Cement & Concrete Composites 27 (2005) 848–853

www.elsevier.com/locate/cemconcomp

Mechano-chemical modification of cement with

high volumes of blast furnace slag
Konstantin Sobolev *

Facultad de Ingenieria Civil, Universidad Autonoma de Nuevo Leon, Apartado Postal #17 Ciudad Universitaria,
San Nicolas de los Garza, N.L. Mexico 66450, Mexico

Received 1 March 2004; accepted 29 March 2005

Abstract

The application of chemical admixtures significantly improves the performance of cement-based materials. Some admixtures can
also be used to modify the cement grinding process and induce changes in the structure of cement minerals due to mechano-chemical
activation. A reactive silica-based complex admixture was developed for the modification of cement grinding. This paper examines
the effect of grinding on the strength of a modified cement containing granulated blast furnace slag in high volumes. According to
the test results, mortars based on the modified cement possess a compressive strength of up to 91.7 MPa, a 62% increase over the
reference.

2005 Elsevier Ltd. All rights reserved.

Keywords: Chemical admixtures; Reactive silica, Cement; Grinding; Mechano-chemical activation; Strength; Durability; Granulated blast furnace
slag; High-performance cement

1. Introduction technique, a number of cement products have been

developed. Well known examples include air-entraining,
The application of advanced cements with high hydrophobic, and plasticized cements; and, also, the
strength and superior durability is an attractive method family of cements manufactured with grinding aids. It
of controlling concrete properties and the design of has been suggested that the action of these modifiers is
high-performance concrete (HPC) [1–8]. These cements governed by mechano-chemical activation [1,3–8]. Gen-
are often based on special clinkers or blended cements: erally, the theory of mechano-chemical activation
for example, shrinkage compensating or expansive (MCHA) has been applied to processing nanopowders,
cements, high-early strength cements, and regulated-set pigments, fillers, binders, ceramic, and ferromagnetic
cements [1,2]. Among these advanced cements are the materials [12–19].
newly developed Ultimax cement, blended Pyrament, Mechano-chemical activation is used to describe the
energetically modified cement, and silica fume cements chemical conversions in solids induced by a mechanical
[5–11]. process such as milling or grinding [14,15]. The mechan-
A novel approach to improving cement performance ical processing usually results in the formation of dislo-
includes the application of chemical admixtures (or cations and other defects in the structure of the material
modifiers) at the stage of cement grinding [3]. Using this [15]. In the case of MCHA, the mechanical impacts
cause the development of elastic, plastic, and shear
deformations leading to fracture, amorphization, and
*
Tel.: +52 81 83524969x201; fax: +1 925 663 0491. even chemical reactions in the solid state. Characteristic
E-mail address: sobolev@fic.uanl.mx features of MCHA are summarized in Fig. 1. Ball

0958-9465/$ - see front matter
2005 Elsevier Ltd. All rights reserved.
doi:10.1016/j.cemconcomp.2005.03.010
 K. Sobolev / Cement & Concrete Composites 27 (2005) 848–853 849

Fig. 2. Scanning electron micro-photograph of HP cement.

Fig. 1. The mechanism of mechanically induced solid state reactions.

(Courtesy of V. Balema, Iowa State University of Science and
Technology, reproduced with permission from [15].) of certain proportions of clinker, gypsum, a complex
admixture (RSA) and, optionally, a mineral additive
of industrial (IBPW) or natural origin [3]. The applica-
milling breaks down the crystallinity of solid reactants tion of HP cement imparts high strength and extreme
and provides a transfer of mass required for chemical durability to the concrete or mortar [3]; and its high
reactions. In addition, high pressure and shear stress strength can be used for engineering of a cement with
facilitate both the phase transitions and the chemical a high volume of mineral additives (HVMA) [20,21].
transformations of solids. The energy in the form of In this case, relatively large amounts (up to 70%) of
various lattice defects, accumulated by the solid during portland cement clinker can be replaced by inexpensive,
the mechanical processing, can support or even trigger locally available mineral additives. Natural pozzolanic
various chemical transformations [15]. materials, sand, limestone, granulated blast furnace slag,
Based on this theory, a significant improvement of fly ash, glass cullet and ceramic waste, all can be used as
cement strength has also been reported [1,3–8]. For mineral additives in HVMA cements [20].
example, low water demand binder (LWDB) is pro-
duced by intergrinding cement and a dry modifier at a
high energy [8]. In spite of a similar process, previously
described [1,10], only the application of a specially 2. Research significance
selected admixture (modifier) at a relatively high dosage
(about 4%) resulted in a cement with both reduced water The effect of silica fume (SF) on the behavior of ce-
demand and high strength. ment-based materials has been extensively reported in
Complex admixtures for application in cement tech- the literature [2]. A number of reports on SF cement
nology have been developed [7]. Generally, these admix- evaluate its positive effect on cement grinding and the
tures contain a reactive silica-based sorbent, an effective consequent improvement of the final product [4–6]. Less
surfactant, and some minor corrective components [3,7]. information is available on the behavior of RSA at the
Supersilica, a reactive silica-based complex admixture grinding stage. It was expected that SF and RSA would
(RSA), was produced using this principle. Although all improve the properties of HVMA cements containing
the effects of RSA have not been completely investi- granulated blast furnace slag. The evaluation of the
gated, it is hypothesized that, when added during the effect of SF, RSA, and the parameters of the mechano-
cement grinding process, RSA modifies the surface of chemical activation on the cement properties is impor-
cement particles and also promotes the formation of tant for a better understanding of the processes involved
highly reactive phases [3]. Scanning electron microscopy in MCHA and, therefore, for the development and pro-
(SEM) helps to reveal some details of interaction duction of high-performance cement-based materials.
between RSA and cement, such as nano-indentations
on the surface of cement particles (Fig. 2). Further, silica
component of RSA also acts as a micro-filler and partic- 3. Experimental program
ipates in a pozzolanic reaction. The mechano-chemical
activation of cement with RSA results in a new product, 3.1. Materials used
high-performance (HP) cement.
High-performance cement can be defined as a mate- The reference cement was portland cement CEM-I
rial manufactured by the mechano-chemical activation 42.5 [22] (NPC, Type I according to ASTM C150
850 K. Sobolev / Cement & Concrete Composites 27 (2005) 848–853

[23]). Three additional components were used in the proved to be critical for the evaluation of high strength
research: ground granulated blast furnace slag binders [3,8].
(GGBFS), silica fume (SF), and a reactive silica-based
complex admixture (RSA). The chemical composition 3.3. Preparation of specimens
of the main cementitious materials is presented in Table
1. Ground granulated blast furnace slag was an interme- Samples of SF and HP cement were obtained by
diate product used for the manufacturing of portland- intergrinding the specified mixtures in a laboratory ball
blast furnace slag cement; it had a Blaine specific surface mill. The sample weight was 5 kg and the grinding media
area of 358 m2/kg (compared with NPC having a Blaine weight was 65 kg. Grinding times were 10, 20, or 30 min.
area of 312 m2/kg). Standard RILEM Cembereau sand The resulting fineness data of the investigated cements
[24] and regular tap water were used for the preparation are summarized in Table 2 and Fig. 3. Mortars based
of the mortars. on the cements obtained were prepared following the
procedure of EN 196-1 [26]. The flow table was applied
3.2. Mixture proportioning to obtain the standard flow of 105–115 mm as per
ASTM C109 [25]. Using a jolting table for the mortars,
The strength properties of seven different cements test specimens were cast into three-gang prism molds
were investigated. These included SF and HP cements (40 · 40 · 160 mm) in accordance with EN 196-1 [26].
produced with three durations of intergrinding (10, 20, After the compaction procedure, the molds were placed
or 30 min) and the reference cement. The SF and RSA in a humidity cabinet for 24 h (keeping a relative humi-
were used at the same dosage of 10% (by weight). The dity of 95% and a temperature of 20 C). Following this
constant amount of GGBFS (45%) was used in all period, the specimens were removed from the molds and
cements (except the reference). The composition of the kept in 20 C water until the time of test.
investigated cements is given in Table 2. The mortars
were proportioned following the recommendations [3], 3.4. Tests performed
with W/C adjusted to provide a constant flow and a
sand-to-cement ratio (S/C) of 2.75 (similar to the The experimental program investigated the effect of
requirements of ASTM C109 [25]). This approach mineral additives and the duration of MCHA on:

• fineness of cements;
Table 1 • normal consistency and setting time;
Chemical analysis of cementitious materials • compressive strength.
Chemical Portland Granulated blast Silica fume
composition cement furnace slag The particle size distribution of investigated cements
SiO2 19.4 37.4 90.0 was measured by a SILAB laser diffraction analyzer.
Al2O3 4.8 10.9 0.4 Normal consistency and setting time were determined
Fe2O3 3.6 0.6 0.4 following the procedure of EN 196-3 [27]. Compressive
CaO 63.7 35.9 1.6
strength tests were conducted using the portions of
MgO 1.9 8.1 1.0
SO3 2.7 2.1 0.4 prisms broken in flexure as per EN 196-1 [26]. The com-
Na2O 0.2 0.4 0.5 pressive strength results are the average of the four test
K2O 0.8 1.2 2.3 values. The mortars were tested at the ages of 2, 7, and
Loss of ignition 2.4 – 3.0 28 days.

Table 2
Composition and fineness of investigated cements
Cement typea Cement composition Duration of MCHA (min) Fineness
NPC GGBFS SF RSA Median size (lm) Blaine (m2/kg)
NPC 100 – – – – 15.4 312
SFC-10 45 45 10 – 10 13.0 689
SFC-20 45 45 10 – 20 11.4 710
SFC-30 45 45 10 – 30 10.8 748
HPC-10 45 45 – 10 10 12.5 639
HPC-20 45 45 – 10 20 10.2 679
HPC-30 45 45 – 10 30 8.4 726
a
NPC—reference portland cement; SFC—# SF cement; HPC—# HP cement; the additional number after the main notation identifies the
duration of intergrinding (as 10, 20, or 30 min).
 K. Sobolev / Cement & Concrete Composites 27 (2005) 848–853 851

100 100

80 80
Passing, %

Passing, %
60 60

40 40
NPC NPC
SFC-10 HPC-10
20 SFC-20 20 HPC-20
SFC-30 HPC-30

0 0
1 10 100
1 10 100
Particle Size, µm Particle Size, µm

Fig. 3. Particle size distribution of the investigated cements.

4. Test results and discussion Inverse results were observed when RSA was applied;
due to additional grinding the normal consistency of
4.1. Particle size distribution HP cements was reduced (from 20% to 18.5%); this also
resulted in a reduction of W/C required to produce the
The particle size distributions of investigated cements mortars of the same flow (Table 3). It is proposed that
are presented in Fig. 3. Specific surface area and median there is a strong correlation between these two parame-
size are given in Table 2. It was found that the applica- ters as well as a correlation between the normal consis-
tion of SF and RSA significantly enlarged the specific tency and the duration of grinding for HP cements.
surface area which continued to increase constantly with Therefore, the W/C of mortars can be actually calcu-
grinding. It is important to notice that the specific sur- lated if the normal consistency is known. It was found
face area was higher when SF was applied. At the same that the setting times of the investigated cements were
time, the median size was smaller in the case of RSA significantly extended because of the application of min-
(Fig. 3). eral additives in large volumes (i.e., 55% of clinker was
replaced by the mineral additives). The subsequent inter-
4.2. Normal consistency and setting time grinding of SF cement further increased the initial and
final setting times, which were extended by 47 min each.
The samples of HP cements demonstrated the In contrast, intergrinding helped to decrease the setting
reduced normal consistency compared with the refer- times of HP cement by almost 50%. This is a clear sign
ence cement (Table 3). Following the trend specified of RSA–cement interaction and mechano-chemically
for the surface area, the normal consistency of SF induced changes within the system, which led to an
cement increased with the duration of intergrinding. acceleration of the hydration of HP cement.

Table 3
Compressive strength of mortars
Cement type Normal consistency (%) Setting time (min) S/C W/C Compressive strength (MPa)
at age (days)
Initial Final 2 7 28
NPC 27.1 2:45 3:25 2.75 0.45 31.6 47.9 56.5
SFC-10 27.0 3:55 4:25 2.75 0.45 16.6 33.2 59.4
SFC-20 27.3 4:22 5:02 2.75 0.45 17.0 35.2 62.5
SFC-30 27.6 4:42 6:12 2.75 0.45 18.1 36.3 65.1
HPC-10 20.0 4:55 5:30 2.75 0.30 35.8 61.1 89.1
HPC-20 19.2 2:39 3:04 2.75 0.29 39.0 65.4 91.7
HPC-30 18.5 2:37 2:52 2.75 0.28 34.2 60.0 75.9
852 K. Sobolev / Cement & Concrete Composites 27 (2005) 848–853

100 when RSA was applied: MCHA resulted in a reduc-

90 tion of the normal consistency of HP cements.
2 7 28
80 3. The setting time of the investigated SF and HP
Compressive Strength, MPa

70 cements increased significantly because of the incor-

60
poration of large volumes of mineral additives. It
was recognized that the setting time was further
50
extended with the intergrinding of SF cement. Due
40
to the RSA–cement interaction and mechano-chemi-
30
cal changes in the system, MCHA helped to shorten
20 the setting time of HP cements to the levels of NPC.
10 4. It is important to notice that the application of
0 MCHA helped to improve the properties of cements
NPC SFC-10 SFC-20 SFC-30 HPC-10 HPC-20 HPC-30 containing large volumes of granulated blast furnace
Fig. 4. Compressive strength of the investigated cements. slag and modifying admixtures (SF or RSA). It is
proposed that the properties of cement could be fur-
ther improved by the application of more effective
4.3. Compressive strength of mortars vibro-mills to facilitate MCHA.
5. It was found that the application of SF cannot over-
The results of the compressive strength are summa- come the hardening delay of HVMA cement with
rized in Table 3 and Fig. 4. According to the obtained 45% of BFS; and, therefore, was not effective at early
results, the application of SF helps to improve the 28- ages. Nevertheless, SF helped to improve the 28-day
day compressive strength of cement with high volumes compressive strength of HVMA cement containing
of GGBFS, up to levels exceeding NPC values by 15% 45% of BFS to levels exceeding NPC and a further
(and even further increase after this age could be ex- increase in ultimate strength could be expected.
pected). In spite of this, SFC had a lower strength at 6. The application of RSA was very effective and
early ages compared with NPC: only 57% and 76% at resulted in HVMA cement, which yielded a high com-
the age of 2 and 7 days, respectively. This delay of hard- pressive strength at all ages of hardening. It was
ening could be explained by the use of high volumes of found that when RSA was applied, the duration of
mineral additives (45% of GGBFS and 10% of SF). The MCHA had an optimum of 20 min. The 28-day com-
additional intergrinding helped to increase the 28-day pressive strength of 91.7 MPa was demonstrated by
strength of SF cement by approximately 10% due to the optimized HVMA cement containing RSA. It
MCHA [4–6]. The application of RSA results in high- was demonstrated if excessive intergrinding exceeded
early strength cement with a compressive strength about the optimal level, that the hydration of HVMA
50% higher compared with SF cement. This difference is cements at early stages was accelerated; however, at
even much higher at early ages. The 7-day compressive the cost of reduced strength.
strength of HP cement is about 16% higher than the 7. Further research programs will concentrate on the
28-day strength of reference NPC (56.5 MPa). For HP investigation and quantification of the micro-struc-
cement, it was found that the intergrinding had an opti- tural changes within the RSA–BFS–NPC system
mum duration of 20 min. This resulted in an HP cement due to MCHA.
with a maximum compressive strength at all ages of
hardening (Table 3). Remarkable strength (91.7 MPa)
was obtained by the optimized HP cement at the age Acknowledgements
of 28 days (62% higher than the reference NPC).
The author acknowledges the receipt of samples of
complex admixture Supersilica (RSA) from SCI Con
5. Conclusions Technologies. The author is grateful to the staff of Qual-
ity Control and R&D Laboratories working under the
1. The test results demonstrated that the specific surface umbrella of the Turkish Cement ManufacturersÕ Associ-
area of HVMA cements containing SF or RSA stea- ation for their help in conducting the experiments. Spe-
dily increased with grinding. The specific surface area cial thanks are conveyed to Dr. A. Oner from Turkish
was higher when SF was applied; and the median size Petroleum Corporation for the assistance in SEM inves-
was smaller in the case of RSA. tigations partially presented in this paper. The assistance
2. The application of RSA reduced the normal consis- and suggestions of Dr. C. Podmore are highly appreci-
tency of the investigated HVMA cements. The nor- ated. The kind permission of Dr. V. Balema to use the
mal consistency of SF cement increased with the graphic representation of the mechanism of mechani-
time of intergrinding. Inverse results were observed cally induced solid state reactions is acknowledged.
 K. Sobolev / Cement & Concrete Composites 27 (2005) 848–853 853

References [12] Schönert K. Physical and technical aspects of very and micro fine
grinding. In: Proceedings of the second world congress particle
[1] Royak SM, Royak GS. Special cements. Moscow, Russia: Stroyi- technology. Kyoto, Japan: Society of Technology; 1990. p. 557–71.
zdat; 1983. p. 88–109 [in Russian]. [13] Heinicke G. Tribochemistry. Berlin: Akademik Verlag; 1984.
[2] Neville AM. Properties of concrete. Prentice Hall; 2000. 844 pp. [14] Fernandez-Bertran JF. Mechanochemistry: an overview. Pure
[3] Sobolev K. The effect of complex admixtures on cement properties Appl Chem 1999;71(4):581–6.
and development of a test procedure for the evaluation of high- [15] Balema V. Synthesis of molecular solids: selected topics—III. In:
strength cements. In: Advances in cement research, vol. AC15, Lecture notes for CHEM, vol. 571, Solid State Chemistry. Iowa
No. 2. UK: 2003. p. 65–75. State University of Science and Technology; 2003. p. 1–4.
[4] Ronin V, Jonasson J-E, Hedlund H. Advanced modification [16] Suryanarayana C. Mechanical alloying and milling. Prog Mater
technologies of the portland cement base binders for different high Sci 2001;46(1–2):1–184.
performance applications. In: 10th international congress on the [17] Boldyrev VV. Reactivity of solids and new technologies. In:
chemistry of cement. Göteborg, Sweden: 1997 [publication no. Boldyrev VV, editor. Reactivity of solids: past present and
2ii077]. future. Oxford: IUPAC and Blackwell Science Ltd; 1996.
[5] Jonasson J-E, Ronin V, Hedlund H. High strength concretes with [18] Takacs L. Self-sustaining reactions induced by ball milling. Prog
energetically modified cement and modelling of shrinkage caused Mater Sci 2002;47(4):355–414.
by self-desiccation. In: Proceedings of 4th international sympo- [19] Balema VP, Wiench JW, Dennis KW, Pruski M, Pecharsky VK.
sium on utilization of high strength concrete (BHP-96). Paris: Titanium catalyzed solid-state transformations in LiAlH4 during
1996. p. 245–54. high-energy ball-milling. J Alloys Comp 2001;329(1–2):108–14.
[6] Ronin V et al. Method for producing cement. International Patent [20] Sobolev K, Arikan M. High volume mineral additive ECO-
Application PCT/SE 94/00383. 1994. cement. Am Ceram Soc Bull 2002;81(1):39–43.
[7] Sobolev K, Soboleva S. Complex admixture and method of [21] Sobolev K. High performance cement: solution for next millen-
cement based materials production. In: International Patent nium. Mater Technol 1999;14(4):191–3.
Application PCT/TR98/00008. 1998 [WIPO publication no. [22] EN 197-1:2000. Cement—Part 1: Composition, specifications and
WO98/54108]. conformity criteria for common cements. European Standard,
[8] Ioudovitch BE, Dmitriev AM, Zoubekhine SA, Bashlykov NF, European Committee for Standardization, 2000.
Falikman VR, Serdyuk VN. Low water requirement binders as [23] ASTM C150. Standard specification for portland cement. In:
new generation cements. In: 10th international congress on the Annual book of ASTM standards, vol. 04.01. Philadelphia, USA:
chemistry of cement. Göteborg, Sweden: 1997 [publication no. 1999. p. 123–7.
3iii021]. [24] TS 819: Rilem. Cembureau standard sand. Ankara, Turkey: TSE;
[9] Muszynski LC. Corrosion protection of reinforcing steel using 1989.
pyrament blended cement concrete. In: Swamy RN, editor. [25] ASTM C109. Compressive strength of hydraulic cement mortars
International conference on blended cements in construction, Shef- (using 2-in or 50-mm cube specimens). In: Annual book of ASTM
field, UK. London: Elsevier Applied Science; 1991. p. 442–54. standards, vol. 04.01. Philadelphia, USA: 1999. p. 64–8.
[10] Skvara F, Kolar K, Novotny J, Zadak Z. Cement pastes and [26] EN 196-1:1994. Methods of testing cement—Part 1: Determina-
mortars with low water-to-cement ratio—I. Cem Concr Res, USA tion of strength. European Standard, European Committee for
1980;10(2):253–62. Standardization, 1994.
[11] Al-Manaseer AA, Aquino EB, Kumbardi H. Properties of [27] EN 196-3:1994. Methods of testing cement—Part 3: Determina-
concrete containing ultimax rapid-setting hydraulic cements. tion of setting time and soundness. European Standard, European
ACI Mater J 1999;96(5):529–35. Committee for Standardization, 1994.