The Effect of Initial Rust on the Bond Strength of Reinforcement

KSCE Journal of Civil EngineeringVol. 8, No. 1 / January 2004pp. 35~41

Structural Engineering

Vol. 8, No. 1 / January 2004 − 35 −

The Effect of Initial Rust on the Bond Strength of Reinforcement

By Byung Duck Lee*, Kook Han Kim**, Hwan Gu Yu***, and Tae Song Ahn****

···············································································································································································································································

Abstract

An experimental investigation on the relationship between corrosion of reinforcement and bond strength in pull-out test specimen has beenconducted to establish the allowable limit of rust of reinforcement in the construction field. The reinforcing bars used in this study were rustedbefore embedded in pull-out test specimen. The first component of this experiment is to make reinforcing bar rust electrically based onFaraday’s theory to be 2, 4, 6, 8 and 10% of reinforcing bar weight. For estimation of the amount of rust by weight, Clarke’s solution and shotblasting were adopted and compared. Parameters also include 24 and 45 MPa of concrete compressive strengths and diameter of reinforcingbar (16, 19 and 25 mm). Pull-out tests were carried out according to KS F 2441 and ASTM C 234 to investigate the effect of the degree of ruston bond strength. It is found from the test results that the test techniques for corrosion of bar used in this study is relatively effective andcorrect. Results show that up to 2% of rust increases the bond strength regardless of concrete strength and diameter of reinforcing bar like theexisting data. It might result from the roughness due to rust. As expected, the bond strength increases as compressive strength of concreteincreases and the diameter of bar decreases.Keywords: corrosion, artificial accelerated potentiometric corrosion, bond strength, slip, pullout test

···············································································································································································································································

1. Introduction

The most of reinforcing bar stored at construction field is likelyto corrode due to the direct exposure to outdoors. However, thecurrent specification is based on bond characteristics of cleanreinforcing bar and previous research have been also carried outoften embed in clean reinforcing bars. Therefore, it has been inconflict between contractor and inspector to use corroded reinforcingbar in construction sites. Thus, it is very important problem todecide whether the corroded reinforcing bar can be used fromefficiency standpoint or that should not be used from safetystandpoint.

According to previous research results about corrosion of thereinforcing bar (Al-Sulaimani et al., 1990; Malvar, 1995), when thecorrosion level of reinforcing bar is small, the bond strength betweenthe reinforcing bar and surrounding concrete increases with anincrease of corrosion. Accordingly, if the results of this study givesthe allowable corrosion level with no deterioration of mechanicalbehavior of reinforced concrete, this study will be not only expectedto contribute to the effectiveness of concrete construction but also tocost-saving.

The main purpose of this investigation is to suggest the allowablecorrosion level of reinforcing bar by test without decrease of bondforce between reinforcing bar and surrounding concrete. Thereinforcing bars used in this study are rusted by artificial acceleratedpotentiometric corrosion method based on Faraday’s law in order toinduce exact amount of the rust and to reduce the time of rustproduction. The calculation of degree of rust is conducted withweight loss method in accordance with the ASTM G1-81 Clarke’ssolution method and the Shot blasting method.

2. Experimental Program

2.1. Test ParametersThe adopted test parameters to measure the corrosion amount are

the nominal diameter of reinforcing bars and the amount of corrosionand rust removal methods (Clarke’s solution and Shot blasting).

To test the bond strength between the reinforcing bar andsurrounding concrete, parameters includes compressive strength ofconcrete and diameter of reinforcing bar corresponding developmentlength for pull-out test. Pull-out tests were carried out according toKS F 2441 and ASTM C 234 to investigate the effect of the amountof rust on the bond behavior between reinforcing bar and concrete.The prepared concrete specimens for testing bond strength inconsideration of those variables are totals to 108 specimens. The testparameters used in this study are summarized in Table 1.

2.2. Materials and Mix Proportions2.2.1. Deformed Reinforcing BarThe reinforcing bars used for corrosion test in this study are based

on deformed bar. The diameters of deformed reinforcing bars wereselected among those used in actual construction field. Theirdesignations are D16, 19, and 25, respectively.

The high-strength deformed reinforcing bars were used, whichwere tested in accordance with KS B 0801 (Test pieces for tensiletest for metallic materials) and KS B 0802 (Method of tensile test formetallic materials).

2.2.2. Cement, Aggregate, and AdmixturesType I ordinary portland cement was used. Fine aggregate was

river sand and coarse aggregate was crushed stone with specific

*Chief Researcher, Korea Highway Corporation, Hwaseong, Korea (E-mail: [email protected])**Section Chief, Korea Highway Corporation, Seongnam, Korea (E-mail: [email protected])

***Section Chief, Korea Highway Corporation, Youngcheon, Korea (E-mail: [email protected])****Member, Research Director, Korea Highway Corporation, Hwaseong, Korea (E-mail: [email protected])

Byung Duck Lee, Kook Han Kim, Hwan Gu Yu, and Tae Song Ahn

− 36 − KSCE Journal of Civil Engineering

gravity 2.63, and the maximum aggregate size of 25 mm.AE water reducing agent (Lignin type) and superplasticizer

(Lignin type) was used for high strength concrete. The amount of AEwater reducing agent and superplasticizer used in this study wereinitially determined on the basis of recommended content by thesupplier and then adjusted further to get a required strength throughmix design and actual testing of mixed concrete. These amount were0.3% and 0.8% of the unit cement content.

2.2.3. Mix Proportions of ConcreteIn order to estimate the variation of bond strength with concrete

strength, specified average strength was 24 MPa and 45 MPa. Themix proportion used in this test is listed in Table 2.

2.3. Corrosion Methods of Deformed Bar2.3.1. Test Methods2.3.1.1. Preparing Deformed BarFirst, deformed reinforcing bar was cut into 1 m length using cutter

and cutting surface was finished flat and then the length of thedeformed bar is measured up to 1mm by ruler. Finally, the weight ofdeformed reinforcing bar was measured up to 1/100 g using high-sensitivity electronic balance (range : 4,000~1/100 g). The measuredlength and weight were later used to determine the amount ofcorrosion. After the measurement, deformed bars were epoxy-coatedunder ventilated low-moisture condition as shown in Fig. 1. The total5 cm was left uncoated for electrical connecting to a power source.The actual length to be corroded was 80 cm although the total lengthof deformed bar was 1 m. The length of deformed bar used for the

measurement of the amount of corrosion was 70 cm. The surfacecoated with epoxy was covered with Parafilm®, in order to preventthe possible damage of epoxy coating during storage or moving. Fig.1 below shows the Schematic diagram of deformed bar for corrosiontesting.

2.3.1.2. Corroding Method of Deformed Bar(1) Faraday’s law of electrolysisThe forced corroding method of deformed bar used in this study is

based on Faraday’s law of electrolysis. Faraday’s law of electrolysisis; “The amount of chemical reaction caused by the flow of current isproportional to the amount of electricity passed”. The following Eq.(1) is indicated the Faraday’s equation.

(1)

where,w = mass loss in g, weight of corrosion products removed by

electrolyte solution during time(sec)I = current, At = applying time of the current(A)M = molecular weight of the metal, g/mol, in case of deformed

reinforcing bar, 55.85 g/moln = electron per molecule oxidized or reduced, in case of

deformed reinforcing bar, 2F = Faraday constant, 96,500 C/mol or 96,500A .s/mol

(2) Constitution of current supplying circuit board and corrosioncell

For artificial corrosion of the deformed bar from Faraday’s law ofelectrolysis, the most important part might be the current supplyingcircuit which can apply the constant electric current(A) for certainperiod of time(t). The electric current of 3±0.2A was applied to eachdeformed bar from the circuit used in this study.

The corrosion cell was made of acrylate board for the corrosionreaction to be observed outside. The cell was divided into 18 subsellsto corrode 18 deformed reinforcing bars simultaneously. The size ofcell was 127×64×101 cm and that of each subcell was 20×20×100cm. A waterproofing sealers was applied between subcells to preventthe flow of electrolyte. Fig. 2 shows the corrosion set up in the cell.

The deformed bar used as an anode was immersed into electrolyteby hanging specimen mount using a grip or connector. A coil-typestainless steel was adopted as a cathode as shown in Fig. 2. Theanode was connected to positive terminal(+) of power source whilethe cathode to the negative terminal(−). In the electrolyzation reaction,the cross-section of the cathode should be more than twice for thecathode reaction not to limit the whole corrosion reaction.Accordingly, the stainless steel used for each subcell was the plate of30×5,000×0.3 mm, which was then made into a coil. 5%-Naclsolution was used as electrolyte which was prepared by firstdissolving Nacl 50 g in 900 ml of water and then making 1,000 mlby adding more water.

w ItMnF

---------- = t unFIM

-----------=⇒

Table 1. Test Parameters for Corrosion Measurement and BondStrength

Parameters Variable for corrosion test Bond strength test

Diameter of deformed bar D16, D19, D25 D16, D19, D25

Level of corrosion (%) 2, 4, 6, 8, 10% 0, 2, 4, 6, 8, 10

Removal methods of rust Claker’s solutionShot blasting −

Concrete compressivestrength (MPa) − 24, 45

Table 2. Mix Proportion of Concrete

Target strength(MPa)

Gmax(mm)

Slump(cm)

W/C(%)

S/A(%)

Unit mix content (kg/m3)

Water Cement Fine aggregate Coarseaggregate

24 25 13.2 41 43 167 406 761 1095

45 25 16.3 36 43 167 463 741 1066

Fig. 1. Schematic Diagram of Deformed Bar for Corrosion Testing


Vol. 8, No. 1 / January 2004 − 37 −

The specified amounts of corrosion were 2, 4, 6, 8, 10%, each ofwhich was the ratio of weight loss to original deformed bar. Theduration of constant current supply was calculated in accordance toan Eq. (1). Table 3 shows the expected amount of corrosion andsupply duration for each deformed bar. The deformed reinforcing barafter the complition of forced corrosion reaction was removed fromNacl solution, washed with clean water and dried about one hour inthe shade before the fabrication of the specimen for bond strengthtest. The amount of corrosion of each dried bar was measured rightbefore the preparing of the specimen.

2.3.1.3. Corrosion Measurement and Rust Removal Methods ofDeformed Reinforcing Bar

(1) Calculation of theoretical corrosion amount by Faraday’s lawThe amount of rust of deformed reinforcing bar was calculated by

weight in according to Eq. (1) based on the Faraday’s law. Therefore,the levels of corrosion by Faraday’s equation law are calculated asthe following Eq. (2).

(2)

where, Cfr = the corrosion ratio, %Wf = mass loss in g, weight of corrosion products removed

by Clarke’s solution or Shot blastingW = origin metal weight, g

(2) Removal of rust and calculation of rust amount by Clarke’ssolution

Two types of methods were adopted for the removal of rust on the

surface of deformed bar. One of them is the removal method byClarke’s solution (ASTM G1-81 : Standard practice for preparing,cleaning, and evaluating corrosion test specimens), which was inaccordance with ASTM G1-81-7.7.2. The preparing method ofClarke’s solution is given in Table 4 (ASTM G1-81, 1981).

First, the deformed reinforcing bars are dipped in the preparedClarke’s solution for a certain period of time. Then, the deformed barspecimens are rinsed with clean water, with a non-polishingtool(dried patch) and weighed. The weight of deformed bar ismeasured up to 1/100 g by using high-sensitivity electron balance.Therefore, the amount of rust removed by Clarker’s solution wascalculated according to Eq. (2).

(3) Removal of rust and calculation of rust amount by Shot blastingAnother method of the rust removal of deformed bar was Shot

blasting machine. The metal balls used in this study were made bycutting piano steel wire with 0.8 mm pieces. The jet velocity of Shotblasting is 3,000 rpm and the operation time of rust removal for thecorroded deformed bar was determined by the preliminary operationwith 2 minutes. The level of corrosion by Shot blasting wascalculated as the following Eq. (2).

2.4. Bond Test Methods between Deformed Bar and ConcreteThe test of bond strength is basically conducted in accordance with

Pull-out test of “Testing method for comparing concrete on the basisof the bond developed with reinforcing bar” of the standards of theKS F 2441 and ASTM C 234. The test specimens consist of cubeswith the size of 150×150×150 mm. However, in case of testing of thebond strength according to this methods, the cone type failure ofconcrete is apt to be taken place at the point where deformed bar ispulled out. To prevent this failure, a certain length of deformed bar atboth the loaded and unloaded ends of the specimen are prepared topipes not to be occurred adhesion. The dimension and bond length of

CfrWf

W------ 100 %( )×=

Fig. 2. Cell Apparatus for Artificial Accelerated Potentiometric Corrosionof Deformed Bar

Table 3. Expected Amount and Time of Corrosion (with an appliedcurrent of 3A)

DescriptionSize of

deformedreinforcing bar

2% 4% 6% 8% 10%

Corrosionamount,

(g)

D16 24.96 49.92 74.88 99.84 124.80

D19 36.00 72.00 108.00 144.00 180.00

D25 63.68 127.36 191.04 254.72 318.40

Corrodingtime,(sec)

D16 28751 57502 86253 115004 143755

D19 41468 82936 124404 165872 207340

D25 73353 146706 220059 293412 366765

Table 4. Preparing of Clarke’s Solution

Hydrochloric acid (HCl, Specific gravity 1.19, 38%) 1 l§

Antimony trioxide (Sb2O3) 20 g

Stannous chloride (SnCl2) 50 g

Temperature room

Dip in time up to 25 min

Fig. 3. Schematic Drawing of the Specimen for Bond Strength



these specimens are listed as following in Fig. 3 and Table 5.The test set-up for bond strength is made by modifying the

measurement apparatus of KS F 2441 and ASTM C 234 (ASTM,1991), which is illustrated in Fig. 4. To measure the amount of slip ofembedded deformed reinforcing bar, the displacement transducerwith 0.01 mm resolution is set up on both the unloaded and loadedend (see Fig. 4). The pull-out test is conducted using UniversalTesting Machine(UTM) of 100 tonf capacity and loading wasapplied at a rate of 1 mm/min through displacement control.

3. Test Results and Discussion

3.1. Corrosion of Deformed Reinforcing BarIn this study, 3 specimens for each experimental variable were

tested and the results are shown in Table 6. As shown in Table 6, theratio of rust formed by artificial accelerated corroding method waslittle higher than theoretical one irrespective of the rust removalmethods such as Clarke’s solution and Shot blasting and the nominaldiameters of deformed reinforcing bars. The reason is thought to bethat additional source of corrosion other than D. C. voltage was inartificial accelerated corroding method seemed to exist.

3.2. Bond Test Results3.2.1. Bond Stress-slip Relationship with Various FactorsGenerally, all deformed reinforcing bars show bond stress-slip

relationship as shown in Fig. 5. However, the slip values s1 and s2

with bond stresses τs1 and τs2 are varying in accordance with thenominal diameter of reinforcing bar. Bond stress τs0 is attributed tochemical adhesion and friction at interface between the deformationspart of reinforcing bar and the surrounding concrete while theincrement of bond stress from τs0 to τs1 to mechanical interactionbetween deformations of the bar and the surrounding concrete,before the failure surface of concrete is cracked.

Therefore, such increment between τs0 and τs1 becomes apparentonly when deformation spacing is small. The increase of bond stressfrom τs1 to τs2 is ascribed to the mechanical interaction betweendeformations of the reinforcing bar and the surrounding concreteafter loading is exerted on failure surface of concrete. Slip extensionrate defined as (s2-s1)/(s1-s0) will increase with the increase ofconcrete strength and so will the increasing rate of bond stress.

The test results are shown in Figs. 6(a)~(f). Yield plateau for D-16deformed reinforcing bar was not shown because of narrow spacingbetween s1 and s2. For D-19 deformed bar, ideal bond stress-slipcurve is obtained and slip at ultimate bond stress is high. For D-25deformed reinforcing bar, s1 is very small to be superimposed on s0

actually. In general, ultimate bond stress and slip extension rateincrease with decrease of nominal diameter of bar irrespective ofconcrete strength or the amount of rust.

Ultimate bond stress for the same nominal diameter of deformedreinforcing bar increases with increasing of concrete strengthirrespective of the amount of rust and so does the slip value(s2) at thepoint of ultimate bond stress due to increase of slip extension rate.The bond stress-slip relationship with the amount of rust ofreinforcing bar is shown in Figs. 7(a)~(f), which show a littlediscrepancies among various nominal diameters of reinforcing bars.For D-16 deformed reinforcing bar, ultimate bond stresses of 2% and4% corroded reinforcing bars are greater than that of 0% corrodedreinforcing bar. For D-19 deformed reinforcing bar, ultimate bondstress of 2% corroded reinforcing bar is greater than that of 0%corroded one irrespective of concrete strength. For D-25 deformedreinforcing bar, similar trends are observed.

3.2.2. The Relationship between the Amount of Rust and BondStress

The ultimate bond stresses of 3 different deformed reinforcing barswith the amount of rust are shown in Table 7. The ratios of ultimatebond stress of the deformed reinforcing bars with various amount of

Table 5. Dimension and Bond Length for Bond Strength Specimens

Description Size of concrete specimens(cm)

Length of bonded part(4D, cm)

D 16 15×15×15 6.40

D 19 15×15×15 7.64

D 25 15×15×15 10.20

Fig. 4. Test Apparatus for Bond Strength

Table 6. The Variation of Corrosion Ratio with Two Rust RemovalMethods

DescriptionAmount of rust (Clarke’s solution/Shot blasting), (%)

2 4 6 8 10

D 16 2.39/2.36 4.53/4.47 6.38/6.46 8.21/8.17 10.48/10.28

D 19 2.31/1.98 4.94/4.90 6.90/6.72 8.31/8.60 10.64/10.61

D 25 2.76/2.90 4.94/4.89 6.78/6.99 8.94/8.66 10.63/10.55

Fig. 5. Bond Stress-slip Curve in Reinforced Concrete


Vol. 8, No. 1 / January 2004 − 39 −

rust to that with 0% corrosion are shown in Figs. 8(a)~(c).Ultimate bond stress of 2% corroded deformed bars is greater than

that of 0% corroded deformed bars irrespective of nominal diametersof deformed bar and concrete strength. For D-19 deformed barembedded normal strength concrete, ultimate bond stress of 4%corroded deformed bars is lower than that of 0% corroded deformedbars. Although, ultimate bond stress of bars with corrosion more than6% is sometimes greater than that of 0% corroded deformed bars, itis general tendency that bond strength of bars with corrosion morethan 6% is lower than that of 0% corroded deformed bars. For D-25deformed bar embedded in high strength concrete, ultimate bondstress of 2, 4, 6, 8, 10% corroded deformed bars become higher thanthat of 0% corroded deformed bars. It might resulted from themechanical interaction between deformations and surroundingconcrete in high strength concrete which is different with that innormal strength concrete. However, further study for local interactionbetween deformations and surrounding concrete to know the

movement precisely is seemed to be needed.For the effects of concrete strength on bond strength, the rate of

change in bond stress with the amount of rust in high strengthconcrete is lower than that in normal strength concrete irrespective ofnominal diameter of reinforcing bar. It might be concluded from theresults that there is no reduction in bond stress of the bars with lessthan 2% corrosion irrespective of nominal diameters and concretestrength. However, with further corrosion, the bond stressdeclines consistently until it becomes negligible for about 4, 6, 8,10% corrosion. It is similar to previous research results (Al-Sulaimani et al., 1990). This can be explained on the basis ofincreased surface roughness of reinforcing bar with the growth offirm rust, which tends to enhance the holding capacity of thereinforcing bar.

3.2.3. Variation of Failure Mode with the Amount of RustIn this study, in high strength reinforced concrete, concrete failure

Fig. 6. (a) ~ (f)



is not observed only for 8% and 10% corroded D-16 deformed bar.However, in normal strength reinforced concrete, concrete failurewas not occurred for D-16 deformed bar with 6, 8 and 10% corrosion

and D-19 deformed bar with 0% corrosion.

Fig. 7. (a) ~ (f)

Table 7. The Ultimate Bond Stress at Maximum Load

DescriptionBond stress(MPa), (Normal strength / High strength)

0% 2% 4% 6% 8% 10%

D 16 13.1/22.7 14.6/23.5 16.5/24.2 11.5/22.5 10.4/22.0 8.2/22.5

D 19 9.9/15.8 14.5/20.1 9.0/16.2 8.9/15.8 8.9/13.8 7.5/13.0

D 25 7.3/8.0 8.2/9.1 7.4/8.9 6.3/9.0 6.5/10.2 7.4/9.4


Vol. 8, No. 1 / January 2004 − 41 −

4. Conclusions

(1) The amount of rust formed by artificial accelerated corrodingmethod was a little higher than theoretical one irrespective ofthe nominal diameters of deformed bars and the rust removalmethods such as Clarke’s Solution and Shot Blasting. It mightresult from the additional source of corrosion other than D. C.voltage was in artificial accelerated corroding method.

(2) The amount of rust produced by two different removalmethods (dipping in Clarke’s Solution and Shot Blasting) wasalmost same. The difference between the measured amount ofrust and the theoretical one became smaller with a decrease ofnominal diameters of bars and its decreased average ratio totheoretical one was about be 13%.

(3) For the effects of nominal diameter on bond stress-sliprelationship, ultimate bond stress and slip extension rateincreased with a decrease of nominal diameters regardless ofconcrete strength or the amount of rust.

(4) For the effects of concrete strength on bond stress-sliprelationship, ultimate bond stress of deformed bar with samenominal diameter increased with increase of concrete strengthirrespective of the amount of rust of deformed bar.

(5) The effects of the amount of rust on bond stress-sliprelationship show a little difference for different nominaldiameters of deformed bars. For D-16 deformed bar embeddedin both high strength and normal strength, ultimate bondstresses of 2% and 4% corroded deformed bar are greater thanthat of 0% corroded deformed bar. For D-19 deformed bar,ultimate bond stress of 2% corroded deformed bar is greaterthan that of 0% corroded deformed bar irrespective of concrete

strength. For D-25 deformed bar embedded in high strengthconcrete, ultimate bond stress of 2, 4, 6, 8, 10% corrodeddeformed bars become higher than that of 0% corrodeddeformed bars. A proper amount of rust may increase the bondstress by increasing roughness of the bar surface while a largeamount of rust may decrease the bond stress due to loose rust.The amount of rust less than 4% seem to play a role inincreasing roughness rather than loosening which resulted inincrease of bond stress.

References

ACI Manual of Concrete Practice (1994). “Materials and General Propertiesof Concrete.” ACI Manual of Concrete Practice, PART 1, pp. 222R 1-30.

Al-Sulaimani, G.J., Kaleemullah, M., Basunbul, I.A., and Rasheeduzzafa.(1990). “Influence of Corrosion and Cracking on Bond Behavior andStrength of Reinforced Concrete Members.” ACI Structural Journal,Technical Paper, pp. 220-231.

ASTM G1-81 (1981). “Preparing, Cleaning, and Evaluating Corrosion TestSpecimens.” ASTM Standards, pp. 829-834.

ASTM C 234-91a (1991). “Standard Test Method for Comparing Concreteson the Basis of the Bond Developed with Reinforcing Steel.” ASTMStandards, pp. 153-157.

Malvar, L.J. (1995). “Tensile and Bond Properties of GFRP ReinforcingBars.” ACI Materials Journal, Technical Paper, Title No. 92-M30, pp.276-284.

Soroushian, P., Choi, K.B., and Park, G.H. (1991). “Bond of Deformed Barsto Concrete: Effects of Confinement and Strength of Concrete.” ACIMaterials Journal, Technical Paper, Title No. 88-M27, pp. 227-232.

(Received on May 9, 2003 / Accepted on October 6, 2003)

Fig. 8. Relative Ultimate Bond Stress to That of 0% Corrosion

Documents

The Effect of Initial Rust on the Bond Strength of Reinforcement