Chloride Ingress and Reinforcement Corrosion in Concrete under De-Icing Highway Environment - A study after 10 years' field exposure.

Full text

(1)Chloride Ingress and Reinforcement Corrosion in Concrete under De-Icing Highway Environment – A study after 10 years’ field exposure. SP Technical Research Institute of Sweden. Tang Luping Peter Utgenannt. Building Technology and Mechanics SP Report 2007:76.

(2)

(3) Chloride Ingress and Reinforcement Corrosion in Concrete under De-Icing Highway Environment – A study after 10 years’ field exposure Tang Luping Peter Utgenannt.

(4) 4. Abstract Chloride Ingress and Reinforcement Corrosion in Concrete under De-Icing Highway Environment – A study after 10 years’ field exposure This report presents the results from a research project financed by the Swedish Road Administration. In this project about 100 chloride and moisture profiles have been measured from various types of concrete specimens exposed to a de-icing highway environment for about 10 years. A newly developed rapid non-destructive technique, RapiCor, for corrosion measurement was used to assess the conditions of steel embedded in concrete beams with different types of binder and water-binder ratios. Both the DuraCrete and ClinConc model were used to predict chloride ingress in concrete. The results show that chloride profiles measured after 10 years’ exposure under the deicing highway environment are, in some cases, lower than those measured after the first 1~2 years’ exposure. Application of new techniques for spreading de-icing salts and the coverage of specimens by the shovelled snow lumps might be possible reasons for the low chloride ingress. The DuraCrete model, when the current input data given in the guidelines are used, may significantly underestimate chloride ingress, while the ClinConc model in general gives better prediction results, but it contains a number of empiric parameters or factors which need to be further verified. Owing to the large scatter in chloride profiles, none of the present models can, so far, properly describe the chloride ingress under such a de-icing highway environment. Non-destructive corrosion measurement by RapiCor instrument is in general reasonably in fairly good agreement with chloride ingress. The corrosion rust observed from the destructive examination verified again that the non-destructive technique RapiCor is a useful tool for detection of ongoing corrosion of steel in concrete.. Key words: Chloride, concrete, corrosion, durability, moisture, reinforcement. SP Sveriges Tekniska Forskningsinstitut SP Technical Research Institute of Sweden SP Report 2007:76 ISBN 978-91-85829-09-5 ISSN 0284-5172 Borås 2007.

(5) 5. Sammanfattning Kloridinträngning och armeringskorrosion i betong exponerad för tösaltad vägmiljö - Resultat efter 10 års exponering vid Riksväg 40 Bakgrund Armeringskorrosion är en av de vanligaste orsakerna till skador på betongbroar. Normalt anses korrosion initieras av antingen kloridinträngning eller karbonatisering. För bägge dessa mekanismer är betongens motstånd mot klorider och dess fuktnivå av avgörande betydelse för om och när korrosion startar. För att kunna dimensionera betongkonstruktioner på ett ekonomiskt och säkert sätt samt göra livslängdsbedömningar på redan befintliga konstruktioner behövs modeller för hur klorider och fukt transporteras i betong. För att dessa modeller skall bli robusta och trovärdiga måste de baseras på resultat från forskning på betongkonstruktioner exponerade i fält. Ett sådant forskningsprojekt är det nationella projektet BMB ”Beständighet Marina Betongkonstruktioner” som startade 1991 med syfte att få fram beständighetsdata för betong exponerad i marint klimat. Detta projekt, som Vägverket aktivt deltog i, resulterade bland mycket annat i ett användarvänligt datorprogram (Clinconc) för beräkning av kloridinträngning i betong i marint klimat. Under 1996 startade ytterligare ett nationellt projekt BTB ”Beständighet Tösaltade Betongkonstruktioner” som har finansierats av Vägverket och Cementa för att studera hur det aggressiva vägklimatet vid Riksväg 40 påverkar betongen. Ett drygt 30-tal betongkvaliteter med olika cementtyper, restmaterial som slagg och flygaska, olika vattenbindemedelstal och olika lufthalter tillverkades och placerades ut på en fältprovplats utmed Riksväg 40 strax väster om Borås. Den första uppmätningen av kloridinträngning utfördes efter två vintersäsonger och den påföljande efter fem säsonger på tre utvalda betongkvaliteter. Efter tio års exponering stöttade Vägverket detta forskningsprojekt för att utföra en omfattande kartering av kloridinträngningen och fuktnivåerna hos samtliga betongkvaliteter. Dessa mätningar har kompletterats med uppmätning av pågående armeringskorrosion med en icke-förstörande mätteknik (RapiCor) samt genom att armeringsjärn frilagts för att kunna bedöma korrosionsomfattningen. Syftet med detta projekt är att • ta fram underlag för livslängdsdimensionering av betongkonstruktioner utsatta för tösaltmiljö, • förbättra kunskap om armeringskorrosion i samband med kloridinträngning och fuktvandring hos betong, • kalibrera befintliga beräkningsmodeller för kloridinträngning, • öka kunskapen om armeringskorrosion, och. verkliga. tröskelvärden. för. kloridinducerad. • kalibrera icke-förstörande metoder för mätning av pågående armeringskorrosion. Denna rapport redovisar resultaten från forskningsprojektet..

(6) 6. Betongprovkroppar och fältexponering Två typer av betongprovkroppar – ett ren betongblock med storlek 400×300×300 mm och en armerad betongbalk med storlek 1200×300×300 mm – var gjutna på SP. Huvudvariationerna i betongsammansättningarna inkluderar • vattenbindemedeltal (vbt = 0,3; 0,35; 0,4 och 0,5 samt en upp till 0,75); • bindemedel (åtta typer med olika tilläggningar av kalksten, kiselstoft och slagg); och • lufthalt (5 % luftpor och naturlig lufthalt). Betongprovkropparna placerades nära körbanan på en fältprovplats vid Riksväg 40 strax väster om Borås och exponerades för den aggressiva miljö som omger en tösaltad motorväg. Ett flertal av provkropparna har ingjuten armering med olika täckskikt och många av dessa är också tillverkade med sprickor. Riksväg 40 mellan Borås och Göteborg är en av Sveriges mest tösaltade vägar. Enligt Vägverkets informationer har en genomsnittlig mängd tösalt 2,2 kg per m² vägyta använts under 1996-1999 men den genomsnittliga mängden tösalt halverades efter 2000 tack vara en ny metod för saltning med mindre mängder salt i saltlaken.. Icke-förstörande mätteknik för bedömning av korrosionstillstånd Den nyutvecklade icke-förstörande mättekniken s k RapiCor har använts i detta projekt för att bedöma armeringsjärnets korrosionstillstånd. Tekniken är baserat på senare års forskningsresultat från ett par FoU projekt, bl a ett Vägverksprojekt. Liksom andra metoder för mätning av armeringskorrosion är RapiCor baserad på galvanostatisk polarisationsteknik. På armeringsytan finns det ett tunt ytskikt av järnoxid som i vanliga fall har mycket högt polarisationsmotstånd och skyddar armeringen från korrosion. När skyddsskiktet har brutits ner på grund av kloridangrepp eller karbonatisering, bildas ett rostigt område på armeringsytan. Den rostiga arean har då lågt polarisationsmotstånd och således fortsätter korrosionsprocessen. Genom att leda en känd galvanostatisk strömstyrka från betongytan till armeringsjärnet och samtidigt mäta potentialsignaler får man veta betong-armeringsjärnkretsens polarisationsmotstånd. Korrosionshastighet kan sedan beräknas ur polarisationsmotståndet. Inte bara korrosionshastigheten utan även armeringsjärnets halvcellpotential och betongens resistivitet kan uppskattas genom RapiCor tekniken. Halvcellpotentialen är en indikation på korrosionssannolikhet enligt ASTM C 876 och resistiviteten återspeglar betongens fuktighet. Ur de tre parametrarna – korrosionshastighet som huvudparameter, halvcellpotential och resistivitet som kompletterande parametrar – får man en säkrare bedömning av armeringsjärnets korrosionstillstånd med uttryck korrosionsgrad enligt nedanstående tabell..

(7) 7. Tabell: Kriterier för uppskattning av korrosionsgrad. Korrosionsgrad Korrosions- HalvcellResistivitet Kriterier hastighet potential [kΩ⋅cm] [µm/år] [mV(CSE)] < XL XL = 1 µm/år 1 Försumbar XL ∼ XM ≥ Ecr ≥ ρcr XM = 3 ~ 5 µm/år* XL ∼ XM < Ecr < ρcr XH = 10 µm/år 2 Lite XM ∼ XH ≥ Ecr ≥ ρcr Ecr = -200 mV(CSE) XM ∼ XH < Ecr < ρcr ρcr = 100 ~ 120 kΩ⋅cm† 3 Måttlig > XH ≥ Ecr ≥ ρcr < Ecr < ρcr 4 Påtaglig > XH * 3 för medelvärde av mätarea och 5 för enskild mätning. † beroende av betongkvalitet, ytbehandling, väder, etc. Korrosionsmätningar utfördes på fältplatsen under två väderförhållanden – torrt och vått. Enligt fältmätningsresultaten togs fyra betongbalkar ut, d v s 202 BB1(svenskt anläggningscement, vbt 0,5), 223 B1 (finskt snabbt cement, vbt 0,5), 236 AB1 och 236 BB1 (svenskt anläggningscement, vbt 0,75). De fördes till SP för förstörande mätningar inklusive okulär undersökning av korrosionstillståndet på de frigörliga armeringsjärnen, karbonatiseringsdjup och kloridhalt på täckskiktsnivåer (10, 20 och 30 mm) för att verifiera de resultat som icke-förstörande tekniken har tagit fram. Resultaten visar god överenskommelse mellan den icke-förstörande mätningen och den förstörande undersökningen. Rostiga fläckar eller märken upptäcktes på armeringsjärnet där korrosionsgraden var 3-4 (måttlig respektive påtaglig korrosion pågår) enligt den icke-förstörande mätningen. Enligt den okulära undersökningen på de ospruckna betongbalkarna initierades korrosionen på armeringsjärnets undersida, där det ofta finns stora luftporer eller cementpasta med hög vattenhalt på grund av möjlig segregation under armeringsjärnet. På de spruckna betongbalkarna skedde korrosion på platser med bredda sprickor. Korrosionsmätningen under vått väder visar generellt högre korrosionsgrad än under torrt väder. Skillnaden i korrosionsgrad mätt mellan det torra och våta vädret är liten för den ospruckna betongen men relativt stor för den spruckna betongen. Det kan förstås att vatten är en av nödvändiga förutsättningar för korrosionsprocessen. Torrt väder kan lätt torka vattnet i sprickor och minska eller tillfälligt stoppa korrosionsprocessen. Karbonatiseringsdjupet är ca 5 mm på betongbalkar med vbt 0,75 (236 AB1 och BB1) respektive 2 mm i betongbalkar med vbt 0,5 (202 BB1 och 223 B1) efter 10 års exponering under vägmiljön på fältplatsen. Placering nära marken där fuktigheten är relativt hög är en möjlig förklaring till det relativt låga karbonatiseringsdjupet. Det framgår från kloridmätningsresultaten att högre kloridinträngning upptäcktes på övre delen än undre delen av betongbalkens vertikala yta. Möjlig orsaken kan vara att de plogade snövallarna har blockerat vidareskvätt från bilar.. Mätning av klorid- och fuktprofiler I denna undersökning har ett 100-tal klorid- och fuktprofiler mätts upp ur 34 st betongblock på betongsammansättningar med olika bindemedel och olika.

(8) 8. vattenbindemedelstal. På varje betongblock mättes två kloridprofiler – en från den övre horisontella exponeringsytan och en från den vertikala exponeringsytan – och en fuktprofil från den vertikala exponeringsytan. Mätning av kloridprofiler utfördes av SP med hjälp från Chalmers Tekniska Högskola för profilfräsning. Fuktmätning utfördes av Lunds Tekniska Högskola. Resultaten visar att kloridinträngningen efter 10 års exponering i några fall är mindre än vid uppmätningen efter två år. En orsak till detta kan vara att det nu används andra spridningstekniker som medför att mindre mängd salt sprids på vägarna idag jämfört med för 10 år sedan. En annan orsak kan vara att kloriderna i ytan lakas ur under sommarhalvåret. Ytterligare en orsak kan vara att det finns lokala variationer, t ex snövallar och olika exponering mot skvätt från bilar vilket medför att kloridinträngningen kan variera något beroende på provkropparnas placering på provplatsen. Som förväntat har betongens vbt stor betydelse för kloridinträngningen. Betongen med lägre vbt visar i allmänhet mindre kloridinträngning. Normalt skulle pozzolantillsatsmaterial såsom kiselstoft och slagg bidra till bättre porstruktur och få till följd mindre kloridinträngning. Resultaten från detta projekt visar emellertid inte tydligt pozzolaneffekten. Sannolikt har den minskande saltspridningen bromsat vidare kloridinträngning men regnväder har lett till mer urlakning av klorider från betongen med enkelt cement än med pozzolantillsatsmaterial p g a den senares låga permeabilitet. Det framgår från uppmätta fuktprofiler att betongen med tillsatsmaterial visade lägre fuktprofiler än utan betongen med svenskt anläggningscement och finskt snabbt standardcement.. Modellering av kloridinträngning under tösaltad vägmiljö Resultat från uppmätt kloridinträngning har jämförts med olika matematiska modeller för kloridinträngning såsom enkel ERFC modell, DuraCrete modell och ClinConc modell. ERFC modellen är baserad på enkel Ficks 2:a lag med ett komplement till fel-funktionen (ERFC). Det är beprövat att denna modell ofta leder till en ”för konservativ” prediktion av kloridinträngning. Därför har ERFC modellen använts endast för beräkning av skenbar diffusionskoefficient och kloridhalt på betongytan. DuraCrete modellen togs fram i EUprojektet DuraCrete som avslutades 2000 och är baserad på ERFC modellen med modifikationer av flera parametrar. ClinConc modellen har utvecklats på Chalmers i mitten av 1990-talet och är baserad på de fysikaliska och kemiska processer som involveras i kloridtransportsteorier. Denna jämförelse visar att DuraCrete modellen i vissa fall kraftigt undervärderar kloridinträngningen. Jämförelse med ClinConc modellen ger generellt bättre överensstämmelse. Denna modell innehåller ett flertal empiriska parametrar som behöver verifieras. Beroende på den relativt stora spridningen i uppmätta kloridprofiler kan ingen existerande modell på ett tillräckligt säkert sätt beskriva kloridinträngningen i denna tösaltade vägmiljö.. Slutsatser och rekommendationer Från jämförelsen mellan korrosionsmätningsresultaten och kloridinträngningsprofiler kan man dra följande slutsatser och rekommendationer: •. Liksom i tidigare undersökningar visade friläggning av korroderade armeringsjärn att den snabba icke-förstörande tekniken RapiCor stämmer bra med den förstörande mätningen. Resultaten från den icke-förstörande korrosionsmätningen uppvisar också bra överensstämmelse med uppmätt.

(9) 9. kloridinträngning. Därför är denna snabba teknik ett användbart redskap och kan rekommenderas för bedömning av pågående armeringskorrosion. •. Resultaten från denna undersökning tyder på att kloridtröskelvärdet inte nödvändigtvis är 0,4 % av bindemedelsvikten vilket är ett vanligt antagande. I denna undersökning uppvisar inte armering i betong med pozzolantillsatsmaterial (slagg och kiselstoft) någon tendens att börja korrodera vid lägre kloridtröskelvärden än vad som antas gälla för betong med rena Portlandcement som bindemedel. Å andra sidan uppvisar den icke-förstörande tekniken en viss grad av korrosion i betongen med byggcement och finskt snabbt cement även när kloridhalt i närheten av armeringsjärnet är lägre än 0,4 % av bindevikt. Därför behöver kloridtröskelvärden för betong med och utan tillsatsmaterial undersökas vidare innan några säkra slutsatser skall kunna dras.. •. Tillsats av 50 % slagg i finskt snabbt cement visar betydlig förbättring av betongens resistens mot kloridinträngning och armeringskorrosion medan tillsats av 10~15 % slagg i finskt standardcement inte visar sådan förbättring. Å andra sidan måste frostbeständighet hos betong med tilläggning av hög volym slagg provas innan denna typ av betong används för tösaltade vägmiljöer.. Ett positivt resultat är att tösalter tränger in i betongen med en reducerad hastighet efter bara ett par års exponering. För att få bekräftat att dessa resultat är representativa även för andra tösaltade vägmiljöer föreslås följande: •. Undersökning av några existerande motorvägsbroar för att mäta upp kloridinträngningen och på så sätt verifiera den reducerade inträngningshastigheten.. •. Fortsätta verifiera och modifiera existerande modeller för kloridinträngning så att dessa stämmer med uppmätt kloridinträngning på verkliga betongkonstruktioner och så att modellerna kan användas vid livslängdsbedömning.. •. Påbörja forskning kring kloridtröskelvärden där inte enbart kloridhalten beaktas utan även andra faktorer så som karbonatisering och lakning av kalcium som påverkar kloridbindningen och alkaliteten. Även inverkan av fuktinnehåll, luftblåsor i betongen och defekter på armeringsjärnets yta bör studeras..

(10) 10. ..

(11) 11. Contents Abstract. 4. Sammanfattning. 5. 1. Introduction. 13. 2. Concrete Specimens and Exposure Conditions. 14. 2.1 2.2. Concrete specimens Exposure conditions at the highway 40 field site. 14 16. 3. Corrosion Measurement. 19. 3.1 3.2 3.3 3.4 3.4.1 3.4.2 3.4.3 3.4.4 3.4.5. Technique for corrosion measurement Technique for corrosion measurement Field measurements Laboratory measurements Concrete beam 202 BB1 Concrete beam 223 B1 Concrete beams 236 AB1 and BB1 Carbonation depths in concrete covers Chloride ingress at the cover levels. 19 20 21 22 23 23 26 29 30. 4. Measurements of Chloride and Moisture Profiles. 32. 4.1 4.2 4.3 4.4 4.4.1 4.4.2 4.5 4.5.1 4.5.2. Sampling Measurement of chloride profiles Measurement of moisture profiles Measured chloride profiles Effect of binder types Effect of water-binder ratios Measured moisture profiles Effect of binder types Effect of water-binder ratios. 32 32 33 33 33 38 40 40 41. 5. Modelling of Chloride Ingress. 43. 5.1 5.2 5.3 5.3.1 5.3.2 5.3.3 5.4. 43 45 46 46 49 49. 5.6 5.6.1 5.6.2 5.6.3 5.6.4. Curve-fitting to the ERFC model DuraCrete model ClinConc model Modelling of free chloride ingress Calculation of total chloride content Input parameters used in the modelling Modelled results in comparison with the results from the field measurement of corrosion Other chloride profiles in comparison with the results from the field measurement of corrosion Discussions DuraCrete model ClinConc model Uncertainty in chloride profiles Chloride threshold value for corrosion. 6. Concluding remarks and suggestions. 71. 7. References. 73. 5.5. 50 60 67 67 67 68 70.

(12) 12. Appendix 1 – Results of corrosion from the field measurement. 75. Appendix 2 – Chloride profiles from the Highway 40 field site. 81. Appendix 3 – Moisture data from the Highway 40 field site. 149. Appendix 4 – Curve-fitted results of apparent diffusion coefficient and surface chloride content. 154.

(13) 13. 1. Introduction. Reinforcement corrosion is one of the most common reasons for the damage of concrete bridges. It is generally believed that reinforcement corrosion is initiated by either chloride ingress in concrete or carbonation of concrete. Moisture is a prerequisite for both mechanisms of chloride transport and carbonation. In order to design concrete structures based on the expected service life and performance, models are needed to predict chloride ingress and moisture conditions in concrete. In the middle of 1990’s, the first mechanismbased model ClinConc has been developed at Chalmers University of Technology [Tang & Nilsson, 1994; Tang 1996]. This model has been verified with the field data after exposure under a marine environment over 10 years Tang, 2003b], under the financial support by the Swedish Road Administration. In the beginning of 2000’s, also under the financial support by the Swedish Road Administration, a rapid technique for corrosion measurement was developed at SP Technical Research Institute of Sweden [Tang, 2002]. This technique has been verified on the laboratory specimens and the reinforced concrete slabs exposed under a marine environment for over 13 years [ Tang et al, 2005]. Since 1996 a large number of reinforced concrete specimens with different qualities have been exposed at the field station by Highway 40. Measurements of chloride and moisture profiles have been made after 1 and 2 years exposure. In 2006 the specimens were exposed for 10 years. In order to investigate these specimens and obtain the “first-hand” information about the long term behaviour of concrete with regard to chloride ingress and reinforcement corrosion under the de-icing highway environment, the Swedish Road Administration financed this project. It was a good opportunity to use these unique specimens for measurements of chloride profiles and moisture profiles, and also for the verification of prediction models and non-destructive techniques. Therefore, the primary objectives of this project include • • • •. to form the basis for service life design of concrete structures exposed to de-icing environment; to improve the knowledge of reinforcement corrosion related to chloride ingress and moisture condition in concrete; to verify the existing models for prediction of chloride ingress; and to verify the non-destructive technique for measurement of on-going corrosion.. This report presents the results from the above mentioned research project (Contract No. AL 90 B 2005:16860), financed by Vägverket – Swedish Road Administration..

(14) 14. 2. Concrete Specimens and Exposure Conditions. 2.1. Concrete specimens. The relevant mixture proportions of concrete are summarised in Table 2.1 and the detailed information about the raw materials and hardened properties of each mixture proportion was published elsewhere [Utgenannt, 1998]. The main variations include water-binder ratio (0.3, 0.35, 0.4 and 0.5, one up to 0.75), binder type (eight types of binder with different additions of limestone, silica fume and blast-furnace slag), and air content (5% entrained air and non-AEA). Two types of concrete specimens, one plain concrete block of 400×300×300 mm and another reinforced concrete beam of 1200×300×300 mm, were cast at SP Technical Research Institute of Sweden (previously Swedish National Testing and Research Institute). The plain concrete blocks were designed for sampling of chloride penetration profiles and the reinforced concrete beams were designed for testing corrosion resistance under uncracked and pre-cracked conditions. A typical structure of the reinforced concrete beam is shown in Figure 2.1 and the detailed information about the reinforcement placement in each beam was published elsewhere [Nordström et al, 1998]. The specimens were cured in the laboratory for 35 to 70 days before placed at the field site.. Reinforcement bars. B. Stainless steel 60 30 15. B Pre-cracks. Exposure zone. B–B Epoxy coating. Notes: 1) In most of the beams there was no rebar at the level with cover 60. 2) Each steel bar was soldered with a copper wire whose another end was connected to the plinth in the electronic box outside the specimen. This wire supplies the connection to the rebar for electrochemical measurement.. Figure 2.1 – Schematic of reinforcement beams exposed to the Highway 40 field site..

(15) 15. Table 2.1 Mixture proportions of concrete exposed at the Highway 40 field site. Mix No.. 201 202 203 204 205 236 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232. Binder type Binder kg/m3. 100% Anl5). 95% Anl +5% SF6). 100% FinStd7). 100% SliteStd8) 56% FinRpd9) +44% SL10) 100% FinRpd 90% Anl +10% SF. P Kalk C11). 420 380 500 450 380 260 420 380 500 450 380 420 380 540 390 420 390 520 410 420 370 540 420 380 540 420 500 450 420 390 530 470 400. WaterFine Coarse binder aggreg. aggreg. ratio1) 0-8 mm 8-16 mm kg/m3 kg/m3 0.4 0.5 0.3 0.35 0.5 0.75 0.4 0.5 0.3 0.35 0.5 0.4 0.5 0.3 0.5 0.4 0.5 0.3 0.5 0.4 0.5 0.3 0.4 0.5 0.3 0.4 0.3 0.35 0.4 0.5 0.3 0.35 0.5. 886.4 890.2 833.5 880.3 938.6 1007.4 860 865.5 806 846.5 905.8 863 885 767.7 915.2 880.3 874.1 799.9 892.3 858 891.8 761.9 863 885 767.7 851.5 796.5 844.1 880.3 874.1 781.2 851 903.8. 851.6 821.8 978.5 953.7 866.4 791.6 860 831.5 985.1 954.5 870.2 863 817 938.3 844.8 845.7 806.9 939.1 823.7 858 823.2 931.2 863 817 938.3 851.5 973.5 951.9 845.7 806.9 954.8 922 834.2. Sp2) % of binder 0.97. AEA3) Air % of % of binder vol. 0.028 0.012. 3.25 2.27. 1.2 1.7 1.5 0 1.2 0 3.6 0 1.7 0 3.8 0 1.15 0 3.8 2 0 4.15 1.42 2.1 1.7 1.7 0 3.6 2.62 0. 0.012 0.04 0.022 0 0 0 0.02 0.008 0 0 0.027 0.01 0 0 0.027 0.01 0.01 0.027 0.013 0 0.08 0 0 0.03 0.014 0 0 0. 4.8 4.5 1.3 0.9 1.3 4.5 4.7 4.4 1.2 1.1 0.9 4.9 4.5 2.5 1.2 4.9 4.4 2.3 1.4 4.5 4.7 1.6 4.7 4.9 2.4 4.8 0.5 0.9 4.8 4.7 2.2 1.8 1.2. 28d compr. strength4) MPa 65.4 50.8 100.6 91.3 54.8 30 78.1 58.2 119.7 103.5 62.6 48.8 42.1 66.9 47.4 60.7 46 85.7 59.9 55 41.5 82 57.6 43.4 66.5 81.4 126.9 112.7 68.9 51.8 98.5 86.3 60.9. 1) Mass ratio of water to binder without consideration of the efficiency factor for silica fume or blast-furnace slag. 2) Sp – Super-plasticizer, Cementa Melcrete 3) AEA – Air entraining agent, Cementa L17 4) According to SS 13 72 10 5) Anl – Anläggningscement (Swedish SRPC, CEM I) 6) SF – Silica fume (Elkem. Norway) 7) FinStd – Finnish standard Portland cement with 10-15% blast-furnace slag (CEM III/C) 8) SliteStd – Swedish standard Portland cement with 5-8% limestone filler, made in Slite (CEM II/A-LL) 9) FinRpd – Finnish rapid Portland cement 10) SL – Finnish blast-furnace slag 11) P Kalk C – Swedish Portland cement with 10-15% limestone filler (CEM II/A-LL).

(16) 16. 2.2. Exposure conditions at the highway 40 field site. The field exposure site at highway 40 was established in the autumn of 1996. It consists of a 200 meter long and a couple of metres wide gravel area along the highway, with specimens mounted in steel frames at road level, as shown in Figure 2.2. A guard rail was installed to separate the exposure site from the traffic. It was placed in such a way as to ensure the traffic safety and to have the specimens fully exposed to the splash water from the traffic. The climate around the specimens is moist, and the specimens are exposed during the winter to low temperatures and de-icing salts, producing a climate corresponding to exposure class XD 3/XF 4 in EN 206-1 (2001).. Concrete blocks & beams. 0.45 m. 0.76 m. Guard rail. Gravel layer. Figure 2.2 – Field exposure site at Highway 40. Specimens placed in steel frames behind a guard rail (right)..

(17) 17. Highway 40 leads from Gothenburg to the east, through Borås and towards Jönköping. Over the year, the daily average number of vehicles passing the field exposure site is around 12000, of which 1250 are heavy vehicles (data from measurements carried out by the Swedish Road Administration in 2000). For safety reasons, de-icing salts are used during the winter in many parts of Sweden to keep road surfaces free from snow and ice. The de-icing agent used is sodium chloride, which is spread either in the form of a solution (about 24% NaCl) as a preventive measure, or as crystals when spread on snow. In this region, de-icing salts are normally used between October and April. Table 2.2 shows an estimate of the total amount of salt spread on the highway per square metre and year. The figures in Table 2.2 are based on the data from the Swedish Road Administration. The table also shows the number of occasions de-icing salts were spread on the highway each winter season. As can be seen in Table 2.2, the amount of salt spread on the road was markedly reduced around year 2000. This was due to the introduction of the new method of applying salt, as a solution, which uses a smaller amount of salt. Table 2.2 –. Estimated total annual amount of salt spread on Highway 40 per square metre and the number of occasions de-icing salts were spread on the road each winter season.. Winter. 96-97 97-98 Amount of salt (kg/m ) 1.9 2.4 Number of occasions 126 157 Winter 02-03 03-04 Amount of salt (kg/m2) 1.1* 1.3* Number of occasions 128 156 * Estimated from the number of occasions. 2. 98-99 2.3 151 04-05 1.3* 151. 99-00 2.1 141 05-06 1.2* 141. 00-01 1.1 117 06-07 1.1* 123. 01-02 1.2 148. The annual precipitation data between 1996 and 2002 obtained from the climate station about 10 km from the exposure site, run by the Swedish Meteorological and Hydrological Institute (SMHI), are shown in Table 2.3. It can be seen that the average annual precipitation is about 1170 mm.. Table 2.3 – Annual precipitation data from SMHI. Year Precipitation, mm. 1996 796. 1997 905. 1998 1256. 1999 1859. 2000 1323. 2001 904. 2002 1127. The monthly air temperature registered at the climate station near the field exposure site is shown in Figure 2.3. From these data the annual average temperature of 7 ºC can be estimated. If the freezing period is excluded, the average temperature will be about 10 ºC. It should be noticed that the actual chloride concentration in the highway environment is unknown, although some data of salt spread is available as shown in Table 2.2. Wirje &.

(18) 18. Offrell (1996) investigated chloride ingress into mortar specimens placed at different locations along the road. The results showed that the ingress of chlorides decreases with increasing height above the road level. Tang & Utgenannt (2000) present results from collecting splash water at different locations around the exposure site. The results confirm the findings by Wirje & Offrell (1996).. Monthly air temperature, °C. 20 15 10 5 Winter 95/96 0 Winter 96/97 -5 Winter 97/98. Figure 2.3 – Temperature data near the field exposure site at Highway 40.. Sep.. Aug.. July. Jun.. May. Apr.. Mar.. Feb.. Jan.. Dec.. Nov.. Oct.. -10.

(19) 19. 3. Corrosion Measurement. 3.1. Technique for corrosion measurement. In this project, a handheld instrument RapiCor was used for measurement of corrosion of rebars in concrete, see Figure 3.1. The detailed descriptions of this measurement technique have been published elsewhere [Tang 2002; Tang & Fu 2006]. The instrument measures half-cell corrosion potential, and then generates two galvanostatic pulses for corrosion rate and resistivity measurements. The measurement is quick and only needs a few seconds to obtain three parameters: corrosion potential, corrosion rate of steel and resistivity of concrete. These three parameters contribute a more accuracy estimation of the corrosion status of steel.. Figure 3.1 − Handheld instrument RapiCor.. The measurement principle of the new rapid technique is illustrated in Figure 3.2. In order to facilitate the modelling of electrical current distributions using 2-D numerical model, the rectangular shape of counter and guard electrodes were used in this new technique. A wet sponge is place on concrete surface in order to improve the contact between concrete and the electrodes unit. Similar to the typical galvanostatic pulse measurement, the instrument firstly measures the corrosion potential Ecorr by the reference electrode placed at the centre of the electrodes unit and afterwards imposes galvanostatic currents ICE and IGE through the counter electrodes “CE” and the guard electrodes “GE” to the steel bar embedded in concrete. Immediately after having imposed the currents ICE and IGE, the data acquisition system starts to record the signal responses of potential ∆Ea(t) at a time interval of less than 0.02 seconds. The recorded potential-time curve is directly displayed on the screen of the instrument and can be used for calculation of various parameters such as ohmic resistance RΩ, polarization resistance Rp, etc. For an “endless” long steel bar embedded in concrete, the imposed total current Itot = ICE + IGE will disperse along the steel bar to a certain distance, depending on the conductivity of.

(20) 20. concrete, σc, the thickness of concrete cover, lc, and the conductivity of the surface film of the steel, σf. Therefore, it does not necessarily mean that the current ICE is equal to the polarization current Ip through the specified polarization length Lp. In order to calculate the proper polarization current Ip through the specified polarization length Lp the numerical modelling must be used. In this new technique, a 2-D FEM (2-Dimensional Finite Element Method) was employed to model the current distributions under the galvanostatic measurement conditions. With the help of modelling the effective polarization current Ip flowing through the specified polarization length Lp can be estimated from the ratio of polarization potential to Ohmic drop, both of which can be obtained from the measurement. Therefore, the true Ohmic and polarization resistances, RΩ and Rp, can be calculated using Ohm’s law and, consequently, the resistivity of concrete and the corrosion rate can be obtained.. Results display, data storage/output. Ip Current. Data treatment Pulse generator IGE. Data acquisition. ICE. ∆Ep = IpRp. E. Potential signals. ∆E0 = IpR0. t. RE GE. CE. CE. R0 Concrete Randles model. Lp. Cdl. GE. Wet sponge Surface layer. Cover. lc. Rp. Steel bar. Figure 3.2 − Measurement principle of RapiCor.. 3.2. Technique for corrosion measurement. Thanks its rapidity, this new technique can be used for mapping a large area of the structure in a short time. For the purpose of structural assessment the corrosion conditions can be classified into four levels: negligible, low, moderate and high corrosion [Rodríguez & Andrade, 2002]. Since the instrument measures not only corrosion rate, but also the half-cell potential and ohmic resistivity of concrete, it would reduce the uncertainty if all these three parameters are utilised in the assessment. From the previous investigations it has been shown that, among these three parameters, the corrosion rate correlates best to the actual chloride content in concrete near reinforcement steel [Tang & Malmberg, 2005; Tang & Utgenannt, 2007]. It has, therefore, been suggested to take corrosion rate as main parameter, and take half-cell potential and resistivity as complementary parameters in the assessment of corrosion level. The criteria of each parameter, especially resistivity, may be dependent on the type of concrete structures, the surface treatment and the weather when the measurement is carried out. An example of criteria is shown in Table 3.1..

(21) 21. Table 3.1 – Example of criteria for classification of corrosion level. Corrosion level. Corrosion rate [µm/yr] 1 Negligible < XL XL ∼ XM 2 Low XL ∼ XM XM ∼ XH 3 Moderate XM ∼ XH > XH 4 High > XH. Half-cell potential [mV(CSE)] ≥ Ecr < Ecr ≥ Ecr < Ecr ≥ Ecr < Ecr. Resistivity [kΩ⋅cm]. Criteria. ≥ ρcr < ρcr ≥ ρcr < ρcr ≥ ρcr < ρcr. XL = 1 µm/yr XM = 3 ~ 5 µm/yr* XH = 10 µm/yr Ecr = -200 mV(CSE) ρcr = 100 ~ 120 kΩ⋅cm†. * 3 for average measurement and 5 for single measurement. † depending on the type of concrete, surface treatment, weather, etc. According to Table 3.1, the parameter corrosion rate is used for primary classification. If the two complementary parameters are not lower than the criteria for possible corrosion, the corrosion condition will be classified as one level lower. For example, if the corrosion rate is >10 µm/yr corresponding to “high”, but the half-cell potential is higher than -200 mV CSE or the resistivity is larger than 120 kΩ⋅cm, the corrosion level will be classified as “moderate”. In this way the assessment would be safer than that based on only one parameter, considering the complication of reinforcement corrosion in the real structures.. 3.3. Field measurements. The field measurements were carried out under two weather conditions, one in June after a few days of sunny weather (dry condition) and another in August and October after a few days of rainy weather. The measured results in the order of the beam placement from the east to the west at the field exposure site are listed in Appendix 1. The average values from the same type of concrete are summarised in Table 3.2. It should be kept in mind that the actual corrosion of steel in concrete is dependent on many factors, such as chloride content, moisture content, cover thickness, crack wideness, etc.. The average values give us a rough comparison only..

(22) 22. Table 3.2 – Average values of corrosion from each type of concrete. w/b. Mix. Corr index. Measured on 06-08-23-24 & 10-27 (wet weather) Half-cell potential Corrosion rate Resistivity mV[CSE] µm/yr kΩ.cm. Corr index. 0.3. 203. -123. 2.1. 113. 1.1. -65. 4.5. 91. 1.3. 0.35. 204. -39. 1.2. 95. 1.0. -38. 1.7. 86. 1.0. 0.4. 201. -91. 3.8. 87. 1.0. -71. 4.0. 69. 1.3. 0.5. 202. -138. 4.0. 78. 1.3. -123. 6.6. 78. 2.1. 0.5. 205. -110. 6.2. 63. 2.0. -215. 10.6. 66. 2.5. 0.75. 236. -270. 11.5. 106. 3.3. -342. 12.1. 144. 3.1. 208. -73. 0.6. 277. 1.0. 37. 2.1. 235. 1.0. 209. -34. 0.7. 118. 1.0. -16. 1.3. 107. 1.0. 0.4. 206. -95. 1.5. 129. 1.0. -75. 2.2. 120. 1.1. 0.5. 207. -46. 2.6. 167. 1.0. 0. 2.5. 189. 1.0. 0.5. 210. -156. 4.8. 69. 1.5. -192. 6.0. 73. 2.0. 213. -104. 2.0. 87. 1.3. -146. 2.7. 85. 1.5. 0.3. 0.5. 214. -195. 6.5. 73. 2.3. -285. 8.9. 79. 2.8. 0.3. 217. -75. 1.6. 86. 1.0. -114. 2.3. 97. 1.3. 0.4. 215. -50. 1.7. 86. 1.0. -144. 2.2. 84. 1.0. 0.5. 216. -124. 3.0. 55. 1.0. -139. 7.8. 62. 2.0. 0.5. 218. -155. 2.8. 43. 1.0. -187. 3.3. 54. 1.5. 0.3. 221. -66. 0.7. 234. 1.0. -170. 1.0. 208. 1.0. 0.4. 219. -46. 0.9. 228. 1.0. -108. 0.9. 194. 1.0. 0.5. 220. -34. 1.3. 283. 1.0. -207. 0.8. 185. 1.0. 0.3. 224. -55. 1.9. 91. 1.0. -105. 3.4. 87. 1.0. 0.4. 222. -85. 2.1. 77. 1.0. -189. 9.0. 58. 2.5. 0.5. 223. -191. 5.0. 52. 2.0. -249. 14.3. 52. 3.8. P Kalk C. 90%Anl +10%SF. 100%Fin Rpd. 100%FinStd. 0.3 0.35. 56%FinRp 100% SliteStd d +44%SL. 95%Anl + 5%SF. 100%Anl. Binder. Measured on 06-06-09 & 13 (dry weather) Half-cell Corrosion Resistivity potential rate mV[CSE] µm/yr kΩ.cm. 0.4. 211. -203. 0.6. 127. 1.0. -158. 3.2. 152. 1.2. 0.5. 212. -260. 3.7. 67. 1.8. -285. 4.3. 76. 2.0. 0.3. 226. 31. 0.9. 317. 1.0. -11. 3.5. 185. 1.3. 0.35. 227. 75. 0.2. 319. 1.0. -64. 2.1. 176. 1.3. 0.4. 225. 28. 1.1. 195. 1.0. -14. 2.4. 152. 1.0. 0.3. 230. -104. 0.8. 141. 1.0. 6. 2.6. 188. 1.0. 0.35. 231. -108. 0.9. 129. 1.0. -33. 1.9. 159. 1.0. 0.4. 228. -225. 1.9. 135. 1.3. -138. 4.9. 138. 1.5. 0.5. 229. -144. 0.4. 90. 1.0. -149. 2.4. 117. 1.0. 0.5. 232. -254. 1.7. 98. 1.5. -229. 5.1. 151. 1.5. From Table 3.2 it can be seen that more corrosion was detected from the measurement under the wet weather than under the dry weather. This should be understandable because under the wet weather the concrete contains more electrolyte in the pores which is prone to corrosion process. For concrete mixes 236 (w/b 0.75) and 214 (w/b 0.5), significant or moderate corrosion could be detected both under dry and wet weather. Generally, if any corrosion could be detected under the dry weather, the corrosion would be confirmed under the wet weather. Some comparisons between chloride ingress and corrosion will be presented later in Chapter 5 “Modelling of chloride ingress”.. 3.4. Laboratory measurements. The purpose of laboratory measurements was to verify the field measurement using the non-destructive technique RapiCor. Since the copper wires connecting each steel bars in the adjacent concrete beams are connected to the same joint box, both the results from the.

(23) 23. field measurements and the possibility for removal of concrete beams without destroying the connections of the adjacent beams have to be considered in the selection of concrete beams for the laboratory investigation. Finally, four concrete beams were selected, that is, 202 BB1(Swedish structural cement, w/b 0.5), 223 B1 (Finnish rapid cement, w/b 0.5), 236 AB1 (Swedish structural cement, w/b 0.75) and 236 BB1 (ditto). These four concrete beams were transport to the laboratory at SP. further non-destructive measurements were carried out at a distance of 10 cm from the left side of each rebar. Afterwards the rebars were released from each concrete beam by sawing and splitting. Since the ribbed steel was used in the concrete beam, it is difficult to carry out any gravimetric measurement for a quantitative evaluation. The corrosion condition on each rebar was, therefore, visually examined only.. 3.4.1. Concrete beam 202 BB1. From the field measurement concrete beam 202 BB1 showed a low corrosion under the dry weather and a moderate corrosion under the wet weather. The results from the laboratory measurements and visual examinations are shown in Figure 3.3. It can be seen from Figure 3.3 that there is significant corrosion rust on Rebar 1 at the distance about 10~15 cm from the left side, which is in a good agreement with the non-destructive measurements. From the non-destructive measurement, rebar 2 showed a moderate corrosion at the distance 40 cm from the left side. A corrosion mark can also be seen at the distance around 44~46 cm, implying a good agreement between non-destructive measurement and destructive observation. It is surprising that the corrosion, no matter by the non-destructive measurements or visual observations, was more severe in rebar 1 with 30 mm cover than in rebar 2 with 15 mm cover. One possible explanation could be that the lower part of concrete beam was covered by the snow lump under the winter, which hindered the splashing water in direct contact with the concrete, resulting in less chloride ingress. This will be discussed later in 3.3.5.. 3.4.2. Concrete beam 223 B1. From the field measurement concrete beam 223 B1 showed a low corrosion under the dry weather and a high corrosion under the wet weather. This is a pre-cracked beam with three major vertical cracks at 20~22 cm (wideness 0.1~0.2 mm), 31~33 cm (wideness 0.2~0.3 mm), and 37~39 cm (wideness 0.1~0.2 mm), respectively, from the left side. The results from the laboratory measurement and visual examination are shown in Figure 3.4. Corrosion marks can be seen at corresponding cracked positions, especially on rebar 2 at position 38 cm. The results from the visual observation are in a fairly good agreement with the results from the non-destructive measurement. The corrosion at cracked positions could be more sensitive to the weather changes. This could be the reason why the results measured on the beam 223 B1 under the dry and the wet weather vary from low to high corrosion. This may also explain the large variation in the results measured on the other cracked beams under the dry and the wet weather, as shown in Appendix 1..

(24) 24. Concrete 202, SRPC w/b 0.5. Corrosion rate, µ m/yr. 100. 10. High Moderate Low. 1 Negligible. Rebar 1, Cover 30 mm Rebar 1, After wet sawing Rebar 2, Cover 15 mm. 0.1 0. 10. 20. 30. 40. 50. 60. Distance from left side, cm Figure 3.3 − Corrosion of steel in concrete beam 202 BB1, with Swedish SRPC (CEM I), w/b 0.50, no pre-cracking..

(25) 25. Concrete 223, Finnish Rapid C, w/b 0.40. 100 Corrosion rate, µ m/yr. Crack 0.2∼ 0.3 mm Crack ∼0.1 mm. Crack 0.1∼ 2 mm. High. 10. Moderate Low Negligible. 1. Rebar 1, Cover 30 mm Rebar 1, After wet sawing Rebar 2, Cover15 mm. 0.1 0. 10. 20. 30. 40. 50. 60. Distance from left side, cm Figure 3.4 − Corrosion of steel in concrete beam 223 B1, with Finnish rapid cement, w/b 0.50, pre-cracked..

(26) 26. 3.4.3. Concrete beams 236 AB1 and BB1. From the field measurement concrete beams 236 generally showed a high corrosion under both the dry and the wet weathers. There was no signal response on beam 236 BB1 in the field measurement (see Appendix 1). It was believed that the wire connection was broken due to severe corrosion. In the laboratory, each rebar in beam 236 BB1 was directly connected for the non-destructive measurement using RapiCor. Concrete beam 236 AB1 is a pre-cracked beam with four major vertical cracks at 6~10 cm (wideness about 0.1 mm), 10~15 cm (wideness 0.1~0.3 mm), 18~23 cm (wideness 0.2~1.5 mm), and 30~32 cm (wideness 0.1~0.5 mm), respectively, from the left side. The results from the laboratory measurements and visual examinations are shown in Figures 3.5 and 3.6. It can be seen from Figure 3.5 that corrosion in beam 236 AB1 is more localised than in beam 236 BB1 (see Figure 3.6). This more localised corrosion in beam 236 AB1 is surely attributed to the pre-cracks which initiated corrosion before the non-cracked parts of concrete were contaminated by chlorides ingress. The results from the visual observation are, again, in a fairly good agreement with the results from the non-destructive measurement. Concrete beam 236 BB1 was not subjected to pre-cracking. The corrosion of steel in this beam should be induced by either carbonation or chloride ingress through the bulk concrete. The carbonation depth in this beam is, however, about 5 mm only, as shown in Table 3.3 and Figure 3.7 in 3.3.4. The corrosion must, therefore, be induced by chloride ingress. Figure 3.6 shows that severe corrosion occurred on rebar 1, while relatively less severe corrosion on rebar 2. This is in good agreement with the non-destructive measurements. Again, the corrosion was more severe in rebar 1 with 30 mm cover than in rebar 2 with 15 mm cover, as will be discussed later in 3.3.5..

(27) 27. Concrete 236, SRPC w/b 0.75. Corrosion rate, µ m/yr. 100. High. 10. Moderate Low Negligible. 1. Rebar 1, Cover 30 mm Rebar 1, After wet sawing Rebar 2, Cover 15 mm. 0.1 0. 10. 20. 30. 40. 50. 60. Distance from left side, cm. Figure 3.5 − Corrosion of steel in concrete beam 236 AB1, with Swedish SRPC (CEM I), w/b 0.75, pre-cracked..

(28) 28. Crack due to splitting. Crack due to splitting. Concrete 236, SRPC w/b 0.75. 100. Crack 0.2∼ 1 mm. Corrosion rate, µ m/yr. Crack 0.1 ∼ 0.2 mm. High 10 Moderate Low. 1. Negligible Rebar 1, Cover 30 mm, after wet sawing Rebar 2, Cover 15 mm, after wet sawing. 0.1 0. 10. 20. 30. 40. 50. 60. Distance from left side, cm. Figure 3.6 − Corrosion of steel in concrete beam 236 BB1, with Swedish SRPC (CEM I), w/b 0.75, no pre-cracking..

(29) 29. 3.4.4. Carbonation depths in concrete covers. After removal of rebars from the concrete beams, carbonation depths were measured on the split surfaces using colourimetric method with phenolphthalein solution. The results are summarised in Table 3.3. Some of the photos were shown in Figures 3.7 and 3.8. It can be seen from the measured results that carbonation depths in the concrete cover are in general very low. Even for the concrete with the highest water-binder ratio (mix 236, w/b 0.75), the carbonation depth was about 5 mm after over 10 years exposure under the highway environment. The placement of concrete near the ground where the average moisture level is relatively high could be one of the reasons to the low carbonation depth. As expected, the cracked zone can easily be carbonated and also give the paths for chloride ingress, initiating corrosion at an early age. Figure 3.8 shows the carbonation front in the concrete cover of 15 mm for rebar 2 in beam 223 B1 at about 38 mm distance from the left side, where there is a pre-crack with wideness of 0.1~0.2 mm. The corresponding corrosion can be seen in Figure 3.4. Table 3.3 – Carbonation depths [mm] in concrete covers Concrete beam 202 BB1 223 B1. 236 AB1. 236 BB1. 0~1. 2~4. 3~6. Rebar 2, cover 15 mm 1~2 0~1* * 15 mm around a cracked zone, see Figure 3.8.. 3~5. 4~6. Rebar 1, cover 30 mm. 1~2. Figure 3.7 − Carbonation depth in concrete cover of beam 236 BB1..

(30) 30. Figure 3.8 − Carbonation front in concrete cover of beam 223 B1.. 3.4.5. Chloride ingress at the cover levels. It has been observed from concrete beams 202 BB1 and 236 BB1 (both without precracking) that corrosion of rebar with 30 mm cover is more severe than that with 15 mm cover, see Figures 3.9, in contrast to the conventional knowledge of steel protection in concrete. One possible explanation could be that the lower part of concrete beam was covered by the snow lump under the winter, which hindered the splashing water in direct contact with the concrete, resulting in less chloride ingress. To verify this, small cores of diameter 50 mm were taken from these two beams at the heights where rebars were embedded. Chloride contents at the depths of 10, 20 and 30 mm were determined using the same methods as will be described in 4.2. The results are shown in Figure 3.10. It can be seen that the chloride ingress in the upper part of the beam where rebar was embedded with cover 30 mm is more than in the lower part of the beam where rebar was embedded with cover 15 mm, especially in the concrete beam 236 BB1. Figure 3.9 shows that, for non-cracked concrete beams 202 BB1 and 236 BB1, corrosion were initiated from the undersides of rebars, where often exist voids or high water content paste due to segregation under the rebars. For the pre-cracked concrete beams, the corrosion should be initiated at the positions with wider cracks..

(31) 31. Figure 3.9 − Corrosion in rebars. upper photo: upsides of rebars; lower photo: undersides of rebars. 1.2 202 BB1, upper cover 30 mm. Cl [% by wt of binder]. 1. 202 BB1, lower cover 15 mm. 0.8 0.6. 236 BB1, upper cover 30 mm. 0.4. 236 BB1, lower cover 15 mm. 0.2 0 0. 10. 20. 30. 40. 50. Depth [mm]. Figure 3.10 − Chloride ingress in upper and lower parts of concrete beams..

(32) 32. 4. Measurements of Chloride and Moisture Profiles. 4.1. Sampling. Totally 34 concrete blocks of size 400×300×300 mm were taken to the laboratory at SP for measurements of chloride and moisture profiles. Three cores of diameter 100 mm, two from the vertical exposure surface and one from the upper horizontal exposure surface, were taken at the positions as shown in Figure 4.1. When the cores became surface dry after coring, they were individually sealed in double thick plastic bags to prevent from further evaporation of moisture. One of the cores from the vertical surface was sent to Lund Institute of Technology for moisture measurement and the rest were divided between Chalmers University of Technology (the 1st set) and SP (the rest sets) for chloride profiling. Φ100, L: ≈ 100-120 Marking: X-T. Direction to highway Φ100, L: ≈ 100-120 Marking: X-S2. Drilling arranged in three sets in order to facilitate the moisture measurements at Lund Institute of Technology. 120. Φ100, L: ≈150 Drill first! Marking: X-S1. 120. 120. X = Concrete block number, e.g ”201 DK3” (see right side). 1st set 201 DK3 202 CK2 202 CK6 203 GK2 203 EK1 204 BK2 205 K2 206 DK3 207 BK2 208 AK1 208 CK1 209 BK2. 2nd set 210 K2 211 BK2 211 BK6 212 K2 213 BK2 214 K2 217 BK2 218 K2 219 K2 220 K2 221 BK2 222 K2. 3rd set 223 K2 224 BK2 225 K2 226 BK2 227 BK2 228 BK3 230 BK2 231 BK2 232 K2 236 CK2. Figure 4.1 − Illustration of sampling positions at concrete block.. 4.2. Measurement of chloride profiles. The same techniques as used in the previous investigations (e.g. Tang, 2003b) were used in this project for measurement of chloride profiles. Powder samples were taken from each core by means of dry-grinding on a lathe with a diamond tool, successively from the exposed surface to a certain depth. The depth of each sample was measured from the lathe with an accuracy of 0.5 mm. After the grinding, the powder samples were immediately dried at 105 °C and then stored in a desiccator for later chloride and calcium analysis. The acid soluble chloride content in each sample was determined principally in accordance with AASHTO T260 using potentiometric titration on an automatic titrator Metrohm Titranor 716 with chloride selective electrode and Ag/AgCl reference electrode. A sample size of about 1 gram was used to facilitate the parallel calcium analysis. The soluble calcium content in each powder sample was determined parallel to the determination of chloride content, using the technique reported by Tang (2003a)..

(33) 33. 4.3. Measurement of moisture profiles. The measurement of moisture profiles was carried out at the department of Building Materials, Lund Institute of Technology. After arrival of the cores individually sealed in double thick plastic bags, they were stored in the laboratory at the room temperature not longer than a few days prior to sampling. A slice of about 20 mm thick was split from each concrete core at depths of about 20~40, 40~60, 60~80 and 80~100 mm starting from the exposure surface, with the help of a compression jack. A large piece of sample of about 10~30 g and a number of small pieces of sample were immediately taken, using hammer and chisel, from the central portion of the freshly split slice. The large piece was immediately weighed and then placed in a box for measurement of degree of capillary saturation, while the small pieces were stored in a glass test tube for measurement of RH (Relative Humidity). The technique for measurement of RH has been well described by Nilsson (1980) and for degree of capillary saturation by Hedenblad and Nilsson (1985). After the above sampling, another slice was successively split and samples were taken. The above sampling process was repeated until all the samples were taken from each core.. 4.4. Measured chloride profiles. The results of chloride and calcium profiles in each core are given in Appendix 2.. 4.4.1. Effect of binder types. The chloride profiles from concrete with different binders are summarised in Figures 4.2 to 4.9. It can be seen from these figures that, for concrete with low water-binder ratios (w/b <0.4), the pozzolanic additions reveal reduction of chloride penetration, while for concrete with w/b 0.4~0.5, the pozzolanic additions reveal unclear effect and, in some cases, the chloride ingress in the concrete with pozzolanic additions is even deeper than in the Portland cement concrete (see figures 4.7 to 4.9). Addition of blast-furnace slag tends to increase chloride binding, while addition of limestone as in the cement “P Kalk C” seems decrease chloride binding. Normally, the addition of pozzolanic materials, such as silica fume, flyash and blastfurnace slag, in concrete will improve the pore structures through the secondary hydration, resulting in less permeable concrete. As a consequence, the concrete with pozzolanic additions should have better resistance to chloride ingress. The chloride profiles in figures 4.7 to 4.9 do not show the positive effect of pozzolanic additions on chloride ingress. One possible reason is probably due to the application of new techniques for salt spreading which results in less splashing of de-icing salts to the sides of highway. With the decreased sources of chlorides, the previously penetrated chlorides in concrete may be washed out during the non-freezing period. More permeable the concrete is, more chlorides may be washed out. However, more data are needed to confirm this explanation. Another possible reason could be the large scatter in chloride ingress in such an environment. The chloride profiles taken from the same type of concrete reveal significant differences, e.g. the replicate profiles in figures 4.7 and 4.9. The chloride profiles taken from earlier exposure periods also showed very large scatters [Lindvall, 2002], as will be discussed later in 5.6..

(34) 34. 1.5 100%Anl ditto, replicate 95%Anl+5%SF ditto, replicate 100%FinStd 100%SliteStd 56%FinRpd+44%SL 100%Fin Rpd 90%Anl+10%SF P Kalk C. Cl% of binder. Concrete w/b 0.30 from upper horizontal surface. 1. 0.5. 0 0. 10. 20. Depth (mm). 30. 40. 50. Figure 4.2 – Summary of chloride profiles in concrete with water-binder ratio 0.30, from the upper horizontal exposure surface.. 1.5 100%Anl ditto, replicate 95%Anl+5%SF ditto, replicate 100%FinStd 100%SliteStd 56%FinRpd+44%SL 100%Fin Rpd 90%Anl+10%SF P Kalk C. Cl% of binder. Concrete w/b 0.30 from vertical surface. 1. 0.5. 0 0. 10. 20. Depth (mm). 30. 40. 50. Figure 4.3 – Summary of chloride profiles in concrete with water-binder ratio 0.30, from the vertical exposure surface..

(35) 35. 1.5 Concrete w/b 0.35 from upper horizontal surface. 100%Anl. Cl% of binder. 95%Anl+5%SF 1 90%Anl+10%SF P Kalk C 0.5. 0 0. 10. 20. Depth (mm). 30. 40. 50. Figure 4.4 – Summary of chloride profiles in concrete with water-binder ratio 0.35, from the upper horizontal exposure surface.. 1.5 Concrete w/b 0.35 from vertical surface. 100%Anl. Cl% of binder. 95%Anl+5%SF 1 90%Anl+10%SF P Kalk C 0.5. 0 0. 10. 20. Depth (mm). 30. 40. 50. Figure 4.5 – Summary of chloride profiles in concrete with water-binder ratio 0.35, from the vertical exposure surface..

(36) 36. 1.5 100%Anl. Concrete w/b 0.40 from upper horizontal surface. 95%Anl+5%SF. Cl% of binder. 100%FinStd ditto, replicate. 1. 56%FinRpd+44%SL 100%Fin Rpd 90%Anl+10%SF P Kalk C. 0.5. 0 0. 10. 20. Depth (mm). 30. 40. 50. Figure 4.6 – Summary of chloride profiles in concrete with water-binder ratio 0.40, from the upper horizontal exposure surface.. 1.5 100%Anl. Concrete w/b 0.40 from vertical surface. 95%Anl+5%SF. Cl% of binder. 100%FinStd ditto, replicate. 1. 56%FinRpd+44%SL 100%Fin Rpd 90%Anl+10%SF P Kalk C. 0.5. 0 0. 10. 20. Depth (mm). 30. 40. 50. Figure 4.7 – Summary of chloride profiles in concrete with water-binder ratio 0.40, from the upper horizontal exposure surface..

(37) 37. 1.5 100%Anl. Concrete w/b 0.50 from upper horizontal surface. ditto, replicate. Cl% of binder. 95%Anl+5%SF 100%FinStd. 1. 100%SliteStd 56%FinRpd+44%SL 100%Fin Rpd P Kalk C. 0.5. 0 0. 10. 20. Depth (mm). 30. 40. 50. Figure 4.8 – Summary of chloride profiles in concrete with water-binder ratio 0.50, from the upper horizontal exposure surface.. 1.5 100%Anl. Concrete w/b 0.50 from vertical surface. ditto, replicate. Cl% of binder. 95%Anl+5%SF 100%FinStd. 1. 100%SliteStd 56%FinRpd+44%SL 100%Fin Rpd P Kalk C. 0.5. 0 0. 10. 20. Depth (mm). 30. 40. 50. Figure 4.9 – Summary of chloride profiles in concrete with water-binder ratio 0.50, from the vertical exposure surface..

(38) 38. 4.4.2. Effect of water-binder ratios. The chloride profiles from concrete with Swedish SRPC and with addition of 5% silica fume are summarised in Figures 4.10 to 4.13. As expected, the higher water-binder ratios often result in the more chloride ingress, especially when water-binder ratio is higher than 0.4. When water-binder ratio is lower than 0.4, the effect seems not significant. From the figures it can also be seen that chloride profiles vary even in the same type of concrete, depending on the local exposure positions and availability of chlorides.. 1.5. 203 EK1-T, w/b 0.3 203 GK2-T, w/b 0.3 204 BK2-T, w/b 0.35 201 DK3-T, w/b 0.4 202 CK2-T, w/b 0.5 202 CK6-T, w/b 0.5 205 K2-T, w/b 0.5 236 CK2-T, w/b 0.75. Cl% of binder. Concrete with 100% SRPC from upper horizontal surface. 1. 0.5. 0 0. 10. 20. 30. 40. 50. Depth (mm) 1.5. 203 EK1-S, w/b 0.3 203 GK2-S, w/b 0.3 204 BK2-S, w/b 0.35 201 DK3-S, w/b 0.4 202 CK2-S, w/b 0.5 202 CK6-S, w/b 0.5 205 K2-S, w/b 0.5 236 CK2-S, w/b 0.75. Cl% of binder. Concrete with 100% SRPC from vertical surface. 1. 0.5. 0 0. 10. 20. 30. 40. Depth (mm) Figure 4.10 – Summary of chloride profiles in concrete with SRPC (CEM I) cement.. 50.

(39) 39. 1.5. 208 AK1-T, w/b 0.3 Concrete with 95% SRPC + 5% SF from upper horizontal surface. 208 CK1-T, w/b 0.3 209 BK2-T, w/b 0.35. Cl% of binder. 206 DK3-T, w/b 0.4 1. 207 BK2-T, w/b 0.5 210 K2-T, w/b 0.5. 0.5. 0 0. 10. 20. 30. 40. 50. Depth (mm) 1.5. 208 AK1-S, w/b 0.3 Concrete with 95% SRPC + 5% SF from vertical surface. 208 CK1-S, w/b 0.3 209 BK2-S, w/b 0.35. Cl% of binder. 206 DK3-S, w/b 0.4 1. 207 BK2-S, w/b 0.5 210 K2-S, w/b 0.5. 0.5. 0 0. 10. 20. 30. 40. 50. Depth (mm) Figure 4.11 –Summary of chloride profiles in concrete with addition of 5% silica fume..

(40) 40. 4.5. Measured moisture profiles. The results of moisture measurements are given in Appendix 3.. 4.5.1. Effect of binder types. The moisture profiles from concrete with different binders are summarised in Figures 4.12 to 4.15. The results show that the concrete with addition of silica fume and blastfurnace slag, as well as with Finnish rapid cement, reveals a lower profile of relative humidity. The concrete with Swedish SRPC or Finnish standard cement reveals, however, a higher profile of relative humidity. 1. 100. 100%Anl. Concrete w/b 0.30 0.95. Degree of Capillary Saturation. Relative Humidity %. 95. 90. 85. 80. 75. 95%Anl+5%SF 100%FinStd. 0.9. 100%SliteStd 0.85. 56%FinRpd+44%SL 100%Fin Rpd. 0.8. 90%Anl+10%SF 0.75 P Kalk C 0.7 0.65. 70. 0.6 20. 40. 60. 80. 20. 100. 40. Depth (mm). 60. 80. 100. Depth (mm). Figure 4.12 – Summary of moisture profiles in concrete with in concrete with waterbinder ratio 0.30. 1. 100. Concrete w/b 0.35 0.95. Degree of Capillary Saturation. Relative Humidity %. 95. 90. 85. 80. 100%Anl. 0.9 95%Anl+5%SF 0.85 90%Anl+10%SF 0.8 P Kalk C 0.75 0.7. 75 0.65 70. 0.6 20. 40. 60. 80. Depth (mm). 100. 20. 40. 60. 80. 100. Depth (mm). Figure 4.13 – Summary of moisture profiles in concrete with in concrete with waterbinder ratio 0.35..

(41) 41. 100. 1 0.95. Degree of Capillary Saturation. Relative Humidity %. 95. 90. 85. 80. 75. 100%Anl. Concrete w/b 0.40. 95%Anl+5%SF. 0.9. 100%FinStd 56%FinRpd+44%SL. 0.85. 100%Fin Rpd. 0.8. 90%Anl+10%SF 0.75. P Kalk C. 0.7 0.65. 70. 0.6 20. 40. 60. 80. 100. 20. 40. Depth (mm). 60. 80. 100. Depth (mm). Figure 4.15 – Summary of moisture profiles in concrete with in concrete with waterbinder ratio 0.40.. 100. 1 0.95. Degree of Capillary Saturation. Relative Humidity %. 95. 90. 85. 80. 75. 100%Anl. Concrete w/b 0.50. 95%Anl+5%SF. 0.9. 100%FinStd 100%SliteStd. 0.85. 56%FinRpd+44%SL. 0.8. 100%Fin Rpd 0.75. P Kalk C. 0.7 0.65. 70. 0.6 20. 40. 60. Depth (mm). 80. 100. 20. 40. 60. 80. 100. Depth (mm). Figure 4.15 – Summary of moisture profiles in concrete with in concrete with waterbinder ratio 0.50.. 4.5.2. Effect of water-binder ratios. The moisture profiles from concrete different water-binder ratios are summarised in Figures 4.16 to 4.17. As expected, the moisture level in concrete increased as waterbinder ratio increased. It seems that the moisture profiles vary not significantly in the.

(42) 42. concrete with a low water-binder ratio (e.g. concrete 203, w/b 0.3), while vary significantly in the concrete with a high water-binder ratio (e.g. concrete 202, w/b 0.5).. 100 203 EK1, w/b 0.3. 0.95. Degree of Capillary Saturation. Relative Humidity %. 95. 90. 85. 80 Concrete with Swedish SRPC 75. 70. 203 GK2, w/b 0.3 204 BK2, w/b 0.35. 0.85. 201 DK3, w/b 0.4 202 CK2, w/b 0.5. 0.75. 202 CK6, w/b 0.5 205 K2, w/b 0.5. 0.65. 236 CK2, w/b 0.75 0.55. 0.45 20. 40. 60. 80. 100. 20. 40. Depth (mm). 60. 80. 100. Depth (mm). Figure 4.16 – Summary of moisture profiles in concrete with in concrete with 100% SRPC (CEM I).. 100. 1 208 AK1, w/b 0.3 0.95. Degree of Capillary Saturation. Relative Humidity %. 95. 90. 85. 80. Concrete with 95% SRPC + 5% silica fume. 75. 208 CK1, w/b 0.3 0.9 209 BK2, w/b 0.35 0.85 206 DK3, w/b 0.4 0.8 207 BK2, w/b 0.5 0.75. 210 K2, w/b 0.5. 0.7 0.65. 70. 0.6 20. 40. 60. Depth (mm). 80. 100. 20. 40. 60. 80. 100. Depth (mm). Figure 4.17 – Summary of moisture profiles in concrete with in concrete with 95% SRPC + 5% silica fume..

(43) 43. 5. Modelling of Chloride Ingress. 5.1. Curve-fitting to the ERFC model. The ERFC model is the simplest model using the complementary error-function (ERFC) to Fick’s 2nd law to describe chloride ingress in concrete [Collepardi et al, 1972]:. ⎛ ⎞ x ⎟ C ( x, t ) = C i + (C s − C i ) ⋅ erfc⎜ ⎜ 2⋅ D ⋅t ⎟ a ⎝ ⎠. (5.1). where Ci is the initial chloride content in the concrete (usually this chloride content is negligible), Cs is the surface chloride content, x is the depth, Da is the apparent diffusion coefficient, t is the exposure duration. In this model the parameters Cs and Da are assumed constant during the whole period of exposure. This model has been proven too conservative due to the lack of consideration of chloride binding effect and other time-dependent effects. Therefore, this model will not be used for prediction but used for curve-fitting the measured chloride profiles to obtain the parameters Da and Cs. The curve-fitted results are listed in Appendix 4 and summarised in Figures 5.1 to 5.3. From the results it can be found that • Curve-fitted apparent diffusion coefficient Da reveals a certain degree of exponential relation to water-binder ratio, but the values of Da vary very much (see Figure 5.1). For concrete with the same type of binder and water-binder ratio, the curve-fitted Da may vary by one order of magnitude, indicating the large variation of chloride profiles in the similar quality of concrete under the similar exposure environment. •. There is no clear relationship between the curve-fitted surface chloride content Cs and water-binder ratio, probably due to the fact that the chloride content has already be expressed by mass of binder (see Figure 5.2). Therefore, the average value of Cs from the same type of binder may be taken as an indicator of chloride binding (see Figure 5.3).. •. The addition of slag shows significantly highest average value of Cs, indicating a strong chloride binding capacity, which is in agreement with the previous studies.. •. The addition of silica fume also shows notable higher average value of Cs, when compared with the reference binder (Anl C – Swedish structural cement). This is probably not attributed to the higher chloride binding, but to the relatively low rate of calcium leaching or carbonation in the concrete with silica fume.. •. The addition of limestone powder (P Kalk C) reveals lowest average value of Cs, probably indicating a lower chloride binding capacity in this type of cement, but the verification by laboratory studies of chloride binding is needed..

(44) 44. 1E-10 y = 4.56E-15e11.9x R2 = 0.927. Curve-fitted D a, m /s. 1E-11. 100%Anl 95%Anl+5%SF 100%FinStd. 2. y = 1.74E-15e15.7x R2 = 0.856. 100%SliteStd. y = 1.60E-14e9.82x R2 = 0.901. 56%FinRpd+44%SL. 1E-12. 100%Fin Rpd 90%Anl+10%SF. 1E-13. P Kalk C y = 3.08E-15e8.38x R2 = 0.802. 1E-14 0.3. 0.35. 0.4. 0.45. 0.5. 0.55. 0.6. 0.65. 0.7. 0.75. 0.8. Water-binder ratio, w/b. Figure 5.1 – Results of the curve-fitted apparent diffusion coefficient. 3. Curve-fitted C s , Cl% of binder. 100%Anl y = -29.15x 2 + 23.74x - 2.7025 R2 = 0.0846. 2.5. 95%Anl+5%SF 100%FinStd. 2. 100%SliteStd 56%FinRpd+44%SL. 1.5. 100%Fin Rpd 90%Anl+10%SF. 1. P Kalk C. 0.5 y = 6.0232x 2 - 6.884x + 2.5498 R2 = 0.5872 0 0.3. 0.35. 0.4. 0.45. 0.5. 0.55. 0.6. 0.65. Water-binder ratio, w/b. Figure 5.2 – Results of the curve-fitted surface chloride content.. 0.7. 0.75. 0.8.

(45) 45. Average value of C s, Cl% of binder. 2.5. 2. 1.5. 1. 0.5. SL pd +4 4% nR. n. d. Rp d. 56 % Fi. 10 0% Fi. nS t 10 0% Fi. C Ka lk P. St d ite 10 0% Sl. 95 % An l+ 5% SF 90 % An l+ 10 % SF. 10 0% An l. 0. Figure 5.3 – Summary of average values of curve-fitted surface chloride content.. 5.2. DuraCrete model. The DuraCrete model is a modification of the ERFC Model by adding some extra parameters to modify the apparent diffusion coefficient Da [DuraCrete, 2000], that is, n. ⎛t ⎞ Da = ke ⋅ kc ⋅ DRCM,0 ⋅ ⎜ 0 ⎟ ⋅ γ Da ⎝t ⎠. (5.2). where DRCM,0 is the chloride migration coefficient measured under the standard conditions at the age t0, ke is the environmental factor, kc is the factor considering the influence of curing on D0, t0 is the reference period (normally 28 days) at which DRCM,0 is measured, n is the age factor describing the time-dependency of the effective diffusion coefficient, and γDa is the partial factor for the apparent diffusion coefficient. The surface chloride content Cs can be calculated using the following equation:. ⎛ w⎞ Cs = ACs ⋅ ⎜ ⎟ ⋅ γ Cs ⎝b⎠. (5.3). where ACs is the constant and γDa is the partial factor for the surface chloride content. This model has been adopted by the Swedish Association of Concrete in the guidelines for durability design of concrete structures (Betongrapport nr 12, 2007). In this study, only the concrete with 100% SRPC and with 95% SRPC + 5% SF will be modelled. The input parameters, if not otherwise stated, were taken from the Swedish guidelines for the modelling, as listed in Tables 5.1 and 5.2. The modelled results will be presented in 5.4..

(46) 46. Table 5.1 – Input parameter DRCM, 28d [×10-12 m2/s] for the DuraCrete model. w/b 0.3 0.35 0.4 0.5 1). 100% SRPC. 4.2. 7. 95% SRPC + 5% SF. 1.4. 2.91). 0.75. 9.8. 19. 632). 4.4. 10. -. 1) Interpolated; and 2) Extrapolated.. Table 5.2 – Other input parameters for the DuraCrete model. Binder type 100% SRPC Limit. Upper1). Lower. Curing factor kc. 95% SRPC + 5% SF Upper1). Lower 0.79. Environmental factor ke,. 0.27. 0.27. Age factor n. 0.67. 0.60. ACs, mass% of binder. 1.4. 1.4. Partial factor γCs. 1. 1.4. 1. 1.4. Partial factor γDa. 1. 2.35. 1. 2.35. Initial Ci, mass% of binder. 0.02. 0.02. 1) Assuming that the costs for reparation is normal.. 5.3. ClinConc model. The ClinConc model is a mechanistic model, which was developed at Chalmers University of Technology in the middle of 1990’s. In the past years, this model was further developed and expressed in an engineer-friendly way, that is, there is no need for special software to carry out numerical iterations. An Excel-workbook is enough for all the calculations. The model predicts the free chloride penetration through the pore solution in concrete using a flux equation based on the principle of Fick’s law, and then converts the free chloride concentration to total chloride content. The model includes a series of equations but can simply be calculated in Excel worksheets.. 5.3.1. Modelling of free chloride ingress. The first step in the predictions with the ClinConc model is to determine the free chloride content in the concrete at depth, x. This is done with the following expression: ⎛ ⎜ ⎜ c − ci x = 1 − erf ⎜ 1− n 1− n n ⎜ cs − c i ′ ⎞ ⎡⎛ t ex ′ ⎞ ′ ⎞ ⎤ ⎛ t ex ⎜ 2 ξ D D6m ⋅ ⎛⎜ t 6m ⎟ ⋅ ⎢⎜1 + ⎟ − ⎜ ⎟ ⎥ ⋅ t ⎜ 1 − n ⎝ t ⎠ ⎣⎢⎝ t ⎠ ⎝ t ⎠ ⎦⎥ ⎝. ⎞ ⎟ ⎟ ⎟ ⎟ ⎟ ⎟ ⎠. (5.4). where c, cs and ci are the concentration of free chlorides in the pore solution at depth x, at the surface of the concrete and initially in the concrete, respectively, D6m is the diffusion coefficient measured by the RCM test at the age of t’6m, ξD is.

(47) 47. the factor bridging the laboratory measured D6m to the initial apparent diffusion coefficient for the actual exposure environment, n is the age factor accounting for the decrease of diffusivity with age, t’ex is the age of concrete at the start of exposure and t is the duration of the exposure. The factor ξD is expressed by the following expression:. ξD =. (0.8 ⋅ a. 2 t. ). − 2 ⋅ a t + 2.5 ⋅ (1 + 0.59 ⋅ K b6m ) ⋅ e. ED ⎛ 1 1 ⎞ − ⎟ ⎜ R ⎝ 293 T ⎠. (5.5) ⋅ kD Eb ⎛ 1 1 ⎞ β b −1 ⎟ ⎜ − ⎛ cs ⎞ 1 + k OH6m ⋅ K b6m ⋅ f b ⋅ β b ⋅ ⎜ ⋅ e R ⎝ T 293 ⎠ ⎟ 35 . 45 ⎝ ⎠ where at is a factor which describe how the chloride binding changes over time, fb and βb are chloride binding constants, ED and Eb are activation energy chloride diffusion and binding, respectively, kD is the expansion factor depending on the type of binder and water-binder ratio, and kOH6m and Kb6m are factors accounting the effects of hydroxide concentration in the pore solution, gel content and water accessible porosity at the age t’6m.. k OH6m = e K b6m =. ⎛ 0.043 0.59 ⎜⎜ 1− ⎝ [OH ]6m. ⎞ ⎟ ⎟ ⎠. Wgel6m 1000 ε 6m. (5.6). (5.7). where [OH]6m, Wgel6m and ε6m are the hydroxide concentration in mol/m3pore3 solution, gel content in kg/m concrete and water accessible porosity at the age t’6m. According to the experience obtained from 10 years’ exposure in the seawater at Swedish west coast Tang, 2003b], the expansion factor, kD, can be estimated by ⎧1 + 8(0.4 − w / b ) + 7 SF + 3800(SF ⋅ FA) ⋅ (SF + FA) kD = ⎨ ⎩1. 0.25 ≤ w / b ≤ 0.4 w / b > 0.4 (5.8). where w/b is the water-binder ratio, SF and FA is the mass fraction of silica fume and fly ash to the total binder, respectively. This expansion factor describes the ratio of diffusion coefficient in the field concrete to that in the laboratory test. Similarly, the binding factor, at, can be estimated by ⎧0.36 + 1.4(0.4 − w / b ) + 0.4 SF + 38(SF ⋅ FA) ⋅ (SF + FA) at = ⎨ ⎩0.36 + 1.4(0.4 − w / b ). 0.25 ≤ w / b ≤ 0.4 0.4 < w / b ≤ 0.6 (5.9). In equations (5.8) and (5.9), the effects of pozzolanic additions become insignificant when w/c > 0.4, probably because of the sufficient volume of capillary network in the concrete with sufficient high w/c (> 0.4). This capillary.

Chloride Ingress and Reinforcement Corrosion in Concrete under De-Icing Highway Environment - A study after 10 years&apos; field exposure.

Chloride Ingress and Reinforcement Corrosion in Concrete under De-Icing Highway Environment - A study after 10 years' field exposure.