Silica Fume and Nanosilica Effects on Mechanical and Shrinkage Properties of Foam Concrete for Structural Application

Adding the appropriate amount of silica fume (SF) or nanosilica (NS) can improve the strength and reduce the shrinkage of foam concrete (FC), thereby widening its application.-is paper reports on a study of FCmechanical and shrinkage properties when varying SF andNS proportions were used to replace cement. Scanning electron microscopy (SEM) and nonmetallic ultrasonic detection were employed to investigate the mechanism of strength changes. -e result showed that when the SF and NS content was 15% and 4%, in the 28 days, the compressive strength of FC peaked at 32.8MPa and 35.8MPa, and the flexural strength maximized at 7.9MPa and 9.1MPa, respectively. -e autogenous shrinkage reached the minimum value, separately, i.e., 741.2×10 and 797.1× 10. -e dry shrinkage value was the highest (862.5×10 and 1387.5×10). -e results indicated that the compressive and flexural strength of FC first increased and then decreased as SF andNS content increased. Autogenous shrinkage first decreased and then increasedwith the increase of SF or NS content. An opposite trend of dry shrinkage could be observed. FC made with NS had higher shrinkage than it did with SF. -e results also indicated that there was a correlation between microstructure and sound velocity in the presence of SF or NS.


Introduction
Nowadays, foam concrete (FC) has attracted more and more attention. FC is an economical and environmentally friendly material [1]. It is light and possesses acceptable thermal insulation and low elasticity [2,3]. It absorbs shocks and has acceptable fire resistance [4]. It is currently widely used in civilian, underground, and municipal engineering and sports construction and the military [5]. But its shortcomings are low strength and high shrinkage [6], which limit its application. erefore, improving FC strength and reducing its shrinkage can help widen FC application.
Contemporary research into FC mechanical and shrinkage properties has primarily concentrated on nonstructural FC. Nambiar and Ramamurthy [7] reported on how various aggregates influenced FC properties. As aggregate sand grain size decreased, FC strength increased. Kearsleya and Wainwright [8,9] studied fly ash (FA) content effects on FC strength. When FA replaced more than 67% of the cement, compressive strength decreased significantly.
Nambiar and Ramamurthy [10][11][12] examined the influence of stomatal characteristics on FC density and compressive strength. Stomatal apertures and spacing between pores influenced FC density and strength. As stomatal distribution and pore diameter distribution uniformed, intensity increased. Stomatal shape had little effect on the intensity. Jones and McCarthy [13] reported the effects of untreated low-calcium FA on FC properties and found it had an acceptable strength. Kearsley and Wainwright [14,15] discovered that FC pore structure closely relates to strength, water absorption, and dry shrinkage. Hou et al. [16] found that adding a proper amount of nanoparticles in foaming agent and precast foam can improve the mechanical properties of foamed concrete. He et al. [17] found that the size and distribution of pores have a great influence on physical and mechanical properties. Nguyen et al. [18] studied the effects of pore characteristics and mortar properties on the mechanical properties of FC by micromechanics. Falliano and et al. [19] studied factors influencing compressive strength of FC, for example, dry density, moisture content, curing conditions, cement type, and foaming agent type. Sun et al. [20] studied the effects of different foaming agents on dry shrinkage of FC.
Little research has been done on FC for structural applications. Bing et al. [21] determined that high-strength FC with a density range of 1000∼1500 kg/m 3 and compressive strength of approximately 20∼50 MPa could be obtained by using SF, polypropylene fiber, and a high-efficiency water reducer. Kearsley and Wainwright [8,9] developed a sandfree, high-strength FC with a low water-binder ratio using SF and FA. Jones and McCarthy [22] studied the influence of FA on high-strength FC performance by substituting cement, or fine aggregate, for FA. ose substitutions significantly improved FC performance.
Higher shrinkage [23] of FC than ordinary concrete has been a critical barrier for the past seven decades. Wan et al. [24] verified that the shrinkage of FC with a dry density of 1500 kg/m 3 approximately equaled that of the ordinary concrete. e shrinkage of FC with a dry density of 1600 kg/m 3 increased with the increase of the amount of polyvinyl alcohol (PVA) fiber [25]. e shrinkage of FC with a density of 1600 kg/m 3 could be reduced efficiently by incorporating propylene glycol (PG), triethylene glycol (TEG), and dipropylene glycol tert-butyl ether (DPTE) [26]. However, rare research has been carried out to compare NS and SF on FC strength or to investigate their influence on shrinkage.
is paper reports on a study of using mechanical tests, shrinkage tests, scanning electron microscopy (SEM), and ultrasonic nondestructive tests to compare FC mechanical and shrinkage properties, hydration products, and micromorphology of FC using different dosages of SF and NS. e potential application of FC in the presence of SF or NS for structural elements was analyzed and discussed.

Materials.
Ordinary Portland Cement 42.5 produced by Hunan South Cement Plant was used. Its chemical composition and basic performance indexes were given in Table 1 and Table 2. Quartz sand produced by Heyuan Wanchuan Quartz Development Limited Company, with a 40-70 mesh size and density of 2.65 g/cm 3 , was used. Its grading curve was given in Figure 1. SF produced by Shanxi Yizhou Ferroalloy Factory, with an average particle diameter ranging from 0.1 to 0.3 microns and density of 2.2 g/cm 3 , was used. Its chemical composition was given in Table 3. NS produced by Shanghai Yuanjiang Chemical Limited Company, with a diameter of 30 nm and a density of 2.2 g/cm 3 , was used. Grade I FA produced by Henan Sitong Chemical Construction Limited Company, with a density of 2.6 g/cm 3 , was used. Its chemical composition was given in Table 4. Lime powder with a density of 1.2 g/cm 3 was used and its composition was given in Table 5. A polycarboxylic acid superplasticizer (water-reducing rate: 31%) produced by Nanjing Sitaibao Company, with a density of 500 g/L, was used. A foaming agent (vegetable protein anionic surfactant) produced by Guangdong Yunfu Qingzhou Cement Limited Company, with a density of 1.01 g/L, was used to induce foam.

Mix
Proportion. An orthogonal test was carried out on the mix proportion of FC with a dry density of 1600 kg/m 3 . e influence of various factors on FC strength was investigated. e optimum mix proportion was determined through range analysis. e factors and levels of the orthogonal test were selected as shown in Table 6. e optimum mix proportion and strength were shown in Table 7.

Test Methods.
e flexural and compressive strength and elastic modulus were evaluated in accordance with the specification in China (GB/T 11969-2008 [27]). e sample sizes for compressive strength and flexural strength were 100 mm × 100 mm × 100 mm and 100 mm × 100 mm × 400 mm, respectively. Sample size for elastic modulus was 100 mm × 100 mm × 300 mm. A method of combination of noncontact probes and corrugated pipes was used to test the autogenous shrinkage. e pipes were 340 ± 5 mm long, with inner and outer diameters of 20 mm and 30 mm, respectively. Each test group had two test samples. A mean value of autogenous shrinkage for the two samples was adopted for each group. e dry shrinkage was determined in accordance with JGJ/T70-2009 [28]. Each group had three 40 mm × 40 mm × 160 mm samples. Comparator (SP-175) was used to measure the length of samples at 0, 3, 7, 14, 21, 28, and 56 days. SEM (JSM-649LV) was used to detect the microstructure of FC. Nonmetallic ultrasonic nondestructive testing equipment (ZBL-V520) was used to measure the ultrasound speed of FC in accordance with ASTM C597-02 [29]. Figure 2, FC compressive strength initially increased and then decreased as SF or NS ratios increased. As SF ratios increased from 0% to 15% and NS ratios increased from 0% to 4%, FC compressive strength at 7 and 28 days gradually increased. FC compressive strength at 7 and 28 days reached its highest level of 21.7 MPa and 32.8 MPa when SF was 15% whereas it was 24.9 MPa and 35.8 MPa when 4% NS was blended. Generally, the micro-or nanoparticles in SF or NS may densify greatly the FC mortar. Additionally, secondary hydration of SF or NS improves strength radically [30]. However, at 4% NS, FC compressive strength at 7 and 28 days was 14.7% and 9.1% higher than 15% SF. is can be attributable to the much smaller particle size and higher specific surface area of NS, which enables NS to hydrate vigorously, and the resulting hydration products efficiently densify FC mortar [31]. FC compressive strength decreased as SF exceeded 15% and NS exceeded 4%. ere are two possible reasons. First, the foam in FC intends to collapse and merge in the presence of excessive SF or NS which may weaken the compressive strength of FC. Second, the increased water absorption of SF or NS inhibits the hydration. Figure 3 shows the influence of SF and NS ratios on FC flexural strength. FC flexural strength shows a similar pattern with compressive strength. When SF and NS percentages increased from 0% to 15% and 0% to 4%, respectively, FC flexural strength at 7 and 28 days gradually increased. FC flexural strength at 7 and 28 days reached the highest level of 6.6 MPa and 8.7 MPa when SF was 15% and 7.0 MPa and 9.1 MPa when NS was 4%. At 4% NS, FC flexural strength at 7 and 28 days was 6.1% and 4.6% higher than 15% SF. When the added amount of SF and NS was more than 15% and 4%, the flexural strength decreased.

Elastic Modulus.
Elastic modulus is an important parameter for designing FC in structural elements. As mentioned above, the compressive strength and flexural strength of FC peaked when SF and NS content was 15% and 4%, respectively. e elastic modulus of FC was measured when the optimum SF or NS content was blended. In the presence of 15% SF or 4% NS, the elastic modulus of FC was 18.5 GPa and 19.4 GPa, respectively. Investigation on the elastic modulus of foam concrete, with a similar strength grade of 30 MPa to 35 MPa, has rarely been reported. However, the results indicated an obvious E FC increase when compared to the ordinary foam concrete [32,33]. When compared to lightweight concrete in the same strength grade range, obvious elastic modulus difference, that is, 2.5 GPa to 5.5 GPa, could also be found [34][35][36].
According to JGJ 51-2002 [37], LC15 lightweight aggregate concrete is allowed for load-bearing components or structures building. e corresponding elastic modulus is 11.6 GPa when the apparent density and strength grade are 1600 kg/m 3 and 15 MPa, respectively. Consequently, the manufactured foam concrete, whether it is blending with SF or NS, characterized with a compressive strength of 32.8 MPa or 35.8 MPa, and elastic modulus of 18.5 GPa or 19.4 GPa, respectively, can be recommended for potential structural application. Figure 4 shows the changes in FC autogenous shrinkage over time, under different SF and NS ratio conditions. As time passed, FC autogenous shrinkage values increased rapidly from 0 to 24 h. ey remained stable after 24 h.

Autogenous Shrinkage.
When SF increased from 0% to 15%, FC autogenous shrinkage decreased. When 15% SF was added, after 72 h, FC          Advances in Materials Science and Engineering autogenous shrinkage was 53.7% lower than when SF was absent. FC strength may develop rapidly when SF is added [38], which results in higher strength and stiffness and strong resistance to autogenous shrinkage. e law of strength corresponds well with that of a contraction driving force. e strength growth rate was higher than that of capillary negative pressure [39], which limited contraction.
When SF increased from 15% to 25%, FC autogenous shrinkage values increased. e reasons may be that, first, SF particles are fine and their filling properties refine FC pore structure; second, early stage-generated Ca(OH) 2 is scattered and small. After adding SF, it easily reacts with SF to form hydrated calcium silicate with a low C/S ratio. is makes the interface between unhydrated cement particles and hydrated calcium silicate more compact and uniform [40]. ird, when SF quantities are high, excess SF continues to react with the C-S-H to generate an outer layer of hydrated calcium silicate which has a lower C/S ratio. is makes the SF cement base material structure more compact [41]. SF reduces FC porosity and average pore diameter. e autogenous shrinkage is related to FC capillary pressure. Capillary pressure is mainly related to diameter. More SF results in smaller capillary diameters, higher capillary pressure, and greater FC autogenous shrinkage.
As NS increased from 0% to 4%, FC autogenous shrinkage gradually decreased. At 4% NS dosage, after 72 h, it was 50.2% lower than the virgin FC without SF. And it was 7.5% higher when compared to FC in the presence of 15% SF. e reason for the decrease of autogenous shrinkage is similar to SF. Due to the higher specific area, NS has a negative effect on FC autogenous shrinkage when compared to SF. Figure 5 shows the change in FC dry shrinkage values over time under different SF and NS quantities. As time elapses, FC dry shrinkage values increase rapidly from 0 to 14 days but remain stable after 14 days.

Dry Shrinkage.
When FC amounts increased from 0% to 15%, FC dry shrinkage values increased. At 15% SF, FC dry shrinkage values after 56 days increased 68.3% more than FC which lacked SF. FC dry shrinkage values increased as SF quantity increased for the following reasons. First, SF is highly active and can accelerate early FC hydration reaction. Rapid internal water consumption leads to an early FC capillary tension growth rate. Growth rates increased as SF amounts increased causing great early shrinkage [42]. Secondly, SF requires large amounts of water. As SF increases, slurry fluidity decreases and a large number of cement particles clump into groups. During mixing, foam deformation and bursting are likely to occur. After hardening, the structure is not dense, the internal defects are increased, and there are many pores connected, which are prone to dry shrinkage.
As SF increased from 15% to 25%, FC dry shrinkage values decreased because excessive SF led to the nonhydration of some other SF particles. SF particles are very fine and effectively fill the pores between concrete particles which improves FC particle gradation and compacts the structure while reducing dry shrinkage values. is plays a dominant role when the quantity of SF is high.
As NS increased from 0% to 4%, FC dry shrinkage increased. After 56 days, dry shrinkage of FC with 4% NS was 170.7% higher than non-NS FC and 60.9% higher than FC with 15% SF. e cause of FC with NS dry shrinkage first increase was similar to that of FC with SF. FC had a larger dry shrinkage value when NS was added compared to SF because NS particles are smaller, have a larger specific surface area, hydrate more completely, and have a quicker internal water consumption compared to SF. is results in increasing FC dry shrinkage. FC dry shrinkage value was decreased with extra NS similar to that of SF. Advances in Materials Science and Engineering 3.3. Microstructure. SEM was used to observe the resulting FC microstructure of different quantities of SF (5%, 15%, and 25%) and NS (1%, 4%, and 5%). Figure 6 shows FC microstructures under different percentages of NS (1%; 4%; 5%). e 50x magnification figures in Figure 6 show that FC with 4% NS had the smallest pore diameter, minimum interconnected pores, and the lowest breaking rate, whereas FC blended with 1% and 5% NS had larger pore diameters, more interconnected pores, and higher breaking rates. e filling effect of NS can improve FC pore structure. But excessive NS may cause foam to burst, causing the FC pore size, the number of interconnected pores, and pore breakage rates to increase. erefore, adding NS to FC in a certain amount can improve pore structure and strength. e 5000x magnification figures in Figure 6 reveal that, compared to FC with 1% and 5% NS, FC with 4% NS had fewer unhydrated particles and flaky calcium hydroxide. It had more C-S-H gels, a denser ettringite, and a more compact structure. At 10,000x magnification ( Figure 6) the ettringite in FC with 4% NS was more compact than in either FC with 1% or 5% NS. is indicated that a certain amount of NS may promote cement hydration, while excessive NS may increase water absorption and restrict hydration reactions. Flake calcium hydroxide is quite brittle structurally and FC strength reduced due to that. NS can partially, or wholly, consume the calcium hydroxide generated by the cement hydration products. It will generate C-S-H in order to make full use of cement and increase concrete internal structural compactness and improve concrete strength [43]. As a result, the strength of FC with 4% NS was greater than either FC with 1% or 5% NS. erefore, FC with 4% NS had the most efficient microstructure and the highest strength. Figure 7 presents the microscopic structure of FC mortar when 5%, 15%, and 25% SF was blended, respectively. It showed that when 15% SF was used, the lowest porosity, minimum interconnected pores, and the lowest breaking rate in FC could be observed. More C-S-H gels, fewer unhydrated particles, and flaky calcium hydroxide could be found. In addition, more densified microstructure also occurred in SEM. us, the strength peaked in the presence of 15% SF.

Comparison between NS and SF.
Compared to the FC with 4% NS shown in Figure 6, FC with 15% SF had larger pore diameters, more interconnected pores, more unhydrated particles, less ettringite, fewer C-S-H gels, and a less compact structure because NS particles are finer than SF particles and can more efficiently fill FC pore structures. NS particles can make the pore structure more compact and the pore size distribution more uniform. NS has a larger specific surface area than SF, which may make the reaction more effective. erefore, FC with 4% NS is stronger than FC with 15% NS.

Relationship between Ultrasonic Speed and FC Strength.
Nonmetallic ultrasonic testing equipment measures FC for SF at 0%; 5%; 10%; 15%; 20%; 25% and for NS at 0%; 1%; 2%; 3%; 4%; 5%. Measurement results appear in Figure 8. Figure 8 shows that FC ultrasonic speed first increased and then decreased as SF and NS amounts increased. FC ultrasonic speed reached its highest levels for 15% SF and 4% NS. is may be due to the fact that when a small amount of SF is added to FC, the cement slurry becomes well covered by foam, making the pore diameter distribution more uniform and decreasing average pore diameter which, in turn, decreases porosity. However, when additional SF is present, the extra SF is not completely hydrated via cement hydration products. is may burst some foam structures

Conclusion
Improving the strength and reducing the shrinkage are critical for the potential structural application of FC. e influence of SF and NS on strength and shrinkage was investigated and compared in detail. SF and NS, in varying amounts, were measured and compared in terms of their mechanical properties, shrinkage properties, microstructure, and ultrasonic velocity with the following results and conclusions.
(1) e compressive strength of FC peaked at 35.8 MPa and 32.8 MPa, when 4% NS or 15% SF was blended, indicating that it was a potential alternative for structural cases. e shrinkage was higher than that of ordinary concrete but acceptable. (2) FC flexural and compressive strength increased at first and then decreased as SF, or NS, was added. When SF and NS were 15% and 4%, respectively, FC compressive and flexural strength maximized. e amount of NS added, as of this moment, was 73.3% less than SF, but the resulting strength was higher compared to when FC was added to SF. (3) FC autogenous shrinkage value decreased at first and then increased as SF, or NS, was added. FC dry shrinkage value increased at first and then decreased as SF, or NS, increased. FC shrinkage was larger in FC with 4% NS than FC with 15% SF.

Data Availability
e data used to support the findings of this study are available from the corresponding author upon request.

Conflicts of Interest
e authors declare that they have no conflicts of interest.