98
DOI: 10.1002/maco.201005707
Materials and Corrosion 2011, 62, No. 2
Ten-year results of galvanic sacrificial anodes in steel
reinforced concrete
G. Sergi*
Zinc sacrificial anodes have been included in patch repairs to steel reinforced
concrete structural elements suffering from corrosion since the mid-1990s. A
number of these anode-containing repairs have been monitored with time. One
of the first monitored sites was of a locally repaired cross beam of a bridge
structure in Leicester, UK, which has now completed 10 years since its original
repair and anode installation. This paper reviews the performance of the anodes
installed at the Leicester site in terms of anode current output and steel
reinforcement polarisation and corrosion rate over the period. It also presents
results of analysis of recovered anodes exposed for 10 years which still show
electrolyte continuity, uniform consumption of the zinc and coherent encasing
mortar. The knowledge gained from the 10 year results has enabled the
development of new, higher current output anodes, which are now trialled in
this and other sites.
1 Introduction
Chloride induced corrosion of steel reinforcement in concrete
structural elements is a major problem in many countries.
Chlorides can be introduced into the concrete via deicing salts or
seawater. This leads to localised breakdown of the normally
passive steel reinforcement in the form of pitting corrosion.
Patch repairs of only the damaged concrete are rarely, if ever,
successful where chlorides are present in sufficient concentration
in the remaining concrete. Under such conditions, treatment of
the symptoms merely serves to send the corrosion cell into areas
adjacent to those that have been repaired. This phenomenon is
known as the incipient anode or ring effect [1–3]. Consequently,
traditional repair techniques, where chlorides are present above
certain levels, are only likely to be effective for a limited period
after which time further repairs will be necessary.
It is essential that some form of intentional ‘cathodic
prevention’ be reinstated within the patch repair region so that the
adjacent areas remain cathodic and corrosion initiation is
prevented. This can be accomplished by embedding sacrificial
anodes around the perimeter of the repair patch.
Galvanic anodes of an appropriate design were developed in
the late 1990s. Their success under laboratory conditions and in
small-scale controlled trials was demonstrated over a period of
about a year [1, 2]. The puck-like anode was produced from zinc
metal encased in a specially formulated porous cementitious
G. Sergi
Vector Corrosion Technologies Ltd., 3 Bodmin Close, Park Hall, Walsall
WS5 3HZ (UK)
E-mail: georges@vector-corrosion.com
ß 2011 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
mortar (Fig. 1) saturated with lithium hydroxide (pH > 14.5).
Such an environment, with a reservoir of excess LiOH
maintaining a constantly high pH, was shown to sustain the
zinc in an active condition producing soluble zinc corrosion
products that do not stifle the corrosion process of the metal [1].
Several trials were set up as part of the repair regime
undertaken for the rehabilitation of structural elements over
the last 10 years both in the UK and abroad. The anodes were used
for two types of application, viz., around the perimeter of patch
repairs in a cathodic prevention mode, and in a grid configuration
in susceptible areas of reinforced concrete in a corrosion control
mode. Some of the results have been reported elsewhere [3].
Current densities of the order of 0.8–10 mA/m2 of steel surface
were recorded depending on the type of application.
The oldest site trial, at a bridge in Leicester, UK (Fig. 2), has
completed 10 years of life this year, a milestone, as the anodes
were designed for a minimum life of 10 years. It is the results of
this trial that are reported in this publication.
2 Setting up of the trial
As verification for the performance of the anodes in actual
structures, a total of 12 commercial anodes were installed in an
otherwise conventional patch repair [4] on spalled and cracked
areas of a beam section of a bridge in Leicester, UK (Fig. 2). The
trial was set on the soffit of a section of the beam between
Columns 6 and 7 on the west pier (Fig. 3). The performance of
these anodes was monitored with time. This formed part of a
repair scheme for the whole bridge which contained two
abutments and two piers each consisting of a long cross-beam
sitting on eight columns. The overall repair system was made
wileyonlinelibrary.com
www.matcorr.com
Materials and Corrosion 2011, 62, No. 2
Figure 1. Puck-type anode used for the trial
Figure 2. Installation of anodes within the repaired area of a beam
showing also control box and wiring
up of patch repairs containing anodes around their perimeter
and included impressed current cathodic protection with
discrete anodes at the more deteriorated areas at the top of the
abutments.
Galvanic sacrificial anodes in steel reinforced concrete
The extent of deterioration of the beam section that was
chosen for the trial is summarised in Figs. 4 and 5 which show the
chloride contamination at increasing depths and the potential
map prior to repair. The chloride concentration at the depth of the
steel was over 2% by weight of cement in location A, close to the
most negative recorded potential within the area. In location B,
the chloride concentration was just below 1%. Both concentrations are overall higher than the range of chlorides found in this
particular part of the pier outside the area of repair and the mean
for the west pier but lower than the maximum concentration
found in the pier (Fig. 4).
The potential map suggested the presence of corrosion
activity of the steel reinforcement which was accompanied by
some cracking and delamination. The range of potentials within
the test area is similar to the whole west pier (Fig. 6), a significant
proportion of the potentials indicating the presence of some
corrosion activity. The concrete was broken out to behind the steel
and to beyond any corroding steel as shown in Figs. 2 and 3. No
attempt was made to ensure that chloride contaminated concrete
areas were removed, the level of chloride in the adjacent unremoved concrete and at depth beyond the steel was known to be
significant, as suggested by Fig. 4. These are classic conditions
for the formation of new anodic sites at the periphery of a
conventionally repaired area causing corrosion of the steel and
cracking of the concrete within a few years [1, 2].
A total of eight anodes were inserted around the perimeter of
the left hand repair area and four anodes were inserted in the
right hand area at between 600 and 700 mm centres (Figs. 2
and 3). The anodes were specially adapted to enable monitoring.
A single wire from each anode was connected to a control
box so that connection could be made individually to the
steel reinforcement via the box. All other similarly cracked,
delaminated or spalled areas of the pier were likewise repaired
with anodes positioned at approximately 600–700 mm centres
around the perimeter of each area. These were connected directly
to the steel reinforcement using the four tie wires available [4]
(Fig. 1).
Figure 3. Schematic of the repaired area of the beam soffit
www.matcorr.com
ß 2011 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
99
Sergi
Materials and Corrosion 2011, 62, No. 2
Figure 4. Chloride concentration profiles at positions A and B within the subsequently repaired area of the beam compared to around the beam
(no. 1–4) and the maximum and mean for the whole west pier prior to repair
Monitoring of the 12 anodes was by a combination of current
output measurements for each installed anode and, on occasions,
a depolarisation potential over 4 or 24 h periods after disconnection of the anodes. A further measurement was the ‘instant-on’
current output of each anode following reconnection to the steel.
Monitoring started in April 1999.
3 Results and discussion
The total estimated current output of each anode could be
converted to a charge as shown in Table 1. Using Faraday’s law,
and assuming an efficiency and utilisation rate for the metal, the
total consumption of the zinc metal could be estimated. For an
efficiency of around 85% the level of consumed metal is as shown
in Fig. 8. By simple extrapolation, the service life of each anode
can be determined. According to these results, a range between
24 and 37 years service life can be achieved for these anodes with
60 g zinc mass.
The 10 year results of the current output of each anode are
presented in Fig. 7. They indicate a variable current depending
on the moisture content in the concrete but primarily on
temperature (see also next section). For example, the same anode
could generate up to 400–600 mA of current during hot periods
and less than 100 mA during cold spells. Corrosion of the steel
is expected to have similarly varying corrosion rates so that the
current output of the anodes is thought to be self-regulating,
producing higher levels when the steel is corroding most.
0
-100-0
-200--100
-300--200
0.5
-400--300
0
0.5
1
1.5
2
Grid (m)
100
1
2.5
Grid (m)
Figure 5. Potential map of the left repaired area of the beam prior to
repair (mV vs. Cu/CuSO4)
ß 2011 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
Figure 6. Range of potential values recorded within the area of the
soffit of the beam chosen for repair compared to all the values recorded
over the whole west pier
www.matcorr.com
Materials and Corrosion 2011, 62, No. 2
Galvanic sacrificial anodes in steel reinforced concrete
Figure 7. Current output of the 12 individual anodes and the mean
output with time
Confirmation of the approximate efficiency was provided by
assessing removed anodes from an adjacent patch repair in the
west pier (Fig. 9). These removed anodes were installed 10 years
ago and the exposure conditions were similar to the 12 monitored
anodes. Exposure of the zinc core revealed a thickness of zinc-rich
pseudomorphic corrosion product at the interface of the zinc and
encasing mortar. No more than 25–30% of the original zinc metal
was lost, a level within the range calculated for the 12 monitored
anodes (Fig. 8), suggesting that the assumed efficiency of 85% is
likely to be a fair estimate. There was evidence also that the pores
of the encasing mortar extending several millimetres away from
the zinc/mortar interface were partly filled with white zinc
oxide and zinc hydroxide corrosion products. This supports the
assertion that the zinc corrosion product remains soluble, owing
to the pH of the pore solution exceeding 14 as a result of saturated
solution of lithium hydroxide present in the encasing mortar,
and can travel through the pores before super-saturation
and precipitation occurs. The porosity of the encasing mortar
is deliberately designed to be high enough to accommodate this
kind of corrosion product movement. The mechanism ensures
that no stresses and no expansive forces build up around the zinc
core and cracks are avoided. It also allows continual exposure of
the zinc substrate to the alkali pore solution ensuring its high
activity.
The level of depolarisation was recorded at 12 points within
the repaired area and at 8 points outside the repair. In the repaired
area this was a 3 4 grid at 500 mm intervals along the length of
the beam and 250 mm intervals across the beam but away from
Table 1. Total charge produced by each individual anode in 10 years of
exposure
Anode
number
Charge
(coulombs)
Anode
number
Charge
(coulombs)
63 141
55 696
46 621
41 135
61 228
60 825
7
8
9
10
11
12
Mean
65 194
43 619
59 459
56 333
57 205
48 989
54 954
1
2
3
4
5
6
www.matcorr.com
Figure 8. Approximate amount of zinc consumed based on the charge
produced and 85% efficiency
the anodes. Outside the repair, the points were along two lines on
the vertical face of the beam at 50 and 300 mm from the edge of
the repair (Fig. 3). The values recorded were the difference
between the potential at the nodes whilst the anodes were
connected (not the instant off potential) and the depolarised
potential at the same nodes 4 h after disconnection of all the
anodes (24 h for the 3400 day results). The actual values are
tabulated in Table 2.
The values are very low initially but approach or exceed
100 mV after 9 years (3400 days) of polarisation of the steel. It is
interesting to note that the level of depolarisation tends to be
higher outside the repaired area and appears to be still significant
up to 300 mm away from the edge of the repair.
Systems such as this are designed to cathodically prevent the
onset of corrosion of the reinforcement and not to control existing
corrosion as is the case in cathodic protection systems. As such, a
much lower current density (0.2–2 mA/m2) is necessary for
cathodic prevention, as reported by Bertolini et al. [5] and Pedeferri
[6] and adapted in the European Standard EN 12696:2000 [7].
Estimation of the steel surface area within the repaired and
affected adjacent area shows that the mean current density
ranged between 0.6 and 3.0 mA/m2 with an overall mean of
around 1.4 mA/m2, generally within the suggested range for
cathodic prevention. Considering that the depolarisation level
from the monitored site increased substantially with time,
indicating a good level of protection of the steel reinforcement,
and that after 10 years life the patch repair is still intact and
operating well, it is reasonable to suggest that the 100 mV
depolarisation criterion is not applicable for this type of cathodic
prevention system, especially early on. Other criteria for
determining the effectiveness of a cathodic prevention system
may be necessary.
Monitoring the depolarised potential of the steel in the
vicinity of the repair with time may be a more effective way of
determining the effectiveness of the system. Figure 10, showing
the mean depolarised potential with time both within and outside
the repaired area, indicates this. It is clear that the recorded mean
potential is continuously moving to a more noble level with time
indicating increasing passivation of the steel.
The corrosion current density of the steel reinforcement, a
direct measure of the state of the steel, is probably the best
parameter available for monitoring the performance of the
system. This was simply estimated using Equation (1) which
ß 2011 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
101
Sergi
Materials and Corrosion 2011, 62, No. 2
Figure 9. (a) Removed 10-year-old anode showing corrosion product around zinc and intact encasing mortar. (b) Zinc core with part of the
corrosion product broken off exposing the zinc substrate
requires only measurement of the applied current density and
equivalent depolarisation potential.
icorr ¼ iappl =½expð2:3h=bc Þ expð2:3h=ba Þ
(1)
where icorr is the corrosion current density, iappl the applied
current density, h the observed potential shift (polarisation
potential), bc the cathodic Tafel slope (assumed as 120 mV) and ba
is the anodic Tafel slope (assumed as 60 mV).
The corrosion current density was shown to diminish with
time. It averaged around 0.7 mA/m2 up to day 112 and reduced
substantially to 0.2 mA/m2 after 9 years. These levels are
considerably lower than 1–2 mA/m2 which is assumed to be
the limit above which corrosion becomes significant.
Another interesting parameter that could be useful as a
criterion for establishing the continued effectiveness of galvanic
anodes is the current output at switch on. In this case, it was
measured five seconds after reconnecting each individual anode
to the steel while all other anodes were disconnected following the
4 or 24 h depolarisation period. The mean current appears to have
increased with time right up to 10 years (Fig. 11). The magnitude
of this current is likely to be related to the potential difference
between the zinc anode and the steel, which may be thought of as
Table 2. Mean depolarisation values at 4 h (24 h at 3400 days)
No. of
days from
switch-on
21
41
50
112
3400
West vertical face of beam
at shown distance from
edge of repair (mV)
Beam soffit
within the
repaired area,
midway between
anodes (mV)
50 mm
300 mm
56
27
22
24
95
58
47
55
48
184
56
31
28
11
Not determined
ß 2011 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
a ‘drive potential’. The zinc potential was not measured but a plot
of the ‘instant-on’ current versus the mean potential of the steel
within and around the repaired area just before switch-on, a
parameter related to the ‘drive potential’ if the potential of the zinc
is assumed to be reasonably stable, suggests a good correlation
(Fig. 12). Both Figs. 11 and 12 also suggest that the anodes are still
able to perform to a very high level after 10 years of operation. The
generally decreasing current output of the anodes shown in Fig. 7
may thus be related primarily to the changing conditions of the
concrete, as drying of the pier has been occurring owing to the
repair of leaking joints, and possible changes to the steel concrete
interface such as increased alkalinity and enhancement of the
passive oxide film or localised deposition of solid phases owing to
the continuous polarisation of the steel [8].
4 New developments
Knowledge gained from this and other trials [3] enhanced by
additional research has enabled the further development of
300mm outside repair
Steel potential (mV vs Cu/CuSO4)
102
50mm outside repair
Within repair
100
0
-100
-200
-300
-400
10
100
1000
10000
Time (days)
Figure 10. Mean depolarised steel potentials with time (4 or 24 h after
disconnection of the anodes)
www.matcorr.com
Materials and Corrosion 2011, 62, No. 2
Galvanic sacrificial anodes in steel reinforced concrete
'Instant-on' Current (mA)
2000
1500
1000
500
0
10
100
1000
10000
Time (days)
Figure 11. Variation of mean current output with time of anodes at
switch-on after depolarisation
Figure 13. Cross-sectional detail of the abutment rehabilitation system
concrete was removed; long lengths of the anode were connected
to the existing reinforcing steel and embedded in a new concrete
layer along with additional epoxy coated reinforcement down the
whole face of the abutment wall. The purpose of the anode
network was to protect the existing steel from chloride-induced
corrosion allowing un-cracked chloride-contaminated concrete to
remain in place and thus reduce concrete breakout. The cross
sectional configuration of the repaired abutment wall and
adjoining structural elements are shown in Fig. 13.
The current output, shown in Fig. 14, is seen to be strongly
related to temperature. Its magnitude varied considerably on an
annual basis with temperature but the mean current density has
been gradually reducing year by year. After an initial level of over
35 mA/m2 of steel area in the first few days, it averaged over
8 mA/m2 during the first year lowering gradually to around 5 mA/
m2 in the fourth year. These levels of current density are within
the design limits of 2–20 mA/m2 of steel area for cathodic
protection as specified in EN 12696:2000 [7]. Current densities in
impressed current cathodic protection systems are also normally
reduced with age as the steel becomes easier to polarise.
Depolarisation levels were measured to be well in excess of
100 mV as specified in the same standard, suggesting that the
galvanic system was deemed to satisfy the criteria for cathodic
protection of steel reinforcement.
100
60
120
0
-100
-200
-300
50
1 00
1 00 0
100 00
'Instant-on' current (mA)
Figure 12. Relationship between mean ‘instant-on’ current and mean
rest potential of steel within and around the repair area
www.matcorr.com
Temperature
40
80
35
30
60
25
20
40
15
10
5
-400
100
45
20
Temperature, degree F
55
Galvanic Current, mA
Rest Potential (mV vs Cu/CuSO4)
galvanic anodes. Single anodes with modified geometry and
encasing mortar composition have been developed with double
and four times the current output capability for use in more
severe conditions. Some of these have now been installed in the
same UK bridge as a new trial, following removal of a number of
original anodes for analysis. The new repaired area contains three
‘double-output’ anodes and seven ‘quadruple-output’ anodes at a
maximum spacing of 300 mm. Chloride levels in the adjacent
undamaged concrete was found to be in the range 1.0–2.6%
chloride by weight of cement. Normally, this highly contaminated
concrete should be removed but it provides an opportunity to test
the capability of the improved anodes. Monitoring is at its infancy
but very early results show a mean ‘instant-on’ current of 1450 mA
per ‘double-output’ anode and 2620 mA per ‘quadruple-output’
anode, both considerably higher than the current of the original
anodes found to be of the order of 800–1000 mA at the equivalent
time.
Other configurations of galvanic anodes have also been
developed for more global corrosion control methodologies. One
such configuration is the system installed at a bridge abutment in
Ohio, USA. The abutment had been contaminated with chlorides
causing localised corrosion of the reinforcing steel.
As part of the rehabilitation, which also included enlargement and strengthening of the abutment, the cracked and spalled
Current
0
0
May-Aug-Oct- Jan- Apr- Jul- Oct- Jan- Apr- Jul- Oct- Jan- Apr- Jul- Oct- Jan- Apr05 05 05 06 06 06 06 07 07 07 07 08 08 08 08 09 09
Date
Figure 14. Current output of anode system and its relationship to
temperature
ß 2011 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
103
104
Sergi
5 Conclusions
Zinc galvanic anodes, activated by a lithium hydroxide saturated
mortar, were shown to be successful in providing adequate
cathodic current to the steel reinforcement around the periphery
of a patch repair for a period of 10 years. This ensured that no
incipient anodes were formed on the steel adjacent to the repaired
area and the repair as a whole remained intact and free from
corrosion of the steel reinforcement. Extrapolation of the results
also showed that a service life of between 24 and 37 years can be
expected from these 60 g zinc anodes.
Depolarisation levels of the steel reinforcement after
disconnection of the anodes for periods of either 4 or 24 h
showed a diminishing trend over the first 112 days, rarely
exceeding 50 mV around the periphery of the repair, but then
increased to over 100 mV after 9 years. The 100 mV depolarisation
criterion, which applies to cathodic protection systems, is unlikely
therefore to apply for cathodic prevention systems of this type.
An alternative more realistic criterion should be developed.
It is suggested that the change in the rest potential of
the steel, following periods of depolarisation over a constant
time (4 or 24 h), be considered as a criterion for establishing the
performance of a corrosion prevention system. This can be aided
by estimation of the corrosion current density of the steel from
knowledge of the depolarised potential and the equivalent applied
current density. In this particular case, it was seen that the
steel rest potential gradually moved in a positive direction while
the corrosion current density diminished to a very low level
signifying improved passivity of the steel.
Although the current output followed an overall decreasing
trend, the driving power of each anode did not show any evidence
of diminishing. To the contrary, the current output at switch-on,
following a period of depolarisation, was seen to increase with
time, possibly because the potential of the steel gradually moved
Materials and Corrosion 2011, 62, No. 2
in the positive direction thus increasing the ‘drive voltage’
between the anode and the steel.
Lessons learned from this and other trials and from further
research have enabled the production of enhanced performance
anodes using a better surface area to volume ratio and improved
chemical composition of the encasing mortar. Anodes with
double or quadruple the current output capability have been used
for a new trial at the same site. Early results confirm their higher
capacity.
The technology was shown to be very flexible and by utilising
a distributed current-type anode set-up consisting of long anodes
affixed along the steel reinforcement, it was possible to provide
depolarisation levels exceeding 100 mV, and to achieve current
densities compatible to conventional impressed current cathodic
protection systems.
6 References
[1] G. Sergi, C. L. Page, presented at EUROCORR’99, Aachen,
Germany, August 29–September, 1999.
[2] C. L. Page, G. Sergi, J. Mater. Civil Eng. Sp. Issue, February
2000, 8.
[3] G. Sergi, D. Simpson, J. Potter, presented at EUROCORR
2008, Edinburgh, UK, 2008.
[4] D. Whitmore, Field Guide to Concrete Repair Application
Procedures ACI RAP Bulletin 8, 2005.
[5] L. Bertolini, F. Bolzoni, A. Cigada, T. Pastore, P. Pedeferri,
Corros. Sci. 1993, 35, 1633.
[6] P. Pedeferri, Constr. Build. Mater. 1996, 10(5), 391.
[7] British Standard BS EN 12696:2000, 2000.
[8] G. K. Glass, B. Reddy, in: R. Weydert (Ed.), COST 521: Final
Reports, Luxembourg University of Applied Sciences, Luxembourg, 2002, p. 227.
(Received: March 6, 2010)
(Accepted: April 8, 2010)
ß 2011 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim
W5707
www.matcorr.com