Galvanic corrosion and cathodic protection of re-grouted, post-tensioned (PTd) concrete systems

. Grouted post-tensioned (PTd) concrete systems are widely used in long-span segmental bridges with a target service life of 100+ years. However, the usage of inadequate grout materials and grouting practices have resulted in the formation of unwanted air voids in the duct, which in turn led to premature corrosion (say, within about 20 years) of strands and failure of tendons. Also, the re-grouting/repairing of void regions have led to localized corrosion of strands at the interface between the dissimilar base-grout (usually carbonated) and repair-grout. This study aims (i) to quantify the galvanic corrosion at the void region in a PTd system re-grouted with a dissimilar grout and (ii) to develop cathodic protection system to protect PTd anchorage regions. Specimens simulating the re-grouted strand-grout-air (SGA) interface were prepared with prestressing steel wires and cementitious grout. The macro-cell current (galvanic current) between the prestressing steels embedded in carbonated base-grout and repair-grout indicated that galvanic corrosion can be possible at the SGA interface – reducing the long-term structural reliability of re-grouted PTd bridges. In addition, the feasibility of galvanic anode cathodic protection system to protect PTd anchorage regions was assessed. For this, a proof-of-concept study was conducted to validate that a thin layer of grout around the strand will be sufficient for a galvanic anode (connected to the end of the strand at outside the tendon anchorage) to protect the strand portions inside the duct/anchorage.


Introduction
Grouted post-tensioned (PTd) concrete systems are widely used in long-span segmental bridges with a target service life of 100+ years. However, premature strand corrosion within about 10 to 20 years has been observed in many such PTd concrete systems due to inadequate grouting materials and practices [1][2][3][4][5]. Figure 1 shows the photo and schematic of the anchorage region from a PTd system with exposed prestressed strands due to inadequate grouting materials and practices. Figure 2 shows the schematic of the side elevation of such inadequately grouted PTd anchorage region with the void and strand-grout-air (SGA) interface. These voids expose the prestressing strands and the SGA interface to atmospheric humidity, CO2 and chlorides.  In a typical internal PTd concrete system, the prestressing strands are embedded inside the duct, which is then embedded inside the concrete. Hence, the condition of strands is not visible from the outside. Therefore, visual inspections cannot be used as a tool to assess the condition of strands in the anchorage region. If left unnoticed, the corrosion of strands might lead to a reduction in the load-carrying capacity and/or catastrophic failures. The major challenge in repairing such inadequately grouted PTd system is the difficulty in accessing the strands which are embedded inside the concrete. The repair work must be done from outside the anchorage. The objective of this work is to understand corrosion of strands in anchorage regions and then propose a feasible electrochemical repair method. The remaining paper is organized as follows: first, a review of possible techniques to repair PTd systems and their associated challenges are presented. Then, an experimental programme assessing the severity of galvanic corrosion due to the electrochemical incompatibility between the base-grout and repair-grout in a re-grouted PTd system is presented. Finally, the feasibility of galvanic anode cathodic protection system in protecting the PTd anchorage region is discussed.

Possible repair techniques and their challenges
Following are some of the possible techniques to control corrosion of prestressing strands in PTd systems, which can be implemented from outside the anchorage region. Re-greasing is a method by which grease can be used to protect the exposed anchorages by the formation of a barrier to water and other corrosive contaminants. However, the removal of old grease is difficult, and if not properly cleaned, the old grease will cover the existing rust and can allow the pits to grow deeper [6], if sufficient moisture and oxygen are available.
In chemical impregnation, corrosion-inhibiting chemicals (hydrocarbon and silicon-based) can be impregnated by pressure through the interstitial spaces between the wires of each strand. These chemicals form a thin film around the steel surface; improve the corrosion resistance of the existing grout and inhibit the corrosive environment. However, it is difficult to fill all the voids [7].
Neutral-pH rust remover solutions can be pumped through the interstitial spaces in the strands. By the process of selective chelation, the solution reacts with the rust surface and eventually gets detached from the metal beneath. This method may not provide much protection for severely corroded steel and has to be combined with other techniques (say, chemical impregnation) [6].
In cable drying/de-humidification, inert noncorrosive gas like nitrogen is passed into the ducts under pressure. The inert gases will displace the oxygen around the tendons; dry the wet grouts (say, RH < 40%); and inhibit the corrosive environment. However, it is difficult to maintain the required pressure all the time. Also, this method may not be feasible for highly impervious grouts [6,8].
Re-grouting the voids with cementitious grouts is one of the feasible methods to repair a PTd anchorage system. Much research has been done on the initial grouting of PT ducts, but only limited research has been done on the repair-grouting of the PT ducts. The voids are generally filled/re-grouted with repair-grouts without addressing/repairing the carbonated SGA interface, as shown in Figure 2. In such situations, an electrochemical incompatibility can arise due to variations in the physical and chemical properties of the base-grout and repair-grout, resulting in the formation of a corrosion cell at the re-grouted SGA interface. The case study on the corrosion failure of a PTd tendon in a bridge within four years after re-grouting emphasize the severity of galvanic corrosion [9]. Due to the concerns of possible galvanic corrosion after re-grouting, supplemental methods to control corrosion in re-grouted PTd systems are required.
The idea of cathodic protection (CP) of prestressed concrete bridges has been adopted from the CP of prestressed concrete cylinder pipes used for water and sewer transmission services [10]. CP is commonly applied in prestressed concrete bridges primarily to protect mild steel reinforcements from corrosion. The application of CP to the prestressed strand is still developing due to the following concern.
The cold-drawn prestressed steel can be susceptible to hydrogen embrittlement when they are cathodically over-protected. In other words, when the prestressed steel is polarized to potentials equal to or more negative than the hydrogen evolution potential, then the metal can be susceptible to hydrogen embrittlement. Several authors have studied this mechanism using slow strain rate experiments and recommended the threshold potential for CP should be ₋ 900 mV (SCE) to prevent hydrogen embrittlement [11][12][13]. However, CP using galvanic anodes is considered safe from the standpoint of hydrogen embrittlement due to the limited polarization and ohmic losses in concrete structures [13].

Galvanic corrosion in re-grouted post-tensioned concrete systems
This study attempted to quantify the galvanic corrosion between the carbonated base-grout and repair-grout in a re-grouted PTd system. For this, specimens simulating the re-grouted SGA interface were prepared with prestressing steel wires and site-batched grout (w/c = 0.45 and plasticized expansive admixture dosage of 0.45 percent by weight of binder). The specimen preparation involves casting the bottom portion simulating the carbonated base-grout and then the top portion simulating the repair-grout, which is explained next.
Prestressing steel of 5.2 mm diameter and 10 mm length was extracted from the central king-wire of a 15.2 mm diameter prestressing strand. The steel specimens were drilled and tapped at one end to enable an electrical connection for the testing. The specimens were ultrasonically cleaned with distilled water for 15 minutes and then wiped with a cotton cloth soaked with ethanol. A stainless-steel rod of 3 mm diameter was fastened to the prestressing steel at one end. The junction between the prestressing steel and the stainless steel rod was coated with epoxy to eliminate the presence of humidity to initiate galvanic corrosion between the prestressing steel and stainless steel. Figure 3 shows the schematic of the bottom portion of the specimen. The specimens were cast and demoulded after 24 hours and were kept in a 65% RH and 25 °C environment for seven days to enable the passivation of the prestressing steel. After that, the specimens were allowed to carbonate in a 3% CO2 environment (± 25 °C and 65 % RH) in a carbonation chamber until the depassivation of the embedded prestressing steel wire. Depassivation behaviour of prestressing steel was assessed using electrochemical impedance spectroscopy (EIS) tests on the 3-electrode corrosion cell with the prestressing steel as the working electrode (WE), a Nickel-Chromium mesh as the counter electrode and a saturated calomel electrode as the reference electrode ( Figure 3). The testing was carried out with an AC perturbation signal of ±10 mV amplitude applied over a frequency range of 10 5 to 0.01 Hz at open circuit potential (OCP) with 10 points per decade. Figure 4 shows the Nyquist and Bode responses of one of the tested specimens. Incomplete and overlapping arcs of three semicircles are observed. As shown in Figure 4(a), an equivalent circuit with three Resistor-Constant Phase (R-Q) elements in series was used to model the response from the specimen. The first semicircle corresponds to the solution (say, water + grout; Rs and Qs). The second semicircle corresponds to the double layer (Rdl and Qdl). The third semicircle corresponds to the oxide layer (passive layer) of the prestressing steel (Rox and Qox). A schematic of an ideal representation of the EIS response is shown in the inset of Figure 4(a) for clarity and easy understanding.
Depassivation of the specimen was assessed based on the comparison of the oxide layer characteristic (Rox). The slope of the arc in the low-frequency region (10 ˗1 to 10 ˗2 Hz) in the Nyquist representation qualitatively indicates the condition of passive layer. A larger slope represents a larger diameter semicircle, indicating a larger Rox. The EIS testing was conducted on Day 30 and Day 45 of carbonation exposure. A reduction in the slope of the arc in the low-frequency region was observed in the Nyquist representation at Day 45. The reduction in slope indicates a decreased Rox, corresponding to a depassivated system. This could be evident from Bode frequency and magnitude responses, as shown in Figures 4b and c. A gradual reduction in the magnitude and phase angle corresponding to the low-frequency region (0.01 Hz) indicates a change in the passive layer. The |Z|0.01 reduced from 95 kΩ on Day 30 to 60 kΩ on Day 45. The φ0.01 also reduced from 66 to 62 degrees -ascertaining that the specimen was de-passivated.
After that, the specimens were removed from the carbonation chamber and the repair-grout was cast over the base-grout. It must be noted that a gap of 1 mm was maintained between the circular end faces of two prestressing steel wire pieces. After one day, the specimen was demoulded, and the two prestressing steel wire pieces were connected externally, as shown in Figure 5. The specimens were maintained at 95% RH and 25 °C until steady currents were observed. A pico-ammeter was used to measure the current between the two prestressing steel wire pieces within the specimen.  An average galvanic current density of 1.5 µA/cm 2 was measured between the base-grout and the repair-grout. The direction of the galvanic current indicates that the prestressing steel in the base-grout was the anode and the prestressing steel in the repair-grout was the cathode. Then, the theoretical mass loss of the prestressing steel was calculated using Faraday's law of electrolysis and the measured galvanic current. The analysis estimated a 5% reduction in mass of the prestressing wire within 20 years of service, emphasising the severity of such galvanic corrosion. Hence, this study recommends the prohibition of re-grouting of voids without treating the carbonated SGA interface.
This study also proposes the re-alkalization of the carbonated SGA interface with alkaline solutions before re-grouting. However, it could be difficult to achieve complete re-grouting of voids; and hence, galvanic anodes can be used as a supplemental method to protect the strands from corrosion which is explained next.

Cathodic protection of strands in the anchorage regions
This study proposes the use of galvanic anode cathodic protection systems to protect the anchorage region (portions of strands lying inside and outside the bearing plate) of a PTd concrete system. Figure 6 shows the schematic of the proposed corrosion protection system with a galvanic anode attached to the end of the strand outside the bearing plate. It is important to understand that every metallic system protected by a galvanic anode needs the following two components: a path for the transfer of electrons and a path for the transfer of ions. The strands, wedges, and bearing plate are all in contact with each other, making them all electrically connected. Hence, it can be well understood that the galvanic anode can protect the metallic components outside the bearing plate due to the availability of an ionic path. However, there arises a question on the protection of strands inside the bearing plate by the galvanic anode, and this study attempted to answer this question by a proof-of-concept experiment.

Fig. 6. Schematic of a cathodically protected PTd anchorage region
A proof-of-concept experiment was conducted to validate the feasibility of galvanic anode to protect the strands lying inside the bearing plate. Figure 7 shows the schematic of the experimental setup simulating the inside and outside portions of a PTd anchorage region. A prestressing wire coated with a thin layer of grout (~ 1 to 2 mm) was positioned between two containers, A and B. The containers were filled with simulated concrete pore solution mixed with 3.5 % chlorides. A discrete zinc-based galvanic anode was immersed in container B and was electrically connected to the prestressing wire through a switch arrangement. Initially, the switch was kept in the OFF position. One end of the prestressing steel was connected to a potentiostat, and the open circuit potential (OCP) of the prestressing steel in both containers was measured. To achieve this, a saturated calomel reference electrode was kept inside container A, and the OCP was measured as -250 mV. Similarly, the saturated calomel reference electrode was kept inside container B, and the OCP was measured as -255 mV. Then, the experiment was continued with the reference electrode placed inside Container A. At this point, the switch was turned ON, and a sudden increase in the OCP towards the negative side was observed, as shown in Figure 8. It can be inferred that as soon as the switch was turned ON, the system became a coupled system (prestressing steel + galvanic anode), and the potentiostat started recording the mixed potential. From this, it is evident that a thin layer of grout over the strand is enough to conduct ions from the outside to the inside of the bearing plate. Hence, galvanic anode cathodic protection is an electrochemically feasible repair solution to protect inadequately grouted PTd anchorages. The use of inadequate grouting materials and practices has resulted in the formation of voids and premature corrosion in PTd concrete systems. A review of the possible techniques to repair such PTd systems and their challenges was presented. This study quantified the galvanic corrosion strands at the interface of the carbonated base-grout and repair-grout in a re-grouted PTd system. A galvanic current density of 1.5 µA/cm 2 was determined at 95% RH and 25 °C. The analysis revealed that such galvanic current density could result in about 5% reduction in mass of the prestressing wire within about 20 years of exposure/service, emphasising the severity of such galvanic corrosion while considering the century-long protection needed. Hence, this study recommends the re-alkalization of carbonated Strand-Grout-Air (SGA) interface followed by re-grouting of voids. However, it could be difficult to achieve complete re-grouting of voids; and hence, galvanic anodes can be used as a supplemental method to protect the strands from further corrosion. A proof-ofconcept experiment proved that a thin layer of grout around the strand would be sufficient for a galvanic anode (connected to the strand-end outside the tendon anchorage) to protect the strand portions inside the duct/anchorage, and hence extend the service life of PTd concrete systems.