Materials and Structures / Matériaux et Constructions, Vol.

37, November 2004, pp 623-643

RILEM TECHNICAL COMMITTEES

RILEM TC 154-EMC: ‘Electrochemical Techniques for Measuring Metallic Corrosion’

Recommendations

Test methods for on-site corrosion rate measurement of

steel reinforcement in concrete by means of the
polarization resistance method
Prepared by C. Andrade and C. Alonso with contributions from J. Gulikers, R. Polder,
R. Cigna, Ø. Vennesland, M. Salta, A. Raharinaivo and B. Elsener

The text presented hereafter is a draft for general consideration. Comments should be sent to the TC Chairlady: Dr. Carmen
Andrade, CSIC - Instituto de Ciencias de la Construccion "Eduardo Torroja", Serrano Galvache s/n - Aptdo 19.002, 28033
Madrid, Spain; Tel.: +34 1 302 04 40; Fax: +34 1 302 07 00; Email: andrade@ietcc.csic.es, by 30 May 2005.

TC Membership – Chairlady: C. Andrade, Spain; Secretary: B. Elsener, Switzerland/Italy; Members: C. Alonso, Spain;
R. Cigna, Italy; J. Galland, France; J. Gulikers, The Netherlands; U. Nürnberger, Germany; R. Polder, The Netherlands; V.
Pollet, Belgium; M. Salta, Portugal; ‡. Vennesland, Norway; R. Weydert, Germany/Luxemburg; Corresponding
members: C. Page, UK; C. Stevenson, South Africa.

1. SCOPE 2. SIGNIFICANCE AND USE

This Recommendation covers the description of non- The test methods described in the present
destructive electrochemical test methods for the estimation Recommendation are suitable for on-site condition
in large size concrete structures of the instantaneous assessment of steel reinforced concrete structures.
corrosion current density, icorr, expressed in PA/cm2, by The methods can be applied regardless of the thickness
means of the so-called Polarization Resistance technique, of concrete cover and the rebar size or detailing. However,
Rp, in order to assess the condition of embedded steel when the bars are electrically connected, only the corrosion
reinforcement related to its corrosion. of the closer layer of reinforcements facing the counter
The values of icorr, can be used to assess the rate of electrode, CE, can be measured. This closer reinforcement
degradation of concrete structures affected by reinforcement layer practically shields the penetration of the polarizing
corrosion. However, they cannot give information on the current to deeper lying reinforcements. When rebars are
actual loss in steel cross section which, at present, only can electrically isolated, the steel bar connected to the
be assessed by means of direct visual observation. instrument will be measured irrespective of its depth
Values of the free corrosion potential or half-cell (depths higher than 1 m have been tested).
potential, Ecorr [V], of the embedded reinforcing steel and of The test methods can be used at any time during the service
the electrical concrete resistance, Re [:], are obtained as life of the structure, and in any kind of climate, providing the
preliminary steps of the Rp measurements. Values of the temperature is higher than 0ºC. A very dry concrete surface of
concrete resistivity, U [:m], can be calculated from Re U > 1000 :m makes the measurement difficult. Some pre-
values providing the geometrical arrangement of the wetting of the concrete surface is always necessary.
electrodes enables this calculation. The Icorr results obtained by the test methods can be used
Both parameters, Ecorr and Re (or U) may be used to for one of the following purposes:
complement the reliability of the icorr measurements.

1359-5997/04 © RILEM 623

RILEM TC 154-EMC

1) To assess the present corrosion condition of the 3.4 Instantaneous corrosion current density,
reinforcement, that is, to discriminate between corroding icorr
and non-corroding (passivated) zones.
2) To evaluate the effectivity of a repair work. The instantaneous corrosion current density, icorr, is
3) To calculate the loss in rebar cross section by means of obtained by dividing a constant, B, by the Rp value [2]:
integration of Icorr during the propagation period (providing
the initiation time is known). This calculation enables the B
Icorr values to be implemented into structural models in icorr (2)
order to assess the further development of the structural Rp
performance with respect to cover cracking, loss of bond
and loss of load-bearing capacity. where Rp is expressed in :.cm2 and B in V is a constant
The corrosion current values in addition to the resulting from a combination of the anodic and cathodic
measurements of the corrosion potential, Ecorr and of the Tafel slopes:
concrete resistivity, U, may be complemented by other data
from the concrete: rebar diameter, chloride profile, depth of ba • bc
B (3)
carbonation, porosity, temperature, cover thickness, exposure 2.303˜(b a  b c )
conditions, crack pattern, etc, in order to help in the evaluation
and prediction of future performance of the structure. Its value for steel reinforcement embedded in concrete
The corrosion current values must be interpreted by has been determined by means of calibration against mass
specialists or skilled engineers experienced in the field of loss measurements [1]. For on-site measurements the
corrosion testing and structural evaluation. recommended value of B= 26mV.
The standard units of icorr are PA/cm2 (also mA/m2 is
currently used).
3. DEFINITIONS
Non uniform corrosion current, Icorr
A non uniform corrosion current, Icorr, (written in capital I
3.1 Polarization Resistance, Rp instead of i) may also be obtained through expression (2) [1,
The Polarization Resistance, Rp, of a reinforcement 5]. It is used when corrosion attack is expected to be localized.
embedded in concrete is defined as the ratio between A corrosion current density, icorr, can only be properly
defined when the attack is uniform. In consequence, in the
applied voltage '( (shift in potential from Ecorr) and the
case localized corrosion prevails, Icorr is preferred instead of
step of current 'I, when the metal is slightly polarized
icorr in order to emphasize the non-uniformity of the
(about 20-50mV) from its free corrosion potential, Ecorr, [1].
corrosion process.
It can also be defined as the slope of the potential-current
Due to the impossibility of direct visual observation of the
polarization curve at the corrosion potential, Ecorr [2, 3]:
morphology of corrosion, Icorr is the most feasible expression
of the corrosion current when measuring in concrete.
§ 'E ·
Rp ¨ ¸ (1)
© ' i ¹ 'E o 0 3.5 Corrosion rate, Vcorr
The corrosion rate or corrosion velocity, Vcorr, represents
where i (current density) = 'I /S, with S being the steel area
the volumetric loss of metal by unit of area and unit of time.
polarized. The dimensions of Rp are :.cm2 or : depending In the present Recommendation it is expressed in mm/year,
on whether the polarized area is taken into account or not. although other units may also be used.
3.2 True Polarization Resistance, Rp,true Vcorr expressed in mm/year is obtained from the
corrosion current, (either icorr or Icorr) in PA/cm2 through
The True Polarization Resistance, Rp,true, [:.cm2 ] [4] is Faraday’s law and the density of the metal. For the steel, 1
defined for large structures as the ratio 'E/'I multiplied by PA/cm2 is equivalent to a corrosion rate of 0.0116 mm/year
a reliable estimation of the steel area (smaller than the total for uniform attack.
area of the reinforcement) effectively polarized by the
current. It can be obtained by means of the methods Vcorr (mm/y)= 0.0116 icorr (PA/cm2) (4)
described in chapter 5 of the present Recommendation.
4. GENERAL CONSIDERATIONS ON Rp
3.3 Apparent Polarization Resistance, Rp,app
TECHNIQUE RELEVANT FOR ON-SITE
The Apparent Polarization Resistance, Rp,app, [4] is termed MEASUREMENTS
the value obtained when a non-accurate determination of the
metallic surface polarized by the current is made. This occurs
when: 1) when the ratio 'E/'I is not multiplied by any area, or 4.1 Basic aspects
2) when the area taken into account is that of the counter
The calculation of the corrosion rate from Polarization
electrode. Consequently, Rp,app results into an erroneous
Resistance, Rp, technique is well established [2, 3, 6-9]. The
estimation of Icorr. In order to avoid misinterpretations the
application to the measurement of the corrosion of steel
surface area used for calculating Rp, should always be
reinforcement started about 1973, [1, 5, 10-23] and the
mentioned in the data presentation report.
agreement between gravimetrically determined weight loss

624
Materials and Structures / Matériaux et Constructions, Vol. 37, November 2004

4.1.1 Compensation of IR drop

Current (mA)
The relatively high electrical resistivities of the concrete
result into too low Icorr values, if compensation of IR drop is
neglected or not adequately performed [1, 5, 10]. This is
due to the fact that the Rp calculated is the sum of the
-10 10 resistance associated with the actual corrosion process and
the resistance associated with the electrolyte resistance
(concrete).
Potential (mV)
Rp (calculated) = Rp (corrosion) + Re (electrical resistance)
The potentiostats to be used for on-site measurements
have to be able to calculate the ohmic drop, Re, or to
compensate for its influence during the recording of the Rp
measurement. Currently, this compensation can be
Fig. 1 - Linear plot of the polarization curve around Ecorr in the automatically made by means of two methods: positive
anodic direction. feed-back, electronically made by the potentiostat, and
and the electrochemical measurements has been largely current interruption, usually as well performed
demonstrated [1, 5, 10]. electronically. Both may be equally accurate if made
The measurement is made by a three-electrode correctly.
arrangement using the reinforcing steel as Working 4.1.2 Range of linearity
Electrode (WE) as Fig. 1 shows. A Counter Electrode, CE, The range of linearity of the current-voltage curve
at least of equal size of the rebar and of a material well (Fig. 1) of steel rebars embedded in concrete has been
dispersing the polarizing current, and a Reference verified [1, 5] for potential ranges around 20-30 mV [1, 15]
Electrode, RE, serve to apply an electrical signal inducing a of Ecorr. This relationship is linear for ranges of even
shift of about 20 mV from the corrosion potential, either in 100 mV in the case of very high corrosion rates.
the anodic or in the cathodic direction. However, the range of linearity may be smaller than
The Rp value in :.cm2 is calculated from the ratio 'E/'I 20 mV in some conditions of extremely low corrosion rates
multiplied by the exposed metallic area (expression (2)). and for galvanized rebars when they are passive or normal
Rp values can be also obtained from the analysis of the steel in presence of certain inhibitors.
transitory period (transient or coulostatic methods) [24-29].
These methods, which will be described in paragraph 5.3, 4.1.3 Polarization time
The waiting time or the sweep rate needed to achieve the
call for very high speed potentiostats with precise
correct conditions for obtaining a reliable Rp measurement
measurement of the ohmic drop, RI, and they require the
varies for passive and active states. For most of the
assumption of an appropriate equivalent electrical model. In
measurements the optimum conditions, [1, 5, 24] are
specimens with uniform distribution of the applied current,
achieved by means of (Fig. 2):
the Randles model (see Fig. 18) is used for the analysis of
a) Waiting times between 15 (corroding) to 60 (passive)
transitory periods, however in on-site measurements this
seconds in potentiostatic modes of operation and
analogue model does not properly represent the
between 30 to 100 s in galvanostatic ones.
phenomenon and therefore, the analysis based on the
b) Sweep rates between 2.5-10 mV/min in
Randles model of transitory periods recorded in large
potentiodynamic modes of operation.
structures, yields unreliable results.
The Polarization Resistance may also be obtained from
Electrochemical Impedance techniques, [30-39]. It is not
the aim of the present Recommendation to describe the
procedure because its application to on-site measurements
proves to be very complex. This is due to the fact that the
polarized steel area changes with the frequency and in
consequence, the calculation of Rp, true is not feasible. The
model for its interpretation needs further development.
In addition, for the correct measurement of Rp of metals
embedded in concrete the following aspects need to be
considered:
1) compensation of IR drop,
2) verification of the range of linearity around Ecorr,
3) optimization of the response time by means of a
sufficiently long waiting period or proper sweep rate,
4) localized character of the corrosion attack,
5) distinction between galvanic current and corrosion rate
6) comparison with the corresponding gravimetric losses
and Fig. 2 - Values of Icorr obtained at different polarization times
7) range of Ecorr values in which Rp determination of (upper scale) or sweep rates (bottom scale). The range indicated in
passive steel is reliable. the window refers to the optimum conditions.

625
RILEM TC 154-EMC

TOTAL EXPOSED AREA

Icorr = CURRENT CALCULATED FROM Rp
TOTAL EXPOSED AREA

LIKELY CORRODING negligeable contribution to the

CATHODIC AREA corrosion due to the distance to the pit Fig. 4 - Distinction between "corrosion rate" and "local attack
penetration". Difference between maximum pit depth (Ppit) or
maximum attack penetration and the averaged corrosion (Px):
Ppit= D·Px

corrosion rate at the pitting site will be the result of

multiplying the averaged Icorr by D (see Equation (12))
TOTAL EXPOSED AREA [41]:

Fig. 3 - Localized attack: Relative error in Icorr due to sample area. In Ipit= Icorr · D (5)
the case of localized attack the relative error in determination of Icorr is In consequence, using Equation (13) the maximum pit
smaller, as smaller is the sample size.
depth will be :

4.1.4 Localized corrosion Ppit = Px ˜ D (6)

When the value of the steel area introduced in expression
(2) is that of the total area polarized by the applied current, 4.1.5 Distinction between galvanic current and corrosion
the corrosion current calculated implicitly assumes a rate
uniform distribution and consequently it represents an In a corroding zone, both anodic and cathodic areas
average corrosion current density, icorr [1, 5]. develop simultaneously. That is, the areas rusting in a
When the corrosion is localized in only a small zone of corroding reinforcement are not pure anodes, but they have
the metallic surface tested, the corrosion current calculated microcell activitiy. In addition, a macrocell may be present
through (2) does not represent a uniform density, but the between the rusting and the clean (cathodes) steel areas [43]
corrosion current refers to the total area composed by the (Fig. 5).
corroding zone plus the non-corroding part contributing
(cathode) or not (rest of area) to the corrosion current
recorded, as Fig. 3 (upper part) depicts. That is why, the
corrosion current is frequently referred to as Icorr instead of
icorr, indicating a probability that the corrosion morphology
is not uniform.
The smaller the metallic area tested, the higher the
reliability of the Rp value in the case of localized corrosion.
The first recommendation in order to minimize this error
is the use of small size specimens [1, 5] which in on-site Fig. 5 - Anodic sites SA present microcell activity in addition
measurements means the need to polarize small regions of to the macrocell formed with the adjacent non-corroding
the infinite rebar (to confine the current). Another zones, SC. SA  SC S (total area). Galvanic current may be
possibility is to refer the measurement, not to the total steel only a fraction of the total corrosion current.
area, but to the area visually identified to suffer from
corrosion (apparent anodic zone) [40, 61]. However, this Therefore the galvanic current, Igalv, flowing between
procedure implies the breaking of the concrete cover for corroding and passive zones represents only a part of the total
visual observation or the measurement of this area only at corrosion activity. The galvanic current, only equates Icorr
the end, area which obviously will be larger than at when the corroding part constitutes a pure anode with no
intermediate steps. microcell activity. This situation might only occur in
In order to avoid the tedious measurement of the exact concrete when a completely oxygen-free atmosphere exists
surface area actually affected by corrosion, statistical around the corroding areas, or when the corroding area is
studies have been made for several metals in order to comparatively very small (small pits) [7, 8].
identify the ratio between general and localized corrosion. The fact that microcell activity always may exist, even in
That is, the ratio between the averaged penetration depth small pits, prevents a direct correlation to be made between
and the maximum pit depth, Px/Ppit (Fig. 4). In the case of the corrosion rate and galvanic current. That is Icorr t Igalv.
reinforced concrete structures [41] a ratio of 10 between Px The ratio between them will depend on variables such as
and Ppit is recommended to be used as a conservative limit, moisture content (concrete resistivity), size of corroding
i.e. this is considered a maximum value. This implies the area and the corrosion rate itself [44].
multiplication of Icorr values derived from Rp measurements It is worth noting that the corrosion rate is expressed
by a pitting factor, D, having a value of 10, when localized referred to the whole (SA + SC) surface area (Fig. 5), while
attack is produced (chloride-induced corrosion). Thus the galvanic current Igalv is usually referred only to the
corroding (SA = anodic) zone [43, 44].

626
Materials and Structures / Matériaux et Constructions, Vol. 37, November 2004

4.1.6 Comparison with gravimetrically determined weight values, although in large structures the determination of the
losses gravimetric loss is only feasible if the reinforcement bar is
The metallic mass loss after suffering corrosion is carefully cut in small portions.
measured by the difference in weight of the metal cleaned
4.1.7 Range of Ecorr for reliable Rp determination in
of any oxide before been submitted to the corrosion process
passivated steel
and after being removed from the concrete. This
It has been recently realized [45-48] that Rp values cannot
gravimetric loss has to be the same than that calculated
be correctly determinated if the steel is passive and exhibits
through Faraday’s law from the integration of the Icorr-time
Ecorr values more cathodic than around -300 mV (SCE) due
curve [1, 5, 10]. The instantaneous values of the Icorr are
to a restricted access of oxygen, as it may happen in
represented along the time as shown in Fig. 6. for mortars
submerged structures. In these cases, the Rp values measured
with several amounts of chlorides added in the mix. The
may be too low (too high corrosion) and in consequence, to
integration of this type of plot is what is compared with the
mislead the correct determination of Icorr. The most likely
gravimetric losses.
explanation is that at these cathodic potentials in passive
The graph showing the comparison between
state, the measurement of Rp seems to give the current
gravimetrically determined and electrochemically derived
metal losses is depicted by Fig. 7. The results must lie exchange of the redox process (Fe (II) l Fe(III)) in the
around the line of equality with a maximum error factor of passive layer. This fact has not been so clearly noticed in
two (3). This error factor is depicted by the two lines lying actively corroding steel because it is likely to be masked by
parallel to the line of equivalence in the figure. This type of the relatively higher importance of the faradaic process in
checking is essential to validate any measurement of Icorr comparison to the redox one.
As a practical consequence, the values of Rp recorded on
passive reinforcements or where a lack of oxygen is
10 suspected (as may be in concrete being very saturated by
2.12% total Cl 3.12% total Cl
0.48% total Cl 1.19% total Cl water or chlorides), have to be carefully interpreted by
specialists.
4.2 Particularities of on-site measurements:
Icorr (uA/cm^2)

1
3.12 determination of the polarised area or current
1.19 confinement
2.12
0.1 Beside the basic aspects of the measurement of Rp
aforementioned, the main characteristic of real size
structures regarding corrosion monitoring, is the quasi-
0.48 infinite length of the rebar. This fact calls for using well-
0.01 defined methodologies, not affected by the rebar size,
0 20 40 60 80 100 which enable the calculation of the actual metallic surface
Time (days) being effectively polarized during the measurement.
In small specimens the uniform distribution of the
current applied between auxiliary and working electrodes is
Fig. 6 - Examples of Icorr- time plots of rebars embedded in usually guaranteed. However in large structures, the
mortar with different proportions of chlorides added in the mix. auxiliary electrode is much smaller in size than the working
electrode (the rebar). This situation gives rise to an
essentially non-uniform distribution of the applied current
along the reinforcement as shown in Fig. 8 [4]. The
electrical signal tends to vanish with the distance from the

Fig. 8 - Lateral spreading of the current when applied through

a small counter electrode.

Fig. 7 - Comparison between gravimetrically determined

losses and electrochemical ones (obtained from the
integration of the Icorr – time plots). The dotted parallel lines Fig. 9 - The length of the rebar polarized to a significant level
delimitate the range of accuracy obtainable (a factor of two by the externally applied current is termed the “critical length”,
times the actual value). Lcrit. RE = Reference electrode, CE = counter electrode.

627
RILEM TC 154-EMC

counter electrode, CE. The required uniform distribution is, values have been confirmed in experiments made in large
therefore, not met and the 'E/'I slope cannot be referred to slabs. In the case of non-corroding rebars the error factor
any specific rebar surface. easily exceeds a value of 100, whereas for corroding rebars,
In consequence, either the so-called critical length, Lcrit, it may be about 10 or smaller. The ratio Rp/Rp,app will
(see Fig. 9) reached by the electrical field has to be depend on the value of the particular Rp itself and of the
calculated or the current has to be confined within a well- concrete resistivity (level of concrete moisture content).
defined delimited area. A general or fixed value of this error factor cannot be
given due to the variability of moisture conditions and
4.2.1 Critical length degree of chloride contamination of the concrete which
The critical length, Lcrit, [4, 49] is named the distance affect both Re and Rp. That is, it does not produce a single
reached by approximately 90% of the current applied by value for the error factor relating Rp, true and Rp, app values.
means of a small auxiliary electrode placed on the surface
of the concrete as Fig. 9 depicts. It is the distance reached 4.2.3 Confinement of the polarizing current
by the current relative to the border of the external auxiliary Another way of delimitation of the area actually
electrode. polarized by the applied current consists in using an
This critical length is a function of the square root of the external circular counter electrode concentric to the CE
ratio Rp/Re [25] and independent of the size of the counter termed“guard ring”, GR [49-51]. This arrangement enables
electrode. It is not a fixed length, L, value. It can be the current applied through the central CE to be confined to
calculated by means of the transmission line model [4] a defined area under the central CE and between it and the
(Fig. 10). external GR, providing the polarization achieved by this
GR is adjusted precisely to counterbalance the central
electrical field (Fig. 12). If the counter-field from the GR is
not correctly controlled and adjusted, the area being
polarized may greatly differ (being smaller or larger) from
the predetermined one. This control is achieved by means
of two additional reference electrodes, S1 and S2, located
between CE and GR as will be commented in paragraph
5.1. Fig. 13 shows an example of a lack of correct
confinement due to the absence of the electrical field
Fig. 10 - Transmission line model (electrical analogue)
representing the lateral distribution of the current along the controllers, S1 and S2
reinforcement bar (see Figs. 8 and 9).

The calculation or measurement of this Lcrit enables to

obtain the Rp, true, because it enables the calculation of the
polarized area and therefore, the Rp, true can be accurately
expressed in :cm2.
4.2.2 Apparent Rp error factor
If the ratio 'E/'I is obtained without referring to Lcrit but,
for instance, is calculated only from the CE area, an Rp,app is
obtained, [4, 49] which results in serious errors if used for
quantitative calculations, because large differences between
Rp, true and this Rp, app may exist.
Thus, in Fig. 11 the error factor (Rp/Rp, app) is shown. The

Fig. 12 - Theoretical confinement of the applied current introduced

by the central counter electrode, when an external guard ring is
used to produce another counterbalancing electrical field.

Fig. 11 - Error factor Rp, true/Rp, app as a function of the area of the
counter electrode when this area is used for calculating Rp,app.
Fig. 13 - Incorrect confinement of the current when the guard ring
does not have an independent control or modulation.

628
Materials and Structures / Matériaux et Constructions, Vol. 37, November 2004

4.2.4 Detection of localized corrosion c) The use of a pair of large size auxiliary electrodes,
When the corrosion is very localized, an error in the which minimize the effect of the critical length. This
calculation is produced which is proportional to the relative method has the inconvenience that, as the counter
size ratio of the corroding area to the total area (see electrodes have to be large, the method is not
paragraph 4.1.4). The larger the total polarized area, the sufficiently sensitive to localized attack. Therefore, this
higher will be the relative error [40, 51]. method will not be described in the present
In the case of on-site measurements, the locally Recommendation.
corroding spots change the current lines from the central
CE resulting into a non uniform distribution as Fig. 14 5.1 Modulated confinement of the applied
depicts. This effect invalidates not only the calculation of current (guard ring) method
Icorr but also the assessment of the location of the pits. The
This method [4, 49] is based on the calculation of the Rp,true
non-confining techniques are not able to correctly localize
using effective confinement of the polarizing current within a
the isolated corroding zones, because they draw the applied
specific area through a second circular counter electrode This
current tens of cm away from the CE. On the opposite, the
so-called “guard ring” is modulated by means of two twin
technique using a modulated guard ring enables the correct
reference electrodes, S1 and S2, located between the central CE
identification of the localized corrosion spots as was shown
and the external GR (Fig. 15) in order to achieve the required
in Fig. 12.
counterbalancing electrical field. These twin electrodes
permanently control the external ring by means of detecting
the current lines coming from the CE in order to adjust them
within the predetermined area of diameter D. The method then
promotes an electrical delimitation of the area instead of
determining it.
Therefore, the auxiliary electrode for confined
measurement consists of (Fig. 15):
a) a central small counter CE,
b) a ring-shaped counter electrode, GR, which surrounds
the central one,
c) a main reference electrode, RE, placed in the center of
the central CE and
d) two auxiliary reference electrodes, S1 and S2, placed
between the central CE and the guard ring.
The calculation of the Rp,true can then be made directly
from the 'E/'I ratio multiplied by the steel surface area
below the central counter electrode until a point halfway the
Fig. 14 - Effect of localized corrosion spots on the current lines two CE (central and ring), as shown in Fig. 15.
between the counter electrode and the reinforcement showing
As was mentioned before, the method of modulated
localised corrosion attack.
confinement of the current is the most suitable for cases of
localized attack, because it delimitates the area polarized and
therefore, it is able to reduce the inherent error due to the
5. METHODS FOR ON-SITE relative area sizes (Fig. 3). Moreover it is the only method
MEASUREMENT OF THE
POLARIZATION RESISTANCE
In order to solve the uncertainty of the rebar surface
really polarized by the applied current, the methods of
measurement developed until now are based on the
following principles:
a) The confinement of the applied current to a
predetermined area by means of a "modulated"
guard ring (second circular counter electrode
applying a modulated counter current).
b) The calculation of the critical length reached by the
current when applied from a small auxiliary
electrode placed on the surface of the concrete.
This calculation might be made through the
following methods: (1) Multiple electrode or
potential attenuation method and (2) Galvanostatic
pulse or transient analysis. This last one presents Fig. 15 - Arrangement of electrodes in the modulated confinement of the
the inconvenience of the need to use several applied current. The reference electrodes S1 and S2 are used to modulate
electrical analogue models for being accurate. the outer guard ring in order to effectively confine the current to the
circular area indicated by the dotted line.

629
RILEM TC 154-EMC

which is able to minimize the effect of macrocells or to a Reference Electrode at its centre. The use of a non-
notice active/passive region transition. modulated guard ring has been tried as well [28]. A current
pulse of several tens of PA to several mA is applied by means
5.2 Multiple electrode or potential attenuation of the counter electrode, and the potential-time response is
method recorded (Fig. 17).
This method [54] is based on the calculation of the For the calculation of the Rp an “analogue or electrical
critical length, Lcrit, which allows calculation of the steel model” is used which simulates the steel/electrolyte interface
area really polarized by the applied current by means of the (double layer) through a capacitance and the faradaic corrosion
transmission line model. (weight loss) through a resistance. The model used is the well
The method makes use of a small counter electrode, CE known Randles circuit shown in Fig. 18.where the capacitor,
which is placed on the concrete surface, as depicted in
Fig. 16. In the center of this CE electrode the main
Reference electrode, RE, is placed which measures the
local corrosion potential Ecorr, of the reinforcing steel. Other

Fig. 17 - Response of potential to a galvanostatic pulse.

Fig. 16 - Electrode arrangement of the potential attenuation
method.

auxiliary reference electrodes (R1, R2,and R3) are located

aligned with the CE and the main RE as Fig. 16 shows.
When the current is applied through the CE, the rebar is
polarized until a certain distance (Lcrit) as shown in Fig. 9.
The auxiliary reference electrodes R1, R2 and R3 serve to
measure this Lcrit by recording the electrical potential at
various distances from CE (potential attenuation) before
Fig. 18 - Randles circuit or electrical analogue model of
and after the application of the current. The position where metal/electrolyte interference. The circuit is too simple for
this potential difference is smaller than approximately 10% modelling the dispersion of the applied current in real size
of the potential difference calculated from RE indicates Lcrit structures.
and therefore, allows to calculate the geometrical rebar area
really polarized by the signal. C, represents the double layer and the resistor, Rt, the
As soon as the polarized area is known and the polarization resistance, that is the resistance to corrosion or to
geometrical characteristics of the concrete element are the loss of metal integrity.
determined, the corrosion current, Icorr, can be calculated. The Rp-value can be calculated by solving the circuit of
This method is the only one able to correctly determine a Fig. 18 resulting [24] into:
Icorr in submerged structures. The modulated guard ring
ǻE
method cannot be applied due to the low resistivity in these
conditions. ǻI
 
Rp 1  e  t / RpC  Re (7)

5.3 Galvanostatic pulse or transient analysis This technique seems very attractive as it is very quick
methods (it tries to measure during the transitory period), however it
has several problems, which are mentioned below:
Several procedures [25, 27, 50-53] have made use of the
a) to accurately record the response in so short times is very
recording of the transient response after the application of a
difficult as very quick response potentiostats are needed,
galvanostatic pulse. The method uses the typical arrangement
which have the inconvenience of not being able to
of unconfining techniques of a single counter electrode having
measure the very low currents recorded in passive state,

630
Materials and Structures / Matériaux et Constructions, Vol. 37, November 2004

b) to accurately measure the ohmic drop, because it may sensitivity to detect variations of 0.5mV in a potential range
represent an important proportion of the potential during from -1.0V to +1.0V. The current circuit must have a
the transitory period, sensitivity of at least 0.05PA in the full range between 0.05
c) the need to assume a value of the capacitance for to 104PA.
calculating Rp, which in the case of concrete is not a fixed The potential or the current is applied, resulting into:
value and even pseudo-capacitances may develop, and a) a sweep potentiodynamic polarization around Ecorr.
d) that unconfined or inadequately modulated techniques are b) a stepwise polarization using single small potential or
unable to precisely locate the corroding spots and pits current steps.
(Fig. 14). Potentiodynamic or galvanostatic measurements allow
Due to these difficulties, up to now pulse techniques for the calculation of the slope 'E/'I, which has to be
have not succeeded in measuring accurately in large multiplied by the rebar surface effectively polarized by the
structures. The use of a guard ring has not improved the electrical signal in order to obtain the True Rp in ohm cm2.
results as it presents the additional inconveniences The procedure was explained in chapter 5 for the different
described in paragraphs 4.2.3 and 4.2.4 (Figs 13 and 14) measurement methods.
related to the difficulty of modulating the guard ring during Concerning the optimum polarization time, corroding
the transitory period. rebars need shorter polarization periods while passive steel
needs longer waiting times to reach a stable value. In
5.4 Large size auxiliary electrode potentiostatic measurements, waiting times of 15-60
Other procedures to measure on site are based on the use seconds are enough to achieve the quasi-stationary regime
of unguarded techniques with a single relatively large necessary to obtain the 'I value to be used in the expression
central counter electrode and its surface is used as a 'E/'I= Rp. This ratio also has to be multiplied by the
reference area for the corrosion measured (multiplying the metallic area really/effectively polarized. In galvanostatic
Rp by the area of the counter) [54]. measurements, longer waiting times may be necessary to
As the critical length Lcrit, is not dependent on the size of record a quasi-stationary 'E value: periods of 30-100
the counter electrode, one method of measurement might be seconds are the most appropriate. As has been said
the use of large size electrodes in order to make the (paragraph 4.1.3), the shorter waiting times are
polarized area below them, comparatively bigger than that recommended for an active corrosion state, while the longer
corresponding to the critical length. Thus, for passive times apply for passive rebars.
conditions, auxiliary electrodes of areas t 1000 cm2 may
minimize the error of not accounting for the polarized area 6.1.2 Auxiliary sensor (holding RE, CE, GR)
encountered in the critical length, while 500 cm2 may be The apparatus used has to be equipped with a sensor
enough in the case of active corrosion. The Rp is then (electrode), containing all the counter and references
multiplied by the area of the counter electrode which should electrodes. Although the counter electrodes can be of any
be mentioned in the measurement report. metallic material able to produce the current, stainless steel
However, the use of large auxiliary counter electrodes has is the most cost-effective material enabling an easy
the important inconvenience of not being able to detect maintainability.
localized corrosion (Fig. 14) at all. However, this method may The reference electrodes can also be any of the
be a practical way of measuring in very wet conditions, where traditional ones: calomel, silver/silver chloride or
a uniform corrosion morphology might be expected and other copper/copper sulphate. This last one has demonstrated to
methods may fail or give erroneous results. be one of the most suitable due to its measurement range,
accuracy and precision, although it has to be maintained
and its junction membrane cleaned periodically.
6. PRACTICAL EXECUTION OF
MEASUREMENTS 6.1.3 Contact sponge between auxiliary sensor and
concrete surface
In order to provide a low electrical resistance path
6.1 Apparatus between the sensor and the concrete surface, a wet sponge
or any other conductive substance or cloth has to be used.
6.1.1 Electronic equipment The sponge and the concrete surface have to be prewetted
The equipment to be used with any of the methods with water or any other low electrical resistance contact
described in chapter 5 has to be based on a potentiostat or solution or gel. The concrete prewetting should be intensive
galvanostat as a means of controlling and measuring enough to establish a good electrolytic contact (see
potential and current. paragraph 6.2.1.)
The potential measuring circuit of the instrument should
be able to maintain an electrode potential within 1mV over 6.2 Calibration and Standardization
a wide range of currents and should have a high input
impedance, i.e. higher than 10M:, in order to minimize 6.2.1 Equipment
current drawn from the corroding system during The calibration of the portable potentiostat/galvanostat
measurements. can be made in a similar way to non-portable ones. The
The current circuit should have a sensitivity such that the equipment can be checked by means of carrying out
Icorr could be determined at least of the order of measurements with a dummy cell. This electrical analogue
0.05PA/cm2. Furthermore the instrument should have a cell can be fabricated by using a Randles circuit of the type

631
RILEM TC 154-EMC

of Fig. 18. Orientative values of the elements of the circuit electrode, making the electrolytic connection by means of a
are: Re = 10k:, for wet concrete and around 50-100k: for wet sponge. Carry out a Rp measurement with a normal
dry concrete, Rp=5k: for representing an active state and potentiostat and calculate the 'E/'I ratio of this central
Rp= 100k: for reproducing a passive state of the steel The rebar. Multiply the “ohms” so obtained by the area of the
capacitance representing the steel may be C= 50PF. working electrode (area= SDL, where D= rebar diameter
Also a transmission line model of the type reproduced in and L= embedded length of the rebar) in order to obtain Rp
Fig. 10 can be used. The values of the elements can be the in ohm˜cm2.
same as those for the Randles circuit. Repeat the measurements placing the reference electrode
In all cases the potentiostat has to give the values of the at two additional points along the rebar and calculate the
circuit within the range of error given by the electrical arithmetic mean of the three results. Give this value as the
elements used to build the dummy cell. reference one: Rp, ref.
Now, make connection through an external wire between
6.2.2 Reference electrodes all the rebars in order to reproduce reality. Take the
The reference electrodes of the auxiliary sensor used as instrument to be calibrated and place the Auxiliary sensor
counter electrode have to be checked periodically in order on the concrete surface, providing electrolytic adequate
to detect any improper functioning due to drying of the interfacial contact has been achieved. Make the connection
porous plug or leakage of the solution. This checking is to one of the rebars providing the whole mat is connected.
performed by comparing their corrosion potential values Carry out the Rp measurement. Give the value in ohm·cm2
with those obtained at the same location with other following the instructions of the equipment in order to
independent electrodes of the same nature. consider the area really polarized by the current.
In the case of the slabs contaminated with chlorides the
6.3 Standard reference measurements value of the Rp obtained with the instrument has to be
The whole apparatus has to be, at least initially, between double and half of Rp, ref:
standardized by comparing its results with those obtained 0.5 Rp, ref < Rp < 2 Rp, ref (8)
by means of a normal potentiostat in reference concrete
slabs (under laboratory conditions). In the case of the passive rebars the Rp obtained through
the on-site instrument should fulfil the requirement (8) or
Procedure give Icorr- values lower than 0.2 PA/cm2.
In order to make this standardization the following
procedure is recommended: fabricate two concrete slabs
6.4 Precision and bias
(Fig. 19), one without admixtures and the other with an The repeatability of the Icorr results, requires more than
addition of 3% CaCl2 by weight of cement in the mixing 10 measurements taken at the same location and
water. The minimum size of the slab without admixtures comparable environmental conditions should be such that
should be of 1.20x1.20m2 by 10cm in thickness. The 90% of the measurements should not differ not more than
minimum size of the chloride contaminated one should be four times the minimum value recorded. For instance, a
of 0.50x0.50m2 by 5-10cm in thickness. A size of measurement can vary between 0.25 PA/cm2 and 1 PA/cm2
150x150x15cm3 is recommended for all cases. without considering the result erroneous. This margin
Then, 5 to 10 isolated rebars have to be embedded as is appears as a consequence of the factor of two inherent to
shown in Fig. 19. Means to make an external electrical the Polarization Resistance method [3].
connection between rebars have to be provided. Simple However, Icorr values taken on the same rebar but in
different points do not have to fulfil this criterion, because
localized corrosion may give different Icorr values with only
a few centimetres of distance along the rebar.
Deviations from these criteria indicate improper
functioning of the instrument.
Time to recover between successive measurements for a
stable Ecorr potential should be considered. In general, the
recovery of the potential varies from some few seconds to
several minutes.
6.5 Practical procedure
This section contains information on the practical
Fig. 19 - Slab type for making reference measurements for
calibration of portable corrosion-rate-meters. procedure for carrying out the measurements on site, as
well as on the sources of error and the recommended
wires with plugs are a suitable method for the verification frequency (geometrical and in time) of measurements.
of proper electrical conductivity. 6.5.1 Sequence of operations
In order to make the reference measurements the bars
have to be initially unconnected. Take one of the central The sequence of operations necessary for measuring the
bars as working electrode and the adjacent one as counter corrosion rate on-site is:
auxiliary (denoted CE in Fig. 19). Place a reference ƒ Preparatory considerations on concrete conditions due
electrode (denoted RE in the figure) above the working to weather effects

632
Materials and Structures / Matériaux et Constructions, Vol. 37, November 2004

ƒ Selection and identification of measurement location minute to about 5 minutes depending upon the actual
ƒ Identification of rebar location corrosion conditions and the method of measurement. The
ƒ Coordinate system for measurements physical processing of placing the sensor may also take a
ƒ Preparation of the concrete surface by prewetting time of 2-5 minutes. So the operator must consider 5 to 15
ƒ Placement and fixation of the auxiliary sensor at the minutes per location to obtain a corrosion rate
measurement location. measurement.
ƒ Connection between equipment and reinforcement. 6.5.1.3 Identification of rebar location
ƒ Execution of the measurement.
The actual geometry of the rebar arrangement is made by
From these operations, the most important details are
using a steel detector. If needed the bar pattern can be
commented in the following paragraphs.
marked on the concrete surface, as well as the cover depth
6.5.1.1 Preparatory considerations on concrete registered. The bar diameter and their distances are needed
conditions due to weather effects for the calculation of the steel area to be polarized during
Corrosion rate equipment, as other electronic devices, does the measurements
not work in extreme conditions of temperature or humidity. 6.5.1.4 Coordinate system for measurements
The devices should not be at operated at temperatures below
When it is decided to map a zone or element,
2ºC (36ºF), or allowed to get over 50ºC (122ºF). If the
measurements should be taken on a grid. It is recommended
environment exceeds these limits, then the meter has to be
that readings are taken over a rebar, so the grid size is partly
operated from an air conditioned enclosure or vehicle.
dependent on the rebar spacing. A 0.25 m grid spacing is
It should be noted that below freezing, the wetted sponge
recommended, except on small structures or elements with
may freeze giving misleading or unstable readings.
severe changes in exposure conditions. An example of a Icorr
However, if it is still considered essential to collect data
under these conditions then an alcohol solution (10-30% of
alcohol by weight) will reduce the freezing point. It should
also be noted that below 5ºC the corrosion rate may reduce
to low values which may mislead the interpretation (Re
may be very high/Icorr very low).
Complete water saturation of concrete may also lead into
measurement difficulties due to the high conductivity path
that may be established through the concrete surface. This
is particularly clear when deicing salts have been used. The
extremely conductive concrete surface may facilitate the
dispersion of the current lines to very long distance which
makes it difficult to obtain a reliable or reproducible Rp Fig. 20 - Map of corrosion rate values in a slab.
value. That is why measurements in submerged structures
may give very unreliable results, unless the method map is shown in Fig. 20.
described in 5.2 (attenuation of potential) is used. Measurements may also be taken on a simple straight
line, if the corrosion condition is likely to vary with
6.5.1.2 Selection of number and measurement locations
distance along an element.
Before starting the survey it is necessary to select the
6.5.1.5 Preparation of the concrete surface by prewetting
number and location of points where corrosion rates will be
measured. The number of points will depend upon: a) the The concrete surface has to be well wetted prior to
amount of time available, access, and size of structure and, applying the Auxiliary sensor. Care has to be taken to avoid
b) the aim of the inspection (see chapters 6.3 and 8). contamination of the reference electrodes with alkaline
The operator must first consider the time which is substances from the concrete. This is achieved by simply
required to get access to each location, to perform other placing a clean wetted sponge between the sensor and the
relevant measurements and other logistical factors concrete surface.
associated with site work. Obviously experience is On coated or hydrophobically treated surfaces, trials
important in being able to collect the most useful data for a have to be made in order to verify the feasibility of the
reasonable expenditure of time and effort. The previous electrolytic (ionic) connection. Hydrophobic treatments
measurements of chloride concentration, rust staining, may not result into correct contact.
cover, carbonation depths, corrosion potential and There must be complete electrolytic contact between the
resistivity, etc. can also be used as indications for selecting sensor and the concrete surface. Any local deformation or
the most appropriate measurement points. That is, insulating layer must be avoided or removed by grinding or
measurements may be taken at strategic locations chosen choosing another location. Small deformations in the surface
because they represent one or more of the following: can be, if needed, "ironed out" by using additional sponges.
- High or low readings from half cell potentials or In case of excessive superficial chloride contamination
concrete resistivity. or very conductive concrete surface layer a correct
- Locations of structural importance (different elements, measurement may be difficult This can be checked by
construction joints, sources of water or chloride, measuring concrete resistivity. If the values obtained are
ground, water level etc.). below 1k:·cm an unacceptably high conductive concrete
With regard to the duration of each corrosion rate surface can be expected. It is then recommended to clean
measurement, each reading may take from less than 1 the surface from salts and to wash the contact sponge

633
RILEM TC 154-EMC

(6.1.3) very well with uncontaminated water. In very wet the same span or concrete member. Otherwise, the
concrete, it may happen that measurements are very connection must be moved from bar to bar for each
unreliable (too high corrosion rates or improper additional test point. The checking is most accurately done
functioning). In these cases, the method suggested for by exposing two or more rebars across the structure and
measuring in submerged structures (chapter 5.2) or the use measuring the potential difference between them with a
of a large auxiliary counter electrode (chapter 5.4) may be high impedance voltmeter (as used for half cell
the only feasible methods. measurements). If the potential difference is less than 1mV,
Measurements can be performed in cracked concrete. then continuity is likely, if it is greater than 3mV,
However, locations with major voids, delaminations or continuity is unlikely, if it is between 1mV and 3mV,
large cracks (>1mm) within the concrete must be avoided, continuity is uncertain. In addition, the DC resistance has to
in particular if the concrete is wet, because these defects be measured, and then the leads reversed and measurement
may cause the signal to deviate from the required repeated or the AC resistance tested. The resistance should
electrolytic path, resulting in erroneous readings. be less than 1 ohm in all cases.
6.5.1.6 Placement and fixation of the auxiliary sensor in It is essential that there is good electrical continuity
the measurement location between the rebar connection and the steel being measured.
Reasons for discontinuity include construction joints with
The auxiliary sensor must be located preferentially
separated rebar cages, excessive corrosion and light
directly over a rebar of known diameter, either a single bar
reinforcement content. If discontinuities are found then it
or at a crossover. Metallic (electronic) short circuits
may be necessary to make regular rebar connections rather
between this sensor and the bar caused by tie wire, nails etc.
than just one or two.
must be avoided as these will invalidate the reading.
The sponge beneath the auxiliary sensor has to be well 6.5.1.8 Execution the measurement
wetted in order to enable an adequate correct electrolytic A correct connection with the reinforcement may be
contact with the prewetted concrete surface. verified by checking the stability of the corrosion potential,
If the surface is vertical or horizontal overhead, the Ecorr. Its stability is necessary to assure a correct
auxiliary sensor has to be secured with appropriate fixing measurement of the voltage shift after (applying) the
tools (plastic straps, screws or some kind of pressing current. The stability is insufficient if the reading fluctuates
devices). more than 0.5 mV every 5 seconds.
The steel area below the counter electrode or that area After switching the device on, the measurement has to be
which will be polarized during the measurement has then to carried out until completion of the testing time. If repetition
be calculated taken into account whether two or more bars of the measurement is necessary, it has to be taken into
account that the time interval between consecutive
measurements shall be long enough to allow the corrosion
Rebars potential to recover its stability.
Finally, the auxiliary sensor has to be removed and a
new location identified.
Auxiliary
6.6 Source of errors in the practical
Sensor
measurement
Rebar Surface There are several sources of error. Some of them can be
Polarized detected by unsuitability or instability of the Ecorr. The most
frequent can be due to:
a) A lack of correct electrical contact between the
equipment and the reinforcement. In order to avoid this, the
Fig. 21 - The metallic area to be taken into account is that bar has to be well cleaned by removing all rust. A surface
facing the auxiliary sensor. with metallic bright appearance is the suitable finishing to
make proper contact with the plug of the equipment.
intersect in that area (see Fig. 21). b) A lack of correct electrolytic contact between auxiliary
6.5.1.7 Connection between equipment and reinforcement sensor and concrete surface. Several factors may be the
In order to complete the measurement circuit, an reason:
electrical connection has to be made between the equipment - The concrete surface is still too dry. Longer and more
and the reinforcement. A window to the rebar must be efficient wetting has to be tried.
opened by coring, excavation or a connection through an - Contamination of the sponge by salts or by the liquid
exposed steel connected to the reinforcement and through a of the reference electrode. Removal of salts from the
cable attached to the rebar. The rebar must be cleaned (for concrete surface and washing of the sponge are needed. As
instance by brushing with a metallic brush or by means of a well cleaning of the reference electrode membrane.
screw) to ensure a good electrical contact. - A lack of moisture in the reference electrode
In order to avoid opening of several holes to have membrane due to long time of storage. Wetting of the
electrical contact with the reinforcement network, it is membrane is necessary to re-establish the electrolytical
necessary to check the electrical continuity between rebars. connection.
If the reinforcement is electrically continuous then this c) Too wet concrete surfaces. Superficial electrolytical
connection can remain in place for testing at other points on shortcircuit (too conductive concrete surface) is produced.

634
Materials and Structures / Matériaux et Constructions, Vol. 37, November 2004

Cleaning of salts or removing of liquid water on the surface c) The record of the weather conditions at the moment of
is necessary. In extreme cases the concrete has to be the measurements (T, RH and rain). If possible, the weather
considered as in submerged conditions. of the previous week should also be included.
d) The existence of stray currents. It is very difficult to d) The grid space used
eliminate their influence due to the instability induced in e) Groups (lots) if made. Criteria used for identifying the lots
Ecorr. The sensorized confinement method may be the only f) The steel reinforcement area used to calculate the Rp, true
possibility of measuring, preferably by moving the g) The method of measurement used.
auxiliary sensor around to measure with the two confining h) Graph of Icorr-U if recorded
reference electrodes in all directions. i) The Ecorr Re and Icorr results recorded; morphology of
e) Care has to be taken in interpreting which bar layer is attack: pitting or uniform.
corroding. The best is to place the auxiliary sensor nearer to j) Cover depth and carbonation; chloride profile.
the layer expected to be corroding. If the first layer is Concerning the presentation of Icorr values themselves,
corroding, it shields the applied current from arriving at either histograms, simple graphs or iso-Icorr maps can be
deeper positions. If the first layer is passive and the second used. The maps may be coloured by giving a colour to the
corroding, an intermediate value between passivity and four ranges of Table 1.
corrosion will be recorded.
Table 1 - Ranges of corrosion current values
6.7 Frequency of measurements related to the significance in terms of service life
Two types of frequencies have to be considered: the of the reinforcement
geometrical (how many readings for the same structure) Icorr (PA/cm2) Vcorr (mm/y) Corrosion level
and the temporal (whether a single visit is enough or a d 0.1 d 0.001 Negligible
sequence of several is necessary). 0.1 – 0.5 0.001-0.005 Low
0.5 – 1 0.005-0.010 Moderate
6.7.1 Geometrical frequency >1 > 0.010 High
With regard to the geometrical frequency, the structure
has to be subdivided in lots (grouping) considering: a) the The Icorr-U graph (Fig. 25) is also a useful tool for the
aggressiveness of the environment, b) the structural interpretation/evaluation and enables the extrapolation for
typology and c) the aim of the inspection. The groups of predicting the maximum corrosion current, I corr, max.
structural elements will be then treated statistically in order Additionally, a statistical treatment may be made, either for
to obtain a representative value for each element group or, each individual lot or with the whole set of values. Simple
on the opposite, they will be mapped to obtain a pattern of cumulative or differential statistical analysis may help to
the corrosion current. identify corroding and non corroding zones or to establish the
REP
averaged I corr to be used for making predictions of future
6.7.2 Temporal frequency evolution of the damage to help in the assessment. Cross
The temporal frequency will depend on the aim of the relations between Icorr, U or chloride contents at the level of the
inspection. If this is only the identification of the corroding reinforcing steel will also be of benefit.
zones or to evaluate the effectivity of a repair work, a single
visit may be enough. The lack of measurements along the
time is recommended to be balanced by taken a larger 8. INTERPRETATION AND USE OF Icorr
possible number of readings at different locations of the RESULTS
structure (geometrical frequency), selected regarding
several degrees of apparent damage or exposure
aggressivity. 8.1 Definitions
In all cases, plotting of the results in a Icorr-U graph as For a correct interpretation of the Icorr results it is
explained in chapter 8.2.3 will help to evaluate the validity required that some variables will be defined:
of the results obtained.
When the appraisal of the load-bearing capacity of the 8.1.1 Representative value of Icorr or Vcorr
structure is aimed at, then, in order to obtain a sufficiently REP REP
REP A representative value of Icorr or Vcorr, ( I corr , Vcorr ) [55,
accurate representative value of the corrosion rate, I corr , it
is necessary to perform several measurements along, at 59, 60] is defined as an averaged value along a certain
least, a whole year following the seasonal exposure period of time for every single measurement location. The
REP
variations (see paragraph 8.2.5). V corr values may then be implemented into structural
calculation models of the residual load-bearing capacity.
The averaged value can be obtained in two manners [64]:
7. DATA PRESENTATION REPORT a) by calculating the arithmetic mean of Icorr values
registered during a period of time and performed during
Basically, the data presentation report has to contain seasons of practical relevance (n is the total number of
(relating only to the corrosion current): measurements and should be higher than 4).
a) The date of testing
b) Description of the measurement site (location and visual
aspect)

635
RILEM TC 154-EMC

n
P
¦ I corr D pit
(12)
I REP
corr
0
(9) Px
n
8.1.4 Maximum pit depth, Ppit
REP
b) by averaging a value I corr measured on-site, Icorr, single,
The maximum pit depth, Ppit (Fig. 23) represents the
with another one Icorr, lab, obtained in cores drilled from the local maximum loss in bar diameter. It is calculated from
structure and conditioned in the laboratory to a certain the expression (11) multiplied by D:
water saturation degree:
REP
I corr , lab  I corr , sing Ppit Px·D Vcorr · tp ·D Vpit · tp (13)
REP
I corr
(10)
2 Therefore, when the attack is localized, the pitting
corrosion rate Vpit (mm/y) can be calculated by multiplying
8.1.2 Penetration of attack, Px the uniform corrosion rate Vcorr by D.
The penetration of attack, Px [63] (Fig. 21) at a certain In the case of steel embedded in concrete D has been found
moment of the propagation period tp, is the loss in radius of to vary from 3 to 10 [41, 61, 63]. The value of 3 can be used
the bar produced at a certain tp. It is also expressed as for the case of several scattered and coarse pits, while the value
“corrosion penetration damage function” when used for of 10 applies to very localized pitting corrosion. In the case of
further calculation of the structural performance. It can be uniform corrosion D takes a value of 2 to represent the uniform
calculated provided the initiation time is established, from loss in diameter around the bar (Fig. 22).
the gravimetric loss divided by the density of the metal and Figs. 22 and 23 also indicate the recommended residual
the total area of attack which gives the equivalence given in cross section (dashed circle within the original bar cross
paragraph 3.5: an averaged corrosion current density of section) to be considered when implementation of the residual
1 PA/cm2 is equivalent to an averaged penetration rate of diameter into structural models is going to be made.
11.6 Pm/year. 8.1.5 Rebar diameter, 'Øt
It may also be obtained by integration of the corrosion
current, Icorr, (expressions (9) and (10)) or corrosion rate, The loss of rebar diameter, 'Øt, can be calculated from
Vcorr, values until tp [1] by means of the expression: expression [63]:

Uniform corrosion: 'Øx(mm)=Øo- Øt= Øo- Px · D (14)

REP REP
Px ( mm) 0.0116 I corr • tp Vcorr • tp (11) Where Øo is the initial diameter and 'Øx is the diameter
loss after a certain period t of corrosion. The value of D is
where tp is the propagation period >year@. equal to 2 (Fig. 22) for uniform corrosion and for localized
corrosion (Fig. 23) may vary from 3 to 10. The value of 10
8.1.3 Pitting factor, D
is usually taken in order to have conservative predictions.
The pitting factor, D, [41] or corrosion concentration
factor, represents the compensation of the relative error due 8.2 Interpretation of corrosion rate results
to the differences in actual size of corroding area due to
localized attack. It results from dividing the maximum pit 8.2.1 Classification of corrosion rate results
depth (localized attack), Ppit, obtained by visual inspection Values of Icorr below 0.1 PA/cm2 indicate negligible
by the attack penetration averaged to the full circumference, corrosion from a practical point of view, and therefore, the
Px, (see Fig. 23): steel reinforcement can be classified as “passive”. The range
between 0.1-0.2 PA/cm2 can be considered the transition
region between passive and active corrosion. This range is
due to the inherent uncertainty of Stern’ s formula
which assumes an intrinsic error factor of two.
Expressing Icorr in PA/cm2, and Vcorr in mm/y, the
following classification has been established [1, 10, 63].
Values of Icorr above 1 PA/cm2 are seldomly
measured in real size structures and values higher than
10 PA/cm2 have almost never been recorded. In

Table 2 - Ranges of corrosion current values

related to the significance in terms of service life
of the reinforcement
Icorr (PA/cm2) Vcorr (mm/y) Corrosion level

d 0.1 d0.001 Negligible

Fig. 22 - Attack penetration: case of Fig. 23 - Residual cross section 0.1 – 0.5 0.001-0.005 Low
uniform corrosion. in the case of pitting (localized 0.5 – 1 0.005-0.010 Moderate
> 1 >0.010 High
attack). >D # 10@ .

636
Materials and Structures / Matériaux et Constructions, Vol. 37, November 2004

consequence the most common values for actively corroding Table 3 - Resistivity ranges related to the
rebars range from 0.1 to 1 PA/cm2. risk of corrosion
8.2.2 Relation between corrosion current and corrosion Concrete resistivity Corrosion risk
potential values (k:·cm)
>100 Negligible
No unambiguous mathematical relationship has been – 100 Low
found between the corrosion potential, Ecorr, and Icorr as is 10 - 50 Moderate
shown in Fig. 24 where results of numerous inspections are < 10 High
presented. This lack of a clear correlation has been
attributed to the fact that both parameters respond presented in Fig. 25 [58]. Although a wide scatter is
differently to the same variables. Particularly moisture registered, a certain relation between both variables seems
(oxygen availability), temperature and concrete resistivity to exist. This relation follows the general trend of the
values seem to affect both electrochemical parameters, expression:
however in different proportion.
Icorr # 3 ˜104·/U (15)
2
with Icorr expressed in PA/cm and U in ohm·cm.
This empirical relation results from the fact that concrete
resistivity is a direct function of the moisture content of the
concrete and more precisely, from its degree of water
saturation. A ranking of resistivity values has been
established [58] which is shown in Table 3:
An Icorr-U graph can be established from Tables 2 and 3.
In the regions with Icorr values above 0.1 PA/cm2, the results
should follow a parallel slope to the main line for different
values of U.
8.2.4 Evolution with time of corrosion current values

Fig. 24 - Plot of Icorr versus Ecorr. A general correlation does When the steel is passive, the Icorr values usually remain
not exist. below 0.1 PA/cm2, however just after mixing or with
prerusted rebars, the Icorr values may be significantly higher
On the other hand, it has been noticed that rust without indicating a significant loss of base metal.
composition has an influence on the Ecorr values [46-48] and When the carbonation front or the chloride threshold
thus, for the same corrosion rate a different aging of the rust reaches the steel surface, local depassivation occurs with a
(proportion Fe+2/Fe+3) will induce different Ecorr values. noticeable increase in the Icorr values [10]. After
8.2.3 Relation between corrosion current and resistivity depassivation, the Icorr may vary in function of the intrinsic
(Icorr-U graph) values variability of any corrosion process or to extrinsic factors
(exposure conditions or external weather).
On the other hand, Icorr values present a better Usually, depassivation does not occur instantaneously
relationship with the concrete resistivity values, as is but takes some time during which events of
activation/repassivation occur until the whole perimeter of
the bar has become depassivated. During depassivation the
corrosion is always produced locally, either due to the
generation of pits or to the fact that the carbonation front
has reached the upper surface of the rebar only. Then the
initial pits or corroding zones may grow or repassivate and
new corroding zones may generate, until the advance of the
carbonation front finally encloses the whole perimeter.
In consequence the values of Icorr during the
depassivation process may not remain constant, but vary
significantly over time and the location of the structure.
When fully depassivated, the corrosion process develops by
generating iron oxides which may induce cracking of the
cover It is important to note that cracking not always
induces an acceleration of the corrosion process because of
the fact that cracks may result in a quicker drying. In
consequence, the whole process is in continuous evolution
and the corrosion rate may also vary from place to place
and time to time.
Fig. 25 - Plot of Icorr values versus concrete resistivity The Icorr may as well change due to the weather conditions
diagonal line indicates the relation usually found between [59, 60]. The degree of water saturation Sw is the main
these two parameters. The corrosion regions are marked factor influencing Icorr, together with temperature. The
following Tables 1 and 2.

637
RILEM TC 154-EMC

degree of water saturation, Sw, is a consequence of the rain Due to the daily temperature cycles, it is recommended
periods (in concrete non sheltered from rain) or of the to obtain at least 2 values per day, related to the maximum
precise conditions in T and RH in sheltered concrete. The (mid-day) and minimum (mid-night) expected temperature
knowledge of these climatic events is of importance to values. Four values per day seems to be an optimum in
interpret the values of Icorr recorded on-site. The Sw can be order not to overload the recording, but enables registration
determined by also measuring the concrete resistivity during the most important weather events (e.g. rainfall).
(graph Icorr-U) or by weighing a concrete core taken from 8.2.5.2 Discontinuous measurements
the structure in the same moisture conditions as the
In the case of non-permanent installations, it is
measurement points.
recommended to take several measurements along a year
8.2.5 Calculation of a representative value of Icorr period. A minimum of 4 readings per year is recommended.
They will be taken during the most extreme climatic
The natural scatter in the evolution of Icorr values in time conditions which, for the sake of the corrosion process, may
and place does not prevent the obtainment of a be defined in the following manner:
representative value to be used for assessing the condition 1. Dry periods with low temperatures.
of the structure, but calls for rules or methodologies for its 2. In the periods of lower temperatures after several
calculation [61, 62]. consecutive events of raining (humid period).
REP
The accuracy in the calculation of a I corr will depend 3. Dry periods and high temperatures.
on the number of individual Icorr measurements recorded. In 4. Periods with high temperature but after long rainfall
on-site measurements, there are basically three possibilities: (humid and hot periods).
1) continuous monitoring by embedded or permanent Measurements performed during these periods will
sensors, 2) the recording of several values at equal time enable the recording of nearly minimum and maximum
intervals along a year period and 3) the obtainment of values of Icorr exhibited by the particular structure at the
REP
single isolated values of Icorr. The general procedure is measurement points. The representative I corr will be
shown in Fig. 26. obtained through expression (9). Its value will be more
accurate as the number of measurements, n, increases. Not
only the mean value, but also the standard deviation may be
used for making reliable predictions (chapter 8.3.3.1.)
8.2.5.3 Single measurements
When only one isolated measurement can be performed,
obtaining a representative Icorr is more uncertain. In order to
interpret the readings in the most accurate way, the
procedure recommended is based on the averaging of the
site measurements with those obtained in cores drilled from
the structure. In these cores either the corrosion rate of
embedded pieces of rebar can be averaged with the site
values, or the Icorr values can be obtained through the
relation between concrete resistivity and Icorr measured in
the cores.
In the most normal case that the cores have not pieces of
rebars, the procedure recommended is shown in Fig. 27. In
this figure it is represented by the line AB the averaged
general relation between Icorr and U when plotted on a log –
log diagram (Fig. 25). The linear relation presents a slope
of –1 (Equation (15)). As a consequence, the procedure
proposed is as follows:
x After having measured on-site the concrete resistivity
and corrosion current, cores should be drilled close to the
measurement points. Cores are returned to the laboratory.
Then they are conditioned to a moisture content
corresponding to 85% RH (for structures sheltered from
rain) or to water saturation after being air/vacuum (for non
sheltered or submerged ones). When the cores are
Fig. 26 - General procedure for obtaining the Representative Icorr equilibrated to the moisture conditions their minimum
when the drilled cores in the right branch (single measurements electrical resistivity, Umin is measured.
on-site) do not have pieces of rebar. x The cores should be in such a condition that they are
representative of the actual condition of the structure (no
8.2.5.1 Continuous monitoring additional cracks/damage due to drilling are allowed).
When continuous recording of data from permanently installed Preferentially, the concrete cores to be investigated in the
REP
sensors is possible, I corr can be obtained from the averaging laboratory should be retrieved at the locations where the
of the Icorr-t curve as indicated by expression (9) or (10). actual measurements on site were performed.

638
Materials and Structures / Matériaux et Constructions, Vol. 37, November 2004

Fig. 29 - Core with a piece of rebar that can be used

to obtain the Icorr, max value through the Rp
determination after conditioning the core in the
laboratory at a predetermined humidity (85% RH or
water saturated).

a piece of rebar (see Fig. 29). If the cores

themselves contain the steel bar, the use of the
U- Icorr –relationship will not be necessary. In
the case of cores with reinforcements, the cores
would be as well conditioned in the laboratory
Fig. 27 - Procedure suggested for averaging results measured in a single visit on to the maximum selected humidity and Icorr
site with values deduced from resistivity measured in drilled cores conditioned would be measured directly in the pieces of
in the laboratory.
rebar in the core by means of a potentiostat and
an external counter and reference electrodes.
The actual condition of the embedded steel
section should be checked after the
measurements and characterized by type of
attack (uniform, localised, penetration depth).
The real diameter loss should be measured as
well after removing the rust.
In summary, when only a singe visit on site
REP
is feasible the value of I corr is calculated by
taking two set of values: those measured on site
(Icorr, single ) and those measured in the laboratory
in cores drilled from the structure. That is, from
the average of the single on-site values, Icorr, single
and the maximum value achieved under
laboratory conditions, Icorr, max, for the particular
condition selected (85% RH or saturated) [62].
The cores drilled may or not have embedded
pieces of rebar. If they have, the Icorr, max can be
measured directly by means of a potentiostat. If
the cores do not contain isolated pieces of rebar,
the Icorr-U graph must be used in order to deduce
Icorr, max from Umin.
Fig. 28 - Alternative to the general procedure shown in Fig. 27 for the case that I co rr , s in g  I co rr , m ax
the drilled cores in the on-site single measurements have pieces if rebar I coR ErrP (16)
embedded. 2
Then, the set or clusters of values of Icorr and - U However, caution has to be taken if the
registered on-site are plotted on the graph (points P on values are to be implemented into structural models for the
Fig. 27) and the Umin measured in the cores is identified in calculation of the load carrying capacity. The lack of
the graph (point R). Then a regression line is drawn through measurements along the different seasons should be
the measured, P, values. This line is extrapolated to reach compensated with the most practically possible larger
sampling in different locations of the structure as was
the vertical of R (Umin) (point C) which is then extrapolated
indicated in paragraph 6.7.
to D, to obtain the Icorr,max (maximum) value coherent with
the minimum resistivity values of the cores.
x An alternative derived from this methodology is given
in Fig. 28. It consists in drilling concrete cores containing

639
RILEM TC 154-EMC

8.2.5.4 Calibration by measurement of the diameter loss treatment is recommended and special care has to be taken
of the reinforcement to obtain representative results before and after inhibitor
Whatever the method used for obtaining a representative treatment, taking changes in temperature and moisture
REP
I corr value it is convenient to calibrate it with condition into account.
In the particular case of non-permanent electrochemical
measurements of the real penetration of attack, Px. This can
repair techniques, particularly realkalization and chloride
be made from visual inspection of the steel bars through
extraction, sufficient time for recovery of the electrochemical
expressions (11, 12 and 13).
potential to the natural values has to be allowed for, because
However, when the corrosion attack is only very slight
if the rebar is polarized the current determined is not related
(superficial), the measurement of a diameter decrease may be
to the true corrosion current Icorr.
very inaccurate. In such cases, a more accurate, although
destructive method, is the measurement of the weight loss of 8.3.3 Implementation into structural calculations
a small portion of the rebar. That is, cutting a small part of (damage functions)
the bar and removing the rust oxides by proper cleaning. The
The corrosion process results into four main detrimental
weight of the bar, Wf, gives the average penetration of attack,
consequences [63-68] (Fig. 30):
Px (expression (11)) by using the following expression:
1. Reduction of bar cross section;
2. Reduction of steel ductility;
W f  S LG ª¬ r i 2   r i 2  P x2  2r i P x  º¼ (17)
S r i 2 LG 3. Cracking of concrete cover;
TTL..G P x (2r i  P x ) 4. Reduction of steel/concrete bond (composite effect).
The degradation of these structural characteristics call
where, ri is the initial nominal radius, L is the length of the for the establishment of several damage functions able to
cut-piece of reinforcement, Wf the final (actual) weight and link the degree of deterioration or cross section loss, Px,
G = 7.85, the theoretical iron density. with the loss of structural load-bearing capacity.
A more accurate result can be obtained if G is measured
by weighing and measuring the volume of an identical
unrusted piece of the same reinforcement.
Concerning the location of these measurements of the
penetration depth, if possible, the points of measurement
should be selected close to those where the Icorr have been
recorded. A distinction has to be made between uniform
and localized corrosion (Vcorr or Vpit).
8.3 Use of corrosion rate results
As was mentioned before, the Icorr values can be used for
the following different purposes:
8.3.1 Identification of actively corroding zones Fig. 30 - Consequences of rebar corrosion which lead into the
decrease in load-bearing capacity of the structure.
These areas will be deduced from the delimitation of
zones with different Icorr values taken Table 1 as indicative
8.3.3.1 Calculation of loss in rebar cross section (damage
of corrosion levels.
function of attack penetration, Px)
Corrosion rate maps will help to identify the corroding
zones in the same manner as potential mapping. Fig. 20 The primary damage function [42, 57, 63] to be obtained
shows an example. from Icorr values recorded on site is that concerning the
As was mentioned in paragraph 8.2.4, when analyzing penetration depth Px (definitions in paragraph 8.1).
the corrosion rate data, account has to be taken regarding: Attention has to be paid on whether the corrosion attack is
a) the particular circumstances of evolution of the whole localized (expression (13)).or rather uniform. The value of
REP
electrochemical process and b) the prevailing moisture Icorr to be introduced in these expressions is I corr .
conditions of the concrete at the measurement location and REP
As was described in chapter 8.2.5, I corr to be used for
time. calculating the future evolution of the damage or of the
8.3.2 Evaluation of the effectivity of repair techniques residual load-bearing capacity, has to be accurately obtained.
Either by calculating an average and standard deviation of
In order to evaluate the effectivity of a repair technique, REP
the I corr from several measurements in the case of isolated
measurements prior to and after the treatment should be
measurements or by using the actual value recorded in the
carried out.
structure and a second value obtained through any of the
Concerning the measurements after the repair, attention
methods described in 8.2.5.3.
has to be paid to the different electrochemical conditions
that a repair may induce in the structure involving evolution 8.3.3.2 Implementation into structural models
of the macrocell activity. Concerning the rest of structural consequences, (cover
In the case of the application of inhibitors, care has to be cracking, decrease in bond, and load-bearing capacity) their
taken on the time elapsed between treatment and Icorr particular damage function formulation is out of the scope of
measurements in order to allow the inhibitor to become the present Recommendation. They can be found in [64-68].
effective. More than one single measurement after the

640
Materials and Structures / Matériaux et Constructions, Vol. 37, November 2004

SYMBOLS U Electrical resistivity [:m]

B Stern constant [V]
C Electrical capacitance [F]
CE Counter or auxiliary electrode REFERENCES
E Electrical potential [V]
[1] Andrade, C. and González, J.A., ‘Quantitative measurements of
'E Potential step [V] corrosion rate of reinforcing steels embedded in concrete using
Ecorr Corrosion or mixed potential [V] polarization resistance measurements’, Werkstoffe und
icorr Instantaneous corrosion current density Korrosion 29, (1978) 515-519.
[PA/cm2] [2] Stern, M. and Geary, A.L., ‘Electrochemical Polarization: I.
Icorr Non-uniform instantaneous corrosion current A. theoretical analysis of the shape of polarization curves’,
Density [PA/cm2] Journal of Electrochemical Soc. 104 (1) (1957) 56-63.
'I Current step [A] [3] Stern, M. and Weisert, E.D., ‘Experimental observations on the
relations between polarization resistance and corrosion rate’,
Ipit Instantaneous corrosion current in a pit or Proc. American Society Testing, Materials 59 (1958) 1280.
localized corroding spot [PA/cm2] [4] Feliú, S., González, J.A., Andrade and C., Feliú, V., ‘On-site
Igalv Galvanic corrosion current between determination of the Polarization Resistance in a reinforced
corroding and cathodic zones [PA/cm2 or PA] concrete beam’, Corrosion (USA) 43 (Sept. 1987) 1-9.
Icorr,lab Icorr determined in small specimens with finite [5] Andrade, C., ‘New electrochemical technique for the
reinforcement area [PA/cm2] corrosion measurement in reinforced and presstressed
Icorr,sing Icorr determined on site only one time concretes. Use of inhibitor admixtures as prevention
method’, Ph.D Thesis - Faculty of Chemistry, Complutense
[PA/cm2] Univ. of Madrid, July 1973.
Icorr,tn Averaged Icorr obtained from integrating or [6] Lorenz, W.J. and Mansfeld, F., ‘Determination of corrosion
averaging several Icorr measurements obtained rates by electrochemical DC and AC methods’, Corrosion
during a period of time tn [PA/cm2] Science 21 (9) (1981) 647-672.
IcorrREP Representative-Icorr value [PA/cm2] [7] Epelboin, I., Gabrielli, C., Keddam, M. and Takenouti, H.,
Lcrit Critical length polarized in on-site ‘Alternating-current impedance measurements applied to
measurements [cm] corrosion studies and corrosion-rate determination’,
Electrochemical Corrosion Testing, ASTM STP 727. F.
Px Penetration depth of corrosion attack at a Mansfeld and U. Bertocci, Eds., American Society for
certain time [mm] Testing and Materials (1981) 150-166.
Ppit Maximum pit or localized penetration depth [8] Gabrielli, C., Keddam, M., Takenouti, H., Vu Quang Kinh
[mm] and Bourelier, F., ‘The relationship between the impedance
Re Electrical resistance [:] of corroding electrode and its polarization resistance
R1,R2, RB Reference electrodes for critical length determined by a linear voltage sweep technique’,
measurement hold in the auxiliary electrode Electrochimica Acta 24 (1979) 61-65.
of potential attenuation method [9] McDonald, D.D. and McKubre, M.C.H., ‘Electrochemical
Impedance Techniques in corrosion science’,
RE Reference electrode for measurement of Ecorr
Electrochemical Corrosion Testing ASTM STP 727 - F.
Rp Polarization Resistance [: or :cm2] Mansfeld and U. Bertocci Eds. (1981) 110-149.
Rp,true True Rp [:cm2] [10] González, J.A. and Andrade, C., ‘Effect of carbonation,
Rp,app Apparent Rp [: or :cm2] chlorides an relative ambient humidity on the corrosion of
Rpref Rp obtained in a reference reinforced concrete galvanized rebars embeded in concrete’, British Corrosion
slab for calibrating portable corrosion-rate- Journal 17 (1) (1982) 21-28.
[11] Gouda, V.K., Shater, M.A. and Mikhail, R. Sh., ‘Hardened
meters [:cm2]
portland blast-furnace slag cement pastes, II. The corrosion
S Area of the reinforcement to be measured behaviour of steel reinforcement’, Cement and Concrete
[m2] Research 5 (1975) 1-13.
SA Anodic or corroding area [m2] [12] Glass, G.K., Page, C.L. and Short, N.R., ‘Factors affecting
SC Cathodic area [m2] the corrosion rate steel in carbonated mortars’, Corrosion
S1, S2 Reference electrodes to control the Science 32 (12) (1991) 1283-1294.
confinement in the guard ring auxiliary [13] Hardon, R.G., Lambert, P. and Page, C.L., ‘Relationship
electrode time or timelife t between electrochemical noise and corrosion rate of steel in
ti Initiation period in service life model [year] salt contaminated concrete’, British Corrosion Journal 23
tp Propagation or corroding period in service (4) (1988) 225-228.
[14] Lambert, P., Page, C.L. and Vassie, P.R.W., ‘Investigations
life model [year] of reinforcement corrosion. 2. Electrochemical monitoring
Vcorr Instantaneous corrosion rate [mm/year or of steel in chloride-contaminated concrete’, Materials and
Pm/year] Structures/Matériaux et Constructions 24 (1991) 351-358.
VcorrREP Representative Vcorr value [Pm/year or [15] Polder, R, Tondi, A. and Cigna, R., ‘Concrete resistivity and
mm/year] corrosion rate of reinforcement’, TNO report 93-BT-r0170,
D Pitting factor TNO Delft (1993).
[16] Pollet, V. and Raharinaivo, A., ‘Assessment of damage by
‡x Loss in reinforcement diameter after a certain
corrosion: Techniques for predicting the extension of rebar
tp [mm] corrosion’, International Symposium on Bridge Engineering
‡o Initial nominal reinforcement diameter [mm] and Management in Asian countries, PIARC, Jakarta
‡t Residual reinforcement diameter after a (Indonesia) 10-13 September (1996).
certain tp [mm]

641
RILEM TC 154-EMC

[17] Raharinaivo, A.L. and Carpio, J.J., ‘The steeping down the polarization effects’, Corrosion 89, New Orleans (USA),
current method: a new corrosion control for cathodic paper 25, April 17-21, 1989.
protection of steel’, NACE Conference Corrosion '92, [35] Andrade C., Soler L. and Nóvoa, X.R., ‘Advances in
Nashville, Paper 228, (1992) 9. electrochemical impedance measurements in reinforced
[18] Dhouibi-Hachani, L, Raharinaivo, A., Triki, E. and Fiaud, concrete’, 5th Intern. Symposium on electrochemical
C., ‘Assessing the corrosion of rebars in concrete methods in Corrosion Research, EMCR’94, Sesimbra,
deteriorated by sulfates and carbonation’, Int. Conference on Portugal, Sept. 1994.
Corrosion and Corrosion Protection of steel in Concrete. Ed. [36] Epelboin, I., Gabrielli, C., Keddam, M. and Takenouti, H.,
R.N. Swamy, Sheffield, July (1994) 258-267. ‘Alternating-current impedance measurements applied to
[19] Berke, N.S., Shen, D.F. and Sundberg, K.M., ‘Comparison corrosion studies and corrosion rate determinations’, ASTM
of the polarization resistance technique to the macrocell STP 727 “Electrochemical Corrosion Testing”, F. Mansfeld
corrosion technique’, Corrosion Rates of steel in Concrete, and U. Bertocci, Eds., American Society for Testing and
ASTM-1065, N. Berke, V. Chacker and D. Whiting Eds. Materials (1981) 150-166.
(1990)38-51. [37] Hachani, L., Carpio, J., Fiaud, C., Raharinaivo, A. and Triki,
[20] Hope, B.B. and Ip, A.K.C., ‘Corrosion of steel in concrete E., ‘Steel corrosion in concretes deteriorated by chlorides
made with slag cement’, ACI Materials Journal (Nov.-Dec. and sulphates: electrochemical study using impedance
1987) 525-531. spectrometry and stepping down the current method, Cement
[21] Hansson, C.M., ‘Comments on electrochemical and Concrete Research 22 (1992) 55-66.
measurements of the rate of corrosion of steel in concrete’, [38] Feliú, S., Galvan J.C., Feliú Jr., S., Bastidas, J.M., Simancas,
Cement & Concrete Res. 14 (1984) 574-584. J., Morcillo, M. and Almeida, E.M., ‘An electrochemical
[22] Gulikers, J., ‘Numerical simulation of corrosion rate impedance study of the behaviour of some pretreatments
determination by linear polarization’, Rilem PRO 18, applied to rusted steel surfaces’, Corrosion Science 35 (5-8)
Workshop on Measurement and interpretation of on-site (1993) 1351-1358.
corrosion rate (Mesina), C. Andrade, C. Alonso, J. Fullea, J. [39] Mansfeld, F., ‘Recording and analysis of AC impedance data
Polimón and J. Rodríguez Eds., Madrid, Feb. (1999). for corrosion studies. I. Background and methods of analysis’,
[23] Pedeferri, P., ‘Corrosion and Protection of Metallic Corrosion (NACE) 36 (5), Mayo (1981).
Materials’, Librería Politécnico de Milano, Italy (1978) [40] Elsener, B., Hug, A., Bürchler, D. and Böhni, H., ‘Evaluation
(only in Italian). of localized corosion of steel in concrete by galvanostatic
[24] González, J.A., Molina, A., Escudero, M.L. and Andrade, pulse technique’, Conference on “Corrosion of Reinforcement
C., ‘Errors in the electrochemical evaluation of very small in Concrete Construction”, C.L. Page, P.B. Bamforth and J.L.
corrosion rates. Part I. Polarization resistance method Figg Eds. SCI, Cambridge (1996) 264-272.
applied to corrosion of steel in concrete’, Corrosion Science [41] González, J.A., Andrade, C., Alonso, C. and Feliú, S.,
(UK) 25 (1985) 917-930. ‘Comparison of rates of general corrosion and maximum pitting
[25] Glass, G.K., Page, C.L., Short, N.R. and Yu, S.W., ‘An penetration on concrete embedded steel reinforcements’,
investigation of galvanostatic transient methods used to Cement & Concrete Research 25 (2) (1995) 257-264.
monitor the corrosion rate of steel in concrete’, Corrosion [42] Andrade, C., Alonso, C. and González, J.A., ‘An initial
Science 35 (5-8) (1993) 1585-1592. effort to use corrosion rates measurements for estimating
[26] Newton, C.J. and Sykes, J.M., ‘A galvanostatic pulse rebar durability’, Corrosion Rates of Steel in Concrete
technique for investigation of steel corrosion in concrete’, ASTM STP-1065, N. Berke, V. Chacker and D. Whiting
Corrosion Science 28 (1988) 1051-1074. Eds. (1990) 29-37.
[27] Pollet, V., Grimaldi, G. and Raharinaivo, A., ‘Corrosion rate [43] Mansfeld, F., ‘The relationship between galvanic current and
of steel in carbonated concrete measured with polarization dissolution rates’, Corrosion (NACE) 29 (10) (1973) 403-405.
resistance method and a new transient technique’, Paper I- [44] Andrade, C., Maribona, I.R., Feliú, S., González, J.A. and
OR14, EUROCORR'96, Nice, France 24-27 Sept. (1996). Feliú Jr., S., ‘The effect of macrocells between active and
[28] Elsener, B., Klinhoffer, O., Frolund, T., Rislund, E., passive areas of steel reinforcements’, Corrosion Science 33
Schiegg, Y. and Böhni, H., ‘Assessment of reinforcement (2) (1992) 237-249.
corrosion by means of glavanostatic pulse technique’, Int. [45] Andrade, C., Merino, P., Nóvoa, X.R., Pérez, M.C., M.C.
Conference on Repair of Concrete Structures, Svolvær, and Soler, L., ‘Passivation of reinforcing steel in concrete’,
Norway, Edited by A. Blackvoll, May (1997) 391-400. Materials Science Forum 192-194 (1995) 891-898.
[29] Feliú, S., González, J.A., Andrade, C. and Feliú, V., ‘The [46] Andrade, C., Bolzoni, F., Cabeza, M., Nóvoa, X.R. and Pérez,
determination of the corrosion rate of steel in concrete by a M.C., ‘Measurement of steel corrosion in concrete by
non-stationary method’, Corrosion Science 26 (1986) 961- electrochemical techniques: influence of the redox processes in
970. oxide scales’, in ‘Electrochemical Approach to Selected
[30] Hladky, K., Callow, L.M. and Dawson, J., ‘Corrosion rates Corrosion and Corrosion Control Studies, European Federation of
from impedance measurements: an introduction’, British Corrosion Pub., no. 28, P.L. Bonora and F. Deflorian Eds., The
Corrosion J. 15 (1) (1980) 20-25. Institute of Materials, London, UK, Cap. 25, (2000) 332-343.
[31] Wenger, F., Galland, J., Lemoine, L., ‘Méthode de contrôle [47] Andrade, C., Keddam, M., Nóvoa, X.R., Pérez, M.C.,
de la corrosion des armatures de béton en milieu marin’, Rangel, M.C. and Takenouti, H., ‘Electrochemical
International Symposium on Behaviour of offshore concrete behaviour of steel rebars in concrete: influence of
structures, Brest, France, October (1980). environmental factors and cement chemistry’, Electrochimica
[32] Andrade, C. and Castelo, V., ‘Practical measurement of the Acta 46 (24-25) (2001) 3905-3912.
A.C. Impedance of steel bars embedded on concrete by means [48] Alonso, C., Andrade, C. Izquierdo, M., Nóvoa, X.R. and
of a Spectrum Analyzer, Fast Fourier Transform’, British Pérez, M.C., ‘Effect of protective oxide scales in the
Corrosion Journal (UK) 19 (1984) 93-100. macrogalvanic behaviour of concrete reinforcements’,
[33] Sagüés, A., ‘Electrochemical impedance of corrosion Corrosion Science 40 (8) (1998) 1379-1389.
macrocells on reinforcing steel in concrete’, NACE [49] Feliú, S., González, J.A., Andrade, C. and Feliú, V.,
Corrosion 90, Las Vegas, USA, Paper 132 (1990). ‘Determining polarization resistance in reinforced concrete
[34] Sagüés, A.A., ‘Evaluation of corrosion rate by slabs’, Corrosion Science 29 (1) (1989)105-113.
electrochemical impedance in a system with multiple

642
Materials and Structures / Matériaux et Constructions, Vol. 37, November 2004

[50] Seghal, A., Kho, Y.T., Osseo-Asare, K. and Pickering, [60] Andrade, C., Sarria, J. and Alonso, C., ‘Relative humidity in the
H.W., ‘Comparison of corrosion rate-measuring devices for interior of concrete exposed to natural and artificial weathering’,
determining corrosion rate of steel in concrete systems’, Cement and Concrete Research 29 (1999) 1249-1259.
Corrosion Engineering 48 (1992) 871-880. [61] Andrade, C. and Alonso, C., ‘On-site measurements of
[51] Andrade, C. and Alonso, C., ‘Corrosion rate monitoring in corrosion rate of reinforcements’, Construction and Building
the laboratory and on-site’, Construction and Building Materials 15 (2001) 141-145.
Materials 10 (5) (1996) 315-328. [62] Andrade, C., Fullea, J. and Alonso, C., ‘The use of the graph
[52] Elsener, B. and Böhni, H., ‘Galvanostatic pulse measurements. corrosion rate- resistivity in the measurement of the corrosion
Rapid on-site corrosion monitoring’, Int. Conference on current’, Workshop on Measurement and Interpretation of On-
Corrosion and Corrosion Protection of steel in Concrete, Site Corrosion Rates, MESINA, RILEM PRO 18, C.
Sheffield, UK, Ed. R.N. Swamy, July (1994) 236-246. Andrade, C. Alonso, J. Fullea, J. Polimon and J. Rodríguez,
[53] Mietz, J. and Isecke, B., ‘Electrochemical potential Eds., RILEM Pub., Madrid (Feb. 1999) 157-165.
monitoring on reinforced concrete structures using anodic [63] Andrade, C., Alonso, C., González, J.A. and Rodríguez, J.,
pulse techniques’, Conference on ‘Non-destructive Testing ‘Remaining service life of corroded structures’, Proceedings
in Civil Engineering’, H. Bungey Ed., The British Institute of IABSE Symposium on Durability of Structures, Lisbon,
of NDT, 2 (1993) 567. September 1989, 359-363
[54] Feliú, S., González, J.A. and Andrade, C., ‘Multiple- [64] Rodríguez, J., Aragoncillo, J., Andrade, C. and Izquierdo,
electrode method for estimating the polarization resistance D., ‘Manual for assessing corrosion- affected concrete
in large structures’, Journal of Applied Electrochemistry 26 structures’, CONTECVET IN30902I,
(1996) 305-309. www.ietcc.csic.es/public_elec/Formulario_Contecvet.html.
[55] Andrade, C., Sarria, J. and Alonso, C., ‘Statistical study on [65] Rodríguez, J., Ortega, L.M. and Casal, J., ‘Corrosion of
simultaneous monitoring of rebar corrosion rate and internal reinforcing bars and service life of reinforced concrete
relative humidity in concrete structures exposed to the structures: corrosion and bond deterioration’, International
atmosphere’, Conference on Corrosion of Reinforcement in Conference on Concrete Across Borders, Odense (Denmark)
Concrete Constructio, C.L. Page, P.B. Bamforth and J.L. Figg Vol. 1 (1994) 215-226.
Eds. SCI, Cambridge (1996) 233-242. [66] Rodríguez, J., Ortega, L.M., Aragoncillo, J., Izquierdo, D. and
[56] Tuutti, K., ‘Corrosion of steel in concrete’, Swedish Cement and Andrade, C., ‘Structural assessment for residual life
Concrete Research Institute (CBI) No. 4-82, Stockholm (1982). calculation of concrete structures affected by reinforcement
[57] Andrade, C. and Alonso, C., ‘Values of corrosion rate of steel corrosion’, Int. RILEM Workshop on Life Prediction and
in concrete in order to predict service life of concrete aging management of concrete structures, D. Naus Ed., Rilem
structures’, ASTM STP-1194 ‘Application of accelerated Pub., Cannes, France (October 2000) 97-111.
corrosion tests to service life prediction of materials’, G. [67] Rodríguez, J., Ortega, L.M., Casal J. and Díez, J.M.,
Cragnolino and U. Sridhan Eds. (1994) 282-295. ‘Corrosion of reinforcement and service life of concrete
[58] Alonso, C., Andrade, C., González, J.A., ‘Relation between structures’, in ‘Durability of Building Materials and
concrete resistivity and corrosion rate of the reinforcements in components’, Vol. I, C. Sjöström Ed., E&FN Spon Pub.,
carbonated mortar made with several cement types’, Cement London (1996) 117-126.
and Concrete Res. 18 (1988) 687-698. [68] Rodríguez, J., Ortega, L.M., Casal, J. and Díez, J.M., ‘Assessing
[59] Andrade, C., Sarria, J. and Alonso, C., ‘Relation between structural conditions of concrete structures with corroded
climate and corrosion rate’, Workshop on Measurement and reinforcements’, Conference on Concrete Repair, Rehabilitation
Interpretation of on-site corrosion rates, MESINA, RILEM and Protection, Dundee, U.K., R.K. Dhir and M.R. Jones Eds.,
PRO 18, C. Andrade, C. Alonso, J. Fullea, J. Polimon and J. E&FN Spon Pub., London (June 1996) 65-77.
Rodríguez, Eds., RILEM Pub., Madrid (Feb. 1999) 123-141.

643