Kingsley Lau, Florida International University, Miami, Florida
Ivan Lasa and Mario Paredes,
FDOT State Materials Office, Gainesville, Florida

The assessment of recently failed external post-tensioned tendons from a Florida bridge provided insight on the conditions that led to the failure. The corrosion was attributed to the presence of segregated grout, characterized by high moisture content and enhanced sulfate concentrations. Accelerated corrosion was caused by macrocell coupling of local anodes in the strand and extended cathodes throughout the tendon. Differential aeration, vastly different moisture contents, and crevices caused corrosion failure within a short time frame.

Lau-1

Location of external PT tendon failure

Corrosion failure of longitudinal external post-tensioned (PT) tendons was identified in 2011 in a Florida bridge that had been in service for approximately eight years. The bridge was a PT segmental bridge with internal and external tendons. The external PT tendons were sheltered in the bridge segment (Figure 1). The corrosion location on the two failed tendons was in the upper portion of the sloped region of the tendon immediately adjacent to the upper deviator. The external PT tendons were nominally 102-mm in internal diameter with 22 16-mm diameter seven-wire strands. Substantial grout segregation with severe corrosion was observed in numerous other tendons. It was apparent that the corrosion condition was not isolated and major repair operations at several tendon locations were in progress. Corrosion-related failures of PT strand have been documented in several Florida bridges. The alkaline (chloride-free) cementitious environment in grout promotes steel passivation that typically prevents any significant corrosion; however, the presence of voids in the grout of tendon ducts would provide conditions whereby corrosion may occur. The causes of voids have been attributed to bleed water during grouting and subsequent reabsorption or evaporation from the grout. Because of corrosion development in the field associated with accumulated bleed water, new PT grout material specifications required low bleed water formation. Florida specifications for PT grout at the time the bridge being studied was constructed required 0% bleeding at 3 h determined by a modified wick-induced bleed test (ASTM C9407). Recent field observations (from a bridge construction project in Texas) have shown indications of grout segregation and stratification leaving part of the grout material in a wet plastic state. The objective of this investigation was to resolve the environmental and material conditions and mechanistic issues on the corrosion failure of the bridge. The assessment of the failed tendons included detailed examination of extracted segments along the length of the tendon. Grout material from the anchors and deviator were tested as well. Laboratory work on the tendons included visual inspection of the ducts, assessment of the grout material, and coverage and corrosion characteristics of the strands.

Experimental Methodology

Lau-2

Schematic of corrosion macrocell test sample

Tendon Section Sampling : For tendon sections selected for study, a 200- to 250-mm long portion of the high-density polyethylene (HDPE) duct, 150 to 200 mm away from the cut edge of the section, was removed to expose the grout and strand material.
Visual examinations for corrosion development, grout placement, grout consistency, duct condition, and presence of voids were made.
Grout Analysis: Bulk petrographic analysis of the grout was done to identify deficiencies in the grout material.
Corrosion Analysis: During the tendon autopsy, corrosion assessment was made by visually observing the corrosion development of the embedded strands, measuring the open circuit potential (OCP) of individual strands when possible by placing a copper/copper sulfate (Cu/CuSO4) reference electrode (CSE) on exposed grout; and measuring the corrosion rate by linear polarization resistance (LPR) and electrochemical impedance spectroscopy (EIS) on test cells constructed from select tendon locations. The corrosion test cells were made from short tendon segments cut 76 to 102 mm in length from an actual tendon (Figure 2). Corrosion rate and OCPs were monitored for up to one week after cement paste casting.

Results and Discussion

Lau-3

Figure 3: Strand corrosion at failure location

Field Corrosion Assessment: The location of the tendon failures exhibited severe corrosion (Figure 3) concentrated mostly on strands in the upper portion of the tendon cross section. Except for the heavily corroded strands there, failure of the remaining strands typically appeared to be ductile as characterized by necking of the wire cross section. Significant surface rust accumulation was also observed on the strands from an autopsied segment of the failed tendon near the failure location in the high point horizontal region adjacent to the pier, ~7 ft (2 m) from the failure location. Severe corrosion was also identified at low elevation anchor caps.

Tendon Section Autopsy Results

Grout-Tendon Characterization: Grout fill deficiency was seen in the first tendon. The void space along the tendon length and elevation was not consistent with the grout fill deficiency being attributed to drainage of the grout when the mix was still fluid. Rather, the grout in the upper horizontal region appeared to have subsided possibly due to segregation of grout material in the sloped region. It was apparent by the presence of a groove or bleed trail and remnants of bubbles at the top portion of the duct that a large amount of entrapped air and water was transported along the length of the tendon. A second failed tendon did not contain significant voids, but significant portions of the tendon contained segregated grout.

Lau-4

Figure 4: Wet segregated plastic grout. The pink color shows the phenolphthalein indicator. The upper portion of the picture shows a screwdriver embedded in soft grout. The lower right picture shows the surface condition prior to autopsy

Grout Consistency: The grout in the bottom portion of the tendon sections from the upper horizontal region was well-consolidated and hardened. The grout from some sections from the upper sloped region of the tendon was moist and soft throughout. Grout from several locations in these sections appeared to have large pores that were filled with ettringite crystals. This material was readily visible in greatly segregated grout. The recovered segregated grout from the first tendon was typically soft, white, and chalky. But the material there did appear to have set. An insufficient amount of wet plastic grout observed in the field was available for testing. In contrast, it was apparent that there were significant amounts of wet plastic segregated grout in the second tendon in both the depth and length of the tendon adjacent to both sides of the high deviator where tendon separation was initially observed. Figure 4 shows the typical appearance of the wet segregated plastic grout.
Grout Laboratory Analysis: The water content of hardened grout was typically ~20% by mass. It was apparent that there was high water content in the segregated grout material that was in the upper portion of the tendon cross section. The water content exceeded 60% by mass in grout characterized as having a wet and plastic consistency and was as high as 50% in grout characterized as soft and chalky. The total chloride content in the segregated grout was low to moderate but there was an indication of enhanced chloride accumulation in the wet segregated grout.
The range and general trends of ionic concentrations of Na+, K+, Ca2+, and OH– in grout locations from the vicinity of deficient grout were similar to those of cement paste specimens ranging from 0 to 30% silica fume cement replacement as measured by Page and Vennesland.10 The apparent lower concentrations of Na+, K+, and OH– and the higher Ca2+ concentration in the deficient grout than hardened grout were consistent with the pore water makeup of cement with higher silica fume cement replacement. There was, however, indication of enhanced Na+ and K+ concentrations in leachate from the greatly segregated grout containing high moisture content.
A significant amount of sulfur was detected in segregated grout in both tendons, but the aqueous sulfate ion concentration was relatively high throughout the tendon. Enhanced sulfate content was observed in the greatly segregated grout. The high concentration of sulfate ions is expected to facilitate the depassivation of steel strands. The sulfate ion concentrations here were significantly higher compared to those in a pore water analysis of silica fume cement paste conducted by Page and Vennesland. Page and Vennesland measured SO42– as high as 3,800 ppm in cement paste, whereas as much as 9,700 ppm SO42– was measured for grout leachate from the bridge. The mechanism for the accumulation of sulfate ions could not be determined.Pore water pH for select locations from the tendons was generally greater than 13 and did not give any indication of possible mechanisms that may cause a decrease in pH, such as carbonation.

Strand Corrosion

Open Circuit Potential: The cumulative fraction of OCP for individual strands within tendon sections are shown in Figure 5. Severe active corrosion was apparent in some tendon segments and, as expected, distinctly negative potentials were measured there. In these sections with active corrosion, a wide range of potentials was measured for strands in the various forms of segregated grout, which signifies differences in corrosion activity even within a small area. The differences were attributed to the amount of moisture, concentration of chloride and sulfate ions, and oxygen present in the grout.

Lau-5

Figure 5: OPC of individual strands. The legend lists the tendon section numbers

Macrocell Corrosion: The tendency for corrosion macrocell activity in tendons with segregated grout was tested by coupling tendon sections with wet plastic grout and tendon sections with hardened grout. Recovered tendon segments from the second tendon, which contained these modes of grout deficiencies, were used to construct test samples as described earlier. The top tendon sections contained hardened grout and the bottom sections contained deficient grout. Electrochemical testing of these sections showed that the deficient grout had a much lower electrical resistivity than the hardened grout, which is consistent with the high moisture and high porosity of the segregated grout. The tests also showed that the steel in the segregated grout did have active corrosion.
The macrocell current after ~1 day was ~60 to 120 μA but decreased to ~40 μA after ~3 months (Figure 6). The decrease in macrocell activity was thought to be in part due to continual hydration of the introduced cement paste. The macrocell at the time of coupling was generally consistent with the potential difference between the top and bottom section and the measured solution resistances.

Durability Implications

Lau-6

Figure 6: Macrocell current after galvanic coupling of strands in hardened and segregated grout. Positive values represent the bottom sample section acting as a net anode

The large localized area of significantly segregated grout in the tendon at the failure locations is of utmost concern because of its detrimental impact on the tendon’s corrosion resistance. The area of significant grout segregation appeared to create a localized environment with a large amount of free moisture. Although relatively high alkalinity was maintained in the segregated grout, the corrosion mechanism proposed by Bertolini and Carsana and elsewhere,11-12 where active corrosion was maintained from high pH and large cathodic polarization, may not have been the primary cause here. It was apparent that significant silica fume sedimentation and accumulation in the segregated grout may have slightly reduced the pH.The assessment of the failed tendon from the Florida bridge provided some insight on the material and environmental conditions of the tendon components; no direct conclusions can be made at this time as to whether the failure was due to poor construction practices or poor grout material.

Conclusions

• Segregated grout material was observed in localized portions of the tendons. Grout segregation was not apparently caused solely by gravimetric sedimentation. The segregated grout was characterized in three modalities: wet plastic, white chalky, and sedimented grout.
• High water content (up to 70 wt%) was measured in grout from localized portions of the tendons where corrosion occurred. No significant drop in pore water pH was observed in a limited amount of sampling. Slight depression of pore water pH (>11) from the wet plastic grout was apparent, however.
• Chloride accumulation was apparent in grout with high water content, but concentrations (<1 lb/yd3) were lower than conventional critical chloride threshold concentrations. Sulfate concentrations in the grout pore water were as high as 9,700 ppm.
• Corrosion in segregated grout (signified by high moisture content, high porosity, and high sulfate concentrations) was likely accelerated because of macrocell coupling with an extended cathode throughout the tendon.

Acknowledgments

The work and assistance by Dennis Baldi, Will Blanchard, Matt Brosman, Jason Burchfield, Pat Carlton, Charles Ishee, Marc Knapp, Awilda Merced, Richard Nalli, Juan Rafols, Nikita Reed, and Yongyang Tang are acknowledged here. Cooperative work with Michael Ahern,Richard Lewis, Randy Poston, and Phillip Sharff and assistance by Marcus Lee and John Newton is also acknowledged. The opinions and findings are those of the authors and not necessarily those of the supporting agencies.

References

1 R.G. Powers, “Corrosion Evaluation of Post-Tensioned Tendons on the Niles Channel Bridge,” Florida Department of Transportation, 1999.
2 J. Corven,“Mid Bay Bridge Post
Tensioning Evaluation,” Florida
Department of Transportation, 2001.
3 “Sunshine Skyway Bridge Post-
Tensioned Tendons Investigation,” Florida Department of Transportation, 2002.
4 A. Sagües, “Corrosion of Post-Tensioning Strands,” Florida Department of Transportation, 2005.
5 D. Trejo, et al., “Effect of Voids in Grouted, Post-Tensioned Concrete Bridge Construction: Vol. 1— Electrochemical Testing and Reliability Assessment,” Texas Department of Transportation, FHWA/TX-09/04588-1, Vol. 1, September 2009.
6 H. Minh, et al., “Influence of Grouting Condition on Crack and Load-Carrying Capacity of Post-Tensioned Concrete Beam due to Chloride-Induced
Corrosion,” Construction and Building Materials 21, 1568-75, 2007.
7 ASTM C940-10a, “Standard Test Method for Expansion and Bleeding of Freshly Mixed Grouts for Preplaced-
Aggregate Concrete in the Laboratory” (West Conshohocken, PA: ASTM).
8 “Standard Specifications for Road and Bridge Construction,” Section 938: Posttensioning Grout, Florida Department of Transportation, 2013.
9 B. Merrill, “Memo, Grout Testing and Analysis,” Texas Department of
Transportation, September 14, 2010.
10 C.L. Page, O. Vennesland, “Pore
Solution Composition and Chloride Binding Capacity of Silica-Fume
Cement Paste,” Matériaux et Constructions 16, 1 (1983): pp. 19-25.
11 L. Bertolini, M. Carsana, “High pH Corrosion of Prestressing Steel in
Segregated Grout,” Modeling of
Corroding Concrete Structures, RILEM Bookseries, Vol. 5 (Dordrecht, The Netherlands: Springer, 2011).
12 E. Blactot, C. Brunet-Vogel, F. Farcas, L. Gailet, I. Mabille, T. Chaussadent,
E. Sutter, “Electrochemical Behavior and Corrosion Sensitivity of Prestressed Steel in Cement Grout,” WIT Transactions on Engineering Sciences, Simulation of Electrochemical Processes II (Southampton, U.K.: WIT Press, 2007): pp. 267-76.

This article is based on CORROSION 2013 paper no. 2600, presented in Orlando, Florida.

KINGSLEY LAU is an assistant professor at Florida International University, Civil and Environmental Engineering Department., 10555 W. Flagler St., Miami, FL 33174. He formerly worked as a corrosion research scientist at the Florida Department of Transportation State Materials Office. He has a Ph.D. in civil engineering and has been a member of NACE International since 2004.
IVAN R. LASA is a state corrosion mitigation technologist at the Florida Department of Transportation, 5008 NE 39th Ave., Gainesville, FL 32609. He has degrees in civil engineering and surveying with undergraduate studies in electrical engineering. He has more than 25 years of experience in the field of corrosion, specializing in corrosion control of reinforcing steel in concrete and rehabilitation of marine structures. He has been a member of NACE since 1994.
MARIO A. PAREDES is a state corrosion engineer at the Florida Department of Transportation. He is a materials engineer with 18 years of corrosion and durability experience in the transportation infrastructure sector.

original source of publication: Materials Performance
copyright holder: NACE International