 130 7 days 2330 2880 80 1500 120 160 2 weeks 2390 2930 160 1540 150 200 3 weeks 2410 2960 180 1590 150 190 4 weeks 2410 2980 180 1529 160 210 5 weeks 2410 2980 180 1600 160 220 6 weeks 2420 2960 180 1610 160 220 7 weeks 2410 2960 180 1600 – – “– ” = Not enough specimens were available for testing at this age. 0 1 2 3 4 5 All CLSM samples Pe rc en t m as s l os s Mixture 23 Mixture 21 Figure 3.13. Box plot of percent mass loss values. Table 3.36. Results of triaxial shear tests. 7 days 28 days Mixture Friction angle φ' (°) Cohesion c' (kPa) Friction angle φ' (°) Cohesion c' (kPa) I-1 36.14 31.8 42.81 40.1 I-2 36.07 96.1 38.73 174.7 I-3 39.35 251.7 47.86 346.2 I-4 21.99 43.9 23.92 43.1 I-5 19.48 89.8 18.47 130.0 I-6 37.30 44.4 33.86 93.4

variables—fly ash type, fine aggregate type, and w/cm—as shown below: where γ = 1.13, 1.07, 1.31, and 0.0 for bottom ash, concrete sand, foundry sand, and no fine aggregate, respectively λ = 0.47, 0.61, 0.69, and 0.0 for Class C, Class F, high-carbon, and no fly ash, respectively. The logarithm of the percent mass loss data is the response variable. The adjusted RP2P for this second model is 67 per- cent and its MSE is 0.0916. Appropriate coefficients should be used to predict the expected mean percent mass loss for spe- cific CLSM mixtures. The coefficients for the fine aggregate type and the fly ash type are significant at the 95 percent con- fidence level and the coefficient for the w/cm is significant at the 89 percent confidence level. Many field investigations on the corrosion of metals embed- ded in soils have reported that resistivity is a major controlling parameter affecting corrosion activity of the embedded metal (Spickelmire 2002, Kozhushner et al. 2001). Prior corrosion research in soils reported a non-linear relationship between mass loss and resistivity (Edgar 1989, Palmer 1989). How- ever, the evidence for such a non-linear relationship for the CLSM data in this study is very weak. The sand used in the control samples exhibited a resistivity of 3.1 × 104 Ω-cm and the average percent mass loss for the control group was 0.39 percent. Ductile iron coupons embedded entirely in CLSM exhibited lower corrosion activity than the ductile iron coupons embedded in the control sand even though the re- sistivity of the control sand material was higher than the re- sistivity of all the CLSM mixtures. This result is contradictory to conventional soil corrosion studies. Some utility agencies have voiced concern that the use of fly ash in CLSM could be detrimental to the corrosion per- formance of metals embedded in CLSM, because fly ash may cause a reduction in the pH, which could further result in higher corrosion activity. The results of this study indicate that the logarithm of the mean percent mass loss of mixtures without fly ash is statistically significantly higher than the mixtures with fly ash. This result indicates that the benefits of the fly ash on the microstructure and long-term passivation characteristics, as reported by Cao et al. (1994), likely have a more significant impact on corrosion performance than the relatively limited reduction in pH. The mean pH of the pore solution from the CLSM mix- tures evaluated in this study was 11.35. Although this high pH value was expected to decrease the corrosion activity of the ductile iron coupons, the results do not indicate a significant log % . . ( .10 0 056 0 0312 3 4mass loss w cm ( ) = − − + ⋅γ λ ) decrease in the percent mass loss as a result of the increased pH. As such, the pH of the pore solution alone does not seem to reliably estimate the corrosion performance of ductile iron coupons embedded in CLSM. Statistical analysis of the data indicated that the mean log- arithm of percent mass loss values for mixtures containing bottom ash, concrete sand, and foundry sand were statisti- cally not different from each other. However, the mean loga- rithm of percent mass loss data for the coupons embedded in mixtures without fine aggregates were statistically different and higher than the other mixtures containing fine aggre- gates. The decrease in percent mass loss could be due to re- ductions in the diffusivity, permeability, and/or porosity of the CLSM mixtures containing fine aggregates. As noted, the amount of cement used in CLSM mixtures is very low compared to the amount of water used. The sta- tistical analysis indicates that the cement content had no sig- nificant effect on the percent mass loss of the ductile iron coupons embedded in the CLSM mixtures. However, the re- sults indicated a slight increase in the logarithm of percent mass loss values with increasing water–cementitious materi- als ratio. Phase I, Coupled Samples. To evaluate the mass loss (i.e., corrosion performance) of the coupled ductile iron coupons embedded in both CLSM and sand, a similar statis- tical analysis as described in the Phase I uncoupled samples study was performed. This analysis indicated that a good pre- diction of mass loss using the explanatory variables—cement content, fine aggregate type, fly ash type, etc.—was not pos- sible for ductile iron coupons embedded in two different en- vironments (i.e., the coupled sample). The corrosion of uncoupled coupons was likely due to the formation of micro-galvanic corrosion cells on the sur- face of a single coupon. However, the major driving force of the corrosion of ductile iron coupons coupled in two dif- ferent environments was likely the formation of macro- galvanic corrosion cells due to the potential difference between the ductile iron coupons. Because these macro- galvanic cells were the major driving force of the corrosion of coupled coupons, factors that significantly affected the cor- rosion of uncoupled coupons were insignificant for coupled coupons. Figure 3.14 compares the logarithm of the distribution of percent mass loss of uncoupled coupons, coupled coupons, and the control group. The distributions are grouped by fly ash type. The figure indicates that the coupling of the ductile iron coupons has a significant impact: the mass loss of duc- tile iron coupons embedded in sand and CLSM (i.e., coupled) can be expected to be significantly larger than the mass loss of the coupons completely embedded in CLSM and the control group samples. 51

In general, the results of the Phase I study indicated the following: • The corrosion activity for ductile iron pipe coupons com- pletely embedded in CLSM was significantly lower than that of ductile iron pipe embedded in sand. • CLSM may provide more protection against corrosion initiation and propagation when metallic structures are completely embedded in CLSM compared to compacted sand. • Examination of the effects of the constituent materials on corrosion with a limited number of samples indicated that there was no significant difference between the fly ash types and the fine aggregate types used in this study. However, the corrosion of metal coupons in uncoupled samples that contained a fine aggregate or a fly ash was lower compared to the coupons in uncoupled samples without a fine aggregate or a fly ash. Phase II, Uncoupled Samples. Figure 3.15 shows the box plot showing the distribution and the median of the percent mass loss data of the 361 galvanized steel and ductile iron coupons embedded in CLSM mixtures exposed to distilled water and chloride solution. A multiple regression analysis and analysis of variance were performed on the data. The percent mass loss data were used as the response variable and the environment, fine aggregate type, fly ash type, resistivity, pH, metal type, water–cementitious materials ratio, percent chloride con- tent, and cement content were used as the explanatory vari- ables. Different possible models consisting of main effects and single interaction effects of the explanatory variables were applied to the data to find the best parsimonious model. Different models were compared using their adjusted coef- ficient of multiple determination (R2) and root mean square values. Models were applied to the observed percent mass loss values, to their square root transformation, and to their logarithm. Trials indicated that a logarithmic trans- formation was more effective in decreasing the observed de- pendence of variability of residuals on the values of re- sponse variable. Among the models evaluated for the logarithm of percent mass loss (LPML) values, the follow- ing model had the highest R2 value and smallest root means square error: The model includes the following relationships: • The main effects of classification variables: environment (α), fine aggregate type (β), fly ash type (γ), and metal type (φ) • The main effects of continuous variables: logarithm of resis- tivity (δ), pH (ε), and water-cementitious material (w/cm) ratio (τ) • The interaction effects of classification variables with clas- sification variables: fine aggregate type with metal type (ϕ), fly ash type with metal type (η), environment with metal type (λ) and fly ash type with environment (σ) • The interaction effect of a classification variable with a con- tinuous variable: logarithm of electrical resistivity with metal type (κ) and w/cm with metal type (ω). However, further evaluation of the model indicated that the assumptions of residuals being normally distributed log % . log 10 10 1 844mass loss r ( ) = + + + + +( ) • α β γ δ κ esistivity pH w cm ( ) + + + +( ) + + + + • • ε φ τ ω ϕ η λ σ ( . )3 5 52 -2.5 -2.0 -1.5 -1.0 -0.5 0.0 0.5 1.0 C-C C F-C F HC-C HC None-C None Control Lo g 1 0 (P erc en t M ass L os s) C: Type C Fly Ash F: Type F Fly Ash HC: High Carbon Fly Ash None: No Fly Ash Control: Sand -C: Indicates Coupled Samples Figure 3.14. Uncoupled versus coupled log mass loss versus control group.

and being independent of the predicted values of LPML were not satisfied very well. Because the assumption of con- stant variance was not satisfied, a weighted regression analysis was performed. The factors that had the largest effect on the LPML values were the environment and the metal type. The variances of the four groups obtained by separating the data by environment and metal type were calculated. The reciprocals of variances of these four groups were used as a weight variable for the weighted regression analysis. Evalu- ation of the studentized residuals of the weighted regres- sion indicated that the normality assumption of residuals was satisfied much better compared to the earlier regres- sion analysis. The R2 value for the weighted regression analysis is 67 percent and the root mean square error value is 0.98. All of the factors included in the model were statis- tically significant. The parameters defined in the model for the main effects of classification variables represent the expected value of the response variable for different levels of the corresponding classification variable, all other factors being the same. The parameters defined in the model for the main effects of con- tinuous variables represent the amount of change in the ex- pected value of the response variable for each unit change of the corresponding continuous variable, all other factors being the same. The interaction parameters in the model define how the response reacts to one variable based on the value or level of another variable. In the case of an interaction of a clas- sification variable with a continuous variable, the coefficient of the continuous variable is changed based on the level of the classification variable. The values of the parameters are given in Appendix A. In addition to the regression and variance analyses, com- parisons of LPML values for the different levels of classifica- tion variables were performed using Tukey’s comparison of means method at the mean values of three continuous vari- ables and at the 64 selected combinations of these three con- tinuous variables. Detailed information on the comparisons and the selection of combinations is given in Appendix A. Analysis indicated that pH was significantly and in- versely correlated to the observed LPML values. Environ- ment was also a significant variable for all the samples. The samples exposed to a chloride solution exhibited signifi- cantly higher LPML values compared to the samples ex- posed to the distilled water. The effect of environment for galvanized steel coupons was larger compared to the duc- tile iron coupons. There was a significant difference in the LPML values of dif- ferent metal types. For low water–cementitious materials ra- tios and logarithm of resistivity values, ductile iron coupons exhibited significantly lower LPML values. However, at higher water–cementitious materials ratios and with increasing log- arithm of resistivity, the difference in values became smaller and, at high enough values of these continuous variables, duc- tile iron coupons exhibited higher LPML values. The effects of different fly ash types and fine aggregate types were more important for samples with ductile iron coupons. Samples that contained a fine aggregate exhibited lower LPML values compared to the samples without fine ag- gregates regardless of the type of the fine aggregate. The dif- ference between the mean LPML values of samples contain- ing bottom ash and sand as fine aggregates was statistically not significant. The samples containing foundry sand as fine aggregate exhibited a mean LPML value between that of the samples with bottom ash or sand and the samples without fine aggregates. Because of the high LPML variability of sam- ples containing foundry sand, the difference between these samples and the samples without fine aggregates was not sta- tistically significant. 53 0 10 20 30 40 50 Pe rc en t M as s L os s Uncoupled Samples Figure 3.15. Percent mass loss distribution of metallic coupons.

Based on the materials used for this study, the results, from this phase only, indicate that the use of fly ash as a supple- mentary cementitious material may have adverse effects on the corrosion of embedded galvanized steel or ductile iron coupons, especially for the ductile iron coupons. Samples containing high-carbon fly ash or Class F fly ash exhibited higher LPML values compared to the samples without fly ashes, but the samples without fly ashes exhibited much larger variation. The mean LPML value of the samples con- taining Class C fly ash was lower than that of the samples with Class F or high-carbon fly ash but higher than that of the sam- ples without fly ash. However, due to the high variance of the samples without fly ash, the difference between the samples containing Class C fly ash and samples without fly ash was not statistically significant. Phase II, Coupled Samples. The histogram showing the percent mass loss of ductile iron and galvanized steel coupons embedded in CLSM and soil sections of coupled samples ex- posed to distilled water and chloride solution are shown in Figure 3.16. Analyses indicate that the percent mass loss values of metallic coupons embedded in CLSM and soil were signifi- cantly correlated and the mass loss values of coupons em- bedded in the soil section of samples were higher compared to the mass loss values of coupons embedded in the CLSM section of samples. For the coupled coupons, the mass loss is believed to be mainly due to galvanic corrosion taking place between the metallic coupons embedded in different sections. The significantly higher mean percent mass loss values ex- hibited by the metallic coupons in the soil section indicate that these coupons were anodes and the coupons in the CLSM section were cathodes. Because the metallic coupons embedded in the soil sections of coupled samples represent the critical anodic areas of pipes for corrosion damage, fur- ther statistical analysis was performed on the percent mass loss data of these coupons. The explanatory variables evaluated for the percent mass loss of coupons included environment, metal type, soil type, fine aggregate type, fly ash type, resistivity of CLSM, resis- tivity of soil, pH of CLSM, chloride content of the CLSM, and chloride content of the soil. Different possible models consisting of main effects and single interaction effects of the explanatory variables were applied to the data to find the best parsimonious model. Different models were compared using their adjusted coefficient of multiple determination (R2) and root mean square values. Models were applied to the observed percent mass loss values, to their square root transformation, and to their logarithm. Trials indicated that a logarithmic transformation was more effective in decreas- ing the observed dependence of variability of residuals on the values of response variable. Among the models evalu- ated for the logarithm of percent mass loss values, the fol- lowing model had the highest R2 value and smallest root mean square error: The coefficients α, β, δ, ε, and γ are assigned values for the different levels of the classification variables: environment (α), soil type (β), fine aggregate type (δ), fly ash type (ε), and metal type (γ), respectively. The coefficients φ, ϕ, η, λ are as- signed values for the two-factor interactions of classification variables: environment with metal type (φ), environment with soil type (ϕ), fly ash type with metal type (η), and fine log % . ( 10 0 97 3 mass loss( ) = + + + + + + + + + α β δ ε γ φ ϕ η λ . )6 54 0 100 200 300 400 500 600 0 12 24 36 48 60 72 84 Fr eq ue nc y Percent Mass Loss Figure 3.16. Percent mass loss of metallic coupons.

aggregate type with soil type (λ). The values of these coeffi- cients are given in Appendix A. Analysis showed that the overall model and all the factors included in the model were significant with the exception of metal type. However, because interactions of other variables with metal type were significant, this factor was left in the model for a complete hierarchy. All assumptions of the re- gression analysis were satisfied and the coefficient of deter- mination (R2) of the model was 35 percent. This low R2 value indicates that a model solely built from these variables can- not be used to estimate the corrosion of metallic coupons with great accuracy. Results indicated that samples exposed to chloride environ- ment exhibited higher mean LPML values compared to the samples exposed to distilled water. The disturbance of the pas- sive layer formation on the steel surface by chloride ions and the low resistivity of CLSM and soil samples exposed to chlo- ride solution could both be the reasons for the higher mean LPML values of samples exposed to chloride environment. Water–cementitious materials ratio had a statistically sig- nificant but small correlation with the chloride content in CLSM and a negative correlation with the logarithm of resis- tivity of CLSM. Results indicated that coupons embedded in clay (soil sec- tion of coupled samples) exhibited statistically significantly higher LPML values compared to the coupons embedded in sand in both chloride and distilled water environments. How- ever, the effect of environment was greater on the coupons that were embedded in sand compared to the coupons em- bedded in clay. Analysis also indicated that the resistivity and pH of clay samples were lower compared to the resistivity and pH of sand. Although the metal type was overall not a statistically sig- nificant factor, analysis indicated that galvanized steel cou- pons exhibited a significantly higher mean LPML value com- pared to ductile iron coupons in a chloride environment. The observed effect of fly ash on the LPML was contradic- tory to the findings of the uncoupled samples. Results indi- cated that, among the coupled samples, CLSM sections with fly ashes exhibited lower mean LPML values compared to the CLSM sections without any type of fly ash (similar to the Phase I study). However, the difference between the LPML values of CLSM sections containing fly ash and without fly ash was only statistically significant for the ductile iron coupons. CLSM sections with bottom ash or foundry sand exhibited significantly higher LPML values compared to the CLSM sec- tions with sand or without fine aggregates. Among the cou- pled samples that had clay in their soil section, CLSM sections with bottom ash exhibited the highest mean LPML value and among the coupled samples that had sand in their soil sec- tion, CLSM sections with foundry sand exhibited the highest mean LPML value. In general, results from the Phase II study indicated the following: • pH was significantly and inversely correlated to the ob- served LPML values. • Environment was a significant variable for all the samples. • The samples exposed to a chloride solution exhibited sig- nificantly higher LPML values compared to the samples exposed to the distilled water. • The effect of environment for galvanized steel coupons was larger compared to the ductile iron coupons. • There was a significant difference in the LPML values of different metal types. • For low w/cm and logarithm of resistivity values, ductile iron coupons exhibited significantly lower LPML values. • At higher w/cm and with increasing logarithm of resistivity, the difference in values became less and, at sufficiently high values, ductile iron coupons exhibited higher LPML values. • The effects of different fly ash types and fine aggregate types were more important for samples with ductile iron coupons. • Samples that contained a fine aggregate exhibited lower LPML values compared to the samples without fine aggre- gates regardless of the type of the fine aggregate. • The difference between the mean LPML values of samples containing bottom ash and sand as fine aggregates was sta- tistically not significant. • The samples containing foundry sand as fine aggregate ex- hibited a mean LPML value between the samples with bot- tom ash or sand and the samples without fine aggregates. • Because of the high LPML variability of samples contain- ing foundry sand, the difference between these samples and the samples without fine aggregates was not statisti- cally significant. • The use of fly ashes may have adverse effects on the corro- sion of embedded galvanized steel or ductile iron coupons, especially for the ductile iron coupons. • Samples containing a high-carbon fly ash or Class F fly ash exhibited higher LPML values compared to the samples without fly ashes, but the samples without fly ashes exhib- ited much larger variation. • The mean LPML value of the samples containing Class C fly ash was lower than that of the samples with Class F or high- carbon fly ash but higher than that of the samples without fly ash. However, because of the high variance of the sam- ples without fly ash, the difference between the samples con- taining Class C fly ash and samples without fly ash was not statistically significant. The following general conclusions were obtained from both phases of the study: • The metallic coupons embedded in the soil section of cou- pled samples exhibited significantly higher percent mass 55

loss values compared to the coupons embedded in uncou- pled samples. • Because the main driving force of corrosion is the poten- tial difference in the coupled samples, the significance of the factors that affected the corrosion in uncoupled sam- ples was generally lower for coupled samples. Service Life of Ductile Iron and Galvanized Steel Coupons Completely Embedded in CLSM ASTM G 1 provides a formula to predict the corrosion rate of metallic samples. By placing the LPML values ob- tained from the statistical model shown in Equation 3.5 into the formula given in ASTM G 1, a service life model for duc- tile iron and galvanized steel pipes completely embedded in CLSM can be derived. Assuming that the useful service life of the pipe will be over at the first perforation of the pipe wall due to corrosion, the service life of a pipe completely embedded in CLSM can be calculated using the following formula: where SL = the service life (years) D = the outside radius (cm) t = the pipe wall thickness (cm) LPML() = the logarithm of percent mass loss obtained from Equation. 3.5 To obtain the LPML value from Equation 3.5, the values of the classification variables and the values of the three continuous variables (w/cm, resistivity, and pH) must be specified. However, only the values of the classification variables, such as fly ash type, fine aggregate type, environ- ment, etc., can be specified. The values of the continuous variables are dependent values, i.e., they cannot be speci- fied; they can only be measured from the samples of de- signed CLSM mixtures. Therefore, Equation 3.7 cannot be used to calculate a specific service life for a designed CLSM mixture. However, the formula can be used to perform a risk analysis by using different combinations of the contin- uous variables in the LPML formula. The data obtained in this study can be used to obtain an estimate of the expected range of the continuous variables for different levels of the classification variables. It should also be noted that the coefficients of the LPML model were determined using a weighted regression analysis. The variance of each residual group that was used to deter- mine the weight variables of the analysis can be used to ob- SL D t x D D tLPML = ⋅ ⋅ − −( )⎡⎣ ⎤⎦ −7978 10 10 3 7 2 2 2() ( . ) tain a distribution around the obtained service life value. Equation 3.8 shows how to obtain the required percentile of the LPML value using the variance and the LPML value ob- tained from Equation 3.5: where LPMLPr. = the LPML for which probability of LPML < LPMLPr. is Pr. −1 = the inverse standard normal distribution function After the values of the classification variables for a specific CLSM design are determined and a combination of the levels of continuous variables is chosen, a service life distribution graph can be generated for a ductile iron or galvanized steel pipe completely embedded in the specific CLSM mixture as shown in Figure 3.17. Calculation of service life estimates for the galvanized and ductile iron pipes embedded in the specific CLSM mix- tures that were evaluated in this study indicated that prop- erly designed CLSM mixtures can provide a service life for ductile iron pipes similar to that in conventional backfill materials. Therefore, in selecting between CLSM and conventional backfill materials, factors other than service life—such as material cost, construction cost, construction time, and long-term settlement—should be considered. However, results indicated that the galvanized steel pipes completely embedded in CLSM can be expected to have service life values comparable to the galvanized steel pipes embedded in severely or moderately corrosive soils with low resistivity and pH values. Therefore, backfilling bare galvanized steel pipes with CLSM mixtures is likely not warranted. Freezing and Thawing Two studies were performed to evaluate the freeze-thaw resistance of CLSM mixtures. In the first study, six CLSM mixtures from the initial mixture series were tested using a modified version of ASTM D 560, originally developed for the assessment of soil-cement. The second study used the same method to assess a wider range of CLSM mixtures (D-series) and evaluated the effects of freeze-thaw damage on permeability. Figure 3.18 shows the measured percent mass loss values plotted against the number of freeze-thaw cycles. The per- cent mass loss values shown in the figure were calculated as- suming that the moisture content of all the specimens were constant throughout the test. According to the “soil-cement LPML LPML VariancePr. Pr. ( . )= + ( ) ×−Φ 1 3 8 56

laboratory handbook” the mass loss after 12 freeze-thaw cy- cles should not exceed 14 percent for a Group A-1 soil (PCA 1992). Soils in group A-1 are coarse grained and low in con- tent of fines. Most of the CLSM mixtures (especially the ones that were cured for 28 days) containing high amounts of air satisfied this criterion. ASTM D 560 (with minor modifications) was found to be an effective and easy-to- perform method to assess the freeze-thaw resistance of CLSM mixtures. The permeability or hydraulic conductivity of the D- series CLSM mixtures before and after exposure to freeze- thaw cycles is shown in Table 3.38. It is interesting to note that mixture D-1 (high-air mixture with 30 kg/m3 of cement) did not survive the 12 cycles, whereas a similar mixture with 45 kg/m3 of cement (D-11) did survive the en- tire 12 cycles, suggesting that both air-void system and strength contribute to freezing and thawing resistance. Mixture D-10 survived all 12 cycles, most likely due to its higher strength (contributed from the Class C fly ash). The remaining mixtures (D-2 through D-9) did not survive all 12 cycles. Mixtures were selected that would likely suffer freezing and thawing damage, allowing for the measure- ment of changes in permeability (before and after testing, as shown in Table 3.38). However, the effects of freezing 57 0 20 40 60 80 100 0 50 100 150 200 250 Pr (S erv ice lif e < X ) Service Life (years) Figure 3.17. Probability distribution of service life. -10 0 10 20 30 40 50 60 0 2 4 6 8 10 12 14 #4 cured for 7 days #6 cured for 7 days #22r cured for 7 days #26 cured for 7 days #4 cured for 28 days #23 cured for 28 days #6 cured for 28 days #22r cured for 28 days #25 cured for 28 days M as s L os s ( %) Number of Cycles Figure 3.18. Mass losses vs. freeze-thaw cycles for selected CLSM mixtures.

and thawing damage on water permeability were somewhat inconclusive, with some mixtures showing increased per- meability and others showing decreased permeability. This result was most likely due to the test setup, which was designed to keep the samples intact, thus allowing for sub- sequent permeability testing. However, keeping the sam- ples intact (and confined) may not have allowed for an ac- curate estimate of in-situ permeability. Because the samples were confined, the expansion due to freezing and thawing may have actually compacted the samples, resulting in an apparent reduction in permeability. More work is needed to elucidate the effects of freezing and thawing damage on permeability. Initially, the composition of the water flowing through the sample was to be analyzed to deter- mine if freezing and thawing damage increased the leach- ing of constituent materials, specifically heavy metals. However, after analyzing the raw materials used in the study (as discussed in the next section), the researchers de- termined that the materials used were intrinsically non- toxic. Thus, the effluent from the freeze-thaw samples was not analyzed. In general, the results of the freeze-thaw testing indicated that CLSM mixtures can be efficiently tested for freeze-thaw resistance following the modified ASTM D 560 with 12 cycles. Results also indicated that CLSM mixtures with high air con- tent and high compressive strength exhibited good freeze- thaw resistance. Leaching and Environmental Impact Table 3.39 summarizes the total concentration of heavy metals present in the by-product materials used in this study. These results represent the total concentration of the eight key heavy metals. A “rule of thumb” that some practi- tioners use is that the concentration of total heavy metals can be up to 20 times the standard TCLP limits. In this study, arsenic concentration in bottom ash, Class C fly ash, and Class F fly ash exceeded this “rule of thumb” value. Thus, additional testing was performed (using the TCLP method) to determine the actual amount of heavy metals that are available to leach from these materials. Because the foundry sand and high-carbon fly ash did not have signifi- cant amounts of total heavy metals, the materials were clas- sified as non-toxic, and no subsequent leaching tests were performed. The TCLP results for Class C fly ash, Class F fly ash, and bottom ash are shown in Table 3.40. The concentration of heavy metals that leached from each material was well below the EPA-recommended TCLP limits; therefore, the materials were classified as non-toxic and suitable for use in CLSM. If any of the by-product materials had exhib- ited significant leaching of heavy metals (above the TCLP limits), the last step would have been to assess the actual leaching of heavy metals from CLSM containing the mate- rial(s) using the American Nuclear Society leachate test (ANS 16.1). 58 Table 3.38. Frost resistance of CLSM (using modified ASTM D 560). MixtureMeasurement D-1 D-2 D-3 D-4 D-5 D-6 D-7 D-8 D-9 D-10 D-11 Original mass (kg) 1.38 1.77 1.70 1.86 1.53 1.47 1.46 1.50 1.55 1.56 1.33 Moisture content (%) 9.8 10.5 13.6 10.1 30.7 35.5 32.4 27.4 28 25.1 7.3 28-day strength (MPa) 0.13 1.02 0.79 1.47 0.11 0.12 0.20 0.38 0.25 0.53 0.34 1 cycle (%) 105.2 99.0 100.1 99.9 D 92.8 93.0 98.9 98.1 99.2 103.5 2 cycle (%) 102.3 95.3 97.3 99.8 D D 94.9 94.8 98.1 105.9 3 cycle (%) 84.4 85.4 78.4 99.1 91.5 87.1 98.0 108.1 4 cycle (%) 74.1 75.7 73.2 96.4 88.0 83.6 98.1 110.3 5 cycle (%) 66.4 72.4 71.1 88.4 84.6 79.1 97.7 110.3 6 cycle (%) 59.6 55.4 63.8 61.0 81.4 73.5 97.3 110.3 7 cycle (%) 51.5 54.9 52.1 59.7 76.9 69.2 97.1 108.8 8 cycle (%) 45.8 46.3 54.3 57.9 73.7 62.1 95.8 108.2 9 cycle (%) 37.3 30.1 48.6 43.8 68.1 56.3 94.4 104.6 10 cycle (%) 33.6 D D 40.3 56.1 D 92.2 103.6 11 cycle (%) 25.8 D 49.8 89.4 100.6 12 cycle (%) D D 85.1 97.5 Final moisture content (%) 28.7 21.3 Dry mass loss (%) 11.0 11.9 Permeability before 12 F-T cycles (× 10-2 mm/s) 7.38 0.35 0.21 13.02 6.60 3.87 1.63 1.60 3.08 3.63 8.06 Permeability after 12 F-T cycles (× 10-2 mm/s) 14.21 1.90 1.61 8.94 0.08 0.94 1.72 0.42 0.73 0.36 5.94 D = Damaged.

This systematic approach to testing leaching potential and environmental impact can be followed for any material being considered for use in CLSM. Although all the materials used in this study were deemed non-toxic, it may be possible that certain materials considered for a given CLSM application may be more of an environmental concern. Summary This chapter described a comprehensive laboratory pro- gram focusing on CLSM, with emphasis on developing/ recommending appropriate test methods to assess key CLSM properties and understanding the impact of materials, mix- ture proportions, and curing regime on performance. Based on the results presented in this chapter, the following general conclusions can be drawn: • Suitable test methods exist to measure most of the key CLSM properties affecting performance in the four target applications. The findings discussed in this chapter, cou- pled with the results from the field testing program (Chap- ter 4), helped to develop the test methods (Appendix B) and specifications (Appendix C). • Models were developed to predict the water demand and compressive strengths for a range of CLSM mixtures. This information can be helpful in designing mixtures for ap- plications where strength may be a key limiting factor, such as in the use of excavatable CLSM. • Improvements were made to the ASTM D 4832 (uncon- fined compressive strength) test method to increase its ac- curacy and improve its user-friendliness. • The effects of temperature on strength gain of CLSM mix- tures can be very pronounced, especially when using Class C fly ash. One should be aware of this increased strength gain, especially when CLSM is being used in a hot climate. Keeping this strong temperature dependence in mind and accounting for it in design can help to effectively produce excavatable CLSM. Trial batching and testing at elevated temperatures helps to gain insight into long-term strength gain in field applications, especially when fly ash or other supplementary cementing materials are used. • There is no single parameter that adequately predicts exca- vatability. Compressive strength can serve as a useful sur- rogate value in some cases, but one should try to capture the long-term strength gain when applying strength as a pre- dictive tool. Basing long-term strength gain on short-term laboratory testing can be problematic for some CLSM mix- tures (especially those containing fly ash). Calculating a re- movability modulus (RE) shows promise in predicting ex- cavatability. Lastly, the dynamic cone penetrometer (DCP) was found to be a valuable method of assessing CLSM in the field and estimating ease of excavatability. 59 Table 3.39. Analysis of heavy metal concentration of raw material extracts. Element TCLP Limit (mg/L) 20 x TCLP Limit (mg/L) Bottom Ash (mg/L) Foundry Sand (mg/L) Class C Fly Ash (mg/L) Class F Fly Ash (mg/L) High-Carbon Fly Ash (mg/L) Arsenic 5.0 100 170.0a 7.7 280.0a 160.0a 58.0 Barium 100.0 2000 2000.0 240.0 1300.0 320.0 1200.0 Cadmium 1.0 20 0.23 0.28 1.55 2.1 0.51 Chromium 5.0 100 10.0 18.0 87.0 96.0 16.0 Lead 5.0 100 <0.2 18.0 <0.2 37.0 <0.2 Mercury 0.2 4 <0.2 <0.2 <0.2 <0.2 <0.2 Selenium 1.0 20 <0.2 <0.2 <0.2 2.4 <0.2 Silver 5.0 100 <0.2 <0.2 <0.2 <0.2 <0.2 aConcentration exceeded the “rule of thumb” value of 20 times the TCLP limit. Table 3.40. TCLP test results for Class C, Class F, and bottom ash. Element TCLP Limit (mg/L) Bottom Ash (mg/L) Class C Fly Ash (mg/L) Class F Fly Ash (mg/L) Arsenic 5.0 0.12 0.074 0.37 Barium 100.0 3.61 0.30 0.17 Cadmium 1.0 0.001 0.004 0.024 Chromium 5.0 0.01 0.29 0.11 Lead 5.0 <0.01 <0.01 <0.01 Mercury 0.2 <0.01 <0.01 0.11 Selenium 1.0 <0.01 0.37 0.02 Silver 5.0 <0.01 <0.01 <0.01

• Significant research was performed on the corrosion of metallic pipe materials embedded in CLSM. In general, CLSM was found to be beneficial in reducing corrosion (compared to typical compacted fill) when pipes are com- pletely embedded in CLSM. The reduced permeability of CLSM can reduce the ingress of chlorides and the micro- structure of CLSM can improve corrosion resistance through changes in the pH and resistivity of the pore solu- tion. There is a potential for corrosion when pipes are em- bedded in both CLSM and surrounding soil or conventional fill, setting up a galvanic cell than can increase corrosion ac- tivity. This situation is similar in nature to metals embedded in dissimilar soils, and similar precautions can be taken to ensure the desired service life. • The by-product materials tested in this study were found to be non-toxic. However, a testing program was pro- posed to evaluate other by-product materials that might be more of a concern with regard to leaching and envi- ronmental impact. This method involves the testing of total heavy metals, possibly followed by TCLP (if the total heavy metals are above certain threshold values), and pos- sibly followed by leachate testing from CLSM containing the subject material (if the TCLP values exceed certain thresholds). 60

Next: Chaper 4 - Field Evaluations of CLSM »
Development of a Recommended Practice for Use of Controlled Low-Strength Material in Highway Construction 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 597: Development of a Recommended Practice for Use of Controlled Low-Strength Material in Highway Construction explores the use of controlled low-strength material (CLSM) in highway construction applications, in particular, as backfill, utility bedding, and void fill and in bridge approaches. The report also examines a recommended practice for the use of CLSM that was developed through a series of full-scale field experiments.

This report presents the full text of the contractor’s final report of the project and three of the five appendices, which present the test methods (Appendix B), specifications (Appendix C), and practice (Appendix D) recommended for implementation. The corrosion study (Appendix A) and implementation plan (Appendix E) are available online as NCHRP Web-Only Document 116.

There is a summary document, Paths to Practice, available.

  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!