You are on page 1of 66

Section 9.

1 - XRD

CHAPTER 9: THE USE OF X-RAY DIFFRACTION AND


SCANNING ELECTRON MICROSCOPY IN
HIGH VOLUME BLENDED CEMENTS
SECTION 9.1:

USE OF X-RAY DIFFRACTION IN HIGH


VOLUME BLENDED CONCRETE SPECIMENS

9.1.1

INTRODUCTION

The Chapter 9 covers both X-ray Diffraction (XRD) and Scanning Electron Microscopy
in high volume blended cements, while the section 9.1 deals with the use of XRD
analysis in high volume cements. This section is structured in to three sub-sections
and this subsection 9.1.1 covers the literature review, while the second subsection
9.1.2 describes the experimental results, and the third subsection 9.1.3 explains the
discussion of results. The XRD is a technique used to determine the crystalline phases
present in the concrete such as calcium hydroxide (portlandite) and ettringite. This
technique is also being extended to determine the early hydration reactions of cement,
for example in synthesis of ettringite. However, the identification of cement phases
present in concrete is more complicated because hydration phases of concrete are
mostly amorphous, which results in weak diffraction peaks. In addition, the peaks from
aggregate minerals may also interfere in the XRD analysis of concrete. Although, few
studies in the past indicated some limitations in the use of XRD on the accuracy of
quantitative phase analysis of cement, however with the use of modern computerized
diffractometers, the analysis is easier and faster. In addition, XRD analysis have many
applications in concrete such as measurement of glass content in pozzolanic material,
degree of hydration, and also in predicting the strength of slag cements. However, in
this study, XRD analysis was applied only for determination of calcium hydroxide in
high volume blended concrete specimens.

The detailed procedure and the principle involved in conducting the XRD analysis in
concrete specimens are described in Chapter 3 (section 3.6.12). In this study, high

553

Section 9.1 - XRD

volume blended concrete specimens were analysed using XRD after 365 days of drying
and also for few concrete specimens after 180 days of drying using two w/b ratios of
0.38 and 0.48.

9.1.1.1

APPLICATIONS

OF

XRD

ANALYSIS

IN

THE

CONCRETE LITERATURE REVIEW


The XRD is useful technique for the direct identification of crystalline phases present in
the cement (Klug and Alexander 1974). However, the identification of the cement
phases present in concrete and mortars is more complicated because of the poorly
crystalline nature of calcium silicate hydrates which provide weak diffraction peaks and
tend to be swamped by the diffraction peaks from calcium hydroxide. In addition,
peaks from minerals present in the aggregates may also interfere (St. John et al. 1998).
The practical limits for detecting phases depend on the crystallinity and ordering of the
phase. Quartz require as little as 0.1% for the detection, however, for many cement
hydrates as much as 5% may be required for detection. XRD analysis, as a rough guide,
may have difficulty in detecting the phase, if it is less than 2% (St. John et al. 1998).

The XRD methods in research studies for cement hydration have been extended by the
use of synchrotron energy dispersive XRD techniques. This powerful technique can
provide real-time investigations of early hydration reactions such as the synthesis of
ettringite (Muhamad et al. 1993) and the rapid conversion of calcium aluminate cement
hydrates (Rashid et al. 1992). The cement is comprised of predominantly silicate
phases, consequently, the interstitial phase (i.e. aluminate and ferroaluminate phases)
cannot be detected by the XRD method.

Therefore, Aitcin and Mindess (1998)

evaluated the fundamental factors underlying the production of modern highperformance concrete.

They proposed a specific treatment called salicyclic acid

treatment for analysing the interstitial and sulfate phases. The salicylic acid dissolves
the silicate phases and the x-ray diffractograms on the non-dissolved part gives the
exact nature of the interstitial and the sulfate phases present in the cement.

554

Section 9.1 - XRD

The quantity of C-S-H can be determined easily in Portland cement paste using standard
indirect techniques, such as loss on ignition or quantitative X-ray analysis of unreacted
cement and calcium hydroxide. The C-S-H is very difficult to measure directly due to
both amorphous nature (lack of crystallinity) and its indefinite composition.

The

pozzolanic reaction between CH and mineral admixture (typically a glassy or partly


glassy phase which cannot be quantified by XRD) produces C-S-H which fails to
distinguish indirect techniques, because the pozzolanic reaction generates C-S-H and
consumes unknown amounts of calcium. In addition, the Ca/Si ratio of C-S-H in
presence of mineral admixture will be lower than the neat cement paste (Olson and
Jennings 2001).

The calculation of compound or phase composition from oxide is made either by


theoretical method (Bogue potential oxide composition) or experimental method (using
X-ray diffraction method or Rosiwal methods). XRD has been successfully used to
determine the mineral concentrations for a number of years, and since 1958 there have
been several attempts to use XRD analysis for determining the compound composition
of cement (Bye et al. 1975). However, the accuracy of this method is open to serious
question. On the other hand, the alternative Bogue method is also subjected to serious
error as cited by (Aldridge 1982a).

Nevertheless, Aldridge (1982a) stated that in

principle, the X-ray method is better analytical method.

XRD technique is used to evaluate each phase of cement or clinker to predict the
performance of Portland cement. The performance of Portland cement in terms of
setting, heat of hydration, strength, and resistance to sulfate attack generally depends on
its composition such as its phases and their proportions. The methods to quantitatively
determine phase proportions include Quantitative X-ray Diffraction (QXRD),
quantitative microscopical analysis by Optical Microscopy (OM), Scanning Electron
Microscopy (SEM), and calculations from chemical analysis (Bogue method).
However, the QXRD and microscopical analysis are direct methods, while the
calculation of the potential phase composition from the chemical analysis is an indirect
method (Struble 1991).

There are no standard procedures for QXRD analysis,

therefore, the details of the experimental procedure vary considerably depending on the
material analyzed and the instrument used for analysis. Taylor (1990) reported that

555

Section 9.1 - XRD

although, despite of the difficulties associated with QXRD, however, with adequate
experimental procedures, the major clinker phases can be determined with an absolute
accuracy of 2 to 5% for alite & belite, and 1 to 2% for aluminates and ferrite. In another
study, Struble (1991) made an attempt to evaluate the standard test method for QXRD
analysis of Portland cement clinker and the results concluded that quantitative phase
analysis of cement and clinker by XRD method exhibited accurate analysis.

XRD technique can be used to measure glass content in fly ash and slag. Several
studies have shown the importance of glass content for the compressive strength in slag
concrete to mention a few: Schwiete and Dolbor (1983) modified quenching conditions
for each of 30 slags and found that the glass content was predominant factor affecting
the strength of slag concrete and the concrete comprising 30 to 40 per cent glass
performed satisfactorily.

On the other hand, Smolczyk (1980) found a linear

relationship between glass content and strength. Nevertheless, Demoulian et al. (1980)
reported that the strength of slag concrete with glass content in excess of 95 per cent
was reduced. Hooton (1981) evaluated various methods of glass content determination
in glass-ceramics, polymers, slags and other waste materials. However, the results
indicated that the extent to which glass content can be measured with precision is
doubtful, and the results between various techniques are not always in agreement.

Roode et al. (1987) developed an accurate procedure to measure glass content in blast
furnace slags and in fly ashes to predict their performance in concrete.

The

investigation involves using QXRD method which is used to compute the mass
percentages of crystalline phases such as quartz, mullite, magnetite and hematite, and
the glass content is computed by difference. However, they were not convinced with
the results whether the QXRD approach to the glass content determination gives the
most accurate values, this is because the potential sources of error depends on the initial
identification of the crystalline phases, preparation of the standard curves and preferably
several for each crystalline phase, although it is a time-consuming operation.
Nevertheless, the QXRD method is most useful for the routine determination of glass
contents of fly ashes, slag and other waste material from specific sources which are not
subject to the variations in chemical and mineralogical compositions observed between
samples from different sources. The author of this study has used SIEMENS D501

556

Section 9.1 - XRD

diffractometer equipped with monochromator and scintillation detector using CuK


radiation. The quantitative estimates of mineral phases in various blended cements
were determined.

The XRD technique is the direct method of finding the degree of hydration by
measuring the fraction of unreacted phases (Parotte et al. 1990). On the other hand, the
indirect determination of the degree of hydration may be carried out by measuring some
property of the product such as chemically combined water, the heat of hydration,
compressive strength or the change in specific volume of the saturated sample called as
chemical shrinkage. Nevertheless, an established correlation between the property of
the product and the true degree of hydration determined by means of QXRD is often
needed in using an indirect method (Parotte et al. 1990).

Aldridge (1982a) evaluated the accuracy and precision of X-ray diffraction method for
analysing Portland cements by using a modified method developed by Berger et al.
(1964) after analysing 150 cement samples from the New Zealand. The results of XRD
were compared with Bogue method and by wet chemical analysis methods. The results
indicated that both Bogue and X-ray compositions were equally good estimates of the
true composition with same degree of accuracy with no significant difference between
the accuracy of composition determined by the Bogue and XRD methods for the New
Zealand cements. Aldridge (1982a) concluded that the present method of XRD analysis
cannot give accurate absolute estimates of composition, however it can give the relative
compositions within a group of cements with reasonable accuracy.

In another study, Aldridge (1982b) evaluated the analysis of Portland cement & clinker
by Bogue, Microscopic, and X-ray diffraction methods and found errors between and
within laboratory, in addition neither the Bogue nor XRD methods gave accurate
analysis of the cement. The errors in the XRD analysis of cement is attributed to three
causes: The variation of intensity of X-rays, and this error includes alignment of the
diffractometer, packing of the sample, insufficient grinding of the sample and the use of
the unsuitable internal standard, while the other error is the difficulty of separating the
overlapping peaks free from interference. Thirdly, the standardisation error as the XRD

557

Section 9.1 - XRD

patterns of the same cement mineral can vary significantly because of the solid solution
replacement of cations.

Aldridge (1982b) stated that the present methods of XRD analysis are in general,
unsatisfactory for analysing cement. The XRD however, is the only analytical method
which can directly determine the amount of the compounds present in cement. Finally,
He concluded that at this moment there is no method of accurately determining the
phase composition of cement.

Nevertheless, Sarkar and Roy (1988) reported that the benefits provided by modern
computerized diffractometers in quantitative phase analysis of cement and clinker are
immense. They make quantitative phase analysis easier and faster, compared to the past
experimental procedures which were too time consuming. For example, multipeaks can
be used for each phase, and computerized profile-fitting algorithms allow more
reproducible intensity measurements in the case of partially overlapping peaks. It is the
authors experience that identification of phase analysis in cement is possible with the
use of XRD in the present computers. However, the analysis results should be guided
by an expert because of the amorphous phases in cement.

X-ray diffraction can also be used to predict the strength of Portland slag cement
mortars. Goswami et al. (1994) developed a very rapid method to predict the 3 and 28days strength of Portland slag cements (PSC) mortars with negligible variations from
the conventionally tested values with actual testing time around 20 minutes. The data is
from the display of the XRD unit and therefore this method avoids any element of
personal judgment as involved in microscopic method. The prediction is based on the
two indices X n and S n , derived from the ratios of XRD pulse counts of the test
samples. The parameter X n is the ratio between the pulse counts for belite and alite and
S n depends on the slag present in the sample. The equation derived for commercial
PSC samples from a particular plant:

558

Section 9.1 - XRD

3-day compressive strength (when X n is >1);

205 + 30( X n 1) ( S n 1.24)500 kg/cm2.Eqn. (9.1)

3-day compressive strength (when X n is <= 1);

205 + 30( X n ) ( S n 1.24)500 kg/cm2...Eqn. (9.2)


28-day compressive strength = C +

C * In

, where C is the 3-day compressive strength calculated from

100

(I) or (II); and I n = 140 12.5 * X n * S n . Nevertheless, no attempt was made by the author to evaluate
the compressive strength from the X-ray diffraction in Portland slag cement mortars.

9.1.1.3

SUMMARY OF THE LITERATURE REVIEW ON


XRDA IN CONCRETE SPECIMENS

X-ray diffraction is the direct method for finding the crystalline phases present
in the concrete.

This technique is used to evaluate each phase of cement or clinker to predict the
performance of Portland cement.

It is used in the determination of glass content in fly ash and slag cements.

It is used in determining the degree of hydration of cement.

It is used in the rapid strength determination of Portland slag cement mortars.

It is observed that quantitative phase analysis of cement using modern


computers is easier and faster compared to the analysis in the past.

559

Section 9.1 - XRD

USE OF X-RAY DIFFRACTION IN HIGH VOLUME


BLENDED CONCRETE SPECIMENS

9.1.2

EXPERIMENTAL RESULTS

9.1.2.1

INTRODUCTION

This subsection 9.1.2 covers only the results and general conclusions of XRD
analysis in high volume blended concrete specimens, while the critical discussion of
results is presented in sub section 9.1.3. The cementitious powder was collected from
the remnants of concrete specimen from compressive strength in various percentage
blends of pozzolanic specimen after 365 days of drying and analysed using X-ray
Diffraction (XRD).

The XRD results are presented in the form of peaks for the

following concrete blends cast using 0.38 and 0.48 w/b ratios: Plain and Silica fume
concrete specimens; various percentage blends of fly ash which includes binary and
ternary concrete blends; various percentage blends of BFS which includes binary and
ternary concrete blends. The XRD peaks were identified with mineral phases for all
blended concrete specimens. In addition, the quantitative percentage* of portlandite in
fly ash and BFS concrete specimens of various percentage blends were presented
separately for both w/b ratios. The amount of Ca(OH)2 present in the blended cements
represents an indication of the degree of hydration. The basic purpose of this study was
to evaluate the percentage of Ca(OH)2 (portlandite) in plain and high volume blended
concrete specimens cast using two w/b ratios namely 0.38 and 0.48 for better
understanding of results. Lastly, the results of XRD in the form of peaks are presented
for a typical ponded slab specimen at three depths: 5-25 mm; 25-45 mm; and 45-60 mm
to see the formation of hydration products in sodium chloride solution.

___________________________________________
*

The summation of all mineral phases observed in XRD analysis is 100%. The expressed percentage

is the fraction of portlandite present in the specimen.

560

Section 9.1 - XRD

9.1.2.2

Quantitative Estimate of all Mineral Phases present in


both Fly Ash and BFS Concrete Specimens in Percentages

The Tables 9.1 and 9.2 are the quantitative estimates of mineral phases observed in
different blends of fly ash and BFS, respectively cast using 0.38 w/b ratio, using the
program Siroquant 2.5 in percentages observed after 365 days of drying.
Table 9.1: Quantitative estimate of the mineral phases in plain and silica fume; and various percentage
blends of fly ash specimens in percentages cast using 0.38 w/b ratio after 365 days of drying
Wt (%)

Chemical Formula

3OPC

3S1

3F2

3F2S1

3F5

3F5S1

3F7

3F7S1

portlandite

Ca(OH)2

10.1

5.7

10.5

4.3

1.4

6.1

quartz

SiO2

79.7

73.8

83.3

76.7

69.7

81.3

64.2

79.3

albite

NaAlSi3O8

3.8

13.4

5.9

11

10.5

9.9

microcline
calcium
silicate
silicon
carbide

KAlSi3O8

anorthite

CaAl2Si2O8

corundum

Al2O3

paragonite

NaAl2(AlSi3)O10(OH)2

mullite

Al(Al1.272Si0.728O4.864)

2.3

Ca3SiO5

4.2

SiC

2.2

7.1

0.3

5.7

13.4
5.7
9.3

6.8

2.7

14.2

11.1

Table 9.2: Quantitative estimate of the mineral phases in various percentage blends of BFS specimens in
percentages cast using 0.38 w/b ratio after 365 days of drying
Wt (%)

Chemical Formula

3OPC

3S1

3B2

3B2S1

3B5

3B5S1

3B7

3B7S1

portlandite

Ca(OH)2

10.1

5.7

15.6

4.2

7.3

3.9

2.6

quartz

SiO2

79.7

73.8

74.2

81.8

75

84

83.5

77.4

albite

NaAlSi3O8

3.8

13.4

10.2

13.6

15.7

12

13.9

16.4

microcline
calcium
silicate
silicon
carbide

KAlSi3O8

anorthite

CaAl2Si2O8

corundum

Al2O3

paragonite

NaAl2(AlSi3)O10(OH)2

mullite

Al(Al1.272Si0.728O4.864)

rutile

TiO2

Ca3SiO5

4.2

SiC

2.2

7.1

6.1
0.4

561

Section 9.1 - XRD

The mineral phases were mostly portlandite, quartz, and albite; the silicon carbide
observed in some concrete specimens is from the abrasion of the drill bit used for
extraction of cementitious powder from the piece of concrete. The mineral phase
calcium silicate was observed in the control plain concrete specimen while it was absent
in (3S1) specimens. This can be attributed to the calcium silicate hydrates when formed
with pozzolanic material (SiO2) is less reactive than the hydration of calcium silicate
compounds contained in the cement as cited by Kasselouri et al. (2001). Therefore, the
calcium silicate is more prominently observed in the control OPC cement compared to
the silica fume specimens. The Tables 9.3 and 9.4 are the quantitative estimate in both
blends of fly ash and BFS, respectively cast using 0.48 w/b ratio, using the program
Siroquant 2.5 in percentages observed after 365 days.
Table 9.3: Quantitative estimate of the mineral phases in various percentage blends of fly ash concrete
specimens in percentages cast using 0.48 w/b ratio after 365 days of drying
Wt (%)

Chemical Formula

4F2

4F2S1

4F5

4F5S1

portlandite

Ca(OH)2

22.9

14

9.6

8.1

quartz

SiO2

62.4

71.6

78.9

77

albite

NaAlSi3O8

11.6

11.2

10.2

11.1

microcline
calcium
silicate

KAlSi3O8

2.6

1.3

Ca3SiO5

1.5

calcite

CaCO3

1.5

andalusite

Al2(SiO4)O

deerite

Fe6(Fe,Al)3Si6O20(OH)5

ettringite

Ca6Al2(SO4)3(OH)12.26H2O

stellerite

Ca2Al4Si14O36.14H2O

cordierite

Mg2Al4Si5O18

0.6
1.7
2.1

562

Section 9.1 - XRD

Table 9.4: Quantitative estimate of the mineral phases in various percentage blends of Blast furnace slag
concrete specimens in percentages cast using 0.48 w/b ratio after 365 days of drying
Wt (%)

Chemical Formula

4B2

4B2S1

4B5

4B5S1

4B7

4B7S1

portlandite

Ca(OH)2

9.6

10.9

8.8

6.2

3.7

quartz

SiO2

77.4

75.8

75.8

76.9

78.6

76.6

albite

NaAlSi3O8

8.8

7.4

11.8

13.9

10.5

16.8

microcline
calcium
silicate

KAlSi3O8

calcite

CaCO3

andalusite

Al2(SiO4)O

deerite

Fe6(Fe,Al)3Si6O20(OH)5

ettringite

Ca6Al2(SO4)3(OH)12.26H2O

stellerite

Ca2Al4Si14O36.14H2O

2.6

cordierite

Mg2Al4Si5O18

1.6

Ca3SiO5

5.9

7.3
6.6
3.5

The mineral phases were mostly portlandite, quartz, and albite. The control and silica
fume concrete specimens cast using 0.48 w/b ratio were not analysed and it was simply
missed out for the XRD analysis. The Table 9.5 shows the estimate of mineral phases
in few specimens of both w/b ratios after 180 days.

Table 9.5: Quantitative estimate of the mineral phases in few various percentage blends of fly ash and
BFS concrete specimens in percentages cast using both 0.38 and 0.48 w/b ratios after
180 days
Wt (%)

Chemical Formula

3F5

3F5S1

3F7

3B7

4B7

4B7S1

portlandite

Ca(OH)2

6.2

2.7

2.5

1.9

1.4

quartz

SiO2

78.8

82.2

88.1

88.3

88.4

92.3

albite

NaAlSi3O8

14.5

11.4

9.3

7.8

9.7

6.4

microcline

KAlSi3O8

0.6

calcite

CaCO3

3.8

3.8

The portlandite in the binary concrete blend of 50% fly ash (3F5) and its ternary blend
with the addition of 10% silica fume (3F5S1) at 180 days was 6.2 and 2.7%,
respectively, while the same concrete specimens after 365 days exhibited portlandite of
5.0 and 1.4%. This indicates that percentage of portlandite in these fly ash concrete
blends at 180 days was higher, while a decrease was observed in these concrete
specimens after 365 days. On the contrary, the percentage of portlandite in the fly ash

563

Section 9.1 - XRD

concrete specimens with 70% replacement of cement (3F7) at 180 and 365 days was 2.5
and 6.1%, respectively. This shows that an increase in amount of portlandite was
observed in 70% fly ash concrete specimens with increase in period of time. Although,
various researchers have used XRD analysis to evaluate portlandite: To state a few
studies (Tarekuddin et al. 2004; Tarekuddin et al. 2002) have used XRDA for the
mineral phases in concrete specimen after marine exposure for their presence or absence
only; while Hill and Sharp (2002) have evaluated the mineral phases in higher volume
blended cement paste specimens and looked into the presence or absence of mineral
phases. However, according to the author to this date there is no mention of studies to
evaluate the quantification of percentage of mineral phases in the increasing percentage
of 25, 50, and 70% of pozzolanic material.

564

Section 9.1 - XRD

9.1.2.3

QUANTIFICATION OF PORTLANDITE USING XRDA


IN THE PLAIN AND SILICA FUME CONCRETE
SPECIMENS CAST USING 0.38 W/B RATIO

Fig. 9.1 shows the XRD peaks for the control (3OPC) and silica fume (3S1) specimens.

Fig. 9.1: X-ray Diffraction Peaks of the Plain and Silica Fume Concrete Specimens; A: albite, P:
portlandite, Q:quartz, CS: calcium silicate

The X-ray diffraction peaks of control specimen (3OPC) indicate peaks of albite,
portlandite, calcium silicate and predominantly quartz. The mineral phases present in
(3OPC) concrete specimens were also observed in silica fume concrete specimen (3S1)
with the exception of calcium silicate. The percentage of portlandite in control and
silica fume concrete specimens was 10.1 and 5.7%, respectively.

565

Section 9.1 - XRD

9.1.2.4

QUANTIFICATION OF PORTLANDITE USING XRDA


IN THE FLY ASH CONCRETE SPECIMENS CAST
USING 0.38 W/B RATIO

Fig. 9.2 shows the X-ray diffraction peaks in the control and binary blends of fly ash
concrete specimens with 0, 25, 50, and 70% replacement of cement, designated as
(3OPC), (3F2), (3F5), and (3F7), respectively.

MU

MU

MU

Fig. 9.2: X-ray Diffraction Peaks of the Plain and the Binary Blends of Fly Ash Concrete with 25, 50, and
70%, replacements. A: albite (NaAlSi3O8), P: portlandite (Ca(OH)2), Q: quartz (SiO2) , CS: calcium
silicate (Ca3SiO5), M: moassanite (SiC); MI: microcline (KAlSi3O8), CR; corundum (Al2O3), MU: mullite
(3Al2O3.2SiO2)

This figure clearly indicates that smaller peaks of portlandite were observed at higher
dosage of 50 and 70% replacement of cement with fly ash at 18.1 and 34.1 2. The
Fig. 9.3 shows the percentage of portlandite in fly ash concrete specimens of various
percentage blends observed after 365 days using Siroquant software.

566

Section 9.1 - XRD

365-3F7S1

365-3F7

365-3F5S1

365-3F5

365-3F2S1

365-3F2

365-3S1

18
16
14
12
10
8
6
4
2
0
365-3OPC

Percentage Portlandite (%)

Percentage Portlandite in Fly ash Concrete Blends:


W/B:0.38

Plain and Fly ash Concrete Specimens


Fig. 9.3: Percentage of portlandite in Plain and Fly ash Concrete Specimens

The percentage of portlandite in the fly ash specimens with 25, 50, and 70%
replacement of cement was 10.5, 5.0, and 6.1%, respectively. This indicates that the
percentage of portlandite decreased with increase in the replacement of cement with fly
ash, however, concrete with 70% replacement of cement with fly ash showed increase
in the percentage of portlandite compared to the binary blend of 50% fly ash. In the
authors study, only one concrete specimen was used for the XRD analysis. However,
the binary blend of 50% fly ash (3F5) was analysed twice only by re-running siroquant
software. The quantitative percentage was 5.2 and 5.0%, respectively, similarly, the
binary blend of 70% fly ash (3F7) was re-analysed twice and the quantitative percentage
was 7.4 and 6.1%, respectively. This indicates that an increasing trend of portlandite in
the 70% fly ash compared to the 50% fly ash specimen was observed in the authors
study and the trend remained similar when the specimen was analyzed twice. The
reason for this will be presented in the discussion section 9.1.3.3. The results
indicate that portlandite in the ternary fly ash blends with the addition of silica fume
showed further decrease compared to the binary blends of fly ash.

The X-ray

diffraction peaks in the control and ternary blends of fly ash is shown in Fig. 9.4.

567

Section 9.1 - XRD

Fig. 9.4: X-ray Diffraction peaks of the Plain, Binary Blends of Fly Ash Concrete with 25, 50, and 70%,
replacements, and the Ternary Blends of Fly Ash with 25, 50, and 70% blends with the addition of 10%
Silica Fume. A: albite (NaAlSi3O8), P: portlandite (Ca(OH)2), Q: quartz (SiO2) , CS: calcium silicate
(Ca3SiO5), M: moassanite (SiC); MI: microcline (KAlSi3O8), CR; corundum (Al2O3)

The ternary blends comprising silica fume at 10% replacement of total cementitious
content and irrespective of any fly ash content showed decrease in portlandite compared
to its respective binary blend of fly ash. It is important to note that no portlandite peak
was observed in the ternary fly ash concrete mix comprising 70% fly ash and 10% silica
fume.

568

Section 9.1 - XRD

9.1.2.5

QUANTIFICATION OF PORTLANDITE USING XRDA


IN THE BLAST FURNACE SLAG SPECIMENS CAST
USING 0.38 W/B RATIO

Fig. 9.5 shows the X-ray diffraction peaks in the binary blends of Blast furnace slag
with 25, 50, and 70%, replacements.

Fig. 9.5: X-ray Diffraction Peaks of the Plain and Binary Blends of Blast Furnace Slag Concrete with 25,
50, and 70%, replacements A: albite (NaAlSi3O8), P: portlandite (Ca(OH)2), Q: quartz (SiO2) , CS:
calcium silicate (Ca3SiO5)

The peaks of portlandite at 18.1 and 34.1 2 in control and binary blends of 25% BFS
were comparable qualitatively, while some reduction in portlandite was observed in the
50% dosage of BFS, and virtually the peaks of portlandite were absent in the specimens
comprising replacement of cement with 70% BFS. Fig. 9.6 shows the percentage of
portlandite formed in binary and ternary blends of Blast furnace slag concrete
specimens.

569

Section 9.1 - XRD

365-3B7S1

365-3B7

365-3B5S1

365-3B5

365-3B2S1

365-3B2

365-3S1

18
16
14
12
10
8
6
4
2
0
365-3OPC

Percentage Portlandite (%)

Percentage Portlandite in BFS Concrete Blends:


W/B:0.38

Plain and BFS Concrete Specimens

Fig. 9.6: Percentage of portlandite in the Plain Concrete, Binary, and Ternary Blends of BFS Concrete
Specimens

The percentage of portlandite in the control and binary blends of 25% BFS specimens
was 10.1 and 15.6%, respectively. This indicates that the percentage of portlandite in
25% BFS was more compared to the normal concrete. However, with increase in the
replacement of cement with 50 and 70% BFS showed decrease in the percentage of
portlandite. The replacement of cement with blast furnace slag at 25, 50, and 70%
levels showed portlandite percentage of 15.6, 7.3, and 2.6%, respectively. On the other
hand, the portlandite in the ternary BFS blends with the addition of silica fume showed
further decrease compared to the binary blends of BFS. The percentage of portlandite
in the ternary blends of BFS in the replacements of 25, 50, and 70% cement with BFS
and addition of 10% silica fume was 4.2, 3.9, and 0%, respectively. In summary,
decrease in the percentage of portlandite was noticed with the increase in the
replacement of cement by blast furnace slag. Fig. 9.7 shows the X-ray diffraction peaks
in the binary and ternary blends of BFS and compared with control specimens.

570

Section 9.1 - XRD

Fig. 9.7: X-ray Diffraction Peaks of the Plain, Binary Blends of Blast Furnace Slag Concrete with 25, 50,
and 70%, replacements, and the Ternary Blends of Blast Furnace Slag with 25, 50, and 70% blends with
the addition of 10% Silica Fume. A: albite (NaAlSi3O8), P: portlandite (Ca(OH)2), Q: quartz (SiO2) , CS:
calcium silicate (Ca3SiO5), T: troilite (FeS)

The percentage of portlandite in the ternary blends comprising 25, 50, and 70% BFS
with the addition of 10% silica fume showed significant decrease compared to the
binary blends of BFS with 25, 50, and 70% levels. The percentage of portlandite
observed in the plain and blended specimens cast using 0.38 w/b ratio from Siroquant
software is shown in Table 9.6.

571

Section 9.1 - XRD

Table 9.6: Percentage of portlandite observed in plain and blended specimens cast using 0.38 w/b ratio

9.1.2.6

S.No.

Concrete mix

Portlandite(%)

365-3OPC

10.1

365-3S1

5.7

365-3F2

10.5

365-3F2S1

4.3

365-3F5

5.0

365-3F5S1

1.4

365-3F7

6.1

365-3F7S1

0.0

365-3B2

15.6

10

365-3B2S1

4.2

11

365-3B5

7.3

12

365-3B5S1

3.9

13

365-3B7

2.6

14

365-3B7S1

0.0

SUMMARY OF RESULTS OF X-RAY DIFFRACTION


IN THE PLAIN, FLY ASH, AND BFS CONCRETE
SPECIMENS CAST USING 0.38 W/B RATIO

9.1.2.6.1

PLAIN AND SILICA FUME CONCRETE SPECIMENS

A significant decrease in the percentage of portlandite was observed in the silica


fume specimens compared to the control specimens.

9.1.2.6.2

FLY ASH CONCRETE SPECIMENS

The percentage of portlandite in the specimens with 25% fly ash was
comparable with the Portland cement

The percentage of portlandite in the fly ash specimens decreased with increase in
the percentage addition of fly ash, except in the 70% fly ash content specimen
which showed slightly higher compared to 50% fly ash specimens.

The ternary fly ash blends showed lower amounts of portlandite formation
compared to their respective binary fly ash counterparts.

572

Section 9.1 - XRD

9.1.2.6.3

BLAST FURNACE SLAG CONCRETE SPECIMENS

The percentage of portlandite in the specimens with 25% BFS was higher
compared to the Portland cement specimens.

The percentage of portlandite in the BFS specimens decreased with increase in


the percentage addition of BFS.

The ternary BFS blends showed lower amounts of portlandite formation


compared to their respective binary BFS counterparts.

573

Section 9.1 - XRD

9.1.2.7

QUANTIFICATION OF PORTLANDITE USING XRDA


IN THE PLAIN AND SILICA FUME SPECIMENS CAST
USING 0.48 W/B RATIO

The percentage of portlandite in plain and silica fume concrete specimens cast using
0.48 w/b ratio could not be evaluated because it was simply missed out for the XRD
analysis.

9.1.2.8

QUANTIFICATION OF PORTLANDITE USING XRDA


IN THE FLY ASH SPECIMENS CAST USING 0.48 W/B
RATIO

Fig. 9.8 shows the X-ray diffraction peaks in the binary blends of 25% fly ash (4F2)
and 50% fly ash (4F5) specimens.

Fig. 9.8: X-ray Diffraction Peaks of the Binary Blends of Fly Ash with 25 and 50%, replacement cast
using 0.48 w/b ratio. A: albite (NaAlSi3O8), P: portlandite (Ca(OH)2), Q: quartz (SiO2), CS: calcium
silicate (Ca3SiO5); MI: microcline (KAlSi3O8)

574

Section 9.1 - XRD

The Fig. 9.8 clearly indicates that the peaks of portlandite at 18.1 and 34.1 2 was
higher in the binary blends of 25% fly ash (4F2) specimen, while the reduction of peaks
was observed in the binary blends of 50% fly ash (4F5) specimen. This indicates that
the percentage of portlandite decreased with increase in the replacement of fly ash. The
Fig. 9.9 shows the X-ray diffraction in the binary blends of 25% fly ash (4F2) and its
ternary blend (4F2S1), binary blends of 50% fly ash (4F5) and its ternary blend
(4F5S1), peaks indicated from the bottom upwards.

Fig. 9.9: X-ray Diffraction Peaks of the Binary Blends of 25% Fly Ash (4F2) and its Ternary Blend
(4F2S1), Binary Blends of 50% Fly Ash (4F5) and its Ternary Blend (4F5S1) cast using 0.48 w/b ratio.
A: albite (NaAlSi3O8), P: portlandite (Ca(OH)2), Q: quartz (SiO2), CS: calcium silicate (Ca3SiO5); MI:
microcline (KAlSi3O8), AND: andalusite Al2(SiO4)O

The ternary blends of fly ash exhibited lower amounts of portlandite compared to their
respective binary blends of fly ash. The percentage of portlandite observed in the fly
ash blended specimens using Siroquant software cast with 0.48 w/b ratio is shown in
Fig. 9.10.

575

Section 9.1 - XRD

25
20
15
10
5

365-4F5S1

365-4F2S1

365-4F5

0
365-4F2

Percentage Portlandite (%)

Percentage Portlandite in the Binary and Ternary


Blends of Fly Ash Specimens : W/B:0.48

Binary and Ternary Blends of Fly ash Specimens

Fig. 9.10:

The percentage of portlandite in the Binary and Ternary Blends of Fly Ash Concrete
Specimens cast using 0.48 w/b ratio

The percentage of portlandite in the binary blends of 25 and 50% fly ash was 22.9 and
9.6%, respectively, while the ternary blends comprising 25% fly ash and 10% silica
fume (4F2S1) and 50% fly ash and 10% silica fume (4F5S1) were 14 and 8.1%,
respectively. The results again showed that the ternary blends showed reduction in the
percentage of Portlandite compared to the binary blends of fly ash concrete specimens.
The author did not evaluate portlandite in the 70% fly ash content specimen (4F7) cast
using 0.48 w/b ratio. Else, the data of 4F7 and 4F5 in the specimens cast using 0.48 w/b
ratio could have strengthened the trend of portlandite in the same blended specimens
(3F7 and 3F5) observed in 0.38 w/b ratio.

576

Section 9.1 - XRD

9.1.2.9

QUANTIFICATION OF PORTLANDITE USING XRDA


IN THE BLAST FURNACE SLAG SPECIMENS CAST
USING 0.48 W/B RATIO

Fig. 9.11 shows the X-ray diffraction peaks in the binary blends of 25% BFS (4B2),
50% BFS (4B5), and 70% BFS (4B7) cast using 0.48 w/b ratio.
Q

Fig. 9.11: X-ray Diffraction Peaks of the Binary Blends of BFS with 25, 50, and 70% additions,
designated as 4B2, 4B5, and 4B7, respectively cast using 0.48 w/b ratio. A: albite (NaAlSi3O8), P:
portlandite

(Ca(OH)2),

Q:

quartz

(SiO2),

STE:

stellerite

(Ca2Al4Si14O36.14.H2O);

Deerite:

[Fe6(Fe,Al)3Si6O20(OH)5]

A decrease in the formation of portlandite was observed in the BFS specimens with
increase in the replacement of cement with BFS. Fig. 9.12 shows the X-ray diffraction
peaks in the 25% BFS (4B2) and its ternary blend with addition of silica fume (4B2S1);
binary blends of 50% BFS (4B5) and its ternary blend (4B5S1); binary blend of 70%
BFS (4B7) and its ternary blend (4B7S1).

577

Section 9.1 - XRD

Fig. 9.12: X-ray Diffraction Peaks in the Concrete Specimens of 4B2, 4B2S1, 4B5, 4B5S1, 4B7, and
4B7S1 cast using 0.48 w/b ratio. A: albite (NaAlSi3O8), P: portlandite (Ca(OH)2), Q: quartz (SiO2), STE:
stellerite (Ca2Al4Si14O36.14.H2O); D; deerite: [Fe6(Fe,Al)3Si6O20(OH)5]

The ternary concrete specimens comprising 70% addition of BFS and addition of 10%
silica fume showed complete consumption of portlandite. The percentage of portlandite
in the BFS concrete specimens using Siroquant software is shown in Fig. 9.13.

578

Section 9.1 - XRD

12
10
8
6
4
2
365-4B7S1

365-4B7

365-4B5S1

365-4B5

365-4B2S1

0
365-4B2

Percentage Portlandite (%)

Percentage Portlandite in the Binary and Ternary


Blends of BFS Specimens: W/B:0.48

Binary and Ternary Blends of BFS Specimens

Fig. 9.13:

The Percentage Blends of portlandite in the Binary and Ternary Blends of BFS Concrete
Specimens Cast using 0.48 w/b ratio

The percentage of portlandite in the concrete specimens of 25, 50, and 70% BFS were
9.6, 8.8, and 3.7%, respectively.

This indicates that the percentage of portlandite

decreased with increase in the replacement of cement with BFS. This decrease in
portlandite can be attributed to the pozzolanic and carbonation reactions (Hill and Sharp
2002) in the binary blends of BFS. These results are discussed in sub-section 9.1.3.4.
On the other hand, the percentage of portlandite in the ternary concrete blends 4B2S1,
4B5S1, and 4B7S1 were 10.9, 6.2, and 0%, respectively. An interesting observation
was the percentage of portlandite in the ternary blends of silica fume and any percentage
binary blends of BFS was lower compared to their respective binary blends of BFS.
This is because CH released by tricalcium silicate and dicalcium silicate (C3S and C2S)
from the cement quickly reacts with silica fume to form C-S-H. In addition, CH in
solution is necessary for the dissolution of silica fume particles. Therefore in these
ternary blends, silica fume rapidly consumes the CH and makes the slag impossible to
obtain CH for its hydration. However, it is reported that hydration of slag in the form of
rims in these ternary blends become visible at 7 days (Sarkar et al. 1990).

579

Section 9.1 - XRD

In summary, a decrease in the percentage of portlandite was observed in the Blast


furnace slag concrete mixes with increase in the addition of Blast furnace slag, while it
was completely consumed in the specimens of 70% BFS and 10% silica fume.
Nevertheless, the same trend was observed in the blast furnace slag concrete specimens
cast using 0.38 w/b ratio. The percentage of porlandite observed in the plain and
blended concrete specimens cast using 0.48 w/b ratio from Siroquant software is shown
in Table 9.7.

Table 9.7: Percentage of portlandite observed in the different dosage of fly ash and BFS concrete blends
cast using 0.48 w/b ratio

9.1.2.10

S.No.

Concrete mix

Portlandite (%)

365-4F2

22.9

365-4F2S1

14

365-4F5

9.6

365-4F5S1

8.1

365-4B2

9.6

365-4B2S1

10.9

365-4B5

8.8

365-4B5S1

6.2

365-4B7

3.7

10

365-4B7S1

0.0

SUMMARY OF RESULTS OF X-RAY DIFFRACTION


IN THE PLAIN, FLY ASH, AND BFS CONCRETE
SPECIMENS CAST USING 0.48 W/B RATIO

9.1.2.10.1

Fly ash concrete specimens cast using 0.48 w/b ratio

The percentage of portlandite in the fly ash specimens decreased with increase in
the replacement of cement with fly ash.

The ternary fly ash blends exhibited lower amounts of portlandite compared to
their respective binary blends of fly ash.

580

Section 9.1 - XRD

9.1.2.10.2

Blast furnace slag concrete specimens cast using 0.48 w/b


ratio

The percentage of portlandite in the BFS specimens decreased with increase in


the replacement of cement with BFS

The ternary BFS blends exhibited lower amounts of portlandite compared to


their respective binary blends of BFS.

9.1.2.11

XRDA

IN

THE

PLAIN

CONTROL

CONCRETE

SPECIMENS PONDED WITH SODIUM CHLORIDE


SOLUTION
Fig. 9.14 shows the X-ray diffraction peaks of concrete powder collected from the
ponded plain concrete slab 3OPC at three depths namely 5-25, 25-45, and 45-60 mm
cast using 0.38 w/b ratio.

Fig. 9.14: X-ray Diffraction Peaks of the Ponded Plain Concrete Slab Specimen Cast using 0.38 w/b
ratio at three various Depths. A: albite (NaAlSi3O8), P: portlandite (Ca(OH)2), Q: quartz (SiO2), ILL: illite
(K,H3O)Al2Si3AlO10(OH)2,

CLI:

clinochlore

(Ca,Na)(Al,Si)2Si2O8

581

(Mg5Al)(Si,Al)4O10(OH)8;

AN:

anorthite

Section 9.1 - XRD

There was no difference in the peaks of portlandite in ponded control concrete slab
3OPC collected at three different depths of 5-25 mm, 25-45 mm, and 45-60 mm as
observed at 18.1 and 34.1 2. This indicates that the percentage of portlandite in
control concrete slab remained uniform.

9.1.2.12

STATISTICAL SIGNIFICANCE

The statistical significance cannot be performed for these concrete specimens as one
value of response variable (percentage of portlandite) was obtained from the software.

582

Section 9.1 - XRD

USE OF X-RAY DIFFRACTION IN HIGH VOLUME


BLENDED CONCRETE SPECIMENS

9.1.3

DISCUSSION OF RESULTS

9.1.3.1

INTRODUCTION

Cementitious powder was extracted from a piece of concrete specimen and analysed for
the XRD analysis. The first part presents the discussion of XRD analysis for the
concrete specimens cast using 0.38 w/b ratio and exposed for a period of 365 days. In
the latter part, the results of XRD analysis of concrete specimens cast using 0.48 w/b
ratio are discussed. The prime objective of this study was to evaluate the quantitative
percentage of crystalline hydration product of cement namely portlandite (Ca(OH)2)
produced from the hydration reactions in the concrete specimens.

9.1.3.2

Portlandite in the Plain and Silica Fume concrete


specimens cast using 0.38 w/b ratio

The quantitative percentage of portlandite in the control and silica fume concrete
specimens was 10.1 and 5.7%, respectively. The reduction of portlandite in silica fume
concrete specimens is due to the pozzolanic reaction. Sarkar et al. (2001) reported that
silica fume in concrete promptly consumes Ca(OH)2 and the products of pozzolanic
reaction fills the pores of concrete; consequently, the silica fume specimens as shown in
Fig. 5.20 exhibited lower total pore volume compared to the control concrete specimen.
Therefore, the microstructure of silica fume concrete becomes more dense and compact
than the normal concrete.

583

Section 9.1 - XRD

9.1.3.3

Portlandite in the Fly ash concrete specimens cast using


0.38 w/b ratio

The results of this study indicated that the percentage of portlandite in the specimens of
20 and 70% fly ash was 10.5 and 6.1%, respectively. This indicates that decrease in
portlandite was observed with increase in the replacement of cement with fly ash. The
results observed in this study are in agreement with the study conducted by Fu et al.
(2002). They evaluated Ca(OH)2 in paste specimens by extraction, reaction and titration
in the glycol and ethanol blending solution. Their results indicate that control cement
paste specimens showed progressive increase in Ca(OH)2 until monitoring period of 28
days, indicating the degree of hydration of cement.

However, the cement paste

specimens with 65% fly ash showed decrease in Ca(OH)2 after 7 days, this is because of
the lower amount of cement content and Ca(OH)2 is rapidly consumed by fly ash.
Nevertheless, Fu et al. (2002) have used activators and gypsum in fly ash blends in their
study. However, the author of this study has not used any of those activators in any
pozzolanic blends. Thomas (1989) evaluated Ca(OH)2 content in paste specimens cured
in saturated Ca(OH)2 water using thermogravimetric analysis. Fly ash paste with 30%
replacement of cement with 0.44 w/b ratio was compared with control OPC pastes. The
results of his study indicate that an increase in Ca(OH)2 was noticed in fly ash
specimens until seven days and after this period fly ash shows significant reaction
indicated by progressive consumption of Ca(OH)2 until 180 days. Similarly, Halse et
al. (1984) stated that maximum CH in fly ash paste specimen was observed after 7 days
from thermogravimetric analysis with heavy etching of fly ash particles also observed
after seven days from SEM analysis.

Fu et al. (2002) reported that XRD pattern irrespective of the amount of fly ash showed
the diffraction peaks of AFt, C-S-H, Ca(OH)2 and -SiO2 crystalline phases in all the
blends of fly ash at seven days. However, at 90 days, the main hydrates in higher
dosage of fly ash were AFt and C-S-H gel, while, the amount of Ca(OH)2 was lesser in
the higher dosage of fly ash concrete blends. In another study, Jiang et al. (1999)
evaluated the hydration products of HVFA pastes using XRD. The results indicated
that most of the products of hydration were non-crystalline and Ca(OH)2 was the main
product of hydration detected by XRD. In addition, many compositions of fly ash were
584

Section 9.1 - XRD

detected even at 90 days. However, Regourd (1985) reported that hydration products in
hardened Portland cement paste such as ettringite, Ca(OH)2, and AFm in concrete
cannot be observed for longer period of time. In this study, XRD in the fly ash concrete
specimens was investigated after 365 days, therefore, most of the early hydration
reactions such as the ettringite were not observed with the exception of Ca(OH)2.
However, portlandite Ca(OH)2, mullite (3Al2O3.2SiO2), and corundum (Al2O3) was
observed in the high volume fly ash specimens. The mineral phases mostly from the
sand such as albite (NaAlSi3O8) and microcline (KAlSi3O8) were also observed from
the XRD in the fly ash specimens.

The results of this study indicated that Ca(OH)2 content in the concrete specimens with
50% fly ash was 5.0%, while the concrete specimens with 70% fly ash was 6.1%, this
indicates that an increase of 22% was observed in the concrete specimens with 70% fly
ash compared to 50% fly ash. The author has checked the casting details and found
nothing unusual in the blends of 3F2, 3F5, and 3F7. The increase in CH in the 70% fly
ash specimens can be attributed from the study of Feldman (1983) who reported an
increase of 38% Ca(OH)2 at 1.5 years compared to 1.0 year using thermogravimetric
analysis. Feldman stated that Ca(OH)2 in fly ash blends depends on the degree of
hydration, ratio of CaO/SiO2, possible carbonation in some specimens, and formation of
ettringite with some combination of Ca(OH)2. Therefore, the increase in Ca(OH)2 in the
authors study of 70% fly ash at 365 days compared to 50% fly ash can be attributed to
the above mentioned factors. However, in another study, Lam et al. (2000) evaluated
Ca(OH)2 content in high volume fly ash cement pastes using thermogravimetric analysis
in the increasing levels of 0, 25, and 55% replacement of cement with 0.3 and 0.5 w/b
ratio. The results of their study indicate a decrease in Ca(OH)2 with increase in fly ash
content and with age of curing. On the other hand, an interesting observation is that
ternary concrete blend comprising fly ash and silica fume showed lower amount of
portlandite compared to their binary blends of fly ash irrespective of any percentage
dosage of fly ash, while the calcium hydroxide was completely consumed in the ternary
concrete blend of 70% fly ash and 10% silica fume.

Therefore, the use of two

supplementary cementitious material such as fly ash and silica fume are complementary
in the sense that silica fume contributes to the early age strength, which was not
possible with the use of single fly ash alone. Consequently, with the use of combination

585

Section 9.1 - XRD

of both silica fume and low calcium fly ash results in both improved early and longterm strength (Thomas et al. 1999).

The results of this XRD analysis study is supported by Mehta and Gjorv (1982) who
proposed the use of both blends of fly ash and silica fume compared to the fly ash alone
from the point of early concrete strength. This was based on evaluating the free lime in
both binary and ternary blends of cementitious pastes, the free lime in the ternary
cementious blend was 9.5% compared to the 13.4% in the binary blends of fly ash
cementitious paste (Mehta and Gjorv 1982). The decrease in free lime is indicative of
the more consumption of calcium hydroxide in ternary fly ash cementitious paste which
makes the concrete dense compared to the binary fly ash cementitious paste. Therefore,
the use of two supplementary cementitious materials such as silica fume and fly ash
would be of reaping benefits in terms of economy and the durability of concrete. On the
other hand, it is interesting to note from SEM analysis that concrete comprising binary
blends of fly ash exhibited more open structure compared to both ternary blends of
silica fume and fly ash, and the binary blends of silica fume (Scali et al. 1987). This
strongly indicates that pozzolanic reaction in the ternary blends of fly ash and silica
fume is predominant compared to the pozzolanic reaction in the binary blends of fly
ash, consequently, a drastic decrease in the portlandite was observed in the ternary
concrete blend compared to the binary blends of fly ash.

Another way of explaining the mechanism of decrease in portlandite in the ternary fly
ash blends compared to the binary fly ash blends is that addition of silica fume in the
former blends further promptly consumes Ca(OH)2 compared to the latter.
Consequently, the ternary blend comprising higher volume of fly ash and further
addition of silica fume with improper curing makes the concrete porous.

Jun-yuan et al. (1984) evaluated the hydration of fly ash pastes by replacing 35% of
cement with class C fly ash and compared with 100% OPC cement paste using X-ray
diffraction technique supported by the SEM study. The results indicated that Ca(OH)2,
C-S-H, and unreacted clinker phases were often present throughout the hydration of all
the cement pastes studied. Ettringite was observed converting to monosulfate at early

586

Section 9.1 - XRD

hydration stages in the fly ash pastes, however, this conversion was observed at later
stages in the control 100% OPC pastes.

Hill and Sharp (2002) have conducted a study on the minerology and microstructure
using pastes comprising replacement of cement with 75% fly ash and compared the
results with 100% OPC paste. Cement paste prisms of 100 mm x 10 mm x 10 mm were
cast with w/b ratio of 0.35 using distilled water. The samples were air cured at 23 2
C and 86 2% RH in an environmental cabinet with the saturated solution of
potassium sulfate to control the humidity. The results indicate that appreciable amount
of calcium hydroxide was observed in 100% OPC paste specimens after a period of 3
days and remained throughout until 180 days. The increase in Ca(OH)2 is due to the use
of 100% Portland cement which caused increase in the alite Ca3SiO5 from the hydration
reaction

of

cement.

However,

smaller

definite

peaks

of

ettringite

(Ca6[Al(OH)6]2(SO4)3.26H2O) persisted through out the investigation in 100% OPC


paste specimens. On the other hand, the Ca(OH)2 was only detected until 28 day, while
no traces were observed at 90 and 180 days in the cement paste specimens comprising
75% fly ash. The calcium hydroxide in the fly ash paste specimens was consumed
rapidly by the pozzolanic reaction and carbonation phenomenon. The latter occurs from
the carbonation of calcium hydroxide which occurred as the specimens were cured in
air. In addition, the ettringite traces were present in 75% fly ash paste specimens after a
period of three and 28 days and later it was not persistantly present in these specimens.

The portlandite formed in the control and fly ash concrete specimens with 25%
replacement of cement in the authors study was 10.1 and 10.5%, respectively. This
indicates that there was no significant difference in the formation of Ca(OH)2 in these
concrete specimens. However, the ternary blends of 25% fly ash and 10% silica fume
(3F2S1) showed a significant decrease in the percentage of Ca(OH)2 compared to the
binary blend of 25% fly ash (3F2). The results of X-ray diffraction of fly ash concrete
specimens observed from the peaks and software Siroquant software were comparable.

587

Section 9.1 - XRD

9.1.3.4

Portlandite in the Blast Furnace Slag concrete specimens


cast using 0.38 w/b ratio

In this study, the percentage of portlandite decreased with increase in the replacement of
cement by blast furnace slag. The percentage of portlandite in the concrete comprising
25, 50, and 70% replacements of cement with BFS was 15.6, 7.3, and 2.6%,
respectively.

Hill and Sharp (2002) evaluated mineralogy and microstructure in high replacements of
75 and 90% BFS cement pastes, and compared with 100% OPC paste. The results from
the XRD analysis indicated that portlandite formed initially in the 90% BFS cement
paste was lesser compared to the paste comprising 75% BFS, because the former paste
had 10% of cement compared to the 25% cement in the latter. Hill and Sharp (2002)
reported that paste comprising 100% OPC exhibited the highest amount of portlandite,
while in the cement paste with 90% BFS, it was not available after 90 or 180 days
because of pozzolanic and carbonation reactions. Consequently, the alkali activation
from calcium hydroxide in 90% BFS paste was drastically reduced and as a result little
slag hydration occurred with no visible sign of rims around the slag particles. In
summary, the replacement of cement with 90% BFS had the weakest microstructure
with higher level of porosity and prolonged setting time. Therefore, it is important that
more amount of calcium hydroxide should be available to continue the pozzolanic
reaction which increases the hydration reactions of cement and makes the concrete
dense. This is possible only with prolonged period of curing specially with pozzolanic
cements. The amount of portlandite formed in these paste in decreasing order was
100%OPC, 75% BFS, and 90% BFS. Therefore, the results of XRD observed in this
study are in agreement with (Hill and Sharp 2002).

9.1.3.5

Portlandite in the Pozzolanic specimens cast using 0.48 W/B

The water/binder (w/b) ratio represents porosity of concrete and the durability of
concrete depends on this ratio.

The objective of this study was to determine the

588

Section 9.1 - XRD

hydration product mainly calcium hydroxide using two w/b ratios to better
understanding of results. Table 9.8 shows the percentage of portlandite formed in few
fly ash and BFS concrete specimens with both w/b ratios.

Table 9.8: Formation of portlandite in fly ash and BFS concrete specimens expressed in percentage in
two w/b ratios concrete.

Concrete Mix
(W/B : 0.38)
portlandite
(%)
Concrete Mix
(W/B : 0.48)
portlandite
(%)

3F2

3F2S1

3F5

3F5S1

3B5

3B5S1

3B7

3B7S1

10.5

4.3

5.0

1.4

7.3

3.9

2.6

4F2

4F2S1

4F5

4F5S1

4B5

4B5S1

4B7

4B7S1

22.9

14

9.6

8.1

8.8

6.2

3.7

The results of this study indicate that concrete specimens when cast with increase in w/b
ratio of 0.48 exhibit increase in the percentage of portlandite compared to the lower w/b
ratio of 0.38. The portlandite has no strength giving nor cementing properties because
of considerable lower surface area of its structure, it is more readily soluble in water and
may slowly leach out during the exposure, resulting in increased permeability of
concrete. Therefore, one of the factors of high strength concrete is casting with lower
w/b ratio, high cement content, silica fume, and high dosage of superplasticizer which
results in complete consumption of portlandite.

Tarekuddin et al. (2004) evaluated the mineralogy of the concrete specimens comprising
OPC, High Early strength (HES), Type B slag cement (SCB - Slag content is 30 to 60%
of cement mass), and Alumina Cement (AL) exposed until 30 years of marine exposure
with water-cement ratio (w/c) ratio varying from 0.52 to 0.53. Cylindrical concrete
specimens of 150 mm diameter and 300 length were cast and exposed to the submerged
zone in a pool located beside the sea after 28 days of standard curing. The cement
mortar samples were collected from 5 to 15 and 65 to 75 mm from the surface of
cylinders for the mineralogical investigation using XRD at the inner and outer regions.
The results indicated that ettringite peaks were clearly observed at the outer and inner
regions of OPC, the same peaks were also clearly observed only on the outer regions of
slag cement. The portlandite formation was observed at the inner region of the slag

589

Section 9.1 - XRD

cement only. Although no justification was given for the formation of portlandite,
however, the author of this study presumes that the presence of portlandite is the sign of
continuing pozzolanic reaction in the slag cement even after 30 years of exposure in sea
water.

9.1.3.6

CONCLUSIONS OF XRD ANALYSIS VALID FOR


BOTH FLY ASH AND BFS SPECIMENS

It is important that fly ash concrete specimens with replacements of 50 or 70%


should be cured long enough to continue release of lime from the Portland
cement hydration which can react with fly ash.

The formation of portlandite was lower in the ternary fly ash blend comprising
fly ash and silica fume compared to the binary blends of fly ash. This is because
silica fume further promptly reduces portlandite and makes the concrete dense.

The fly ash blends with 25% replacement exhibited higher amount of portlandite
because it has 75% cement content compared to the higher replacement of
cement with fly ash.

Concrete specimens cast using higher w/b ratio exhibit higher amounts of
calcium hydroxide compared to the concrete cast using lower w/b ratio.

The following section covers the use of scanning electron microscopy in plain and
blended cements specimens for hydration characteristics.

References

Aitcin, P. C., and Mindess, S., 1998, 'High Performance Concrete: Science and
Applications', Materials Science of Concrete - Vol. V, J. Skalny, S. Mindess, ed.,
The American Ceramic Society, pp. 477-511.
Aldridge, L. P., 1982a, 'Accuracy and Precision of an X-Ray Diffraction Method for
Analysing Portland Cements', Cement and Concrete Research, vol. 12, no. 4, pp.
437-446.
Aldridge, L. P., 1982b, 'Accuracy and Precision of Phase Analysis in Portland Cement
by Bogue, Microscopic, and X-ray Diffraction Methods', Cement and Concrete
Research, vol. 12, no. 3, pp. 381-398.

590

Section 9.1 - XRD

Berger, R. L., Frohnsdorff, G. J. C., Harris, P. H., and Johnston, P. D., 1964, 'Highway
Research Board', Special Report, vol. 90, pp. 234-253.
Bye, G. C., Varma, S. P., and Moore, A. E., 1975, 'Quantitative X-ray Diffraction
Analysis of Portland Cement Clinker', The State of the Art - 1975 - Mineral
Composition of Clinker, 2nd Cembureau Sub-Committee.
Demoulian, E., Gourdin, P., Hawthorn, F., and Vernet, C., 1980, Proceedings of the 7th
International Congress on the Chemistry of Cement, Paris, vol. 2, no. III, pp.
89-94.
Feldman, R. F., 1983, 'Significance of Porosity Measurements on Blended Cement
Performance', Fly Ash, Silica Fume, Slag, and other Mineral By-Products in
Concrete, V. M. Malhotra, ed., American Concrete Institute, ACI (SP-79),
Montebello, Canada, pp. 415-433.
Fu, X., Wang, Z., Tao, W., Yang, C., Hou, W., Dong, Y., and Wu, X., 2002, 'Studies on
Blended Cement with a Large Amount of Fly Ash', Cement and Concrete
Research, vol. 32, no. 7, pp. 1153-1159.
Goswami, G., Panigrahy, P. K., and Panda, J. D., 1994, 'Rapid Strength Prediction of
Portland Slag Cement by X-Ray Diffractometry', Proceedings of 16th
International Conference on Cement Microsocopy, Richmond, Virginia, USA,
pp. 257-268.
Halse, Y., Pratt, P. L., Dalziel, J. A., and Gutteridge, W. A., 1984, 'Development of
Microstructure and other Properties in FlyAsh OPC Systems', Cement and
Concrete Research, vol. 14, no. 4, pp. 491-498.
Hill, J., and Sharp, J. H., 2002, 'The Mineralogy and Microstructure of Three Composite
Cements with High Replacement levels', Cement and Concrete Composites, vol.
24, no. 2, pp. 191-199.
Hooton, R. D. 1981, 'Ph.D. Thesis in Civil Engineering', McMaster University,
Hamilton, Ontario, Canada.
Jiang, L., Baoyu Lin, and Cai, Y., 1999, 'Studies on Hydration in High Volume Fly ash
Binders', ACI Materials Journal, vol. 96, no. 6, pp. 703-706.
Jun-yuan, H., Scheetz, B. E., and Roy, D. M., 1984, 'Hydration of Fly Ash-Portland
Cements', Cement and Concrete Research, vol. 14, no. 4, pp. 505-512.
Kasselouri, V., Kouloumbi, N., and Thomopoulos, T., 2001, 'Performance of Silica
Fume-Calcium Hydroxide Mixture as a Repair Material', Cement and Concrete
Composites, vol. 23, no. 1, pp. 103-110.
Klug, H. P., and Alexander, L. E., 1974, 'X-Ray Diffraction Procedures for
Polycrystalline and Amorphous Material' Wiley Interscience, New York.
Lam, L., Wong, Y. L., and Poon, C. S., 2000, 'Degree of Hydration and Gel/Space ratio
of High-Volume Fly Ash/Cement Systems', Cement and Concrete Research,
vol. 30, no. 5, pp. 747-756.
Mehta, P. K., and Gjorv, O. E., 1982, 'Properties of Portland Cement Concrete
containing Fly Ash and Condensed Silica Fume', Cement and Concrete
Research, vol. 12, no. 5, pp. 587-595.
Muhamad, M. N., Barnes, P., Fentiman, C. H., Hausermann, D., Pollmann, H., and
Rashid, S., 1993, 'A time Resolved Synchrotron Energy Dispersive Diffraction
Study of the Dynamic Aspects of the Synthesis of Ettringite During Mine
Packing', Cement and Concrete Research, vol. 23, no. 2, pp. 267-272.
Olson, R. A., and Jennings, H. M., 2001, 'Estimation of C-S-H content in a Blended
cement paste using water adsorption', Cement and Concrete Research, vol. 31,
no. 3, pp. 351-356.

591

Section 9.1 - XRD

Parotte, L. J., Geiker M., Gutteridge, W. A., and Killoh, D., 1990, 'Monitoring Portland
Cement Hydration: Comparison of Methods', Cement and Concrete Research,
vol. 20, no. 6, pp. 919-926.
Rashid, S., Barnes, P., and Turillas, X., 1992, 'The Rapid conversion of High Alumina
Cement Hydrates, as revealed by Synchroton Energy-Dispersive X-ray
Diffraction', Advances in Cement Research, vol. 4, no. 14, pp. 61-67.
Regourd, M., 1985, 'Microstructure of High Strength Cement Paste Systems',
Proceedings of the Materials Research Society Symposium, vol. 42, Pittsburgh,
pp. 3-16.
Roode, M. V., Esther Douglas, and Hemmings, R. T., 1987, 'X-Ray Diffraction
Measurement of Glass Content in Fly Ashes and Slags', Cement and Concrete
Research, vol. 17, no. 2, pp. 183-197.
Sarkar, S. L., Aimin, X., and Dipayan, J., 2001, 'Scanning Electron Microscopy, X-Ray
Microanalysis of Concretes', Handbook of Analytical Techniques in Concrete
Science and Technology, V. S. Ramachandran, J. J. Beaudoin, ed., William
Andrew Publishing, Noyes Publications, New York, U.S.A, pp. 231-274.
Sarkar, S. L., Aitcin, P. C., and Djellouli, H., 1990, 'Synergistic roles of Slag and Silica
fume in Very High-Strength Concrete', Cement, Concrete, and Aggregates, vol.
12, no. 1, pp. 32-37.
Sarkar, S. L., and Roy, D. M., 1988, 'QXRD in Cement and Clinker Phase Analysis: Its
Progress and Limitations', Proceedings of the 10th International Conference on
Cement Microscopy, San Antonio, Texas, U.S.A, pp. 285-297.
Scali, M. J., Chin, D., and Berke, N. S., 1987, 'Effect of Microsilica and Fly ash upon
the Microstructure and Permeability of Concrete', Proceedings of the Ninth
Annual International Conference on Cement Microscopy, Reno, Nevada USA,
pp. 375-397.
Schwiete, H. E., and Dolbor, M. A., 1983, 'Forchungs-Bericht des Landes Nordrhein',
Westaflen No. 1186.
Smolczyk, H. G., 1980, 'Slag Structure and Identification of Slag', Proceedings of the
7th International Congress on Chemistry of Cement, Paris, pp. III-1/3-1/17.
St. John, D. A., Poole, A. W., and Sims, I., 1998, 'Spectroscopic Methods of
Identification', Concrete Petrography: A Hand book of Investigative Techniques,
D. A. St John, A. B. Poole, and I. Sims, ed., Arnold Publishers, pp. 44-45.
Struble, L. J., 1991, 'Quantitative Phase Analysis of Clinker using X-ray Diffraction',
Cement, Concrete, and Aggregates, vol. 13, no. 2, pp. 97-102.
Tarekuddin, M., Hidenori Hamada, and Yamaji, T., 2004, 'Concrete After 30 years of
Exposure-Part I: Minerology, Microstructures, and Interfaces', ACI Materials
Journal, vol. 101, no. 1, pp. 3-12.
Tarekuddin, M., Toru Yamaji, and Hamada, H., 2002, 'Chloride Diffusion,
Microstructure, and Mineralogy of Concrete after 15 years of Exposure in Tidal
Environment', ACI Materials Journal, vol. 99, no. 3, pp. 256-263.
Taylor, H. F. W., 1990, 'Cement Chemistry' Academic Press, San Diego, CA.
Thomas, M. D. A., 1989, 'The Effect of Curing on the Hydration and Pore Structure of
Hardended Cement Paste', Advances in Cement Research, vol. 2, no. 8, pp. 181188.
Thomas, M. D. A., Shehata, M. H., Shashiprakash, S. G., Hopkins, D. S., and Cail, K.,
1999, 'Use of Ternary Cementitious Systems containing Silica Fume and Fly ash
in Concrete', Cement and Concrete Research, vol. 29, no. 8, pp. 1207-1214.

592

Section 9.2 - SEM

THE USE OF SCANNING ELECTRON MICROSCOPY IN


HIGH VOLUME BLENDED CEMENTS

9.2.1

INTRODUCTION

In this chapter a brief review of the use of Scanning Electron Microscopy (SEM) in
concrete is presented. Typical hydration products of cement, namely ettringite, and
ettringite related phases will be briefly described. This chapter also covers an overview
of the microstructural development in cement paste hydration, pozzolanic blends, and
the interface between the aggregate and cement. A review of pozzolanic blends is made
in this section with respect to the use of SEM. This chapter further covers different
morphologies of C-S-H and literature review on the usage of SEM by several
researchers.

9.2.2

Scanning Electron Microscopy

Scanning Electron Microscopy (SEM) is a technique that provides both topographic and
compositional analysis of materials.

Many SEMs are equipped with an Energy-

Dispersive Spectrometer (EDS) which detects x-rays emitted by the sample during
electron-beam excitation. The usage of SEM and the EDS does not fall in any standard
testing procedure of concrete and therefore, this technique is relatively new to be used
as investigative tool. The characteristic features of SEM are enhanced resolution, high
magnification, and large depth of field which results in a three-dimensional appearance
of the texture surfaces, and has many applications in cement and concrete. For example,
Rodriguez and Hooton (2002) have used SEM technique to evaluate the chloride ion
ingress in cracks, Hill and Sharp (2002) have used it to evaluate the microstructure in
high volume cements comprising 75% fly ash, 75% BFS, and 90% BFS. Tarekuddin et
al. (2004) have used it for studying the microstructure of pozzolanic concrete blends
after 30 years of marine exposure.

In addition, several researchers Denes and Buck

(1987) and Thaulow and Grelk (1993) have utilized this technique for investigating the
chloride ion penetration in concrete.

Furthermore, the understanding of basic

593

Section 9.2 - SEM

microstructure is important at micro level when water is added to the cement and
aggregate to improve the mechanical properties of concrete, ultimate distribution of
pores, porosity of hydrated cement paste, reaction kinetics of different pozzolanic
materials which can be possible by this technique.

This technique also helps in

knowing the cause, extent, and mechanism of concrete deterioration (Sarkar et al. 2001).
All elements of the periodic table starting from boron and heavier can be detected using
EDS. The output of EDS appears as a spectrum that displays energy peaks of all the
elements proportional to their concentration as detected in the concrete (Rodriguez and
Hooton 2002). Fig. 9.15 shows a typical EDS trace of the microstructure.

Fig. 9.15: Typical EDS Trace of 70% Fly Ash Concrete observed in SEM

The performance of microscopy depends on the resolution of the microscope and its
ability to provide an image which has sufficient contrast to distinguish the observed
phases from the background (Sarkar et al. 2001). The SEM has two modes of operation
that are of prime importance. It has the ability to produce images with surface details in
the range (1 5 nm) with sufficient depth of field to give a three-dimensional effect
(Brink 2005). Secondly, it can be used for the electron beams production of X-rays
which facilitates in analysis of volumes as small as 1 m in diameter (St John et al.
1998d). It is important to note that unlike X-ray diffraction, which gives the analysis of
the mineral phases, the SEM gives the analysis of elements only.

594

Section 9.2 - SEM

The results of SEM are similar to optical microscopy in that they are mainly imaging
output. However, the former technique has the advantages of higher resolution. There
are two types of images produced using SEM namely back scattered electron (BSE)
images and secondary electron (SE) images which provide imaging facility. The BSE
images are generated from the primary high-energy electrons being inelastically
scattered back, thereby generating a contrast that is composition differentiating
(Roode-Gutzmer and Ballim 2001). Both BSE and SE image modes provides imaging
facility, nevertheless, BSE is more oriented for detecting atomic density which can be
related to the atomic number and density of grains forming the object (Sarkar et al.
2001). The specimens for BSE image analysis should be flat and because they are
higher energy, the resultant micrographs does not show details for deep layer e.g. a hole.
It is known that material having higher atomic number shows higher reflection. In
cement, the C4AF phase is brighter due to the presence of iron whose atomic number is
26 compared to calcium and aluminum in C3A phase (Sarkar et al. 2001). However, the
SE mode is more focussed and is used for the topographic contrast. Roode-Gutzmer
and Ballim (2001) stated that the most important limitations of any microscopic
technique is that increased magnification reduces the sampling area and therefore
sample representivity decreases.

9.2.3

Application of SEM in the Microstructure of Concrete

The understanding of microstructure is important in assessing concrete performance.


Concrete in microstructural terms is an extremely complex system of solid phases,
pores, and water with a high degree of heterogeneity.

Ordinary Portland cement

contains oxides known as basic oxides, such as lime (CaO), and other oxides, such as
silica (SiO2), alumina (Al2O3) and iron oxides (Fe2O3) which are called as acidic oxides.
In fact these four oxides comprise of major oxides and together constitute about 90% by
weight of cement. In addition to these four oxides, cement also contains magnesia
(MgO), alkali oxides, (Na2O and K2O), titanium (TiO2), phosphorous pentoxide, and
gypsum which comprise of remaining 10% are called the minor oxides of cement
(Rasheeduzzafar 1994).

The four principal compounds of cement are alite (impure

tricalcium silicate C3S), belite (impure dicalcium silicate C2S); and the aluminate and
ferrite phases which have average compositions of C3A and C4AF, respectively. These
595

Section 9.2 - SEM

minerals react with water to give a variety of hydrates. The calcium silicates react to
give calcium hydroxide Ca(OH)2 referred as CH and calcium silicate hydrate usually
written as (C-S-H) with variable mixture of CaO and SiO2 ratio usually in the order of
1.2 to 2, while the aluminate and ferrite phases react with the added calcium sulfate to
give two groups of product referred to as AFt and AFm (Scrivener 1989). It is known
that hydration of C4AF is similar to C3A, however, it proceeds slowly. Moir (2003)
reported that iron enters into solid solution in the crystal structures of ettringite and
monosulfate substituting for aluminum, therefore, the products of reaction of C4AF is
called as AFt (alumino-ferrite trisulfate hydrate) called as iron substituted ettringite,
while subsequently formed AFm (alumino-ferrite monosulfate hydrate) called as iron
substituted monosulfate, while the products of reaction of C3A are ettringite which
subsequently reacts to form monosulfate (Moir 2003).

The constituents of hydration products are grouped into crystalline and quaziamorphous or amorphous phases.

The crystalline compounds likely to be present

include calcium hydroxide, ettringite (calcium aluminate trisulfate hydrate or high


sulfate) (3CaO.Al2O3.3CaSO4.26-32H2O), calcium aluminate monosulfate hydrate or
low sulfate (3CaO.Al2O3.CaSO4.12H2O), (C-A-H) probably C4AH13, and residual C3S,
-C2S, and C4AF (Diamond 1976). The ettringite in primary form is relatively long (45 m) narrow rods with no branching, whereas secondary ettringite, produced by sulfate
attack from exterior sources of sulfate, tends to be shorter, thicker, and have hexagonal
cross-sections. On the other hand, calcium aluminate monosulfate hydrate is readily
recognizable in the cement paste by their characteristic hexagonal platy microstructure.
The Fig. 9.16 shows the formation of ettringite in air bubbles from the study of Lachemi
et al. (1998).

596

Section 9.2 - SEM

(a)

(b)

Fig. 9.16: (a): The Formation of ettringite in Air Bubbles; (b): Air Bubble filled with ettringite; From the
Study of Lachemi et al. (1998)

Calcium hydroxide (CH) is claimed to be the only hydration product to have


well-defined crystal structure and generally forms as hexagonal plates.

This

morphology changes with hydration process and it is sensitive to the added reagents
(Berger and McGregor 1972). On the other hand, the calcium silicate hydrate in cement
paste is a gel with short range crystallinity and this phase has wide range of
morphologies (Diamond 1976).

The term AFt denotes the phases related to ettringite-calcium aluminum trisulfate, also

written as (C3A.3C S .32H2O) where S represents SO3 and F indicates possible


substitution of iron for aluminum in the structure. This phase is distinguished by the
morphology of hexagonal rods and forms when the concentration of sulfate ion in the
solution is relatively higher (Scrivener 1989), while the AFm phases are isostructural

with calcium aluminum monosulfate (C3A.C S .12H2O). These phases may possibly
substitute aluminum by iron, while hydroxide, carbonate, or chloride ions may replace
the sulfate and they are characterized by hexagonal plate morphology (Scrivener 1989).
The microstructural study conducted by Lessard et al. (1992) on silica fume concrete
after 10 years of exposure in de-icing environment in Canada showed the signs of AFt
with needle morphology.

They attributed the formation of AFm to be natural

phenomenon arising from the hydration of ASTM Type I/II cement. Fig. 9.17 shows
the formation of AFt needles and AFm plates from the study of Lessard et al. (1992).

597

Section 9.2 - SEM

(a)

(b)

Fig. 9.17: (a): AFt Needles in a Vacant Space; (b): AFm Plates in a Large Pore. From the study of
Lessard et al. (1992).

Indirect techniques such as thermogravimetry (TG) and X-ray diffraction (XRD) are
also used to study the microstructure to determine the amount of certain phases in the
sample, while direct or microscopical techniques provide information about the
component phases arranged in the microstructure. The information derived from direct
techniques usually takes the form of images.

Scrivener (1989) reported that

microstructure for a given paste depends on the phases present and their distribution
within the grains, the grains size distribution, the amount of reaction, and the w/c ratio.
Nevertheless, studies indicate that amorphous nature of hydration products of concrete
would likely affect the morphology during the specimen preparation (Scrivener 1989).

In this study, The SEM was used for the morphological observations in concrete
specimens after 28, 91, 180, and 365 days of being in a controlled environmental room
(drying shrinkage conditions) or (36, 99, 188, 373 days after casting) were analyzed for
the elemental composition using the Electron Probe Micro Analysis (EPMA). The
pozzolanic materials such as fly ash and blast furance slag were evaluated using
systematic increase of 25, 50, and 70% replacement, and silica fume at 10% of
replacement of total cementitious material. Concrete specimens of both binary and
ternary blends of pozzolanic material of various percentage blends were evaluated to
determine the optimum percentages that may be replaced with cement. The basic
objective of this study was to identify changes in the hydration products as a result of
blending cement with other pozzolanic materials such as silica fume, fly ash, and blast
furnace slag, The C-S-H gel, ettringite, and ettringite related phases present in the
hydration products were identified and furthermore, the importance of interfacial
transition zone is presented. In summary, the results of SEM were supplemented with
598

Section 9.2 - SEM

pore size distribution and mechanical properties in various percentage blends of


concrete used in this study. The microstructure development in cement paste hydration,
pozzolanic blends, and the interface between the cementitious blend and aggregates are
described briefly below:

9.2.4

Summary of Microstructural development of Cement


Paste Hydration

When water is added to portland cement the minerals begin to ionize, and the ionic
species form hydrated products of lower solubility and precipitate out of the solution.
The hydration products have lower density compared to anhydrous minerals and fills
into previously filled water spaces leading to progressive decrease in consistency. This
process with passage of time consequently leads to increase in strength and decrease in
permeability (Mehta 1983).

The hydration of Portland cement is dictated by the reaction of calcium silicates and
water. The hydrated reaction depends on the relative hydrated proportions of C3A,
sulfates, and CH. The C3A and calcium sulfate reacts rapidly to form ettringite (calcium
aluminate trisulfate hydrate or high sulfate). The ettringite (3CaO.Al2O3.3CaSO4.2632H2O), where (3CaO.Al2O3) is known as C3A, ettringite is produced by a reaction that
requires an excess of sulfate ion over the aluminate phase in the pore solution (St John
et al. 1998b).

The formation of this phase during the setting in plastic stage is

expansive (Neville 1995a; St John et al. 1998b). However, the volume increase of
concrete at that stage is easily accommodated. The ettringite envelops the unhydrated
C3A and prevents rapid setting of cement. The ettringite with time is replaced by
calcium monosulfoaluminate (St John et al. 1998a). As the hardening proceeds more of
the aluminate phase moves into pore solution converting ettringite to monocalcium
sulfo-aluminate hydrate or low sulfate (3CaO.Al2O3.CaSO4.12H2O) (St John et al.
1998b). The other way of explaining the formation of ettringite and monosulfate is that
ettringite is the initial cement hydration product of C3A and water in the presence of
calcium sulfate. The ettringite crystallizes out in lime-saturated pore fluid as short
prismatic rods. After all the calcium sulfate is converted to ettringite, the excess C3A

599

Section 9.2 - SEM

reacts with ettringite itself to form calcium aluminate monosulfate hydrate or low
sulfate
C3A + CaSO4 + H2O 3CaO.Al2O3.3CaSO4.26-32H2O (ettringite)Eqn. (9.3)
C3A

3CaO.Al2O3.3CaSO4.26-32H2O

(ettringite)

3CaO.Al2O3.CaSO4.12H2O

(monosulfate)..Eqn. (9.4)

The formation of ettringite is accompanied by an increase in volume with lower density


of 1.73 g/cm3, compared to an average of 2.5 for other hydration products of ettringite
(Rasheeduzzafar 1994).

The portlandite and hydrated calcium aluminate form

hexagonal platelets, while the ettringite and monosulfate form acicular crystals (St John
et al. 1998a). French (1991) stated that portlandite crystals formed using higher-w/b
ratio are well defined, relatively coarse, and predominantly situated along the aggregate
boundaries and void edges. In addition, Jepsen and Christensen (1989) stated that the
unusually large Ca(OH)2 crystals are indicative of high w/c ratio resulting from
bleeding. However, portlandite crystals formed using lower-w/b ratio are irregular in
shape, smaller, and uniformly distributed through the paste. This study is to evaluate
the hydration products using pozzolanic material such as fly ash and slag in the
percentage blends of 25, 50, and 70%, and silica fume at 10% of total cementitious
material.

The Ca(OH)2 can be dissolved by percolating water and it is re-deposited in open spaces
within the concrete. Therefore, the secondary portlandite appears as a lining to the void
or cracks. Studies by St. John et al. (1998c) indicate that leaching of portlandite often
occurs with secondary ettringite deposits. The formation of ettringite due to the sulfate
ion (Neville 1995b) or the situations conducive when concrete is steam cured and later
reacted with CO2 with monocalcium sulfo-aluminate which also forms ettringite (Kuzel
1990) is not potentially damaging when compared to the situations whereby percolation
of water through concrete to cause conditions that allow the ettringite to reform in the
pores, cracks and fissure which is always potentially damaging (Min and Mingshu
1994).

600

Section 9.2 - SEM

In summary, the tri-sulfate hydrate (C6A S 3H32), popularly known as ettringite


precipitates first out of the system and is responsible for the initial setting and very early
strength up to three days (Mehta 1983). Subsequently with time and depending on the
availability and reactivity of the sulfate, alkali, and aluminate bearing phases in the
cement, either monosulfate hydrate (C4A S H18) or combination of monosulfate hydrate
and calcium aluminate hydrate (C4AH19) forms (Mehta 1983). The reaction products of
the aluminate-sulfate interaction, comprising ettringite, monosulfate hydrate, calcium
aluminate hydrates, and their iron analogs, are capable of contributing strength to
hardened cement paste (Mehta 1983).

The hydration of cement depends on the w/c ratio and degree of curing.

The

compressive strength of concrete is dominated by cement content and degree of


hydration of cement. A very low w/b ratio is beneficial in the mix provided the
concrete is compactable and can lead to under hydration of cement, while the high w/c
ratio can lead to complete hydration but cause large capillary pore volume in the cement
paste and reduces both strength and impermeability of concrete.

9.2.5

Effect of Mineral replacements on Microstructure of


Concrete

9.2.5.1

Microstructure of Fly Ash on Hydration

Fly ash consists of approximately 80% reactive glass comprising calcium, aluminum,
iron, and other minor oxides.

Xu and Sarkar (1996) summarized the hydration

characteristics of fly ash and reported that when cement starts hydration, substantial
amount of ions diffuse to and adsorbed on fly ash forming uniform calcium-rich coating
on fly ash at early stage. This reduces the concentration of hydration products in the
vicinity of cement grains; consequently, the ash particle provides more nucleation sites
for cement hydrates to grow. The SiO44- and AlO2- ionic masses in the glass phase of
fly ash form double layer with Ca2+ which increases the cement hydration process to a
certain extent.

601

Section 9.2 - SEM

During the pozzolanic reaction, fly ash dissociates in alkaline solution and reacts with
calcium ions to form CaO based hydrates. Fraay et al. (1989) reported that solubility of
aluminum and silica ions in fly ash is significantly increased when the pH of solution is
higher than 13.4, and this requires higher amount of cement to hydrate.

Fly ash

spherical particle of fly ash are surrounded by abundant CH crystals, while the surface
is eroded and exposed with mullite crystals which was subsurface. A thin layer of
C-S-H appears between the ash particles and CH crystals indicating start of pozzolanic
reaction.

In another study, Scrivener (1989) stated that glassy components of fly ash

react with portlandite from the cement to produce C-S-H and the surface of fly ash
particles exhibits deposition of C-S-H and AFt from the cement within minutes of
hydration. Halse et al. (1984) reported that crystalline mineral phases such as mullite
(3Al2O3.2SiO2) and quartz (SiO2) appear after several days from the dissolution of glass
phase due to the pozzolanic reaction. Fig. 9.18 (a) shows the appearance of crystalline
material on the fly ash particles indicated by arrow; (b) several particles of fly ash in the
massive region of calcium hydroxide.

(a)

(b)

Fig. 9.18: (a) Appearance of Crystalline Phases indicated by arrow such as mullite and quartz from
within glass due to the Pozzolanic Reaction taken from Halse et al. (1984); (b); Several
Particles of Fly Ash in a massive region of calcium hydroxide and some of which show the
signs of reaction cited by Scrivener (1989).

9.2.5.2

Microstructure of Blast furnace Slag on Hydration

The Blast furnace slags comprise of typically 90% glass, although the BFS are
inherently hydraulic, however, the portlandite is needed to activate the hydration

602

Section 9.2 - SEM

reaction (Scrivener 1989). This is because the amount of cementitious products formed
and the rate of formation due to slag hydration alone are insufficient for application for
structural purposes. Therefore, when slag is used with cement, the hydration of slag is
accelerated by the presence of Ca(OH)2 and CaSO4. The Ca(OH)2 particles grow
around the slag particles and these particles react slowly with Ca(OH)2 over a period of
several months; consequently, darker reaction rims comprising mostly MgO are
observed surrounding the slag particles compared to the hydration products around the
cement grains (Tanaka et al. 1983). However, Harrison et al. (1986) reported that the
reaction rim composed of C-S-H and hydrated magnesium aluminate. Fig. 9.19 (a)
shows the BSE image of cement-slag paste (30/70) hydrated for one day; (b) hydration
reaction rims around the cement and slag particles.

Cement

Slag

(a)

(b)

Fig. 9.19: BSE Image of the Cement-Slag Hydrated for One Day; (b) Hydration Rims around the Cement
and Slag Particles. Cited by (Scrivener 1989).

9.2.5.3

Microstructure of Silica fume on Hydration

Silica fume concrete generally requires superplasticizer to disperse the silica fume
grains, however, some clumps of silica fume still persist and identified as Si on the
microstructure shown in Fig. 9.20. In addition formation of large hollow pores, shown
603

Section 9.2 - SEM

in Fig. 9.20, are the consequence of hydration of large cement grains containing
unreacted ferrite phase (Scrivener 1989).

Si
Si

Si

Fig. 9.20: Clumps of Silica Fume visible on the Microstructure of Mature Cement/Silica fume Paste
(85/15) and the Large Hollow Pores is the consequence of Hydration of Cement Particles
(Scrivener and Pratt 1984)

9.2.6

Interfaces of concrete

The interfacial regions are significantly different from the bulk cement paste in terms of
morphological composition, and density.

A large number of crystals of calcium

hydroxide are generally present in the interfacial zone with a preferential orientation.
Scrivener and Pratt (1986) reported that large CH crystals precipitate at two occasions:
Firstly, when there is relative movement of cement and sand during mixing and settling
of sand grains before setting cement; secondly, areas of localized bleedings at the
interfaces are believed to be the sites at which large CH crystals precipitates. The
thickness of interfacial zone varies from 40 to 50 m. The very weakest part of this
interfacial zone lies 5 to 10 m away from the physical interface within the paste
fraction (Mindess 1989). Fig. 9.21 shows the interfacial transition zone.

604

Section 9.2 - SEM

(a)

(b)

Fig. 9.21: Schematic Representation of Transition zone and Bulk Cement Paste in Concrete (Mehta
1986); (b) Schematic Interface between Non-Reactive Silica Substrate and Type I Cement
Paste (Langton and Roy 1980)

9.2.7

Bonding within the HCP System

The available evidence suggests that strength in HCP is controlled primarily by a


combination of total porosity and stress concentrations due either to pores or other flaws
in the microstructure. The microstructure of cement paste containing a pozzolanic
material should therefore, be superior when compared to plain Portland cement paste.
In the 100% OPC concrete, the microstructure is heterogeneous and characterized by
local areas of high concentration of large pores as well as large crystals of calcium
hydroxide from which microcracks usually originate (Mehta and Aitcin 1990).

Sarkar and Aitcin (1987) made a comparative study of microstructures of normal and
Very High Strength (VHS) concrete. The proportion of portlandite in the VHS concrete
is considerably lower and the portlandite grains in the VHS also contain silicon
indicating the transformation of CH to C-S-H. The main characteristic of very high
strength concrete is the addition of silica fume which forms dense microstructure
resulting from the usage of lower w/b ratio. The C-S-H gel in silica fume absorbs more
amount of alkalies, especially potassium, (Sarkar and Aitcin 1987) and higher presence
of sulphur ions, maintaining lower Ca/Si ratio of (1.4 to 1.8) (Regourd 1985; Sarkar and
Aitcin 1987).

The very low water content cannot hydrate all the cement grains,

consequently, some of the grains remain unreacted and they act as finer aggregate albeit
with higher cost compared to the fine aggregate (sand). The direct contact between the
605

Section 9.2 - SEM

C-S-H and aggregate makes the aggregate-mortar bonding stronger in the VHS
concrete.

Sarkar et al. (1990) evaluated the effect of blending silica fume and BFS on the
microstructure and strength development in which an active pozzolanic blend was
blended with the slow cementitious material. The slag cements exhibits lower early age
strength which is attributed to the slow growth of C-S-H. The lower age compressive
strength is improved by the addition of silica fume, while the progressive hydration of
slag over long period of time increases the later-age strength and the dissolution of
silica fume converts the larger pores into finer pores. The synergistic effects of mixing
slag and silica fume results in the inhibition of ettringite which is due to the reduced
amount of CH formation; consequently, the bond between the aggregate and cement
will be higher.

9.2.8

The Structure of Calcium-Silicate-Hydrate

The different phases present in hardened cement paste in proportion by weight : C-S-H
gel about 70%; Ca(OH)2 about 20%; ettringite and calcium aluminate monosulfate
hydrate about 7% together, unhydrated clinker residue and other minor constituents
perhaps 3% by weight (Diamond 1976).

Diamond (1976) reported that C-S-H gel has a variable composition with typically high

C/S mole ratio ranging from 2 to 3 and the extensive incorporation of S , Al, and Fe
ions and small amounts of K, Na, and other ions. The true structure of C-S-H gel is
unknown and probably locally variable. The surface area of this gel is high (> 900
m2/g) in the saturated condition, but it apparently loses about 2/3 of this internal surface
on d-drying; however, the lost area is recovered on re-saturation. Diamond (1976)
summarized an overview of hydraulic cement paste at several levels and indicated that
there are four different morphological structures of C-S-H. The varieties of C-S-H gel
are Type I - fibrous particles, Type II reticular network, Type III small equant
grains, and Type IV inner product.

An explanation of different morphological

features of C-S-H gel is presented in the following section.

606

Section 9.2 - SEM

9.2.9 Different Morphologies of C-S-H gel particles in cement paste


The most leading C-S-H gel at early ages of cement grain hydration is Type I C-S-H
gel. This is described by various names such as spines, acicular particles, and
fibrous crystals, however, it is mostly referred as fibrous crystals. These fibrous
crystals typically may be anywhere from 0.5 m to about 2 m in length, and are
usually less than 0.2 m thick. Their outlines are not parallel and moreover, their outer
tip is branched into two or more portions. Fig. 9.22 shows the morphology of Type I
C-S-H gel.

(a)

(b)

Fig. 9.22: (a): C-S-H Gel of Type I; (b): Type I C-S-H Gel and ettringite (elongated rods) in Portland
Cement Paste Hydrated for Two Days (Diamond 1976).

The Type II C-S-H occurs at the same period of time with Type I C-S-H. Fig. 9.23
shows the morphology of Type II C-S-H structure. The microstructure shows several
small cement grains enmeshed in the network. Diamond (1976) referred to the
morphology as a reticular network, while the others have cited the names as
interlocking structure and honeycomb morphology. However, reticular network is
retained as the preferred nomenclature. Type II morphology occurs almost universally
in hardened cement paste.

607

Section 9.2 - SEM

Fig. 9.23: Reticular Morphology of Type II C-S-H gel (Diamond 1976).

A third type of C-S-H gel is common in the hardened cement pastes, and comprises
reasonable proportion of total hydration product. It is often not recognized in the
literature because of its non descriptive character. The C-S-H particles of this form is
designated as Type III and described as equant grain morphology with no more
than 0.3 m. Fig. 9.24 is a composite figure showing the appearance of such particles
in a variety of cement pastes examined by Diamond (1976).

(a)

(b)

Fig. 9.24: (a): Equant Grain Morphology of Type III C-S-H gel; b): Mature Cement Paste Showing
mostly Type III C-S-H gel and Ca(OH)2 Deposits in Portland Cement Paste Hydrated for 179
days (Diamond 1976).

Many Researchers have observed significant proportion of volume of C-S-H gel in the
cement grains and termed it as inner product. The inner product morphology is

608

Section 9.2 - SEM

designated as Type IV C-S-H gel shown in Fig. 9.25 and has a dimpled appearance
with either regular pores or close packed equant grains.

Fig. 9.25: Type IV C-S-H gel microstructure hydrated for two days (Diamond 1976)

In summary, the Type I form of C-S-H gel is prominent at early stage of hydration (not
mature pastes), while the Type II morphology occurrs occasionally in the paste. Type
III morphology appears especially as hydration proceeds, although some equant grains
are visible at scattered locations in the paste even at early ages. It is important to note
that observation of the details of the hydration would be difficult as the hydration
proceeds to maturity (Diamond 1976).

Aitcin et al. (1986) evaluated the principal characteristics of cement and aggregates
necessary to produce very high strength concrete. Concrete specimens with and without
silica fume with w/b ratio of 0.24 and 0.27 were cast to achieve the objective of their
study.

The mercury intrusion porosimetry curves in HSC showed similar shapes,

indicating that in practise the properties such as w/c ratio, nature of the coarse
aggregate, and the presence or absence of silica fume do not significantly affect
porosity. Further, these high strength concrete specimens exhibited higher amount of
entrapped air in the fresh concrete. In addition, Delage and Aitcin (1983) observed that
silica fume concretes containing higher dosage of water reducer or superplasticizer
exhibited coarse pores. The morphology of C-S-H in HSC exhibited Type III C-S-H,
while Type II and IV were also observed and this is in agreement with the study of
Diamond (1976). Studies indicate that Type I cement mostly occurs in immature pastes

609

Section 9.2 - SEM

(Diamond 1976). It is known that C-S-H gel comprises of wide range of impurities in
its structure, while Aitcin et al. (1986) observed Na, Mg, Al, S, K, Cl, and Fe in varying
proportions. The portlandite in silica fume is not well crystallized compared to OPC
concrete. The study of Aitcin et al. also indicated cases of unreacted or partly reacted
silica fume spheres with surface cracking indicating desiccation at some stage of
hydration in very high strength concrete.

9.2.10

The use of SEM in blended cements

The microstructure of ready-mix concrete with compressive strength of 135 MPa at six
months was investigated by Aitcin et al. (1983). In their study, the selection of material
was mostly based on macrostructural tests by optimizing the composition with local
available materials. The coarse aggregates were selected based on the tests such as
compressive strength and modulus of elasticity on rock cores from different quarries.
The best coarse aggregate was selected on the highest compressive strength and the
lowest elastic modulus. They proposed that elastic modulus of aggregate should be
similar or close to the mortar paste to have more homogenous materials to avoid
differential strains at the interface, while the fine aggregate with fineness modulus of
3.0 is preferable in very high strength concrete. The microstructural study of concrete
after six months indicated dense C-S-H with some un-reacted grains of C3S, portlandite,
and silica fume particles. This was attributed to the very low w/b ratio and due to very
high capacity of the hydrated cement paste. The results of SEM indicated that some
unreacted silica fume spheres persisted in the cement paste and not all the silica fume
grains reacted with the portlandite liberated by the hydration of the C3S and C2S.
However, Fig. 9.26 shows unreacted silica fume particles in the microstructure.

610

Section 9.2 - SEM

Fig. 9.26: The Unreacted Silica Fume Particles Observed in the Microstructure (Aitcin et al.
1983)

It is important to note that portlandite crystals in silica fume concrete are strongly
reacted as shown in Fig. 9.27 compared to the well defined portlandite crystal structure
in Portland cement concrete. In addition, the microstructure of silica fume concrete also
indicated un-hydrated C3S grains at various places.

Fig. 9.27: Strong Reaction of Silica Fume on the Portlandite Crystals in Silica Fume Concrete (Aitcin et
al. 1983)

The effective bonding between the aggregates and the C-S-H in silica fume concrete is
due to the absence of portlandite which is due to both pozzolanic action of silica fume
and usage of low w/b ratio indicated in Fig. 9.28.

611

Section 9.2 - SEM

Fig. 9.28: The Reaction of Silica Fume with Aggregates at Some Places making the Concrete Dense
(Aitcin et al. 1983).

Bache (1981) and Birchdall et al. (1981) have stated that maximum density in HPC
concrete can be achieved provided it is free from macro defects which reflects in either
higher compressive or tensile strength testing. Therefore, it is necessary to limit the
water/binder ratio of concrete to very minimum. This minimum amount of water
should be necessary only to hydrate the cement grains participating in the initial setting
of the cement and should not completely hydrate the cement grains (Aitcin et al. 1983).
Aitcin et al. proposed that water can be added later for complete hydration of cement
grains through the pore system after setting concrete.

The results of compressive

strength using 100 mm diameter and 200 mm length at 28 and 180 days were 107.7 and
134.9 MPa, respectively. The silica fume concrete mix comprised of Type III cement
using 0.26 w/b ratio with slump of 105 mm by adding 2.8% superplasticizer and 1.4%
of retarding agent.

The pore size distribution of very high strength concrete was

characterized by total absence of intermediate pores in the range of (350 to 27,000)


and refinement of the very fine pores and may have fewer macropores (Aitcin et al.
1983). The results of pore size distribution in silica fume concrete indicated that there
is less intrusion of mercury per unit mass of concrete when the w/b ratio decreased to
0.25. Delage and Aitcin (1982) reported the disappearance of the intermediate pore size
family is closely related to the pozzolanic reaction of condensed silica fume.

Regourd (1985) reported that crystalline products such as ettringite, portlandite, and
AFm phase appear on specimens of few days old and they disappear with age. He
stated that concrete with higher amount of silica fume showed a lower C/S ratio, while

612

Section 9.2 - SEM

they contain more alkalis.

Consequently, the Ca/Si ratio of C-S-H measured by

Electron Probe Microanalysis after 200 days of hydration using mixes of Portland
cement, Portland cement with 13% silica fume, and Portland cement with 28% silica
fume was 1.6, 1.3, and 0.9, respectively. The characteristics of high strength cement
paste system is indicated by the presence of amorphous C-S-H and virtually no pores in
the range of 500 to 5000 and refinement of pores in the range of 50 to 500 . Studies
in the literature as cited by Regourd (1985) indicated that high strength concrete was
observed when the aggregate and cement paste have similar mechanical properties and
when the coarse and fine aggregates are produced from the same rock.

Lessard et al. (1992) evaluated the long-term behaviour of silica fume concrete cast on
walkway section. Silica fume was added in the percentage replacements ranging from
10 to 20% and cement content varying from 195 kg/m3 to 405 kg/m3 were used. The
properties as compressive strength, chloride ion permeability, and microstructural
characteristics were evaluated after 10 years. In their study, four silica fume concrete
mixes with each two of mixes with low and higher w/b ratio were cast. The objective of
their study was to know if any deterioration occurred over a period of 10 years.

The results of Lessard et al. indicated formation of well crystallized Ca(OH)2 in the
silica fume concrete even after a period of ten years. The trend of CH formation
indicates that it is function of cement content and that it is somewhat interdependent on
the w/b ratio. Silica fume concrete comprising higher amount of cement caused micro
fissures and partly hydrated cement grains in this concrete. One of the silica fume
concrete mixes showed the formation of AFm hexagonal plates which is said to be
natural phenomenon arising from the hydration of Type I/II cement, while in another
silica fume concrete mix chloroaluminate was observed. Lessard et al. attributed the
formation of chloroaluminate to the percolation of Cl ions from the de-icing salts.
However, the appearance of the AFt hexagonal rods was due to the unstable nature of
chloroaluminate. An interesting observation of microstructure of silica fume concrete
from one year to 10 years had shown mineralogical alterations. One of the silica fume
concrete mixes showed the presence of AFm phase after ten years, however, it was not
observed at one year. On the other hand, another silica fume concrete specimen showed
the evidence of AFt and chloroaluminate, the latter mineral was most likely formed due

613

Section 9.2 - SEM

to the percolation of de-icing salts.

The aggregates indicated no evidence of

deterioration and the paste-aggregate bonding was still strong and there was no evidence
of cement components leaching out due to the increased permeability. The four field
concretes in which silica fume was used were still in stable conditions with satisfactory
durability properties even after 10 years of service.

The normal hardened Portland cement paste is heterogeneous and porous material.
Most of the hydrate appears as an amorphous C-S-H gel which takes variety of
morphologies (Diamond 1976). Of the crystalline products of hydration, ettringite is in
the form of rods or fibers with Ca(OH)2 and AFm phase as hexagonal platelets. These
morphologies are observed most clearly on the fracture surface of specimens a few days
old and they become less distinct with age. The overall porosity of 25 to 30% in
volume comprises of two pore families such as gel poles related to the silicate C-S-H
structure with few nanometers in size; capillary pores between the various hydrates
formed of air bubbles and the microcracks from several hundred nanometers to a few
millimetres in size or length.

The pozzolanicity of silica fume can be characterized by its ability to react with
Ca(OH)2. In the concrete mix comprising 30% replacement of silica fume all the
Ca(OH)2 liberated by the hydration of clinker silicates has been consumed by the silica
fume. The concrete with higher amounts of silica fume showed lower C/S ratio and
inversely they contain more alkalis. The Ca/Si ratio of C-S-H measured by Regourd
(1985) using Electron Probe Microanalysis (EPM) after 200 days of hydration in
mixtures of Portland cement, Portland cement with 13% silica fume and Portland
cement with 28% silica fume was 1.6, 1.3, and 0.9%, respectively.
Bache (1981) developed compressive strength of 110-160 MPa cast using 480 kg/m3 of
total cementitious content with 16.7% silica fume and superplasticizer of 1-4%
corresponding to w/b ratio of 0.16 and 0.18 in densified system concrete. The most
important characteristics of densified concrete is the presence of an amorphous C-S-H
which occupies majority of the space in the matrix decreasing the capillary porosity and
increasing the finest porosity with practically no pores between (500 5000 ) and the
refinement of pores in the range between (50 to 500 ). The elemental composition of

614

Section 9.2 - SEM

C-S-H determined by EPM analysis in silica fume concrete exhibits lower Ca/Si ratios
than in Portland cement. Bache (1981) reported that concrete comprising silica fume
with 40% replacement becomes brittle and some silica particles remain un-reacted
(Regourd 1985).

Jiang et al. (1999) evaluated the morphological and microchemical features in high
volume fly ash paste binders at 3, 7, 28, and 90 days by using SEM with an Energy
Dispersive X-ray Analyser (EDXA). The results indicated that fly ash particles undergo
significant morphological changes with curing. Fly ash with lower replacement and
lower w/b ratio exhibited dense structure, while, fly ash in high volume fly ash concrete
was not fully hydrated. This is attributed from the fact that even after 90 days, there
were many fly ash particles embedded in hydration products. The products of hydration
observed by SEM were C-S-H, Ca(OH)2, and AFt.
Sarkar (1993) investigated the long-term properties of HPC cast in field exposed to the
effects of environmental and climatic conditions using silica fume concrete.

Two

concrete columns were cast in the basement of a 26-storey building in Montreal,


Quebec, Canada.

The two columns were identical in geometry, composition, and

structure, however, only one column was load-bearing.

The compressive strength

measured on 100 mm diameter and 200 mm length showed an increase in compressive


strength from 28 days. The core specimens from the column tested after two, four, and
seven years ranged from 83 to 87.5 MPa.

The microstructure after seven years

indicated that the paste had retained the dense C-S-H, however, with a small
development of CH crystals and ettringite needles, although their appearance (CH and
ettringite) is rare in such concretes (Aitcin et al. 1986). Sarkar (1993) attributed this unusual appearance to non-homogeneous dispersion of silica fume causing some parts of
concrete devoid of silica fume showing the characteristics of normal concrete with very
low w/c ratio, resulting in the generation of some CH and ettringite. In addition, the
paste-aggregate bonding was strong as observed after two and four years.

References

615

Section 9.2 - SEM

Aitcin, P. C., Regourd, M., and Bedard, C., 1983, 'Microstructural Study of a 135 MPa
(19500 psi) Ready mix Concrete', Proceedings of the Fifth Annual International
Conference on Cement Microscopy, Nashville, Tennessee USA, pp. 164-179.
Aitcin, P. C., Sarkar, S. L., and Diatta, Y., 1986, 'Microstructural Study of Different
Types of Very High Strength Concretes', Symposium on Microstructural
Development During the Hydration of Cement, Materials Research Society, vol.
85, pp. 261-272.
Bache, H. H., 1981, 'Densified Cement/Ultra-fine Particle-based Materials', Second
International Conference on Superplasticizers in Concrete, Ottawa, pp. 1-35.
Berger, R. L., and McGregor, J. D., 1972, 'Influence of Admixtures on the Morphology
of Calcium Hydroxide Formed During Tricalcium Silicate Hydration', Cement
and Concrete Research, vol. 2, no. 1, pp. 43-55.
Birchdall, J. D., Howard, A. J., and Kendall, K., 1981, 'Flexural Strength and Porosity
of Cements', Nature, vol. 289, no. 5796, pp. 388-390.
Brink, F., 2005, 'Discussion on SEM', Personal Communication, Senior Technical
Officer, Australian National University, Electron Microscopy Unit.
Delage, P., and Aitcin, P. C., 1982, 'Effect of Condensed Silica Fume on the Porosity of
Field Concrete', 84th Annual Meeting of the American Ceramic Society,
Cincinatti.
Delage, P., and Aitcin, P. C., 1983, 'Industrial and Engineering Chemistry - Product
Research and Development', vol. 22, pp. 286-290.
Denes, T., and Buck, A. D., 1987, 'An Experiment to Investigate Chloride Intrusion on
Construction Joints in Concrete', Cement, Concrete, and Aggregates, vol. 9, no.
2, pp. 80-81.
Diamond, S., 1976, 'Cement Paste Microstructure - An Overview at Several Levels',
Hydraulic Cement Pastes: Their structure and Properties, Cement and Concrete
Association , Slough, U. K., pp. 2-30.
Fraay, A. L. A., Bijen, J. M., and de Haan, Y. M., 1989, 'The Reaction of Fly Ash in
Concrete - A Critical Examination', Cement and Concrete Research, vol. 19, no.
2, pp. 235-246.
Halse, Y., Pratt, P. L., Dalziel, J. A., and Gutteridge, W. A., 1984, 'Development of
Microstructure and other Properties in FlyAsh OPC Systems', Cement and
Concrete Research, vol. 14, no. 4, pp. 491-498.
Harrison, A. M., Winter, N. B., and Taylor, H. F. W., 1986, 'An Examination of Some
Pure and Composite Portland Cement Pastes using Scanning Electron
Microscopy with X-ray Analytical Capability', Proceedings of the 8th
International Congress on the Chemistry of Cement, Finep, Rio de Janeiro,
Brazil, pp. 170-175.
Hill, J., and Sharp, J. H., 2002, 'The Mineralogy and Microstructure of Three Composite
Cements with High Replacement levels', Cement and Concrete Composites, vol.
24, no. 2, pp. 191-199.
Jepsen, B. B., and Christensen, P., 1989, 'Petrographic Examination of Hardened
Concrete', Bulletin of the International Association of Engineering Geology, vol.
39, no., pp. 99-103.
Jiang, L., Baoyu Lin, and Cai, Y., 1999, 'Studies on Hydration in High Volume Fly ash
Binders', ACI Materials Journal, vol. 96, no. 6, pp. 703-706.
Kuzel, H. J., 1990, 'Reactions of CO2 with Calcium Aluminate Hydrates in Heat
Treated Concrete', Proceedings of the Twelfth Annual International Conference
On Cement Microscopy, Vancour, Canada, pp. 219-227.

616

Section 9.2 - SEM

Lachemi, M., Li, G., Tagnit-Hamou, A., and Aitcin, P. C., 1998, 'Long-term
Performance of Silica Fume Concretes', Concrete International, vol. 20, no. 1,
pp. 59-65.
Langton, C. A., and Roy, D. M., 1980, 'Morphology and Microstructure of Cement
Paste/Rock Interfacial Regions', 7th International Congress on the Chemistry of
Cement, pp. VII-127-VII-132.
Lessard, M., Sarkar, S. L., Ksinsik, D. W., and Aitcin, P. C., 1992, 'Long-term Behavior
of Silica Fume Concrete', Concrete International, vol. 14, no. 4, pp. 25-30.
Mehta, P. K., 1983, 'Pozzolanic and cementitious By-products as Mineral Admixtures
for Concrete - A Critical Review', First International Conference on Fly Ash,
Silica fume, Slag, and Other Mineral By-Products in Concrete, ACI (SP-79),
Montebello, Canada, pp. 1-46.
Mehta, P. K., 1986, 'Concrete: Structures, Properties and Materials' Prentice-Hall,
Englewood Cliffs, NJ.
Mehta, P. K., and Aitcin, P. C., 1990, 'Principles underlying Production of High
Performance Concrete', Cement, Concrete, and Aggregates, vol. 12, no. 2, pp.
70-78.
Min, D., and Mingshu, T., 1994, 'Formation and Expansion of Ettringite Crystals',
Cement and Concrete Research, vol. 24, no. 1, pp. 119-126.
Mindess, S., 1989, 'Interfaces in Concrete', Materials Science of Concrete I, J. P.
Skalny, ed., The American Ceramic Society, Westerville, OH, pp. 163-180.
Moir, G., 2003, 'Hydration of C3A and C4AF', Advanced Concrete Technology Constituent Materials, J. Newman, B. S. Choo, ed., Elsevier Publishers, pp.
1/16-1/17.
Neville, A. M., 1995a, 'Shrinkage-Compensating Concrete', Properties of Concrete, A.
M. Neville, ed., Addison Wesley Longman Limited, pp. 448.
Neville, A. M., 1995b, 'Sulfate Attack on Concrete', Properties of Concrete, A. M.
Neville, ed., Addison Wesley Longman Limited, pp. 508-511.
Rasheeduzzafar, 1994, 'Lecture Notes', Personal Communication, Deceased - Formerly
Professor, Dept. of Civil Engineering, King Fahd University of Petroleum and
Minerals, Dhahran; Saudi Arabia.
Regourd, M., 1985, 'Microstructure of High Strength Cement Paste Systems',
Proceedings of the Materials Research Society Symposium, vol. 42, Pittsburgh,
pp. 3-16.
Rodriguez, O. G., and Hooton, R. D., 2002, 'Scanning Electron Microscopy in studying
Chloride Ingress into Cracked Concrete', Proceedings of the Twenty-Fourth
International Conference on Cement Microscopy, San Diego, California, pp.
286-303.
Roode-Gutzmer, Q. I., and Ballim, Y., 2001, 'Phase Composition and Quantitative XRay Powder Diffraction Analysis of Portland Cement and Clinker', Materials
Science of Concrete VI, S. Mindess, and Skalny, J., ed., The American Ceramic
Society, Ohio, pp. 1-45.
Sarkar, S., Aitcin, P. C., and Djellouli, H., 1990, 'Synergistic roles of Slag and Silica
Fume in High Strength Concrete', Cement, Concrete, and Aggregates, vol. 12,
no. 1, pp. 32-37.
Sarkar, S. L., 1993, 'Performance of a High-Strength Field concrete at 7 Years',
Concrete International, vol. 15, no. 1, pp. 39-42.

617

Section 9.2 - SEM

Sarkar, S. L., Aimin, X., and Dipayan, J., 2001, 'Scanning Electron Microscopy, X-Ray
Microanalysis of Concretes', Handbook of Analytical Techniques in Concrete
Science and Technology, V. S. Ramachandran, J. J. Beaudoin, ed., William
Andrew Publishing, Noyes Publications, New York, U.S.A, pp. 231-274.
Sarkar, S. L., and Aitcin, P. C., 1987, 'Comparative Study of the Microstructures of
Normal and Very High-Strength Concretes', Cement, Concrete, and Aggregates,
vol. 9, no. 2, pp. 57-64.
Scrivener, K. L., 1989, 'The Microstructure of Concrete', Materials Science of Concrete
I, J. P. Skalny, ed., The American Ceramic Society, pp. 127-161.
Scrivener, K. L., and Pratt, P. L., 1984, 'Microstructural Studies of the Hydration of C3A
and C4AF Independently and in Cement Paste', The Chemistry and Chemically
Related Properties of Cement, pp. 207-220.
Scrivener, K. L., and Pratt, P. L., 1986, 'A Preliminary Study on the Microstructure of
the Cement/Sand Bond in Mortars', 8th International Congress on the Chemistry
of Cement, Finep, Rio de Janeiro, Brazil, pp. 466-471.
St John, D. A., Poole, A. B., and Sims, I., 1998a, 'C-S-H and the Microstructure of
Cement Paste', Concrete Petrography: A Handbook of Investigative Techniques,
D. A. St John, A. B. Poole, and I. Sims, ed., Arnold Publishers, pp. 94-96.
St John, D. A., Poole, A. B., and Sims, I., 1998b, 'Ettringite', Concrete Petrography : A
Handbook of Investigative Techniques, D. A. St John, A. B. Poole, and I. Sims,
ed., Arnold Publishers, pp. 192-193.
St John, D. A., Poole, A. B., and Sims, I., 1998c, 'Portlandite (Ca(OH)2)', Concrete
Petrography : A Handbook of Investigative Techniques, D. A. St John, A. B.
Poole, and I. Sims, ed., Arnold Publishers, pp. 99-101.
St John, D. A., Poole, A. W., and Sims, I., 1998d, 'Scanning Electron Microscopy',
Concrete Petrography; A Handbook of Investigative Techniques, D. A. St John,
A. B. Poole, and I. Sims, ed., Arnold Publishers, pp. 46.
Tanaka, H., Totani, Y., and Saito, Y., 1983, 'Structure of Hydrated Glassy Blast Furnace
Slag in Concrete', Proceedings of the 1st Conference on Fly Ash, Silica Fume,
Slag, and Natural Pozzolans in Concrete, pp. 963-977.
Tarekuddin, M., Hidenori Hamada, and Yamaji, T., 2004, 'Concrete After 30 years of
Exposure-Part I: Minerology, Microstructures, and Interfaces', ACI Materials
Journal, vol. 101, no. 1, pp. 3-12.
Thaulow, N., and Grelk, B., 1993, 'Micro Defects Influence on Chloride Ingress into
Concrete', Phase 1.1: Preliminary Tests on Chloride Ingress by Diffusion and
Pentration; Storebeltsforbindelsen A/S (Great Belt Link A/S), Copenhagen.
Xu, A., and Sarkar, S. L., 1996, 'Fly Ash and its Hydration Characteristics in
Cementitious Matrix', Fourth International Conference on Concrete Technology
in Developing Countries, Gazimagusa, Turkish Republic of Northern Cyprus,
pp. 109-119.

618

You might also like