Corrosion Resistance and Physical-Mechanical Properties of Reinforced Mortars with and without Carbon Nanotubes

Following the evolution of currently enforced Performance Based Design standards of reinforced concrete (RC) structures for durability, the designer, ra-ther than complying with given prescriptive limits, may instead specify a cementitious mix design that is proven to exhibit a code prescribed resistance level (class) to a given exposure environment. Such compliance will lead to the protection of the steel reinforcement from corrosion and the cementitious mortar from degradation, during the design lifespan of the structure, under aggressive environmental exposure conditions such as, marine or deicing salts and carbonation. In this context, the enhancement of the physical and durability properties of common cement-based mortars under chloride exposure are experimentally investigated herein. In particular, the experimental program reported herein aims to evaluate the influence of incorporating mul-ti-walled carbon nanotubes on the physical and mechanical properties of reinforced mortars against chloride ions. Furthermore, the anticorrosion protection of cementitious composites prepared with nanomaterials at 0.2% w/w is further investigated, by comparing all test results against reference specimens prepared without any additive. Electrochemical (Half-cell potential, corrosion current) and mass loss of reinforcement steel measurements were performed, while the porosity, capillary absorption and flexural strength were measured to evaluate the mechanical and durability characteristics of the mortars, following a period of exposure of eleven months; SEM images coupled with EDX analysis were further recorded and used for microstructure observation. The test results indicate that the inclusion of the nanomaterials in the mix improved the durability of the mortar specimens, while the nano-modified composites exhibited higher chloride penetration resistance and flexural strength than the corresponding values of the reference mortars. The test results and the comparison between nanomodified and reference mortars showed that the use of CNTs as addition led to protection of steel reinforcing bars against pitting corrosion and a significant improvement in flexural strength and porosity of the mortars.

noted that the action of chlorides and carbon dioxide (CO 2 ) is synergistic. Thus, in the case of coastal structures, while an amount of Cl − is bound by AFM phase to form chloride salts and reduce the permeability of the concrete, CO 2 reacts with the hydrated lime (CH) to form calcium carbonate CaCO 3 with the direct effect of pH reduction and, consequently, corrosion of the reinforcement [1].
From the above, it is obvious that in concrete works that are exposed to strongly chloride and carbonate corrosive environment, such as sea water, antifreeze salts, corrosive soils, industrial areas, etc., it is necessary and, nowadays, mandatory in currently enforced design [2] and product standards [3] to improve the physical-chemical properties of concrete and its corrosion protective role of the reinforcement. So far, the main methods of protection of RC elements include the use of coatings, cathodic protection, corrosion inhibitors, addition of finer pozzolanic materials, etc. The above methods have several disadvantages such as 1) the high maintenance costs of electrochemical methods, 2) the possible toxicity of corrosion inhibitors, and 3) the relatively slow effectiveness regarding the use of pozzolans. Following the evolution of nanomaterials, nanotechnology can also be used to produce new composite mortars with significantly increased strength and durability properties, for use in RC structures exposed to highly corrosive environments and/or extreme operating conditions.
There are two different structures of CNTs, the monofilament carbon nanotubes (SWCNTs) [4] [5] and the multilayer carbon nanotubes (MWCNTs) [6] [7]. Monofilament carbon nanotubes consist of a graphite sheet wrapped in a specified direction in a cylindrical shape. SWCNTs can be closed at their ends by "caps" with a hemispherical structure; their diameter does not exceed 2 nm, while their length often reaches 5 μm. Multilayer carbon nanotubes consist of a series of graphite sheets, which are wrapped concentrically into each other. The diameter of MWCNTs usually ranges between 3 nm and 250 nm. The distance between their walls is close to the distance between two graphite sheets (0.335 nm). Of particular research interest in recent years are MWCNTs comprising two graphite sheets (Double-Walled Carbon Nano-Tubes-DWCNTs) whose properties are similar to those of monofilament CNTs. They have the advantage that they are more chemically modified than monofilaments, to which some double bonds need to be broken in order to, chemically, add a group. In this way gaps are created in the structure of monofilament CNTs and therefore their electrical and mechanical properties are changed, in contrast to DWCNTs in which only the outer wall is modified.
Carbon nanotubes, in addition to their excellent electrical/thermal properties, exhibit very high tensile strength [8] and modulus of elasticity [9] [10] due to the covalent bonds (sp2 bonds) between the carbon atoms embedded in their grid. It has been reported [11] that the modulus of elasticity of CNTs exceeds 1 ΤPa while, in comparison, steel (high strength) reaches a maximum of 200 GPa. In addition, the tensile strength of CNTs is between 65 to 93 GPa, compared to that of conventional high strength structural steel ranging between 1 to 2 GPa [12]. Therefore, by mixing carbon nanotubes in the mortar mix, a significant improvement in the tensile strength of the mortar would be expected, which, in conven-Journal of Materials Science and Chemical Engineering tional mortar is extremely low (approximately ≈ 1:10 of the corresponding compressive strength); such an increase will also enhance the resistance in other tension related modes of failure, such as bending [13].
MWCNTs have a specific gravity of ≈0.18, a specific surface area > 200 m 2 /g and are added to the mortar mixture in small dosages [14]. The improvement of the mechanical properties of nanomodified mortar is due to the development of Van Der Waal forces between the nanomaterial and the hydration products (C-S-H and C-H) that prevent the material from breaking during tensile rupture through bridging between the cracked surfaces [15]. But apart from the excellent mechanical properties, carbon nanotubes have the ability to fill the gap between the hydration products and the unhardened cement grains, thus reducing the porosity at the micro-and nano-scales, as well as the permeability of the mortar.
These factors above lead to an increase in the density of the structural material and a corresponding reduction in the corrosion of the reinforcement [16].
However, the experimental results in the literature on the effect of introducing CNTs on the corrosion rate of steel embedded in mortar samples, are contradictory. The addition of CNTs caused higher corrosion intensities for cement mortars with up to 0.5% CNT subject to simulated sea water and accelerated carbonation exposure [17]. On the other hand, Gdoutos et al. [18] reported that the inclusion of 0.1 wt% CNTs decreased the corrosion rate and significantly increased the resistance to corrosion by delaying the onset of the corrosion reaction (initiation period). The authors further studied the influence of high CNTs content (0.5 wt%) on the corrosion behavior of mortars and showed that by adding higher amounts of CNTs, the reinforced mortars exhibited higher corrosion rate compared to that of mortars with 0.1 wt% CNTs. According to the authors, the observed delay in the onset of corrosion was due to a marginal decrease in permeability, leading to a reduction in the ingress of aggressive agents.
It was concluded, however, that this increased tendency to corrosion needed further investigation. Hassan et al. [19] also showed that the corrosion of reinforcing steel embedded in CNT mortars was affected by its diameter: more specifically, steel reinforcement of 16 mm diameter had a lower corrosion resistance than that of steel reinforcement of diameter 12 mm. It should be noted, however, that the behavior of CNT modified mortars also depends on the physio-mechanical characteristics of the CNTs, something which is investigated herein by using CNTs with different mechanical characteristics compared to Gdoutos et al. [18], as described further on.
Consequently, the main objective of the present experimental investigation is the study of the utilization of multi-walled carbon nanotubes (MWCNTs) for the production of high-performance cementitious mortars with improved physicalmechanical properties and chloride penetration resistance.

Experimental Set-up
In order to investigate the physical-mechanical properties and chloride penetra-

Materials, Chemical Properties, and Specimen Preparation
CEM I 42.5N cement was used as a binder, while calcareous fine aggregates (0 -4 mm) and water from the potable supply network were utilized in the preparation of the mixtures. In all mixture groups a constant ratio of raw material weights was used, namely cement: sand: total water = 1:3:0.65, which, for the aggregate used, corresponded to an active water cementitious (w:c) ratio of 0.59. The choice of a relatively high w:c ratio was intentional, so as to produce relatively low strength and high porosity mortars, representative of common mortar production, which would highlight the effect of CNTs in this type of common structural material. The steel reinforcement bars to be concreted were cleaned according to ISO/DIS 8407.3 with a solution of hydrochloric acid (HCl) containing an organic inhibitor for a period of twenty (20) minutes in order to remove the oxides and impurities on their surface. The bars were then immersed in deionized water and acetone to remove grease and oil from their surface and were subsequently placed in a glass desiccator up to concreting to prevent corrosion by atmospheric air. Finally, the reinforcement steel bars were weighed on a digital balance (0.1 mg) and connected with a copper wire so as to ensure the electrical continuity. The specimens were removed from the molds after 24 hours, remained in water for seven (2) days and were then partially immersed in 3.5 wt% solution.
For the physico-mechanical tests and the electrochemical measurements the following specimens were constructed in the lab: • Half-cell potential (HCP): 100 mm reinforced mortars with a diameter of φ50 mm in which a φ10 mm diameter, 100 mm long Tempcore B500C reinforcing steel bar was axially secured.
• Corrosion current: 100 mm reinforced mortars with a diameter of φ50 mm in which a φ10 mm diameter, 100 mm long Tempcore B500C reinforcing steel bar was axially secured.  Table 2. As can be seen, the nanomaterials used in the present study are shorter and have higher surface specific comparing to Gdoutos et al. [18] research; this is beneficial to fill the cracks and pores in cement matrix thus resulting a solid mortar with increased density and reduced porosity. For compaction, mechanical vibration was performed during casting to reduce the air content of the mixtures. The reinforcing bars were 20 mm away from the base of the mold so as not to be directly exposed to corrosion conditions (Figure 1(b)). The surface on the top of the mortar specimens and the protruding part of the reinforcement were then covered with Araldite epoxy resin to prevent galvanic corrosion. The specimens were water-cured for seven days with relative humidity RH ≥ 99% ± 1% and temperature T = 25˚C ± 1˚C to avoid cracking due to hydration heat release.

Three-Point Bending Test
The bending strength of nanocomposite and conventional mortars was measured by loading prismatic beams of dimension 40 × 40 × 160 mm 3 ( Figure 2) with a span length 120 mm, namely three times of the width (40 mm). The flexural strength of the mortar was obtained using Equation (1): where, fl σ is the flexural strength (MPa), F is the maximum force at fracture (N), L is the span length (mm) and b, d are the width and the height (mm) of the specimen.

Porosity and Capillary Absorption Measurements
The apparent density and open porosity of mortars were estimated using vacuum saturation; for the tests, cubic specimens with dimensions 50 × 50 × 50 mm 3 were used, as depicted in Figure 3. The test samples were full-dried in an oven at 110˚C ± 5˚C, before being subjected to vacuum saturation with water for 24 h. Following, the open porosity (P) was determined using Equation (2) below: where, W s is the liquid-saturated mass of sample (g), W d the oven dried mass (g) and W h the mass of the immersed liquid-saturated sample (g).
The capillary absorption is considered as a function of the density, the viscosity, and the surface tension of the liquid used for the test, and the pore structure of the test sample. In the present experimental study, the sorptivity was measured on cubic specimens 50 × 50 × 50 mm 3 using the method proposed by Gummerson et al. and Hall et al. [21] [22]. During the test, the level of methanol was constantly 2 -3 mm over the bottom side of the specimens, while the specimen's temperature was continuously recorded. The volume of liquid absorbed per square unit of absorbing surface area i was estimated using Equation (3) below.
The sorptivity (S i ) of the specimens was estimated as the slope of the corresponding i vs t graph, in accordance with Equation (4)   where, d m is the mass gain in g, a is the surface area of the immersed specimen side in mm 2 and d is the density of absorbing liquid in g/mm 3 .

Half-Cell Potential Measurements
Half-cell potential measurements were conducted for a period of 11 months according to ASTM C-867 [23], in order to evaluate the corrosion rate of steel reinforcing bars. The half-cell potential difference was measured using a highimpedance voltammeter and a reference electrode (Ag/AgCl), using the test setup shown in Figure 4.
In Table 3

Corrosion Current Measurements
A Potensiostat/Galvanostat Model 263A from EG&G Princeton Applied Research was used for the test, with the associated software for data acquisition and analysis, in order to analyze the test data. The potential scan range was ±10 mV  The Linear Polarization Resistance (LPR) Technique is a rapid and non-destructive electrochemical method for monitoring corrosion rate in real time. In LPR measurements the reinforcing steel is perturbed by a small amount from its equilibrium potential. This can be accomplished potentiostatically by changing the potential of the reinforcing steel by a fixed amount, ΔΕ, and monitoring the current decay, ΔΙ, after a fixed time [25].
Based on the Stern-Geary method, the polarization resistance is calculated using Equation (5).
From Equation (5), the corrosion rate I corr can be calculated by Equation (6): where β α ·β c are the anodic and the cathodic Tafel slopes, respectively. The current (I) was calculated using Equation (7), where A is the surface area of steel that has been polarized.
The mass loss was calculated by Equation (8): 1 IMt zF where β is the mass loss of the steel rebar (g), I is the corrosion rate (A), M is the atomic mass of the metal (equal to 56 g for Fe), t is the time of exposure (s), z is the ion chance (=2 for Fe → Fe2 + + 2e − ), and F is the Faraday's constant 96.500 (A × s).

Scanning Electron Microscopy (SEM)
The microscopical analyses were caried out on the fresh fractured prismatic specimens after the flexural strength test. As a final means of investigation of the nano modified mortars, and in order to establish the suitability of the CNT dispersion method and the progress of the hydration products with time, portions of the mortar samples after strength testing were obtained and analyzed visually using a JEOL, JSM-6610 LV SEM with a maximum resolution of 3.0 nm at 30 kV (in the high vacuum mode) and a magnification of 5x -300.000x. The optical observations were performed using a Backscattered Electron Detector (BSE).
Prior to the analysis, the samples were sputter coated with gold (50 nm thick) to make their surface conductive. The size, shape and morphology of the crystals were observed, and the structure of samples was investigated. The comparisons of the two mortar groups are given in the subsequent section.

Porosity and Sorptivity Measurements
The results of the porosity measurements performed on cement mortars using vacuum saturation at 7, 28, 120 and 240 days are shown in Figure 5, where it can be seen that both examined groups exhibit a gradual reduction in their porosity from 7 to 240 days. From the results it is also evident that the initial capillary porosity of the nano modified composites is lower than that of the REF mortars

Flexural Strength
The  The strength improvement in flexure may be the evidence of an improved bonding and cohesion between CNTs and the cement matrix. In addition, CNTs additions in mortar enhance its resistance through initial and gradual micro-cracking limitation; at the same time, the nanomaterials bridge the macro-cracks, thus delaying the crack propagation and improving the toughness [38] [39] [40] of the hardened paste. CNTs due to their chemical composition and surface area most probably interact with cement, leading to the observed strength enhancement. In addition to the above, due to its fineness (>500 m 2 /kg), CNTs probably act as a filler in the mixture, thus further reducing the pore size of the hardened composite and generating more nucleation sites to accelerate hydration reactions [41]; this nucleation effect influences the hydration process of the mortar by enhancing the crystallinity of the cementitious materials and leads to improved mechanical properties of the nanomaterial enriched cement composites. Makar et al. [42] [43] in their experimental study reported that the incorporation of CNTs in concrete accelerated the hydration process at early age. The authors considered that the latter effect was attributed to the polarization and adsorption between cement particles and CNTs, where the nano-size effect of CNTs provided the nucleation points for the hydration products. The researchers also proved the CNTs' addition effect on the hydration process and the morphology of hydration products (C 3 S and C 3 A).
These findings are also corroborated in the common mortar strength modified with CNTs of the present study: in Figure 9, the microscopy images by SEM of the microstructure of the nanocomposite mortars obtained on flat, polished surfaces of the CNT mortars used in this study, are depicted. These SEM analyses indicate that the nanomaterials demonstrate a network structure that can allow the bridging between narrow cracks in mortar (Figure 9(c)); in addition, it is clearly evident that the interfacial transition zone (ITZ) between cement and aggregates fills with CNTs resulting in a dense and solid structure of cement mortar (Figure 9(a)).

Half-Cell Potential (HCP)
The diagram in Figure 10 illustrates the half-cell corrosion potential (HCP) measurements, using a Ag/AgCl reference electrode for both test specimen groups  The results of the HCP on the CNT specimens in the present study were compared in Figure 10 with the corresponding HCP measurements on the REF specimens, established previously in [44], where mortar specimens with the same Journal of Materials Science and Chemical Engineering reference mix design were exposed to similar wet-dry cycles. From the compari-  Figure 11 shows the corrosion current (I corr ) for the two mortar groups under investigation; from the beginning of the immersion in NaCl 3.5 wt%, the reinforcement bars seem to be in an active state of corrosion, being in the depassivation region of the reinforcement corrosion state (  The addition of carbon nanotubes to mortars affects the durability of mortars in several ways [49]: the conductivity of mortar increases, the contact of steel with carbon leads to a galvanic element leading to increased oxidation of steel, the diameter of the mortar pores is reduced, and the mortar cracks are reduced and bridged and, generally, the micromorphology of the mortar is smoothed.

Corrosion Current
In terms of corrosion of the reinforcement, the conductivity of the mortar is also affected. When the CNTs are properly bonded with the reinforcement they can restrain crack formation, thereby improving load transfer and increased ductility of the composite structural material. However, the corrosion of the steel reinforcing bars embedded in cement mortars increases, especially when the CNT concentration is above a limit and percolation initiates, changing drastically the material properties. In the experimental investigation of Hall et al. [19], it was proven that, when the concentration of CNTs in the mortar exceeds the value of 0.5% by weight of cement, cement, corrosion increases dramatically. For this reason, a concentration of CNTs 0.2% by weight of cement was chosen in this study.

Chloride Content in Mortars
From the results illustrated in Figure 12, the REF and CNT specimens have similar chloride content values, measured by titration of AgNO 3 . In addition, after 9 months of partial immersion in chloride salt, there is a slight increase in the amount of total chlorides. In particular, the REF group exhibits 7.4% higher value of Cl − content than that of the CNT group at 9 months, while the corresponding value at 6 months was 8%. At the same time, it can be seen from the test results that the chloride content has increased from 6 to 9 months for the two mixtures; more precisely, the relative increase of the REF specimens at 6 months was 3.1%, while for the CNTs specimens was 3.6%. Journal of Materials Science and Chemical Engineering Figure 12. Total chloride contents (wt%) in mortars partially immersed in 3.5 wt% NaCl solution for 9 months. Figure 13 depicts the average mass loss of three (3) ), and are also reported in Figure 13.

Mass Loss of the Steel Reinforcing Bars
It can be seen from both measurements of mass loss (Gravimetric/Electrochemical) that the REF specimens are more corroded than the CNT specimens. For the 6-month period in a corrosive environment, the mass loss of the REF specimens was 50% greater than the mass loss of the CNT specimens. At the same time, for the period of 9 months exposure in NaCl, the mass loss of the reinforcement was 28% higher in the REF group compared to the CNT group.
The corrosion behavior of the specimens is similar when calculating the electrochemical loss based on the Faraday law. The electrochemical mass loss compared to gravimetric mass loss differs in the result by about 15 mg, being greater than the gravimetric mass loss (9 months, Figure 13). This is systematic between these two types of measurements of the mass loss, and within the limits obtained in similar mass loss comparisons between Faraday's law and the gravimetric measurement.
The electrochemical measurements and gravimetric measurements of the mass loss are also confirmed by the total chloride content test results shown in Figure   12. The total chlorides in the REF specimens are higher than those of the CNT specimens, confirming the reduction of the diameter of the mortar pores with CNTs compared to the reference specimens.
N. Chousidis et al. Figure 13. Gravimetric and electrochemical mass loss (mg) vs time, of the steel reinforcing bars embedded in mortars and partially immersed in NaCl for 9 months.
In conclusion, both the mass loss measurements shown in Figure 13 are and the chloride content measurements of the two types of mortars in Figure 12, demonstrate systematically that the steel reinforcing bars embedded in the CNT mortars exhibit higher chloride penetration resistance than the conventional reinforced specimens. Therefore, the addition of 0.2 wt% in the mortar mix also provides anticorrosion protection in the mortar reinforcement.

Conclusions
The results of an experimental study of the physical-mechanical and durability properties of common strength mortars with and without the inclusion of CNTs were presented and compared. Based on the findings of the tests and the discussion of the results, the following can be summarized in conclusion: 1) The addition of multi-walled CNTs improves the porosity and capillary absorption of cement mortars with time after immersion in NaCl solution.
2) In terms of measured half-cell potential, the introduction of CNTs on the mortars resulted in a reduction of 50% in the exponential decay factor of the measured half-cell potential.
3) The increase in the flexural strength ( Figure 8) against time of exposure in corrosive environment is attributed to the change in the density of test specimens and the porosity ( Figure 6). Electrochemical (corrosion current and half-cell potential) and chloride content measurements indicate the chloride penetration resistance and durability of CNTs mortars comparing with OPC specimens.
4) The mass loss of steel bars embedded in CNT-modified mortars is reduced, compared to those in the reference specimens; this confirms the anticorrosive effect of CNTs used at 0.2 wt% content in reinforced CNT modified mortar composites.
cial Fund-ESF) through the Operational Programme "Human Resources Development, Education and Lifelong Learning 2014-2020" in the context of the project "nanotechnology and metal fibers applications for the resistance of reinforced concrete exposed to highly corrosive environment" (MIS 5049426). Furthermore, the authors greatly appreciate the assistance of Professor Dr. I. Ioannou of the University of Cyprus, for providing us the facilities for performing the physico-mechanical tests and microscopical analysis. The views, statements and opinions presented in this paper are solely those of the authors.