Effect of Chloride Content on Bond Behavior Between FRP and Concrete

Tran s.Tian jin Univ.2010,16:405-410DOI 10.1007/s 12209-010-1399-xAccept ed d at e:2009-09-29.Supported by National Natural Science Foundation of Chi na (No.50538086),National Key Basic Research and Development Program of China (“973”,N B63),x Y T f S U y N ,6,,D ,f N ,j @Effect of Chloride Content on Bon d BehaviorBetween FRP and Con crete *PAN Jinlong(潘金龙)1，HUANG Y ifa ng (黄毅方)1，X ING Fe ng(邢锋)2(1.Key Laboratory of Concrete &Prestressed Concrete Structures of Ministry of Education of China,College of Civil Engi-neering,Southeast University,Nanjing 210096,China ；2.College of Civil Engineering,Shenzhen Key Laboratory on Durability of Civil Engineering,Shenzhen University,Shenzhen 518060,China)Tianjin University and Springer-Verlag Berlin Heidelberg 2010Ab stract ：For reinforced concrete structures located along the seaside,the penetration of chloride ions into concrete may be a threat to the durability of the structures.Experimental investigations were carried out to study the effect of chloride content on the bond behavior between concrete and fiber reinforced polymer (FRP)plates.Direct shear tests were conducted on the FRP strengthened concrete members.Before testing,the specimens were immersed in NaCl solutions with concentrations ranging from 3%—15%for different time (0—120d).Then,the specimens were dried and tested to obtain the initial and ultimate debonding loads,together with strain distributions along the FRP plates of different load values.The correlations between chloride content and debonding parameters are established.Test results show that the debonding parameters are closely related to the immersing time rather than the chloride content of the solution.Keyword s ：fiber reinforced polymer (FRP);concrete;chloride content;direct shear test;debonding behaviorWith the increasing need for strengthening and ret-rofitting concrete structures,external bonding of fiber reinforced polymer (FRP)to the tensile side of concrete members has been established as an efficient and effec-tive method to increase both strength and stiffness [1-3].In real applications,many structures may be located at theseaside,and the penetration of chloride ions(Cl -)into concrete from the seawater poses a threat to the durability of the FRP strengthened concrete structures.Many stud-ies have been performed to investigate the effect of steel corrosion on the strength and mechanical behavior of reinforced concrete (RC)members.It is found that lower degrees of corrosion,ranging from 0to 4%mass loss of steel bars,result in an increase in bond strength between the steel and concrete in the structures,while higher de-grees of corrosion,above 4%mass loss,lead to cracks and rapid decrease in bond strength [4-6].Fortunately,for the concrete structures strengthened by FRP composites,the bonding of FRP plates to the concrete member can stop or slow down the penetrationof Cl -into concrete from the tensile surface.For exam-ple,for the strengthened concrete beams,the bridgingforce in the bonded FRP plates can limit the opening of cracks and play an important role in improving the dura-bility of the strengthened concrete members [7-9].For a carbon fiber reinforced polymer(CFRP)wrapped con-crete beam,CFRP wraps can reduce the steel mass loss as well as the corresponding current [10].One test result shows that the CFRP wrapped RC beams,with 4.6%mass loss of steel,achieve a higher bond stress compared with the uncorroded beams [11].Another experiment indi-cates that the ultimate monotonic strength of the CFRP strengthened corroded RC beams increases to a level be-tween 37%and 87%above the predicted strength of simi-lar unstrengthened uncorroded beams [12].However,forthe unwrapped concrete members,Cl -can still diffuse to the steel surface from the side surfaces or cracked part of the concrete member [13].For the FRP strengthened con-crete members,the bond behavior between FRP and con-crete should be an important issue deserving to be fo-cused on.Based on experimental results,Yang et a l [14]summed up that the bond capacity of FRP-to-concrete interface decreased with the immersing time in the sea-Program o.2009C 2200and the E cellent oung eachers Program o outheast niversit .PA Jinlong born in 197male r associate Pro .Correspondence to PA Jinlong E-mail:i nlongp .Transactions of Tianjin University V ol.16No.62010—6—water,but the bond capacity had little correlation with the concrete compressive strength.Soudki et al [15]reported the results of 11RC beams strengthened with CFRP sheets and subjected to a severe environment (300wet-ting -drying cycles with 3%NaCl ),and found that the ultimate capacity of the CFRP strengthened beams de-creased by 11%—28%,but the stiffness and yield load of the RC beams were not affected by the environmental exposure [15].For the strengthened concrete structures,the durability of adhesive and CFRP composites should also be important problems for their durability.It is found that the extensibility of CFRP plates decreased a little but Young ’s modulus or tensile strength can hardly be af-fected by corrosive environments,such as high and low temperatures,freeze -thaw cycles,dry -humidity cycles [16].Epoxy resin is often used as adhesive for structural strengthening,and its compressive strength decreases with the immersing time in alkaline solution and natural water [17].In this paper,the direct shear test was conducted to investigate the correlation between FRP/concrete bond behavior and chloride content.1Experimental1.1Material propertie sIn this experiment,two batches of concrete prisms were cast for the direct shear tests,and the mixing pro-portions of the components were 1∶0.5∶1.5∶2(ratio of cement to water to sand to aggregate,C ∶W ∶S ∶A in short )and 1∶0.6∶1.5∶2by weight respectively.As for the dimensions,the specimens were 150mm in height,150mm in width,and 550mm in length.Some other concrete cubes were cast to test the compressive strength of concrete for the beam specimens.The elastic modulus of concrete (E C )is listed in Tab.1.Tab.1Material properties of concreteMixing proportion (C ∶W ∶S ∶A)Compressive strength/MPaYoung ’s modulus/GPaBatch 11∶0.5∶1.5∶217.2523.7Batch 21∶0.6∶1.5∶216.0622.9The FRP composites employed in the tests are the carbon fibers from HT300,Nippon Oil Corporation,Ja-pan.The Young ’s modulus of the FRP was tested to be 234GPa and the thickness was 0.165mm per ply accord-ing to the production specifications.The tensile modulus of the epoxy -based adhesive (TATSUHI Co.,Ltd.)used G 1.2Pre pa ration of spe cime nsFourteen specimens were tested for investigating the effect of chloride content on the bond behavior between FRP and concrete.Before applying the FRP composites,the surfaces of concrete prisms were roughened to expose the aggregate for increasing the bond between FRP and concrete.Epoxy resin,with resin/hardener ratio of 2to 1by weight,was used as both matrix for the FRP layers and adhesive between FRP and concrete.In succession,two layers of FRP sheets were bonded to the concrete prisms layer by layer by using the epoxy resin.The con-figuration and geometric dimensions of the prism speci-mens are shown in Fig.1.Four solutions with different concentrations of sodium chloride (NaCl )(3%,6%,10%,and 15%)were applied to the tests,which were much higher in concentration than NaCl in seawater (approxi-mately 2.72%),to speed up the test procedure.The con-crete prisms were entirely immerged into the solutions to ensure that the interface between FRP and concrete could be eroded.After certain days (0d,15d,30d,60d,90d and 120d ),as listed in Tab.2,the concrete prisms were taken out and totally dried for testing.For convenience of description,the notation of the specimens is described as follows.For example,B1-30-10,the first part “B1”de-notes that the specimen was cast with concrete of Batch 1,the second number “30”denotes that the specimen was immersed in the solution for 30d,and the last number “10”denotes that the specimen was immersed in a solu-tion with the Cl -concentration of10%.Fig.1Specimen dimensions1.3T e st setup a nd procedureBefore testing,the specimen should be carefully aligned to make sure that the pulling force is along the vertical plane of the FRP plate so as to reduce any possi-ble peeling effect (Fig.2).The test was conducted under f 540in the tests was tested to be 2.49Pa.displacement control with a loading rate o 0.0mm/min.PAN Jinlong et al:Effect of Chloride Content on Bond Behavior Between FRP and Concrete——Tab.2Environmental exposure situation of all specimensNaCl concentration15d30d60d90d120d0d3%B1-90-3,B2-90-36%B1-90-6,B2-90-610%B1-15-10*B1-30-10*,B1-30-10B1-60-10B1-90-10,B2-90-10B1-120-1015%B1-90-15,B2-90-15B1-0-0*Note:*represen ts that the direct shear test of specimens was conducted on the 120th day,while others on the 380th day.The load value from the material testing system (MTS )was automatically collected.To record the strain varia-tions along the FRP plate during the loading process,eight strain gauges were attached to the FRP plate with a center -to -center spacing of 30mm (Fig.2).Finally,the bonded strain gauges were connected to an automatic data logger (TDS303)to collect the strains during the loadingprocess.Fig.2Test setup2T est results and discussion2.1Fa ilure cha racterizationIn the experiments,the debonding failure of FRP plate was observed for all specimens.With an unbonded zone of 50mm at the pulled end of the FRP plate,wedge failure of concrete was avoided.With increasing external load,debonding started to occur from the loaded end of the FRP plate.After final failure,the FRP plate was en-tirely separated from the concrete prism with a thin layer of concrete attached to its surface.The initial and ulti-mate debonding loads of each specimen were recorded.To illustrate the debonding behavior during the test,the specimen B1-60-10was used as an example.The typical curve of load v ersus global displacement is shown in Fig.3,as well as the curve of load versus relative slip be-tween FRP and concrete collected from linear variable differential transformer (LVDT).The specimen B1-60-10was immersed for 60d in a solution with the chloride concentration of 10%.Before initial debonding (the mo-F 3),x increases linearly with the global displacement of the bonded FRP plate.At this stage,there are no cracks along the concrete/FRP interface and the stresses are transferred from the FRP plate to the concrete prism by the shear deformation of the adhesive layer.After the external load reaches around 13.7kN (the dashed line in Fig.3),initial debonding occurs at the concrete/FRP interface near the pulled end.Then,the external load continues to increase with the debonded zone due to the presence of interfacial friction and interlocking between the failure surfaces.At this stage,the external load increases with the global dis-placement with a smaller slope.The size of the debonded zone grows steadily with the increase of the external load until the load reaches 15.59kN.After that,the global displacement increases significantly,but the external load shows a sudden drop due to the occurrence of unstable debonding,indicating that the ultimate stage is reached.Similar observations are also obtained for other speci-mens during the loadingprocess.Fig.3Load versus global displacement curve for specimen B1-60-10Fig.4shows strain distributions along the FRP plates for specimen B1-60-10.Each curve corresponds to the strain distribution along the FRP plate at a certain load value.It is observed that when the load is lower than 13.7kN,the strains in the FRP plate decrease quickly with the distance from the loaded end,which can be ascribed to much smaller axial stiffness of the bonded FRP plate with respect to that of the concrete prism.Prior to initial debonding,the increase of applied load will cause the curve to shift upward.In this situation,there is no inter-f FR 407ment as the circles shown in ig.the e ternal loadacial crack between the P plate and the concreteTransactions of Tianjin University V ol.16No.62010——prism and the stresses are transferred from the FRP plate to the concrete by the shear deformation of the adhesive.However,from the curves corresponding to the load val-ues beyond 13.7kN,the shape of the curve starts to change and the slope of the curve near the loaded end tends to decrease due to interfacial debonding failure.Since the slope of the curve determines the rate of strain change in the FRP plate,which is proportional to the shear stress at the bottom of the FRP plate,the decrease of the slope represents shear softening at the debonded interface.As the load increases,the interfacial debonding tends to propagate towards the free end of the FRP plate,which can be indicated by the shifting of strain distribu-tion curves towards the free end of the FRP plate.After the external load reaches 15.59kN,the maximum strain in the FRP plate seems to remain approximately constant,indicating that unstable debonding occurs.Fig.4Strain variations at different loading values for speci-Fig.4men B1-60-102.2Effe ct of immersing tim e and conce ntra tion on com pr essive stren gth of concreteTo investigate the correlation between compressive strength of concrete and concentration or immersing time,concrete cubes were immersed in the same solu-tions as the concrete prisms and for the same time.After taken out and totally dried in the atmosphere,the cubes were tested to obtain the compressive strength which was correlated with the immersing time and chloride concen-tration,as shown in Fig.5.From Fig.5(a ),the compressive strength of concrete increased by 57.5%of the standard compressive strength (standard curing for 28d )after being cured for 380d in the atmosphere.When the cubes were cured in the salt solutions,the compressive strength increased linearly from 27.2MPa to 32MPa along with the immersing time.Due to the incompleteness of hydration of cement at the early stage,the strength of concrete increased with curing age and immersing time.Moreover,the hydrationcan be accelerated when the cube was immersed in water.Fig.5(b )shows that there is little correlation between concentration and compressive strength,indicating that the concentration of Cl -has imperceptible effect on the strength of concrete.(a)Concrete in10%solution(b)Concrete immersed for 90dFig.5Correlation between immersing time/concen tration an d compressive strength of concrete2.3Effect of imm ersin g time and chloride conce n-tration on initia l and ultimate debonding load Based on test results,the correlations between im-mersing time and initial debonding load (P i )or ultimate load (P ult )are plotted in Fig.6(a ).When the immersing time is less than 30d,P i and P ult decrease with the im-mersing time due to the degradation of adhesive in the salt solutions [18].In the cases with immersing time of 30d to 90d,P i and P ult both increase with the immersing time because of the increase of concrete strength result-ing from further hydration of cement.At this stage,the degradation of the adhesive becomes a secondary factor.After 90d,P i and P ult decrease again with the immersing time due to the degradation of both the adhesive and con-crete.Fig.6(b)shows the correlation between the con-centration,P i and P ult .According to the test results,the concentration of chloride seems to have little effect on the initial and ultimate debonding loads.Further experi-ments on specimens with a longer corrosion time are needed to clarify this effect.408PAN Jinlong et al:Effect of Chloride Content on Bond Behavior Between FRP and Concrete——(a)Concrete in 10%solution(b)Concrete immersed for 90dFig.6Correlation between immersing time/concentrationan d load2.4Effe ct of immersing tim e and conce ntra tion on interfa cial fracture ener gyIn most existing models,the FRP/concrete bond ca-pacity can be expected from the interfacial fracture en-ergy as well as FRP stiffness,thickness and length,etc.According to the model proposed by T ljsten [19],the in-terfacial fracture energy G F can be given by2u l t F 2P P P(1)2P G E b t α+=(1)where E P is the Young ’s modulus of FRP;α=E P t P b P /(E C t C b C )(E C is the Young ’s modulus of concrete );b P ,b C are the width of the FRP plate and concrete prism respec-tively;t P is the thickness of the FRP plate;t C is the depth of the concrete prism.G F for each specimen can be calcu-lated from Eq.(1).On the other hand,Pan et a l [20]proposed that the bond capacity is mostly related to the concrete surface tensile strength (f t ).And the general relationship between G F (N/mm)and f t is assumed to take the following form:F t bG af =(2)where a,b are constant parameters;f t can be calculated bycompressive strength of concrete (f c )with empirical formula;and G F can be calculated by Eq.(1).A database of 25test results from direct shear test is employed to obtain the constants in Eq.(2)through the approach of least squares fitting.Based on the theory of ,f then calculated as a=0.387,and b=0.622.Hence,the resulting empirical equation for G F is0.622F t 0.387G f =(3)According to Eq.(1)and Eq.(3),the ratio of the predicted fracture energy (calculated by Eq.(3))to the experimental interfacial fracture energy (calculated by Eq.(1))can be figured out,as shown in Fig.7.Fig.8shows the correlation between G F and the immersing time,which is similar to that between P ult and the im-mersing time.G F decreases at the initial stage because of the deterioration of adhesive,and increases appreciably during 30d to 90d due to the hydration of cement,and thereafter decreases again until 120d as a result of the degradation of the adhesive and concrete.Since the con-centration of Cl -has little effect on the concrete strength and the ultimate bond capacity between FRP and con-crete,it also shows little correlation with the interfacial fracture energy.Fig.7Comparison of fracture energy from empirical equa-tion and experimental resultFig.8Correlation between immersing time and G F3ConclusionsIn this paper,direct shear test was conducted to in-vestigate the effect of chloride content on the bond be-havior between concrete and the FRP plates.Based on test results,the following conclusions can be drawn.(1)The concrete compressive strength increases f y N 409multiple nonlinear regressions the values o constants aresigni icantl with the immersing time in the aCl solu-Transactions of Tianjin University V ol.16No.62010——tions.The concentration of the chloride ions has little effect on the compressive strength of concrete.(2)Both the initial debonding load (P i )and ulti-mate debonding load (P ult )show a small decrease due to the degradation of the adhesive in the NaCl solution at the early stage.Then,both of them increase with the im-mersing time due to further hydration of cement.Finally,P i and P ult decrease with the immersing time because of the degradation of the adhesive and concrete in severe conditions.(3)Based on test results,the correlation between G Fand the immersing time shows the same trend as that be-tween P i (or P ult )and immersing time.The interfacial fracture energy has little correlation with the concentra-tion of Cl -in the NaCl solution.Referen ces［1］Meier U,Deuring M,Me ier H et a l.Strengthening ofstructures with CFRP laminates:Research and applications in Switzerland [J].Advanced Composite Materials in Bridges and Structures,1992,58(3):33-77.［2］Ritchie P A,Thomas D A,Lu L W et al.External strength-ening of concrete beams using fiber-reinforced plastics [J].ACI Struct J,1991,88:490-500.［3］Saadatmanesh H,Ehsani M R.RC beams strengthenedwith GFRP plates (I):Experimental study [J].J Struct Eng,1991,117(11):3417-3433.［4］Almusallam A A,Al-Gahtani A S,Rasheeduzzafar A R.Effect of reinforcement corrosion on bond strength[J].Construction and Building Materials,1996,10(2):123-129.［5］Al-Sulaimani G J,Kaleemullah M,Basunbul I A et al.Influence of corrosion and cracking on bond behaviour and strength of reinforced concrete members[J].ACI Struc-tural Journal,1990,87(2):220-231.［6］Cabrera J G.Deterioration of concrete due to reinforce-ment steel corrosion[J].Cement and Concrete Compos-ites,1996,18(1):47-59［7］Teng J G,Chen J F,Smith S.FRP-Strengthened RC Struc-tures [M].John Wiley and Sons,Ltd,2001.［8］Arduini Marco,Di Tommaso Angelo,Nanni Antono.Brit-tle failure in FRP plate and sheet bonded beams [J].A CI Structural Journal,1997,94(4):363-370.［9］Norris T,Saadatmanesh H,Ehsani M R.Shear and flexuralstrengthening of RC beams with carbon fiber sheets [J].Journal of Structural Engineering,A SCE,1997,123(7):903-911.［10］EI Maaddawy Tamer,Chahrour Ali,Soudki Khaled.Effectof fiber-reinforced ploymer wraps on corrosion activity and concrete cracking in chloride-contaminated concrete cylinders [J].Journal of Composites for Construction,ASCE,2006,10(2):139-147［11］Craig B C,Soudki K A.Confining effects of CFRP lami-nates on corroded concrete members [C].In:Brahim Benmokrane,Proceedings of the 2nd International Con-ference (CDCC ′02).Montreal,Canada,2002.383-394.［12］Masoud Sobhy,Soudki Khaled,Topper Tim.CFRP-strengthened and corroded RC beams under monotonic and fatigue loads [J].Journal of Composites for Construction,ASCE,2001,5(4):228-236.［13］Hanaoka D,Miyazato S,Moriwake A.Fundamental studyon repair technology of reinforced concrete [C].In:Pro-ceedings of the 2nd International Conference on Protec-tion of Structures Against Hazards.Singapore,2004.183-190.［14］Yang Yongxin,Guo Chunhong,Cai Peng et al.The effectof sea water on CFRP/concrete interface [J].Industrial Construction,2006,136(18):10-12(in Chinese).［15］Soudki Khaled,EI-Salakawy Ehab,Craig Brent.Behaviorof CFRP strengthened reinforced concrete beams in corro-sive environment [J].Journal of Composites for Construc-tion,A SCE,2007,11(3):291-298.［16］Yang Yongxin,Wang Jing,Zhang Xiaodong.Research fordurability of concrete structures strengthened with carbon fiber reinforced polymer sheets [J].Port Engineering Technology,2002(2):25-27(in Chinese).［17］Cai Guangting,Chen Junqiu,Zou Yue et al.Study on thedurability of structural adhesive used to adhere steel for strengthening concrete [J].Concrete,2000(8):37-40.［18］Hu Haitao,Yang Yongxin,Guo Chunhong et al.The effectof the sea water on the durability of resin [J].Industrial Construction,2006,136(18):13-15(in Chinese).［19］T ljsten B.Strengthening of concrete prisms using theplate bonding technique [J].Int J Fract,1996,82(1):253-266.［20］Pan Jinlong,Leung C K Y.Effect of concrete compositionon FRP/concrete bond capacity [J].Journal of Composites for Construction,A SCE,2007,11(6):611-618.410。