evaluation of corrosion protection for internal

Evaluation of Corrosion Protectionfor Internal Prestressing Tendonsin Precast Segmental Bridges

Jeffrey S. West, Ph.D., P.E.Assistant Professor

Department of Civil EngineeringUniversity of Waterloo

Waterloo, Ontario, Canada

John E. Breen, Ph.D., P.E.Nasser I. Al-Rashid Chairin Civil EngineeringThe University of Texas at AustinAustin, Texas

René P. Vignos, P.E., S.E.Associate

Forell/Elsesser Engineers, Inc.San Francisco, California

A research program utilizing modified macrocellcorrosion specimens was conducted to investigatecorrosion protection for internal tendons insegmental bridges. Test variables were segmentaljoint type, duct type, joint precompression andgrout type. Specimens were subjected to ‘2 yearsof exposure testing, after which selectedspecimens were removed for destructiveexamination. Dry joint specimens performed verypoorly, as evidenced by corrosion currentsmeasured during exposure testing and by tendonand duct corrosion observed upon destructiveexamination. Epoxy joints limited chloridepenetration, preventing tendon corrosion andreducing duct corrosion. Plastic post-tensioningducts performed very well, limiting strandcorrosion to negligible levels. The researchindicates that epoxy joints are required for theprotection of internal tendons in aggressiveenvironments, and that plastic post-tensioningducts provide a significant improvement in tendoncorrosion protection. Many of the conclusions andrecommendations presented are not only specificto segmental construction but are applicable to allforms of internal, grouted post-tensioning tendons.

precast concrete segmental bridge construction in NorthAmerica normally consists of match-cast box girdersegments post-tensioned for continuity. Post-tension

ing may be internal bonded tendons, external tendons, or acombination of both. Current specifications1require the useof match-cast epoxy joints with internal tendons.

Epoxy joints were introduced to enhance force transferacross the segmental joint and to seal the joint against

76 PCI JOURNAL

moisture entry. More recently, epoxyjoints have been recognized as an absolute requirement for durability wheninternal tendons are used. The post-tensioning ducts for internal tendonsare not continuous across the segmental joints in North American practice,and no special coupling of tendonducts is made with match-cast joints.

Corrosion protection for bonded internal tendons in precast segmentalconstruction can be very good. Withinthe segment, internal tendons are wellprotected by high quality concrete,duct and cement grout. General information on corrosion protection forprecast, prestressed concrete is provided in a number of sources, including References 2, 3, 4 and 5. The potential weak link in corrosionprotection for internal tendons in segmental construction is at the joint between segments, where the joint represents a preformed crack at the locationof a discontinuity in the duct.

In saltwater exposures or in areaswhere deicing salts are used, the jointcould possibly allow moisture andchlorides to reach the tendon andcause corrosion. Since the tendonsprovide structural continuity, failure ofa tendon due to corrosion could lead toserious distress or collapse of thebridge.

The overall performance of precastsegmental bridges in North Americahas been very favorable,6 and therehave been no reported cases of corrosion of internal tendons in precast segmental construction resulting frommoisture penetration at epoxy joints.However, it has been argued that thelack of duct continuity at the segmental joint leaves the potential for moisture and chlorides to reach the prestressing tendon.

Recently, a number of post-tensioning tendon corrosion problems havecome to light in North America, withthe most notable occurring in severalFlorida bridges. While much of the information related to these tendon corrosion problems has not been formallypublished, some documentation isavailable directly from the Florida Department of Transportation. In addition, some discussion of the grout-related aspects of the corrosionproblems is provided by Ronald.7

Fig. 1. Macrocell specimen details.

The recently discovered tendon corrosion problems in Florida and elsewhere involved a number of differentforms of post-tensioned bridge construction. However, it is important tonote that none of the problems wererelated to corrosion of internal tendonsin precast segmental construction.

The tendon corrosion problems inFlorida highlight the importance of attention to corrosion protection forgrouted post-tensioning systems. Theobjective of the research describedherein was to use laboratory corrosiontests to evaluate the potential for corrosion of bonded internal post-tensioning tendons with details typical ofNorth American precast segmentalconstruction.

This article provides a brief description of the test specimens and van-

ables for the research study. Test datafrom 4’2 years of exposure testing arepresented and discussed. One-half ofthe macrocell corrosion specimenswere subjected to a complete destructive examination after the 41/2 yeartesting period. A detailed descriptionof the autopsy process and findings isprovided. Conclusions and recommendations suitable for implementationare presented based on the exposuretesting and destructive examination.

TEST PROGRAM

The test program enlisted 38 modifled macrocell corrosion specimens in19 pairs. Four categories of variableswere selected to evaluate typical details and protection measures in segmental construction.

3% NaCI solution

plexiglass

R

(when applicable)

in. dia.)

2- #4

match castsegmental joint

Longitudinal Section

0.25 in. end cover

—Duct:1.18 in. O.D. steel1.3 in. O.D. plastic

un. -.l

foam gasket:1/4 thick &1/4 wide

ductopening

match-castsegment face

End View

Note: lin.25.4mm

Section Through EDoxyJoint with Gasket

September-October 2002 77

Table 1. Specimen designation. Macrocell Corrosion Specimens

Test specimens were based on thestandard macrocell described in theAmerican Society for Testing and Materials (ASTM) Standard G109.8 Thejoint and tendon details investigatedrepresent modern precast segmentalbridge construction in North America.The modified macrocell specimenconfiguration is shown in Fig. 1.

Each specimen consists of twomatch-cast segments, with continuitybetween the segments provided by ain. (12.7 mm) diameter, seven-wireprestressing strand inside a groutedduct. The duct is not continuous acrossthe joint, providing details typical ofbonded internal tendons in precast segmental bridge construction. Due to thesmall size of the specimens, the tendoncould not be prestressed effectively,and, as a result, stress corrosion wasnot considered in this research.

The pairs of match-cast segmentsare stressed together using externalloading frames to simulate precompression across the joint due to post-tensioning. Two No. 4 (12.7 mm) mildsteel reinforcing bars are used in thebottom of the specimen to representsegment reinforcement. The mild steelbars do not cross the transverse joint,which is consistent with precast segmental construction.

Following ASTM G109, the exposed length of the prestressing strandand mild steel bars was limited to 5 in.(125 mm) using epoxy paint. Electrical contact between the two layers of

0.030 steel, necessary for macrocell corro

002

__________

sion, is achieved by wiring the pro-•truding ends of the steel together as

0.020

____________

shown in Fig. 1. A resistor is placed in0.015 the wire connection to allow assess0.010 ment of the corrosion current by mea

0005 suring the voltage drop across the resistor (Ohm’s Law).

0.000

___

Exposure conditions consist of a-0.005 two-week cycle of two weeks dry and-0.010 two weeks wet. During the wet period

of the cycle, the top surface of the specimen is ponded with a salt solution.

-0.020

-0.025Variables

.0.030800 1000 1200 1400 1800 1800 The variables selected for investigaTIme (days) tion in this program cover four com

ponents of the precast concrete segmental bridge related to corrosion of

DJ-S-L-NG

Joint Type 4 Grout TypeDuct Type 4 Joint Precompression

Joint Type:DJ =

SE =

EG =

Dry JointStandard EpoxyEpoxy with Gasket

Duct Type:S SteelP = Plastic

Joint Precompression:L = Low: 5 psi (35 kPa)M = Medium: 50 psi (345 kPa)H = High: 3,jipsi (190 psi, 1310 kPa)

Grout Type:NG= Normal GroutSF = Silica Fume AddedCI = Corrosion Inhibitor

4)

C-)C

I0 200 400 600 800 1000

TIme (days)

1200 1400 1600 1800

Fig. 2. Macrocell corrosion current: dry joint, steel duct, normal grout.

— DJ-S-L-CI-1—DJ-S-M.CI-1

I

0 200 400 600

Fig. 3. Macrocell corrosion current: dry joint, steel duct, corrosion inhibitor in grout.

78 PCI JOURNAL

internal tendons: joint type, duct type,joint precompression and grout type.

Joint type refers to the type ofpreparation or bonding agent used atthe match-cast segmental joint. Variables investigated were dry joint (nobonding agent), epoxy joint and epoxyjoint with a gasket around the ductconnection. Dry joints with internaltendons are prohibited by the guidespecifications for segmental bridges,’and were included only as a worst caseexample for comparison purposes.

The epoxy jointed specimens wereassembled according to standard practice. Both match-cast faces were coatedwith epoxy and the segments werepushed together. The joint was precompressed at 50 psi (345 kPa) for 48hours, after which the specimens wereunloaded and reloaded to the desiredlevel of precompression (see below).

In the epoxy/gasket joint, a foamgasket was glued to the face of onesegment around the duct opening, asshown in Fig. 1, prior to application ofthe epoxy. The primary purpose of thegasket is to prevent epoxy from entering the duct during segment placementand initial stressing. In the epoxy jointwithout a gasket, the duct wasswabbed out immediately after stressing to 50 psi (345 kPa) to prevent theepoxy from blocking the duct. Bothtechniques (gasket and no gasket) areused in current practice.

Duct types investigated were galvanized steel duct and plastic duct.Polyvinyl-chloride (PVC) pipe wasused for the plastic duct due to sizelimitations. Although PVC pipe wasused in the testing program, it is generally not permitted for use in concretestructures due to the potential forlong-term breakdown of the PVC andassociated release of chlorides overthe service life of the structure.

Joint precompression refers to thelevel of prestress provided by the internal and/or external tendons in thebridge. Three levels of precompression were selected; 5, 50 and 190 psi(35, 345 and 1310 kPa). The lowestlevel of 5 psi (35 kPa) could representthe level of precompression encountered in a precast segmental columnunder self-weight. The precompression of 50 psi (345 kPa) is based onthe AASHTO Guide Specifications.’

DJ-P.LNG-1— DJ-P-L-NG.2--. DJ-P-M.NG-1

-DJ-P-M.NG-2 -

o 200 400 600 800 1000 1200 1400 1600 1800

Time (days)

o 200 400 600 800 1000 1200 1400 1600 1800

Tme (days)

Fig. 5. Macrocell corrosion current: standard epoxy joint, steel duct, normal grout.

The highest precompression value of190 psi (1310 kPa) corresponds to3J psi (7.88f kPa).

Three cement grout types were selected for evaluation: normal grout[plain cement grout, water-cementratio (wic) of 0.40], grout with silicafume [13 percent cement replacementby weight, water-cementitious material ratio (wlcm) of 0.32, superplasticizer added] and grout with a calciumnitrite corrosion inhibitor [water-cement ratio (w/c) of 0.40]. Details ofthe grout mix proportions are providedby West et al.9

0.025

0.020

0.015

0.010

0.005 -

0.000

.2 -0.00500

-0.0100

-0.015

.0.020

.0.025 --- --- ---- --

fl fm

Fig. 4. Macrocell corrosion current: dry joint, plastic duct, normal grout.

j-=-SE-SM.NG-21

.

w

C.,

0

eC.,

0.030

0.025

0.020

0.015

0.010

0.005

0.000

-0.005

-0.010

.0.015

.0.020

.0.025

O.030

A total of 19 specimen types wereselected to address all of the variables.Each specimen type was duplicatedfor a total of 38 specimens. The notation used in the specimen designationis shown in Table 1.

Measurements DuringExposure Testing

Two forms of regular measurementswere taken to evaluate corrosion activity. Macrocell corrosion current wasdetermined by measuring the voltagedrop across a known resistance in the


corrosion cell, as described previously. Regular half-cell potential measurements, following ASTM C876,1°were recorded over time to aid in thedetection of the onset of corrosion.

EXPERIMENTAL RESULTSThis article summarizes test data

gathered over a period of four years

and five months. Macrocell corrosioncurrent data indicated that 12 of the 38specimens experienced some amountof corrosion activity during this period. Macrocell corrosion current datafor these specimens are plotted inFigs. 2 through 5.

The “polarity” of the corrosion current determines the direction of electron flow, thus indicating which layer

of steel is corroding or experiencingmetal loss. Negative corrosion currents in Figs. 2 through 5 indicate thatthe mild steel reinforcing bars at thebottom of the specimen are activelycorroding while the prestressing strandis acting as the cathode.

Of the 12 corroding specimens, onlyseven showed continued corrosion activity after four years and five months.Eleven of the 12 specimens with signsof corrosion had dry segmental joints,illustrating the poor level of protectionprovided by dry joints. One specimenwith a match-cast epoxy joint showedan initiation of corrosion on the mildsteel reinforcement, as indicated by thenegative corrosion current in Fig. 5.

Time to Corrosion

The time to corrosion for each ofthe 12 specimens displaying corrosioninitiation is plotted in Fig. 6. The timeto corrosion ranged from 128 days fora dry joint specimen with galvanizedsteel duct and low joint precompression to 1330 days for the epoxy jointspecimen with corrosion activity.Note that time to corrosion values arefor the first detected corrosion activity, whether it occurred on the prestressing strand or the mild steel bars.

Fig. 7. Calculatedmetal loss for allspecimens.

I

Dry JointSteel DuctNormal Grout

Dry JointSteel DuctCorr. Inhib. Grout

Dry JointPVC DuctNormal Grout

Epoxy Joint

__________________________________

Steel DuctNormal Grout

0 200 400 600 800 1000 1200 1400

Time to Corrosion (Days)

JointPrecomp.:

D5 psi•50 psiI19O psi

Fig. 6. Time to corrosion for specimens with corrosion activity.

Joint Type: Duct Type: Joint Precompresslon: Grout Type:DJ = Dry Joint S = Steel L = 5 psi (35 kPa) NG = Normal GroutSE = Epoxy Joint P = Plastic M = 50 psi (345 kPa) SF = Silica Fume GroutEG = Epoxy/Gasket H = 190 psi (1310 kPa) Cl Corrosion Inhibitor

250 —

200

E150

U,0-J. 1004-a)

50

0Ioz z

; ‘ ‘ ;i

C? 04.4 Or o.-, -, 0 —0 0 0 o

Dry Joints

I Strand CorrodingD Bar Corroding

I

0.0088

0.0071

N0

0.0053U,0-j

0.00354-a,

0.0018

0.00000z‘I

-,

0

0000zz zC? LI? C? OrC? ILl Lii LII Lii LLII LU 0 0 0 00 LII 0

0 0 0

.I

C?0 0U uj U

Epoxy Joints Epoxy &Gasket

80 PCI JOURNAL

The times to corrosion in Fig. 6 donot appear to indicate any trends in theeffect of the variables. Conceptually, ahigher level of joint precompressionmight be expected to limit the entry ofmoisture and chlorides, providing bettercorrosion protection and longer times tocorrosion. This trend is not supportedby the data for the three levels of jointprecompression investigated. In addition, the duct type and grout type do notappear to affect the time to corrosionbased on the data to date.

Corrosion Severity

Calculation of the metal loss duringcorrosion allows a relative comparisonof corrosion severity between specimens. The amount of steel consumedby macrocell corrosion is directly related to the total number of electronsexchanged between the anode andcathode. The amount of metal loss canbe computed by numerically integrating the macrocell corrosion currentover the duration of exposure. Basedon corrosion science, one amp of corrosion current consumes 1.04 grams ofsteel (iron) per hour.’1

Computed values of metal loss forall specimens are shown in Fig. 7.ASTM G109 defines failure for themacrocell corrosion specimen as aweighted average corrosion current of10 mA (average calculated over theduration of testing). For an exposureduration of four years and fivemonths, this would correspond to ametal loss of 0.014 oz (400 mg).

The most severe corrosion occurredin specimens with dry joints, galvanized steel ducts and normal grout,having a calculated metal loss of lessthan 250 mg (0.0088 oz). In general,the calculated values of metal loss arewell below the failure limit of 0.014 oz(400 mg), suggesting corrosion activityto date is minor in almost all cases.

DESTRUCTIVEEXAMINATION

One specimen from each identicalpair was removed from testing for destructive examination or autopsy afterfour years and five months of exposure testing. The objectives of the destructive examination were to obtain a

Fig. 8. Saw cut and chloride sample locations.

visual evaluation of corrosion damageon the duct, strand and mild steel reinforcement, and to assess chloride ionpenetration at locations adjacent toand away from the segmental joint.

Procedure

The procedure for the destructiveexamination began with sampling ofthe concrete from selected specimensfor chloride analysis. This was followed by cutting of the specimens toallow careful extraction of the duct,tendon and mild steel bars. Specimencondition was documented at eachstage of the process.

Chloride Analysis — Concretepowder samples were collected to determine chloride ion profiles adjacentto the joint and away from the joint toexamine the influence of joint type onchloride penetration. Sampling locations are shown in Fig. 8. Concretepowder samples were collected using arotary hammer and following a procedure based on AASHTO T 260.12 Samples were analyzed for acid solublechlorides using a specific ion probe.

Grout samples were also collectedfrom selected specimens for chlorideanalysis. Samples were carefully removed from the strand at the locationof the joint and at a distance of 2 in.(51 mm) from the joint. The groutpieces were crushed and ground into

powder using a mortar and pestle.Grout powder samples were analyzedfor acid soluble chlorides using a specific ion probe.

Longitudinal Saw Cuts — Twolongitudinal saw cuts were made oneach side of the specimens to facilitateremoval of the duct/strand unit andmild steel bars, as shown in Fig. 8. Anadditional pair of longitudinal sawcuts was made at the midheight of selected epoxy joint specimens to evaluate the joint condition. Joint sectionswere examined for indications ofvoids in the epoxy, or the presence ofmoisture, salt, or corrosion products.

Autopsy Program

One specimen from each duplicatepair of specimen types was selectedfor destructive examination. Details ofthe 19 specimens selected for autopsyare listed in Table 2. This table alsoindicates the specimen condition at thetime of autopsy, and whether chloridesampling was performed.

Evaluation and Rating ofCorrosion Found DuringDestructive Examination

A generalized evaluation and ratingsystem was developed in this researchprogram to quantify the severity andextent of corrosion damage in the test

chloride samplelocations

saw cut lineat strand level

“saw cut lineat bar level

Note: lin.=25.4mm

Chloride Samole Locations:A - 2.0 in. from jointB - 0.5 in. from jointC - 4.0 in. from joint


Table 2. Specimens selected for forensic examination.

Specimen Time to corrosion Corrosion location Corrosion activity Chloride samples Midheight cutDJ-S-L-NG-1 128 days Strand Inactive A, B, C n/aDJ-S-M-NG-1 1110 days Strand Inactive A, B n/aDJ-S-H-NG-1 615 days Bars Active A, B n/aDJ-P-L-NG-1 1250 days Bars Active A, B n/aDJ-P-M-NG-1 565 days Bars Inactive None n/aDJ-S-L-CI-1 580 days Strand Active A, B n/aDJ-S-M-CI-1 835 clays Bars Inactive A, B n/aSE-S-L-NG-2 n/a n/a n/a A, B, C Yes

SE-S-M-NG-2 1330 days Bars Active A, B YesSE-S-H-NG-2 n/a n/a n/a A, B Yes

SE-P-L-NG-2 n/a n/a n/a None NoSE-P-M-NG-2 n/a n/a n/a None No

SE-S-L-CI-2 n/a n/a n/a None NoSE-S-M-CI-2 n/a n/a n/a None NoSE-S-H-CI-2 n/a n/a n/a None NoSE-S-L-SF-2 n/a n/a n/a None No

EG-S-L-NG-2 n/a n/a n/a I A, B 1 YesEG-S-M-NG-2 n/a n/a n/a None YesEG-S-H-NG-2 n/a n/a n/a None Yes

specimens. The procedure is presentedin a universal form with the intentionof applying the same rating system toother situations.

The rating procedure involved subdividing the length of strand, mildsteel reinforcement or galvanizedsteel duct into eight increments. Thesteel was examined at each increment, and a rating was assigned to describe the corrosion severity withinthat increment. The ratings for theeight increments were summed to

give a total corrosion rating for the element that could be compared for different specimens. By assigning a corrosion severity at eight locations,both the extent and severity of corrosion is considered.

The corrosion severity rating systemis essentially the same for prestressingstrand, mild steel reinforcement andgalvanized duct, with some modifications to reflect unique corrosion aspects of each type of steel. In general,the evaluation system doubles the

severity rating for each category of increasing corrosion damage. A detaileddescription of the rating system and itsapplication is provided by West et al.9

DESTRUCTIVEEXAMINATION RESULTS

Destructive examination of eachspecimen type revealed that corrosiondamage to prestressing strand and mildsteel reinforcement was not severe.Only one prestressing strand wasfound to have pitting corrosion, and nomild steel bars were found to havemeasurable area reduction.

Similar testing programs usingASTM G109 type macrocell corrosionspecimens normally report severe corrosion damage and specimen failure ina test duration of less than 41/2 years.This observation highlights the overallexcellent performance of the groutedpost-tensioning system in this testingprogram. A detailed summary of the destructive examination results is provided for each specimen in Reference 9.

Corrosion Ratings

The strand, bar and duct corrosionratings for all specimens are plotted inFigs. 9 through 11. A “Threshold ofConcern” was assigned in the figuresat a corrosion rating of 50 to indicatecorrosion related deterioration deemedsevere enough to warrant concern. Thethreshold of concern is useful to illus

125

100

C,

e 5°0

C)

25

0

Fig. 9. Strand corrosion ratings for all specimens.

82 PCI JOURNAL

trate that in most cases the observedcorrosion was negligible from a practical standpoint. In general, corrosionratings greater than 50 correspondedto pitting corrosion for strands andbars, and holes in the galvanized steelduct caused by corrosion.

Specimen DJ-S-L-CI-1 [dry joint,steel duct, 5 psi (35 kPa) precompression, corrosion inhibitor in grout] hadthe most severe strand corrosion, witha corrosion rating of 114 (see Fig. 9)compared to the average of 19.5 andmedian of 12. This was the only specimen with a strand corrosion ratinggreater than 50.

Specimen DJ-S-H-NG-l [dry joint,steel duct, 3J precompression (190psi, 1310 kPa), normal grout] had themost severe mild steel reinforcementcorrosion with a rating of 60 (see Fig.10) compared to the average of 9.1and median of 1. This was the onlyspecimen with a bar corrosion ratinggreater than 50.

Specimen DJ-S-L-NG-l [dry joint,steel duct, 5 psi (35 kPa) precompression, normal grout] had the worst ductcorrosion with a rating of 528 (seeFig. 11) compared to the average of122.9 and median of 79.

In each case, the specimen with thelargest corrosion rating was severaltimes higher than the average and median values. The average rating islarger than the median rating for allthree ratings. The difference is largestfor the mild steel bars, where the average is almost an order of magnitudelarger than the median. These trendsillustrate that the worst performancegenerally occurred in a limited numberof specimens.

Chloride Analysis Results

Concrete powder samples were collected from six dry joint specimensand four epoxy joint specimens forchloride analysis. In addition, sampleswere collected from the grout in thesespecimens for chloride analysis.

The chloride ion profiles in the concrete revealed distinct trends in chloride ion penetration in dry joint andepoxy joint specimens. In general, thedry joint specimens showed significantly higher chloride contents adjacent to the joint in comparison to mea

surements away from the joint. In theepoxy joint specimens, the chlorideprofiles were essentially the same nearand away from the joint.

Typical profiles for dry joints andepoxy joints are shown in Fig. 12 andFig. 13, respectively. Values plotted inthe figures are acid soluble chloridelevels, expressed as a percentage ofconcrete weight (weight of sample).

1

The chloride threshold for corrosion isindicated in the figures at 0.033 percent. This value is intended as a guideline only, and is based on the widelyaccepted chloride threshold value of0.2 percent of the weight of cement.13

In the dry joint specimens, the chloride contents were well above the corrosion threshold over the depth of thespecimen. In some cases, chloride con-

C)C

Fig. 10. Mild steel reinforcing bar corrosion ratings for all specimens.

125

100

a,C

75

5000

25

.,7R.s25 15114

—237—

8]of Concern 78

hEF. ••8 ••1012

•LLI II.LLIII04rCC

ZZZZZ . LZZZ

‘?‘?‘?4°r4!- - - *3a33

Epoxy &Dry Joints - Epoxy Joints Gasket

Fig. 11. Duct corrosion ratings for all specimens.


tents at 2 in. (51 mm) from the jointwere higher in the dry joint specimensin comparison to those with epoxyjoints. Samples collected at LocationC, 4 in. (102 mm) from the joint,showed negligible chloride levels inboth dry and epoxy joint specimens.

The chloride profile for SpecimenSE-S-M-NG-2 [epoxy joint, steel duct,50 psi (345 kPa) precompression, normal grout] displays a discontinuity inthe measurements adjacent to the joint,as shown in Fig. 14. Chloride measure-

ments decrease to zero by midheight ofthe specimen, but increase dramatically at the level of the mild steel barsnear the joint. This profile correlateswith the observed corrosion on themild steel at this location.

The most probable explanation forthis result is that saltwater leakagefrom the ponded area ran down theexterior of the specimen to the bottomwhere it entered the concrete. The topsurface and sides of the specimen aresealed with epoxy according to

ASTM G109, while the bottom is not.This phenomenon is common in

bridges, where moisture runs down theside of a member, and chlorides collect at the bottom of the member. Theepoxy sealant on the top and sides ofthe macrocell corrosion specimenswould amplify this effect, accountingfor the increased chloride levels nearthe bottom surface.

The results of the chloride analysison grout samples are shown in Fig. 15.The values are plotted as acid solublechlorides, as a percentage of the groutweight. The chloride threshold for corrosion in grout is taken as approximately 0.14 percent by weight of sample. This threshold is based on achloride threshold of 0.2 percent byweight of cement’3 and a water-cement ratio of 0.44. Note that theweight of cement is approximately 69percent of the sample weight for aplain grout with a water-cement ratioof 0.44.

The dry joint specimens show veryhigh chloride contents, particularly inthe vicinity of the joint. The two dryjoint specimens with steel ducts andlow precompression (Specimens DJ-SL-NG-1 and DJ-S-L-Cl-l) also showlarge chloride contents inside the duct,2 in. (51 mm) from the joint.

The dry joint specimen with a plastic duct, Specimen DJ-P-L-NG-1,showed a high chloride content at thejoint, but only negligible chlorides2 in. (51 mm) inside the duct.

The four epoxy joint specimens analyzed show very low or unmeasurablechlorides at the joint. At a distance of2 in. (51 mm) inside the duct, all samples showed unmeasurable chloridelevels for the epoxy joint specimens.

ANALYSIS ANDDISCUSSION OF RESULTSFour years and five months of expo

sure test data and destructive examination of 19 specimens provided awealth of data to assess the variablesinvestigated in this testing program.An analysis and discussion of the laboratory results are provided below interms of the four groups of variables,namely, joint type, duct type, jointprecompression and grout type. Additional observations are also discussed.

— I I

Acid Soluble Chloride Content(% by weight of concrete)

0.0% 0.1% 0.2% 0.3% 0.4% 0.5%0 . . 0

1 25

51

o

___________

76

U C)lO20 oC.) C)

5 Level of Bars 127_cr Threshold

for Corrosion

6 — 152

Levelof Strand

in. from joint II

I—.—0.5 in. from joint‘

—

Fig. 1 2. Typical concrete chloride ion profiles for dry joint specimens.


0.0% 0.1% 0.2% 0.3% 0.4% 0.5%U— I I I U

I 25

— Level of StrandE

.2 I 51E/ I—--4in.fromjoint I

‘C‘a ( I—”—21n.fromjoint Io I 03 I—.—0.5 in. from jointl 76

U C)l020 oC) C)

5 Levelof Bars 127_crThresholdfor Corrosion

152

Fig. 13. Typical concrete chloride ion profiles for epoxy joint specimens.

84 PCI JOURNAL

Effect of Joint Type

Of the four variable groups investigated, joint type appears to have themost significant effect on the performance of the specimens. In general,dry joints performed very poorly, withcorrosion currents for 78 percent ofthe dry joint specimens indicating corrosion activity. The effect of joint typeon the measured and observed resultsis described below.

Galvanized Steel Duct Corrosion— The extent and severity of duct corrosion was significantly affected bythe joint type. The photos in Fig. 16show typical corrosion of the galvanized steel duct in each of the threejoint types, with two epoxy/gasketjoint specimens shown to illustrate thevaried conditions observed for thisjoint type.

The specimens in Fig. 16 have beencut open at the level of the duct, andthe photo shows the top surface of thespecimen and a top view of the ductstill embedded in the concrete. The topsurface of the concrete had a longitudinal crack due to corrosion in three ofthe four specimens shown. Note thatthe crack has been highlighted in thephoto, and the black arrow indicatesthe joint location.

In general, the duct corroded areaand corrosion severity were less forepoxy joints and corrosion inducedcracking on the concrete surface wasmore severe for dry joints. Duct corrosion was centered on the segmentaljoint in all of the dry joint specimens.Corrosion was not centered on thejoint in the standard epoxy joint, suggesting corrosion was caused by moisture and chloride migration throughthe concrete with no discernible influence from the joint.

Two of the three epoxy/gasket jointspecimens autopsied indicated that thegasket interfered with epoxy coveragein the vicinity of the duct. When thejoint was sound, as shown in the lowerleft photograph in Fig. 16, the ductcorrosion in the epoxy/gasket jointwas less severe than the dry joints andwas not centered on the joint, similarto the standard epoxy joint. However,when epoxy coverage was not complete, severe duct corrosion and concrete cracking occurred, centered on

the joint, suggesting that moisture andchlorides penetrated at the joint.

These results indicate that the standard epoxy joint consistently providesthe best corrosion protection. The results also indicate that the complications introduced in the process byadding gaskets could be counterproductive since corrosion protection wasreduced when compared to the epoxyjoint without a gasket.

Prestressing Strand Corrosion —

Macrocell corrosion current data mea

DJ-S-L-NG.1

DJ-S-M-NG-1

DJ-S-H-NG-1

DJ-P-L-NG-1

DJ-S-L-CI-1

DJ-S-M-CI.1

SE.S-L-NG-2

SE-S-M-NG-2

SE-S-H-NG-2

EG-S-L-NG-2

sured during exposure testing indicated that corrosion of the prestressing strand was only occurring in fourspecimens, all with dry joints. Theprestressing strand corrosion foundduring the destructive examinationwould be considered very mild ornegligible for all specimens with theexception of Specimen DJ-S-L-CI-1[dry joint, steel duct, 5 psi (35 kPa)precompression, corrosion inhibitorgrout].

In general, the strand corrosion

0.0% 0.1%

0

I

. I I

‘aa,

a,

0C0

C)

3

4


0.2% 0.3% 0.4% 0.5%0

25

E51 E

__________

‘aa,

76a,

0ILI

0C.)

127

152

Level of Strand

—.--2 in. from joInt——0.5 In. from joint

..‘

-‘S.5.

5’

‘55.

cr ThresholdLevel of Bars

for Corrosion

5

6

Fig. 14. Concrete chloride ion profiles for Specimen SE—S—M—NG—2.

0.0% 0.1% 0.2%

Acid Soluble Chloride Content(% by weight of grout)

0.3% 0.4% 0.5% 0.6% 0.7%

••At Joint

in. from Joint

ICoosion Threshold 0.14%h

Fig. 15. Measured chloride contents in post-tensioning grout.


Fig. 16. Galvanized steel duct corrosion: effect of joint type.

found in the dry joint specimens wasworse than in the epoxy joint specimens, as illustrated previously in Fig.9. Light to moderate surface corrosionwas found on the strand in all of thedry joint specimens where galvanizedsteel ducts were used. In the standardepoxy and epoxy/gasket joints, corrosion ratings for most specimens werevery low (less than 10) and were attributed to patches of discoloration orvery light strand corrosion.

Mild Steel Reinforcement Corrosion — Corrosion current data indicated that corrosion of mild steel reinforcement was occurring in seven dryjoint specimens and one standardepoxy joint specimen. Destructive examination revealed reinforcing bar corrosion in all of the dry joint specimens,one small area of discoloration in two

epoxy joint specimens and light corrosion in two epoxy joint specimens.

Two-thirds of the epoxy joint specimens had no discoloration or corrosionof the mild steel bars. The highest mildsteel corrosion rating for epoxy jointspecimens occurred in Specimen SE-SM-NG-2 [epoxy joint, steel duct, 50psi (345 kPa) precompression, normalgrout]. This was the only epoxy jointspecimen where corrosion currents indicated activity during exposure testing. As discussed previously, the measured chloride profile for this specimen(see Fig. 14) strongly suggests that elevated chloride levels at the bottom ofthe specimen resulted from an externalsource of moisture and chlorides ratherthan from penetration at the epoxyjoint or through the concrete.

Chloride Penetration — Chloride

penetration was higher for dry jointspecimens in all cases, as illustratedpreviously in Figs. 12 and 13. Chlorideprofiles in the epoxy joint specimenssuggested no influence from the joint.

Chloride analysis performed onsamples from the grout showed veryhigh chloride contents for dry jointspecimens, even at distances of 2 in.(51 mm) from the joint. Grout chloridecontents in the epoxy joint specimenswere very low or negligible.

The measured chloride profiles forthe dry joint specimens illustrate anincreased potential for corrosion of themild steel reinforcement within thesegment. Although dry joints are notpermitted with internal tendons, dryjoints are commonly used in precastsegmental construction with externaltendons. The chloride test data suggest

Dry joint Epoxy joint

Sound epoxy/gasket joint Poor epoxy/gasket joint

86 PCI JOURNAL

Fig. 1 7. Severe duct corrosion damage.

that corrosion protection of mild steelreinforcement adjacent to the jointface should not be overlooked whendry joints are used.

Grouting — Grout leaked into thejoint region in five of the seven dryjoint specimens. The extent of the leakranged from very minor around theduct opening to almost 80 percent ofthe joint face covered with grout inone specimen. No grout leakage wasfound in the standard epoxy joint andepoxy/gasket joint specimens.

Effect of Duct Type

Duct Corrosion — Galvanizedsteel ducts were corroded in all cases,with an example of typical corrosiondamage shown in Fig. 17. Duct corrosion led to concrete cracking along theline of the duct on the top surface ofthe specimen in eight of the fifteenspecimens with galvanized steel ducts.No cracks were found in specimenswith plastic ducts.

Galvanized steel ducts were perforated by corrosion action (corrodedthrough) in nine of fifteen specimens,allowing direct ingress of moistureand chlorides. Plastic ducts were notaffected by exposure testing, and remained intact as a barrier in the corrosion protection system.

The concrete cover in the macrocellcorrosion specimens was lower thanwould be allowed by specification,and this contributed to the severe galvanized duct corrosion in a short period of time. However, the test resultsindicate the potential for corrosion

problems when using galvanized ductsin aggressive exposures, and the relative performance of the galvanizedand plastic ducts is not affected by thelow cover.

Prestressing Strand Corrosion —

Little or no strand corrosion wasfound in dry joint and epoxy jointspecimens with plastic ducts. Strandcorrosion ratings for the four plasticduct specimens autopsied were all lessthan 10, and were attributed to discoloration found on the strands. Light tomoderate surface corrosion and somepitting was found on the strands ingalvanized steel duct specimens withdry joints.

Reversed Macrocell — Macrocellcorrosion current data for the four dryjoint specimens with plastic ducts indicated that the mild steel bars werecorroding instead of the prestressingstrand. Destructive examinations performed on two of the plastic ductspecimens confirmed that the mildsteel reinforcement was the primarycorrosion site (anode). This data suggests that the plastic ducts providedimproved corrosion protection for theprestressing strand in the dry jointspecimens, and, as a result, the mildsteel reinforcement became the preferential site for corrosion.

Effect of Joint Precompression

The three levels of joint precompression investigated show no clear,consistent trends in strand, duct ormild steel reinforcement corrosion.

Effect of Grout Type

Measured macrocell corrosion currents indicated that the prestressingstrand was corroding in four specimens. Three of these four specimenswere autopsied, with the observedstrand corrosion ratings listed in Table3. The most severe corrosion, including the only pitting corrosion, wasfound in the specimen with a corrosion inhibitor admixture in the grout.Based on this limited data, there doesnot appear to be any improvement incorrosion protection when a calciumnitrite corrosion inhibitor is used incement grout.

The dosage of corrosion inhibitorused in this testing program was thesame dosage normally used for concrete (approximately 4 gal. per cu ydof concrete or 20 Liters/m3 of concrete). This dosage was based on thedirect recommendations of the manufacturer’ s representative at the time ofspecimen construction.

The effectiveness of calcium nitrite

(a) Outside surface of duct (joint location at left end)

;iW!nhluT’lllhI

0 10 20 30 40 90 60 70 90 90 100 110 120

Millimot an

(b) Inside surface of duct (joint location at left end)

Table 3. Effect of grout type — strand corrosion ratings.

Specimen

___

Grout type Strand corrosion rating Comments

DJ-S-L-NG-LJ Norma1_gput

__________

26

________

Light to moderate corrosion

DJ-S-M-NG- iJ_ Normal grout 43 Light to moderate corrosion

Light to moderate corrosionDJ-S-L-C1-1 Corrosion inhibitor 114 .

with pitting on three wires


Fig. 18. Typical grout voids.

corrosion inhibitor is related to theratio of calcium nitrite solids to cement solids. Due to the higher cementcontent of grout in comparison to concrete, the dosage used in this testingprogram may be too low for the corrosion inhibitor to be effective. In spiteof this, it is very concerning that calcium nitrite appears to have worsenedcorrosion in comparison to plaingrout.

Other research has found calciumnitrite corrosion inhibitor to be detrimental to corrosion protection whenused in cement grouts. Koester’4 performed anodic polarization tests ongrouted prestressing strand to investigate the corrosion protection providedby various cement grouts.

In these tests, it was found that calcium nitrite significantly reduced thetime to corrosion in comparison toplain grout, and had no effect on corrosion rate after the initiation of corrosion. The calcium nitrite dosage wasadjusted to account for the higher cement content in grout for those tests.Calcium nitrite has shown good resultswhen used in concrete.5’11’15 However,further investigation may be warrantedbefore calcium nitrite corrosion inhibitor should be used in cement grout.

The grout containing 13 percent silica fume was used only in specimenswith a standard epoxy joint. Macrocellcorrosion currents did not indicate aninitiation of corrosion in these specimens. Destructive examination ofSpecimen SE-S-L-SF-2 [epoxy joint,steel duct, 5 psi (35 kPa) precompression, silica fume grout] found smallareas of light corrosion on the pre

stressing strand and a total corrosionrating of 12. These data do not indicate a positive or negative effect ofusing silica fume in cement grout forthe reported exposure period.

Grout Voids

Voids were found in the grout of all19 specimens autopsied. In somecases, the shape and appearance of thevoids suggested that the voids resultedfrom insufficient fluidity. In othercases, voids appear to have resultedfrom air pockets, bleed water collection, or as a result of incomplete filling of the duct during grouting. Inmost cases, observed voids were smallor shallow. However, in several cases,voids were extensive and deep and theprestressing strand was exposed. Examples of typical voids are shown inFig. 18.

Normally, a grout void that does notexpose the prestressing tendon wouldnot be deemed a concern. However,during the destructive examination itwas discovered that five specimenshad holes corroded through the galvanized steel duct at the location of avoid. In two of these specimens, largeholes in the duct corresponded directlyto the voids in shape and size, asshown in Fig. 19.

These findings suggest that the presence of a void in the grout may lead tomore severe corrosion of the galvanized steel duct. The duct is intendedto provide corrosion protection for thetendon, and thus holes in the duct willeffectively eliminate the duct as a protective barrier for the tendon.

The importance of grouting to overall corrosion protection of the post-tensioning system often receives comparatively little attention during thedesign, construction and inspection ofpost-tensioned structures. Grout voidsand poor grout quality are frequentlysignificant factors in tendon corrosionproblems, as evidenced recently bysome high-profile post-tensioningproblems in prestressed concretebridges in Florida.

The findings of this research program illustrated that even under “laboratory conditions,” grout voids may beencountered if proper materials andprocedures are not followed. Recentlypublished guide specifications forgrouting of post-tensioned structures’6and new research efforts17’18 shouldhelp to mitigate grouting problems.

Mechanism for ReversedCorrosion Macrocell

The polarity of the macrocell corrosion current data indicates that eight ofthe 12 specimens with corrosion activity have developed reversed corrosionmacrocells where the mild steel reinforcing bars are corroding (anodic reaction) instead of the prestressingstrand. The development of a reversedmacrocell in typical macrocell specimens is not common, and may be attributed to the transverse segmentaljoint. The use of a dry joint is particularly severe, as indicated by the experimental data.

A possible mechanism for reversedmacrocell corrosion is shown in Fig.20. The dry joint allows rapid penetra

Si’-P-L-NG.2(,routcd I’S ‘rand

I View

SE- 1’- M - N G-2,rout,d I’” Si .i

I III> ‘. It’,

Void caused by entrapped air or incomplete filling Void caused by lack of grout fluidity

88 PCI JOURNAL

Fig. 19. Hole in duct corresponding to grout void.

tion of chlorides to the bottom layer ofsteel. The small end cover for the bottom bars [0.25 in. (6 mm)] provideslittle protection from lateral migrationof the chlorides, and the steel becomesquickly depassivated. It is assumedthat the prestressing steel benefitsfrom the additional protection provided by the grout and duct, and thecorrosion macrocell develops.

It is likely that the added protectionfor the prestressing tendon is primarilydue to the extra thickness of the groutover the strand in comparison to theend cover for the bars. Although theduct is discontinuous at the joint, it mayalso contribute to corrosion protection.

The occurrence of a reversed macro-cell was confirmed by destructive examination. Corrosion of the mild steelreinforcement was found in each of thefive autopsied specimens that hadshown negative macrocell corrosioncurrents during exposure testing. Chloride profiles (where available) also indicated that chloride levels were in excess of the corrosion threshold at thelevel of the bars in these specimens.

SUMMARY OFTEST RESULTS

The majority of corrosion activityafter ‘2 years of extreme, acceleratedexposure testing has occurred in specimens with dry joints (11 of 12 speci

mens with corrosion). Exposure testing is continuing for 19 specimens(one of each specimen type). Continued exposure testing may provide additional results to assist comparison ofvariables, and may change the conclusions presented in this paper. The primary findings of the research programare listed below.

Overall Performance

• Overall performance of the segmental macrocell corrosion specimensin this program is very good, withonly minor corrosion detected in alimited number of specimens.

• Metal loss calculations indicate thatprestressing strand corrosion overthe exposure duration is minor ornegligible.

• Possible strength degradation in the

form of pitting corrosion on prestressing strand was found in onlyone specimen.

Segmental JointsAll significant corrosion occurred inspecimens with dry joints. Seventy-eight percent (11 of 14) dry jointspecimens displayed corrosion activity. Specimens with dry jointsshowed increased chloride penetration and increased corrosion of galvanized steel duct, prestressingstrand and mild steel reinforcement.The mild steel reinforcement is corroding instead of the prestressingstrand in seven of the 11 dry jointspecimens with corrosion activity.This occurrence is attributed to penetration of chlorides at the dry segmental joint and indicates a possible

DJ-S-M-NG-1Dud - Segment B

Top Outside Surface

II 11111111111111111111111 ii 11111 11111 111111 tIll liii I50 40 30 20 10 0

Milllmeters

Corroded duct with large hole corresponding togrout void

• 111111111111111111 111111111111111111111 111111 Ii

50 40 30 20 10 0

Grout surface with large void — Note corrosion productaccumulated in void

NaCIstrand “protected” by . solutionduct & grout grouted

prestressingdry joint strandprovidespathway for mild steel

chlorides to barsI

_

reach bars —

______

Fig. 20. Probable mechanism for “reversed” macrocell corrosion.


increased corrosion threat for mildsteel reinforcement within the segment when dry joints are used. Thiscould occur in bridges with externaltendons, and highlights the importance of clear cover over the ends oflongitudinal bars in the segments.One out of 24 specimens with epoxyjoints has shown corrosion activity.Autopsy of this specimen confirmedthat the mild steel reinforcement wascorroding rather than the prestressing strand. Measured chloride profiles for this specimen suggest thatcorrosion resulted from an externalsource of moisture and chlorides(leakage of dam on top of specimen)rather than from penetration at theepoxy joint or through the concrete.Corrosion of the galvanized steelduct was reduced in extent andseverity in specimens with epoxyjoints. Only very minor prestressingstrand corrosion was found in specimens with epoxy joints. The experimental data indicate that thin epoxyjoints provide substantially improvedcorrosion protection for internal tendons in segmental construction.

• The use of gaskets in epoxy jointsmay interfere with epoxy coverageon the joint. Autopsied epoxy/gasketjoint specimens revealed incompleteepoxy coverage near the duct openings, leading to increased chloridepenetration and duct corrosion. Theobserved deficiencies occurred incarefully controlled laboratory conditions, and could possibly be worseunder field conditions.

Ducts for Internal Post-Tensioning

• Strand corrosion was not detectedduring exposure testing in anyepoxy joint specimens with plasticducts. Reversed macrocell corrosiondeveloped in the four dry joint specimens with plastic ducts, indicatingthat the plastic duct is providing improved corrosion protection for theprestressing strand (tendon), evenwhen penetration of chlorides at thedry joints has caused corrosion ofthe mild steel bars.

• Destructive examination revealedonly very minor corrosion or discoloration on the prestressing strandfrom specimens with plastic ducts.

• Galvanized steel ducts were corroded in all cases, leading to concrete cracking along the line of thetendon in many specimens. Ductswere corroded through in nearlytwo-thirds of the specimens, eliminating the duct as corrosion protection for the prestressing tendon. Theconcrete cover in the test specimenswas lower than specification, contributing to the poor performance ofthe galvanized duct in a short periodof time. However, test results indicate the potential for durabilityproblems when using galvanizedducts in aggressive exposures.

• Specimens with plastic ducts andepoxy joints had the best overallperformance in the testing programin terms of strand, mild steel andduct corrosion.

Joint Precompression

• The range of joint precompressioninvestigated did not affect the timeto corrosion or corrosion severity.

Grouts for Bonded Post-Tensioning

• The most severe corrosion of theprestressing tendon was foundwhere a calcium nitrite corrosion inhibitor was used in the grout.

• Two specimens with silica fume inthe grout (and epoxy joints) did notshow corrosion activity.

CONCLUSIONSAfter 41/2 years of severe exposure

testing and destructive examination ofone-half of the test specimens, a number of significant conclusions can bedrawn. While the focus of this research was on precast segmental construction, several of the conclusionsare not strictly related to segmentalconstruction, but are equally relevantto all applications of internal, groutedpost-tensioning tendons. The primaryconclusions from this research are asfollows:

1. Dry joints permit moisture andchloride penetration at the joint between segments, compromising corrosion protection for internal tendons.

2. Penetration of moisture and chlorides at dry joints increases the extentand severity of corrosion of mild steel

reinforcement and galvanized steelpost-tensioning ducts within the segment.

3. Properly constructed thin epoxyjoints prevent chloride penetration atthe segmental joint.

4. Thin epoxy joints substantiallyimproved corrosion protection for internal prestressing tendons, and forgalvanized steel ducts and mild steelreinforcement within the segments.

5. The use of a gasket around theduct opening in a thin epoxy joint mayinterfere with thorough coating of thejoint face with epoxy, and could compromise corrosion protection.

6. Galvanized steel post-tensioningducts provide only limited corrosionprotection for internal tendons, andmay corrode through in severe exposure conditions or when low concretecover has been provided.

7. Plastic ducts provide a significantimprovement in corrosion protection,limiting prestressing tendon corrosionto negligible levels and eliminatingconcrete cracking due to duct corrosim.

8. Joint precompression up to 3 f’psi (7.88/ kPa) does not appear toinfluence corrosion activity.

9. Test results suggest that the useof a calcium nitrite corrosion inhibitorshould be carefully investigated priorto use in cement grouts.

10. The presence of grout voids mayincrease the extent and severity ofboth tendon and duct corrosion.

RECOMMENDATIONSBased on the results of this labora

tory research program, the followingrecommendations are made regardingcorrosion protection for internal post-tensioning tendons in precast segmental bridges. Similar to the researchconclusions, some of the recommendations are not specific to segmentalconstruction, but are applicable to allforms of internal, grouted post-tensioning tendons.

1. Match-cast dry joints should notbe used with internal tendons in anyenvironment where exposure to chlorides may occur, including applications where chloride-bearing deicingchemicals are used and marine (saltwater) environments.

2. Where dry joints are used with ex

90 PCI JOURNAL

ternal tendons, the face of the segmentshould be considered as an exposedface. Accordingly, concrete cover formild steel segment reinforcement adjacent to the joint face should meet thesame cover requirements as requiredelsewhere in the structure to minimizecorrosion due to penetration of moisture and chlorides at the dry joint.

3. Match-cast thin epoxy jointsshould be used in all applicationswhere corrosion of the internal tendons is a concern.

4. Proper application of epoxy to thejoint faces during the construction process is critical to corrosion protection atthe joint. Thus, this step in the construction process must receive an appropriatelevel of attention from all constructionand inspection personnel involved.

5. The use of gaskets around theduct opening in epoxy joints should beavoided. As an alternative, ductsshould be swabbed immediately following segment placement and initialstressing to prevent epoxy from blocking the duct.

6. Plastic post-tensioning ductsshould be used for all internal tendonsin applications where corrosion is aconcern, including marine exposuresand environments where chloride-bearing deicing chemicals are used.

7. Proper grout materials and procedures should be utilized to minimizegrout voids and improve corrosionprotection. Significant advances in thestate-of-the-practice for grouting havebeen made recently. Guide specifications are provided in Reference 16.

ACKNOWIEDGMENTSThe research in this project was

sponsored by the Texas Department ofTransportation. Any opinions expressed in this paper are those of theauthors and not necessarily those ofthe sponsor. The authors would like tothank Dr. H. R. Hamilton, III, for hisassistance in the development of thetesting program and with monitoringof the specimens. Thanks are also dueto Mr. Blake Stasney at the Phil M.Ferguson Structural Engineering Laboratory, and to the many graduate research assistants who helped withspecimen monitoring over the years.

The authors wish to express theirappreciation to the PCI JOURNAL reviewers for their constructive comments.

REFERENCES1. AASHTO, Guide Specifications for Design and Construction

of Segmental Concrete Bridges, Second Edition, American Association of State Highway and Transportation Officials,Washington, DC, 1999.

2. ACT Committee 222, Corrosion of Prestressing Steels, ACT222.2R-01, American Concrete Institute, Farmington Hills,MI, 2001.

3. Sherman, M. R., McDonald, D. B., and Pfeifer, D. W., “Durability Aspects of Precast Prestressed Concrete — Part 1 Historical Review,” PCI JOURNAL, V. 41, No. 4, July-August1996, pp. 62-74.

4. Sherman, M. R., McDonald, D. B., and Pfeifer, D. W., “Durability Aspects of Precast Prestressed Concrete — Part 2: Chloride Permeability Study,” PCI JOURNAL, V. 41, No. 4, July-August 1996, pp. 76-95.

5. Pfeifer, D. W., Landgren, J. R., and Zoob, A., “Protective Systems for New Prestressed and Substructure Concrete,”FHWA/RD-86/193, Federal Highway Administration, Washington, DC, April 1987.

6. Miller, M. D., “Durability Survey of Segmental ConcreteBridges,” PCI JOURNAL, V. 40, No. 3, May-June 1995, pp.110-123.

7. Ronald, H. D., “Design and Construction Considerations forContinuous Post-Tensioned Bulb-Tee Girder Bridges,” PCIJOURNAL, V. 46, No. 3, May-June 2001, pp. 44-66.

8. ASTM, “Standard Test Method for Determining the Effects ofChemical Admixtures on the Corrosion of Embedded Steel Reinforcement in Concrete Exposed to Chloride Environments,”ASTM G109-92, American Society for Testing and Materials,Philadelphia, PA, 1992.

9. West, J. S., Vignos, R. P., Breen, J. E., and Kreger, M. E.,“Corrosion Protection for Bonded Internal Tendons in PrecastSegmental Construction,” Research Report 1405-4, Center forTransportation Research, Bureau of Engineering Research,

The University of Texas at Austin, Austin, TX, October 1999.10. ASTM, “Standard Test Method for Half-Cell Potentials of Un-

coated Reinforcing Steel in Concrete,” ASTM C876-9l,American Society for Testing and Materials, Philadelphia, PA,1991.

11. Virmani, Y. P., Clear, K. C., and Pasko, T. J., “Time-to-Corrosion of Reinforcing Steel in Concrete Slabs, V. 5: Calcium Nitrite Admixture or Epoxy-Coated Reinforcing Bars as Corrosion Protection Systems,” Report No. FHWA/RD-83/012,Federal Highway Administration, Washington, DC, September1983, 71 pp.

12. AASHTO, “Sampling and Testing for Chloride Ion in Concrete and Concrete Raw Materials,” AASHTO T 260-94,American Association of State Highway and TransportationOfficials, Washington, DC, 1994.

13. ACT Committee 222, Corrosion of Metals in Concrete, ACI222R-96, American Concrete Institute, Farmington Hills, MI,1996.

14. Koester, B. D., “Evaluation of Cement Grouts for Strand Protection Using Accelerated Corrosion Tests,” M. S. Thesis, TheUniversity of Texas at Austin, Austin, TX, December 1995.

15. Berke, N. S., Dallaire, M. P., Hicks, M. C., and Hoopes, R. J.,“Corrosion of Steel in Cracked Concrete,” Corrosion, V. 49,No. 11, November 1993, pp. 934-943.

16. PTI Grouting Committee, Guide Specification of Grouting ofPost-Tensioned Structures, Post-Tensioning Institute, Phoenix,AZ, 2001, 69 pp.

17. Schokker, A. J., Hamilton III, H. R., and Schupack, M., “Estimating Post-Tensioning Grout Bleed Resistance Using a Pressure-Filter Test,” PCI JOURNAL, V. 47, No. 2, March-April2002, pp. 32-39.

18. Schokker, A. J., Breen, J. E., and Kreger, M. E., “Grouts forBonded Post-Tensioning in Corrosive Environments,” ACI Materials Journal, V. 98, No.4, July-August 2001, pp. 296-305.


evaluation of corrosion protection for internal

Documents