January 2007
Research Report: UCPRC- RR- 2005- 10
Laboratory Evaluation of Corrosion
Resistance of Steel Dowels
in Concrete Pavement
Authors:
Mauricio Mancio, Cruz Carlos Jr., Jieying Zhang,
John T. Harvey, and Paulo J. M. Monteiro
Work Conducted as part of Partnered Pavement Research Center Strategic Plan
Element No. 4.8: Dowel Bar Retrofit Rehabilitation of Rigid Pavements
PREPARED FOR:
California Department of Transportation
( Caltrans), Division of Research and
Innovation
PREPARED BY:
University of California
Pavement Research Center
UC Davis and Berkeley
ii
DOCUMENT RETRIEVAL PAGE RR No: UCPRC- RR-
2005- 10
Title: Laboratory Evaluation of Corrosion Resistance of Steel Dowels in Concrete Pavement
Authors: Mauricio Mancio, Cruz Carlos Jr., Jieying Zhang, John T. Harvey, and Paulo J. M. Monteiro
Prepared for:
Caltrans Division of Research and
Innovation.
FHWA No.
S/ CA/ RI- 2006/ 27
Client Reference No:
Strategic Plan Element Number 4.8
Date:
January 2007
Strategic Plan Element No:
4.8
Status:
Final
Version No:
Stage 5
Abstract: An extensive investigation was carried out on the corrosion performance of several types of steel dowels
embedded in concrete beams, to develop recommendations for their use in different environmental conditions.
Specimens were exposed to concentrated chloride solutions to accelerate corrosion. The types of dowel bars
evaluated in the study were: bare carbon steel, stainless steel clad, grout- filled hollow stainless steel, microcomposite
steel, carbon steel coated with flexible epoxy ( green color code), and carbon steel coated with non- flexible epoxies ( two
types: purple and gray color codes). Half- cell potential measurements ( indicative of the probability of corrosion
activity), and Linear Polarization Resistance, LPR ( for determining corrosion rate) were used. Every epoxy- coated bar
examined had one or more defects on the coating, especially along the edges at the ends.
The study was conducted in three phases. Phase I: Dowels were cast in concrete beams with joints and then
subjected to a corrosive environment. Half- cell potential was monitored for six months. Phase II: similar to phase I, but
using a more permeable concrete and accelerated corrosion for a period of 18 months. Half- cell potential tests, LPR
curves, visual inspections, chloride content analyses, and scanning electron microscopic investigations were carried out
to evaluate the corrosion performance. Phase III: consisted in testing concrete slabs with transverse joints from a DBR
project in Washington State. The slabs and dowels were subjected to half- cell potential tests, LPR curves, chloride
content tests, and visual inspections. Measurement of chloride contents of these and other concrete cores we used for
comparison of the laboratory and field conditions. The following main recommendations are drawn. 1. Uncoated
carbon steel dowels should not be used. 2. Epoxy dowels present some risk of corrosion. It is recommended that: a)
Quality control checks to control holidays should be implemented, b) Bar ends should be coated with epoxy, and care
must be taken during shipping, storage, and installation. 3. It is recommended that the use of stainless steel clad, hollow
stainless steel, or microcomposite steel dowels should be considered for locations with high risk of high chloride
exposure ( such as on mountain passes and marine environments).
Keywords: Dowel Bar Retrofit, corrosion, chloride exposure, steel dowel bars, stainless steel clad, hollow stainless
steel, microcomposite, epoxy coated steel dowel, half- cell potential measurements, linear polarization resistance
Proposals for implementation: Uncoated carbon steel dowels should not be used. Epoxy dowels present some risk of
corrosion. It is recommended that: a) Quality control checks to control holidays be implemented, b) Bar ends should be
coated with epoxy, and care must be taken during shipping, storage, and installation. Stainless steel clad, hollow
stainless steel, or microcomposite steel dowels should be considered for locations with high risk of high chloride
exposure.
Related documents: Test Plan for project 4.8 “ Dowel Bar Retrofit Rehabilitation of Rigid Pavements”. Ukiah and
Palmdale HVS Tests Reports.
Signatures:
Mauricio Mancio
1st Author
Erwin Kohler
Technical Review
D. Spinner
Editor
John Harvey
Principal
Investigator
Michael Samadian
Caltrans Contract
Manager
iii
DISCLAIMER
The contents of this report reflect the views of the authors who are responsible for the facts and
accuracy of the data presented herein. The contents do not necessarily reflect the official views or policies of
the State of California or the Federal Highway Administration. This report does not constitute a standard,
specification, or regulation.
PROJECT OBJECTIVES
The work presented in this document is part of UCPRC Strategic Plan Item 4.8: “ Dowel Bar Retrofit,
Rehabilitation of Rigid Pavements.” The goal of this research is to establish whether the dowel bar retrofit
( DBR) rehabilitation technique provides adequate performance relative to its cost for the rehabilitation of rigid
pavements. In order to obtain the expected outcome, the program addresses the following four areas of
research:
1. Accelerated pavement testing with the Heavy Vehicle Simulator ( HVS)
2. Live field traffic testing ( collection of field data on a long- term basis)
3. Laboratory testing of DBR materials ( steel and fiber- reinforced polymer dowel bars)
4. Modeling/ analysis ( finite element analysis of doweled concrete pavement joint, DBR
performance, and life- cycle cost analyses).
This document addresses part of Task No. 3.
The objective of this particular study was to perform a laboratory investigation of the corrosion
performance of several types of steel dowels embedded in concrete beams. The concrete beams and dowels
were exposed to concentrated chloride solutions intended to accelerate corrosion and simulate environmental
conditions. The purpose of the investigation was to develop recommendations for use of different types of
dowels for different environmental risk conditions. To provide an indication of the aggressiveness of the
laboratory conditions relative to field conditions, a set of field slabs from Washington State was examined, and
chloride- content analyses were performed on the concrete beams used in the laboratory studies and on cores
taken from nine locations in Washington State.
iv
EXECUTIVE SUMMARY
Dowels are used in jointed concrete pavements to provide load transfer across transverse joints. Their
use reduces vertical deflections that cause faulting and stresses that cause corner and longitudinal cracking by
transferring part of the load to the unloaded slab. However, if corrosion of the dowels occurs, a number of
problems can arise that can compromise the performance of the pavement and lead to premature failure. These
problems include ( a) the loss of dowel cross- section, which reduces load transfer capability, and ( b) the
accumulation of corrosion products, which restricts the free expansion and contraction of the slabs.
Dowel bar corrosion has been investigated in the field and laboratory in the past, which has
led to the widespread use of epoxy coatings for steel dowels in concrete pavements in place of bare carbon
steel. Steel reinforcement in sound concrete is protected from corrosion by a passive film formed due to the
high pH ( 12.5– 13.5) of concrete pore solutions. This thin protective film slows the corrosion reaction rate to
very low levels. However, if the passive layer is broken or dissolves, then the metal reverts to active behavior
and rapid corrosion can occur. The process of steel corrosion in reinforced concrete structures is subdivided
into an initiation stage and a propagation stage.
The initiation stage is the time necessary for depassivation of the protective passive layer as a
result of the penetration and concentration of aggressive agents such as carbon dioxide and chloride ions. For
dowels in concrete pavements, the initiation stage is very short because of easy access to the dowels by
aggressive agents through the joints; therefore the corrosion performance of the system depends largely on the
properties of the steel dowel being used. Aggressive agents also potentially access the full length of the dowels
because the bond between the dowels and concrete is designed to be tight but to have low friction, which
probably permits easier diffusion of the agents along the dowel than along other reinforcing materials ( rebar).
The cyclic horizontal movement of the pavement slabs and the dowels is also likely to increase the risk of
damaging protective coatings such as epoxy.
Most state agencies, including the California Department of Transportation ( Caltrans), seal
the joints of concrete pavements in order to minimize the ingress of water and fine debris into the joint. The
effectiveness of this joint- sealing practice in preventing aggressive agents from accessing dowels is unknown.
Further, sealing the joint from the surface cannot completely seal the joint from exposure to water and debris
because these can still reach the dowel from the sides of the joint or from beneath it. Aggressive agents can
also access the dowels by penetrating the concrete.
Dowel Types Evaluated
Seven kinds of steel dowels were evaluated in this study: bare carbon steel, stainless steel- clad, grout-filled
hollow stainless steel, microcomposite steel, carbon steel coated with flexible epoxy ( green color code,
Designation ASTM A775), and carbon steel- coated with non- flexible epoxies ( two types: purple and gray
color codes, Designation ASTM A934).
The stainless- clad bars have a core of carbon steel covered by an outer layer ( approximately 5
mm thick) of stainless steel. The ends of the stainless- clad dowels do not have stainless steel cladding, but they
do have a protective paint coat. Epoxy- coated bars were also epoxy- coated at the ends. The stainless hollow
dowels consisted of a hollow stainless steel cylinder with a wall thickness of approximately 5 mm, filled with a
cementitious grout.
In this study, microcomposite steel refers to microstructurally designed steels with a
dislocated lath structure ( laths of martensite alternating with thin films of austenite) in which the formation of
carbides is avoided. Microcomposite dowels are anticipated to be more resistant to corrosion than carbon steel.
v
Test Methods
Half- cell potential measurements are indicative of the probability of corrosion activity of the
reinforcing steel located beneath the half- cell. The procedure is described by ASTM C876. Half- cell potential
measurement has been widely used in the field due to its simplicity and general agreement that this technique
is a good indicator of the existence of active corrosion along the steel reinforcement in concrete.
The Linear Polarization Resistance ( LPR) technique is a well- established method for
determining corrosion rate by using electrolytic test cells. The corrosion rate, expressed as the corrosion
current density, is inversely related to the polarization resistance. A major concern with the LPR technique is
uncertainty about the area of the steel bar that is affected by the current from the counter electrode. In the
present study, the counter electrode was placed over a fabricated joint filled with NaCl solution and located
above the dowel. Since virtually all the current will flow through the salt solution — which represents a path of
very low resistance compared to concrete — it has been assumed that the area polarized corresponds to that
part of the dowel exposed to the NaCl solution inside the joint. Epoxy- coated bars, however, could not be
evaluated quantitatively using the LPR technique. It was found that the corrosion that occurred in the epoxy-coated
dowels was localized and concentrated in small holidays ( pinholes, voids, defects, etc.) and defects on
the coating, and therefore the assumption about the exposed area does not hold for these bars.
In order to facilitate the identification of corroded areas in the epoxy- coated dowels and the
evaluation of the role of defects in the development of localized corrosion, the epoxy- coated dowels were
checked for holidays. This was achieved using a low- voltage holiday detector tester before casting the dowels
in concrete beams. This mapping of coating defects was used to check against locations of corrosion, identified
during the visual inspections of corroded dowels after conditioning. Every epoxy- coated bar examined had one
or more defects on the coating, especially along the edges at the ends. These dowels were shipped from the
manufacturer directly to the laboratory and were subjected to careful handling in the laboratory.
Chloride analyses were performed on concrete cores extracted from the laboratory beams
from around the dowels in pavement slabs obtained from an early dowel bar retrofit project in Washington
State, and from slab corners of field slabs at six locations in Washington State. The samples taken from
different levels in the cores were tested by the Caltrans chemistry laboratory or Construction Testing
Laboratory in Illinois.
Experiment Design
This study was conducted in three phases. Details of the three phases are as follows.
Phase I
Four types of dowels were cast in concrete beams with joints. The four types of dowels investigated
were: carbon steel; stainless steel- clad; stainless hollow; and carbon steel coated with flexible epoxy. The
dowels ( 1.5 × 18 in.), were cast in concrete beams measuring ( 6 × 6 × 22 in.). Electrical connections were
made to the steel bars before casting, expansion end caps were installed on the assembly, and the dowels were
mounted on plastic chairs. In the middle of the concrete beam, a joint was simulated by using a polystyrene
foam spacer that was removed after 30 days.
The specimens were subjected to a corrosive environment ( weekly wet and dry cycling with 3% NaCl
solution ponded on top of the beams, permitting access of the corrosive solution through the simulated joint) at
two temperatures: cold ( 4.4 º C) and hot ( 40– 43 º C). No mechanical loading was placed on the beams.
Half- cell potential was monitored for six months, and visual inspection of the corroded dowels was
made at the end of testing. Three replicates for each dowel type were tested.
vi
Phase II
Seven types of dowels were cast in concrete beams with joints: carbon steel; microcomposite steel;
stainless steel- clad; stainless steel hollow; carbon steel coated with flexible epoxy ( green color- code); and
carbon steel coated ( two types) with non- flexible epoxies ( purple and gray color codes). The dimensions of the
specimens are the same as those in Phase I, however, a more permeable concrete was used, with water- to-cement
ratio of 0.65 instead of 0.42, plus a different cement type and larger aggregate size. Corrosion was
accelerated by exposing the samples to cycles of a 3.5% NaCl solution at room temperature for a period of 18
months. Half- cell potential tests, Linear Polarization Resistance ( LPR) curves, visual inspections, chloride-content
analyses, and scanning electron microscopic ( SEM) investigations were carried out to evaluate the
corrosion performance of the dowels. Four replicates of each type of steel dowel were tested in this phase.
Phase III
Three concrete slabs with two transverse joints were extracted from a dowel bar retrofit project in
Washington State. The joints showed loss of load transfer efficiency after 13 years of service. The slabs and
dowels were shipped to Richmond, California, and subjected to half- cell potential tests, LPR curves, chloride-content
tests, and visual inspections. This phase of the study also included measurement of chloride contents of
concrete cores taken from the transverse joints of field slabs in various climate regions in Washington provided
by the Washington State Department of Transportation ( WSDOT), and of cores taken from the laboratory
beam specimens from the Phase II testing, for comparison of the laboratory and field conditions.
Conclusions
The following sections present conclusions from the various phases of the study.
Phase I Testing
• Carbon steel dowels present the shortest corrosion initiation period — when chlorides have direct access to
the bar through the joint, the initiation stage can be disregarded and the corrosion propagation phase
begins immediately. Epoxy- coated dowels exhibited a considerably lengthened initiation period, while the
stainless hollow and stainless clad dowels provided the highest resistance to the onset of corrosion.
• From the visual inspections after six months of cyclic ponding, it was observed that the carbon steel
dowels exhibited uniform corrosion along the bar. Epoxy- coated dowels had localized corrosion at defects
— mostly at the ends of the bars where the coating is most vulnerable to damage. No visible corrosion was
observed on either the stainless steel hollow bars or stainless clad bars.
Phase II Laboratory Testing
• In coated specimens, such as the epoxy- coated specimens included in this study, corrosion is not uniform,
but is instead concentrated at localized defective areas ( e. g., pinholes, voids, etc.). Given that epoxy is an
electrical insulator, polarization only happens at very small locations ( defective areas) that cannot be
accounted for in the calculation of the polarization resistance ( Rp) term. Therefore, the epoxy- coated
dowels cannot be quantitatively evaluated with the other dowels and must be evaluated qualitatively.
vii
• The carbon steel dowels exhibited the lowest values of Rp and therefore have the smallest resistance to
charge transfer across the interface. Carbon steel dowels are therefore expected to have the fastest rate of
corrosion propagation among the types included in this study.
• Microcomposite steel dowels exhibited polarization resistance approximately 35 times larger than carbon
steel dowels, while stainless clad and stainless hollow bars had about 73 times greater polarization
resistance. This observation indicates that the microcomposite steel dowels exhibit much greater resistance
to corrosion propagation than carbon steel dowels, but not as much as the stainless clad and hollow bars.
• Based on corrosion current density results, it was verified that the carbon steel dowels exhibited very rapid
corrosion while microcomposite steel exhibited a moderate level and stainless steel- clad and stainless steel
hollow proceeded at low rates of corrosion.
• Visual inspections of the corroded dowels revealed heavy and mostly uniform corrosion along the carbon
steel dowels, light corrosion in the microcomposite steel dowels, and no visible corrosion in the stainless
steel- clad and stainless steel hollow bars. For the epoxy- coated dowels, the visual inspections generally
revealed that visible corrosion was not widespread, but did occur at a few localized defective areas,
generally at holidays and at the edges of the bar ends. No significant difference was observed on the
performance of non- flexible and flexible epoxy- coated dowels.
• In general, the microscopic investigation by scanning electron microscopic ( SEM) matches well the results
anticipated by the electrical measurements and visual inspections. However, the analysis has focused
mostly on the corroded areas of each sample, and revealed corroded areas that were not visible to the
naked eye.
• Statistical analyses of the results show that in all cases, the type of steel dowel has a statistically significant
effect on the quantitative parameters studied ( i. e., half- cell potential, polarization resistance, and corrosion
current density).
Phase III Testing
• In the extracted slabs from which cores were taken at the joints, a considerable amount of corrosion
product was verified by means of visual inspection beneath the epoxy coating on the central region of the
dowel located below the joint. The corrosion is likely to have contributed to the loss of load transfer
efficiency ( LTE) of the joint because of the low strength corrosion products at the interface between the
concrete and the dowel. Lack of centering of one dowel over the transverse joint is also likely to have
contributed to the low LTE.
• Half- cell potential and Linear Polarization Resistance results match the visual observations, indicating the
presence of active corrosion.
Chloride Concentration
The following conclusions are drawn from the relationship among the concentrations of chlorides in
the laboratory samples, extracted WSDOT field slabs, and cores extracted from slabs at various locations in
Washington State.
• Chloride concentrations close to the pavement joints are significantly higher than in other regions of the
pavement. At the joint, easier access and accumulation of chlorides leads to higher, localized
concentrations.
• When a joint is present, the chloride ions do not diffuse through the concrete ( or grout) from the top;
instead, they migrate through the open joint to the dowel.
viii
• In the field cores, it was verified that the chloride threshold for carbon steel was exceeded in five out of six
projects.
• In the laboratory samples, with open joints located above the dowels, the chloride concentrations are more
constant along the depth profile, as compared to the field conditions in which the chlorides have to diffuse
through the concrete or migrate through a narrower joint.
• The use of a 3.5% NaCl solution for laboratory experiments may lead to higher chloride concentrations
than those found in the field specimens, greatly accelerating the corrosion process compared to the field.
As a result of this aggressive environment, corrosion could be observed in nearly all samples in only 18
months of exposure.
• Laboratory results can be used to comparatively evaluate the corrosion resistance of different materials
when exposed to the same aggressive environment. However, the chloride concentration analyses indicate
that the actual field conditions and local environment should be taken into account when choosing the
appropriate material for a given project.
Recommendations
The following recommendations are based on the conclusions presented above.
• The presence of corrosion at the bar ends and along the bar from ponding water on dowels cast in concrete
in the laboratory indicates that chlorides can pass all the way to the bar ends from the joint along the
horizontal interface between the dowel and the concrete, or through the concrete. For this reason it is
recommended that uncoated carbon steel dowels not be used.
• Epoxy dowels present some risk of corrosion, primarily localized at holidays and the ends of the bars.
Based on this finding, it is recommended that:
1. Quality control checks to control holidays should be implemented.
2. Bar ends should be coated with epoxy, and care must be taken with epoxy- coated dowels during shipping,
storage, and installation. Corrosion will be exacerbated if the bar ends are not coated ( observed on various
Caltrans construction sites) or if the coated ends are damaged during storage, transport, and installation.
• It is recommended that the use of stainless steel- clad, hollow stainless steel, or microcomposite steel
dowels be considered for locations with high risk of high chloride exposure ( such as on mountain passes
and marine environments), where exposure to corrosive water is anticipated. The selection of a specific
corrosion- resistant dowel should be based on further field investigations and cost differences.
• It is recommended that a field study be performed at several mountain pass locations to measure the
chloride content of snowmelt after sand/ salt application for comparison with the chloride content of the
solution used in the laboratory testing in Phases I and II and the core results from Phase III. The results of
this study should be used to further refine the risk assessment in these critical locations.
ix
TABLE OF CONTENTS
Executive Summary........................................................................................................................ .................... iv
List of Figures........................................................................................................................ ............................. xi
List of Tables ............................................................................................................................... ..................... xiv
Abbreviations and Terms Used in the Text ........................................................................................................ xv
1 Introduction ............................................................................................................................... .................. 1
1.1 Background..................................................................................................................... ................... 1
1.1.1 Corrosion Vulnerability ................................................................................................................. 1
1.1.2 Corrosion Prevention ..................................................................................................................... 3
1.2 Objectives and Scope.......................................................................................................................... 4
2 Experiment Design and Test Procedures ...................................................................................................... 5
2.1 Phase I: Laboratory Testing................................................................................................................ 5
2.2 Phase II: Laboratory Testing .............................................................................................................. 7
2.3 Phase III: Evaluation of Field Slabs and Chloride Testing of Cores from Field Slabs....................... 8
2.3.1 Evaluation of Field Slabs ............................................................................................................... 9
2.3.2 Chloride Testing of Cores from In- Service Pavements................................................................ 14
2.4 Detection Techniques ....................................................................................................................... 14
2.5 Holiday Check ............................................................................................................................... .. 15
2.6 Chloride Analyses of Laboratory and Field Cores ........................................................................... 15
3 Results and Discussion ............................................................................................................................... 16
3.1 Phase I Results........................................................................................................................ ......... 16
3.1.1 Half- Cell Potential and Linear Polarization Resistance ............................................................... 16
3.1.2 Visual Inspection..................................................................................................................... .... 21
3.2 Phase II Results ............................................................................................................................... 26
3.2.1 Half- Cell Potential and Linear Polarization Resistance ............................................................... 26
3.2.2 Statistical Analysis of Results...................................................................................................... 34
3.2.3 Visual Inspection..................................................................................................................... .... 35
3.2.4 Microstructural Analysis of Corroded Areas ............................................................................... 39
3.3 Phase III Results ............................................................................................................................... 46
3.3.1 Extracted WSDOT Pavement Slabs............................................................................................. 46
3.4 Analysis of Chloride Contents.......................................................................................................... 49
3.4.1 Chloride Analyses from WSDOT Slab Extracted from Interstate 90 .......................................... 49
3.4.2 Chloride Analyses of Cores from WSDOT Pavements in Various Locations ............................. 53
3.4.3 Chloride Analyses from Laboratory Samples .............................................................................. 57
x
4 Conclusions and Recommendations ........................................................................................................... 62
4.1 Phase I Testing Conclusions............................................................................................................. 62
4.2 Phase II Testing Conclusions............................................................................................................ 62
4.3 Conclusions from Phase III Testing.................................................................................................. 63
4.4 Conclusions from Chloride Analysis................................................................................................ 63
4.5 Recommendations ............................................................................................................................ 64
5 References ............................................................................................................................... .................. 65
Appendix A: Detection Techniques.................................................................................................................. A- 1
Appendix B: Linear Polarization Results of Individual Specimens.................................................................. B- 1
Appendix C: Concrete Mix Proportions ........................................................................................................... C- 1
Appendix D: Chloride Thresholds .................................................................................................................... D- 1
Appendix E: Half- cell Potentials and Linear Polarization Resistance Laboratory Results............................... E- 1
Appendix F: Chloride Test Results.................................................................................................................... F- 1
Appendix G: Concrete Technology Laboratories ( CTL) Results ( Raw Data).................................................. G- 1
Appendix H: Characteristics of Microcomposite Steel Used in the Research.................................................. H- 1
xi
LIST OF FIGURES
Figure 1. Schematic representation of steel corrosion sequence in concrete. ( 6) ................................................ 2
Figure 2. Aggressive agents have free access to dowels and are easily dispersed along the length of the dowel.
............................................................................................................................... ..................................... 3
Figure 3. Steel dowel types investigated in Phase I. From top to bottom: carbon steel, stainless steel clad,
carbon steel coated with bendable epoxy, and stainless steel hollow. ......................................................... 6
Figure 4. Sequential preparation of dowels before casting in the concrete beams. .............................................. 6
Figure 5. Half- cell potential test using a high impedance ( 10 Meg- Ohm) voltmeter, and a copper/ copper
sulfate reference cell. ............................................................................................................................... ... 7
Figure 6. Different types of steel dowels investigated. From left to right: microcomposite steel, stainless steel
hollow, stainless steel- clad, non- bendable epoxy- coated dowel ( gray coating), non- flexible epoxy- coated
dowel ( purple coating), flexible epoxy- coated dowel ( green coating), carbon steel................................... 8
Figure 7. Electrical connections made to the dowels ( a) before casting them in the concrete in order to perform
half- cell potential and Linear Polarization Resistance experiments. Connections were sealed with epoxy
( b) before casting. ............................................................................................................................... ...... 10
Figure 8. Electrical connections made to the stainless steel hollow specimens. Connections were made at the
side of the dowel ( a) and sealed with epoxy ( b) before casting. ................................................................ 11
Figure 9. Experiment setup to accelerate corrosion using chloride ponding. ..................................................... 12
Figure 10. Picture of the experimental setup illustrated in Figure 9. .................................................................. 12
Figure 11. WSDOT slabs at the UC Pavement Research Center lab, located at the University of California,
Berkeley Richmond Field Station.............................................................................................................. 13
Figure 12. Half- cell potentials of carbon steel bars in chloride ponding at 40– 43 º C. ........................................ 17
Figure 13. Half- cell potentials of carbon steel bars in chloride ponding at 4.4 º C. ............................................. 17
Figure 14. Half- cell potentials of epoxy- coated bars in chloride ponding at 40– 43 º C. ...................................... 18
Figure 15. Measured potentials of epoxy- coated bars in chloride ponding at 4.4 º C........................................... 18
Figure 16. Half- cell potentials of stainless hollow steel bars in chloride ponding at 40– 43 º C........................... 19
Figure 17. Half- cell potentials of stainless hollow steel bars in chloride ponding at 4.4 º C................................ 19
Figure 18. Half- cell potentials of stainless clad steel bars in chloride ponding at 40– 43 º C. .............................. 20
Figure 19. Half- cell potentials of stainless clad steel bars in chloride ponding at 4.4 º C. ................................... 20
Figure 20. Linear polarization of the carbon and stainless steels at 40– 43 º C..................................................... 22
Figure 21. Extraction of dowels for visual inspection. ....................................................................................... 23
Figure 22. Corrosion along the carbon steel bars................................................................................................ 23
Figure 23. Corrosion on the epoxy- coated bars. ................................................................................................. 24
Figure 24. Appearance of stainless steel hollow bars after tests......................................................................... 25
Figure 25. Appearance of stainless clad bars after tests...................................................................................... 25
xii
Figure 26. Linear Polarization Resistance results for carbon steel dowels. ........................................................ 27
Figure 27. Linear Polarization Resistance results for microcomposite steel dowels. ......................................... 27
Figure 28. Linear Polarization Resistance results for stainless steel- clad dowels. ............................................. 28
Figure 29. Linear Polarization Resistance results for stainless steel hollow dowels. ......................................... 28
Figure 30. Linear Polarization Resistance results for non- flexible purple epoxy- coated dowels. ...................... 29
Figure 31. Linear Polarization Resistance results for non- flexible gray epoxy- coated dowels. ......................... 29
Figure 32. Linear Polarization Resistance results for flexible green epoxy- coated dowels................................ 30
Figure 33. Summary plot showing the variation of potential and current density about the half- cell potential,
for different dowels......................................................................................................................... .......... 31
Figure 34. Detail of the region around the half- cell potential. The greater the slope, the higher the corrosion
resistance..................................................................................................................... .............................. 31
Figure 35. Comparative column plots showing average results of polarization resistance ( Rp) and corrosion
current density ( icorr) for non- coated dowels.............................................................................................. 33
Figure 36. Carbon steel dowels. ( a) Center region below joint; ( b) End with electrical connection. ................. 36
Figure 37. Microcomposite steel dowels. ( a) General view of the center region below joint; ( b) End region. .. 36
Figure 38. Stainless clad dowels. ( a) General view of the center region below joint; ( b) End region showing
corrosion between the carbon steel core and the stainless steel outer layer............................................... 37
Figure 39. Stainless hollow dowels. ( a) General view of the center region below joint; ( b) End region showing
grouted core. ............................................................................................................................... .............. 37
Figure 40. Epoxy- coated dowel ( purple). ( a) General view of the center region; ( b) Corrosion was verified at
the end region, underneath the epoxy seal. ................................................................................................ 37
Figure 41. Epoxy- coated dowel ( gray). ( a) General view of the center region; ( b) Corroded edges at the end of
dowels......................................................................................................................... .............................. 38
Figure 42. Epoxy- coated dowel ( green). ( a) General view of the center region; ( b) Corroded edges and
corrosion underneath coating at the end of dowels.................................................................................... 38
Figure 43. Corrosion product concentrated in a defective area in the epoxy coating ( a); part of the epoxy
coating was removed in order to observe the state of the carbon steel beneath the coating ( b)................. 38
Figure 44. Carbon steel samples: ( a) heavy corrosion along the dowel surface, magnification = 100×; ( b) same
region, 200×; ( c) different region at the surface, 200×; and ( d) corrosion at the bar end, 100×................ 40
Figure 45. Microcomposite steel samples: ( a) aspect of corrosion along at the surface, 100×; ( b) same region,
205×; ( c) and ( d) details of characteristic corrosion sites, 205×. ............................................................... 41
Figure 46. Hollow stainless steel samples: ( a) view at the surface, 100×; ( b) same region, 300×; and ( c) detail
of the surface condition, 1,000×. The surface appears rough, but no signs of corrosion damage were
observed....................................................................................................................... ............................. 42
xiii
Figure 47. Stainless steel- clad samples: ( a) aspect of corrosion along the surface, in a region close to the end of
the dowel, 100×; ( b) zoom around same region, 200×; ( c) details of corrosion at the surface, interface
between sound and corroded area, 2,390×; and ( d) detail of corrosion at an edge, 2,320×....................... 43
Figure 48. Gray epoxy- coated samples: ( a) view of a defect in the coating, with corroded area inside, 50×; ( b)
another region, at an edge, where coating is lifted 75×; ( c) general view of the surface, 100×; and ( d)
detail of a pinhole present in image ( c) ( see arrow), 1,500×...................................................................... 44
Figure 49. Green epoxy- coated samples: ( a) general view of the surface, 100×; ( b) region where part of the
epoxy coating was removed, 100×; ( c) condition of the steel underneath the epoxy, 302×; and ( d) detail
of corroded area and pits present under the coating, 1,330×. .................................................................... 45
Figure 50. Purple epoxy- coated samples: ( a) general view of the surface with corrosion products accumulated
in a holiday, 100×; ( b) zoom in the same region, 200×; and ( c) condition of the steel underneath the
epoxy, 200×. ............................................................................................................................... .............. 46
Figure 51. Linear Polarization Resistance, WSDOT slab, Specimen A. ............................................................ 47
Figure 52. Linear Polarization Resistance, WSDOT slab, Specimen B. ............................................................ 48
Figure 53. Linear Polarization Resistance, WSDOT slab, Specimen C. ............................................................ 48
Figure 54. Samples extracted from the WSDOT pavement slab for chloride analyses. ..................................... 49
Figure 55. View of the interior of a recently made core hole showing corrosion products underneath the epoxy
coating........................................................................................................................ ............................... 50
Figure 56. Chloride profile of the slab concrete ( Core # 1). ................................................................................ 51
Figure 57. Chloride profile of the slab concrete ( Core # 2). ................................................................................ 51
Figure 58. Chloride profile of the grout used for dowel bar retrofit ( Core # 3)................................................... 52
Figure 59. Chloride profile of the grout around the pavement joint ( Core # 5)................................................... 52
Figure 60. Chloride profile of the grout around the pavement joint ( Core # 6)................................................... 53
Figure 61. Chloride profile for I- 5 MP 166.00 ( downtown Seattle). .................................................................. 55
Figure 62. Chloride profile for I- 90 MP 91.324 ( Elk Heights)........................................................................... 55
Figure 63. Chloride profile for I- 90 MP 61.304 ( Price Creek). ......................................................................... 56
Figure 64. Chloride profile for SR 82 MP 10.617 ( Military Road). ................................................................... 56
Figure 65. Chloride profile for SR 82 MP 43.524 ( Wapato). ............................................................................. 57
Figure 66. Chloride profile on the concrete of a carbon steel specimen. ............................................................ 58
Figure 67. Chloride profile on the concrete of a stainless clad specimen. .......................................................... 59
Figure 68. Chloride profile on the concrete of a stainless hollow specimen....................................................... 59
Figure 69. Chloride profile on the concrete of a flexible ( green) epoxy- coated specimen. ................................ 60
Figure 70. Chloride profile on the concrete of a non- flexible ( purple) epoxy- coated specimen. ....................... 60
Figure 71. Chloride profile on the concrete of a microcomposite steel specimen. ............................................. 61
xiv
LIST OF TABLES
Table 1. Relationship between Half- cell Potential and Corrosion of Steel Embedded in Concrete ( ASTM
C876) ............................................................................................................................... ......................... 14
Table 2. Relationship between Corrosion Current Density and Corrosion Rate ( 9)........................................... 15
Table 3. Name Code and Environmental Conditions Used During Phase I ....................................................... 16
Table 4. Summary Results of Half- Cell Potential, Polarization Resistance, and Corrosion Current Density .... 32
Table 5. Analysis of Variance ( ANOVA) of Half- Cell Potential Results of Non- Coated Samples ................... 34
Table 6. Analysis of Variance ( ANOVA) of Half- Cell Potential Results of All Samples, Including Epoxy-
Coated Bars........................................................................................................................... .................... 35
Table 7. Analysis of Variance ( ANOVA) of Polarization Resistance ( Rp) Results ........................................... 35
Table 8. Analysis of Variance ( ANOVA) of Corrosion Current Density ( icorr) Results ..................................... 35
Table 9. Samples Extracted from In- Service Pavements for Chloride Analyses ( Information Provided by
WSDOT)......................................................................................................................... .......................... 54
xv
ABBREVIATIONS AND TERMS USED IN THE TEXT
California Department of Transportation ( Caltrans)
Construction Technology Laboratories, Inc. ( CTL)
Falling Weight Deflectometer ( FWD
Linear Polarization Resistance ( LPR)
load transfer efficiency ( LTE)
relative humidity ( RH)
Scanning Electron Microscope ( SEM)
University of California Pavement Research Center ( UCPRC)
Washington State Department of Transportation ( WSDOT)
Corrosion current density ( icorr)
Polarization Resistance ( Rp)
xvi
1
1 INTRODUCTION
The performance of jointed concrete pavements depends to a large extent on adequate load transfer at
the joint. Traffic loads must be effectively transferred across the transverse joints to minimize differential
vertical movement between slabs, and the total movement of each slab. Minimization of vertical slab
movement through the use of dowels has been found to be the most effective method of slowing the
development of faulting, which is a primary cause of roughness on concrete pavements in California. Effective
load transfer at transverse joints also reduces stresses in the slabs responsible for corner cracking. ( 1)
Dowels are available in many materials, including stainless steel, epoxy- coated steel, and fiber-reinforced
polymer. The most common type is the epoxy- coated steel variety. Dowel bars provide a
mechanical connection between slabs to limit differential vertical movement without restricting horizontal
joint movement. Dowels are usually placed at mid- depth in the slab and coated with a bond- breaking substance
to prevent bonding to the concrete and allow for the aforementioned horizontal movement.
Corrosion of dowels can compromise the performance of the dowels and of the pavement in which
they are installed and lead to premature failure. In general, concrete seals the steel dowels from the corrosive
effects of weather and environmental exposure, allowing the dowel bars to function effectively as a long- term
reinforcement. ( 2) However, if corrosion occurs in a steel- reinforced concrete structure, the expansive steel
corrosion products build up tensile stress in the concrete, often large enough to lead to cracking and
deterioration of the structure. ( 3)
In the case of doweled pavements, the dowel must not have any “ play” in the concrete to achieve
maximum load transfer and restriction of vertical movement. The effect of the loss of cross- section of a dowel
due to corrosion introduces “ play” and reduces the dowel’s ability to transfer load and restrain vertical
movement. Davids et al. have used finite element analysis to show that a low level of loose fit between the
dowel and the concrete can substantially reduce load transfer efficiency ( LTE). ( 4) Finally, the accumulation of
the corrosion products may restrict the free expansion and contraction of the slabs, causing pavement lock up
and potentially inducing cracks in the pavement.( 5)
1.1 Background
The combination of concrete and steel is usually regarded as optimal for both mechanical performance
and durability. Theoretically, this combination should be highly durable, as the concrete cover provides a
chemical and physical protection barrier to the steel and can potentially eliminate steel corrosion problems.
Steel reinforcement in sound concrete is protected from corrosion by a passive film formed due to the
high pH ( 12.5- 13.5) of concrete pore solutions. This thin protective film lowers the corrosion reaction rate to
very low levels. However, if the passive layer is broken or dissolves, then the metal reverts to active behavior
and rapid corrosion can occur. In reinforced concrete, two major factors cause the passive coating to break
down:
1. Carbonation ( reaction with CO2), and
2. The presence of chlorides.
Dowel bar corrosion has been investigated in the field and the laboratory in the past, which has lead to
the widespread use of epoxy coatings for steel dowels in concrete pavements in place of bare carbon steel.
1.1.1 Corrosion Vulnerability
In 1982, Tuutti proposed a conceptual model to represent the process of steel corrosion in reinforced
concrete structures. In the model, the service life is subdivided into an initiation stage and a propagation stage.
( 6) Figure 1 illustrates this model.
2
The initiation period is the time necessary for “ depassivation” ( disruption) of the protective passive
layer. The time to corrosion initiation is determined by how rapidly the depassivation process occurs as a result
of the penetration and concentration of aggressive agents such as carbon dioxide ( CO2) and chloride ions ( Cl-).
It is important to note that the initiation stage in Tuutti’s model assumes that the steel reinforcement or
dowels are completely embedded in the concrete, with no direct access by aggressive agents to the steel except
through the intact concrete. However, for dowels in concrete pavements there is a drastically reduced initiation
stage because the access of the aggressive agents to a dowel bar in a pavement is much easier than to
reinforcing steel ( rebar) in a structural concrete element. This is due to the unique functions and design
requirements of dowels:
1. The joints through which dowels pass typically allow for free penetration of aggressive agents
such as oxygen, moisture, and de- icing salts to the bar’s surface ( Figure 2). It is recognized that it
is not possible to construct and maintain a completely water- and airtight joint. Water and air can
enter the joint from below, above, and from the sides. Joints are open the widest when
temperatures are low, which in California is the time when most rainfall occurs.
2. Unlike steel placed in concrete for structural reinforcement ( rebar), the bond between the dowels
and concrete is designed to be tight, but to have low friction, which likely permits easier diffusion
of the aggressive agents along the dowel than along a rebar.
3. The cyclic horizontal movement between the pavement slabs and the dowel would increase the
risk of damaging protective materials like epoxy- coating.
Thus, aggressive agents such as CO2 and Cl-, plus water and oxygen, have free access to the steel
surface of a dowel, which essentially reduces the time necessary to complete the initiation stage.
Figure 1. Schematic representation of steel corrosion sequence in concrete. ( 6)
3
Figure 2. Aggressive agents have free access to dowels and are easily dispersed along
the length of the dowel.
The beginning of the corrosion process starts at the propagation stage, and the length of this stage is
determined by the rate of corrosion, which is mainly influenced by the moisture content of the concrete, the
temperature, the permeability of the concrete, the chemical composition of the pore solution, and the thickness
of the concrete cover.( 6)
1.1.2 Corrosion Prevention
Several methods for reducing the risk of corrosion of steel embedded in concrete exist. Some are
based on the properties of the concrete and others on the characteristics of the steel itself.
Over the years, the general approach to improving the durability of reinforced concrete structures has
focused primarily on the improvement of concrete performance. Improvement of the cracking resistance of the
concrete and reduction of concrete permeability, both of which slow the access of aggressive agents to the
passive layer, have produced considerable benefits and will continue to improve the performance of reinforced
concrete structures. However, in the case of steel- doweled pavements, as discussed in the previous section,
aggressive agents have greater access to the steel than in typical steel- reinforced concrete and therefore the
corrosion performance of the system depends largely on the properties of the steel dowel being used.
Most state agencies, including the California Department of Transportation ( Caltrans), seal the joints
of concrete pavement to minimize the entrance of water and fine debris into the joint. Sealing is often
performed at the time of construction, and joints are often resealed at intervals in the pavement life. The cost
effectiveness of joint sealing is a subject of discussion in the field of concrete pavements, although most states
continue to seal joints. Little research is available in the literature evaluating the effect of joint sealing on
dowel corrosion performance. Significant research has also improved the corrosion resistance of concrete
reinforcement, for example, the development of coated steel dowel bars and use of steels with higher corrosion
resistance, as well as several alternatives to regular carbon steel dowels which are now available.
4
1.2 Objectives and Scope
The main objective of this study was to investigate in the laboratory the corrosion performance of
several types of steel dowels embedded in concrete beams and subjected to environmental conditions intended
to accelerate corrosion by exposure to concentrated chloride solutions. In this study, seven types of steel
dowels were evaluated for their corrosion performance:
• Bare carbon steel,
• Stainless steel- clad,
• Grout- filled hollow stainless steel,
• Microcomposite steel,
• Carbon steel coated with flexible epoxy ( green color code, ASTM Designation A775), and
• Carbon steel coated with non- flexible epoxies ( purple and gray color codes, ASTM Designation A934).
Chapter 2 of this report provides an overview of the experimental work and the description of the test
procedures. Chapter 3 presents the results and discussions of the different phases of the current study, and
Chapter 4 presents the conclusions and recommendations. A review of the detection techniques used, a brief
discussion on chloride thresholds, and detailed test data and raw data are presented in the appendices.
5
2 EXPERIMENT DESIGN AND TEST PROCEDURES
This study included three phases, as follows:
1. Placement of four types of dowels ( carbon steel, stainless steel- clad, hollow stainless steel, and
epoxy- coated steel dowels) in concrete beams with joints, exposed to a corrosive environment
with chlorides, under two temperature extremes ( 4 º C and 40- 43 º C). The evaluation included half-cell
potential tests over time for the determination of the corrosion initiation period, and visual
inspection at the completion of testing. Three replicates of each type of dowel were tested.
2. Placement of seven types of dowels cast in concrete beams with joints, exposed to an accelerated
corrosive environment with chlorides. These specimens were evaluated for corrosion performance
using half- cell potential tests, Linear Polarization Resistance ( LPR) experiments, visual
inspections, chloride analyses, and microscopic investigations by Scanning Electron Microscope
( SEM). In this phase, the in- situ corrosion rate was measured in order to determine how fast
corrosion occurred in the propagation stage. Four replicates of each type of dowel were tested.
3. Evaluation of three concrete slabs with two transverse joints extracted from a dowel bar retrofit
project in Washington State showing loss of load transfer efficiency ( LTE) after 13 years of
service. This phase of the study also included measurement of chloride contents of concrete cores
taken from transverse joints of field slabs in various climate regions in Washington provided by
the Washington State Department of Transportation ( WSDOT) for comparison of the laboratory
and field conditions.
2.1 Phase I: Laboratory Testing
In the first phase, the four types of dowels investigated were:
• Carbon steel. Uncoated and untreated, ASTM A615.
• Stainless steel hollow. The stainless steel hollow dowels consisted of a hollow type 316 stainless steel
cylinder approximately 5 mm thick, filled with a cementitious grout.
• Stainless steel- clad. The stainless clad bars consisted of a core of carbon steel covered by an outer layer
( approximately 5 mm thick) of stainless steel type 316L. The ends of the dowels did not have stainless
steel cladding, but did have a protective paint coat.
• Carbon steel coated with flexible epoxy. Epoxy- coated dowels with epoxy patch coating on the ends.
Figure 3 illustrates the dowels used Phase I. All four types of dowels measured 38.1 mm ( 1.5 in.) in
diameter and 460 mm ( 18 in.) in length. They were cast in concrete beams measuring 150 × 150 × 560 mm ( 6
× 6 × 22 in.). As shown in Figure 4, before casting, electrical connections were made to one end of the steel
bars, expansion end caps were placed, and the dowels were mounted on plastic chairs. In the middle of the
concrete beam, a joint was simulated by using a polystyrene foam spacer, which was carefully removed after
30 days of age.
The water- to- cement ratio ( w/ c) of the concrete was 0.42, and the cement: sand: aggregate ratio was
1: 1.84: 2.76. Calcium sulfoaluminate cement was used, and the maximum size of the coarse aggregate was 38
mm ( 1.5 in.). The specimens were demolded 24 hours after casting and cured at 23 º C and 100 percent relative
humidity ( RH) for 7 days. These specimens were subjected to a corrosive environment ( weekly wet and dry
cycling with 3 percent NaCl solution ponded on top of the beams, permitting access of the corrosive solution
through the simulated joint) at two temperatures: cold ( 4 º C) and hot ( 40– 43 º C). These temperatures
represented average low and high temperatures in different regions in California. No mechanical loading was
placed on the beams.
6
Figure 3. Steel dowel types investigated in Phase I. From top to bottom: carbon steel, stainless steel clad,
carbon steel coated with bendable epoxy, and stainless steel hollow.
Figure 4. Sequential preparation of dowels before casting in the concrete beams.
Half- cell potential was monitored using a copper/ copper sulfate ( Cu/ CuSO4) electrode placed on the
surface of the concrete for six months in order to determine the corrosion initiation period. Visual inspection of
the corroded dowels was made after the experiments were completed. The salt solution was removed from the
7
concrete surface before the electrical measurements were performed. As previously noted, three replicates for
each type of steel dowel were tested. Figure 5 illustrates the experimental setup. Details of the half- cell
potential test are included in APPENDIX A: Detection Techniques.
Figure 5. Half- cell potential test using a high impedance ( 10 Meg- Ohm) voltmeter,
and a copper/ copper sulfate reference cell.
2.2 Phase II: Laboratory Testing
Phase II of the testing considered seven different types of dowels:
• Carbon steel. Uncoated and untreated, ASTM A615.
• Microcomposite steel. Microcomposite steel refers to microstructurally- designed chromium- containing
steels with a dislocated lath structure ( laths of martensite alternating with thin films of austenite) that is
virtually carbide- free. Conventional steels have ferrite- carbide microstructures and in corrosive
environments these carbides become cathodic to ferrite and can develop microgalvanic corrosion cells,
which can lead to accelerated corrosion. On the other hand, microcomposite steels are produced in a way
that avoids the formation of carbides and thus they are anticipated to be more resistant to corrosion. In this
study, the microcomposite steel specimens presented about 9 percent chromium, according to information
provided by the manufacturer ( Appendix H).
• Stainless steel hollow. The stainless steel hollow dowels consisted of a hollow Type 316 stainless steel
cylinder approximately 5 mm thick, filled with a cementitious grout.
• Stainless steel clad. The stainless clad bars have a core of carbon steel covered by an outer layer
( approximately 5 mm thick) of stainless steel Type 316L. The ends of the dowels do not have stainless
steel cladding, but do have a protective paint coat.
• Carbon steel coated with flexible epoxy ( green color code). Epoxy- coated bars were also epoxy- coated at
the ends.
8
• Carbon steel coated with non- flexible epoxy ( two types). Two types were studied: purple and gray color
code epoxy.
The dimensions of the specimens are the same as those in Phase I, however a more permeable
concrete was used with a water- to- cement ratio of 0.65 and mix proportions 1: 3.0: 3.25 ( cement: sand: gravel).
Type I/ II cement was used, and the maximum aggregate size was 12.8 mm ( 0.5 in.). The specimens were
demolded 24 hours after casting and cured at 23 º C and 100 percent relative humidity ( RH) for 28 days.
In the Phase II testing, corrosion was accelerated by exposing the samples to weekly wet- and- dry
cycles of a 3.5 percent NaCl solution at room temperature for a period of 18 months. Half- cell potential tests,
Linear Polarization Resistance ( LPR) curves, visual inspections, chloride analyses, and microscopic
investigations were carried out to evaluate the corrosion performance of the dowels. Four replicates of each
type of steel dowel were tested in this study. Figure 6 illustrates the different types of dowels used. The in- situ
corrosion rate was measured in order to permit estimation of the service life of the dowels under the
accelerated conditions. Details of the detection techniques are included in Appendix A. Figures 7 and 8 show
the electrical connections. Figures 9 and 10 demonstrate the experiment setup.
Figure 6. Different types of steel dowels investigated. From left to right:
microcomposite steel, stainless steel hollow, stainless steel- clad, non- bendable
epoxy- coated dowel ( gray coating), non- flexible epoxy- coated dowel
( purple coating), flexible epoxy- coated dowel ( green coating), carbon steel.
2.3 Phase III: Evaluation of Field Slabs and Chloride Testing of Cores from Field Slabs
In addition to the laboratory experiments described in Section 2.2, further experiments were
performed on concrete pavement slabs obtained from a dowel bar retrofit project from the Washington
Department of Transportation ( WSDOT). These slabs were constructed in 1964; dowel bar retrofit was later
performed in 1994. The slabs were extracted from Interstate 90, near the town of Cle Elum in Washington
State, and shipped to the University of California Pavement Research Center ( UCPRC) at the Richmond Field
9
Station for testing and evaluation. They were selected because load transfer efficiencies ( LTE) measured using
the Falling Weight Deflectometer ( FWD) indicated that the joints were experiencing a decrease in LTE. This
was the first WSDOT dowel bar retrofit project that has shown significant loss of LTE.
According to information provided by WSDOT, until the early 1980s the approximate use of rock salt
was 120– 150 days per year at a typical rate of 180 pounds per mile using a 5: 1 ratio ( one scoop of salt to five
scoops of sand). Each day typically had at least two applications, and salting typically occurs from mid-
November to mid- April. From the early 1980s to 1997 or so, rust inhibitor type de- icers/ salts have been used.
Since 1997, liquid magnesium chloride has been used at a typical rate of 35 gallons per mile with a range of
15 to 50.
2.3.1 Evaluation of Field Slabs
The slabs used in this study were sawed and sealed during initial construction with a rubberized joint
sealer product called “ Seal Target 164.” In 1994, during the dowel bar retrofit, joints were also sealed with
rubberized joint sealant. Figure 11 shows the slabs tested.
The location from which these slabs were removed is at an elevation of 543 m. This location is subject
to 568 mm average annual rainfall, 2,070 mm average annual snowfall, and 76 average annual freeze- thaw
cycles ( air temperatures) ( 7). The slabs were taken from the outside lane and were subjected to approximately
1 million Equivalent Standard Axle Loads ( ESALs) annually during their service life. The dowels are 38 mm
( 1.5 in.) in diameter and the concrete slabs 228 mm ( 9 in.) thick. The dowels were carbon steel coated with
flexible ( green) epoxy. After removing the slabs, half- cell potential tests, Linear Polarization Resistance
experiments, chloride analyses, and visual inspection work were performed in order to provide comparison
with results obtained in the laboratory under accelerated conditions. All testing was performed at the
University of California Pavement Research Center.
10
( a)
( b)
Figure 7. Electrical connections made to the dowels ( a) before casting them in the
concrete in order to perform half- cell potential and Linear Polarization Resistance
experiments. Connections were sealed with epoxy ( b) before casting.
11
( a)
( b)
Figure 8. Electrical connections made to the stainless steel hollow specimens.
Connections were made at the side of the dowel ( a) and sealed with epoxy ( b) before casting.
12
CROSS -
SECTION
SIDE- VIEW
L evel of NaCl solution
Dowel bar
Open- bottom box Open joint
Electrical
Connection
3 ½ ”
1”
Figure 9. Experiment setup to accelerate corrosion using chloride ponding.
Figure 10. Picture of the experimental setup illustrated in Figure 9.
13
( a)
( b)
Figure 11. WSDOT slabs at the UC Pavement Research Center lab, located at
the University of California, Berkeley Richmond Field Station.
14
2.3.2 Chloride Testing of Cores from In- Service Pavements
In addition to the testing performed on the WSDOT slab noted in Section 2.3.1, WSDOT extracted
two cores from near- joint corners of in- service slabs at various locations in Washington. These cores were
tested for chloride content by Construction Technology Laboratories, Inc. ( CTL) for comparison with the
laboratory results. ( Section 3.4.2 describes the locations from which cores were extracted, including years of
construction, salt usage, and joint sealing practices.)
2.4 Detection Techniques
Several electrochemical methods have been used to evaluate the corrosion activity of steel
reinforcement. The half- cell potential and the LPR methods are among the most commonly used and accepted
test methods for in- situ measurements. These tests are easy to perform in the field, and commercial instruments
are readily available. As Carino observes, among the many methods that have been investigated for measuring
the in- situ corrosion rate of steel embedded in concrete, the Linear Polarization Resistance appears to be
gaining most acceptance. ( 8)
The half- cell potential measurements are indicative of the probability of corrosion activity of the
reinforcing steel located beneath the half- cell, and the procedure is described in ASTM C876. The setup
basically consists of an external electrode ( half cell), connecting wires, and a high impedance voltmeter
( impedance > 10MΩ). A high impedance voltmeter is used so that there is a very little current through the
circuit. Copper/ copper sulfate ( Cu/ CuSO4) electrodes were used throughout this research.
The half- cell potential measurement has been widely used in the field due to its simplicity and general
agreement among researchers that this technique effectively indicates the existence of active corrosion along
the steel reinforcement in concrete. Table 1 illustrates the relationship between measured potential and the
likelihood of corrosion activity.
Table 1. Relationship between Half- cell Potential and Corrosion of Steel
Embedded in Concrete ( ASTM C876)
Half- Cell Potential ( mV)* Corrosion Interpretation
> - 200 Low probability ( 10%) of corrosion
- 200 to - 350 Corrosion activity uncertain
< - 350 High probability ( 90%) of corrosion
* Measurements made with a copper/ copper sulfate electrode.
Thus, the half- cell potential method provides an indication of the likelihood of corrosion activity at the
time of the measurement. However, it does not give any indication of the rate of corrosion of the steel.
On the other hand, the LPR technique is a well- established method for determining the rate of
corrosion by using electrolytic test cells. The technique basically involves measuring the change in the open-circuit
potential of the electrolytic cell when an external current is applied to the cell. For a small perturbation
about the open- circuit potential, there is a linear relationship between the change in applied current per unit
area of electrode ( Δi) and the change in the measured voltage ( ΔE). The ratio ΔE/ Δi is called the polarization
resistance ( Rp). The corrosion rate, expressed as the corrosion current density ( i), is inversely related to the
polarization resistance, as indicated by the Stern- Geary relationship i = B/ Rp, where B is a constant ( ASTM
G59). Some guidelines have been developed to establish a relationship between corrosion current density and
corrosion rate, as shown in Table 2. ( 9)
15
Table 2. Relationship between Corrosion Current Density and Corrosion Rate ( 9)
Corrosion Current Density, I corr ( μA/ cm2) Corrosion Rate
< 0.1 Negligible
0.1 – 0.5 Low
0.5 – 1.0 Moderate
> 1.0 High
A major concern with the LPR method is the uncertainty about the area of the steel bar that is affected
by the current from the counter electrode. Usually, it is assumed that the current flows in straight lines
perpendicular to the bar and the counter electrode, and the affected bar area is understood to be the bar
circumference multiplied by the length of the bar below the counter electrode, which in fact is not the case, as
illustrated in Appendix A. However, in the laboratory experiments performed in this study ( Phases I and II),
this concern is not justified, since virtually all the current will flow through the NaCl solution present in the
open joint, which represents a very low- resistivity path compared to concrete. As previously stated, the salt
solution was removed from the concrete surface in order to perform the electrical measurements, but the joint
remained filled with solution to provide this low- resistivity path for current flow. Note that the counter
electrode was placed just above the fabricated joint. Thus, it has been assumed that the area polarized
corresponds to that part of the dowel exposed to the NaCl solution inside the fabricated joint.
The Linear Polarization Resistance experiments were performed using a Princeton Applied Research
Model 263 Potentiostat- Galvanostat. A detailed description of these detection techniques, along with the
explanation of the principles involved, is presented in Appendix A.
2.5 Holiday Check
During Phase II of this research, in order to facilitate the identification of corroded areas in the epoxy-coated
dowels and the evaluation of the role of defects on the development of localized corrosion, the epoxy-coated
dowels were checked for holidays ( pinholes, voids, defects, etc.). This was done before casting the
dowels in the concrete beams using a low- voltage holiday detector tester following ASTM G62 and CTM 685
( brand Tinker & Rasor, Model M/ 1 Holiday Detector). This mapping of coating defects was used to check
against locations of corrosion identified during the visual inspections of corroded dowels after conditioning
( Section 3.2.2). It is worth mentioning that every single epoxy- coated bar used in Phase II presented one or
more defects on the coating, especially along the edges at the dowel ends. These dowels were shipped from the
manufacturer directly to the laboratory and were subjected to reasonably careful handling in the laboratory. It
would be expected that dowel bars would be subjected to much rougher handling on a construction site. The
holidays found in the mapping were not caused by being dropped or banged in the laboratory.
2.6 Chloride Analyses of Laboratory and Field Cores
Several chloride analyses were performed on concrete cores extracted from the laboratory beams and
from the pavement slabs obtained from WSDOT in order to evaluate the concentration of chlorides in the
region surrounding the steel dowels, particularly in the vicinities of the open joints. It is believed that over
time, chloride ions may build up around the joint area and close to the dowels, increasing the chloride
concentration to considerable levels, perhaps way above the chloride threshold of the steel being used. The
extracted samples were shipped to the Caltrans chemistry laboratory where the chemical analyses were
performed. In addition, cores from slab corners were obtained from WSDOT pavement at various locations
and were analyzed for chloride content by Construction Technology Laboratories, Inc. Chloride tests are
performed according to ASTM C1152.
16
3 RESULTS AND DISCUSSION
The results of Phase I are presented in Section 3.1 of this report. The results of Phase II are presented
in Section 3.2. The Phase III results obtained for the WSDOT slabs from the dowel bar retrofit project on
Interstate 90 and the chloride analyses from the WSDOT cores are presented in Section 3.3.
3.1 Phase I Results
In the following text and in Figures 12– 19, the specimens are identified as RS (“ regular steel,” i. e.,
carbon steel), SS ( stainless steel hollow), EX ( epoxy- coated [ flexible epoxy, green code]), and XX ( stainless
clad bars), as illustrated in Table 3.
Table 3. Name Code and Environmental Conditions Used During Phase I
Identification of the Specimens
Type of Dowel Hot Room ( 40– 43 º C) Cold Room ( 4 º C)
Carbon Steel RS- 4, RS- 5, RS- 6 RS- 7, RS- 8, RS- 9
Stainless Hollow SS- 4, SS- 5, SS- 6 SS- 7, SS- 8, SS- 9
Epoxy- coated EX- 4, EX- 5, EX- 6 EX- 7, EX- 8, EX- 9
Clad XX- 4, XX- 5, XX- 6 XX- 1, XX- 2, XX- 3
3.1.1 Half- Cell Potential and Linear Polarization Resistance
Figure 12 through Figure 19 show the measured half- cell potentials with time for the four types of
dowels investigated in the first phase of this study. As can be observed in Figure 12, for two of the three
replicates of the carbon steels at 40– 43 º C, the potentials moved to the active corrosion region lower than - 350
mV, indicating a 90 percent probability of corrosion as described in Table 1) after about 75 days. For
Specimen RS- 5, the potentials were more negative than - 350 mV from nearly the beginning of the test.
17
0 15 30 45 60 75 90 105 120 135 150 165 180
- 950
- 900
- 850
- 800
- 750
- 700
- 650
- 600
- 550
- 500
- 450
- 400
- 350
- 300
- 250
- 200
- 150
- 100
- 50
0
Age ( days)
Half- cell Potential ( mV, CSE)
RS- 4 ( Carbon Steel)
RS- 5 ( Carbon Steel)
RS- 6 ( Carbon Steel)
Above this line (- 0.2v),
90% probability that
no corrosion has occurred
Under this line (- 0.35V),
90% probability that
corrosion has occurred.
Figure 12. Half- cell potentials of carbon steel bars in chloride ponding at 40– 43 º C.
0 15 30 45 60 75 90 105 120 135 150 165 180
- 950
- 900
- 850
- 800
- 750
- 700
- 650
- 600
- 550
- 500
- 450
- 400
- 350
- 300
- 250
- 200
- 150
- 100
- 50
0
Age ( days)
RS- 7 ( Carbon Steel)
RS- 8 ( Carbon Steel)
RS- 9 ( Carbon Steel)
Half- cell Potential ( mV, CSE)
Above this line (- 0.2v),
90% probability that
no corrosion has occurred
Under this line (- 0.35V),
90% probability that
corrosion has occurred.
Figure 13. Half- cell potentials of carbon steel bars in chloride ponding at 4.4 º C.
18
0 15 30 45 60 75 90 105 120 135 150 165 180
- 950
- 900
- 850
- 800
- 750
- 700
- 650
- 600
- 550
- 500
- 450
- 400
- 350
- 300
- 250
- 200
- 150
- 100
- 50
0
Age ( day)
Above this line (- 0.2v),
90% probability that
no corrosion has occurred
Under this line (- 0.35V),
90% probability that
corrosion has occurred.
EX- 4 ( Epoxy- Coated)
EX- 5 ( Epoxy- Coated)
EX- 6 ( Epoxy- Coated)
Half- cell Potential ( mV, CSE)
Figure 14. Half- cell potentials of epoxy- coated bars in chloride ponding at 40– 43 º C.
0 15 30 45 60 75 90 105 120 135 150 165 180
- 950
- 900
- 850
- 800
- 750
- 700
- 650
- 600
- 550
- 500
- 450
- 400
- 350
- 300
- 250
- 200
- 150
- 100
- 50
0
Age ( days)
EX- 7 ( Epoxy- Coated)
EX- 8 ( Epoxy- Coated)
EX- 9 ( Epoxy- Coated)
Above this line (- 0.2v),
90% probability that
no corrosion has occurred
Under this line (- 0.35V),
90% probability that
corrosion has occurred.
Half- cell Potential ( mV, CSE)
Figure 15. Measured potentials of epoxy- coated bars in chloride ponding at 4.4 º C.
19
0 15 30 45 60 75 90 105 120 135 150 165 180
- 950
- 900
- 850
- 800
- 750
- 700
- 650
- 600
- 550
- 500
- 450
- 400
- 350
- 300
- 250
- 200
- 150
- 100
- 50
0
Age ( days)
Half- cell Potential ( mV, CSE)
SS- 4 ( Stainless Hollow)
SS- 5 ( Stainless Hollow)
SS- 6 ( Stainless Hollow)
Above this line (- 0.2v),
90% probability that
no corrosion has occurred
Under this line (- 0.35V),
90% probability that
corrosion has occurred.
Figure 16. Half- cell potentials of stainless hollow steel bars in chloride ponding at 40– 43 º C.
0 15 30 45 60 75 90 105 120 135 150 165 180
- 950
- 900
- 850
- 800
- 750
- 700
- 650
- 600
- 550
- 500
- 450
- 400
- 350
- 300
- 250
- 200
- 150
- 100
- 50
0
Age ( days)
SS- 7 ( Stainless Hollow)
SS- 8 ( Stainless Hollow)
SS- 9 ( Stainless Hollow)
Above this line (- 0.2v),
90% probability that
no corrosion has occurred
Under this line (- 0.35V),
90% probability that
corrosion has occurred.
Half- cell Potential ( mV, CSE)
Figure 17. Half- cell potentials of stainless hollow steel bars in chloride ponding at 4.4 º C.
20
0 15 30 45 60 75 90 105 120 135 150 165 180
- 950
- 900
- 850
- 800
- 750
- 700
- 650
- 600
- 550
- 500
- 450
- 400
- 350
- 300
- 250
- 200
- 150
- 100
- 50
0
Age ( days)
Above this line (- 0.2v),
90% probability that
no corrosion has occurred
Under this line (- 0.35V),
90% probability that
corrosion has occurred.
Half- cell Potential ( mV, CSE)
XX- 4 ( Stainless Clad)
XX- 5 ( Stainless Clad)
XX- 6 ( Stainless Clad)
Figure 18. Half- cell potentials of stainless clad steel bars in chloride ponding at 40– 43 º C.
0 15 30 45 60 75 90 105 120 135 150 165 180
- 950
- 900
- 850
- 800
- 750
- 700
- 650
- 600
- 550
- 500
- 450
- 400
- 350
- 300
- 250
- 200
- 150
- 100
- 50
0
XX- 1 ( Stainless Clad)
XX- 2 ( Stainless Clad)
XX- 3 ( Stainless Clad)
Above this line (- 0.2v),
90% probability that
no corrosion has occurred
Under this line (- 0.35V),
90% probability that
corrosion has occurred.
Half- cell Potential ( mV, CSE)
Age ( days)
Figure 19. Half- cell potentials of stainless clad steel bars in chloride ponding at 4.4 º C.
21
As discussed in Section 1.1.1, compared to reinforcing steel in concrete structures in which the steel is
completely encased in concrete, the initiation period may be greatly reduced in the case of steel dowels due to
the easier access of aggressive agents to the bars.
At 4.4 º C ( Figure 13), potentials moved to the active corrosion zone after about 30 days for RS- 7 and
RS- 9, and not until 75 days for RS- 8. It is interesting to notice that for some specimens corrosion was initiated
earlier for carbon steels in the 4.4 º C room than those in 40– 43 º C. The reason may be the difference in oxygen
availability at different temperatures. The higher the temperature, the lower the oxygen solubility is in water.
For instance, oxygen solubility in saltwater is about 11.7 mg/ L at 4 º C, and the value drops to 6.5 mg/ L at 30 º C.
It is important to mention that the polystyrene foam sheet used to create the joints was removed after
30 days of age. After the polystyrene sheet was removed, corrosion was immediate for the carbon steel in the
4.4 º C room, which indicates that the pavement joint is the dominant access of chloride, oxygen, and water to
the dowel surface. This suggests that joint sealing may potentially reduce the likelihood of or retard the rate of
corrosion of dowels by limiting access of corrosive solutions. However, joints are typically only sealed at the
surface and corrosive solutions may be able to enter the joint from the sides and bottom of the joint in addition
to the surface.
For the epoxy- coated steel bars at high temperatures ( 40– 43 º C) ( see Figure 14 and Figure 15), the
measured potentials for Specimens EX- 4 and EX- 5 experienced an abrupt decrease at around 40 days to
approximately - 650 mV in the high- corrosion risk region. Specimen EX- 6, on the other hand, measured zero
and stayed in the inactive region throughout the test.
For cold temperature specimens ( 4 º C), the measured potential of Specimen EX- 9 remained at zero
throughout the test, possibly indicating absence of defects in its epoxy coating or perhaps some problem with
the electrical connection at the dowel. The potentials of Specimen EX- 7 oscillated within the intermediate
corrosion risk ( uncertain) range, while Specimen EX- 8 entered the high- corrosion risk region (< - 350mV) at
around 70 days. Note that, in Phase I of the study, the epoxy coating was not checked for defects before
casting. This procedure was adopted in the second phase, as described in Section 2.2.
Both stainless hollow and stainless clad dowels ( Figure 16 to Figure 19) generally had measured
potentials above the - 350 mV line at both test temperatures, indicating low probability of corrosion.
In the first phase, emphasis was placed on the determination of the initiation period based on the
measurement of half- cell potentials over time. However, an LPR test to evaluate corrosion rate was also
performed on the carbon and stainless steel dowels. Figure 20 shows the results obtained ( specimens were
stored at 40– 43 º C). The slope of a polarized potential and corresponding current supply curve is the
polarization resistance, Rp. Its value is inversely proportional to the corrosion rate ( i) indicated in Stern- Geary
relationship i = B/ Rp, where B is a constant with the value from 25 mV to 56 mV. The corrosion rate of carbon
steel can be 10 times as high as that of the stainless steel. Note that the corrosion rate of RS4 was low while its
half- cell potential was located in the active corrosion region (<- 350mV), indicating that the half- cell potential
measurement is not a quantitative method. Results from the other two specimens ( RS5 and RS6) matched the
half- cell potential measurements well.
3.1.2 Visual Inspection
At the end of six months, the concrete beams were broken open and the dowels were retrieved and
inspected. Figure 21 shows one of the concrete beams at the moment the dowel was being retrieved.
22
- 0.10 - 0.05 0.00 0.05 0.10
- 80
- 60
- 40
- 20
0
20
40
60
80
Current ( mA)
SS4 ( Stainless Hollow)
SS5 ( Stainless Hollow)
SS6 ( Stainless Hollow)
RS4 ( Carbon Steel)
RS5 ( Carbon Steel)
RS6 ( Carbon Steel)
Hotroom Age: 130 Days
Potential ( mV)
Figure 20. Linear polarization of the carbon and stainless steels at 40– 43 º C.
For the carbon steels exposed to chloride ponding, corrosion products were distributed along the
length and perimeter of the bars for both the cold and hot conditions, as shown in Figure 22. The epoxy- coated
bars had localized corrosion mostly at the ends of the bars where the coating is the most vulnerable to damage.
All the epoxy- coated bars exposed to Chloride ponding, except EX- 9, were corroded at the ends, as shown in
Figure 23. The results matched the half- cell potential in the sense that once corrosion occurred and the coating
was lifted, the measured potential was not zero anymore. No visible corrosion was found on either stainless
steel hollow bars or stainless clad bars, as can be observed in Figure 24 and Figure 25.
23
Figure 21. Extraction of dowels for visual inspection.
Figure 22. Corrosion along the carbon steel bars.
24
Figure 23. Corrosion on the epoxy- coated bars.
25
Figure 24. Appearance of stainless steel hollow bars after tests.
Figure 25. Appearance of stainless clad bars after tests.
26
3.2 Phase II Results
This section presents the results obtained in the second phase of the research, in which half- cell
potential tests, Linear Polarization Resistance curves, visual inspections, chloride analyses and microscopic
investigation were obtained at room temperature on the seven types of steel dowels investigated [ carbon steel,
microcomposite steel, stainless hollow, stainless clad, carbon steel coated with flexible epoxy ( green colored),
and carbon steel coated with non- flexible epoxies ( purple and gray colored)].
3.2.1 Half- Cell Potential and Linear Polarization Resistance
In this section, individual graphs present the LPR results for each type of dowel investigated. The
plots present the average results of four replicates normalized with respect to the half- cell value. Current
density ( μA/ cm2) and potential ( mV) are given as variations from the values obtained at equilibrium, half- cell
potential. Normalizing the data facilitates comparison among the different samples. Non- normalized results for
each type of steel dowel are presented in Appendix B.
As described in Section 2.1, the slope of the curves ( ΔE/ Δi) provides the value of the polarization
resistance, Rp. Notice that the greater the slope, the more difficult the charge transfer across the
metal/ electrolyte interface, and therefore the corrosion current density is lower and, consequently, the
corrosion reaction rate is slower.
Figure 26 to Figure 29 present the Linear Polarization Resistance plots for the carbon steel,
microcomposite, hollow stainless steel, and stainless steel- clad bars, respectively. It can be seen that for the
same variation in potential, carbon steel bars exhibit a much larger variation in the corrosion current density
compared to the other materials. Consequently, the carbon steel samples will exhibit a smaller slope, smaller
polarization resistance, and larger corrosion rate, followed by microcomposite steel, stainless steel- clad, and
stainless steel hollow, respectively.
Figure 30 through Figure 32 present LPR plots for the epoxy- coated bars, respectively non- flexible
( purple and gray) and flexible ( green) epoxy coatings. Note that the LPR results obtained for the epoxy- coated
bars are not to be interpreted quantitatively. In coated specimens, corrosion is not uniform. Rather, it is
concentrated at localized defective areas ( such as pinholes, voids, etc.). Given that epoxy is an electrical
insulator, polarization only occurs at these defective areas, which cannot be accounted for in the calculation of
the polarization resistance term. For this reason, Rp and icorr calculations for epoxy- coated bars are not
presented here.
As mentioned in Section 2, it has been assumed that the polarized area corresponds to the surface of
the dowels exposed to the NaCl solution inside the fabricated joints. If the polarized area decreases in size, the
corrosion current density ( given in μA/ cm2) will increase.
The surface area used in the current density calculations was 7.18 cm2, which corresponds to the
dowel bar circumference ( 119.69 mm) multiplied by the thickness of the joint ( 6 mm). The microcomposite
steel dowels had smaller diameter ( 1.25 in. or 31.75 mm), and the polarized area was thus 5.98 cm2.
27
y = 2.9143x - 0.2403
R2 = 0.9902
- 30
- 20
- 10
0
10
20
30
- 10 - 8 - 6 - 4 - 2 0 2 4 6 8
Current Density ( μA/ cm2)
Potential Variation ( mV)
Figure 26. Linear Polarization Resistance results for carbon steel dowels.
y = 41.802x + 2.3608
R2 = 0.9515
- 30
- 20
- 10
0
10
20
30
- 1.0 - 0.8 - 0.6 - 0.4 - 0.2 0.0 0.2 0.4 0.6
Current Density ( μA/ cm2)
Potential Variation ( mV)
Figure 27. Linear Polarization Resistance results for microcomposite steel dowels.
28
y = 279.31x + 4.6152
R2 = 0.9204
- 30
- 20
- 10
0
10
20
30
- 0.15 - 0.12 - 0.09 - 0.06 - 0.03 0.00 0.03 0.06
Current Density ( μA/ cm2)
Potential Variation ( mV)
Figure 28. Linear Polarization Resistance results for stainless steel- clad dowels.
y = 258.13x + 1.6974
R2 = 0.9312
- 30
- 20
- 10
0
10
20
30
- 0.15 - 0.12 - 0.09 - 0.06 - 0.03 0.00 0.03 0.06 0.09
Current Density ( μA/ cm2)
Potential Variation ( mV)
Figure 29. Linear Polarization Resistance results for stainless steel hollow dowels.
29
y = 91.692x - 0.2301
R2 = 0.9988
- 30
- 20
- 10
0
10
20
30
- 0.30 - 0.25 - 0.20 - 0.15 - 0.10 - 0.05 0.00 0.05 0.10 0.15 0.20 0.25
Current Density ( μA/ cm2)
Potential Variation ( mV)
Figure 30. Linear Polarization Resistance results for non- flexible purple epoxy- coated dowels.
y = 366.55x + 0.2249
R2 = 0.9657
- 30
- 20
- 10
0
10
20
30
- 0.10 - 0.08 - 0.06 - 0.04 - 0.02 0.00 0.02 0.04 0.06 0.08
Current Density ( μA/ cm2)
Potential Variation ( mV)
Figure 31. Linear Polarization Resistance results for non- flexible gray epoxy- coated dowels.
30
y = 409.59x + 2.9279
R2 = 0.9348
- 30
- 20
- 10
0
10
20
30
- 0.09 - 0.07 - 0.05 - 0.03 - 0.01 0.01 0.03 0.05 0.07
Current Density ( μA/ cm2)
Potential Variation ( mV)
Figure 32. Linear Polarization Resistance results for flexible green epoxy- coated dowels.
The fact that electrical contact can be made between the reference electrode ( copper/ copper sulfate)
and the working electrode ( steel dowel), and that a half- cell potential can be measured for epoxy- coated bars,
indicates the presence of defects in the coating. The Linear Polarization Resistance results may indicate the
presence of corrosion in the defective areas, but an evaluation of the extent of corrosion must rely on visual
inspections and microscopic investigations.
Figure 33 and Figure 34 present a comparison of the Linear Polarization Resistance results of various
dowels. It can be easily seen that the carbon steel specimens performed much worse than the other dowels
tested. On the other hand, it can be observed that the stainless steel hollow bars exhibited the best performance,
followed by stainless steel- clad and microcomposite steel dowels. Table 4 summarizes the individual results
obtained for each sample including half- cell potential, polarization resistance ( Rp) and corrosion current
density ( icorr). ( Plots with calculations are presented in Appendix B.) The criteria used to interpret the results
presented in Table 4 are per ASTM C876 and ASTM G59 and are presented in Table 1 and Table 2,
respectively. Figure 35 presents a comparative summary of Linear Polarization Resistance and current density
results.
Regarding the half- cell potential values, note that the stainless steel hollow and stainless steel- clad
specimens are, on average, located in the uncertain corrosion region, while all the other samples are in the
active corrosion region ( refer to Table 1).
31
- 25.0
- 20.0
- 15.0
- 10.0
- 5.0
0.0
5.0
10.0
15.0
20.0
25.0
- 9.0 - 8.0 - 7.0 - 6.0 - 5.0 - 4.0 - 3.0 - 2.0 - 1.0 0.0 1.0 2.0 3.0 4.0 5.0 6.0 7.0 8.0
Current Density ( μA/ cm2)
Potential Variation ( mV)
Carbon Steel
Microcomposite Steel
Stainless Steel Clad
Stainless Steel Hollow
Figure 33. Summary plot showing the variation of potential and current density about
the half- cell potential, for different dowels.
- 25.0
- 20.0
- 15.0
- 10.0
- 5.0
0.0
5.0
10.0
15.0
20.0
25.0
- 1.5 - 1.3 - 1.1 - 0.9 - 0.7 - 0.5 - 0.3 - 0.1 0.1 0.3 0.5 0.7 0.9 1.1 1.3 1.5
Current Density ( μA/ cm2)
Potential Variation ( mV)
Carbon Steel
Microcomposite Steel
Stainless Steel Clad
Stainless Steel Hollow
Figure 34. Detail of the region around the half- cell potential. The greater the slope, the higher the
corrosion resistance.
32
Table 4. Summary Results of Half- Cell Potential, Polarization Resistance, and Corrosion Current
Density
Specimen # Half- Cell Potential
( mV)
Polarization Resistance
( Rp)
Corrosion Current Density
( icorr)
1 - 602 1544.3 16.84
2 - 640 6066.5 4.29
3 - 694 2568.2 10.12
4 - 629 5410.8 4.81
Avg. - 641.3 3897.5 9.01
Carbon
steel
Std. Dev. 38.6 2183.2 5.84
1 - 567 23495 1.11
2 - 571 23146 1.12
3 - 292 367410 0.07
4 - 278 124620 0.21
Avg. - 427.0 134667.8 0.63
Micro-composite
steel
Std. Dev. 164.1 162343.7 0.57
1 - 286 311560 0.08
2 - 179 249520 0.10
3 - 216 361130 0.07
4 - 345 216530 0.12
Avg. - 256.5 284685.0 0.09
Stainless
clad
Std. Dev. 73.8 64414.8 0.02
1 - 210 229850 0.11
2 - 569 192840 0.13
3 - 193 481700 0.05
4 - 320 233050 0.11
Avg. - 323.0 284360.0 0.10
Stainless
hollow
Std. Dev. 173.4 132819.5 0.03
1 - 476 - -
2 - 582 - -
3 - 464 - -
4 - 811 - -
Avg. - 583.3 - -
Purple
epoxy
coating
( non-flexible)
Std. Dev. 160.8 - -
1 - 630 - -
2 - 405 - -
3 - 648 - -
4 - 597 - -
Avg. - 570.0 - -
Gray
epoxy
coating
( non-flexible)
Std. Dev. 112.0 - -
1 - 420 - -
2 - 760 - -
3 - 558 - -
4 - 643 - -
Avg. - 595.3 - -
Green
epoxy
coating
( flexible)
Std. Dev. 143.2 - -
33
3897.45
134667.75
0
50000
100000
150000
200000
Carbon Steel Microcomposite Steel Stainless Steel Clad Stainless Steel Hollow
Dowel Type
Linear Polarization Resistance
( a)
0.6
0.1 0.1
0.0000
1.0000
2.0000
3.0000
4.0000
Carbon Steel Microcomposite Steel Stainless Steel Clad Stainless Steel Hollow
Dowel Type
Current Density ( icorr)
( b)
Figure 35. Comparative column plots showing average results of polarization resistance ( Rp) and
corrosion current density ( icorr) for non- coated dowels.
284685.0 284360.0
9.01
34
With respect to the average corrosion current densities, one can observe that the carbon steel dowels
exhibit very rapid corrosion while microcomposite steel exhibits a moderate level of corrosion. Stainless steel-clad
and stainless steel hollow dowels proceed at a low corrosion rate ( refer to Table 1 and Table 2 for criteria
used). Individual results show that two out of four microcomposite samples were in the “ uncertain” corrosion
region ( as per half- cell potential results), and presented low corrosion rate.
The average values presented for the epoxy- coated bars also indicate a low corrosion rate and, in fact,
visual inspection ( Section 3.2.3) revealed no major signs of corrosion for these dowels, except at a few
localized defective areas. However, LPR results for epoxy- coated bars cannot be interpreted quantitatively. As
is further discussed in Section 3.2.4 ( microstructural analysis), localized corrosion was identified at defects in
the coating. Additional corrosion was also observed under the coating.
3.2.2 Statistical Analysis of Results
The Analysis of Variance ( ANOVA) model was used to evaluate the results obtained and to verify if
there is a statistically significant effect of the type of steel dowel on the parameters investigated. Basically, the
Analysis of Variance table takes into account the effect of the factor( s) being studied and the variability of the
experiments ( i. e., the experimental error) to determine whether the results obtained are statistically significant
at a confidence interval of 95 percent. The analysis is based on the so- called “ F- test,” in which the calculated
value of the parameter F (“ Fcalc”) is compared to the critical value “ Fcrit”. If Fcalc > Fcrit., it is said that the
effect is statistically significant.
In the case of the research presented herein, the factor or independent variable being analyzed is the
type of steel dowel, and the dependent variable is the experimental result obtained. Separated analyses were
performed in order to evaluate the half- cell potential, Linear Polarization Resistance, and corrosion current
density results as functions of the type of dowel. Table 5 through Table 8 present the ANOVA results obtained.
In these tables, SS means Sum of Squares, df is degrees of freedoms, and MS is the Mean Square. By
comparing the Fcalc and Fcrit values obtained, one can observe that in all cases the type of steel dowel had a
statistically significant effect ( Fcalc > Fcrit) on the parameters studied.
Table 5. Analysis of Variance ( ANOVA) of Half- Cell Potential Results of Non- Coated Samples
Material Count Sum Average Variance Std. Dev.
Carbon steel 4 - 2565 - 641.3 1491.58 38.62
Microcomposite steel 4 - 1708 - 427.0 26920367 164.08
Stainless steel clad 4 - 1026 - 256.5 5449.67 73.82
Stainless steel hollow 4 - 1292 - 323.0 30064.67 173.39
ANOVA
Source of Variation SS df MS Fcalc Fcrit
Type of dowel 339527 3 113175.7 7.08 3.49
Error 191780 12 15981.65
Total 531307 15
35
Table 6. Analysis of Variance ( ANOVA) of Half- Cell Potential Results of All Samples, Including Epoxy-
Coated Bars
Material Count Sum Average Variance Std. Dev.
Carbon steel 4 - 2565 - 641.3 1491.6 38.6
Microcomposite steel 4 - 1708 - 427.0 26920.7 164.1
Stainless steel clad 4 - 1026 - 256.5 5449.7 73.8
Stainless steel hollow 4 - 1292 - 323.0 30064.7 173.4
Purple ( non- flexible) epoxy 4 - 2333 - 583.3 25864.9 160.8
Gray ( non- flexible) epoxy 4 - 2280 - 570.0 12546.0 112.0
Green ( flexible) epoxy 4 - 2381 - 595.3 20507.6 143.2
ANOVA
Source of Variation SS df MS Fcalc Fcrit
Type of dowel 541069 6 90178.14 5.14 2.57
Error 368535 21 17549.3
Total 909604 27
Table 7. Analysis of Variance ( ANOVA) of Polarization Resistance ( Rp) Results
Material Count Sum Average Variance Std. Dev.
Carbon steel 4 15589.8 3897.45 4766409 2183.2
Microcomposite steel 4 538671 134667.8 2.64E+ 10 162343.7
Stainless steel clad 4 1138740 284685 4.15E+ 09 64414.8
Stainless steel hollow 4 1137440 284360 1.76E+ 10 132819.5
ANOVA
Source of Variation SS df MS Fcalc Fcrit
Type of dowel 2E+ 11 3 7.32E+ 10 6.08 3.49
Error 1+ 11 12 1.2E+ 10
Total 4E+ 11 15
Table 8. Analysis of Variance ( ANOVA) of Corrosion Current Density ( icorr) Results
Material Count Sum Average Variance Std. Dev.
Carbon steel 4 36.051 9.0127 34.162 5.845
Microcomposite steel 4 2.509 0.6273 0.320 0.566
Stainless steel clad 4 0.380 0.0949 0.000 0.021
Stainless steel hollow 4 0.413 0.1034 0.001 0.035
ANOVA
Source of Variation SS df MS Fcalc Fcrit
Type of dowel 229.78 3 76.59247 8.88 3.49
Error 103.45 12 8.621028
Total 333.23 15
3.2.3 Visual Inspection
After the electrical experiments were finished, one beam of each dowel type was cored at the joint in
order to analyze the concentration of chlorides in the concrete surrounding the dowels. The remainder of the
36
specimens were cut open and broken to retrieve the dowels for visual inspection of the corroded areas. At this
moment, specimens were selected for the microstructural SEM ( Scanning Electron Microscope) analyses
presented in Section 3.2.3.
Figure 36 through Figure 42 illustrate the general appearance of the inspected dowels. In general,
carbon steel dowels exhibited generalized corrosion along the bars. Microcomposite steel dowels showed
significantly less corrosion ( as compared to carbon steel specimens). Stainless clad dowels had no visible
corroded areas along the bars ( corrosion was verified at the ends between the carbon steel core and outer
stainless steel layer). Stainless hollow dowels showed basically no visible corrosion. Localized corrosion was
verified in the defective regions on the epoxy- coated dowels, but no major signs of corrosion could be
observed visually.
( a) ( b)
Figure 36. Carbon steel dowels. ( a) Center region below joint; ( b) End with electrical connection.
( a) ( b)
Figure 37. Microcomposite steel dowels. ( a) General view of the center region below joint; ( b) End
region.
37
( a) ( b)
Figure 38. Stainless clad dowels. ( a) General view of the center region below joint; ( b) End region
showing corrosion between the carbon steel core and the stainless steel outer layer.
( a) ( b)
Figure 39. Stainless hollow dowels. ( a) General view of the center region below joint; ( b) End region
showing grouted core.
( a) ( b)
Figure 40. Epoxy- coated dowel ( purple). ( a) General view of the center region; ( b) Corrosion was
verified at the end region, underneath the epoxy seal.
38
( a) ( b)
Figure 41. Epoxy- coated dowel ( gray). ( a) General view of the center region; ( b) Corroded edges at the
end of dowels.
( a) ( b)
Figure 42. Epoxy- coated dowel ( green). ( a) General view of the center region; ( b) Corroded edges and
corrosion underneath coating at the end of dowels.
( a) ( b)
Figure 43. Corrosion product concentrated in a defective area in the epoxy coating ( a); part of the epoxy
coating was removed in order to observe the state of the carbon steel beneath the coating ( b).
39
As seen in Figure 36 through Figure 42, the carbon steel samples were heavily corroded. The
microcomposite steel dowels exhibited less corrosion. The stainless steel- clad and stainless steel hollow
dowels showed essentially no visible signs of corrosion. Defects were observed in the epoxy- coated bars and,
as expected, localized corrosion in the defective regions was verified.
3.2.4 Microstructural Analysis of Corroded Areas
The microstructural investigation was performed with a digital Leo 430 Scanning Electron
Microscope ( SEM) using secondary electrons. The SEM imaging consists of bombarding the sample with an
electron beam with sufficient energy to excite and eject secondary electrons, which are detected and recorded
using a CCD camera.
In this study, samples of each corroded dowel type were collected during the visual inspection and
then analyzed by SEM.
In the SEM microstructural analysis, preference was given to the inspection of the corroded areas in
each sample ( e. g., Figure 43a). In addition, selected regions such as the one illustrated in Figure 43b were also
investigated. Figure 44 through Figure 50 present characteristic SEM pictures for each material investigated.
In general, the severity of the corrosion observed with the electron microscope matches the results
anticipated by the electrical measurements and visual inspections. However, one should keep in mind that the
microstructural investigation focused mostly on the corroded areas of each sample, and does not represent the
general state of the dowel bars. The stainless steel- clad dowels, for example, exhibited excellent performance
on the corrosion rate measurements, and no corrosion was identified by visual inspection; however, relatively
severe localized corrosion was observed at the ends of the dowels close to the interface with carbon steel ( see
Figure 38b and Figure 47).
A similar situation was observed with the analysis of the epoxy- coated samples. It has been observed
that the electrical measurements do not provide a reliable quantitative estimate of the corrosion state of such
specimens, and in the visual inspection it was found that the bars were still in relatively good condition,
especially compared to carbon steel dowels. However, several defects on the epoxy coating have been
presented in this section, and the relation between defects and the occurrence of localized corrosion has been
explored.
40
( a) ( b)
( c) ( d)
Figure 44. Carbon steel samples: ( a) heavy corrosion along the dowel surface, magnification = 100×; ( b)
same region, 200×; ( c) different region at the surface, 200×; and ( d) corrosion at the bar end, 100×.
41
( a) ( b)
( c) ( d)
Figure 45. Microcomposite steel samples: ( a) aspect of corrosion along at the surface, 100×; ( b) same
region, 205×; ( c) and ( d) details of characteristic corrosion sites, 205×.
42
( a) ( b)
( c)
Figure 46. Hollow stainless steel samples: ( a) view at the surface, 100×; ( b) same region, 300×; and ( c)
detail of the surface condition, 1,000×. The surface appears rough, but no signs of corrosion damage
were observed.
43
( a) ( b)
( c) ( d)
Figure 47. Stainless steel- clad samples: ( a) aspect of corrosion along the surface, in a region close to the
end of the dowel, 100×; ( b) zoom around same region, 200×; ( c) details of corrosion at the surface,
interface between sound and corroded area, 2,390×; and ( d) detail of corrosion at an edge, 2,320×.
44
( a) ( b)
( c) ( d)
Figure 48. Gray epoxy- coated samples: ( a) view of a defect in the coating, with corroded area inside,
50×; ( b) another region, at an edge, where coating is lifted 75×; ( c) general view of the surface, 100×; and
( d) detail of a pinhole present in image ( c) ( see arrow), 1,500×.
45
( a) ( b)
( c) ( d)
Figure 49. Green epoxy- coated samples: ( a) general view of the surface, 100×; ( b) region where part of
the epoxy coating was removed, 100×; ( c) condition of the steel underneath the epoxy, 302×; and ( d)
detail of corroded area and pits present under the coating, 1,330×.
46
( a) ( b)
( c)
Figure 50. Purple epoxy- coated samples: ( a) general view of the surface with corrosion products
accumulated in a holiday, 100×; ( b) zoom in the same region, 200×; and ( c) condition of the steel
underneath the epoxy, 200×.
3.3 Phase III Results
Six cores from the WSDOT slabs extracted and transported to California were also tested in the
Caltrans laboratory. Ten cores extracted from the corners of in- service pavements ( two from each of five
locations in Washington State) were submitted to Construction Technology Laboratories, Inc. for testing as
well. The chloride contents of the concrete at different depths of the cores were compared between the
laboratory specimens and the field slabs to compare the severity of the laboratory chloride environment to that
of the field slabs. This comparison helps to relate the results from the accelerated laboratory testing to the long-term
field conditions.
3.3.1 Extracted WSDOT Pavement Slabs
This section presents the results and discussion of half- cell potential tests, LPR experiments, chloride
analyses, and visual inspection performed on the slabs from the WSDOT dowel bar retrofit project on
Interstate 90.
47
3.3.1.1 Half- Cell Potential and Linear Polarization Resistance
Figure 51 through Figure 53 present the results of LPR tests done on the extracted WSDOT slabs. In
these plots, the current density ( in μA/ cm2) is plotted versus potential ( in mV), and the X- axis crosses the Y-axis
at the half- cell potential value. The slope of the curves ( ΔE/ Δi) provides the value of the polarization
resistance ( Rp), with flatter slopes indicating faster corrosion rate. The location of specimens A, B, and C are
illustrated in Figure 54.
Although the polarization resistance appears to be large when estimated by the Linear Polarization
Resistance method ( which would lead to a small corrosion current density), it is important to highlight that
accelerated, localized corrosion may occur at defects and/ or underneath the epoxy coating. One has to keep in
mind that this corrosion detection technique cannot be interpreted quantitatively for epoxy- coated rebar, since
in this case corrosion is most likely localized at the coating holidays ( pinholes, voids, etc.). Given that epoxy is
an electrical insulator, polarization occurs only at very small areas ( defective areas) that cannot be accounted
for in the calculation of the polarization resistance term.
y = 920.11x - 651.2
R2 = 0.9728
- 460
- 450
- 440
- 430
- 420
- 410
- 400
0.21 0.22 0.23 0.24 0.25 0.26 0.27
Current Density ( μA/ cm2)
Potential ( mV)
Figure 51. Linear Polarization Resistance, WSDOT slab, Specimen A.
48
y = 570.17x - 450.75
R2 = 0.9818
- 430
- 420
- 410
- 400
- 390
- 380
- 370
0.04 0.05 0.06 0.07 0.08 0.09 0.10 0.11 0.12 0.13
Current Density ( μA/ cm2)
Potential ( mV)
Figure 52. Linear Polarization Resistance, WSDOT slab, Specimen B.
y = 738x - 430.4
R2 = 0.9908
- 420
- 410
- 400
- 390
- 380
- 370
- 360
0.02 0.03 0.04 0.05 0.06 0.07 0.08
Current Density ( μA/ cm2)
Potential ( mV)
Figure 53. Linear Polarization Resistance, WSDOT slab, Specimen C.
49
Figure 54. Samples extracted from the WSDOT pavement slab for chloride analyses.
Regarding the half- cell potential results, one can observe that the different dowels presented similar
values ( respectively - 430mV, - 400mV, and - 392mV), being all located in the high probability of corrosion
region ( Table 1).
3.3.1.2 Visual Inspection
Figure 55 shows the interior of a core hole located at the joint immediately after coring. The top of the
corroded dowel can be observed at the bottom of the hole. After scraping away some of the loose epoxy
coating, a considerable amount of corrosion product was found. The core was removed using water for cooling
the core bit, however, no time was permitted for corrosion to occur because of the coring water.
The coring also revealed that one of the retrofit dowel bars was not centered over the transverse joint,
and only extended approximately 100 mm into the slab on one side of the joint. It is certain that this
misplacement of the dowel contributed to the low load transfer efficiency ( LTE) of the joint. It is likely that the
observed corrosion, and presence of the weak layer of corrosion product between the concrete and the dowel
also contributed to the loss of LTE.
3.4 Analysis of Chloride Contents
As mentioned in Section 2.6, several chloride analyses were performed on concrete cores cut from: ( 1)
extracted WSDOT pavement slabs, ( 2) near- joint corners of in- service slabs at five different locations in
Washington State, and ( 3) the laboratory beams from Phase II testing exposed to 3.5% NaCl solution. Such
chloride profiles are used to evaluate the concentration of chlorides in the region surrounding the steel dowels,
particularly in the vicinities of the open joints, and to compare field and laboratory exposure conditions.
3.4.1 Chloride Analyses from WSDOT Slab Extracted from Interstate 90
Six samples were obtained from the extracted WSDOT slabs from Interstate 90 for chloride analyses.
Cores 1 and 2 were made through the concrete only; cores 3 and 4 were cut through the retrofit grout and the
end of the dowel, and cores 5 and 6 were obtained at the joint through grout and the dowel. The samples were
submitted to the Caltrans chemistry laboratory for the analyses to be performed.
Core 4 ( grout) could not be completely removed on the same day due to problems with the coring
machine, and was left out of the analyses. Results for Core 3 ( Figure 58) showed that the concentration of
chlorides in this region was low, and attention was focused on the study of chloride concentrations in the
jointed region.
50
Figure 55. View of the interior of a recently made core hole showing
corrosion products underneath the epoxy coating.
Figure 56 through Figure 60 show the chloride profiles obtained for this pavement slab. The pavement
was built in 1964, and deicing salts were applied 120– 150 days per year, with each icy day typically having at
least two applications. Details of location, salt usage, and joint sealing practices are presented in Table 9 ( SR-
90 MP 76.88, extracted slab).
The steel dowels are located approximately at 3 in. depth. It can be noticed that the chloride
concentration at this depth is around 1.0 lb./ cu. yd., which is near the chloride threshold for regular carbon
steel found in the literature ( 10). Notice that sometimes a smaller concentration may be verified at the top layer
due to Cl- being “ washed out” by surface water.
Figure 58 shows a profile of the chloride concentrations on the grout used for the dowel- bar retrofit,
which was done in 1994. Even approximately 10– 11 years after the project, one can observe that the
concentration of Cl- ions is considerably below the minimum value to initiate corrosion, even for regular
unprotected carbon steels. However, in Figure 59 and Figure 60 it can be seen that the chloride concentrations
close to the joint are significantly higher, reaching values as high as 11.38 lb./ cu. yd. along the profile and 8.43
lb./ cu. yd. in a region close to the dowel ( Figure 60).
The results presented in Figure 56 through Figure 60 indicate that the chloride ions can more easily
penetrate the grout through the open joint. Thus, at the joint, easier access and accumulation of chlorides
results in higher localized concentrations close to the dowels. It is important to notice that in this case the
chloride ions do not infiltrate the concrete ( or grout) from the top, but instead they migrate through the open
joint to the dowel.
51
0 2 4 6 8 10 12 14 16 18 20
0.16
0.31
0.47
0.63
0.79
0.94
1.10
1.26
1.42
1.57
1.73
1.89
2.05
2.20
2.36
2.76
Depth ( in.)
[ Cl-], lb./ cu. yd.
Figure 56. Chloride profile of the slab concrete ( Core # 1).
0 2 4 6 8 10 12 14 16 18 20
0.16
0.31
0.47
0.63
0.79
0.94
1.10
1.26
1.42
1.57
1.73
1.89
2.05
2.20
2.36
2.52
2.68
2.95
Depth ( in.)
[ Cl-], lb./ cu. yd.
Figure 57. Chloride profile of the slab concrete ( Core # 2).
52
0 2 4 6 8 10 12 14 16 18 20
0.16
0.31
0.47
0.63
0.79
0.94
1.10
1.26
1.42
1.57
1.73
1.89
2.05
2.20
2.76
Depth ( in.)
[ Cl-], lb./ cu. yd.
Figure 58. Chloride profile of the grout used for dowel bar retrofit ( Core # 3).
0 2 4 6 8 10 12 14 16 18 20
0.16
0.31
0.47
0.63
0.79
0.94
1.10
1.26
1.42
1.57
1.73
2.76
Depth ( in.)
[ Cl-], lb./ cu. yd.
Figure 59. Chloride profile of the grout around the pavement joint ( Core # 5).
53
0 2 4 6 8 10 12 14 16 18 20
0.16
0.31
0.47
0.63
0.79
0.94
1.10
1.26
1.42
1.57
1.73
1.89
2.05
2.20
2.76
Depth ( in.)
[ Cl-], lb./ cu. yd.
Figure 60. Chloride profile of the grout around the pavement joint ( Core # 6).
3.4.2 Chloride Analyses of Cores from WSDOT Pavements in Various Locations
This section presents the results of the chloride analyses performed on cores extracted from slab
corners, next to the transverse joint of in- service slabs at five different locations in Washington State. These
cores were extracted by WSDOT and tested for chloride content by Construction Technology Laboratories,
Inc. for comparison with the laboratory results. Table 9 describes the locations from which cores were
extracted, including years of construction and salt usage and joint- sealing practices. Figure 61 through Figure
65 present the chloride concentration profiles for the different locations described in Table 9. The error bars
represent minimum and maximum values obtained at each location.
It is interesting to notice that the concentration of 1.0 lb./ cu. yd., regarded as the chloride threshold for
carbon steel, is exceeded in nearly all cases, except for the samples from Wapato
Table 9. Samples Extracted from In- Service Pavements for Chloride Analyses ( Information Provided by WSDOT)
Route
Number
Location
( MP) Year Depth
( in) Salt Usage Joint Sealing
I- 90 76.88
( extracted
slab)
1964 9 Up until the early 1980’ s, rock salt was used
approximately 120- 150 days per year at a typical rate
of 180 pounds per mile using a 5: 1 ratio ( 1 scoop of
salt to 5 scoops of sand). Each day typically had at
least two applications. Salting typically occurs from
mid November to mid April. From the early 1980’ s
to 1997 or so, rust inhibitor type deicers/ salts have
been used. Since 1997 or so liquid magnesium
chloride has been used at a typical rate of 35 gallons
per mile with a range of 15 to 50. The average rate
seems to be coming down with the use of better flow
controls.
Sawed and sealed during initial construction with a
rubberized joint sealer product called “ Seal Target
164.” Joint sealed during 1994 DBR with rubberized
joint sealant.
I- 5 166.00
( Downtown
Seattle)
1967 9 No data yet. Sawed and sealed during initial construction with a
rubberized joint sealer product called “ Seal Target
164.” No sealing since original construction.
I- 90 91.324 ( Elk
Heights)
1967 9 See I- 90 above. Same. Sawed and sealed during initial construction with a
rubberized joint sealer product called “ Seal Target
164.” Joint sealed during 1994 DBR with rubberized
joint sealant
I- 90 61.304
( Price
Creek)
1959 9 See I- 90 above. Same. Sawed and sealed during initial construction with a
rubberized joint sealer product called “ Seal Target
164.” Joint sealed during 1997 DBR with rubberized
joint sealant.
SR 82 10.617
( Military
Road)
1971 9 Same as I- 90 above except that the use of rock salts
was about 100 days per year. Most days had two
applications. Rust inhibitor deicers/ salts were used
from the early 1980’ s to 1997 or so. Since 1997
liquid magnesium chloride has been used at the
typical rates reported above.
Sawed and sealed during initial construction with a
rubberized joint sealer product called “ Seal Target
164.” Joint sealed during 1997 DBR with rubberized
joint sealant.
SR 82 43.524
( Wapato)
1981 9 Same as I- 90 above, however the applications of
rock salt were about 60- 70 days per year until about
1985. Each day often had two applications. Rock salt
was used at a 15: 1 to a 20: 1 blend. From 1985 to
about 2000 rust inhibitor salts were used at the same
rate. Since 2000, magnesium chloride has been used.
Sawed and sealed during initial construction with a
rubberized joint sealer product called “ Seal Target
164.” No sealing since original construction.
55
0 - ¾ "
¾ " - 1 ½ "
1 ½ " - 2 ¼ "
2 ¼ " - 3"
3" - 3 ¾ "
3 ¾ " - 4 ½ "
0 2 4 6 8 10 12 14 16 18 20
Depth ( in.)
[ Cl-], lb./ cu. yd.
Figure 61. Chloride profile for I- 5 MP 166.00 ( downtown Seattle).
0 - ¾ "
¾ " - 1 ½ "
1 ½ " - 2 ¼ "
2 ¼ " - 3"
3" - 3 ¾ "
3 ¾ " - 4 ½ "
0 2 4 6 8 10 12 14 16 18 20
Depth ( in.)
[ Cl-], lb./ cu. yd.
Figure 62. Chloride profile for I- 90 MP 91.324 ( Elk Heights).
56
0 - ¾ "
¾ " - 1 ½ "
1 ½ " - 2 ¼ "
2 ¼ " - 3"
3" - 3 ¾ "
3 ¾ " - 4 ½ "
0 2 4 6 8 10 12 14 16 18 20
Depth ( in.)
[ Cl-], lb./ cu. yd.
Figure 63. Chloride profile for I- 90 MP 61.304 ( Price Creek).
0 - ¾ "
¾ " - 1 ½ "
1 ½ " - 2 ¼ "
2 ¼ " - 3"
3" - 3 ¾ "
3 ¾ " - 4 ½ "
0 2 4 6 8 10 12 14 16 18 20
Depth ( in.)
[ Cl-], lb./ cu. yd.
Figure 64. Chloride profile for SR 82 MP 10.617 ( Military Road).
57
0 - ¾ "
¾ " - 1 ½ "
1 ½ " - 2 ¼ "
2 ¼ " - 3"
3" - 3 ¾ "
3 ¾ " - 4 ½ "
0 2 4 6 8 10 12 14 16 18 20
Depth ( in.)
[ Cl-], lb./ cu. yd.
Figure 65. Chloride profile for SR 82 MP 43.524 ( Wapato).
( Figure 65), which is a more recently built pavement where deicing salts were applied only
60- 70 days/ year at lower rates. Even so, at this location [ Cl-] reached 0.6– 0.7 lb./ cu. yd. On the other hand,
only the older pavement built in 1959 ( Figure 63), which was exposed to two daily applications of deicing salts
for 120– 150 days/ year, reached a concentration that was high enough to overcome the chloride threshold of 7.7
lb/ cu. yd. found in the references for the microcomposite steel ( 8- 10).
3.4.3 Chloride Analyses from Laboratory Samples
As described in Section 2.6, cylindrical cores were extracted from the laboratory concrete beams at the
central, jointed region. The laboratory specimens were subjected to wet- and- dry cycles with a 3.5% NaCl
solution, which is typically used to simulate seawater, over an 18- month period. The extracted samples were
shipped to the Caltrans chemistry laboratory where the chloride analyses were performed. Figure 66 through
Figure 71 present the results from these analyses.
In general, it can be seen that the chloride concentrations are much more constant along the depth
profile compared to situations in which the chloride has to diffuse through the concrete ( e. g., Figure 57 and
Figure 58) or migrate through a narrow joint ( e. g., Figure 60 and Figure 63). This is due to the fact that in the
laboratory specimens, a large ( approximately 6 × 50 mm) open joint was cast above the dowels.
In addition, it was verified that the use of a 3.5% NaCl solution may lead to high chloride
concentrations, as compared to the results obtained in the field specimens, greatly accelerating the corrosion
process. As a result of this aggressive environment, corrosion could be observed in nearly all samples after
only 18 months of exposure, except for the stainless clad and stainless hollow specimens. Comparatively,
based on the previous results, it can certainly be said that the corrosion resistance of carbon steels is
considerably lower than that of microcomposite steels, which in turn are less corrosion- resistant than stainless
hollow and stainless steel- clad dowels. However, one should keep in mind that the chloride concentration
58
analyses indicate that the actual field conditions and local environment should be taken into account when
choosing the appropriate material for a given project.
0 4 8 12 16 20 24 28 32 36 40
0.16
0.31
0.47
0.63
0.79
0.94
1.10
1.26
1.42
1.57
1.73
1.89
2.05
2.20
2.36
Depth ( in.)
[ Cl-], lb./ cu. yd.
Figure 66. Chloride profile on the concrete of a carbon steel specimen.
59
0 4 8 12 16 20 24 28 32 36 40
0.16
0.31
0.47
0.63
0.79
0.94
1.10
1.26
1.42
1.57
1.73
1.89
2.05
2.20
2.36
2.52
2.68
2.83
Depth ( in.)
[ Cl-], lb./ cu. yd.
Figure 67. Chloride profile on the concrete of a stainless clad specimen.
0 4 8 12 16 20 24 28 32 36 40
0.16
0.31
0.47
0.63
0.79
0.94
1.10
1.26
1.42
1.57
1.73
1.89
2.05
2.20
Depth ( in.)
[ Cl-], lb./ cu yd.
Figure 68. Chloride profile on the concrete of a stainless hollow specimen.
60
0 4 8 12 16 20 24 28 32 36 40
0.16
0.31
0.47
0.63
0.79
0.94
1.10
1.26
1.42
1.57
1.73
1.89
2.05
Depth ( in.)
[ Cl-], lb./ cu yd.
Figure 69. Chloride profile on the concrete of a flexible ( green) epoxy- coated specimen.
0 4 8 12 16 20 24 28 32 36 40
0.16
0.31
0.47
0.63
0.79
0.94
1.10
1.26
1.42
1.57
1.73
1.89
2.05
2.20
2.36
2.52
Depth ( in.)
[ Cl-], lb./ cu. yd.
Figure 70. Chloride profile on the concrete of a non- flexible ( purple) epoxy- coated specimen.
61
0 4 8 12 16 20 24 28 32 36 40
0.16
0.31
0.47
0.63
0.79
0.94
1.10
1.26
1.42
1.57
1.73
1.89
2.05
2.20
2.36
2.52
Depth ( in.)
[ Cl-], lb./ cu. yd.
Figure 71. Chloride profile on the concrete of a microcomposite steel specimen.
62
4 CONCLUSIONS AND RECOMMENDATIONS
This study investigated through the use of accelerated laboratory tests the corrosion performance of
several types of steel dowels cast in concrete beams. The accelerated testing consisted of ponding water with a
chloride solution on the top of the beams and including a “ joint” in each beam that provided direct access to a
short segment of the dowels. No force loading was applied to the beams. Results are presented from the tests.
Conclusions are based on the results of half- cell potential tests, LPR curves, visual inspections, and
microscopic investigations by Scanning Electron Microscope ( SEM), all of which are presented in this report.
4.1 Phase I Testing Conclusions
The following conclusions are from the Phase I testing consisting of the placement of four types of
dowels ( carbon steel, epoxy- coated steel, stainless steel hollow, and stainless steel- clad) cast in concrete beams
with joints in different corrosive environments, and evaluation using half- cell potential tests for corrosion
initiation and visual inspection at the completion of testing.
1. Carbon steel dowels present the shortest corrosio