Citation
Detection of Microcracks in Concrete Cured at Elevated Temperature

Material Information

Title:
Detection of Microcracks in Concrete Cured at Elevated Temperature
Creator:
SHAH, VISHAL SANJAY ( Author, Primary )
Copyright Date:
2008

Subjects

Subjects / Keywords:
Cements ( jstor )
Concretes ( jstor )
Electrons ( jstor )
High temperature ( jstor )
Hypertension ( jstor )
Image analysis ( jstor )
Image processing ( jstor )
Microcracks ( jstor )
Room temperature ( jstor )
Specimens ( jstor )

Record Information

Source Institution:
University of Florida
Holding Location:
University of Florida
Rights Management:
Copyright Vishal Sanjay Shah. Permission granted to University of Florida to digitize and display this item for non-profit research and educational purposes. Any reuse of this item in excess of fair use or other copyright exemptions requires permission of the copyright holder.
Embargo Date:
12/18/2004
Resource Identifier:
57731751 ( OCLC )

Downloads

This item has the following downloads:


Full Text












DETECTION OF MICROCRACKS IN CONCRETE CURED AT ELEVATED
TEMPERATURE















By

VISHAL SANJAY SHAH


A THESIS PRESENTED TO THE GRADUATE SCHOOL
OF THE UNIVERSITY OF FLORIDA IN PARTIAL FULFILLMENT
OF THE REQUIREMENTS FOR THE DEGREE OF
MASTER OF SCIENCE IN BUILDING CONSTRUCTION

UNIVERSITY OF FLORIDA


2004



























Copyright 2004

By

Vishal Sanjay Shah

































The thesis is dedicated to my loving parents Sanjay and Rekha and to my best friends
Reshma, Akshay, Piyush and Anushree for their unconditional support and caring
throughout my academic endeavors. It is with the support and love of my family and
friends that I am able to reach to my goals.















ACKNOWLEDGMENTS

The author would like to thank all of the members of his supervisory committee for

their help and ideas throughout this effort especially Dr. Abdol Chini, committee chair,

for his advice, patience and dedication as my advisor and also for the financial support

provided by him.

The author would also like to express gratitude to Dr. Larry Muszynski and Dr.

Andrew Boyd for their support and contribution as members of my advisory committee.

Acknowledgements are also owed to Tanya Reidhammer for her help in explaining and

conducting scanning electron microscopy. In addition, I would like to thank Richard

DeLorenzo for his help in the experimental process.

My thanks also go to everyone within M.E.Rinker Sr. School of Building

Construction for their continued help during the last two years. The author would like to

thank his friends at the University of Florida for their support and help they offered

during the research and writing of this thesis.
















TABLE OF CONTENTS

page

A CK N O W LED G M EN T S ....................................................................... iv

LIST OF TABLES .................. ........................................... ...... .. ........vii

LIST OF FIGURES .................. .......................................... ...... ........ viii

ABSTRAC T ............. ............ ................. ........ ........x

CHAPTER

1 IN TRODU CTION .............. ............................... ............ ..............1

Background......................................................... .................. ...............1
Curing of Concrete .............. ...... ................................2
H eat G generation in M ass Concrete..........................................................................3
Objective and Scope ................................................ .... ....5

2 LITERATURE REVIEW .............................................................7

Overview of Sam ple Preparation Techniques ........................................ .............. 8
Radiography and Acoustic Techniques................... .......................9
Replica Technique ........... .......... .... .......................... .................. 10
Impregnation Techniques .............................. ....................... ...10
Dye Impregnation Methods.............................. ..........................11
Epoxy Impregnation Techniques ............................................11
Thin Samples................................. ........ ........ 13
Impregnation by Wood's Metal .................................. .. ......14
High Pressure Epoxy Impregnation Technique ........................................ 15
Comparison of Sample Preparation Techniques............................ ...............16
M icroscopic Instrum ents ................................................ ............... 18
O ptical M icroscope ............................................... .............. 18
Scanning Electron M icroscope............. ................ .................... .......... 19
Variable Pressure Scanning Electron M icroscope .....................................21
Image Processing Techniques..............................................23
Summary ............................... ..................24






v










3. RESEARCH M ETHODOLOGY ......................................................... ... ............ 25

Introduction .................. .... ................. ........ .25
Sample Selection .............................................................25
Specim en Preparation M ethod.................................. ....................... ..............27
Procedure for Sample Preparation......................................... ..........28
Preparation for Image Processing............................ ......... ......... 32
Image Analysis .......................... ...........................33
Sum m ary .................. ........... ....................................... ........ 34

4. IMAGE PROCESSING AND ANALYSIS ..................................... .....35

Introduction .................. ............................................................... 35
Im age Processing T technique .................... ..........................................36
Sample 95BOOP Room Temperature Sample 1 ..........................................37
Sample 95BOOP Room Temperature Sample 2 .....................................41
Sample 95BOOP High Temperature Sample 1 ...................................... 45
Sample 95BOOP High Temperature Sample 2..................................... 49
Com parative A analysis of Sam ples................................................................... 53
Sum m ary ..................................................................................... 54

5 SUMMARY, RESULTS AND RECOMMENDATIONS ...........................................55

Sum m ary ................................................................................. 55
Results and Conclusion................ .... ............... 57
Recommendations.......................... .......... 57

LIST OF REFERENCES ..................................... ............... .....................59

BIOGRAPHICAL SKETCH ............................................... 62
















LIST OF TABLES


Table page

3-1 Mix Design Data................. ............................. 26

3-2 Adiabatic Temperature Rise Data.................................... .................. 27

3-3 Mixing Schedule for Epoxy........................................ 30

4-1 Crack Quantification Analysis for Mix 95 B 00 P RT-1 ................... ..............39

4-2 Comparison of Odd and Even Number Images.......... ............................40

4-3 Crack Quantification Analysis for Mix 95 B 00 P RT-2................................43

4-4 Comparison of Odd and Even Number Images................... ...............44

4-5 Crack Quantification Analysis for Mix 95 B 00 P HT-1.................................47

4-6 Comparison of Odd and Even Number Images.................................48

4-7 Crack Quantification Analysis for Mix 95 B 00 P HT-2................................51

4-8 Comparison of Odd and Even Number Images............................52

4-9 Comparative Analysis of 95 B OOP RT1 and 95BOOP RT2 .....................................53

4-10 Comparative Analysis of 95 B OOP HT1 and 95BOOP HT2.................................53

5-1 Comparative Analysis of Samples....................... ... .....................57
















LIST OF FIGURES


Figure page

2-1 Working of Scanning Electron Microscope .................................... .....19

3-1 D iam ond W after Saw ..............................................................29

3-2 U ltra Low Epoxy K it by SPI .............................................. ............... 30

3-3 Metaserv Grinder and Polisher............... ....... ......... ..........................31

3-4 Polished Sam ple ................................................. ....... .31

3-5 D iam ond Suspension Paste.................................................... 32

3-6 Prepared Samples ....................................... .... ... .... ...............32

3-7 Specim en w ith Copper Tape...................................................... 33

3-8 Specim ens with 4x4 m atrixes .......................................................... .............33

4-1 Representative Images of Mix 95 B 00 P RT-1 before Image Processing ..................37

4-2 Representative Images of Mix 95 B 00 P RT-1 after Image Processing....................38

4-3 Variations in Crack Density of 95 B 00 P RT1 ................. ................. ...........40

4-4 Representative Images of Mix 95 B 00 P RT-2 before Image Processing .................41

4-5 Representative Images of Mix 95 B 00 P RT-2 after Image Processing.....................42

4-6 Variations in Crack Density of Sample 95 B 00 P RT2.......................................44

4-7 Representative Images of Mix 95 B 00 P HT-1 before Image Processing................45

4-8 Representative Images of Mix 95 B 00 P HT-1 after Image Processing.....................46

4-9 Variations in Crack Density of Sample 95 B OOP HT- 1.......................................48

4-10 Representative Images of Mix 95 B 00 P HT-2 before Image Processing................49

4-11 Representative Images of Mix 95 B 00 P HT-2 after Image Processing................ 50









4-12 Variations in Crack Density of Sample HT 2......................................52
















Abstract of Thesis Presented to the Graduate School
of the University of Florida in Partial Fulfillment of the
Requirements for Master of Science in Building Construction

DETECTION OF MICROCRACKS IN CONCRETE CURED AT ELEVATED
TEMPERATURE

By

Vishal Sanjay Shah

December 2004

Chair: Abdol Chini
Cochair: Larry Muszynski
Major Department: Building Construction

Proper curing is a necessity for durable concrete. Exothermic reaction during curing

process generates heat in concrete creating thermal differential between outer and inner

core of concrete. In mass concrete pores at Florida Department of Transportation,

internal temperatures of concrete have been reported as much as 1800F to 2000F, which

leads to a concern that higher curing temperature might initiate higher number of

microcracks during its early age.

Phase I of this study indicated the presence of higher curing temperature in

concrete which is made from cement having higher heat of hydration, higher placing

temperature and does not contain any pozzlonic material in mix design. For comparison

of curing conditions, four samples were selected from this type of concrete for

quantification of microcracks. Out of these four samples, two of the samples were cured

at room temperature and the other two samples were cured at high temperature.









A literature review revealed that various specimen preparation methods have been

used in past for viewing concrete microstructure. However, a counter diffusion method

with epoxy impregnation was used in this study as this method does not require any pre

heating of specimen and pressure or vacuum in sample preparation. After specimen

preparation, samples were viewed under variable pressure scanning electron microscope,

and a total sixteen images from each sample were taken.

Image analysis was done with standard image processing software by converting

images into binary mode by thresholding it until features of interest (microcracks and

voids) were correctly highlighted. After differentiating voids from cracks by shape

analysis method, microcracks were manually mapped and quantified by calculating crack

density in each image. A statistical analysis was performed to indicate representative

nature of image quantity taken in sample.

The results showed no significant difference between the number of microcracks

present in samples cured at room temperature and high temperature in concrete made by

cement having heat of hydration up to 76.2 cal/g, placing temperature of 950F and

without any pozzlonic material. Thus it could be suggested that high temperature curing

upto 1800F does not play any significant role in the development of microcracks.















CHAPTER 1
INTRODUCTION

Background

Concrete is one of the oldest engineering material and certainly one of the key

integrants in development of modern infrastructure. When properly designed and

carefully produced with good quality control, concrete is inherently a durable material.

However, under adverse environment and exposure conditions, concrete interacts with its

surrounding environment and undergoes chemical changes. These processes involve

movement of water or other gaseous substances transporting aggressive agents through

pore structures of concrete, thus making concrete potentially vulnerable to deleterious

attacks.

Reinforced cement concrete structures exposed to hostile environment require highly

durable concrete to provide long lasting performance with minimal maintenance. Modern

day infrastructure has necessitated construction of numerous bridges, dams, tunnels and

concrete pavements. Massive concrete structures of this nature would contain large

amount of concrete, most of these structures could be termed as "mass concrete" work.

According to American Concrete Institute Committee 116, report SP-19 (1985), mass

concrete is "any volume of concrete with dimensions large enough to require that

measures be taken to cope with generation of heat of hydration of the cement and

attendant volume change to minimize cracking." The Florida Department of

Transportation (2004) further specified dimensions of mass concrete as "when the









minimum dimension of the concrete exceeds 36-inches and the ratio of volume of

concrete to the surface area is greater than 12-inches, provide for mass concrete."

Curing of Concrete

Curing is widely perceived as being an important factor in achieving durable

concrete structure. Curing ensures that concrete maintains a sufficient amount of water to

reach a proper degree of hydration in the surface which is required to continue chemical

reaction between cement and its constituents. Proper Curing allows hydration of cement

to continue which is expected to reduce capillary porosity, thereby strengthening the

concrete to desired strength and increasing its resistance to penetration by aggressive

gases and liquids. The higher the curing temperature, the faster is the reaction between

cement and water, thus shorter setting time. It can reduce durability and useful life span

of structure by hardening concrete faster, reducing strength and causing excessive

seepage. Curing measures are necessary to maintain a satisfactory moisture and

temperature condition in concrete, because internal temperature and moisture directly

influence early and ultimate concrete properties.

During curing process, concrete element undergoes exothermic chemical reaction.

This heat generation is of little consequences and can be easily dissipated in surrounding

environments. However, in mass concrete, heat generated within the body of member

does not get dissipated easily. This temperature rise causes expansion while the concrete

is hardening. Temperature in inner core of concrete is much greater than outer core of

concrete due to loss of heat to surroundings by exposed surfaces. If the temperature rise is

significantly high and concrete undergoes non-uniform or rapid cooling, stresses due to

thermal contraction in conjunction with structural restraint can result in cracking before

or after the concrete eventually cools to surrounding temperature. Due to faster cooling of









outer surface, concrete tends to shrink and causes tensile stress in outer core. When these

tensile stresses are greater than tensile strength of concrete, cracks are developed.

Heat Generation in Mass Concrete

Heat generated in mass concrete increases temperature of concrete to a varying

degree depending on surrounding temperature, size of structural element, mix design

configurations, and curing temperature. Current steps to counter high temperature rise in

concrete includes choosing a low heat cement, external curing system, and cooling of

aggregates prior to mixing. Constructing large structures in small stages, or lifts, can be

done, but it is slow and the finished structure is weakened at the cold joints in the

concrete. Replacement of cement by pozzlonic materials results in a decreased

temperature rise in fresh concrete, which is a particularly important issue in mass

concrete where a large temperature rise can lead to cracking.

Currently, any mass concrete pour for FDOT requires thermal control plan which

describes the measures and procedures intended to maintain temperature differential of

350F (200C) or less between interior and exterior portion of designated mass concrete

element during its curing phase; there is however, no requirement on the maximum

temperature rise within the mass concrete. Internal temperatures of 1800F to 2000F have

been reported in mass concrete pours in FDOT projects. There is a concern that such

higher curing temperature would initiate microcracks during its early age which would

accelerate deterioration and compromise durability of concrete by microcracks acting as a

conduit of harmful minerals and gases, causing reduction in durability of concrete.

Study conducted by Maggenti (2001) of California Department of Transportation

indicated no irregular, externally visible cracks in mass concrete work. However, it

would be important to find out changes occurring at the micro level, which will obviously









grow in future and hamper durability of concrete. The capillary effect of the

microcracks increases probability of exchange of substances between the environments

and microstructure of concrete. This exchange of substances may lead to transport,

solution and enrichment processes in concrete microstructure which can damage the

concrete by increasing the volume.

Concrete, when fully hardened, contains a variety of micro structural features on a

variety of scales. To the unaided eye, it appears from a polished section to be essentially a

two phase composite-a matrix of hardened mortar plus coarse aggregate that occupies

about 75 % of its volume. Recent studies of concrete microstructure have played

important part in understanding behavior of concrete matrix. It is necessary to study the

effect of elevated temperature rise in microstructure specially comparing it with normal

room temperature cured samples and find out if any additional microcracks have been

developed. Formation of microcracks in the surface can lead to decrease of durability,

especially in exposed outdoor structures as bridge superstructures but also concrete

pavement and structures in direct contact with the ground.

Many methods ranging from destructive to non-destructive have been used in past

to understand microstructure of concrete. A problem common with many of the

techniques used to examine the microstructure of concrete is that they can alter the

sample. For example, many techniques require that the specimen be prepared by

mechanical polishing or to dry concrete specimen before examination by putting in

vacuum or heat. Such Specimen preparation methods would invariably alter the original

artifacts and induce additional microcracks.









Advancement in science has assisted in developing equipment and techniques

capable of studying the microstructure which facilitate in-depth study and analysis of

specimen. While many existing techniques and equipment require drying the specimen

and in a way altering the specimen, evolution of Variable Pressure Scanning Electron

Microscope (VPSEM) techniques allows the study of unaltered wet specimen in a low

vacuum.

Objective and Scope

This study is an extension of a research project "Adiabatic Temperature Rise in

Mass Concrete" which is aimed to develop adiabatic temperature rise curves for concrete

mixes used in FDOT projects and investigate properties of high temperature cured

concrete. Total of 20 mixes of concrete were made as a part of the study by changing

various parameters such as the source of cement, the percentage of pozzlonic materials

and the initial placing temperature. Concrete samples were kept in heating curing

chambers to simulate conditions of mass concrete for monitoring rise in temperature. At

14 and 28 days, samples kept in heating chambers as well as samples in normal room

temperature were tested for compressive strength.

The objective of this part of the study is to focus on detecting and quantifying

additional microcracks developed due to concrete cured at elevated temperature. This

would be done by comparing room temperature cured samples with samples cured at

elevated temperature to find out if there is any difference in quantity of microcracks.

The Study would be performed as given below:

1. A literature search to find out current methods used in identification of microcracks
and sample preparation techniques. It should also include possible image
processing techniques to enhance analysis capability.






6


2. Prepare concrete specimens according to an identified sample preparation
technique and view them with selected Microscopy equipment.

3. Use Image Processing Techniques for further identification and quantification of
microcracks. Compare the quantity of microcracks in samples cured at elevated
temperature with those cured at room temperature to determine if any additional
microcracks have been developed due to high temperature curing.














CHAPTER 2
LITERATURE REVIEW

Presence of voids, cracks, and other defects play an important role in determining

the mechanical performance of the concrete. If concrete surface is pre-damaged by

microcrack formation in the near-surface area and exposed to open weathering, capillary

effect would increase the transfer of substances from the environment and increase the

concrete volume which can lead to several mechanisms capable of negatively influencing

durability of concrete.

Studies have indicated that the development of cracks and connected crack network

contributes to the increase in permeability and diffusivity of concrete. The presence of

preexisting microcracks of 10 [im or greater on the surface can lead to reduced durability,

especially of exposed outdoor structures such as bridge superstructures, but also concrete

pavements and structures in direct contact with the ground. Patel et al., (1995) found that

concretes cured at higher temperatures exhibited a coarser microstructure than that of a

typical concrete cured at 200C, particularly with respect to ettringite. The presence of

microcracks will increase moisture mobility within the concrete, which may produce

density gradients within the matrix leading to further microcracking. The domination of

the matrix by a network of microcracks is also conducive to the formation of secondary

ettringite.

Recorded curing temperature of 1800F 2000F inside core of FDOT Mass concrete

elements has raised concerns over the initiation of microcracks. Slate and Hover (1984)

defined microcracks as cracks having width of less than 100 [im, while Jansen defined









microcracks as extended faults with a width of less than 10 rim. According to Kjellsen

and Jennings (1996), these differences in definition appeared to be closely related to

experimental techniques and to the orientation of the studies. For the scope of this study,

reasonable and practical interpretations of microcracks were set as cracks having widths

less then 10 jim. While macrostructure of concrete can be seen unaided, microstructure

(200 rm or smaller) must be observed with the aid of a microscope. It would be fair to

say, as the documented research indicates, that cracks and crack propagation is

understood well from macroscopic point of view but unavailability of precise specimen

preparation techniques hinder studies of concrete microstructure.

Overview of Sample Preparation Techniques

Sample preparation is a key to microscopic analysis of concrete. Proper preparation

of concrete samples so that microcracks and voids develop a distinct contrast against the

body of concrete is a pre-requisite for the application of the modern day image processing

and analyzing techniques (Soroushian et al., 2003). Poor preparation methods can lead to

erroneous diagnoses of problems associated with concrete specimen. According to

Hornain et al., (1996), ideal specimen preparation techniques should not induce any

cracks during the preparation of samples, thus techniques which involves prior drying of

specimen should not be used. The ideal specimen preparation should be simple,

economic, rapid, and should be able to detect very fine cracks. And finally, a sample

preparation technique will not be useful without accurate image analysis and processing

techniques. The sample preparation would also depend on the type of equipment used for

image collection, required resolution of image, and objective of the study.









For the past few decades, studies have been performed to find out a method which

fulfills all definitions for ideal specimen preparation technique. New equipment and

methods developed by technological advances have helped researchers in their analysis

of concrete specimens. As reported by Ringot and Bascoul (2001), two methods are

necessary for the characterization of microcracks, one for sample preparation and one for

quantification of cracks. Sample preparation techniques also depend on the type of

microscopy used (S. Marusin 1995). Acoustic method, ultra sonic and laser sparkle

methods are useful for studies related to crack propagations but they are unequipped to

monitor initial state of samples crack quantification. Thus, the rest of the sample

preparation techniques can be briefly classified in three main categories: radiography

techniques, replica techniques, and impregnation techniques.

Radiography and Acoustic Techniques

Earlier Methods by Slate and Olsefski (1963) described the use of X-radiography to

study the internal features of concrete and crack formation process of mortar and concrete

sample. Thin samples of 4 mm thickness were exposed to X-rays flux normal to the plane

of the sample. Cracks were identified from normal constituent of concrete as penetration

of X-rays was greater in surrounding area. A comparative study carried out by Najjar et

al., (1986) indicated that due to poor resolution of images, X-radiography systematically

overlooks thin cracks. Neutron radiography described by Samaha and Hover, (1992),

which is principally quite similar to X-ray radiography approach, is in fact a better

alternative for increasing image resolution. However, in this process the samples need to

be air dried beforehand in order to impregnate specimen with gadolinium nitrate, which

in turn might induce microcracks in samples.









Replica Technique

Replica method developed by Ollivier in 1985 is one of the methods which do not

require any pre drying of the sample. In this method, the film of acetylcellulose is placed

on concrete specimen by methyl acetate. The film is taken off after the solvent (methyl

acetate) is evaporated in air and observed under optical or scanning electron microscope.

As there is no disturbance made to the specimen surface itself, crack propagation studies

can be carried out with replica method. However, in order to peel the acetylcellulose film

safely, the area of film has to be limited to 2 cm2. Thus too many replicas need to be

prepared for covering significant portion of specimen, which could be time consuming.

Impregnation Techniques

Impregnation techniques are one of the oldest techniques used to study

microstructure of concrete. In this method, concrete samples are impregnated with dye or

epoxy to facilitate detection and identification of cracks. In order to reduce difficulty in

viewing cracks in dense microstructure, fluorescent dye was used by Knab et al., (1984).

However, pre drying of specimen in an electric oven before impregnation, which could

alter the specimen condition, remained a major inhibition to this process. Methods

developed by Struble et al., (1989) and Gran (1995) overcome drying of specimen by

using counter diffusion method for replacement of pore water by dye impregnated

organic solution. In this method, a thin sample of 15 mm was cut and kept in solution

made by dissolved dye and ethanol. After 4 days of counter diffusion process between

pore water present in concrete specimen and dye induced ethanol, specimen was taken

out of the solution to remove excess dye present on the surface by polishing. However,

this process could take several days and increases the duration of the test.









Dye Impregnation Methods

Hornain et al., (1996) modified dye impregnation techniques to reduce the time

taken to prepare specimen. In this Method, 4 grams of a water soluble red powder,

commercially known as Irgacete was dissolved in 100 ml ethanol solution. After keeping

specimens for 5 minutes in dyed solution, a second impregnation was done. After

completing these 2 steps, the specimen was taken out and polished under water with 6

lm diamond paste to remove excess dye. In order to achieve proper polishing, the

specimen was again polished with 3[im and 1 [im of diamond paste. The polished

specimens were observed under optical microscope at 100X. The study reported that

cracks of 1 .im width and less could be easily distinguishable, providing contrast level

between specimen and cracks. During sample microscopic studies, cracks going through

the hydrated cement phase were distinguished from discontinuity at paste-aggregate

interface. However, the author reported that dye impregnation method was not useful in

finding out cracks in highly porous areas.

Epoxy Impregnation Techniques

Struble and Stutzman (1989) developed a new technique which involved three steps

procedure to replace pore water from concrete sample. Epoxy impregnation would not

only support the microstructure of specimen by filling the voids and cracks up on curing,

but also support fragile pores and matrix phases by restraining it against disintegration

during the different stages of preparation. This could be a major issue if the samples are

to be viewed under scanning electron microscope which would generate high pressure

inside the chamber. Another advantage of using epoxy is to enhance contrast between the

pores, hydration products and cementitious material. The selection of the type of epoxy

depends on the objective of the study. Low viscous epoxy was used for relatively highly









permeable materials or cementitious powders, while ultra low viscosity epoxy aids in

rapid infiltration for less permeable cement pastes and concrete samples.

The first step for this sample preparation method was to cut the specimen of 1 inch

size and immerse specimen in ethanol solution at 600C. The blades of saw were

immersed in propylene glycol to prevent it from drying. A companion specimen was

prepared by Struble and Stutzman (1989) which was kept in an ethanol solution with red

dye to see the extent of time taken for the replacement of ethanol in pore structure. For a

given concrete sample, the time taken for replacement depends essentially on the

thickness of the sample. After 50% of alcohol replacement in depth by visual inspection

in companion specimen, the solution is replaced by epoxy at room temperature and cured

according to the manufacture's specifications. After proper curing, the samples were

polished and lapped with abrasion papers and diamond paste in decreasing order and

viewed under Backscattered Electron imaging technique of Scanning Electron

Microscope.

The samples should be grinded and polished perfectly in order to remove raw saw

cuts on the specimen. Insufficient polishing would leave a disturbed surface on specimen

and the cracks would not be easily discernible during the microscopic observations. On

the other hand, an excessive polishing would contribute to the removal of microstructure,

and could affect the subsequent quantification operations. It was believed that higher

heating temperature would hasten the process of pore replacement. Higher temperature

for heating was not used in subsequent studies due to limited measure of quantifying the

pore water alcohol exchange and quick evaporation of alcohol. Modified epoxy









impregnation method does not require any heating or vacuum on samples and would not

induce any specimen preparation related microcracks.

However as reported by Soroushian (2003), epoxy impregnation can highlight more

porous areas of cement paste, but generally do not yield crisp boundaries and sharp

contrast between microcracks and air voids against body of concrete. A new two stage

technique was developed as Ink- Epoxy Impregnation. In the first stage, fine capillary

pores were first impregnated with parker blue ink and then microcracks and voids were

impregnated with fluorescent ultra low viscosity epoxy solution. Specimen were cut to

20-30 mm or 0.8-1 inch, and lapped by abrasive liquid under 3 PSI pressure After

cleaning the specimen for loose debris by first water bath and then ultrasonic bath,

specimen was dried at 1400 F for 3-4 hours. Ink impregnation for first stage was done by

keeping the container with ink and specimen under nitrogen pressure at 280 psi for 18-24

hours. Specimen was removed and heated at 1400 F for 24 hours to let ink dry.

Stage 2 was performed for fluorescent epoxy penetration in samples. The samples

were kept in a vacuum chamber with 0.38 psi for 1 hour. Epoxy solution was mixed and

poured in a glass container and kept in nitrogen pressure of 0.0133 psi for 3-4 hours. To

cure epoxy, samples underwent heat in an electric oven at 150 OF for 18-24 hours

following by polishing operations. While comparing with normal epoxy replacement

technique, ink-epoxy replacement technique gives better results. However, the

application of heat as well as vacuum pressure could instigate microcracks in samples.

Thin Samples

A study conducted by Anderson (1989) introduced preparing thin section of

concrete samples. In this method, a larger sample of 45 mm x 30 mm x 20 mm was cut

from a section of concrete, selected from the portion which was few centimeter away









from to be investigated area. The cut section was kept under ethanol for 12 hours to

reduce its tendency for cracking. The concrete specimen was kept in a vacuum oven at

85 'F to 90 OF to dry for two hours. A homogenized mixture of low viscosity epoxy

prepared with 1.1% by volume of dye was mixed in magnetic stirrer for 24 hours in

advance. Dried concrete specimen was kept in the solution for 1 hour for vacuum

impregnation. It was observed that concrete was impregnated by 1-2 mm in depth. Stage

2 covered mounting of glass slide on the selected face of specimen with Ultraviolet (UV)

hardening glue. The specimen was then grinded by a 15-20 [im diamond disc in order to

get smooth surface. After cleaning and drying the specimen, it was re-impregnated with

epoxy. After curing, a thin sample was cut from the specimen by diamond saw and final

grinding and polishing was done. Although the results from this method are widely used

in European counties, lengthy sample preparation time and dependability on skilled and

experienced technician were primary requirements.

Impregnation by Wood's Metal

Nemati (1997) developed a new approach in impregnation techniques using

Wood's metal instead of epoxy. The idea of this research was to preserve the

microstructure of the concrete samples which are kept under compressive stress to

analyze microcracks as they exist under loading. In this three-phased study, cracks were

introduced in the concrete specimen by a new test setup which allowed the application of

axial stress on concrete and impregnation by wood's metal simultaneously. Wood's metal

was impregnated within the sample for providing stability and better contrast for the

identification of microcracks. However, sample preparation method involved drying of

concrete cylinder by 109.50F as well as gradual heating of test assembly by 1220F, 1670F

and 204.60F for unified initiation of molten wood's metal in the specimen. As explained









earlier, heating or drying operations before impregnation could introduce microcracks in

concrete. Author reported that in no loading conditions, few cracks were observed which

could be present due to cracks introduced in sample preparation or drying shrinkage.

Wood's Metal impregnation technique was further developed by Soroshian et al.,

(2003) for concrete samples without initiating microcracks through stress introduction. In

this method, 2 inch thick slices were cut and washed to remove any loose debris attached.

The cleaned sample was kept in electrical oven at 150 OF for 24 hours to remove water

present. The dried sample was than kept in a steel mold for impregnation with Wood's

metal. To facilitate impregnation and liquidification of Wood's metal, steel mold is kept

inside a vacuum pressure chamber by 0.95 psi for 30- 40 minutes. After keeping the oven

temperature for 200 'F for 1-2 hour, the air vacuum was replaced by nitrogen pressure of

280-300 psi for 3-4 hours. The specimen was allowed to cool down, followed by cutting

a 6 mm sample and then the surface was prepared for viewing under Backscattered

technique in Environmental Scanning Electron Microscope (ESEM). Results indicated

replacement by Wood's Metal gave desired contract for microcrack identification against

concrete surface. However, heating and vacuum pressure application on concrete would

develop additional microcracks which can be quantified with original microcracks.

High Pressure Epoxy Impregnation Technique

Previous methods involving florescent epoxy replacement being time consuming,

Chen (2002) applied high pressure of> 20 bars to florescent epoxy impregnation in

concrete sample. Samples of 2x2x1.5 cm were prepared and after initial drying of 176 'F

for 2 days, samples were kept in a steel cylinder in epoxy resin at least 5 cm higher than

the top surface of specimen. The top and bottom of steel cylinder were blocked by

Teflon, desired pressure ( 145,725, 1160, 1450, 2900 and 5800 psi) was applied on the









blocks through a hydraulic piston and held under pressure for 90 minutes followed by

curing of epoxy by heating it upto 1220F for 24 hours. This robust impregnation

procedure results in a uniform and vast distribution procedure of the florescent epoxy

resin throughout the material structure. Prepared samples were viewed under optical

microscope using UV light. The experiment proved high pressure impregnation to be a

more effective and faster method to induct epoxy resin in cementitious materials. This

method show higher brightness and more effective penetration of epoxy in material and

when pressure and vacuum impregnated samples kept in optical microscope to analyze,

high pressure impregnated samples took less time for reaching a threshold value of

photon. Pressure impregnation allows higher tensile and flexural strength in samples

compared to vacuum process, thus implying greater stability of cementitious specimen

during the preparation phase. However, no reports clearly show or address the issue of

the introduction of cracks during sample impregnation.

Comparison of Sample Preparation Techniques

Microscopic image is a combination of original material from which the specimen

was taken, cumulative effects of all the procedures required to prepare the specimen for

examination, the examination technique itself and our interpretation of the image.

Specimen preparation method depends on the type of microscope used for viewing the

specimen. The review of all above techniques reveals that there is no standard specimen

preparation for viewing microcracks in cement and cementitious materials. Almost all

researchers identified pre-drying of specimen as a major source of developing cracks

prior to microscopic examination. However, most of the studies dried specimens in order

to remove pore water of concrete to facilitate the introduction of dye, epoxy or wood's

metal. While some techniques required polished concrete section or broken fracture









surfaces, some required thin sections dried at 122 'F or oven dried in NO2 atmosphere.

Some samples are cut with a diamond impregnated wire saw without a coolant, while

some samples prepared by epoxy impregnated are cut by diamond grit or lapped by

alumina grit and then polished by diamond grit/ abrasive liquid of decreasing order

(Marusin S., 1995).

Some researchers believe that application of improper sample preparation methods

would erode and destroy existing microstructure. It can also introduce elements which are

not part of the original specimen which would lead to false diagnosis of damage

mechanism. Most of the specimen preparation techniques requires to dry sample till a

certain degree which would affect the pore structure and can induce more microcracks.

As reported by Hornain et al., (1996), crack density increases with intensity of drying.

Thus such methods would not be able to give true quantitative analysis of microcracks.

Vacuum impregnation methods results in excessive drying of the sample and therefore

cracks are introduced during sample preparations (Chen 2002). Techniques involving

mechanical polishing to produce either a flat surface or thin specimen could be

problematic for concrete as different phases of concrete polish at different rates. Similarly

putting concrete samples into a vacuum would lead to loss of water and change of

structure.

Polishing and lapping operation to smooth the specimen surface would lead to

spreading of components over surface, thus preventing effective use of dot map studies

for elements. Polishing would also destroy delayed ettemdite crystals and cracks

originating from it. While pressure or vacuum would introduce more cracks in the

sample, original cracks would heal or widened by epoxy impregnation techniques









(Marusin, 1995). However, sawn and unpolished approach could be better if the study is

related to DEF analysis but for proper crack identification and quantification, sample

preparation techniques would be required.

In brief, Bisschop and Van Mier (2002) and other studies have indicated

impregnation of whole sample and then cutting thin samples from fully impregnated

samples would introduce less microcracks. Specimen preparation techniques involve

cutting, drying, lapping, grinding and polishing. Improper handling of these operation

results in induction of additional microcracks. For the scope of this study, epoxy

impregnation technique developed by Struble and Stutzman (1989) was performed as it

did not involve any drying of samples in any form. In this two-stage counter diffusion

process, pore water is replaced by ethanol, and ethanol is substituted by epoxy without

involvement of preheating or use of pressure techniques.

Microscopic Instruments

Optical Microscope

Optical microscopy is one of the favored techniques for concrete petrography. In

this instrument, image is viewed in full color and generally at a lower magnification, thus

making it a more suitable technique for observing features at the millimeter scale and

larger. Concrete and cementitious specimens prepared by thin-sectioning and fluorescent

microscopy techniques can be viewed under optical microscope with the use of reflected

ultraviolet lights. The image acquisition is done by placing a Tri-CCD camera which is

linked to a personal computer acquisition system. However, Optical microscopes have

shallow depth of field and limited resolution capacity, so a highly smooth and polished

surface is required to produce a focused image under this technique. Furthermore, the










intensity of the vacuum-impregnated sample is not stable after 2 min of exposure to UV

light, and decreases in a continuous fashion with the elapse of time (Chen 2002).

Scanning Electron Microscope

Scanning electron microscopy/ x-ray microanalysis is comparatively more useful in

studies which investigate quantification of microstructural properties such as microcracks

and voids. Scanning electron microscopes has greater depth of field and high spatial

resolution which produces focused images of poor specimens and analytical data for

elemental composition analysis for the features seen on those images. As scanning

electron microscopy and optical microscope petrography are different tools, different

sample preparations are used for these different approaches.

As described in figure 2.1, SEM scans a focused beam of electron across the

specimen and measures any of several signals resulting from electron beam interaction

with surface of concrete specimen. Images are monochrome in nature because they

reflect the electron or x-ray flux resulting from beam/specimen interaction. Three major

types of signals generated as a result are Secondary electrons, backscattered electrons and

X-rays analysis.





1e-sys(= l









specimen d&t1or
to pumps

Figure 2-1 Working of Scanning Electron Microscope, (Perkes, 1999)









Secondary Electrons (SE) are low energy electrons resulting from an inelastic

collision of a primary beam electron with an electron of a specimen atom. Because of

their low energy, they are readily absorbed and only those produced near the surface

escape, resulting in an image of surface topography. SE imaging is principally applied in

the examination of early age paste microstructure, high magnification imaging of

microstructural features and for examining texture. Knowledge of the morphological and

compositional characteristics of the hardened cement paste constituents is invaluable for

their identification. As the hardened cement paste matures, filling of the void spaces

eliminates the well-formed crystals and backscattered electron and x-ray imaging are

more useful in examination of these microstructures (Stutzman, 2001)

Backscattered Electron are high energy beams and capable of reflecting difference

in atomic numbers. It can distinguish between the particles present in matrix on the basis

of the variation in brightness of their images. In Backscattered Electron imaging

technique contrast is generated by different phase compositions relative to their average

atomic number and is observed by differential brightness in the image. SEM analysis

using backscattered electron and X ray imaging requires highly polished surface for

optimum imaging and X ray microanalysis.

X radiation is produced when a specimen is bombarded by high energy electrons.

X-ray microanalysis systems generally employ an energy dispersive detector with the

other detector type being a wavelength detector. The x-ray energy level is displaced as

the number of counts of each energy interval and appears as a set of peaks on a

continuous background. X-ray radiation allows elements compositions to be obtained as









the qualitative analyses prints like a dot map images. Each element is identified on a

continuous spectrum by the position of its pick.

Any sample can be examined by SE imaging but best use of BE imaging is for

examining flat, polished surfaces (Struble and Stutzman, 1989). One of the disadvantages

of Scanning Electron Microscope is that samples must be coated to allow discharge of

electron build-up on the examined surface. A thin metal coating decreases build-up of

negative charge by forming a conducting path for electrons to avoid distorted images. If

images are accompanied by X-ray analysis, carbon coating can not be used if a

carbonation process is to be studied. A gold coating gives a peak almost at a same

position as sulfur (Oberhostler 1992). Thus, in some samples, it might be appropriate to

coat the specimen due to concerns for altering the microstructure. High vacuum pressure

inside the specimen chamber is also one of the factors which limits the use of SEM as

some samples might break or disintegrate under high vacuum pressure.

Variable Pressure Scanning Electron Microscope

Hanke L, (1999) described that many applications where SEM/EDS evaluation

could be useful involve samples that are not electrically conductive. These samples have

traditionally required pretreatment, by coating with a conductive film, before SEM

examination. Nonconductive samples are subject to a buildup of electrons on the

examined surface. This buildup of electrons eventually causes scattering of the incoming

electron beam, which interferes with imaging and analysis. Furthermore, samples that

contain substantial water or other materials that volatilize in high vacuum also present

challenges for SEM examination. These samples require controlled drying to allow the

SEM chamber to reach high vacuum and to prevent deformation of the sample at the

SEM vacuum.









As the SEM is increasingly used for routine evaluations, there is increasing demand

for examination without pretreatment. An answer to the problems of charging and

volatile samples is the development of scanning electron microscopes that operate

without exposing the sample to high vacuum. These microscopes are referred to

alternately as environmental, low-vacuum, or variable-pressure SEM. In variable pressure

SEM, the chamber at the electron gun is maintained at high vacuum, while a controlled

amount of gas is allowed into the sample chamber. A fine aperture separates the gun and

sample chambers to prevent excessive gas entrance into the gun chamber. Separate

vacuum systems control the vacuum in the sample chamber and at the gun. The

advantages of a higher pressure in the sample chamber are obvious for wet and volatile

samples. The higher pressure decreases the rate of volatilization. This decreases the

drying and deformation of wet samples.

For nonconductive samples, the advantage of higher pressure is less obvious. When

gas molecules in the sample chamber are struck by the electron beam, the gas is ionized.

These positive ions are attracted to and neutralize the negative charge build-up on the

nonconductive specimens. By controlling the pressure in the sample chamber, the number

of gas molecules intercepting the electron beam is maintained at a level that is sufficient

to prevent charging, but does not deflect the beam sufficiently to prevent imaging and

microanalysis.

Thus, the use of variable pressure scanning electron microscope holds advantage

over traditional scanning electron microscope as no coating is required over the samples

and samples with moisture could also be imaged.









Image Processing Techniques

Important steps for any microstructural images are to acquire the image and to

process microstructure image to get a quantitative analysis of the sample. For this study,

quantitative analysis is to identify and measure the number of cracks and crack density

over number of images. The image processing should be able to differentiate between

voids and cracks, and boundary cracks around aggregates or between phases. Darwin

(2001) researched on quantification of microcracks of cement mortar samples by

backscatter electron technique of scanning electron microscope. He reported that the use

of edge detection, gradients and other filters failed to identify cracks passing through low

density phases as its intensity depends on adjacent and underlying phases. In his

approach, the identification of cracks was based on local changes in grey level. A line

scan was taken perpendicular to the images for identification of potential cracks.

Perimeter to squared area was taken to differentiate between cracks and voids. Phase

identification was also carried out based on pixel intensity. He conclude that as grey level

of image is affected by the density of adjacent and underlying phases, cracks identified

based on change in grey level would not be correct. Thus the correct practice for crack

identification should establish minimum gradient in grey level adjacent to cracks.

Soroshian (2003) used segmentation of images from grey level to binary images by

thresholding approach for viewing microcracks in samples. The research compared

manual thresholding operations with three automated thresholding operation such as

factorization, entropy and moment. In manual thresholding, low threshold level was set to

zero or default for auto thresholding and best threshold level image was determined. The

optimum high threshold level was determined by comparing original gray image and the

image after the application of manual threshold on binary image. When features of









interest (microcracks and voids) were correctly highlighted, the most distant contrast

between microcracks/voids and concrete background was obtained. After training the

program with images of different resolutions, three automated thresholding operation

were performed and compared results with manual thresholding. The study reported that

there was no significant difference between manual thresholding and automated

thresholding operation by factorization method. There was a difference of 5% and 2%

respectively, for entropy and moment methods of automated process. After identifying

the cracks, form/shape factor of 3.5 was used to distinguish cracks from voids. It was

shown that automated process was different by 6.39% and 3.02% from manual process.

However, the author concluded that due to the ease of operation and efficiency given by

automated process, this error was reasonably small.

Summary

Literature review of methods for preparation of concrete samples for viewing

microstructure was performed. As described before, many sample preparation methods

involved the use of pressure or heat, which might cause concerns over additional cracks

in concrete and might lead to erroneous results. Epoxy impregnation of concrete samples

was identified as a process which uses counter diffusion process as sample preparation

means for this study. As optical microscope would have inadequate depth of field and

resolution for microstructure, scanning electron microscopy was preferred for image

acquisition, Variable pressure electron scanning microscope, which requires no special

coating over samples and ability to view samples with moisture was found to be more

suitable for image analysis. After acquiring images, manual thresholding approach to

map microcracks was selected for quantification of microcracks in a given sample














CHAPTER 3
RESEARCH METHODOLOGY

Introduction

This chapter describes the methods for sample selection, specimen preparation

method and image analysis techniques used in this research. As described in Chapter 2,

the objective of this study was to find if the concrete cured at elevated temperature shows

any additional microcracks compared to samples cured at room temperature. This

investigative study was carried out in three phases. Phase one was to identify the samples

which shows tendency for microcracks, while phase two was to select an appropriate

sample preparation methods. Phase three comprised of sample preparation and image

analysis technique to quantify and compare results.

Sample Selection

Samples were selected from an ongoing research project "Adiabatic Temperature

Rise in Mass Concrete" conducted by University of Florida for Florida Department of

Transportation (FDOT). In this study, twenty different mixes were prepared using

AASHTO Type II cement made by two manufactures with different percentages of

pozzlonic materials and two different placing temperatures of 730F and 950F were used

to study adiabatic temperature rise in laboratory simulated mass concrete conditions. As

higher placing temperature increases the concrete curing temperature, samples made from

higher of two placing temperatures, 950F were selected for microstructural analysis. The

field of investigation was further narrowed down further by selecting the mix made with

cement which has higher heat of hydration. A Heat of Hydration test by Construction










Technologies Ltd. (CTL) revealed higher heat of hydration in AASHTO Type II cement

(Cement B) manufactured by CEMEX (78.2 cal/g at 7 days) was greater compared to

AASHTO Type II cement ( Cement A) manufactured by Florida Rock (66.2 cal/g at 7

days). Thus, mixes made with Cement B and 95 'F placing temperature were selected.

This included five different mixes were prepared with varied pozzlonic contents.

* 100% Cement

* 20% Fly Ash + 80 % Cement

* 35% Fly Ash + 65 % Cement

* 50% Slag + 50% Cement

* 70% Slag + 30% Cement

Table 3-1 Mix Design Data


AASHTO Type II CEMEX Cement ( Cement B ) (Mix proportions per cubic yard)

Cement Fly Slag Water Fine Coarse Air Water
Cement Slag Water
Mix Ash Aggregate Aggregate Entertainer Admixture
(lb) (lb) (lb) (oz) (oz)
95BOOP 760.00 0.00 0.00 279.00 1033.00 1736.00 2.30 60.86

95B20F 608.00 152.00 0.00 279.00 1033.00 1708.00 4.00 30.40

95B35F 494.00 266.00 0.00 279.00 980.00 1687.00 4.00 68.20

95B50S 380.00 0.00 380.00 279.00 1021.00 1724.00 2.30 38.06

95B70S 228.00 0.00 532.00 279.00 1021.00 1724.00 2.30 68.20


Prepared concrete samples were kept in two different types of containers. For

samples to be kept at room condition, concrete was poured in 6x12 plastic containers and

for samples to be kept at high temperature cured condition, specially acquired thermal

curing chambers were used. Thermal curing chamber maintained temperature inside

chamber in an adiabatically condition. The chambers were connected with thermo










couples to a micro controller system for maintaining temperature and to document the

temperature rise in the concrete samples.

Data given in following Table 3.2 shows the first four day temperature rise in

concrete specimen kept in thermal curing chambers.

Table 3-2 Adiabatic Temperature Rise Data

Cement B (950F Placing Temperature)
Time Plain Cement 20% Fly Ash 35% Fly Ash 50% Slag 70% Slag
(Day) Temp. Rise Temp. Rise Temp. Rise Temp. Rise Temp. Rise
(OF) (OF) (OF) (OF) (OF)
0.0 0.00 0.00 0.00 0.00 0.00
0.5 2.60 56.80 50.35 58.60 45.00
1.0 72.15 72.90 61.90 76.70 69.40
1.5 84.20 77.40 67.85 80.70 73.90
2.0 85.60 77.60 68.90 81.80 76.30
2.5 85.65 77.60 69.20 82.00 77.90
3.0 85.95 77.60 69.30 82.00 78.90
3.5 85.95 77.60 69.60 82.00 79.50
4.0 85.95 77.60 69.90 82.00 79.90

As it can be seen in Table 3-2, data indicates that sample 95 B 00 P had the highest

temperature rise of 85.95 'F over placing temperature of 99 'F, thus effectively curing

temperature in the core is 1860F. Therefore, sample 95 B 00 P was selected for

microscopic studies due to the following reasons.

* Higher placing temperature had higher curing temperature and therefore gives
higher probability of presence of additional microcracks

* Cement B had higher heat of hydration (78.2 cal/g at 7 days) compared to Cement
A (66.2 cal/g at 7 days), which might lead to higher curing temperature

* As indicated in Table 3.2, sample 95 B 00 P with plain cement showed higher
temperature rise compared to samples with varied percentages of fly ash or blast
furnace slag.

Specimen Preparation Method

As described in Chapter 2, many different approaches have been used in the past to

view and document microcracks in cement paste or concrete. However, most of the









methods indicate the use of heat, vacuum or pressure along the course of sample

preparation or prior to impregnation, which might lead to additional microcracks due to

sample preparation widely known as secondary microcracks. Samples cured at elevated

temperature as well as those cured at room temperature would undergo epoxy

impregnation method for Scanning Electron Microscopy as established by Struble and

Stutzman (1989). In this method they used counter diffusion method to replace the pore

water of concrete first by ethyl alcohol and then low viscosity epoxy. Neither of these

processes requires any prior pre-heating of surface to dry, nor does it use any vacuum for

impregnation of epoxy. The only drawback of this method is that sample preparation

would take longer time to complete the process. Considering all methods and their

advantages and disadvantages, this method was selected for sample preparation in current

study.

Procedure for Sample Preparation

A small 1/2" thick sample was cut with diamond saw from the samples cured at

room temperature and those cured in thermal curing chambers. During this operation, in

order to reduce the heat generation, cooling water was circulated through water pump to

diamond blade to aid cutting. Once the required thickness of the sample is obtained, other

cutting operation was carried out on a diamond wafer saw machine (figure 3-1). It was

used to reduce any artifacts and required polishing time in the samples during the

preparation period. In this machine, blade was passed through cutting fluid

(polypropylene glycol) to keep the samples from drying. After cutting by diamond

wafer, average size of specimen was 2.15"x 1.8". The samples then went through a two

stage counter diffusion process.













i I I


Figure 3-1 Diamond Wafer Saw

In this process, first pore water of concrete is replaced by 200% Ethanol made by Sigma-

Aldrich Inc., Montana, and then in the second stage, ethanol was replaced by ultra low

viscosity epoxy solution made by Structure Probe Inc., Pennsylvania.

For the first stage of alcohol pore water replacement stage, the specimen was

placed in a lidded jar filled with ethanol (200 Proof) for the alcohol pore solution

replacement stage. A companion specimen was used to determine the depth of

replacement of pore water by alcohol. This companion specimen was a remnant from the

specimen after trimming. This specimen was now placed in a jar filled with ethanol dyed

by a deep red alcohol-miscible dye by Poly science Inc. By splitting or sawing the

companion specimen after a period of time, the depth of replacement is seen by the depth

of dye coloration. On the average, alcohol- pore water replacement was found to be

Imm/day in concrete samples. When this depth reaches to half the section thickness, the

pore solution replacement by alcohol was stopped. The section is then placed in a









container with the ultra low-viscosity epoxy (see Figure 3-2) for the same time as alcohol

pore water replacement process. Curing of epoxy was performed as specified by product

manual. The epoxy was made up of 4 different chemicals as described in Table. 3-3

















Figure 3-2 Ultra Low Epoxy Kit by SPI

Table 3-3 Mixing Schedule for Epoxy
No. Name of Chemical Mixing Schedule
1. Vinylcyclohexene dioxide (VCD) 10g
2. n-Octenyl succinic anhydride (n-OSA) 20 g
3. Butanediol Diglycidyl Ether (BDE) 0.3 g
4. Dimethylaminoethanol (DMAE) 0.3 g

Ethanol was replaced by epoxy at two hour intervals with the proportion of ethanol

to epoxy was changed from 3:1, 1:1, 1:3 at each interval and finally 100% epoxy

solutions was used for keeping it in embedding medium overnight. For curing of epoxy,

samples were kept in a heating chamber at 1310F for six hours. At the end of six hours,

the specimens were ready for the cutting and polishing stages.

Grinding and polishing are important steps in sample preparation. Grinding is used

to expose new surface layer and to remove excessive epoxy coating on the surface.

However, excessive grinding would remove the depth of epoxy layer and could damage

microstructure. Grinding was done on a Metaserv 2000 Grinder Polisher Machine (see









figure 3-3 and 3-4) on 150 RPM. Abrasive papers of 180,240, 320,400 and 600 were

used for the removal of material from specimen surface. Approximately 2 minute of time

was given to each grinding cycle.


















Figure 3-3 Metaserv Grinder and Polisher Figure 3-4 Polished Sample


Polishing is done for undoing the damage done due to sawing and grinding of

surface. After grinding by 600 grits, surface becomes smooth and ready for further

polishing. Polishing operation involves successively decreasing size of diamond

suspension paste from 6 m to 3pm, lm and 0.25 im manufactured by Buhler (see figure

3-5). A polishing cloth was attached to rotating wheel of Metaserv 2000 Grinder Polisher.

This method helps in reducing the damage done by previous suspension by removal of

finer scratches and improving the surface for microscopic study. At the end of this

operation, the sample is wiped with a plain cloth and kept in glass jar (See figure 3-6).

During polishing process, it is important to understand that cutting by diamond saws

would damage the samples up to 30% of its thickness. So polishing should be done till

30% of blade thickness in order to remove all artifacts.























Figure 3-5 Diamond Suspension Paste


Figure 3-6 Prepared Samples


Preparation for Image Processing

Microstructural analysis was performed at Advance Material Characteristic Lab at

the Department of Civil and Coastal Engineering, University of Florida, Gainesville. The

facility has a Hitachi S 3000N, variable pressure Scanning Electron Microscope

(VPSEM). As described earlier, VPSEM would allow the viewing of the specimen

without covering it with a conductive surface. A double sided carbon tape was used for

attaching samples to specimen holder. This conductive tape avoids any charge built-up

and provides a good background surface.

Images were viewed at 50 Pa (0.4 Torr), accelerating voltage of 10.0 KV and on the

scale of 200 [im at 250 X magnification of specimen surface. At the start of this study,

polished samples were viewed directly under microscope and five representative images

were taken from different location of specimen. However, during the course of study, it

was decided to increase the number of images to 16 to have good representative nature of

the sample, which was formulated on basics of 4x4 matrixes. The following figures 3-7

and 3-8 shows the grid system.











1 2 3 4




2.15"

9 10 11 12


13 14 15 16


1.80"1


Figure 3-7 Specimen with Copper Tape Figure 3-8 Specimens with 4x4 matrixes




The samples were divided into 16 equal parts using a copper conductive tape to

make grid on the sample to identify the points of interesting while viewing under

microcracks.

Image Analysis

Images taken from 16 different square position provides information related to

microstructure of concrete. A manual mapping of cracks was done in order to find out the

length of the microcracks and density per micrometer for comparing all the samples. To

identify proper microcracks from the image, shape factor was used. Shape factor is ratio

of length to width. If this ratio was greater then 3.5, the object of consideration termed as

microcracks or if the ratio was less than 3.5, the object was termed as Void. Each image

was manually mapped as well as by viewing on ImageJ software to find out the exact

number of cracks to avoid any bias from operator. The method of image analysis is

described in more details in Chapter 4.









Summary

In this study, samples were taken from concrete mix which showed maximum

probability of crack development due to high curing temperature. Specimen preparation

method was selected to avoid any preheating of the samples and to avoid any use of

vacuum or pressure techniques for impregnation. Once specimen were impregnated and

prepared with grinding and polishing, a grid made from conductive copper tape was used

to locate the points of interests, while viewing the sample under microscope. At the end, a

statistical analysis was performed to investigate total representative nature of images and

the rationality for use in this study.














CHAPTER 4
IMAGE PROCESSING AND ANALYSIS

Introduction

This chapter describes the image processing and analysis carried out in this

research. A total of four samples were made, two of which cured in room temperature

and the other two cured at elevated temperature. As described earlier, the samples were

divided into 16 equal parts by conductive copper tape. The images were taken around the

center of square in 250X resolution Thus for each sample a total of 16 images were taken

in 1028 x 960 pixel resolution and 50 Pa working pressure.

In SEM observation, finer surface structure images can generally be obtained with

lower accelerating voltages. At higher accelerating voltages, the beam penetration and

diffusion area become larger, resulting in unnecessary signals (e.g., backscattered

electrons) being generated from within the specimen. These signals reduce the image

contrast and veils fine surface structures. It is especially desirable to use low accelerating

voltage for observation of low-concentration substance.

After acquiring the images, image processing was performed to view microcracks

in concrete matrix. ImageJ software developed by National Institute of Health was used

for the purpose of image processing in this study. Based on java platform, this software

has the capability to perform wide ranges of operations such as binary thresholding, area

measurement, filtering noise and mathematical operations.









Image Processing Technique

First step of image processing was to perform segmentation of image into binary

image (black and white) by brightness thresholding approach. In this process, software

would change the intensities of image pixels which satisfy the given condition, thus

highlighting the regions of interests. However, as it can be imagined, there would be

variations in grey level of matrix in different images, and there could be bias in operator

visions. Investigating this issue, Soroshian (2003) performed a comparative analysis of

manual and automated thresholding operation on images taken by Scanning Electron

Microscope, where three types of automated thresholding operation factorizationn,

entropy and moment) were performed. However, study reported that there were no

significant differences between manual and factorization mode of thresholding in

automated process. The study also reported 5% and 2% difference in Entropy and

Moment mode of thresholding respectively which could be regarded as less significant.

Therefore for the scope of this study, manual thresholding by brightness approach

was performed. Images were acquired with variable pressure scanning electron

microscope and PCI-quartz picture acquisition system. Images were transferred and

viewed with ImageJ software. For each image, low threshold level was set to zero or

default for thresholding and best threshold level image was determined by comparing the

original gray image and the image after application of manual threshold on binary image.

Total of sixty-four images were converted one after another to binary mode. After binary

operation, images were manually mapped to find out total number of crack lengths and

crack density across specimens.











Sample 95BOOP Room Temperature Sample 1


A. Image 95 B 00 P -RT 1 (2)


C. Image 95 B 00 P -RT 1 (11)


B. Image 95 B 00 P -RT 1 (10)


D. Image 95 B 00 P -RT 1 (16)


Figure 4-1 Representative Images of Mix 95 B 00 P RT-1 before Image Processing

After highlighting features of interest (microcracks and voids) to find the most

distinct contrast with concrete background, image analysis was performed to distinguish

microcracks and voids. As it is known that microcracks are long elongated structure,

length to width (also known as aspect/ form/ shape) ratio was applied. The objects having

shape factor of 3.5 or greater was termed as microcracks, while the remaining objects

were termed as voids. A software plugin Shape descriptor from ImageJ was used to










identify voids in the images. Shape descriptor takes minimum and maximum pixel values

and plots the void concentration in a separate image. After completing void removal from

image, microcracks are mapped manually and counted.


A. Image 95 B 00 P RT 1 (2)


C. Image 95 B 00 P -RT 1 (11)


B. Image 95 B 00 P RT -1 10 )


SUPA 21-OCt-U4 RT1 WDUl.I2m 1UURkV X25U ZIJUU7


D. Image 95 B 00 RT -1 (16)


Figure 4-2 Representative Images of Mix 95 B 00 P RT-1 after Image Processing

As the sample was divided into 16 equal parts by copper conductive double-sided

tape, 16 representative images were taken from the center of every part. After processing

the image, cracks were manually mapped to find out the total length of cracks in each


i









image. As it can be seen in table 4-1, density of cracks over the given area was found by

converting crack lengths and dividing it by image area.

Table 4-1 Crack Quantification Analysis for Mix 95 B 00 P RT-1

No. Image Scale Cracks Cracks Image Field Crack
(cm) length Density
(lm)_ (m/4m2
1 RT 1-1 200Om/11 cm 72 cm 1309.091 183 pm x 250 im 0.028614
2 RT 1-2 200Om/11 cm 71 cm 1290.909 183 pm x 250 im 0.028217
3 RT 1-3 200nm/11 cm 65 cm 1181.818 183 mr x 250 mr 0.025832
4 RT 1-4 200im/11 cm 82 cm 1490.909 183 rm x 250 pm 0.032588
5 RT 1-5 200nm/11 cm 61 cm 1109.091 183 mr x 250 mr 0.024242
6 RT 1-6 200Cim/11 cm 66 cm 1200.000 183 rm x 250 rm 0.026230
7 RT 1-7 200nm/11 cm 64 cm 1163.636 183 rm x 250 pm 0.025435
8 RT 1-8 200rm/11 cm 61 cm 1109.091 183 mx250 m 0.024242
9 RT 1-9 200rm/11 cm 73 cm 1327.273 183 rmx 250 pm 0.029011
10 RT 1-10 200um/11 cm 92 cm 1672.727 183 mr x 250 mr 0.036562
11 RT 1-11 200pm/11 cm 72 cm 1309.091 183 pm x 250 pm 0.028614
12 RT 1-12 200m/11 cm 68 cm 1236.364 183 mr x 250 mr 0.027024
13 RT 1-13 200um/11 cm 98 cm 1781.818 183 mx250 m 0.038947
14 RT 1-14 200um/11 cm 63 cm 1145.455 183 mr x 250 mr 0.025037
15 RT 1-15 200pm/11 cm 65 cm 1181.818 183 pm x 250 pm 0.025832
16 RT 1-16 200pm/11 cm 58 cm 1054.545 183 pm x 250 pm 0.023050
Avg. 1285.227 0.028092
SD. 203.7358 0.004453

Data analysis shows that for 95 B OOP RT-1 samples, average lengths of cracks

were 1285.227 rim. Maximum length of cracks in an image was found to be 1672[im,

while minimum length of cracks in an image was 1054.545 rim, giving standard deviation

of 203.227. As it can be seen in Table 4.1, average crack density in an image was

0.028092 nm/pm,2 the maximum density of cracks was 0.038947 im/m2 and minimum

density of cracks was 0.02305 rm/pm2, giving standard deviation of 0.004463. Figure

4.3 describes variations in crack density of room temperature cured sample by each


image.










96 B 00 P Room Temprature Sample 1
0.05
> 0.04
S0.03
0.02 -
S0.01

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
Image Number


Figure 4-3 Variations in Crack Density of 95 B 00 P RT1

To find out if the sample size was representative in nature, a statistical study was

performed. Odd and Even numbers of images were selected for calculating average

density of lengths. As it can be seen in Table 4-2, Crack density of odd number images

were 0.02832 gm/gm2 with standard deviation of 0.005281, while crack density of even

number images were 0.02787 gm/gm2 with standard deviation of 0.004982. After

comparing it with average crack density of all 16 images of 0.02809 gm/gm2, it can be

seen that crack density is averaged almost equally. Thus it could be deduced that crack

density is representative in nature throughout the sample and the number of images are

adequate for giving representative portrait of the sample.

Table 4-2 Comparison of Odd and Even Number Images

No Crack Density of Odd Number Image Crack Density Even Number Images
No.(m/gm2) (m/gm2)
1 0.028614 0.028217
2 0.025832 0.032588
3 0.024242 0.026230
4 0.025435 0.024242
5 0.029011 0.036562
6 0.028614 0.027024
7 0.038947 0.025037
8 0.025832 0.023050
Avg. 0.028316 0.027869
SD 0.004643 0.004563









Sample 95BOOP Room Temperature Sample 2


Figure 4-4 Representative Images of Mix 95 B 00 P RT-2 before Image Processing

Figure 4.4 shows images from sample number two which was cured at room

temperature. It can be seen from images that extensive cracking is present in almost all

parts of representative images. As described earlier, a total of 16 images were taken from

each sample. The images were then processed by ImageJ software for quantification of

microcracks.




























50Pa 15-Oct-04 wD17,0o 10.0kV x250 200um


Image 95 B 00 P-RT2 (4)


Image 95 B 00 P-RT2 (15)


Image 95 B 00 P-RT2 (10)


Image 95 B 00 P-RT2 (16)


Figure 4-5 Representative Images of Mix 95 B 00 P RT-2 after Image Processing

After processing the image by converting it into binary image, cracks were

manually mapped to find out the total length of cracks in each image. As it can be seen in

table 4-3, the density of cracks over the given area was found by converting crack lengths

and dividing it by image area. Another statistical analysis was carried out to see if the

number of images constituent representative area for the sample.


I









Table 4-3 Crack Quantification Analysis for Mix 95 B 00 P RT-2


No. Image Scale Cracks Cracks Image Field Crack
length Density
(cm) (nm/m2) (Gm/4m2)
1 RT 2-1 200pm/11 cm 64 cm 1163.636 183 pm x 250 pm 0.025435
2 RT 2-2 200Om/11 cm 68 cm 1236.364 183 pm x 250 Lm 0.027024
3 RT 2-3 200m/l 1 cm 56 cm 1018.182 183 m x 250 m 0.022255
4 RT 2-4 200m/11 cm 89 cm 1618.182 183 mx250 m 0.035370
5 RT 2-5 200m/11 cm 90 cm 1636.364 183 pm x 250 pm 0.035768
6 RT 2-6 200m/11 cm 65 cm 1181.818 183 pm x 250 Lm 0.025832
7 RT 2-7 200m/11 cm 82 cm 1490.909 183 pm x 250 pm 0.032588
8 RT 2-8 200(m/11 cm 69 cm 1254.545 183 mx250 m 0.027422
9 RT 2-9 200um/11 cm 83 cm 1509.091 183 pmx250 pm 0.032986
10 RT 2-10 200um/11 cm 67 cm 1218.182 183 pm x 250 pm 0.026627
11 RT 2-11 200pm/11 cm 64 cm 1163.636 183 pm x 250 pm 0.025435
12 RT 2-12 200um/11 cm 61 cm 1109.091 183 pm x 250 pm 0.024242
13 RT 2-13 200um/11 cm 67 cm 1218.182 183 mx250 m 0.026627
14 RT 2-14 200um/11 cm 82 cm 1490.909 183 pmx250 pm 0.032588
15 RT 2-15 200um/11 cm 84 cm 1527.273 183 pm x 250 pm 0.033383
16 RT 2-16 200pm/11 cm 68 cm 1236.364 183 pm x 250 pm 0.027024
Avg. 1317.045 0.028788
SD. 194.692 0.004256

Data analysis shows that for 95 B OOP RT-2 samples, average lengths of cracks was

1317.045 rim. Maximum length of cracks in an image was found to be 1636km, while

minimum length of cracks in an image was 1018.182 rim, giving standard deviation of

194.692. As it can be seen in figure 4-6, average crack density in an image was 0.028788

inm/im2, maximum density of cracks was 0.035768 im/m2 and minimum density of

cracks was 0.0222 im/pm2, giving standard deviation of 0.004256.

The above data shows presence microcracks in the given sample. However, to see

if the images were sufficient to draw conclusions from it, another statistical analysis was

carried out. In this technique, images were divided into two formations. The formation

was varied such as diagonals, odd-even or lines to prove representative nature of the


samples.




























Figure 4-6 Variations in Crack Density of Sample 95 B 00 P RT2

Figure 4-6 demonstrates variations in crack density of sample 95 B OOP RT2 by

each image. The data given in Table 4-4 shows that the crack density of odd number

images were 0.029309 gm/gm2 with standard deviation of 0.0049, while crack densities

of even number images were 0.028266gm/gm2 with standard deviation of 0.0037.

Table 4-4 Comparison of Odd and Even Number Images

No Crack Density of Odd Number Image Crack Density Even Number Images
(gm/gm2) (gm/gm2
1 0.025435 0.027024
2 0.022255 0.035370
3 0.027422 0.032588
4 0.025832 0.035768
5 0.032986 0.026627
6 0.025435 0.024242
7 0.026627 0.027024
8 0.033383 0.032588
Avg. 0.029309 0.028266
SD 0.004922 0.003737



After completing analysis of samples cured at room temperature, crack density

calculations were performed on samples cured at high temperature. As described earlier,

samples were divided in 16 equal parts with conductive copper tape and images were


C 95 B 00 P Room Temperature

a 0.04
c 0.035
k 0.03
0.025
D 0.02 -
e 0.015 -
0.01
s
0.005 -
T 0
Y 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
Number of Images









taken from the center of rectangle. Figure 4-7 shows images from high temperature cured

sample one.

Sample 95BOOP High Temperature Sample 1


Figure 4-7 Representative Images of Mix 95 B 00 P HT-1 before Image Processing

Maximum heat gain during high temperature curing process of concrete reached up

to 85.5 'F. The adiabatic temperature condition was generated and measured in the

laboratory with a microcontroller based SURE-CURE curing system.

























Image 95 B 00 P HT 1 (2)


UvFa 2-O- GU WD01.M 1U.U'W ZwU MVUW
Image 95 B 00 P -HT 1(3)


dUFA 2U-uct-U GCU W OIU.J=n laU.wCY LV X UUuM
Image 95 B 00 P -HT 1(7)


Figure 4-8 Representative Images of Mix 95 B 00 P HT-1 after Image Processing

After preparing the samples and subsequent polishing and lapping operations,

copper metal grid system was mounted. Samples were kept in variable pressure scanning

electron microscope to analyze microscopic features. As it can be seen in Figure 4-8,

images were initially processed by manual thresholding by brightness approach. Lower

thresholding was kept at zero, while upper thresholding was moved till the best contrast

between images and objects of interest developed. Voids were removed from the binary

picture and manual mapping of microcracks were done for crack quantification.


Image 95 B 00 P HT-1 (1)









Table 4-5 Crack Quantification Analysis for Mix 95 B 00 P HT-1

No. Image Scale Cracks Cracks Image Field Crack
length Density
(cm) (Gm) (Gm/4m2
1 HT 1 -1 200m/11 cm 105 cm 1909.091 183 pm x 250 Lm 0.041729
2 HT 1-2 200m/11 cm 106 cm 1927.273 183 pm x 250 Lm 0.042126
3 HT 1-3 200m/11 cm 94 cm 1709.091 183 mx250 m 0.037357
4 HT 1-4 200im/11 cm 52 cm 945.4545 183 pm x 250 pm 0.020666
5 HT 1-5 200um/11 cm 58 cm 1054.545 183 pm x 250 pm 0.023050
6 HT 1-6 200pm/11 cm 53 cm 963.6364 183 pm x 250 pm 0.021063
7 HT 1-7 200m/11 cm 48 cm 872.7273 183 pm x 250 pm 0.019076
8 HT 1-8 200(m/11 cm 54 cm 981.8182 183 mx250 m 0.021461
9 HT 1-9 200um/11 cm 58 cm 1054.545 183 mx250 m 0.023050
10 HT 1-10 200um/11 cm 55 cm 1000.000 183 pm x 250 pm 0.021858
11 HT 1-11 200pm/11 cm 71 cm 1290.909 183 pm x 250 pm 0.028217
12 HT 1-12 200m/11 cm 65 cm 1181.818 183 pm x 250 pm 0.025832
13 HT 1-13 200um/11 cm 62 cm 1127.273 183 mx250 m 0.024640
14 HT 1-14 200um/11 cm 75 cm 1363.636 183 mx250 m 0.029806
15 HT 1-15 200um/ll cm 65 cm 1181.818 183 pm x 250 pm 0.025832
16 HT 1-16 200pm/11 cm 82 cm 1490.909 183 pm x 250 pm 0.032588
Avg. 1253.409 0.027397
SD. 339.4677 0.007420

Data analysis from Table 4-3 shows that for 95 B OOP HT-1 samples average

lengths of cracks were 1253.409 rim. Maximum length of cracks in an image was found

to be 1927 rim, while minimum length of cracks in an image was 872.73 rim, giving

standard deviation of 339.4677. As it can be seen in Table 4-5, average crack density in

an image was 0.027397 [im/pm2, maximum density of cracks was 0.042126 im/m2 and

minimum density of cracks was 0.019076 rm/pm2' giving standard deviation of 0.00742.

The above results determine presence of microcracks in the given sample, to find if

the sample size is sufficient to draw conclusion, another statistical analysis was carried

out. In this process, images were divided into two groups in various formations and

average crack density was calculated for each of them. Results shown in table 4-6 shows

that average crack density is present throughout the concrete sample.











95 B 00 P High Temprature Sample 1


0.05
S0.04
C 0.03
g 0.02
0 0.01
0


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
Image Number


Figure 4-9 Variations in Crack Density of Sample 95 B OOP HT- 1

As it can be seen in Table 4-2, Crack density of odd number images were 0.027869

nm/gm2 with standard deviation of 0.007721, while crack density of even number images

were 0.026925 gm/gm2 with standard deviation of 0.007676. After comparing it with

average crack density of all 16 images of 0.027397 nm/gm2, it can be seen that crack

density is averaged almost equally.

Table 4-6 Comparison of Odd and Even Number Images

Crack Density of Odd Number Image Crack Density Even Number Images
No. (nm/nm2) (mnm/nm2)
1 0.041729 0.042126
2 0.037357 0.020666
3 0.023050 0.021063
4 0.019076 0.021461
5 0.023050 0.021858
6 0.028217 0.025832
7 0.024640 0.029806
8 0.025832 0.032588
Avg. 0.027869 0.026925
SD 0.007751 0.007576



Coefficient of variation being less than 15% proves that results are concentrated within

15% of the mean. Thus average crack density is averaged equally all across the sample.










Sample 95BOOP High Temperature Sample 2


Image 95 B 00 HT -2 (15)


Image 95 B 00 P -HT 2 (15)


Image 95 B 00 HT -2 (7)


Image 95 B 00 P -HT 2 (16)


Figure 4-10 Representative Images of Mix 95 B 00 P HT-2 before Image Processing

The above four images are representative images from mix 95 B 00 P cured at

high temperature, taken by Hitachi S 3000 N variable pressure scanning electron

microscope. A 4x4 matrix made by conductive copper tape was used to take 16

representative images across the sample.


I




























WD17.0mB 10.OWkV 250 200Dum


Image 95 B 00 P -HT 2 (4)


Un


Image 95 B 00 P -HT 2 (10)


Image 95 B 00 P -HT 2


Image 95 B


UOL.1m HTU.U2V xzU zUum


00 P -HT 2


Figure 4-11 Representative Images of Mix 95 B 00 P HT-2 after Image Processing

All the images were processed in a java based free source ImageJ software,


developed by National Institute of Health, for image processing. Manual thresholding


was done till the perfect contrast between regions of interest with background was


developed. Images were then processed to distinguish voids from cracks. As cracks have


higher length to width ratio, shape factor of 3.5 was used to filter out voids from the


cracks. After that, cracks lengths were manually mapped on the image.


50P 15-oct-04


50Pa 15-OGt-04


W17.2- 10O.kV .250 20Dum


buFa i5-ut-UG









Table 4-7 Crack Quantification Analysis for Mix 95 B 00 P HT-2


No. Image Scale Cracks Cracks Image Field Crack
length Density
(cm) (m) (Gm/4m2)
1 HT 2-1 200Cim/11 cm 58 cm 1054.545 183 rm x 250 rm 0.023050
2 HT 2-2 200rm/11 cm 98 cm 1781.818 183 pm x 250 im 0.038947
3 HT 2-3 200nm/11 cm 88 cm 1600.000 183 mr x 250 rm 0.034973
4 HT 2-4 200m/11 cm 64 cm 1163.636 183 mx250 m 0.025435
5 HT 2-5 200nm/11 cm 57 cm 1036.364 183 mr x 250 mr 0.022653
6 HT 2-6 200m/11 cm 69 cm 1254.545 183 pm x 250 im 0.027422
7 HT 2-7 200nm/11 cm 57 cm 1036.364 183 mr x 250 mr 0.022653
8 HT 2-8 200(m/11 cm 54 cm 981.8182 183 mx250 m 0.021461
9 HT 2-9 200rm/11 cm 75 cm 1363.636 183 rmx250 pm 0.029806
10 HT 2-10 200nm/11 cm 65 cm 1181.818 183 mr x 250 mr 0.025832
11 HT 2-11 200rm/11 cm 70 cm 1272.727 183 pm x 250 pm 0.027819
12 HT 2-12 200nm/11 cm 62 cm 1127.273 183 mr x 250 mr 0.024640
13 HT 2-13 200rm/11 cm 65 cm 1181.818 183 mx250 m 0.025832
14 HT 2-14 200rm/11 cm 76 cm 1381.818 183 mx250 m 0.030204
15 HT 2-15 200nm/11 cm 95 cm 1727.273 183 mr x 250 mr 0.037755
16 HT 2-16 200rm/11 cm 61 cm 1109.091 183 pm x 250 pm 0.024242
Avg. 1265.909 0.027670
SD. 238.8743 0.005393

Data analysis from Table 4-8 shows that for 95 B OOP HT-2 samples average

lengths of cracks was 1265.909 rim. Maximum length of cracks in an image was found to

be 1781.818 rim, while minimum length of cracks in an image was 981.8182 rim, giving

standard deviation of 238.8743. As it can be seen in Table 4.5, average crack density in

an image was 0.027670 [im/pm2, maximum density of cracks was 0.038947 im/m2 and

minimum density of cracks was 0.021461 rm/pm2, giving standard deviation of

0.005393. Figure 4-12 shows variations in microcrack density in the sample made from

high temperature cured concrete mix 95 B 00 P, i.e., concrete made with cement B,

(manufactured by Cemex AASHTO type II, heat of hydration 78.2 cal/g at 7 days),

placing temperature of 950F, and without any pozzlonic contents added to mix. The


higher temperature reached was 1800F.











95 B 00 P High Temprature

0.045
0.04
>, 0.035
*3 0.03
0.025 -
0.02
8 0.015
6 0.01 -
0.005

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16
Number of Images

Figure 4-12 Variations in Crack Density of Sample HT 2

Figure 4-12 shows maximum value of crack density was 0.3894 with lowest crack

density appeared to be 0.214.

Table 4.8 shows crack density of odd number images from the sample having crack

density of 0.028068 gm/gm2, while crack density of even number images from samples

was 0.027273 gm/gm2. From Statistical analysis of image patterns and coefficient of

ratio within 20% suggests that, average image density is present throughout the images

and given sample size is true representative in nature. Similar results were found with

random or diagonal image patterns with same images.

Table 4-8 Comparison of Odd and Even Number Images

Crack Density of Odd Number Image Crack Density Even Number Images
No. m/m2 m/m2

1 0.023050 0.038947
2 0.034973 0.025435
3 0.022653 0.027422
4 0.022653 0.021461
5 0.029806 0.025832
6 0.027819 0.024640
7 0.025832 0.030204
8 0.037755 0.024242
Avg. 0.028068 0.027273
SD 0.005775 0.005349









Comparative Analysis of Samples

To analyze the differences between both samples, which are cured at room

temperature, a comparative analysis was performed.

Table 4-9 Comparative Analysis of 95 B OOP RT1 and 95BOOP RT2

Sample Avg. Crack Length (jim) Avg. Crack Density (inm/jim2)

RT1 1285.227 0.028092

RT2 1317.045 0.028788

Average 1301.136 0.02844

% Change 2.48% 2.48%


The results shown in Table 4-9 show that the difference between average crack

lengths of samples cured at room temperatures was 32 inm/jm2; the difference between

average crack densities was 0.0006. The differences between average crack lengths in

high temperature cured samples were 12.5 jim/jim2 and difference in their crack density

were 0.0002.

Table 4-10 Comparative Analysis of 95 B OOP HT1 and 95BOOP HT2

Sample Avg. Crack Length (jim) Avg. Crack Density (inm/inm2)


HT1 1253.409 0.027397

HT2 1265.909 0.027670

Average 1259.659 0.027535

% Change 0.957% 0.99%









Summary

Four samples were taken from concrete mix 95 B 00 P, two of which were cured at

room temperature and the two remaining samples were cured at high temperature. After

sample preparation, the samples were viewed under Hitachi S 3000 N, variable pressure

scanning electron microscope. Each sample was coated with copper conductive tape as

4x4 matrixes which divided the samples in sixteen equal parts. Thus, a total of 16 images

were taken, each one from the center of the parts. The images were later analyzed with

ImageJ, image processing software. Segmentation of images was done by manually

thresholding the image till the best contrast developed between background and points of

interests. After thresholding, cracks were manually mapped for quantification of

microcracks in sample.

The results show that the average crack density of concrete sample cured at room

temperature was 0.02844 mm/mm2, while crack density of sample cured at elevated

temperature was 0.027535 mm/mm2. These results are in accordance with research study

conducted by Ammouche et al., (2000) in which average crack density of undamaged

samples was found to be 0.00271 mm/mm2. From the above analysis, it could be

concluded that room temperature cured sample had approximately 3% more microcracks

than the sample cured at high temperature. Thus it could be said that there is no or very

small difference between the quantity of microcracks found in concrete cured at room

temperature and concrete cured at elevated temperature for concrete samples made

without pozzlonic materials.














CHAPTER 5
SUMMARY, RESULTS AND RECOMMENDATIONS

Summary

The purpose of this research was to identify if any additional microcracks are

developed due to concrete cured at elevated temperature. To accomplish that following

tasks were performed.

* Identification of the samples which had highest curing temperature and therefore
may develop additional microcracking.

* Performed a literature review on methods available for diction of microcracks and
finding out their advantages and disadvantages.

* Prepared specimens in order to view them under Scanning Electron Microscope for
investigating microcracks.

* Processed images to quantify and compare the intensity of microcracks in both
samples

Samples were selected out of 20 different concrete mixes made with different

percentages of pozzlonic materials, two AASHTO type II cement with two different

placing temperatures. Phase I of this project showed that, concrete made from a cement

with higher heat of hydration and without adding any pozzlonic material generate highest

curing temperature. Therefore, a mix placed at 950F and made with AASTO type II

cement having heat of hydration of 78.2 cal/g at 7 days without any pozzlonic material

was selected for this study.

Selection of appropriate sample preparation method was one of the important issues

as improper sample preparation would cause additional microcracks in sample and lead to

erroneous results. Most of the sample preparation methods used in past involved use of









preheating the specimen for removal of pore water or putting specimen under vacuum for

impregnation. There is a concern that by use of heat or vacuum, samples might develop

additional microcracks. Epoxy impregnation method by Struble and Stutzman (1989) was

used in this study as this method uses counter diffusion process to impregnate samples.

Apart from supporting the microstructure of specimen, Epoxy impregnation also provides

better contrast in cementitious matrix. Longer sample preparation time is the only

drawback for performing this experiment.

Images were acquired by a Hitachi S 3000 N, variable pressure scanning electron

microscope (VP-SEM) in backscattered scattered electron (BSE) mode which is a high

energy beam capable of distinguishing the particles present in matrix by variation of

brightness in images. The specimens were divided as 4x4 matrixes by a copper

conductive tape to aid viewing of points of interest from predefined positions.

Total of 16 images were taken from each sample and analyzed in ImageJ software

for quantification of microcracks. Images were converted to binary mode by manual

thresholding operation. When features of interest (microcracks and voids) were correctly

highlighted, the most distinct contrast between microcracks/voids and concrete

background was obtained. After getting binary image, next task performed was to

distinguish between voids and microcracks. As its known that microcracks are long

elongated structure, thus length to width (also known as aspect/ form/ shape) ratio was

applied. Objects having shape factor of 3.5 or greater was termed as microcracks and rest

other objects were termed as voids. A software plugin Shape descriptor from ImageJ was

used to identify voids in the images. Shape descriptor takes minimum and maximum

pixel values and plots the void concentration in a separate image. After completing void









removal from image, microcracks are mapped manually and counted. Results of

quantification analysis were later converted into crack density per area to give

comparative analysis.

Results and Conclusion

Results from this studies showed that, crack density of concrete sample cured at

room temperature was 0.02844 while, crack density of sample cured at elevated

temperature was 0.027535.

Table 5-1 Comparative Analysis of Samples
Avg. Crack Density
Sample Avg. Crack Length(rm) (&m m2)

Room Temperature Cured samples 1301.136 0.028440

High Temperature Cured samples 1259.659 0.027535

% Change 3.3% 3.28%



This analysis indicated that there is very small difference between quantity of

microcracks found in concrete cured at room temperature or concrete cured at elevated

temperature for samples made with no pozzlonic materials and placed at higher

temperature. Thus it can be concluded that high temperature curing conditions upto

1800F may not play significant part in formation of additional cracks.

Recommendations

From research of this study, the following recommendations were made for

conducting future work in this area.

* In future, studies involving mixes made with different percentage of pozzlonic
materials, water/cement ratios, aggregates and placing temperature should be
investigated to find out if mix design changes have any impact on density of
microcracks in concrete samples.






58


* Epoxy solution used in this study had to be cured at 1310F according to
manufacturer's instruction. Curing of the sample at high temperature might alter
the microstructure of sample, thus efforts should be directed toward finding an ultra
low viscosity epoxy solution, which does not require heating during its curing
process.

* As cutting operations would create damaged zone in concrete microstructure upto
30% of blade thickness in sample, proper grinding and polishing operations should
be followed to expose an undamaged surface.

* A more intrinsic grid system of conductive tape should be created to acquire more
images from concrete sample which can be useful if automated image processing is
used.













LIST OF REFERENCES


American Concrete Institute (ACI) Committee 116, Report SP-19, Farmington Hills,
Michigan, 1995

Advances in Materials Problem solving with Electron Microscope Proceedings of
Material Research Society, Volume 589, 1999

Ammouche, A. Breysse, D. Hornain, H. Didry, 0., and Marchand, J., "New image
analysis technique for the quantitative assessment of microcracks in cement-based
material," Cement and Concrete Research Volume 30, No.1, Page 25-35, 2000

Bisschop, J., and Van Mier, J.G.M., "How to study drying shrinkage microcracking in
cement-based materials using optical and scanning electron microscopy," Cement
and Concrete Research Volume 32, No. 2, Page 279-287, February 2002

Florida Department of Transportation, Structural Design Guide, Section 3.9 Mass
Concrete, Tallahassee, Florida

Gran, H. C., "Fluorescent liquid replacement technique: A means of crack detection and
water-binder ratio determination in high-strength concretes." Cement and Concrete
Research Volume 25, No 5, Page 1063-1074, 1995

Hanke, L. Variable Pressure Scanning Electron Microscopy for Nonconductive and
Volatile Samples, Handbook of Analytical Methods, 1999

Hornain H., Marchand J., Ammouche A., Commene J.P., and Moranville M.,
"Microscopic observation of cracks in concrete-a new sample preparation technique
using dye impregnation." Cement and Concrete Research, Volume 26, No.4 Page
573-583, 1996

Kjellsen, K. 0., and Jennings, H M., "Observations of microcracking in cement paste
upon drying and rewetting by environmental scanning electron microscopy,"
Advanced Cement Based Materials Volume 3, No.1, Page 14-19, January 1996

Knab L.I., Walker H.N., Clifton J.R., and Fuller E.R., "Fluorescent thin sections to
observe the fracture zone in mortar," Cement and Concrete Research, Volume 14,
No. 3, Page 339-344, 1984

Maggenti R., Mass Concrete Report, San Francisco-Oakland Bay Bridge East Spans
Safety Project. 04-Ala/SF-80-Var, E.A. 04-012021, Available online at
http://www.dot.ca.gov/hq/esc/tollbridge/SFOBB/EastSpan/012024/MaterialsHando
ut/cd2/concretl.pdf, last seen, Oct 12, 2004









Marusin S., "Sample preparation-the key to ESM studied of failed concrete," Cement and
Concrete. Composites Volume 17, No.4, Page 311-318, 1995

Najjar, W.S., Aderhold, H.C., and Hover, K.C., "Application of neutron radiography to
the study of microcracking in concrete Cement", Concrete and Aggregates, Volume
8, No. 2, Page 103-109, 1986

Oberholster, R. E., Maree, H. and Brand, J. H. B, "Cracked prestressed concrete railway
sleepers: alkali silica reaction or delayed ettringite formation," Proceedings of the
9th International Conference on Alkali-Aggregate Reaction in Concrete, London,
Page 740-749, 1992

Ollivier, J.P., Cement and Concrete Research, Volume 15, No.6, Page 1055-1060, 1985

Patel, H.H., Bland, C.H., and Poole, A.B., "Microstructure of concrete cured at elevated
temperatures," Cement and Concrete Research Volume 25 No. 3 Page 485-490,
April 1995

Perkes, P., ACEPT W group, Department of Physics and Astronomy, Arizona State
University, http://acept.la.asu.edu/PiN/rdg/elmicr/elmicr.shtml Last Visited
October 2004

Ringot, E., and Bascoul, A., "About the analysis of microcracking in concrete," Cement
and Concrete Composites Volume 23, No. 2-3, Page 261-266, April 2001

Samaha, H.R., and Hover, K.C., "Influence of microcracking on the mass transport
properties of concrete," ACI materials Journal, Volume 89, No.4, Page 416-424,
1992

Slate, F.O., and Hover, K.C., Fracture Mechanics of Concrete: Edited by F.H. Whitmann,
Published by Elsevier Science Page 85-93, 1984

Slate F.O., and Olsefski, S., "X-Rays for Study of Internal Structure and Microcracking
of Concrete," Journal of American construction Institute, Volume 60, No. 5, Page
575-588, 1963

Soroushian, P., Elzafraney, M., and Nossoni, A., "Specimen preparation and image
processing and analysis techniques for automated quantification of concrete
microcracks and voids," Cement and Concrete Research Volume 33, No. 12, Page
1949-1962, December 2003

Saint John, D., Poole A., and Sims A, Concrete Petrography-A Handbook of
Investigative Techniques Published by Arnold, 1998

Struble L. J., and Stutzman P.E., "Epoxy impregnation of hardened cement for
microstructural characterization," Journal of Material Science Letters, Volume 8,
Page 632-634, 1989






61


Stutzman P, Scanning Electron Microscopy in Concrete Petrography, Materials Science
of Concrete Special Volume: Calcium Hydroxide in Concrete (Workshop on the
Role of Calcium Hydroxide in Concrete). Proceedings of The American Ceramic
Society, edited by J. Skalny, J. Gebauer and I. Odler, November 1-3, 2000, Anna
Maria Island, Florida, Page 59-72, 2001.















BIOGRAPHICAL SKETCH

Vishal Shah was born in Surendranagar, India, on March 21st, 1980, to Sanjay and

Rekha Shah and grew up in Mumbai, located in Western India. He graduated in April of

1997 from Mithibai College and joined the Maharashtra Institute of Technology,

University of Pune, to complete his undergraduate degree in civil engineering. After

graduating with first class in 2002, Vishal continued his graduate studies at the

M.E.Rinker Sr. School of Building Construction, University of Florida. During his stay at

the University of Florida, Vishal worked as a graduate assistant for Dr. Abdol Chini.

Upon his graduation in December 2004, Vishal plans to pursue a career in the exciting

and challenging field of construction.




Full Text

PAGE 1

DETECTION OF MICROCRACKS IN CONCRETE CURED AT ELEVATED TEMPERATURE By VISHAL SANJAY SHAH A THESIS PRESENTED TO THE GRADUATE SCHOOL OF THE UNIVERSITY OF FLOR IDA IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF MASTER OF SCIENCE IN BUILDING CONSTRUCTION UNIVERSITY OF FLORIDA 2004

PAGE 2

Copyright 2004 By Vishal Sanjay Shah

PAGE 3

The thesis is dedicated to my loving parent s Sanjay and Rekha and to my best friends Reshma, Akshay, Piyush and Anushree fo r their unconditional support and caring throughout my academic endeavors. It is with the support and love of my family and friends that I am able to reach to my goals.

PAGE 4

ACKNOWLEDGMENTS The author would like to thank all of the members of his supervisory committee for their help and ideas throughout this effort especially Dr. Abdol Chini, committee chair, for his advice, patience and dedication as my advisor and also for the financial support provided by him. The author would also like to express gratitude to Dr. Larry Muszynski and Dr. Andrew Boyd for their support and contribution as members of my advisory committee. Acknowledgements are also owed to Tanya Reidhammer for her help in explaining and conducting scanning electron microscopy. In addition, I would like to thank Richard DeLorenzo for his help in the experimental process. My thanks also go to everyone within M.E.Rinker Sr. School of Building Construction for their continued help during the last two years. The author would like to thank his friends at the University of Florida for their support and help they offered during the research and writing of this thesis. iv

PAGE 5

TABLE OF CONTENTS page ACKNOWLEDGMENTS .................................................................................................iv LIST OF TABLES ............................................................................................................vii LIST OF FIGURES .........................................................................................................viii ABSTRACT .........................................................................................................................x CHAPTER 1 INTRODUCTION............................................................................................................1 Background...................................................................................................................1 Curing of Concrete.......................................................................................................2 Heat Generation in Mass Concrete...............................................................................3 Objective and Scope.....................................................................................................5 2 LITERATURE REVIEW.................................................................................................7 Overview of Sample Preparation Techniques..............................................................8 Radiography and Acoustic Techniques.................................................................9 Replica Technique...............................................................................................10 Impregnation Techniques....................................................................................10 Dye Impregnation Methods..........................................................................11 Epoxy Impregnation Techniques.................................................................11 Thin Samples................................................................................................13 Impregnation by Woods Metal...................................................................14 High Pressure Epoxy Impregnation Technique...........................................15 Comparison of Sample Preparation Techniques.........................................................16 Microscopic Instruments............................................................................................18 Optical Microscope......................................................................................18 Scanning Electron Microscope.....................................................................19 Variable Pressure Scanning Electron Microscope.......................................21 Image Processing Techniques.....................................................................................23 Summary.....................................................................................................................24 v

PAGE 6

3. RESEARCH METHODOLOGY...................................................................................25 Introduction.................................................................................................................25 Sample Selection........................................................................................................25 Specimen Preparation Method....................................................................................27 Procedure for Sample Preparation..............................................................................28 Preparation for Image Processing...............................................................................32 Image Analysis...........................................................................................................33 Summary.....................................................................................................................34 4. IMAGE PROCESSING AND ANALYSIS..................................................................35 Introduction.................................................................................................................35 Image Processing Technique......................................................................................36 Sample 95B00P Room Temperature Sample 1................................................37 Sample 95B00P Room Temperature Sample 2................................................41 Sample 95B00P High Temperature Sample 1..................................................45 Sample 95B00P High Temperature Sample 2..................................................49 Comparative Analysis of Samples..............................................................................53 Summary.....................................................................................................................54 5 SUMMARY, RESULTS AND RECOMMENDATIONS.............................................55 Summary.....................................................................................................................55 Results and Conclusion...............................................................................................57 Recommendations.......................................................................................................57 LIST OF REFERENCES...................................................................................................59 BIOGRAPHICAL SKETCH.............................................................................................62 vi

PAGE 7

LIST OF TABLES Table page 3-1 Mix Design Data..........................................................................................................26 3-2 Adiabatic Temperature Rise Data................................................................................27 3-3 Mixing Schedule for Epoxy.........................................................................................30 4-1 Crack Quantification Analysis for Mix 95 B 00 P RT-1.............................................39 4-2 Comparison of Odd and Even Number Images...........................................................40 4-3 Crack Quantification Analysis for Mix 95 B 00 P RT-2.............................................43 4-4 Comparison of Odd and Even Number Images...........................................................44 4-5 Crack Quantification Analysis for Mix 95 B 00 P HT-1.............................................47 4-6 Comparison of Odd and Even Number Images...........................................................48 4-7 Crack Quantification Analysis for Mix 95 B 00 P HT-2.............................................51 4-8 Comparison of Odd and Even Number Images...........................................................52 4-9 Comparative Analysis of 95 B 00P RT1 and 95B00P RT2........................................53 4-10 Comparative Analysis of 95 B 00P HT1 and 95B00P HT2......................................53 5-1 Comparative Analysis of Samples...............................................................................57 vii

PAGE 8

LIST OF FIGURES Figure page 2-1 Working of Scanning Electron Microscope................................................................19 3-1 Diamond Wafer Saw...................................................................................................29 3-2 Ultra Low Epoxy Kit by SPI.......................................................................................30 3-3 Metaserv Grinder and Polisher....................................................................................31 3-4 Polished Sample..........................................................................................................31 3-5 Diamond Suspension Paste..........................................................................................32 3-6 Prepared Samples........................................................................................................32 3-7 Specimen with Copper Tape........................................................................................33 3-8 Specimens with 4x4 matrixes......................................................................................33 4-1 Representative Images of Mix 95 B 00 P RT-1 before Image Processing..................37 4-2 Representative Images of Mix 95 B 00 P RT-1 after Image Processing.....................38 4-3 Variations in Crack Density of 95 B 00 P RT1...........................................................40 4-4 Representative Images of Mix 95 B 00 P RT-2 before Image Processing..................41 4-5 Representative Images of Mix 95 B 00 P RT-2 after Image Processing.....................42 4-6 Variations in Crack Density of Sample 95 B 00 P RT2...........................................44 4-7 Representative Images of Mix 95 B 00 P HT-1 before Image Processing..................45 4-8 Representative Images of Mix 95 B 00 P HT-1 after Image Processing.....................46 4-9 Variations in Crack Density of Sample 95 B 00P HT1.............................................48 4-10 Representative Images of Mix 95 B 00 P HT-2 before Image Processing................49 4-11 Representative Images of Mix 95 B 00 P HT-2 after Image Processing...................50 viii

PAGE 9

4-12 Variations in Crack Density of Sample HT 2............................................................52 ix

PAGE 10

Abstract of Thesis Presented to the Graduate School of the University of Florida in Partial Fulfillment of the Requirements for Master of Science in Building Construction DETECTION OF MICROCRACKS IN CONCRETE CURED AT ELEVATED TEMPERATURE By Vishal Sanjay Shah December 2004 Chair: Abdol Chini Cochair: Larry Muszynski Major Department: Building Construction Proper curing is a necessity for durable concrete. Exothermic reaction during curing process generates heat in concrete creating thermal differential between outer and inner core of concrete. In mass concrete pores at Florida Department of Transportation, internal temperatures of concrete have been reported as much as 180F to 200F, which leads to a concern that higher curing temperature might initiate higher number of microcracks during its early age. Phase I of this study indicated the presence of higher curing temperature in concrete which is made from cement having higher heat of hydration, higher placing temperature and does not contain any pozzlonic material in mix design. For comparison of curing conditions, four samples were selected from this type of concrete for quantification of microcracks. Out of these four samples, two of the samples were cured at room temperature and the other two samples were cured at high temperature. x

PAGE 11

A literature review revealed that various specimen preparation methods have been used in past for viewing concrete microstructure. However, a counter diffusion method with epoxy impregnation was used in this study as this method does not require any pre heating of specimen and pressure or vacuum in sample preparation. After specimen preparation, samples were viewed under variable pressure scanning electron microscope, and a total sixteen images from each sample were taken. Image analysis was done with standard image processing software by converting images into binary mode by thresholding it until features of interest (microcracks and voids) were correctly highlighted. After differentiating voids from cracks by shape analysis method, microcracks were manually mapped and quantified by calculating crack density in each image. A statistical analysis was performed to indicate representative nature of image quantity taken in sample. The results showed no significant difference between the number of microcracks present in samples cured at room temperature and high temperature in concrete made by cement having heat of hydration up to 76.2 cal/g, placing temperature of 95F and without any pozzlonic material. Thus it could be suggested that high temperature curing upto 180F does not play any significant role in the development of microcracks. xi

PAGE 12

CHAPTER 1 INTRODUCTION Background Concrete is one of the oldest engineering material and certainly one of the key integrants in development of modern infrastructure. When properly designed and carefully produced with good quality control, concrete is inherently a durable material. However, under adverse environment and exposure conditions, concrete interacts with its surrounding environment and undergoes chemical changes. These processes involve movement of water or other gaseous substances transporting aggressive agents through pore structures of concrete, thus making concrete potentially vulnerable to deleterious attacks. Reinforced cement concrete structures exposed to hostile environment require highly durable concrete to provide long lasting performance with minimal maintenance. Modern day infrastructure has necessitated construction of numerous bridges, dams, tunnels and concrete pavements. Massive concrete structures of this nature would contain large amount of concrete, most of these structures could be termed as mass concrete work. According to American Concrete Institute Committee 116, report SP-19 (1985), mass concrete is any volume of concrete with dimensions large enough to require that measures be taken to cope with generation of heat of hydration of the cement and attendant volume change to minimize cracking. The Florida Department of Transportation (2004) further specified dimensions of mass concrete as when the 1

PAGE 13

2 minimum dimension of the concrete exceeds 36-inches and the ratio of volume of concrete to the surface area is greater than 12-inches, provide for mass concrete. Curing of Concrete Curing is widely perceived as being an important factor in achieving durable concrete structure. Curing ensures that concrete maintains a sufficient amount of water to reach a proper degree of hydration in the surface which is required to continue chemical reaction between cement and its constituents. Proper Curing allows hydration of cement to continue which is expected to reduce capillary porosity, thereby strengthening the concrete to desired strength and increasing its resistance to penetration by aggressive gases and liquids. The higher the curing temperature, the faster is the reaction between cement and water, thus shorter setting time. It can reduce durability and useful life span of structure by hardening concrete faster, reducing strength and causing excessive seepage. Curing measures are necessary to maintain a satisfactory moisture and temperature condition in concrete, because internal temperature and moisture directly influence early and ultimate concrete properties. During curing process, concrete element undergoes exothermic chemical reaction. This heat generation is of little consequences and can be easily dissipated in surrounding environments. However, in mass concrete, heat generated within the body of member does not get dissipated easily. This temperature rise causes expansion while the concrete is hardening. Temperature in inner core of concrete is much greater than outer core of concrete due to loss of heat to surroundings by exposed surfaces. If the temperature rise is significantly high and concrete undergoes non-uniform or rapid cooling, stresses due to thermal contraction in conjunction with structural restraint can result in cracking before or after the concrete eventually cools to surrounding temperature. Due to faster cooling of

PAGE 14

3 outer surface, concrete tends to shrink and causes tensile stress in outer core. When these tensile stresses are greater than tensile strength of concrete, cracks are developed. Heat Generation in Mass Concrete Heat generated in mass concrete increases temperature of concrete to a varying degree depending on surrounding temperature, size of structural element mix design configurations, and curing temperature. Current steps to counter high temperature rise in concrete includes choosing a low heat cement, external curing system, and cooling of aggregates prior to mixing. Constructing large structures in small stages, or lifts, can be done, but it is slow and the finished structure is weakened at the cold joints in the concrete. Replacement of cement by pozzlonic materials results in a decreased temperature rise in fresh concrete, which is a particularly important issue in mass concrete where a large temperature rise can lead to cracking. Currently, any mass concrete pour for FDOT requires thermal control plan which describes the measures and procedures intended to maintain temperature differential of 35F (20C) or less between interior and exterior portion of designated mass concrete element during its curing phase; there is however, no requirement on the maximum temperature rise within the mass concrete. Internal temperatures of 180F to 200F have been reported in mass concrete pours in FDOT projects. There is a concern that such higher curing temperature would initiate microcracks during its early age which would accelerate deterioration and compromise durability of concrete by microcracks acting as a conduit of harmful minerals and gases, causing reduction in durability of concrete. Study conducted by Maggenti (2001) of California Department of Transportation indicated no irregular, externally visible cracks in mass concrete work. However, it would be important to find out changes occurring at the micro level, which will obviously

PAGE 15

4 grow in future and hamper durability of concrete. The capillary effect of the microcracks increases probability of exchange of substances between the environments and microstructure of concrete. This exchange of substances may lead to transport, solution and enrichment processes in concrete microstructure which can damage the concrete by increasing the volume. Concrete, when fully hardened, contains a variety of micro structural features on a variety of scales. To the unaided eye, it appears from a polished section to be essentially a two phase composite-a matrix of hardened mortar plus coarse aggregate that occupies about 75 % of its volume. Recent studies of concrete microstructure have played important part in understanding behavior of concrete matrix. It is necessary to study the effect of elevated temperature rise in microstructure specially comparing it with normal room temperature cured samples and find out if any additional microcracks have been developed. Formation of microcracks in the surface can lead to decrease of durability, especially in exposed outdoor structures as bridge superstructures but also concrete pavement and structures in direct contact with the ground. Many methods ranging from destructive to non-destructive have been used in past to understand microstructure of concrete. A problem common with many of the techniques used to examine the microstructure of concrete is that they can alter the sample. For example, many techniques require that the specimen be prepared by mechanical polishing or to dry concrete specimen before examination by putting in vacuum or heat. Such Specimen preparation methods would invariably alter the original artifacts and induce additional microcracks.

PAGE 16

5 Advancement in science has assisted in developing equipments and techniques capable of studying the microstructure which facilitate in-depth study and analysis of specimen. While many existing techniques and equipments require drying the specimen and in a way altering the specimen, evolution of Variable Pressure Scanning Electron Microscope (VPSEM) techniques allows the study of unaltered wet specimen in a low vacuum. Objective and Scope This study is an extension of a research project Adiabatic Temperature Rise in Mass Concrete which is aimed to develop adiabatic temperature rise curves for concrete mixes used in FDOT projects and investigate properties of high temperature cured concrete. Total of 20 mixes of concrete were made as a part of the study by changing various parameters such as the source of cement, the percentage of pozzlonic materials and the initial placing temperature. Concrete samples were kept in heating curing chambers to simulate conditions of mass concrete for monitoring rise in temperature. At 14 and 28 days, samples kept in heating chambers as well as samples in normal room temperature were tested for compressive strength. The objective of this part of the study is to focus on detecting and quantifying additional microcracks developed due to concrete cured at elevated temperature. This would be done by comparing room temperature cured samples with samples cured at elevated temperature to find out if there is any difference in quantity of microcracks. The Study would be performed as given below: 1. A literature search to find out current methods used in identification of microcracks and sample preparation techniques. It should also include possible image processing techniques to enhance analysis capability.

PAGE 17

6 2. Prepare concrete specimens according to an identified sample preparation technique and view them with selected Microscopy equipment. 3. Use Image Processing Techniques for further identification and quantification of microcracks. Compare the quantity of microcracks in samples cured at elevated temperature with those cured at room temperature to determine if any additional microcracks have been developed due to high temperature curing.

PAGE 18

CHAPTER 2 LITERATURE REVIEW Presence of voids, cracks, and other defects play an important role in determining the mechanical performance of the concrete. If concrete surface is pre-damaged by microcrack formation in the near-surface area and exposed to open weathering, capillary effect would increase the transfer of substances from the environment and increase the concrete volume which can lead to several mechanisms capable of negatively influencing durability of concrete. Studies have indicated that the development of cracks and connected crack network contributes to the increase in permeability and diffusivity of concrete. The presence of preexisting microcracks of 10 m or greater on the surface can lead to reduced durability, especially of exposed outdoor structures such as bridge superstructures, but also concrete pavements and structures in direct contact with the ground. Patel et al., (1995) found that concretes cured at higher temperatures exhibited a coarser microstructure than that of a typical concrete cured at 20C, particularly with respect to ettringite. The presence of microcracks will increase moisture mobility within the concrete, which may produce density gradients within the matrix leading to further microcracking. The domination of the matrix by a network of microcracks is also conducive to the formation of secondary ettringite. Recorded curing temperature of 180F 200F inside core of FDOT Mass concrete elements has raised concerns over the initiation of microcracks. Slate and Hover (1984) defined microcracks as cracks having width of less than 100 m, while Jansen defined 7

PAGE 19

8 microcracks as extended faults with a width of less than 10 m. According to Kjellsen and Jennings (1996), these differences in definition appeared to be closely related to experimental techniques and to the orientation of the studies. For the scope of this study, reasonable and practical interpretations of microcracks were set as cracks having widths less then 10 m. While macrostructure of concrete can be seen unaided, microstructure (200 m or smaller) must be observed with the aid of a microscope. It would be fair to say, as the documented research indicates, that cracks and crack propagation is understood well from macroscopic point of view but unavailability of precise specimen preparation techniques hinder studies of concrete microstructure. Overview of Sample Preparation Techniques Sample preparation is a key to microscopic analysis of concrete. Proper preparation of concrete samples so that microcracks and voids develop a distinct contrast against the body of concrete is a pre-requisite for the application of the modern day image processing and analyzing techniques (Soroushian et al., 2003). Poor preparation methods can lead to erroneous diagnoses of problems associated with concrete specimen. According to Hornain et al., (1996), ideal specimen preparation techniques should not induce any cracks during the preparation of samples, thus techniques which involves prior drying of specimen should not be used. The ideal specimen preparation should be simple, economic, rapid, and should be able to detect very fine cracks. And finally, a sample preparation technique will not be useful without accurate image analysis and processing techniques. The sample preparation would also depend on the type of equipment used for image collection, required resolution of image, and objective of the study.

PAGE 20

9 For the past few decades, studies have been performed to find out a method which fulfills all definitions for ideal specimen preparation technique. New equipments and methods developed by technological advances have helped researchers in their analysis of concrete specimens. As reported by Ringot and Bascoul (2001), two methods are necessary for the characterization of microcracks, one for sample preparation and one for quantification of cracks. Sample preparation techniques also depend on the type of microscopy used (S. Marusin 1995). Acoustic method, ultra sonic and laser sparkle methods are useful for studies related to crack propagations but they are unequipped to monitor initial state of samples crack quantification. Thus, the rest of the sample preparation techniques can be briefly classified in three main categories: radiography techniques, replica techniques, and impregnation techniques. Radiography and Acoustic Techniques Earlier Methods by Slate and Olsefski (1963) described the use of X-radiography to study the internal features of concrete and crack formation process of mortar and concrete sample. Thin samples of 4 mm thickness were exposed to X-rays flux normal to the plane of the sample. Cracks were identified from normal constituent of concrete as penetration of X-rays was greater in surrounding area. A comparative study carried out by Najjar et al., (1986) indicated that due to poor resolution of images, X-radiography systematically overlooks thin cracks. Neutron radiography described by Samaha and Hover, (1992), which is principally quite similar to X-ray radiography approach, is in fact a better alternative for increasing image resolution. However, in this process the samples need to be air dried beforehand in order to impregnate specimen with gadolinium nitrate, which in turn might induce microcracks in samples.

PAGE 21

10 Replica Technique Replica method developed by Ollivier in 1985 is one of the methods which do not require any pre drying of the sample. In this method, the film of acetylcellulose is placed on concrete specimen by methyl acetate. The film is taken off after the solvent (methyl acetate) is evaporated in air and observed under optical or scanning electron microscope. As there is no disturbance made to the specimen surface itself, crack propagation studies can be carried out with replica method. However, in order to peel the acetylcellulose film safely, the area of film has to be limited to 2 cm 2 Thus too many replicas need to be prepared for covering significant portion of specimen, which could be time consuming. Impregnation Techniques Impregnation techniques are one of the oldest techniques used to study microstructure of concrete. In this method, concrete samples are impregnated with dye or epoxy to facilitate detection and identification of cracks. In order to reduce difficulty in viewing cracks in dense microstructure, fluorescent dye was used by Knab et al., (1984). However, pre drying of specimen in an electric oven before impregnation, which could alter the specimen condition, remained a major inhibition to this process. Methods developed by Struble et al., (1989) and Gran (1995) overcome drying of specimen by using counter diffusion method for replacement of pore water by dye impregnated organic solution. In this method, a thin sample of 15 mm was cut and kept in solution made by dissolved dye and ethanol. After 4 days of counter diffusion process between pore water present in concrete specimen and dye induced ethanol, specimen was taken out of the solution to remove excess dye present on the surface by polishing. However, this process could take several days and increases the duration of the test.

PAGE 22

11 Dye Impregnation Methods Hornain et al., (1996) modified dye impregnation techniques to reduce the time taken to prepare specimen. In this Method, 4 grams of a water soluble red powder, commercially known as Irgacete was dissolved in 100 ml ethanol solution. After keeping specimens for 5 minutes in dyed solution, a second impregnation was done. After completing these 2 steps, the specimen was taken out and polished under water with 6 m diamond paste to remove excess dye. In order to achieve proper polishing, the specimen was again polished with 3m and 1 m of diamond paste. The polished specimens were observed under optical microscope at 100X. The study reported that cracks of 1 m width and less could be easily distinguishable, providing contrast level between specimen and cracks. During sample microscopic studies, cracks going through the hydrated cement phase were distinguished from discontinuity at paste-aggregate interface. However, the author reported that dye impregnation method was not useful in finding out cracks in highly porous areas. Epoxy Impregnation Techniques Struble and Stutzman (1989) developed a new technique which involved three steps procedure to replace pore water from concrete sample. Epoxy impregnation would not only support the microstructure of specimen by filling the voids and cracks up on curing, but also support fragile pores and matrix phases by restraining it against disintegration during the different stages of preparation. This could be a major issue if the samples are to be viewed under scanning electron microscope which would generate high pressure inside the chamber. Another advantage of using epoxy is to enhance contrast between the pores, hydration products and cementitious material. The selection of the type of epoxy depends on the objective of the study. Low viscous epoxy was used for relatively highly

PAGE 23

12 permeable materials or cementitious powders, while ultra low viscosity epoxy aids in rapid infiltration for less permeable cement pastes and concrete samples. The first step for this sample preparation method was to cut the specimen of 1 inch size and immerse specimen in ethanol solution at 60C. The blades of saw were immersed in propylene glycol to prevent it from drying. A companion specimen was prepared by Struble and Stutzman (1989) which was kept in an ethanol solution with red dye to see the extent of time taken for the replacement of ethanol in pore structure. For a given concrete sample, the time taken for replacement depends essentially on the thickness of the sample. After 50% of alcohol replacement in depth by visual inspection in companion specimen, the solution is replaced by epoxy at room temperature and cured according to the manufactures specifications. After proper curing, the samples were polished and lapped with abrasion papers and diamond paste in decreasing order and viewed under Backscattered Electron imaging technique of Scanning Electron Microscope. The samples should be grinded and polished perfectly in order to remove raw saw cuts on the specimen. Insufficient polishing would leave a disturbed surface on specimen and the cracks would not be easily discernible during the microscopic observations. On the other hand, an excessive polishing would contribute to the removal of microstructure, and could affect the subsequent quantification operations. It was believed that higher heating temperature would hasten the process of pore replacement. Higher temperature for heating was not used in subsequent studies due to limited measure of quantifying the pore water alcohol exchange and quick evaporation of alcohol. Modified epoxy

PAGE 24

13 impregnation method does not require any heating or vacuum on samples and would not induce any specimen preparation related microcracks. However as reported by Soroushian (2003), epoxy impregnation can highlight more porous areas of cement paste, but generally do not yield crisp boundaries and sharp contrast between microcracks and air voids against body of concrete. A new two stage technique was developed as InkEpoxy Impregnation. In the first stage, fine capillary pores were first impregnated with parker blue ink and then microcracks and voids were impregnated with fluorescent ultra low viscosity epoxy solution. Specimen were cut to 20-30 mm or 0.8-1 inch, and lapped by abrasive liquid under 3 PSI pressure After cleaning the specimen for loose debris by first water bath and then ultrasonic bath, specimen was dried at 140 F for 3-4 hours. Ink impregnation for first stage was done by keeping the container with ink and specimen under nitrogen pressure at 280 psi for 18-24 hours. Specimen was removed and heated at 140 F for 24 hours to let ink dry. Stage 2 was performed for fluorescent epoxy penetration in samples. The samples were kept in a vacuum chamber with 0.38 psi for 1 hour. Epoxy solution was mixed and poured in a glass container and kept in nitrogen pressure of 0.0133 psi for 3-4 hours. To cure epoxy, samples underwent heat in an electric oven at 150 F for 18-24 hours following by polishing operations. While comparing with normal epoxy replacement technique, ink-epoxy replacement technique gives better results. However, the application of heat as well as vacuum pressure could instigate microcracks in samples. Thin Samples A study conducted by Anderson (1989) introduced preparing thin section of concrete samples. In this method, a larger sample of 45 mm x 30 mm x 20 mm was cut from a section of concrete, selected from the portion which was few centimeter away

PAGE 25

14 from to be investigated area. The cut section was kept under ethanol for 12 hours to reduce its tendency for cracking. The concrete specimen was kept in a vacuum oven at 85 F to 90 F to dry for two hours. A homogenized mixture of low viscosity epoxy prepared with 1.1% by volume of dye was mixed in magnetic stirrer for 24 hours in advance. Dried concrete specimen was kept in the solution for 1 hour for vacuum impregnation. It was observed that concrete was impregnated by 1-2 mm in depth. Stage 2 covered mounting of glass slide on the selected face of specimen with Ultraviolet (UV) hardening glue. The specimen was then grinded by a 15-20 m diamond disc in order to get smooth surface. After cleaning and drying the specimen, it was re-impregnated with epoxy. After curing, a thin sample was cut from the specimen by diamond saw and final grinding and polishing was done. Although the results from this method are widely used in European counties, lengthy sample preparation time and dependability on skilled and experienced technician were primary requirements. Impregnation by Woods Metal Nemati (1997) developed a new approach in impregnation techniques using Woods metal instead of epoxy. The idea of this research was to preserve the microstructure of the concrete samples which are kept under compressive stress to analyze microcracks as they exist under loading. In this three-phased study, cracks were introduced in the concrete specimen by a new test setup which allowed the application of axial stress on concrete and impregnation by woods metal simultaneously. Woods metal was impregnated within the sample for providing stability and better contrast for the identification of microcracks. However, sample preparation method involved drying of concrete cylinder by 109.5F as well as gradual heating of test assembly by 122F, 167F and 204.6F for unified initiation of molten woods metal in the specimen. As explained

PAGE 26

15 earlier, heating or drying operations before impregnation could introduce microcracks in concrete. Author reported that in no loading conditions, few cracks were observed which could be present due to cracks introduced in sample preparation or drying shrinkage. Woods Metal impregnation technique was further developed by Soroshian et al., (2003) for concrete samples without initiating microcracks through stress introduction. In this method, 2 inch thick slices were cut and washed to remove any loose debris attached. The cleaned sample was kept in electrical oven at 150 F for 24 hours to remove water present. The dried sample was than kept in a steel mold for impregnation with Woods metal. To facilitate impregnation and liquidification of Woods metal, steel mold is kept inside a vacuum pressure chamber by 0.95 psi for 3040 minutes. After keeping the oven temperature for 200 F for 1-2 hour, the air vacuum was replaced by nitrogen pressure of 280-300 psi for 3-4 hours. The specimen was allowed to cool down, followed by cutting a 6 mm sample and then the surface was prepared for viewing under Backscattered technique in Environmental Scanning Electron Microscope (ESEM). Results indicated replacement by Woods Metal gave desired contract for microcrack identification against concrete surface. However, heating and vacuum pressure application on concrete would develop additional microcracks which can be quantified with original microcracks. High Pressure Epoxy Impregnation Technique Previous methods involving florescent epoxy replacement being time consuming, Chen (2002) applied high pressure of > 20 bars to florescent epoxy impregnation in concrete sample. Samples of 2x2x1.5 cm were prepared and after initial drying of 176 F for 2 days, samples were kept in a steel cylinder in epoxy resin at least 5 cm higher than the top surface of specimen. The top and bottom of steel cylinder were blocked by Teflon, desired pressure ( 145,725, 1160, 1450, 2900 and 5800 psi) was applied on the

PAGE 27

16 blocks through a hydraulic piston and held under pressure for 90 minutes followed by curing of epoxy by heating it upto 122F for 24 hours. This robust impregnation procedure results in a uniform and vast distribution procedure of the florescent epoxy resin throughout the material structure. Prepared samples were viewed under optical microscope using UV light. The experiment proved high pressure impregnation to be a more effective and faster method to induct epoxy resin in cementitious materials. This method show higher brightness and more effective penetration of epoxy in material and when pressure and vacuum impregnated samples kept in optical microscope to analyze, high pressure impregnated samples took less time for reaching a threshold value of photon. Pressure impregnation allows higher tensile and flexural strength in samples compared to vacuum process, thus implying greater stability of cementitious specimen during the preparation phase. However, no reports clearly show or address the issue of the introduction of cracks during sample impregnation. Comparison of Sample Preparation Techniques Microscopic image is a combination of original material from which the specimen was taken, cumulative effects of all the procedures required to prepare the specimen for examination, the examination technique itself and our interpretation of the image. Specimen preparation method depends on the type of microscope used for viewing the specimen. The review of all above techniques reveals that there is no standard specimen preparation for viewing microcracks in cement and cementitious materials. Almost all researchers identified pre-drying of specimen as a major source of developing cracks prior to microscopic examination. However, most of the studies dried specimens in order to remove pore water of concrete to facilitate the introduction of dye, epoxy or woods metal. While some techniques required polished concrete section or broken fracture

PAGE 28

17 surfaces, some required thin sections dried at 122 F or oven dried in NO 2 atmosphere. Some samples are cut with a diamond impregnated wire saw without a coolant, while some samples prepared by epoxy impregnated are cut by diamond grit or lapped by alumina grit and then polished by diamond grit/ abrasive liquid of decreasing order (Marusin S., 1995). Some researchers believe that application of improper sample preparation methods would erode and destroy existing microstructure. It can also introduce elements which are not part of the original specimen which would lead to false diagnosis of damage mechanism. Most of the specimen preparation techniques requires to dry sample till a certain degree which would affect the pore structure and can induce more microcracks. As reported by Hornain et al., (1996), crack density increases with intensity of drying. Thus such methods would not be able to give true quantitative analysis of microcracks. Vacuum impregnation methods results in excessive drying of the sample and therefore cracks are introduced during sample preparations (Chen 2002). Techniques involving mechanical polishing to produce either a flat surface or thin specimen could be problematic for concrete as different phases of concrete polish at different rates. Similarly putting concrete samples into a vacuum would lead to loss of water and change of structure. Polishing and lapping operation to smooth the specimen surface would lead to spreading of components over surface, thus preventing effective use of dot map studies for elements. Polishing would also destroy delayed etterndite crystals and cracks originating from it. While pressure or vacuum would introduce more cracks in the sample, original cracks would heal or widened by epoxy impregnation techniques

PAGE 29

18 (Marusin, 1995). However, sawn and unpolished approach could be better if the study is related to DEF analysis but for proper crack identification and quantification, sample preparation techniques would be required. In brief, Bisschop and Van Mier (2002) and other studies have indicated impregnation of whole sample and then cutting thin samples from fully impregnated samples would introduce less microcracks. Specimen preparation techniques involve cutting, drying, lapping, grinding and polishing. Improper handling of these operation results in induction of additional microcracks. For the scope of this study, epoxy impregnation technique developed by Struble and Stutzman (1989) was performed as it did not involve any drying of samples in any form. In this two-stage counter diffusion process, pore water is replaced by ethanol, and ethanol is substituted by epoxy without involvement of preheating or use of pressure techniques. Microscopic Instruments Optical Microscope Optical microscopy is one of the favored techniques for concrete petrography. In this instrument, image is viewed in full color and generally at a lower magnification, thus making it a more suitable technique for observing features at the millimeter scale and larger. Concrete and cementitious specimens prepared by thin-sectioning and fluorescent microscopy techniques can be viewed under optical microscope with the use of reflected ultraviolet lights. The image acquisition is done by placing a TriCCD camera which is linked to a personal computer acquisition system. However, Optical microscopes have shallow depth of field and limited resolution capacity, so a highly smooth and polished surface is required to produce a focused image under this technique. Furthermore, the

PAGE 30

19 intensity of the vacuum-impregnated sample is not stable after 2 min of exposure to UV light, and decreases in a continuous fashion with the elapse of time (Chen 2002). Scanning Electron Microscope Scanning electron microscopy/ x-ray microanalysis is comparatively more useful in studies which investigate quantification of microstructural properties such as microcracks and voids. Scanning electron microscopes has greater depth of field and high spatial resolution which produces focused images of poor specimens and analytical data for elemental composition analysis for the features seen on those images. As scanning electron microscopy and optical microscope petrography are different tools, different sample preparations are used for these different approaches. As described in figure 2.1, SEM scans a focused beam of electron across the specimen and measures any of several signals resulting from electron beam interaction with surface of concrete specimen. Images are monochrome in nature because they reflect the electron or x-ray flux resulting from beam/specimen interaction. Three major types of signals generated as a result are Secondary electrons, backscattered electrons and X-rays analysis. Figure 2-1 Working of Scanning Electron Microscope, (Perkes, 1999)

PAGE 31

20 Secondary Electrons (SE) are low energy electrons resulting from an inelastic collision of a primary beam electron with an electron of a specimen atom. Because of their low energy, they are readily absorbed and only those produced near the surface escape, resulting in an image of surface topography. SE imaging is principally applied in the examination of early age paste microstructure, high magnification imaging of microstructural features and for examining texture. Knowledge of the morphological and compositional characteristics of the hardened cement paste constituents is invaluable for their identification. As the hardened cement paste matures, filling of the void spaces eliminates the well-formed crystals and backscattered electron and x-ray imaging are more useful in examination of these microstructures (Stutzman, 2001) Backscattered Electron are high energy beams and capable of reflecting difference in atomic numbers. It can distinguish between the particles present in matrix on the basis of the variation in brightness of their images. In Backscattered Electron imaging technique contrast is generated by different phase compositions relative to their average atomic number and is observed by differential brightness in the image. SEM analysis using backscattered electron and X ray imaging requires highly polished surface for optimum imaging and X ray microanalysis. X radiation is produced when a specimen is bombarded by high energy electrons. X-ray microanalysis systems generally employ an energy dispersive detector with the other detector type being a wavelength detector. The x-ray energy level is displaced as the number of counts of each energy interval and appears as a set of peaks on a continuous background. X-ray radiation allows elements compositions to be obtained as

PAGE 32

21 the qualitative analyses prints like a dot map images. Each element is identified on a continuous spectrum by the position of its pick. Any sample can be examined by SE imaging but best use of BE imaging is for examining flat, polished surfaces (Struble and Stutzman, 1989). One of the disadvantages of Scanning Electron Microscope is that samples must be coated to allow discharge of electron build-up on the examined surface. A thin metal coating decreases build-up of negative charge by forming a conducting path for electrons to avoid distorted images. If images are accompanied by X-ray analysis, carbon coating can not be used if a carbonation process is to be studied. A gold coating gives a peak almost at a same position as sulfur (Oberhostler 1992). Thus, in some samples, it might be appropriate to coat the specimen due to concerns for altering the microstructure. High vacuum pressure inside the specimen chamber is also one of the factors which limits the use of SEM as some samples might break or disintegrate under high vacuum pressure. Variable Pressure Scanning Electron Microscope Hanke L, (1999) described that many applications where SEM/EDS evaluation could be useful involve samples that are not electrically conductive. These samples have traditionally required pretreatment, by coating with a conductive film, before SEM examination. Nonconductive samples are subject to a buildup of electrons on the examined surface. This buildup of electrons eventually causes scattering of the incoming electron beam, which interferes with imaging and analysis. Furthermore, samples that contain substantial water or other materials that volatilize in high vacuum also present challenges for SEM examination. These samples require controlled drying to allow the SEM chamber to reach high vacuum and to prevent deformation of the sample at the SEM vacuum.

PAGE 33

22 As the SEM is increasingly used for routine evaluations, there is increasing demand for examination without pretreatment. An answer to the problems of charging and volatile samples is the development of scanning electron microscopes that operate without exposing the sample to high vacuum. These microscopes are referred to alternately as environmental, low-vacuum, or variable-pressure SEM. In variable pressure SEM, the chamber at the electron gun is maintained at high vacuum, while a controlled amount of gas is allowed into the sample chamber. A fine aperture separates the gun and sample chambers to prevent excessive gas entrance into the gun chamber. Separate vacuum systems control the vacuum in the sample chamber and at the gun. The advantages of a higher pressure in the sample chamber are obvious for wet and volatile samples. The higher pressure decreases the rate of volatilization. This decreases the drying and deformation of wet samples. For nonconductive samples, the advantage of higher pressure is less obvious. When gas molecules in the sample chamber are struck by the electron beam, the gas is ionized. These positive ions are attracted to and neutralize the negative charge build-up on the nonconductive specimens. By controlling the pressure in the sample chamber, the number of gas molecules intercepting the electron beam is maintained at a level that is sufficient to prevent charging, but does not deflect the beam sufficiently to prevent imaging and microanalysis. Thus, the use of variable pressure scanning electron microscope holds advantage over traditional scanning electron microscope as no coating is required over the samples and samples with moisture could also be imaged.

PAGE 34

23 Image Processing Techniques Important steps for any microstructural images are to acquire the image and to process microstructure image to get a quantitative analysis of the sample. For this study, quantitative analysis is to identify and measure the number of cracks and crack density over number of images. The image processing should be able to differentiate between voids and cracks, and boundary cracks around aggregates or between phases. Darwin (2001) researched on quantification of microcracks of cement mortar samples by backscatter electron technique of scanning electron microscope. He reported that the use of edge detection, gradients and other filters failed to identify cracks passing through low density phases as its intensity depends on adjacent and underlying phases. In his approach, the identification of cracks was based on local changes in grey level. A line scan was taken perpendicular to the images for identification of potential cracks. Perimeter to squared area was taken to differentiate between cracks and voids. Phase identification was also carried out based on pixel intensity. He conclude that as grey level of image is affected by the density of adjacent and underlying phases, cracks identified based on change in grey level would not be correct. Thus the correct practice for crack identification should establish minimum gradient in grey level adjacent to cracks. Soroshian (2003) used segmentation of images from grey level to binary images by thresholding approach for viewing microcracks in samples. The research compared manual thresholding operations with three automated thresholding operation such as factorization, entropy and moment. In manual thresholding, low threshold level was set to zero or default for auto thresholding and best threshold level image was determined. The optimum high threshold level was determined by comparing original gray image and the image after the application of manual threshold on binary image. When features of

PAGE 35

24 interest (microcracks and voids) were correctly highlighted, the most distant contrast between microcracks/voids and concrete background was obtained. After training the program with images of different resolutions, three automated thresholding operation were performed and compared results with manual thresholding. The study reported that there was no significant difference between manual thresholding and automated thresholding operation by factorization method. There was a difference of 5% and 2% respectively, for entropy and moment methods of automated process. After identifying the cracks, form/shape factor of 3.5 was used to distinguish cracks from voids. It was shown that automated process was different by 6.39% and 3.02% from manual process. However, the author concluded that due to the ease of operation and efficiency given by automated process, this error was reasonably small. Summary Literature review of methods for preparation of concrete samples for viewing microstructure was performed. As described before, many sample preparation methods involved the use of pressure or heat, which might cause concerns over additional cracks in concrete and might lead to erroneous results. Epoxy impregnation of concrete samples was identified as a process which uses counter diffusion process as sample preparation means for this study. As optical microscope would have inadequate depth of field and resolution for microstructure, scanning electron microscopy was preferred for image acquisition, Variable pressure electron scanning microscope, which requires no special coating over samples and ability to view samples with moisture was found to be more suitable for image analysis. After acquiring images, manual thresholding approach to map microcracks was selected for quantification of microcracks in a given sampl

PAGE 36

CHAPTER 3 RESEARCH METHODOLOGY Introduction This chapter describes the methods for sample selection, specimen preparation method and image analysis techniques used in this research. As described in Chapter 2, the objective of this study was to find if the concrete cured at elevated temperature shows any additional microcracks compared to samples cured at room temperature. This investigative study was carried out in three phases. Phase one was to identify the samples which shows tendency for microcracks, while phase two was to select an appropriate sample preparation methods. Phase three comprised of sample preparation and image analysis technique to quantify and compare results. Sample Selection Samples were selected from an ongoing research project Adiabatic Temperature Rise in Mass Concrete conducted by University of Florida for Florida Department of Transportation (FDOT). In this study, twenty different mixes were prepared using AASHTO Type II cement made by two manufactures with different percentages of pozzlonic materials and two different placing temperatures of 73F and 95F were used to study adiabatic temperature rise in laboratory simulated mass concrete conditions. As higher placing temperature increases the concrete curing temperature, samples made from higher of two placing temperatures, 95F were selected for microstructural analysis. The field of investigation was further narrowed down further by selecting the mix made with cement which has higher heat of hydration. A Heat of Hydration test by Construction 25

PAGE 37

26 Technologies Ltd. (CTL) revealed higher heat of hydration in AASHTO Type II cement (Cement B) manufactured by CEMEX (78.2 cal/g at 7 days) was greater compared to AASHTO Type II cement ( Cement A) manufactured by Florida Rock (66.2 cal/g at 7 days). Thus, mixes made with Cement B and 95 F placing temperature were selected. This included five different mixes were prepared with varied pozzlonic contents. 100% Cement 20% Fly Ash + 80 % Cement 35% Fly Ash + 65 % Cement 50% Slag + 50% Cement 70% Slag + 30% Cement Table 3-1 Mix Design Data AASHTO Type II CEMEX Cement ( Cement B ) (Mix proportions per cubic yard) Mix Cement (lb) Fly Ash (lb) Slag (lb) Water (lb) Fine Aggregate (lb) Coarse Aggregate (lb) Air Entertainer (oz) Water Admixture (oz) 95B00P 760.00 0.00 0.00 279.00 1033.00 1736.00 2.30 60.86 95B20F 608.00 152.00 0.00 279.00 1033.00 1708.00 4.00 30.40 95B35F 494.00 266.00 0.00 279.00 980.00 1687.00 4.00 68.20 95B50S 380.00 0.00 380.00 279.00 1021.00 1724.00 2.30 38.06 95B70S 228.00 0.00 532.00 279.00 1021.00 1724.00 2.30 68.20 Prepared concrete samples were kept in two different types of containers. For samples to be kept at room condition, concrete was poured in 6x12 plastic containers and for samples to be kept at high temperature cured condition, specially acquired thermal curing chambers were used. Thermal curing chamber maintained temperature inside chamber in an adiabatically condition. The chambers were connected with thermo

PAGE 38

27 couples to a micro controller system for maintaining temperature and to document the temperature rise in the concrete samples. Data given in following Table 3.2 shows the first four day temperature rise in concrete specimen kept in thermal curing chambers. Table 3-2 Adiabatic Temperature Rise Data Cement B (95F Placing Temperature) Plain Cement 20% Fly Ash 35% Fly Ash 50% Slag 70% Slag Time (Day) Temp. Rise (F) Temp. Rise (F) Temp. Rise (F) Temp. Rise (F) Temp. Rise (F) 0.0 0.00 0.00 0.00 0.00 0.00 0.5 2.60 56.80 50.35 58.60 45.00 1.0 72.15 72.90 61.90 76.70 69.40 1.5 84.20 77.40 67.85 80.70 73.90 2.0 85.60 77.60 68.90 81.80 76.30 2.5 85.65 77.60 69.20 82.00 77.90 3.0 85.95 77.60 69.30 82.00 78.90 3.5 85.95 77.60 69.60 82.00 79.50 4.0 85.95 77.60 69.90 82.00 79.90 As it can be seen in Table 3-2, data indicates that sample 95 B 00 P had the highest temperature rise of 85.95 F over placing temperature of 99 F, thus effectively curing temperature in the core is 186F. Therefore, sample 95 B 00 P was selected for microscopic studies due to the following reasons. Higher placing temperature had higher curing temperature and therefore gives higher probability of presence of additional microcracks Cement B had higher heat of hydration (78.2 cal/g at 7 days) compared to Cement A (66.2 cal/g at 7 days), which might lead to higher curing temperature As indicated in Table 3.2, sample 95 B 00 P with plain cement showed higher temperature rise compared to samples with varied percentages of fly ash or blast furnace slag. Specimen Preparation Method As described in Chapter 2, many different approaches have been used in the past to view and document microcracks in cement paste or concrete. However, most of the

PAGE 39

28 methods indicate the use of heat, vacuum or pressure along the course of sample preparation or prior to impregnation, which might lead to additional microcracks due to sample preparation widely known as secondary microcracks. Samples cured at elevated temperature as well as those cured at room temperature would undergo epoxy impregnation method for Scanning Electron Microscopy as established by Struble and Stutzman (1989). In this method they used counter diffusion method to replace the pore water of concrete first by ethyl alcohol and then low viscosity epoxy. Neither of these processes requires any prior pre-heating of surface to dry, nor does it use any vacuum for impregnation of epoxy. The only drawback of this method is that sample preparation would take longer time to complete the process. Considering all methods and their advantages and disadvantages, this method was selected for sample preparation in current study. Procedure for Sample Preparation A small 1/2 thick sample was cut with diamond saw from the samples cured at room temperature and those cured in thermal curing chambers. During this operation, in order to reduce the heat generation, cooling water was circulated through water pump to diamond blade to aid cutting. Once the required thickness of the sample is obtained, other cutting operation was carried out on a diamond wafer saw machine (figure 3-1). It was used to reduce any artifacts and required polishing time in the samples during the preparation period. In this machine, blade was passed through cutting fluid (polypropylene glycol) to keep the samples from drying. After cutting by diamond wafer, average size of specimen was 2.15x 1.8. The samples then went through a two stage counter diffusion process.

PAGE 40

29 Figure 3-1 Diamond Wafer Saw In this process, first pore water of concrete is replaced by 200% Ethanol made by Sigma-Aldrich Inc., Montana, and then in the second stage, ethanol was replaced by ultra low viscosity epoxy solution made by Structure Probe Inc., Pennsylvania. For the first stage of alcohol pore water replacement stage, the specimen was placed in a lidded jar filled with ethanol (200 Proof) for the alcohol pore solution replacement stage. A companion specimen was used to determine the depth of replacement of pore water by alcohol. This companion specimen was a remnant from the specimen after trimming. This specimen was now placed in a jar filled with ethanol dyed by a deep red alcohol-miscible dye by Poly science Inc. By splitting or sawing the companion specimen after a period of time, the depth of replacement is seen by the depth of dye coloration. On the average, alcoholpore water replacement was found to be 1mm/day in concrete samples. When this depth reaches to half the section thickness, the pore solution replacement by alcohol was stopped. The section is then placed in a

PAGE 41

30 container with the ultra low-viscosity epoxy (see Figure 3-2) for the same time as alcohol pore water replacement process. Curing of epoxy was performed as specified by product manual. The epoxy was made up of 4 different chemicals as described in Table. 3-3 Figure 3-2 Ultra Low Epoxy Kit by SPI Table 3-3 Mixing Schedule for Epoxy No. Name of Chemical Mixing Schedule 1. Vinylcyclohexene dioxide (VCD) 10g 2. n-Octenyl succinic anhydride (n-OSA) 20 g 3. Butanediol Diglycidyl Ether (BDE) 0.3 g 4. Dimethylaminoethanol (DMAE) 0.3 g Ethanol was replaced by epoxy at two hour intervals with the proportion of ethanol to epoxy was changed from 3:1, 1:1, 1:3 at each interval and finally 100% epoxy solutions was used for keeping it in embedding medium overnight. For curing of epoxy, samples were kept in a heating chamber at 131F for six hours. At the end of six hours, the specimens were ready for the cutting and polishing stages. Grinding and polishing are important steps in sample preparation. Grinding is used to expose new surface layer and to remove excessive epoxy coating on the surface. However, excessive grinding would remove the depth of epoxy layer and could damage microstructure. Grinding was done on a Metaserv 2000 Grinder Polisher Machine (see

PAGE 42

31 figure 3-3 and 3-4) on 150 RPM. Abrasive papers of 180,240, 320,400 and 600 were used for the removal of material from specimen surface. Approximately 2 minute of time was given to each grinding cycle. Figure 3-3 Metaserv Grinder and Polisher Figure 3-4 Polished Sample Polishing is done for undoing the damage done due to sawing and grinding of surface. After grinding by 600 grits, surface becomes smooth and ready for further polishing. Polishing operation involves successively decreasing size of diamond suspension paste from 6m to 3m, 1m and 0.25m manufactured by Buhler (see figure 3-5). A polishing cloth was attached to rotating wheel of Metaserv 2000 Grinder Polisher. This method helps in reducing the damage done by previous suspension by removal of finer scratches and improving the surface for microscopic study. At the end of this operation, the sample is wiped with a plain cloth and kept in glass jar (See figure 3-6). During polishing process, it is important to understand that cutting by diamond saws would damage the samples up to 30% of its thickness. So polishing should be done till 30% of blade thickness in order to remove all artifacts.

PAGE 43

32 Figure 3-5 Diamond Suspension Paste Figure 3-6 Prepared Samples Preparation for Image Processing Microstructural analysis was performed at Advance Material Characteristic Lab at the Department of Civil and Coastal Engineering, University of Florida, Gainesville. The facility has a Hitachi S 3000N, variable pressure Scanning Electron Microscope (VPSEM). As described earlier, VPSEM would allow the viewing of the specimen without covering it with a conductive surface. A double sided carbon tape was used for attaching samples to specimen holder. This conductive tape avoids any charge built-up and provides a good background surface. Images were viewed at 50 Pa (0.4 Torr), accelerating voltage of 10.0 KV and on the scale of 200 m at 250 X magnification of specimen surface. At the start of this study, polished samples were viewed directly under microscope and five representative images were taken from different location of specimen. However, during the course of study, it was decided to increase the number of images to 16 to have good representative nature of the sample, which was formulated on basics of 4x4 matrixes. The following figures 3-7 and 3-8 shows the grid system.

PAGE 44

33 Figure 3-7 Specimen with Copper Tape Figure 3-8 Specimens with 4x4 matrixes 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 2.15 1.80 The samples were divided into 16 equal parts using a copper conductive tape to make grid on the sample to identify the points of interesting while viewing under microcracks. Image Analysis Images taken from 16 different square position provides information related to microstructure of concrete. A manual mapping of cracks was done in order to find out the length of the microcracks and density per micrometer for comparing all the samples. To identify proper microcracks from the image, shape factor was used. Shape factor is ratio of length to width. If this ratio was greater then 3.5, the object of consideration termed as microcracks or if the ratio was less than 3.5, the object was termed as Void. Each image was manually mapped as well as by viewing on ImageJ software to find out the exact number of cracks to avoid any bias from operator. The method of image analysis is described in more details in Chapter 4.

PAGE 45

34 Summary In this study, samples were taken from concrete mix which showed maximum probability of crack development due to high curing temperature. Specimen preparation method was selected to avoid any preheating of the samples and to avoid any use of vacuum or pressure techniques for impregnation. Once specimen were impregnated and prepared with grinding and polishing, a grid made from conductive copper tape was used to locate the points of interests, while viewing the sample under microscope. At the end, a statistical analysis was performed to investigate total representative nature of images and the rationality for use in this study.

PAGE 46

CHAPTER 4 IMAGE PROCESSING AND ANALYSIS Introduction This chapter describes the image processing and analysis carried out in this research. A total of four samples were made, two of which cured in room temperature and the other two cured at elevated temperature. As described earlier, the samples were divided into 16 equal parts by conductive copper tape. The images were taken around the center of square in 250X resolution Thus for each sample a total of 16 images were taken in 1028 x 960 pixel resolution and 50 Pa working pressure. In SEM observation, finer surface structure images can generally be obtained with lower accelerating voltages. At higher accelerating voltages, the beam penetration and diffusion area become larger, resulting in unnecessary signals (e.g., backscattered electrons) being generated from within the specimen. These signals reduce the image contrast and veils fine surface structures. It is especially desirable to use low accelerating voltage for observation of low-concentration substance. After acquiring the images, image processing was performed to view microcracks in concrete matrix. ImageJ software developed by National Institute of Health was used for the purpose of image processing in this study. Based on java platform, this software has the capability to perform wide ranges of operations such as binary thresholding, area measurement, filtering noise and mathematical operations. 35

PAGE 47

36 Image Processing Technique First step of image processing was to perform segmentation of image into binary image (black and white) by brightness thresholding approach. In this process, software would change the intensities of image pixels which satisfy the given condition, thus highlighting the regions of interests. However, as it can be imagined, there would be variations in grey level of matrix in different images, and there could be bias in operator visions. Investigating this issue, Soroshian (2003) performed a comparative analysis of manual and automated thresholding operation on images taken by Scanning Electron Microscope, where three types of automated thresholding operation (factorization, entropy and moment) were performed. However, study reported that there were no significant differences between manual and factorization mode of thresholding in automated process. The study also reported 5% and 2% difference in Entropy and Moment mode of thresholding respectively which could be regarded as less significant. Therefore for the scope of this study, manual thresholding by brightness approach was performed. Images were acquired with variable pressure scanning electron microscope and PCI-quartz picture acquisition system. Images were transferred and viewed with ImageJ software. For each image, low threshold level was set to zero or default for thresholding and best threshold level image was determined by comparing the original gray image and the image after application of manual threshold on binary image. Total of sixty-four images were converted one after another to binary mode. After binary operation, images were manually mapped to find out total number of crack lengths and crack density across specimens.

PAGE 48

37 Sample 95B00P Room Temperature Sample 1 A. Image 95 B 00 P RT 1 (2) B. Image 95 B 00 P RT 1 (10) C. Image 95 B 00 P RT 1 (11) D. Image 95 B 00 P RT 1 (16) Figure 4-1 Representative Images of Mix 95 B 00 P RT-1 before Image Processing After highlighting features of interest (microcracks and voids) to find the most distinct contrast with concrete background, image analysis was performed to distinguish microcracks and voids. As it is known that microcracks are long elongated structure, length to width (also known as aspect/ form/ shape) ratio was applied. The objects having shape factor of 3.5 or greater was termed as microcracks, while the remaining objects were termed as voids. A software plugin Shape descriptor from ImageJ was used to

PAGE 49

38 identify voids in the images. Shape descriptor takes minimum and maximum pixel values and plots the void concentration in a separate image. After completing void removal from image, microcracks are mapped manually and counted. A. Image 95 B 00 P RT 1 (2) B. Image 95 B 00 P RT -1 10 ) C. Image 95 B 00 P -RT 1 (11) D. Image 95 B 00 RT (16) Figure 4-2 Representative Images of Mix 95 B 00 P RT-1 after Image Processing As the sample was divided into 16 equal parts by copper conductive double-sided tape, 16 representative images were taken from the center of every part. After processing the image, cracks were manually mapped to find out the total length of cracks in each

PAGE 50

39 image. As it can be seen in table 4-1, density of cracks over the given area was found by converting crack lengths and dividing it by image area. Table 4-1 Crack Quantification Analysis for Mix 95 B 00 P RT-1 No. Image Scale Cracks (cm) Cracks length (m) Image Field Crack Density (m/m 2 ) 1 RT 1-1 200m/11 cm 72 cm 1309.091 183 m x 250 m 0.028614 2 RT 1-2 200m/11 cm 71 cm 1290.909 183 m x 250 m 0.028217 3 RT 1-3 200m/11 cm 65 cm 1181.818 183 m x 250 m 0.025832 4 RT 1-4 200m/11 cm 82 cm 1490.909 183 m x 250 m 0.032588 5 RT 1-5 200m/11 cm 61 cm 1109.091 183 m x 250 m 0.024242 6 RT 1-6 200m/11 cm 66 cm 1200.000 183 m x 250 m 0.026230 7 RT 1-7 200m/11 cm 64 cm 1163.636 183 m x 250 m 0.025435 8 RT 1-8 200m/11 cm 61 cm 1109.091 183 m x 250 m 0.024242 9 RT 1-9 200m/11 cm 73 cm 1327.273 183 m x 250 m 0.029011 10 RT 1-10 200m/11 cm 92 cm 1672.727 183 m x 250 m 0.036562 11 RT 1-11 200m/11 cm 72 cm 1309.091 183 m x 250 m 0.028614 12 RT 1-12 200m/11 cm 68 cm 1236.364 183 m x 250 m 0.027024 13 RT 1-13 200m/11 cm 98 cm 1781.818 183 m x 250 m 0.038947 14 RT 1-14 200m/11 cm 63 cm 1145.455 183 m x 250 m 0.025037 15 RT 1-15 200m/11 cm 65 cm 1181.818 183 m x 250 m 0.025832 16 RT 1-16 200m/11 cm 58 cm 1054.545 183 m x 250 m 0.023050 Avg. 1285.227 0.028092 SD. 203.7358 0.004453 Data analysis shows that for 95 B 00P RT-1 samples, average lengths of cracks were 1285.227 m. Maximum length of cracks in an image was found to be 1672m, while minimum length of cracks in an image was 1054.545 m, giving standard deviation of 203.227. As it can be seen in Table 4.1, average crack density in an image was 0.028092 m/m, 2 the maximum density of cracks was 0.038947 m/m 2 and minimum density of cracks was 0.02305 m/m 2 giving standard deviation of 0.004463. Figure 4.3 describes variations in crack density of room temperature cured sample by each image.

PAGE 51

40 95 B 00 P Room Temprature Sample 100.010.020.030.040.0512345678910111213141516Image NumberCrack Density Figure 4-3 Variations in Crack Density of 95 B 00 P RT1 To find out if the sample size was representative in nature, a statistical study was performed. Odd and Even numbers of images were selected for calculating average density of lengths. As it can be seen in Table 4-2, Crack density of odd number images were 0.02832 m/m 2 with standard deviation of 0.005281, while crack density of even number images were 0.02787 m/m 2 with standard deviation of 0.004982. After comparing it with average crack density of all 16 images of 0.02809 m/m 2 it can be seen that crack density is averaged almost equally. Thus it could be deduced that crack density is representative in nature throughout the sample and the number of images are adequate for giving representative portrait of the sample. Table 4-2 Comparison of Odd and Even Number Images No. Crack Density of Odd Number Image (m/m 2 ) Crack Density Even Number Images (m/m 2 ) 1 0.028614 0.028217 2 0.025832 0.032588 3 0.024242 0.026230 4 0.025435 0.024242 5 0.029011 0.036562 6 0.028614 0.027024 7 0.038947 0.025037 8 0.025832 0.023050 Avg. 0.028316 0.027869 SD 0.004643 0.004563

PAGE 52

41 Sample 95B00P Room Temperature Sample 2 A. Image 95 B 00 PRT2 (4) B. Image 95 B 00 PRT2 (10) C. Image 95 B 00 P -RT2 (15) D. Image 95 B 00 P -RT2 (16) Figure 4-4 Representative Images of Mix 95 B 00 P RT-2 before Image Processing Figure 4.4 shows images from sample number two which was cured at room temperature. It can be seen from images that extensive cracking is present in almost all parts of representative images. As described earlier, a total of 16 images were taken from each sample. The images were then processed by ImageJ software for quantification of microcracks.

PAGE 53

42 Image 95 B 00 P-RT2 (4) Image 95 B 00 P-RT2 (10) Image 95 B 00 P-RT2 (15) Image 95 B 00 P-RT2 (16) Figure 4-5 Representative Images of Mix 95 B 00 P RT-2 after Image Processing After processing the image by converting it into binary image, cracks were manually mapped to find out the total length of cracks in each image. As it can be seen in table 4-3, the density of cracks over the given area was found by converting crack lengths and dividing it by image area. Another statistical analysis was carried out to see if the number of images constituent representative area for the sample.

PAGE 54

43 Table 4-3 Crack Quantification Analysis for Mix 95 B 00 P RT-2 No. Image Scale Cracks (cm) Cracks length (m/m 2 ) Image Field Crack Density (m/m 2 ) 1 RT 2-1 200m/11 cm 64 cm 1163.636 183 m x 250 m 0.025435 2 RT 2-2 200m/11 cm 68 cm 1236.364 183 m x 250 m 0.027024 3 RT 2-3 200m/11 cm 56 cm 1018.182 183 m x 250 m 0.022255 4 RT 2-4 200m/11 cm 89 cm 1618.182 183 m x 250 m 0.035370 5 RT 2-5 200m/11 cm 90 cm 1636.364 183 m x 250 m 0.035768 6 RT 2-6 200m/11 cm 65 cm 1181.818 183 m x 250 m 0.025832 7 RT 2-7 200m/11 cm 82 cm 1490.909 183 m x 250 m 0.032588 8 RT 2-8 200m/11 cm 69 cm 1254.545 183 m x 250 m 0.027422 9 RT 2-9 200m/11 cm 83 cm 1509.091 183 m x 250 m 0.032986 10 RT 2-10 200m/11 cm 67 cm 1218.182 183 m x 250 m 0.026627 11 RT 2-11 200m/11 cm 64 cm 1163.636 183 m x 250 m 0.025435 12 RT 2-12 200m/11 cm 61 cm 1109.091 183 m x 250 m 0.024242 13 RT 2-13 200m/11 cm 67 cm 1218.182 183 m x 250 m 0.026627 14 RT 2-14 200m/11 cm 82 cm 1490.909 183 m x 250 m 0.032588 15 RT 2-15 200m/11 cm 84 cm 1527.273 183 m x 250 m 0.033383 16 RT 2-16 200m/11 cm 68 cm 1236.364 183 m x 250 m 0.027024 Avg. 1317.045 0.028788 SD. 194.692 0.004256 Data analysis shows that for 95 B 00P RT-2 samples, average lengths of cracks was 1317.045 m. Maximum length of cracks in an image was found to be 1636m, while minimum length of cracks in an image was 1018.182 m, giving standard deviation of 194.692. As it can be seen in figure 4-6, average crack density in an image was 0.028788 m/m 2 maximum density of cracks was 0.035768 m/m 2 and minimum density of cracks was 0.0222 m/m 2 giving standard deviation of 0.004256. The above data shows presence microcracks in the given sample. However, to see if the images were sufficient to draw conclusions from it, another statistical analysis was carried out. In this technique, images were divided into two formations. The formation was varied such as diagonals, odd-even or lines to prove representative nature of the samples.

PAGE 55

44 95 B 00 P Room Temperature Figure 4-6 Variations in Crack Density of Sample 95 B 00 P RT2 Figure 4-6 demonstrates variations in crack density of sample 95 B 00P RT2 by each image. The data given in Table 4-4 shows that the crack density of odd number images were 0.029309 m/m 2 with standard deviation of 0.0049, while crack densities of even number images were 0.028266m/m 2 with standard deviation of 0.0037. Table 4-4 Comparison of Odd and Even Number Images No. Crack Density of Odd Number Image (m/m 2 ) Crack Density Even Number Images (m/m 2 ) 1 0.025435 0.027024 2 0.022255 0.035370 3 0.027422 0.032588 4 0.025832 0.035768 5 0.032986 0.026627 6 0.025435 0.024242 7 0.026627 0.027024 8 0.033383 0.032588 Avg. 0.029309 0.028266 SD 0.004922 0.003737 After completing analysis of samples cured at room temperature, crack density calculations were performed on samples cured at high temperature. As described earlier, samples were divided in 16 equal parts with conductive copper tape and images were 0 0.005 0.01 0.015 0.02 0.025 0.03 0.035 0.04 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 Number of Images Crack Den s I T Y

PAGE 56

45 taken from the center of rectangle. Figure 4-7 shows images from high temperature cured sample one. Sample 95B00P High Temperature Sample 1 Image 95 B 00 HT -1 (1) Image 95 B 00 HT -1 (2) Image 95 B 00 HT -1 (3) Image 95 B00 HT -1 (16) Figure 4-7 Representative Images of Mix 95 B 00 P HT-1 before Image Processing Maximum heat gain during high temperature curing process of concrete reached up to 85.5 F. The adiabatic temperature condition was generated and measured in the laboratory with a microcontroller based SURE-CURE curing system.

PAGE 57

46 Image 95 B 00 P HT-1 (1) Image 95 B 00 P HT 1 (2) Image 95 B 00 P HT 1(3) Image 95 B 00 P HT 1(7) Figure 4-8 Representative Images of Mix 95 B 00 P HT-1 after Image Processing After preparing the samples and subsequent polishing and lapping operations, copper metal grid system was mounted. Samples were kept in variable pressure scanning electron microscope to analyze microscopic features. As it can be seen in Figure 4-8, images were initially processed by manual thresholding by brightness approach. Lower thresholding was kept at zero, while upper thresholding was moved till the best contrast between images and objects of interest developed. Voids were removed from the binary picture and manual mapping of microcracks were done for crack quantification.

PAGE 58

47 Table 4-5 Crack Quantification Analysis for Mix 95 B 00 P HT-1 No. Image Scale Cracks (cm) Cracks length (m) Image Field Crack Density (m/m 2 ) 1 HT 1 -1 200m/11 cm 105 cm 1909.091 183 m x 250 m 0.041729 2 HT 1-2 200m/11 cm 106 cm 1927.273 183 m x 250 m 0.042126 3 HT 1-3 200m/11 cm 94 cm 1709.091 183 m x 250 m 0.037357 4 HT 1-4 200m/11 cm 52 cm 945.4545 183 m x 250 m 0.020666 5 HT 1-5 200m/11 cm 58 cm 1054.545 183 m x 250 m 0.023050 6 HT 1-6 200m/11 cm 53 cm 963.6364 183 m x 250 m 0.021063 7 HT 1-7 200m/11 cm 48 cm 872.7273 183 m x 250 m 0.019076 8 HT 1-8 200m/11 cm 54 cm 981.8182 183 m x 250 m 0.021461 9 HT 1-9 200m/11 cm 58 cm 1054.545 183 m x 250 m 0.023050 10 HT 1-10 200m/11 cm 55 cm 1000.000 183 m x 250 m 0.021858 11 HT 1-11 200m/11 cm 71 cm 1290.909 183 m x 250 m 0.028217 12 HT 1-12 200m/11 cm 65 cm 1181.818 183 m x 250 m 0.025832 13 HT 1-13 200m/11 cm 62 cm 1127.273 183 m x 250 m 0.024640 14 HT 1-14 200m/11 cm 75 cm 1363.636 183 m x 250 m 0.029806 15 HT 1-15 200m/11 cm 65 cm 1181.818 183 m x 250 m 0.025832 16 HT 1-16 200m/11 cm 82 cm 1490.909 183 m x 250 m 0.032588 Avg. 1253.409 0.027397 SD. 339.4677 0.007420 Data analysis from Table 4-3 shows that for 95 B 00P HT-1 samples average lengths of cracks were 1253.409 m. Maximum length of cracks in an image was found to be 1927 m, while minimum length of cracks in an image was 872.73 m, giving standard deviation of 339.4677. As it can be seen in Table 4-5, average crack density in an image was 0.027397 m/m 2 maximum density of cracks was 0.042126 m/m 2 and minimum density of cracks was 0.019076 m/m 2, giving standard deviation of 0.00742. The above results determine presence of microcracks in the given sample, to find if the sample size is sufficient to draw conclusion, another statistical analysis was carried out. In this process, images were divided into two groups in various formations and average crack density was calculated for each of them. Results shown in table 4-6 shows that average crack density is present throughout the concrete sample.

PAGE 59

48 95 B 00 P High Temprature Sample 100.010.020.030.040.0512345678910111213141516Image NumberCrack Density Figure 4-9 Variations in Crack Density of Sample 95 B 00P HT1 As it can be seen in Table 4-2, Crack density of odd number images were 0.027869 m/m 2 with standard deviation of 0.007721, while crack density of even number images were 0.026925 m/m 2 with standard deviation of 0.007676. After comparing it with average crack density of all 16 images of 0.027397 m/m 2 it can be seen that crack density is averaged almost equally. Table 4-6 Comparison of Odd and Even Number Images No. Crack Density of Odd Number Image (m/m 2 ) Crack Density Even Number Images (m/m 2 ) 1 0.041729 0.042126 2 0.037357 0.020666 3 0.023050 0.021063 4 0.019076 0.021461 5 0.023050 0.021858 6 0.028217 0.025832 7 0.024640 0.029806 8 0.025832 0.032588 Avg. 0.027869 0.026925 SD 0.007751 0.007576 Coefficient of variation being less than 15% proves that results are concentrated within 15% of the mean. Thus average crack density is averaged equally all across the sample.

PAGE 60

49 Sample 95B00P High Temperature Sample 2 Image 95 B 00 HT -2 (15) Image 95 B 00 HT -2 (7) Image 95 B 00 P HT 2 (15) Image 95 B 00 P HT 2 (16) Figure 4-10 Representative Images of Mix 95 B 00 P HT-2 before Image Processing The above four images are representative images from mix 95 B 00 P cured at high temperature, taken by Hitachi S 3000 N variable pressure scanning electron microscope. A 4x4 matrix made by conductive copper tape was used to take 16 representative images across the sample.

PAGE 61

50 Image 95 B 00 P HT 2 (4) Image 95 B 00 P HT 2 (10) Image 95 B 00 P HT 2 Image 95 B 00 P HT 2 Figure 4-11 Representative Images of Mix 95 B 00 P HT-2 after Image Processing All the images were processed in a java based free source ImageJ software, developed by National Institute of Health, for image processing. Manual thresholding was done till the perfect contrast between regions of interest with background was developed. Images were then processed to distinguish voids from cracks. As cracks have higher length to width ratio, shape factor of 3.5 was used to filter out voids from the cracks. After that, cracks lengths were manually mapped on the image.

PAGE 62

51 Table 4-7 Crack Quantification Analysis for Mix 95 B 00 P HT-2 No. Image Scale Cracks (cm) Cracks length (m) Image Field Crack Density (m/m 2 ) 1 HT 2-1 200m/11 cm 58 cm 1054.545 183 m x 250 m 0.023050 2 HT 2-2 200m/11 cm 98 cm 1781.818 183 m x 250 m 0.038947 3 HT 2-3 200m/11 cm 88 cm 1600.000 183 m x 250 m 0.034973 4 HT 2-4 200m/11 cm 64 cm 1163.636 183 m x 250 m 0.025435 5 HT 2-5 200m/11 cm 57 cm 1036.364 183 m x 250 m 0.022653 6 HT 2-6 200m/11 cm 69 cm 1254.545 183 m x 250 m 0.027422 7 HT 2-7 200m/11 cm 57 cm 1036.364 183 m x 250 m 0.022653 8 HT 2-8 200m/11 cm 54 cm 981.8182 183 m x 250 m 0.021461 9 HT 2-9 200m/11 cm 75 cm 1363.636 183 m x 250 m 0.029806 10 HT 2-10 200m/11 cm 65 cm 1181.818 183 m x 250 m 0.025832 11 HT 2-11 200m/11 cm 70 cm 1272.727 183 m x 250 m 0.027819 12 HT 2-12 200m/11 cm 62 cm 1127.273 183 m x 250 m 0.024640 13 HT 2-13 200m/11 cm 65 cm 1181.818 183 m x 250 m 0.025832 14 HT 2-14 200m/11 cm 76 cm 1381.818 183 m x 250 m 0.030204 15 HT 2-15 200m/11 cm 95 cm 1727.273 183 m x 250 m 0.037755 16 HT 2-16 200m/11 cm 61 cm 1109.091 183 m x 250 m 0.024242 Avg. 1265.909 0.027670 SD. 238.8743 0.005393 Data analysis from Table 4-8 shows that for 95 B 00P HT-2 samples average lengths of cracks was 1265.909 m. Maximum length of cracks in an image was found to be 1781.818 m, while minimum length of cracks in an image was 981.8182 m, giving standard deviation of 238.8743. As it can be seen in Table 4.5, average crack density in an image was 0.027670 m/m 2 maximum density of cracks was 0.038947 m/m 2 and minimum density of cracks was 0.021461 m/m 2, giving standard deviation of 0.005393. Figure 4-12 shows variations in microcrack density in the sample made from high temperature cured concrete mix 95 B 00 P, i.e., concrete made with cement B, (manufactured by Cemex AASHTO type II, heat of hydration 78.2 cal/g at 7 days), placing temperature of 95F, and without any pozzlonic contents added to mix. The higher temperature reached was 180F.

PAGE 63

52 95 B 00 P High Temprature00.0050.010.0150.020.0250.030.0350.040.04512345678910111213141516Number of ImagesCrack Density Figure 4-12 Variations in Crack Density of Sample HT 2 Figure 4-12 shows maximum value of crack density was 0.3894 with lowest crack density appeared to be 0.214. Table 4.8 shows crack density of odd number images from the sample having crack density of 0.028068 m/m 2 while crack density of even number images from samples was 0.027273 m/m 2 From Statistical analysis of image patterns and coefficient of ratio within 20% suggests that, average image density is present throughout the images and given sample size is true representative in nature. Similar results were found with random or diagonal image patterns with same images. Table 4-8 Comparison of Odd and Even Number Images No. Crack Density of Odd Number Image m/m 2 Crack Density Even Number Images m/m 2 1 0.023050 0.038947 2 0.034973 0.025435 3 0.022653 0.027422 4 0.022653 0.021461 5 0.029806 0.025832 6 0.027819 0.024640 7 0.025832 0.030204 8 0.037755 0.024242 Avg. 0.028068 0.027273 SD 0.005775 0.005349

PAGE 64

53 Comparative Analysis of Samples To analyze the differences between both samples, which are cured at room temperature, a comparative analysis was performed. Table 4-9 Comparative Analysis of 95 B 00P RT1 and 95B00P RT2 Sample Avg. Crack Length (m) Avg. Crack Density (m/m 2 ) RT1 1285.227 0.028092 RT2 1317.045 0.028788 Average 1301.136 0.02844 % Change 2.48% 2.48% The results shown in Table 4-9 show that the difference between average crack lengths of samples cured at room temperatures was 32 m/m 2 ; the difference between average crack densities was 0.0006. The differences between average crack lengths in high temperature cured samples were 12.5 m/m2 and difference in their crack density were 0.0002. Table 4-10 Comparative Analysis of 95 B 00P HT1 and 95B00P HT2 Sample Avg. Crack Length (m) Avg. Crack Density (m/m 2 ) HT1 1253.409 0.027397 HT2 1265.909 0.027670 Average 1259.659 0.027535 % Change 0.957% 0.99%

PAGE 65

54 Summary Four samples were taken from concrete mix 95 B 00 P, two of which were cured at room temperature and the two remaining samples were cured at high temperature. After sample preparation, the samples were viewed under Hitachi S 3000 N, variable pressure scanning electron microscope. Each sample was coated with copper conductive tape as 4x4 matrixes which divided the samples in sixteen equal parts. Thus, a total of 16 images were taken, each one from the center of the parts. The images were later analyzed with ImageJ, image processing software. Segmentation of images was done by manually thresholding the image till the best contrast developed between background and points of interests. After thresholding, cracks were manually mapped for quantification of microcracks in sample. The results show that the average crack density of concrete sample cured at room temperature was 0.02844 mm/mm 2 while crack density of sample cured at elevated temperature was 0.027535 mm/mm 2 These results are in accordance with research study conducted by Ammouche et al., (2000) in which average crack density of undamaged samples was found to be 0.00271 mm/mm 2 From the above analysis, it could be concluded that room temperature cured sample had approximately 3% more microcracks than the sample cured at high temperature. Thus it could be said that there is no or very small difference between the quantity of microcracks found in concrete cured at room temperature and concrete cured at elevated temperature for concrete samples made without pozzlonic materials.

PAGE 66

CHAPTER 5 SUMMARY, RESULTS AND RECOMMENDATIONS Summary The purpose of this research was to identify if any additional microcracks are developed due to concrete cured at elevated temperature. To accomplish that following tasks were performed. Identification of the samples which had highest curing temperature and therefore may develop additional microcracking. Performed a literature review on methods available for diction of microcracks and finding out their advantages and disadvantages. Prepared specimens in order to view them under Scanning Electron Microscope for investigating microcracks. Processed images to quantify and compare the intensity of microcracks in both samples Samples were selected out of 20 different concrete mixes made with different percentages of pozzlonic materials, two AASHTO type II cement with two different placing temperatures. Phase I of this project showed that, concrete made from a cement with higher heat of hydration and without adding any pozzlonic material generate highest curing temperature. Therefore, a mix placed at 95F and made with AASTO type II cement having heat of hydration of 78.2 cal/g at 7 days without any pozzlonic material was selected for this study. Selection of appropriate sample preparation method was one of the important issues as improper sample preparation would cause additional microcracks in sample and lead to erroneous results. Most of the sample preparation methods used in past involved use of 55

PAGE 67

56 preheating the specimen for removal of pore water or putting specimen under vacuum for impregnation. There is a concern that by use of heat or vacuum, samples might develop additional microcracks. Epoxy impregnation method by Struble and Stutzman (1989) was used in this study as this method uses counter diffusion process to impregnate samples. Apart from supporting the microstructure of specimen, Epoxy impregnation also provides better contrast in cementitious matrix. Longer sample preparation time is the only drawback for performing this experiment. Images were acquired by a Hitachi S 3000 N, variable pressure scanning electron microscope (VP-SEM) in backscattered scattered electron (BSE) mode which is a high energy beam capable of distinguishing the particles present in matrix by variation of brightness in images. The specimens were divided as 4x4 matrixes by a copper conductive tape to aid viewing of points of interest from predefined positions. Total of 16 images were taken from each sample and analyzed in ImageJ software for quantification of microcracks. Images were converted to binary mode by manual thresholding operation. When features of interest (microcracks and voids) were correctly highlighted, the most distinct contrast between microcracks/voids and concrete background was obtained. After getting binary image, next task performed was to distinguish between voids and microcracks. As its known that microcracks are long elongated structure, thus length to width (also known as aspect/ form/ shape) ratio was applied. Objects having shape factor of 3.5 or greater was termed as microcracks and rest other objects were termed as voids. A software plugin Shape descriptor from ImageJ was used to identify voids in the images. Shape descriptor takes minimum and maximum pixel values and plots the void concentration in a separate image. After completing void

PAGE 68

57 removal from image, microcracks are mapped manually and counted. Results of quantification analysis were later converted into crack density per area to give comparative analysis. Results and Conclusion Results from this studies showed that, crack density of concrete sample cured at room temperature was 0.02844 while, crack density of sample cured at elevated temperature was 0.027535. Table 5-1 Comparative Analysis of Samples Sample Avg. Crack Length(m) Avg. Crack Density (m/m 2 ) Room Temperature Cured samples 1301.136 0.028440 High Temperature Cured samples 1259.659 0.027535 % Change 3.3% 3.28% This analysis indicated that there is very small difference between quantity of microcracks found in concrete cured at room temperature or concrete cured at elevated temperature for samples made with no pozzlonic materials and placed at higher temperature. Thus it can be concluded that high temperature curing conditions upto 180F may not play significant part in formation of additional cracks. Recommendations From research of this study, the following recommendations were made for conducting future work in this area. In future, studies involving mixes made with different percentage of pozzlonic materials, water/cement ratios, aggregates and placing temperature should be investigated to find out if mix design changes have any impact on density of microcracks in concrete samples.

PAGE 69

58 Epoxy solution used in this study had to be cured at 131F according to manufacturers instruction. Curing of the sample at high temperature might alter the microstructure of sample, thus efforts should be directed toward finding an ultra low viscosity epoxy solution, which does not require heating during its curing process. As cutting operations would create damaged zone in concrete microstructure upto 30% of blade thickness in sample, proper grinding and polishing operations should be followed to expose an undamaged surface. A more intrinsic grid system of conductive tape should be created to acquire more images from concrete sample which can be useful if automated image processing is used.

PAGE 70

LIST OF REFERENCES American Concrete Institute (ACI) Committee 116, Report SP-19, Farmington Hills, Michigan, 1995 Advances in Materials Problem solving with Electron Microscope Proceedings of Material Research Society, Volume 589, 1999 Ammouche, A. Breysse, D. Hornain, H. Didry, O., and Marchand, J., New image analysis technique for the quantitative assessment of microcracks in cement-based material, Cement and Concrete Research Volume 30, No.1, Page 25-35, 2000 Bisschop, J., and Van Mier, J.G.M., How to study drying shrinkage microcracking in cement-based materials using optical and scanning electron microscopy, Cement and Concrete Research Volume 32, No. 2, Page 279-287, February 2002 Florida Department of Transportation, Structural Design Guide, Section 3.9 Mass Concrete, Tallahassee, Florida Gran, H. C., Fluorescent liquid replacement technique: A means of crack detection and water-binder ratio determination in high-strength concretes. Cement and Concrete Research Volume 25, No 5, Page 1063-1074, 1995 Hanke, L. Variable Pressure Scanning Electron Microscopy for Nonconductive and Volatile Samples, Handbook of Analytical Methods, 1999 Hornain H., Marchand J., Ammouche A., Commene J.P., and Moranville M., "Microscopic observation of cracks in concrete-a new sample preparation technique using dye impregnation. Cement and Concrete Research, Volume 26, No.4 Page 573-583, 1996 Kjellsen, K. O., and Jennings, H M., Observations of microcracking in cement paste upon drying and rewetting by environmental scanning electron microscopy, Advanced Cement Based Materials Volume 3, No.1, Page 14-19, January 1996 Knab L.I., Walker H.N., Clifton J.R., and Fuller E.R., Fluorescent thin sections to observe the fracture zone in mortar, Cement and Concrete Research, Volume 14, No. 3, Page 339-344, 1984 Maggenti R., Mass Concrete Report, San Francisco-Oakland Bay Bridge East Spans Safety Project. 04-Ala/SF-80-Var, E.A. 04-012021, Available online at http://www.dot.ca.gov/hq/esc/tollbridge/SFOBB/EastSpan/012024/MaterialsHandout/cd2/concret1.pdf, last seen, Oct 12, 2004 59

PAGE 71

60 Marusin S., Sample preparation-the key to ESM studied of failed concrete, Cement and Concrete. Composites Volume 17, No.4, Page 311, 1995 Najjar, W.S., Aderhold, H.C., and Hover, K.C., Application of neutron radiography to the study of microcracking in concrete Cement, Concrete and Aggregates, Volume 8, No. 2, Page 103-109, 1986 Oberholster, R. E., Maree, H. and Brand, J. H. B, Cracked prestressed concrete railway sleepers: alkali silica reaction or delayed ettringite formation, Proceedings of the 9th International Conference on Alkali-Aggregate Reaction in Concrete, London, Page 740-749, 1992 Ollivier, J.P., Cement and Concrete Research, Volume 15, No.6, Page 1055-1060, 1985 Patel, H.H., Bland, C.H., and Poole, A.B., Microstructure of concrete cured at elevated temperatures, Cement and Concrete Research Volume 25 No. 3 Page 485-490, April 1995 Perkes, P., ACEPT W group, Department of Physics and Astronomy, Arizona State University, http://acept.la.asu.edu/PiN/rdg/elmicr/elmicr.shtml Last Visited October 2004 Ringot, E., and Bascoul, A., About the analysis of microcracking in concrete, Cement and Concrete Composites Volume 23, No. 2-3, Page 261-266, April 2001 Samaha, H.R., and Hover, K.C., Influence of microcracking on the mass transport properties of concrete, ACI materials Journal, Volume 89, No.4, Page 416-424, 1992 Slate, F.O., and Hover, K.C., Fracture Mechanics of Concrete: Edited by F.H. Whitmann, Published by Elsevier Science Page 85-93, 1984 Slate F.O., and Olsefski, S., X-Rays for Study of Internal Structure and Microcracking of Concrete, Journal of American construction Institute, Volume 60, No. 5, Page 575-588, 1963 Soroushian, P., Elzafraney, M., and Nossoni, A., Specimen preparation and image processing and analysis techniques for automated quantification of concrete microcracks and voids, Cement and Concrete Research Volume 33, No.12, Page 1949-1962, December 2003 Saint John, D., Poole A., and Sims A, Concrete Petrography-A Handbook of Investigative Techniques Published by Arnold, 1998 Struble L. J., and Stutzman P.E., Epoxy impregnation of hardened cement for microstructural characterization, Journal of Material Science Letters, Volume 8, Page 632-634, 1989

PAGE 72

61 Stutzman P, Scanning Electron Microscopy in Concrete Petrography, Materials Science of Concrete Special Volume: Calcium Hydroxide in Concrete (Workshop on the Role of Calcium Hydroxide in Concrete). Proceedings of The American Ceramic Society, edited by J. Skalny, J. Gebauer and I. Odler, November 1-3, 2000, Anna Maria Island, Florida, Page 59-72, 2001.

PAGE 73

BIOGRAPHICAL SKETCH Vishal Shah was born in Surendranagar, India, on March 21 st 1980, to Sanjay and Rekha Shah and grew up in Mumbai, located in Western India. He graduated in April of 1997 from Mithibai College and joined the Maharashtra Institute of Technology, University of Pune, to complete his undergraduate degree in civil engineering. After graduating with first class in 2002, Vishal continued his graduate studies at the M.E.Rinker Sr. School of Building Construction, University of Florida. During his stay at the University of Florida, Vishal worked as a graduate assistant for Dr. Abdol Chini. Upon his graduation in December 2004, Vishal plans to pursue a career in the exciting and challenging field of construction. 62