alonso 1995

8/3/2019 Alonso 1995

http://slidepdf.com/reader/full/alonso-1995 1/8

P e r g a m o n

ii~!~!:i:iii!i:i!ii~!:!i~i!:

Cement and Concrete Research, Vol. 25, No . 2 . pp. 25%264, 1995 iiiiii!iii!!iiiiiiiiiiiiiiii

Copyright © 1995 Elsevier Science Ltd !iiiiiiiii!iii!i!i!iii!i!i!iPrinted in the US A. All rights reserved !!iiiiiiiiiiiiiiii!ii;!iii!

0008-8846/95 $9.50+.00 i!i!iiiililililili!i!ilili

0 0 0 8 - 8 8 4 6 ( 9 5 ) 0 0 0 0 6 - 2

COMPARI SON OF RATES OF GENERAL CORROSION AND MAXIMUM P ITTING

PENETRA TION ON CONCRETE EMBEDDED STEEL REINFO RCEME NT

J.A. Gonz~lez*, C. Andrade**, C. Alonso** and S. Feliu*

ili!iiiiii!i!iiiiii!i~iiiiiiiii!iiiiiii

!iiii!M!i!iiiii!i!i!iii

;!i!!!i!i!!!ii!!i!i!!~i!i[iiiili!ilili~i~i~i~iii~ii;ilililiiiiiiililil:iiii~i!iliiiiiiiiii~ii;)i!il;!ililr!iiiiiiiiii~ii:::::::::::::::::::::::::::::::::::::::::::::::::

Centro Nacional de I nvestigaciones Metal~rgicas (CENIM), A vda. ii!i~i!iiiiiiiiiiiiii~

Gre gor io del Amot 8, 28040 Madrid, Spain. Fax: 34 1 534 7425

** Instit uto de Ciencias de la Const rucci dn "Eduardo Torroja"

(ICCET), Se rran o Gal vac he s/n, 28033 Madri d, Spain.

Fax: 34 1 302 0700

(Comm~icatedbyC.D. Pomcroy)

( R e c e iv e d A u g m t ~ , l~4 )

ABSTRACT

Local attack of reinforcement is observe d usually in

chlorid e-conta minated concrete. When mechanic al weak ening

of the concrete structure is affec ted by the bar section

loss in the places where corr osion is intensified , the

maxim um penetr ation of the deepest pits happens to be a

very rele vant datum. In this work average corrosion values

are compared with maximum pit depth values. Using natural

corros ion tests and acce lerat ed tests, it has been foun dthat the maximum pen etration of localized attack on steel

embedded in concrete containing chlorides is equivalent to

about four to eight times the average general penetration.

Key words: Chloride-c ontaminat ed concrete, reinforc ement

deterioration, ma ximum penetratio n of pits, average general

corrosion, damage prediction.

INTRODUCTION

It is well known the harmful role that corr osion of reinf orce ment

plays in the service life of reinfor ced concrete structu res (1,2),

This calls for routine assessment of any corrosion process

developed on them, preferab ly by using non-destruc tive techniques.

Nowadays, the Continuous or periodic monito ring of corrosio n is

main ly based on electrochemica l measurements. These measure ments

are ref erre d both to corr osion p otenti al (Eeoc) and Corro sion

curr ent (Icon). Eeorr s rel ate d to the lik elih ood that rei nfo rce men ts

suffer acti ve corro sion (3), and theref ore fails to prov ide an

accurate measureme nt of corrosion rate. This problem is solved with

Ieo mea sure men ts (4,5), a lthough other problems arise in this case,

such as those related to the correct interp retation of the

measurements.

The values of Ico= expe rimen tally deter mined in reinfo rcemen ts, even

the highe st ones, do not seem to imply at first sight too much

257

8/3/2019 Alonso 1995


258 J.A. Gonz~ilez et al. Vol. 25. No. 2

danger. Valu es of 1-3 ~A/cm 2 are frequent in active corr osio n and

only seldom values of the order of i0 ~A/c m 2 or h igher are

det erm ine d (6). Values of 1-3 ~ A / c m 2 are equivalent to corrosion

rates of 11-33 ~m per year, so that it is diffi cult to und ers tan d

how from such a low figures a structure becomes unus able in a

rel ativ ely short period of time, as so happens in some instances.

However, such values are only appar ently small. Extra pol ati ng from

an assum ed constant rate of 20 ~m/yea r up to 50 years, an averag e

atta ck pen etra tio n of 1 mm will be obtained, which me ans a loss of

material equivalent to the 21 per cent of the cross-s ectiona l area

of a rebar of 9 mm diameter, which may ex ceed the sa fety

coefficie nt. The prese nce of localiz ed corrosion, with pit depths

several times grea ter than the average corr osion attack, ma kes the

situation worse, especial ly in prestressed structures. Similar

consequences concernlng structural safety and servic eabil ity come

from the point of view of the concrete cover cracking. So, it has

been estab lish ed (7) that the amount of iron oxide gene rat ed duri ngthe corrosio n of 10-50 ~m of metal (depending on rebar dia meter)

is enough to crack a concrete cover of 2-3 cm. This means that many

concre te covers will be cracked after 1-3 years as a result of

corr osion rates of only 10-30 ~m/year.

The values of Ico= are average values refer red to the overal l

reinfor cement surface. The integration of these values over the

exposure time allows estimate of an average pen etrat ion of

corr osion (Pay), but it does not inf orm at all about the max imu m

pe ne tra ti on (Pmax) of the deepes t pits when the at tac k is ver y

uneven. For life predictio ns, this Pmax value may have more

importance than the average value when the mechanic al weake ning ofthe structu re is affe cted by the bar section loss in those places

wher e c orro sion is intensified. P~x could be esti mat ed from Ico

pr ovi de d that a ratio (R) bet wee n Pmax and Pay coul d be esta bli she d.

Despite the practical interest of the matter, little syste matic

work has so far been done to the writers' knowl edge on pit tin g of

reinf orcem ents. A study by Tuutti (8), who det erm ine d values of

R = 4-10, make s an exception, althoug h he did not include the

effect of chloride additions to cement and used an external voltage

source as a corr osio n acce lerator. Some more in form atio n on pit

depths is found in a number of papers on atmosp her ic and soil

corr osion of bare metals. From these data, values of R = 3-5 for

mil d and low-a lloy steel (9-13) and R = 8-10 for alum iniu m allo ys

(14), expos ed to differ ent atmospheres, as well as values of R

3.5-15 for burie d uncoated steel and cast-i ron pip e samples

(15,16), have been estimated.

Determ inati on of R in natural reinforced concrete corrosion entails

impor tant expe rime ntal difficulties, such as to have (a) to brea k

up the concrete sample in order to examine pitt ing on the

re in for ce me nt surface, (b) to wait for a long time (many years) to

get, th roug h a natural corros ion process, a suffi cient attac k to

facil itate the survey, and (c) to handle concr ete samples of large

dime nsio ns (big enough for the exami nati on of an exten sive

reinfo rcemen t surface) in order to arrive at statist ically soundconclusions regarding the distribution and size of pits.

8/3/2019 Alonso 1995


iiiiiii:i:!ililiVol. 25, No. 2 CORROSION RATE, STEEL, P1TFING, CHLORIDES, PENETRATION 259

: : : : : : : : :

::: ::::

~i~iiiiiiijiiiii

:::::::::::

These difficu ltie s have, no doubt, contr ibute d to delays in test ing

this partic ular aspect.::::::::

i:::: ::i:;:: ;..;

Both natural and accelerated corrosion testing will be used in this

investigation, the latter to verify the feasibility of using a

small impr esse d current (of the order of Ico=) to acceler ate pit tin gin laboratory testing of chloride containing reinfor ced concrete

samples.

:i:i:i:i:i:i~i~i!i

E x P E R I E N TA

For the natural testing, two types of concrete samples wer e made

of 50 cm x 50 cm x i0 cm and 20 cm x 15 cm x i0 cm dime nsi on s,

respectively, having embedded 0.8 cm diameter steel re inforcement. .............

The concrete used in the first type of samples cont aine d 3% of

CaCI 2 per wei ght of cement, in order to provok e an active attac k on

reinforcements. This sample was subjected to repeated periods of

wetting-drying , by keeping a water moistened pad on its surface,which were followed by drying periods in the laboratory atmospher e

(50-60% RH approx.). The second type of concrete samples cont ained

no admixtures, and were sub merged into a natural sea wate r tank

located at the Instit uto de Ciencias del Mar (CSIC) of Barcelona .

The mean salini ty of water was 28 g/l, te mper atur e vari ed from 14-°C

in wint er to 24-°C in summer, and oxygen mean conten t was 5 ppm

(17).

Afte r 6 years testing, both types of samples wer e broken in order

to assess the corrosi on suffered by the steel reinforceme nt, i.e.

weight loss, attack morphol ogy and pit depth distribution. Weight

loss was obtain ed by comparing the weig ht per unit length of

attacked reinforcem ent (once all adhered concrete and corrosion

products were removed), with that of the unattack ed reinforcement.

The depth of pitting was measured in different ways, accordin g to

convenience. Where pits were big enough a microm eter was directly

used. In other cases, the depth was deter min ed from the

displacem ent of an optical microscope objective focussing the

extern al surface and the botto m of the pit successivel y.

For the acc eler ated testing, two types of samples were also

prep ared : (a) mor tar samples 2 cm x 5.5 cm x 8 cm, wit h mild steel

bars e mbed ded of 0.6 cm in diameter and a mort ar cover of 0.7 cm,

and (b) co nc re te sam ples 15 cm x 15 cm x 40 cm, wi th bar s of 1.6

cm in dia me ter and concr ete covers of 2 to 5 cm. Bot h series w erepre par ed with 2 and 3% by weight of admixed CaCl2 to prov oke active

corro sion from the begi nnin g of test. Anodi c current s rang ing from

i0 to I00 ~ A/cm 2 were applied to the rei nforc ement. A central

grap hite bar in samples "a" and an external me tal lic plate pl aced

on the concrete surface (through a wet pad) in samples "b" were

used as counter electr odes. At the end of the test period, when

cracking of mortar or concrete cover occur red due to the

acc umul atio n of corros ion products (oxide), the samples were broken

and pits examin ed similar ly to natural testing.

RESULTS

Natural testingThe examina tion of the reinf orcem ent that had been part of the

samp le of dim ens ion s 50 cm x 50 cm x i0 cm expos ed to the

8/3/2019 Alonso 1995


260 J.A. Gonz~ lez et al. Vol . 25, No. 2

laboratory atmosphere disclosed a great number of pits, distri buted

all over their surface, wh ich is normal with highly c hlori de-

contaminate d concretes. The study of this specimen was mainl y based

on two reinforcements: one reinfo rcemen t "A" located in front of

a crack in the concrete cover (the crack was det ecte d on the third

year of testing), and other reinfor cement "B" that was all the timeunder an undam aged concrete cover. In the case of rein forc emen t

"A", the weig ht loss after the 6 years was equi valent to an averag e

pene trat ion of the attack on all the surface of 0.265 mm (Table i).

Table i. Values of R obtain ed in natural conditions, wh en the

concrete samples containin g 3% CaCl 2 by weig ht of cement

were submitted to wetting-dr ying cycles in the atmosphere.

Conditions Average penetration Maximum pit R value

of attack after depth after

6 yea rs (mm) 6 ye ar s (mm)

Rebar in front of

a crac k in the

concrete cover 0.265 1.20 4.4

Uncracked

cover 0.085 0.50 5.9

The deepes t pits showed penetratio ns of the order of 1.20 mm. In

this bar, pitti ng was more abund ant and deeper along the

reinforceme nt generatrix coincident with the position of the crackon the concrete. If the average value (0.265 mm) is compare d with

the d ept h of the deep est pits, a ratio R betwe en P~x and Pay of 4.4

is obtained. In the case of reinfo rcement "B", the weig ht loss

after the 6 years was equi valent to an average p enet rati on of 0.085

mm, which is lower than for reinfor cement "A", most likely due to

the higher prote ctin g ca pacity of the undam aged concrete cover. The

dee pes t pit s s howe d depths in the order of 0.50 mm. In this case,

R= 5.9.

During the six years of testing, Ieo meas urem ents were highl y

influenced by wetting and drying conditions of concrete, ranging

from about 0.I ~A/cm 2 for dry concrete up to some 7 ~A/cm 2 for very

wet concrete.

Table 2 summarizes the charac teris tics of the attack in samples

that were c ontin uousl y submerg ed in sea water. The averag e value

of R for all the reinforce ments examined gave a figure of 4.5 whic h

is of the order of those prev ious ly found. Scatter of data seems

to be normal in pit tin g attack. In these samples the value of I~o~

incre ased gr adual ly with time as the chlorides p enetr ate into the

concrete, reachi ng values of 10-20 ~ A/cm 2 after 5-6 years. The

total corros ion estimates from curves Ic~ vs. time gave values in

reasonable agreement with the measured weight losses.

Accelerated testing

Tables 3 and 4 show the values of R obtai ned when an anodic curren t

is applied. They range from 3.2 to 16.1 with an average v alue of

8/3/2019 Alonso 1995


ii:iii:i~iii~iii~i

Vol. 25, No. 2 CORROSION RATE, STEEL, PrITIN G, CHLORIDES, PENETRATION 261

: : : : : : : : : :

Table 2. value s of R obtai ned in natural condit ions when the

concr ete samples were immers ed in natural sea water.

:::::::::::

C o n c r e t e R e i n f o r - A v e r a g e p e n e t r a - M a x i m u m p i t R v a l u e

t y p e c e m e n t tion of attack depth after

after 6 years 6 year s

(ram) (mm)

iiiiilZ!i!iii!iiiiii

~iii!:i!~!i?i!~

OPC -40 0 kg/m 3 IN 0.62 5.50 8.9

2N 0.41 1.51 3.7

3N 0.41 1;51 3.7

w/c= 0.3 8 5N 0.43 2.15 5.0

OPC-4 00 kg/m 3 17N 0.53 2.2 0 4.2

19N 0.42 1.15 2.7

w/c=0 .6 20N 0.47 2.50 5.3

OPC -40 0 kg/m 3 28N 0.38 1.72 4.6w / c = 0 . 5

0PC -30 0 kg/m 3 32N 0.54 2.19 4.1

w/ c=0 ;6 33N 0.40 I.I0 2.8

iiii~iiiiii

i~ii!iiiiii!iiii

Table 3. Value s of R obtai ned from acce lera ted tests in mor tar

s a m p l e s c o n t a i n i n g 2 % C a C l 2 b y w e i g h t o f ce m e nt . A p p l i c a t i o n o f a

cur ren t of i0 ~A /cm 2.

C o n d i t i o n s T e s t d u r a t i o n R m e a n v a l u eof 6 reba rs

s a m p l e s i m m e r s e d

in wat er 1-3 months 3.2

S a m p l e s e x p o s e d

to RH > 90% 1-3 mo nt hs 3.7

Table 4. Valu es of R obta ined from acce lera ted tests in concr ete

samp les co nt ain in g 3% of CaCl 2. Test du ration , 1 month .

C o n c r e t e C u r r e n t N u m b e r A v e r a g e M a x i m u m R v a l u e

cover (~A/cm 2 of bars pen etra tion pit

(cm) tested of attack depth

(mm) (mm)

2 i0 2 0.0 95 1.20 12.6

i00 3 0.276 1.68 6.1

3 100 5 0.314 1.86 5.9

5 I0 1 0.090 1.50 16.1

I00 3 0.263 2.15 8.2

8/3/2019 Alonso 1995


262 J.A. Gonz~lez et al. Vol. 25, No. 2

R = 8.2. In spite of the applied current, surface exam inat ion

reve aled a very uneven corrosion, with areas not corr oded and

others intensely attacked, reproducing in some way the appearance

of natural corros ion of reinfo rcemen t in concrete.

DISCUSION

The fre quen t obt ai nin g of R = R ax/Pay valu es ran gin g from abo ut 4 to

8 in natural corrosion, and from 5 to 13 in acce lera ted te stin g of

rein forc ed concrete, is a fact of a great practical impor tance for

predict ion of residual life of concrete structures from

ele ctro chem ical Ico measurem ents. Accordingly, it can be expecte d

that pit growt h proceeds 4-8 times or so more rapid ly than the

inst antan eous Ico. (average rate) measured. Similarly, the maxim um

total de pth of localize d attack will exce ed by about 4-8 times the

estimated average general penetration deduced from loss in weight

after a given exposure time.

Certainly, the above conclusions are provisional whil e a meticul ous

statis tical study on a large number of samples is not accomplished.

It is interesting to note that the furnished data appear to agree

among thems elves and, as mentioned, with other data repo rted for

bare metals in the atmospher e or in the soil. The resemb lanc e am ong

R values in different circumstances recalls other similarities

between the mechanisms of concrete corrosion and atmosphe ric

corrosion alr eady emphasized (18).

A rigorous approach to pitting corrosion should involve the

comb ined use of a great amount of field measur ements , statis tical

analysis and mechan istica lly based modelling. Extreme value

statistics may provide an approximate method of predictin g the

max imu m pit depth in a given area (19). Analysis of man y sam ples

is required, wh ere the number of pits and the depths of all of them

have to be measured. In the statistical analysis, confl ict with the

physi cal limits on pit growth kinetics can exist and signi fican t

progr ess is still needed in the devel opment of realis tic models

(20). In a general treatment, progress of pits with ti me s hould

also be considered; espec ially the effect of time on the number,

distrib ution, total area and volume of all pits.

The va lue of R depends on the pen etr ati on rate (Vmax) for the

deepe st pits and the average corrosion rate (Vav) for the wholeexposed surface. The complex nature of the pitting process makes

it difficult to predict the effect of the different variabl es on

R. It is for ese eab le that Vma be mor e or less influenced , acc ord ing

to exper iment al circumstances, by (i) diffu sion of corro sion

produ cts and species conce rned in the pit prop agat ion process, (ii)

acti vati on energi es of the anodic and cathodic processes, and (iii)

electrical resistance between the anodic and cathodic areas. On the

oth er hand, Vav is a compl ica ted func tion of (i) pre sen ce of

hete roge neit ies th at act as nuclei for pit initiation, (ii)

alteration of potential distr ibution around a establis hed pit in

a sense unfavour able to the starting of new pits (anodic

dissolut ion of metal within a pit renders the surroun ding regioninsu ffi cien tly anodic for the begin ning of fresh pitting), (iii)

resu ltin g total number of pits on the metal surface and

8/3/2019 Alonso 1995


Vol. 25, No. 2 CORROSION RATE, STEEL,PITHNG,CHLORIDES,PENETRATION 263

dist ribu tion of pit shapes and sizes, (iv) pit growth kinetics, and

(v ) ra te of at ta ck of th e re ma in in g me ta l su rf ac e un af fe ct ed by th e ::::::::::::p it s.

:::: :::::

ii:i:i:!iiiA t h e o r e t i c a l j u s t i f i c a t i o n o f t h e o b se r v e d v a l u e s o f R sh o u l d

include a study in depth of all the above points, and for this

pu rp os e e ss en ti al in fo rm at io n is lacking . Fo rt un at el y, th e .........

empiri cal fact that R values of 4-8 are often obtained, in spite

of the numb er o f va r i a b les tha t in theor y wi l l a f fec t R , may be ://::

used as a basis for appr oxi mat e pred ict ion s of resi dual servi ce ~;;<

l i f e .iiiiiii[ii~i~: ::::::::: :

CONCLUS IONS

i. It has been found that the maxim um pen etra tion of the l ocaliz ed

attack on steel embedded in concrete containing chlorides is

equiv alent to about four to eight times the average p enet rat ion of

the attac k on the overall rei nforcem ent surface. This relation doesnot diff er too much from similar relations repo rted for the

corro sion of metals in the atmosphere or in the soil.

2. The study of pitt ing corrosion on reinf orce d concre te can be

accelerated by using an impressed anodic current. Values of R

rather similar, althou gh sli ghtly higher, to those found under

natural conditions have been obtained with the appl icat ion of i0-

i00 ~A/c m 2 to reinf orceme nts in concrete conta ining chlorides.

3. The knowledge of R may be useful for estimation of the residual

life of concrete structures from on-site Ic~ measur ements .

ACKNOWLEDGMENTS

This work was suported by the Comisidn Interministerial de Ciencia

y Tecno logia (CICYT) of the Mini stry of Educ ation and Science of

Spain. The authors are indebted to M. Teresa Gutie rrez of "E.

Torroja" Institute for careful and patien t meas ure ment of pits in

rein forc ed concrete samples exposed to sea water.

R E F E R E N C E S

i. B. Boyard, C. Warren, S. Somaya ji and R. Heid ersb ach, Cor ros ionRates of Steel in Concrete, AST M STP 1065, N.S. Berke, V.

Cha ker and D. Whiting , Eds., Philadel phia, P.A. (1990), p. 174.

2. Tranporta tion Research Board, Strategic Highway Research

Prog ram Plans. Final Report. NCHRP Project 20-20. Washingto n,

D.C. (1986).

3. Standa rd Test Meth od for Half Cell Potentia l of Rein for cing

Steel in Concrete, AST M C 876-87, Phila delph ia, P.A.

4. C. An dr ad e and J.A. Gonz ~lez , Werk s. u. Korr., 29, 515 (1978).

5. J.A. Go nz~ le z, S. Al ga ba an d C. An dr ad e, Br. Corro s. J., 2__55,

135 (1980).

8/3/2019 Alonso 1995


264 J.A. Gonz~lez et al. Vol. 25, No. 2

6. C. Andrade, M.C. Alonso and J.A. Gonz~lez, Corr osio n Rates of

Steel in Concrete, A STM STP 1065, N.S. Berke, V. Chak er and D.

Whiti ng, Eds., Philadel phia, P.A. (1990), p. 29.

7. C. Andr ade, C. Alo nso and J. Molina, Ma ter ial s and Struc tures ,

26 453 (1994).

8. K. Tuutti, C orr osi on of Steel in Concrete. Swed ish Ceme nt and

Concret e Resea rch Institute, Stock holm (1982).

9. S. Feliu and M. Morcillo, Corrosion and Prot ecti on of Metals

in the Atmo sphe re (in Spanish), Ediciones B ella ter ra S.A.,Barcelona (1982).

I0. C.R. Southwell, J.D. Bultman and A.L. Alexander, Mat eria lsPerformance, 15, No. 7, 9 (1976).

Ii. H.R. Copson, Proceed. ASTM, 52, 1005 (1952).

12. T. Mo ro is hi a nd J. Satake, J. Iron Steel Inst. (Japan), 59, No.2, 293 (1973).

13. P. Merino, Me tall ogra phic Study of Diffe rent Types of

Corrosion: Depth of Pits (in Spanish), Docto ral Thesis,

Uni vers ida d de Santiago de Compostela, Spain, 1987.

14. E. Otero, R. Liza rbe and S. Feliu, Rev. de Me ta lu rg ia (Madrid),!, 359(1971).

15. M. Romanoff, Under grou nd Corrosion. Nati onal Burea u of

Standards, Circula r 579. US. Governmen t Print ing Office,Washington (1957).

16. J.C. H udson and K.O. Watkins, BISRA Open Report No. MG/B/3/68.

17. C. Andr ade, C. Alonso, S. Go~i and J.A. Gonz~ lez, Cor ros ion of

Rein forc eme nts in Concrete Immersed in Sea Water, 7th

International Congress on Marine Corrosion and Fouling.

Uni vers ida d Polit6cnica, Valencia, Spain (1988).

18. C. Andr ade, C. Alonso, J.A. Gonz~ lez and S. Feliu, Cor ros ion

of Rei nfo rce men ts in Concrete, Ed. C.L. Page, K.W.J. Treada way,

P.O. Bamforth. Society of Chemical Industry, Birmingham, U.K.(1990), p 39.

19. P.M. Aziz, Cor rosi on, 12, 35 (1956).

20. A. Tur nbul l, Br. Corros. J., 28, 297 (1993).

alonso 1995

Documents