Journal of Advanced Concrete Technology Vol. 4, No.

3, 357-367, October 2006 / Copyright 2006 Japan Concrete Institute

357

Scientific paper

XRD/Rietveld Analysis of the Hydration and Strength Development of

Slag and Limestone Blended Cement
Seiichi Hoshino1, Kazuo Yamada2 and Hiroshi Hirao1
Received 1 July 2006, accepted 5 October 2006

Abstract
The X-ray diffraction (XRD)/Rietveld method was applied to analyze the hydration progress of cement paste prepared
with blast furnace slag powder (BFS) and limestone powder (LSP). The Rietveld method can be used to quantify the
amount of amorphous phase as well as crystalline phases. However, in cement paste containing BFS, two kinds of
amorphous phases, unhydrated BFS and C-S-H, are included and the Rietveld method cannot distinguish them. In order to
discriminate these two phases, a selective dissolution method was used to quantify the amount of unhydrated BFS. By
combining these two methods, it became possible to quantify the amount of BFS and C-S-H in hydrated cement. The
analysis results show that BFS accelerates the hydration of C3S, C3A, and especially C4AF. The strength increasing effect
of LSP is more pronounced in the ordinary Portland cement (OPC) BFS system than in the OPC system. This is attributed to the higher amount of calcium aluminate hydrates generated in the BFS contained system compared to the
system without BFS. The additional calcium aluminate hydrates fill pores as calcium carboxaluminate by the reaction
with LSP, increasing the compressive strength. These results show the effectiveness of the XRD/Rietveld method as an
analysis tool for cement hydration relating to various properties of cement materials.

1. Introduction
The X-ray diffraction (XRD)/Rietveld method has been
used and applied for the quantification of crystalline
phases such as cement minerals. Recently, it has also
been applied for the quantitative analysis of cement
phases containing various mineral admixtures (Hoshino
et al. 2005) and hydration analysis of cement paste
(Scrivener et al. 2004, Hoshino et al. 2006). One characteristic point of these recent applications is the quantification of amorphous phases.
In most previous studies of hydration analysis using
the Rietveld method, the materials analyzed consisted
mainly of pure cement or cement containing a crystalline
mineral admixture such as limestone powder (LSP) but
were very limited for cement containing amorphous
phases such as blast furnace slag powder (BFS). In general, the quantification of amorphous phases by the
Rietveld method is done by the addition of a specific
amount of internal standard such as corundum (Al2O3).
The quantified amount of corundum is greater than the
amount added when the sample contains an amorphous
phase, and the amount of amorphous phase is estimated
from this measured value (Jones et al. 2000). The estimated value is the total amount of amorphous phases.
Therefore, when the sample contains more than two
phases of amorphous materials, it is generally impossible
to estimate the amount of each amorphous phase by this
1

Researcher, R&D Center, Taiheiyo Cement Corp.,

Japan.
2
Manager, R&D Center, Taiheiyo Cement Corp., Japan.
E-mail: kazuo_yamada@taiheiyo-cement.co.jp

method.
When cement paste contains BFS, two kinds of
amorphous phases are considered to exist. One of them is
C-S-H generated by the hydration of cement and BFS,
and the other is the unhydrated BFS. Both phases show a
diffraction halo at a similar position around 2 = 31
(CuK) in XRD profiles. Therefore, it is difficult to
distinguish these two phases only by the Rietveld method
and the hydration analysis of BFS blended cement requires advanced techniques such as ones that combine
multiple methods. As for the quantitativity of the Rietveld method, the accuracy of quantification has to be
verified for every phase separately through comparison
with other reliable methods because of the large number
of adjustable parameters and several factors not usually
considered such as solid solution, polymorphism, and
differences in X-ray adsorption depending on chemical
composition.
Regarding the hydration reaction of BFS blended
cement, some studies have reported about the hydration
process of individual cement minerals. BFS is reported to
accelerate the hydration of C3S (Ogawa et al. 1980,
Uchikawa 1986, Imoto et al. 2001), to delay the hydration of C3A estimated from the analysis of C3A-mellilite
glass system (Hanada and Ito 1969), and to delay the
hydration of C4AF at an early age estimated from the
analysis of the C4AF-BFS system (Asaga and Kuga
1996). However, studies on the hydration analysis of
BFS blended cement have been very limited. One objective of this study is the establishment of a method to
quantify BFS blended cement hydration.
Another objective of this study is the investigation of
the mechanism of LSP influence on the strength of cement paste. The effect of LSP on the hydration of C3S has

358

S. Hoshino, K. Yamada and H. Hirao / Journal of Advanced Concrete Technology Vol. 4, No. 3, 357-367, 2006

been reported (Ramachandran 1986), as has the interaction between calcium aluminate phase and LSP (Jambor
1980). Based on these studies, the hydration of BFS
blended cement containing LSP was analyzed in this
study. It has been reported that LSP increases the strength
of cement and that this effect is especially significant in
the system of BFS blended cement (Shimobayashi et al.
1979). Although this kind of phenomena has been reported, the mechanism of this strength increase is not
known and there are no studies from the standpoint of
hydration analysis to the authors knowledge.
Based on this background, a combined
XRD/Rietveld-selective dissolution method was adopted
to quantify the hydration of cement containing BFS and
LSP. The measurement results were used to estimate the
mechanism of strength development. Through this investigation, the authors tried to show the effectiveness of
this combined analytical method for the properties of
cement based materials.

2. Experiments
2.1 Sample preparation
The materials used in this study were ordinary Portland
cement (OPC), BFS, and LSP with 3250 cm2/g, 4480
cm2/g, and 5270 cm2/g of Blaine specific surface area,
respectively. The chemical compositions of OPC and
BFS are shown in Table 1. The mineral composition of
OPC is shown in Table 2. The mix proportions of the
cement and additives prepared in this study are shown in
Table 3. Four kinds of blend were examined, OPC, OPC
with 4 mass% of LSP by substitution, OPC replaced by
40 mass% BFS, and 40 mass% BFS-OPC with 4 mass%
of LSP by substitution.
Cement paste with a water-to-cement ratio of 0.50 was
mixed in a Hobart mixer for 3 min. The mixed paste was
kept in a sealed container at 20C for 3, 7, and 28 days. At
each age, hydration was stopped by immersion into
acetone and the specimen was kept under RH11% for 7
days before being used for analysis.
2.2 Measurements
(1) Thermal gravity and differential thermal
analysis (TG-DTA)
Calcium hydrate (CH) and ignition loss (1000C, ig.loss)
was quantified by using TG-DTA. In the analysis done in
this study, the value was calibrated to eliminate the effects of mass loss by the decarbonation of calcite and the
dehydration of gypsum.

Table 2 Mineral composition of OPC.

C3 S
59.6

C2 S
12.6

C3 A
6.8

C4AF
14.1

Gypsum
4.0

Clinker Minerals: by Rietveld Method, Gypsum: by

TG-DTA
Table 3 Mix proportion of cement.

Mixing Ratio (mass%)

LSP (subOPC BFS
stitution)
OPC
100

OPC-LSP
100
4

OPC-BFS
60
40

OPC-BFS-LSP 60
40
4
Sample Name

(2) XRD/Rietveld method

The sample measured by XRD was prepared by mixing a
paste whose hydration was stopped through immersion
into a large amount of acetone and 10 mass% -Al2O3
(corundum) with an average particle size of 0.4 m. The
measurement conditions of XRD were a tube voltage of
50 kV, tube current of 250 mA, scan range of 2 = 5 to
65, step width of 0.02, and scan speed of 2/min. The
software used for Rietveld analysis was TOPAS (Bruker
AXS). The phases that were quantified were calcite,
gypsum, bassanite, portlandite, ettringite (Ett), monosulfate (Ms), monocarbonate (Mc), hemicarbonate (Hc)
depending on the sample, as well as basic minerals such
as alite (C3S), belite (C2S), cubic-C3A
orthorhombic-C3Aand C4AF. Regarding the amount of
C3A, although two kinds of C3A were measured by the
Rietveld method, for further investigation, both phases
were treated as a sum.
Structural models for clinker minerals (alite, belite,
cu-C3A, or-C3A, and C4AF) were taken from the NIST
Technical Report (Stutzman et al. 2002). Those for calcite, gypsum, bassanite, CH, Ett, Ms, Mc and corundum
were taken from the ICSD database (Fachinformationszentrum 2006). There is no definite structure model
established for Hc phase. Therefore, in this study, the
structural data of synthesized Hc refined by Rietveld
analysis was used. Hc crystal was synthesized by mixing
of stoichiometric proportions of C3A, CaCO3, Ca(OH)2
and water according to Li et al. (1997). The crystal sys-

Table 1 Chemical composition of OPC and BFS.

OPC
BFS

SiO2 Al2O3 Fe2O3 CaO MgO

18.56 6.61 3.74 63.23 1.66
32.31 13.70 0.30 42.47 6.10

SO3
3.02
2.01

Na2O
0.36
0.21

K2O
0.40
0.41

TiO2
0.37
0.56

P2O5
0.42
0.02

359

S. Hoshino, K. Yamada and H. Hirao / Journal of Advanced Concrete Technology Vol. 4, No. 3, 357-367, 2006

100

where G is the amount of amorphous phase (mass%), R is

the added ratio of corundum (mass%), and A is the
measured value of corundum (mass%).
Regarding the quantified values of each phase in the
hydrated sample, the mass that the ig.loss amount was
subtracted from the mass of the original sample is shown,
in other words, the value is the mass ratio against the
unhydrates. In this study, LSP was added as substitution,
and the amount of LSP and carbonate hydrates generated
from the reaction with LSP is expressed as the substitution of the system.
The quantification accuracy of cement minerals, hydrates and amorphous phase has been reported in other
articles (Hoshino et al. 2005, 2006). In the investigation
using unhydrated cement with blast furnace slag as the
amorphous phase, a maximum difference from the true
value of the quantified amount of C2S was 2.2 mass%
and that of C3S, C3A and C4AF was 1.2 mass%. Also, the
accuracy of quantification of the amorphous phase was
2 mass%. In the investigation using unhydrated cement
with synthesized hydrates, the maximum difference from
the true value of the quantified amount of Ett, Ms, Mc,
and Hc was 1.0 mass%, 1.2 mass%, 0.6 mass%, and 1.0
mass%, respectively.
(3) Evaluation of the reaction ratio of BFS
The salicylic acid-acetone-methanol method was
adopted to quantify the amount of BFS (Kondo and
Ohsawa, 1969, Luke and Glasser, 1987). 10 g of Salicylic
acid, 140 ml of acetone, and 60 ml of methanol were
mixed and 2 g of sample was added to this solution and
mixed for one hour by a stirrer. The mixture was kept
until the age of 24 hours. Then, the insoluble residue was
separated by suction filtration on a paper filter. The
residue was heated to 850C for 15 min to incinerate the
paper and the insoluble residue was then weighed. Simultaneously, the insoluble content of plain cement paste
without BFS at the same age was evaluated and this value
was subtracted from the measured value of cement paste
with BFS to obtain the content of unhydrated BFS.

Intensity

(1)

3.1 Profile fitting by Rietveld method to XRD

pattern
As examples of profile fitting by the Rietveld method to
measured XRD patterns, the results of profile fitting of
cement paste at the age of 7 days are shown in Fig. 1. In
the cases of OPC and OPC-BFS, besides some of the

OPC Minerals
Ca(OH)2
Ettringilte
Monosulfate
Corundum (internal standard)

10

20

30

2()

40

50

60

(a) OPC (7days)

Intensity

100 ( A R )
A ( 100 R )

3. Hydration analysis by Rietveld method

OPC Minerals
Calcite
Ca(OH)2
Ettringite
Monocarbonate
Hemicarbonate
Corundum Internal Standard

10

20

30

40

50

60

2()

(b) OPC-LSP (7days)

Intensity

G=

(4) Compressive strength of cement paste

The compressive strength of cement paste was measured
on rectangular prisms measuring 4 4 16 cm. The W/C
of the paste was 0.50 and methylcellulose-type thickener
was added to control material segregation. The paste was
mixed in a Hobart mixer for 4 min and poured in a mold
on a table vibrator.

OPC Minerals
Ca(OH)2
Ettringilte
Monosulfate
Corundum (internal standard)

10

20

30

2()

40

50

60

(c) OPC-BFS (7days)

Intensity

tem, space group, and a axis and c axis of the synthesized

Hc obtained in this study were trigoal, R3c, 0.579 and
4.88 nm, respectively. In the analysis, the lattice parameter, orientation parameter, crystallite size parameter
and scale factor of each mineral were refined as appropriate. The orientation parameters of C3S, gypsum, calcite, CH, Ms, Mc and Hc, which were judged to require
in consideration of an accuracy of profile fitting, were
refined appropriately. A halo pattern of amorphous phase
was also refined as a background function.
The amount of amorphous phase is calculated by Eq.
(1).

OPC Minerals
Calcite
Ca(OH)2
Ettringite
Monocarbonate
Hemicarbonate
Corundum Internal Standard

10

20

30

2()

40

50

60

(d) OPC-BFS-LSP (7days)

Fig. 1 Examples of profile fitting results of hydrated cement paste by Rietveld method.

S. Hoshino, K. Yamada and H. Hirao / Journal of Advanced Concrete Technology Vol. 4, No. 3, 357-367, 2006

remaining cement minerals, peaks of Ett, Ms, and CH are

recognized as hydrate phases and the profile fitting is
successful. Halos of amorphous phases, which can be
estimated as C-S-H from OPC and BFS and unhydrated
BFS, are also fit as background curves.
In the cases of paste containing LSP, Ett, Mc, Hc, and
CH are recognized as hydrate phases and no Ms is identified. These results also indicate successful profile fitting by the Rietveld method to measured XRD patterns
of peaks of crystalline phases and halos of amorphous
phases.

Reaction Ratio of C3S (%)

80
60
40
OPC
OPC-LSP
OPC-BFS
OPC-BFS-LSP

20
0

10
20
Age (days)

30

Quantitative Value of Amorphous Phase (%)

Fig. 2 Reaction ratio of C3S.

100
80
60
40
OPC
OPC-LSP
OPC-BFS
OPC-BFS-LSP

20
0

10
20
Age (days)

30

Fig. 3 Total amorphous phase.

80

80
Reaction Ratio of BFS (%)

3.2 Hydration reaction of calcium silicates and

BFS
The evolution of the reaction ratio of C3S is shown in Fig.
2. The reaction ratio of C3S of OPC increases with age
and was estimated as 72, 82, and 89% at the ages of 3, 7,
and 28 days, respectively. In the case of the paste containing LSP, the reaction ratio of C3S is similar to the case
of OPC.
In the case of OPC-BFS and OPC-BFS-LSP, except
for the data of OPC-BFS-LSP at 28 days that can be
judged as an experimental error, the hydration ratios are
higher than OPC at every age, which indicates that BFS
accelerates the hydration of C3S. Ogawa et al. (1980)
reported that BFS accelerates the hydration of C3S based
on the hydration analysis of the C3S-BFS system using
synthesized C3S from reagents. This result agrees with
the results obtained in this study.
The development of the total amount of amorphous
phases quantified by the Rietveld method is shown in Fig.
3. At every age, the amounts of amorphous phases of the
systems containing BFS were greater than those of the
systems without BFS. This is thought to be caused by
unhydrated BFS being counted as amorphous phases
besides C-S-H.
The development of the reaction ratio of BFS and the
remaining amount of unhydrated BFS are shown in Fig.
4. This figure shows the average of 2 measurement results. In both the cases of OPC-BFS and OPC-BFS-LSP,
the reaction ratio of BFS increased with age, reaching
53% at 28 days.
Because the majority of amorphous phases in hydrated
OPC paste can be assumed to be C-S-H, the calculated
amount of amorphous phase is defined as the amount of
C-S-H produced in this study and the results are shown
in Fig. 5.
The amounts of C-S-H produced in the case of OPC
were 44, 52, and 63% at the ages of 3, 7, and 28 days,
respectively. LSP showed little effect of the produced
amount of C-S-H although the amount of C-S-H in the
case of OPC-LSP was slightly larger at the early ages.
In the cases containing BFS such as OPC-BFS and
OPC-BFS-LSP, the amounts of C-S-H produced were
less than in the cases of OPC and OPC-LSP at every age.
This phenomenon can be explained by the fact that the
produced amount of C-S-H from BFS is less owing to the
slower reaction progress of BFS compared to calcium

100

OPC-BFS
OPC-BFS-LSP

60

60

40

40

20

20

10
20
Age (days)

0
30

Remaining Amount of Unhydrated BFS (%)

360

Fig. 4 Reaction ratio and remaining amount of BFS

measured by selective dissolution method.

silicates.
The relationship between the reaction amount of cal-

361

OPS
OPC-LSP
OPC-BFS
OPC-BFS-LSP

80
Production Amount of C-S-H (%)

60

40
OPC
OPC-LSP
OPC-BFS
OPC-BFS-LSP

10
20
Age (days)

40

20
[Production Amount of C-S-H in Pastes without BFS]
0.991[Reaction Amount of Calcium Silicate ]
R 2 0.996

0
0

30

Fig. 5 Production amount of C-S-H.

cium silicates, which is the sum of C3S and C2S, and the
produced amount of C-S-H, is shown in Fig. 6. In the
cases of OPC and OPC-LSP, the produced amount of
C-S-H increases proportionally with increases in the
reaction amount of calcium silicate following a linear
line that crosses the origin. In this study, the amounts of
calcium silicates and amorphous phase were quantified
simultaneously by the Rietveld method. This method
gave a good correlation between the reaction amount of
calcium silicate and the produced amount of C-S-H.
Thus the accuracy of this method can be judged high
enough to warrant further investigation.
In the cases containing BFS of OPC-BFS and
OPC-BFS-LSP, the amounts of C-S-H produced for the
same reaction amount of calcium silicates were greater
than in the cases of OPC and OPC-LSP. This is caused by
the formation of C-S-H from BFS and the difference
between the two systems represents the produced amount
of C-S-H from BFS.
Based on this discussion, the development of the total
amount of C-S-H and C-S-H produced from calcium
silicates in the case of OPC-BFS is shown in Fig. 7. In
this result, the difference between the amount of C-S-H
from the total and that from calcium silicates is assumed
as the produced amount of C-S-H from BFS. The amount
of C-S-H from calcium silicates increases rapidly until
the age of 3 days and then rises at a slower rate until the
age of 28 days. On the other hand, although the amount
of C-S-H from BFS is limited at the age of 3 days and the
ratio to the total amount of C-S-H is low, the produced
amount increases with age and the ratio becomes significant at the age of 28 days.
The relationship between the total reaction amount of
calcium silicates and BFS and the produced amount of
C-S-H are shown in Fig. 8. There is a linear correlation
crossing the origin, which indicates that similar amounts
of C-S-H are produced from the same amounts of calcium silicates and BFS.

20

40

60

Total Reaction Amount of Calcium Silicates

[C3S+C2S] (%)

Fig. 6 Relationship between the reacted amount of

calcium silicates and production amount of C-S-H.

80
Production Amount of C-S-H (%)

20

60

OPC-BFS
Total Amounut of C-S-H
C-S-H Produced from
Calcium Silicates

60
Amount
C-S-H Produced from BFS

40

20

C-S-H Produced
from Calcium Silicate

10
20
Age (days)

30

Fig. 7 Estimation of the production amount of C-S-H form

calcium silicates and BFS.

80
Production Amount of C-S-H (%)

Production Amount of C-S-H (%)

S. Hoshino, K. Yamada and H. Hirao / Journal of Advanced Concrete Technology Vol. 4, No. 3, 357-367, 2006

60

OPS
OPC-LSP
OPC-BFS
OPC-BFS-LSP

40
20
0
0

20

40

60

80

Total Reaction Amount of

Calcium Silicates and BFS (%)

Fig. 8 Relation between the total reaction amounts of

calcium silicates and BFS and production amount of
C-S-H.

362

S. Hoshino, K. Yamada and H. Hirao / Journal of Advanced Concrete Technology Vol. 4, No. 3, 357-367, 2006

The development of the amount of CH is shown in Fig.

9. In the case of OPC-BFS, the produced amount of CH
is less than in the case of OPC at every age. This is
caused by the consumption of CH as a reaction stimulator for BFS. When LSP is added such as in the cases of
OPC-LSP and OPC-BFS-LSP, the produced amounts of
CH were less than in the cases without LSP such as OPC
and OPC-BFS. This is thought be caused by the consumption reaction of CH by the formation of carbonate
hydrates when LSP was added, as shown in Eq. 2 (Li et al.
1997). This is discussed in greater detail in section 3.3.
C3A 0.5CaCO3 0.5Ca(OH)2 12H2O
C3A0.5CaCO30.5Ca(OH)212H2O
(2)

Production Amount of CH (%)

In order to verify the accuracy of the quantified values

shown in Fig. 9, the relationship between the quantified
values of CH by the Rietveld method and those by
TG-DTA is shown in Fig. 10. Both results show reasonable coincidence, confirming the accuracy of the
Rietveld method for CH.
The development of the Ca/Si ratio of C-S-H was
calculated based on the reaction amount of C3S and C2S
and the produced amount of CH, as shown in Table 4. In
30

20
15
10
5
0
0

10
20
Age (days)

Age
Ca/Si molar ratio

3d

7d

28d

1.98

1.92

1.83

this table, only the OPC system is described because this

kind of calculation is difficult for BFS containing systems owing to the consumption reaction of CH to produce C-S-H and for LSP containing systems because of
the consumption reaction of CH to produce Hc.
The Ca/Si ratio of OPC obtained by the Rietveld
method was 1.98 at the age of 3 days and decreased with
age, dropping to 1.92 and 1.83 at the ages of 7 and 28
days, respectively. There are many reports on the Ca/Si
ratio of C-S-H. Sakai et al. (1998) reported that the Ca/Si
ratio of OPC paste at the age of 28 days was 1.73 by
XRD-TG measurement and 1.83 by TEM-EDS measurement, and that it decreases with age. The values obtained in this study were similar to those obtained by
earlier research.
3.3 Hydration reaction of calcium aluminate
phases
The development of the reaction ratio of C3A is shown in
Fig. 11. The reaction ratios of C3A of OPC systems increased with age and were 64, 80, and 92% at the ages of
3, 7, and 28 days, respectively. By the addition of BFS,
the reaction ratio of C3A increased especially at early
ages before 7 days. Thus BFS is thought to accelerate the
hydration of C3A.
The development of the reaction ratio of C4AF is
shown in Fig. 12. The reaction ratios of C4Af of OPC
were 24, 38, and 56% at the ages of 3, 7, and 28 days,
respectively. The reactions of C4AF in OPC-BFS and
OPC-BFS-LSP were 1.5 times higher than that of OPC
and OPC-LSP. This indicates that BFS accelerates the
hydration reaction of C4AF.

OPC
OPC-LSP
OPC-BFS
OPC-BFS-LSP

25

Table 4 Ca/Si molar ratio of C-S-H in hydrated OPC

paste calculated by Rietveld method.

30

Fig. 9 Production amount of CH by Rietveld method.

100

20

OPC
OPC-LSP
OPC-BFS
OPC-BFS-LSP

Reaction Ratio of C3A (%)

CH Amount [by Rietveld] (%)

25

15
10
5
0
0

10

15

20

25

CH Amount [by TG] (%)

Fig. 10 Quantitative accuracy of CH by Rietveld method.

80
60
40
OPC
OPC-LSP
OPC-BFS
OPC-BFS-LSP

20
0

10
20
Age (days)

Fig. 11 Reaction ratio of C3A.

30

363

S. Hoshino, K. Yamada and H. Hirao / Journal of Advanced Concrete Technology Vol. 4, No. 3, 357-367, 2006

In previous studies regarding the hydration reaction of

C3A coexisting with BFS, for example the C3A-melilite
glass-gypsum system, the rapid hardening of C3A was

Reaction Ratio of C4AF (%)

100
OPC
OPC-LSP
OPC-BFS
OPC-BFS-LSP

80
60
40
20
0

10 15 20
Age (days)

25

30

Fig. 12 Reaction ratio of C4AF.

OPC

20

Amount of Minerals (%)

C3A
Gypsum

Ettringite
Monosulfate

Amount of Minerals (%)

20

retarded by melilite glass (Hanada and Ito 1969). Regarding the hydration of C4AF, based on the measurement of hydration heat development of the
C4AF-BFS-Gypsum system, BFS was reported to delay
the hydration of C4AF (Asaga and Kuga 1996). However,
in this study on C3A and C4AF, after the age of 3 days,
hydration of both phases was accelerated by BFS. More
accurate evaluation of these contradicting results is
needed.
The hydration reactions of the calcium aluminate
phases of OPC, OPC-LSP, OPC-BFS, and
OPC-LSP-BFS are shown in Fig. 13. The quantification
errors of these phases were confirmed to be less than 1.2
mass% in a separate study (Hoshino et al. 2006).
In the case of OPC, the amounts of C3A and gypsum
decreased and the amount of Ms increased with age. The
amount of Ett showed a peak at the age of 3 days. In the
case of OPC-LSP, although Ett formation at early ages
was the same as in the case of OPC, later hydration became different, with Ett not transforming into Ms but
remaining a constant amount from 3 to 28 days, while
Mc and Hc were produced and their amounts increased

15

10

10

20
Age (days)

15

Ettringite
Monosulfate
Monocarbonate
Hemicarbonate

10

30

OPC-BFS

20

Ettringite
Monosulfate

Amount of Minerals (%)

C 3A
Gypsum

Amount of Minerals (%)

C3A
Gypsum
LSP

20

OPC-LSP

15

10

10
20
Age (days)

30

OPC-BFS-LSP
C 3A
Gypsum
LSP

15

Ettringite
Monosulfate
Monocarbonate
Hemicarbonate

10

0
0

10

20
Age (days)

30

10
20
Age (days)

Fig. 13 Amounts of hydrates relating to calcium aluminate.

30

40
OPC
OPC-LSP
OPC-BFS
OPC-BFS-LSP

20
0

10
20
Age (days)

30

30
25
20
15
10
OPC
OPC-LSP
OPC-BFS
OPC-BFS-LSP

5
0
0

20

40

60

80

100

Total Reaction Amount of OPC and BFS (%)

Fig. 15 Relationship between the total reaction amount of
OPC+BFS and ig.loss.
28 days
7 days
3 days

60

40

OPC-BFS-LSP

20

OPC-BFS

The quantification of hydrated cements containing BFS

and LSP was discussed in section 3. The authors also
reported a study on the analysis of strength development
using the Rietveld method (Hoshino et al. 2006). That
study showed that the strength increasing effect of LSP
addition on cement was the increase in the produced
amount of C-S-H by the hydration acceleration of calcium silicates at an early age around 3 days and the pore

60

OPC-LSP

4. Application of quantification by Rietveld

method for analysis of strength
development

80

OPC

3.4 Hydration reaction ratio of cement coexisting with BFS

The hydration reaction ratio of cement was calculated
from the total reaction amount of C3S, C2S, C3A, and
C4AF. The development of the hydration reaction ratio is
shown in Fig. 14. In the cases of OPC and OPC-LSP, the
hydration reaction ratio of cement increased gradually
60%, 70%, and 80% at 3, 7, and 28 days, respectively. In
the cases of OPC-BFS and OPC-BFS-LSP, the hydration
reaction ratios were 8% to 10% higher than in the cases
of OPC and OPC-LSP except for one data of OPC-BFS
at 28 days. Thus BFS has an accelerating effect on cement hydration. In general as an index of hydration reaction of cement, ig.loss is used in many cases. In order
to verify the accuracy of the results shown in Fig. 14, the
total reaction ratios of cement and BFS were plotted
against the ig.loss and the relationship is shown in Fig.
15. Regardless of the existence of BFS and LSP, there is a
linear relationship crossing the origin between the total
reaction amount of cement and BFS and ig.loss, indicating that the quantification of the reaction rate of cement and BFS by this method can be considered adequately accurate.

100

Fig. 14 Reaction ratio of matrix cement.

ig.loss (%)

with age. In the case of OPC, Ett was generated at an

early age and at a later age, Ett changed into Ms by reaction with C3A. However, in the case of OPC-LSP, the
amount of Ett remained constant in the long term through
the effect of LSP, because C3A does not affect the
transformation of Ett into Ms, instead generating carbonate hydrates (Li et al. 1998).
The hydration reaction of aluminate phases in
OPC-BFS is similar to the case of OPC. At early ages, Ett
was generated and later the amount of Ett decreased and
the amount of Ms increased. In the case of
OPC-BFS-LSP, the general tendency is similar to the
case of OPC-LSP. Ett generated at early ages did not
transform into Ms and the amount remained constant,
while Mc and Hc were generated. Compared to the case
of OPC-LSP, the amount of carbonate hydrates, especially the amount of Hc, was greater. Regarding the
reaction speed of calcium carboxaluminates, Hc is generated earlier than Mc. At later ages, the amount of Hc
becomes constant but the amount of Mc is continued to
increase.

Reaction Ratio of Matrix Cement (%)

S. Hoshino, K. Yamada and H. Hirao / Journal of Advanced Concrete Technology Vol. 4, No. 3, 357-367, 2006

Compressive Strength of Cement Paste (N/mm2)

364

Fig. 16 Compressive strength of cement paste.

28days
7days
3days

365

S. Hoshino, K. Yamada and H. Hirao / Journal of Advanced Concrete Technology Vol. 4, No. 3, 357-367, 2006

Table 5 Density data used for the calculation of

theoretical porosity.

Phase
OPC
LSP
C-S-H
CH
Ett
Ms
Mc
Hc

Density
(g/cm3)
3.15
2.72
1.90
2.24
1.78
2.01
2.17
1.98

OPC

0d

OPC

3d

Ms
C-S-H

CH Ett

7d
28d
0d

OPC-LSP

4.2 Relationship between porosity and strength

The theoretical volume of each hydrate phases were
calculated using the quantification results of each hydrate
phase obtained in section 3 and the density data of each
mineral listed in Table 5 (Taylor 1997). The development of theoretical volume of each hydrate is shown in
Fig. 17. In this figure, the total volume of unhydrated
phases is expressed as OPC but not as individual phases
for easier understanding, and the initial total volume of
unhydrated cement and mixing water is assumed to be
100%. Although LSP was added as substitution, carboaluminate hydrates were counted as part of the total
volume this time.
In every case, the amount of unhydrated cement decreased and hydrates such as C-S-H, CH, and aluminate
hydrates were generated and the theoretical volume increased with age. Although the percentage by C-S-H was
most significant in the total volume occupied by hydrates,
aluminate hydrates also occupied a significant volume
and served to fill the texture of hydrates.
In the cases of OPC and OPC-BFS without LSP, the
total volume of aluminate hydrates were constant or
slightly decreasing with age. The phase ratio of alumi-

LSP

OPC

3d

C-S-H

Mc Hc
CH Ett

7d
28d
0d

OPC-BFS

4.1 Paste strength

The paste strength of each cement is shown in Fig. 16.
With the addition of LSP, the paste strengths of OPC and
OPC-BFS increased at the ages of 3, 7, and 28 days. The
strength increasing effect of LSP was more pronounced
in the cases containing BFS than in the cases without
BFS at all ages. BFS enhances the strength increasing
effect of LSP. This tendency has been reported already
(Shimobayashi et al. 1979) and the results obtained in
this study confirms this effect. In the following sections,
the mechanism of strength change by BFS and LSP additions is discussed.

nate hydrates grew richer in Ms at later ages, unlike Ett,

which was richer at early ages.
When LSP was contained, the total volume of aluminate hydrates increased with age. The aluminate hydrates
were composed of Ett, Mc, and Hc, and the carboaluminate hydrates contributed to the increase in volume.
In the case of OPC, as the authors have reported
(Hoshino et al. 2005), the addition of LSP contributes to
the filling of pore systems through the increase in the
produced amount of C-S-H by the acceleration of the
reaction of calcium silicates and the increase in the produced amount of carboaluminate hydrates.
On the other hand, in the cases containing BFS, although there were insignificant differences in the total
aluminate hydrates at 3 days due to the existence of LSP,
after the age of 7 days, total aluminate hydrates increased
more in the case of OPC-BFS-LSP than in the case of
OPC-BFS because BFS contains a higher amount of
Al2O3 than OPC. This result indicates that the packing of
pores was promoted by the greater volume of aluminate
hydrates. The increase in the volume occupied by aluminate hydrates with the addition of LSP was more significant in the case containing BFS than in the case
without BFS at the ages of 7 and 28 days. When BFS
blended cement is used, by adding LSP, the packing of
pores by aluminate hydrates is more significant than in
the case of simple Portland cement.
Based on the theoretical volume shown in Fig. 17, the
development of theoretical porosity was calculated and is
shown in Fig. 18. In every case, the theoretical porosity
decreases with age. With the addition of LSP, in both the
cases of OPC and OPC-BFS, the theoretical porosity
decreased.

OPC-BFS-LSP

filling effect of carboaluminate hydrates generation.

Based on that study, the same technique was applied in
cement containing BFS and LSP in this study in order to
clarify the effect of LSP and BFS on strength development.

OPC

3d

BFS
C-S-H

Ms

CH Ett

7d
28d
0d

OPC

3d

BFS
C-S-H

LSP
CH Ett

7d
28d

Mc Hc

20
40
60
Theoretical Volume Occupancy (Vol%)

Fig. 17 Theoretical volume estimated by Rietveld

method.

S. Hoshino, K. Yamada and H. Hirao / Journal of Advanced Concrete Technology Vol. 4, No. 3, 357-367, 2006

60

OPC
OPC-LSP
OPC-BFS
OPC-BFS-LSP

55
48
46
44
42
40
38

0 3

Age (days)

28

Fig. 18 Theoretical porosity by Rietveld method.

The relationship between the theoretical porosity and

paste strength is shown in Fig. 19. Regardless of the
existence of BFS and LSP, the paste strength shows a
linear correlation with the theoretical porosity. This result indicates that the strength increasing effects of BFS
and LSP are caused by the filling of pores by hydrates
generated by the addition of these materials. In general it
is well know that there is a close relationship between
measured porosity by some method such as mercury
intrusion porosimetry and compressive strength, and this
correlation was detected even in the cases BFS and LSP
added (Ozu et al. 2002). The theoretical porosity obtained by Rietveld method shows a close correlation with
compressive strength and it can be concluded that the
Rietveld method is an effective tool for the analysis of
hydration reaction and understanding of the strength
development.

5. Conclusions
The analysis of hydration reaction of cement paste containing BFS and LSP by applying the XRD/Rietveld
method was investigated.
(1) By the combination of the Rietveld method and
selective dissolution method, the hydration reaction
of cement containing two amorphous phases of
C-S-H and unreacted BFS was successfully quantified.
(2) By the addition of BFS, after the age of 3 days, every
major phase such as C3S, C2S, C3A, and C4AF in
cement was accelerated to hydrate. The acceleration
was most significant for C4AF.
(3) By the addition of LSP, carboaluminate hydrates
such as monocarbonate and hemicarbonate were
generated. The produced amount of carboaluminate
was greater when BFS was added.

Strength of Cement Paste (N/mm2)

Theoretical Porosity by Rietveld Method (%)

366

60

OPC
OPC-LSP
OPC-BFS
OPC-BFS-LSP

40

20

0
36 38 40 42 44 46 48 50
Theoretical Porosity by Rietveld Method (%)

Fig. 19 Relationship between theoretical porosity and

strength of cement paste.

(4) The strength increasing effect of LSP was more

significant in the case of OPC-BFS than in the case
of OPC. This is caused by the greater filling ratio of
pores through greater production of aluminate hydrates by the reaction of BFS rich in Al2O3 with LSP.
(5) The theoretical porosity calculated by the quantification results proposed in this study showed a linear
correlation with compressive strength as expected
from previous studies and this new method is expected to be an effective tool for the hydration reaction analysis.
(6) One very important aspect the authors wish to emphasize is that the minor addition of LSP, less than 5
mass% to BFS blended cement, cannot be considered as a simple inert filler or extending agent but as
a significant admixture that affects the properties of
cement. Aluminates in BFS react with LSP as well
as those in OPC and this reaction contributes to
many of the properties of cement. Further investigation is in order.
References
Asaga, K. and Kuga, H. (1996). Effect of inorganic
material powder on the hydration speed of 4CaO
Al2O3Fe2O3. Cement and Concrete Research, 50,
68-73.
Bonavetti, V. L., et al. (2001). Studies on the
carboaluminate formation in limestone in limestone
filler-blended cements. Cement and Concrete
Research, 31, 853-859.
Fachinformationszentrum Karlsruhe and National
Institute of Standards and Technology (2006).
Inorganic Crystal Structure Database, ICSD.
http://icsd.ill.fr/icsd/index.html
Hanada, M. and Ito, M. (1969). Hydration reaction of
glassy melillite. Semento-Gijutu-Nenpo, 23, 72-75.

S. Hoshino, K. Yamada and H. Hirao / Journal of Advanced Concrete Technology Vol. 4, No. 3, 357-367, 2006

(in Japanese)
Hoshino, S., Hirao, H. and Yamada, K. (2005).
Application of XRD/Rietveld method for the
qualification minerals of cement containing
amorphous phases. Cement Science and Concrete
Technology, 59, 14-21. (in Japanese)
Hoshino, S., Hirao, H. and Yamada, K. (2006).
Hydration analysis of cement paste by XRD/Rietveld
method. Proceedings of JCI, 28, 41-46. (in Japanese)
Imoto, H., Sakai, E., Osawa, E. and Daimon, M. (2001).
Effect of the fineness of blast furnace slag and
chemical composition on the hydration of blast
furnace slag blended cement. Proceedings of JCA, 55,
44-45. (in Japanese)
Jambor, J. (1980). Influence of 3CaOAl2O3CaCO3
nH2O on the structure of cement paste. Proceedings
of the 7th International Congress on the Chemistry of
Cement, 4, 487-492.
Jones, R. C., et al. (2000). Estimation of the total
amorphous content of Hawaii soils by the Rietveld
method. J. Am. Soil Sci. Soc., 64, 1100-1108.
Kondo, R. and Ohsawa, S. (1969). 5th International
Symposium on Cement and Concrete, 4, 255.
Li, S., Oba, Y., Sakai, E., and Daimon, M. (1997).
Synthesis of hydrates in 3CaOAl2O3-CaSO42H2OCaCO3- Na2SO4-H2O system. Inorganic Materials, 4,
196-204. (in Japanese)
Li. S., et al. (1998). Effect of gypsum on the hydration
of 3CaOAl2O3-CaSO42H2O-CaCO3 system. Inorganic Materials, 5, 194-199. (in Japanese)
Ogawa, K., Uchikawa, H. and Takemoto, K. (1980).
The mechanism of the hydration in the system
C3S-pozzolana. Cement and Concrete Research, 10,

367

683-696.
Luke, K. and Glasser, F. P. (1987). Selective dissolution
of hydrated blast furnace slag cements. Cement and
Concrete Research, 17, 273-282.
Ozu, H., Yamada, K., Kobayakawa, M. and Ichikawa, M.
(2002). Effect of curing time on the neutralization
speed of cement mortar. Cement Science and
Concrete Technology, 56, 463-470. (in Japanese)
Ramachandran, V. S. and Chun-mei, Z. (1986).
Influence of CaCO3 on hydration and microstructural
characteristic of tricalcium silicate. Il Cement, 3,
129-152.
Sakai, E., Kato, M., Asaga, K. and Diamon, M. (1998).
A model of phase composition of cement hydrates.
Proceedings of JCI, 20, 101-106. (in Japanese)
Scrivener, K. L., et al. (2004). Quantitative Study of
Portland
Cement
Hydration
by
X-ray
Diffraction/Rietveld Analysis and Independent
Methods. Cement and Concrete Research, 34,
1541-1547.
Shimobayashi, S. and Iwabuchi, T. (1969). Effect of
limestone
powder
on
cement
strength.
Semento-Gijutu-Nenpo, 33, 84-87. (in Japanese)
Stutzman, P. and Leigh, S. (2002). Phase Composition
Analysis of the NIST Reference Clinker by Optical
Microscopy and X-ray Powder Diffraction. NIST
Technical Note 1441.
Taylor, H. F. W. Cement Chemistry 2nd Edition.
Thomas Telford, 1997.
Uchikawa, H. (1986). Effect of blending components on
hydration and structure formation. Proceedings of the
8th International Congress on the Chemistry of
Cement, 1, 249-280.