Few studies have been carried out focusing on the effect ofCNFs–CNTs on the mechanical properties of cementitious composites(8–13). A summary is presented by the authors elsewhere (14). Thestudies have shown that CNFs–CNTs can improve properties such astensile and compressive strength, although marginally in most cases.A major problem that arises when CNFs–CNTs are used to reinforceany kind of material is dispersion. CNFs–CNTs strongly attract eachother due to van der Waals forces. This attraction results in the for-mation of agglomerations in the form of entangled ropes and clumpsthat are very difficult to disentangle. The dispersion problem has beencombated by methods like surface modification of the fibers and byusing surfactants, usually in combination with sonication, and alsoimplanting or growing the fibers directly on nonhydrated cementgrains. In this work, past experiments on CNT-reinforced cementitiouscomposites and the approaches chosen to address the problem ofdispersion are summarized. Then, an experimental program is pre-sented. The program was aimed at qualifying the effectiveness ofsuperplasticizers (high-range water-reducing admixtures) and ultra-sonic processing for the purpose of dispersing CNFs in water andfinally cement paste. In addition, the role of cement particle size indispersion and the advantages and shortcomings of using fine-graincement were studied.

MOTIVATION AND OBJECTIVES

The problem of dispersing carbon nanoparticles, in particular CNFs,in cement paste is new and a critical hurdle to using such materials incementitious composites. The advantageous properties of CNFs–CNTscannot be harnessed without controlling dispersion. DispersingCNFs–CNTs in liquids and solid composites has always been chal-lenging, but the challenge may be even greater with cementitiouscomposites. Hydration of cement can be affected negatively by manyof the well-known surfactants, which have been shown to be effectivein dispersing CNFs–CNTs in water and polymeric media. Moreover,the relatively short period of time between adding water to cement andthickening of the paste makes it difficult to exfoliate and distributeCNFs–CNTs during paste mixing. This study explores the problemof dispersing CNFs through an experimental program. The resultswill show how superplasticizers, which are hydration-compatiblesurfactants, can enhance the distribution of CNFs in water and cementpaste. The objective of this study is to demonstrate the obstacleslimiting uniform distribution of CNFs in cement paste. Through thereview of past work and an experimental program, problems such asthe resistance of CNF–CNT bundles to exfoliation, the effectivenessand limitations of hydration-compatible surfactants, the role of therelatively large size of cement particles on CNF–CNT distribution,and the negative effects of using fine-grain cement are examined.

Distribution of Carbon Nanofibers andNanotubes in Cementitious Composites

Ardavan Yazdanbakhsh, Zachary Grasley, Bryan Tyson, and Rashid K. Abu Al-Rub

89

Carbon nanofibers (CNFs) and nanotubes (CNTs) are known to beextremely strong and stiff, and their potential as reinforcement has beenof interest to many investigators in the past decade. One of the mostimportant keys for fully harnessing the properties of any type of fiber isto control the distribution in the material matrix. As far as CNFs–CNTsare concerned, the strong attraction among nanoscale fibers due tovan der Waals forces makes this task difficult. This study focuses on someof the problems that prevent a uniform distribution of CNFs–CNTs incement paste and the methods used in the past to enhance dispersion.The first phase of the experimental program investigates the effect ofusing superplasticizers accompanied by sonication on the dispersionof CNFs in water and paste. The second phase focuses on the problemof cement grain size and limitations that the use of fine grain cement causes.Finally, on the basis of results and past studies, suggestions are made forachieving enhanced dispersion of CNFs–CNTs in cement paste.

Because of their excellent mechanical properties, carbon nanotubes(CNTs) have been the subject of many investigations in the past decadeas reinforcement for several composite applications. Mechanically,CNTs exhibit elastic moduli of more than 1 TPa (1.5 × 108 psi) (1).Their theoretical strength is 100 times that of steel, at only one-sixththe specific gravity (2). Values as high as 60 GPa (8.7 × 106 psi) forultimate strength and 6% for ultimate strain have been reported (3, 4).Salvetat and Kulik reported an elastic strain capacity of 12%, whichis 60 times greater than that of steel (1). CNTs are also highly flexible,capable of bending in circles and forming knots. Like macroscopictubes, they can buckle or flatten under appropriate loadings (5).Yakobson and Avouris further summarized the mechanical behaviorof CNTs (6). CNTs are extremely small and their diameter is usuallyless than 20 nm. Carbon nanofibers (CNFs), on the other hand, arerelatively large; their diameter can be as large as 200 nm. Recently,Ozkan et al. performed direct mechanical measurements on CNFs (7).The CNFs they investigated had a tensile strength between 2 and 5 GPa(2.9 × 105 to 7.3 × 105 psi) with an average modulus of elasticity of300 GPa (4.4 × 107 psi).

A. Yazdanbakhsh, 503 CE/TTI Building; Z. Grasley, 503C CE/TTI Building; and B. Tyson, 705 CE/TTI Building, 3135 TAMU, Department of Civil Engineering,Texas A&M University, College Station, TX 77843-3135. R. K. Abu Al-Rub,710B CE/TTI Building, 3136 TAMU, Department of Civil Engineering, Texas A&MUniversity, College Station, TX 77843-3136. Corresponding author: Z. Grasley,zgrasley@civil.tamu.edu.

Transportation Research Record: Journal of the Transportation Research Board,No. 2142, Transportation Research Board of the National Academies, Washington,D.C., 2010, pp. 89–95.DOI: 10.3141/2142-13

DISPERSION OF CNFs–CNTs IN CEMENTITIOUSCOMPOSITES AND PAST EFFORTS

A common way to improve CNF–CNT dispersion is with surfactants.Dispersion of CNTs in liquid media is a well-researched topic andmany surfactants, combined with sonication, have been shown to beeffective (15). The surfactants add charges or hydrophilic ends tothe surface of CNTs while their bundles are exfoliated by sonication.During sonication, ultrasonic waves are usually transmitted from aprobe into a liquid and produce alternate expansions and compressions.The pressure fluctuations give birth to microscopic bubbles (cavities),which expand during the negative pressure excursions and implodeviolently during the positive excursions. As the bubbles collapse,millions of shock waves, acoustic streaming, and a combination ofpressure and temperature extremes are generated at the implosionsites. The cumulative amount of energy produced is extremely highand significantly accelerates chemical reactions and breaks the clumpsand agglomerations of particles.

Some investigators have first dispersed CNTs in water by usingsurfactants and sonication and then added the dispersion to cementfor mixing. Cwirzen et al. dispersed CNTs in water by sonicationwhile using polyacrylic acid polymers as a surfactant (10). Trettinand Kowald (9) reported an experiment in which they used differenttypes of polycarboxylate-based superplasticizers as surfactants toproduce cementitious CNT-reinforced composites. When the distri-bution of CNFs–CNTs in cement paste is considered, applicabilityof surfactants is limited. As reported earlier by the authors, many ofthe surfactants that can effectively disperse CNFs–CNTs in liquidsare incompatible with cement hydration; they can retard or preventhydration, entrap substantial air in the paste, or undergo reactionswith water-reducing admixtures, resulting in reagglomeration (14).

An alternative approach to the use of surfactants to enhance dis-persion is functionalization. In this method, polar impurities such ashydroxyl or carboxyl end groups are added to the outer surface ofCNFs–CNTs (16–18). A basic and common method of functional-ization is to oxidize the fibers in nitric acid or a mixture of nitric andsulfuric acids while the acid-treated CNF–CNT solution is being heatedor sonicated (19). Although acid treatment results in CNFs–CNTsthat are more soluble than pristine fibers, if not controlled properlyit can damage or shorten CNFs–CNTs or even dissolve them. Thesenegative effects are probably why the first efforts to use acid-treatedCNTs did not result in remarkable improvements in the mechanicalproperties of hardened cement paste (8, 20). Moreover, sonication(especially if it is done for a long period) can damage and shorten theCNFs–CNTs (7 ). The experiments of Cwirzen et al. have shownthat incorporating a small dose of properly functionalized CNTs incement paste (CNT–cement weight ratio of 0.045%) could improvethe compressive strength of hardened paste by up to 50%, which isa sign of good dispersion of the CNTs in the paste (10).

Another way to enhance dispersion of CNTs in cementitiouscomposites is to distribute and attach the nanotubes to the surface ofcement particles before hydration. Makar and Beaudoin dispersedCNTs (single-walled, diameter 1.4 nm) in ethanol by sonication andthen added cement to the liquid to form a slurry, which was furthersonicated for hours (21). The slurry was then allowed to dry andground to form a powder. Scanning electron microscope (SEM)images of the powder show that grains of cement powder havebeen well coated with CNT bundles. Makar et al. produced cementpaste using CNT-coated cement produced by a similar method (22).Through SEM imaging of the fractured surface of dry paste, they

90 Transportation Research Record 2142

found that distribution of the CNTs in the hydrated samples is notthe same as observed on the nonhydrated cement grains. CNT bundleswere smaller in apparent diameter and more widely distributed inthe hydrated matrix. The smallest bundles imaged had diametersless than 5 nm, suggesting that they were composed of only a few1.4-nm-diameter single-walled CNTs. Nasibulin et al. recently devel-oped a method to grow CNFs–CNTs directly on the surface of cementparticles (23). On the basis of experimental measurements, they notedthat the fibers were homogeneously dispersed in the paste made withthe produced cement and intermingled with the products during thehydration process. Using this cement resulted in a more than 100%increase in the compressive strength of hardened paste.

One obstacle that can interfere with a good dispersion of nano-particles in cement paste is the size of nonhydrated cement particles.The average size of a Type I cement grain is between 10 and 20 µm.Moreover, due to environmental moisture, cement starts to hydratevery slowly after production even in an airtight package. As a result,some of the cement grains are bonded to each other and form largerparticles or clumps. A typical cement particle (or clump) is very largecompared with the typical size of multiwall CNTs (diameter 10 nm,length 1.5 µm). The presence of large cement grains or clumps createsregions where no CNFs–CNTs can be present because they cannotpenetrate the cement grains or clumps. The lack of CNFs–CNTs incertain zones forces a higher concentration of CNFs–CNTs in theremaining material, resulting in an increased probability of clumpingof the CNFs–CNTs and poor dispersion. Using fine-grain cementcould solve this problem. The second phase of the experimentalprogram will investigate the feasibility of this solution.

EXPERIMENTAL

In this section, the summary of a two-phase experimental investigationis presented. The first phase focuses on using superplasticizer todisperse CNFs in water by ultrasonic processing and mixing theresulting dispersion with cement to produce paste. This experimentwas done with the objective of finding out whether a good disper-sion of CNFs in an aqueous solution containing cement-hydration-compatible surfactant can guarantee a uniform distribution of CNFsin the cement paste. The second part of the research presents observa-tions on the actual size of cement grains and shows the consequencesof using cements with very fine particles for improving the dispersionof CNFs–CNTs in cement paste.

Materials and Instruments

The CNFs used in this experiment are vapor grown; they were pro-duced by exposing gaseous hydrocarbons or carbon dioxide to hightemperatures (24, 25). The CNFs have a diameter between 60 and150 nm, a length between 30 and 100 µm, and a specific surface areabetween 50 and 60 m2/g. Type I portland cement was used. The cementwas fresh when purchased and was stored in air-tight containers. Thesurfactant was a high-range polycarboxylate-based water-reducingagent (superplasticizer). Dispersions were made by ultrasonicallyprocessing the CNFs in a water–superplasticizer solution. A 20-kHzsonicator with a 1⁄2-in.-diameter titanium alloy probe was used atan amplitude setting of 50%. Optical and electron microscopes wereused to image CNFs in dispersion and hardened cement paste. Foroptical microscopy, a Zeiss Axiophot microscope was used with a×40 dry objective lens. For electron microscopy, observation was

performed with an FEI Tecnai F20 transmission electron microscope(TEM) and a JEOL JSM-7500F SEM. For the second phase, a ballmill was used to break cement grains into smaller particles.

Method

In the first phase, a CNF dispersion was made by 15 min of ultra-sonic processing of CNFs in water–superplasticizer solution. The superplasticizer–water and CNF–water weight ratios for dis-persion specimens were 1.71% and 1.14%, respectively. The opticalmicroscope was used in the transmitted mode to observe the CNFdispersion. The imaging sample was prepared by placing a small dropof dispersion on a glass slide and covering it with a thin transparentslip. The dispersion was also observed by TEM, using the “cryo”method in which the diluted dispersion was rapidly frozen to lessthan −175°C and kept at this temperature throughout the imagingprocess.

The same dispersion was added to cement to make paste with awater–cement ratio of 0.35, which resulted in the CNF–cementweight ratio of 0.4%. A small multispeed planetary mixer was usedfor mixing paste by a method similar to that of ASTM C305-06but with a longer mixing time (total of 7 min) to enhance pasteuniformity. After the mixing was completed, the paste was cast intosmall beam molds. The beams were demolded after 24 h and thencured in lime-saturated water. Three weeks after casting, the beamswere broken. Their fractured surfaces were imaged by SEM afterbeing coated with a 4-nm-thick platinum–palladium (80:20) layer toenhance surface conductivity.

In the second phase, the cement grains were broken into smallerparticles by ball milling. Two cement specimens were milled; oneof them was plain cement and the other was cement containing 1.0%by weight of CNFs. Because ball milling is more effective in liquidmedia, a slurry was made by mixing cement, or cement–CNF powder,with isopropyl alcohol (cement–alcohol weight ratio of 70%). Halfthe cylindrical container was filled with steel balls and the gap betweenthe balls was filled with the slurry. The container was tightly closedand horizontally rotated for 15 h. Then, the milled slurry was separated

Yazdanbakhsh, Grasley, Tyson, and Abu Al-Rub 91

from the balls by using a sieve. The slurry was carefully dried in anoven at 100°C. The result, which was a soft and cracked cementcake, was easily powderized with a gloved hand. SEM was used toimage both untreated and milled cement grains. The two milledplain and fibrous cement specimens were used to make paste. Toexamine the viability of using fine-grain cement to produce paste, amix design with the same water–cement and superplasticizer–cementratios as in the first phase was chosen as the starting attempt to producepaste. However, because the water demand of fine cement is greater,more water and superplasticizer were added during mixing to reacha workable paste. The final amounts of water and superplasticizerwere calculated and are presented in the following section.

Results and Discussion

The optical microscopy images of the CNF dispersion show that thefiber clumps and agglomerations were effectively disentangled dueto the application of superplasticizer and sonication. Figure 1a showsthe CNFs in water–superplasticizer solution without ultrasonic pro-cessing. Large and small agglomerations can be seen in this image.Similar agglomerations were observed when the CNFs were sonicatedin water without the presence of a superplasticizer or any kind of surfactant. However, as shown in Figure 1b, the fibers are welldispersed in the solution after 15 min of sonication. TEM observa-tions showed that the dark areas in optical microscope images aremostly amorphous carbon or impurities and occasionally longitudinalbundles of CNFs. The TEM images also showed that many of theCNFs were broken and shortened after sonication. The bond betweenmatrix and fibers decreases when the fibers are shortened. This detailis significant because some of the important properties of fibrousmaterials, such as enhanced toughness, are related to the elastic andfrictional bond between fiber and matrix. Therefore, ultrasonic pro-cessing should be controlled and optimized for best dispersion andleast damage.

The SEM investigation of the fractured surface of CNF-reinforcedhardened paste suggests that the CNFs are not uniformly distributed inthe paste. In fact, no fiber could be seen in a large portion of the

20 um 20 um

(a) (b)

FIGURE 1 Effect of sonication on dispersion of pyrograph CNFs in aqueous media. In both specimens the proportions are the sameand fibers were added to a water–superplasticizer solution: (a) hand-shaken only and (b) sonicated for 15 min.

fractured surface area. Figure 2a shows some areas on the fracturedsurface; CNFs are absent even in crack openings. On the other hand,in some other areas CNFs could be seen either agglomerated or closelypacked (Figure 2b). These observations indicate that a uniform dis-tribution of CNFs in the dispersion used to make cement paste doesnot guarantee a uniform distribution in the paste matrix. The reasonsfor the poor dispersion of CNFs in the paste are unknown. One reasonmay be the adsorption of superplasticizer by cement during mixing.According to Banfill, layers of superplasticizer with a thickness

92 Transportation Research Record 2142

up to 60 nm are adsorbed on the cement grains (26). As a result, theremaining superplasticizer might not be enough to keep the CNFsseparated.

One of the least investigated causes of the poor dispersion ofCNFs–CNTs in cement paste is the size of cement grains. The averagesize of Type I portland cement is between 10 and 20 µm. Therefore,cement particles are very large compared with CNTs or even CNFs.The SEM observations in the second phase show the geometry oftypical cement particles (Figure 3). Many of the cement particles are

1 um 1 um

(a) (b)

FIGURE 2 (a) Fractured surface of hardened cement paste reinforced with CNFs (CNF–cement ratio of 0.4%). Image represents mostof the fractured surface; CNFs are not present, meaning they are not uniformly distributed in the paste. The vast area of the fracturedsurface without CNF shows that the size of cement particles cannot be the only reason for the nonuniform distribution of CNFs in paste.(b) Agglomerated and closely packed CNFs in some areas of the fractured surface of hardened cement paste made by mixing cementand CNF dispersion with uniformly distributed fibers.

10 um

(a) (b)

FIGURE 3 Type I portland cement particles. (a) Several particles are as large as 10 �m. (b) A particle can be seen at left that is >100 �m. Cement was purchased fresh and kept in an airtight container. Particles are much larger than expected average size ofcement grains. These large dimensions can be a cause of the poor dispersion of CNFs in paste.

as large as 10 µm and particles larger than 100 µm can be observed.These observations are in agreement with the measured cementparticle size distribution by laser diffraction techniques (27 ). Twoparameters affect the role of cement particle size on the distributionof fibers in cement paste. The first parameter is the size of the fibersrelative to cement particles. CNFs–CNTs are much smaller thantraditional fibers, so uniform dispersion at a certain volume fractionrequires much smaller spacing between fibers. Because CNFs–CNTsare much smaller than cement grains, the minimum consistentspacing between fibers is limited by the size of the cement grains;that is, the CNFs–CNTs cannot penetrate the cement grains and arethus separated by a larger than optimal spacing in the presence oflarge grains. By having larger than optimal spacing in certain regionsof the material, smaller than optimal spacing between CNFs–CNTsoccurs in other regions; the result is a greater potential for poordispersion.

The second parameter dictating the role of cement particle size onCNF–CNT distribution is the distance between adjacent fibers givena uniform distribution of fibers in the matrix. If this distance is largerthan the average particle size, then cement grains do not disruptuniform dispersion even if the grains are much larger than fibers.To make an approximate but reliable measurement of the distancebetween well-dispersed fibers, it was assumed that all fibers are alignedin one direction; that is, all the fibers are parallel. On the basis of TEMand SEM images of the CNFs, it was also assumed that the internaldiameter of a CNF tube is 1.5 times its wall thickness and that thehollow part of CNF remains empty during and after paste mixing.With these assumptions, using CNFs with an average diameter of100 nm in a paste with water–cement = 0.35 and CNF–cement = 0.004(same as in the paste made in Phase I), the distance between twoadjacent CNFs would be about 5 µm, which means that if the averagecement grain size were 4 µm, it would not have a negative effect ondispersion. However, when the average size of cement particles is10 µm or more, a uniform dispersion of CNFs with the mentioned

Yazdanbakhsh, Grasley, Tyson, and Abu Al-Rub 93

mix design and fiber size is not possible. Moreover, when a higherdose or fibers with smaller diameters are used, given a uniformdistribution, the space between the fibers is reduced. For example,for the CNF–paste weight ratio of 1.0, the space between adjacentfibers would be 2.7 µm. As another example, if CNTs with an averagediameter of 10 nm and a CNT–cement ratio of 0.5 are used, thespace will be as small as 440 nm.

The experiment in the second phase was carried out to find outwhether it is possible to use fine-grain cement to produce CNF–cementitious composites. Figure 4a shows the SEM image of thecement produced by ball-milling the Type I portland cement shownin Figure 3. The milled particles are very fine and their averagediameter is less than 1 µm. When this cement was used to make apaste with water–cement = 0.35, after adding water–superplasticizersolution to cement, mixing resulted in only a wet powder. The mixdid not turn into a paste until the water–cement and superplasticizer–cement ratios were increased to 0.52 and 0.03, respectively. Theproblem was more severe with the ball-milled cement–CNF mix;water–cement and superplasticizer–cement ratios of 0.80 and 0.02were used to achieve a paste mix. In both cases, the paste was veryviscous and had rapid exothermic heat output. However, despite thevery high water–cement ratio of the fibrous paste, and as a result thehigh magnitude of shrinkage stresses, the cracks were very fine anduniformly distributed throughout the fractured surface, indicatingthe viability of CNFs in effective distribution of stress in cementitiouscomposites (Figure 4b).

A main reason for the high demand of water by fine cement isthat breaking cement grains into smaller particles increases the sur-face area of cement. Higher surface area results in more reactivityand a larger amount of water required to be adsorbed on the surfaceof cement particles to produce a workable paste. In general, thesurface area of particles is inversely proportional to their dimension.In other words, if the particles of a given mass of cement are milledso that the nominal average diameter of the new particles is half

1 um 1 um

(a) (b)

FIGURE 4 (a) Type I portland cement ball-milled for 15 h. Average size of particles is smaller than 1 �m, and therefore the surface areaof the milled cement is significantly larger than the original cement. (b) Fractured surface of hardened cement paste made by ball-milledcement–CNF mix. The CNF–cement ratio is 1.0%. Although the water–cement ratio in this paste is very high (0.80), and therefore thepaste was subjected to significant shrinkage stresses, large cracks were not observed when the surface was scanned by SEM. Multipleclosely spaced nanocracks indicate the role of a high content of CNFs in stress distribution.

the original average particle diameter, the surface area becomestwice as large. Therefore, in the case of this experimental program,when the average size of the milled cement grains is less than1/10th the size of original particles, the new surface area is morethan 10 times greater, and therefore the required amount of adsorbedwater and superplasticizer becomes significantly greater. Usingfine-grain cement has other disadvantages. Mehta pointed out thatusing fine cement can have negative effects on durability (28). Thelarge amount of heat released during hydration of fine cement canlead to thermal cracking. Also, finer cements generally producemore chemical and autogenous shrinkage (27 ). As a result, millingcement or using very fine cement may not be a practical approachto enhance the dispersion of CNFs–CNTs in cementitious matrix.It should therefore be understood that CNFs–CNTs may not be ableto be distributed uniformly in a cement paste above a certain volumefraction, depending on the particle size distribution of the cement.This fact should be considered when the behavior of CNF–CNTreinforced cementitious composites is theoretically modeled andanalyzed.

Finally, the SEM observations of the fractured surface of hardenedfibrous paste made with nonmilled cement show that there are areaswithout any CNF with dimensions as large as hundreds of microm-eters between the areas that contain CNFs. This amount is muchlarger than the size of the largest cement particle. Therefore, the sizeof cement particles is not the only reason for poor dispersion. Twoapproaches can be used to improve dispersion. The first approach isto attach or grow fibers directly on nonhydrated cement grains. Asmentioned earlier, Makar et al. could attach CNTs to cement particlesby sonicating a solution containing isopropanol, cement, and CNTsfor 4 h (22). However, this method is not practical for CNFs becausethey will be severely damaged after a long period of sonication. Themethod of Nasibulin et al. (23) to grow CNF–CNT directly on cementparticles, although very promising, is novel and its feasibility issuesare not well studied. The other approach is to use functionalizedCNFs–CNTs. These fibers are much more dispersible and theirphysical properties—for example, surface tension and wettability—can be controlled during functionalization by parameters such asthe period of treatment and the concentration or intensity of thefunctionalizing agent.

CONCLUDING REMARKS

A summary of past efforts to disperse CNTs in cement paste waspresented. Although CNFs have lower strength and elasticity modulusthan CNTs, they are very strong and stiff compared with materialssuch as steel. The first phase of the experimental program showed that,when a superplasticizer is used as surfactant, CNFs can be uniformlydispersed in water by ultrasonic processing. However, a uniformdistribution of CNFs in cement paste will not be obtained by mixinga water–superplasticizer–CNF dispersion with cement. The SEMimages of the fractured surface of such a fibrous paste revealed theabsence of CNFs in a large portion of the surface and the presenceof closely spaced and packed CNFs in some other areas. To reach abetter distribution in paste, either functionalized and highly dispersibleCNFs should be used or the CNFs should initially be implantedor grown on cement particles. The second phase of the project wasdedicated to the relationship between cement particle size and thedispersion of CNFs–CNTs in paste. It was shown that large cementparticles prevent a uniform distribution, particularly when fibers are

94 Transportation Research Record 2142

very small or are used in high doses. However, breaking cement grainsinto smaller particles was shown to be impractical and problematic.Fine-grain cement is very reactive and consumes a lot of water at ahigh rate. As a result, producing paste with typical water–cementratios such as 0.40 or 0.45 would be very difficult or impossible. It istherefore important to use fresh cement with a minimal amount of largegrains and clumps for making CNF–CNT reinforced cementitiouscomposites.

ACKNOWLEDGMENTS

This study was sponsored in part by FHWA and by the SouthwestUniversity Transportation Center.

REFERENCES

1. Salvetat, J. P., and A. J. Kulik. Electronic and Mechanical Properties ofCarbon Nanotubes. Advanced Materials, Vol. 22, 1997, pp. 1–7.

2. Wong, E. W., P. E. Sheehan, and C. M. Lieber. Nanobeam Mechanics:Elasticity, Strength, and Toughness of Nanorods and Nanotubes. Science,Vol. 277, No. 5334, 1997, pp. 1971–1975.

3. Yu, M.-F., O. Lourie, M. J. Dyer, K. Moloni, T. F. Kelly, and R. S.Ruoff. Strength and Breaking Mechanism of Multiwalled CarbonNanotubes Under Tensile Load. Science, Vol. 287, No. 5453, 2000,pp. 637–640.

4. Walters, D. A., L. M. Ericson, M. J. Casavant, J. Liu, D. T. Colbert,K. A. Smith, and R. E. Smalley. Elastic Strain of Freely Suspended Single-Wall Carbon Nanotube Ropes. Applied Physics Letters, Vol. 74, No. 25,1999, pp. 3803–3805.

5. Lourie, O., D. M. Cox, and H. D. Wagner. Buckling and Collapse ofEmbedded Carbon Nanotubes. Physical Review Letters, Vol. 81, No. 8,1998, pp. 1638–1641.

6. Yakobson, B., and P. Avouris. Mechanical Properties of Carbon Nano-tubes. In Carbon Nanotubes (M. S. Dresselhaus, G. Dresselhaus, andP. Avouris, eds.), Vol. 80, Springer, Berlin, 2001, pp. 287–327.

7. Ozkan, T., Q. Chen, M. Naraghi, and I. Chasiotis. Mechanical and Inter-face Properties of Carbon Nanofibers for Polymer Nanocomposites.Society for the Advancement of Material and Process Engineering,Memphis, Tenn., 2008.

8. Li, G. Y., P. M. Wang, and X. Zhao. Mechanical Behavior and Micro-structure of Cement Composites Incorporating Surface-Treated Multi-walled Carbon Nanotubes. Carbon, Vol. 43, No. 6, 2005, pp. 1239–1245.

9. Trettin, R., and T. Kowald. Nanotubes fur Hochleistungsbetone (Nano-tubes for High-Performance Concretes). Betonwerk und Fertigteil-Technik(Concrete Precasting Plant and Technology), Vol. 71, No. 2, 2005,pp. 20–21.

10. Cwirzen, A., K. Habermehl-Cwirzen, and V. Penttala. Surface Decorationof Carbon Nanotubes and Mechanical Properties of Cement/CarbonNanotube Composites. Advances in Cement Research, Vol. 20, No. 2,2008, pp. 65–73.

11. Wang, X., J. Ye, Y. Wang, and L. Chen. Reinforcement of CalciumPhosphate Cement by Bio-Mineralized Carbon Nanotube. Journal of theAmerican Ceramic Society, Vol. 90, No. 3, 2007, pp. 962–964.

12. Yakovlev, G., J. Keriene, A. Gailius, and I. Girniene. Cement BasedFoam Concrete Reinforced by Carbon Nanotubes. Materials Science,Vol. 12, No. 2, 2006, pp. 147–151.

13. Yakovlev, G., Y. Keriene, and V. Krutikov. Nanoreinforcement of FoamConcrete. ETH Honggerberg, Budapest, Hungary, 2006, pp. 222–223.

14. Yazdanbakhsh, A., Z. Grasley, B. Tyson, and R. K. Abu Al-Rub. Car-bon Nanofibers and Nanotubes in Cementitious Materials: Some Issueson Dispersion and Interfacial Bond. ACI Special Publication, Vol. 267,2009, pp. 21–34.

15. Moore, V. C., M. S. Strano, E. H. Haroz, R. H. Hauge, R. E. Smalley, J. Schmidt, and Y. Talmon. Individually Suspended Single-Walled CarbonNanotubes in Various Surfactants. Nano Letters, Vol. 3, No. 10, 2003,pp. 1379–1382.

16. Gogotsi, Y., N. Naguib, and J. A. Libera. In Situ Chemical Experimentsin Carbon Nanotubes. Chemical Physics Letters, Vol. 365, Nos. 3–4,2002, pp. 354–360.

17. Gogotsi, Y., J. A. Libera, A. Guvenc-Yazicioglu, and C. M. Megaridis.In Situ Multiphase Fluid Experiments in Hydrothermal Carbon Nanotubes.Applied Physics Letters, Vol. 79, No. 7, 2001, pp. 1021–1023.

18. Yao, N., V. Lordi, S. X. C. Ma, E. Dujardin, A. Krishnan, M. M. J. Treacy,and T. W. Ebbesen. Structure and Oxidation Patterns of CarbonNanotubes. Journal of Materials Research, Vol. 13, No. 9, 1998, pp. 2432–2437.

19. Tchoul, M. N., W. T. Ford, G. Lolli, D. E. Resasco, and S. Arepalli.Effect of Mild Nitric Acid Oxidation on Dispersability, Size, and Structureof Single-Walled Carbon Nanotubes. Chemistry of Materials, Vol. 19,No. 23, 2007, pp. 5765–5772.

20. Kowald, T. Influence of Surface-Modified Carbon Nanotubes on Ultra-high Performance Concrete. International Symposium on Ultra HighPerformance Concrete, 2004, pp. 195–203.

21. Makar, J. M., and J. J. Beaudoin. Carbon Nanotubes and Their Applicationin the Construction Industry. 1st International Symposium on Nano-technology in Construction, Paisley, United Kingdom, 2004, pp. 331–341.

22. Makar, J., J. Margeson, and J. Luh. Carbon Nanotube/Cement Composites—Early Results and Potential Applications. 3rd InternationalConference on Construction Materials: Performance, Innovations andStructural Implications, Vancouver, British Columbia, Canada, 2005.

Yazdanbakhsh, Grasley, Tyson, and Abu Al-Rub 95

23. Nasibulin, A. G., S. D. Shandakov, L. I. Nasibulina, A. Cwirzen, P. R.Mudimela, K. Habermehl-Cwirzen, D. A. Grishin, Y. V. Gavrilov,J. E. M. Malm, U. Tapper, Y. Tian, V. Penttala, M. J. Karppinen, andE. I. Kauppinen. A Novel Cement-Based Hybrid Material. New Journalof Physics, Vol. 11, 2009, pp. 1–11.

24. Suresh, G. A. Processing and Properties of Nanocomposites, 1st ed.World Scientific Publishing Co., Singapore, 2007.

25. Baker, R. T. K. Synthesis, Properties and Applications of GraphiteNanofibers. WTEC Workshop Report on R&D Status and Trends inNanoparticles, Nanostructured Materials, and Nanodevices in the UnitedStates. World Technology Evaluation Center, Lancaster, Pa., 1997.

26. Banfill, P. F. G. A Discussion of the Papers “Rheological Properties ofCement Mixes” by M. Daimon and D. M. Roy. Cement and ConcreteResearch, Vol. 9, No. 6, 1979, pp. 795–796.

27. Bentz, D. P., E. J. Garboczi, C. J. Haecker, and O. M. Jensen. Effects ofCement Particle Size Distribution on Performance Properties of PortlandCement-Based Materials. Cement and Concrete Research, Vol. 29,No. 10, 1999, pp. 1663–1671.

28. Mehta, P. K. Durability—Critical Issues for the Future. Concrete Inter-national, Vol. 19, No. 7, 1997, pp. 27–33.

The Nanotechnology-Based Concrete Materials Task Force peer-reviewed thispaper.