examining cost-saving measures in material selection for

CIVIL ENGINEERING STUDIES Illinois Center for Transportation Series No. 21-014

UILU-ENG-2021-2014 ISSN: 0197-9191

Examining Cost-Saving Measures in

Material Selection for Continuously

Reinforced Concrete Pavement: Volume 4

Prepared By Luca Montanari

Prannoy Suraneni Mehdi Khanzadeh Moradllo

Cameron Wilson Armen Amirkhanian Marisol Tsui Chang

Chiara Villani Steven R. Reese W. Jason Weiss

Oregon State University

Research Report No. ICT-21-014

A report of the findings of

ILLINOIS STATE TOLL HIGHWAY AUTHORITY PROJECT Innovative Structural and Material Design for Continuously Reinforced Concrete Pavement

https://doi.org/10.36501/0197-9191/21-014

Illinois Center for Transportation

May 2021

TECHNICAL REPORT DOCUMENTATION PAGE

1. Report No. ICT-21-014

2. Government Accession No.

N/A

3. Recipient’s Catalog No.

N/A

4. Title and Subtitle

Examining Cost-Saving Measures in Material Selection for Continuously Reinforced Concrete Pavement: Volume 4

5. Report Date

May 2021

6. Performing Organization Code

N/A

7. Authors

Luca Montanari, Prannoy Suraneni, Mehdi Khanzadeh Moradllo, Cameron Wilson, Armen Amirkhanian, Marisol Tsui Chang, Chiara Villani, Steven R. Reese, W. Jason Weiss

8. Performing Organization Report No.

ICT-21-014

UILU-ENG-2021-2014

9. Performing Organization Name and Address

Oregon State University

School of Civil & Construction Engineering

111 Kearney Hall, Corvallis, OR 97331

10. Work Unit No.

N/A

11. Contract or Grant No.

12. Sponsoring Agency Name and Address

Illinois State Toll Highway Authority

2700 Ogden Ave

Downers Grove, IL 60515

13. Type of Report and Period Covered

Final Report

14. Sponsoring Agency Code

N/A

15. Supplementary Notes

16. Abstract See executive summary.

17. Key Words Superabsorbent Polymers, Pore Solution, Absorption, Desorption Kinetics, Lightweight Aggregates, Design Methodology, Concrete Pavements, Autogenous Shrinkage, Relative Humidity, Superabsorbent Polymers, Hydration, Desorption, Autogenous Shrinkage, Relative Humidity, Internal Curing, Neutron Radiography, Internal Curing, Hydration

18. Distribution Statement

No restrictions. This document is available through the National Technical Information Service, Springfield, VA 22161.

19. Security Classif. (of this report) Unclassified.

20. Security Classif. (of this page)

Unclassified

21. No. of Pages

110

22. Price

N/A

Form DOT F 1700.7 (8-72) Reproduction of completed page authorized

i

ACKNOWLEDGMENT, DISCLAIMER, MANUFACTURERS’ NAMES

This publication is based on the results of Illinois State Toll Highway Authority Project titled Innovative Structural and Material Design for Continuously Reinforced Concrete Pavement. This study was funded by the Illinois State Toll Highway Authority. Acknowledgement is given to Mr. Steve Gillen and Mr. Dan Gancarz.

The contents of this report reflect the view of the authors, who are responsible for the facts and the accuracy of the data presented herein. The contents do not necessarily reflect the official views or policies of the Illinois Center for Transportation or Illinois State Toll Highway Authority. This report does not constitute a standard, specification, or regulation.

ii

EXECUTIVE SUMMARY

The Illinois State Toll Highway Authority (Tollway) is examining various techniques to produce a concrete pavement with an extended performance life, minimal maintenance, and lower long-term cost. Continuously reinforced concrete pavements (CRCP) may offer one option to achieve these performance criteria. The research team investigated several options to reduce costs and improve the performance, which included:

1. Examining mixture costs by evaluating current specifications for binders to determine if more effective binders may be able to be used;

2. Examining the role of material design on the curling/warping behavior which, will impact thickness and potential serviceability; and

3. Performing field trials with internal curing to be prepared for potential implementation.

This project examined the use of internal curing as a way to improve the performance and reduce the costs of CRCP. Two test sections were prepared using conventional internal curing mixture design methods. Observations made during the design and the construction of these two pavements that have led to the research outlined in this report. It was noted that a large volume of lightweight aggregate was used in the pavement mixtures. While the proportions are similar to that used in bridge decks, the volume of concrete pavements may provide some batching challenges for the producers. First, this required a relatively large laydown area. Second, sufficient moisture conditioning and quality control assessment of the lightweight aggregate were essential. Observations indicated that there was potential for inadequate moisture distribution in the aggregate pile which would alter the properties of the lightweight aggregate making it more difficult to batch during the course of the day.

It was hypothesized that the amount of aggregate could actually be reduced while obtaining the same benefits thereby improving both the construction process and cost benefit analysis for the concrete mixtures. Therefore, section 3 focuses on a new mixture design methodology for internally cured concrete containing fine lightweight aggregate (FLWA). This method divides the water into its use in the concrete and therefore provides details can result in a reduction in the volume of FLWA needed. In addition, an alternative to lightweight aggregate was pursued which consisted of superabsorbent polymers (SAP). The application of this approach would eliminate the cost and time required for pre-wetting materials, as the SAP can be added to the mixer in a dry form. Section 2 characterized the superabsorbent polymers. Section 4 examined the mixture design philosophy (from Section 3) associated with SAPs.

In Section 3, the current mixture design approach for internally cured concrete provides a volume of internal curing water, which is equal to the chemical shrinkage of the paste. This internal curing water reduces autogenous shrinkage, reducing the potential for early age cracking, increases the degree of hydration of cement and, possibly, the degree of reaction of supplementary cementitious materials (SCM), with respect to non-internally cured systems. An alternative methodology for determining the FLWA replacement volume was developed based on providing internal curing water to maintain a targeted relative humidity by considering the pore size distribution of the paste. The measurements

iii

on the mortar systems show benefits in terms of increased relative humidity and reduction in autogenous shrinkage, even for the lowest replacements of FLWA, with higher benefits coming from higher volumes of FLWA. The pore size distribution of the matrix was measured using desorption isotherms and used to establish the volume of internal curing water needed to fill the porous medium up to a target pore radius. An expression was developed in order to predict the shrinkage reduction, which was obtained through the addition of partial volumes of internal curing water, with respect to the commonly used approach.

In Section 2 a commercially available SAP was characterized for use as an internal curing agent in concrete. While many of the measured properties are specific to this particular SAP, the procedures adopted in this study can be extended to other SAP used as an internal curing agent. The ionic concentration of the pore solution influences the absorption of the SAP. More highly concentrated ionic solutions result in lower absorption. A relatively modest increase in the SAP absorption was observed when comparing the ordinary Portland cement (OPC) system and the OPC-fly ash systems (up to 17% with the 60% fly ash replacement).

When the mixture design approach developed in Section 3 was extended to SAP in Section 4, fractional volumes of SAP provide considerable benefits. Using SAP that contains 25% of the chemical shrinkage volume results in 55% of autogenous shrinkage reduction in the first seven days. This indicates that a reduction on the amount of SAP to include in the mixture is possible while still maintaining a large portion of the benefits coming from a full replacement.

The research team also investigated the role that internal curing may have on reduced curing times and the reduced curling of concrete pavements. Neutron radiography was used to measure moisture gradients in real time and this was coupled with the simple elastic analysis. It has been speculated that internal curing would reduce curling. However, unlike other studies this is not observed in the mixtures containing FLWA in this study. Rather it was observed that the samples containing FLWA had an increase in the amount of curling. It is believed that this may be due to the aggressive drying conditions exhibited for the samples as they were exposed to 50% relative humidity for an extended period. Additional testing is currently underway to better understand this observation. In practice, the evaporation rate from the slabs will be critical in the evaluation of curling.

Neutron radiography was used to quantify the extent of hydration at various distances from the drying surface. The influence of curing methods (internal/external curing) on the hydration of cement in concrete was examined. This is particularly useful in determining the ‘curing affected zone’ (CAZ). In the mixture exposed to drying after 1 day, the top 12.5 mm (1/2 inch) of the mortar was dramatically impacted by the loss of water to evaporation. While the top 5 mm of the surface of the sample exposed to drying at 1 day had a degree of hydration that was 32% less than the 14-day moist cured sample. As such, the duration of curing is important, especially with supplementary cementitious materials. This shows a technique to better quantify the required curing times. Finally, this shows the potential use of internal curing to create a more well hydrated and durable surface.

Additionally, the use of supplementary cementitious materials was investigated for both internal curing and its role in helping to reduce thermal gradients. The use of supplementary cementitious materials could create a graduation as well as reactions with deicing salts. It is recommended that the

iv

Illinois Tollway continues to utilize supplementary cementitious and admixtures because they are beneficial.

v

TABLE OF CONTENTS

CHAPTER 1: INNOVATIVE STRUCTURAL AND MATERIAL DESIGN FOR CONTINUOUSLY

REINFORCED CONCRETE PAVEMENT—PHASE 1 EXAMINING COST-SAVING MEASURES IN

MATERIAL SELECTION ............................................................................................................. 1

1.1 INTRODUCTION ...................................................................................................................... 1

1.2 CHLORIDE INGRESS FROM FIELD SAMPLES............................................................................. 1

1.3 FREEZE-THAW INCLUDING SUBGRADE ANALYSIS ................................................................... 3

1.4 FIELD TEST SECTIONS .............................................................................................................. 5

1.4.1 Test Section 1 – August 2016 .......................................................................................... 5

1.4.2 Test Section 2 – April 2017 .............................................................................................. 7

CHAPTER 2: ABSORPTION AND DESORPTION OF SUPERABSORBENT POLYMERS FOR USE IN

INTERNALLY CURED CONCRETE ............................................................................................. 18

2.1 INTRODUCTION .................................................................................................................... 18

2.1.1 Research Objectives ...................................................................................................... 19

2.2 MATERIALS AND EXPERIMENTAL PROCEDURES ................................................................... 19

2.2.1 Materials ...................................................................................................................... 19

2.2.2 Experimental Methods.................................................................................................. 20

2.2.2.1 Pore Solution Extraction and Simulated Pore Solution Preparation ........................................... 20

2.2.2.2 Paste Set Time .......................................................................................................................... 21

2.2.2.3 Isothermal Calorimetry ............................................................................................................. 21

2.2.2.4 XRF Analysis of Chemical Composition of Ionic Solutions .......................................................... 21

2.2.2.5 Degree of Hydration and Powers’ Model .................................................................................. 21

2.2.2.6 SAP Absorption......................................................................................................................... 22

2.2.2.7 Desorption Behavior of SAPs .................................................................................................... 23

2.3 RESULTS AND DISCUSSION ................................................................................................... 24

2.3.1 Absorption of SAP as a Function of Time ....................................................................... 24

2.3.2 Absorption of SAP as a Function of pH .......................................................................... 26

2.3.3 Influence of Inclusion of SCMs on SAP Absorption ........................................................ 27

2.3.4 Influence of External Relative Humidity on SAP Desorption .......................................... 30

2.3.5 Changing Ionic Concentration to Study SAP Desorption ................................................ 31

2.3.6 Hydration of the System and Its Influence on SAP Desorption....................................... 33

vi

2.4 CONCLUSIONS ...................................................................................................................... 35

CHAPTER 3: TOWARD A DESIGN METHODOLOGY FOR PARTIAL VOLUMES OF INTERNAL

CURING WATER BASED ON THE REDUCTION OF AUTOGENOUS SHRINKAGE ....................... 37

3.1 INTRODUCTION .................................................................................................................... 37

3.2 EXPERIMENTAL PROCEDURE ................................................................................................ 38

3.2.1 Materials ...................................................................................................................... 38

3.2.2 Mixture Design and Mixing Procedure .......................................................................... 38


3.3 EXPERIMENTAL RESULTS ...................................................................................................... 40

3.3.1 Proposed IC Mixture Design Methodology .................................................................... 45

3.4 CONCLUSIONS ...................................................................................................................... 52

CHAPTER 4: ACCOUNTING FOR WATER STORED IN SUPERABSORBENT POLYMERS IN

INCREASING THE DEGREE OF HYDRATION AND REDUCING THE SHRINKAGE OF INTERNALLY

CURED CEMENTITOUS MIXTURES ......................................................................................... 53

4.1 INTRODUCTION .................................................................................................................... 53

4.2 MATERIALS AND EXPERIMENTAL PROCEDURES ................................................................... 53

4.2.1 Materials ...................................................................................................................... 53

4.2.2 Mixture Design and Mixing ........................................................................................... 54

4.2.3 Experimental Procedures .............................................................................................. 55

4.2.3.1 Pore Solution Extraction ........................................................................................................... 55

4.2.3.2 SAP Absorption......................................................................................................................... 56

4.2.3.3 Desorption Behavior of SAPs .................................................................................................... 56

4.2.3.4 Internal Relative Humidity Measurements ................................................................................ 57

4.2.3.5 Autogenous Shrinkage Measurements ..................................................................................... 57



4.3.1 Absorption and Desorption Measurements of SAP ....................................................... 58

4.3.2 Internal Relative Humidity of Mortar Systems............................................................... 60

4.3.3 Autogenous Shrinkage in Mortars ................................................................................. 61

4.3.4 Increase in Degree of Hydration through Isothermal Calorimetry ................................. 62

4.3.5 Internally Cured Mixtures Design Model ....................................................................... 63

vii

4.4 CONCLUSIONS ...................................................................................................................... 69

CHAPTER 5: ROLE OF MIXTURE CONSTITUENTS ON CURL/WARP ......................................... 70

5.1 INTRODUCTION .................................................................................................................... 70

5.2 EXPERIMENTAL PROCEDURE ................................................................................................ 75

5.2.1 Beam Curling ................................................................................................................ 75

5.2.2 Moisture Gradient by Neutron Radiography ................................................................. 76

5.2.3 Calculated Curling Deflections ...................................................................................... 77


CHAPTER 6: EXAMINING CURING EFFIECIENCY USING NEUTRON RADIOGRAPHY................ 84

6.1 INTRODUCTION .................................................................................................................... 84

6.1.1 Research Objective ....................................................................................................... 86

6.2 EXPERIMENTAL PROGRAM ................................................................................................... 86

6.2.1 Materials and Mixture Proportions ............................................................................... 87

6.2.2 Casting and Curing ........................................................................................................ 88


6.2.3.1 Using LOI to Determine the Non-Evaporable Water Content .................................................... 89


6.2.3.3 Neutron Radiography (NR) ....................................................................................................... 89

6.3 EXPERIMENTAL RESULTS AND DISCUSSION ......................................................................... 92

6.3.1 Future Studies............................................................................................................... 95

6.4 CONCLUSIONS ...................................................................................................................... 95

REFERENCES .......................................................................................................................... 97

1

CHAPTER 1: INNOVATIVE STRUCTURAL AND MATERIAL DESIGN FOR CONTINUOUSLY REINFORCED CONCRETE PAVEMENT—PHASE 1 EXAMINING COST-SAVING MEASURES IN MATERIAL SELECTION

1.1 INTRODUCTION

The Illinois State Toll Highway Authority (Tollway) is examining various techniques to produce a concrete pavement with an extended performance life, minimal maintenance, and lower long-term cost. Continuously reinforced concrete pavements (CRCP) may offer one option to achieve this performance criteria. The research team investigated several options to reduce costs and improve the performance, which included:

1. Examining mixture costs by evaluating current specifications for binders to determine if more effective binders may be able to be used;

2. Examining the role of material design on the curl/warp behavior which will impact thickness and potential serviceability; and

3. Perform field trials with internal curing to be prepared for potential implementation.

Re-engineering, innovating, and building a more cost effective CRCP could occur with the use of higher supplementary cementitious materials, alternative reinforcement type/location and internal curing.

The goal of this research is to examine the potential for using materials and construction practices to examine the binder that is used in the CRCP and to reduce the potential for curling and warping related to differential moisture movement as well as thermal effects. This work examined fundamental material parameters related to the stresses generated by thermal and hygral gradients. The work measured moisture distribution using small scale tests and material scale tests for the purpose of quantifying stress in sections of conventional and internally cured concrete. While the majority of the research focused on the binder composition and the potential use for internal curing, two additional items were examined including chloride ingress and potential issues associated with freeze-thaw. These additional items are discussed in the following section while the remainder of the report deals with aspects of internal curing.

1.2 CHLORIDE INGRESS FROM FIELD SAMPLES

Preliminary chloride ingress evaluations were performed using core samples obtained from the Tollway and provided to Purdue/Oregon State. The core samples were obtained from ARA at various locations along the Tollway. Figure 1-1 provides photos of the tested cores. The goal of testing the chloride ingress of these samples was to provide information about the depth of chloride penetration that may be expected to occur over a certain period of service. Comparing the obtained chloride ingress profiles with the historical data provides historical conditions for which these pavements are

2

designed. This analysis provided information about the quality of the current concrete (From a durability perspective) and the depth of penetration that the chlorides penetrate.

Figure 1-1. Five core samples obtained from concrete pavements on the Tollway.

Chloride ingress profiles were measured on concrete powders that were obtained by grinding the cores in 2 mm sections from the surface of the concrete up to the depth of 10 cm. Powered concrete samples were dried for a week according to the procedure proposed by Delagrave et al. (1997). A sample of at least 10 g of powder was ground and prepared in accordance with the testing procedure described in ASTM C1543-10 (the samples were tested as received without exposure to chloride solution (e.g., ponding) in the lab). Figure 1-2 shows the measured chloride depth of penetration from tested core samples.

Typical chloride profile starts with a maximum chloride concentration at the surface and decreases over the sample depth toward the center of the pavement (Fickian diffusion profile). Figure 1-2 illustrates that near the surface the chloride content is actually lower than it is in the core. This phenomenon has been referred to as the “maximum phenomenon” in the literature. The absorption–desorption/evaporation process (convection), skin effect, carbonation, washout during the rain, and calcium hydroxide leaching are the main reasons that have been stated as a cause for the formation of maximum phenomenon in concrete skin layer (Moradllo et al. 2018). It can be seen that the chloride ingress is the greatest for the samples from I-80 within the first inch (25 mm) of the surface. The 6-inch diameter cores show much more gradual chloride ingress throughout the course of the depth. However, the depth of chloride penetration is higher for 6-inch diameter cores compared to the cores from I-80. It should be noted that in both cores the chloride content is not sufficient enough

3

to reach a level that would initiate chloride corrosion at the 4.9 inch depth of the reinforcing steel (reinforcement cover depth in CRCP).

Figure 1-2. Chloride profile from concrete cores taken from CRCP pavements.

It is important to note that the crack development on CRCP can lead to the higher chloride concentration. Currently, there is a question as to whether epoxy coated or “black” rebar is needed or whether a particular rebar can be used. Reducing the application of epoxy-coated rebars might be desirable due to the cost associated with its use. Regarding the influence of cracking on corrosion, it is known that the cracks will accelerate the ingress of the chlorides. However, the cracks in CRCP are relatively tight. In addition, the reinforcement cover depth is substantially greater in CRCP compared to other reinforced concrete elements. This results in the cracks being relatively tight across the depth make it difficult for the chlorides to transport through the cracks.

Further, based on stress analysis, it is assumed that there is not substantial debonding along the length of the reinforcing bar. However, additional research is needed to confirm this hypothesis. This would imply that corrosion is taking place primarily at the crack and not propagating along the bar which appears to be a major factor that would influence reduction in service life. In a discussion with both ARA and Illinois Tollway representatives, it has been reported that the corrosion of the reinforcing steel is not a significant issue that is limiting the life of the CRCP. If that is true, the need for a particular reinforcing steel may be questioned. If the epoxy coated steel was removed, substantial reductions in cost may be obtained. Research is needed to better understand the shape of the cracks when epoxy coated black steel is used.

1.3 FREEZE-THAW INCLUDING SUBGRADE ANALYSIS

During this task, research was performed to evaluate two aspects of the mixture design in terms of the freeze-thaw performance and the susceptibility to salt damage. The goal was to provide a high

4

level of probability that the mixtures will not experience durability related freeze-thaw damage. The results will also include an analysis of the potential role of the subgrade in altering the capillary absorption of water as different subgrade treatments are being considered.

Distress has recently been observed in some concrete pavements throughout the Midwest, primarily at the joints of joined plain concrete pavements in the wet-freeze states. This distress often begins in longitudinal joints, followed by transverse joints and results in the significant loss of material from the joint area. This deterioration greatly reduces the service life and increases maintenance costs of the pavements. Primary issues that emerged from studies on this phenomenon include the importance of the timing of joint sawing, the width of the joint opening, the sealing of the concrete or joint sealing, the drainage and degree of saturation of the concrete at the joint, the quality of the air void system, the role of deicing chemicals, the quality of curing, and the degree of restraint at the joint.

During the last decade, several research projects have been conducted which were focused on identifying the mechanisms responsible for premature joint deterioration in concrete pavements. While a number of potential mechanisms responsible for the observed distress have been proposed, they can generally be classified as either:

1) Classic freeze-thaw damage due to increased levels of saturation for pavement joints with no or low salt concentrations; or

2) Chemical reactions between chloride bearing salts (especially CaCl2 and MgCl2) and the cementitious matrix (specifically calcium hydroxide, Ca(OH)2) for high salt concentrations.

Early in this project, it was determined that this issue was not being substantially observed in continuously reinforced pavements. However, the steps and design changes that are being developed for jointed plain concrete pavements should be considered in the mixture. The following list provides a short summary of the main findings and recommendations (INDOT SPR 3808), followed by implications for the Illinois Tollway.

1) Increase the Specified Volume of Air Entrainment and Reduce the Variation in Air Content. The recommendation for the Tollway is to continue to require a total volume of air of 6.5% and to encourage contractors to reduce the variation through the quality control processes.

2) Reduce the Volume of Cementitious Paste in Concrete Pavements. The recommendation for the Tollway is to encourage mixtures that have optimized aggregate gradations that enable the paste volume to be reduced below 25%. Any minimum cement contents should be considered to be removed from the specification.

3) Reduce the Water-to-Cementitious Materials Used in Concrete Pavements. As it is related to the Tollway, it is suggested that the water to binder (w/b) ratio be limited to 0.42 or lower.

4) Use of a formation factor to specify the transport properties of concrete. This recommendation is not ready to be implemented at the current time with the Tollway. However, it should be considered in the future in coordination with AASHTO PP-84.

5

5) Use Supplementary Cementitious Materials (SCM) to reduce susceptibility to salt damage/ use a performance test for mixture design to limit calcium oxychloride damage. This recommendation is not ready to be implemented at the current time with the Tollway.

6) Use a topical treatment for concrete that repels water or seals the concrete. This recommendation is not ready to be implemented at the current time with the Tollway.

7) Reduce the tie bar size and spacing to the necessary level. This recommendation is only for jointed plain concrete pavement and should not be implemented at the current time with the Tollway.

8) Remove the backer rod or a cavity in the design of pavement joints. This recommendation is only for jointed plain concrete pavement and should not be implemented at the current time with the Tollway.

9) Consider the use of unsealed Joints. This recommendation is only for jointed plain concrete pavement and should not be implemented at the current time with the Tollway.

10) Use a capillarity break below the pavement. This recommendation suggests the use of an open graded sub-base below the pavement with proper draining. This should be considered with the complete pavement section design.

11) Reduce the strength required to open a pavement to traffic. This recommendation is not ready to be implemented at the current time with the Tollway.

12) Increase the use of maturity to accept concrete pavement at early ages while long term strength is used in design. This recommendation is not ready to be implemented at the current time with the Tollway.

13) Improve the use of methods to detect water ponding in concrete pavements. This recommendation is not ready to be implemented at the current time with the Tollway.

14) Examine the proportion of salts in blended systems. This recommendation refers to deicing practice rather than mixture proportioning or pavement design.

1.4 FIELD TEST SECTIONS

There were three test sections cast during the course of this project with two of the sections employing internal curing. The first IC test section occurred in August of 2016 and the second one in April of 2017. The sections were heavily instrumented by researchers from the University of Illinois Urbana-Champaign and Texas A&M University. The personnel from Oregon State University were onsite to provide technical assistance with materials quality control during construction. This section contains reports of observations from the two visits during paving.

1.4.1 Test Section 1 – August 2016

Two researchers from Oregon State were present for the construction of the IC/MF sections on August 6. They assisted the construction team at the IC/MF construction site and in the batch plant in running the centrifuge test for the FLWA. A few pictures of the construction including the piles of materials and testing devices are shown in Figures 1-4 to 1-13.

6

During construction, there was a minor issue with the batching of six trucks due to operator error, but the mistake was caught. The gauges embedded in the pavement were well protected by hand placing concrete around the gauges. The lead wires for the gauges were carefully avoided by the construction personnel and the location of the gauges was marked on the base layer adjacent to the construction so that the location could be recorded at a later time.

The batch plant was located at 1765 Armitage Ct in Addison. Two different devices were used to run the centrifuge test to determine the moisture content of the lightweight aggregates in the batch plant (the centrifuges from STATE Testing and Prairie, Figures 1-9 to 1-13). The moisture measurement results from two centrifuges are summarized in Table 1-1. Based on results, there is a discrepancy between measured moisture contents from two different centrifuges. The inspection team from Oregon State believes that this deviation in results is attributed to the limitations of the centrifuge used by Prairie. Based on ITM 222-14T, the extraction apparatus should be a “centrifuge, in accordance with AASHTO T 164 Method A, with controls for the time of operation and maximum speed.” AASHTO T 164 Method A states that the extraction apparatus should be a “centrifuge extractor with a 3000 g sample capacity in which the bowl may be revolved at a controlled, variable speeds up to 3,500 RPM. However, the capacity of centrifuge bowls used by Prairie was limited to 1500 g with diameter of 10 inches (Figure 1-12). This can be cause of deviation in moisture content measurements since the different centrifugal forces on the aggregates in the different size bowls will require different centrifuge speeds and spin times. In addition, the centrifuge used by Prairie did not include a way to measure time of operation. The maximum speed in Prairie centrifuge can only be controlled on an analogue dial that has units 1-10. These numbers needed to be correlated to the appropriate centrifuge speed in RPM which was done only after the construction. It is suggested that Prairie follow the testing requirements outlined in (Miller et al. 2014b).

Table 1-1. Moisture measurement results during the course of construction on Aug 6, 2016

Time STATE Testing Prairie

Surface Absorption Total Surface Est. RPM

05:00 11.5 11.1 23.9 9.9 1925

06:00 6.4 11.8 18.9

07:20 7.7 5.4 1925

09:11 12.0 11.9 25.3

09:24 11.8

10:30 11.2

11:00 11.4

11:20 10.8 12.7 2450

11:35 12.3

12:10 10.6 8.8 2100

12:30 12.1 11.5 2275

The pore sizes being emptied for both the STATE Testing and Prairie centrifuges can be calculated based on the speed. From (Miller et al. 2014b), the pore size, r, that is being emptied can be calculated from Equation 1-1 where γ is the surface tension of water [0.072 N/m], θ is the contact angle of the pore solution [assumed to be 0], ρ is the density of water at 25°C [1000 kg/m3], σ is the

7

angular velocity in rads/sec, and R is the radius of the centrifuge bowl in meters. The RPM dependent behavior can be plotted (Figure 1-3) and it is seen that the Prairie centrifuge, for a given RPM, would fail to empty some of the pores that the STATE Testing centrifuge would. The STATE Testing centrifuge ran at a consistent 2000 rpm and thus pores greater than 262 µm are being emptied. The centrifuge used by Prairie emptied pores between 250 µm and 317 µm, which explains the discrepancy seen between the two devices. In some cases, the Prairie centrifuge removed less water than the specified centrifuge and thus underestimated the true moisture content. That said however, it is clear that they both provide similar general information.

𝒓 = √𝟔𝜸 𝐜𝐨𝐬 𝜽

𝑹𝝆𝝎𝟐 Eq. 1-1

Figure 1-3. Comparison of the pore emptying ability of the two centrifuges used during construction.

1.4.2 Test Section 2 – April 2017

This work was performed with Prairie Material, the Illinois Tollway, and Oregon State University. An Oregon State University researcher was present at the batch plant on the casting day of the Internally Cured test section (IC) on 4-18-17. The main objective was to monitor the aggregate moisture (absorption and surface moisture) of the FLWA used in the 400 cubic yard IC trial section for the IL Tollway (located around 1200 E Thorndale Ave Wood Dale, IL 60191). It should be noted that on the same day as the IC trial section for the IL Tollway, two more internally cured projects were to occur and be mixed at the batch plant (Prairie Material Yard 14 on 1765 Armitage Ct, Addison, IL 60101). A 400 cubic yard bridge deck for the IL Tollway was placed simultaneously, and afterwards a moment

8

slab for the Tollway was batched. The aggregate moistures in the following report are reflected in each of these three projects, but the IC Tollway pavement will be the focus here. The IC Tollway pavement consisted of one lane at 1000 feet long (Figure 4).

The apparatus that had soaked the lightweight aggregate pile was one of the sprinkler type as seen in Figure 1-5. According to Prairie Material, the water that was soaking the lightweight aggregate pile was turned off the night before. The pile [Figure 1-6, Figure 1-7 and Figure 1-8] had been turned as well. By visual and physical inspection, the sand in the pile appeared sufficiently wet at the beginning of the day.

At 5:48 AM the first aggregate moistures were taken by both Prairie Material and STATE Testing. The inspector from Oregon State tested moistures throughout the remainder of the day using equipment supplied by STATE testing as shown in Figure 1-9. Prairie Materials tested aggregate moistures on the lightweight fines two more times throughout the remainder of the day, at 9:10 AM and 12:20 PM. The equipment used by Prairie Material can be seen in Figure 1-10. It should be noted that the centrifuge used by Prairie Material did not correspond to the guidelines outlined in ITM 222-14T and AASHTO T-164 (ITM 222 2014; AASHTO T 164 2014). Specifically, the capacity of centrifuge bowl used by Prairie Material was 1500 g while, the standard procedures require a 3000 g capacity bowl to be used. The STATE Testing equipment complied to these regulations. The relative bowl sizes are illustrated in Figure 1-11 and Figure 1-12. Also, the centrifuge used by Prairie material did not allow the speed of the centrifuge to be controlled to a known RPM as seen in Figure 3. The last moisture measurement by STATE Testing centrifuge was performed at 1:00 PM. It should be noted that the design absorption was 12.5%. The plots from the day of testing can be seen in Figure 1-14, Figure 1-15, and Figure 1-16. In these Figures, “SM” denotes surface moisture, “ABS” denotes absorption, and “TM” denotes total moisture. It should also be noted that sample 13 was sampled by an Oregon State inspector and sample 14 was sampled by Prairie Materials. Samples 13 and 14 were sampled within a foot of each other.

Generally, the surface moistures tended to decrease throughout the day. Perhaps this is due to the single sprinkler used to wet the pile [Figure 1-4]. It appears that one sprinkler was not enough considering the size of the pile. As the day progressed, the temperature increased from around 40° F to 70°F. The temperature increase could also have had some impact on the significant moisture loss, but this is likely not the main cause.

As the loader kept removing sand from the pile, the sand underneath became drier. As such, it is believed that the pile was not sufficiently wet. At the beginning of the day, the surface moistures were at a sufficient level (around 10%), but then they steadily decreased. Some areas of the pile (the areas in the range of the sprinkler) were sufficiently wet, but other areas were not. A drawing of the surface moisture samples and their approximate locations in the pile can be seen in Figure 1-17.

Prairie Material did correct for the surface moisture, as suggested by OSU (three times throughout the day as signified by the orange data points), but they could not correct for all the variations throughout the day in sufficient time. With inconsistency in the SM of FLWA, adequate correction would have been difficult (as operations at the plant must continue without halt). If the sand were

9

more consistently soaked (less variability in the SM), then sufficient SM corrections would have been possible.

The picture of sample SM with respect to their location in the pile is shown in Figure 1-17. It is clear that the pile could have used more water, or multiple sprinklers.

Figure 1-4. IC section cast on 4-18-17 for the Illinois Tollway (picture taken on 4-18-17).

Figure 1-5. Sprinkler that was used to soak the lightweight fines pile (can be seen in upper right corner of photo).

10

Figure 1-6. This is the pile before operations began on the morning 4-18-17.

Figure 1-7. The pile of lightweight fines as material is being taken away from it.

11

Figure 1-8. Toward the end of the day, the pile is losing moisture.

12

Figure 1-9. Equipment supplied by STATE Testing for spinning the lightweight fines to obtain surface moisture.

13

Figure 1-10. Centrifuge used by Prairie Material.

14

Figure 1-11. Bowl used by STATE Testing (3000 gram capacity). The same brush is used for scale.

Figure 1-12. Bowl used by Prairie Material (1500 gram capacity). Note: the bowl has a smaller radius. The bowl with a smaller radius will not empty the same size pores in the FLWA as the larger bowl. In other words, more SM will remain on the lightweight fines if a smaller radius bowl is used rather than the larger bowl (holding all other parameters constant). According to Miller et al, 2016,

“Bowls with a smaller radius will empty smaller pores than bowls with larger radii. To obtain consistent results between laboratories, it would be necessary to standardize the size of the

centrifuge bowl or make an adjustment to the procedure to obtain comparable results.”

15

Figure 1-13. Speed control on the centrifuge used by Prairie Material. The numbers do not directly correspond to a given RPM.

Figure 1-14. Surface Moisture versus time. The blue data points refer to samples taken by Cameron (Oregon State). The orange points refer to samples taken by Prairie Material.

16

Figure 1-15. Surface moisture versus sample number.

Figure 1-16. Sample numbers versus absorption and total moisture. It should also be noted that the absorption values of the lightweight fines were acceptable throughout the day.

12.00%

14.00%

16.00%

18.00%

20.00%

22.00%

24.00%

26.00%

28.00%

30.00%

0 1 2 3 4 5 6 7 8 9 10 11

Per

cen

t

sample #

sample num versus absorption and total moisture

sample num vs abs

sample num vs total mositure

17

Figure 1-17. Drawing of the pile of lightweight aggregate fines and the locations, sample number, and corresponding surface moisture of the sample.

18

CHAPTER 2: ABSORPTION AND DESORPTION OF SUPERABSORBENT POLYMERS FOR USE IN INTERNALLY CURED CONCRETE

2.1 INTRODUCTION

Internal curing was defined by ACI as “supplying water throughout a freshly placed cementitious mixture using reservoirs, via pre-wetted lightweight aggregates, that readily release water as needed for hydration or to replace moisture lost through evaporation or self-desiccation” (RILEM Technical Committee 225 2012; ACI Committees 308, 213, 2013; Castro et al. 2016; Kovler and Jensen 2005; RILEM Technical Committee TC 181-EAS 2003; Cusson and Hoogeveen 2008; Bentz and Weiss 2011). However, superabsorbent polymers (SAP) may be a potential alternative to pre-wetted fine lightweight aggregates (FLWA) as the internal curing agent (RILEM Technical Committee 225 2012; Montanari et al. 2017; Hasholt et al. 2012; Schröfl et al. 2012; Jensen and Hansen 2002; Wang et al. 2009). The use of superabsorbent polymers (SAP) in cementitious mixtures is currently being examined by several groups in the concrete research community for use in internal curing applications (RILEM Technical Committee 225, 2012).

The application of internal curing is of particular interest in high-strength, low water-to-cementitious (w/cm) systems, where early age cracking due to self-desiccation is more frequent (RILEM Technical Committee TC 181-EAS 2003; Shah et al. 1998).

Internal curing is based on the concept that as cement hydrates, vapor filled spaces are formed as a result of chemical shrinkage. The fluid in the internal curing agent is then released to partially fill the vapor spaces. This reduces the effect of self-desiccation and magnitude of autogenous shrinkage (Wang et al. 2009; Shah et al. 1998; Geiker et al. 2004; Snyder and Bentz 1999).

While the original purpose of internal curing was to reduce autogenous shrinkage and early age cracking (Bentz and Weiss 2011; Friggle and Reeves 2008), other benefits have been observed, including an increased degree of hydration (Castro et al. 2016; Bentz and Weiss 2011; Lura et al. 2006; Farzanian et al. 2016), reduced water absorption (Henkensiefken et al. 2009), reduced thermal and drying shrinkage cracking (Henkensiefken et al. 2009; Schlitter et al. 2010; Byard et al. 2010), reduced chloride ingress (Bentz 2009; Di Bella et al. 2012), increased relative humidity at early ages, and reduced autogenous shrinkage (Montanari et al. 2017; Hasholt et al. 2012; Wang et al. 2009; Craeye et al. 2011).

One benefit of using SAP as compared to LWA, is the removal of the need of a pre-wetting stage, due to the fast absorption of the SAP. By removing the need for pre-wetting, substantial reductions in time and effort needed for the preconditioning of materials in the field may be obtained, which may result in simplifications in field production.

While benefits are associated with the use of SAP, their correct usage in cementitious mixtures requires the SAP to be thoroughly characterized. The absorption and desorption kinetics of the SAP

19

are two important parameters for mixture proportioning. An incorrect determination of SAP absorption might influence the w/cm, while a slow desorption might reduce the benefits of internal curing at early ages. Moreover, the absorption of the SAP is a required parameter that is needed for mixture design and quality control operations (Snyder and Bentz 1999; Bentz et al. 2005).

2.1.1 Research Objectives

The objectives of this research are:

• To contribute to the standard practice in performing the teabag method and accurately determining the SAP absorption in simulated or extracted pore solution.

• To examine the influence of the ionic concentrations on the absorption of a commercially available SAP.

• To examine the influence of relative humidity on the SAP desorption.

• To examine the influence of the ionic concentration of the solution on the SAP desorption and to determine the reversibility in absorption and desorption.

• To develop an empirical model to predict the relative impact of desorption due to changes in ionic concentration of the hydrating pastes.

2.2 MATERIALS AND EXPERIMENTAL PROCEDURES

2.2.1 Materials

A commercially available SAP (provided by BASF USA) was used in this study. The SAP was reported to have a specific gravity of 1.4 in dry state. The main constituent of the SAP was crosslinked anionic polyacrylamide. The particle size distribution of the dry SAP was equal to or smaller than 150 μm. The shape of the SAP was angular due to grinding during production.

Several pastes were considered, including a pure ordinary Portland cement (OPC) paste and pastes where OPC was replaced with fly ash.

A type I/II OPC complying with ASTM C150/C150M ̶ 17, was selected for this study. The cement had a Blaine fineness of 383 m2/kg. The phase composition was 60.1% C3S, 10.1% C2S, 8.2% C3A, 8.2% C4AF (using cement chemistry notation, C: CaO, A: Al2O3, F: Fe2O3, and S: SiO2). The specific gravity of the cement was assumed to be 3.15. A Class F fly ash (complying with ASTM C618 ̶ 15) was used in blended cements to study the impact of supplementary cementitious materials (SCMs) on the SAP absorption. The fly ash had a CaO content of 10.5% and it was assumed to have a specific gravity of 2.6. The oxide contents of the cement and the fly ash, as measured using X-ray fluorescence (XRF) (ASTM C114 ̶ 15) are shown in Table 2-1.

20

Table 2-1. Oxide content of the cement and fly ash measured using XRF

Phase Composition Mass, %

Oxide OPC FA

Na2O 0.26 2.46 MgO 3.39 1.00 Al2O3 4.77 23.71 SiO2 19.11 55.70 SO3 3.21 0.25 K2O 0.78 0.81 CaO 62.62 10.50 Fe2O3 2.88 3.75

OPC and OPC-fly ash pastes were designed with a water-to-cementitious ratio (w/cm) of 0.36. Fly ash was used at replacement levels of 0%, 20%, 40%, and 60% (by volume) to prepare the OPC-fly ash cement pastes. The paste samples were vacuum mixed for two 90-second periods at 400 rpm, with a 30 second pause to scrape material from the blade and the sides of the mixer.

2.2.2 Experimental Methods

2.2.2.1 Pore Solution Extraction and Simulated Pore Solution Preparation

Pore solution expression from fresh paste samples was performed using a commercial expression apparatus (from Millipore). After mixing, the fresh paste samples were cast in plastic cylinder molds of approximate size of 2.54 cm x 4 cm, and sealed for 30 minutes, in order to minimize the loss of moisture to the environment. Approximately 30 minutes after the water and cement came into contact, the fresh paste was moved to a fresh pore solution extractor (Rajabipour et al. 2008), as shown in Figure 2-1.

Figure 2-1. Pore solution extractor setup.

21

The pore solution extractor was used at a constant pressure of 200 kPa, for a period of five minutes using nitrogen gas. Depending on the mixture and the volume of fly ash, approximately 10 – 15 g of solution were obtained from these extractions. To minimize solid particles from contaminating the extracted pore solution, a cellulose membrane filter with an average pore diameter of 0.45 μm was used. Once the solutions were extracted, they were sealed inside 20 ml plastic canisters and stored at 5 ± 1°C environmental chamber, to minimize the potential for carbonation (Tsui-Chang et al. 2018).

Pore solution was extracted from fresh OPC paste and OPC-fly ash pastes. For each system, a set of three replicates was extracted and analyzed. The absorption of the SAP was measured using these extracted pore solutions.

Simulated pore solutions were prepared with a range of ionic concentrations by dissolving pellets of NaOH, with a 98% purity, and Ca(OH)2 with a purity of 99%, in DI water.

2.2.2.2 Paste Set Time

The set time of the OPC paste was measured following the procedure described in ASTM C191 – 14, which uses a Vicat needle apparatus to measure the increase in penetration resistance of a hydrating paste sample with time. Initial and final set times were determined according to the standard.

2.2.2.3 Isothermal Calorimetry

The rate of hydration of the OPC paste was studied in the first few hours after initial contact between cement and water using an isothermal calorimeter. This test was performed to relate the rate of hydration and potential pore solution composition changes to the change in the SAP absorption at early ages (due to an increase in the pore solution ionic concentration), and to the setting time of the OPC paste system. The goal was to determine whether it was possible to use the absorption of the SAP in pore solution extracted 30 minutes after mixing began as a representative value of the SAP absorption for mixture proportioning. After mixing, a glass ampoule was filled with approximately 6 g of the cement paste. The ampoules were then sealed and placed in the isothermal calorimeter (TA instruments), which was preconditioned at a temperature of 23 ± 0.1°C. The heat flow was recorded for a period of 6 hours from mixing. Duplicate specimens were tested and the heat flow normalized to the cement content in the paste was recorded.

2.2.2.4 XRF Analysis of Chemical Composition of Ionic Solutions

The chemical composition of the pore solutions extracted from the fresh pastes was analyzed through X-ray fluorescence (Tsui-Chang et al. 2018; Tsui-Chang et al. 2017). A solution sample (4 to 5 g of each of the solutions) was placed in a plastic cup with a 4-μm thick polypropylene film base for the determination of the four primary ionic species (Na, K, S, Ca) using XRF (Tsui-Chang et al. 2018; Tsui-Chang et al. 2017).

2.2.2.5 Degree of Hydration and Powers’ Model

In this study, the degree of hydration of the OPC paste was estimated based on the Parrot and Killoh model (Parrot and Killoh; Lothenbach and Winnefeld, 2006). The degree of hydration was estimated at the same ages at which the pore solution was extracted from the OPC paste (0.1, 1, 3, 5, 7, 14, 24 hours and 3, 7, 14, 28 days). The model uses the water-to-cement ratio, the ambient temperature

22

and the phase composition of the cement in order estimate the evolution of the degree of hydration of the cement paste. Powers’ model is then used to correlate the calculated degree of hydration to the volume fraction of the following phases: capillary water, unreacted cement, chemical shrinkage, gel water and gel solid (Powers and Brownyard, 1946).

2.2.2.6 SAP Absorption

In order to measure the SAP absorption, the gravimetric method known as ‘teabag’ method was used (Miller et al, 2014; Schröfl et al, 2017). The method uses a teabag and fills it with SAP to measure the weight change of the SAP when soaked in solution. It has been shown that the absorption of the SAP is highly dependent on the ionic strength of the solution used (Zhao et al, 2005; Zohuriaan-Mehr and Kabiri, 2008). To better understand the influence of ions on the SAP absorption, the measurements were conducted on solutions with a range of different solutions made with varying ionic concentrations and extracted pore solutions.

The teabag method uses a known quantity of SAP in a dry state, which is introduced in a dry teabag of known mass and absorption. For this test, a quantity of approximately 0.05 g of dry SAP was used in the test. The absorption of the teabag is measured in a similar manner with respect to the SAP, by introducing it inside the solution and measuring its change in weight after a determined period of soaking. After moving the dry SAP inside the teabag, the weight of the teabag with the SAP in a dry state is then recorded. The teabag with the SAP inside is then soaked in the solution that is being investigated. After 10 minutes from initial immersion, the teabag with the SAP is removed from the solution and potential agglomerates of dry SAP are dispersed by gently applying pressure with a finger. Agglomerates of SAP are usually recognized by their white color. The teabag is then reduced to surface dry condition by tapping it on a dry paper towel 8 times, alternating each time the tapped side. The weight is then recorded. The teabag is then moved back inside the solution for 20 more minutes, after which it is removed, reduced to surface dry condition in an equivalent manner to the step at 10 minutes and the weight is recorded. The measured weights are then used to calculate the absorption of the SAP using Eq. 2-1:

𝛷𝑆𝐴𝑃,30 =𝑊𝑆𝐷,𝑇𝐵+𝑆𝐴𝑃 − 𝑊𝑆𝐷,𝑇𝐵−𝑊𝐷,𝑆𝐴𝑃

𝑊𝐷,𝑆𝐴𝑃 (2-1)

where:

ΦSAP,30 = absorption of solution of the SAP at 30 minutes, g/g,

WSD,TB+SAP = weight of the teabag with SAP in the soaked surface dry condition, g,

WSD,TB = weight of the teabag alone in the soaked surface dry condition, g, and

WD,SAP = weight of the dry SAP, g.

The absorption of the SAP was tested using the teabag method in DI water, in simulated pore solutions prepared with different molar concentrations of NaOH ranging from 0.01 to 1 M NaOH, in lime water, and in the extracted pore solutions.

23

2.2.2.7 Desorption Behavior of SAPs

In addition to measuring absorption, the desorption behavior of the SAP was studied through two different experiments: 1) response to a change in relative humidity as measured with a Dynamic Vapor Sorption (DVS) and 2) exposure of the saturated SAP to ionic solutions of greater concentration than the soaked solution. These are expected to be two of the three main driving factors in the SAP desorption when placed in concrete (Mönnig 2009). The third driving factor would be capillary suction (RILEM Technical Committee 225 2012; Farzanian and Ghahremaninezhad 2017). However, this was not analyzed in this study.

The desorption isotherms of the saturated SAP were measured using a dynamic vapor sorption analyzer (TA Q50000; TA Instruments). The changes in the mass of the sample at a set relative humidity were recorded as a function of time. The relative humidity in the DVS device was ranged from a maximum of 97% to 94% and then reduced and stabilized at 0%. These high relative humidity values were chosen as they mimic the early age conditions in cementitious systems.

Before testing, the SAP was soaked in solution similar to the steps in the teabag methods. Three different solutions, characterized by different ionic strengths were used: DI water, simulated pore solution (0.25 M NaOH solution), and extracted pore solution. Thirty minutes after initial immersion began, the teabag was removed from the solution and approximately 35 mg of the saturated SAP was rapidly transferred to a tared quartz pan of the DVS. Once the mass of the pan plus the sample was recorded, it was then moved into the dynamic sorption analyzer, which stabilized at 97% relative humidity. This first step was maintained until the change in mass was recorded to be < 0.001%/15 min. Once the equilibrium criterion was met, the relative humidity was reduced, using 1% humidity steps, while the set relative humidity was maintained until the change in mass was recorded to be < 0.001%/15 min or a period of 1440 minutes had passed (Castro et al. 2011). When the device reached the equilibrium for the 94% relative humidity set point, the relative humidity was then decreased to 0% and maintained for a period of 24 hours, to ensure the total desorption of the SAP. The 94% relative humidity set point was chosen to compare the desorption properties of the SAP to that used for LWA in ASTM C1761 -17.

The second set of experiments measured the influence of changes in ionic concentrations on the desorption of SAP. Samples of dry SAP were first soaked in a solution of known ionic concentration. The absorption of SAP was measured using Eq. 1 and the sample was then transferred to a solution of different ionic concentration. The absorption of the SAP was measured again after the sample had been immerged in the second solution for another 30 minutes. A third step involved the re-placing of the teabag in the first solution and the measurement of the ‘recovered’ absorption after 30 more minutes of soaking. The schematic of this experiment is shown in Figure 2-2.

24

Figure 2-2. Transfer of SAP between solutions with different ionic concentrations.

2.3 RESULTS AND DISCUSSION

2.3.1 Absorption of SAP as a Function of Time

The absorption of the SAP was recorded at four different soaking periods in DI water to study the effect of soaking time on the absorption. The results are shown in Figure 2-3. When soaking in DI water, the absorption is quite high (approximately 210 g/g at 30 minutes). It was observed that after 30 minutes of immersion, the absorption keeps increasing (approximately increasing 10% more at 60 minutes), after which it stabilizes.

Figure 2-3. Absorption of SAP in DI water as a function of soaking time (the error bars represent the coefficient of variation with respect to the average value).

While this is true for DI water, more ionic solutions, such as pore solutions, were observed not to show any increase in absorption after 30 minutes of soaking (Figure 2-3). The pore solution shown in

25

Figure 2-3 was expressed at 30 minutes from mixing (initial contact between cement and water in the OPC paste) and had a pH of 13.2. Absorption values at times greater than 30 minutes did not significantly deviate from the average absorption value of 22.10 g/g (with readings of ± 5% on the measured SAP absorption, which match the observed variability of the teabag test). This suggests that at a soaking time of 30 minutes, the SAP was fully swollen. Therefore, the teabag tests in this study were run with a total soaking time of 30 minutes.

To determine a representative time of extraction of the pore solutions, the correlation between the heat flow, set time, and the SAP absorption was studied. The aim was to determine if SAP absorption would significantly change between 30 minutes and set time and if this was related to the hydration kinetics of the OPC paste. Figure 2-4 shows the correlation between the time of extraction, heat flow, and measured SAP absorption for the OPC paste.

Figure 2-4. SAP absorption as a function of time from mixing compared to isothermal heat flow curve.

By comparing the heat flow to the change in SAP absorption, it was observed that during the induction period, there was no significant change in the SAP absorption. The absorption of the SAP measured with pore solution extracted at 5 hours (after final setting) was 16% lower than the SAP absorption measured 30 minutes from time of mixing. This suggests that the change in ionic concentration has already caused the SAP to partially desorb. The roughly constant value of the SAP absorption in the first hours is due to the relatively stable concentration of the main ionic species in solution at such early ages: Na, K, Ca and SO4.

These results suggest that the SAP absorption and desorption could be reasonably characterized using pore solutions extracted at 30 minutes from initial contact between the cementitious materials and water, likely because the concentration of the pore solution does not vary substantially during the that time till the time of setting.

26

2.3.2 Absorption of SAP as a Function of pH

Figure 2-5 shows the influence of the pore solution pH on the absorption of the SAP. These results are broadly similar to those observed by others in literature, although the absorption kinetics depend on SAP molecular structure and particle size (Schröfl et al. 2012; Esteves 2011) among other factors. It should be noted that for concrete, the pore solutions will have relatively high ionic concentrations (Andersson et al. 1989) and this influences the absorption of the SAP (Schröfl et al. 2012; Zhao et al. 2005). The absorption of the SAP reduces as the ionic concentrations of solutions increase, reducing from an average value of 210.2 g/g measured for DI water at 30 minutes, to a value of 15.5 g/g for the 1 M NaOH solution (pH = 14.0). Such reductions have also been noted by other authors and this has been hypothesized to be due to a screening effect of the ions on the SAP charges (Schröfl et al. 2012; Zhu et al. 2015), combined with a promotion of ionic crosslinking between the polymers (Zhu et al. 2015; Horkay et al. 2000). A logistic sigmoidal curve (Eq. 2-2) was found to fit the absorption data.

𝛷𝑆𝐴𝑃,30(𝑝𝐻) = 𝛷𝑆𝐴𝑃,𝑀𝐼𝑁 +(𝛷𝑆𝐴𝑃,𝑀𝐴𝑋 − 𝛷𝑆𝐴𝑃,𝑀𝐼𝑁)

(1 + 10(𝐿𝑂𝐺(𝑝𝐻0)−𝑝𝐻)∗𝑝) Eq. 2-2

where:

ΦSAP,30(pH) = absorption of the SAP at 30 minutes for a given pH, g/g,

ΦSAP,MIN = minimum SAP absorption measured, g/g,

ΦSAP,MAX = maximum SAP absorption measured, g/g,

LOG(pH0) = pH at which the inflection point of the curve is located, -,

pH = independent variable, log, and

p = fitting parameter, -.

Figure 2-5. SAP absorption as a function of pH.

27

The average absorption value for the extracted OPC pore solution (pH = 13.2) was measured to be 22.1 g/g. This value lies between the measured absorption of the 0.5 M NaOH simulated pore solution (19.5 g/g), which has a calculated pH of 13.5, and the 0.1 M NaOH simulated pore solution (31.5 g/g), which has a calculated pH of 12.7. This shows that the extracted pore solution, while including divalent (Ca2+, SO4

2-) ionic species, has a similar absorption behavior to that of the simulated, monovalent ionic solutions having similar pH. The testing of the SAP absorption in limewater, where the presence of calcium, a divalent ion, generates a pH of 12.4, was tested to verify the accuracy of the previous statement. The determined average SAP absorption of 31.3 g/g in limewater slightly deviates from the empirical trend of the monovalent ionic solutions, but still shows a reasonable agreement. This suggests that for this type of SAP, the ionic strength has a more profound impact than the actual nature of the present ions (monovalent and divalent). This could be seen as unusual, as it is known from literature that for some SAPs, divalent and trivalent ions can have a stronger effect on the absorption, depending on the SAP molecular structure (Schröfl et al. 2012). Further studies on the role of the oxidation number of dissolved ions in the pore solution on the SAP absorption are planned for the future.

2.3.3 Influence of Inclusion of SCMs on SAP Absorption

A potential influence of SCM replacements on the absorption of the SAP was explored. Figure 2-6 shows the influence of increasing volume replacements of SCMs on the ionic concentration of the pore solutions extracted 30 minutes after mixing. The pore solution composition (ionic concentration) measurements were conducted with the use of XRF (Parrot and Killoh 1984). Alkali concentrations decrease with increasing fly ash replacement. While the alkali content in the fly ash is higher than in the cement, the measurements show lower concentrations of alkalis in the pore solutions of the blended systems. This can be explained by a higher percentage of the alkalis being fixed in various nonreactive phases of the fly ash, which prevents them from dissolving in the pore solution (Shehata et al. 1999). Calcium and sulfate are expected to be within the same range of the OPC case, with sulfate having slightly lower concentrations in the pore solution of blended systems (Vollpracht et al. 2016). This is confirmed from the experimental data. Assuming the fly ash alkalis do not dissolve (Shehata et al. 1999), and a low reactivity of the SCM at early ages, the lower concentrations of ions pore solutions of the blended pastes can be studied as a simple dilution effect, where the concentration of a particular ion in the fly ash-OPC system can be estimated by multiplying the OPC system ion concentration by the dilution of the cement (one minus the volume fraction of the fly ash in the system). In Figure 2-6, the theoretical ionic concentrations due to dilution are plotted against the measured ones in dashed lines.

28

Figure 2-6. Ionic concentrations with increasing volume replacements of fly ash.

Figure 2-7 shows the absorption values of the SAP, measured on the pore solutions extracted at 30 minutes from the straight and blended pastes. On the same plot, the calculated pH of each solution is also shown. The pH was calculated from the OH- concentrations using a charge balance of the pore solution ions detected from the XRF.

Figure 2-7. Influence of SCM volume replacements on absorption of SAP.

The average absorption measured for the four systems was 23.7 g/g with a standard deviation of 1.3 g/g. The absorption does not appear to be strongly dependent on the SCM volume replacement and on the change in the composition of the pore solution which occurs for the blended pastes.

29

While the previous experiments showed that the teabag method shows a significant repeatability, with a measured coefficient of variation of 6%, it is of interest to explore the effects of systematic errors and their potential impact on the w/cm of the system.

In Figure 2-8, the possibility of a systematic error associated with the teabag method and the potential impact on the w/cm of the system is explored. An erroneous assumption of the SAP absorption might result in a lower or higher amount of free water in the system at early ages, which would affect the w/cm and, therefore, lead to changes in the porosity.

By assuming a cement content of 313 kg/m3, taken from a mixture design being used in the field (Montanari et al. 2018), and a w/cm of 0.36, the impact of an increasing error in the absorption measurement is calculated in terms of a change in the w/cm of the system. The amount of SAP included in the system is calculated following the Snyder and Bentz approach (Snyder and Bentz 1999), assuming a chemical shrinkage of 0.07 ml/g.cm (or L/kg.cm). Eq. 2-3 is used to calculate the impact on the w/cm:

𝑤

𝑐 𝑚𝑎𝑥=

𝑤 + (1 −𝛷𝑆𝐴𝑃,𝑎𝑐𝑡𝑢𝑎𝑙

𝛷𝑆𝐴𝑃,𝑡𝑒𝑎𝑏𝑎𝑔) ∗ 𝐶𝑆 ∗ 𝑐

𝑐

Eq. 2-3

where:

w = water content based on w/cm of 0.36, kg/m3

c = cementitious content (313 kg/m3), kg/m3

𝛷𝑆𝐴𝑃,𝑡𝑒𝑎𝑏𝑎𝑔 = absorption as mesured from teabag test, g/g

𝛷𝑆𝐴𝑃,𝑎𝑐𝑡𝑢𝑎𝑙 = actual absorption, g/g

CS = chemical shrinkage, 0.07 ml/g.cm (or L/kg.cm)

Eq. 2-4 is used to determine the error in the SAP absorption:

𝛷𝐸𝑟𝑟𝑜𝑟(%) =𝛷𝑆𝐴𝑃,𝑎𝑐𝑡𝑢𝑎𝑙 − 𝛷𝑆𝐴𝑃,𝑡𝑒𝑎𝑏𝑎𝑔

𝛷𝑆𝐴𝑃,𝑡𝑒𝑎𝑏𝑎𝑔∗ 100 Eq. 2-4

where:

𝛷𝐸𝑟𝑟𝑜𝑟 = error in the measured absorption, %

𝛷𝑆𝐴𝑃,𝑡𝑒𝑎𝑏𝑎𝑔 = absorption as measured from teabag test, g/g

𝛷𝑆𝐴𝑃,𝑎𝑐𝑡𝑢𝑎𝑙 = actual absorption, g/g

30

Figure 2-8. Impact on w/cm from a systematic error in the SAP absorption.

From these calculations, a 50% underestimation of the absorption would decrease the w/cm to 0.32; while an overestimation of 50% would increase it to 0.40. As the teabag method has a single operator precision of approximately 6%, in the absence of systematic errors, this leads to a variation of the w/cm between 0.356 and 0.363. This low variation in the method ensures a very limited influence from the presence of SAP on the effective w/cm of the system.

2.3.4 Influence of External Relative Humidity on SAP Desorption

Figure 2-9 shows the degree of saturation (DOS, the normalized moisture content, or the ratio of the fluid content of the SAP at any RH and the maximum fluid content of the SAP for any particular fluid), as measured from the DVS for saturated SAP (exposed to DI water, 0.25 M NaOH solution (with pH 13.4) and extracted pore solution (with pH 13.3)). An assumption of full saturation of the SAP (DOS = 100%) at the start of testing is made. The degree of desorption of the SAP is influenced by the presence of ions in the solution. For a given relative humidity, solutions with a lower ionic concentration will release more water. The presence of ionic species causes the SAP to retain an amount of solution for a given relative humidity as the vapor pressure of the solution itself will come to equilibrium with the external environment and not lose additional water (for example, a saturated NaCl solution at an ambient temperature of 23°C, will be in equilibrium at 75% RH (Greenspan 1977)). This, in some extreme cases, might limit the efficiency of the SAP which will not be able to provide sufficient solution to mitigate the process of self-desiccation (Castro et al. 2011). It was observed that the SAP conformed to the requirement of ASTM C1761-17 (ASTM Standard 2015), when tested with extracted pore solution, which requires the internal curing agent to release 85% or more of the water at a relative humidity of 94%.

31

Figure 2-9. Degree of saturation of SAP as a function of relative humidity.

As discussed earlier, a lower degree of desorption of the SAP in concentrated ionic solutions is generally observed. However, this data shows that the desorption may be limited with SAP. The equilibrium relative humidity and the higher concentration of ions, which promote ionic crosslinking in the SAP, might explain this phenomenon (Schröfl et al. 2012; Zhu et al. 2015; Horkay et al. 2000). In order to account for the partial desorption of internal curing agents at high relative humidity, Castro et al. 2011 proposed a correction to be applied to the Snyder and Bentz 1999 equation, used in the design of internal curing mixtures, which can be extended to SAP applications.

2.3.5 Changing Ionic Concentration to Study SAP Desorption

Using the teabag method, the kinetics of desorption due to an increase in the solution ionic concentration were studied. It was observed that moving the SAP from a solution with a lower ionic concentration to a higher ionic concentration results in rapid desorption of the SAP. The results of several such experiments are shown in Table 2-2 (with 30 minutes of soaking in each solution (Figure 2-2, Page 33)).

32

Table 2-2. SAP absorption values after being moved into more concentrated solutions (soak time in each solution is 30 minutes)

Test

Solution 1 Absorption (g/g)



1 DI water

0.01 M NaOH

DI water

203.79 145.85 190.30

2 DI water

0.25 M NaOH

DI water

207.35 25.5 79.01

3

0.01 M NaOH

0.25 M NaOH

0.01 M NaOH

83.24 34.36 56.43

Moving the teabag with the SAP to the second solution, with a higher ionic concentration, caused the SAP to release some of the water. When the sample is then moved back to the original solution, there is a limited recovery of the initial absorption. This difference is higher in tests number 2 and 3, where the second solution has higher ionic concentration (0.25 M NaOH) with respect to test number 1 (0.01 M NaOH): the difference with respect to initial absorption is only – 6.8% in test 1, – 61.9% in test 2, and -32.2% in test 3.

This partial recovery of the absorption might be caused by the transfer of some ions to the initial solution, which would increase its ionic concentration and pH and cause a lower absorption of the SAP. Another explanation involves the potential ionic crosslinking of the polymers which might have occurred in the second ionic solution.

To determine the actual cause, an experiment was conducted where the teabag and SAP were moved from a 0.25 M NaOH to an “infinite volume” of DI Water, which was simulated with a large beaker filled with DI water. Additionally, to remove the influence of time, the soaking time was extended to 120 minutes. This was done with the aim of inducing an infinite dilution of any potential ion transferred to the first solution.

The results showed that, even when an ‘infinite’ dilution volume was provided and the soaking time was extended to 120 minutes, the SAP did not recover the initial absorption with an average reduction of approximately 47% (111 g/g from 210 g/g). This suggests possible formation of ionic crosslinking sites between the sodium ions and the hydrolyzed acrylamide segments of the polymer network during the soaking of the SAP in the second solution. These crosslinks may persist inside the SAP after moving to the third (original) solution due to an insufficient osmotic pressure gradient (partially caused by the limited soaking time).

33

In order to further study the interactions between ions and SAP, experiments were conducted to verify whether the SAP, when soaked in solution, would change the ionic concentration of the solution. The teabag method was used and the SAP was exposed to the 30 minutes extracted fresh pore solution for a period of 5 hours. Periodically, the solution concentration was tested using XRF (Tsui-Chang et al. 2017) and the influence of the SAP on the ionic concentration was analyzed. Figure 2-10 shows the results of the experiment.

Figure 2-10. Changes in ionic concentration of solution as a function of SAP soaking time.

From the measurements, it appears that the SAP absorbs the solution without changing the ionic concentration, as this remains constant throughout the duration of the experiment.

Based on this observation, we can expect the SAP to absorb a fluid with the same ionic concentration as in the surrounding solution. Since for SAP it is more appropriate to talk about ‘fluid absorption’ rather than ‘water absorption’, it is important to understand whether the presence of these ions in the SAP could impact calculations of SAP absorption.

In particular, the presence of ions could have potentially added extra mass to the teabag that would result in overestimating the amount of water which was actually absorbed.

However, by analyzing the pore solution ionic concentration and calculating the relative mass fraction of the ions with respect to the total weight of the solution, it was calculated that the presence of ions impacts the measured fluid absorption by less than 2%, and was therefore considered to be negligible.

2.3.6 Hydration of the System and Its Influence on SAP Desorption

To determine how the pore solution changes as a function of the degree of hydration, the pore solution of the OPC system was extracted at several ages from the time of mixing up to 28 days.

34

The composition of the pore solutions was measured using XRF. Figure 2-11a shows the evolution of the ionic species in the pore solution as a function of age. As the cement hydrates, the ionic concentration of the pore solution increases, primarily due to a reduction in the volume of the liquid phase. This observation was confirmed by combining Powers’ model prediction of the volume fraction of available free water (Powers and Brownyard 1946) with the calculated degree of hydration as a function of time based on the work of Parrot and Killoh 1984. To predict the change in concentration of sodium and potassium as a function of time, the first measured ionic concentrations of sodium and potassium at 10 minutes were normalized by the reduced fraction of free water for a given degree of hydration, which is indicated in Powers’ model as the sum of gel water and capillary water (Powers and Brownyard 1946). Figure 2-11b shows the comparison between the concentration predictions and the measured concentrations for potassium and sodium ions.

The pH was calculated based on the hydroxyl ions concentration. The change in pH in the first 72 hours is primarily due to the depletion of the sulfate ions and the increase in the potassium and sodium ionic concentrations. Using Eq. 2, the theoretical absorption of the SAP was calculated for each pH. In Figure 2-11c, the SAP absorption and the pore solution pH are plotted as a function of the paste age. The plot shows that changes in the SAP absorption that can occur due to an increase in the ionic concentration of the pore solution occur between 5 and 72 hours. During this period of time, it was calculated that approximately 31% of the initially absorbed solution is released from the SAP due to an increase in the pH. This will occur without capillary suction or a change in the relative pore pressure.

35

Figure 2-11. (a) Ionic concentrations evolutions as a function of time from mix, (b) Prediction in ionic concentration changes using Powers’ model, (c) Changes in SAP absorption and pH as a

function of time from mix.

2.4 CONCLUSIONS

A commercially available SAP was characterized for use as an internal curing agent in concrete. While many of the measured properties are specific to this particular SAP, the procedures adopted in this

36

study can be extended to any SAP used as an internal curing agent. The following conclusions can be drawn:

The teabag method can be used to assess the SAP absorption. This was done using pore solution extracted from fresh paste at an age of 30 minutes since the pore solution is not typically observed to change significantly during the dormant period (i.e., from thirty minutes after the initial contact between cement and water and the time of setting).

The ionic concentration of the solution influences the absorption of the SAP. More highly concentrated ionic solutions result in lower absorption. A relatively modest increase in the SAP absorption was observed when comparing the OPC system and the OPC-fly ash systems (up to 17% with the 60% fly ash replacement).

The ionic concentration of the solution was observed to influence the desorption of the SAP. A slower rate of desorption was observed when the ionic concentration is increased, for a given relative humidity. As the relative humidity is changed, the SAP releases some of the water, until the solution reaches a salt concentration which is in equilibrium again with the environment. The SAP had greater than 85% desorption as prescribed in ASTM 1761-17 for FLWA (Montanari et al 2017). The partial desorption is particularly important in the case of internal curing applications, where the agent needs to release most of its internal curing water volume at high relative humidity.

Soaking the SAP in a solution and moving it to a solution with a higher ionic concentration resulted in desorption. This occurred until the SAP reached equilibrium with the ionic concentration of the second solution. Moving the SAP to a more concentrated solution causes an irreversible decrease in its absorption capacity.

By studying the ionic concentration of a pore solution from a hydrating OPC system, the SAP absorption was determined as a function of age (i.e., degree of hydration). It was observed that the increase in the ionic concentration is mainly caused by a consumption of free water from the cement hydration. The change in ionic concentration that occurs between the time of setting and 72 hours corresponds to 31% of the fluid being desorbed from the SAP. This is a desorption mechanism that has not been widely explored and should be considered more carefully when SAP is used.

37

CHAPTER 3: TOWARD A DESIGN METHODOLOGY FOR PARTIAL VOLUMES OF INTERNAL CURING WATER BASED ON THE REDUCTION OF AUTOGENOUS SHRINKAGE

3.1 INTRODUCTION

Internal curing was developed to improve the performance of concrete by reducing the risk of early age cracking in low water-to-cement ratio (w/c) concrete mixtures (Bentur 2003; Bentz and Weiss 2011; Geiker et al. 2004; Kovler and Jensen 2005). A commonly used mixture design methodology for internally cured mixtures is based on replacing a portion of fine aggregates with pre-wetted fine lightweight aggregates (FLWA) (Snyder and Bentz 1999). This approach provides internal curing water that has an equivalent volume to that of the chemical shrinkage in the mixture. The procedure, outlined in ASTM C1761 ̶ 15 (ASTM Standard 2015), estimates the mass of FLWA, MFLWA, as described by Eq. 3-1:

𝑀𝐹𝐿𝑊𝐴 =𝐶𝑓 𝐶𝑆 𝛼𝑚𝑎𝑥

𝑆 𝛷𝐹𝐿𝑊𝐴 Eq. 3-1

where: Cf is the amount of cement or cementitious material in the mixture (kg/m3), CS is the chemical shrinkage of the cement or cementitious blend (ml/gcm), αmax is the expected maximum degree of hydration (-), between 0 and 1, S is the level of saturation of the FLWA (-), between 0 and 1, and ΦFLWA is the absorption capacity of the FLWA (g/g). This procedure was modified (Castro et al. 2011) to account for the FLWA desorption and time dependent effects of aggregate absorption.

While higher w/c systems (w/c ≥ 0.42) do not experience the same magnitude of autogenous shrinkage, incorporation of FLWA into such mixtures still increases the strength (Espinoza-Hijazin and Lopez 2011) and can reduce the curling in concrete pavements (Amirkhanian and Roesler 2017; Rao and Darter 2013; Wei and Hansen 2008). Internal curing and FLWA have been implemented in bridge decks, water tanks, repair materials, and pavements with promising results (Barrett et al. 2015; Bentz and Weiss 2011; Friggle and Reeves 2008).

It has been hypothesized that the partial replacement of the volumes of FLWA calculated through the ASTM C1761 ̶ 13 approach can provide a majority of the benefits that would come from the total replacement (Henkensiefken et al. 2009). Reducing the volume of FLWA would reduce the cost of the concrete and improve staging for aggregate preparation and construction, while still maintaining similar benefits to design using Eq. 3-1.

The shrinkage properties of a cementitious system are dependent on the pore structure of the cement paste. The procedure shown in Eq. 3-1, however, does not account for the pore size and determines a sufficient amount of water to fill all the voids created by self-desiccation. This study explores a revision of the design approach, which considers the pore size distribution in the determination of the mass of FLWA to be included. Autogenous shrinkage measurements on partially internally cured mortar systems are performed.

38

3.2 EXPERIMENTAL PROCEDURE

3.2.1 Materials

The cement used in this study was an ordinary Type I/II portland cement (ASTM 2017). The Blaine fineness of the cement was 383 m2/kg and the potential phase composition from the mill sheet was 47% C3S, 21% C2S, 7% C3A, 9% C4AF. A Class C fly ash (complying with ASTM C618 ̶ 15 (ASTM 2015b)) and a Grade 100 slag (complying with ASTM C989 ̶ 16 (ASTM 2016a)) were used in some mixtures. The fly ash had a CaO content of 25.45%, a 12.5% by mass retained on a #325 sieve, and a specific gravity of 2.74. The slag had a Blaine fineness of 573 m2/kg and a specific gravity of 2.89. The fine aggregate used was crushed river gravel, with a specific gravity of 2.67, a fineness modulus of 2.8, and an absorption capacity of 1.4%. The fine lightweight aggregate was a water-cooled expanded blast furnace slag, with a fineness modulus of 3.4, a 24-hour absorption capacity of 10.5%, and a 72-hour absorption capacity of 12%, which were measured through the centrifuge method (Miller et al. 2014).

3.2.2 Mixture Design and Mixing Procedure

A total of fifteen different mortar mixtures were cast. Three different series of mixtures were evaluated with five different FLWA replacement levels. The mixtures are similar to those used in paving applications.

Table 3-1. Mixture proportions for the mortar systems examined

MIXTURE DESIGNATION

w/c or w/cm

Cement (kg/m3)

Fly ash (kg/m3)

Slag (kg/m3)

Water (kg/m3)

Sand (kg/m3)

FLWA, dry (kg/m3)

Water FLWA (kg/m3)

Water Sand (kg/m3)

Seri

es 1

OPC 0% 0.36 638 - - 230 1493 0 0 21

OPC 25% 0.36 638 - - 230 1343 114 12 19

OPC 50% 0.36 638 - - 230 1193 228 24 17

OPC 75% 0.36 638 - - 230 1044 342 36 15

OPC 100% 0.36 638 - - 230 894 456 48 13

Seri

es 2

Ternary 0% 0.36 374 176 73 224 1493 0 0 21

Ternary 25% 0.36 374 176 73 224 1347 111 12 19

Ternary 50% 0.36 374 176 73 224 1201 222 23 17

Ternary 75% 0.36 374 176 73 224 1054 334 35 15

Ternary 100% 0.36 374 176 73 224 908 445 47 13

Seri

es 3

Ternary 0% 0.417 346 162 67 240 1493 0 0 21

Ternary 25% 0.417 346 162 67 240 1377 103 11 19

Ternary 50% 0.417 346 162 67 240 1223 206 22 17

Ternary 75% 0.417 346 162 67 240 1087 308 32 15

Ternary 100% 0.417 346 162 67 240 952 411 43 13

39

Series 1 consisted of an ordinary portland cement (OPC) binder with a w/c of 0.36. Series 2 and 3 consisted of ternary systems, with water-to-cementitious ratio (w/cm) 0.36 and 0.417, respectively. These systems had fly ash and slag volumetric replacement levels of 31% and 12% with respect to the total volume of cementitious materials. The fine aggregate represented 57% of the mortar mixture volume. The FLWA modified mixtures were cast using five different levels of replacement: 0%, 25%, 50%, 75% and 100% with respect to the volume calculated by Eq. 3-1.

The FLWA were oven dried to constant mass before being soaked in the total volume of the mixture design water for a period of 24 ± 1 hours before mixing. Fine aggregate (sand) was oven dried to constant mass and conditioned at room temperature for a period of at least 24 hours before mixing. The paste samples were used to determine the pore size distribution and for chemical shrinkage measurements. The mortar samples were used for the autogenous shrinkage and relative humidity measurements. The mortars and pastes were mixed in accordance with ASTM C305 ̶ 14. The pastes were mixed using the same proportions as in the mortar paste fractions.


The desorption isotherms of the FLWA and pastes were measured using a dynamic vapor sorption (DVS) analyzer (Q5000SA, TA Instruments). The DVS measured the changes in the mass of the sample as a function of relative humidity at a constant temperature. The relative humidity was decreased stepwise from a maximum of 97.5% to a minimum of 0%, which corresponds to oven dried conditions. For the FLWA, the sample size was approximately 50 mg and it was introduced in a prewetted surface dry state. The DVS was first stabilized at a relative humidity of 97.5% for a period of 24 hours. The relative humidity was decreased to 80% through 1% relative humidity steps. Each relative humidity set value was maintained until the change of mass was recorded to be less than 0.001%/15 minutes or a period of time of 12 hours had passed (Castro et al. 2011). In the final step, the relative humidity was dropped to 0% and the sample was stabilized for a period of 48 hours, to ensure all the water was released. The pastes’ desorption isotherms were measured in a similar manner, with the main difference being that the relative humidity was dropped of a 7.5% first step, followed by 10% steps from 90% until 0%.

Autogenous shrinkage measurements were performed on each mortar mixture following ASTM C1698 ̶ 13. Three replicate specimens were measured for each mortar mixture. The length change was measured with spring loaded linear variable differential transformers (LVDTs). Data from each gauge was recorded at one-minute intervals for seven days. Mass was recorded at the beginning of testing and the end to ensure that the mortars had been properly sealed. Data measurements were recorded from set time which was measured for each mixture following ASTM C403/C403M ̶ 08.

Internal relative humidity measurements of the mortar mixtures were measured using crushed specimens at specific ages (Castro et al. 2016). The mortar samples were cast and sealed inside cylindrical plastic containers for the first 24 hours at a constant temperature of 23 ± 0.1°C. After 24 hours, the samples were demolded and crushed using a mortar and pestle. Approximately 5 grams of crushed mortar was collected after being sieved through a #8 sieve and placed in 12 mm deep plastic cups. The cups were then placed inside water-jacketed stainless-steel cylinders which provided temperature control of the samples in the containers and insulation from the external environment.

40

The relative humidity sensors (HygrClip2S, Rotronics) were mounted on the stainless cylinders which were then hermetically sealed. The sensors were calibrated using saturated salts solutions (Greenspan 1977). A linear correction curve based on the calibration was used to correct the relative humidity measurements (Castro et al. 2016). The sensors recorded the temperature and the relative humidity of the crushed samples at five-minute intervals. This procedure was repeated for each recorded measurement: each day for the first 7 days, at 21 days, and at 28 days.

3.3 EXPERIMENTAL RESULTS

The desorption behavior of the FLWA used in this study was measured using the DVS (Figure 3-1). The aggregate source used meets the requirements of ASTM C1761 ̶ 15 with the exception of the ability to release at least 85% of its water at 94% RH as measured by the DVS procedure, where the aggregate was shown to release only approximately 76% of the absorbed moisture (Figure 3-1). It has been shown previously that water-cooled expanded blast-furnace slag aggregates desorb less at high relative humidity compared to other fine lightweight aggregates (House et al. 2014).

Figure 3-1. Desorption isotherm of FLWA sample and the ASTM C1761 specification.

The internal relative humidity measurements (Figure 3-2) show that the relative humidity increases with the FLWA content, similar to measurements reported elsewhere (Castro et al. 2016). The influence of internal curing water is more pronounced for the lower w/c or w/cm systems, where the systems self-desiccate rapidly. The systems including supplementary cementitious materials (SCMs) show higher relative humidity at early ages. The presence of fly ash and slag reduces the initial consumption of water as the pozzolanic reactions are rather slow at early ages (at least for fly ash (Mounanga et al. 2011)). This results in higher relative humidity readings.

42

Figure 3-2. Internal relative humidity measurements for (a) Series 1, (b) Series 2, and (c) Series 3 (Accuracy: ± 1% RH).

Autogenous shrinkage measurements were performed on all the mixtures (Figure 3-3). An initial expansion occurs for all the internally cured mixtures. This expansion may be due to the increase of the system temperature. However, the temperature rise was measured to be than 1°C. The sudden release of internal curing water from the lightweight aggregates may also be contributing to the expansion (Cusson and Hoogeveen 2008).

43

Figure 3-3. Autogenous shrinkage measurements for (a) Series 1, (b) Series 2, (c) Series 3.

Autogenous shrinkage measurements for the OPC system (Series 1) show generally higher values of shrinkage with respect to the ternary system with the same w/cm (Series 2), due to higher self-desiccation, which leads to a lower internal relative humidity. For the purpose of evaluating the benefits of internal curing in terms of autogenous shrinkage, a reduction factor was defined, which calculates the normalized reduction in shrinkage at 7 days of the internally cured mixtures with respect to the 0% system:

44

𝑅 = −휀𝐼𝐶,𝑋% − 휀0%

휀0% Eq. 3-2

where R is the reduction factor (-), εIC,X% is the shrinkage of the internally cured mixture with X% replacement of FLWA at 7 days (µε), ε0% and is the shrinkage of the non-internally cured mixture at 7 days (µε).

Table 3-2 shows the calculated shrinkage reduction at 7 days of the internally cured mortars, when the initial expansion is included.

Table. 3-2. Shrinkage reduction of internally cured mortars at 7 days

FLWA % ε 7D (με) R

Seri

es 1

OPC 0% -277 - OPC 25% -183 0.34 OPC 50% -106 0.62 OPC 75% -10 0.96 OPC 100% 65 1.23

Seri

es 2

Ternary 0% -125 - Ternary 25% -83 0.34 Ternary 50% -44 0.65 Ternary 75% -34 0.73 Ternary 100% -11 0.91

Seri

es 3

Ternary 0% -57.7 - Ternary 25% -30 0.48 Ternary 50% -11 0.81 Ternary 75% 10.7 1.19 Ternary 100% 51 1.88

In general, due to the high w/cm of the system and the presence of large volumes of SCMs, the shrinkages of Series 2 and 3 have relatively small values.

Desorption isotherms were obtained on paste mixtures. The pore size distribution was obtained using the Kelvin-Laplace equation (Eq. 3-3), which correlates the relative humidity to the Kelvin pore radius (Brunauer et al. 1967),

𝑟 =2𝛾𝑉𝑚

𝑅𝑇𝑙𝑛(𝑅𝐻) Eq. 3-3

where γ is the surface energy of the water (0.072284 N/m), Vm is the molar volume of water (1.8·105 mol/m3), R is the ideal gas constant (8.3145 N·m/mol·K), T is the temperature (K), and RH is the relative humidity in decimal form. For the ternary system, the paste was tested at three different ages to observe the change in the pore size distribution as a function of time. As expected, older specimens have a higher proportion of fine pores and a lower total porosity due to increased

45

hydration. When comparing the pore size distribution for the three different paste systems analyzed at an age of seven days, the OPC system has the finest pore size distribution, characterized by smaller pores. This is partly due to the higher degree of hydration of the OPC system with respect to the ternary systems at an age of seven days. Furthermore, in comparing the two ternary systems, the paste with the higher w/cm contains a higher volume of pores, mainly due to the higher volume of capillary pores (Figure 3-4).

Figure 3-4. Cumulative particle size distribution for the three paste systems, at an age of seven days, calculated from DVS data using Kelvin-Laplace equation.

3.3.1 Proposed IC Mixture Design Methodology

Mackenzie (Mackenzie 1950) developed an expression to predict the shrinkage of a solid body with spherical holes, εp, which can be applied to a cementitious paste using Eq. 3-4,

휀𝑃 =𝜎𝑐𝑎𝑝

3(

1

𝐾−

1

𝐾𝑆) Eq. 3-4

where σcap is the capillary pressure that develops in the solution (MPa), K is the bulk modulus of the paste (MPa), and KS is the skeleton modulus of the paste (MPa). The capillary pressure can be calculated as a function of the pore radius using Eq. 3-5.

𝜎𝑐𝑎𝑝 =−2𝛾𝑐𝑜𝑠𝜗

𝑟 Eq. 3-5

46

where γ is the surface tension of the fluid (0.072284 N/m for water), 𝜗 is the contact angle of the fluid and can be assumed to be 0° (Selander 2010; Villani et al. 2014). Bentz modified Eq. 3-4 by adding a saturation term (Bentz 2005), S (-), so that Eq. 3-4 now becomes Eq. 3-6.

휀𝑃 = −2

3 𝑆

𝛾

𝑟(

1

𝐾−

1

𝐾𝑆) Eq. 3-6

This equation uses the radius of curvature for the meniscus in the pores to quantify the capillary stress in the matrix. Practically, the radius of the meniscus corresponds to the radius of the largest fluid filled pore. In order to determine the degree of saturation of the system, S, which is defined as the ratio of the cumulative fluid filled pore volume Vc and the total pore volume Vtot, the pore size distribution and the relative humidity of the sample are needed. It is possible to fit the cumulative pore volume at any arbitrary filled state with a function. For this study, Vc was fitted using a logistic function to the pore volume data calculated from the cement paste desorption results. This function uses physical quantities as fitting values, and is shown in Eq. 3-7:

𝑉𝑐 = 𝑉𝑚𝑖𝑛 +𝑉𝑚𝑎𝑥 − 𝑉𝑚𝑖𝑛

(1 + (𝑟𝑟𝑖

)𝑠

)𝑐

Eq. 3-7

where Vc is the fluid filled cumulative pore volume at a given relative humidity (ml/gcem), Vmin and Vmax are the minimum and maximum pore volumes in the paste (ml/gcem), r is the radius of largest filled pore (nm), ri is the pore radius corresponding to the inflection point of the fitting curve (nm), s and c are fitting parameters which describe the sigmoidal behavior of the curve (-). For the OPC system, ri = 0.85 (nm), s = 0.325 (-), and h = -3.3 (-). To include Eq. 3-7, we can change how we express the degree of saturation in Eq. 3-6, so that we obtain Eq. 3-8:

휀𝑃 = −2

3

𝑉𝑐

𝑉𝑡𝑜𝑡 𝛾

𝑟(

1

𝐾−

1

𝐾𝑆) Eq. 3-8

where Vtot represents the total pore volume (ml/gcem), and the ratio between Vc/Vtot represents the degree of saturation of the system (-).

The proposed design methodology aims to predict the shrinkage reduction that comes from the provision of additional internal curing water to the cementitious system. By calculating the ratio between the shrinkage of the internally cured system and the shrinkage of the non-internally cured system, we obtain Eq. 3-9:

휀𝑃,𝑥

휀𝑃,0=

𝑉𝑐,𝑥

𝑉𝑐,0(

1𝑟𝑥

1𝑟0

) (

1𝐾𝑥

−1

𝐾𝑆,𝑥

1𝐾0

−1

𝐾𝑆,0

) Eq. 3-9

where the subscript ‘x’ refers the volume of FLWA present in the mixture and ‘0’ indicates the 0% FLWA system. The ratio between the two terms provides the fraction of the shrinkage of the 0% FLWA system that is obtained through partial internal curing. It can be assumed that, at early ages,

47

the modulus of the matrix of the mixture with no FLWA is similar to the modulus of the matrix of the internally cured system, so that the K terms cancel out in Eq. 3-9. This was considered reasonable especially when dealing with small FLWA replacements. The ratio can be redefined as a reduction factor, R, and thus, Eq. 3-9 can be simplified to Eq. 3-10.

𝑅 = 1 −𝑉𝑐,𝑥

𝑉𝑐,0

𝑟0

𝑟𝑥 Eq. 3-10

The aforementioned relationship assumes that the water held by the FLWA is completely desorbed at the internal relative humidity measured at seven days. For most FLWA, this assumption holds true (Castro et al. 2011). However, as an example of non-ideal behavior, the FLWA used in this study does not release all of its water at the relative humidity measured at seven days (Figure 3-1).

The determination of the r0 parameter is based on the internal relative humidity measurements of the non-internally cured sealed specimens at an age of seven days (Figure 3-2). In this section, the 0.36 w/c OPC system will be used as an example for the application of the methodology. For the OPC system, the 0% mixture has an internal relative humidity of 80.9% at 7 days. Using Eq. 3-2, the pore radius, r0, at the given relative humidity is 5 nm. If a reduction of shrinkage of 50% is desired, the pore radius that needs to be filled to, rx, can be calculated (11 nm) as seen in Figure 3-5.

Figure 3-5. Schematic of r0 and rx determination and calculation of water demand for a 50% reduction in autogenous shrinkage for OPC system.

Then, the water demand required to maintain that filled pore volume can be calculated by subtracting VF,x(rx) from VF,0(r0) and normalizing the obtained volume to the content of cementitious

48

material in the sample. The cementitious content can be obtained by accounting for the chemically bound water in the paste sample which was used for the desorption isotherm, as described in Eq. 3-11:

𝑔𝑐𝑚

𝑔𝑠𝑎𝑚𝑝𝑙𝑒=

1

(1 + 𝐷𝑂𝐻 ∙ 0.23) Eq. 3-11

where gcm/gsample is the normalized cementitious weight per sample weight (g/g), and DOH is the degree of hydration (-). The DOH was based on the chemical shrinkage at 7 days, which was measured to be 70% of the final value of 0.075 ml/gcm (Montanari 2017). The chemically bound water in the sample at 100% hydration was assumed to be 0.23 ml/gcm (Lam et al. 2000). This procedure yields a water demand of 0.014 mL water per gram of cementitious material for a target 50% shrinkage reduction.

The reduction factor, R, based on the water demand to fill up the pore volume up to 42 nm, is compared with the experimental shrinkage reductions, with measurements starting at 1 day, in order to exclude the initial expansion (Figure 3-6).

Figure 3-6. Comparison of experimental data with predicted reduction in shrinkage for OPC series.

The value of 42 nm was chosen as it represents the biggest radius that can be measured through DVS, approximately corresponding to a relative humidity of 97.5%. Nevertheless, filling the pore volume up to this radius size, was calculated to be enough to reduce the shrinkage by almost 90%.

Figure 3-6 shows that the predicted shrinkage reductions, calculated using Eq. 3-10, are much larger than the ones experimentally measured. This suggests that only a certain fraction of the internal curing water is filling in the vapor filled pores. This then leads to the obvious question, where does

49

the rest of the water go? One possible explanation of the discrepancy is that a portion of the internal curing water is used up in increasing the degree of hydration. This phenomenon is known from literature (Bentz and Stutzman 2008; Castro et al. 2016; Lura et al. 2006; Shin et al. 2010), with increases in degree of hydration between 5 and 10% being reported. It can be then assumed that this part of the internal curing water will become chemically bound and therefore unavailable to fill the vapor filled pores in the paste matrix. Other explanations include a lack of complete desorption from the FLWAs used in this study.

In general, therefore, it is then possible to assume that the internal curing water, at any time t, will be shared as described in Eq. 3-12:

𝑉𝐼𝐶 = 𝑉𝑝𝑣 + 𝑉𝛼 + 𝑉𝜓 + 𝑉𝑜𝑡ℎ𝑒𝑟 Eq. 3-12

where VIC is the total volume of chemical shrinkage water (ml/gcem), Vpv is the volume of internal curing water ending up filling the pores (ml/gcem), Vα is the volume of internally cured water chemically bound, due to an increase in the degree of hydration (ml/gcem), Vψ is the partial non desorbed water from the FLWA (ml/gcem), and Vother includes other unknown factors, such as possible reabsorption of water into other phases in the cementitious system (e.g., ettringite), slower than predicted desorption kinetics, etc (ml/gcem).

By using the experimental values of reduction in shrinkage with respect to the 0% FLWA system, it is possible to calculate the amount of internal curing water which ends up filling the pores, using Eq. 3-10. Assuming two values of an increase in degree of hydration, 5% and 10%, the amount of water which is chemically bound at 7 days was calculated as described by Eq. 3-13 (Lam et al. 2000):

𝑉𝛼 = 0.23 ∙ 𝐷𝑂𝐻𝑖 Eq. 3-13

where DOHi is the increased degree of hydration (-).

The volume of undesorbed water was measured by using the experimentally found filled volume and the increased degree of hydration, as described by Eq. 3-14:

𝑉𝜓 =(𝑉𝑝𝑣 + 𝑉𝛼)

(1 − 𝐷𝑂𝑆𝐹𝐿𝑊𝐴(𝑅𝐻)) 𝐷𝑂𝑆𝐹𝐿𝑊𝐴(𝑅𝐻) Eq. 3-14

where DOSFLWA(RH) is the degree of saturation of the FLWA for a set relative humidity (-).

Each filled volume has an associated relative humidity value, which can be used to determine the corresponding degree of saturation of the FLWA on the desorption isotherm (Figure 3-1).

Figure 3-7 shows the impact of the new additional terms in the sharing of internal curing water for the 50% FLWA case.

50

Figure 3-7. Impact of the inclusion of Vψ and Vα on the distribution of the internal curing water for the 50% FLWA volume replacement.

The chart illustrates how additional hydration and partial desorption of the internal curing agent has a large impact on the fraction of available water in the system. The ‘other water’ fraction on the pie chart indicates the portion of the internal curing water, which is not accounted for, even with the inclusion of Vψ and Vα. From these calculations, the ‘other water’ fraction is highly dependent on the expected increase in degree of hydration, with larger volumes of unaccounted water being present for larger FLWA volume replacements.

Although a portion of the internal curing water might not be available at the age of seven days, it is expected that the lightweight aggregates will keep desorbing at later ages, as suggested by the relative humidity measurements. It is shown that, especially for higher replacement levels of FLWA, relative humidity readings at 21 and 28 days are sufficiently high to continue hydration (Figure 3-2). The internal curing water released after the initial seven days will continue to increase the degree of hydration of the cement and reduce the autogenous shrinkage by filling the vapor filled pores in the cementitious matrix.

On the basis of the aforementioned considerations, the model was modified. In addition to the volume necessary to fill the target pore volume, a volume of water, Vα, due to an increase in the degree of hydration (assuming increases in degree of hydration of 0%, 5%, and 10%), was added. An additional volume comes from the partial desorption of the FLWA, Vψ. This was calculated from the desorption isotherm of the FLWA (Figure 3-1), as previously described. Finally, since chemical

51

shrinkage has not fully developed at such early ages, the total volume of water was normalized to the chemical shrinkage at 7 days, which from measurement was 70% of the final value of 0.075 ml/gcem.

Ignoring Vother, which cannot be directly calculated, the volume of internal curing water normalized to the chemical shrinkage, necessary for a given filled pore volume is calculated as shown in Eq. 3-15:

𝑉𝐼𝐶

𝑉𝐶𝑆,𝐹𝑖𝑛𝑎𝑙=

𝑉𝑝𝑣 + 𝑉𝛼 + 𝑉𝜓

𝑉𝐶𝑆,7 𝐷𝑎𝑦 Eq. 3-15

where VCS,Final is the ultimate chemical shrinkage (ml/gcem), and VCS,7 Day is the measured chemical shrinkage at 7 days (ml/gcem). Chemical shrinkage tests were performed as a part of this study, but are not shown here; they are detailed elsewhere (Montanari 2017).

Figure 3-8 shows the predicted shrinkage reduction and a comparison to the experimental data.

Figure 3-8. Shrinkage reduction prediction from model compared to experimental data for OPC series, considering the impact of additional hydration.

The three curves are representative of the three different assumptions that were made on the increased degree of hydration of the cement: 0%, 5%, and 10%. In order to calculate the volume of internal curing water, which is required for a target shrinkage reduction, the model starts from measuring the volume of water that is needed to fill the pores, as shown in Figure 3-5. To this volume of water, a second volume, which accounts for the increased hydration of the cement, is added. Finally, the volume of unabsorbed water is calculated, as described in Eq. 3-14, and added to the previously calculated two terms. The sum of these volumes is then normalized to the measured chemical shrinkage at 7 days, as the vapor filled space has not yet fully developed.

52

The comparison of the modified model with the experimental data suggests that the inclusion of the additional terms for additional hydration and partial desorption, together with the normalization with respect to the chemical shrinkage at 7 days, increases the reliability of the prediction of the expected shrinkage reduction. The predictions also suggest that the increase in the degree of hydration might depend on the FLWA replacement level. This hypothesis was confirmed when using SAP as internal curing agent in a parallel study (Montanari et al. 2017).

3.4 CONCLUSIONS

The current mixture design approach for internally cured concrete provides a volume of internal curing water, which is equal to the chemical shrinkage of the paste. This internal curing water reduces autogenous shrinkage, reducing the potential for early age cracking, increases the degree of hydration of cement and, possibly, the degree of reaction of SCMs, with respect to non-internally cured systems.

The increasing interest in the use of internal curing in large volume applications like concrete pavements has led to questions about the potential to reduce the volume of FLWA, while still maintaining the benefits of internal curing.

This study investigates an alternative methodology for determining the FLWA replacement volume, based on providing internal curing water to maintain a targeted relative humidity by considering the pore size distribution of the paste. The measurements on the mortar systems show benefits in terms of increased relative humidity and reduction in autogenous shrinkage, even for the lowest replacements of FLWA, with higher benefits coming from higher volumes of FLWA. The pore size distribution of the matrix was measured using desorption isotherms and used to establish the volume of internal curing water needed to fill the porous medium up to a target pore radius. An expression was developed in order to predict the shrinkage reduction that we obtain through the addition of partial volumes of internal curing water, with respect to the commonly used approach. The equation uses the additional volume of internal curing water to predict the reduction in shrinkage, as a function of the total filled pore volume and the Kelvin radius for the largest filled pore. Measurements of autogenous shrinkage and relative humidity demonstrate that only a fraction of the internal curing water, provided by the FLWA, ends up filling the pores; the remaining water is accounted for by other phenomena, such as increased degree of hydration, partial desorption of the internal curing agent and partial chemical shrinkage at early ages. The method developed based on some of these factors shows a better agreement with the experimental data, but work still needs to be done in order to explore additional causes that play a role in the partial availability of internal curing water at early ages.

53

CHAPTER 4: ACCOUNTING FOR WATER STORED IN SUPERABSORBENT POLYMERS IN INCREASING THE DEGREE OF HYDRATION AND REDUCING THE SHRINKAGE OF INTERNALLY CURED CEMENTITOUS MIXTURES

4.1 INTRODUCTION

Internal curing was developed to minimize the impact of self-desiccation by providing curing water throughout the cross section of low water-to-cement ratio (w/c) cementitious systems. This reduces the potential for early age cracking due to autogenous shrinkage, which is a common problem in high performance concrete (Henkensiefken et al. 2009). The curing water can be provided through the partial replacement of (fine) aggregates with pre-wetted lightweight aggregates (Schlitter et al. 2010; Castro et al. 2016; Kovler and Jensen 2005; Bentur 2003; Cusson and Hoogeveen 2008; Bentz and Weiss 2011), through the use of superabsorbent polymers (SAP) (Hasholt et al. 2012; Schröfl et al. 2012; Jensen and Hansen 2002), or through the use of absorptive fibers (Kawashima and Shah 2011).

Internal curing through the addition of SAP to concrete typically results in a reduction in the autogenous shrinkage (Hasholt et al. 2012; Wang et al. 2009; Craeye et al 2011) and an increase in the internal relative humidity (Shen et al. 2015; STAR 225-SAP 2012).

Typically, mixture proportioning for internal curing using SAP is done by providing a volume of internal curing water that is equivalent to the total chemical shrinkage (ASTM Standard 2015). In this work, the effect of providing fractional amounts of the internal curing water was studied. This is done with the objective of reducing the amount of SAP used, which could be beneficial in applications such as pavements where high volumes of materials are used. A design methodology for internally cured mixtures, based on calculating the amount of internal curing water required on the basis of the predicted shrinkage reduction, based on the paste pore size distribution (Montanari et al. 2017), is applied to SAPs.

It is shown that an increased degree of hydration and partial desorption of the internal curing agents, are factors that must be considered when calculating the volume of water, which is available to fill the (partially) vapor filled pores in the cementitious matrix.

4.2 MATERIALS AND EXPERIMENTAL PROCEDURES

4.2.1 Materials

A type I/II ordinary portland cement (OPC), complying with ASTM C150/C150M ̶ 17 (ASTM Standard 2017), was used in this study. Table 4-1 shows the oxide content of the cement, measured using X-ray fluorescence (XRF) (ASTM C114 ̶ 15) (Snyder and Bentz 1999). The cement had a Blaine fineness of 383 m2/kg and the phase composition from the mill sheet was 47% C3S, 21% C2S, 7% C3A, 9% C4AF (using cement chemistry notation, C: CaO, A: Al2O3, F: Fe2O3, and S: SiO2). The specific gravity of the cement was assumed to be 3.15.

54

Table 4-1. Oxide content of the cement measured through XRF

Phase Composition

Oxide Mass, %

Na2O 0.20 MgO 3.72 Al2O3 4.84 SiO2 19.91 SO3 4.05 K2O 1.06 CaO 60.93 Fe2O3 2.96

The sand used for the mortar samples was a typical river sand, with a specific gravity of 2.67 and an absorption of 1.4%. The SAP used in the experiments was a commercially available product (provided by BASF USA). The SAP were angular and had a particle size less than or equal to 150 μm in dry conditions. The main constituent was acrylamide. The specific gravity of the SAP was assumed to be 1.

4.2.2 Mixture Design and Mixing

The paste and mortar systems were designed to have a w/c of 0.36. The amount of SAP to include in the mixtures was calculated based on the method proposed by Snyder and Bentz 1999, as shown in Eq. 4-1. Equation 4-1 is based on the principle of providing a volume of water in SAP that is equal to the volume of chemical shrinkage:

𝑀𝑆𝐴𝑃 =𝐶𝑓 𝐶𝑆 𝛼𝑚𝑎𝑥

𝛷𝑆𝐴𝑃,30 Eq. 4-1

where:

MSAP = the mass of the dry SAP (kg/m3),

Cf = the cement content of the mixture (kg/m3),

CS = the chemical shrinkage of the cement (mL water/g cement),

αmax = the expected maximum degree of hydration (0 to 1) (unitless), and

ΦSAP = the absorption of the SAP in pore solution, here at 30 minutes absorption, (kg of solution/kg of dry SAP).

Five different mixtures were designed for use in this study, and each mixture included a different volume of SAP. The 100% SAP mixture included the amount of SAP calculated through Eq. 4-1. For the other mixtures, a portion of the 100% SAP was used. Table 4-2 shows the proportions of materials used for each of these mixtures.

55

Table 4-2. Mortar mixture proportions used in relative humidity and autogenous shrinkage measurements

Mixture Designation

w/c Cement, kg/m3

Water, kg/m3

Sand, kg/m3

Water in Sand, kg/m3

IC Water1, kg/m3

Dry SAP, kg/m3

0% SAP 0.36 705 254 1299 18 0 0

25% SAP 0.36 696 250 1281 18 13 0.5

50% SAP 0.36 687 247 1264 18 26 0.9

75% SAP 0.36 678 244 1248 17 38 1.3

100% SAP 0.36 670 241 1232 17 50 1.7 1IC water is the extra internal curing water that is added to the mixture, based on Eq 4-1.

Mortars were mixed following a modified version of ASTM C305 ̶ 14 (ASTM Standard 2014). The procedure was modified in order to take into account the presence of the SAP. After mixing the cement with water and the addition of dry aggregates, the SAP was introduced into the mixture in a dry state. While SAP may also be added directly to the dry mix, in that case, they may absorb mixing water and result in a reduction of the mortar workability. When SAP was added after water in the experiments reported here, issues with dispersion of the SAP and formation of conglomerates were not observed. An additional 60 seconds of mixing were added to the standard mixing steps, after the introduction of the dry SAP, to facilitate a good dispersion of the SAP.

In addition to the mortars, cement pastes with the same w/c were mixed in a vacuum mixer. The paste samples were used for pore solution extraction and isothermal calorimetry experiments. Samples for the isothermal calorimetry were designed with different volumes of SAP, following the same principle of the mortar mixtures, with SAP volumes of 0%, 25%, 50%, 75%, and 100% with respect to the Snyder and Bentz 1999 approach, as expressed by Eq. 4-1.

4.2.3 Experimental Procedures

4.2.3.1 Pore Solution Extraction

Pore solution was extracted from the fresh paste to measure the absorption of the SAP when soaked in the same. After mixing, the fresh paste was sealed in a plastic cylinder for 30 minutes, in order to minimize the loss of moisture to the environment. Then, the fresh paste was moved to a pore solution extractor (Rajabipour et al 2008).

The pore solution extractor uses compressed nitrogen at a pressure of 200 kPa to extract the pore solution from the fresh paste. A cellulose membrane filter with pore diameter of 0.45 μm prevents the cement grains from contaminating the extracted pore solution during the extraction procedure. The pressure was maintained for a period of five minutes. Once extracted, the solution was sealed inside plastic cylinders of 22 mm x 50 mm and stored inside an environmental chamber at a constant temperature of 5 °C, in order to minimize the potential for carbonation.

56

4.2.3.2 SAP Absorption

Absorption measurements of the SAP were conducted following the gravimetric method known as ‘teabag’ method (Schröfl et al. 2017; Miller et al. 2014). To perform the teabag test, a known quantity of dry SAP was introduced in a dry teabag of known mass and absorption. The mass of the teabag with the SAP was then recorded. The teabag was then submerged in the prepared extracted pore solution and initial soaking time was recorded. Ten minutes after initial immersion, the teabag was removed from the solution and patted dry using paper towels (to remove the surface moisture from the teabag and the SAP), and its mass was recorded. During this step, potential agglomerates of dry SAP were dispersed by gently applying pressure with a finger. The teabag was then re-introduced into the solution and after 30 minutes from the initial moment of immersion. Next, the teabag was again removed and its mass is recorded. The masses were then used to calculate the absorption of the SAP through Eq. 4-2:

𝛷𝑆𝐴𝑃,30 =𝑀𝑆𝐷,𝑇𝐵+𝑆𝐴𝑃 − 𝑀𝑆𝐷,𝑇𝐵−𝑀𝐷,𝑆𝐴𝑃

𝑀𝐷,𝑆𝐴𝑃 Eq. 4-2

where:

ΦSAP,30 = the absorption of the SAP at 30 minutes (g/g),

MSD,TB+SAP = the mass of the teabag with SAP in the soaked surface dry condition (g),

MSD,TB = the mass of the teabag alone in the soaked surface dry condition (g), and

MD,SAP = the mass of the dry SAP (g).

In this study, a 30-minute time of immersion was chosen based on the observation that longer soaking periods did not result in a substantial increase in the SAP absorption (average increase in absorption at 60 minutes = +2.1%, for pore solutions). While this is true for this particular SAP, absorption, in other studies has been shown strongly time dependent (Schröfl et al. 2012). The results from the teabag method appeared to quite consistent, with single operator coefficient of variation being 5.83%. Some concerns about the method may exist, specifically due to the possibility of trapping of capillary water between SAP gel particles (Kang et al. 2017; Jensen 2011), however researchers have indicated that the absorption values from the teabag method match those from other methods and the water trapped between the SAP particles is likely negligible (Farzanian et al. 2016).

4.2.3.3 Desorption Behavior of SAPs

The desorption behavior of the SAP was quantified through dynamic vapor sorption (DVS). A dynamic vapor sorption analyzer (TA Q50000; TA Instruments) was used to measure the desorption isotherms of the SAP. Changes in the mass of the specimen and relative humidity were recorded as a function of time. The relative humidity ranged from a maximum of 97.5% to a minimum of 0%.

Two different specimens were tested. These specimens consisted of SAPs that had been soaked in DI water (absorption of 210 g/g at 30 minutes) and simulated pore solution (0.25 M NaOH solution with

57

absorption of 28.5 g/g at 30 minutes), for 30 minutes. It should be noted that the desorption and desorption kinetics in extracted pore solutions and a simulated pore solution may be different due to the presence of divalent and trivalent ions (Schröfl et al. 2012). However, to understand these differences, detailed studies on extracted pore solutions need to be carried out. The simulated pore solution was chosen as it had ionic strength and SAP absorption values similar to the extracted pore solution (at 30 minutes).

From the soaked SAP sample, a specimen of approximately 50 ± 3 mg was selected and placed in a tared quartz pan, and its mass recorded. The pan was then moved into the dynamic sorption analyzer, stabilized at 97.5% relative humidity. This relative humidity was maintained until the change in mass was recorded to be < 0.001%/15 min. Once equilibrium was reached, the relative humidity was progressively reduced, using 1% relative humidity steps, until the equilibrium for the 90.5% set point was reached. At this point, relative humidity was decreased and maintained at 0% for a period of 24 hours, to ensure the total desorption of the SAP.

4.2.3.4 Internal Relative Humidity Measurements

The internal relative humidity at early ages was measured on the mortar specimens using HygroClips2-S(3)-heated sensors (with RH and temperature sensitivity of ± 0.8% RH and ± 0.1 °C at 23 °C, respectively). The sensors were calibrated using supersaturated salt solutions (Castro et al. 2016), comparing the read values with the theoretical relative humidity of the salt solutions (potassium sulfate: 97.42 ± 0.48% RH at 23 °C; potassium nitrate: 94.00 ± 0.59% RH at 23 °C; potassium chloride: 84.65 ± 0.27% RH at 23 °C; and sodium chloride: 75.36 ± 0.13% RH at 23 °C (Greenspan 1977)). A linear correction curve was used to calibrate the relative humidity sensors (Castro et al. 2016).

The mortar samples were cast and sealed inside plastic cylinder molds, 22 x 50 mm, and were stored at a constant temperature of 23 ± 0.1 °C. After 24 hours from casting, the samples were demolded from the cylinders and crushed through the use of a mortar and pestle. The crushed mortar was then sieved through a 2.38 mm sieve and approximately 5 g of the sieved material was placed inside a 12 mm deep plastic cup. The cups were then moved inside water-jacketed stainless steel cylinders, which provided control on the temperature of the specimen and prevented loss of moisture to the environment. The sensors were then mounted on top of the hermetically sealed cylinders. Every 24 hours, the specimens inside the cylinders were replaced by newly crushed material. This was done in order to minimize the loss of moisture to the environment and have more realistic measurements for sealed conditions. The temperature and the relative humidity were measured with a frequency of 1 data point every 5 minutes.

4.2.3.5 Autogenous Shrinkage Measurements

Autogenous shrinkage tests were performed on mortar specimens following ASTM C1698 ̶ 09 (ASTM Standard 2014). Two duplicates of each mixture were cast in corrugated polyethylene tubes and sealed through the use of end lids. The linear shrinkages were measured with the use of spring loaded DC LVDTs (with sensitivity ± 0.508 mm) installed on stainless steel frames. The test was conducted at a constant temperature of 23 ± 0.1 °C for a period of seven days from the time of mixing. The mass of the mortar filled tubes was recorded at the beginning and at the end of the experiment, to verify that there were no moisture losses during the test. Autogenous shrinkage

58

measurements are shown starting from final set, measured for each mixture following ASTM C403/C403M ̶ 16 [30] where the penetration resistance of a cylindrical mortar specimen of standard dimensions is periodically tested from the time of mixing (set coincides with the time specimen has a penetration resistance of 27.58 MPa for a 2.54 cm penetration depth).


Approximately 7 g of the paste material, after mixing, was poured into a glass ampoule, which was gently tamped to ensure the consolidation of the specimen. The glass ampoule was then sealed with a lid to prevent the evaporation of any moisture during the experiment. The specimens were then moved to an isothermal calorimeter (TA Instruments), which was preconditioned at 23°C. The heat flow and the cumulative heat release were measured for a period of seven days after mixing. The cumulative heat release at seven days was normalized to the cement content of the paste and compared to the 0% SAP system, in order to quantify the increase in the degree of hydration of the internally cured systems.


4.3.1 Absorption and Desorption Measurements of SAP

The absorption of the SAP in the extracted pore solution from the teabag method was reported to be, on average, 28.90 (g/g SAP) (the absorption in DI water was 210 (g/g SAP) at 30 minutes immersion). This absorption value of the SAP in the pore solution was used in Eq. 4-1 in order to quantify the amount of SAP needed by the Snyder and Bentz 1999 approach.

Figure 4-1 shows the degree of saturation (DOS, the normalized moisture content), for the desorption of SAP soaked in DI water, and simulated pore solution (0.25 M NaOH solution). At the start of the test, the SAPs are fully saturated, corresponding to a DOS of 1. For the two solutions, most of the moisture is lost during the first relative humidity step from 100% to 97.5%. This is an important requirement for internal curing applications, where internal curing agents are expected to desorb at high relative humidity (Jensen 2013). Similar to the absorption behavior, the desorption behavior is also influenced by the solution composition (Jensen 2013; Mechtcherine et al. 2015). Stronger ionic solutions reduce the maximum desorption of the SAP at each relative humidity level as shown by the higher degree of saturation for the 0.25 M NaOH solution when compared to DI water. In addition, the relative humidity equilibrium of the solution (8.24% at 23°C (Greenspan 1977) for a saturated NaOH solution) likely limits the maximum desorption of the SAP soaked in 0.25 M NaOH. The inset in Figure 4-1 shows the behavior at lower relative humidity values and magnifies the difference between the two solutions in their desorption behavior.

59

Figure 4-1. Degree of saturation of SAP as a function of relative humidity. The inset in the figure shows the behavior at low relative humidities in higher detail.

This lower degree of desorption in the simulated pore solution has been observed in literature and is likely due to the higher concentration of ions which promote crosslinking in the SAP (Schröfl et al. 2012) and influence the osmotic pressure of the SAP (Wang et al. 2015). The equilibrium relative humidity for desorbing the sample to a certain degree is then expected to be lower for the strong ionic solutions. Figure 4-2 shows the desorption (expressed as a mass change) of the SAP soaked with the two different solutions and exposed to a constant RH of 97.5% as a function of time. The speed at which the solution is released by the SAP is slower for the highly ionic solution. Therefore, it can be stated that increasing the ionic strength of the soak solution causes a reduction in the maximum desorption for a set relative humidity and causes a slower rate of desorption.

60

Figure 4-2. Desorption of SAP as a function of time.

4.3.2 Internal Relative Humidity of Mortar Systems

Internal relative humidity measurements of the mortars with various fractional volumes of the SAP are shown in Figure 4-3. As the SAP replacement increases, the relative humidity increases at a given age. At the highest SAP replacement (100%), the relative humidity is almost constant (approximately 97%) and does not drop significantly in the first seven days of measurement. These results indicate that the internal relative humidity of the system, even at early ages, is affected by the total volume of SAP (Snoeck et al. 2015), with higher replacements providing larger benefits to the system (Jensen and Hansen 2002). However, using fractional amounts of SAPs also provides benefits in terms of relative humidity. As an example, using 75% of the full replacement results in a relative humidity value which drops very slowly and is still above 92% at seven days. These results are consistent with previous studies where it was shown that higher SAP replacements helped slow down the reduction in RH at early ages (Shen et al. 2015).

This increase in relative humidity would be especially effective in the case of low w/c mixtures, where the self-desiccation process is faster, and the available volume of free water necessary to fill the pores is lower (with respect to higher w/c mixtures) (Castro et al. 2016).

61

Figure 4-3. Internal relative humidity of mortar specimens as a function of age and SAP fractional replacement.

4.3.3 Autogenous Shrinkage in Mortars

Figure 4-4 shows the autogenous shrinkage in the mortars with various fractional replacements of the SAP. These systems show an initial expansion, which is typical of internally cured concrete (Schröfl et al. 2012), and is likely caused by a combination of thermal expansion (Montanari et al. 2017) and sudden release of internal curing water from the SAP after set occurs. As with the relative humidity measurements, higher replacements provide larger beneficial effects to the system. However, even with lower replacements, a large fraction of the benefits coming from full replacement are obtained. The systems with 50% and 75% of the full replacement only show a very small autogenous shrinkage at seven days, with a reduction of 98% and 78% in autogenous shrinkage, respectively, compared to the 0% SAP mixture, when the initial expansion is included. The 25% SAP reduced the final shrinkage value by almost 55%, if the initial expansion is included, when compared to the mortar without SAP. The internal relative humidity and autogenous shrinkage follow very similar trends, suggesting a strong correlation between the two (Jensen and Hansen 2002). For both, a lower volume of internal curing water increases the rate at which the system runs out of available water. This is manifested as a more rapid reduction in relative humidity and a higher measured shrinkage for lower SAP fractional replacements.

62

Figure 4-4. Autogenous shrinkage of mortar specimens as a function of age and SAP fractional replacement.

Considerations on the desorption mechanism of SAP in cementitious systems can be made from combining the results from the desorption experiments, the relative humidity measurements, and the autogenous shrinkage measurements. For the purpose of mixture design, the absorption value of the SAP in the pore solution at 30 minutes (28.90) was used. If the SAPs initially absorb a value similar to this value, then their desorption as the ionic concentration of the pore solution increases would not be very large, considering the absorption value in a much stronger solution (the 1 M NaOH) is not that much smaller (22.85). However, a reduction in the relative humdity would cause a continuous desorption of the SAP. A continous desorption of the SAP is consistent with the relative humidity and autogenous shrinkage results, which indicate that the SAPs continue to desorb through the duration of the tests. Therefore, one may tentatitively conclude that, at least for this SAP, the desorption is more likely driven by a relative humidity reduction than by an increase in the external solution ionic strengths (Whebe and Ghahremaninezhad 2017). However, the desorption of SAP in cement pastes is more complicated than in solutions and can be influenced by other factors (Castro et al. 2010; Wehbe and Ghahremaninezhad 2017; Farzanian and Ghahremaninezhad 2017) so further experiments need to be carried out to test the validity of this hypothesis.

4.3.4 Increase in Degree of Hydration through Isothermal Calorimetry

Figure 4-5 shows the effect of increasing volume of SAP (and therefore, the internal curing water as calculated through the Snyder and Bentz method (Snyder and Bentz 1999)) on the increase in degree of hydration seven days from mixing. The increase in the degree of hydration is dependent on the volume of SAP and it plateaus for volumes greater than 75%. Others have also reported an increase in the degree of hydration of the cement paste at early ages due to the inclusion of SAP (Justs et al. 2015; Justs et al. 2014).

63

Figure 4-5. Increase in degree of hydration of internally cured paste as a function of SAP volume.

4.3.5 Internally Cured Mixtures Design Model

Snyder and Bentz 1999 proposed a mixture design methodology for internally cured concrete mixtures where the volume of internal curing water provided in the mixture is equal to the volume of chemical shrinkage. This study examines an alternative methodology based on providing a sufficient volume of water to maintain fluid filled pores of a specific size (42 nm, the pore volume that corresponds to 97.5% RH) (Montanari et al. 2017). Specifically, the amount of water to be provided the water needed to fill the volume of pores between the size of the pores between those emptied by self-desiccation and 42 nm. Toward this effort, the pore size distribution was measured for the corresponding paste at one, seven, and 28 days. The seven-day pore size distribution was used for the calculations shown here. The relative humidity was measured for the 0% mortar (i.e., without internal curing agents). This relative humidity was used to determine the size of pores emptied due to self-desiccation. While the pore radius is estimated through Kelvin Laplace equation, the pore volume in between the radii intervals is known (Montanari et al. 2017). Filling the pores up to 97.5% RH was calculated to be adequate to reduce the shrinkage by 90% (Montanari et al. 2017). This water will in principle minimize the amount of water used for the greatest amount of shrinkage reduction. The larger pores and their associated volumes are ignored in this methodology as shown in Figure 4-6.

64

Figure 4-6. Schematic representation of the new proposed design methodology based on pore size distribution calculations.

In theory, by providing a target volume of internal curing water, it is assumed that the entire volume would desorb and end up filling in an equivalent volume of vapor filled space. If this is true, it was calculated that a volume of water equal to 25% of the chemical shrinkage (Montanari et al. 2017) should be enough to reduce the shrinkage by almost 90%, which is not true, according to the results (Figure 4-4). This suggests that only a portion of the provided internal curing water is available to fill in the vapor filled pores generated by self-desiccation. This reduced volume of available internal curing water (towards pore filling) is also confirmed by the relative humidity measurements. Figure 4-7 shows the relative humidity that is measured in the mortars up to an age of seven days, as a function of the SAP volume. A secondary axis on the graph defines, for each measured relative humidity, the corresponding Kelvin radius (for the filled pores) calculated through Kelvin Laplace equation. For the calculation of the pore radius, Eq. 4-3 was used (Brunauer et al 1967):

𝑟 =2𝛾𝑉𝑚

𝑅𝑇𝑙𝑛(𝑅𝐻) Eq. 4-3

where:

γ = the surface energy of the water (0.072284 N/m),

Vm = the molar volume of water (1.8·105 mol/m3),

R = the ideal gas constant (8.3145 N·m/mol·K),

T = the temperature (K), and

RH = the relative humidity in decimal form (unitless).

65

None of the systems actually reach the target radius of 42 nm and there is a large difference between the expected filled pore radius and the actual filled pore radius, confirming that only a fraction of the additional internal curing water provided by the SAP is available as free water in the system. Nevertheless, for all the volume replacements we see benefits in terms of autogenous shrinkage reduction and increased relative humidity (Figure 4-3 and Figure 4-4).

Figure 4-7. Measured relative humidity and Kelvin pore radius for the filled pores at 7 days for the different volume replacements of SAP.

This discrepancy can be resolved by considering several additional factors which contribute in the sharing of the internal curing water, as described in Eq. 4-4:

𝑉𝐼𝐶 = 𝑉𝑝𝑣 + 𝑉𝛼 + 𝑉𝜓 + 𝑉𝑜𝑡ℎ𝑒𝑟 Eq. 4-4

where:

VIC = the total volume of chemical shrinkage water (ml/gcem),

Vpv = the volume of internal curing water filling the pores (ml/gcem),

Vα = the volume of internally cured water chemically bound, due to an increase in the degree of hydration, (ml/gcem),

Vψ = the volume of internal curing water not desorbed by the agent (ml/gcem), and

Vother = a volume which includes other unknown factors that participate in the consumption of internal curing water (ml/gcem).

66

The model was modified to include the additional internal curing water consumption as described in Eq. 4-4. An additional volume of water, based on three different assumed increases in the degree of hydration, 0%, 5% and 10%, was calculated assuming a consumption of 0.23 ml per g of cement, as described by Eq. 4-5:

𝑉𝛼 = 0.23 ∙ 𝐷𝑂𝐻𝑖 Eq. 4-5

where:

DOHi = the increased degree of hydration (-).

These values for increased degree of hydration are consistent with the isothermal calorimetry measurements showing increases in degree of hydration between 4-12%, depending on the SAP fractional replacement.

The volume of undesorbed internal curing water from the SAP at the age of seven days, was calculated using the desorption isotherm for the simulated pore solution (Figure 4-1) as shown in Eq. 4-6:

𝑉𝜓 =(𝑉𝑝𝑣 + 𝑉𝛼 )

(1 − 𝐷𝑂𝑆𝑆𝐴𝑃(𝑅𝐻)) 𝐷𝑂𝑆𝑆𝐴𝑃(𝑅𝐻) Eq. 4-6

where:

DOSSAP(RH) = the degree of saturation of the SAP at a set relative humidity (-).

The isotherm associates a degree of saturation of the SAP to each relative humidity level. Finally, the sum of these three volumes was normalized to the chemical shrinkage at seven days, measured to be 70% of the final value, in order to take into account the partial self-desiccation that occurs at seven days. Equation 4-7 shows the calculation of the fraction of chemical shrinkage water for a target shrinkage reduction:

𝑉𝐼𝐶

𝑉𝐶𝑆,𝐹𝑖𝑛𝑎𝑙=

𝑉𝑝𝑣 + 𝑉𝛼 + 𝑉𝜓

𝑉𝐶𝑆,7 𝐷𝑎𝑦 Eq. 4-7

where:

VCS,Final = the ultimate chemical shrinkage (ml/gcem), and

VCS,7 Day = the partial chemical shrinkage at seven days (ml/gcem).

The aforementioned corrections help to understand the reason behind the limited volume of available internal curing water, as suggested by the relative humidity and the autogenous shrinkage measurements. The inclusion of these factors provides a more realistic prediction of the impact that

67

increasing volumes of internal curing water have on the reduction of the shrinkage in the system (Figure 4-8), with respect to simply accounting for the volume of vapor filled pores.

Figure 4-8. Comparison between the predictions of the model and experimental measurements.

In addition to the previously discussed method, an alternative approach was investigated in order to predict the theoretical shrinkage reduction. This approach uses the experimentally measured relative humidity values at seven days, in combination with the shrinkage reduction factor, R, in order to estimate the predicted shrinkage, as shown in Eq. 4-8:

휀𝑇(𝑥%, 7𝐷) = 휀(0%, 7𝐷)(1 − 𝑅) = 휀(0%, 7𝐷) (1 − 1 +𝑃𝑉(𝑟𝑥)𝑟0

𝑃𝑉(𝑟0)𝑟𝑥)

= 휀(0%, 7𝐷)𝑃𝑉(𝑟𝑥)𝑟0

𝑃𝑉(𝑟0)𝑟𝑥

Eq. 4-8

where:

휀𝑇(𝑥%, 7𝐷) = the theoretical shrinkage for x% volume replacement with respect to the total chemical shrinkage (microstrain),

휀(0%, 7𝐷) = the strain at 7 days for the non-internally cured mixture (microstrain), and

R = the shrinkage reduction function, as shown in Eq. 4-9 (Montanari et al. 2017):

68

𝑅 = 1 −𝑃𝑉(𝑟𝑥)𝑟0

𝑃𝑉(𝑟0)𝑟𝑥 Eq. 4-9

where:

x = the subscript indicates the amount of SAP present and 0 subscript indicates the system without SAP,

PV(r) = the cumulative filled volume of the paste (ml/gcem), and

r = the Kelvin radius representing the largest filled pore (nm).

Figure 4-9 shows the autogenous shrinkage measurements of the mixtures at seven days as a function of the fractional SAP volumes, excluding the initial expansion. The figure shows experimental data along with predicted values that are obtained by using modified models with three different assumptions regarding the additional degree of hydration, 0%, 5% and 10%. In addition, the values predicted through Eq. 4-9 are shown. The plot shows an increasing reduction of the shrinkage as the volume of SAP replacement increases, which is expected. At 100% SAP replacement, there is still some shrinkage. It is clear from the figure that using fractional replacements of the SAP results in significant reduction of the autogenous shrinkage. While the predictions from the models considering increased degree of hydration show better agreement with the experimental data, work still needs to be done in order to investigate the role of additional components that contribute to the consumption of the internal curing water provided.

Figure 4-9. Autogenous shrinkage at seven days (excluding the initial expansion) for increasing SAP volume.

69

4.4 CONCLUSIONS

In this study, properties of cementitious mixtures with various amounts of a superabsorbent polymer (SAP) were studied.

The ionic concentration of the solution has a strong influence on the desorption behavior of the SAP, with higher ionic concentration causing a slower and only partial desorption of the SAP for a set relative humidity.

The SAP, by desorbing, provides the system with internal curing water, which results in higher internal relative humidities and lower autogenous shrinkage in mortars.

Fractional volumes of SAP provide considerable benefits. Using SAP that contains 25% of the chemical shrinkage volume results in 55% of autogenous shrinkage reduction in the first seven days, when compared to the 0% SAP case and including the initial expansion. This indicates that a reduction on the amount of SAP to include in the mixture is possible while still maintaining a large portion of the benefits coming from a full replacement (which substitutes the chemical shrinkage volume with an equivalent volume of internal curing water).

The isothermal calorimetry measurements showed an increase in the degree of hydration at seven days for the internally cured pastes. The increase in the degree of hydration depended on the volume of SAP included, with a plateau observed when providing 75% of the SAP volume as calculated through the Snyder and Bentz methodology.

Autogenous shrinkage data at seven days was compared with a previously developed model which aims to predict the shrinkage reduction that comes from partial internal curing. The model is based on providing a sufficient volume of water to maintain fluid filled pores of a specific size. Experimental data indicates that only a fraction of the internal curing water ends up filling the vapor filled pores: other factors participate in the consumption of the available internal curing water, such as increased degree of hydration, partial desorption of the SAP and partial chemical shrinkage at early ages. The model was modified to take into account these additional volumes and it results in a better agreement with the experimental data.

An additional method, which uses the shrinkage reduction factor ‘R’ and the relative humidity measurements at seven days can also be used to accurately predict experimental autogenous shrinkage data at seven days.

70

CHAPTER 5: ROLE OF MIXTURE CONSTITUENTS ON CURL/WARP

5.1 INTRODUCTION

Concrete pavements undergo deformation as water is removed from the pores due to chemical reaction (self-) or evaporation (drying). Several factors have been identified in the literature as having a significant effect on the magnitude of deformation including w/c (secondary effort), paste volume, extent of reaction/hydration pore refinement, moist curing duration, and internal curing (inclusion of saturated fine lightweight aggregates) (Weiss 99; Amirkhanian and Roesler 2017; Hajibabaee et al. 2016; Hajibabaee and Ley 2015).

Temperature and moisture gradients can result in curling and warping deformations. When these occur at early ages, it can be referred to as built-in curling (Wang 99, Beckemeyer et al. 2002; Hansen et al. 2006; Rao and Roesler 2005; Wells et al. 2006). Restrained curling strain gradients produce tensile stress in the pore structure, which may increase the probability of early or long-term slab cracking. Curling, due to moisture gradient, is driven primarily by external drying environment adjacent to the free surfaces. As water is lost to the environment the internal capillary pressure causes a contraction in the pat (Pickett 1956), leading to high internal material strains. These strains can be reduced by maintaining a high level of pore saturation.

Maintaining high levels of pore saturation away from the exposed surface can be achieved through the use of internal curing. Water from a dispersing saturated lightweight aggregate usually will only travel 2 to 3 mm (Henkensiefken 2008; Trtik et al. 2011) unless the pore network is still percolating, in which case the water can travel over 20 mm (Bentz et al. 2007).

One of the first uses of FLWA in concrete pavements was in Texas in the 1960s (Ledbetter et al. 1965; Ledbetter and Buth 1970). While not specifically used for internal curing purposes, the researchers found that the incorporation of FLWA into the mixture, along with coarse lightweight aggregate, was enhanced the strength, and the design thickness was calculated to be lower than the normal weight concrete (Figure 5-1). The benefits of internal curing were seen but not necessarily attributed correctly.

71

Figure 5-1. Calculated design thicknesses for normal and lightweight continuously reinforced concrete pavement (CPCR) and jointed plain concrete pavement (CPJ) sections after

(Ledbetter et al. 1965).

Another comparison site was constructed in Texas in 1964 as part of State Highway 610. The CRCP sections were short (<500 m) but the normal weight and lightweight pavements were constructed side-by-side with the same base and subgrade characteristics as well as similar traffic levels. After 18 years of traffic, a crack survey was performed (Bissett 1984) and the results indicated significant performance benefits in the lightweight concrete pavements sections in terms of cracking (Figure 5-2). The lightweight section had much lower total number of cracks and of those a significant amount were hairline cracks no larger than 3.75 mm. The presence of numerous hairline cracks likely reduced the pavement stress levels and prevented the development of larger cracks and associated deterioration. A follow up study at 24 years (Won et al. 1989) and 34 years (Sarkar 1999) found that the section made using lightweight aggregate was still performing significantly better than the normal weight section.

72

Figure 5-2. Comparison of lightweight section (left) and normal weight section (right) on state highway 610 (a). Cracking survey results from two comparison sections (b) after (Bissett 1984).

The aforementioned field test sites did not specifically examine or characterize the effects of internal curing. More recently, projects have been constructed to specifically look at the effects the additional moisture has on the pavement deformations. The Texas Department of Transportation has incorporated lightweight aggregate into several projects. A 1,000 m3 project on State Highway 121 was constructed in November of 2006. The mixtures containing lightweight aggregate of 178 kg/m3 of FLWA which was sufficient amount to satisfy the estimated autogenous shrinkage amount. This site is relatively unique in that a companion section was cast alongside without the incorporation of FLWA. A crack width study by (Friggle and Reeves 2008) showed that the widths of the cracks in the FLWA section were significantly smaller than the normal concrete companion section (Figure 5-3). A smaller crack width reduces the moisture and crack spalling potential due to incompressible material, which should increase the service life of the section.

(a) (b)

73

Figure 5-3. Crack width comparison between the FLWA and normal concrete pavement section on Texas State Highway 121. Adapted from (Friggle and Reeves 2008).

One of the largest internally cured pavement projects in the United States was the Union Pacific Railway intermodal yard near Dallas, Texas. The project used over 191,000 m3 of internally cured concrete to pave a nearly 1.5 km2 area (Babcock and Taylor 2015; Rao and Darter 2013; Villarreal and Crocker 2007). After eight years in service, the pavement has minimal cracking. Additionally, a random selection of 20 slabs were measured for curling and none of the sampled slabs had measurable amounts of curling (Figure 5-3) which indicates a high level of performance (Rao and Darter 2013).

Figure 5-4. Upward curling measurements being made across the diagonal of a randomly selected slab at the Dallas UP intermodal yard. From (Rao and Darter 2013).

74

In August of 2013, a proof of concept section consisting of three panels of JPCP were constructed on state highway 24 in Topeka, Kansas (Carter et al. 2016). While the study did not examine the effects of FLWA in concrete pavements, it tested and validated curl measurement techniques to be utilized on future projects. In May and July of 2014, the Kansas DOT constructed a comparison test section on US-54 outside of Iola, Kansas. The measured mechanical concrete properties were similar between the FLWA and normal concrete mixtures.

Accelerated pavement testing has also been performed on internally cured concrete pavements (Tia et al. 2015). The project was designed to compare FLWA and non-FLWA mixtures in terms of loading and fatigue performance using a heavy vehicle simulator (HVS). After fatigue loading of a 20,000 kg equivalent axle, the FLWA section had no visible cracking while the normal section had significant amounts of visible cracking (Figure 5-4).

Mixtures can still experience significant deformations, strains, and uplift solely because of the formation of a moisture gradient through external drying (Yang). These gradients can be strong, especially in severe drying environments with high airflows and low humidities (Almusallam 2001). Large curling deformations in concrete pavements have been linked to rideability (Johnson et al. 2010), cracking (Beckemeyer et al. 2002), and durability issues (Lange and Shin 2001). In industrial floor slabs, large curling deflection can cause joint spalling because of steel wheels from fork-lifts impacting the joints (Farny 2001; Mailvaganam et al. 2001). It is hypothesized that the addition of FLWA should delay the depth of drying at early ages through the extra reservoir of internal water and thus reduce the moisture gradient and subsequent curling deflection magnitudes. Since hydration progresses under external drying via the internally held water, the surface porosity should be better refined and reduce the rate of water diffusion to the environment.

FLWA has already been shown to provide a higher degree of hydration at the same curing time (Espinoza-Hijazin and Lopez 2011; Henkensiefken et al. 2009) for mixtures with little to no autogenous shrinkage. Thus, an added benefit is similar to higher strength and lower permeability at the surface at earlier ages, where it is critical for long-term performance.

Relatively long moist curing durations have been noted as a preferred method for ensuring adequate strength development in most concrete applications (ACI International 2016). While it is true that longer moist curing durations increase strength and improve surface properties such as hardness and permeability, the effect on the curling behavior of slab systems may be detrimental. Longer curing durations has been noted to increase slab curling deflections (Farny 2001; Suprenant 2002) likely due to pore refinement and a reduction in diffusion coefficient. For industrial floor design, it has been recommended to moist cure for a duration of three to seven days and then begin a slow drying process (Farny 2001). Since surface durability is an important factor, industrial floors need some moist curing but also a slow drying process to reduce the effect of the moisture gradient. Similarly, recent literature has shown that when the moist curing duration is significantly long, between seven and 14 days, large deformations can develop within the concrete specimen upon external drying (Amirkhanian and Roesler 2017; Hajibabaee and Ley 2015; Ley et al. 2013; Perenchio 1997).

75

5.2 EXPERIMENTAL PROCEDURE

This study was divided into two parts with one focusing on physical deformations and the other focusing on quantifying moisture gradients.

5.2.1 Beam Curling

The mortar mixture designs used for this portion of the study were based on the designs from the Section 3. Two mortar mixtures (unary and ternary) with a w/c(m) of 0.42 were investigated with and without internal curing (IC). The IC mixtures contained 30% FLWA replacement of sand by weight. The ternary mixture was a blend of cementitious materials comprising of 60% portland cement, 28% fly ash, and 12% slag, by weight of total cementitious content. The unary mixture, with 100% portland cement, was adjusted so that the two mixtures (unary and ternary) were volumetrically identical. A summary of the test specimens is show in Table . Two replicates were cast and tested for each mixture shown.

Table 5-1. Summary of test specimens used for beam curling experiments

Mix FLWA, % fine aggregate

Unary or Ternary Blend Sealed Curing, days

V-U-1 0 100% PC 1

F-U-1 30 100% PC 1

V-U-7 0 100% PC 7

F-U-7 30 100% PC 7

V-T-1 0 60% PC, 28% FA, 12% Slag 1

F-T-1 30 60% PC, 28% FA, 12% Slag 1

V-T-7 0 60% PC, 28% FA, 12% Slag 7

F-T-7 30 60% PC, 28% FA, 12% Slag 7

The mixtures were cast into formwork on top of steel channel pieces. A steel channel was chosen to obtain consistent and measurable deformations upon drying (Figure 5-5). The dimensions of the concrete specimen were 100 × 44 × 815 mm. This channel piece acts as an effective infinitely stiff base. To increase the absolute deformation and reduce the number of required sensors to one, one end of the beam was anchored to the channel piece with the other end free to curl upward.

To ensure the formation of a 1D drying front, the formwork was lined with plastic prior to casting. After the concrete had cured for at least 24 hours, the formwork was removed. The plastic was then sealed at the top edges of the beam using a polymer adhesive. In this way, the beam was completely sealed on five sides, including the bottom which would become exposed after the beam has curled upward. Additionally, the plastic portion on the bottom of the beam acted as a bond-breaker to minimize base restraint. After the prescribed sealed curing duration, the beams were placed into an environmental chamber at 25 ± 1°C and 50 ± 2% RH.

76

Measurements were taken with an LVDT (TE Connectivity DC-EC 500) placed at the end of the curling beam at 1-minute intervals over a period of 12 days (Figure 5-5). The values from the two replicate specimens were averaged and the standard deviation was estimated.

Figure 5-5. Schematic of beam curling setup.

5.2.2 Moisture Gradient by Neutron Radiography

The mortar mixtures used for the beam curling measurements were also cast and exposed to curing for measuring moisture gradient using neutron radiography (NR). A prismatic Teflon mold (with inside dimensions of 60 (width) × 50 (length) × 20 (depth) mm) was used to cast the samples. Teflon was selected since it has low neutron attenuation and it does not react with cementitious materials. After the mortar placement and consolidation, the mortar was screeded in order to make a flat and smooth surface. The samples were cured in a sealed condition at 23 ± 1 °C for 1 d or 7 d after casting in two 1 mm thick plastic bags.

Upon removal from the sealed environment, the specimens were placed into an environmental chamber which maintained conditions of 25 ± 1 °C and 50 ± 2% RH for 14 days. The environmental chamber was positioned in the beamline so that neutron imaging could be conducted without the need to remove the specimens from the drying environment. A precision linear motion stage was used to ensure all specimens were imaged in the same spot over the course of the test period. More details of the NR setup are presented in Section 6.

Neutron imaging was conducted at the Neutron Radiography Facility (NRF) at Oregon State University (OSU) Radiation Center. This facility utilizes a 1.1 MW water-cooled research reactor that uses uranium/zirconium hydride fuel elements in a circular grid array. The beam within the NRF has a collimation ratio (L/D) of 115 ± 4 and a thermal neutron flux of 9.4×105 ± 1.6×104 cm–2 s–1. The NRF at OSU uses a neutron sensitive micro-channel plate (MCP) detector with delay-line anodes as a neutron detection scheme. The detector set-up is equipped with a LiF/ZnS scintillator and charge coupled device (CCD) camera to capture images with a spatial resolution of approximately 30 µm. The field of view (FOV) is approximately 45 mm × 45 mm. Three images of each sample were taken with an exposure time of 120 seconds. A median filter was used to combine three images of a sample to eliminate the artifacts due to the gamma rays. In addition, multiple images were taken from the background (detector’s FOV) with open beam (flat-field image) and with closed beam (dark-field

LVDT

Channel

Concrete beam

77

image) for the background correction. The images from NR were processed and analyzed with ImageJ software. This process includes a background correction and calculation of intensity in the region of interest. The attenuation coefficient of water (µW = 0.1259 mm-1) was calculated by scanning the empty and water-filled stepped cells and developing a relationship between water thickness (xw) and optical density (OD = ln(I/I0)) based on the Beer–Lambert law, previously described in (Hickner et al. 2008) (Eq. 5-1).

WW

O

xI

IOD −=

= ln Eq. 5-1

where xW is the water thickness, I is the measured intensity; and I0 is the intensity of the flat field.

The degree of saturation (DoS) in the mortar is calculated from the radiographs using the Eq. 5-2 as previously described in (Lucero et al. 2015, 2017).

drysaturated

drymortarDoS

−

−= Eq. 5-2

where µmortar=OD/xsample is the time-dependent attenuation of the mortar; and µdry and µsaturated are the attenuation coefficients of the mortar in the oven-dried and saturated conditions, respectively.

The samples were oven-dried at 65 ± 1 ºC after the drying experiment to measure the µdry. The calculated µmortar from the initial radiograph of the sample before exposure to the drying condition was considered as a µsaturated. The calculated DoS based on Eq. 5-2 can be slightly different than the actual DoS in the sample since all the pores in the sample are not saturated at the end of sealed curing period. Additional testing is currently underway to quantify this difference.

5.2.3 Calculated Curling Deflections

Previous work (Rodden 2006; Rodden et al. 2007; Šelih 1995; Šelih et al. 1996; Weiss et al. 1998) has been done to develop various frameworks in which to calculate the cross-section strain in a drying cementitious material. The most extensive framework (Šelih 1995; Šelih et al. 1996) requires a significant number of parameters to initialize and in most cases, assumptions must be made of the materials and hydration process. Based on previous work (Weiss et al. 1998), the following procedure is presented to allow for the calculation of beam curling deflections based on moisture measurements obtained from NR. The neutron images provide a direct measure of the degree of saturation of the specimen over a depth. Due to the inherent noise of the measurements, a locally weighted scatterplot smoothing (LOWESS) algorithm was applied to the data using a 5% span. At the same time, the DOS can be converted into an equivalent internal relative humidity based on prior work (Bentz et al. 1998; Rodden et al. 2007) using,

0.98√1 −1−𝐷𝑂𝑆

0.75

3= 𝑅𝐻 Eq. 5-3

The shrinkage strain from the RH data can be calculated using a simple approximation,

78

휀𝑠(𝜉) = 휀𝑆𝐶 [1 − 𝑅𝐻] Eq. 5-4

where, εsc is defined as 2500µε (Weiss et al. 1998). A stress gradient incorporating creep compliance and aging can be calculated using,

휀𝑘(𝑖 ⋅ Δt) = ∑ [(1

𝐸𝜎(𝑚⋅Δ𝑡)+

𝜙(𝑖Δ𝑡,𝑚Δ𝑡)

𝐸𝑐) (𝜎𝑘(𝑚 ⋅ Δ𝑡) − 𝜎𝑘((𝑚 − 1) ⋅ Δ𝑡)) + 𝛼𝑘(𝑚Δ𝑡) ⋅ Δ𝑡]𝑖

𝑚=1 Eq. 5-5

where, 𝐸𝜎 is the time dependent modulus, 𝐸𝑐 is the reference modulus at 28 days of hydration, 𝜙(𝑖Δ𝑡, 𝑚Δ𝑡) is the creep coefficient as estimated by the CEB MC90-99 model, 𝜎𝑘 is the internally generated stress, and 𝛼𝑘 is the differential shrinkage. Since the specimen is free to shrink without restraint, the differential shrinkage term goes to zero.

By integrating over the strain gradient for each point in time, the stress present in the specimen after relaxation can be calculated. The overall axial load can be calculated by integrating over the stress gradient. The eccentricity of the equivalent axial load can be calculated by taking the second moment of the stress gradient (i.e. integrating over the axial load gradient and multiplying by the thickness of the specimen). An example of the analysis is shown in Figure.

Figure 5-6. Example data analysis starting with degree of saturation (a), converting to an equivalent relative humidity (b), calculating the strain from the relative humidity (c), applying creep relaxation

and calculating a stress (d), and finally calculating the eccentricity of the axial load (e).

The eccentricity and axial load were approximated by the end deflection (δ) in the beam assuming the simple beam theory (this does not account for aging, viscoelasticity or microcracking),

𝛿 =𝑃𝑒

𝐸𝐼

𝐿2

2 Eq. 5-6

where, P is the equivalent axial load, e is the eccentricity, E is the elastic modulus at the specified point in time, I is the moment of inertia, and L is the length of the beam.

(a) (b) (c) (d) (e)

79


The curling behavior of the mortar beams sealed cured for 1 and 7 days are shown in Figure 5-7. The plotted data is the average of two replicates with the estimated standard deviation shown at discrete points. There is a distinct difference between the Unary and Ternary cementitious blends. This is expected due to the slower hydration process associated with the ternary blend.

Comparing the FLWA and virgin mixtures, there is a clear trend of the FLWA modified mixtures exhibiting a higher degree of curling, irrespective of the cementitious composition (i.e. unary or ternary). This appears to be an unexpected outcome since a previous study (Amirkhanian and Roesler 2017) suggested there should be a reduction, as much as 50%, in curling behavior with the inclusion of saturated FLWA. However, that study examined the behavior of saturated FLWA in concrete beams and not mortar beams, as used for this project. Volumetrically, the FLWA in the Amirkhanian and Roesler (2017) study was 6% of the total specimen volume whereas the FLWA volumetric content for this study is approximately 21% of the total specimen volume. The larger volume of FLWA provides substantially more water to the system. As has been shown by numerous studies (Amirkhanian and Roesler 2017; Hajibabaee et al. 2016; Hajibabaee and Ley 2015, 2016), the presence of extended moist curing can cause significantly higher curling deformations to develop compared to conditions of drying only with minimal additional curing water. Further, the higher curling in IC mixtures can be attributed to the aggressive drying conditions exhibited for the samples as they were exposed to 50% relative humidity for an extended period of time. The evaporation from drying generates a capillary suction in the pore fluid. This pressure is a driving force for the embedded FLWA to steadily provide water during the initial hours of drying (Wyrzykowski et al. 2015). Additional testing is currently underway to better understand this observation. In practice, the evaporation rate from the slabs will be critical in the evaluation of curling.

Figure 5-7. Beam curling measurements for specimens seal cured for 1 day (a) and 7 days (b).

The measured time-dependent DoS profiles from NR are presented in Figures 5-8 to 5-11 for mortar samples with and without IC. The neutron radiography analysis indicates that there are substantial differences in the drying behavior of the mortars with and without FLWA. Examining the Unary

(a) (b)

80

mixtures with FLWA, the most noticeable difference is the higher degree of saturation that remains at the surface during drying for the FLWA mixture. The additional water provided by the FLWA keeps the surface portion at a higher DoS for a longer period of time. This can improve the surface durability as the hydration process near the surface continues until the degree of saturation is lower than approximately 85%.

Another interesting feature of the FLWA mixture is that the drying process for the first seven days creates a gradient cross section that is drier and drier. However, by the fourteenth day of drying, the internal water has redistributed in such a way that the deeper portions of the gradient are at saturation levels similar to the early stages of drying. It has been previously seen that the drying process can cause a significant migration of water from the edges of the specimen and cause this snapback effect. It is not clear the length scale that is needed to reduce this effect from the measurements. Nevertheless, the FLWA maintains a higher overall degree of saturation in the specimen under drying conditions compared to a virgin aggregate mixture.

Figure 5-8. Degree of saturation at various stages of drying for the unary mixture with virgin (a) and FLWA (b) aggregates after 1 day of sealed curing.

Figure 5-9. Degree of saturation at various stages of drying for the unary mixture with virgin (a) and FLWA (b) aggregates after 7 days of sealed curing.

(a) (b)

(a) (b)

81

Figure 5-10. Degree of saturation at various stages of drying for the ternary mixture with virgin (a) and FLWA (b) aggregates after 1 day of sealed curing.

Figure 5-11. Degree of saturation at various stages of drying for the ternary mixture with virgin (a) and FLWA (b) aggregates after 7 days of sealed curing.

Figures 5-12 to 5-15 compare the experimental and theoretical curling deflections for the various mixtures. The experimentally measured and the theoretically calculated (based on DoS profiles from NR) deflections show good agreement at early ages while, there is a discrepancy between them at later ages. As mentioned, the calculated DoS based on NR can be slightly different than the actual DoS in the sample since the calculated µmortar from the initial radiograph of the sample before exposure to the drying condition was considered as a µsaturated. This difference in DoS can cause the deviation between the experimental and theoretical curling deflections. Additional testing is currently underway to better understand this observation.

With the higher degree of saturation, it would be expected that the beams containing FLWA would curl at a lesser magnitude than a similar mixture with no FLWA. However, the sharpness or steepness of the gradient also dictates the curling behavior. While the bulk of the FLWA cross section remains at

(a) (b)

(a) (b)

82

a high DOS, the surface loses similar amounts of water compared to the virgin mixtures. However, the virgin mixtures, when confined to one surface, as is the case for the beams investigated in this study, the shrinkage strain induces a curling deformation due to the gradient in strain from the varying sizes of pores that are emptying.

Figure 5-12. Comparison of experimental and theoretical curling deflections of the unary mixtures with virgin (a) and FLWA (b) aggregates undergoing 1 day of sealed curing.

Figure 5-13. Comparison of experimental and theoretical curling deflections of the ternary mixtures with virgin (a) and FLWA (b) aggregates undergoing 1 day of sealed curing.

(a) (b)

(a) (b)

83

Figure 5-14. Comparison of experimental and theoretical curling deflections of the unary mixtures with virgin (a) and FLWA (b) aggregates undergoing 7 days of sealed curing.

Figure 5-15. Comparison of experimental and theoretical curling deflections of the ternary mixtures with virgin (a) and FLWA (b) aggregates undergoing 7 days of sealed curing.

(a) (b)

(a) (b)

84

CHAPTER 6: EXAMINING CURING EFFIECIENCY USING NEUTRON RADIOGRAPHY

6.1 INTRODUCTION

State highway agencies (SHAs) specify a series of steps to be taken after concrete is placed to ensure that the concrete is cured. Curing consists of maintaining satisfactory moisture content and temperature of concrete for a sufficient period of time during and immediately following placing so that desired properties may develop (Kosmatka et al. 2011). Curing is an important step in developing long-lasting concrete since it promotes hydration of the cement leading to strength and durability. Curing in practice largely consists of supplying additional water through water ponding or wet fabrics (e.g., burlap) or providing a curing compound that minimizes water loss from the system through a plastic membrane or a membrane forming curing compound (ASTM C309-11 2015). Due to its importance in concrete technology, much has been written about curing (ACI 308R-16 2016; Taylor 2013).

Many specifications provide a required duration during which curing should be provided. For example, AASTHO recommends three days of curing for pavements and seven days for bridge decks, which extends to ten days when supplementary cementitious materials are used (AASHTO T23 2017). Poole 2006 found that the duration of curing required was frequently based on attaining some fraction of overall strength. For example, ACI 301 recommends a minimum curing duration for the concrete to attain 70 % of the specified compressive strength which often corresponds to 7-day curing (ACI 301R-16 2016). Australian specifications (AS 3600-2009 2009) relate curing requirements with the environment to which the concrete will be exposed. Carino and Clifton 2000 identified that the required curing depends on the type of concrete being used and performed research to examine the curing needed for high performance concrete. Other researchers have examined the quality of the materials used to cure concrete. Vandenbossche 1999 reviewed curing compounds and application techniques used by Minnesota Department of Transportation. Hajibabaee et al. 2016 examined the influence of different curing compound compositions on the curling of paste samples.

Recently, the concept of internal curing (IC) has been explored. IC is provided by partially replacing the fine aggregates with an equivalent volume of pre-wetted fine lightweight aggregates (Bentz and Weiss 2011; Jensen and Kovler 2007). Others have internally cured concrete using superabsorbent polymers (SAP) that absorb water (pore fluid) quickly during mixing and release it slowly into the hydrating cement paste which then maintains high relative humidities over time (Schroefl et al. 2015). Relatively little work has been performed to understand how the duration of curing may change when internally cured mixtures are used. However, work has examined the use of IC for pervious mixtures (Kevern 2014) and IC mixtures have shown great promise in field applications (Weiss and Morian 2017; Barrett et al. 2015; Rupnow 2017). This work will examine the potential impact of using internally cured concrete with differing external curing environments.

It is important to note that while the entire cross section of concrete being placed requires curing, the surface of the concrete is the most sensitive to moisture loss, and therefore is the most sensitive

85

to curing (ACI308R-16 2016). Cather 1992 defined a zone at the surface of concrete (i.e., the skin of the concrete), called the curing-affected zone (CAZ). While estimates of the size of the CAZ may vary, the CAZ can be thought of as extending from the surface to a depth of approximately a half inch to an inch and a half (12.5 to 37.5 mm).

A variety of test methods have been used to assess curing efficiency by either measuring moisture/moisture loss or surface/near-surface physical properties. The simplest test appears to be to visually verify that the surface of the concrete remains wet (Carrier and Cady 1970). Menzel 1954 provided early work to examine the influence of evaporation rates on concrete. Jensen and Hansen 1995 proposed the use of a novel curing sensor that consisted of a capillary tube that is placed on the concrete surface. Hanson 1968 measured the relative humidity (RH) near the surface of concrete during drying. However, it should be noted that currently many of the RH sensors are relatively large (Villani 2014) and have a larger variability at high RH levels. Powers et al. 1965 studied how the duration of wet curing pastes of differing water-to-cement ratio (w/c) impacted the permeability of the paste. Surface absorption was used to measure the effectiveness of water curing (Chan et al. 1991). Others (Hilsdorf and Kropp 1995; Carino and Meeks 2000) used a variety of methods to measure permeability, water absorption, relative humidity, abrasion resistance, hardness of the surface concrete, and also a notched cylindrical test specimen to measure the tensile strength of concrete at different depths. Poole 2006 asked whether it may be a ‘better approach’ to ‘rely totally on performance testing’ citing that curing could be ceased once concrete achieved a specific strength level.

In the current study, a non-destructive technique1 is used to measure the quantity of water that has reacted with the cementitious material. This non-destructive technique is called neutron radiography (NR) (Berger 1971; Vontobel et al. 2006). NR consists of passing a neutron beam through a material, in this case a cementitious mortar. A portion of this neutron beam passes through the sample while a portion is absorbed or reflected. The natural logarithm of the ratio of the intensity of the neutron that is transmitted through the concrete (IT) and the original beam (IO) is related to an absorption coefficient for the sample (µs) and the thickness of the sample (xs) as shown in Equation 6-1 (Hussey et al. 2012).

SS

O

T xI

I−=

ln Eq. 6-1

Unlike X-ray radiography, which is based on the density of a material, NR is primarily related to the neutron atomic cross section (a term used to express the likelihood that an incident neutron will interact with the target nucleus). The neutron atomic scatter and absorption cross sections are unique to each atom, but are particularly large for the first isotope of hydrogen (H-1), which composes 99.9885% of all hydrogen found in nature (Sears 1992). For example, H-1 has a total scattering cross section of 82.02 barns as compared with other materials frequently found in concrete that are an order of magnitude lower (e.g., sodium 3.28, calcium 2.90, magnesium 4.03, oxygen 4.23,

1 The sample can be tested repeatedly over time at different conditions using neutron radiography.

86

and silicon 2.17) (Sears 1992). As a result NR is used to quantify water content in materials and has been widely used for fuel cell testing (Hickner et al. 2008).

NR has been used to study cementitious materials. Najjar et al. 1986 provided early use of NR in studies of microcracking in concrete. De Beer et al. 2004 studied the water movement during the drying process of concrete. Villani et al. 2015 investigated drying in concrete containing shrinkage reducing admixtures. Wyrzykowski et al. 2015 used neutron tomography to examine the loss of water from fresh concrete. NR has been used to investigate fluid absorption in cementitious materials, especially at sawcuts and cracks (Li et al. 2012; Li et al. 2016; Lucero et al. 2015; Lucero et al. 2017). Zhang et al. (2010, 2010) and Kanematus et al. 2009 examined fluid transport and the role of cracking on fluid transport. Jones et al. 2017 employed neutron tomography to quantify the crack filling ability of polymers in cracked concrete. Trtik et al. (2011, 2010) used neutron tomography to examine the movement of water from lightweight aggregate and SAP for IC applications. Livingston 2017 has focused on using neutron tomography to measure deleterious reactions like delayed ettringite formation. While the vast majority of the aforementioned studies have focused on the qualitative movement of ‘free water’ in concrete, Lucero et al. 2015 quantified the attenuation coefficients of both the constituent and hydrated materials. This paper describes the potential use of NR to quantify the water that has reacted with hydration products and to utilize the spatial resolution capabilities of the technique to quantify curing effectiveness.

6.1.1 Research Objective

While curing is related to wind speed, relative humidity, concrete temperature and air temperature,

this paper focuses only on the impact of a single environment on curing (23 ± 1C, 50 ± 2% RH). This paper has three objectives. First, this paper examines whether it is feasible to use NR to quantify the degree of hydration (DOH). Second, this paper examines the role of curing on the DOH, particularly in the CAZ. Third, this paper examines the role of SAP in increasing the DOH.

6.2 EXPERIMENTAL PROGRAM

This paper investigated a series of experiments to evaluate whether NR could be used to measure the DOH and the impact of different practices on curing efficiency. The DOH was measured using isothermal calorimetry on mortars, non-evaporable water based on loss on ignition (LOI) (Fagerlund 2009), and NR. The latter two measurements were performed on sealed samples that were oven dried at 105 °C at ages from 6 h to 14 days to remove the evaporable water. The evaluation of curing efficiency was measured on samples that were water ponded in a curing room or had wet burlap applied for mortar mixtures with and without SAP. The samples in each of these conditions were exposed to drying at prescribed times as shown in Table 6-1 and were oven dried at 14 days to remove the evaporable water. The following section describes the details of the sample program, conditioning, and testing.

87

Table 6-1. Curing regimes for prismatic and cubic samples

Mixture Curing age

(days)

Prismatic samples

(7.6 × 7.6 × 30.0 cm)

Cubic samples

(5.1 × 5.1 × 5.1 cm)

Water

ponded

Burlap then

drying Sealed Sealed

OPC

0.25 √***

0.5 √** √

1 √ √

3 √* √ √

7 √ √ √

14 √ √*** √

SAP

0.25 √

0.5 √

1 √ √

3 √ √ √

7 √ √ √

14 √ √ √

* Duration of wet curing

** The age that burlap was removed and sample was exposed to 50% RH until age of 14 days

*** Duration of sealed curing

6.2.1 Materials and Mixture Proportions

Two mortar mixtures were prepared, a plain mortar mixture and a mortar mixture containing SAP using the proportions given in Table 6-2.

Table 6-2. Mortar mixture proportions (in SSD conditions)

Mixture w/c

Cement (kg/m3,

m3/m3)

Water

(kg/m3, m3/m3)

Aggregate (kg/m3,

m3/m3)

IC Water

(kg/m3, m3/m3)

Dry SAP

(kg/m3)

OPC 0.42 676.6, 0.215 284.2, 0.285 1242.4, 0.500 0 0

SAP 0.42 644.2, 0.205 270.6, 0.271 1188.5, 0.480 45.1, 0.045 1.675

A Type I ordinary portland cement conforming to ASTM C 150 was used in this study. The chemical composition of the cement from the mill sheet was CaO 63.20%, SiO2 20.40%, Al2O3 4.70%, Fe2O3

88

3.30%, SO3 2.80%, MgO 2.50%, Na2Oeq 0.54% and a LOI of 1.30%. The fine aggregate is a natural river sand with a specific gravity of 2.49, and an absorption of 3.10%. A commercially available SAP used in the experiments.

The amount of SAP for IC was chosen to provide a volume of water that was equivalent to the volume of chemical shrinkage developed by the cement, as described in (Bentz and Snyder 1999; Montanari et al. 2017). The fluid absorption capacity of the SAP was determined using a pore solution extracted from a cement paste with the same w/c of the mortar mixture 30 minutes after mixing. The fresh paste was placed in a pore solution extractor with nitrogen gas providing a pressure of 200 kPa for a period of 5 minutes as described in (Montanari et al. 2017). The extractor used a cellulose membrane filter with an average pore size of 0.45 μm, in order to minimize the potential for contamination of the solution from cement particles. The extracted pore solution was sealed in a glass vial and stored at 5 °C to minimize the potential for carbonation. The absorption capacity of the SAP was measured using the ‘teabag method’ described in (Montanari et al. 2017) and an absorption of 26.9 (g/g) was determined.

6.2.2 Casting and Curing

The mixtures were prepared following a procedure similar to ASTM C 305. However, after mixing the cement with water and adding the dry aggregates, the SAP was introduced into the mixture in a dry state. An additional 120 seconds of stirring were added to the standard mixing steps, after the introduction of the dry SAP, to ensure a good dispersion of the polymers. Water in the mixture accounted for the absorption from aggregate.

Three sample geometries were prepared. Cubic samples, 5.1 × 5.1 × 5.1 cm, were prepared for determination of DOH through LOI and NR. Prismatic samples (7.6 × 7.6 × 30.0 cm) were cast to investigate the impact of different practices on curing efficiency by NR.

These samples were prepared by tamping and vibrating the fresh mortar. The sealed samples were placed under plastic during curing and then in two thick plastic bags 6 h after casting. These samples were maintained in a sealed condition at 23 ± 1 °C until their time of testing (6 h to 14 days) using LOI or NR.

The prismatic samples were demolded after 12 h from casting and then sealed with aluminum tape on all sides but the finished surface. Three different curing regimes were implemented for curing prism samples (described in Table 6-1): (1) water ponded in moist room at 23 ± 1 °C for 3 days, 7 days, and 14 days; (2) wet curing with saturated wet burlap for different periods of time and then drying at 50 ± 2 % RH and 23 ± 1 °C until age of 14 days; and (3) sealed curing for 14 days at 23 ± 1 °C. Wet-cured samples in moist room were stored in a sealed condition at 23 ± 1 °C after termination of the curing until age of 14 days in order to test all samples at the same age.

89


6.2.3.1 Using LOI to Determine the Non-Evaporable Water Content

After achieving the required age, mortar samples were crushed into smaller pieces using a hammer and were placed in an oven at 105 °C for 3 days. After drying in the oven, approximately 10 g of crushed pieces were placed in a furnace at 1000 °C for a minimum of 6 h. The cement powder and the fine aggregate were dried in an oven at 105 °C and then ignited in a furnace at 1000 °C, in order to correct non-evaporable water content based on the LOI of the cement and aggregate. The non-evaporable water content was calculated as the difference between the 1000 and 105 °C mass measurements for the crushed pieces. Then, this measurement was corrected for the LOI of the cement powder and the fine aggregate (Fagerlund 2009). The DOH of the sample was found by dividing the measured non-evaporable water content of the sample with the amount of non-evaporable water in the same sample when it is fully hydrated, which is assumed 0.23 g/g cement (Fagerlund 2009; Powers and Brownyard 1948).


The mortars were tested using isothermal calorimetry. After mixing, approximately 10 g of the mortar was poured into a glass ampoule, and then tamped to ensure good consolidation of a sample. The glass ampoule was then sealed with a lid to prevent the evaporation of any moisture during the experiment. Next, the samples were placed in an isothermal calorimeter (TA Instruments), which was preconditioned at 23 °C. The cumulative heat release was measured for the mortar samples until 7 days after mixing. For all samples, testing was performed in duplicate, and the results are presented as the average of the values from the two samples. The difference in the cumulative heat release between the internally cured sample and the non-internally cured sample is used to calculate the change in the DOH in these samples.

6.2.3.3 Neutron Radiography (NR)

After the curing period, one-inch (≈25mm) slices were cut from cubic and prismatic samples using a wet saw for the NR experiment. The samples were then dried at 105 °C to remove the evaporable water. Neutron imaging was conducted at the Neutron Imaging Facility (NRF) at Oregon State University (OSU) Radiation Center. This facility utilizes a 1.1 MW water-cooled research reactor that uses uranium/zirconium hydride fuel elements in a circular grid array. The beam within the NRF has a collimation ratio (L/D) of 115 ± 4 and a thermal neutron flux of 9.4×105 ± 1.6×104 cm–2 s–1. The NRF at OSU uses an image intensifier coupled with a light-tight right-angle adapter to a charge-coupled device (CCD) camera as a neutron detection scheme. The scintillation screen is composed of 5 μm diameter needle cesium iodide crystals doped with Gd. A Nikon D610 camera with 50 mm f/1.2 lens was used to capture images through the scintillator with a spatial resolution of approximately 90 µm. The field of view (FOV) is approximately 150 mm × 150 mm. Three images of each sample were taken with an exposure time of 2 seconds. A motorized translating stage was used to image multiple samples (Figure 6-1). A median filter was used to combine three images of a sample to eliminate the artifacts due to the gamma rays. In addition, multiple images were taken from the background (detector’s FOV) with open beam (flat-field image) and with closed beam (dark-field image) for the background correction. The images from NR were processed and analyzed with ImageJ software. This

90

process includes a background correction and calculation of intensity in the region of interest. The attenuation coefficient of water (µW = 0.1267 mm-1) was calculated by scanning the empty and water-filled stepped cells (Figure 1d) and developing a relationship between water thickness (xw) and optical density (OD = ln(I/I0)) based on Equation 6-1 (Hussey et al. 2012). In addition, the attenuation coefficients of the oven-dried fine aggregate (µA = 0.0179 mm-1) and unhydrated cement powder (µC = 0.0159 mm-1) were determined using Equation 6-1. The thicknesses of these materials were corrected by their estimated packing densities. The following section describes the details of DOH calculation from NR measurements.

Figure 6-1. Experimental set up and typical image from NR.

Equation 6-1 can be re-written to account for the different phases present in a hydrated cement paste. Figure 6-2 shows an example of the volumetric proportions of the raw materials and reacted phases plotted against DOH following the Powers and Brownyard model (Livingston 2017) for a mortar with w/c of 0.42.

( ) ( ) SCSCSCWCWGWGWGSGSCCAA

n

i

SiiSS

O

T xVVVVVVxVxI

I +++++−=−=−=

=1

ln Eq. 6-2

91

where µi is the attenuation coefficient and Vi is the volume fraction of component “i” where the components are referred to as aggregate (A), unhydrated cement (C), hydrated gel solids (GS), gel water (GW), capillary water (CW) and chemical shrinkage (CS). In this formulation, the phase volumes vary as a function of the DOH. It should be noted that this formulation does not include entrained or entrapped air, but this could be added.

(a) (b)

(c)

Figure 6-2. (a) Volumetric proportions of the raw materials and reacted phases following the Powers and Brownyard model, (b) a volumetric examination of the water bound in the hydrated products, and (c) a schematic representation of materials volume for use in Equations 1 and 3.

92

However, the change in beam intensity after interacting with a sample can be examined in a different manner as shown in Figure 6-2b. One can remove the water components that are lost on heating to 105 °C (CS, CW, GW) and consider that the volume of the aggregate and cement do not change during the test (i.e., cement and aggregate are not lost from the system). In that case one can assume that a portion of gel solid is the original cement volume (Vc-Original) (hydrated products excluding water and the water in the hydrated product) and a portion is the water that has reacted with the cement (non-evaporable water, VW-Hydration Products). In doing so, the attenuation coefficient can be determined as shown in Equation 6-3.

( ) SoductsHydrationWWOriginalCCAA

O

T xVVVI

IPr ln −− ++−=

Eq. 6-3

Equation 6-3 can be rearranged to solve for the volume of the water in the hydrated products by NR.

−−

−= −− OriginalCCAA

S

O

T

W

oductsHydrationW VVx

I

I

V

ln1

Pr Eq. 6-4

6.3 EXPERIMENTAL RESULTS AND DISCUSSION

The non-evaporable water content (as measured on sealed samples by LOI) can be used to determine the DOH as plotted in Figure 6-3a. The DOH increases over time as may be expected. The DOH of the mortars with SAP is similar to mortars without SAP at early ages. However, the mixture with SAP has a higher DOH at 14 days (7.8%). This is comparable to the increase in the DOH measured by isothermal calorimetry at 7 days which is 7.2%. The DOH from LOI is plotted against the volume fraction of water in the hydrated materials (Equation 6-4) obtained using NR for sealed samples in Figure 6-3b. A linear relationship exists between the DOH and the volume fraction of water in the hydrated products as expected (Powers and Brownyard 1946) (Figure 6-2). This water in the hydrated products was determined by calculating the water that chemically reacted with the cement (i.e., the original water content minus the capillary and gel water). Therefore, Equation 6-5 was used to determine the DOH from the calculated fraction of water in hydrated products which is determined using Equation 6-4.

oductsHydrationWVCDOH Pr1 −= Eq. 6-5

where C1 is a constant predicted using the Powers and Brownyard model (930 %/unitless). The best fit line of the experimental data corresponds to approximately 15% more water in the hydrated products (783 %/unitless) than that predicted using Powers and Brownyard model (Figure 6-3b). While the exact reason for this discrepancy is not known, it may be due to differences in modern cements from those used by Powers and Brownyard including changes in cement chemistry, the inclusion of limestone, and grinding aids. Further, this is predicated based on the attenuations of each

93

of the constituent materials and the accuracy of the non-evaporable water content that decreases with tests on mortar. Nevertheless, Equation 6-5 can be a powerful tool to convert neutron radiograph measurements to the DOH.

(a) (b)

Figure 6-3. (a) Calculated DOH over time by LOI method, and (b) DOH vs. volume fraction of water in hydrated materials by NR.

Figure 6-4 shows the gradient in the DOH as a function of depth for plain and SAP samples exposed to differing external curing environments. The DOH values were calculated using Equations 6-4 and 6-5. The calculated C1 (783 %/unitless) was used to develop Figure 6-4. Based on Figure 6-4a and b, wet curing with saturated burlap showed a significant improvement in DOH of both plain and SAP systems at the top 20 mm of the sample surface. For example, the DOH increased about 24% in 5 mm of the surface of the plain and SAP samples by extending the wet burlap duration from 1 to 3 days. The sealed plain samples had 3.2% greater hydration at the core than the samples exposed to drying at 1 day while the water cured samples had 12.5% greater hydration at the core than the samples exposed to drying at 1 day (Figure 6-4c and 6-4d). In addition, the sealed and 14 day water cured SAP samples showed 8.5% and 14% higher DOH at the core compared to the sample exposed to drying at 1 day.

An increase in DOH in samples containing SAP is presented in Figure 6-4e. Information is segmented into the cover material (CAZ) in the top 15 mm, the core material 15 to 75 mm, and in sealed mortar. The SAP samples exhibited small changes in DOH when exposed to drying at 1 day compared to the plain system. However, the SAP increased DOH more for samples exposed to drying at 3 and 7 days as well as the sealed samples. In addition, while defining the CAZ is difficult, this work found that the depth of the plain sample at which 70% DOH was obtained at 14 days was 12 mm for exposure to dry at 1 day and 7.5 and 5.5 mm for exposure to drying at 3 and 7 days, respectively. The mixture containing SAP decreased the CAZ by nearly 4 mm (1/6 in) for the 3 day and 7 day exposures to drying.

94

(a) (b)

(c) (d)

95

(e)

Figure 6-4. DOH over depth by NR for (a) plain mortar exposed to drying (b) SAP mortar exposed to drying (c) comparison between various curing practices – plain mixture (d) comparison between various curing practices – SAP mixture (symbols are at every 100th data point) and (e) increase in

DOH over depth by implementing SAP. (Note: “wet curing” represents water ponded curing in moist room and “drying” shows wet curing with saturated wet burlap for different periods of time

and then drying at 50 ± 2 % RH and 23 ± 1 °C until age of 14 days.)

6.3.1 Future Studies

While this paper has outlined that neutron radiography is an emerging technology that can be used to evaluate curing efficiency, further work is needed to develop the database of information needed to improve curing specifications. For example, the current paper only considered one environmental exposure (23 ± 1 °C, 50 ± 2% RH) and evaporation rate. Additional evaporation rates should be considered especially when the temperature is lower due to the reduced rate of hydration. Further, the current study only considered mixtures made with ordinary portland cement and as such the role of supplementary materials should be considered in future work. With respect to the use of neutron radiography, the current approach relied on relatively thin samples (25 mm thick). However, thicker samples could be evaluated for mortar (up to 60 mm) and for concrete (up to 75 or 100 mm) depending on mixture proportions. It is also recommended that in order to make sure the sample surface is parallel to the beam, additional steps need to be taken using a rotational table as described by Henkensiefken et al. 2011.

6.4 CONCLUSIONS

In conclusion, this study has shown that neutron radiography is an experimental technique that can be used to measure the water contained in hydrated cements and can be used to determine the

96

degree of hydration in a cementitious system. The results have indicated that the volume of water in the hydrated products (water in the gel solids) can be related to the non-evaporable water determined by loss on ignition. This can also be correlated with the degree of hydration as assessed using isothermal calorimetry. Neutron radiography can quantify the extent of hydration at various distances from the drying surface. It was shown in this work that obtaining the spatial information provides the ability to assess the influence of curing methods on the hydration of cement in concrete. This is particularly useful in determining the ‘curing affected zone’ (CAZ). In the mixture exposed to drying after 1 day, the top 12.5 mm (1/2 inch) of the mortar was dramatically impacted by the loss of water to evaporation. While the top 5 mm of the surface of the sample exposed to drying at 1 day had a degree of hydration that was 32% less than the 14-day moist cured sample, it should be noted that the overall degree of hydration in the core mortar was 11 % less than a water cured sample (3% less than a sealed sample). This clearly indicates transport of water to and from the core of the concrete (Villani 2014; Villani et al. 2015). The use of superabsorbent polymers demonstrated benefits in terms of both an increase in the overall degree of hydration (this can conservatively be assumed to be approximately 3.7-7.8%) and a reduction in the depth of the CAZ by 4 mm (1/6 in) when SAP is used. Further work would be necessary to develop the quantifiable correlation between the use of SAP and equivalent moist curing durations.

97

REFERENCES

AASHTO T23. Standard Method of Test for Making and Curing Concrete Test Specimens in the Field. American Association of State and Highway Transportation Officials, Washington, D.C., 2017.

ACI 301R-16. Specifications for Structural Concrete. American Concrete Institute, Michigan, 2016.

ACI 308R-16, Guide to External Curing of Concrete. American Concrete Institute, Michigan, 2016.

ACI Committees 308, 213 “(308-213)R-13 Report on internally cured concrete using prewetted absorptive lightweight aggregate,” American Concrete Institute, 2013. https://www.concrete.org/store/productdetail.aspx?ItemID=30821313&Language=English

ACI International. (2001). ACI 308R-01: Guide to Curing Concrete.

ACI International. (2016). ACI 308R-16: Guide to Curing Concrete.

Almusallam, A. A. (2001). “Effect of environmental conditions on the properties of fresh and hardened concrete.” Cement and Concrete Composites, Elsevier, 23(4), 353–361.

Amirkhanian, A. N., and Roesler, J. R. (2017). “Unrestrained curling in concrete with fine lightweight aggregates.” Journal of Materials in Civil Engineering, 29(9).

Andersson, K., B. Allard, M. Bengtsson, and B. Magnusson, “Chemical composition of cement pore solutions,” Cem. Concr. Res., vol. 19, no. 3, pp. 327–332, 1989. https://doi.org/10.1016/0008-8846(89)90022-7

AS 3600-2009. Concrete Structures. Standards Australia, Sydney, Australia, 2009.

ASTM C309-11. Standard Specification for Liquid Membrane-Forming Compounds for Curing Concrete. American Society of Testing and Materials, Pennsylvania, 2015.

ASTM, “Standard Test Method for Time of Setting of Hydraulic Cement by Vicat Needle,” C191, West Conshohocken, PA., 2014.

ASTM, “Standard test methods for chemical analysis of hydraulic cement,” C114, West Conshohocken, PA., 2015.

ASTM. (2007). “Standard Test Method for Chemical Shrinkage of Hydraulic Cement Paste.” ASTM C1608 - 12, 1–5.

ASTM. (2008). “Standard Test Method for Time of Setting of Concrete Mixtures by Penetration Resistance.” ASTM C403/C403M - 16, 1–7.

ASTM. (2010). “Standard Specification for Coal Fly Ash and Raw or Calcined Natural Pozzolan for Use in Concrete.” ASTM C618 - 15, 4(2), 3–6.

ASTM. (2011a). “Standard Specification for Portland Cement.” ASTM C150/150M - 17, 4(2), 1–7.

ASTM. (2011b). “Standard Practice for Mechanical Mixing of Hydraulic Cement Pastes and Mortars of Plastic Consistency.” ASTM C305 - 14, 1–3.

ASTM. (2013a). “Standard Specification for Slag Cement for Use in Concrete and Mortars.” ASTM C989/C989M - 16, 44(0), 1–8.

98

ASTM. (2013b). “Standard Test Method for Autogenous Strain of Cement Paste and Mortar.” ASTM C1698 - 09 (2014), i, 1–8.

ASTM. (2013c). “Standard Specification for Lightweight Aggregate for Internal Curing of Concrete.” ASTM C1761/C1761M - 15, i, 1–8.

ASTM. (2014a). “Standard practice for mechanical mixing of hydraulic cement pastes and mortars of plastic consistency.” C305, West Conshohocken, PA.

ASTM. (2014b). “Standard test method for autogenous strain of cement paste and mortar.” C1698, West Conshohocken, PA.

ASTM. (2015a). “Standard specification for lightweight aggregate for internal curing of concrete.” C1761, West Conshohocken, PA.

ASTM. (2015b). “Standard specification for coal Fly Ash and raw or calcined natural pozzolan for use in concrete.” C618, West Conshohocken, PA.

ASTM. (2016a). “Standard specification for slag cement for use in concrete and mortars.” C989/C989M, West Conshohocken, PA.

ASTM. (2016b). “Standard test method for time of setting of concrete mixtures by penetration resistance.” C403/C403M, West Conshohocken, PA.

ASTM. (2017). “Standard specification for portland cement.” C150/150M, West Conshohocken, PA.

Babcock, A., and Taylor, P. (2015). Impacts of Internal Curing on Concrete Properties. InTrans Project 14-499.

Barrett, T., Miller, A., and Weiss, W. J. (2015). Documentation of the INDOT Experience and Construction of the Bridge Decks Containing Internal Curing in 2013. FHWA/IN/JTRP-2015/10.

Beckemeyer, C., Khazanovich, L., and Thomas Yu, H. (2002). “Determining Amount of Built-in Curling in Jointed Plain Concrete Pavement: Case Study of Pennsylvania 1-80.” Transportation Research Record: Journal of the Transportation Research Board, Transportation Research Board of the National Academies, (1809), 85–92.

Bentur, A. (Ed.). (2003) “Report 25: early age cracking in cementitious systems-Report of RILEM technical committee TC 181-EAS: early age cracking shrinkage induced stresses and cracking in cementitious systems.,” RILEM Pubblications, vol. 25, 2003.

Bentz D., and J. Weiss. Internal Curing: A 2010 State-of-the-Art Review. NISTIR 7765, National Institute of Standards and Technology, Washington, D.C., 2011.

Bentz, D. P. (2005). “Curing with Admixtures Beyond drying shrinkage reduction.” Concrete International, 27(10), 55–60.

Bentz, D. P. (2007). “Internal curing of high-performance blended cement mortars.” ACI Materials Journal, 104(4), 408–414.

Bentz, D. P. and W. J. Weiss, “Internal Curing : A 2010 state of the art review,” Natl. NISTIR 7765, p. 82, 2011.

Bentz, D. P., “Influence of internal curing using lightweight aggregates on interfacial transition zone

99

percolation and chloride ingress in mortars,” Cem. Concr. Compos., vol. 31, no. 5, pp. 285–289, 2009. https://doi.org/10.1016/j.cemconcomp.2009.03.001

Bentz, D. P., and K. A. Snyder. Protected Paste Volume in Concrete: Extension to Internal Curing Using Saturated Lightweight Fine Aggregate. Cement and Concrete Research, Vol. 29, No. 11, 1999, pp. 1863–1867.

Bentz, D. P., and Stutzman, P. E. (2006). “Curing, hydration, and microstructure of cement paste.” ACI Materials Journal, 103(5), 348–356.

Bentz, D. P., and Stutzman, P. E. (2008). “Internal curing and microstructure of high-performance mortars.” ACI SP-256, Internal curing of high-performance concretes: laboratory and field experiences, 81–90.

Bentz, D. P., and Weiss, W. J. (2008). “REACT: Reducing Early-Age Cracking Today.” Concrete Plant International, 3, 56–61.

Bentz, D. P., and Weiss, W. J. (2011). “Internal Curing : A 2010 State-of-the- Art Review.” National Institute of Standards and Technology (NIST) Report 7765, 82.

Bentz, D. P., Garboczi, E. J., and Quenard, D. A. (1998). “Modelling drying shrinkage in reconstructed porous materials: application to porous Vycor glass.” Modelling and Simulation in Materials Science and Engineering, IOP Publishing, 6(3), 211.

Bentz, D. P., Koenders, E. A. B., Mönnig, S., Reinhardt, H. W., van Breugel, K., and Ye, G. (2007). “Materials science-based models in support of internal water curing.” RILEM report, 41, 29–43.

Bentz, D. P., Koenders, E. A. B., Mönnig, S., Reinhardt, H. W., van Breugel, K., and Ye, G. (2007). “Materials science-based models in support of internal water curing.” RILEM report, 41, 29–43.

Bentz, D. P., P. Lura, and J. W. Roberts, “Mixture proportioning for internal curing,” Concr. Int., vol. 27, no. 2, pp. 35–40, 2005. https://www.concrete.org/publications/internationalconcreteabstractsportal.aspx?m=details&ID=14237

Berger, H. Neutron Radiography. Annual Review of Nuclear Science, Vol. 21, No. 1, 1971, pp. 335–364.

Bissett, J. (1984). Crack Survey Comparison of Lightweight Aggregate and Conventional Gravel Aggregate Concrete Pavement. Hercules Engineering and Testing Services, Inc., Private Report.

Brunauer, S., Mikhail, R. S., and Bodor, E. (1967). “Some remarks about capillary condensation and pore structure analysis.” Journal of Colloid and Interface Science, 25(3), 353–358.

Byard, B. E., A. K. Schindler, and R. W. Barnes, “Cracking tendency of lightweight concrete,” ACI Mater. J., vol. 111, no. 2, pp. 179–188, 2010. https://www.concrete.org/publications/internationalconcreteabstractsportal.aspx?m=details&ID=51686499

Carino, N. and K. Meeks. Curing of High Performance Concrete: Report of the State-of-the-Art. National Institute of Standards and Technology, Building and Fire Research Laboratory, Gaithersburg, Maryland, 2000.

100

Carion, N., and J. Clifton. Outline of a National Plan on High Performance Concrete: Report on the NIST/ACI Workshop. Gaithersburg, MD, 2000.

Carrier, R., and P. Cady. Evaluating Effectiveness of Concrete Curing Compounds. Journal of Materials, American Society for Testing and Materials, Vol. 5, No. 2, 1970, pp. 294–302.

Carter, N., Jenkins, A., and Meggers, D. (2016). Curling and Warping of Concrete Pavement: An Investigation and Proof of Concept Study. Kansas Department of Transportation, FHWA-KS-16-04, 55pp.

Castro, J. Peled, A., Weiss, WJ, “Full and partial internal curing in mortar in terms of internal relative humidity, hydration and performance,” Proceedings of the 11th International Society of Concrete Pavements, San Antonio, 2016.

Castro, J., Lura, P., Rajabipour, F., Henkensiefken, R., and Weiss, W.J., “Internal Curing : Discussion of the Role of Pore Solution on Relative Humidity Measurements and Desorption of Lightweight Aggregate ( LWA ),” ACI Spec. Publ., vol. 270, no. 1, pp. 89–100, 2010.

Castro, J., and Weiss, W. J. (2005). “LWA absorption and desorption : the influence on the microstructure and transport properties of internally cured mortars.” (1999), 7–9.

Castro, J., Keiser, L., Golias, M., and Weiss, J. (2011). “Absorption and desorption properties of fine lightweight aggregate for application to internally cured concrete mixtures.” Cement and Concrete Composites, 33(10), 1001–1008.

Castro, J., Peled, A., and Weiss, W. J. (2016). “Full and Partial Internal Curing in Mortar in Terms of Internal Relative Humidity, Hydration and Performance.” International Society of Concrete Pavements, San Antonio, Texas.

Cather, R. How to Get Better Curing. The Journal of Concrete Society, Vol. 26, No. 5, 1992, pp. 22–25.

Chan, R., P. Dhir, Y. Hewlett. Near-Surface Characteristics of Concrete: Abrasion Resistance. Materials and Structures, Vol. 24, No. 122, 1991.

Craeye, B., M. Geirnaert, and G. D. Schutter, “Super absorbing polymers as an internal curing agent for mitigation of early-age cracking of high-performance concrete bridge decks,” Constr. Build. Mater., vol. 25, no. 1, pp. 1–13, 2011. https://doi.org/10.1016/j.conbuildmat.2010.06.063

Cusson, D., and Hoogeveen, T. (2008). “Internal curing of high-performance concrete with pre-soaked fine lightweight aggregate for prevention of autogenous shrinkage cracking.” Cement and Concrete Research, 38(6), 757–765.

De Beer, F., W. Strydom, and E. Griesel. The Drying Process of Concrete: A Neutron Radiography Study. Applied Radiation and Isotopes, Vol. 61, No. 4, 2004, pp. 617–623.

Di Bella, C., Villani, C., Phares, N., Hausheer, E., and Weiss, J. (2012). “Chloride Transport and Service Life in Internally Cured Concrete.” Structures Congress 2012, 686–698.

Espinoza-Hijazin, G., and Lopez, M. (2011). “Extending internal curing to concrete mixtures with w/c higher than 0.42.” Construction and Building Materials, 25(3), 1236–1242.

Esteves, L. P., “Superabsorbent polymers: On their interaction with water and pore fluid,” Cem. Concr. Compos., vol. 33, no. 7, pp. 717–724, 2011.

101

https://doi.org/10.1016/j.cemconcomp.2011.04.006

Fagerlund, G. Chemically bound water as measure of degree of hydration: method and potential errors. TVBM; Vol. 3150. Division of Building Materials, Lund University, 2009.

Farny, J. A. (2001). Concrete floors on ground. Portland Cement Assn.

Farzanian, K. and A. Ghahremaninezhad, “The effect of the capillary forces on the desorption of hydrogels in contact with a porous cementitious material,” Mater. Struct., vol. 50, no. 5, pp. 216, 2017. https://link.springer.com/article/10.1617/s11527-017-1068-9

Farzanian, K., Pimenta Teixeira, K., Perdigão Rocha, I., De Sa Carneiro, L., and Ghahremaninezhad, A., “The mechanical strength, degree of hydration, and electrical resistivity of cement pastes modified with superabsorbent polymers,” Constr. Build. Mater., vol. 109, pp. 156–165, 2016. https://doi.org/10.1016/j.conbuildmat.2015.12.082

Friggle, T., and Reeves, D. (2008). “Internal curing of concrete paving: laboratory and field experience.” ACI SP-256, Internal curing of high-performance concretes: laboratory and field experiences, 71–80.

Fu, T. (2011). “Autogenous Deformation and Chemical Shrinkage of High Performance Cementitious Systems.” Oregon State University.

Geiker, M., Bentz, D., and Jensen, O. (2004). “Mitigating autogenous shrinkage by internal curing.” ACI Special Publication, 218, 143–148.

Greenspan, L. (1977). “Humidity fixed points of binary saturated aqueous solutions.” Journal of Research of the National Bureau of Standards - A. Physics and Chemistry, 81A(1).

Hajibabaee, A., and Ley, M. T. (2015). “Impact of Wet and Sealed Curing on Curling in Cement Paste Beams from Drying Shrinkage.” ACI Materials Journal, 112(1).

Hajibabaee, A., and Ley, M. T. (2016). “The impact of wet curing on curling in concrete caused by drying shrinkage.” Materials and Structures, Springer Netherlands, 49, 1629–1639.

Hajibabaee, A., Khanzadeh Moradllo, M., and Ley, M. T. (2016). “Comparison of curing compounds to reduce volume change from differential drying in concrete pavement.” International Journal of Pavement Engineering, Taylor & Francis, 1–10.

Hajibabaee, A., M. Khanzadeh Moradllo, and T. Ley., "Comparison of Curing Compounds to Reduce Volume Change from Differential Drying in Concrete Pavement". International Journal of Pavement Engineering, 2016, pp. 1–10. http://dx.doi.org/10.1080/10298436.2016.1210442.

Hansen, W., Wei, Y., Smiley, D. L., Peng, Y., and Jensen, E. A. (2006). “Effects of paving conditions on built-in curling and pavement performance.” International Journal of Pavement Engineering, Taylor & Francis, 7(4), 291–296.

Hanson, J. Effects of Curing and Drying Enviroments on Splitting Tensile Strength of Concrete. Journal Proceedings, Vol. 65, No. 7, 1968, pp. 535–543.

Hasholt, M. T., O. M. Jensen, K. Kovler, and S. Zhutovsky, “Can superabsorent polymers mitigate autogenous shrinkage of internally cured concrete without compromising the strength?,” Constr. Build. Mater., vol. 31, pp. 226–230, 2012.

102

Henkensiefken, R. (2008). “Internal curing in cementitious systems made using saturated lightweight aggregate.” Masters Thesis, Purdue University.

Henkensiefken, R., Castro, J., Bentz, D., Nantung, T., and Weiss, J. (2009). “Water Absorption in Internally Cured Mortar made with Water-Filled Lightweight Aggregate.” Cement and Concrete Research, 39(10), 883–892.

Henkensiefken, R., D. P. Bentz, T. Nantung, and W. J. Weiss, “Volume change and cracking in internally cured mixtures made with saturated lightweight aggregate under sealed and unsealed conditions,” Cem. Concr. Compos., vol. 31, no. 7, pp. 427–437, 2009. https://doi.org/10.1016/j.cemconcomp.2009.04.003

Henkensiefken, R., Sant, G., Nantung, T., and Weiss, W. J. (2008). “Comments on the Shrinkage of Paste in Mortar Containing Saturated Lightweight Aggregate.” International RILEM Symposium on Concrete Modelling, Delft, The Netherlands, 707–714.

Henkensiefken, R., T. Nantung, and J. Weiss. Saturated Lightweight Aggregate for Internal Curing in Low w/c Mixtures: Monitoring Water Movement Using X-Ray Absorption. Strain, Vol. 47, No. s1, 2011, pp. e432–e441.

Hickner, M., N. Siegel, K. Chen, D. Hussey, D. Jacobson, and M. Arif. In Situ High-Resolution Neutron Radiography of Cross-Sectional Liquid Water Profiles in Proton Exchange Membrane Fuel Cells. Journal Electrochemical Society, Vol. 155, No. 4, 2008, pp. B427–B434.

Hilsdorf, H., and J. Kropp. Performance Criteria for Concrete Durability. CRC Press, 1995.

Horkay, F., I. Tasaki, and P. J. Basser, “Effect of monovalent−divalent cation exchange on the swelling of polyacrylate hydrogels in physiological salt solutions,” Biomacromolecules, vol. 2, no. 1, pp. 2195–199, 2000. https://pubs.acs.org/doi/abs/10.1021/bm0056153

House, M., Di Bella, C., Sun, H., Zima, G., Barcelo, L., and Weiss, W. J. (2014). “Influence of slag aggregate production on its potential for use in internal curing.” Transportation Research Record: Journal of the Transportation Research Board, 2441(1), 105–111.

Hussey, D., D. Spernjak, A. Weber, R. Mukundan, J. Fairweather, E. Brosha, J. Davey, J. Spendelow, D. Jacobson, and R. Borup. Accurate Measurement of the Through-Plane Water Content of Proton-Exchange Membranes Using Neutron Radiography. Journal of Applied Physics, Vol. 112, No. 10, 2012.

Jensen O., and P. Hansen. Autogenous Relative Humidity Change in Silica Fume-Modified Cement Paste. Advanced Cement Research, Vol. 7, No. 25, 1995, pp. 33–38.

Jensen, O. M. , and Hansen, P. F. ,“Water-entrained cement-based materials: II. Experimental observations,” Cem. Concr. Res., vol. 32, no. 6, pp. 973–978, 2002. https://doi.org/10.1016/S0008-8846(02)00737-8

Jensen, O. M., “Water absorption of superabsorbent polymers in a cementitious environment,” Int. RILEM Conf. Adv. Constr. Mater. Through Sci. Eng., no. September, pp. 22–35, 2011.

Jensen, O. M., “Use of Superabsorbent Polymers in Concrete concrete admixtures,” Concr. Int., pp. 48–52, 2013.

103

Jensen, O. M., and Lura, P. (2007). “Materials and methods for internal curing.” Internal Curing of Concrete, (January 2007), 45–55.

Jensen, O., and K. Kovler. Internal Curing of Concrete - State-of-the-Art Report of RILEM Technical Committee 196-ICC. rep041, 2007. ISBN: 978-2-35158-009-7.

Johnson, A. M., Smith, B. C., Johnson, W. H., and Gibson, L. W. (2010). Evaluating the Effect of Slab Curling on IRI for South Carolina Concrete Pavements. South Carolina Department of Transportation, FHWA-SC-10-04, 30pp.

Jones, S., D. Bentz, J. Davis, D. Hussey, D. Jacobson, J. Molloy, and J. Sieber. Measurement and Modeling of the Ability of Crack Fillers to Prevent Chloride Ingress into Mortar. Cement and Concrete Composities, Vol. 81, 2017, pp. 109–121.

Justs, J., Wyrzykowski, M., Bajare, D., and Lura, P., “Internal curing by superabsorbent polymers in ultra-high performance concrete,” Cem. Concr. Res., vol. 76, pp. 82–90, 2015.

Justs, J., Wyrzykwoski, M., Winnerfeld, D., and Lura, P., “Influence of Superabsorbent Polymers on Hydration of Cement Pastes with Low Water-to-Binder Ratio,” J. Therm. Anal. Calorim., vol. 115, no. 1, pp. 425–432, 2014.

Kanematsu, M., I. Maruyama, T. Noguchi, H. Iikura, and N. Tsuchiya. Quantification of Water Penetration into Concrete Through Cracks by Neutron Radiography. Nuclear Instruments Methods Phys. Res. Sect. A Accel. Spectrometers, Detect. Assoc. Equip., Vol. 605, No. 1–2, 2009, pp. 154–158.

Kang, S. H., Hong, S. G., and Moon, J., “Absorption kinetics of superabsorbent polymers (SAP) in various cement-based solutions,” Cem. Concr. Res., vol. 97, pp. 73–83, 2017.

Kawashima, S. and S. P. Shah, “Early-age autogenous and drying shrinkage behavior of cellulose fiber-reinforced cementitious materials,” Cem. Concr. Compos., vol. 33, no. 2, pp. 201–208, 2011.

Kevern, J., D. Brown. Internal Curing of Pervious Concrete using Lightweight Aggregate. Pervious Paradise, San Diego, 2014. https://www.slideshare.net/DaleFisher4/internal-curing-of-pervious-concrete-using-lightweight-aggregate. Accessed July 29, 2017.

Kosmatka, S. H., W. C. Panarese, and B. Kerkhoff. Design and Control of Concrete Mixtures. Portland Cement Association, Skokie, IL, 2011.

Kovler, K. and Jensen, OM, “Novel techniques for concrete curing,” Concr. Int., vol. 27, no. 9, pp. 39–42, 2005. https://www.concrete.org/publications/internationalconcreteabstractsportal.aspx?m=details&ID=14681

Lam, L., Wong, Y. L., and Poon, C. S. (2000). “Degree of hydration and gel/space ratio of high-volume fly ash/cement systems.” Cement and Concrete Research, 30(5), 747–756.

Lange, D., and Shin, H.-C. (2001). “Early age stresses and debonding in bonded concrete overlays.” Transportation Research Record: Journal of the Transportation Research Board, Transportation Research Board of the National Academies, (1778), 174–181.

Ledbetter, W. B., and Buth, E. (1970). “TTI Study Explores Durability Of Structural Lightweight

104

Concrete For Highway Building.” Highway Research News, (40), 31–38.

Ledbetter, W. B., Perry, E. S., Houston, J. T., and Thompson, J. N. (1965). Critical Mechanical Properties of Structural Lightweight Concrete and the Effects of These Properties on the Design of the Pavement Structure. Center for Highway Research, The University of Texas, Research Report Number 55-3F, 94pp.

Ley, T., Hajibabbee, A., Kadam, S., Frazier, R., Aboustait, M., Ebisch, T., and Riding, K. (2013). Development and implementation of a mechanistic and empirical pavement design guide (MEPDG) for rigid pavements phase I.

Li, W., M. Pour-Ghaz, J. Castro, and J. Weiss. Water Absorption and Critical Degree of Saturation Relating to Freeze-Thaw Damage in Concrete Pavement Joints. Journal of Materials in Civil Engineering, Vol. 24, 2012, pp. 299–307.

Li, W., M. Pour-Ghaz, P. Trtik, M. Wyrzykowski, B. Münch, P. Lura, P. Vontobel, E. Lehmann, and J. Weiss. Using Neutron Radiography to Assess Water Absorption in Air Entrained Mortar. Construction and Building Materials, Vol. 110, 2016, pp. 98–105.

Livingston, R. A. Neutron Tomography Measurement of Delayed Ettringite Formation in Concrete. 14th ISNDCM, 2015. http://slideplayer.com/slide/6190161. Accessed January 1, 2017.

Lothenbach, B. and F. Winnefeld, “Thermodynamic modelling of the hydration of Portland cement,” Cem. Concr. Res., vol. 36, no. 2, pp. 209–226, 2006. https://doi.org/10.1016/j.cemconres.2005.03.001

Lucero, C. L., Bentz, D. P., Hussey, D. S., Jacobson, D. L., and Weiss, W. J. (2015). “Using neutron radiography to quantify water transport and the degree of saturation in entrained air cement based mortar.” Physics Procedia, Elsevier, 69, 542–550.

Lucero, C., R. Spragg, D. Bentz, D. Hussey, D. Jacobson, and J. Weiss. Neutron Radiography Measurement of Salt Solution Absorption in Mortar. ACI Materials Journal, Vol. 114, 2017, pp. 149–159.

Lura, P., Felix, D., Loukili, A., Kovler, K., and Jensen, O. M. (2006). “Compressive Strength of Cement Pastes and Mortars with Superabsorbent Polymers.” International RILEM Conference on Volume Changes of Hardening Concrete: Testing and Mitigation.

Mackenzie, J. (1950). “The Elastic Constants of a Solid containing Spherical Holes.” Proceedings of the Physical Society, 63(1)(2).

Mailvaganam, N. P., Springfield, J., Repette, W. L., and Taylor, D. A. (2001). “Curling of concrete floor slabs on grade-causes and repairs.” Journal of performance of constructed facilities, American Society of Civil Engineers, 15(1), 11–16.

Mechtcherine, V., Secrieru, E., and Schröfl, C., “Effect of superabsorbent polymers (SAPs) on rheological properties of fresh cement-based mortars - Development of yield stress and plastic viscosity over time,” Cem. Concr. Res., vol. 67, pp. 52–65, 2015.

Menzel, M. Causes and Prevention of Crack Development in Plastic Concrete. PCA Annual Meeting, 1954, pp. 130–136.

105

Miller, A. E., Barrett, T. J., Zander, A. R., and Weiss, W. J. (2014b). “Using a Centrifuge to Determine Moisture Properties of Lightweight Fine Aggregate for Use in Internal Curing.” Advances in Civil Engineering Materials, 3(1), 142–157.

Miller, A., Spragg, R., Antico, F., Ashraf, W., Barrett, T., Behnood, A., Bu, Y., Chiu, Y., Desta, B., Farnam, Y., Jeong, H., Jones, W., Lucero, C., Luo, D., Macobatti, F., Nickel, C., Panchmatia, P., Pin, K., Qiang, S., Qiao, C., Shagerdi, H., Tian, Q., Tokpotayeva, R., Vilani, C., Wiese, A., Woodard, S., and Weiss, W. J. (2014). “Determining the moisture content of pre-wetted lightweight aggregate: assessing the variability of the paper towel and centrifuge methods.” 4th International Conference on the Durability of Concrete Structures, (July), 1–5.

Miller, A.E., T. J. Barrett, and W. J. Weiss, “Evaluation of superabsorbent polymers for use in cementitious systems for the purpose of mitigating autogenous shrinkage,” Proceedings of International RILEM Conference on the Application of Superabsorbent Polymers and Other New Admixtures in Concrete Construction, Dresden, 2014.

Mönnig, S., “Superabsorbing additions in concrete – applications, modelling and comparison of different internal water sources,” Ph.D. Thesis, University of Stuttgart, 2009.

Montanari, L, Suraneni, P., and Weiss, W.J., “Accounting for water stored in superabsorbent polymers in increasing the degree of hydration and reducing the shrinkage of internally cured cementitious mixtures,” Adv. Civ. Eng. Mater., vol. 6, no. 1, pp. 583–599, 2017. https://doi.org/10.1520/ACEM20170098.

Montanari, L., Amirkhanian, A. N., Suraneni, P., & Weiss, J. (2018). Design Methodology for Partial Volumes of Internal Curing Water Based on the Reduction of Autogenous Shrinkage. Journal of Materials in Civil Engineering, 30(7), 04018137.

Mounanga, P., Khokhar, M. I. A., El Hachem, R., and Loukili, A. (2011). “Improvement of the early-age reactivity of fly ash and blast furnace slag cementitious systems using limestone filler.” Materials and Structures, 44(2), 437–453.

Parrot, L. J. and D. C. Killoh, “Prediction of cement hydration,” Br Ceram Proc, vol. 35, pp. 41–53, 1984. https://www.scopus.com/record/display.uri?eid=2-s2.0-0021492266&origin=inward&txGid=6edecdd83257ab118224542c265dd98d

Perenchio, W. F. (1997). “The Drying Shrinkage Dilemma-Some Observations and Questions about Drying Shrinkage and its Consequence.” Concrete Construction, 42(4), 379–383.

Perenchio, W. F. (1997). “The Drying Shrinkage Dilemma-Some Observations and Questions about Drying Shrinkage and its Consequence.” Concrete Construction, 42(4), 379–383.

Pickett, G. (1956). “Effect of Aggregate on Shrinkage of Concrete and a Hypothesis Concerning Shrinkage.” ACI Journal, 52(1), 581–590.

Poole, T. S. Guide for Curing of Portland Cement Concrete Pavements. FHWA Publication FHWA-RD-02-099, Turner-Fairbank Highway Research Center, McLean, VA, 2006.

Powers, T. C. and T. L. Brownyard, “Studies of the physical properties of hardened Portland cement paste,” J. Am. Concr. Inst., vol. 43, no. 9, pp. 101–132, 1946. https://www.concrete.org/publications/internationalconcreteabstractsportal/m/details/id/15302

106

Powers, T. C., and T. L. Brownyard. Studies of the Physical Properties of Hardened Portland Cement Paste. Portland Cement Association Research Development Lab. Chicago, IL, 1948.

Powers, T. C., and T. L. Brownyard. Studies of the Physical Properties of Hardened Portland Cement Paste Part 3. Theoretical Interpretation of adsorption Data. Journal of American Concrete Institute, Vol. 18, No. 469, 1946, pp. 469–504.

Powers, T. C., L. E. Copeland, and H. M. Mann. Capillary Continuity or Discontinuity in Cement Pastes. Portland Cement Association Research Development Lab., Vol. 1, No. 2, 1959, pp. 38–48.

Rajabipour, F., G. Sant, and J. Weiss, “Interactions between shrinkage reducing admixtures (SRA) and cement paste’s pore solution,” Cem. Concr. Res., vol. 38, no. 5, pp. 606–615, 2008. https://doi.org/10.1016/j.cemconres.2007.12.005

Rao, C., and Darter, M. (2013). Evaluation of internally cured concrete for paving applications.

Rao, C., and Darter, M. (2013). Evaluation of internally cured concrete for paving applications. Applied Research Associates, Private Report.

Rao, S., and Roesler, J. R. (2005). “Characterizing effective built-in curling from concrete pavement field measurements.” Journal of Transportation Engineering, American Society of Civil Engineers, 131(4), 320–327.

RILEM Technical Committee 196-ICC. (2007). Internal Curing of Concrete - State-of-the-Art.

RILEM Technical Committee 225, “Application of superabsorbent polymers ( SAP ) in concrete construction,” RILEM, 2012. https://doi.org/10.1007/978-94-007-2733-5

RILEM Technical Committee TC 181-EAS, “Report 25: Early age cracking in cementitious systems,” RILEM, 2003. https://books.google.com/books/about/Report_25_Early_Age_Cracking_in_Cementit.html?id=-2XJogXvrvMC

Rodden, R. (2006). “ANALYTICAL MODELING OF ENVIRONMENTAL STRESSES IN CONCRETE SLABS.” University of Illinois Urbana-Champaign.

Rodden, R., Roesler, J., and Lange, D. (2007). “A Micromechanical Approach to Computing Linear and Nonlinear Moisture and Temperature Stresses in Concrete Pavements.” Proc. of International Conference on Advanced Characterization of Pavement and Soil Engineering Materials, 20–22.

Rupnow, T. Lightweight Aggregate and Internally Cured Concrete (ICC) Prove Their Value on Louisiana Bridge. Trinity LightWeight, Lafayette, LA, 2015. http://www.trinitylightweight.com/index.php/case-studies/internal-curing/item/west-congress. Accessed July 29, 2017.

Sarkar, S. (1999). “Durability of Lightweight Aggregate Pavement.” Concrete International, 21(5), 32–36.

Schlitter, J., R. Henkensiefken, J. Castro, K. Raoufi, and J. Weiss “Development of internally cured concrete for increased service life,” JTRP Technical Report FHWA/IN/JTRP-2010/10, 2010.

Schneider, C. A., Rasband, W. S., and Eliceiri, K. W. (2012). “HISTORICAL commentary NIH Image to ImageJ: 25 years of image analysis.” Nature Methods, Nature Publishing Group, 9(7), 671–675.

107

Schroefl, C., V. Mechtcherine, P. Vontobel, J. Hovind, and E. Lehmann. Sorption Kinetics of Superabsorbent Polymers (SAPs) in Fresh Portland Cement-based Pastes Visualized and Quantified by Neutron Radiography and Correlated to the Progress of Cement Hydration. Cement and Concrete Research, Vol. 75, 2015, pp. 1–13.

Schröfl, C., D. Snoeck, and V. Mechtcherine, “A review of characterisation methods for superabsorbent polymer (SAP) samples to be used in cement-based construction materials: Report of the RILEM TC 260-RSC,” Mater. Struct., vol. 50, no. 4, p. 197, 2017. https://link.springer.com/article/10.1617/s11527-017-1060-4

Schröfl, C., Mechtcherine, V., and Gorges, M., “Relation between the molecular structure and the efficiency of superabsorbent polymers (SAP) as concrete admixture to mitigate autogenous shrinkage,” Cem. Concr. Res., vol. 42, no. 6, pp. 865–873, 2012. https://doi.org/10.1016/j.cemconres.2012.03.011

Sears, V. Neutron Scattering Lengths and Cross Sections. Neutron News, Vol. 3, No. 3, 1992, pp. 26–37.

Selander, A. (2010). “Hydrophobic Impregnation of Concrete Structures – Effects on Concrete Properties.” KTH.

Šelih, J. (1995). “Movement of water during drying of fully saturated concrete.” University of New Brunswick.

Šelih, J., Sousa, A. C. M., and Bremner, T. W. (1996). “Moisture transport in initially fully saturated concrete during drying.” Transport in porous media, Springer, 24(1), 81–106.

Shah, S. P., Weiss, W. J., and Yang, W., “Shrinkage cracking - Can it be prevented?,” Concr. Int., vol. 20, no. 4, pp. 51–55, 1998. https://www.concrete.org/publications/internationalconcreteabstractsportal.aspx?m=details&ID=132

Shehata, M. H., M. D. A. Thomas, and R. F. Bleszynski, “The effects of fly ash composition on the chemistry of pore solution in hydrated cement pastes,” Cem. Concr. Res., vol. 29, no. 12, pp. 1915–1920, 1999. https://doi.org/10.1016/S0008-8846(99)00190-8

Shen, D., T. Wang, Y. Chen, M. Wang, and G. Jiang, “Effect of internal curing with super absorbent polymers on the relative humidity of early-age concrete,” Constr. Build. Mater., vol. 99, pp. 246–253, 2015.

Shin, K. J., Castro, J., Schlitter, J., Golias, M., Pour-Ghaz, M., Henkensiefken, R., Peled, A., and Weiss, J. (2010). “The role of internal curing as a method to improve durability.” Handbook on Concrete Durability, H. Kim and K. Y. Ann, eds., Middleton Publishing, 379–428.

Snoeck, D., Jensen, O.M., and De Belie, N., “The influence of superabsorbent polymers on the autogenous shrinkage properties of cement pastes with supplementary cementitious materials,” Cem. Concr. Res., vol. 74, no. August, pp. 59–67, 2015.

Snyder, K. A., and Bentz, D. P. (1999). “Protected Paste Volume in Concrete : Extension to Internal Curing Using Saturated Lightweight Fine Aggregate.” Cement and Concrete Research, 29(April 2017), 1863 – 1867.

108

STAR 225-SAP, “Application of Superabsorbent Polymers ( SAP ) in Concrete Construction,” RILEM, 2012.

Suprenant, B. A. (2002). “Why Slabs Curl: Part 2-Factors Affecting the Amount of Curling.” Concrete International, 24(4), 59–64.

Taylor, P. C. Curing Concrete. CRC Press, Boca Raton, FL, 2013.

Tia, M., Subgranon, T., Kim, K., Rodriguez, A., and Algazlan, A. (2015). Internally Cured Concrete for Pavement and Bridge Deck Applications. Florida Department of Transportation, BDV31-977-11, 175pp.

Trtik, P., B. Muench, J. Weiss, G. Herth, A. Kaestner, E. Lehmann, and P. Lura. Neutron Tomography Measurements of Water Release From Superabsorbent Polymers in Cement Paste. International Conference Material Science, Aachen-MATSCI, Vol. III, 2010, pp. 1–10.

Trtik, P., B. Münch, W. J. Weiss, A. Kaestner, I. Jerjen, L. Josic, E. Lehmann, and P. Lura. Release of Internal Curing Water from Lightweight Aggregates in Cement Paste Investigated by Neutron and X-ray Tomography. Nucl. Instruments Methods Phys. Res. Sect. A Accel. Spectrometers, Detect. Assoc. Equip., Vol. 651, No. 1, 2011, pp. 244–249.

Trtik, P., Munch, B., Weiss, W. J., Kaestner, A., Jerjen, I., Josic, L., Lehmann, E., and Lura, P. (2011). “Release of internal curing water from lightweight aggregates in cement paste investigated by neutron and X-ray tomography.” Nuclear Instruments and Methods in Physics Research A, 651, 244–249.

Tsui Chang, M., P. Suraneni, O. B. Isgor, D. Trejo, and W. J. Weiss, “Using X-ray fluorescence to assess the chemical composition and resistivity of simulated cementitious pore solutions,” Int. J. Adv. Eng. Sci. Appl. Math., vol. 9, no. 3, pp. 136–143, 2017. https://link.springer.com/article/10.1007/s12572-017-0181-x

Tsui-Chang, M., P. Suraneni, L. Montanari, and W. J. Weiss, “Determination of chemical composition and electrical resistivity of expressed cementitious pore solution using X-ray fluorescence,” ACI Mater. J., 2018, submitted.

Vandenbossche, J. A Review of the Curing Compounds and Application Techniques Used by the Minnesota Department of Transportation for Concrete Pavements. Minnesota Department of Transportation, 1999.

Villani, C. Transport Processes in Partially Saturated Concrete: Testing and Liquid Properties. Purdue University, IN, 2014.

Villani, C., C. Lucero, D. Bentz, D. Hussey, D. Jacobson, and J. Weiss. Neutron Radiography Evaluation of Drying in Mortars with and without Shrinkage Reducing Admixtures. ACI Spececial Publication, 2015, pp. 1–24.

Villani, C., Spragg, R., Pour-Ghaz, M., and Jason Weiss, W. (2014). “The influence of pore solutions properties on drying in cementitious materials.” Journal of the American Ceramic Society, 97(2), 386–393.

Villarreal, V., and Crocker, D. (2007). “Better Pavements Through Internal Hydration.” Concrete

109

International, 2, 32–36.

Vollpracht, A., B. Lothenbach, R. Snellings, and J. Haufe, “The pore solution of blended cements: A review,” Mater. Struct., vol. 49, no. 8, pp. 3341–3367, 2016. https://link.springer.com/article/10.1617/s11527-015-0724-1

Vontobel, P., E. Lehmann, R. Hassanein, and G. Frei. Neutron Tomography: Method and Applications. Physica B: Condensed Matter, Vol. 385–386 I, 2006, pp. 475–480.

Wang, F., Yang, J., Cheng, H., Wu, J., and Liang, X., “Study on mechanism of desorption behavior of saturated superabsorbent polymers in concrete,” ACI Mater. J., vol. 112, no. 3, pp. 463–469, 2015.

Wang, F., Zhou, Y., Peng, B., Liu, Z., and Hu, S., “Autogenous shrinkage of concrete with super-absorbent polymer,” ACI Mater. J., vol. 106, no. 2, pp. 123–127, 2009. https://www.concrete.org/publications/internationalconcreteabstractsportal.aspx?m=details&i=56458

Wedding, P., W. Najjar, H. Aderhold, and K. Hover. The Application of Neutron Radiography to the Study of Microcracking in Concrete. Cement Concrete Aggregates, Vol. 8, No. 2, 1986, pp. 103.

Wehbe, Y., and Ghahremaninezhad, A., “Combined effect of shrinkage reducing admixtures (SRA) and superabsorbent polymers (SAP) on the autogenous shrinkage, hydration and properties of cementitious materials,” Constr. Build. Mater., vol. 138, pp. 151–162, 2017.

Wei, Y., and Hansen, W. (2008). “Pre-soaked lightweight fine aggregates as additives for internal curing in concrete.” American Concrete Institute, 256(January), 35–44.

Weiss, B. J., Bentz, D., Schindler, A., and Lura, P. (2012a). “Internal Curing - Making Concrete More Robust During Construction.” Structure Magazine, Selection Invited and Reviewed by Editors, (January), 10–14.

Weiss, J., and D. Morian. Review of Internally Cured Concrete as It Relates to Pavements. Transportation Research Board 96th Annual Meeting, Washington DC, 2017.

Weiss, J., Schlitter, J., Bella, C. Di, and Carboneau, N. (2012b). Documenting the Construction of a Plain Concrete Bridge Deck and an Internally Cured Bridge Deck.

Weiss, W. J. (2016). “Internal Curing for Concrete Pavements.” Tech Brief, (July), 1–7.

Weiss, W. J., Yang, W., and Shah, S. P. (1998). “Shrinkage cracking of restrained concrete slabs.” Journal of Engineering Mechanics, American Society of Civil Engineers, 124(7), 765.

Wells, S. A., Phillips, B. M., and Vandenbossche, J. M. (2006). “Quantifying built-in construction gradients and early-age slab deformation caused by environmental loads in a jointed plain concrete pavement.” International Journal of Pavement Engineering, Taylor & Francis, 7(4), 275–289.

Won, M., Hankins, K., and McCullough, F. (1989). A Twenty-four Year Performance Review of Concrete Pavement Sections Made with Siliceous and Lightweight Coarse Aggregates. Center for Transportation Research at the University of Texas, Austin, 50pp.

Wyrzykowski, M., P. Trtik, B. Münch, J. Weiss, P. Vontobel, and P. Lura. Plastic Shrinkage of Mortars

110

with Shrinkage Reducing Admixture and Lightweight Aggregates Studied by Neutron Tomography. Cement and Concrete Research, Vol. 73, 2015, pp. 238–245.

Zhang, P., F. Wittmann, T. Zhao, E. Lehmann, and P. Vontobel. Neutron Radiography, a Powerful Method to Determine Time-Dependent Moisture Distributions in Concrete. Nuclear Engineering and Design, Vol. 241, No. 12, 2011, pp. 4758–4766.

Zhang, P., F. Wittmann, T. Zhao, E. Lehmann, L. Tian, P. Vontobel. Observation and Quantification of Water Penetration into Strain Hardening Cement-based Composites (SHCC) with Multiple Cracks by Means of Neutron Radiography. Nuclear Instruments Methods Phys. Res. Sect. A Accel. Spectrometers, Detect. Assoc. Equip., Vol. 620, No. 2–3, 2010, pp. 414–420.

Zhao, Y., H. Su, L. Fang, and T. Tan, “Superabsorbent hydrogels from poly(aspartic acid) with salt-, temperature- and pH-responsiveness properties,” Polymer, vol. 46, no. 14, pp. 5368–5376, 2005. https://doi.org/10.1016/j.polymer.2005.04.015

Zhu, Q., C. W. Barney, and K. A. Erk, “Effect of ionic crosslinking on the swelling and mechanical response of model superabsorbent polymer hydrogels for internally cured concrete,” Mater. Struct., vol. 48, no. 7, pp. 2261–2276, 2015. https://link.springer.com/article/10.1617/s11527-014-0308-5

Zohuriaan-Mehr, M.J., and K. Kabiri, “Superabsorbent polymer materials: A review,” Iran. Polym. J., vol. 17, no. 6, pp. 451–477, 2008. https://www.researchgate.net/publication/242582443_Superabsorbent_Polymer_Materials_A_Review

examining cost-saving measures in material selection for

Documents