The influence of silica fume, nano silica and mixing method on the strength and durability of concrete

The present work was accomplished to investigate silica fume (SF) and nano silica (NS) effects on the compressive strength (fcu28) of concrete made by cement contents (CC) of 300, 400, 500 and 600 kg/m3. Two NS types having purity of 89% (Type I) and 99% (Type II) were partially replaced CC by the percentages of 1.5% and 3%. The influence of replacing cement by 5% and 10% SF was also studied. Nano silica was mixed by two techniques, mechanically and by ultrasonic device. The influence of SF, NS and CC on fcu28 and water absorption (WA) was found. The obtained results indicated that fcu28 increased and WA% reduced by replacing part of CC by either NS or SF. For both of NS types, the enhancement ratio in fcu28 was higher when replacing cement by 1.5% NS compared to replacement ratio of 3% and both ratios recorded greater values of compressive strength when compared to that without NS. At 1.5% NS ratio, fcu28 increased by about 14.81% for type I and by 41.33 for type II while at 3% NS, the enhancement ratios were 5.86% for type I and 35.46 for type II respectively. Mixing Type I NS by ultrasonic mixing method recorded higher values for fcu28 as compared to those recorded by the mechanical method.


Introduction
Concrete is considered the most familiar construction materials around the world. The recent advance in concrete industry is to reduce cement content by using supplementary cementitious materials (SCMs). Adding the SCMs such as silica fume (SF), micro-silica (MS), slag or fly ash (FA) to concrete can improve its properties. Regarding to the physical and chemical effect of MS, the use of MS in concrete effectively improved its short and long term properties [1,2].
Nowadays, utilization of mineral additives such as SF or FA has increased due to improvements in rheological, mechanical, durability and because of environmental concerns [3,4]. When compared with type I OPC, SF shows particle two orders finer and highly pozzolanic reactive chemistry. However, the micro-level currently fails to provide enough visions into building materials so, Nano scale gaining increasing attention. The SiO 2 nanoparticles increased concrete strengths [5] and enhanced its resistance to WA [6]. Also, leakage of calcium related to concrete degradation is measured by the utilization of SiO 2 [7]. Nano silica (SiO 2 ) also accelerates the hydration of C 3 S owing to extremely reactive surface of the nanoparticles [8]. Mechanical properties, Environmental resistance and durability of ordinary, geopolymer concrete and high performance concrete can be improved using NS [9][10][11][12].
The results obtained from [14] assured that Regardless of the W/CM ratio, concrete age or the replacement level, SF mixes have always exhibited a higher pozzolanic strength activity index compared to MK mixes. Moreover, The test results reported in [15] indicated that the replacement of cement by 0.5% NS and 0.75% NS increased the compressive strength by 14.2% and 20.5% for 7 days, 8.7% and 24.4% for 28 days and 8% and 18.1% for 90 days respectively over that of the control mix. Otherwise the strength decreased for specimens incorporating 1.5% NS, due to the agglomeration occurred by NS.
The mechanical strengths of concretes increased when adding NS. For specimens cured in water, 1% NS gave the maximum enhancement [16]. The influence of NS on the concrete strengths tested at 7, 28 and 90 days was discussed [17,18]. In these works, cement content was partially substituted with 0%, 0.5%, 1%, 1.5% and 2% NS. They found that, for specimens cured in water, the peak compressive strength was obtained at 1% NS while specimens cured in saturated waterline showed compressive strength greater than that cured in water. The optimum amount of NS provided the peak strengths was 4% as illustrated in [19]. Zhang and Li [20] verified a growth in the compressive strength with adding NS. Also, concrete strength containing 1% NS was greater than that incorporating 3% NS. In the other hand, the combined of MS and NS addition increased the concrete compressive strength and elastic modulus than using NS or MS only. Moreover, the cementing efficiency factor of NS is significantly higher than that of MS [21].
From the previous studied related to the use of silica fume and nano silica it was assured that, Silica fume belongs to the category of highly pozzolanic materials as it consists of silica in non-crystalline form with particles of a high specific surface, and thus exhibits great pozzolanic activity [25]. Also nano silica acts as the micro-filler of the cement particles. The Nano materials can reduce the amount of water that filled the voids of the blending materials [26].
From the above researches, the effect of SF and NS for different cement contents, NS purity, time of curing and mixing technique on the concrete properties are still limited. In This paper is aimed to evaluate the concrete properties casted with different cement contents incorporating SCMs in micro scale and nano-scale. Cement contents of 300, 400, 500 and 600 kg/m 3 were used. Two NS types having purity of 89% (Type I) and 99% (Type II) are partially replaced cement content by the percentages of 1.5% and 3%. The influence of replacing cement by 5% and 10% SF was also studied. Nano silica was mixed by two techniques, mechanically and by ultrasonic device to study the effect of mixing techniques on the compressive strength. The influence of SF, NS and CC on water absorption (WA) was found.

Experimentation
In this paper, the properties measured were compressive strength and WA for 180 cubes tested at 28 or 56 days. Four values of CCs, 300, 400, 500 and 600 kg/m 3 , were studied. The cement was partially replaced by two percentages of NS, 1.5% and 3%. NS was mixed by either mechanical method or ultrasonic method. The effect of replacing cement by 5% or 10% SF was also investigated. Superplasticizer was used for consistency adjustment. Trial mixes were prepared on the different investigated mixes to find the appropriate quantity of superplasticizer required for every mix to keep its slump value within the designed range of 80-100 mm.

Materials
Concrete cubes were casted using Ordinary Portland Cement (OPC) having a grade CEM I 42.5 N (BS EN 197-1/2000). Highly effective water-reducing agent (Sikament-163, supplied by Sika Egypt for construction chemicals) with synthetic type dispersion base and density equal 1.2 kg/l (ASTM C-494 Type A&F and BS 5075-3) was used. Two percentages of cement content (5 and 10%) were replaced by SF. The physical properties and chemical configuration of SF are listed in Table 1 as obtained from the manufacture sheet. Two kinds of Nano-SiO 2 (Type I and Type II) in powder form were examined in this study. The first type (Type I) was commercial. The second type of NS (Type II) was prepared in the Chemical Department, Faculty of Science, Zagazig University. Physical properties of Type I and Type II NS are found in Table 1. Natural sand and local natural dolomite of 14 mm NMS were used as fine and coarse aggregate respectively. Physical properties of aggregate are reported in Table 1. The grading of sand and dolomite were adjusted to follow the ACI recommendation as shown in Fig. 1.

Mix design and mixing procedures
The design of all mixtures was achieved to follow ACI method. Required quantities for concrete constituents' materials to produce a cubic meter for each mix are scheduled in Table 2.
The legend of specimen refers to content of CM (3 for 300 kg/m 3 and 4 for 400 kg/m 3 and so) followed by NS percent (N-%) then SF percent (S-%) and in ultrasonic method the samples names tailed by letter (U). The mixtures code for second type of NS samples tailed by letter (T). For example: (5N3.0S0T) is the code for samples containing cementitious materials = 500 kg/m 3 , NS = 3.0%, SF = 0 and T is the code for NS type II.
Drum mixer was used in order to mix concrete constituents and NS mechanically (mechanical method). Coarse aggregate, sand, cement and NS were mixed during 2 min in dry form. Then 50% of mixing water enclosing the total quantity of SP was combined and mixed till 3 min. After that, about 1 min rest was allowed and finally the remaining water was added inside the mixture and mixed during investigated mixes are within the range from 80 to 100 mm. The fresh concrete was cast in cubes of 100 mm side length and stored at room temperature till 24 h. After being de-molded the cubes were cured using water at 20 ± 1 °C for 28 days. The concrete samples were removed from curing tank at the specified testing age and any deposits on the specimens faces were removed before testing.

Test procedures for hardened concrete
Cube specimens 100 mm side length was prepared to evaluate the compressive strength (BS 1881-116). The cured cubes were tested after 28 and 56 days of curing (Three cubes for each mix). Also, WA test was performed at 28 days according to ASTM C 642-06 by draying cube  samples in oven at a temperature of 100-110 °C for 24 h and then weighted (W 1 ). After that, the samples were immersed inside water for 48 h followed by drying their surface and then weighted again (W 2 ). The WA% was implemented as follows:

Results and discussion
The compression test results measured at 28 and 56 days (f cu28 and f cu56 ) and WA test results measured at 28 days for the cubes are found in Table 4.

Fresh properties of NS concrete
In general, all slump values are ranged between 80 and 100 mm for all investigated mixes. As NS percentage increases, the superplasticizer dosage was increased to regulate the slump of mixture. The results are agreed with those reported in [6,27,28]. It is detected that mixes containing high content of SF required also high SP dosages. The high demand of superplasticizer with the concrete containing SF may be accredited to its very fine particle that acts as micro-filler of the cement particles. This may lessen the water that fill the blending materials voids and thus causes some of the SP being absorbed on its surface [29,30].

Concrete incorporating SF and Type I NS
The effect of CC on the compressive strength measured at 28 and 56 days (f cu28 and f cu56 ) is cleared in Fig. 2. The figure illustrates that f cu28 increased by about 28.4% with increasing CC from 300 to 400 kg/m 3 . The enhancement in f cu28 with increasing CC to 500 kg/m 3 is 37% compared to concrete with CC of 400 kg/m 3 . By increasing CC from 500 to 600 kg/m 3 , f cu28 enhanced by about 11% (Fig. 2a).
At 56 days, f cu56 increased as the cement content increased (Fig. 2b). The ultimate improvement in f cu56 was about 14.81% recoded at cement content of 300 kg/m 3 as the testing age increased from 28 days to 56 days. The concrete strength improvement ratio increased as testing age increased while it decreased when CC increased (see Table 4). Figure 3 shows SF effects on f cu28 and f cu56 of concrete made with CMs of 300, 400, 500 and 600 kg/m 3 respectively. In general, it is clear that, replacement of CC by SF enhanced f cu28 for different CMs contents (see Fig. 3a). The maximum enhancement was observed at SF% equals 5%, after that f cu28 decreased but still greater than that of the control mix (SF = 0). Silica fume increases the concrete properties by two ways: first, its small particles act as filler for the spaces between cement and aggregates particles. Second, SF reacts with CH to produce a greater solid volume of C-S-H gel, tends to an additional reduction in capillary porosity during hydration [31,32]. Herein, the maximum f cu28 was recorded for concrete having 5% SF partially replaced instead of CC. This result is matched with Nili and Ehsani [33]. Also, Youm et al. [34] found that 3.5% SF gave the maximum f cu28 compared to 7% SF for both normal and light-weight aggregate concretes. For concrete incorporating SF and tested after 56 days, f cu56 increased in a similar manner to that of f cu28 (Fig. 3b) but with higher enhancement ratios than those of f cu28 . A high improvement ratio of f cu56 was 34.61% at CMs of 400 kg/ m 3 (Cubes 4N0.0S05, SF = 5%). The enhancement ratio of f cu56 increased as the CMs (Cement + SF) increased from 300 to 400 kg/m 3 then decreased after words (see Table 4). Figure 4 shows the impact of Type I NS (1.5% and 3%) on f cu28 and f cu56 for concrete made with different CCs (300, 400, 500, 600 kg/m 3 ). The figure clearly indicate that spare of cement by 1.5% NS enhanced f cu28 at the different CCs.  This can be attributed to the reaction of silica with calcium hydroxide (pozzolanic reaction) resulted from cement hydration besides the formation of C-S-H gel. Also, the high specific surface of NS enables it to make very well [7,8,11]. On increasing NS from 1.5 to 3%, f cu28 decreased but it still greater than the control mix compressive strength (NS = 0), (Fig. 3a). Givi et al. [35] recorded a rise in concrete strength with adding NS till 1.5% replacement and after that it decreased. This was explained as follows: as the quantity of SiO 2 nano-particles added to the mix is greater than that needed to syndicate with the liberated lime through the hydration process, the extra silica, which does not contribute to concrete strength, replaces portion of the CMs materials and thus decreased the strength. When the NS percentage is large, the weak zone in concrete increases then f cu decreases. This decrease is owing to the agglomeration and defects generated in dispersion of NS particles [8,11].
The 56 days compressive strength, f cu56 , of concrete incorporating type I NS increased in a similar manner to that of f cu28 (Fig. 3b) with higher improvement ratios than those of f cu28 . A high improvement ratio of f cu56 /f cu28 of about 22.22% was recorded for concrete containing CMs of 300 kg/m 3 (Cubes 3N1.5S00, NS = 1.5%). The improvement ratio of f cu56 reduced as the CMs (Cement + NS) increased (see Table 4).

Concrete incorporating Type II NS
This type was included to investigate its influence on the concrete strength having CMs of 400, 500 and 600 kg/ m 3 and tested 28 days later. As shown in Fig. 5, the maximum value for f cu28 was attained at 1.5% NS partially added instead of CC. As the NS ratio increased to 3%, f cu28 decreased but still greater than f cu28 of control mix (NS = 0) for the different CCs. Based on the results illustrated in Figs. 4 and 6 Fig. 6 Effect of CMs on f cu28 of concrete containing Type I NS% mixed by ultrasonic method investigated. For example at CMs equals 400 kg/m 3 , with the utilization of Type II NS, f cu28 was 29% and 32% higher than those of Type I NS at NS equals 1.5% and 3% respectively. These percentages of enhancement in f cu28 with the usage of Type II NS decreased to about 9.5% as the CMs increased to 500 and 600 kg/m 3 . The purity of silica in Type II NS is around 99% (wt%) leads to increasing f cu28 .

Influence of mixing method for NS
In this section, the results concerning to ultrasonic technique effects to disperse Type I NS in water were assessed and likened to traditional method. This mixing method was applied on concrete mixes having CMs of 400, 500 and 600 kg/m 3 and tested 28 days later. Figure 6 shows Type I NS effects (1.5% and 3%) mixed by ultrasonic method on f cu28 . The maximum f cu28 was recorded at 1.5% NS. After increasing NS amount to 3%, f cu28 decreased but still greater than the control mix strength at the different CCs. However, NS have great surface area with high surface energies; agglomeration would appear at high ratios of the added powder, which prevents uniform distribution of NS particles within the mortar [36]. The mixing effects by mechanical and ultra-sonication on f cu28 of concrete containing 1.5% and 3% Type I NS partially added instead of CC at different CMs can be compared using the results reported in the two figures (Figs. 4a, 6). It is clear that the ultrasonic mixing method does not show a significant improvement on f cu28 for different CMs (maximum of about 5%). As stated above, owing to their extremely small particle sizes, NS particles formed agglomeration easily and uniform dispersion is hampered. This result agreed with that reported in [36]. It was found that f cu28 of sonicated mortars reduced by 5-10% compared to mortars prepared by mechanical mixing. Figure 7 shows CC effects (300, 400, 500 and 600 kg/m 3 ) on WA% of concrete at SF = 0 and NS = 0 tested 28 days later. It is clear that with increasing CC, WA% decreased with a decreasing rate. Using 400 kg/m 3 CC in concrete decreased WA% by about 24.8% compared to concrete containing 300 kg/m 3 CC. By increasing CC to 500 kg/m 3 , WA% decreased by about 23.6% compared to 400 kg/m 3 CC. Rising CC from 500 to 600 kg/m 3 decreased WA from 3.57 to 3.34 with reduction percentage of about 6.4%.

Water absorption
The SF partially replaced CC and their effect on WA% is demonstrated in Fig. 8. Silica fume improved concrete resistance for WA as its clear from the figure. This is because SF has finer particles and creates concrete further denser [28]. In general, SF decreasing pore size thus reduced WA [37]. The highest reduction in WA% for concrete containing 300 and 400 kg/m 3 was recorded at 10% SF. After rising CM to 500 and 600 kg/m 3 , the highest reduction in WA% was observed at 5% SF. This may be owing to agglomeration which appears at high ratios of SF, which prevents uniform distribution of particles within the matrix. Figure 9 shows the type I NS effect (1.5% and 3.0%) on WA% of concrete made with different CCs at 28 days. Using NS in general improved concrete resistance to WA. As particles of NS were finer than cement particles it fills the cement paste spaces and replace amount of CH by C-S-H leading to upgrading the interface of structure. Therefore, the permeability of the concrete would decreased by Nano particles reducing the pores at Nano level [38].
From Fig. 9, it is clear that, the reduction in WA increases as NS increases to 3%. This may be because NS can plays as a nucleus to tightly bond with C-S-H gel particles. Thus, the concrete durability during long-term may be increased as NS improved the hydration stability [39]. They stated that with increasing NS till 4%, the total specific pore volumes of concretes decreased, and the voids diameters of   Fig. 9 Effect of type I NS% on WA% of concretes at different CMs content