Cement and Concrete Research 31 (2001) 7 ± 12

Influence of silica fume on long-term strength of mortars containing

different aggregate fractions
G. Appa Rao*
Department of Civil Engineering, Indian Institute of Science, Bangalore 560 012, India
Received 16 March 1999; accepted 15 June 2000

Abstract

An experimental investigation on the influence of high - strength cement with silica fume on long - term strength of mortar is reported.
The effect of aggregate size and specific surface of aggregate on the variation of compressive strength has been studied. It has been
observed that the early strength development was very significant with the addition of silica fume in mortars. It was also found that the
size of the aggregate and its specific surface play a very significant role on the strength of the mortar. The strength of mortar increases
initially and then gradually decreases as the grain size and the specific surface of aggregate increases. The modulus of elasticity increases
as the compressive strength of the mortar increases. Significant strength losses have been observed in both silica fume and non - silica
fume mortars at the age of 180 days. D 2001 Elsevier Science Ltd. All rights reserved.

Keywords: Aggregate; Elastic moduli; Fineness modulus; Mortar; Silica fume; Surface area

1. Introduction ages [7 ± 10,14,15]. This has been due mainly to the nuclea-
tion of CH crystals around silica fume particles around the
When silica fume is incorporated, the rate of cement aggregate [13]. The increase in curing temperature at early
hydration increases at the early hours due to the release of hours results in very significant strength development with
OH ÿ ions and alkalis into the pore fluid [1]. The increased silica fume [8,12]. The pore structure of cement paste
rate of hydration may be attributable to the ability of silica depends on the aggregate content and the amount of pores
fume to provide nucleating sites to precipitating hydration in silica fume cementitious product increases as the aggre-
products like lime, C ± S ± H, and ettringite. It has been gate content increases [11]. Using scanning electron micro-
reported that the pozzolanic reaction of silica fume is very scope (SEM) and X - ray diffraction techniques, the
significant and the non- evaporable water content decreases formations of CH and ettringite have been observed
between 90 and 550 days at low water /binder ratios with the [14,16,17] in the transition zone. The thickness of this zone
addition of silica fume [2,3]. The presence of sand in mortars seems to be function of the size and shape of the sand
acts as a sink for crystallized Ca(OH)2 and lowers the particles. During the last two decades, few controversial
permeability of the cement paste [3]. It has been generally reports [18,21 ± 30] have evolved and expressed concern
reported [4± 6,19,20] that the strength development in the over the strength retrogression of high -strength cementitious
cement composites incorporated with silica fume is due products. In this study, an effort is made to investigate the
mainly to the pore size refinement and matrix densification, influence of silica fume, size of sand particles, and its
reduction of Ca(OH)2 content, and strong cement paste ± specific surface on the long - term strength of mortar using
aggregate interfacial bond. The formation of dense C ± S± H high -strength cement.
gel and more homogeneous product at the interfacial zone
leads to the rapid strength development of mortar at the early
2. Research significance

* Tel.: + 91 - 80 - 309 - 2330. During the last two decades, the concept of high -perfor-
E-mail address: [email protected] (G.A. Rao). mance cementitious materials has been on the rise. Applica-

0008-8846/01/$ ± see front matter D 2001 Elsevier Science Ltd. All rights reserved.
PII: S 0 0 0 8 - 8 8 4 6 ( 0 0 ) 0 0 3 4 6 - X
8 G.A. Rao / Cement and Concrete Research 31 (2001) 7±12

Table 1 four silica fume mortar mixes. The water / cementitious

Physical properties of cement materials ratio was 0.5 in Series II mortar mixes. Tables 3
Property of cement Result and 4 show the mix proportions of various mortar mixes.
Fineness, % 2.9 Cube specimens of size 70.7 mm were used for the
Standard consistency, % 31.80 determination of compressive strength of mortar. The
Setting times strength was calculated as the average of strength on three
(a) Initial (min) 125
cube specimens. In each mix, the fineness modulus and
(b) Final (min) 305
Specific gravity 3.12 specific surface of aggregates were changed to distinguish
Compressive strength, MPa the effect of the aggregate size, specific surface, and fine-
(a) At 3 days 23.00 ness modulus on the strength of mortars with and without
(b) At 7 days 34.50 silica fume. The surface area of the aggregates was calcu-
(c) At 28 days 52.50
lated by assuming that all the sand particles as spherical.
The surface area = (6 / rdmean), where r is the density of
aggregate (2.65 for quartzite) and dmean the mean diameter
tion of high -strength cementitious products for the field of the aggregate fraction.
applications is increasing rapidly. Moreover, incorporation The cube specimens were demoulded after 24 h and cured
of high range water reducing (HRWR) agents and use of for 28 days in water. Subsequently, the specimens were
pozzolanic materials like pulverized fuel ash and silica fume exposed for the laboratory environment at 27 ‹ 3°C and
are proven to be useful with low water /cementitious mate- relative humidity of 60 ‹ 5% for 180 days. The test results
rials ratio. Field engineers, academicians, and scientists in are shown in Tables 5 and 6. In obtaining the modulus of
the research laboratories have found the long - term strength elasticity according to ASTM C 469 -94, cylindrical speci-
reduction in mortar / concrete with the addition of silica mens 100 mm in diameter and 200 mm in height were cast
fume. The information on the effect of silica fume on the and the stress ± strain relations were drawn. The curing
long - term strength of mortar is limited. The influence of conditions were the same as that for cube specimens.
grain size of sand particles on the strength of mortar is
negligible. In this study, the influence of these parameters
has been studied on the long - term strength of mortar. 4. Test results and discussion

Tables 5 and 6 show the test results of the program for

3. Experimental program Series I and II mortar mixes, respectively. In Series I
mortars, the strength was observed at different ages of 7,
3.1. Materials 28, and 180 days. While in Series - II mortars, the strength
was observed at the age of 1 day. Fig. 1 shows the
3.1.1. Cement variation of compressive strength at 28 days for the two
Ordinary Portland cement was used with the properties series of mortar mixes with size of aggregate. From the
shown in Table 1. observations, it was clear that the compressive strength
initially increases and then gradually decreases as the
3.1.2. Silica fume grain size of sand increases. For all the mortar mixes,
Silica fume with a specific surface of 14 ±16 m2 / g and with and without silica fume, it is apparent that the
specific gravity of 2.05 was used for the program. The highest strength values were achieved with grain sizes
chemical composition of silica fume is shown in Table 2. between 1.5 and 2.0 mm. It is also clear that the silica
Silica fume was used as cement replacement material at 10 fume mortars are ahead of the plain cement mortars in
wt.% of cement throughout the program. compression at 28 days. At early days of curing, i.e. at 1
day and 7 days, the development of strength in silica
3.1.3. Fine aggregate fume mortars is very significant. This has been possible
The fine aggregate was non -porous quartzite from the due mainly to the pozzolanic action of silica fume [1± 6].
River Swarnamukhi, Tirupathi, AP, India. The natural sand
was sieved through different sieves to obtain required Table 2
particle size fractions. The cement/fine aggregate ratio was Chemical composition of silica fume
maintained at 1:3 in each of the mortar mixes. The ratios of Chemical compound Result (%)
different size fractions are presented in Tables 3 and 4.
Silica (SiO2) 86.00
Two series of mortar mixes were made. Series I consisted Alumina oxide (Al2O3) 1.0 (max)
of 14 mortar mixes with plain Portland cement mortar as Iron oxide 2.0 ± 3.5
reference mix, along with silica fume cement mortar at Silica + alumina + iron oxide (SiO2 + Al2O3 + Fe2O3) 87.00 ± 91.00
water /cementitious materials ratio of 0.417. Series II mortar Calcium oxide (CaO) 1.0 ± 1.5
Loss on ignition 4.0 ± 7.0
mixes consisted of eight mixes, four plain mortar mixes, and
 G.A. Rao / Cement and Concrete Research 31 (2001) 7±12 9

Table 3
Mix proportions and preliminary properties of Series I (W / C = 0.417) mortars
Ratios of grain size fractions in the mix
4.75 ± 2.36 2.36 ± 1.18 1.18 ± 0.60 0.60 ± 0.30 Average grain Fineness Specific surface,
Mix designation mm mm mm mm size, mm modulus m2 / kg
MPC I 1 0 0 0 3.55 5.0 0.637
MSF I 1 0 0 0 3.55 5.0 0.637
MPC II 0 1 0 0 1.71 4.0 1.273
MSF II 0 1 0 0 1.71 4.0 1.273
MPC III 0 0 1 0 0.89 3.0 2.544
MSF III 0 0 1 0 0.89 3.0 2.544
MPC IV 0 0 0 1 0.45 2.0 5.03
MSF IV 0 0 0 1 0.45 2.0 5.03
MPC V 1 1 1 1 1.65 3.50 1.372
MSF V 1 1 1 1 1.65 3.50 1.372
MPC VI 0 4 1 1 1.364 3.50 1.662
MSF VI 0 4 1 1 1.364 3.50 1.662
MPC VII 0 2 1 1 1.19 3.25 1.903
MSF VII 0 2 1 1 1.19 3.25 1.903

Table 4
Mix proportioning and preliminary properties of Series II (W / C = 0.50) mortars
Ratios of grain size fractions in the mix
4.75 ± 2.36 2.36 ± 1.18 1.18 ± 0.60 0.60 ± 0.30 Average grain Fineness Specific surface,
Mix designation mm mm mm mm size, mm modulus m2 / kg
PCM I 1 0 0 0 3.55 5.0 0.637
SFM I 1 0 0 0 3.55 5.0 0.637
PCM II 0 1 0 0 1.71 4.0 1.273
SFM II 0 1 0 0 1.71 4.0 1.273
PCM III 0 0 1 0 0.89 3.0 2.544
SFM III 0 0 1 0 0.89 3.0 2.544
PCM IV 0 0 0 1 0.45 2.0 5.03
SFM IV 0 0 0 1 0.45 2.0 5.03

However, the strength development at the early age is researchers concluded that the increase in the strength
more likely due to a nucleation effect: CSH formed from with the addition of silica fume results in the improved
OPC tends to precipitate on small SF particles, allowing a interfacial zone between cement paste and aggregate,
further dissolution of OPC in the liquid phase. Some which is more impermeable and compact with the silica
fume [7 ± 17]. Percentage compressive strength loss in
Table 5 various mortar mixes is shown in Tables 5 and 6,
Test results on compressive strength, modulus of elasticity, and strength respectively, in Series I and II mortar mixes.
loss in Series I mortar mixes In Fig. 2, it has been observed that as the specific
Compressive strength, surface of aggregate increases, the strength increases up to
MPa Modulus of Strength Percent
Mix 7 28 180 elasticity, loss, strength Table 6
designation days days days GPa MPa loss Test results on compressive strength, modulus of elasticity, and strength
MPC I 18.00 28.00 38.20 1.54 ÿ 10.20 ÿ 36.42 loss in Series II mortar mixes
MSF I 28.20 36.80 46.00 1.55 ÿ 9.20 ÿ 25.00 Compressive strength,
MPC II 44.00 54.20 50.00 3.6 4.20 7.73 MPa
Modulus of Strength Percent
MSF II 48.00 59.23 56.00 2.11 3.23 5.45
Mix 1 7 28 180 elasticity, loss, strength
MPC III 39.60 44.60 44.00 3.44 0.60 1.35
designation Day Days Days Days GPa MPa loss
MSF III 41.80 50.20 48.00 3.64 2.20 4.38
MPC IV 21.25 35.20 42.00 0.88 ÿ 6.80 ÿ 19.32 PCM I 8.60 24.00 39.20 34.80 5.94 4.40 11.22
MSF IV 26.60 40.60 39.00 2.86 1.60 3.94 SFM I 11.60 25.40 41.20 38.00 3.44 3.20 7.78
MPC V 48.60 50.70 50.00 2.90 0.70 1.38 PCM II 12.20 22.40 40.00 39.00 3.75 1.0 2.50
MSF V 50.20 60.70 52.00 4.20 8.70 14.33 SFM II 13.60 25.60 43.60 41.00 1.20 2.60 6.05
MPC VI 31.13 53.20 50.00 3.0 3.20 6.02 PCM III 10.80 16.70 29.20 30.20 2.50 ÿ 1.0 ÿ 3.40
MSF VI 39.33 58.20 58.00 3.69 0.20 0.34 SFM III 16.80 20.40 39.00 40.40 1.85 ÿ 1.4 ÿ 3.59
MPC VII 30.60 48.00 43.00 2.53 5.0 10.42 PCM IV 6.85 17.20 30.00 29.20 0.50 0.80 2.70
MSF VII 36.00 56.00 50.00 2.70 6.0 10.71 SFM IV 12.30 23.20 36.52 35.60 1.06 3.92 2.52
10 G.A. Rao / Cement and Concrete Research 31 (2001) 7±12

Fig. 1. Variation of compressive strength with grain size of sand particles Fig. 3. Variation of compressive strength with fineness modulus of sand
in mortar. particles in mortar.

certain value and then gradually decreases as the specific decrease in the grain size. As the specific surface in-
surface of sand particles increases. It is a known fact that creases, the water demand is obviously greater. As a
the specific surface of sand particles increases with result, the cement and silica fume binder was supplied

Fig. 4. Variation of modulus of elasticity with compressive strength of

Fig. 2. Variation of compressive strength with specific surface of aggregate. mortar.
 G.A. Rao / Cement and Concrete Research 31 (2001) 7±12 11

Fig. 5. Percentage strength loss in different mortar mixes.

with insufficient water for complete hydration, which thus the strength loss seems to be similar in all the mortar
lead to a weaker bond between cement paste and grain mixes, with and without silica fume, and this has been due
surface. In the case of larger sand particles, which results mainly to the drying effect.
in lesser specific surface, the excess water around the
sand particles increases the porosity and the interfacial
bond between cement paste and aggregate surface be- 5. Conclusions
comes weak. One more reason regarding this is that the
smaller sand particles are nearly of spherical in shape. The following conclusions may be drawn from the
The increased smooth surface of sand particles reduces experimental test results:
the aggregate interlocking, whereas, this effect is very
significant in the case of larger sand particles. It has been 1. The grain sizes of sand particles influence the strength
observed that as the fineness modulus of sand increases, of mortar. The compressive strength increases gradu-
the strength increases initially up to certain value and ally and then decreases as the size of sand particles
then decreases gradually. This trend has been shown in increased. The highest strength was observed with
Fig. 3 for both the mortar mixes in Series I and II. The grain sizes between 1.50 and 2.0 mm.
variation of modulus of elasticity of mortar with the 2. The addition of silica fume has significant influence
compressive strength has been shown in Fig. 4. It has on the early strength development of mortar.
been observed that the highest compressive strengths 3. In mixes with larger size grains, fractured surfaces
achieved using sand with fineness modulus in the range revealed that crack have crossed the grains, whereas
between 3 and 4. The modulus of elasticity of mortar cracks could not propagate across the grains with
increases as the strength increases. smaller size.
The test results on the long - term strength of mortar 4. The increase in specific surface of aggregate
show a negative effect after 28 days curing on mortars, results in the increase in strength, initially, then
with and without silica fume. It has been observed that gradually decreases.
strength retrogression was noticed in most of the mortar 5. Similar trend was observed with the fineness modulus
mixes. Very few results have been published on the of aggregate on strength of mortar. The highest
influence of temperature and relative humidity on the strength was achieved with fineness modulus in the
strength of mortars. However, extreme drying causes sig- range 3± 4.
nificant drop in the strength. In addition, severe strength 6. The strength loss was observed in mortar mixes, with
reduction may be possible when the specimens are inter- and without silica fume. This may be due mainly to
rupted during the first few hours after casting. In the drying effect.
present case, the ratio of drying surface area to volume
of test specimen is quite high and the drying effect seems
to be significant. Recently, the loss of strength due to
References
drying effect has been explained by few researches
[21,23,24,26]. Various mortar mixes with percentage [1] J.A. Larbi, A.L.A. Fraay, J.M.J.M. Bijen, The chemistry of the pore
strength losses are shown in Fig. 5. From the observations fluid of silica fume blended cement systems, Cem Concr Res 20 (4)
made on various mortar mixes, it has been observed that (1990) 506 ± 516.
12 G.A. Rao / Cement and Concrete Research 31 (2001) 7±12

[2] M.H. Zhang, O.E. Gjorv, Effect of silica fume on cement hydration in [16] X. Ping, J.J. Beaudoin, Modification of transition zone microstruc-
low porosity cement pastes, Cem Concr Res 21 (5) (1991) 800 ± 808. ture ± silica fume coating of aggregate surfaces, Cem Concr Res 22 (4)
[3] H.C. Yi, R.F. Feldman, Hydration reactions in Portland cement ± silica (1992) 597 ± 604.
fume blends, Cem Concr Res 15 (4) (1985) 585 ± 592. [17] M.D. Cohen, A. Goldman, W.F. Chen, The role of silica fume in
[4] R.F. Feldman, H.C. Yi, Properties of Portland cement silica fume mortar: Transition zone versus bulk paste modification, Cem Concr
pastes: II. Mechanical properties, Cem Concr Res 15 (6) (1985) Res 24 (1) (1994) 95 ± 98.
943 ± 952. [18] V. Yogendran, B.W. Langan, M.N. Haque, M.A. Ward, Silica fume in
[5] M.D. Cohen, M. Klitsikas, Mechanisms of hydration and strength high strength concrete, ACI Mater J 90 (1989) 124 ± 129.
developments in Portland cement composites containing silica fume [19] Z. Bayasi, J. Zhou, Properties of silica fume concrete and mortar, ACI
particles, Indian Concr J, (November) (1986) 296 ± 299. Mater J 90 (4) (1993) 349 ± 356.
[6] M.D. Cohen, A look at silica fume and its actions in Portland Cement [20] X. Cong, S. Gong, D. Darwin, S.L. McCabe, Role of silica fume in
Concrete, Indian Concr J, (September) (1990) 429 ± 438. compressive strength of cement paste, mortar, and concrete, ACI Ma-
[7] B. Matkovic, B. Grzeta, M.P. Jevic, V. Rogic, D. Dasovic, D. Dimic, ter J 89 (4) (1992) 375 ± 387.
Hydrated fly ash with SiO2 fume and / or Portland Cement addition: [21] F.D. Larr, P.C. Aitcin, Apparent strength retrogression of silica fume
Reactions in pastes and strength development in mortars, Cem Concr concrete, ACI Mater J 90 (6) (1993) 581 ± 585.
Res 20 (3) (1990) 475 ± 483. [22] M. Buil, A.M. Paillere, B. Russel, High strength mortars con-
[8] S. Wild, B.B. Sabir, J.M. Khatib, Factors influencing strength devel- taining condensed silica fume, Cem Concr Res 14 (5) (1984)
opment of concrete containing silica fume, Cem Concr Res 25 (7) 693 ± 704.
(1990) 1567 ± 1580. [23] F.D. Larrard, J.L. Bostvironnois, On the long - term strength losses of
[9] H.A. Toutanji, T. El - Korchi, The influence of silica fume on the silica fume high strength concretes, Mag Concr Res 43 (155) (1991)
compressive strength of cement paste and mortar, Cem Concr Res 109 ± 119.
25 (7) (1995) 1591 ± 1602. [24] R.D. Hootan, Influence of silica fume replacement of cement on phy-
[10] H.C. Yi, R.F. Feldman, Influence of silica fume on the microstructural sical properties and resistance to sulfate attack, freezing and thawing,
development in cement mortars, Cem Concr Res 15 (2) (1985) 285 ± and alkali ± silica reactivity, ACI Mater J 90 (2) (1993) 143 ± 151.
294. [25] S.L. Wood, Evaluation of the long - term properties of concrete, ACI
[11] R.F. Feldman, The effect of sand / cement ratio and silica on the mi- Mater J 88 (6) (1991) 630 ± 642.
crostructure of mortars, Cem Concr Res 16 (1) (1986) 31 ± 39. [26] M.N. Haque, Strength development and drying shrinkage of high -
[12] S.L. Mak, K. Torii, Strength development of high strength concretes strength concretes, Cem Concr Compos 18 (1996) 333 ± 342.
with and without silica fume under the influence of high hydration [27] T.C. Hansen, Long - term strength of high flyash concrete, Cem Concr
temperatures, Cem Concr Res 25 (8) (1995) 1791 ± 1802. Res 20 (2) (1990) 193 ± 196.
[13] J.A. Larbi, A. Bijen, J.M.J.M. Bijen, Orientation of calcium hydroxide [28] B. Persson, Long - term effect of silica fume on the principal properties
at the Portland cement paste aggregate interface in mortars in the of low - temperature cured ceramics, Cem Concr Res 27 (11) (1997)
presence of silica fume: A contribution, Cem Concr Res 20 (3) 1667 ± 1680.
(1990) 461 ± 470. [29] P.C. Aitcin, P. Laplante, Long - term compressive strength of silica
[14] P.J.M. Monterio, J.C. Maso, J.P. Ollivier, The aggregate mortar inter- fume concrete, J Mater Civ Eng 2 (3) (1990) 164 ± 170.
face, Cem Concr Res 15 (6) (1985) 953 ± 958. [30] S.A. Khedr, M.N.A. Zeid, Characteristics of silica fume concrete, J
[15] M.H. Zhang, O.E. Gjorv, Microstructure of the interfacial zone bet- Mater Civ Eng 6 (3) (1994) 357 ± 375.
ween light weight aggregate and cement paste, Cem Concr Res 20 (4)
(1990) 610 ± 618.