Stress Corrosion Behavior of Ungrouted Pretensioned Concrete Beams

Prestressed concrete beams of size 150×150×1000mmwere designed, and two bonded cold-drawn 7mm steel wires were stressed at 70% UTS under service conditions before concreting. /e beams were cast with M40 grade concrete mix with various percentages of chlorides ranging from 0, 1, 2, and 3% by weight of cement and cured for 28 days. After 28 days, the stretching forces were released, the prestressing steel wire was allowed to regain its original length, the tensile stresses were transformed into a compressive stress in the concrete, and the stress corrosion behavior was assessed. Stress corrosion cracking (SCC) is due to the simultaneous action of stress, corrosive media, and material properties. /e stress corrosion behavior of ungrouted pretensioned steel was assessed by using various electrochemical techniques such as electrochemical noise, open-circuit potential measurement, AC impedance, and potentiodynamic polarization measurements. /e same experiments were conducted for rebars embedded in the concrete beam with various percentages of chlorides ranging from 0, 1, 2, and 3% by weight of chloride. After 30 days of exposure, the beams were tested for their flexural strength measurements to find out the load-bearing capacity.


Introduction
Prestressed concrete (PC) structures are commonly used for long-span bridges, dams, silos and tanks, industrial pavements, and nuclear containment structures [1]. Compressive stresses are induced in prestressed concrete either by pretensioning or posttensioning. In pretensioning, the steel is stretched before concreting. Pretensioned concrete is used in precast members such as roof slabs, piles, poles, bridge girders, wall panels, and railroad ties. In posttensioning, the steel is stretched after the concrete hardens [2]. Posttensioned concrete is used for bridges, large girders, floor slabs, shells, roofs, and pavements. PC is used in school auditoriums, gymnasiums, and cafeterias because of its acoustical properties and its ability to provide long, open spaces. One of the most widespread uses of PC is parking garages [3]. PC has high ability to resist the impact, high fatigue resistance, and high live load-carrying capacity and is highly crack resistant. e PC structures have smaller deformation and can absorb more tension than the nonprestressed concrete structure [4]. PC structures are usually constructed by using highperformance concrete containing high-strength steel under very high levels of tensile load. e damage due to microcracking leads to a loss of elastic modulus. And the presence or ingress of chloride ions may cause the corrosion of steel wires which will lead to sudden or brittle failure of steel wires causing premature failure of concrete structures [5][6][7][8][9].
High-strength steel is susceptible to stress corrosion cracking, hydrogen embrittlement, or a combination of both mechanisms called hydrogen-induced stress corrosion cracking in the presence of aggressive environments. And SCC is caused due to the combined effect of mechanical loading, aggressive environment and some defects in the material [10][11][12][13]. e defects present in the steel wires are the weaker spots from which corrosion initiates, and they act as a stress concentrated area where diffusion of hydrogen contributes towards the SCC damage process [14,15]. Pitting corrosion is considered as a signi cant cause of brittle failure of prestressed concrete structures [16,17]. Pitting corrosion reduces the cross-sectional area of the prestressing steel and hence the load-carrying capacity of the structure [18]. e reported failures due to SCC are the collapse of the Berlin Congress Hall [19], accident at the Uster indoor swimming pool in Switzerland, collapse of a railway overpass at Berghausen, Germany [20], and collapse of the Bickton Meadows bridge, the Ynys-Y-Gwas bridge in UK [21], and the Saint Stefano bridge in Italy [22]. Many researchers have studied the corrosion behavior of PC structures [23][24][25]. Studies revealed that strand corrosion decreases the ultimate strength and the ductility of pretensioned members [26,27]. Strand corrosion will a ect the structural stability in terms of prestress loss, strength degradation, structural cracking, and failure of structures [28][29][30][31][32]. e e ect of strand corrosion on structural behavior was studied by many researchers [33][34][35]. Li et al. [36] reported that the combined e ect of fatigue and corrosion behavior decreased the bond with increased fatigue stress under a constant action of corrosion. It was reported that the corrosion of the strand in the pretensioned PC structure was found to degrade the tensile strength of the strand instead of its bond strength [37]. It has been found that the stress corrosion is more dominant than pitting corrosion [18]. Pits may be an initiation site for SCC or hydrogen embrittlement (HE) [5,6].
In the present investigation, prestressed concrete beams of size 150 × 150 × 1000 mm were cast with two pretensioned steel wires and M40 grade concrete with 1%, 2%, and 3% chloride by weight of cement. e stress corrosion behavior of ungrouted pretensioned steel wires was assessed by using various electrochemical techniques such as open-circuit potential, AC impedance, linear polarization, Tafel extrapolation, noise measurements, and weight loss measurements. All these measurements were made by using the fabricated embedded corrosion rate monitoring probe (ECMP). e sensor was fabricated as per the procedure reported elsewhere [38]. e schematic representation of the ECMP used for the investigation is given in Figure 1 [39]. e experiments were carried out for the rebars embedded in the concrete beam, and the results are discussed in detail.

Materials Used in the Test.
Ordinary Portland Cement 43 grade as per the Indian standard (IS) 8112 was used for the investigation. e speci c gravity of the cement used was 3.14. e properties of OPC are shown in Table 1. Natural river sand of size below 4.75 mm conforming to zone III of IS 383-1970 was used as ne aggregates. Coarse aggregates used in this study consist of crushed stone of size 12 mm and below. e speci c gravity of the ne and coarse aggregates used was 2.64 and 2.63. Cold-drawn stress-relieved prestressing plain steel wire of 7 mm diameter having an ultimate tensile strength of 1470 N/mm 2 conforming to IS 1785, Part 1, was used for the study. e chemical composition of the prestressing steel used was given in Table 2. AR (analytical grade) NaCl purchased from Fischer Scienti c was used for the investigation.   In general, there are two ways in which prestressing of concrete by steel elements can be accomplished, namely, pretensioning and posttensioning. In pretensioning, the strands are tensioned before the casting of the concrete, and in posttensioning, the strands are tensioned after the concrete hardens enough to support the stress. In this work, the pretensioning technique was adopted.
e prestressed concrete beam of size 150 × 150 × 1000 mm was designed, and two bonded steel wires of 7 mm diameter were stressed at 70% UTS (ultimate tensile strength) under service conditions before concreting. e ultimate tensile strength of the 7 mm steel wire was 1470 N/mm 2 . e average yield strength of the 10 mm TMT steel rebar was 454 N/mm 2 and 6 mm stirrup was 210 N/mm 2 . e elastic modulus of the prestressing steel wire and the TMT rebar was 210 kN/mm 2 and 200 kN/mm 2 , respectively. e resultant stress diagram is given in Figure 2. e beams are reinforced with two 10 mm diameter rebars at the bottom, two 12 mm diameter rebars at the top, and 6 mm stirrups at 150 mm spacing. ECMP was embedded very near to the pretensioned steel. e schematic of the reinforcement concrete details is given in Figure 3. e beams were cast with M40 grade concrete mix of 1 : 0.98 : 2.96 with 0.4 w/c ratio. While casting, the concrete mix was admixed with 1%, 2%, and 3% by weight of cement of NaCl to accelerate the corrosion process. e cast specimens were allowed to remain in the steel mould for 24 hours at room temperature. After 24 hrs, the specimens were demolded with care so that no edges were broken and were subjected to cure for 28 days by covering with the wet sack. For each system, triplicate beams were cast. After 28 days of curing, the beams were subjected to corrosion over the period of 30 days, the stress was released, and the measurements were made. At the end of 30 days, the prestressed steel wire and the rebar embedded in the concrete beams were subjected to electrochemical noise, AC impedance, and potentiodynamic polarization measurements. But the open-circuit potential measurements were made after the curing period was over until the end of the 30th day.

Flexure Test.
e prestressed beams were subjected to the exural test to nd out the load-bearing capacity. After 28 days of curing, the beams were allowed to corrode over a period of 30 days, and the stress was released. After that, the beams were dried in the open atmosphere for 24 hours and subjected to the exural testing in the UTM (universal testing machine) of 100T capacity under two-point loading, and the load versus de ection curve and the initial crack load versus nal crack load were recorded. e experimental  Advances in Materials Science and Engineering setup is shown in Figure 4. Modulus of rupture (F b ) was calculated using the formula: where P is the maximum load at failure, L is the span length of the beam, b is the breadth of the beam, and d is the depth of the beam.

Electrochemical Noise (ECN).
e ECN is one of the nondestructive and nonintrusive techniques being used in monitoring the corrosion process of various metallic materials [40], and it can provide information about the corrosion mechanism [41]. It was found that the sensitivity of the ECN is much higher than that of the other techniques in identifying the localized corrosion process [42]. e ECN is achieved by simultaneous measurement of fluctuation of current and potential caused by the spontaneous electrochemical reactions [43]. e technique of measuring the electrochemical noise uses no applied external signal for the collection of experimental data. e ECN technique measures the signal perturbations with low-level fluctuations of the corrosion potential between two nominally identical electrodes which can be used in the automatic determination of corrosion type and speed. e noise corresponds with the low-level frequency noise (differential of the ZRA) signal but has much lower amplitude when general corrosion is involved [44]. e technique is widely used within the corrosion engineering as a useful corrosion monitoring technique. ECN measurements were carried out for the prestressed steel and embedded rebars in the chloride-contaminated concrete.

Open-Circuit Potential
Measurements. After 28 days of curing, the prestressing steel under stressed conditions and rebars embedded in the concrete beams with and without various percentages of chloride were subjected to open-circuit potential measurements with respect to ECMP. e positive terminal of the microammeter was connected to the prestressing steel and the negative terminal to the ECMP, and the potential versus time plot was plotted for the prestressed steel and rebars. Potential measurements were taken for a period of 30 days.

AC Impedance
Spectroscopy. An electrochemical impedance measurement is an appropriate method for corrosion studies, particularly for corrosion rate determinations, passivation, and passivity process and for the evaluation of different inhibitors. Here, the prestressing steel acts as a working electrode, and the stainless steel and ECMP act as a counterelectode and a reference electrode, respectively. Impedance measurements were carried out for the prestressed steel and rebars embedded in the chloridecontaminated concrete. A time interval of 10 to 15 min was given for the OCP to reach a steady state value. e impedance spectroscopy measurement was carried out by using ACM Instruments, UK. e real part (Z′) and the imaginary part (−Z″) of the cell impedance was measured for various frequencies (30 kHZ to 10 mHz). Plots Z′ versus −Z″ were made. Impedance measurements were carried out at the end of the exposure period of 30 days.

Potentiodynamic Polarization Technique.
e same experimental setup used in the AC impedance technique was used here also. e ACM instrument has provisions for programs to evaluate corrosion kinetic parameters such as I corr and E corr . e potentiodynamic condition corresponds to a potential sweep rate of 0.1 mVs −1 and a potential range of +200 to −200 mV from OCP. Polarization studies were carried out for all the systems. All the experiments were carried out at a room temperature of 30 ± 2°C.

Electrochemical Noise (ECN).
e ECN measurement was performed using ACM instruments, UK, at the end of the 30-day exposure period. e two-electrode system was used for taking the measurements. Two identical prestressing steels were used as the working electrode 1 and working electrode 2, and the ECMP was used as the reference electrode. e electrochemical current noise and the potential noise between the coupled working electrodes and the reference electrode were measured simultaneously. ECN records were taken for 1000 s containing 1000 data points, recorded with a datasampling interval of 1.0 s (1 Hz). e current versus time noise and potential versus time noise plots for the rebars and prestressed steel are given in Figures 5 and 6. From Figure 5, it is observed that the potential and current noise exhibit different characteristics during the experimental period. e potential signal is found to have less fluctuation when compared to the current noise signal. Subsequently, the amplitude of fluctuation was found to increase with the increase in chloride content in both the prestressed steel and the rebar. e current signal fluctuation was found to vary between 0.5 µA and 5.5 µA. e potential fluctuation shift was found to be −190 mV, −290 mV, −375 mV, and −440 mV for 0, 1, 2, and 3% chloride level, respectively, in the rebar embedded in concrete. It indicates that, in the rebars embedded in 3% chloride only, the corrosion was initiated. In the case of the prestressed steel (Figure 6), the potential uctuation was found to be −175, −280, −410, and −510 mV, respectively, at 0, 1, 2, and 3%, respectively, indicating the active condition of the steel wire at 2 and 3% chloride levels. e potential noise uctuation at 3% chloride was found to be 2.91 times higher than that at 0% chloride. As the chloride level increases, the potential noise uctuation was also found to increase. e current uctuation was varied from 0.50 µA to 10.0 µA. e current noise uctuation at 3% chloride level was found to be 20 times higher than that at 0% chloride level which indicates the severity of the corrosion in the stressed condition. ere is a sudden shift observed in both the potential and current noise at 1 to 2% chloride levels, and from 2 to 3% chloride, the potential and current noise uctuation is found to be lesser when compared to 1 to 2% chloride levels. Zhao et al. [45] observed the same trend of ECN behavior in chloride-contaminated cement mortar-containing reinforcing steel.

Open-Circuit Potential Measurement.
e opencircuit potential measured for the prestressed steel against the ECMP in various chloride concentrations is illustrated in Figure 7. From the gure (Figure 7(a)), it is observed that the potential of the prestressed steel (without chloride) is found to vary between −121 and −225 mV versus ECMP indicating the passive condition of the steel wire throughout the exposure period of 30 days. However, in 1% chloridecontaminated concrete, the potential was found to vary between −242 mV during the starting period and it crossed the threshold limit of −275 mV after the 4th day of exposure, and the potential was shifted towards the negative direction In 3% chloride, the rebar has shown a more negative potential of −500 mV versus ECMP at the end of the 30th day indicating the active condition of the rebar. Figure 7(b) represents the potential versus time plot of the prestressed steel with various chloride levels. From the results, it is found that the prestressed steel in control concrete is under passive condition showing the less negative potential of −250 mV versus ECMP at the end of 30 days. Whereas in 1%, 2%, and 3% chloride-contaminated concrete, the prestressed steel has shown more negative potentials than the unstressed steel indicating the active nature of the steel after 30 days of exposure. In 3% chloride, the prestressed steel has shown a more negative potential of −540 mV versus ECMP after 30 days of the exposure indicating that the tendon has corroded severely.

AC Impedance.
e Nyquist plot of the prestressed steel and the rebar embedded in concrete with di erent % of chloride levels is given in Figure 8. e parameters obtained from the impedance technique for the prestressed steel and the rebar in various chloride-contaminated concrete after 30 days of exposure are shown in Table 3. From the table, it is found that the stressed steel has shown higher corrosion rate than the unstressed steel in all the chloride levels. After 30 days of exposure in chloride-contaminated concrete, the prestressed steel has 1.06, 1.03, and 1.48 times higher corrosion rate than the rebar embedded in 1%, 2%, and 3% chloride-contaminated concrete. For example, I corr value for the rebar and the stressed steel in 3% NaCl was 1.224 × 10 −3 mA·cm 2 and 2.098 × 10 −3 mA·cm 2 , respectively.
R ct values of the rebar and prestressed steel in 3% NaCl were 2.131 ×10 4 and 1.441 ×10 4 Ω·cm 2 , respectively. e corrosion rate of the rebar and the prestressed steel in 3% NaCl was 1.419 × 10 −2 mmpy and 2.098 × 10 −2 mmpy, respectively. In both the prestressed steel and rebar, the R ct values were found to decrease and C dl values were found to increase with the increase in chloride concentration, respectively. e I corr values were found to increase with the chloride concentrations irrespective of the stressed conditions. Compared to the stressed steel, the rebar has lesser corrosion rate, higher R ct values, and lower C dl values. Figure 9 depicts the impedance modulus curve for the rebar and prestressed steel with di erent percentages of chloride levels after 30 days of exposure. It is found that the impedance values were found to decrease with the increase in chloride concentration. When compared to the prestressed steel, the rebar has lesser impedance values than the prestressed steel in all the chloride levels. In the stressed condition, as the chloride level increases, the impedance value was found to decrease by one order of magnitude at each chloride level. e lowest corrosion resistance of the steel was observed in 3% NaCl in both the prestressed steel and the rebar. Slight capacitive behavior with the lowfrequency domain is observed in both the stressed and unstressed steel wires at all chloride levels.
ese results could be explained by taking into consideration that the chlorides present in the surrounding concrete reached the steel wire and activated/initiated the corrosion process [46].

Potentiodynamic Polarization Studies.
e potentiodynamic polarization curve for the rebar and the prestressed steel under various chloride concentration levels is depicted in Figure 10. e corrosion kinetic parameters obtained from the potentiodynamic polarization studies are given in   Control 1% NaCl 2% NaCl 3% NaCl   Figure 9: Impedance modulus curve for the rebar (a) and prestressed steel (b) in chloride-contaminated concrete.
Advances in Materials Science and Engineering prestressed steel were found to be higher when compared to the rebar. e corrosion rate of the rebar was 0.721, 1.167, 1.524 × 10 −3 mmpy in 1, 2, and 3% chloride levels, respectively. e corrosion rate of the prestressed steel was 0.771, 1.374, 1.586 × 10 −3 mmpy in 1, 2, and 3% chloride levels, respectively. e stressed steel has 1.04, 1.18, and 1.04 times higher corrosion rate than the rebar at 1, 2, and 3% chloride levels, respectively. From the potentiodynamic polarization curve, it is observed that the stressed steel has corroded severely than the rebar. From the abovementioned electrochemical studies, it is found that the shift in the corrosion rate was higher at 2% chloride level when compared to 1% and 3% chloride levels. is may be due to the fact that initially during the hydration, at 1% chloride levels, the chlorides combined with the cement to form a complex called the Friedel salt [47,48], and it reduced the availability of free chloride ions. At 2% chloride, the complex formation was found to be lower, and the free chloride ions may be higher to cause more corrosion both in the case of the rebar and prestressed steel. Figure 11 shows the load versus displacement behavior of 0, 1, 2, and 3% chloride-admixed prestressed concrete beams tested under two-point loading. Table 5 shows the initial and nal crack load and the modulus of rupture for the various chloride-admixed concrete beams after 30 days of exposure. From Figure 10, it is found that the cracks appeared in the midspan. An additional increase in load under various stages resulted in the formation of additional cracks and widening of some earlier formed cracks. e cracks from the tension zone traversed up to the neutral axis, and cracking of the concrete took place at later stages. e failure of the specimen occurred due to the fast propagation of the cracks. From the load versus displacement curve, it is found that the maximum ultimate load for 0, 1, 2, and 3% of the chloridecontaminated concretes is 120.6, 98.5, 90.6, and 80.9 kN, respectively. From the results, it is observed that when compared to the control concrete, the chloridecontaminated concrete has reduced the load-carrying capacity of 18.3, 24.9, and 32.9% for 1, 2, and 3% chlorideadmixed concrete. e modulus of rupture was also found to

Conclusions
(1) e exural strength of the prestressed concrete beam was decreased by 32.9% as the chloride content was increased to 3%. (2) e corrosion rate of the prestressed steel is found to be higher than that of the rebar. (3) e present investigation further demonstrates that the ECN measurement combined with other electrochemical techniques is used to study the corrosion process of the prestressing steel. e ECN results are in good agreement with those of the OCP and EIS measurements. (4) e ECN clearly indicated the potential and current shift with chloride contents. e potential shifts observed were found to be −190 mV for 0% Cl − and −510 mV for 3% Cl − in the stressed condition. ese data showed 2.68 times shift for chloridecontaminated prestressed concrete beam. (5) From the abovementioned experiments, it is found that the ECN clearly indicated the corrosion status of the chloride-contaminated and -uncontaminated prestressed steel and rebar.

Conflicts of Interest
e authors declare that there are no con icts of interest in publishing this paper. Control 1% NaCl 2% NaCl 3% NaCl Figure 11: Load versus displacement behavior of corroded prestressed concrete beams after 30 days of exposure under various chloride levels.