CONCRETE EXPOSURE CLASSES CONCERNINGCORROSION INDUCED BY CHLORIDES – CRITERIONS –EXSPERIMENTAL RESULTS AND POSSIBLE MISLEADS
Elica Marušić*, Nedjeljko Akrap*, Tin Dragović*, Inga Čujić*Marija Jurašin*
*Institut IGH, d.d., Laboratory IGH
Regional unit in Split, Matice hrvatske 15, 21000 Split, Croatia
e-mail: , web page:
Keywords:concrete, chlorides, exposure classes, criterions
Abstract: Chloride migration coefficient is one of the main parameters for describing the durability of concrete constructions concerning the risk of reinforcement corrosion caused by chlorides. Exposure classes related to corrosion induced by chlorides from sea water are defined in Croatian regulation, but criterions for satisfying exposure classes aren't. In Croatian practice these criterions are usually defined by values of chloride migration coefficient tested on concrete according to standard NT BUILD 492 from non-steady-state migration experiment There are also some values of mentioned coefficient that are accepted in practice for defining exposure classes, but they aren't officially approved.
During performance of laboratory tests it was noticed that results obtained by test method NT BUILD 492 on same concrete were significantly different depending on moisture content of concrete although preconditioning was prescribed in the test method. So,results obtained by tests could mislead and put the same concrete in different exposure classes.
In order to find the improvement for test method, interlaboratory test was organized. Results are shown in this article together with suggestion for improvement of standard NT BUILD 492 and suggestion for defining state of saturation in project requirements together with required value of chloride migration coefficient.
Besides, as permeability properties, especially of concrete with mineral additives, depend on age of concrete, it is necessary to prescribe age in which chloride migration coefficient should have required value.
1.INTRODUCTION
In this study suggestions for defining criterions for exposure classes of concrete will be presented. Before defining criterion values it is necessary to choose test method. Criterion value will depend on it.
2.CHLORIDE MIGRATION COEFFICIENT –TEST METHOD
Since 2005 in IGH laboratory for testing concrete in Split chloride migration coefficient has been tested according to NT BUILD 4928on more then 300 concrete specimens. For part of these specimens reliable data exist about concrete composition because mixes were prepared in laboratory. In some cases tests were performed on specimens which were dried before preconditioning in Ca(OH)2 according to NT BUILD 492. On such specimens it was noticed that results had surprisingly high values.
The conclusionwas that it had been happening because preconditioning according to NT BUILD 492 was insufficient for obtaining such specimen saturation which wouldn't cause significant differences in test results on different specimens made from same concrete but cured in different conditions.
2.1Interlaboratory test results
In 2007 we organized interlaboratory test7 to exclude the possibility of mistake in performing test method. Three laboratories took participation in interlaboratory test: IGH laboratory in Split, IGH laboratory in Zagreb and Laboratory of Faculty of Civil Engineering in Zagreb. In Croatia they are the only laboratories which were performing that method. All three laboratories have similar equipment produced by Germann Instruments, which is one modification of equipment prescribed in NT BUILD 492. One mix of high strength concrete (68,1 MPa after 28 days) was prepared, ten samples were made: six cubes 15x15x15 cm and four cylinders with diameter of 100 mm and height 200 mm. Finally, after drilling and sawing there were twenty four specimens in total (cylinders with diameter of 100 mm and height of 50 mm) for testing. Each laboratory had eight specimens to test. Specimens were distributed in laboratories so that each laboratory had similar combination of specimens and positions. After curing of more then forty days in water, half of total number of specimens in each laboratory were dried to constant mass in oven on temperature of 105±5°C and then cooled in dessicator, so that before start of testing in each laboratory there were four specimens dried to constant mass and four specimens saturated in water to constant mass. After that, each specimen was tested according to test method NT BUILD 492. Preconditioning for surface dried specimens in saturated Ca(OH)2 with vacuum treatment were performed for all specimens, no matter they were dried or water saturated. Preconditioning takes place for twenty to twenty four hours.
Test results confirmed presumptions that preconditioning defined in test method was insufficient and that state of saturation before preconditioninghad significant influence on test results.
2.2Chloride migration coefficient – influence of moisture content
Short analyze of interlaboratory test results is presented on figure 1 and in table 1.Diagram on figure 1 shows chloride migration coefficient test results grouped according to state of saturation before preconditioning, in one group there are specimens which were dried and in the other group there are specimens which were water saturated. In each group there are two subgroups according to specimen origin: specimens drilled from cubes and then sawn and specimens sawn from cylinders.
Figure 1: Results of chloride migration coefficient grouped according to state of saturation before preconditioning7
Origin of specimen (diameter 100 mm and 50 mm) / Chloride migration coefficientAverage value / Standard deviation / Repeata-bility / Reproduce-bility
repeata-bility / reproduce-bility / repeata-bility / reproduce-bility
sr / sR / sr / sR / r / R
10-12 m2/s / 10-12 m2/s / % / %
Specimens water saturated before preconditioning
Specimens taken from cube / 3,08 / 0,40 / 0,46 / 13,0 / 14,9 / 36,8 / 42,1
Specimens taken from cylinder / 2,39 / 0,11 / 0,24 / 4,6 / 10,0 / 13,0 / 28,3
Specimens dried before preconditioning
Specimens taken from cube / 24,19 / 1,06 / 4,32 / 4,4 / 17,9 / 12,4 / 50,6
Specimens taken from cylinder / 23,35 / 4,60 / 8,43 / 19,7 / 36,1 / 55,7 / 102,1
Table 1: Summary of test results
Average value of chloride migration coefficient for all water saturated specimens is 2,73x 10-12 m2/s.
Average value of chloride migration coefficient for all dried specimens is 23,77x 10-12 m2/s.
As in test method there are no data about precision, obtained precision could be an indication. Experimental results show that precision of test method for specimens which were water saturated before preconditioning according to the method is better then that for dried specimens. Test results of dried specimens sometimes show total penetration of chlorides through specimen. In such cases test result is questionable. Real coefficient is probably greater.
Average value of chloride migration coefficientfor samples which were dried before preconditioning according to the method (23,77x10-12 m2/s) is 8,7 times greater then that for samples which were water saturated before preconditioning (2,73x10-12 m2/s). Such significant difference in values of chloride migration coefficient leads to conclusion of insufficiency of prescribed preconditioning in the method. It is especially important in case when criterions which concrete should satisfy are defined.
Specimens which were taken from cylinder show slightly better results then that taken from cube.
Presently, new program is in progress where different moisture contents were taken into account, not only extreme states of saturationas totally water saturated and oven dried concrete specimens. Those tests haven’t yet been completed.
2.3Chloride migration coefficient – influence of age of concrete
It isn’t recommended to test early age concrete because chloride migration coefficient decreases with age of concrete due to filling of cement stone structure. It is especially noticeable in concrete with addition of micro silica. As chloride attack is type of aggression which develops in time, only thin layer of concrete will be attacked in early age of concrete. Early age chloride migration coefficient should be used as indication of coefficient in later ages.
According to experimental results, chloride migration coefficient tested in age of 7 days is two times greater then that tested in age of 30 to 40 days. With further aging decrease of chloride migration coefficient is lower. After three months it decreases for further 10 to 20 % in average.
2.4Results of Round Robin test1
Precision of method NT BUILD 492 obtained in Round Robin Test between 27 laboratories is represented through repeatability of 18 % and reproducibility of 36 %. Results of interlaboratory test presented in this study shows that precision for water saturated specimens taken from cylinder is better then that obtained by Round Robin test.
3.CRITERIONS,EXPOSURE CLASSESAND CONCRETE COMPOSITION
Requirements for concrete exposed to chloride ingress in Croatian regulation are considered only in HRN EN 2062 where concrete classes of exposure for concrete are defined. Recommended limiting values for composition and properties of concrete are defined, such asminimum cement content, maximumwater-cement ratio and minimum strength class. Any technologist knows that it isn’t enough and that concrete which only satisfies such minimum requirements should hardly satisfy requirements for increased resistance to chloride attack. Concrete designed for significant constructions exposed to chloride ingress has to satisfy requirements for durability. So, it is necessary to recommend test method and criterions in order to protect reinforcement from chloride attack and corrosion as its result.
In HRN 11283 which is national addition to HRN EN 206 there isn't any movement forward. In practice,small number of designers in cooperation with technologists prescribes in most cases two coefficients9: 6x10-12 m2/s for XS3 exposure class and 9x10-12 m2/s for XS2 exposure class. Recently, requirement for chloride migration coefficient of 3x10-12 m2/s appears in some projects.
Some designers define only class of exposure without criterions. They can't say anything about predicted lifetime of designed construction.Situation is much better when some permeability parameter is used as criterion.
Designers who don’t know much about technology of concrete aren’t aware of the fact that recommended limiting values in HRN EN 206-1 can hardly enable resistance to chloride ingress. On the other side, requiring toosmall value of chloride migration coefficient causes another kind of problems.Such requirement will result in significant increase of expenses because in that case high performance concrete is designed. Companies which make the offers for execution works usually forget to include all expenses.
As in four years in our laboratory more then 300 concrete specimens were tested according to NT BUILD 492, and for more then 50specimens there are reliable data about composition, we analyzed part of these results regarding composition and classed them. In table 2 recommended limiting values for composition and properties of concrete according to HRN EN 206-1 are presented.
Exposure classes for chloride induced corrosion (sea water) / XS1 / XS2 / XS3Max. water-cement ratio (w/c) / 0,50 / 0,45 / 0,45
Minimum strength class / C30/37 / C35/45 / C35/45
Minimum cement content (kg) / 300 / 320 / 340
Table 2: Recommended limiting values for composition and properties of concrete according to HRN EN 206-1
Cement content (kg) / 365 / 330 / 360 / 410Water-cement ratio / 0,58 / 0,50 / 0,45 / 0,40
Compressive strength (on cube) (MPa) / 40-46 / 60 / 63 / 70
Chloride migration coefficient (x10-12m2/s) / 18-15 / 12 / 12-10 / 10-9
Exposure class according to table 2 / - / XS1 / XS3 / XS3
Suggested class / - / XS1 / XS1 / XS1
Note / Normal concrete / High strength concrete
Table 3: Few examples6 from practise with parameters close tothose recommended in HRN EN 206-1
In table 3 few concrete compositions are classified according to criterions shown in table 2. Changing the type of cement from CEM I 42,5to cement with addition of slag (CEM II/A-S 42,5N) decreases chloride migration coefficient for more then 20 %6.
Further increase of cement content and decrease of water-cement ratio doesn’t result in significant decrease of chloride migration coefficient. In the most number of cases it will be lowered to value from 6 to 9 x10-12m2/s.
Cement content (kg) / 430-480 / 420-440 / 430-480Water-cement ratio / 0,33-0,36 / 0,34-0,36 / 0,33-0,34
Compressive strength (MPa) / 74-88 / 69-83 / 77-81
Difference in constituents / Addition of micro silica / - / Different aggregate
Chloride migration coefficient (x10-12m2/s) / < 3 / 7 - 9 / 9 – 10,5
Exposure class according to table 2 / XS3 / XS3 / XS3
Suggested class / XS3 / XS2 / XS1
Note / High strength concrete
Table 4: Results of experimental program with ten mixes4
It was confirmed in another test program with ten mixes4 which have water-cement ratio from 0,33 to 0,36, cement content from 420 to 480 kg per cubic meter of concrete and compressive strength from 69 to 88 MPa that only three mixes which were prepared with addition of micro-silica have chloride migration coefficient lower then 3x10-12 m2/s. Value of chloride migration coefficient for five mixes ranges from 7 to 9x10-12 m2/s and for two mixes ranges from 9 to 10,5 x10-12 m2/s. All these concrete compositions can be declared as high strength concrete.
Cement content (kg) / 490 / 430Water-cement ratio / 0,33 / 036
Compressive strength (MPa) / 85 / 72
Difference in constituents / - / Addition of micro silica
Chloride migration coefficient (x10-12m2/s) / 8 / < 3
Exposure class according to table 2 / XS3 / XS3
Suggested class / XS2 / XS3
Note / - / -
Table 5: Results of one more experimental program5
These are typical examples which illustrate different classes of concrete concerning resistance to chloride ingress. Classification made in HRN EN 206-1 is insufficient. As type of cement and addition of mineral additions significantly decreases value of chloride migration coefficient it is clear that classes of exposure should be defined through one of chloride transport parameters in concrete, directly, not through concrete composition.
From tables 3, 4 and 5 it can be seen that increasing of cement content and decreasing of water-cement ratio can improve resistance of concrete to chloride ingress only to certain limit. Suggested classes of exposure are given.
4.SUMMARY
In case of defining criterion for chloride migration coefficient in concrete design it is necessary to define: 1.test method, 2.age of concrete specimen in the time of testing, 3.state of saturation before testing.
We suggest using of NT BUILD 492 test method because test equipment is easy to get and test is short and reliable if test specimen is water saturated before start of preconditioning according to test method. Test specimens in age of more then 28 days should be tested.
To make decision about value of chloride migration coefficient as requirement in project it is necessary to know: class of exposure, projected lifetime of construction, kind of reinforcement, thickness of concrete protective layer and concrete surface protection. Considering all these parameters together should lead to define optimal chloride migration coefficient.
Taking all this into account it can be concluded that we are at very beginning of setting up the solution for this problem in regulations. Results of Round Robin tests show that situation isn’t better in other parts of the world.
In this study it is tried to facilitate designers to decide about value ofchloride migration coefficient as criterion for exposure classes of concrete. It was done by analyzing different concrete compositions together with chloride migration coefficient test results and their grouping.
Designers have to be aware of the fact that each level of coefficient of migration requires new type of concrete.
5.CONCLUSIONS
Experimental results obtained by testing different concrete specimens in laboratory and confirmed by interlaboratory tests lead to conclusion that conditions of preconditioning in test method NT BUILT 492 are insufficient. Previous state of saturation has significant influence on test results so that specimens from same concrete mix have significantly different test results, depending on previous state of saturation. In this study it is proposed to test completely saturated specimens, because test results are more reliable, more precise and problems with total penetration of chlorides through specimen don’t exist.
Analyse of experimental results leads to conclusion that compositions of concrete could be grouped in following classes concerning chloride migration coefficient: <3x10-12 m2/s, 3-6 x10-12 m2/s, 6-9x10-12 m2/s, 9-12x10-12 m2/s and >12x10-12 m2/s. First two groups could satisfy exposure class XS3, third group could satisfy XS2 and forth group could satisfy class XS1.
REFERENCES
[1] Castellote, M. & Andrade, C. 2006. Round-Robin Test on methods for determining chloride transport parameters in concrete.Materials and Structures, vol.39, 955-900. RILEM.
[2] HRN EN 206-1. 2006, Beton – 1.dio: Specifikacije, svojstva, proizvodnja I sukladnost (uključuje amandmane A1:2004 I A2:2005), (EN 206-1:2000+A1:2004+A2:2005, Concrete – Part 1: Specification, performance, production and conformity (includes amendments A1:2004 and A2:2005).
[3] HRN 1128, 2007. Beton – Smjernice za primjenu norme HRN EN 206-1 (Concrete – Guidelines for the implementation of HRN EN 206-1).
[4] Institut IGH, Regional Unit in Split, 2005. Ispitivanje tehničkih značajki svježeg i očvrslog betona klase C 40/50 visoke trajnosti za sastav betona kupole i temeljnog prstena sportske dvorane Višnjik u Zadru za Lavčević Inženjering, Split (Testing of properties of fresh and hardened concrete C40/50 of high durability for compositions of concrete for dome and foundation ring of sports hall Višnjik in Zadar for Lavčević Inženjering, Split), Test report 3611-B-635/08, 2005-06-10.
[5] Institut IGH, Regional Unit in Split, 2007. Istraživanje sastava betona razreda najmanje C 25/30, C 40/50 i C 50/60, različitih razreda izloženosti za Mucić&Co. (Exploring the composition of concrete, class at least C 25/30, C 40/50 i C 50/60, of different exposure classes for Mucić&Co.), Test report 3610-0081/07, 2007-03-01.
[6] Institut IGH, Regional Unit in Split, 2008. Ispitivanje primjene raznih vrsta cementa u betonu različitih sastava za Dalmacijacement(Exploring the effect of usage of different kinds of cement (in different compositions of concrete) for Dalmacijacement), Test report 3610-B-0035/08, 2008-01-23.
[7] Institut IGH, Regional Unit in Split, 2008. Međulaboratorijsko ispitivanje koeficijenta migracije klor iona u betonu(Interlaboratory test of chloride migration coefficient in concrete), Test report 3610-0053/08, 2008-09-24.
[8] NT BUILD 492, Nordtest Method 1999. Cncrete, mortar and cement-based repair materials: Chloride migration coefficient from non-steady-state migration experiments. Nordtest.
[9] Port Authority of Dubrovnik2005. PassengerPortDubrovnik, Rehabilitation and Construction of Berth 10-16, Tender Documents, Vol.I: The Tender.
.