当前位置:首页 >> 电子/电路 >>

Studies on corrosion inhibition of steel reinforcement by phosphate and nitrite


Materials' and Structures / Matrriaux et Constructions, Vol. 36, October 2003, pp 530-540

Studies on corrosion inhibition of steel reinforcement by phosphate and nitrite
L. Dhouibi 1, E. Triki 1, M. Salta 2, P. Rodrigues 2 and A. Raharinaivo 3
(1) U.R. Corrosion et Protection des Mrtalliques, ENIT, B.P. 37 le Belvrdrre 1002, Tunis, Tunisie (2) Laboratorio National de Engenhafia Civil, Departemento de Materials de Construcc2ao,Av. do Brasil, 101 PT- 1700-066, Lisboa, Portugal (3) Laboratoire Central des Ponts et Chaussres, 58 Bd Lefebvre, 75732 Paris cedex 15, France

ABSTRACT
This work deals with the effectiveness of sodium phosphates and nitrites used as inhibitors against steel corrosion in concrete reinforcement. First, concrete pore water was simulated with several alkaline solutions. Then, reinforced concrete specimens with inhibitors added in fresh concrete mix are immersed in chloride solution. The effectiveness of inhibitors was assessed by applying electroclaemical techniques, namely Eleclrochemical hnpedance Spectroscopy (E[S), Polafisation Curves (PC), and Open Circuit Potential (OCP) measurements. The final concrete condition was analysed with Scanning Electronic Microscopy (SEM), X- Ray Diffraction (XRD) and infrared Spectroscopy (FT-IR). In solutions simulating concrete contaminated with chloride, the influence of the it~aibitors on the steel corrosion was assessed by (PC) and (EIS). The results obtained show that phosphate prevents pitting corrosion when its content is equal to chloride concentration, and that nitrite only contributes to inc~ease the value of pitting potential. Corrosion rate is reduced with both inhibitors at the different contents tested. For reinforced concrete specimens immersed in chloride solution, theft conditions were assessed by applying EIS. The results indicate that after 1 year of immersion with the two inhibitors the corrosion rate decreased. Then, aIter 3 years of immersion no influence ofinhibitors on the corrosion rate was observed. However a significant increase in concrete electrical resistance was obsewed when inhibitors were present. Visual examinations showed that all steel bars were corroded in the presence of chlorides. Results from analytical tests done on the concrete after 3 years of irranersion show that as far as the final concrete condition is concerned, the test~ inhibitors did not change the type of compounds in concrete. But the final fi'ee Chloridecontent remained higher than the critical chloride tN~shold. The results of FT-IR technique showed that nitrites are likely washed out of concrete during immersion in chloride solution and phosphates interfere with the equilibrium between CO32 and HCO3"in concrete. The main conclusion of this study is: the efficiency of the tested inhibitors decreases with time, after two years of immersion in chloride solution.

I~SUMI~
Le prOsent travail eoncerne l'~valuation de l'efficacitO du nitrite de sodium et du phosphate sodique contre la corrosion des armatures" du bOton. L ~tude exp~rimentale a ~tk men~e en milieu simulcmt la solution interstitielle c& b~ton et dam' le b~ton durci additionn~ de l'inhibiteur lots de son malaxage. Plusieurs techniques" ont Otk appliquOes, Olectrochimiques telles clue : la Spectroscopie dTmpkdcmce FJectrochimique (SIE), les Courbes de Polarisation (CP), et la mesure du potentiel de corrosion (PC'), et analytiques, notamment la Microscopie Electronique g~ Balayage (MEB), la Di.ffi'aetion des rayons" X (DRX) et la Spectroscopie Infi~arouge (IR). En milieu simulant le b~ton contamin~ par les chlorures, l'influence de l'inhibiteur sur la corrosion a ~t~ dvalu~e par (CP) et (SIE), Les r~sultats ~ obtenus ont montr~ que les pho~phat~ empkchent la corrosion par pi@re torsque leur teneur est ~gate g7 la concentration en chlorure et clue les nitrites anoblissent uniquement le potentiel de piq~ration. La vitesse de eoJ~osion est rdduite avec les dewc inhibiteurs. Pour le b~ton durci additionnd de t'inhibiteur et immerge en milieu chlorurk, la rOsistance dt la corrosion a Ot8 dvatude par SIE. Les rg~'ultats ont indiqud que pour des courtes durde d'immersion (une annde), la vitesse de corrosion diminue. A k)ng terme (3 arts~ d'immersion), l'inhibiteur augmente la rdsis'tance ~lectrique du bdton mais son influeme swr la ~rrosion s 'estompe. L bbservation vis'uelle des"armatures montrent que celles-ci sont toutes corrodOes. Les r~sultats relate' d t'analyse du bdton de l 'interface aprds' 3 ans d'immersion, ont rgvdl~ clue les inhibiteurs test& ne changent pas la nature chimique des p r o f i t s d'hydratation ~t ciment durci. Cependant; ta teneur des chlorures libres demeure supdrieure gt [a concentration critique. Les rdsuttats de la spectroscopie inJ?arouge stipulent que les nitrites sont su'sceptibles. de sortir du b~ton durant son immersion en milieu chlorur~ et clue les phosphates s 'inter~rent avec l'~quilibre entre C032 et HCO f darts le b~ton. La conclusion principale de ce travail est qu'en presence des ehlorures, l'efficaeit6 de l'inhibiteur diminue h long terme (aprbs 2 aTzs'd'immersion).

Editorial Note Dr. Manuella Salta is a RILEM Senior Member. She worLs' at LNEC, a RILEM Titular Member. Dr, A. Raharinaivo work~ at LCPC, a RILEM Titular Member. Both participate in RILEM TC 178-TMC 'Testing and modelling chloride penetration in concrete'.

1359-5997/03 9 RILEM

530

Materials and Structures / Matdriaux et Constructions, Vol. 36, October 2003

1. I N T R O D U C T I O N The application of corrosion inhibitor is a technique for preventing steel corrosion in concrete exposed to chloride environments. Several papers have been published on the effectiveness of inhibitors mixed in fresh concrete [1-6]. This paper is a complement to previous works [7-9], which have shown the inhibiting effect of sodium phosphate and sodium nitrite tested in simulated concrete solutions. The efficiency of these two inhibitors was assessed by using steel specimens in solution simulating concrete pore water and in concrete. The electrochemical tests applied fbr steel in solution were Electrochemical Impedance Spectroscopy (EIS), Polarisation Curves (PC), and Open Circuit Potential measurements (OCP). For reinforced concrete specimens with mixed-in inhibitors, EIS and OCP tests were used: the steel condition was monitored after 1, 2 and 3 years of immersion in chloride solution and in distilled water. After 3 years of exposure, concrete near steel was analysed by applying X- Ray Diffraction (XRD), infrared Spectroscopy (FT-IR) and chloride content determination. The effect of the inhibitors on concrete microstructure was assessed by Scanning Electronic Microscopy (SEM) and Energy Dispersion Spectroscopy (EDS). 2. E X P E R I M E N T A L PROCEDURES

- a transfer function analyser (Solartron FILa. SI 1250) and a filter (Kemo VBF8), a PC computer with FRACOM software for plotting impedance by controlling the analyser of transfer function and a SIMPLEX software fbr modelling experimental data by an "Equivalent Electric Circuit". The best fitting has been obtained by introducing a symmetrical Cole-Cole distribution in the time constants. In order to check the validity of the calculated results the experimental and calculated Nyquist diagrams were plotted in the same graph. In all cases, the relative experimental error was less than 5 % of the impedance value. 2.2 T e s t s o n r e i n f o r c e d c o n c r e t e The specimens were prisms made of reinforced concrete, 40 mm x 40 m m x 160 mm in dimensions. Reinforcement was a rod 6.5 mm in diameter and 120 mm long, made of plain carbon steel (C = 0.22%). The steel surface was brushed before concrete casting. A copper wire was welded in the middle of the bar for electrical connection. This weld was protected with epoxy resin. Concrete specimens were made with ordinary Portland cement with a high resistance to sulphates referred to as CP HRS (calcium aluminate content C3A -- 2.3 %). The water cement ratio (w/c) was equal to 0.4, and Table 1 shows the concrete proportions. The inhibitors were added to the concrete mixing water. Their weight contents were 0% (reference specimens), and 6% against cement weight. The concrete specimens were stored in the moulds for 48 hours in a chamber at 60 % relative humidity. Then, they were taken out off the moulds and immersed in water at 20~ for 28 days. Table I - Concrete mix, w/e =0.4 Component Maximum diameter (mm) Content kg. m -3 400 635 Portland cement Sand Gravel 8 1290

2.1 T e s t s in c h l o r i d e s o l u t i o n A chloride-contaminated concrete was simulated with an alkaline solution of saturated calcium hydroxide with 0.5 M sodium chloride. Different concentrations of sodium phosphate and sodium nitrite were added to this solution. The inhibitor concentrations Ci~ (Ci,1 [PO43-] or Cin = [NO2]) were determined by using the ratio R = Cin /Ccv, where Ccl = [CI] is the chloride concentration. Three values of R were studied: 0, 0.4, and 1. The pH values of the different solutions were about 11.6 ? 0.2. Polarisation curves and OCP measurements were made by using a three-electrode cell containing 250 ml of solution. The working electrode was a polished piece of carbon steel (33 mm 2 area); the reference electrode was a saturated calomel electrode, and the counter-electrode was a platinum wire. The solution was slightly agitated and maintained at a temperature of 25 ? I~ during tests. The potentiostat-galvanostat used was a Radiometer-Tacussel| PGP201 potenstiostat-galvanostat monitored by a Voltarnaster- 1| software. Before drawing polarisation curves, the open circuit potential was measured and plotted against time during one hour. Polarisation curves scanning started from the open circuit potential up to an anodic current of 200 gA cm -2, at a rate of 25 mV rain-1, followed by a similar scan in the reverse direction down to -1300 mV/scE. Impedance spectroscopy was performed at the corrosion potential (OCP), after one hour o f immersion in the same cell. Steel was polarised at + 10 mV around OCP, by an alternate current with a frequency ranging between 10 mHz and 65 kHz, 5 fi'equency values were chosen per decade. The equipment used included: - a potentiostat (Solartron SI 1286),
=

After curing, specimens were immersed for three years in a tight vessel containing either distilled water (as a reference condition) or sodium chloride solution (0.5 M). The monitoring tests (EIS and OCP) were carried out once a year during this period. For EIS, the counterelectrode was a stainless steel (18 % Ni, 10 % Cr) plate and the reference electrode was a saturated calomel electrode (SCE). During testing, specimens were removed from the

Fig. 1 - Cell used for electrochemical tests of reinforcement concrete.

531

Dhouibi, Triki, SaRa, Rodrigues, Raharinaivo solution, the counter-electrode and reference electrode were placed on concrete surface (Fig. 1) and the Nyquist diagram has been obtained at open circuit potential. The prismatic reinforced concrete specimens were split by applying a compressive load, then the reinforcement surface and the corrosion products were observed. At the steel /concrete interface in specimens, powder concrete samples were taken for analyses "free" (water soluble) and "total" chloride content dosages, XRD and FTIR analyses. Free chlorides were extracted from concrete with distilled water and total chlorides with HNO3 solution. The chloride content was determined with a specific ion probe and expressed as a percentage of cement weight. Powder samples recovered at the interface steel/concrete were placed into about 60 ml of boiling distilled water, agitated for five minutes and then filtrated. The filtrate was dried by evaporating water in a double boiler and the residue has been placed in a steam room at 100 ~ for ten minutes. A very small amount of this dry residue was embedded in dry KBr. The presence of the inhibitor in concrete was detected by lnli~red Spectroscopy by using a Nicolet Magna 300 FTIR spectrometer from 4000 to 500 cm"~. In order to detect crystalline products, these powder samples were also analysed using a X-ray powder dif~actometer, Joel JDX 8030 with Cu K ~ i filtered) radiation (y = 1.5418 A ~ under 40 kV, 20 mA, a scanning rate of 0.05 -20 ~ per step and a measuring time of 1 sec per step. In order to detect any change of concrete microstructure, some samples were taken from concrete interface with steel. Then, they were observed with Scanning Electronic Microscopy and analysed with Energy Dispersion Spectroscopy (EDS). Table 2 gives the experimental conditions and the type of tests performed for each specimen.
Table 2 2 Experimental conditions and types o f tests p e r f o r m e d for concrete specimens Set I Experimental conditions Tests performed Analytical tests 1 30-day-old specimens 2 3-year immersion in Electrochemical and analytical tests distilled water 3 3-year immersion in 0.5 Electrochemical and analytical tests M NaC1 solution Table 3 - Specific potentials: pitting Ep and repassivatio n Erep. (? 20 m V seE) Inhibitor Ep (mV scE) Erep (mY SCE) Ratio R 01 -237 :595 Nitrite 680 -400 0.4 ,,,Phosphate 580 -570 ,, Nitrite 1 360 -380
-100 i i

+

7 *" Vho~p~i
"' " III ,,

-300 i~

.4o0

-500 4 " ,

.600 :
.700-I................................ ~ ............................ ~ ....................... + ......................... ~ ....................... 0 0.2 0,4 0.6 0.8 1
R = l Z ~ l / lCi]

Fig. 2 - Relationship between zero-current potential Ec and R = [inhib] / [CI] for nitrite and phosphate additions. 10 i
i

2-I -2 -6 i -10 I* ' -1300
10

-800

-300 200 Ee (mVscE)
$

700

~--R=

3. RESULTS AND DISCUSSION 3.1 Tests on solution
Fig. 2 shows the evolution of the OCP of steel in simulated concrete pore solution versus time. Nitrite shows an anodic inhibition behaviour. At low values of ratio R, phosphate acts as a cathodic inhibitor, whereas at higher Rvalues, it becomes a mixed inhibitor. Figs. 3a and 3b display the polarisation curves. As pitting co~osion is important in chloride environment, the pitting potential Ep and the repassivation potential E,.~p, are of interest. It is to be noticed that Ep is the potential where a first pit appears, when no initial pit exists before the potential scanning starts [10]; E~p is the potential below which a pit can grow and it is considered as the potential

6~

If

R = 0.4

2~

-2

-1300

-900

-500

-100

300

700

Ec (mY seE)

Fig. 3 - The effect of inhibitors addition on the polarisation curves: a) Nitrite addition, b) Phosphate addition.

532

Materials and Structures I Mat6riaux et Constructions, Vol. 36, October 2003

30000
6000

0.01 Hz

--I
A

~E

0,01

"m m

~ 4000

20000
Im *W

10000

0
2000 4000, Re (Z) f2~cm~ 6009 SOf~

0

10000

20000
2

30000

Re (Z) fl.cm

I~R=0 *R=0,4 mR=ll 30000 "~
t

20000 ~ d
I0000

~ J ~ - - - - O , 0 1 Hz

o' _~.,...,.~ t0

0

10000 Re (Z) s em 2

20000

300~

Fig. 4 -Nyquist diagrams of steel in simulating concrete solution, a: with nitrite, b: with phosphate, c: simulation results (phosphate; R= 0.4), d: circuit equivalent model used for the sinmlation.

capacitive loops. They were fitted by using the electrical where the forward scan intersects the reverse scan [10]. In equivalent circuit of Fig. 4c. the absence of pitting, forward and reverse scans coincide Fig. 4d compm~esexperimental and calculated plots from one (no hysteresis loop); thus E~p >_ Er Table 3 gathers the specimen. It appears that the calculated impedance consists of: specific potentials: Ep and E~ep. - A first capacitive loop defined in the high frequency (HF) In the presence of phosphate, the pitting potential Ep shills range, which is centered slightly off the real axis. The towards more noble values. For R = 1, this Ep value was higher associated capacitance C~ = (2rlf,,~,R04 is close to the double than 600 mV, the passivation plateau increased to about 1000 layer capacitance (Cdl close to 60 gF. cm'2). So, the resistance mV and the hysteresis loop vanishes. In fact, the observed steel R1 determined from the diameter of this loop is equivalent to surface showed no corrosion. This can correspond to the the charge transfer resistance (Rd and the HF are represents the development of ferric oxide [ 11]. Thus, the sodium phosphate charge transfer process, which leads to the initiation of inhibitor improves the pitting resistance of steel in solution corrosion. simulating concrete, by inhibiting pit growth. With nitrite, Ep and E ~ increased with Table 4 - Fitted specific parameters observed from several tests the inhibitor concentration. The presence of a hysteresis loop, even when R = 1, means ~nhibitor R0 ~L JR1 m . c, g. oq R2 kfL C2.103g C~2 that pitting growth occurred, and this Ratio R cm2 cm2 cm-2 cm2 cm'Z means that sodium nitrite had no inhibiting 0 4.'4 0.71 43 0.92 3.74 5.69 0.43 effect. This was confirmed by examining .................. 0.4 3.5 1.16 42 0.92 30.90 50 0.47 steel surface after every test, and this surface was covered with corrosion i Nitrite 1 2.5 1.45 45 0.91 51.00 100 0.43 products. 0.4 4 3.20 46 0.9 60.00 0.59 0.6 Figs. 4a and 4b give the Nyquist plots of EIS results. They are all formed by two Phosphate 1 2.65 9.04 70 0.92 89.89 0.46 0.68

533

Dhouibi, Triki, Salta, Rodrigues, Raharinaivo - The low frequency (LF) loop is rather flattened and the associated capacitance C, is higher than C~. So, this loop can represent either the propagation corrosion process or the mass transfer. The LF boundary (R2) is equivalent to the polarisation resistance R~. Table 4 gathers these dift~rent parameters, which have been determined by fitting the experimental values of the impedance diagrams to the values calculated on the basis of the equivalent circuit shown in Fig. 4e. It appears that the values of the characteristic interface parameters change with the content of the inhibitors. The electrolyte resistance R0 decreases slightly; the solution becomes more conductive. The capacilanee C~ remains close to the value corresponding to a double layer, but resistance R~ increases, particularly, with a phosphate content at R =1. This indicates that tested inhibitors do not change the interface reactions mechanisms but they can decrease the rates of some reactions leading to the growth of a passive film (likely oxide and / or a hydroxide) on the metal surface. Phosphate decreases capacitance C2 and increases resistance R2 whereas, nitrite increases these two parameters. The increase of R~ indicates that the presence of these two inhibitors diminishes the corrosion rate. Previous studies showed that a high value of C2 and the low values of the dispersion constant o.2 are related to the micro roughness o f metal surface [12]. This is due to the precipitation of some products. So, in the presence of inhibitors, these products are passivating, and the protective effect is higher with phosphate than with nitrite. Results from electrochemical tests in simulated pore solution indicate that phosphate with R equal to I yields the highest protection against steel corrosion
20(X}O0 ,
m--0.01
O

Hz

I

,~,

1ooooo
I a @

eli

o
Re (Z) ~]. c m

2

I 0oooo

ca 1000 i }

q,

"

g

HF.

(Arc) a

t

0

2000 Re (Z) g2. cm~

4000

20000

0.01 Hz
II

-%

g
O,
J~.-OA

10000
,if II'I O o o /$ 9 nI + +

3.2 Tests on r e i n f o r c e d c o n c r e t e
The open circuit potential (OCP) of reinforcing steel in sets 2 and 3 were measured once a year, before applying EIS test. Fig. 5 shows the steel OCP results obtained. It appears thai fbr concrete free from inhibitor, steel OCP is more negative (Ec = -550 ~: 20 mV sc~, when specimens were immersed in chloride solution, than when they were in ~ > 100 0
0 0

20000 Re (Z) K~. cm 2 ~* _(~or~4rotSpecimen
9

40000

6 % Nit.rim * 6 % Phosphate Hz

4(}~M
w-...~.0 . 0 1
I

/65 kHz

*100 ~ "0
I

m
n * ---0A

g
t~

- oo
-200 -300
~ ",
..+

=

4000 R,e (Z) ~. em~

8()(YO

12000

~,

9 -E -400 ~ 9

'"

Q2~.tm!Spc,cimen 9 6 % Nhritq * 6.~Aph3:)sph~ ~i
I0

-500 ~ ' : . . 2 : : : : : : . . : . . . . . . . . . . . . . . . .

9

t200

1

OA 0AH Hz

0

. . . . . . . . . . . . . . . . . . . . . . .

1

2 Time (Years)
............................... ; GTii, i? .......... ]

1200

2400

36@)

4800

6~I

~00

Re (Z) gL cm~

- - I-

- "cl

so~ . ( ~ ' i ~ i

. . . .

cl

sot,.~)~i]

Fig. 5 - Open-circuit-potential versus time of concrete specimens.

Fig. 6 - Nyquist diagrmns of concrete specimens, a: ageingone year in distilledwateT,b: ageing one year in CI solution,c: ageing two years in c r solution,d: ageing three yem~in el- solution.

534

Materials and Structures / Mat4riaux et Constructions, Vol. 36, October 2003

Fig. 7 - Equivalent circuit for modelling the steel-concrete interl'ace.

distilled water (Ec = -120 ? 20 mV seE. After one year of immersion, chloride ions content at steel surface was very high resulting in steel depassivation and corrosion initiation. Moreover, the value of OCP potential becomes more positive only for specimens containing nitrite, and after one-year immersion. With the other tests, the OCP values are similar to those of control specimens, which were immersed in the same solution. Electrochemical hnpedance Spectroscopy (EIS) was used to study the steel long-term behaviour in concrete of sets 2 and 3. The Nyquist plots contain two or three capacitive loops whose centres lie below the real axis, and are more or less separated. Figs. 6 a, b, c and d give typical diagrams for the different periods of immersion. The fitting of EIS Nyquist plots was made according to the model in Fig. 7. Figs. 8a, b and c Show the comparison between typical experi-mental and calculated Nyquist plots. The calculated impedance is composed of the electrolyte resistance R0, in series with other elements, which are: A resistance R~ , 250000 7 which corresponds to the diameter of the high frequency arc. It is related to the ionic conductivity in mortar 0,01 Hz pores [13-15]. The ,oooo corresponding capacitance values (CO obtaift. ~25000 ned from all diagrm~ns are in the order of 1 to ,,, ~G 10 nF. cm"2. These / E values are not fl~ O. 1 compatible with an interthcial film [16]. ~ o 10o00 20000 According to some works [17, 18], this 0 Re (Z) ~. cm2 125000 Re (Z) ~). cm; order of magnitude is indicative of bulk 4000 c ~ 0.01 Hz effects and the high ,, * Exp. " Cal.: frequency process is related to the dielectric 65 kHz I00 ~: 0.1 properties of concrete cover [ 19, 20]. ' Dim iN. .......... I A resistance R,z given by the diameter of the imermediate arc. 0 4000 8000 12000 The corresponding Re (Z) ~). cm 2 calculated capacitor C2 has a value in the 1 to Fig. 8 - Modelling results, a: Control specimen, ageing one year in water, b: Specimen with phosphate, ageing one year in CI solution, c: Specimen with nitrite, ageing two years in CI" solution.
-

10 gF. crn"2 ranges and does not change significantly with time. The magnitude of this capacitance C; is lower than the douNe layer capacitance. So, this arc is not indicative of a charge transfer process but is due to the growth of an interfacial layer at the steel-concrete boundary [21]. R2 deals with the resistance to ionic transport on the interfacial layer. A resistance R~, which corresponds to the low frequency arc. This resistance was used to estimate the polarisation resistance of the steel. The calculated value of (53 capacitance ranges between 0.2 mF. cm-2 and 100 inF. cm-2. This capacitance is high to be associated with a double layer capacitance. So, the low frequency loop could be ascribed to a charge ~ansfer process combined with a mass transfer process [22]. This faradic process takes place on a steel surface free from any film. - A Warburg diffusion W at the low frequency region. The slope of a logarithmic plot of the imaginary impedance component against frequency helps interpreting the results: for pure Warburg diffusion the slope value is equal to 0.5. For specimens with or without the inhibitors mixed in concrete and immersed for one, two or three years in distilled water, two time constants were determined from the experimental curves (Fig. 6a). The inhibitors effects on the concrete cover properties are related to the change of R0 and R~ resistance values (Figs. 9a and 9b). The P~ and R~ values increase with time and with the inhibitors addition. The changes with time are likely due to the ageing (hydration process) of concrete cover. The changes with the inhibitors contents are related to reactions between these compounds and hydrating concrete. Thus, concrete porosity and permeability can decrease. The low frequency parts for these diagrams correspond to a passivation process. Fig. 9c, Shows the polarisation resistance R3 as a function of ageing time for the several tests. At any
-

a

r

0

' ..............................................

..........................................

" .....................................

535

Dhouibi, Triki, Salta, Rodrigues, Raharinaivo

5

4~
R

b
300 : e

2,oe ~ : : 2 : 2; ........................... .. ..................................

........... ZL ;i................

I ( N e .................................................... ~. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2 3 1 2 3

,

Thee (Years'~

Time (Years)

I

2 Time (Yem)

3

6 % N~m~e

Fig. 9 - The evolution of the resistances as function of time of specimens aged in water. time, the R3 values are higher 3~ with the inhibitors mad increase ....... ~ ................................................... m slightly with time. The ~ 2 ~ presence of the inhibitors at the steel concrete interface leads to a passive film more protective i .,-J .J'~" than the products, which are 0 i ............................................ , ........................................................... o ~" i~ ............................................. ................................................ naturally formed in concrete. 1 2 3 ! 2 3 Some studies concerning Time (Years) Time ( Ye~rs] nitrite action confirm this point i ContrOl Spedmell [23]. These results show that ....I 6 % Nitri~ i . . . . . . 6 % Phosphate j corrosion rate is negligible in all samples, and this agrees with the visual observation of .....................................................A gO " } ............, d steel conditions. The shapes of ~ j-. .................. Nyquist diagrams of steel in concrete with or without r inhibitors and aged for one year in chloride solution 0 ............................................. ,..................................................................... *~ ......................................................... *...................... (Fig. 6b) are similar to those 0 ...................................................... ..................................................... 1 2 3 1 2 3 obtained for specimens T i m e (Years) ' ~ m e (Yeea,) immersed in distilled water. However, the low frequency I I " 6 %Ni~nW I 6 % Nilri~c loops are almost complete semi circles with a smaller Fig. 10 - The evolution of the resistances as/'unction of time of specimens aged in C1- solution. diameter as compared to those obtained with chloride free aged during three years of immersion, the highest values of Ro specimens. This is typical of an activated process and it were obtained for concrete with inhibitors. However, these corresponds to the dissolution of the passive layer by chloride values remain lower than those obtained with chloride free ions [21]. After two-year of immersion in the same solution, solutions. Thus, an inhibitor increases concrete resistivity, and the Nyquist plots are formed by three capacitive arcs (Fig. 6c). reversibly chloride ions in concrete pore solution increases its The intermediate frequency region is a complete semi circle, conductivity. The value of resistance RI also increases with time which is much depressed for inhibitor flee concrete. With indicating that steel is covered with hydration product layers, nitrite containing specimens, the shape of the low frequency which have a high ionic resistance. For inhibitor free samples, part of the impedance curve is similar to the usual shape this high resistance drops after a one-year immersion. As shown related to diffusion with a characteristic slope equal to 1. In in Fig. 10c, the value of resistance R2 increases with time. this case, reactions kinetics at the interface are governed by According to Sagoe-Crentsil et al. [21], this indicates either the diffusion processes [24] and the polarisation resistance can growth of the layer and/or some structural or chemical change only be roughly estimated. For the other this slope is equal to in the layer products. The effectiveness of the inhibitors against 0.7. This means that the interface is under an activation steel corrosion in concrete was determined by measuring the process, which takes place on steel surface uncovered by any . . . . . . . . . . resistance R3 versus time (Fig. 10d). For all film. For specimens after three years of immersion, tim immersion times, the highest R~ values correspond to samples Nyquist diagram is tbrmed of three depressive loops (i.e. containing inhibitors. In all the specimenS, these R3 values tend ct<0.5) then a very small capacitive loop indicates that steel is to decrease with time down to very low values, which indicates under a corrosion process. a severe corrosion. Specimens without inhibitors, immet~d in As described 'above, the long term action of inhibitors on the chloride solution, show a severe corrosion before a one,year bulk properties of the concrete cover was evaluated by the immersion (R3 < 13 k.Q. cm2). The presence of inhibitors and resistances R0 mad RI (Figs. 10a and 10b). For all the s ~ i m e n s
<

536

Materials' a n d S t r u c t u r e s 1 Matrriaux

et Constructions, Vol. 36, October 2003

Table 5 - Content of"Free" and "Total" chloride (% of cement weight) at the interface of the reinforced concrete. (Immersion 3 years in 0.5 M NaCI)

Concrete rWithout inhi'igitor With nitrite With phosphate

"Free" chloride 1.9 1.16 1.2

"Total" chloride 2,75 2.1 2.6

Fig. I 1 - Visual observation of the reinforcing steel. Ageing: three years in 3 % NaCI solution, a: without inhibitor, b: with Nitrite, c: with Phosphate. particularly nitrite significantly decreases the corrosion rate for specimens in Chloride solution after one-year immersion (R3 = 70 kfl. cm 2 ). When immersion time exceeds one year, the inhibitor efficiency is significantly reduced (R3 < 3 kf2. cm~ after a three-year immersion) and the steel corrosion rate increases. This was confirmed by a visual observation of the steel condition. Previous works [25-28] are also in agreement with these results. After immersions, reinforced concrete specimens were split by a compressive loading, parallel to the bars. The visual examinations of the exposed steel bars showed no corrosion in set 2. The steel bars in set 3 showed tarnishing and rust pots, mainly for specimens without inhibitor (Figs. 11 a, 1 lb and 1 lc). The broken parts of split reinforced concrete specimens were used tbr determining chloride contents. The free and total chlorides in concrete at steel level were determined for specimens, which had been immersed for three years in chloride solution (set 3). Table 5 shows the results obtained.

When concrete is free from inhibitors, free chloride content is high at the rebar level measured. The presence of inhibitor in concrete tends to reduce chloride content but this value remains higher than the critical chloride corrosion threshold. It is of interest to notice that bound chloride, which is the difference between the total and the free chloride, has a lower content in concrete containing nitrite than in the other specimens. Figs. 12a, 12b and 12c show the infrared spectra of concrete samples taken close to the rebar. The main bands observed in the spectra are those characteristic of carbonate from calcite, nitrite and phosphate. Absorption peaks at 1440, 870 and 710 cm 4 are related to carbonates. The band at 1270 cm 4 is attributed to nitrite and the bands at 1130 and 625 cm 4 to phosphate. These results show that: - Nitrite can be easily detected in 30 day-old concrete and after three years o f immersion in distilled water but not in chloride solution for which the absorption bands significantly decreased. This can be due to the enrichment of the matrix in sodium chloride and / or to the diffusion of nitrite from the concrete to the solution. -However, phosphate can be easily detected only in 30 day-old concrete. The main characteristic absorption band of the carbonate (1440 cm 4) becomes larger and its shape changes when phosphate is added. As concrete pH
b !S

X ....................~ i ~ i ~ ...............................................
3

b
1628.58
............ /,. .......

is ,A Ib

\+

./
I <" /1132.19'

1ii ~s 3448'45 t

62291

i
i~ 3 2 ., ,

86~.45

1413.60

~A ib !
4~>0 3000 2060
1000

!

n
i i

i
4ooo

",
3oo0 2000

A/
...........................................

.

~

t
\ .....

567.t5
p ~..~"~ .........

',.......................................................................

~m~r~_~'~), ............................................................

i ............................................................. ~v~,i~:)

jooo ...........................................

',A

. . . . . .

7~?~e~ .......................................................................................... iA

3"0 . .i .V . .'~ . .tter .......................................

1421.39

I

e' s

3)

~t ~ "\,~

1384,28}

1270.99

715,67
',A
',r

i; ,,0
le

*
~ f~; ................................................................................

'!)

4

CC~i8~8~ )7t.,, ......................................................................

+

....................................................

30 C ~

1635.82
j.....................................................

1447,715

)a

; ,

~. , ....

}C
4ooo 3.000

i;
i

ia )n

............,. . . . .

-/"

"

2ooo

Iooo

I

'.01 0.S.
4ooo

...............
3ooo

................ 2 ~ ' .....................

.............. ..... ,>...1
ioc~

9 6 ~ . 2 1 t,
............. J

Fig. 12 - IR spectrums of concrete near the rebar, a: Sound concrete withom exposure (30 days old), b: Concrete without inhibitors (immersion 3 years), c: Concrete with nitrite (exposure 3 years), d: Concrete with phosphate (exposure 3 years).

537

Dhouibi, Triki, Salta, Rodrigues, Raharinaivo
Table 6 - Main cement hydration products detected by XRD measurement

Set

1 (30-day-old specimens) Product A Portlandite i ++ Calcite + Tr Ettringite Chloroaluminate No
B C

+++ + Tr No

+++ + Tr No

2 (years of immersion A B ++ +++ +++ +++ Tr Tr iNo No

in water) C +++ + Tr No

3 (years of immersion in chloride) A B C ++ +++ +++ +++ +++ + + + +
+ + ++

Notation: A." Inhibitor f r e e sample ; B: Nitrite containing sample ; C: Phosphate containing sample +++: abundant content ; ++: middle content ; + low content " Tr." very low content : No: Absence.

decreases in the presence of phosphate, the equilibrium between CO32 and HCO3- can change, and calcium carbonate can be formed either in distilled water or in chloride solution. Table 6 shows the main crystalline hydration products in samples, as well as their qualitative content. In samples containing inhibitors, X-ray diffraction measurements detected the same products, in concrete at its interface with concrete, as for free inhibitor samples. But the amount of calcite :from cement is lower in concrete containing phosphate than in the other specimens. Concrete samples were scrapped from its interface with steel in specimens of set 2 (3-year immersion in water). They were examined with SEM and the results obtained show that only ettringite, portlandite and calcite have been detected inside the cement pores. In samples scrapped from the concrete interface of set 3 (3-year immersion in chloride solution), other compounds such as chloroaluminates and some corrosion products were observed. Some traces of phosphates were detected (< 0.8 %) by EDS at the interface. The SEM examination showed that corrosion products are amorphous and they probably deal with iron chlorocomplexes [29]. Some corrosion products are crystalline and they are then related to iron oxides. Figs. 13a, 13b and 14a, 14b and 14c give some SEM photos and the EDX microanalysis of concrete samples containing nitrite or phosphate. 4. C O N C L U S I O N The results of electrochemical tests on steel in solution indicate that the sodium phosphate prevents pitting corrosion when its content is not less than the chloride concentration and nitrite only increases the value of pitting potential. The two inhibitors do not change the interface reaction mechanisms but they slow down both the rate of dissolution processes and the rate of corrosion products growth. This leads to the growth of a passive film on the metal surface, which reduces the area of active surthce on steel. In solutions, inhibitor based phosphate added at Cin = Col, yields the highest protection against steel corrosion. Some conclusions can be drawn from tests on reinforced concrete: 9 First, AC impedance spectra make it possible to identify passive state, light corrosion and severe corrosion of steel. They provide information on the electrical resistance of concrete, on the interface layer and on the corrosion process. So, this EIS technique is suitable for studying the action of inhibitors on steel-concrete interface, and for

evaluating their long-term effectiveness. It shows that steel remains in a passive state, when its concrete (with or without inhibitor) cover is immersed in distilled water. The measured values of corrosion potential values and visual examinations o f steel confirmed these EIS results. 9 After one year of immersion in chloride solution the corrosion rate of steel in concrete with the two inhibitors, decreased. But, the corrosion process remains active because chlorides break down the passive film. 9 After two years of immersion in chloride solution, an intermediate frequency arc appears in the EIS Nyquist diagrams. This was attributed to the growth of corrosion product layer at steel surface: The corrosion rate increased, and the reaction at concrete-steel interface is a diffusioncontrolled charge transfer process. 9 After three years of immersion in NaC1 solution, no influence of inhibitors on the corrosion rate was observed, but the electrical resistance of concrete significantly increased in the presence of inhibitor. The measured flee chloride content was high at the rebar level. In the presence of inhibitors chloride content was lower, but its value was higher than the critical chloride threshold. So, C1- ions are likely complexing Fe 2+cations and corrosion rate increases. Tarnishing and rust pots were visually observed on steel surface. The SEM examinations showed that corrosion products were either amorphous, (iron chlorocomplexes) or crystalline (iron oxides). 9 FT-IR spectra indicate that the absorption bands of nitrite significantly decreased in chloride solution. This can be due to a nitrite enrichment o f the matrix immersed in sodium chloride and/or to a leaching out (diffusion of nitrite from the concrete to the solution). The carbonate absorption band becomes larger and its form changes, when phosphate is added. So, phosphate likely interferes with the CO32- / HCO3- equilibrium, and crystalline calcium carbonate is likely formed. 9 The tests applied to concrete samples taken from its interface with steel were SEM observations and the XRD measurements. The results obtained showed no new compound in concrete: only ettringite, portlandite, calcite and chloroaluminates were observed. Some corrosion products and traces of phosphates (< 0.8 %) were observed. 9 The last conclusion is: solutions simulating concrete do not show the same inhibitor performance as in real concrete, but they help to clarify the reactions mechanisms. The results obtained with reinforced concrete specimens showed that none of the two studied mixed-in inhibitors have a long-term effectiveness when concrete is later contaminated with chloride.

538

Materials and Structures / Mattriaux et Constructions, Vol. 36, October 2003

ACKNOWLEDGEMENTS
The authors wish to thank Ana Paula Menezes, LNEC, for her help in SEM and EDS techniques. The authors also gratefully acknowledge the assistance provided by Ant6nio Santos Silva and Ludovina Matos, also LNEC, concerning the execution and interpretation of X-Ray spectra.

REFERENCES
[1] Rosenberg, A.M. and Gaidis, J.M., 'The mechanism of nitrite inhibition of chloride attack on reinforcing steel in alkaline aqueous environments', Mat. Pelformance 18 (11 ) (1979) 45-48. [2] Gaidis, J.M. and Rosenberg, A.M., 'The inhibition of chloride-induced corrosion in reinforced concrete by calcitLm nitrite', Cement, Concrete and Aggregates 9 (1) (1987) 30-33. [3] Andrade, C., Alonso, C. and Gonzalez, J.A., 'Some laboratory experiments on the inhibitor effect of sodium nitrite on reinforcement corrosion', Cement, Concrete and Aggregates 8 (2) (1986) 110-116. [4] Andrade, C., Alonso, C., Acha, M. and Malric, B., 'Preliminary testing of Na2PO3F as a curative corrosion inhibitor for steel reinforcements in concrete', Cement and Concrete Research 22 (1992) 869-881. [5] Berke, N.S. and Sundberg, K.M., 'The effect of admixtures and concrete mix designs on long-term concrete durability in chloride environments', Corrosion (386) (1989). [6] Berke, N.S., 'The effect of calcium nitrite and mix design on the corrosion resistance of steel in concrete (part 2, Long-term Results)', Corrosion (132) (1987). [7] Dhouibi, L., Triki, E., Raharinaivo, A., Trabanelli, G. and Zucchi, F., 'Electrochemical methods for evaluating inhibitors of steel corrosion in concrete', British Corrosion Journal 35 (2) (2000) 145-149. [8] Dhouibi, L., Triki, E. and Raharinaivo, A., 'Inhibiting the convsion of steel in a medium simulating concrete by ttsing sodimaa phosphate', Conference on Understanding Corrosion Mechanisms in Concrete., Cambridge, Massachussets, USA, July 1997. [9] Dhouibi, L., Refait, Pb., Abdelmoula, M, Triki, E. and Gtnin, J.-M.R., 'Influence of phosphate on the corrosion products of iron in chloridepolluted concrete-simulating ~lutions: Ferrihydrke Vs green rust', Corrosion Nace. (In press).

Fig. 13 - SEM photos and EDS analysis of nitrite containing concrete, a: Presence of Portlandite (3 years' immersion in water), b: Presence of corrosion products at the cement pores (3 years' immersion in chloride solution).

Fig. 14 - SEM photos and EDS analysis of phosphate containing concrete, a: Presence of chloroaluminates (3 years immersion in chloride solution), b: Presence of corrosion products at the cement pores (3 years inamersion in chloride solution), c: Cristallized corrosion products.

539

Dhouibi, Triki, Salta, Rodrigues, Raharinaivo [10] Berke, N.S. and Hicks, M.C., 'Electrochemical methods of determining the corrosivity of steel in concrete' Corrosion Testing and Evaluation', ASTM STP 1000, (1990) 425-440. [11] Berke, N.S. and Hicks, M.C., 'Protection mechanisms of calcium nitrite', Conference on Understanding Corrosion Mechanisms in Concrete. Cambridge, USA, July (1997). [12] Roberge, P.R., 'Analysing electrochemical impedance corrosion measurements by the systematic permutation of data points', ASTM, STP (1154) (1992)197-21 t. [13] Monticelli, C.A., Frignani, G., Trabanelli, G. and Brunoro, A., ' Study on the inhibiting efficiency of a glycerophosphatenitrite admixtures against steel corrosion in mortars', Proceedings of the 8t~ European Symposium on Corrosion Inhibitors, Ferrara, (1995) 609-620. [14] Hachani, L., Carpio, C., Fiaud, C., Raharinaivo, A. and Triki, E., 'Steel Corrosion in Concrete Deteriorated by Chlorides and Sulphates: Electrochemical Study Using Impedance Spectrometry and "Stepping Down the Current" Method', Cement and Concrete Research 22 (1992) 56-66. [15] Hachani, L., Fiaud, C., Triki, E. and Raharinaivo, A., 'Characterization of steel-concrete interface by electrochemical impedance spectroscopy', Brit. Corr. J. 29 (2) (1994) 22-217. [16] Dhouibi-Hachani, L., Triki, E., Grandet, J. and Raharinaivo, A., 'Comparing the steel-concrete interface state and its electrochemical impedance', Cement and Concrete Research 26 (2) (1996) 253-266. [17] Keddam, M., Novoa, X.R., Soler, L., Andrade, C. and Takenouti, H., 'An equivalent electrical circuit of macrocell activity in facing electrodes embedded in cement mortar', Corrosion Science 36 (7) (1994) 1155-1166. [18] Scuderi, E., Mason, T.O. and Jennings, H.M., 'Impedance spectra of hydrating cement pastes', in 'Materials Science and Engineering', (Chapman and Hall Ltd, 1991) 349 - 353. [19] Montemor, M.F., Simts, A.M.P., Salta, M.M. and Ferreira, M.G.S., 'The assessment of the electrochemical behaviour of flyash-containing concrete by impedance spectroscopy', Corrosion Science 35 (8) (1993) 1571-1578. [20] Andrade, C., Soler, L., Alonso, C., Novoa, X. R. and Keddam, M., 'The importance of geometrical considerations in the measurements of steel corrosion in concrete by means of AC impedance', Corrosion Science 37(12) (1995) 2013-2023. [21] Sagoe-Crentsi, K.K., Glasser, F.P. and Irvine, J.T.S., 'Electrochemical characteristics of reinfbrced concrete corrosion as determined by impedance spectroscopy', Br. Corr. J. 27 (2) (1992) 113-118. [22] Flis, J.L., Dawson, J., Gill, G.C. and Wood, G., 'Impedance and electrochemical noise measurements on iron and ironcarbon alloys in hot caustic soda', Corrosion Science 32 (8) (1991) 877-892. [23] Berke, N.S., 'Corrosion inhibitors in concrete', Concrete International (1991) 24-27. [24] Dhouibi, L., Triki, E. and Raharinaivo, A., 'Laboratory experiments for assessing the effectiveness of inhibitors against steel corrosion in concrete', Sixth International Symposium on Advances in Electrochemical Science and Technology, Chennai (Madras), India (1998). [25] Dhouibi, L., Triki, E. and Raharinaivo, A., 'Study of sodium nitrite and calcium nitrate as inhibitors of steel corrosion in concrete', Intemational congress of creating with concrete, Dundee Scotland, (1999). [26] Dhouibi, L., Triki, E. and Raharinaivo, A., 'Application of electrochemical impedance spectroscopy for assessing the effectiveness of inhibitors against steel corrosion in concrete', 14th lntemational Corrosion Congress, Cape Town South Africa, (1999). [27] Dhouibi, L., Triki, E. and Raharinaivo, A., 'The application of electrochemical impedance spectroscopy to determine the long-term effectiveness of corrosion inhibitors for steel in concrete', Cement and Concrete Composite 24 (2002) 35-43. [28] Graig, R.J. and Wood, L.E., 'Effectiveness of corrosion inhibitors and their influence on the physical properties of portland cement mortars', Highway Research Record (328) (1970). [29]Berke, N.S. and Thomas, T.G., 'World-wide review of corrosion inhibitors in concrete', Advances in Concrete Technology', CANMET, Editor V.M. Malhotra, (1989) 899924.
Paper received: February 28, 2002; Paper accepted: April 29, 2002

540


相关文章:
...on Mechanical Behavior of Concrete Due to Reinforcement ....ppt
Research on Mechanical Behavior of Concrete Due to Reinforcement Corrosion_建筑/土木_工程科技_专业资料。Research on Mechanical Behavior of Concrete Due to ...
...of Silica Fume on Corrosion Behaviour of Reinforced Steel ....pdf
Influece of Silica Fume on Corrosion Behaviour of Reinforced Steel in Different Media_IT/计算机_专业资料。Electrochemical and corrosion behaviour of ...
Corrosion of steel fibre reinforced concrete from the cracks_....pdf
Cement and Concrete Research 35 (2005) 572 577 Corrosion of steel ...Fibre reinforcement; Corrosion; Steel fibres 1. Introduction Fibre addition in...
...study of corrosion inhibition of steel reinforcement in ....unkown
J Mater Sci (2007) 42:47214730 DOI 10.1007/s10853-006-0880-3 Electrochemical study of corrosion inhibition of steel reinforcement in alkaline ...
...Inhibitors on the Corrosion of Mild Steel Reinforcement in....unkown
1 Comparative Study of Inhibitors on the Corrosion of Mild Steel Reinforcement in Concrete Burubai, W. and Dagogo, G. Department of Agricultural and ...
...Study on Corrosion Protection of Reinforcing Steel by ....unkown
10.3390/ma8010251 Article OPEN ACCESS materials ISSN 1996-1944 www.mdpi.com/journal/materials Comparative Study on Corrosion Protection of Reinforcing Steel ...
...phosphates as corrosion inhibitors for steel reinforcements.unkown
study of three sodium phosphates as corrosion ...inhibition of steel reinforcement corrosion using ... sodium monofluorophosphate (Na2PO3F), disodium...
...Phosphates a Corrosion Inhibitors of Steel Reinforcements ....unkown
The present work studies this objective, analyzing the passivating capacity of... as well as the corrosion behavior of steel reinforcements under the effect...
Surface corrosion of steel fibre reinforced concrete.pdf
Surface corrosion of steel fibre reinforced concrete...reinforcement (E) Concrete skin Corrosion (C) ...Thus a study of concrete porosity and chloride ...
...in-situ measurement of reinforcement corrosion.pdf
reinforcement corrosion_建筑/土木_工程科技_专业资料...steel is one of the most dominant causes of ...However, studies on the relationship between the ...
...with All Soaking and Impressed Current Method Corrosion*_....pdf
Soaking and Impressed Current Method Corrosion*_建筑/土木_工程科技_专业资料。...In order to research the bond properties between corroded reinforcement bars ...
...of Marine Reinforced Concrete with Steel Corrosion and ....pdf
Probabilistic Lifetime Assessment of Marine Reinforced Concrete with Steel Corrosion and Cover_专业资料。ChnOcaiaenEn,12,o2P.031 gVo.5N.,P358 ◎2 1...
...effects of influence factors on probability of corrosion ....pdf
Coupling effects of influence factors on probability of corrosion initiation time of reinforced_专业资料。JCetSuhUnvTehn1(0 :22 .n.ot i.co.21)1232918...
...of resistivity spectra from reinforced concrete corrosion ....pdf
Cement and Concrete Research 31 (2001) 603 ± 607 Validation of resistivity...resistivity spectra measured from corrosion of reinforcing steel bar in ...
Evaluation of Corrosion Inhibiting Systems Used in Reinforced....pdf
solutions to the reinforcement corrosion in concrete...This study was performed by measuring the ...1. Values of 1/Rp measured on carbon steel in...
Phosphate ions as corrosion inhibitors for reinforcement ....unkown
Accepted Manuscript Title: Phosphate ions as corrosion inhibitors for reinforcement steel in chloride-rich environments Author: L. Yohai M. Vazquez M.B...
Corrosion Effects on the Strength Properties of Steel Fibre ....unkown
Hindawi Publishing Corporation International Journal of Corrosion Volume 2014, Article ID 595040, 7 pages http://dx.doi.org/10.1155/2014/595040 Research ...
...Corrosion Behavior of Mild Steel Reinforcement in Concrete....unkown
SCIENCE Effect of Inhibitors and Admixed Chloride on Electrochemical Corrosion Behavior of Mild Steel Reinforcement in Concrete in Seawater ABSTRACT ...
...for the Corrosion Inhibition of Steel Reinforcement in ....unkown
16, No. 2, 99108 (2010) Silane and Siloxane Based Formulations for the Corrosion Inhibition of Steel Reinforcement in Concrete M. Friedel1 *, M. ...
A Study on Corrosion of Reinforcement in Concrete and Effect ....unkown
A Study on Corrosion of Reinforcement in Concrete...inhibition to the corrosion of the rebar increases...is 3-6 times more than the volume of steel. ...
更多相关标签: