U.S. Department of Transportation
Federal Highway Administration
1200 New Jersey Avenue, SE
Washington, DC 20590
202-366-4000
Federal Highway Administration Research and Technology
Coordinating, Developing, and Delivering Highway Transportation Innovations
This report is an archived publication and may contain dated technical, contact, and link information |
|
Publication Number: FHWA-RD-02-086 Date: August 2006 |
Previous | Table of Contents | Next
This site consists of two sections on twin bridges over Raymond Run in Columbus, OH. The site is located 0.8 kilometers (km) (0.5 miles (mi)) south of Roberts Road overpass over I–270. Overall views of both spans are shown in figures 2 and 3. The twin bridges are single span, 6.1 meters (m) (20 feet (ft)) long, and simply supported at the abutments. The deck is 380 millimeters (mm) (15 inches) thick.
Figure 2. Photograph showing test section on I–270 northbound (1995).
Figure 3. Photograph showing test section on I–270 southbound (1995).
The decks were originally constructed in 1969 and were covered with a 6-mm (0.23-inch) thick latex emulsified asphaltic concrete overlay. In 1991, an inspection by state forces indicated the deck to be in generally satisfactory condition with delaminations and patches over 10 percent of the deck surface. There were obvious transverse cracks in the decks. The decks were included for overlayment as part of a larger rehabilitation project (Ohio Department of Transportation (ODOT) 8753-91) that covered 4.3 km (6.9 mi) of I–270. The project included repaving and widening of the bridge structures. The northbound lanes were selected for overlay with silica fume concrete and the southbound lanes with LMC-III concrete. The I–270 site was inspected on the following dates:
Northbound (NB) Lanes—Project records indicate that the materials used for the SFC overlay mix consisted of: a Type I cement from a local supplier; a sub-rounded to rounded natural sand which was a mixture of siliceous and calcareous minerals having a specific gravity of 2.62, absorption of 1.54 percent, and a fine modulus (FM) of 3.36; and a crushed limestone coarse aggregate having a maximum top size of 12 mm (0.5 inch), a specific gravity of 2.62, and an absorption of 1.96 percent. The admixtures used were: Sika-Crete silica fume (slurry form); Sika AER, a neutralized Vinsol resin air-entraining agent; Plastocrete 161 MR, a Type D water reducer/retarder; and Sikament 300, a Type F high range water reducer. The mixture proportions reported for both overlays are shown in table 1. The specification on the fresh properties of the silica fume mix include a slump of 100 to 200 mm (4 to 8 inches) and an air content range of 6 to 10 percent.
Material (lb/y3) | LMC-III | SFC |
---|---|---|
Cement | Type III—658 | Type I—700 |
Fine Aggregate | 1591 | 1470 |
Coarse Aggregate | 1306 | 1292 |
Water | 321 | 223 |
Silica Fume | - | 70 |
Latex Modifier (gallon) | 25 | - |
Water Reducer (ounce) | - | 25 |
High-range water reducer (HRWR) (ounce) | - | 218 |
Air entraining agent (AEA) (ounce) | - | 5 |
Water-to-cement (W/C) ratio | 0.48 | 0.29 |
1 ounce = 29.57 mL
1 gallon =3.785 L
1 lb/y3 = 0.347 kg/m3
Southbound (SB) Lanes—The materials used in LMC-III overlay mix included: a Type III cement; the same aggregates as the SFC mix; and a Latex, Styrofan 1186, which is a styrenebutadiene latex emulsion. The specifications for the fresh concrete were a slump range of 100 to 150 mm (4 to 6 inches) and an air content not to exceed 7 percent.
The asphalt overlay was removed and the surface scarified in preparation for placement of the overlays. Both overlays were placed on May 27, 1992. The SFC was placed first on the passing lane of the northbound deck. Two 3.82-m3 (5-y3) batches were delivered. The deck was wetted 1 hour prior to placement of the overlay. The placement went smoothly with no reported problems. The surface was textured with steel tines transversely on 16 mm (0.6 inch) centers, then covered with wet burlap and polyethylene sheeting and continuously soaked for a period of 48 hours. The LMC-III overlay was placed the afternoon of the same day on the passing lane of the southbound deck. Placement appeared to go smoothly, however the state inspectors questioned the calibration of the mobile mixer during the pour. Improper calibration led to higher amounts of water being used than desired. Calculations showed the water to cement ratio was actually 0.48 in the mix. The same texture was applied to this overlay and it was then moist cured for 48 hours under soaked burlap. Both sections were opened to traffic on June 19, 1992.
The undersides of the two bridges were inspected annually as traffic control was being set up. The structures have no girders or thickened beam portions, but have flat undersides. Some fine cracking was observed on the underside but no efflorescence staining was seen on the underside of either span. No corrosion staining occurred at the crack locations. Some leakage was apparent at the joint between the bridge decks and the abutments. The drains in the abutment headwalls were functioning, with water flowing. A cold joint could be seen on the deck bottom and in the abutment where the bridge had been widened. The only other feature of note was rust staining at the exposed bar chair tips on the underside of the deck.
The test sections studied on both spans include the shoulder and travel (outside) lanes. Examination of the deck overlay surface indicated several narrow longitudinal cracks in the shoulder lanes of both bridges. Cracks were more frequent in the travel lanes. The travel lane of the northbound SFC bridge had mostly single nonconnected cracks running in both the longitudinal and transverse directions with very little branching. The travel lane of the southbound LMC bridge had more and longer longitudinal cracks and also more branching of cracks.
Severe cracking and spalling was noted in the exposed travel surfaces of the abutments of the northbound SFC structure. The overlay at the approach joint was spalled away and had been patched with a soft tar-like material. The joint between the bridge and the abutment on the leave side of the bridge was irregular, with a separate one-foot wide closure pour placed between the deck and the abutment. The joint between the closure pour and the abutment was well-sealed, but the formed joint between the deck and the closure pour was open and packed with debris.
The chain drag and hammer survey technique was used to locate areas of delamination of the overlay from the original concrete deck. Delaminations were found mostly at crack locations. The chain drag was used to initially find the delaminations. They were then more clearly delineated using hammer tapping and loose sand. Table 2 shows the number and areas of delaminations located on the northbound SFC bridge. The area of the northbound test section is 37.2 m2 (400 ft2). The number of delamination areas has increased over the last five years from 0 in 1994 to 6 in 1998. All the delaminations are in the travel lane and are isolated from each other. The largest single delamination area in 1998 is 1.16 m2 (13 ft2). The total area of delaminations is 2.46 m2 (26.5 ft2) or 6.61 percent of the test area and is not considered a serious amount.
Year | Number of Areas |
Total Area (m2) |
Percent of Total Area |
---|---|---|---|
1994 | 0 | 0.00 | 0.00 |
1995 | 2 | 0.14 | 0.38 |
1996 | 3 | 0.93 | 2.50 |
1997 | 4 | 1.48 | 3.98 |
1998 | 6 | 2.46 | 6.61 |
Table 3 shows the number and areas of delaminations located on the southbound LMC bridge. The area of the southbound test section is 40.9 m2 (440 ft2). There are the same number of delaminations on this bridge (six) as on the northbound bridge but the total area is slightly more at 2.87 m2 (31.9 ft2) or 7.02 percent of the test area. The largest single area of delamination is 0.56 m2 (6.0 ft2) but there are two that size and three others slightly smaller. Again all the delaminations are in the travel lane. Some delaminated areas are connected by longitudinal cracks which may be a precursor to further delamination.
Year | Number of Areas |
Total Area (m2) | Percent of Total Area |
---|---|---|---|
1994 | 0 | 0.00 | 0.00 |
1995 | 2 | 0.46 | 1.12 |
1996 | 4 | 0.93 | 2.27 |
1997 | 6 | 1.25 | 3.06 |
1998 | 6 | 2.87 | 7.02 |
The half-cell potential survey was performed using a 0.6-m (2-ft) grid spacing over the entire test area (shoulder and outside travel lane) of each span in 1994 and 1997. The electrical connection to the top layer of reinforcement was made using a 50-mm (2-inch) diameter core hole drilled to the top layer of reinforcing steel. A hole was drilled in the reinforcing bar and a self-tapping screw was used to make the connection. The half-cell readings for the northbound lanes SFC in 1994 ranged from –179 mV to –471 mV, with large potential differences between nearby areas. These readings tend to indicate that some corrosion is taking place in the bridge. Table 4 shows the cumulative frequency of the half-cell potentials for the northbound SFC structure in 1994. Table 5 shows the cumulative frequency of the half-cell potentials measured in 1997. These numbers range from –133 mV to –535 mV. A histogram of the data for the I–270 northbound SFC structure for 1994 and 1997 is shown in figure 4. The potentials tend to be less negative in 1997 than in 1994, however, many areas remain very negative, indicating continued corrosion.
An equipotential map for the northbound SFC test section in 1997 is shown in figure 5. The most negative potentials were seen at the ends of the bridge and near the cracked delaminations. No area of more negative potentials was associated with the longitudinal construction joint. Approximately three-quarters of the readings were more negative than -250 mV, indicating that corrosion is likely occurring over most of the deck. The alignment of the areas of rapid potential changes with the edges of the bridge and the cracked delaminations indicate that the overlay is not providing sufficient protection from water and waterborne chlorides in these locations.
Potential Range (mV) |
Number of Observations |
Cumulative Percentage |
---|---|---|
0 to –149 | 0 | 0.0 |
–150 to –199 | 3 | 2.5 |
–200 to –249 | 14 | 14.2 |
–250 to –299 | 21 | 31.7 |
–300 to –349 | 35 | 60.8 |
–350 to –399 | 25 | 81.7 |
–400 to –449 | 16 | 95.0 |
–450 to –499 | 6 | 100.0 |
–500 to –549 | 0 | 100.0 |
Potential Range (mV) |
Number of Observations |
Cumulative Percentage |
---|---|---|
0 to –149 | 3 | 2.8 |
–150 to –199 | 7 | 9.3 |
–200 to –249 | 16 | 24.1 |
–250 to –299 | 23 | 45.4 |
–300 to –349 | 28 | 71.3 |
–350 to –399 | 15 | 85.2 |
–400 to –449 | 11 | 95.4 |
–450 to –499 | 3 | 98.2 |
–500 to –549 | 2 | 100.0 |
Figure 4. Histogram of half-cell potentials for I–270 NB SFC test section for 1994 and 1997.
Figure 5. Equipotential map of half-cell potentials for I–270 NB (silica fume-modified concrete).
The half-cell potentials for the southbound LMC span in 1994 range from –188 mV to –540 mV. A cumulative percentage summary of the data is shown in table 6. The data for 1997 is shown in table 7. For 1997 the data ranges from –167 mV to –449 mV. A histogram of the data for the southbound LMC test section for the years 1994 and 1997 is shown in figure 6. The potential generally shift less negative (less corrosion) in 1997 versus 1994.
Potential Range (mV) |
Number of Observations |
Cumulative Percentage |
---|---|---|
0 to –149 | 0 | 0.0 |
–150 to –199 | 1 | 0.8 |
–200 to –249 | 1 | 1.7 |
–250 to –299 | 8 | 8.3 |
–300 to –349 | 30 | 33.3 |
–350 to –399 | 38 | 65.0 |
–400 to –449 | 30 | 90.0 |
–450 to –499 | 10 | 98.3 |
–500 to –549 | 2 | 100.0 |
Potential Range (mV) |
Number of Observations |
Cumulative Percentage |
---|---|---|
0 to –149 | 0 | 0.0 |
–150 to –199 | 2 | 1.5 |
–200 to –249 | 15 | 13.1 |
–250 to –299 | 35 | 40.0 |
–300 to –349 | 40 | 70.7 |
–350 to –399 | 28 | 92.3 |
–400 to –449 | 10 | 100.0 |
–450 to –499 | 0 | 100.0 |
–500 to –549 | 0 | 100.0 |
Figure 6. Histogram of half-cell potential readings for the I–270 southbound LMC-III test section for 1994 and 1997.
An equipotential map of the half- cell data for the I–270 southbound test section in 1997 is shown in figure 7. The most negative potentials were seen in the areas of the construction joint, the ends of the bridge, and near cracked delaminations. The majority of the readings were more negative than –250 mV. Also, indicative of corrosion in the structure are the relatively large differences in potential across the area surveyed, as indicated by the closely spaced lines on the contour plot. The most negative potentials along the edges of the bridge, the construction joint, and the cracked delamination indicate that the overlay is not providing sufficient protection from water and waterborne chlorides in these locations.
Figure 7. Equipotential map of half-cell potentials for I–270 southbound (latex-modified concrete).
Past experience with research performed at the contractor site and published in FHWA report FHWA-RD-86-193, Protective Systems for New Prestressed and Substructure Concrete, indicates that corrosion potentials more negative than approximately –230 mV indicate that corrosion is taking place in bridge deck type structures. According to the –230 mV criteria, the majority of the locations surveyed had potentials indicating that corrosion was taking place. Both structures appear to be undergoing corrosion over the majority of their deck area. The ends of the bridges and the areas under cracked delaminations appear to be especially vulnerable. A cumulative frequency diagram of the potentials is shown in figure 8. The curve for both bridge decks is shifted to the left from 1994 to 1997. This indicates a decrease in the overall half-cell potentials. The shift is larger for the southbound lanes which have the latex-modified overlay.
Figure 8. Cumulative frequency diagram for I–270 (NB and SB) for years 1994 and 1997.
In 1994, polarization resistance (PR) tests were performed at six locations on the northbound SFC bridge. The test locations were chosen from one low-potential area and five high-potential areas. Simulated 3LP tests were also performed at each of the test locations. The PR testing was generally more successful than the 3LP testing, with five of the six locations being successfully tested. Only three of the six locations were successfully tested using the 3LP technique. The results are summarized in table 8.
The corrosion rate measurements were in general agreement with the half-cell potential measurements, with the most negative potentials associated with the highest corrosion rates. The location with the least-negative potential, point B8, could not be successfully measured using either the polarization resistance or the 3LP technique. This matches previous experience where low corrosion rate areas are difficult to measure using these techniques. The difficulties may also be due to the high electrical resistance of the SFC overlay. Silica fume is well known for its ability to increase the electrical resistance of concrete, one of the reasons for the very lowAASHTO T 277 "coulomb" values produced by silica fume concretes.
Grid Location |
Equilibrium Potential (mV)* |
PR Testing | 3LP Testing | ||
---|---|---|---|---|---|
Rp (ohm) |
Icorr (mA/cm2) |
Rp (ohm) |
Icorr (mA/cm2) |
||
H12 | –355 | 300 | 0.072 | - | - |
G12 | –400 | 647 | 0.034 | 188 | 0.116 |
B8 | –205 | - | - | - | - |
C6 | –451 | 121 | 0.179 | 149 | 0.145 |
G5 | –439 | 157 | 0.138 | - | - |
F3 | –395 | 320 | 0.068 | 331 | 0.066 |
* Measured using a Cu/CuSO4 reference electrode.
Polarization resistance tests were also performed at six locations on the southbound LMC bridge in 1994. The test locations were chosen from one medium-potential area and five high-potential areas. The PR and 3LP testing was generally more successful on the southbound LMC structure than on the northbound SFC structure, with all six locations being successfully tested using both techniques. The results of the testing are summarized in table 9.
Grid Location |
Equilibrium Potential (mV)* |
PR Testing | 3LP Testing | ||
---|---|---|---|---|---|
Rp (ohm) |
Icorr (mA/cm2) |
Rp(ohm) | Icorr (mA/cm2) |
||
G11 | –497 | 337 | 0.058 | 991 | 0.021 |
C11 | –332 | 261 | 0.083 | 711 | 0.031 |
A9 | –336 | 193 | 0.109 | 469 | 0.046 |
C7 | –411 | 346 | 0.063 | 1081 | 0.020 |
B5 | –269 | 456 | 0.048 | 1123 | 0.019 |
F4 | –536 | 252 | 0.086 | 322 | 0.068 |
* Measured using a copper/copper sulfate (Cu/CuSO4) reference electrode.
On the southbound LMC structure, the corrosion rate measurements were not in general agreement with the half-cell potential measurements, as was seen on the northbound structure. On the southbound structure, the highest corrosion rates did not necessarily correspond to the most negative potentials. The agreement between the corrosion rates estimated using the two techniques was encouraging, as all of the estimates were within a factor of two or three, which is considered good using these techniques.
In 1997, linear polarization measurements were performed on both spans at locations with high, medium, and low corrosion potentials. The testing was performed using a James Instruments GECOR 6 Polarization Resistance Tester. The equipment uses a confinement ring to delineate the area under test. For comparison, tests were also performed using an EG&G PARC Model 270A potentiostat under computer control, as was performed during the initial 1994 tests. The test equipment performed poorly on the southbound LMC side, with the current frequently unconfined and the readings erratic. Readings for which the signal was not fully confined are suspect because the area under test is unknown and cannot be compensated for. The results of the testing for both the northbound and the southbound structures are shown in table 10.
Direction | Grid Location |
Half-Cell Potential (mV) |
Measured Corrosion Rate (μA/cm2) |
Notes |
---|---|---|---|---|
Southbound LMC |
D2 | –374 | 0.725 | Not fully confined, corrosion rate measured as 0.604 μA/cm2 using EG&G PARC equipment |
A3 | –197 | 2.759 | Not fully confined | |
E13 | –338 | 0.055 | Not fully confined | |
B11 | –199 | 0.022 | Not fully confined | |
Northbound SFC | H16 | –176 | 0.052 | Prewet |
H16 | –222 | 0.110 | Prewet | |
B16 | –149 | 0.131 | Prewet | |
B16 | –133 | 1.293 | Prewet, not fully confined | |
C12 | –421 | 1.184 | Uncracked, not fully confined | |
G22 | –468 | 0.630 | At a crack, prewet | |
G22 | –390 | 0.044 | Near joint | |
B10 | –421 | 0.202 | Near bond core, wet from coring | |
F4 | –384 | 0.182 | After coring | |
F4 | –385 | 2.578 | Repeat of previous test, not fully confined | |
F4 | –396 | 0.441 | Same location as Point F4 above, but opposite side of core hole | |
F12 | –430 | 0.789 | After coring | |
F12 | –418 | 0.516 | Adjacent bar | |
F10 | –441 | 0.491 | Surface dry | |
B14 | –250 | 0.442 | Not fully confined, surface dry |
In 1994, three of the cores removed from the southbound structure and four of the cores removed from the northbound structure were tested to determine their bond strength in direct shear. The overlay thicknesses measured ranged from 25 to 102 mm (1 to 4 inches). The observed clear cover over the bars ranged from 76 to 102 mm (3 to 4 inches). The results of the testing are shown in table 11.
Structure / Overlay Type |
Core ID |
Bond Strength MPa (psi) |
Failure Location (percent of failure area) | ||
---|---|---|---|---|---|
Overlay | Bond Line | Substrate | |||
Northbound Silica fume |
N2 | 3.16 (459) | 5 | 0 | 95 |
N4 | 4.09 (593) | 15 | 0 | 85 | |
N8 | 4.52 (656) | 15 | 0 | 85 | |
N9 | 3.42 (496) | 15 | 0 | 85 | |
Southbound LMC-III |
S6 | 3.32 (481) | 5 | 5 | 90 |
S7 | 4.72 (684) | 20 | 0 | 80 | |
S8 | 4.76 (691) | 10 | 0 | 90 |
The average bond strength for the northbound SFC cores was 3.79 megapascal (MPa) (551 pounds per square inch (psi) and for the southbound LMC cores was 4.27 MPa (619 psi). The failure strengths of the two overlays were similar. This is to be expected as both failed primarily in the substrate concrete, which was nominally the same in both concretes. The failure of the cores primarily in the substrate concrete indicates that the surface preparation and application was excellent for these two overlays, as improper installation or poor surface preparation will typically result in failures at the bond line, with the substrate concrete left intact.
The shear bond testing results for 1997 are shown in table 12. The average bond strength for the northbound SFC cores is 5.8 MPa (840 psi) and for the southbound LMC cores is 4.83 MPa (700 psi). The failure locations were primarily within the base concrete. These results indicate that the bond between the overlay and the substrate remains good.
Structure / Overlay Type |
Core ID |
Exposure | Bond Strength MPa (psi) |
---|---|---|---|
Northbound Silica Fume | N1 | Wheelpath, near delamination | 3.6 (520) |
N2 | Wheelpath | 9.1 (1320) | |
N3 | Wheelpath | 6.2 (900) | |
N7 | Shoulder | 4.3 (620) | |
Southbound LMC-III | S1 | Wheelpath | 4.1 (600) |
S2 | Shoulder | 4.6 (670) | |
S3 | Centerline | 5.7 (830) |
Chloride analyses were performed on six cores from each overlay in 1994. The samples were taken from the 13- to 25-mm (0.5- to 1-inch) depth surface region and from the 95- to 102-mm (3.75- to 4-inches) depth where the reinforcing bars are typically located. All of the slices were pulverized and analyzed to determine their acid-soluble chloride content. In addition to the acid-soluble chloride analyses, water-soluble chloride analyses were performed on selected samples. The results of the testing are shown in table 13.
Structure / Overlay Type |
Sample Number |
Acid-Soluble Chloride Content, percent by concrete weight |
Water-Soluble Chloride Content, percent by concrete weight |
||
---|---|---|---|---|---|
0.5 to 1 inch |
3.75 to 4 inches |
0.5 to 1 inch |
3.75 to 4 inches |
||
Northbound Silica fume |
N1 | 0.017 | 0.075 | - | - |
N3 | - | 0.129 | - | - | |
N5 | - | 0.076 | - | - | |
N6 | - | 0.095 | - | - | |
N7 | 0.018 | 0.039 | - | - | |
N8 | - | 0.018 | - | - | |
Southbound LMC-III |
S2 | - | 0.046 | - | 0.032 |
S3 | 0.048 | - | - | 0.032 | |
S4 | - | 0.045 | - | 0.025 | |
S5 | - | 0.093 | - | - | |
S7 | - | <0.007 | - | - | |
S9 | 0.041 | 0.034 | - | - |
All of the locations had moderate to high chloride concentrations at the level of the reinforcing bars, except for sample S7 and N8. The threshold for chloride-induced corrosion of black reinforcing steel is approximately 0.025 percent by concrete weight, so most of the locations in the deck appear to have sufficient chloride to support corrosion. This agrees with the indication of corrosion given by the half-cell potential and corrosion rate readings. The chloride analyses for the overlays in 1998 are shown in table 14. Two samples were taken from each structure. Chlorides in the near surface of the overlays tend to be higher in the LMC concrete than in the SFC concrete. Except for one core from the LMC that had high chlorides throughout its depth, chlorides typically decrease to moderately low levels at the 15.9 to 25.4 mm (0.625 to 1 inch) depth
Overlay Type |
Sample Number |
Acid-Soluble Chloride Content, percent by concrete weight | |||
---|---|---|---|---|---|
0.25 to 0.5 inch |
0.625 to 1 inch |
1.125 to 1.5 inches |
3.75 to 4 inches |
||
Northbound | N1 | 0.210 | 0.032 | 0.019 | 0.059 |
Silica fume | N2 | 0.065 | 0.034 | 0.020 | 0.023 |
Southbound | S1 | 0.278 | 0.202 | 0.218 | 0.094 |
LMC-III | S2 | 0.208 | 0.065 | 0.021 | 0.050 |
The SFC and LMC-III overlays were placed in May 1992. Overall, the overlays were in good condition in 1998. Each overlay had some cracking and delaminations increased for zero to six locations between 1994 and 1998. The total area of delamination was moderate and about 2.5 m2 (26.9 ft2) SFC and 2.9 m2 (31.2 ft2) LMC-III in 1998. This represents 6.6 percent SFC and 7 percent LMC-III of the total deck area. Delaminations have been increasing in a linear or slightly exponential manner since 1994. Shear load tests indicate that the nondelaminated areas maintain high bond strength for both overlays. The percentages of half-cell potentials more positive than –250mV (CuCuSO4) are presented in table 15. The potential data indicate that corrosion of the embedded steel is likely over three-quarters of each deck but that the potentials have shifted slightly more positive (less corrosive) with time.
Year | NB (SFC) | SB (LMC-III) |
---|---|---|
1994 | 14.2 | 1.7 |
1997 | 24.1 | 13.1 |
Corrosion rates of the embedded reinforcing were measured in 1994 and in 1997 using various techniques. Difficulties in measuring the corrosion rates through the overlays were encountered. The rates varied from low to high values. Corrosion of areas of both decks continues. The chloride ion content in the original deck concrete at the level of the reinforcing steel is sufficient to cause corrosion of mild steel. Redistribution of this chloride with time has occurred and is expected. Chloride has also penetrated into the surface of the overlays in large amounts. Future corrosion and possibly continued delamination or other distress should be anticipated on this bridge.