Cast-in-Place Concrete Connections for Precast Deck Systems (2011)

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.

370 Chapter 11 Selection of Durable Closure Pour Materials for Accelerated Bridge Construction 11.0 Introduction For precast bridge deck systems with cast-in-place (CIP) connections, precast elements are brought to the construction site ready to be set in place and quickly joined together. Then, a concrete closure pour (CP) completes the connection. The performance of the CP material is one of the key parameters affecting the overall performance of the bridge system. Longitudinal connections between the flanges of DBTs and between precast panels between girders require that the joints must be able to transfer shear and moment induced by vehicular loads. Shrinkage of CP materials and transverse shortening of precast members further subject the joints to direct tension. Freeze- thaw resistance and low permeability of joints are also important. An ideal CIP connection detail emulates monolithic behavior and results in a more durable and longer lasting structure. Traditionally, different types of grouts have been used as CP materials for precast bridge deck systems with CIP connections. Mrinmay (1986) documented a wide variety of materials used after 1973 to avoid joint failure in closure pours. These materials included sand-epoxy mortars, latex modified concrete, cement- based grout, non-shrink cement grout, epoxy mortar grout, calcium aluminate cement mortar and concrete, methylmethacrylate polymer concrete and mortar, and polymer mortar. Epoxy or polymer modified grouts can have significant advantages, such as a high strength of 10 ksi in 6 hours, better bond, reduced chloride permeability, and lower shrinkage (Issa et al 2003) than different MAP grouts. However, they are often significantly more expensive and less compatible with surrounding concrete. In addition, if the resin is used in too large of a volume, the heat of reaction may cause it to boil, and thereby develop less strength and lose bond. Cementitious grouts have been used more in precast construction than have epoxy or polymer- modified grouts (Matsumoto et al. 2001). A primary disadvantage of cementitious grouts is the shrinkage and cracking that result from the use of hydraulic cement. Non-shrink grout compensates for the shrinkage by incorporating expansive agents into the mix. With non-shrink grout, the effects of shrinkage cracks or entrapped air on the transfer of forces and bond are minimized, though not eliminated. ASTM C 1107 establishes strength, consistency, and expansion criteria for prepackaged, hydraulic-cement, non-shrink grout. Nottingham (1996) reported that the very nature of Portland cement grouts virtually assures some shrinkage cracks in grout joints, regardless of quality control. Prepackaged magnesium ammonium phosphate (MAP) based grout often extended with pea gravel can meet requirements, like high quality, low shrinkage, impermeable, high bond, high early strength, user friendly and low temperature curing ability (Nottingham 1996; Issa et al. 2003). Gulyas et al. (1995) undertook a laboratory study to compare composite grouted keyway specimens using two different grouting materials: non-shrink grouts and magnesium ammonium phosphate (MAP) mortars, in which MAP materials performed better than non- shrink grouts. Gulyas and Champa (1997) further examined inadequacies in the selection of a traditional

371 non-shrink grout for use in shear keyways. The MAP grout outperformed the non-shrink grout in all areas tested, including direct vertical shear, direct tension, longitudinal shear, bond, shrinkage, etc. Menkulasi and Roberts-Wollmann (2005) presented a study of the horizontal shear resistance of the connection between full-depth precast concrete bridge deck panels and prestressed concrete girders. Two types of grout were evaluated: a latex modified grout and a MAP grout. For both types of grout, an angular pea gravel filler was added. The MAP grout developed slightly higher peak shear stresses than the latex modified grout. Grout without coarse aggregate extension is usually referred to as neat grout, while grout with coarse aggregate extension, typically 1/2 or 3/8 in. coarse aggregate, is extended grout. Compared with neat grout, extended grout has the following potential benefits: (1) more compatible with concrete; (2) better interlock between connection components; (3) denser, less permeable; (4) less drying shrinkage and creep; and (5) larger grout volume per bag, hence less expensive. However, it was pointed out by Matsumoto et al. (2001) that the extended grout required more cement paste than available in prepackaged bags, leading to lower strengths and poor workability. As discussed above, numerous products are available for CP materials, and various materials were studied. However, limited research had been previously conducted to provide consistent comparison among a large number of different types of CP materials. Also adequate performance-based criteria need to be developed to ensure appropriate selection of CP materials, particularly for accelerated bridge construction, which is the subject of this chapter. Performance-based specifications focus on properties such as consistency, strength, durability, and aesthetics. They reward quality, innovation, and technical knowledge, in addition to promoting better use of materials, and present an immense opportunity to optimize the design of materials. As part of the process of developing the performance criteria, eight candidate CP materials were selected and evaluated with respect to their potential effectiveness in accelerated bridge construction. In this chapter, accelerated bridge construction is defined with respect to two categories: overnight cure of CP materials and 7-day cure of CP materials. For the overnight cure, published performance data from different grout materials were collected through contacts with material suppliers and users. For the 7-day cure, standard or special concrete mixtures and their performance data were collected through contacts with HPC (High Performance Concrete) showcase states as well as with material suppliers. Based on these initial collected data, four grouts were first selected as candidate overnight cure materials, and four special concrete mixes as candidate 7-day cure materials. The preliminary selection was based on strength tests of selected materials or prediction models to narrow the candidate materials down to two materials in each of the two categories. Then long-term tests were performed on the four final selected materials, including freezing-and-thawing durability, shrinkage, bond, and permeability tests. The final performance criteria for selecting durable CP materials were developed based on results of these long-term tests.

372 11.1. Preliminary Performance Criteria Performance characteristics, compressive strength, shrinkage, chloride penetration, freezing-and-thawing durability and bond strength, were investigated as performance criteria. For the closure pour/precast unit interface, the focus must be on minimizing cracking in this location to reduce intrusion of water that may result in corrosion. And thus, shrinkage, chloride penetration, freezing-and-thawing durability and bond strength need be investigated to control cracking and corrosion. An extensive literature review was performed to develop preliminary performance criteria of overnight and 7-day cure CP materials. The FHWA defined a set of concrete performance characteristics for long-term concrete durability and strength of highway structures. Standard laboratory tests, specimen preparation procedures, and grades of performance were suggested for each characteristic. Because standard test methods sometimes offer different options, Russell and Ozyildirim (2006) modified the FHWA definition. The modified performance characteristic grades for high-performance structural concrete are given in Table 11.1.1. Tepke and Tikalsky (2007) provided a working guide to the design and construction of concrete structures using attainable high standards rather than common practice. An engineering design tool for the development of performance specifications for reinforced concrete highway structures was developed and performance characteristic grades for HPC are given in Table 11.1.2. As shown in Tables 11.1.1 and 11.1.2, the same or similar standard laboratory tests were recommended. Also three grades were suggested in both criteria. The performance criteria identified in Table 11.1.2 have lower requirements for compressive strength and chloride penetration, and higher requirements for shrinkage than the criteria in Table 11.1.1 in all three grades. They have similar grade limits for freezing-and-thawing durability. Both criteria were developed generally for bridges including girders and decks.

373 Table 11.1.1: Proposed performance characteristic grades by Russell and Ozyildirim (2006) Performance characteristic* Test Method Grade 1 Grade 2 Grade 3 Compressive Strength** (CS), ksi AASHTO T22 ASTM C 39 8≤CS<10 10≤CS<14 14≤CS Shrinkage*** (S), με AASHTO T160 ASTM C157 600≤S<800 400≤S<600 S<400 Chloride Penetration (ChP), coulombs AASHTO T277 ASTM C1202 1500<ChP≤2500 500<ChP≤1500 ChP≤500 Freezing-and-thawing Durability (F/T) (relative dynamic modulus of elasticity after 300 cycles) AASHTO T161 ASTM C666 Procedure A 70%≤ F/T<80% 80%≤ F/T<90% 90%≤ F/T *All tests to be performed on concrete samples moist- or submersion-cured for 56 days until otherwise specified. **The 56-day strength is recommended. *** Shrinkage measurements are to start 28 days after moist curing and be taken for a drying period of 180 days. Table 11.1.2: Performance characteristic grades by Tepke and Tikalsky (2007) Performance Characteristic Test Method Grade 1 Grade 2 Grade 3 Compressive Strength (CS), ksi AASHTO T22 3.5≤CS<8.0 @ 28 days 8.0≤CS @ 28 days 3.5≤CS @ early ages Shrinkage (S), με ASTM C157 S≤600 @ 56 days S≤400 @ 56 days S≤200 @ 56 days Chloride Penetration (ChP), coulombs AASHTO T277 ChP≤4000 @ 56 days ChP≤1500 @ 56 days ChP≤800 @ 56 days Freezing-and-thawing Durability (F/T) (relative modulus after 300 cycles) AASHTO T161 Proc. A after 28 days moist curing and 7 days air drying 60%≤F/T 80%≤F/T 90%≤F/T

374 There exist some practical difficulties in implementing the above performance criteria. For example, MAP grouts like EUCO-SPEED MP and Set® 45 HW should be air cured for 8 hours, as overnight cure materials, while HPC should be cured for 7 days, as 7-day cure materials, by both the membrane-forming compound method and the water method with burlap. However, proposed test methods in Table 11.1.1 and 11.1.2 have a very different curing scheme. The test methods need also be modified based on the following considerations. For the shrinkage, when shrinkage occurs after initial moist curing, concrete starts to develop stiffness as measured by the modulus of elasticity. High performance concretes often have low w/cm (water to cementitious materials ratios) and high stiffness as a result. If the shrinkage strains are high enough, they simply crack due to the restraint, the stiffness, and the drying shrinkage. For CP materials in floor systems, the restraint is developed due to the internal reinforcing steel, especially the steel that runs through the construction joint in the existing concrete member into the next cast adjacent concrete member or section. This creates a tremendous “racking restraint” that does not allow the second adjacent slab to shorten during cooling from hydration heat and also due to later developing drying shrinkage. The AASHTO PP34-99 (1998) restrained shrinkage ring test can test the crack potential, and should be used instead of the ASTM C157 test. There are also issues with the ASTM C1202 rapid chloride permeability test (RCPT). The RCPT has some interference problems with materials such nitrate corrosion inhibitors and even Set® 45. Part of the problem is the epoxy coating that must be bonded to the exterior side walls of the core. The coating does not work well with the material and cannot block the chloride from running through the specimen. To avoid this issue, the ASTM C1543 ponding test is recommended to be used to determine the chloride gradient. To aid in the selection of candidate CP materials for long-term durability tests, reasonable preliminary performance limits were specified first based on extensive literature reviews as well as the following considerations. To expedite construction and reduce cost, it is desirable to minimize the width of the joint zone. The headed bar detail with a 6 in. lap length was recommended as the improved longitudinal joint detail for DBT bridges by Li et al. (2010). And it was found that a certain compressive strength was needed to develop the headed bars within a short overlap length (Li et al. 2010). A criterion of 6 ksi was proposed for the compressive strength of the CP materials. CDOT Specifications Committee (2005) specified Class H concrete used for bare concrete bridge decks must not exhibit a crack at or before 14 days in the cracking tendency test (AASHTO PP 34). For the chloride threshold level (CTL) for steel corrosion in concrete, Glass and Buenfeld (1995) summarized various research studies and found CTLs to vary from 0.17 to 2.5% (by mass of cement) with a value of 0.2% chosen as a good prediction of CTL for harsh environments. A depth of 1.5 in., corresponding to the minimum concrete cover for #5 and smaller bars cast in CIP concrete exposed to earth or weather by ACI 318-08, was proposed as the maximum depth of chloride penetration for percent chloride of 0.2% by mass of cement after 90-day ponding. To determine the required bond strength, results of the parametric study on joints of DBT bridges described in Chapter 10 was used. In that study, it was found that the maximum shear stress at joints due to live loads was 84 psi (6.09 kip/ft). Thus, a higher limit of 200 psi was proposed for bond strength. For freezing-and-thawing durability, performance characteristic grades by Russell and Ozyildirim (2006) were used; that is, the relative dynamic modulus of

375 elasticity after 300 cycles was required to be greater than 70, 80, and 90%, for Grades 1, 2 and 3, respectively, as shown in Table 11.1.1. Based on these performance criteria, a preliminary selection was made to narrow down the choices of CP materials in the study from the candidate materials to two different materials in each of the two joint material classifications (i.e., overnight and 7-day cure). Further long-term tests, including freezing-and- thawing durability, shrinkage, bond, and permeability tests, were performed to evaluate the selected four joint materials (two for each cure) in order to validate or finalize the proposed preliminary performance criteria. 11.2. Selection of Candidate Materials for Long-Term Tests 11.2.1. Overnight Cure Materials and Their Preliminary Selection For the overnight cure, different grout materials were considered as candidate materials. As discussed earlier, published performance data from different grout materials were collected through contacts with material suppliers and users. Based on their potentials to meet the proposed preliminary performance criteria, candidate overnight cure materials were selected as shown in Table 11.2.1 with the mixing information. Five Star® Patch was cement-based, while EUCO-SPEED MP, Set® 45 and Set® 45 Hot Weather were all magnesium-phosphate based. Water and aggregate extension amounts used were based on manufacturer recommendations. The aggregate used was 3/8 in. pea gravel, which was tested for fizzing with 10% HCL to avoid calcareous aggregate made from soft limestone.

376 Table 11.2.1: Candidate overnight cure materials including mix proportions Product Name Mixing Quantities per 50-lb, Bag Initial Water, pints Additional Water, pints Aggregate Extension, % by weight Aggregate Extension, lb Yield Volume, cu. ft. Neat Grout EUCO-SPEED MP 3.1 0.5 0 0 0.42 Five Star® Patch 5.00 1.00 0 0 0.40 Set® 45 3.25 0.50 0 0 0.39 Set® 45 HW 3.25 0.50 0 0 0.39 Extended Grout EUCO-SPEED MP 3.1 0.5 60 30 0.57 Five Star® Patch 5.00 1.00 80 40 0.66 Set® 45 3.25 0.50 60 30 0.58 Set® 45 HW 3.25 0.50 60 30 0.58 The preliminary selection was based on strength tests of selected materials to narrow the choices down to two different materials in the overnight cure material classification. For neat grouts, the compressive strength was tested per ASTM C 109 modified. Both ASTM C 109 modified and ASTM C 39 modified were used to obtain the compressive strength for extended grouts to get both the cube strength and the cylinder strength. Both ASTM C 109 and ASTM C 39 required moist curing. However, the manufacturers for EUCO- SPEED MP, Set® 45 and Set® 45 HW did not recommend wet curing their products. Thus two normally used curing methods, air curing and moist curing, were investigated. The compressive strengths using these two curing methods are compared in Figure 11.2.1. The reported strength was the average of three specimens. As shown in Figure 11.2.1, the extended grouts gained strength slower than the corresponding neat grouts. Also, for the extended grouts, there was no significant difference between the strengths achieved by the two curing methods, although for neat grouts, there was a difference for the 4-hour strengths when comparing two curing methods except for Five Star® Patch.

377 (a) (b) (c) (d) (e) (f) (g) (h) Figure 11.2.1: Compressive strength development of the neat grouts per ASTM C 109 and extended grouts per ASTM C 39. (“h”=hour; “d”=day) Flow characteristics for each grout were measured in accordance with ASTM C 1437 modified. Specifications for the flow table and truncated flow cone were found in ASTM C 230. In these tests, the table was dropped 10 times within 15 seconds instead of 25 drops within 15 seconds according to the standard test method. The modification was needed to consider the fact that these particular types of grouts tend to flow better than the average mortars for which this test method was intended. Twenty-five drops would result in the grout spreading across the entire 10 in. diameter of the table and the purpose of the test would be lost. Flow results are presented in Figure 11.2.2. Observations were made regarding the workability of each grout based on the degree of effort required to mix each product as well as their work time and initial set time, as given in Table 11.2.2. Work time was measured from the start of mixing until workability began to decrease. Decreased workability was defined by the inability to move the grout with vibration, or easily finish a surface. Initial set time was measured from the start of mixing until the product showed resistance to the penetration of a thin rod or trowel edge. EUCO-SPEED MP 0 1500 3000 4500 6000 7500 9000 4h 8h 20h Age C om pr es si ve s tr en gt h (p si ) Air Curing Moist Curing Five Star Patch 0 1500 3000 4500 6000 7500 9000 4h 8h 20h Age C om pr es si ve s tr en gt h (p si ) SET 45 0 1500 3000 4500 6000 7500 9000 4h 8h 20h Age C om pr es si ve s tr en gt h (p si ) SET 45 HW 0 1500 3000 4500 6000 7500 9000 4h 8h 20h Age C om pr es si ve s tr en gt h (p si ) EUCO-SPEED MP extended 0 1500 3000 4500 6000 7500 9000 8h 1d 2d Age C om pr es si ve s tre ng th (p si ) Five Star Patch extended 0 1500 3000 4500 6000 7500 9000 8h 1d 2d Age C om pr es si ve s tre ng th (p si ) SET 45 extended 0 1500 3000 4500 6000 7500 9000 8h 1d 2d Age C om pr es si ve s tre ng th (p si ) SET 45 HW extended 0 1500 3000 4500 6000 7500 9000 8h 1d 2d Age C om pr es si ve s tre ng th (p si )

378 Figure 11.2.2: Truncated flow cone spread values per ASTM C 1437 for neat gouts and extended gouts Table 11.2.2 Candidate grout workability observations for neat gouts and extended grouts Grout Work Time, min. Initial Set Time, min. Consistency EUCO-SPEED MP 8 14 medium Five Star® Patch 18 32 medium Set® 45 6 10 medium Set® 45 HW 32 47 runny EUCO-SPEED MP extended 16 21 thick Five Star® Patch extended 13 27 thick Set® 45 extended 10 18 thick Set® 45 HW extended 30 45 thick Among the grout candidates, EUCO-SPEED MP and Set® 45 HW performed better than the remaining grouts based on the flow and workability performance. All of the extended grouts did not perform well in the flow cone spread testing. With regard to the workability, only Set® 45 HW extended had favorable workability results. Five Star® Patch extended 80% and Set® 45 extended 60% were almost impossible to mix with such a high recommended aggregate extension. Their flow suffered because of this. All the extended grouts exhibited lower strength than the corresponding neat grouts at an age of 8 hours. A lower aggregate 0 1 2 3 4 5 6 7 8 9 10 A ve ra g e S p re ad ( in .) EUCO- SPEED MP Five Star Patch SET 45 SET 45 HW EUCO- SPEED MP extended Five Star Patch extended SET 45 extended SET 45 HW extended After 10 Drops Initial

379 extension ratio would be more suitable for use in a precast deck panel system. Comparing the compressive strength and flow and workability performance of both neat and extended grouts, EUCO-SPEED MP and Set® 45 HW were selected for further evaluation. 11.2.2. 7-Day Cure Materials and Their Preliminary Selection For the 7-day cure materials, standard or special concrete and mortar mixtures, including five HPC mixtures, Emaco® T430 mix with latex, LMC-VE and two RSLP mixes, were considered as candidate materials. The candidate HPC mix designs are listed in Table 11.2.3. Mixes 1 to 3 were selected from Russell et al. (2006), and Mix 4 and 5 were developed by working with River Region Cement Division of Lafarge. Emaco® T430 mix with latex was developed by working with BASF Construction Chemicals, LLC, LMC-VE by working with the Virginia DOT, and RSLP mixes by the Virginia DOT and CTS Cement Manufacturing Corporation. The mix designs are given in Table 11.2.4.

380 Table 11.2.3: Candidate HPC mixes and mix proportions MIX NUMBER MIX 1 MIX 2 MIX 3 MIX 4 MIX 5 W/CM Ratio 0.31 0.35 0.31 0.32 0.35 Cement Type I I II I / II (Lafarge Sugar Creek SF) I / II (Lafarge Sugar Creek SF) Cement Quantity, lb/yd3 750 474 490 563* 431* Fly Ash Type C Quantity, lb/yd3 75 221 210 75 58 Slag Quantity, lb/yd3 113 86 Fine Aggregate, lb/yd3 1400 1303 1365 1161 1308 Coarse Aggregate Maximum Size, in 0.5 1 1.25 1.5 0.5 1.5 0.5 Coarse Aggregate Quantity, lb/yd3 1400 1811 1900 1530 270 1520 380 Air Entrainment, fl oz/yd3 5 3.1 3 2.3 Water reducer, fl oz/yd3 30 Retarder, fl oz/yd3 22 28 High-Range Water Reducer, fl oz/yd3 135 122 156 60 46 Shrinkage Reducing Admixture, fl oz/yd3 32 24.7 * 7% of silica fume is included in blended cement.

381 Table 11.2.4: 7-day cure mixes and mixture proportions, lb/yd Mixture 3 Emaco® T430 mix with latex LMC-VE RSLP Mix 1 RSLP Mix 2 Cement/Mortar Type Emaco® T430 CTS Rapid Set® cement CTS RSLP CTS RSLP Cement/Mortar Quantity 2530 658 665 658 Fine Aggregate --- 1600 1200 1695 Coarse Aggregate* Quantity 1380 1168 1800 1454 Latex 30 205 --- --- Water 170 137 280 263 * Coarse aggregate maximum size is 0.5 in. Three HPC mix proportions were first selected from the five candidate HPC mixes in Table 11.2.3 using the worksheet developed by Lawler et al. (2007). A statistically based experimental methodology was used in the worksheet to identify the optimum concrete mixture proportions for a specific set of conditions, as well as to predict performances of hydraulic cement concrete mixtures incorporating supplementary cementitious materials. In the worksheet, “desirability” was introduced. Desirability is a function that converts any test result into a value between “0” and “1”, where “0” means the result is unacceptable, and “1” means the result needs no improvement. Intermediate values show the level of acceptability (desirability) of the result. The overall desirability for each mixture is the geometric mean of the individual desirability for that mixture for each test. “Response” is the measured value from a performance test. According to the criteria employed for the study, the four responses were associated with the compressive strength at 7 day, shrinkage, chloride penetration and freezing-and-thawing durability. The corresponding desirability functions of the four “responses” were selected. Using standard linear regression analysis to obtain the response of a given mix that best fit the testing data, a model was developed by Lawler et al. (2007), which can be used to predict the desirability functions of an untested mixture. Based on this procedure, Mix 1, 4 and 5 were selected, comparing the predicted overall desirability and 7-day compressive strength desirability of the five HPC mixes based on the model. Compressive strength tests were performed on the remaining seven 7-day cure materials, HPC Mix 1, Mix 4 and Mix 5 and Emaco® T430 mix with latex, LMC-VE, RSLP Mix 1 and RSLP Mix 2, for further selection. The compressive strengths at 7 days and 28 days of the remaining seven 7-day cure materials are listed in Table 11.2.5. The reported strength was the average of three cylinders. Based on the compressive strengths, HPC Mix 1 and RSLP Mix 2 were selected as the two 7-day cure materials for the long-term tests.

382 Table 11.2.5: Compressive strengths (psi) per ASTM C 39 Modified HPC Mix 1 HPC Mix 4 HPC Mix 5 Emaco T430 mix LMC-VE RSLP Mix 1 RSLP Mix 2 7-day Compressive Strength 6494 4112 5058 1470 4404 3820 10564 28-day Compressive Strength 8895 5269 7309 2307 4389 4232 11262 11.3. Long-Term Tests Long-term tests were performed on the four candidate materials selected (i.e., EUCO-SPEED MP and Set® 45 HW for the overnight cure materials, and HPC Mix 1 and RSLP Mix 2 for the 7-day cure materials). The long-term tests included freezing-and-thawing durability, shrinkage, bond, and permeability tests. 11.3.1. Bond Strength Test The bond strength test was conducted per ASTM C 882 modified. Scholz et al. (2007) investigated slant cylinder bond strength of eight grouts with varying concrete surface preparations: a) smooth, b) exposed aggregate, c) raked, and d) raked and sandblasted. There was not a particular preparation found that consistently provided the best bond strength for all of the tested grouts. Scholz et al. (2007) concluded that the smooth interface performed better than anticipated, providing the worst or second worst bond strength for only half of the candidate grouts. For the cost involved with the other surface preparations (i.e., exposed aggregate, raked and sand blasting), the smooth interface was used for the NCHRP 10-71 study. The concrete half-cylinders were made using the mold and dummy section shown in Figure 11.3.1-a. After curing for at least 28 days, the half cylinders were inserted into a whole 4 in. by 8 in. cylinder mold. Then for the overnight cure materials, the grout was poured into the mold to complete the cylinder (see Figure 11.3.1-b). For the 7-day cure materials, a layer of cement paste was first applied onto the slanted face of the half-cylinder and then the test material was poured into the mold to complete the cylinder. Specimens for two overnight cure materials were air cured for 8 hours, while specimens for two 7-day cure materials were cured for 7 days by both the membrane-forming compound method and the water method with burlap, a typical practice for curing bridge decks. After curing, cylinders were tested in compression to investigate the bond strength of each material. The test setup is shown in Figure 11.3.1-c. Observations were made to determine whether the cylinders failed along the shear plane or if failure was due to significant cracking in the grout or concrete. The failure modes are shown in Figure 11.3.2 and described in Table 11.3.1. In each case, the maximum load was recorded and converted to stress by dividing by the elliptical area of the bonded interface, as suggested by Scholz et al. (2007). The maximum load was multiplied by the cosine of 30o to obtain the true shear stress component

383 acting along the bonded interface. Results for the slant cylinder tests are presented in Table 11.3.1. The strength results represent the average of three cylinders. All the materials had bond strengths greater than 300 psi, and the lower-bound of the criterion was subsequently increased from 200 to 300 psi. (a) (b) (c) Figure 11.3.1: ASTM C882 Test: (a) Test mold and dummy section (b) Completed slant shear cylinders ready for testing (c) Test setup (a) (b) (c) Figure 11.3.2: ASTM C882 Test Failure Modes (a), (b) and (c) *Descriptions of the three types of failure modes a-c are provided in footnotes to Table 11.3.1 *

384 Table 11.3.1 Slant cylinder bond strength and failure mode Material Type Specimen Number Test Age Shear Stress (psi) Average Shear Stress (psi) Mode of Failure* EUCO-SPEED MP 1 8 hours 456 397 b 2 159 b 3 576 b Set® 45 HW 1 8 hours 1161 1176 b 2 1121 b 3 1240 b HPC Mix 1 1 7 days 1607 1817 c 2 1917 a 3 1925 a RSLP Mix 2 1 7 days 659 705 a 2 634 a 3 823 a *a) clean shearing of bond along slanted interface (Figure 11.3.2-a) b) grout and/or concrete cracking before interface bond failure. However, it was possible to load the specimen until the bonded interface failed. (Figure 11.3.2-b) c) grout cracked and split in a vertical manner so that it was not possible to continue loading the specimen (Figure 11.3.2-c) 11.3.2. Permeability Test As discussed earlier, the ponding test was prepared in accordance with ASTM C 1543 modified instead of the ASTM C1202 test. Three specimens (10×10×3 in.) for each of the selected overnight cure materials, EUCO-SPEED MP and Set® 45 HW, and 7-day cure materials, HPC Mix 1 and RSLP Mix 2, were cast. Specimens for the two overnight cure materials were air cured for 8 hours, while the two specimens for the 7-day cure materials were cured for 7 days by both the membrane-forming compound method and the water method with burlap. After curing, the sides of the specimens were coated with the rubber coating

385 material, 1 in. high closed-cell polystyrene foams were bonded to the specimens with silicone sealant, and then the specimen were subjected to continuous ponding with a 3% sodium chloride solution to a depth of approximate 0.8 in for 90 days, as shown in Figure 11.3.3-a. The specimen surfaces were then brushed with a wire brush to remove the salt, and 4 in. cores were taken as shown in Figure 11.3.3-b. The core cylinders were then cut into slices. Four slices were cut from different depths (Layer 1: 0-0.25 in.; Layer 2: 0.25-0.75 in.; Layer 3: 0.75-1.25 in. and Layer 4: 1.25-1.75 in). The concrete slices were then dried at 105 °C to constant mass and ground with a pulverizer to pass an 850-μm sieve [No. 20] sieve. So for each specimen, powder samples were obtained at four corresponding depths. A solution was made with each powder sample following the ASTM C 1152 modified procedure. (a) Ponding of specimens (b) Specimen coring Figure 11.3.3: Specimen preparation per ASTM C 1543 Figure 11.3.4: Chloride concentration determination with ISE The titration test was introduced in the ASTM C 1152 to determine the chloride concentration. However, this method was very time-consuming. The tests by Ghanem et al. (2008) showed that the chloride ion selective electrode (ISE) matched titration readings, and suggested that the chloride concentration can be

386 taken directly using the ISE. Consequently, the chloride ISE was used rather than the titration test. The ISE was calibrated using chloride solutions with five different concentrations. These solutions were obtained by diluting a 100 ppm solution two, five, ten, and 100 times to get solutions with concentrations ranging from 1 to 100 ppm. A calibration curve was constructed with the measured electrode potential in mV (linear axis) plotted against the concentration (log axis). The mV readings of the sample solutions were taken, as shown in Figure 11.3.4, and the concentration was then determined from the calibration curve. The chloride concentrations were analyzed and are shown in Figure 11.3.5. Figure 11.3.5: Chloride content profile after 90-day ponding test (Layer 1: 0-0.25 in.; Layer 2: 0.25-0.75 in.; Layer 3: 0.75-1.25 in. and Layer 4: 1.25-1.75 in.)

387 Table 11.3.2: Depths (in.) for 0.2% chloride content (by mass of cement) Materials Sample 1 2 3 EUCO-SPEED MP <0.125 <0.125 <0.125 Set® 45 HW <0.125 <0.125 <0.125 HPC Mix 1 <1 <1 <1 RSLP Mix 2 >1.5 >1.5 >1.5 The depths for the 0.2% chloride content (by mass of cement) for the four materials were calculated based on Figure 11.3.5, as listed in Table 11.3.2. Note that the “down arrows” in Figure 11.3.5 for the HPC mixes indicate the depths at which the 0.2% chloride content (by mass of cement) threshold was exceeded. For calculating the depths, average depths of 0.125, 0.5, 1.0, and 1.5 in. were assigned to the respective layers (Layers 1-4). The two overnight cure CP materials had depths for 0.2% chloride content less than 0.125 in. For the 7-day cure CP materials, HPC Mix 1 was less than 1.0 in, and the RSLP Mix 2 was greater than 1.5 in. The two overnight cure and one of the 7-day cure materials, HPC Mix 1, met the criterion (i.e., 0.2% chloride content was not exceeded at a depth of 1.5 in.). 11.3.3. Freezing-and-Thawing Test The freezing-and-thawing test was prepared in accordance with ASTM C 666 Procedure A modified. Specimens for two overnight cure materials were air cured for 8 hours, while specimens for the two 7-day cure materials were cured for 7 days by both the membrane-forming compound method and the water method with burlap. After curing, specimens were moisture-conditioned in saturated lime water at 23.0±2oC (73.4±3oF) for 48 hours prior to testing, as specified by ASTM C 666 for specimens sawed from hardened concrete. After curing and 48-hour moisture-conditioning, the test was started as shown in Figure 11.3.6. After 76 cycles, the RSLP Mix 2 specimens failed as shown in Figure 11.3.6-c. The relative dynamic modulus of elasticity after 300 cycles was 92% for EUCO-SPEED MP, 96% for Set® 45 HW, and 96% for HPC Mix 1. Each result was the average of three specimens. EUCO-SPEED MP, Set® 45 HW and HPC Mix 1 performed very well, exceeding the preliminary criterion (i.e., the relative dynamic modulus of elasticity after 300 cycles was greater than 70%, 80%, and 90%, for Grades 1, 2 and 3, respectively).

388 (a) (b) (c) Figure 11.3.6: ASTM C666 Freezing-and-Thawing Durability Test: (a) Freezing-and-thawing apparatus, (b) Fundamental transverse frequency test, (c) Failure of RSLP Mix 2 specimens 11.3.4. Shrinkage Test For shrinkage, as discussed earlier, the steel ring test was prepared in accordance with AASHTO PP34 (1998) modified as shown in Figure 11.3.7. Strain gages were bonded at four equidistant mid-height locations on the interior of the steel ring and were oriented to measure strain in the circumferential direction. Three ring specimens were fabricated for each material, and were immediately transferred to the curing room with a constant air temperature of 21.0±1.7oC (73.4±3oF) and a relative humidity of 50±4 percent after completion of casting. The strain gages were connected to the data acquisition system to start monitoring the strain development in the steel ring. Specimens for the two overnight cure materials were air cured, while specimens for the two 7-day cure materials were cured by both the membrane-forming compound method and the water method with burlap till the age of 24 hours ± 1 hour. Then the outer ring was removed and the top surface was sealed. The strain development of one specimen is shown in Figure 11.3.8. Cracks were found for specimens of the HPC Mix 1 at the age of 20.5 days. No crack was observed to occur for the EUCO-SPEED MP, Set® 45 HW and RSLP Mix 2 throughout the tests which were terminated at the ages of 58, 62 and 61 days, respectively. All materials met the criterion (i.e., to not exhibit a crack at or before 14 days).

389 Figure 11.3.7: AASHTO PP34 test setup Figure 11.3.8: Steel ring strain versus specimen age for HPC Mix 1 -160 -140 -120 -100 -80 -60 -40 -20 0 20 0 5 10 15 20 25 Age (days) St ee l r in g st ra in (m ic ro st ra in ) HPC1-1-Strain guage #1 HPC1-1-Strain guage #2 HPC1-1-Strain guage #3

390 11.4. Proposed Performance Criteria and Conclusions For accelerated bridge construction, both transverse and longitudinal joints in decks may be required. Appropriate selection of CP materials for these deck-joints is critical for long-term durability considerations. Two categories of CP materials, overnight cure and 7-day cure, were proposed and studied. Candidate materials were compared by laboratory tests and software analysis, and two CP materials were selected for each category, as discussed in Section 11.2. Based on extensive literature reviews and the experimental investigation carried out on the selected candidate materials in association with NCHRP 10-71, performance criteria for selecting durable CP materials were developed as listed in Tables 11.4.1 and 11.4.2. Durability tests used to establish the performance criteria with the selected candidate materials included freezing- and-thawing durability, shrinkage, bond, and permeability tests. Table 11.4.1: Proposed performance criteria of CP materials Performance Characteristic Test Method Performance Criteria Compressive Strength (CS), ksi ASTM C39 modified 6.0≤CS @ 8 hours (overnight cure) @ 7 days (7-day cure) Shrinkagea(S), (Crack age, days) AASHTO PP34 modified 20<S Bond Strength (BS), psi ASTM C882 modified 300<BS Chloride Penetrationb(ChP), (Depth for Percent Chloride of 0.2% by mass of cement after 90-day ponding, in.) ASTM C1543 modified ChP<1.5 Freezing-and-thawing Durability (F/T), (relative modulus after 300 cycles) ASTM C666 Procedure A modified Gradec 1 Grade 2 Grade 3 70%≤F/T 80%≤F/T 90%≤F/T a: No S criterion need be specified if the CP material is not exposed to moisture, chloride salts or soluble sulfate environments. b: No ChP criterion need be specified if the CP material is not exposed to chloride salts or soluble sulfate environments. c: Grades are defined in Table 11.4.2.

391 Table 11.4.2: Application of CP material grades for freezing-and-thawing durability Freezing- and- thawing Durability (F/T) Is the concrete exposed to freezing-and- thawing environments? Yes Is the member exposed to deicing salts? Yes Will the member be saturated during freezing? Yes. Specify F/T-Grade 3 No. Specify F/T-Grade 2 No. Specify F/T- Grade 1 No. F/T grade should not be specified.