Exploratory Investigation of Nanomaterials to Improve Strength and Permeability of Concrete

Celik Ozyildirim and Caroline Zegetosky

Concrete containing various supplementary cementitious materials (SCMs) such as silica fume, y ash, and slag has improved properties. Nanomaterials, new SCMs with possible applications in concrete, have the smallest particle size (less than 100 nm). Nanomaterials are reactive because of the small size and large surface area of the particles, and they have great potential in improving concrete properties such as compressive strength and permeability. This study evaluates the use of a variety of nanomaterials in concrete compared with conventional concrete and concrete containing common SCMs. The potential benets of using nanomaterials over other SCMs are high reactivity and cost-effectiveness; in addition, smaller amounts are necessary, resulting in less cement replacement. Concretes containing nanosilica and nanoclay were prepared in the laboratory. They were compared with concretes containing silica fume, y ash, slag, or only portland cement. Specimens were tested for compressive strength and permeability. The microstructure of selected concretes with improved compressive strength and permeability was analyzed by using an atomic force microscope and nanoindenter to explain the improvements. The results of this study indicate that some of the nanomaterials tested have potential in concrete applications. The microstructure of the nanosilica concrete was denser and more uniform than the conventional concrete microstructure. In addition, the nanosilica had the largest improvement in both compressive strength and permeability among the nanomaterials tested.

The durability of concrete is important, especially when concretes are exposed to the outdoors. Therefore, many new ways to improve concrete properties to ensure longevity are being researched. Nanotechnology is a relatively new technology that is being integrated into many applications, including electronics, telecommunications, and biomedicine. Manipulation at the nanoscale can change chemical reactions, temperature, electricity, and magnetism (1). Nanotechnology pertains to particles having at least one dimension between approximately 1 and 100 nm (ASTM E2456). Recently, nanotechnology has been applied in the production of cementitious systems to reduce the permeability of concrete, which is essential in extending the service life (2). In concrete, the nanomodication can result in improvements in permeability, strength, shrinkage, ductility, and impact resistance (1).
Virginia Transportation Research Council, 530 Edgemont Road, Charlottesville, VA 22903. Corresponding author: C. Ozyildirim, celik@vdot.virginia.gov. Transportation Research Record: Journal of the Transportation Research Board, No. 2142, Transportation Research Board of the National Academies, Washington, D.C., 2010, pp. 18. DOI: 10.3141/2142-01

Various nanoparticles, when used as supplementary cementitious materials (SCMs) in concrete, can improve and densify the cement matrix, leading to improved permeability and strength. The nanoparticles act as nuclei of hydration (3), possess pozzolanic behavior (3), and can ll the voids in the cement matrix (4 ). Pozzolans chemically react with calcium hydroxide liberated during hydration to form cementitious compounds (5). The large surface area of nanoparticles and the abundance because of the small size can facilitate the chemical reactions necessary to produce a dense cement matrix with more calcium silicate hydrate and less calcium hydroxide. This, in turn, should enhance the overall concrete performance. Nanoparticles are in general smaller than the commonly used SCMs, making them more reactive and effective. Different forms of nanosilica and nanoclays in cement paste have been shown to increase compressive strength, reduce permeability, and cause a denser microstructure (2). Nanoparticles can also strengthen the interfacial transition zone between the cement paste and the aggregate, which would lead to improved strength and permeability. For nanoparticles to be a substitute for other SCMs of larger particle size, equal or better performance at lower or equal cost is needed. This result may be achieved by using lower dosages of nanoparticles. Some nanomaterials for use in concrete include nanosilica (nanoSiO2), nanoclays, nanotubes, nanocomposites (example nanoclays), and nanotitanium dioxide (nano-TiO2) (2). There are various nanoproducts from different companies, and it is unknown which nanoproducts will improve the properties of concrete. This study focuses on air-entrained concrete and seeks to determine whether nanomaterials can increase strength, decrease permeability, and cause a denser cement matrix. If the properties of concrete are improved, evaluation of the microstructure should indicate the reasons for the change.

PURPOSE AND SCOPE This study was conducted to determine the impact of nanomaterials on concrete performance, especially strength and permeability. Twenty-six batches of concrete were made, including a nanosilica, six nanoclays, common SCMs (silica fume, Class F y ash, and slag), and only portland cement. The various concretes were analyzed by using the results from the following tests: 1. Fresh concrete tests: air content, density, and slump. 2. Hardened concrete tests: compressive strength, modulus of elasticity, permeability, and length change.
1

Transportation Research Record 2142

3. Microstructure tests: atomic force microscope (AFM) and nanoindenter. On selected concretes the uniformity of microstructure was correlated with the material properties.

TABLE 2

Mixture Proportions Control (lb/yd3) 635 1,804 1,148 286 None 0 0.45 Silica Fume (lb/yd3) 591 1,804 1,131 286 Silica fume 43 0.45 Nanosilica (lb/yd3) 616 1,804 1,145 286 Nanosilica 19 0.45

Material Cement Coarse aggregate Fine aggregate Water SCM Amount of SCM w/cm

METHODS Concretes were prepared in the laboratory, and specimens were cast for tests. The following sections present an overview and explain mixture proportions, standard concrete tests, AFM, and nanoindentation.

Overview Concretes were made with various amounts and types of SCMs (see Table 1). Fresh and hardened concrete properties were determined. The concretes that exhibited improvements in strength and permeability were evaluated at the nanoscale by using the AFM and nanoindenter.

NOTE: 1 lb/yd3 = 0.59 kg/m3.

Mixture Proportions Concretes were mixed in the laboratory in 0.75-ft3 (0.05-m3) batches with a conventional pan mixer. A total cementitious material of 635 lb/yd3 (375 kg/m3) was used at two watercementitious material ratios (w/cm). The w/cm of 0.38 was to represent high-strength concretes [greater than 6,000 psi (41 MPa) at 28 days] mainly used in beams, and the w/cm of 0.45 for the deck concrete [strengths exceeding 4,000 psi (28 MPa) at 28 days]. Two batches with two other w/cm are included in the tables for additional information. The replacement rates of the SCMs are given in Table 1. Replacement rates were selected on the basis of the literature and suggestions from the material providers. Some of the SCM was used only at the w/cm of 0.45 because this amount is more common in bridge structures and there were not enough nanomaterials to make additional batches at the other w/cm. All of the mixtures contained the same materials, except the SCMs. The coarse aggregate used was granite gneiss with a nominal maximum size of 1 in. Natural sand was used for the fine aggregate. The portland cement was Type I/II. All conTABLE 1 SCM Class F y ash Slag Silica fume Nanosilica NC vermiculite NC hydrophobic nano montmorillonite synthetic NC nanoparticle NC synthetic nanoparticle or colloid NC hydrophilic nano montmorillonite NC surface modied nano montmorillonite
NOTE: NC = nanoclay.

cretes contained commercially available air entraining admixture, and all concretes except the ones with Nanoclay (NC) 4 contained polycarboxylate-based, high-range water-reducing admixture. The mixtures with the promising strength and permeability values were selected for microstructure analysis. Table 2 presents a summary of the mixture proportions for the selected concretes that represent conventional concrete, concrete containing an SCM with a particle size of microns (silica fume), and an SCM with a particle size of nanometers (nanosilica), respectively. Concrete Testing Concrete specimens were made according to ASTM C192. Specimens were cured for 28 days at room temperature in the moist room. Permeability specimens were also subjected to accelerated curing [moist cured for 1 week at 73F (23C) and 3 weeks at 100F (38C)]. Accelerated cure indicates long-term permeability (6 months and beyond) at 28 days (6, 7 ). Fresh concrete tests included density (ASTM C138), air content (ASTM C231), and slump (ASTM C143). Hardened concrete tests included compressive strength (ASTM C39), elastic modulus (ASTM C496), permeability (ASTM C1202), and length change (drying shrinkage) (ASTM C157). The specimens were 4 8 in. for the strength and elastic modulus, 2 4 in. for permeability, and 3 3 11 in. for the shrinkage tests.

Information on Supplementary Cementitious Materials Designation FA S SF NS NC1 NC2 NC3 NC4 NC5 NC6 Average Particle Size 25 m <45 m 0.15 m 22 nm <33% more than 45 m; thickness in nm Sieve residue (63 m): max. 55%; thickness in nm 2 m long, diameter 30 100 nm Length and width 100s of nanometers, thickness 1 nm 1520 m, thickness in nm Material Form Powder Powder Densied powder Slurry with water Clay dispersed in water Powder Rod-shaped particles Powder Powder Powder Replacement (%) 21 40 7 3 3 0.5 0.5, 2, 4 0.25, 0.5, 2 0.5, 1, 3 0.5, 1, 3

Ozyildirim and Zegetosky

Atomic Force Microscope The AFM was used to observe visually the microstructure and to generate information on roughness (lack of uniformity) of the paste. Concrete samples for AFM imaging were polished with various diamond disks. Initially, an 80-grit disk was used. This was followed by 220 grit, 600 grit, and 1,200 grit. Once the concrete was polished, it was cut down to samples approximately 5 5 2 mm for use in the AFM. AFM height and deection images were made in contact mode where the tip of the AFM stays in contact with the sample while moving across it. Roughness measurements were made by performing power spectral density (PSD). The Nanoscope software used with the Veeco AFM executed a PSD on whole images. A root-mean-square (RMS) roughness value is obtained by integrating the PSD over a frequency and then taking the square root (8). The RMS roughness value takes into account the more than 200,000 data points in a randomly produced image. The images were 35.8 35.8 m in size. The roughness test enables analysis of the uniformity of the microstructure. A more uniform and dense microstructure (lower roughness value) should relate to a concrete with higher compressive strength, higher elastic modulus, and lower permeability. Because all of the samples were polished in the same manner, the roughness values are expected to indicate the relative impacts that the different SCMs have on the cement paste structures.

RESULTS Concrete Properties The concretes containing nanomaterials were expected to have a significantly increased compressive strength and lower permeability on the basis of the results in the literature. The fresh concrete properties of the 26 mixtures are listed in Table 3, and the hardened concrete properties are in Table 4. The tables are grouped according to the w/cm. Many of the concretes containing nanomaterials were successful in increasing the compressive strength as compared with the control concrete. The results are an average of two samples. The compressive strength of concretes with a w/cm of 0.38 ranged from 6,610 to 8,600 psi (45.6 to 59.3 MPa), and those with a w/cm of 0.45 from 4,420 to 7,080 psi (30.5 to 48.8 MPa). All values were satisfactory. Permeability samples at 28 days subjected to accelerated curing exhibited lower permeability than those subjected to standard curing. SCMs take time to contribute to lowering permeability because they react after hydration reactions. Therefore, the comparison of results of accelerated curing versus standard curing at 28 days indicated the contribution of the SCMs as summarized in Table 4 and displayed in Figure 1. Based on accelerated curing, the control concrete with a w/cm of 0.45 had a high permeability (exceeding 4,000 coulombs). The concrete with silica fume had very low permeability (<1,000 coulombs), and the concretes with fly ash or slag had low

Nanoindentation
TABLE 3 Fresh Concrete Properties Percentage 0 3 0.5 1 0.5 1 4 1.8 0 21 40 7 3 3 0.5 0.5 2 0.25 0.5 2 0.5 1 3 0.5 1 3 w/cm 0.38 0.38 0.38 0.38 0.38 0.38 0.40 0.41 0.45 0.45 0.45 0.45 0.45 0.45 0.45 0.45 0.45 0.45 0.45 0.45 0.45 0.45 0.45 0.45 0.45 0.45 Air (%) 5.8 6.8 2.8 3.2 2.7 4.5 5.5 9.5 6.6 6.9 6.7 8.4 6.8 5.4 3.3 7.4 8 4.7 7.5 4.9 8.5 5.8 2.5 2.3 3 8 Slump (in.) 2.75 9.25 2.5 0.75 3.5 2.25 3.25 1.5 3.75 4 5 4.5 2.75 9.3 2 2.25 2.25 2.5 4.5 8 4 2.5 3 5.5 4 4.5

Nanoindentation was also performed on the polished samples. Two methods were used. First, the samples were indented elastically by using the ramping mode of the AFM. Then, the samples were tested by applying loads to a three-sided pyramid (Berkovich) tip. In this project, the tip was pushed 5 m into each concrete sample, causing permanent deformation. The nanoindenter graphs load applied versus the displacement of the tip. A curve is created during loading and another curve is produced in unloading. The stiffness, S, which is equal to dP/dh (P = load and h = displacement), is found from the unloading curve and is used to nd the elastic modulus of the indented surface (9). Nanoindentation is important because of its ability to calculate the microscale modulus of elasticity, stiffness, and paste uniformity. The nanoindentation results were obtained by indenting three different paste areas on each sample. Four to seven indents were performed on each of the three paste areas. All indentation data for an individual concrete sample were put into one file and averaged. The average microscale elastic modulus (E) values for each type of concrete were compared to determine the relationship among the different microstructures. In addition, the macroscale and nanoscale E-values were compared and analyzed. Initially, indentation was attempted with the ramping mode of the AFM. However, the probes that were used were not stiff enough (k = 0.06 N/m) for the hardened concrete. The cantilever that the tip is attached to was bending instead of exing. Because the cantilever was not strong enough, the data did not accurately represent the indenting process because it was not a measurement of the tip deection. The tip that performed the best had a spring constant of 40 N/m, and it indented up to only 40 nm before bending. This did not yield enough data to nd the modulus of elasticity. A tip with a spring constant on the order of 100 N/m (which was not available) would allow more data to be collected from indentation.

Material Control NC1 NC5 NC5 NC6 NC6 NC3 Nanosilica Control Fly ash Slag Silica fume Nanosilica NC1 NC2 NC3 NC3 NC4 NC4 NC4 NC5 NC5 NC5 NC6 NC6 NC6

NOTE: 1 in. = 25.4 mm.

Transportation Research Record 2142

TABLE 4

Hardened Concrete Properties at 28 Days Compressive Strength (psi) 7,710 6,610 7,430 8,080 7,870 8,600 8,380 7,290 5,620 4,420 6,020 5,920 7,080 5,580 6,230 5,950 6,040 5,450 4,800 5,170 5,220 5,960 6,420 6,450 6,810 4,610 Permeability (coulombs) 3,300 3,200 4,158 3,770 3,887 3,889 3,795 2,606 5,319 5,516 2,614 2,456 3,004 5,086 5,988 3,394 3,960 3,909 5,849 6,613 5,057 5,319 6,858 6,754 5,941 5,082 Accelerated Permeability 2,656 2,665 3,784 3,024 2,761 2,743 3,042 2,312 4,225 1,517 1,151 865 2,280 4,055 5,417 3,038 2,791 3,338 4,326 5,994 4,188 4,163 5,424 5,020 4,395 4,479

Material Control NC1 NC5 NC5 NC6 NC6 NC3 Nanosilica Control Fly ash Slag Silica fume Nanosilica NC1 NC2 NC3 NC3 NC4 NC4 NC4 NC5 NC5 NC5 NC6 NC6 NC6

Percentage 0 3 0.5 1 0.5 1 4 1.8 0 21 40 7 3 3 0.5 0.5 2 0.25 0.5 2 0.5 1 3 0.5 1 3

w/cm 0.38 0.38 0.38 0.38 0.38 0.38 0.4 0.41 0.45 0.45 0.45 0.45 0.45 0.45 0.45 0.45 0.45 0.45 0.45 0.45 0.45 0.45 0.45 0.45 0.45 0.45

E (ksi) 3,955 3,685 3,800 4,270 4,095 4,095 3,515 3,520 3,550 3,255 3,330 3,265 3,595 3,520 3,525 3,515 3,330 3,620 3,350 3,010 3,500 3,580 3,415 3,910 3,695 3,085

NOTE: 1 ksi = 1,000 psi = 6.89 MPa.

permeability (1,000 to 2,000 coulombs). The concretes containing nanomaterials had moderate permeability (2,000 to 4,000 coulombs) to high permeability (>4,000 coulombs). The nanomaterial concrete with the lowest permeability (2,280 coulombs) contained nanosilica. The six concretes with a w/cm of 0.38 all had moderate permeability. In nanoclays at the w/cm of 0.45, for NC3 and NC4 going to a higher percentage of addition resulted in an increase in permeability and for NC5 and NC6 no change in permeability. This behavior can be attributed to the existence of an optimum replacement rate and possibly to inefficient mixing and high variability. Mixing was performed in a conventional 1.75-ft3 (0.05-m3) pan mixer, which has a low shearing action. There are high-intensity, high-shear mixers that can provide more thorough and uniform mixing. In addition, the selection of the proper dispersant would help in obtaining the desired mixing. The lack of improvement with a higher percentage of the nanomaterial was also observed in the NC6 mixtures at the w/cm of 0.38. Among the concretes with a w/cm of 0.45, the nanosilica concrete had the highest compressive strength. The silica fume concrete had the lowest permeability. They were selected with the control concrete for the microstructure evaluation. Figure 2 shows the air content, slump, and compressive strength of the three mixtures. They had similar workability indicated by the slump value, but the air content of the silica fume concrete was higher than the others. An increase in air content of 1% by volume relates to a drop in compressive strength on the

order of about 5% (10). The concrete with nanosilica had a higher compressive strength than the silica fume concrete but also had lower air content. Concretes with nanosilica or silica fume had higher strength than the control even though the control had the lowest air content. The size of the nanoparticle and the chemicals used in the slurry are expected to affect the air entrainment. The effect of the nanomaterials on the air content and air void parameters should be explored. The elastic modulus (E) values for all of the concrete mixes are also listed in Table 4. The values are for an average of two test specimens. They ranged from 3,010 to 4,270 ksi (20.8 to 29.4 GPa). The length change data are shown in Table 5 for 28 days, 16 weeks, and 32 weeks after 28 days of moist curing. The values indicate drying shrinkage taking place. Concretes made with the same materials had less shrinkage with a lower w/cm because they had a lower water content. Shrinkage is primarily related to the water content of the mixture if the other ingredients are kept constant. At 32 weeks, concretes with SCMs at a w/cm of 0.45 had less shrinkage than the control except for NC5 (3%). Similarly, at a w/cm of 0.38, SCM concretes had less shrinkage than the control concrete except for NC5 (0.5%).

Microstructure Testing Evaluation of the microstructure was conducted through the AFM images and nanoindentation.

Ozyildirim and Zegetosky

(a)

(b) FIGURE 1 Contribution of SCMs to concrete permeability for ( a ) w/cm 0.38 and ( b ) w/cm 0.45. Percentage of nanomaterial is given in parentheses.

AFM Images
AFM images of regular concrete, silica fume concrete, and nanosilica concrete were analyzed for the different characteristics of the microstructures. In a visual observation, the regular concrete microstructure appeared very gritty in nature (Figure 3a). The silica fume concrete microstructure appeared to be more mountainous (Figure 3b). In general, the attest, most uniform microstructure visually was that of the nanosilica concrete (Figure 3c). AFM images measuring 35.8 35.8 m were analyzed for roughness. Before the analysis, each image was plane tted and attened to eliminate image distortion from the microscope. The Nanoscope software used with the Veeco Instruments AFM has plane t and atten options that will remove the bow and tilt from images. The third order plane t was used to perform a plane t of the entire image about the x-axis. It is subtracted from the image to remove the bow and tilt (11). The third order atten was used to perform least-squares ts to each

scan line and subtract the t from the scan line to remove the bow and tilt. It also accounts for the Z-offset between the scan lines (11). Once the image distortion was removed, a PSD was performed and RMS roughness values were calculated. A higher roughness value represents a less uniform paste surface. The results are shown in Figure 4. The nanosilica concrete has the overall lowest roughness values, and the silica fume has the overall highest roughness values. Lower roughness values correspond to a more uniform cement paste. The high silica fume values might be caused by the agglomeration of the silica fume particles as a result of the difficulty in separating the densied silica fume particles during mixing.

Indentation
A range of four to seven indents were performed on each of the three paste areas on each sample by using the nanoindenter. E-values

Transportation Research Record 2142

FIGURE 2 Air content and slump for selected mixtures (1 in. 25.4 mm, 1 ksi 6.89 MPa).

(elastic modulus of concrete) were calculated from the following equations (9): S= dP 2 = Er A dh (1)

(1 2 ) + (1 i2 ) 1 = Er E Ei
where

Concrete containing nanosilica exhibited improvements in compressive strength compared with the concrete containing silica fume but was not as effective in lowering the permeability. The nanoindentation results and roughness results indicate that the nanosilica paste is stronger (highest E-value) and has a more uniform microstructure (lowest roughness value) than the silica fume paste and regular conTABLE 5 SCM Control NC1 NC5 NC5 NC6 NC6 NC3 Nanosilica Control Fly ash Slag Silica fume Nanosilica NC1 NC2 NC3 NC3 NC4 NC4 NC4 NC5 NC5 NC5 NC6 NC6 NC6 Shrinkage Data in Microstrain Percentage 0 3 0.5 1 0.5 1 4 1.8 0 21 40 7 3 3 0.5 0.5 2 0.25 0.5 2 0.5 1 3 0.5 1 3 w/cm 0.38 0.38 0.38 0.38 0.38 0.38 0.40 0.41 0.45 0.45 0.45 0.45 0.45 0.45 0.45 0.45 0.45 0.45 0.45 0.45 0.45 0.45 0.45 0.45 0.45 0.45 28 Days 340 300 385 315 280 360 400 315 420 390 350 375 355 320 440 355 385 335 410 485 410 390 515 410 385 370 16 Weeks 470 460 605 525 490 540 465 395 580 530 475 480 490 515 650 495 500 445 590 695 605 585 720 620 565 580 32 Weeks 560 515 600 530 485 530 505 460 690 630 590 575 590 575 650 510 500 440 595 685 625 600 720 645 580 610

(2)

S = experimentally measured stiffness of the upper portion of the unloading curve, A = projected area of the elastic contact, P = load, h = displacement, Er = reduced elastic modulus dened in Equation 2, = Poissons ratio, and i = values for the indenter tip. A total of 15 indents were performed on the silica fume concrete, 17 indents on the nanosilica concrete, and 20 indents on the regular concrete sample. All of the runs for an individual concrete sample were put into one le and averaged. The E-values given in the table below represent the average values (1 ksi = 6.89 MPa). This table includes elastic modulus values determined by using the standard elastic modulus test on the concrete cylinders (macroscale E).
Type of Concrete Control Silica fume Nanosilica Microscale E (ksi) 4,080 4,060 4,930 Macroscale E (ksi) 3,550 3,265 3,595

The two tests show the same trend in elastic modulus values. The nanoindentation values are averages of three paste areas in each sample, and the macroscale values are averages of two cylinders. The nanosilica concrete had the highest E-values, and the silica fume concrete had the lowest values. The macroscale values of E were lower than the microscale values. This result is attributed to the averaging value in the macroscale E, which includes the interfacial transition zone (weak zone).

Ozyildirim and Zegetosky

(a) FIGURE 3

(b)

(c)

AFM images, 35.8 35.8 m, of ( a ) control concrete, ( b ) silica fume concrete, and ( c ) nanosilica concrete.

crete paste. However, the amount and size of nanosilica may not be sufficient to ll in pores and reduce transport of uids within the concrete. Three percent nanosilica was used, which may not be enough or may not be well dispersed. The lack of nanosilica in some areas could lead to places with voids that are not lled, which provide easy transport of uids affecting the permeability. SUMMARY OF FINDINGS This study presents the mechanical properties of different mixtures incorporating various types of SCMs, including nanomaterials. Many of the types of nanomaterials improved the compressive strength and permeability of the concrete as compared with the control mix. In this study, nanosilica was the nanomaterial with the best impact on the concrete performance. Some nanoclays had a negative effect compared with the control concrete. However, the results for the concretes containing these nanoclays were in the same range as the results for the control concrete. The results from this study show that nanosilica does improve the cement paste uniformity, compressive strength, and elastic modulus over regular concrete and concrete containing silica fume. The nanosilica also improved the permeability over the regular concrete. However, the nanosilica when added at 3% did not improve the permeability over the silica fume concrete, which may be a result of the nonoptimized amount of nanosilica or the poor dispersion of the nanosilica particles in the concrete. The dispersion of the nanoparticles is a concern. Even if the nanosilica is added as a slurry to aid in

dispersion, clumping may still occur. High-intensity, high-shear mixing with the use of a proper dispersant would be helpful in thorough mixing with minimal clumping. The AFM images of regular concrete, nanosilica concrete, and silica fume concrete were analyzed for roughness. A lower roughness value correlates to a more uniform microstructure. According to the roughness results and visual observation of the AFM images, the nanosilica concrete has the most uniform microstructure, followed by the regular concrete; the silica fume concrete has the least uniform microstructure. The silica fume concrete may also have dispersion issues, and the densified silica fume particles may have formed agglomerates that are difficult to break.

FIGURE 4

Roughness values of three concretes.

Transportation Research Record 2142

The trend in elastic modulus values corresponds to the trend in roughness values. Because a lower roughness or more uniform paste corresponds to a higher strength, the elastic modulus should be higher for more uniform paste structures. As expected, the nanosilica concrete had the lowest roughness value and the highest elastic modulus.

REFERENCES
1. Birgisson, B. Nanomodication of Cement Paste to Improve Bulk Properties of Concrete. Presented at the National Science Foundation Workshop on Nanomodication of Cementitious Materials, University of Florida, 2006. 2. He, X., and X. Shi. Chloride Permeability and Microstructure of Portland Cement Mortars Incorporating Nanomaterials. In Transportation Research Record: Journal of the Transportation Research Board, No. 2070, Transportation Research Board of the National Academies, Washington, D.C., 2008, pp. 1321. 3. Li, H., H. Xiao, J. Yuan, and J. Ou. Microstructure of Cement Mortar with Nano-Particles. Composites Part B: Engineering, Vol. 35, No. 2, 2004, pp. 185189. 4. Shih, J., T. Chang, and T. Hsiao. Effect of Nanosilica on Characterization of Portland Cement Composite. Materials Science and Engineering A, Vol. 424, No. 12, 2006, pp. 266274. 5. Kosmatka, S. H., B. Kerkhoff, and W. C. Panarese. Design and Control of Concrete Mixtures. Portland Cement Association. Skokie, Ill., 2002. 6. Ozyildirim, C. Permeability Specications for High-Performance Concrete Decks. In Transportation Research Record 1610, TRB, National Research Council, Washington, D.C., 1998, pp. 15. 7. Ozyildirim, C. Effects of Temperature on the Development of Low Permeability in Concretes. VTRC 98-R14. Virginia Transportation Research Council, Charlottesville, 1998. 8. Digital Instruments. Nanoscope II Scanning Probe Microscope Instruction Manual, Version 5. Santa Barbara, Calif., 1990. 9. Oliver, W. C., and G. M. Pharr. Improved Technique for Determining Hardness and Elastic Modulus Using Load and Displacement Sensing Indentation Experiments. Journal of Materials Research, Vol. 7, No. 6, 1992, pp. 15641580. 10. Whiting, D. A., and M. A. Nagi. Manual of Control of Air Content in Concrete. Portland Cement Association, Skokie, Ill., 1998. 11. Digital Instruments, Veeco Metrology Group. MultiMode SPM Instruction Manual, Version 4.31ce. Santa Barbara, Calif., 2004.
The Nanotechnology-Based Concrete Materials Task Force peer-reviewed this paper.

CONCLUSIONS Nanomaterials added in small quantities can improve the properties of concrete. Dispersion of nanomaterials is a concern. AFM provides information on roughness to evaluate the microstructure of the concrete. Nanoindentation provides information on elastic modulus, which can be used to explain the changes in the microstructure of the concrete and also to verify the results of the AFM.

ACKNOWLEDGMENTS The authors thank the Virginia Transportation Research Council and the Federal Highway Administration for their support of this research. The authors recognize the knowledgeable advice provided by Edward Berger, Neal Berke, Surendra Shah, David Mokarem, Stephen Lane, Sakya Tripathy, and John Gaskins. Special thanks to Jae Hong Kim for his help in interpreting the AFM images. The authors appreciate the contributions of Bobby Marshall, Mike Burton, Bill Ordel, Andy Mills, and Aaron DuPont in the preparation and testing of the specimens.