Handbook of Corrosion Engineering Episode 1 Part 3 docx

1.14 that hydrogen reduction canoccur on the reinforcing steel cathode if its potential drops to highlynegative values.Fe Fe2+ Decreasing pHfrom carbonationmakes shift toactive fieldposs

alloying additions or protective coatings for corrosion resistance areassociated with this steel.

In simplistic terms, concrete is produced by mixing cement clinker,water, fine aggregate (sand), coarse aggregate (stone), and other chem-ical additives When mixed with water, the anhydrous cement clinkercompounds hydrate to form cement paste It is the cement paste thatforms the matrix of the composite concrete material and gives it itsstrength and rigidity, by means of an interconnected network in whichthe aggregate particles are embedded The cement paste is porous innature An important feature of concrete is that the pores are filledwith a highly alkaline solution, with a pH between 12.6 and 13.8 atnormal humidity levels This highly alkaline pore solution arises fromby-products of the cement clinker hydration reactions such as NaOH,KOH, and Ca(OH)2 The maintenance of a high pH in the concrete poresolution is a fundamental feature of the corrosion resistance of carbonsteel reinforcing bars

At the high pH levels of the concrete pore solution, without theingress of corrosive species, reinforcing steel embedded in concretetends to display completely passive behavior as a result of the forma-tion of a thin protective passive film The corrosion potential of passivereinforcing steel tends to be more positive than about 0.52 V (SHE)according to ASTM guidelines.9 The E-pH diagram in Fig 1.14 con-

firms the passive nature of steel under these conditions It also cates that the oxygen reduction reaction is the cathodic half-cellreaction applicable under these highly alkaline conditions

indi-One mechanism responsible for severe corrosion damage to ing steel is known as carbonation In this process, carbon dioxide fromthe atmosphere reacts with calcium hydroxide (and other hydroxides)

reinforc-in the cement paste followreinforc-ing reaction (1.6)

Ca(OH)2 CO2→CaCO3 H2O (1.6)The pore solution is effectively neutralized by this reaction.Carbonation damage usually appears as a well-defined “front” parallel

to the outside surface Behind the front, where all the calcium ide has reacted, the pH is reduced to around 8, whereas ahead of thefront, the pH remains above 12.6 When the carbonation front reachesthe reinforcement, the passive film is no longer stable, and active cor-rosion is initiated Figure 1.14 shows that active corrosion is possible

hydrox-at the reduced pH level Damage to the concrete from carbonhydrox-ation-induced corrosion is manifested in the form of surface spalling, result-ing from the buildup of voluminous corrosion products at theconcrete-rebar interface (Fig 1.15)

carbonation-A methodology known as re-alkalization has been proposed as aremedial measure for carbonation-induced reinforcing steel corro-

sion The aim of this treatment is to restore alkalinity around thereinforcing bars of previously carbonated concrete A direct current isapplied between the reinforcing steel cathode and external anodespositioned against the external concrete surface and surrounded byelectrolyte Sodium carbonate has been used as the electrolyte in thisprocess, which typically requires several days for effectiveness.Potential disadvantages of the treatment include reduced bondstrength, increased risk of alkali-aggregate reaction, microstructuralchanges in the concrete, and hydrogen embrittlement of the reinforc-ing steel It is apparent from Fig 1.14 that hydrogen reduction canoccur on the reinforcing steel cathode if its potential drops to highlynegative values.

Fe

Fe2+

Decreasing pHfrom carbonationmakes shift toactive fieldpossible

Potential rangeassociatedwith passivereinforcing steel

Re-alkalizationattempts tore-establishpassivity

HFeO2

-Fe O3 4

Figure 1.14 E-pH diagram of the iron-water system with an emphasis on the

microenviron-ments produced during corrosion of reinforcing steel in concrete.

1.3 Kinetic Principles

Thermodynamic principles can help explain a corrosion situation interms of the stability of chemical species and reactions associated withcorrosion processes However, thermodynamic calculations cannot beused to predict corrosion rates When two metals are put in contact,they can produce a voltage, as in a battery or electrochemical cell (seeGalvanic Corrosion in Sec 5.2.1) The material lower in what has beencalled the “galvanic series” will tend to become the anode and corrode,while the material higher in the series will tend to support a cathodicreaction Iron or aluminum, for example, will have a tendency to cor-rode when connected to graphite or platinum What the series cannotpredict is the rate at which these metals corrode Electrode kineticprinciples have to be used to estimate these rates

1.3.1 Kinetics at equilibrium: the exchange

current concept

The exchange current I0 is a fundamental characteristic of electrodebehavior that can be defined as the rate of oxidation or reduction at anequilibrium electrode expressed in terms of current The term

exchange current, in fact, is a misnomer, since there is no net current

flow It is merely a convenient way of representing the rates of tion and reduction of a given single electrode at equilibrium, when noloss or gain is experienced by the electrode material For the corrosion

oxida-of iron, Eq (1.1), for example, this would imply that the exchange

Stresses due to corrosion product buildup

Voluminous corrosion products

Cracking and spalling of the concrete cover

rent is related to the current in each direction of a reversible reaction,

i.e., an anodic current I arepresenting Eq (1.7) and a cathodic current

I crepresenting Eq (1.8)

Since the net current is zero at equilibrium, this implies that the

sum of these two currents is zero, as in Eq (1.9) Since I a is, by vention, always positive, it follows that, when no external voltage orcurrent is applied to the system, the exchange current is as given by

a function of the following main variables:

1 Electrode composition. Exchange current density depends uponthe composition of the electrode and the solution (Table 1.1) For redoxreactions, the exchange current density would depend on the composi-tion of the electrode supporting an equilibrium reaction (Table 1.2)

TABLE 1.1 Exchange Current Density (i0 )

for Mz+/ M Equilibrium in Different Acidified

Table 1.3 contains the approximate exchange current density for thereduction of hydrogen ions on a range of materials Note that the val-

ue for the exchange current density of hydrogen evolution on platinum

is approximately 102A/cm2, whereas that on mercury is 1013A/cm2

2 Surface roughness. Exchange current density is usuallyexpressed in terms of projected or geometric surface area and dependsupon the surface roughness The higher exchange current density forthe H/H2 system equilibrium on platinized platinum (102 A/cm2)compared to that on bright platinum (103A/cm2) is a result of the larg-

er specific surface area of the former

3 Soluble species concentration. The exchange current is also acomplex function of the concentration of both the reactants and theproducts involved in the specific reaction described by the exchangecurrent This function is particularly dependent on the shape of thecharge transfer barrier  across the electrochemical interface

TABLE 1.2 Exchange Current Density (i0 ) at 25°C for Some Redox Reactions

System Electrode Material Solution log10i0, A/cm 2

Platinum 10%–Rhodium Perchloric acid 9.0

TABLE 1.3 Approximate

Exchange Current Density (i0 ) for

the Hydrogen Oxidation Reaction

4 Surface impurities. Impurities adsorbed on the electrode face usually affect its exchange current density Exchange current den-sity for the H/H2system is markedly reduced by the presence of traceimpurities like arsenic, sulfur, and antimony.

sur-1.3.2 Kinetics under polarization

When two complementary processes such as those illustrated in Fig.1.1 occur over a single metallic surface, the potential of the materialwill no longer be at an equilibrium value This deviation from equilib-

rium potential is called polarization Electrodes can also be polarized

by the application of an external voltage or by the spontaneous duction of a voltage away from equilibrium The magnitude of polar-ization is usually measured in terms of overvoltage , which is a

pro-measure of polarization with respect to the equilibrium potential Eeqof

an electrode This polarization is said to be either anodic, when theanodic processes on the electrode are accelerated by changing the spec-imen potential in the positive (noble) direction, or cathodic, when thecathodic processes are accelerated by moving the potential in the neg-ative (active) direction There are three distinct types of polarization

in any electrochemical cell, the total polarization across an chemical cell being the summation of the individual elements asexpressed in Eq (1.11):

electro- total act conc iR (1.11)where act activation overpotential, a complex function describing

the charge transfer kinetics of the electrochemicalprocesses act is predominant at small polarization cur-rents or voltages

conc concentration overpotential, a function describing themass transport limitations associated with electrochemi-cal processes conc is predominant at large polarizationcurrents or voltages

iR  ohmic drop iR follows Ohm’s law and describes the

polar-ization that occurs when a current passes through anelectrolyte or through any other interface, such as surfacefilm, connectors, etc

Activation polarization. When some steps in a corrosion reaction trol the rate of charge or electron flow, the reaction is said to be underactivation or charge-transfer control The kinetics associated withapparently simple processes rarely occur in a single step The overallanodic reaction expressed in Eq (1.1) would indicate that metal atoms

con-in the metal lattice are con-in equilibrium with an aqueous solution contacon-in-ing Fe2 cations The reality is much more complex, and one would need

contain-to use at least two intermediate species contain-to describe this process, i.e.,

Felattice→Fesurface

Fesurface→Fe2 

surface

Fe2  surface→Fe2 

solution

In addition, one would have to consider other parallel processes,such as the hydrolysis of the Fe2 cations to produce a precipitate orsome other complex form of iron cations Similarly, the equilibriumbetween protons and hydrogen gas [Eq (1.2)] can be explained only byinvoking at least three steps, i.e.,

H →Hads

Hads Hads→H2 (molecule)

H2 (molecule) →H2 (gas) The anodic and cathodic sides of a reaction can be studied individual-

ly by using some well-established electrochemical methods in which theresponse of a system to an applied polarization, current or voltage, isstudied A general representation of the polarization of an electrode sup-porting one redox system is given in the Butler-Volmer equation (1.12):

ireaction i0exp reaction reactionexp  (1  reaction) reaction  (1.12)

where ireaction anodic or cathodic current

reaction charge transfer barrier or symmetry coefficient for the

anodic or cathodic reaction, close to 0.5 reaction  Eapplied  Eeq, i.e., positive for anodic polarization and

negative for cathodic polarization

n  number of participating electrons

R  gas constant

T  absolute temperature

F Faraday

nF RT

nF RT

36 Chapter One

When reactionis anodic (i.e., positive), the second term in the

Butler-Volmer equation becomes negligible and i a can be more simplyexpressed by Eq (1.13) and its logarithm, Eq (1.14):

i a  i0expa a  (1.13)

where b ais the Tafel coefficient that can be obtained from the slope of

a plot of against log i, with the intercept yielding a value for i0

Similarly, when reactionis cathodic (i.e., negative), the first term in

the Butler-Volmer equation becomes negligible and i ccan be more

sim-ply expressed by Eq (1.16) and its logarithm, Eq (1.17), with b c

obtained by plotting versus log i [Eq (1.18)]:

Because the rate of the cathodic reaction is proportional to the face concentration of the reagent, the reaction rate will be limited by adrop in the surface concentration For a sufficiently fast charge trans-fer, the surface concentration will fall to zero, and the corrosionprocess will be totally controlled by mass transport As indicated inFig 1.16, mass transport to a surface is governed by three forces: dif-

RT

nF

i a

i0nF RT

fusion, migration, and convection In the absence of an electric field,the migration term is negligible, and the convection force disappears

in stagnant conditions

For purely diffusion-controlled mass transport, the flux of a species

O to a surface from the bulk is described with Fick’s first law (1.19),

where J O  flux of species O, mol  s1 cm2

D O  diffusion coefficient of species O, cm2 s1

 concentration gradient of species O across the interface,

TABLE 1.4 Solubility of Oxygen in Air-Saturated Water

Temperature, °C Volume, cm 3 * Concentration, ppm Concentration (M), mol/L

D O (1.20)

where z O  the valency of species O

R gas constant, i.e., 8.314 J  mol1 K1

T  absolute temperature, K

F Faraday’s constant, i.e., 96,487 C  mol1

Table 1.6 contains values for D O and Oof some common ions Formore practical situations, the diffusion coefficient can be approximat-

ed with the help of Eq (1.21), which relates D Oto the viscosity of thesolution  and absolute temperature:

Figure 1.16 Graphical representation of the processes occurring at an electrochemical interface.

TABLE 1.6 Conductivity and Diffusion Coefficients of Selected Ions at Infinite Dilution in Water at 25°C

Cation |z| , S  cm 2  mol 1 D 5 , cm 2  s 1 Anion |z| , S  cm 2  mol 1 D 5 , cm 2  s 1

The region near the metallic surface where the concentration dient occurs is also called the diffusion layer Since the concentra-tion gradient C O/ x is greatest when the surface concentration of species O is completely depleted at the surface (i.e., C O  0), it followsthat the cathodic current is limited in that condition, as expressed by

While the ohmic drop is an important parameter to consider whendesigning cathodic and anodic protection systems, it can be mini-mized, when carrying out electrochemical tests, by bringing the refer-ence electrode into close proximity with the surface being monitored.For naturally occurring corrosion, the ohmic drop will limit the influ-ence of an anodic or a cathodic site on adjacent metal areas to a cer-tain distance depending on the conductivity of the environment Fornaturally occurring corrosion, the anodic and cathodic sites often areadjacent grains or microconstituents and the distances involved arevery small

1.3.3 Graphical presentation of kinetic data

Electrode kinetic data are typically presented in a graphical formcalled Evans diagrams, polarization diagrams, or mixed-potential dia-grams These diagrams are useful in describing and explaining manycorrosion phenomena According to the mixed-potential theory under-lying these diagrams, any electrochemical reaction can be algebraical-

ly divided into separate oxidation and reduction reactions with no netaccumulation of electric charge In the absence of an externallyapplied potential, the oxidation of the metal and the reduction of somespecies in solution occur simultaneously at the metal/electrolyte inter-face Under these circumstances, the net measurable current is zeroand the corroding metal is charge-neutral, i.e., all electrons produced

by the corrosion of a metal have to be consumed by one or more

cathod-ic processes (eproduced equal econsumed with no net accumulation

of charge)

It is also important to realize that most textbooks present corrosioncurrent data as current densities The main reason for that is simple:Current density is a direct characteristic of interfacial properties.Corrosion current density relates directly to the penetration rate of ametal If one assumes that a metallic surface plays equivalently therole of an anode and that of a cathode, one can simply balance the cur-rent densities and be done with it In real cases this is not so simple.The assumption that one surface is equivalently available for bothprocesses is indeed too simplistic The occurrence of localized corrosion

is a manifest proof that the anodic surface area can be much smallerthan the cathodic Additionally, the size of the anodic area is ofteninversely related to the severity of corrosion problems: The smaller the

anodic area and the higher the ratio of the cathodic surface S cto the

anodic surface S a , the more difficult it is to detect the problem.

In order to construct mixed-potential diagrams to model a corrosionsituation, one must first gather (1) the information concerning theactivation overpotential for each process that is potentially involvedand (2) any additional information for processes that could be affected

by concentration overpotential The following examples of increasingcomplexity will illustrate the principles underlying the construction ofmixed-potential diagrams

The following sections go through the development of detailed tions and present some examples to illustrate how mixed-potentialmodels can be developed from first principles

equa-1 For simple cases in which corrosion processes are purely controlled

activation-2 For cases in which concentration controls at least one of the sion processes

Activation-controlled processes. For purely activation-controlled

processes, each reaction can be described by a straight line on an E versus log i plot, with positive Tafel slopes for anodic processes and

negative Tafel slopes for cathodic processes The corrosion anodicprocesses are never limited by concentration effects, but they can belimited by the passivation or formation of a protective film

Note: Since 1 mA cm2corresponds to a penetration rate of 1.2 cm peryear, it is meaningless, in corrosion studies, to consider current densi-

ty values higher than 10 mAcm2or 102A cm2.

The currents for anodic and cathodic reactions can be obtainedwith the help of Eqs (1.14) and (1.17), respectively, which generallystate how the overpotential varies with current, as in the followingequation:

 b log10(I/I0)  b log10(I)  b log10(I0)where  E  Eeq

cases 1 to 3, to determine Ecorr and Icorr It is also possible to solvethese problems mathematically, as illustrated in the following trans-formations

The applied potential is

E  Eeq b log10(I)  b log10(I0)and the applied current can then be written as

log10(I)   log10(Io)   log10(I0)or

I  10[(E  Eeq)/b  log10(I

For a corroding metal, one can assume that Eeq  E0.

-0.3 -0.2 -0.1 0

Log (I(A))

2H + + 2e -→ H 2

Fe → Fe 2+ + 2 e

-E corr & I corr

Figure 1.18 The polarization curve corresponding to iron in a pH 0 solution at 25°C (Fig 1.17).

1. 3. 3 Graphical presentation of kinetic data

Electrode kinetic data are typically presented... obtainedwith the help of Eqs (1. 14) and (1. 17), respectively, which generallystate how the overpotential varies with current, as in the followingequation:

 b log10 (I/I0)... penetration rate of 1. 2 cm peryear, it is meaningless, in corrosion studies, to consider current densi-

ty values higher than 10 mAcm2or 10 2A