Chloride ingress in concrete exposed to marine environment - field data up to 10 years exposure

(1)

Chloride Ingress in Concrete Exposed

to Marine Environment

– Field data up to 10 years exposure

SP Swedish National Testing and Research Institute Building Technology and Mechanics

(2)

Chloride Ingress in Concrete Exposed to Marine

Environment – Field data up to 10 years exposure

Abstract

This report presents the results from a research project financed by the Swedish National Road Administration.

In this project about 60 chloride and moisture profiles have been measured from various types of concrete specimens exposed to marine environment for more than 10 years. Totally about 240 chloride profiles measured after 0.5, 1, 2, 5 and 10 years exposure and about 100 moisture profiles after 5 and 10 years exposure have been collected and formatted into similar worksheets in order to facilitate the establishment of database. These collected data have been used for curve-fitting to diffusion functions based on Fick’s second law, for validation of the rapid migration test, and for verification of the prediction model ClinConc.

According to the results it seems that the curve-fitted diffusion coefficient DF2 decreases

with the exposure time only in the first five years period, and afterwards this decrease tendency becomes unclear.

The rapid migration test, or called the CTH method, has been proven applicable to various types of concrete including silica fume and fly ash. There exist reasonably good linear relationships in logarithmic scale between DCTH and DF2. The lowest DCTH value

well corresponds to the lowest chloride ingress in the group of concrete with similar water-binder ratios.

The ClinConc model has also been successfully used to predict the chloride ingress profiles of various types of concrete. The predicted profiles are in general in fairly good agreement with the measured ones, although an extension coefficient has to be introduced to extend the laboratory diffusivity to the field one. Further experimental investigation is needed to verify this extension coefficient.

Key words: concrete, chloride, chloride ingress, durability, field test, moisture.

SP Sveriges Provnings- och Forskningsinstitut

SP Rapport 2003:16 ISBN 91-7848-948-2 ISSN 0284-5172 Borås 2003

SP Swedish National Testing and Research Institute

SP Report 2003:16

Postal address:

Box 857, SE-501 15 BORÅS Sweden

Telephone +46 33 16 50 00 Telex 36252 Testing S Telefax +46 33 13 55 02

(3)

Preface

In the beginning of 1990 I came to Sweden and started my research career on chloride ingress in concrete at the Chalmers University of Technology under the supervision of Prof. Lars-Olof Nilsson. I fortunately participated in the national project on the durability of marine concrete structures –BMB (Beständiga Marina Betongkonstruktioner) and measured the first chloride profiles in concrete sampled from the Träslövsläge field exposure site. It was my real privilege that, under the financial support of Swedish National Road Administration, I

coordinated this project and measured the chloride profiles in the same concrete slabs after over 10 years field exposure.

I am satisfied with the results, especially the good applicability of the CTH rapid method in evaluating the resistance of concrete to chloride ingress. While enjoying my satisfaction, I would like to thank Prof. Lars-Olof Nilsson, who supervised me from the beginning of the long-term field exposure (BMB project) and also gave me constructive comments and suggestions to this report. Ms. Nidal Yousif, Mr. Marek Machowski and Dr. Anders Sjöberg are acknowledged for their contributions to the measurement and summarisation of moisture profiles. Ms. Iman Jasem is acknowledged for her contributions to the work of chloride and calcium analysis.

Special thanks are extended to Mr. Samir Redha for his suggestions to the improvement of the final report. The financial support from Swedish National Road Administration is greatly appreciated.

Tang Luping Borås, 2003

(5)

Sammanfattning (Summary in Swedish)

Denna rapport presenterar resultaten från Vägverksprojektet nr AL 90 AB 2001:24965 med syftet att sammanställa klorid- och fuktprofiler hos betong som har exponerats i fältstationen vid hamnen i Träslövsläge i över tio år.

Armeringskorrosion är en av de vanligaste orsakerna till skador på betongbroar i marin miljö. Kloridtransport i betong är ett internationellt hett ämne på grund av att de besvärliga

kloridjonerna kan inducera armeringskorrosionen i betongkonstruktioner. 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 “Beständiga Marina Betongkonstruktioner” som initierades 1991 för att få fram beständighetsdata för betong exponerad i marint klimat. Sedan slutet av 1991 exponeras ett stort antal provkroppar av ca 40 olika betongkvaliteter vid en fältprovplats vid hamnen i Träslövsläge i Varberg. Genom det nationella

forskningsprojektet BMB har grundliga karteringar av fukt- och kloridprofiler gjorts på utvalda betongkvaliteter efter 0,5 respektive 2 års exponering och på samtliga

betongkvaliteter efter 1 respektive 5 års exponering i det aggressiva, marina klimatet. Förutom de användbara fältmätvärdena utvecklades CTH-metoden för mätning av

kloridinträngningshastighet i betong samt datorprogrammet - ClinConc (Cl in Concrete) - för modellering av kloridinträngning i betong i marinmiljö. Både CTH-metoden och ClinConc-modellen har fått stor internationell uppmärksamhet. Den förra har godkänts som en Nordtest standardmetod – NT BUILD 492 och används i flera länder, medan den senare varit känd i världen som en vetenskaplig eller fysikalisk modell. Till stor skillnad från de empiriska modellerna är ClinConc-modellen baserad på fysikaliska och kemiska processer som sker i samband med kloridinträngning i betong. Efter den senaste utvecklingen kan modellen i princip användas för såväl marin miljö som vägmiljö efter verifiering och eventuell modifiering med tillräckliga data från fältmätning.

Vägverksprojektet är en fortsättning av BMB-projektet för en ytterligare kartering av samtliga betongkvaliteter efter över 10 års exponering i marint klimat för att skapa ännu bättre

underlag för verifiering av olika modeller för kloridinträngning och fuktnivåer hos betong exponerad i marin miljö.

Genom Vägverksprojektet har ca 60 klorid- och fuktprofiler bestämts på de existerande betongkvaliteter som varit exponerade i Träslövslägehamnen under 10 år. Totalt har ca 240 kloridprofiler utmätta efter 0,5, 1, 2, 5 respektive 10 års exponering och ca 100 fuktprofiler efter 5 respektive 10 års exponering sammanställts i likformiga Excel-blad för att underlätta etablering av databasen.

Dessa insamlade data har använts för validering av den snabba CTH-metoden samt för verifiering av prognosmodellen ClinConc. Enligt resultaten verkar det som om den

kurvanpassade diffusionskoefficienten, DF2, minskar med exponeringstiden enbart inom den

första 5-åriga perioden, och därefter blir tendensen för en minskande DF2 oklar. Å andra sidan

(6)

Cs kan dölja en minskande DF2. Därför är hypotesen med en ständigt minskande DF2 i några

empiriska modeller för livslängdprognos mycket tvivelaktig.

Den snabba metoden, s k CTH-metoden, har visats vara lämplig till olika betongkvaliteter inklusive de med kiselstoft och flygaska. Det finns rimligt bra linjära relationer i logaritmisk skala mellan DCTH och DF-värdet. Det lägsta DCTH-värdet motsvarar väl den minsta

kloridinträngningen i gruppen betong med samma vattenbindmedeltal.

Det har också visat sig att de kloridinträngningsprofiler som är beräknade med ClinConc-modellen stämmer för det mesta väl med de mätta profilerna.

Enligt resultaten från projektet kan följande rekommendationer föreslas:

• Reducering av vattenbindmedeltal är ett effektivt sätt att förbättra betongens motståndsförmåga mot kloridinträngning, men när vbt är mindre än 0,3 blir den ytterligare effekten obetydlig.

• Ett alternativt sätt att förbättra betongens motståndsförmåga mot kloridinträngning är tillsättning av restmaterial, t ex 5∼10% kiselstoft eller 20%, eller mer effektivt med en kombination av 5% kiselstoft och 10∼20% flygaska. Tillsättning av stora mängder flygaska eller andra typer av restmaterial kan kraftigt förbättra betongens

beständighet, men mer experimentella undersökningar behövs för att verifiera deras effekt.

• CTH-metoden är ett användbart verktyg för en snabb utvärdering av betongens motståndsförmåga mot kloridinträngning. Provningsresultaten kan användas som indata för prognosticering av kloridinträngning och vidare för prognosticering av livslängd hos armerade betongkonstruktioner.

• Man måste försiktig med att använda de empiriska modeller för livslängdprognos som antar att kloriddiffusionskoefficienten DF2 ständigt minskar, eftersom inga klara

tecken på att DF2 minskar med tiden har upptäckts i undersökningarna som pågått i

över 10 år

• Den fysikaliska modellen ClinConc är ett lovande verktyg för prognosticering av kloridinträngning. Ytterligare utveckling behövs för att modellen lätt ska kunna användas i verkliga dimensioneringar av anläggningskonstruktioner eller vid beräkning av återstående livslängd hos befintliga konstruktioner.

(7)

1 Introduction

Chloride induced reinforcement corrosion is one of the most important degradation processes in reinforced concrete structures exposed to marine environment and road environment where de-icing salt is used in the winter. The degradation of reinforced concrete structures,

especially infrastructures, has very important economical and social consequences due to the need for diverting resources for repairing damaged structures and sometimes the need to close the facility for carrying out the repair work. Owing to its long coastline and intensive

application of de-icing salt, the topic of chloride ingress in concrete has special significance in Sweden. In the beginning of 1990’s, a Swedish national project called “BMB” – Durability of Marine Concrete Structures – was initiated (Sandberg, 1996). As a part of work in the BMB project, some 40 types of concrete specimens were exposed to seawater at the

Träslövsläge field site at the west coast of Sweden. The specimens were periodically sampled for chloride penetration profiles, which serve as the “first hand’s” information about chloride ingress into concrete and are believed valuable for the examination of modelling for chloride penetration. The chloride ingress profiles up to five years exposure to seawater at the field site have been measured during the lifetime of the BMB project.

After the BMB project, many concrete slabs are remained at the field site for continuous exposure. To collect the field exposure data after 10 years exposure, Swedish National Testing and Research Institute (SP) together with Chalmers University of Technology (Chalmers) carried out a project under the financial support of Swedish National Road Administration. The main objectives of the project is

• to measure chloride and moisture profiles in the concrete remained at the field site after 10 years exposure,

• to compile the profiles measured from this project as well as from the previous BMB project,

• to compare the new data with the previous data, and

• to model chloride ingress using the previously developed model ClinConc.

This report presents the results from the above research project (No. AL 90 AB 2001:24965, financed by Vägverket - Swedish National Road Administration).

(8)

2 Concrete Specimens and Exposure Conditions

2.1 Concrete

slabs

The relevant mixture proportions of concrete are summarised in Table 2.1. The main

variations include water-binder ratio (0.25, 0.3, 0.35, 0.4, 0.5, 0.6 to 0.75), binder type (four types of cement with different additions of silica fume and fly ash), and air content (6% entrained air and non-AEA). Concrete slabs of 1000×700×100 mm were cast at the SP Swedish National Testing and Research Institute. After about two weeks moisture curing, the slabs were transported to the Träslövsläge field site and mounted on the sides of the pontoons for the exposure with the bottom side of the slab outward the seawater. The parallel slabs called “noll” (“zero”) specimens were transported to the laboratory at the Chalmers for measurement of chloride transport properties, as will be described in section 5.2. Table 2.1 Mixture proportions of concrete exposed at the Träslövsläge field site.

Mix No.

Binder type Binder kg/m3 Water-binder ratio1) Fine aggreg. 0-8 mm kg/m3 Coarse aggreg. 8-16 mm kg/m3 Sp2) % of binder AEA3) % of binder Air content % 28d compr. Strength4) MPa 1-35 450 0.35 839 839 1 0.041 6.0 70 1-40 100%Anl5) 420 0.40 873 806 0.8 0.03 6.2 58 Ö 100%Anl 430 0.38 813 840 1 0.04 6.2 58 1-50 370 0.50 876 808 - 0.033 6.4 41 1-75 240 0.75 1013 796 - 0.029 6.1 21 2-35 450 0.35 801 868 1.7 0.038 5.7 60 2-40 420 0.40 871 804 1.3 0.029 6.2 54 2-50 100%Slite6) 390 0.50 853 787 - 0.026 5.8 42 2-60 310 0.60 936 797 - 0.022 6.3 35 2-75 250 0.75 999 785 - 0.02 5.8 26 3-35 450 0.35 801 868 1.2 0.08 5.8 72 3-40 95%Anl+5%SF7) 420 0.40 835 835 0.8 0.043 6.1 61 3-50 370 0.50 840 840 - 0.04 6.0 45 3-75 240 0.75 966 823 - 0.039 5.9 21 4-40 90%Anl+10%SF 420 0.40 803 870 1.17 0.043 6.6 65 5-40 95%Anl+5%SF 420 0.40 878 878 1.5 0.006 2.9 81 6-35 95%Anl+5%SF 450 0.35 858 929 1.5 - 2.1 93 6-40 95%Anl+5%SF 420 0.40 898 898 1.5 - 1.7 87 7-35 100%Anl 450 0.35 898 898 1.5 - 2.4 91 7-40 100%Anl 420 0.40 939 867 1 - 2.1 79 7-75 100%Anl 265 0.75 1044 821 - - 1.1 32 8-35 100%Slite 470 0.35 847 918 1.8 - 2.1 73 8-40 100%Slite 440 0.40 882 882 1.5 - 2.1 67 8-50 100%Slite 410 0.50 893 924 - - 1.4 56 8-60 100%Slite 330 0.60 977 833 - - 1.6 45 8-75 100%Slite 270 0.75 1040 817 - - 1.4 37

(9)

Table 2.1 (Continuation)

Mix

No. Binder type

Binder kg/m3 Water-binder ratio Fine aggreg. 0-8 mm kg/m3 Coarse aggreg. 8-16 mm kg/m3 Sp % of binder AEA % of binder Air content % 28d compr. strength MPa 9-40 95%DK8)+5%SF 420 0.40 839 839 1.2 0.037 6.5 63 10-40 78.5%DK+17%FA9)+ 4.5%SF 420 0.40 770 905 1.7 0.063 6.1 69 11-35 85%DK+10%FA+5%SF 450 0.35 781 917 2.33 0.04 5.7 84 12-35 85%Anl+10%FA+5%SF 450 0.35 781 917 1.87 0.055 6.4 73 H1 95%Anl+5%SF 500 0.30 836 942 2.3 - 0.8 112 H2 90%Anl+10%SF 500 0.30 820 963 2.1 - 1.1 117 H3 100%Anl 492 0.30 791 892 2.7 - 3.6 96 H4 95%Anl+5%SF 420 0.40 840 840 0.8 0.055 5.9 63 H5 95%Anl+5%SF 551 0.25 806 946 3 - 1.3 125 H6 95%Anl+5%FA 518 0.30 791 892 2.5 - 2.8 95 H7 95%Deg40010)_+5%SF₅₀₀_{0.30 836 942 2.3 - 1.3 117} H8 80%Anl+20%FA 616 0.30 680 865 2.8 - 3.0 98 H9 100%Deg400 500 0.30 812 916 2.3 - 2.9 102

1) Assuming that the efficiency factor of silica fume is 1 and fly ash is 0.3. 2) Sp – Super-plasticizer. Cementa 92M

3) AEA – Air entraining agent. Cementa L14 4) According to SS 13 72 10

5) Anl - Anläggningscement (Swedish SRPC) 6) Slite - Slite cement (Swedish OPC) 7) SF - Silica fume (Elkem. Norway) 8) DK - Aalborg Lav cement (Danish SRPC) 9) FA - Fly ash (Aalborg . Denmark)

10) Deg400 - Degerhamn 400 cement (another type of Swedish SRPC)

It should be noticed that the slurry silica fume was used in concrete H-series, while the powder one was used in the other types of concrete.

2.2 Exposure conditions at the Träslövsläge field site

An overview of the Träslövsläge field site is shown in Fig. 2.1. The chloride concentration in the seawater varies from 10 to 18 g Cl per litre, with an average value of about 14 g Cl per litre. The typical water temperature is illustrated in Fig. 2.2 and has an annual average +11 °C.

(10)

Fig. 2.1 Overview of the Träslövsläge field site.

In the sine function: Tmax = 20 °C, Tmin = 2 °C, t0 =119 days

1994-12-30 is taken as the original point

Fig. 2.2 Annual temperature in the seawater.

0 5 10 15 20 25 1994-07-01 1994-09-30 1994-12-30 1995-03-31 1995-06-30 Temperature, °C Measured value Measured Mean Sine curve

(11)

3 Measurements of Chloride and Moisture Profiles

3.1 Sampling

In the previous investigations, the concrete slabs were taken back to the laboratory. A 30~50 mm thick slice was cut away from one of the outmost sides of each slab to avoid the influence of two-dimensional penetration, and a prism of size 1000×100×100 mm was then

successively cut as the sample specimen for different penetration zones, as shown in Fig. 3.1. After cutting for the specimens, the slabs were sent back to the field site for continuous

exposure. The cores were drilled in the exposure direction from different zones of each prism. In this investigation, cores were directly drilled at the field site. Two cores, one for chloride profile and another for moisture profile, were taken in respective zones. The sampling positions were chosen in such a way that the least distance between the curved surface of a core and the outmost side of a slab is about 50 mm to avoid the influence of two-dimensional penetration. From the previous investigations it has been known that the chloride profile from splash zone varies very much due to the unstable climate in this zone, but the profile is normally between those from atmospheric zone and submerged zone. To minimise the

laboratory work, no core was taken from splash zone except for three types of concrete (1-40, 2-40 and 3-40), which might be most commonly used in Sweden for infrastructures. Each core was sealed in double thick plastic bags when its surface was slightly dry, and then was transported to the respective laboratory for further sampling.

Fig. 3.1 Sampling illustration of a concrete slab after exposure.

3.2 Measurement of chloride profiles

The cores individually sealed in double thick plastic bags were stored in the laboratory at the room temperature not longer than two weeks prior to sampling. Powder samples were then taken from each core by means of dry-grinding on a lathe with a diamond tool (Fig. 3.2), successively from the exposed surface to a certain depth. The depth of each sample was

1 0 c m fo r sp e c im e n s 3 ~ 5 c m c u t a w a y S p e c im e n N o . 1 8 9 1 0 7 6 5 4 3 2 A tm o s p h e r ic z o n e S p la sh z o n e S u b m e r g e d z o n e 3 0 c m 3 0 c m 4 0 c m 1 0 c m fo r sp e c im e n s 3 ~ 5 c m c u t a w a y S p e c im e n N o . 1 8 9 1 0 7 6 5 4 3 2 A tm o s p h e r ic z o n e S p la sh z o n e S u b m e r g e d z o n e 3 0 c m 3 0 c m 4 0 c m

(12)

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 (Fig. 3.3). A sample size of about 1 gram was used to facilitate the parallel calcium analysis. According to the results from a Nordic comparison test (Tang, 1998) and an international inter-comparison test (Castellote and Andrade, 2001), this potentiometric titration technique reveals a good precision.

The technique for determination of soluble calcium content parallel to the determination of chloride content was originally developed at Chalmers. In the past year, a Nordtest project was carried out to evaluate the precision of this technique (Tang, 2003a). The results from the Nordic inter-laboratory comparison test reveal a satisfactory precision of this technique, whose pooled standard deviation of repeatability is 0.38 mass% of sample and pooled standard deviation of reproducibility is 0.90 mass% of sample.

A more detailed description of the method for determination of chloride and calcium contents is given in Appendix 1. Since similar techniques were employed in the precious investigations even though by different operators, the data from various investigations should be reliable and comparable.

(13)

Fig. 3.3 Device for potentiometric titration for chloride and calcium content in concrete.

3.3 Measurement of moisture profiles

The measurement of moisture profiles was carried out at Chalmers. The cores individually sealed in double thick plastic bags were stored in the laboratory at the room temperature not longer than a few days prior to sampling. A slice of about 10~20 mm thick was split from each concrete core, starting from one of the ends, 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.

(14)

4 Measured Chloride and Moisture Profiles

4.1 Data

format

From the BMB project the chloride ingress profiles were measured after 0.5, 1, 2 and 5 years for the selected types of concrete. In this project the profiles in all the types of concrete except for those with high water-binder ratios (w/b > 0.5) or the specimens are exhausted due to previous sampling. During the 10 years period, the software of personal computer has been updated a number of times and the data format of chloride profiles varied from time to time. In order to facilitate the database establishment, the format of data has to be consistent. Therefore, all the previous data of chloride profiles were re-formatted following the protocol used in this project. The moisture profiles were measured on two occasions, that is, after 5 and 10 years exposure. Therefore, it is relatively easier to keep the consistent format. The collected data of chloride and moisture profiles are compiled in Chalmers Publication P-03:03 (Tang, 2003b).

4.2 Effect of exposure zones

The chloride profiles in concrete with three typical types of binder and water-binder ratio are summarised in Fig 4.1. It should be noticed that, if not otherwise stated, the profile data are hereafter referred to 10 years exposure. It can be seen that the chloride ingress in submerged zone is in general the severest among all the three exposure zones, while the chloride ingress in splash zone may be similar to or less than that in submerged zone. This is in agreement with the results from the previous investigations after 5 years exposure (Tang and Andersen, 2000), where it was shown that chloride ingress in various types of concrete follows the order of submerged > splash > atmospheric zone. From the moisture profiles as shown in Figs. 4.2 and 4.3 it can be seen that the concrete exposed in submerged zone shows highest moisture content, which provides paths for chloride ions to diffuse into concrete. It is interesting to see that the moisture profiles in splash zone and atmospheric zone are not far different, but the chloride ingress in splash zone is remarkably severer than in atmospheric zone. This can be explained by side diffusion and capillary suction.

It can also be seen from Fig. 4.1 that chloride ingress from the backside (corresponding to the topside in casting) of the slab is in general severer than that from the front side

(corresponding to the bottom side in casting), even though the moisture profiles do not show such a tendency. Apparently, this is attributed to segregation, as indicated in Fig. 4.4, where the backside of slabs contains more binder than the front side. This segregation makes the concrete in the topside more permeable than in the bottom side.

The measured degree of capillary saturation is sometimes (e.g. in Fig. 4.3) larger than 1, implying that these specimens lost weight when put in capillary contact water in the laboratory after having been exposed in seawater for ten years at a small depth. Most

probably, it is the air voids that were partly water filled after the exposure, but were emptied when the specimens were taken out of the seawater. Further study is needed to clarify this phenomenon.

When comparing the 10 years profiles with 5 years profiles (Figs. 4.5 and 4.6) from splash zone it can be observed that, even though the moisture contents from splash zone do not significantly increase, the chloride ingress in splash zone significantly increased in the past 5 years, especially in the concrete with Anl cement. This could imply that the chloride ingress

(15)

According to Tang and Sandberg (1996), the chloride penetration profiles in splash zone is, however, very sensitive to the position where the core is taken. Little difference in the coring position may result in large difference in chloride profile. Further observations from a longer exposure period are needed to verify the above phenomenon.

Fig. 4.1 Profiles of chloride ingress in concrete under various exposure zones.

0 2 4 6 8 0 25 50 75 100 C l% of bi nde r

Submered Splash Atmospheric Concrete 1-40, 100%Anl 0 2 4 6 8 0 25 50 75 100 C l% of bi nde r Concrete 2-40, 100%Slite 0 2 4 6 8 0 25 50 75 100 Depth (m m) C l% of bi nde r Concrete 3-40, 95%Anl + 5%SF

(16)

Fig. 4.2 Profiles of relative humidity in concrete under various exposure zones after 10 years exposure. 50 60 70 80 90 100 0 25 50 75 100 Re la ti ve Hu m id ig y, % R H

Submerged Splash Atmospheric Concrete 1-40, 100%Anl 50 60 70 80 90 100 0 25 50 75 100 Re la ti ve Hu m id ig y, % R H Concrete 2-40, 100%Slite 50 60 70 80 90 100 0 25 50 75 100 Depth (m m ) Re la ti ve Hu m id ig y, % R H Concrete 3-40, 95%Anl + 5%SF

(17)

Fig. 4.3 Profiles of degree of capillary saturation in concrete under various exposure zones after 10 years exposure.

0.5 0.6 0.7 0.8 0.9 1 1.1 0 25 50 75 100 D eg ree o f C ap ill ar y S at u ra ti o n

Submerged Splash Atmospheric Concrete 1-40, 100%Anl 0.5 0.6 0.7 0.8 0.9 1 1.1 0 25 50 75 100 D egr ee o f C api lla ry S atu ra ti on Concrete 2-40, 100%Slite 0.5 0.6 0.7 0.8 0.9 1 1.1 0 25 50 75 100 Depth (mm ) D egr ee o f C api lla ry S at u ra ti on Concrete 3-40, 95%Anl + 5%SF

(18)

Fig. 4.4 Distributions of binder in concrete.

Fig. 4.5 Comparison between moisture profiles after 5 and 10 years exposure.

Submerged Zone 0 5 10 15 20 25 30 0 25 50 75 100 Depth (m m) B ind er c o nt en t, % of s ampl e 1-40, 100%Anl 2-40, 100%Slite 3-40, 95%Anl+5%SF 50 60 70 80 90 100 0 10 20 30 40 50 x, m m RH % Concrete 1-40, 100%Anl 5 years exposure 50 60 70 80 90 100 0 10 20 30 40 50 RH % Concrete 1-40, 100%Anl 10 years exposure 50 60 70 80 90 100 0 10 20 30 40 50 x, m m RH %

Submerged Splash Atmospheric

Concrete 3-40, 95%Anl + 5%SF 5 years exposure 50 60 70 80 90 100 0 10 20 30 40 50 x, mm RH % Concrete 3-40, 95%Anl + 5%SF 10 years exposure

(19)

Fig. 4.6 Comparison between chloride profiles after 5 and 10 years exposure. 0 2 4 6 8 0 10 20 30 40 50 x, mm C l% of bi nd er Concrete 1-40, 100%Anl 5 years exposure 0 2 4 6 8 0 10 20 30 40 50 x, mm C l% of bi n d er Concrete 2-40, 100%Slite 5 years exposure 0 2 4 6 8 0 10 20 30 40 50 x, mm C l% of bi nd er Concrete 1-40, 100%Anl 10 years exposure 0 2 4 6 8 0 10 20 30 40 50 x, mm C l% of bi nd er Concrete 2-40, 100%Slite 10 years exposure 0 2 4 6 8 0 10 20 30 40 50 x, mm C l% of bi nd er

Submered Splash Atmospheric Concrete 3-40, 95%Anl + 5%SF 5 years exposure 0 2 4 6 8 0 10 20 30 40 50 x, mm C l% of bi n d er Concrete 3-40, 95%Anl + 5%SF 10 years exposure

(20)

4.3 Effect of water-binder ratios

Since the chloride ingress in submerged zone is severest as shown in section 4.2, the

discussions hereafter will be limited to the submerged zone. The chloride profiles in concrete with different water-binder ratios are summarised in Figs. 4.7 to 4.9. As expected, lower water-binder ratio results in less chloride ingress, especially in the concrete with Anl cement. There exists a limitation: when water-binder ratio is very low, e.g. 0.25, it may not help to reduced chloride ingress, as shown in Fig. 4.9.

Fig. 4.7 Chloride profiles in concrete with Anl cement.

Fig. 4.8 Chloride profiles in concrete with Slite cement.

Submerged Zone, 100%Slite

0 2 4 6 0 10 20 30 40 50 Depth (m m) C l% of bi nd er 2-35, w /b 0.35 2-40, w /b 0.40 2-50, w /b 0.50

Submerged Zone, 100%Anl

0 2 4 6 0 10 20 30 40 50 Depth (mm ) C l% of bi nde r H3, w /b 0.30 1-35, w /b 0.35 1-40, w /b 0.40

(21)

Fig. 4.9 Chloride profiles in concrete with Anl cement blended with 5% silica fume. The moisture profiles in concrete with different water-binder ratios are summarised in Figs. 4.10 to 4.12. Also as expected, less chloride ingress is in general related to a lower moisture profile, even thought it seems difficult to find a clear relationship between chloride and moisture profiles.

Fig. 4.10 Moisture profiles in concrete with Anl cement.

70 80 90 100 0 10 20 30 40 50 Re la ti ve Hu m id it y, % R H 0.7 0.8 0.9 1 1.1 0 10 20 30 40 50 Depth (m m) D eg ree o f C ap ill ar y S atur ati on H3, w /b 0.30 1-35, w /b 0.35 1-40, w /b 0.40

Submerged Zone, 95%Anl + 5%SF

0 2 4 6 0 10 20 30 40 50 Depth (m m) C l% of bi nd er H5, w /b 0.25 H1, w /b 0.30 3-35, w /b 0.35 3-40, w /b 0.40

(22)

Fig. 4.11 Moisture profiles in concrete with Slite cement.

Fig. 4.12 Moisture profiles in concrete with Anl cement blended with 5% silica fume.

70 80 90 100 0 10 20 30 40 50 Re la ti ve Hu m id it y, % R H 0.7 0.8 0.9 1 1.1 0 10 20 30 40 50 Depth (mm ) D eg ree o f C ap ill ar y S atu ra ti o n 2-35, w /b 0.35 2-40, w /b 0.40 2-50, w /b 0.50

70 80 90 100 0 10 20 30 40 50 Re la ti ve Hu m id it y, % R H 0.7 0.8 0.9 1 1.1 0 10 20 30 40 50 Depth (m m) D eg ree o f C ap ill ar y S atu ra ti o n H5, w /b 0.25 H1, w /b 0.30 3-35, w /b 0.35 3-40, w /b 0.40

(23)

4.4 Effect of entrained air

The chloride profiles in concrete with and without addition of AEA (Air Entraining Agent) are summarised in Figs. 4.13 to 4.15, and the corresponded moisture profiles are shown in Figs 4.16 to 4.18. It seems that the 6% entrained air tends to increase chloride ingress. The least chloride ingress in the concrete with 3% entrained air (Fig. 4.15) might be attributed to a better compaction under such an air content, but the conclusion cannot be drawn from this single profile.

Fig. 4.13 Chloride profiles in concrete with and without entrained air, Anl cement.

Fig. 4.14 Chloride profiles in concrete with and without entrained air, Slite cement.

0 2 4 6 0 10 20 30 40 50 Depth (mm ) C l% of bi nde r 1-35, 6%air 1-40, 6%air 7-35, non-AEA 7-40, non-AEA

0 2 4 6 0 10 20 30 40 50 Depth (m m) C l% of bi nd er 2-35, 6%air 2-40, 6%air 2-50, 6%air 8-35, non-AEA 8-40, non-AEA 8-50, non-AEA

(24)

Fig. 4.15 Chloride profiles in concrete with and without entrained air, Anl cement with 5% silica fume.

Fig. 4.16 Moisture profiles in concrete with and without entrained air, Anl cement.

0 2 4 6 0 10 20 30 40 50 Depth (m m) C l% of bi nd er 3-40, 6%air 5-40, 3%air 6-40, non-AEA

70 80 90 100 0 10 20 30 40 50 Re la ti ve Hu m id it y, % R H 0.7 0.8 0.9 1 1.1 0 10 20 30 40 50 Depth (m m) D eg ree o f C ap ill ar y S atur ati on 1-35, 6%air 1-40, 6%air 7-35, non-AEA 7-40, non-AEA

(25)

Fig. 4.17 Moisture profiles in concrete with and without entrained air, Slite cement.

Fig. 4.18 Moisture profiles in concrete with and without entrained air, Anl cement with 5% silica fume.

70 80 90 100 0 10 20 30 40 50 Re la ti ve Hu m id it y, % R H 0.7 0.8 0.9 1 1.1 0 10 20 30 40 50 Depth (m m) D eg ree o f C ap ill ar y S atu ra ti o n

2-35, 6%air 2-40, 6%air 2-50, 6%air

8-35, non-AEA 8-40, non-AEA 8-50, non-AEA

70 80 90 100 0 10 20 30 40 50 Re la ti ve Hu m id it y, % R H 0.7 0.8 0.9 1 1.1 0 10 20 30 40 50 Depth (m m) D eg ree o f C ap ill ar y S atu ra ti o n

(26)

4.5 Effect of binder type

The chloride profiles in concrete with various types of binder are summarised in Figs. 4.19 to 4.21, and the corresponded moisture profiles are shown in Figs 4.22 to 4.24. It can be seen from Fig.4.19 that Slite cement reveals lower chloride profile than Anl cement, but the significance becomes negligible when water-binder ratio is reduced to 0.35. There is no significant difference in moisture profiles between Slite cement and Anl cement.

No significant difference in chloride profile between Swedish Anl cement and Danish cement could be found (Figs. 4.19 and 4.20), even though there exists some differences in their moisture profiles (Figs 4.22 and 4.23).

Addition of silica fume and fly ash in Swedish Anl cement and Danish cement effectively increases the resistance to chloride ingress (Figs. 4.19 to 4.21), but not in Deg400 cement, which already reveals a better resistance to chloride ingress when compared with pure Anl cement (Fig. 4.21). It seems that addition of fly ash by 5% did not significantly improve the resistance, while 20% fly ash can remarkably increase the resistance (Fig. 4.21).

Fig. 4.19 Chloride profiles in concrete with water-binder ratio 0.4.

Subm e rge d Zone , w /b 0.40

0 2 4 6 0 10 20 30 40 50 De pth (m m ) C l% of bi nd e r 1-40, A nl, 0% SF 2-40, Slite, 0% SF 3-40, A nl, 5% SF 9-40, DK, 5% SF 4-40, A nl, 10% SF 10-40, DK, 17% FA +4.5% SF

(27)

0 2 4 6 0 10 20 30 40 50 De pth (m m ) C l% of bi nde r 1-35, A nl, 0% SF 2-35, Slite, 0% SF 3-35, A nl, 5% SF 11-35, DK, 10% FA +5% SF 12-35, A nl, 10% FA +5% SF

0 2 4 6 0 10 20 30 40 50 De pth (m m ) C l% of bi nde r H3, A nl, 0% SF H1, A nl, 5% SF H2, A nl, 10% SF H6, A nl, 5% FA H8, A nl, 20% FA H9, Deg400, 0% SF H7, Deg400, 5% SF

(28)

Fig. 4.22 Moisture profiles in concrete with water-binder ratio 0.4. Submerged Zone, w/b 0.40 70 80 90 100 0 10 20 30 40 50 Re la ti ve Hu m id it y, % R H 0.7 0.8 0.9 1 1.1 0 10 20 30 40 50 Depth (mm ) D egr ee of C api lla ry S at u ra ti on 1-40, Anl, 0%SF 2-40, Slite, 0%SF 3-40, Anl, 5%SF 9-40, DK, 5%SF 4-40, Anl, 10%SF 10-40, DK, 17%FA+4.5%SF

(29)

Fig. 4.23 Moisture profiles in concrete with water-binder ratio 0.35. Submerged Zone, w/b 0.35 70 80 90 100 0 10 20 30 40 50 Re la ti ve Hu m id it y, % R H 0.7 0.8 0.9 1 1.1 0 10 20 30 40 50 Depth (mm ) D eg ree o f C ap ill ar y S at u ra ti o n

1-35, Anl, 0%SF 2-35, Slite, 0%SF 3-35, Anl, 5%SF

(30)

Fig. 4.24 Moisture profiles in concrete with water-binder ratio 0.3. Submerged Zone, w/b 0.30 70 80 90 100 0 10 20 30 40 50 Re la ti ve Hu m id it y, % R H 0.7 0.8 0.9 1 1.1 0 10 20 30 40 50 Depth (mm ) D egr ee of C api lla ry S at u ra ti on

H3, Anl, 0%SF H1, Anl, 5%SF H2, Anl, 10%SF

H6, Anl, 5%FA H8, Anl, 20%FA H9, Deg400, 0%SF

(31)

5 Modelling of Chloride Ingress

5.1 Curve-fitted diffusion coefficient

According to Fick’s second law, the following error function solution can be used to describe diffusion under the semi-infinite boundary:

( )

(

)

                − ⋅ − + = t D x C C C t x C F2 ini s ini 2 erf 1 , (5.1)

where C(x,t) is the chloride concentration at depth x after exposure period t, Cini is the initial

chloride concentration in concrete, Cs is the chloride concentration at the exposure surface,

DF2 is the chloride diffusion coefficient, and erf is the error function. When a concrete slab

with a limited thickness of 2L is exposed to the seawater, as in our case at the Träslövsläge field site, the chlorides from the seawater can penetrate into concrete from the both sides of the slab. If the chlorides have penetrated through the centre of the slab, the above equation cannot be used because the boundary is no longer the semi-infinite. Instead, the following equation should be used for such a case according to Nilsson’s suggestion (2003):

( )

(

)

      ⋅ − + = _ini _s _ini , ₀ , F L x U C C C t x C (5.2) where

( )

₍

₎

(

)

            − +               + − + − − =      

_∑

∞ = L x n F n n F L x U n 1 2 1 2 cos 2 1 2 exp 1 2 1 4 1 , ₀ 2 2 0 2 0 π π π (5.3)

and the Fourier number F0 is equal to

2 F2 0 L t D F = (5.4)

Theoretically, only the gradient of free chloride concentration is contributed to the driving force for chlorides to penetrate into concrete. However, the free chloride profiles can, up to now, hardly be measured. The reported chloride profiles are normally based on the

determination of total chloride content. The total chloride content is not necessarily proportional to the free chloride concentration due to the non-linear behaviour of chloride binding. Therefore, using the gradient of total chloride content as a driving force in Fick’s law may not be theoretically correct, but just empirically convenient for describing the

characteristics of penetration profiles. The models based on the above equations may be regarded as empiric models. Nevertheless, curve-fitting the measured chloride profile to equation (5.1) or (5.2), one can obtain two parameters, DF2 and Cs. These two parameters are

anyway a description of chloride ingress under a specific exposure condition and after a specific exposure period. A MS Excel-based program was used for curve-fitting calculations. For each chloride profile, the first one or two points were omitted prior to the calculation, if the values are significantly out of the diffusion curve. The initial chloride content was assumed as 0.01 mass% of binder. The curve-fitted results are given in Appendix 2 and an

(32)

example of relationships between curve-fitted parameters and exposure time is shown in Fig. 5.1. Although some researchers assume that the curve-fitted diffusion coefficient DF2

decreases with the exposure time, from this investigation it is difficult to find a clear tendency of decreasing DF2. It appears that in the first five years DF2 decreases to some extent, but

afterwards it becomes more or less uncertain. Due to this uncertainty, it may be difficult to use the value of DF2 obtained at an early age to predict the long-term chloride ingress. On the

other hand, the surface chloride content Cs gradually increased in the first 5 years and then

kept more or less unchanged. The increase in Cs is due to the increased chloride binding and

also probably due to the increase in saturation degree of air pores. This increased Cs may

mislead a decreased DF2 when using equations (5.1) or (5.2) for calculating DF2, because in

these equations Cs is assumed as constant, but it is not! From this point of view the hypothesis

of a constantly decreasing DF2 in some empiric models for service life prediction is very

questionable.

Fig. 5.1 Relationships between curve-fitted parameters and exposure time.

0.1 1 10

0 2 4 6 8 10 12

Exposure time, years

D F2 , x 1 0 -1 2 m 2 /s 1-40 & Ö 2-40 3-40 10-40 12-35 H1 H2 H3 H5 H8 0 2 4 6 8 0 2 4 6 8 10 12

Exposure time, years

C s , m a s s % of bi nde r 1-40 & Ö 2-40 3-40 10-40 12-35 H1 H2 H3 H5 H8

(33)

Fig. 5.2 shows the relationships between the curve-fitted parameters and water-binder ratio. It is clear that DF2 significantly decreases as water-binder ratio decreases. This is in agreement

with the observations from the penetration profiles, as shown in Figs. 4.7 to 4.9. Addition of silica fume remarkably decreased the values of DF2, implying an improvement of chloride

resistance. Addition of 20% fly ash has also improved the resistance to chloride ingress. No relationship between Cs and water-binder ratio could be found from Fig.5.2.

Fig. 5.2 Example of uncertain relationship between curve-fitted DF2 and exposure time.

0.1 1 10 0.20 0.30 0.40 0.50 0.60 Water-binder ratio DF2 , x10 -1 2 m 2 /s 0 2 4 6 0.20 0.30 0.40 0.50 0.60 Water-binder ratio Cs , m as s% of binde r SRPC OPC SRPC+5%SF SRPC+10%SF SRPC+5%FA SRPC+20%FA SRPC+5%SF+10-17%FA

(34)

5.2 Chloride migration coefficient from the CTH method

In the BMB project, a rapid test method for determination of chloride transport property in concrete was developed at the Chalmer, called the CTH method (Tang & Nilsson, 1992; Tang, 1996). The test is based on the principle of ionic migration under an external electrical field (Fig. 5.3). This external electrical field can greatly accelerate the transport process of chloride ions in concrete.

Fig. 5.3 Test arrangement of the CTH method.

The method involves applying an external potential of 10 to 60 V across a 50 mm thick disc specimen for certain test duration (in most cases 24 hour), then splitting the specimen and measuring the penetration depth of chlorides by using a colourimetric method (Fig. 5.4). The chloride migration coefficient is then calculated from the measured chloride penetration depth according to the following equations:

t x x U F z L T R D d d CTH α − ⋅ = (5.5) where:       − ⋅ = − 0 d 1 ₁ 2 erf 2 c c U F z L T R α (5.6) DCTH: migration coefficient, m2/s;

z: absolute value of ion valence, for chloride ions, z = 1;

+ - Potential (DC) a. Rubber sleeve b. Anolyte c. Anode d. Specimen e. Catholyte f. Cathode g. Plastic support h. Plastic box a b c d e f g h

(35)

U: absolute value of the potential difference across the specimen, V; R: gas constant, R = 8.314 J/(K·mol);

T: solution temperature, K; L: thickness of the specimen, m; xd: penetration depth, m;

t: test duration, second; erf-1_{: inverse of error function;}

cd: chloride concentration at which the colour changes, cd ≈ 0.07 N;

c0: chloride concentration in the upstream cell, c0 ≈ 2 N.

Fig. 5.4 Example of chloride penetration depth (white colour) after the migration test. This method has been used to measure DCTH on all the types of “noll” concrete specimens,

and also on the specimens of non-chloride contaminated portion (e.g. in the central portion of a core) taken from the field site after having been exposed for one to two years. The test results are summarised in Appendix 3, and the detailed information about the test conditions can be found elsewhere (Tang, 1997). Owing to its simplicity and rapidity, the CTH method is very useful for evaluating the resistance of concrete to chloride ingress. In 1999, the method has been standardised as NT BUILD 492. A comparison between the values of DCTH

(36)

Fig. 5.5 Comparison between DCTH measured at an age of about a half of year and DF2

curve-fitted from the profiles after different exposure periods in submerged zone. It can be seen that there exist fairly good linear relationships in logarithmical scale between DF2 and DCTH. The variation in regression coefficients with different exposure periods is

probably due to the fact that the curve-fitted DF2 varies with exposure time. Since the

parameters DF2 and DCTH involves both the material property and the exposure condition, it is

not expected to compare their absolute values, but the sensitivity to ranking the concrete with regard to chloride ingress. From the above figure and regression results it is easy to know that ∆DCTH/DCTH is always larger than ∆DF2/DF2, implying that the CTH method is really a

sensitive tool for ranking concrete’s resistance to chloride ingress. When DCTH is less than

y = 0.4395x0.586 R2_{= 0.7798} y = 0.8266x0.5681 R2_{= 0.9279} y = 0.5003x0.6319 R2_{= 0.7977} y = 0.8075x0.5063 R2_{= 0.8891} y = 1.7439x0.4522 R2_{= 0.8641} 0.1 1 10 100 0.1 1 10 100

DCTH, x10-12 m²/s (measured at age 6 months)

DF2 , x10 -1 2 m ²/s (c u rve-fi tt e d af te r d if fer en t t) t = 0.6-0.9yr t = 1-1.3yr t = 2-2.3yr t = 5.1-5.4yr t = 10.1-10.5yr Submerged zone 0.1 1 10 100 0.1 1 10 100

DCTH, x10-12 m²/s (measured at age 6 months)

DF2 , x10 -1 2 m ²/s ( c u rve-fi tt e d af te r 10 year s exp o su re ) 1-series 2-series 3-series 4-40 5-40 6-series 7-series 8-series 9-40 10-40 11-35 12-35 H-series Submerged zone

(37)

1 ×10-12 m2/s. Comparing Figs. 4.19 to 4.21 with the DCTH values in Appendix 3 it can be

found that the lowest DCTH value in the w/c 0.4 group is concrete 10-40, in the w/c 0.35 group

is concrete 11-35, and in the w/c 0.3 group is concrete H2. All these three types of concrete reveal the lowest chloride ingress in their group, as shown in Figs. 4.19 to 4.21. This indicates that the CTH method is indeed measuring the resistance of concrete to chloride ingress and is applicable to various types of concrete including silica fume and fly ash.

5.3 ClinConc

model

The model ClinConc (Cl in Concrete) was also developed during the BMB project in the middle of 1990’s (Tang & Nilsson 1994; Tang 1995). The model consists of two main procedures:

1) Simulation of free chloride penetration through the pore solution in concrete using a genuine flux equation based on the principle of Fick’s law with the free chloride concentration as the driving potential, and

2) Calculation of the distribution of the total chloride content in concrete using the mass balance equation combined with non-linear chloride binding. Not like other models, a unique character of the model ClinConc is that the chloride diffusivity, which can be determined by, e.g. the CTH method, is considered as a material property. It changes only when concrete is young, like many other material properties, such as porosity and strength. After an age of a half of year, this diffusivity becomes more or less constant according to the experiments (Tang & Nilsson 1992b; Tang 1996). Another unique character of the model ClinConc is that the climatic parameters, such as chloride

concentration and temperature, are used in both the flux and the mass balance equations. Therefore, the model can well describe the effects of exposure conditions on chloride penetration.

In this study, a MS Excel based version 4b, dated 2001 May, was used for calculation. This version is applicable to various exposure environments including alternative wet-and-dry ones (Tang & Nilsson, 2000; 2000b; 2002). In fact, nothing except for the exposure conditions has been modified in the latest version. The calculation work in this study was limited to

submerged zone and to the types of concrete with w/b ≤ 0.5 and at least two chloride profiles available. Some commonly used parameters that were also used in the previous modelling are listed in Table 5.1. Since this study covers various types of binder, the chloride binding coefficient fb has to be considered individually. Based on the previous experimental study

(Tang & Nilsson, 1993), the basic value of fb was taken as 3.6 for plain Portland cement. It

was found from both the previous studies and this study that this value is also applicable to the concrete with silica fume. For the concrete with fly ash, a higher value of binding coefficient was considered in the modelling, as listed in Table 5.2.

(38)

Table 5.1. Common parameters used in the calculation. Activation energy for chloride binding: Eb = 40000 J/mol

Chloride binding exponent: B = 0.38

Time-dependent factor for chloride binding: f_t =a_tln

(

t_Cl +0.5

)

+1

where at is given in Table 5.2 andtCl is the local chloride contamination time in years.

Activation energy for diffusivity: ED = 42000 J/mol

Depth dependent for diffusivity: None (steel form) Age at which DCTH becomes constant: tDa = 180 days

Age-dependent exponent for diffusivity:  ⋅

                   − + = β − 60. 2 t 0.152 2.098 3.497 b w b SF b SF

where SF, b and w are content of silica fume, binder and water, respectively. Specimen thickness: 100 mm

Exposure boundary: Two sides Exposure zone: Submerged

Average temperature in seawater: +11 ºC

(39)

Mix No. w/b fb at kD Notes 1-35 0.35 0.42 1.5 1-40/Ö 0.40/0.38 3.6 0.36 1 1-50 0.50 0.30 1 2-35 0.35 0.42 2.5 2-40 0.40 3.6 0.25 1 2-50 0.50 0.10 1 3-35 0.35 0.50 2 3-40 0.40 3.6 0.36 1.5 3-50 0.50 0.25 1 4-40 0.40 3.6 0.36 1 9-40 0.40 3.6 0.36 1.5 10-40 0.40 4.6 0.42 5 17%FA+4.5%SF 11-35 0.35 4.6 0.50 6 10%FA+5%SF 12-35 0.35 4.6 0.50 4 10%FA+5%SF H1 0.30 3.6 0.60 2 H2 0.30 3.6 0.50 3 H3 0.30 3.6 0.50 2 H4 0.40 3.6 0.42 1.5 H5 0.25 3.6 0.60 2 H6 0.30 4.2 0.36 1 5%FA H7 0.30 3.6 0.50 5 Deg400+5%SF H8 0.30 5.5 0.50 2 20%FA H9 0.30 3.6 0.36 1 Deg400

where fb – Chloride binding coefficient;

at – Coefficient to the time dependent factor ft in Table 5.1;

kD – Extension coefficient to DCTH.

Coefficient at in Table 5.2 reflects the increment of surface chloride content in concrete. A

high value corresponds a large increment, and vice versa. It can be seen from Table 5.2 that this value is normally high for the concrete with low w/c and/or blended with silica fume or fly ash, but low for the concrete with Slite cement. Probably at value is related to the

long-term hydration process. Slite cement is finer than the other types of cement and its hydration process is relatively quicker. Therefore, the chloride binding process is also shorter.

From the previous studies (e.g. Tang & Nilsson 2002b) it has been found that the calculated results underestimate chloride ingress in concrete with low water-binder ratios or blended with silica fume or fly ash. There might be two possible reasons: 1) the laboratory test method

(40)

underestimates the actual chloride diffusion coefficient, and/or 2) the actual diffusion coefficient in such types of concrete may be larger than the one measured on the laboratory specimens. In order to best fit the measured profiles, an extension coefficient kD was simply

introduced to extend the value of DCTH. The value of kD in Table 5.2 is normally equal to 1 for

plain cement concrete with w/b > 0.35. When w/b ≤ 0.35 or one type of pozzolanic materials, i.e. silica fume or fly ash, is blended in concrete, it will be in a range of 1.5 to 2.5. When both silica fume and fly ash are blended in concrete, the value of kD could be up to 4 to 6. The

large value of kD implies that the diffusion coefficient of such types of concrete was

underestimated by the CTH method, or the concrete has become more permeable after in-field exposure, due to e.g. temperature cycles. The latter effect could even be seen from the values of DCTH measured after exposed at the field site for one to two years, as shown in Fig. 5.6.

These values are in many cases larger than those measured on the specimens stored in the laboratory. Further experiment will be arranged to find more evidences of this phenomenon, so as to be able to verify the extension coefficient kD.

Fig. 5.6 Comparison in DCTH measured on the specimens in laboratory and exposed in the

field (see Appendix for the detailed data).

5.4 Modelled results and discussions

The results calculated by the ClinConc model will be presented and discussed in order of binder type.

Swedish SRPC - Anl cement

The calculated results are shown in Figs. 5.7 to 5.9. It should be noticed that the measured profiles year 1, 2 and 5 in Fig. 5.8 were taken from concrete Ö (the mixture proportion similar to that used in the repair work of the Öland bridge at the east coast of Sweden), which is basically similar to concrete 1-40. It can be seen that the predicted profiles are in fairly good agreement with the measured profiles. The deviations at the points of 40 and 50 mm depths in

0.1 1 10 100 0 0.5 1 1.5 2 2.5 t, ye a r DCT H , x 1 0 -1 2 m ²/s 1-40 (Ö) 1-50 2-40 2-50 3-40 3-50 10-40 12-35 H1 H2 H3 H4 H5 H8 Lab specim ens

(41)

during the casting) where the concrete is normally more permeable due to segregation.

Fig. 5.7 Chloride ingress profiles in concrete 1-35 (Anl cement, w/b 0.35) up to 10 years exposure in submerged zone.

Fig. 5.8 Chloride ingress profiles in concrete 1-40 & Ö (Anl cement, w/b 0.4 or 0.38) up to 10 years exposure in submerged zone.

0 1 2 3 4 5 0 10 20 30 40 50 x , mm To ta l C l% of bi n d e r Predicted 0.8 y 10.3 y 0 1 2 3 4 5 0 10 20 30 40 50 x , mm To ta l C l% o f b inde r Predicted 0.8 y 1 y 2 y 5 y 10.3 y

(42)

Fig. 5.9 Chloride ingress profiles in concrete 1-50 (Anl cement, w/b 0.5) up to 10 years exposure in submerged zone.

The measured profiles year 1 and 2 in Fig. 5.9 overlapped each other, probably due to the quality variation in concrete or the measurement uncertainty in chloride profiling.

Swedish OPC - Slite cement

The results from concrete 2-series are shown in Figs. 5.10 to 5.12. Again, the predicted profiles are also comparable with the measured profiles.

Fig. 5.10 Chloride ingress profiles in concrete 2-35 (Slite cement, w/b 0.35) up to 10 years exposure in submerged zone.

0 1 2 3 4 5 6 0 10 20 30 40 50 x , mm To ta l C l% o f b inde r Predicted 0.8 y 1.2 y 2.2 y 5.2 y 0 1 2 3 4 5 0 10 20 30 40 50 x , mm To ta l C l% o f b inde r Predicted 1.3 y 10.5 y

(43)

0 1 2 3 4 5 0 10 20 30 40 50 x , mm T o ta l C l% of bi n d e r Predicted 1.3 y 2.3 y 5.3 y 10.5 y 0 1 2 3 4 5 0 10 20 30 40 50 x , mm To ta l C l% o f b inde r Predicted 0.9 y 1.3 y 10.5 y

(44)

Addition of silica fume

The results from the concrete blended with silica fume are shown in Figs. 5.13 to 5.17. The predicted profiles are still in good agreement with the measured profiles when a value of kD =

1.5∼2 was used for this type of concrete (see Table 5.2).

Fig. 5.13 Chloride ingress profiles in concrete 3-35 (Anl cement + 5%SF, w/b 0.35) up to 10 years exposure in submerged zone.

0 1 2 3 4 5 0 10 20 30 40 50 x , mm Tot a l C l% of bi nde r Predicted 0.8 y 10.3 y 0 1 2 3 4 5 0 10 20 30 40 50 x , mm Tot a l C l% of bi nde r Predicted 0.8 y 1.2 y 2.2 y 5.2 y 10.3 y

(45)

0 1 2 3 4 5 0 10 20 30 40 50 x , mm Tot a l C l% of bi nde r Predicted 0.8 y 1.2 y 2.2 y 5.2 y 0 1 2 3 4 5 0 10 20 30 40 50 x , mm Tot a l C l% of bi nde r Predicted 0.7 y 10.2 y

(46)

Fig. 5.17 Chloride ingress profiles in concrete 9-40 (Danish SRPC + 5%SF, w/b 0.4) up to 10 years exposure in submerged zone.

Addition of silica fume together with fly ash

The results from the concrete blended with both silica fume and fly ash are shown in Figs. 5.18 to 5.20. It appears that the predicted profiles better fit to the actual profiles measured after 1 to 2 years and after 10 years, but not to the 5-year profiles, which reveal deeper

chloride ingress than the predicted. Considering the fact that the chloride ingress in such types of concrete is less than 30 mm even after 10 years exposure, the discrepancy in 5-years

profiles might be attributed to the variation of samples. It should be noticed that, for these three types of concrete, the significant high value of kD (4∼6) was used in the modelling. If

this is because of the underestimation of diffusion coefficient by the CTH method, it cannot explain the fact that the lowest DCTH always corresponds to the lowest chloride ingress in the

group of concrete with similar water-binder ratios, as discussed in section 5.2. Therefore, the high value of kD is to a great extent attributed to the changes in diffusivity during the field

exposure. The consequence is, however, not serious according to the profiles after 10 years exposure, because these types of concrete still reveal lowest chloride ingress in all types of concrete with similar water-binder ratios, as shown in Figs. 4.19 and 4.20.

0 1 2 3 4 5 0 10 20 30 40 50 x , mm Tot a l C l% of bi nde r Predicted 0.7 y 10.2 y

(47)

Fig. 5.18 Chloride ingress profiles in concrete 10-40 (Danish SRPC + 17%FA + 4.5%SF, w/b 0.4) up to 10 years exposure in submerged zone.

Fig. 5.19 Chloride ingress profiles in concrete 11-35 (Danish SRPC + 10%FA + 5%SF, w/b 0.35) up to 10 years exposure in submerged zone.

0 1 2 3 4 5 0 10 20 30 40 50 x, mm Tot a l C l% of bi nde r Predicted 0.7 y 2.1 y 5.1 y 10.2 y 0 1 2 3 4 5 0 10 20 30 40 50 x , mm Tot a l C l% of bi nde r Predicted 0.7 y 10.2 y

(48)

Fig. 5.20 Chloride ingress profiles in concrete 12-35 (Anl cement + 10%FA + 5%SF, w/b 0.35) up to 10 years exposure in submerged zone.

H-series concrete

This series of concrete was intended to represent high performance concrete with low water-binder ratio and high water-binder content. In this series of concrete the slurry silica fume was used. The predicted results of chloride ingress from this series of concrete are shown in Figs. 5.21 to 5.29. It can be seen that the predicted results are fairly in agreement with the measured profiles, although there exist some discrepancies in the predicted and measured profiles for some of the concrete types. Again, considering the fact that the chloride ingress in this series of concrete is in most cases less than 20 mm, these discrepancies should be thought as small.

0 1 2 3 4 5 0 10 20 30 40 50 x , mm Total Cl% of binder Predicted 0.7 y 2.1 y 5.1 y 10.2 y

(49)

Fig. 5.21 Chloride ingress profiles in concrete H1 (Anl cement + 5%SF, w/b 0.3) up to 10 years exposure in submerged zone.

0 1 2 3 4 5 0 10 20 30 40 50 x , mm Tot a l C l% of bi nde r Predicted 1 y 2 y 5 y 10.2 y 0 1 2 3 4 5 0 10 20 30 40 50 x , mm Tot a l C l% of bi nde r Predicted 1 y 2 y 5 y 10.2 y

(50)

0 1 2 3 4 5 0 10 20 30 40 50 x , mm Tot a l C l% of bi nde r Predicted 1 y 2 y 5 y 10.2 y 0 1 2 3 4 5 0 10 20 30 40 50 x , mm Tot a l C l% of bi nde r Predicted 0.6 y 1 y 2 y 5 y

(51)

Fig. 5.26 Chloride ingress profiles in concrete H6 (Anl cement + 5%FA, w/b 0.3) up to 10 years exposure in submerged zone.

0 1 2 3 4 5 0 10 20 30 40 50 x , mm Tot a l C l% of bi nde r Predicted 1 y 2 y 5 y 10.2 y 0 1 2 3 4 5 0 10 20 30 40 50 x , mm Tot a l C l% of bi nde r Predicted 1 y 10.1 y

(52)

Fig. 5.27 Chloride ingress profiles in concrete H7 (Deg400 cement + 5%SF, w/b 0.3) up to 10 years exposure in submerged zone.

Fig. 5.28 Chloride ingress profiles in concrete H8 (Anl cement + 20%FA, w/b 0.3) up to 10 years exposure in submerged zone.

0 1 2 3 4 5 0 10 20 30 40 50 x , mm Tot a l C l% of bi nde r Predicted 1 y 10.1 y 0 1 2 3 4 5 0 10 20 30 40 50 x , mm Tot a l C l% of bi nde r Predicted 1 y 2 y 5 y 10.1 y

(53)

Fig. 5.29 Chloride ingress profiles in concrete H9 (Deg400 cement + 0%SF, w/b 0.3) up to 10 years exposure in submerged zone.

0 1 2 3 4 5 0 10 20 30 40 50 x, mm Tot a l C l% of bi nde r Predicted 1 y 10.2 y

Chloride ingress in concrete exposed to marine environment - field data up to 10 years exposure

Chloride Ingress in Concrete Exposed

to Marine Environment

– Field data up to 10 years exposure

Chloride Ingress in Concrete Exposed to Marine

Environment – Field data up to 10 years exposure

Abstract

Contents

Preface

Sammanfattning (Summary in Swedish)

1 Introduction

2

Concrete Specimens and Exposure Conditions

2.1 Concrete

slabs

2.2

Exposure conditions at the Träslövsläge field site

3

Measurements of Chloride and Moisture Profiles

3.1 Sampling

3.2

Measurement of chloride profiles

3.3

Measurement of moisture profiles

4

Measured Chloride and Moisture Profiles

4.1 Data

format

4.2

Effect of exposure zones

4.3

Effect of water-binder ratios

4.4

Effect of entrained air

4.5

Effect of binder type

5

Modelling of Chloride Ingress

5.1

Curve-fitted diffusion coefficient

( )

(

)

( )

(

)

( )

(

)

(

)

∑

5.2

Chloride migration coefficient from the CTH method

5.3 ClinConc

model

(

)

5.4

Modelled results and discussions

₍

₎

_∑