Construction and Building Materials 35 (2012) 647–655 Contents lists available at SciVerse ScienceDirect Construction and Building Materials journal homepage: www.elsevier.com/locate/conbuildmat On the aspect ratio effect of multi-walled carbon nanotube reinforcements on the mechanical properties of cementitious nanocomposites Rashid K. Abu Al-Rub ⇑, Ahmad I. Ashour, Bryan M. Tyson Zachary Department of Civil Engineering, Texas A&M University, College Station, TX, USA h i g h l i g h t s " Integration of carbon nanotubes (CNTs) into the cement paste fraction of concrete. " CNT’s aspect ratio effect on mechanical properties of cement is investigated. " Low concentrations of long CNTs are more effective than high volume of short CNTs. " SEM and TEM images are obtained showing crack bridging and CNT pullout. " CNTs’ effect on mechanical properties is investigated at 7, 14, and 28 days. a r t i c l e i n f o Article history: Received 29 January 2012 Received in revised form 28 April 2012 Accepted 29 April 2012 Keywords: Carbon nanotubes Cement Aspect ratio Strength Surface area to volume ratio a b s t r a c t For their novel mechanical, thermal, chemical, and electrical properties, carbon nanotubes (CNTs) are widely used in many fields of nanocomposite materials. In cementitious nanocomposites, CNTs can act as effective bridges to minimize and limit the propagation of micro-cracks through the matrix, under the conditions of well dispersion of the CNTs within the matrix and good bonding between the CNTs and the surrounding hydrated cement matrix. This study focuses on the effect of different concentrations of long multi-walled carbon nanotubes (MWCNTs) – high length/diameter aspect ratios of 1250–3750 – and short MWCNTs – aspect ratio of about 157 – in cement paste. Flexural bending tests were performed to evaluate four major mechanical properties of the cement/CNTs composites at ages of 7, 14, and 28 days. Results show that the flexural strength of short 0.2 wt.% MWCNT and long 0.1 wt.% MWCNT increased by 269% and 65%, respectively, compared to the plain cement sample at 28 days. The highest increase in ductility at 28 days for the short 0.1 wt.% MWCNT and the short 0.2 wt.% MWCNT was 86% and 81%, respectively. It is concluded that nanocomposites with low concentration of long MWCNTs give comparable mechanical performance to the nanocomposites with higher concentration of short MWCNTs. Clear evidence was obtained from scanning electron microscope images for micro-crack bridging; many of the MWCNTs were stretching across the micro-cracks showing CNTs breakage and pull-out. Ó 2012 Elsevier Ltd. All rights reserved. 1. Introduction Concrete is the most widely used construction material in the world. However, the cement paste phase of concrete is a quasi-brittle material that has low tensile strength, low ductility, and early development and propagation of micro-cracks due to shrinkage at early ages. Therefore, there is a great desire to tailor the tensile and flexural mechanical properties of the cement paste in order to improve the damage and fracture resistance of concrete. Typical reinforcement of cementitious materials is usually done at the meso-scale (millimeter scale) and/or at the micro-scale using macrofibers and microfibers, respectively [1]. On the other hand, the ⇑ Corresponding author. Tel.: +1 979 862 6603; fax: +1 979 845 6554. E-mail address: [email protected] (R.K. Abu Al-Rub). 0950-0618/$ - see front matter Ó 2012 Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.conbuildmat.2012.04.086 nano-scale and unique multifunction properties of carbon nanotubes (CNTs) make them promising reinforcements to many engineering materials [2–4]. CNTs occur as single-walled carbon nanotubes (SWCNTs) and multi-walled carbon nanotubes (MWCNTs). SWCNTs are composed of a single graphite sheet rolled into a long hollow cylinder, whereas MWCNTs are nested arrays of SWCNTs. The average diameter of an individual SWCNT is on the order of 1 nm whereas the average diameter of an individual MWCNT is on the order of 10 nm. Typical properties of CNTs include an average Young’s modulus approaching 1000 GPa, exceptional tensile strengths in the range of 20–100 GPa, and an ultimate strain of 12% [5,6]. Therefore, when compared to strongest steel, CNTs have a modulus of elasticity of approximately five times higher, a tensile strength 100 times stronger, and elastic strain capacities 50 times greater, and yet a specific gravity one-sixth that of steel. 648 R.K. Abu Al-Rub et al. / Construction and Building Materials 35 (2012) 647–655 By now, it is well-known that there are two major challenges need to be solved in order to effectively obtain a successful CNTs/cement composite; the well-dispersion of the CNTs within the cement paste matrix, and the bonding or cohesive properties between the surface of the CNTs and the cement paste around them [7,8]. At the nano-scale the van der Waals forces between CNTs are relatively high, and due to their large surface-area-to-volume (SA/V) ratio, CNTs tend to attract to each other and agglomerate, making it difficult to disperse and separate them. The use of ultrasonic mixer with surfactants in aqueous solutions, with a specific amount of energy and time, a good dispersion could be achieved. However, CNTs might dissolve in the solution or broke down into smaller pieces if excessive amount of sonicating energy was used. The compatibility of the surfactant used to disperse the CNTs with cement is another important issue. The hydration and the chemical reactions of cement could be badly affected; it could delay or even stop the hydration and the hardening process of the cement paste [9]. Despite the aforementioned challenges, there has been a growing interest in integrating CNTs in cement paste since 2004. A short review of these developments is presented here. Makar and Beaudoin [10,11] are the first to integrate CNTs in cement paste and to show successful micro-crack bridging by CNTs. Li et al. [12,13] tested both functionalized (i.e. surface-treated) and non-functionalized MWCNTs within cement composites, and showed that the compressive and flexural strengths for the functionalized CNTs/cement composites were slightly higher than the non-functionalized MWCNTs/cement composites. Other researchers tried a novel approach by growing CNTs on the surface of the cement particles directly in order to enhance the dispersion and reduce the time and efforts of mixing and dispersing the CNTs within the cement paste [14,15]. In a series of studies, Shah and co-workers (e.g. [16–19]) have investigated the integration of various concentrations (as low as 0.02 wt.% and as high as 0.1 wt.% by weight of cement) of MWCNTs in the cement paste and consistently reported an improvement in the mechanical properties of the nanocomposite cement. They showed an increase in the flexural strength by 8– 40% and increase in the Young’s modulus by 15–55%. They also showed that low concentrations of well-dispersed MWCNTs lead to the largest enhancement. Through using nanoindentation of the cement paste reinforced with MWCNTs, Konsta-Gdoutos et al. [19] also reported a significant increase in the stiffness of the calcium–silicate–hydrate (C–S–H) suggesting that, besides the reinforcing action, CNTs produce significantly stronger and tougher cementitious materials than traditional reinforcing materials through affecting the nano-scale processes that control the C–S–H formation. Abu Al-Rub and co-workers (see e.g. [7–9,20]) have investigated the integration (dispersion and bond) of both CNTs and carbon nanofibers (CNFs) of different concentrations within the cement paste at different ages. Four mechanical properties have been investigated; strength, ductility, modulus of elasticity, and modulus of toughness, where modulus of toughness in this paper refers to the total area under the stress–strain diagram. In addition, the microstructure of the nanocomposites and dispersion of CNTs and CNFs have been investigated using ultra-resolution scanning electron microscope (SEM). The results showed that, generally, chemically-functionalized MWCNTs lead to degradation in the mechanical properties over time due to formation of expansive phases that weaken the formation of C–S–H, which is also shown in [21]. However, most of the fabricated nanocomposites showed improvements in the ductility compared to the plain cement specimens. Recently, Morsy et al. [22] have investigated the effect of different concentrations of MWCNTs with nano-clay particles on the compressive strength of cement mortar, and concluded that large concentrations of CNTs may agglomerate around the cement grains leading to partial hydration of cement and the production of hydrated products with weaker bonds. Moreover, they have noticed that the nano-clay particles improved the dispersion of CNTs within the mortar. Luo et al. [23] have tested MWCNTs with cement paste with silica fume, at different relatively high concentrations of MWCNTs (0.1, 0.5, and 1 wt.%), and reported 45% increase in the 28 day’s flexural strength. Hunashyal et al. [24] have fabricated hybrid composites of plain cement by integrating both carbon micro-fibers (CMFs) and MWCNTs, and obtained 88% flexural strength increase as compared to the plain cement specimens. Generally, results to date have shown modest improvement in the tensile/flexural strength of CNT/cement composites. Any improvement depends on many factors (e.g. nanotube content and aspect ratio, surfactants, sonication time and power, waterto-cement ratio, nanotube’s surface treatment). Since there has been no special emphasis placed on investigating the effect of aspect ratio (length-to-diameter ratio) of CNTs on the mechanical properties of cementitious nanocomposites, the objective of this study is to investigate the integration of long and short MWCNTs in the cement paste. The size and aspect ratio of CNTs mean that they can be distributed on a much finer scale than commonly used micro-reinforcing fibers. As a result, micro-cracks are interrupted much more quickly and frequently during propagation in a nanoreinforced matrix; producing much lower crack widths at the point of first contact between the moving crack front and the CNT. Fig. 1 shows graphically the relationships between the length-to-diameter aspect ratio and the surface-area-to-volume (SA/V) ratio for different lengths of SWCNTs, MWCNTs, carbon nano-fibers (CNFs), and carbon micro-fibers (CMFs). It is noticed from this chart that only CNTs can provide a very high SA/V ratio, which is one of the most important and desired elements in fiber-reinforced composite systems in order to obtain the best and the most efficient composite materials. A SWCNT is the only material which has a SA/V ratio that exceeds 2.0 nm-1, especially when considering the ultra-long SWCNTs that have aspect ratios that can reach several millions. A higher SA/V ratio means a larger contact area between the fibers and the surrounding matrix, hence higher interaction with the matrix and more efficient reinforcing. On the other hand, Fig. 2 shows further physical relationships for SWCNTs and MWCNTs for different parameters such as the aspect ratio, length, and surface-area-to-volume ratio of the CNTs, in addition to cement paste/CNTs composite parameters such as mass fraction of CNTs by weight of dry cement, and the approximate number of CNTs obtained for that specific length and mass fraction. In fact, Fig. 2 is very useful in showing the effect of several aspects of CNTs (SA/V ratio, CNTs’ concentration, aspect ratio, length, and number of CNTs). Therefore, based on the desired aspect ratio, one can find the corresponding SA/V ratio and the needed concentration and the type of CNTs. Fig. 2 demonstrates in a comprehensive manner the related features of using CNTs as reinforcements. Figs. 1 and 2 show conceptually the importance of aspect ratio and SA/V ratio of CNTs when used, particularly, in cementitious materials. Therefore, the focus of the current study is on investigating the effects of these parameters on the mechanical properties of cementitious materials reinforced with CNTs. In this study, different low and high concentrations of short and long MWCNTs are integrated in the cement paste with 0.4 waterto-cement ratio. Effective dispersion of CNTs within water was achieved by applying ultrasonic energy and with the use of a commercially available surfactant that is commonly used in cementitious materials. Three-point flexural bending tests were performed to evaluate four major mechanical properties (Young’s modulus, flexural strength, ductility, and toughness) of the cement/CNTs composites at ages of 7, 14, and 28 days. The microstructure at the fracture surfaces was investigated using a high-resolution electron microscope images. 649 R.K. Abu Al-Rub et al. / Construction and Building Materials 35 (2012) 647–655 Fig. 1. A chart showing the length-to-diameter (l/d) aspect ratio effect on the surface-area-to-volume (SA/V) ratio for different lengths of SWCNTs, MWCNTs, CNFs, and CMFs. * Reading the mass fraction values shall be through the curves with “*” numbers. Fig. 2. A chart showing the aspect ratio effect on the surface area/volume ratio for different lengths of SWCNTs and MWCNTs, and the relations between surface area/volume ratio with mass fraction of CNTs by cement weight and their number per 1 g of cement. 2. Experimental work 2.1. Materials The short MWCNTs are NC7000 multi-walled carbon nanotubes provided by Nanocyl Inc. with an average aspect ratio (length/diameter ratio) of about 157, whereas the long non-functionalized MWCNTs were provided by Cheap Tubes Inc. with an average aspect ratio of 1250–3750. Both short and long MWCNTs are produced by Catalytic Chemical Vapor Deposition (CCVD) process. The physical properties for both the short and long CNTs as provided by the manufacturers are shown in Table 1. The cement used in the mixtures is the commercial Type I/II Portland cement. A commercial water reducing admixture (polycarboxylate), provided by Grace Corporation, named ‘‘ADVA cast 575’’ was used as a superplasticizer or as a surfactant to aid in the dispersion of CNTs. All materials were used as received. 2.2. Preparation and dispersing of the MWCNTs Three different batches for the short MWCNTs were made at three different concentrations; 0.04%, 0.1%, and 0.2% of the weight of cement. For the long MWCNTs, two batches were made at 0.04% and 0.1% by cement weight. All speci- Table 1 Physical properties for the short and long MWCNTs. Diameter Length Specific surface area Purity Short MWCNTs Long MWCNTs 9.5 nm 1.5 lm 250–300 m2/g P90% <8 nm 10–30 lm >500 m2/g >95% mens had a water/cement ratio of 0.4, and a superplasticizer of 0.4% by cement weight. The reason why one more test of long MWCNTs at concentration of 0.2% is skipped is based on the conclusions of [18] as mentioned before, where less amount of long MWCNTs is needed to achieve the same level of composite characteristics. In addition to that, the dispersion process for a higher percentage of long MWCNTs requires more energy than other specimens, and that may affect the consistency of the testing. On the other hand, the short MWCNTs have been used at all three concentrations of 0.04%, 0.1%, and 0.2%. 650 R.K. Abu Al-Rub et al. / Construction and Building Materials 35 (2012) 647–655 For each batch, in order to get representative measurements for the mechanical characteristics, three to five replicates of each specimen were made and tested at ages of 7, 14, and 28 days, then the average value and the standard error for the replicates were computed at each age. CNTs were dispersed following the method described in [20] using an ultrasonic wave mixer provided from Sonics & Materials (Vibra-Cell), the model is ‘‘VC-505’’, with 20 kHz liquid processor with a ½ in. diameter titanium alloy probe was used to disperse the nano-filaments within an aqueous solution of water with superplasticizer at room temperature. The sonication power used was 78.64 W of 70% amplitude of the maximum (the power per solution volume was 0.4626 W/ml). The mix design parameters and the sonication time are provided in Table 2. The dispersion quality was evaluated at the fracture surface of the hardened CNT/cement paste composite using scanning electron microscope (SEM) imaging. Further information and discussions will be presented in the results subsection of this study. 2.3. Mixing and specimens preparation After dispersion, the CNTs solution was added to the cement at a water/cement ratio of 0.4, and then they were mixed together for 7 min using an electrical multispeed planetary blender. After mixing, the CNTs/cement mixture was placed in vacuuming chamber for 3 min to dispose any air entrained in the mixture in order to reduce the air bubbles. The mixture was then poured into square acrylic molds (of size 6.5 mm 6.5 mm 160 mm). During pouring process the acrylic molds were placed on an electrical vibrator table to get rid of the air bubbles generated during the casting process. All specimens were demolded at age of 7 days, and immersed in water saturated with lime for curing until testing. Plain cement specimens were casted as a reference sample. The plain cement mix was made with a water/cement ratio of 0.4, with addition of 0.1% of superplasticizer by the weight of cement. This amount of superplasticizer is equal to 25% of what has been added to other CNTs/cement composite specimens. This is different than the earlier work by the authors [20] where the same amount of superplasticizer has been used for both the reference and the nanocomposite samples. In this study, the reason why less amount of superplasticizer has been used in the plain cement specimens is that larger amount of superplasticizer is needed in the nanocomposites in order to improve the dispersion of the nano-filaments (see [7,8]). Around 75% of the added superplasticizer to the composites with CNTs is consumed during the dispersion process, and 25% of superplasticizer is left to enhance the workability of the cement paste which is comparable to that consumed by the reference samples. Therefore, this will allow more objective comparisons between the results from the reference and the nanocomposite samples. In fact, comparing the strength results from the reference samples in the current study and that in [20] shows an increase in the order of 100% due to the use of more superplasticizer in the reference samples in [20]. Therefore, the testing results from the nanocomposite specimens in [20] showed less gain in the mechanical properties due to the integration of CNTs as compared to the results from the plain cement reference specimens. This lack of objectivity when making the quantitative comparisons between samples without and with CNTs is avoided in the current study. Also, in the current study, the mixing time for the plain cement specimen has been reduced to 3 min after the addition of the cement to the water. The air vacuuming, casting, demolding, and curing of the plain cement specimen are the same as for all other composite specimens. 2.4. Flexural testing The testing apparatus used was a three-point bending testing frame of a small scale of dimensions (170 mm 25 mm 114 mm), fixed with an actuator (NSA12 from Newport Corp.), load cell (from Strain Measurement Devices of 1 and 2.5 kg capacity) and LVDT (from Macro Sensors) to measure the displacement, connected a data acquisition board (from National Instruments), a controller (Newport NSC200) with a LabView program was used to control the actuator and to measure the load versus the displacement. See Table 3 for the specifications of the employed actuator, which allowed us to get accurate load–displacement diagrams. The loads and displacements were measured and recorded by the software and then converted to stresses and strains using Euler–Bernoulli elastic beam theory in order to calculate some mechanical properties: r¼ LC F 4I and e¼ 12C L2 y ð1Þ Table 3 Newport NSA12 actuator specifications. Feedback Open loop, no encoder Drive screw pitch (mm) Travel range (mm) Minimum incremental motion (lm) Maximum speed (mm/s) Axial load capacity (N) 0.3048 11 0.2 0.9 28 where r is the flexural tensile stress at the extreme tension fibers of the beam, e is the elastic strain at the extreme tension fibers of the beam at each load step, L is the span length of the beam (160 mm), C is the half-depth of the beam cross-section (3.25 mm), I is the second moment of area of the square cross-section beam, F is the applied force measured by the load cell at mid-span length, and y is the deflection measured by the LVDT at mid-span length. 3. Results and discussion 3.1. Flexural testing results Four mechanical properties of the specimens were measured; flexural strength, strain capacity (i.e. strain-to-failure which is referred to here as ductility), modulus of elasticity (Young’s modulus), and modulus of toughness (defined as the area under the stress–strain diagram). Summary of the average values of the results compared to the plain cement reference specimens are shown in Fig. 3 through Fig. 6. The column charts show the mean values and the standard error of the mean for each specimen. At age of 7 days, most of the MWCNT/cement composites show an increase in the flexural strength, ductility, and modulus of toughness when compared to the plain cement (reference) specimens. The highest improvement in the flexural strength was shown in the short 0.04% MWCNT specimens with an increase of 66% compared to the reference specimens. The short 0.2% MWCNT specimens show a decrease in their flexural strength by 38%, while it showed the highest improvement in the ductility with an increase of 130% with respect to the reference specimens. The modulus of elasticity values for most of the specimens were close to the reference samples. However, the short 0.1% and 0.2% MWCNT show a slight increase compared to the reference samples. The highest improvement in modulus of toughness was provided by the short 0.04% MWCNT with a significant increase of 154% from the plain cement samples. On the other hand, considerable changes in the behavior of the composites were observed at age of 14 days. Large drop in the flexural strength of most of the composites occurred, to be less than the reference samples strength. The highest ductility obtained after 14 days was for the short 0.2% MWCNT with an increase of 72% from the corresponding value of the reference samples. No significant changes in the general behavior of the composites regarding the modulus of elasticity values with respect to the reference samples, but most of the specimens show decrease in the modulus of elasticity in general compared to the values at 7 days. Many of the nanocomposites showed degradation in modulus of toughness values in comparing to the reference samples, large decrease is noticed in the long 0.1% MWCNT. Table 2 Mix design of the test specimens. Test specimens Superplasticizer: % weight of cement CNTs: % weight of cement Ultrasonication time (min.) Plain cement (reference) Short CNTs 0.04 Short CNTs 0.1 Short CNTs 0.2 Long CNTs 0.04 Long CNTs 0.1 0.1 0.4 0.4 0.4 0.4 0.4 0.0 0.04 0.1 0.2 0.04 0.1 – 20 30 30 30 20 R.K. Abu Al-Rub et al. / Construction and Building Materials 35 (2012) 647–655 651 Fig. 3. Average flexural strength results for different MWCNTs composite specimens with the standard error of the mean. Fig. 4. Average ultimate strain results for different MWCNTs composite specimens with the standard error of the mean. After 28 days, all of the composites retrieved their strength values to become higher than the values at 14 days; however, the short 0.2% MWCNT showed a significant large increase in the flexural strength, specifically 269% compared to the reference samples’ values. In addition, the long 0.1% MWCNT has also increased by 65%. Most of the composites show a reduction in ductility in general, where the highest composites in ductility were the short 0.1% MWCNT and short 0.2% MWCNT; however, the short 0.04% MWCNT showed almost a constant value for ductility from age of 7 days through age of 28 days. All specimens showed an increase in their modulus of elasticity values compared to 14 days values, and the short 0.2% MWCNT has the highest increase. Due to the large increase in the flexural strength of the short 0.2% MWCNT, it has recorded the highest modulus of toughness value at age of 28 days. Therefore, from the reported results for the nanocomposites at different curing periods, one can notice that in most cases, the strength and Young’s modulus decreased from 7 day curing to 14 day curing and then increased (see Figs. 3 and 5). The opposite is seen for the ductility (see Fig. 4). However, overall, this is almost seen to be the case for the reference samples without the CNTs. Therefore, the authors believe that this observed trend may not be due to CNTs’ integration, but due to the curing process and the resulting properties of the cement paste matrix, which is discussed later in this section. It is interesting to note from Figs. 3–6 that cement nanocomposites with low concentration of long MWCNTs give comparable mechanical performance to the nanocomposites with higher concentration of short MWCNTs. For example, the average modulus of elasticity at 14 and 28 days for specimens with 0.04 wt.% of long MWCNTs is almost the same as that for specimens with 0.1 wt.% of short MWCNTs as shown in Fig. 5. This is in close agreement with the results in [18], which can be attributed to: (a) comparable effective interaction with the cement matrix of long CNTs with low concentrations as compared to that of short CNTs with high concentrations; (b) higher concentrations of short CNTs reduce the CNT-free volume of the cement paste which helps in arresting micro-cracks propagation; generally, one can see that the mechanical properties increase with the increase in the concentration (0.04, 0.1, and 0.2 wt.%) of the short MWCNTs; and (c) high concentration of short CNTs are more effective in filling nano-sized voids, which lead to enhancements in the packing density of the cement paste. The above conclusion of the effect of the CNT’s aspect ratio on the Young’s modulus is emphasized in Fig. 7 which shows the variation of the normalized Young’s modulus (normalized with respect to the Young’s modulus of the reference cement paste specimen) with the CNTs’ concentration for both short and long aspect ratios. The data points plotted in Fig. 7 for the normalized Young’s moduli are obtained from the literature as well as the 652 R.K. Abu Al-Rub et al. / Construction and Building Materials 35 (2012) 647–655 Fig. 5. Average modulus of elasticity results for different MWCNTs composite specimens with the standard error of the mean. Fig. 6. Average modulus of toughness results for different MWCNTs composite specimens with the standard error of the mean. current study and are summarized in Table 4. The data points are fitted using a power-law relation. One can clearly see from Fig. 7 that the Young’s modulus increases as the long-CNTs’ concentration decreases, whereas the Young’s modulus increases as the short-CNTs’ concentration increases. Also, one can notice from Fig. 7 that low concentrations of long-CNTs can lead to a much higher increase in the Young’s modulus as compared to higher Fig. 7. Variation of the normalized Young’s modulus with the CNT’s concentration for two different aspect ratios; long and short. concentrations of short-CNTs. For example, from Fig. 7, 0.04 wt.% of long-CNTs yields a comparable Young’s modulus to a 0.2 wt.% of short-CNTs. Although this conclusion is of a very high importance from the cost point of view; the lower the concentration of CNTs the lower the cost of the nanocomposite, achieving well dispersion of high concentrations of long CNTs in the material matrix is highly desirable for other multifunctional properties (e.g. electrical and thermal conductivity). Note that the rate of increase in the Young’s modulus with decreasing the concentration of long-CNTs is higher than the rate of increase in the Young’s modulus with increasing the concentration of short-CNTs. Therefore, it is more desirable to achieve better dispersion of long-CNTs since it is expected to lead to a much higher increase in the mechanical properties due to the higher SA/V ratio as indicated in Fig. 1. However, Fig. 7 shows that the Young’s modulus decreases at high concentrations of long-CNTs indicating the difficulty in dispersing high concentrations of long-CNTs as compared to short-CNTs. These results show that the aspect ratio of CNTs plays an important role in the reinforcement of cementitious materials. However, it should be emphasized that although the above observations are obvious for the Young’s modulus, but is not as obvious for the other mechanical properties. Moreover, the results from 0.1 wt.% long and 0.2 wt.% short MWCNTs are not similarly related due to the poor dispersion of long MWCNTs with higher concentration than 0.04 wt.%, and the ability of the short MWCNTs 653 R.K. Abu Al-Rub et al. / Construction and Building Materials 35 (2012) 647–655 Table 4 Summary of obtained Young’s modulus values from the literature for CNT/cement composites. CNTs: % weight of cement Nanocomposite Young’s modulus (MPa) Control Young’s modulus (MPa) Water/cement ratio Aspect ratio References Short CNTs 0.048 0.048 0.08 0.1 0.2 0.04 0.1 0.2 19.4 11.3 13 15.9 19.2 12.4 15.2 16.98 16.7 8.8 8.8 14.8 14.8 14.38 14.38 14.38 0.3 0.5 0.5 0.4 0.4 0.4 0.4 0.4 300 300 300 150 150 150 150 150 [16,18–19] [16,18–19] [16,18–19] [20,29] [20,29] Current study Current study Current study Long CNTs 0.025 0.048 0.025 0.048 0.2 0.04 0.1 22 19 13.8 12.7 17.5 15.25 12.99 16.7 16.7 8.8 8.8 23 14.38 14.38 0.3 0.3 0.5 0.5 0.4 0.4 0.4 1600 1600 1600 1600 2500 2500 2500 [16,18–19] [16,18–19] [16,18–19] [16,18–19] Current study Current study Current study Fig. 8. An example of the stress–strain diagrams for some samples of the 0.04% short MWCNT and 0.1% long MWCNT at ages of 14 and 28 days. The multi-peaks behavior can be observed. with higher concentration to reduce the CNT-free volume of the cement paste that help in arresting micro-cracks and filling nanosized voids. Therefore, the integration of 0.2 wt.% of short MWCNTs is more effective than integrating 0.1 wt.% of long MWCNTs. Based on the observation of the stress–strain diagrams of all specimens, one can notice the multi-peaks behavior that is related to the CNTs pull-out from the matrix. This is important for the fracture energy stored in the composite, which will affect the strain capacity (ductility) and modulus of toughness of the composite. An example of the stress–strain diagrams for some specimens is shown in Fig. 8. It has been noticed that although all samples of a batch are identical and casted from the same mix, there is variability in the results from one sample to another of the same composite. The uniformity of the distribution of the MWCNTs within the aqueous solution does not guarantee a uniform distribution of the nano-filaments within the composite, hence the stresses distribution in the composite beams will be non-uniform, and will cause variability in the flexural behavior from sample to sample. Besides the non-uniform dispersion and weak bonding, the degradation in the flexural strength in many specimens could be related to formation of weak hydration products as C–S–H that is a main product of the Portland cement hydration, which has an important contribution to the strength of the cement hydrate in different stages of hydration, [21] and [25], or due to existence of Fig. 9. SEM image showing the micro-crack bridging and breakage of the MWCNTs within the cement paste composite. harmful components like excessive formation of expansive ettringite, more details were discussed in the work of [20]. Another reason that the authors have investigated was the curing methods used to cure all the beam specimens. Leaching of Calcium Hydroxide (CH) [Ca(OH)2] from the cement paste submerged into the lime water, would increase the porosity (capillary voids) of the matrix and hence degrade the mechanical properties of the cement paste. Carde et al. [26] have shown experimentally that the compressive strength of plain cement paste of meso-cylinders (diameters of 10–30 mm) had degraded significantly due to the leaching process of CH. It was proposed that leaching of CH from the external layers (leaching zone) of the plain cement paste specimens and the loss of Calcium [Ca2+] ions content due to the progressive decalcification of the C–S–H, have a major contribution in the degradation of the mechanical properties of the cement 654 R.K. Abu Al-Rub et al. / Construction and Building Materials 35 (2012) 647–655 Fig. 10. SEM image of MWCNTs agglomerations within a small area of cement paste. paste [27]. For this study, the authors believe that although the leaching kinetics is usually slow, but due to the special case in this study of testing a very small specimens’ size (6.5 mm 6.5 mm cross-section) with large surface area to volume ratio – as in our case – will expedite the leaching process due to the increase in the leaching area to total area ratio. Crumbie et al. [28] have tested concrete specimens with w/c ratio of 0.4 and 0.6. The specimens were cured in saturated lime solution after 24 h of casting. The SEM back scattered images and the methanol-exchange porosity testing confirmed that the porosity and the permeability had increased near the surface. Also, they reported that the surface region was depleted in CH content. However, leaching of CH occurred although the specimens were immersed in saturated CH solution. CNTs breakage implies a good bonding between the CNTs surfaces and the surrounding cement paste. However, it is clear that the existence of the CNTs will affect the chemical reaction of the hydrated cement. The SEM images also indicate the high elastic flexural properties of the long MWCNT allow them to sustain large elongations without breakage. Although individual CNTs can be identified on the fracture surface in Fig. 10, relatively large agglomerations of the long MWCNTs can be seen even after the aforementioned dispersion and mixing processes. It was noticed from the SEM scanning of the fracture surface that some regions have no CNTs, while in other regions large number of agglomerated CNTs can be seen. This requires more investigations for improving the current dispersion techniques in cementitious materials. However, in order to effectively relate the SEM images to the aspect ratio effect on the mechanical properties of the nanocomposites, it should be noticed that the actual aspect ratio retained in the composite is a function of the unbroken length of the CNTs, i.e. the sonication process impose large amount of energy on the CNTs in order to well disperse them, but that also could break them into shorter fragments, especially the long ones. Obtaining transmission electron microscope (TEM) images for CNTs within the hardened cement paste is a very challenging task due to the brittleness of cement paste and in turn the difficulty in processing the cement paste sample for TEM imaging. Fig. 11 shows successfully obtained cryogenic-TEM images of the short and long MWCNTs within a thin strip of hardened cement paste. One can clearly see that the MWCNTs are nicely imbedded into the hydration products of the C–S–H phase, but closely spaced and to certain extent agglomerated. Also, one can notice that the long MWCNTs bridge the neighboring hydration products allowing for better crack-bridging and interaction with the cement matrix. This might explain the comparable increase in the mechanical properties of cement nanocomposites with low concentration of long MWCNTs as compared to the nanocomposites with high concentration of short MWCNTs. 3.2. Electron microscopy images 4. Conclusions In order to visually explain and justify the results obtained, many SEM images were taken for the tested samples. An example of the micro-crack bridging by the MWCNTs within the cement paste is shown in Fig. 9 with a scale of 100 nm. The images indicate clearly that many CNTs are bridging the micro-crack. Both pull-out and breakage of CNTs can also be observed from images. In fact, the The results obtained showed improvements in the flexural strength and ductility for all of the nanocomposites at age of 28 days when compared to the plain cement samples. It is observed that nanocomposites with low concentration of long MWCNTs give comparable mechanical performance to the nanocomposites with Fig. 11. Cryo-TEM images of short MWCNTs (left) and long MWCNTs (right) within the hardened cement paste. R.K. Abu Al-Rub et al. / Construction and Building Materials 35 (2012) 647–655 higher concentration of short MWCNTs. Moreover, it was found that the short MWCNT at 0.2% concentration by cement weight has better results than other specimens at age of 28 days. Those observations could be due to that short CNTs result in a relatively better dispersion of filaments within the cement paste, reduction in the filament-free volume of the cement paste, and effective filling of nano-sized voids. At early ages (age of 7 days) all specimens showed higher ductility than the plain cement (reference) samples. It was noticed that at age of 28 days, for the same concentrations of 0.1% and 0.04%, the long MWCNTs composites showed higher strength and less ductility than the corresponding short MWCNTs composites which can be due to breakage of long CNTs during the sonication process that in turn led to a relatively better dispersion of the CNTs. The nanocomposite specimens showed multi-peaks in the stress–strain response, which indicates multiple pull-out actions of the MWCNT from the hydrated cement matrix. However, this multi pull-outs behavior, from the stress–strain diagrams, could not improve significantly the strength or the ductility of the CNTs/cement composites in all cases; this could be due to the low bonding between the CNTs and the cement matrix or due to the formation of low-stiffness hydration products in the cement paste matrix due to the existence of the CNTs. The SEM images clearly show successful crack-bridging by the MWCNTs and pull-outs and breakage. The CNTs breakage implies a good bonding between the CNTs surfaces and the surrounding cement paste. The TEM images show clear embedment of CNTs within the cement hydration products, and bridging of neighboring hydration products by long MWCNTs indicating that MWCNTs with higher aspect ratios are more effective reinforcements if well-dispersed. Extensive studies are needed on the effects of the existence of the CNTs within cement paste on the chemical reactions and hydration process of cement. Improvements on dispersing techniques are also needed in order to effectively utilize long MWCNTs in cement composites. Also, novel multiscale computational modeling of CNT/cement composites that can be used carefully for investigating the effects of aspect ratio and various reinforcing mechanisms by CNTs is highly desirable. Acknowledgements The authors wish to acknowledge the financial support by Texas Department of Transportation through the Southwest University Transportation Center (SWUTC). Also, R.K. Abu Al-Rub would like to acknowledge the financial support provided by Qatar National Research Fund (QNRF) through the National Priority Research Program Project 4-1142-2-440. The QNRF funding supported the microscopy images in Section 3.2. References [1] Balaguru PN, Shah SP. Fiber reinforced cement composites. New York: McGraw-Hill Inc.; 1992. [2] Collins PG, Avouris P. Nanotubes for electronics. Sci Am 2000;2000(12):67–9. [3] Demczyk BG. Direct mechanical measurement of the tensile strength and elastic modulus of multiwalled carbon nanotubes. Mater Sci Eng A Struct Mater 2002;334(1):173–8. [4] Iijima S. Helical microtubules of graphitic carbon. Nature 1991;354(6348):56–8. 655 [5] Yu MF, Lourie O, Dyer MJ, Moloni K, Kelly TF, Ruoff RS. Strength and breaking mechanism of multiwalled carbon nanotubes under tensile load. Science 2000;287(5453):637–40. [6] Wong EW, Sheehan PE, Lieber CM. Nanobeam mechanics: elasticity, strength, and toughness of nanorods and nanotubes. Science 1997;277(5334):1971–5. [7] Yazdanbakhsh A, Grasley Z, Tyson B, Abu Al-Rub RK. Carbon nanofibers and nanotubes in cementitious materials: some issues on dispersion and interfacial bond. ACI Spec Publ 2009;267:21–34. [8] Yazdanbakhsh A, Grasley Z, Tyson B, Abu Al-Rub RK. Distribution of carbon nanofibers and nanotubes in cementitious composites. Transp Res Rec 2010;2142:89–95. [9] Abu Al-Rub RK, Tyson BM, Yazdanbakhsh A, Grasley Z. Mechanical properties of nanocomposite cement incorporating surface-treated and untreated carbon nanotubes and carbon nanofibers. ASCE J Nanomech Micromech 2012;2(1):1–6. [10] Makar JM, Beaudoin JJ. Carbon nanotubes and their application in the construction industry. In: 1st International symposium on nanotechnology in construction, Paisley, Scotland; 2004. p. 331–41. [11] Makar JM, Margeson JC, Luh J. Carbon nanotube/cement composite – early results and potential applications. In: 3rd International conference on construction materials: performance, innovation and structural implications, Vancouver, BC; 2005. p. 1–10. [12] Li GY, Wang PM, Zhao X. Mechanical behavior and microstructure of cement composites incorporating surface-treated multi-walled carbon nanotubes. Carbon 2005;43(6):1239–45. [13] Li GY, Wang PM, Zhao X. Pressure-sensitive properties and microstructure of carbon nanotube reinforced cement composites. Cem Concr Compos 2007;29(5):377–82. [14] Cwirzen A, Habermehl-Cwirzen K, Penttala V. Surface decoration of carbon nanotubes and mechanical properties of cement/carbon nanotube composites. Adv Cem Res 2008;20(2):65–73. [15] Nasibulin AG, Shandakov SD, Nasibulina LI, Cwirzen A, Mudimela PR, Habermehl-Cwirzen K, et al. A novel cement-based hybrid material. New J Phys 2009;11:1–11. [16] Shah SP, Konsta-Gdoutos MS, Metaxa ZS. Highly-dispersed carbon nanotubes– reinforced cement-based materials. USA: Publication USPA; 2009. [17] Shah SP, Konsta-Gdoutos MS, Metaxa ZS, Mondal P. Nanoscale modification of cementitious materials. Berlin: Springer-Verlag Berlin; 2009. [18] Konsta-Gdoutos MS, Metaxa ZS, Shah SP. Highly dispersed carbon nanotube reinforced cement based materials. Cem Concr Res 2010;40(7):1052–9. [19] Konsta-Gdoutos MS, Metaxa ZS, Shah SP. Multi-scale mechanical and fracture characteristics and early-age strain capacity of high performance carbon nanotube/cement nanocomposites. Cem Concr Compos 2009;32(2):110–5. [20] Tyson BM, Abu Al-Rub RK, Yazdanbakhsh A, Grasley Z. Carbon nanotubes and carbon nanofibers for enhancing the mechanical properties of nanocomposite cementitious materials. ASCE J Mater Civ Eng 2011;23(7):1028–35. [21] Musso S, Tulliani J-M, Ferro G, Tagliaferro A. Influence of carbon nanotubes structure on the mechanical behavior of cement composites. Compos Sci Technol 2009;69(11–12):1985–90. [22] Morsy MS, Alsayed SH, Aqel M. Hybrid effect of carbon nanotube and nanoclay on physico-mechanical properties of cement mortar. Constr Build Mater 2011;25:145–9. [23] Luo JL, Duan ZD, Zhao TJ, Li QY. Effect of multi-wall carbon nanotube on fracture mechanical property of cement-based composite. Adv Mater Res 2011;146–147:581–4. [24] Hunashyal AM, Lohitha SJ, Quadri SS, Banapurmath NR. Experimental investigation of the effect of carbon nanotubes and carbon fibres on the behaviour of plain cement composite beams. IES J Part A: Civ Struct Eng 2011;4(1):29–36. [25] Dweck J, da Silva PFF, Buchler PM, Cartledge FK. Study by thermogravimetry of the evolution of ettringite phase during type II Portland cement hydration. J Therm Anal Calorim 2002;69(1):179–86. [26] Carde C, François R, Torrenti J-M. Leaching of both calcium hydroxide and C–S– H from cement paste: modeling the mechanical behavior. Cem Concr Res 1996;26(8):1257–68. [27] Adenot F, Buil M. Modelling of the corrosion of the cement paste by deionized water. Cem Concr Res 1992;22(2–3):489–96. [28] Crumbie AK, Scrivener KL, Pratt PL. The relationship between the porosity and permeability of the surface layer of concrete and the ingress of aggressive ions. In: Roberts LR, Skalny JP, editors. Material research society symposium proceedings, Pittsburgh; 1989. p. 279–84. [29] Tyson BM. Carbon nanotube and nanofiber reinforcement for improving the flexural strength and fracture toughness of Portland cement paste. [Masters thesis]. College Station, Texas: Texas A&M University; 2010.
© Copyright 2026 Paperzz