Sunteți pe pagina 1din 9

CORROSION SCIENCE

Characterization of Activated Titanium


Solid Reference Electrodes
for Corrosion Testing of Steel in Concrete
P. Castro,* A.A. Sags, E.I. Moreno,** L. Maldonado,* and J. Genesc***

ABSTRACT
Small bars of Ti activated with mixed-metal oxide (commercially produced for permanent impressed-current anodes in
cathodic protection) were used as embedded reference electrodes (RE) in concrete. Their electrochemical behavior was
evaluated through measurements and analyses of potential,
electrochemical impedance spectroscopy (EIS), cyclic polarization (CP), and galvanostatic tests in buffer solutions of
pH 4, 7, and 10, saturated calcium hydroxide (Ca[OH]2,
pH ~ 12.5) simulated concrete pore solution (SPS) with
pH ~ 13.5, and various concrete mixes with and without
pozzolanic additions as cement replacement. Effects of deaeration and sodium chloride (NaCl) additions were
evaluated. The potential of the activated Ti rod (ATR) electrodes resembled the expected dependence for the system
Ir2O3 + H2O = 2IrO2 + 2H+ + 2e in aqueous solutions. The
ATR electrode presented generally good stability with time in
concrete for up to 900 days. Anomalous behavior was found
in two concrete mixes with the highest pozzolanic content.
Results from EIS tests revealed a constant phase element
(CPE) behavior, which agreed with results of CP tests that
showed a very large apparent interfacial capacitance. The
apparent capacitance was on the order of 102 F/cm2, resulting in very low impedance, which is advantageous when
using ATR electrodes to conduct EIS or polarization resistance tests. Galvanostatic application of ~ 0.075 A/cm2
caused little variation of potential with time (~ 60 mV in

Submitted for publication May 1995; in revised form, August


1995.
* Depto. Fsica Aplicada, Centro de Investigacin y de Estudios
Avanzados (CINVESTAV)-Unidad Mrida, AP 73 Cordemex, CP
97310, Mrida, Yucatn, Mxico.
** Department of Civil and Environmental Engineering, University of
South Florida, Tampa, FL, 33620.
*** Facultad de Qumica de la Universidad Nacional Autnoma de
Mxico, Cd Universitaria, 04510 Mxico, DF, Mxico.

CORROSIONVol. 52, No. 8

several days), indicating the presence of a finite polarization


resistance. Little short-term susceptibility of the ATR electrode potential to NaCl additions was found. The ATR
electrode potential also showed little short-term sensitivity to
variations in oxygen partial pressure (PO2).
KEY WORDS: activated titanium, constant phase element,
interfacial capacitance, mixed-metal oxide, pH, polarization
resistance, reference electrodes, steel

INTRODUCTION
Monitoring and control of corrosion of reinforcing steel
in concrete requires reliable measurement of electrode
potentials. A reference electrode (RE) provides a metallic surface at a fixed potential away from the electrolyte
(the moisture present in the concrete pores). In a RE,
the difference in potential between its metallic surface
and the electrolyte should be stable, reproducible, and
well known with respect to the standard hydrogen electrode (SHE). The RE should develop only small
overpotentials when subject to the current demand of
ordinary potential measurement instrumentation. Potential recovery from reasonable excursions should be
quick. For polarization measurements of reinforcing
steel, the electrode should have low impedance and
excellent short-term stability.
Use of RE in concrete presents several challenges. Conventional electrodes such as the
copper-copper sulfate (CuSO4) electrode (CSE) can be
placed easily on an external concrete surface. However, significant errors can develop because of
various factors such as the often large resistance
path between the steel and the surface, the presence
of junction and streaming potentials, rapid variation

0010-9312/96/000137/$5.00+$0.50/0
1996, NACE International

609

CORROSION SCIENCE

FIGURE 1. ATR electrode construction.

with time of the moisture of the concrete near the


electrode tip, and overall heterogeneity of the medium. As a result, variations of > 100 mV are not
uncommon when a surface electrode is moved a few
millimeters or when some of the external concrete
surface is chipped away. Ideally, an embedded electrode should be placed close to the steel surface.1
An embedded electrode should satisfy stability
and reproducibility requirements, be adequately
small, and show little sensitivity to changing variables inherent to concrete service.2
Considerable attention has been given to alternative RE for steel in concrete, reflecting the need for
practical and low-cost approaches.1-8 The objective of
the present study was to establish the behavior of
commercially available activated Ti rod (ATR) covered
with mixed-metal oxides when used as material for
RE embedded in concrete.9-12 This type of electrode
has been used successfully for short-term steel polarization measurements, in which low RE impedance
at moderate to high frequencies and excellent stability were achieved routinely.13-14 This investigation
addressed performance of the material for short- and
long-term absolute potential determination, with
emphasis on the open-circuit behavior of the electrodes. The polarizability of the ATR electrodes, the
effect of sodium chloride (NaCl) additions, and the
O activity also were studied.

EXPERIMENTAL PROCEDURE

Reference Electrodes
The RE examined were made from ATR stock
(3-mm [0.12-in.] diam) produced commercially for
permanent impressed-current anodes in cathodic
protection of reinforcing steel in concrete. The
surface of the ATR was covered with a precious
mixed-metal oxide applied using a proprietary process. Energy-dispersive x-ray spectroscopy (EDS) of
the ATR coating identified Ir as the main metallic
species, with others in lower quantities (e.g., Ru, Ta,
and Si). The ATR surface had a flat black appearance.
ATR electrodes (Figure 1) were prepared by cutting the rod in 5-cm (2-in.) segments, drilling a
1.5-mm (0.06-in.)-diam hole 6 mm (0.24 in.) deep at

Trade name.

610

one end, and crimping the stripped end of an insulated braided Cu cable in the hole. The crimped and
opposite ends of the bar were covered by beads of
epoxy resin, leaving a ~ 4-cm (1.57-in.)-long ATR
segment exposed. Surface area was ~ 4 cm2 (0.62 in.2).
For comparison, graphite RE as used in other
investigations were prepared using 5-mm (0.20-in.)diam as-received commercial rod stock for dry-cell
batteries.15 A Cu wire was inserted in a hole in the
end of a graphite rod 5 cm (2 in.) long, and the junction was sealed with epoxy.

Tests in Liquid Solutions


ATR electrodes were placed in buffer solutions of
pH 4 (0.05 M potassium biphthalate buffer), pH 7
(0.05 M potassium phosphate monobasic-sodium
hydroxide buffer), and pH 10 (0.05 M potassium
carbonate-potassium borate-potassium hydroxide
buffer), as well as in saturated calcium hydroxide
(Ca[OH]2, pH ~ 12.5), and simulated concrete pore
solution (SPS) with pH ~ 13.5. The SPS was made by
mixing 7.4 g of sodium hydroxide (NaOH) and 35.6 g
of potassium hydroxide (KOH) per liter of saturated
Ca(OH)2 solution. Unless indicated, all solutions were
at room temperature and exposed to laboratory air.
Excess of Ca(OH)2 was used in the two high-pH solutions to compensate for a thin layer of carbonation
products on the surface in contact with air.
The effect of deaeration using the same solutions
was investigated by placing the ATR electrode in a
polyethylene cell containing the test solution,
through which air or commercial-purity nitrogen gas
(N2) was bubbled at room temperature. The cell contained a moist wooden plug through which potential
measurements could be made against an external
saturated calomel electrode (SCE). A graphite electrode with a connecting wire was placed inside the
deaeration cell for comparisons.
Susceptibility of the open-circuit potential (Eoc) of
ATR electrodes in saturated Ca(OH)2 to the presence
of Cl was evaluated using progressive additions of
NaCl over 3 days until reaching 3.5 wt%. Potential
measurements were made after each addition of NaCl
until stable values were obtained.
All potential measurements in concrete and liquid solutions were taken with electronic voltmeters
with input resistance of 200 M (M.C. Miller, Model
LC-4). No potential drift, indicative of loading effects
by input resistance of the voltmeter, was observed.
Cyclic polarization (CP) tests were performed
with ATR electrodes and graphite electrodes for comparison. The scan rate was 0.1 mV/s. Ohmic drop
effects were negligible in the range of current densities obtained. Potential sweeps went from Eoc to
300 mV in the cathodic direction. The scan direction
was reversed until the initial Eoc was reached. An
activated Ti wire mesh was used as the counter electrode (CE), an ATR electrode as the working electrode

CORROSIONAUGUST 1996

CORROSION SCIENCE

TABLE 1
Concrete Mix and Properties

Series

Cementitious Material(A)

Cementitious
Content
(kg/m3)

Water-toCementitious
Ratio

Admixtureto-Cement
Ratio

Measured Strength
28-day
(MPa)

Prisms
A
B
C
D
E
F

100% PC
80% PC + 20% FA
70% PC + 30% FA
72% PC + 20% FA + 8% SF
62% PC + 30% FA + 8% SF
100% PC

333
444
444
444
444
360

0.55
0.41
0.43
0.39
0.39
0.41

NA(B)
0.25
0.43
0.39
0.61
NA

37
46
42
58
56
34(C)

Concrete Columns
BB(D)
80% PC + 20% FA
DD
75% PC + 20% FA + 5% SF

302
302

0.45
0.45

0.25
0.33

34
35

(A)

PC = Portland cement, Type II; FA = fly ash, Class F; and SF = silica fume.
NA = not available.
(C)
Design, measured not available.
(D)
Three sets (BB1, BB2, and BB3) of identical columns.
(B)

(WE), and another ATR electrode as the RE. The


same kind of cell was used for electrochemical impedance spectroscopy (EIS) tests in liquid solutions.
The EIS tests were conducted in the frequency range
10 kHz to 100 mHz using a CAPCIS-VOLTECH 2001
frequency response analyzer (FRA), and in the
200 Hz to 1 mHz range using a custom system.16
Signal amplitude typically was < 10 mV.
Galvanostatic tests were performed with ATR
electrodes in saturated Ca(OH)2 solutions. A cathodic
direct current (DC) of 0.3 A was applied to the ATR
electrodes for 1 week. After this, the current was
interrupted, and the ATR electrodes were allowed to
recover to their initial Eoc.

Tests in Concrete
The results were limited to ATR electrodes in
moist concrete, using external conventional RE in
temporary contact with permanently moist concrete
surfaces or salt water surrounding the concrete.
Embedded ATR electrodes were cast in the center
of reinforced concrete prisms (15 cm by 7.5 cm by
30 cm [6 in. by 3 in. by 12 in.]) as part of a separate
research program. Each prism had two pieces of no.
4 rebar (1.25 cm [0.5 in.] diam, 20 cm [8 in.] length)
embedded in the concrete. The connecting cable of
the ATR electrode protruded from the concrete surface, allowing easy external contact. The prisms were
made of six types of concrete (Table 1).
Four prisms of each concrete formulation were
kept nearly saturated with fresh water inside plastic
bags following an initial 28-day curing period in a
moist chamber. The ATR electrodes were calibrated
periodically against a CSE that touched the external
concrete surface at one corner of the prism. There
was no evidence of significant potential errors from
corrosion macrocell ohmic drops.

CORROSIONVol. 52, No. 8

EIS tests were performed in the prisms by connecting the ATR electrode as the WE and using one of
the embedded rebar as the CE. The remaining rebar
served as a temporary RE. The tests were performed
in the 200 Hz to 1 mHz range.
ATR electrodes also were cast in reinforced
concrete columns that were part of an investigation
of corrosion in marine substructure conditions.17
The two concrete mix designs used for the columns
are shown in Table 1. The columns were 12.5 cm by
12.5 cm by 250 cm (5 in. by 5 in. by 100 in.) and
stood vertically with the lower 60 cm (24 in.) immersed in 5% NaCl solution. A total of 12 columns
were used. The potentials of ATR electrodes in the
center of the immersed portion of the column were
monitored against an SCE in contact with the
NaCl solution. Potential gradients from corrosion
macrocells were judged to be negligible in the specimens relevant to these tests.

RESULTS

Liquid Solutions
Potentials Figure 2 shows the potential evaluation with time of a typical set of ATR electrodes
immersed in liquid solutions. A trend line was drawn
to each data series, showing a tendency to stabilize
after ~ 1 week of immersion. The trend value corresponding to the seventh day of exposure was
considered representative of the electrode potential
for that solution and electrode set. The average of
those values is given in Table 2. Table 2 also shows
the maximum and minimum potential values of several electrode test sets. The potentials, taken against
a SCE, were converted to the SHE scale.
Effect of Deaeration Figure 3 shows a typical
result from experiments to determine sensitivity of

611

CORROSION SCIENCE

TABLE 2
Eoc (SHE) of ATR Electrodes at the Seventh Day
Following Immersion in Various pH Solutions

No. tests(A)
Maximum
Minimum
Average
(A)

pH 4
(mV)

pH 7
(mV)

pH 10
(mV)

pH 12.5
(mV)

pH 13.5
(mV)

5
649
541
560

5
456
373
392

5
331
118
251

5
275
48
143

2
149
134
141

Each test involved a new specimen and liquid solution for each pH.

FIGURE 3. Potential evolution of graphite and ATR electrodes during


a deaeration experiment in saturated Ca(OH)2 solution (pH ~ 12.5).

FIGURE 2. Typical potential evolution with time of ATR electrodes


following immersion in solutions of various pH values.

the ATR electrode to oxygen partial pressure (PO2)


variations. The control graphite electrode showed an
effect when switching from natural aeration at normal air pressure (PO2 = 0.2 atm) to bubbling
commercial-purity N2 gas (PO2 < 103 atm) through the
test cell and then bubbling air. During deaeration of
the Ca(OH)2 solution, the potential of the graphite
electrode shifted in the less noble direction (shortterm erratic behavior sometimes was observed) until
a plateau was reached at a value ~ 50 mV lower than
the initial aerated condition. Upon air bubbling, the
graphite electrode returned to the initial natural
aeration potential. The graphite electrodes in aerated
solutions showed rest potentials consistent with
those in the literature.10 In contrast, the change of
potential of the ATR electrode upon deaeration and
reaeration typically was several times smaller than
that of the graphite electrode. Figure 4 summarizes
results of duplicate tests using the solutions at various pH values. The potential shift was the average of
the downward and upward potential change experi-

612

enced during the deaeration and reaeration steps,


respectively.
Effect of Cl Addition Two sets of triplicate
specimens in saturated Ca(OH)2 were used to examine the effect of Cl addition at room temperature on
the potential of the ATR electrodes. No clear trends
were observed for individual specimens upon addition of 3.5 wt% NaCl. The overall average potential
changed only ~ 35 mV (in the more positive direction)
after the 3-day addition sequence.
CP Tests Results of the CP tests are illustrated
in Figures 5 and 6 for the ATR and graphite electrodes, respectively. Duplicate experiments provided
essentially the same results. At 0.1 mV/s, the ATR
electrodes demanded similar current densities when
polarized cathodically from Eoc by roughly the same
amount at all pH values examined. All ATR electrode
polarization curves showed significant hysteresis.
The graphite electrodes showed less polarization
from Eoc at pH 10 and pH 12.5 than at pH 7 at the
same current density. The curves showed much less
hysteresis than the ATR electrodes. Ignoring hysteresis at the highest current densities, the curves
showed apparent cathodic polarization slopes on the
order of 100 mV/decade. The graphite electrodes
demanded much lower current densities than the
ATR electrodes to produce a given deviation from Eoc.

CORROSIONAUGUST 1996

CORROSION SCIENCE

FIGURE 5. CP curves of ATR electrodes in solutions of various pH


values (scan rate = 0.1 mV/s). The circle indicates start of the scan.
Specimen area = ~ 4 cm2 (0.62 in.2).

FIGURE 4. Average potential shift during deaeration/reaeration


experiments in liquid solutions.

EIS Figure 7 illustrates the typical appearance


of the high-frequency portion of the Nyquist diagram
for ATR electrodes in liquid solutions (pH 4 to 12.5).
Instrumentation limitations introduced phase-angle
errors at frequencies approaching 10 kHz. Only results for frequencies < 3 kHz were considered. The
diagram, obtained in the saturated Ca(OH)2 medium,
was similar in shape and size to those in the other
solutions. The high-frequency limit impedance value
(~ 10 ) agreed well with the value expected from the
test solution conductivities and cell dimensions used.
The high-frequency arc corresponded to a time constant of ~ 102 s and had a small (4 to 10 )
apparent diameter that did not vary greatly from
solution to solution. The lower-frequency response
(Figure 8, Curve L) was dominated by a constant
phase-angle element (CPE). The CPE had a high admittance coefficient (Yo ~ 0.02 1 sn to 0.07 1 sn)
and a frequency exponent 0.7 n 0.9. A similar
range of parameters was observed for all the test
solutions (Table 3). No other admittance elements
could be identified clearly at the low frequencies.
Galvanostatic Tests Results of the
galvanostatic tests are shown in Figure 9. The potential of the ATR electrodes shifted ~ 60 mV more
negative and approached a limiting value after 1 to 2
days. The potential recovered toward the initial values upon interruption of the current. The system
exhibited a similar apparent time constant (~ 1 day)
during current application and recovery.

Concrete
Potentials Figure 10 shows the ATR electrode
potential (converted to the SCE scale from the actual

CORROSIONVol. 52, No. 8

FIGURE 6. CP curves of graphite electrodes in solutions of various


pH values (scan rate = 0.1 mV/s). Circle indicates start of the scan.
Specimen area = ~ 8 cm2 (1.24 in.2).

CSE measurements by adding 75 mV) as a function


of time for each concrete prism series over a 500-day
test interval. Plotted potentials were the average for
four specimens (except for Series A, in which only
one of the four probes remained functional by the
end of the test because of an early destructive
polarization experiment). The sequential reading-toreading variability in Figure 10 likely reflected
sensitivity of the measurements to positioning of the
external RE tip on the concrete surface more than
shorter-term instability of the ATR electrodes. As part
of an independent investigation involving the prisms,
the potential of steel rods embedded in the concrete
was measured periodically against the ATR electrodes. In most cases, the short-term potential
difference between the steel rods and ATR embedded
in the same concrete prism was highly stable. Varia-

613

CORROSION SCIENCE

FIGURE 7. Typical impedance diagram for an ATR electrode in


saturated Ca(OH)2 solution (pH ~ 12.5) from 3 kHz to 100 mHz.
Specimen area = ~ 4 cm2 (0.62 in.2).

FIGURE 9. Eoc (vs SCE) of ATR electrodes in saturated Ca(OH)2


solution (pH ~ 12.5).

tions typically were on the order of 1 mV/h. This


characteristic, which greatly facilitates polarization
measurements of steel, was the main reason for using the ATR electrodes in the investigation for which
the prisms were built.
Longer-term drift of the measured ATR electrode
potentials was evident in some of the tests. The rate
of drift for Series A through C and Series F became
less pronounced after ~ 100 days of exposure. The
drift was pronounced and continuous in concrete
Series D and E, both with silica fume addition, and
showed ATR electrode potentials ~ 600 mV less noble
than those of the other series by day 500. The difference tended to be greater for Series E, which had the
highest pozzolanic content (Table 1). The addition of
fly ash alone (Series B and C) did not seem to result
in pronounced long-term drift or in significant differ-

614

FIGURE 8. Curve L: Typical impedance diagram for an ATR electrode


in saturated Ca(OH)2 liquid solution (pH ~ 12.5) from 200 Hz to
1 mHz. Curve C: Typical impedance diagram for ATR electrodes in
concrete, prism specimen series F exposed 500 days. Specimen
area in both curves = ~ 4 cm2 (0.62 in.2).

entiation from the plain cement series (Series A and


F). The water-to-cement variation between Series A
and F also had no significant effect on ATR potentials.
Figure 11 shows results of nearly 900 days of
testing with the test columns. Potentials showed
good long-term stability after a systematic drift of
~ 150 mV over the first year of exposure. The potentials near 900 days tended to be similar to the
average for Series A through C and Series F near
500 days in the prism specimens.
EIS Low-frequency EIS tests performed with
ATR electrodes in selected concrete prism specimens
(Series A and F, but not Series D and E) revealed a
behavior similar to that in the liquid solutions
(Figure 8, Curve C) except that solution resistance
was higher because of greater resistivity of the concrete. The admittance coefficient and frequency
exponent values were comparable to those in the
liquid solutions (Table 3).
Preliminary examination of the impedance
response of the ATR electrodes in Series D and E
revealed a finite polarization resistance with magnitude 20 k. This behavior is subject to a
continuing investigation.

DISCUSSION
Figure 12 summarizes the liquid solution and
concrete potential measurements. Results are
overimposed on a simplified Pourbaix diagram of the
Ir-H2O system. The pH dependence of the ATR electrode potential in the liquid solutions was evident.
The slope of the line joining the results was 45 mV.

CORROSIONAUGUST 1996

CORROSION SCIENCE

FIGURE 10. Eoc (vs SCE) of ATR electrodes in the concrete prisms.
Concrete compositions are given in Table 1.

FIGURE 11. Eoc (vs SCE) of ATR electrodes in the concrete columns.
Concrete compositions are given in Table 1.

This slope and the position of the line roughly


approximated the behavior expected from the Ir-H2O
system, shown by the dashed line.18 Ir was a dominant component of the proprietary coating of the ATR
electrode. The observed pH dependence of the potential and its low sensitivity to PO2 suggested the
potential of the ATR electrode was determined (at
least on first approximation) by an equilibrium between Ir oxides at the electrode surface of the type:

CP test results for the ATR electrodes were consistent with the presence of a very large apparent
interfacial capacitance, which dominated the polarization behavior at the scan rate chosen. During the
forward cathodic polarization scan, the interface
stored charge, as determined primarily by the scan
rate and the effective interfacial capacitance (which
may have reflected mostly electrode processes that
involve conversions between oxide forms).19-20,25 As
the scan rate reversed, the charge was released,
causing the hysteresis. Similar behavior has been
reported for other mixed-metal oxide systems.25
Other electrode processes, such as hydrogen
evolution or oxygen reduction, appeared to be secondary since the general shape of the curves and
magnitude of the currents for a given deviation from
the initial Eoc was roughly the same at all pH values
examined.
These observations agreed with results from the
EIS tests, which revealed a CPE behavior with a large
admittance factor (Yo). A CPE subject to a polarization potential ramp develops, unlike a true capacitor,
a current that increases with time (as observed in the
sloping portion of the downward polarization
curves).26 Nevertheless, the electrode behavior can be
considered approximately as that of a large capacitor.
The apparent capacitance of the ATR electrodes can
be computed using the difference in current values
between the downward and upward portions of the
CP curves or the magnitude of the CPE impedance at
low frequencies.26 These estimates resulted in apparent interfacial capacitances on the order of 101 F to
102 F (specimen area ~ 4 cm2 [0.62 in.2]) for the time
domain range of interest (minutes to hours).

Ir2O3 + H2O = 2IrO2 + 2H+ + 2e

(1)

Equation (1) corresponds to the dashed line in


Figure 12. This reaction was proposed in the interpretation of experiments with electrodes of unmixed
Ir oxides on a Ti substrate, showing that the behavior
in purer systems approaches closely the ideal dependence with pH.19-20
Reactions involving dissolved O2 directly (which
affect the graphite electrode potential [Figure 3])21-22
appeared to play only a secondary role in determining the ATR electrode potential in the short term.
This lack of sensitivity, if confirmed over long time
periods, would be a desirable feature for applications
where O2 concentration of the concrete could vary
significantly during service (such as in periodically
immersed concrete structures).23
Potentials of the ATR electrodes in concrete (values from the end of each long-term exposure period)
also are displayed in Figure 12. The potential of the
ATR electrodes in most concrete mixes fell roughly on
the line which joined the results from the liquid solution tests, assuming that the pH of concrete pore
solutions was somewhere in the 12.5 to 13.5 range.24

CORROSIONVol. 52, No. 8

615

CORROSION SCIENCE

TABLE 3
Low-Frequency Parameters of ATR Electrodes in Liquid Solutions and in Concrete
pH 4
Yo(B)
No. tests
Maximum
Minimum
Average
(A)
(B)

pH 7
n

Yo

6
0.045
0.028
0.034

pH 10
n

Yo

5
0.77
0.70
0.73

0.067
0.033
0.049

Yo

6
0.83
0.81
0.82

0.053
0.042
0.047

Concrete(A)

pH 12.5
n

Yo

6
0.85
0.77
0.84

0.030
0.015
0.022

n
6

0.81
0.69
0.73

0.041
0.016
0.023

0.83
0.79
0.81

Average of selected specimens from Series A and F (Table 1).


Yo values in 1 sn.

FIGURE 12. Summary of potential measurements of ATR electrodes


in liquid solutions (circles, average values from Table 2) and concrete
(values at end of each exposure period), with partial Pourbaix
diagram for the Ir-water system.

The high equivalent capacitance results in low


electrode impedance at other than low frequencies,
which is advantageous when conducting steel EIS or
polarization resistance tests. The capacitive behavior
also provides tolerance to brief current delivery to the
electrode, which may result from accidental improper
connection to a control system. For example, computation of the time-domain response of the ATR electrodes using the procedure of Sags, et al., showed
that the electrodes could sustain cathodic currents
on the order of 5 A for 1 min, without polarizing by
more than a few millivolts.26

616

Stabilization of the ATR electrode potential with


time in the galvanostatic tests exemplified in Figure 9
indicated a polarization resistance of ~ 200 k (based
on the applied current and resulting long-term potential change). This suggested the electrodes could
resist adequately the long-term effects of the small
bias currents (typically < 1 nA) required by common
electrochemical control instrumentation.
Except for Series D and E, the long-term stability
of the ATR electrodes in moist concrete was reasonable after a period from a few months to 1 year of
exposure. However, there was no fixed potential offset with respect to the CSE. Calibration of individual
ATR electrodes with respect to an external electrode
is essential if reliable absolute potential measurements are required. The ATR electrode potential
appeared not to be sensitive to Cl contamination of
the immediate environment. In general, electrodes
assembled as shown in Figure 1 proved reliable, with
connection failures or unaccounted sudden potential
excursions affecting a small fraction of hundreds of
specimens in corrosion laboratory and field use.
The potential of ATR electrodes in concrete Series D and E (concrete prisms) deviated notably from
the rest, drifting toward increasingly negative values
throughout the test procedure. Series D and E contained the largest active pozzolanic additions of the
concretes tested and, therefore, could be expected to
have the lowest pore solution pH values. However,
the potential deviation from the electrodes in the
other concretes was too large and was in the wrong
direction to ascribe the effect to differences in the pH
of the solution surrounding the ATR electrodes. The
preliminary EIS observation of a low polarization
resistance in these series, together with the development of very negative potentials, suggested corrosion
reactions were taking place, possibly affecting the Ti
substrate. Causes of corrosion of the ATR substrate
may have included the presence of unidentified trace
components from the silica fume used in Series D
and E, but not in Series A through C and Series F.
However, Series DD in the concrete columns also
contained silica fume and did not show anomalous
behavior. Additional monitoring of these systems is
in progress.

CORROSIONAUGUST 1996

CORROSION SCIENCE

CONCLUSIONS
The ATR electrode potentials in liquid solutions
approximated the absolute values and pH dependence of the reaction Ir2O3 + H2O = 2IrO2 + 2H+ + 2e.
Unlike graphite electrodes, the ATR electrode potential in liquid solutions showed little short-term
sensitivity to variations in PO2 at neutral and high pH
values.
The ATR electrode showed excellent short-term
stability in concrete. This behavior, together with low
impedance at moderate to high frequencies, makes
embedded ATR electrodes useful for steel polarization
measurements.
The ATR electrode showed reasonable long-term
(2- to 3-year tests) stability in most of the concretes
tested under moist conditions with various amounts
of pozzolanic addition and in the presence of chlorides from the surrounding environment.
In concretes with low to moderate pozzolanic addition, the ATR electrode potentials approached values
expected from measurements in liquid solutions and
the estimated pH of the pore water. In two concretes
with high fly ash and silica fume additions as cement
replacement, the potentials showed anomalous behavior, drifting by several hundred millivolts in the
negative direction over a 2-year period.
Results underscored the need to calibrate ATR
electrodes against an external conventional electrode,
especially when dealing with new concrete formulations using silica fume.
The ATR electrode behaved as a pseudocapacitor
that tolerated brief current application without significant polarization.
There was evidence that a finite polarization resistance of ~ 200 k in 4 cm2 (0.62 in.2) electrodes
limited the extent of potential excursion that may be
caused by long-term application of small currents.

ACKNOWLEDGMENTS
The authors acknowledge support of the Consejo
Nacional de Ciencia y Tecnologa (CONACYT), Mxico,
contract no. 0527-A9109, E 120-3052; the National
Science Foundation (NSF), contract no. INT9217441;
and the Florida Department of Transportation. Opinions and findings are those of the authors and not
necessarily of the supporting organizations. The authors acknowledge the assistance of R.G. Powers
(FDOT), M. Balancn, L. Diaz, F. Presuel, and J.
Boodts (Universidade de So Paulo).
REFERENCES
1. J.E. Bennett, T.A. Mitchell, Reference Electrodes for Use with
Reinforced Concrete Structures, CORROSION/92, paper no.
191 (Houston, TX: NACE, 1992).

CORROSIONVol. 52, No. 8

2. F.J. Ansuini, J.R. Dimond, Factors Affecting the Accuracy of


Reference Electrodes, CORROSION/94, paper no. 323 (Houston, TX: NACE, 1994).
3. K. Gurusamy, M.P. Geoghegan, Long-Term Performance of
Embedded Reference Electrodes for Cathodic Protection and
In-Situ Monitoring of Steel in Concrete, in Corrosion of Reinforcement in Concrete, eds. C.L. Page, K.W.J. Treadaway,
P.B. Bamforth (London, U.K.: Elsevier Applied Science, 1990),
p. 333.
4. C.E. Locke, C. Dehghanian, MP 18, 2 (1970): p. 70.
5. H. Arup, B. Sorensen, A New Embeddable Reference Electrode
for Use in Concrete, CORROSION/92, paper no. 208 (Houston,
TX: NACE, 1992).
6. S.A. White, G.E. Stoner, S.R. Taylor, Electrochemical Sensors
to Monitor the Corrosion of Reinforcing Steel in Concrete,
CORROSION/89, paper no. 124 (Houston, TX: NACE, 1989).
7. C. Dehghanian, C.R. Root, C.E. Locke, Review of Embeddable
Reference Electrodes for Portland Cement Concrete, CORROSION/81, paper no. 45 (Houston, TX: NACE, 1981).
8. F. Fernandez, C. Lasarte, O. de Rincn, Comparative Study
Between Graphite and Cu/CuSO4 Electrodes for Corrosion
Detection of Reinforcing Steel in Concrete, II Workshop on
Industrial Coatings, Maracaibo, Venezuela, 1990 (Maracaibo,
Venezuela: Universidad Naional del Zulia, 1990).
9. C.J. Mudd, G.L. Mussinelli, M. Tettamanti, P. Pedeferri, Cathodic Protection of Steel in Concrete with Mixed-Metal Oxide
Activated Titanium Anode Net, CORROSION/88, paper no. 229
(Houston, TX: NACE, 1988).
10. C.J. Mudd, G.L. Mussinelli, M. Tettamanti, P. Pedeferri, New
Developments in Mixed-Metal Oxide Activated Net for Cathodic
Protection of Steel in Concrete, CORROSION/89, paper no.
168 (Houston, TX: NACE, 1989).
11. H.C. Schell, D.G. Manning, MP 24, 7 (1985): p. 18.
12. J.E. Bennett, R.G. Powers, B.L. Martin, Cathodic Protection
System for a Steel-Reinforced Concrete Structure, U.S. patent
no. 5,098,543 (1992).
13. A. Aguilar, A. Sags, R. Powers, Corrosion Measurements of
Reinforcing Steel in Partially Submerged Concrete Slabs, in
Corrosion Rates of Steel in Concrete, STP 1065, eds. N.S.
Berke, V. Chaker, D. Whiting (Philadelphia, PA: ASTM, 1990),
p. 66.
14. A. Sags, H. Perez-Duran, R. Powers, Corrosion 47 (1991):
p. 884.
15. O.T. de Rincn, A.R. de Carruyo, D. Romero, E. Cuicas, A
Three-Year Evaluation of the Effect of the Oxidation Products of
Aluminum Used as a Sacrificial Anode in Reinforced Concrete
Structures, CORROSION/91, paper no. 128 (Houston: TX,
NACE, 1991).
16. A. Sags, A System for Electrochemical Impedance Corrosion
Testing Using a PC with Isaac-Cyborg Interface, CORROSION/
88, paper no. 104 (Houston, TX: NACE, 1988).
17. A.A. Sags, S.C. Kranc, A.K.M. Al-Mansur, S.L. Hierholzer,
Factors Controlling Corrosion of Steel-Reinforced Concrete
Substructure in Seawater, Report FL/DOT/RC/0537-3523,
Natl. Technial Information Service, Springfield, VA (1994).
18. M. Pourbaix, Atlas of Electrochemical Equilibria in Aqueous
Solutions (Houston, TX: NACE, 1974), p. 374.
19. K. Kinoshita, M.J. Madou, J. Electrochem. Soc. 131 (1984):
p. 1,089.
20. S. Ardizzone, A. Carugati, S. Trasatti, J. Electroanal. Chem.
126 (1981): p. 287.
21. F.J. Ansuini, J.R. Dimond, Long-Term Stability Testing of
Reference Electrodes for Reinforced Concrete, CORROSION/
94, paper no. 295 (Houston, TX: NACE, 1994).
22. A.A. Sohanghpurwala, W.T. Scannell, A. LaConti, Improvement
in Graphite Reference Cell for Reinforced Concrete, CORROSION/94, paper no. 307 (Houston, TX: NACE, 1994).
23. R.D. Browne, Mechanisms of Corrosion of Steel in Concrete in
Relation to Design, Inspection, and Repair of Offshore and
Coastal Structures, in Performance of Concrete in Marine
Environment, SP-65, ed. V.M. Malhotra (Detroit, MI: ACI,
1980), p. 169.
24. A. Moragues, A. Macias, C. Andrade, Cem. Concr. Res. 17
(1987): p. 173.
25. B.E. Conway, J. Electrochem. Soc. 138 (1991): p. 1,539.
26. A.A. Sags, S.C. Kranc, E.I. Moreno, Corros. Sci. 37 (1995):
p. 1,097.

617

S-ar putea să vă placă și