99020115

CORROSIONVol. 55, No.
2 115
CORROSION SCIENCE SECTION
Submitted for publication December 1997; in revised form,
September 1998. Presented as paper no. 726 at CORROSION/98,
March 1998, San Diego, CA.
* Center for Marine Materials, Department of Ocean Engineering,
Florida Atlantic University, Boca Raton, FL 33431.
Effect of Velocity on Current Density
for Cathodically Polarized Steel in Seawater
D. Hugus and W.H. Hartt*
ABSTRACT
A series of cathodic polarization experiments was performed
in which sections of UNS G10230 steel pipe were galvani-
cally coupled through an external resistor to an aluminum
anode ring. These cells were incorporated into a seawater
flow loop such that the hydrodynamic conditions were de-
fined quantitatively. Seawater velocities for the experiments
were 0.03, 0.09, and 0.30 m/s. From the results, the trend of
steady-state potential vs current density at a given velocity
was characterized in terms of the recently proposed slope
parameter approach to cathodic protection (CP) design. Also,
an expression for the velocity dependence of the mainte-
nance current density was developed. Anomalous behavior
was encountered for some specimens tested at velocities of
0.09 m/s and 0.30 m/s in that these exhibited a period of
depolarization followed, in some cases, by repolarization.
This behavior was reviewed in terms of transitory calcareous
deposit protectiveness, possibly reflecting the relative
amounts of calcium and magnesium in the deposits at a
particular time.
KEY WORDS: calcareous deposits, cathodic protection,
design, offshore environments, seawater, structural steel,
UNS G10230, velocity
INTRODUCTION
Cathodic protection (CP), with or without coatings,
has served for decades as the corrosion control
method of choice for steel members and structures in
seawater. In the past, incentives for optimizing this
protection have been relatively modest, but the ad-
vent of deep-water petroleum production activities
and environmental concerns have prompted research
that has led to designs for rapid polarization,
1-4
im-
proved understanding of calcareous deposits and
structure current demand,
5-10
and development of the
slope parameter approach to CP design.
11-13
Through-
out these studies and the evolution of various design
practices,
14-16
the maintenance or mean current
density (i
m
), the time-averaged current density over
the design life (T) of the CP system, has remained a
key parameter. This term is represented in current
design practices by:
15-16

i
m
=
NC
A
c
T
(1)
where N is the requisite number of anodes, C is an-
ode current capacity, and A
c
is cathode surface area.
In the slope parameter format, this expression
becomes:
11-13

i
m
=
R
a
wC
TS
(2)
where R
a
and w are the resistance and weight of the
individual anodes, respectively, and S is the slope
parameter. With development of this latter equation,
0010-9312/99/000027/$5.00+$0.50/0
1999, NACE International
116 CORROSIONFEBRUARY 1999
the design process was elevated from an empirical
procedure in which overdesign by some undeter-
mined amount is implicit to one which is
first-principles based.
12
Still to be developed, how-
ever, is an improved method for projecting i
m
, since
this presently is determined only by prior service
experience in comparable ocean locations or by
experiments and test exposures that simulate such
experience. Even information developed from these
may be inadequate or misleading, as evidenced by
the fact that the existing recommended practice lists
i
m
for the Gulf of Mexico as 55 mA/m
2
(5 mA/ft
2
),
15
whereas long-term steady-state values an order of
magnitude or more lower have been reported.
17
It has been established from corrosion principles
that CP current demand for a structure and, hence,
i
m
are governed by oxygen availability or by the rate
of oxygen transport across the diffusional and hydro-
dynamic boundary layers as affected by temperature,
relative water movement (velocity), and surface films
(coatings, calcareous deposits, and fouling). Particu-
larly absent are studies that incorporate seawater
velocity as a control variable and studies where this
parameter has been characterized hydrodynamically.
Exceptions include the research of Wolfson and
Hartt,
18
Smith, et al.,
19
and Gartland, et al.,
20
each of
whom potentiostatically polarized steel specimens
under controlled seawater flow conditions. However,
the potentiostatic nature of these tests was such
that, while information was gained regarding calcare-
ous deposits and the capability for these to reduce
current demand as exposure time progressed, the
results are not quantitatively relatable to galvanic CP
where both potential and current density change
with time.
A procedure for addressing the combined influ-
ences of electrolyte flow and presence of calcareous
deposits upon current density has been proposed
based upon the Sherwood (Sh), Schmidt (Sc), and
Reynolds (Re) numbers.
19-20
Thus, for flow along
a straight, circular cross-section pipe, Re is
expressed as:

Re =
vd
(3)
where v is the average fluid velocity, d is the inside
pipe diameter, and is the kinematic viscosity.
Under these conditions, flow is considered tur-
bulent if Re exceeds 2,100 and laminar at lower
values. Sc, in contrast, characterizes the diffusional
nature of the electrolyte according to:

Sc =

D
(4)
where D in the present case is the oxygen diffusivity.
TABLE 1
Chemical Composition of the Aluminum Anode Material
Element Composition (wt%)
Zn 1.444
In < 0.0010
Hg 0.0433
Si 0.036
Cu 0.00345
Fe 0.038
Cd 0.0016
B < 0.03
Sn 0.0033
Mg 0.0075
Cr 0.002
Ni 0.0025
Mn 0.004
Ti 0.003
V 0.003
Bi 0.0023
Al Bal.
TABLE 2
Chemical Composition of the Steel Cathode Pipe Material
Element Composition (wt%)
C 0.23
Mn 0.39
Si 0.04
Ni 0.01
Cr 0.03
Mo < 0.01
S 0.011
P 0.008
Cu 0.02
(a)
(b)
FIGURE 1. Schematic of: (a) a steel cathode specimen and (b) an
aluminum anode specimen after machining.
CORROSIONVol. 55, No. 2 117
Sh is determined from:

Sh= 0.03 Re
0.8
Sc
0.33
(5)
and this can be incorporated into Ficks first law
such that:
21

i
L
=
DnFc
x/Sh + t/p
(6)
where i
L
is the limiting current density for oxygen
concentration polarization (equivalent to i
m
), n and F
have their normal meanings, c is the bulk dissolved
oxygen concentration, x is the average length of the
pipe, t is the deposit thickness, and p is the porosity
constant of the calcareous deposit (unitless).
In conjunction with development of the slope
parameter approach to design of galvanic CP systems,
an experimental methodology was developed that
facilitates quantitative characterization of polarization
in terms of field-relevant parameters.
11
The objective
of the present research was to extend this approach
to experiments where flow was controlled and quanti-
fied for the purpose of characterizing i
m
in terms of
electrochemical and hydrodynamic parameters.
EXPERIMENTAL
Materials and Specimens Anodes for the ex-
periments were machined into a ring shape from an
Al-Zn-Hg (Galvalum I
) casting. The cathodes were

cylinders that were sectioned from a 3.7-m length of
UNS G10230
(1)
steel pipe 10 cm (4 in.) in nominal
diameter with the internal surface polished to a 600-
grit finish. Interior surface area of the steel cathodes
was 506 cm
2
and was 43 cm
2
for the anodes, such
that the surface area ratio (anode-to-cathode) was
0.085. Both specimen types were degreased after
preparation. Final dimensions are shown in Figure 1.
Chemical compositions of the aluminum and steel
are presented in Tables 1 and 2, respectively.
Test Cells and System A Schedule 80 polyvinyl
chloride (PVC) slip flange was compression-fitted
about each end of the steel cathode specimens, and
the joint between the two was sealed at the outer
surface. A machine screw was soldered to the outer
steel surface at mid-length. Anode rings were drilled
and tapped on the outside surface and fitted with a
machine screw. In both cases, these screws served as
sites for electrical connection.
Figure 2 shows the test cell configuration, which
was comprised of a steel cathode with flanges, an
aluminum anode ring, a PVC pipe section with
flanges, an external resistor between the anode and
cathode, and a silver-silver chloride (Ag-AgCl) refer-
ence electrode. The external resistor was used to
affect a particular value for S. The Ag-AgCl reference
electrodes were fabricated from silver wire 0.5 mm in
diameter and molten AgCl. These were potted in a
threaded PVC cap using silicone and then positioned
into a length of PVC pipe with end flanges of similar
dimensions as for the steel cathode such that the tip
of the reference electrode extended 5 mm below the
interior PVC pipe surface. Different sections were
bolted together across an anode ring such that a
watertight seal was created. All components had the
same internal diameter, and caution was taken to
ensure that these aligned with one another.
Figure 3 illustrates the test system used for the
first set of experiments (Set 1). This consisted of an
upper reservoir tank that gravity fed seawater from
an elevation of 2 m through three flow lines, each
(1)
UNS numbers are listed in Metals and Alloys in the Unified
Numbering System, published by the Society of Automotive
Engineers (SAE) and cosponsored by ASTM.
Trade name.
FIGURE 2. Side-view schematic of the test cell layout.
FIGURE 3. Side-view schematic of the first experimental setup.
of which was inclined at a 22 angle. Each flow line
consisted of four test cells mounted in parallel
(Figure 2). From the flow lines, the water emptied
into a lower reservoir and was pumped back to the
upper reservoir.
Upon completing the first set of experiments and
opening the system, it was determined that corrosion
products from the anode had settled along the bot-
tom of the line and that these extended onto the
initial portion of the adjacent cathode. Also, corro-
sion products on the marginally protected and
underprotected cathodes indicated the flow had a
spiral component, apparently because of the rela-
tively short distance from the upstream cell to the
test sections. Consequently, the test system was
modified for the second set of tests (Set 2). This
modification included positioning the flow lines verti-
cally, adding 13 mm-diameter PVC pipe sections
within the PVC portion of the lines as flow straight-
eners, and repositioning the anodes so that these
were downstream rather than upstream from the
corresponding cathode. Figure 4 is a schematic of
this arrangement.
Flow rate was controlled by a gate valve in each
line such that average velocities of 0.03, 0.09, and
0.30 m/s were achieved, as determined by time mea-
surements of flow volume for each leg. The seawater
was pumped continuously to the laboratory from a
buried offshore wellpoint, which had been shown by
monitoring during an annual cycle to be typical of
semitropical ocean water.
22
In the flow system, this
water was replenished at an exchange rate of five to
six turnovers per day. This was considered sufficient
to maintain the water quality in view of the aeration
that resulted from the pumping and because the
rates of calcium (Ca
2+
) and magnesium (Mg
2+
) deple-
tion via calcareous deposition were calculated to be
small compared to those of replenishment.
Calculation of Re (Equation [3]) indicated that
flow in the 10-cm test line was turbulent at each of
the three velocities. At the same time, flow in the flow
straighteners that were added for Set 2 was laminar.
Consequently, this may have resulted in a laminar-
to-turbulent flow transition along the steel cathode
specimens in the Set 2 experiments.
(a)
(b)
(c)
FIGURE 5. Plot of: (a) vs time, (b) i vs time, and (c) vs i for
Specimen I-32/03 (S = 0.32 -m
2
, and v = 0.03 m/s).
FIGURE 4. Side-view schematic of the second experimental setup.
External resistors were sized such that one
specimen in each flow line had S values of 0.32, 0.63,
1.94, and 3.85 -m
2
during Set 1 and 0.081, 0.17,
0.32, and 0.63 -m
2
for Set 2. Cathode potential and
voltage drop across the external resistors were moni-
tored during the experiments and were recorded
using a computer-based data acquisition system.
RESULTS AND DISCUSSION
Polarization Data
Potential () and current density (i) for all
specimens decreased during the initial 100 h or
so of exposure and approached steady-state
values as illustrated by the example in Figure 5,
which plots vs time, i vs time, and vs i. This
behavior was similar to that reported from similar
galvanic CP experiments and was attributed to
progressive oxygen concentration polarization
in conjunction with formation of calcareous
deposits.
11-12
Previous studies have shown this
-i data trend (Figure 5[c]) to conform to the
relationship:

c
= R
t
A
c
i
c
+
a
(7)
TABLE 4
Steady-State Potentials for the Steel Cathode Specimens in Set 2
S v Steady-State Potential
Specimen (-m
2
) (m/s) (V
Ag-AgCl
)
II-081/03 0.081 0.03 1.05
II-081/09 0.081 0.09 0.98
II-081/30 0.081 0.30
(A)
II-17/03 0.17 0.03 1.04
II-17/09 0.17 0.09 0.99
II-17/30 0.17 0.30
(B,C)
II-32/03 0.32 0.03 1.05
II-32/09 0.32 0.09 0.96
(C)
II-32/30 0.32 0.30
(B,C)
II-63/03 0.63 0.03 1.04
II-63/09 0.63 0.09 0.97
(B,C)
II-63/30 0.63 0.30 0.78
(B)
(A)
Exception 3 (see text)
(B)
Exception 1, Type 1 (see text).
(C)
TABLE 3
Steady-State Potentials for the Steel Cathode Specimens in Set 1
S v Steady-State Potential
Specimen (-m
2
) (m/s) (V
Ag-AgCl
)
I-32/03 0.32 0.03 1.04
I-32/09 0.32 0.09 0.98
I-32/30 0.32 0.30 0.88
(A)
I-63/03 0.63 0.03 1.00
I-63/09 0.63 0.09 0.90
I-63/30 0.63 0.30 0.84
(A)
I-194/03 1.94 0.03 0.89
I-194/09 1.94 0.09 0.78
(A,B)
I-194/30 1.94 0.30 0.67
(B)
I-385/03 3.85 0.03 0.71
I-385/09 3.85 0.09 0.68
(B)
I-385/30 3.85 0.30 0.66
(A)
Exception 2 (see text).
(B)
where
c
is the cathode potential, R
t
is the total
circuit resistance (dominated in the present case by
magnitude of the external resistor), and
a
is the
anode potential.
Thus, the -i trend is projected to be linear with
S = R
t
x A
c
and with intersection of the vertical axis at
a
, provided S and
a
are constant with time. The
departure from linearity in Figure 5(c) probably
reflects a period of anode activation during the initial
stages of the test (upper right portion of the curve)
and partial passivation of the anode once i became
low (lower left).
The above example where and i decreased
monotonically to steady-state values was termed
normal behavior, as contrasted with exceptions that
also were noted. Tables 3 and 4 list steady-state
potentials for the test specimens in experiments in
Sets 1 and 2, respectively, and indicate that such
exceptions to normal polarization occurred mostly at
the highest velocity (0.30 m/s) and, to a lesser
degree, at the intermediate velocity (0.09 m/s). All
tests at the lowest velocity exhibited normal behav-
ior. These occurrences of exceptional behavior were
categorized according to:
Category 1: cathode depolarization followed by
repolarization ( vs i linear);
Category 2: cathode depolarization without
subsequent repolarization ( vs i linear); and
Category 3: cathode depolarization followed by
repolarization ( vs i nonlinear).
In addition, two types of Category 1 behavior
were encountered. The first, which was designated as
Type 1, occurred relatively early in the tests and was
of short duration. The second (Type 2) occurred sub-
(a)
(b)
(c)
(a)
(b)
(c)
Specimen II-63/09 (S = 0.63 -m
2
, and v = 0.09 m/s).
Specimen I-63/30 (S = 0.63 -m
2
, and v = 0.30 m/s).
sequently and lasted longer. Figures 6 through 8
illustrate examples of each category. Thus, in the
case of Figure 6, two depolarization/repolarization
events were noted, where the first (Category 1/Type
1) occurred in the 100-h to 200-h timeframe, and the
second (Category 1/Type 2) occurred at 500 h to
2,500 h. Figure 7 shows an example where both Cat-
egory 1/Type 1 and Category 2 behaviors occurred.
Figure 8 shows an example of Category 3. The -i
plots for the Categories 1 and 2 behavior (Figures 6
and 7) were essentially linear and, as such, indicated
that circuit resistance and
a
remained constant
throughout the experiments. Instances of Category 2
behavior were noted only in Set 1 experiments. Since
the duration of these experiments was shorter than
for Set 2 tests (1,850 h compared to 4,500 h), it was
possible that Category 2 behavior was the same as
for Category 3 but that there was insufficient time for
repolarization in the former case. The nonlinear
trend in the case of Category 3, in contrast, indicated
R
t
,
a
, or both changed (increased) as these experi-
ments progressed. To investigate this,
a
and
c
(current-on values) were recorded periodically over a
2-week period. The potential difference between these
two (
c

a
) was determined and compared with the
voltage drop across the external resistors. Results are
shown in Table 5, which indicates that the magni-
tude of these two parameters was essentially the
same. It was concluded from this that a circuit resis-
tance increase, as could result from corrosion
product accumulation upon the anode calcareous
deposits upon the cathode, or a combination of these
two, was not a factor and that polarization of the
anode caused the -i nonlinearity.
In such experiments, the cathode would be ex-
pected to control with regard to current. That this
was the case in instances of normal and Categories 1
and 2 behavior was indicated by the fact that i
tracked (i.e., a positive excursion was accompa-
nied by an i increase and visa versa [cf. Figures 6
and 7]). If this was the case, then some developmen-
tal aspect in the formation of the calcareous deposit,
whereby its protectiveness was compromised at first
and later restored, must have been responsible.
A similar trend as for Category 1/Type 1 and
Category 2 behavior occurred also in Category 3
(Figure 8) to ~ 1,500 h. Beyond this, i decreased with
time, while
c
remained about the same (slightly
negative to 0.80 V
Ag-AgCl
) to ~ 2,850 h and then in-
creased to 0.70 V
Ag-AgCl
. During this same period,
a
increased to near this same value (0.70 V
Ag-AgCl
).
This trend, coupled with the fact that the peak cur-
rent density upon the anode (1,500 h into the test)
was 10 A/m
2
, suggested the critical i may have been
reached and that the anode partially passivated.
In evaluating data for the various specimens,
consideration was given to the fact that the experi-
mental setup was changed between the Set 1 and Set
2 tests. However, no distinctions in the -i-velocity
trends between the two sets were apparent. Thus, the
data were analyzed assuming they conformed to a
common population.
Velocity Dependence of Polarization
Figure 9 shows -i data for each of the three
velocities in cases where steady-state was achieved.
Also included are results from previous tests that
were performed in a 2-L test cell with a seawater
exchange rate of 150 mL/min, for which velocity was
considered quiescent.
11
The three Set 2 specimens
for which -i behavior was nonlinear (Table 4) were
not included in Figure 9 because steady-state was
not achieved for these. In each case, a trend curve
was added. Thus, the quiescent velocity data exhib-
ited a relatively well-defined sigmoidal trend. The
0.03-m/s and 0.09-m/s data were displaced progres-
sively toward higher velocity but showed a trend
(a)
(b)
(c)
Specimen II-081/30 (S = 0.081 -m
2
, and v = 0.30 m/s).
similar to that of the quiescent curve, although scat-
ter was greater and there were gaps in the data for
the two higher velocities. The 0.30-m/s results indi-
cated a progressive current density increase as
steady-state potential became more negative.
The maximum steady-state current density at
the lower two velocities (quiescent and 0.03 m/s)
occurred near 0.80 V
Ag-AgCl
(that this was so was less
apparent in the 0.03-m/s case), whereas this maxi-
mum was at ~ 0.90 V
Ag-AgCl
for 0.09 m/s and at an
even lower , assuming such a peak occurred at all,
for 0.30 m/s. Such a trend ( at which the peak i
occurred decreasing with increasing velocity above
0.03 m/s) indicated rapid polarization. Achievement
of a low i
m
1-4
may have become more difficult, if not
impossible, at higher velocities.
The data in Figure 9 indicated that some speci-
mens tested at different velocities developed nearly
the same steady-state potential, 0.78, 0.88, and
0.98 V
Ag-AgCl
. This facilitated projection of a steady-
state i-vs-velocity relationship (Figure 10). In doing
this, the quiescent water movement data were
assigned zero velocity.
11
A best fit line, which con-
formed to the relationship:

i
m
= 852v + 48 (8)
was fitted to the data. On this basis, an order of mag-
nitude increase in velocity caused about a fourfold
increase in i.
Calcareous Deposit Properties
Porosity An attempt was made to evaluate the
calcareous deposit porosity constant (p) in Equation
(6). To accomplish this, Re was determined for the
present experimental setup assuming v = 8.81 x
10
7
m
2
/s and Sc (Equation [4]) assuming an oxygen
diffusivity (D) of 2.72 x 10
5
cm
2
/s. From these, the
Sh (Equation [5]) was calculated based upon an as-
sumed bulk oxygen concentration (c) of 5 mg/L.
Deposit thickness values were taken from previous
research for velocities and that were comparable to
those of the present test conditions,
18
and the cur-
rent densities (i
L
) were those measured in the present
study. Table 6 lists these t
values
, and accordingly,
Table 7 indicates the calculated p values. Although
results were limited, they showed that p increased
with increasing velocity and with decreasing . Also,
the ratio of thickness-to-porosity was 3.3 to 5.7 times
greater than the average x-to-Sh ratio. This indicated
the resistance to diffusion afforded by the calcareous
deposit was of greater significance than that of the
boundary layer.
Composition and Morphology Following the
experiments, one specimen from each flow line of Set
1 and all Set 2 specimens were analyzed by scanning
electron microscopy (SEM) and energy-dispersive
x-ray analysis (EDX). Specimens in the former case
(Set 1) were selected such that the steady-state was
approximately the same for each. Figures 11 through
13 illustrate the morphology for these, where for the
first (Figure 11, Specimen I-194/03) the deposit was
relatively coarse, with EDX indicating this was pre-
dominantly calcium. Figure 12, on the other hand,
shows the morphology for Specimen I-63/09 and
that this exhibited an inner deposit similar to that in
Figure 11 and a finer-grain outer one. EDX revealed
the former to be calcium-rich and the latter predomi-
nantly magnesium. Figure 13 shows the deposit for
Specimen I-63/30, which was mostly fine grain and
magnesium-rich. These micrographs suggested there
was a tendency for deposits that formed under these
conditions (steady-state of 0.98 V
Ag-AgCl
) to transi-
TABLE 5
Comparison of
c

a
and Voltage Drop
Across the External Resistors for Set 2 Specimens
V Across
S Time
c

a

c

a
Resistor
(-m
2
) (h) (V) (V) (V) (V)
3,740 0.689 0.700 0.011 0.011
0.081 3,915 0.709 0.720 0.011 0.011
4,035 0.695 0.707 0.012 0.012
4,165 0.695 0.706 0.010 0.011
3,740 0.669 0.691 0.022 0.021
0.17 3,915 0.673 0.693 0.020 0.020
4,035 0.671 0.689 0.018 0.019
4,165 0.67 0.687 0.017 0.016
3,740 0.832 0.952 0.120 0.119
0.32 3,915 0.796 0.926 0.130 0.130
4,035 0.775 0.900 0.125 0.125
4,165 0.777 0.897 0.120 0.119
tion from calcium- to magnesium-rich with increas-
ing velocity. The calcium-rich deposits were
presumably aragonite (CaCO
3
), and the magnesium
ones were brucite (Mg[OH]
2
).
Set 2 specimens tested at 0.03 m/s velocity all
polarized to 1.04 V
Ag-AgCl
or 1.05 V
Ag-AgCl
, irrespective
of S (Table 4). Figure 14 illustrates the deposit mor-
phology for Specimen II-17/03 and reveals this to be
similar to what has been reported historically for
comparable test conditions.
5,9-10
EDX showed the
composition to be calcium-rich with only a trace of
magnesium. The deposits for other Set 2 specimens
tested at this same velocity were similar to this with
regard to morphology and composition.
Steady-state for Set 2, 0.09-m/s velocity speci-
mens were in the range 0.96 V
Ag-AgCl
to 0.99 V
Ag-AgCl
.
An example of these for Specimen II-63/09 is shown
in Figure 15. Here, the morphology was a mixture of
coarse and fine grain particles that appeared similar
to those for Specimen I-63/09 (Figure 12). However,
EDX showed both to be predominantly calcium.
Set 2 specimens tested at 0.30-m/s velocity
exhibited a relatively broad range of steady-state
(Table 4) because some of these depolarized and
others did not. In this regard, Specimens II-081/30
and II-17/30 had final that were positive to
0.70 V
Ag-AgCl
. Consistent with this, relatively little
calcareous deposit was present. Figure 16 shows
examples of the deposit morphology on the other
two specimens (Specimens II-32/30 and II-63/30)
and that these also were a mixture of coarse- and
fine-grain particles. EDX revealed these to be
calcium-rich.
Previous studies of calcareous deposits have
ascribed significance to the relative amounts of cal-
cium and magnesium, which often has been
characterized in terms of the calcium-to-magnesium
ratio.
23
Apparently, the calcium-rich phase (aragonite
or calcite) serves as a more resistant barrier to oxy-
gen diffusion than does the magnesium one (brucite),
although this concept has been challenged in the
case of the thin, inner Mg(OH)
2
film that invariably is
present beneath the CaCO
3
layer but not necessarily
detected.
10
Because of gaps in the polarization data
and because deposit compositional information was
available for only two exposure times, it was not pos-
sible to construct a comprehensive evolutionary
model of the calcareous deposits and to interrelate
this to the -i history. However, the depolarization
and depolarization/repolarization trends in Figures 6
through 8, coupled with the finding that the calcium-
magnesium ratio was often low after 1,850 h but
high at 4,500 h, suggests that, subsequent to initial
rapid polarization, Mg(OH)
2
precipitated in some
cases upon specimens exposed to the higher veloci-
ties and caused current demand for these to
increase, which resulted in depolarization. Subject to
the anode potential remaining sufficiently negative,
TABLE 6
Calcareous Deposit Thickness
at Different Potentials and Velocities
18
Seawater Deposit
Velocity Potential Thickness
(m/s) (V
SCE
) (mm)
1.03 0.18
0.30 0.93 0.07
0.78 0.04
1.03 0.06
0.08 0.93 0.05
0.78 0.03
TABLE 7
Calculated Values for p
Approximate
Potential
(V
Ag-AgCl
) 0.09 m/s 0.30 m/s
0.78 0.035 0.061
0.89 0.086 0.104
p
(unitless)
FIGURE 9. Plot of steady-state vs steady-state i for specimens of
the present tests.
FIGURE 10. Plot of steady-state i as a function of V
Ag-AgCl
.
these specimens subsequently repolarized in re-
sponse to CaCO
3
being reestablished as the principle
deposit phase. The cause of such a velocity depen-
dent deposit compositional cycle was not determined.
Also unclear was why such a cycle has apparently
not been evident from data acquired in conjunction
with monitoring programs on offshore structures for
locations where water movements were comparable
to those in the present study. Possible explanations
for this are that:
Data of this type from actual structures are
limited, and the exposure conditions for where they
do exist may not have been the same as for the
present experiments;
Velocity was continuous in the present experi-
ments; whereas in actual service, it may be near zero
during slack periods. Deposit initiation and growth is
expected to be enhanced during times of low or nil
velocity; and
Macrofouling, which was not present in these
laboratory experiments since the seawater was sand
filtered, may provide an additional barrier to water
movement on actual offshore structures such that
the effective velocity at the steel surface is lower than
was the case for the present specimens.
These same factors may explain also why a
relatively sensitive dependence of i
m
upon velocity
(Equation [8]) was realized from the present experi-
ments, such that calculated values for this
parameter exceeded what has typically been reported
for actual offshore structures. For example, the aver-
age of the three 0.09-m/s data points in Figure 10
(a) (b)
(a) (b)
FIGURE 11. SEM micrographs of the calcareous deposit that formed on Specimen I-194/03 (S = 1.94 -m
2
, and v = 0.03 m/s).
2
, and v = 0.09 m/s).
was 145 mA/m
2
, which exceeds typical steady-state
values for Gulf of Mexico structures by an order of
magnitude or more. Accordingly, deposits that form
during periods of nil water movement may be rela-
tively protective, such that, once these are
established, the velocity dependence of current den-
sity is more modest than projected by Equation (8),
as has been reported.
20
CONCLUSIONS
O In most cases, polarization conformed to what was
termed a normal trend where and i decreased with
time to a steady-state value. Exceptions were noted
at the two higher velocities, however, where the ini-
(a) (b)
(a) (b)
tial polarization was followed by depolarization and,
in some cases, repolarization. Apparently, some as-
pect of the higher flow rates caused the calcareous
deposits to become less protective for some period
and for current demand to increase.
O The trend of vs i at the two lower velocities re-
vealed a sigmoidal curve with a peak current density
being achieved for 0.03-m/s tests at ~ 0.80 V
Ag-AgCl
and for 0.09 m/s at ~ 0.90 V
Ag-AgCl
. At the highest
velocity (0.30 m/s), i increased with decreasing
steady-state for the entire range of S values for
which results were available.
O For specimens that reached steady-state in the
range 0.78 V
Ag-AgCl
to 0.98 V
Ag-AgCl
, i
m
increased with
v according to Equation (8). This velocity dependence
2
, and v = 0.30 m/s).
FIGURE 14. SEM micrographs of the calcareous deposit that formed on Specimen II-17/03 (S = 0.17 -m
2
, and v = 0.03
m/s).
of i
m
exceeds what typically is realized in service,
possibly because velocity was continuous during the
present tests such that there were no nil water move-
ment periods during which a more protective
calcareous deposit could have formed.
O Based upon thickness values for calcareous de-
posits taken from the literature, the deposit porosity
constant was calculated to be in the range 0.035 to
0.104 (unitless). Consequently, the restriction to
oxygen availability afforded by the deposits was 3.3
to 5.7 times greater than otherwise would have been
the case.
O For Set 1 specimens (exposure time = 1,850 h)
tested at the two higher velocities (0.09 m/s and
0.30 m/s), the calcareous deposit was comprised of a
magnesium-rich as well as a calcium-rich phase. In
the case of Set 2 specimens (exposure time = 4,500 h),
the deposit was calcium-rich with only trace
amounts of magnesium. The depolarization/repolar-
ization behavior noted may have been related to an
evolutionary compositional aspect of calcareous de-
posit development where the relative amounts of
calcium and magnesium varied with time.
ACKNOWLEDGMENTS
The authors acknowledge financial support of
the Florida Sea Grant Program of the National
Oceanographic and Atmospheric Administration
(project R/C-D-16).
(a) (b)
(a) (b)
FIGURE 15. SEM micrographs of the calcareous deposit that formed on Specimen II-63/09 (S = 0.63 -m
2
, and v = 0.09 m/s).
FIGURE 16. SEM micrographs of the calcareous deposit that formed on: (a) Specimen II-63/03 (S = 0.63 -m
2
, and v = 0.30
m/s) and (b) Specimen II-32/30 (S = 0.32 -m
2
, and v = 0.30 m/s).
REFERENCES
1. T. Foster, V.G. Moores, Cathodic Protection Current Demand
of Various Alloys in Seawater, CORROSION/86, paper no. 295
(Houston, TX: NACE, 1986).
2. S. Evans, MP 27, 2 (1988): p. 9.
3. C.F. Schrieber, J. Reding, Application Methods for Rapid
Polarization of Offshore Structures, CORROSION/90, paper
no. 381 (Houston, TX, NACE, 1990).
4. K.P. Fischer, T. Sydberger, R. Lye, Field Testing of Deep-Water
Cathodic Protection on the Norwegian Continental Shelf,
CORROSION/87, paper no. 67 (Houston, TX: NACE, 1987).
5. W.H. Hartt, C.H. Culberson, S.W. Smith, Corrosion 40 (1984):
p. 609.
6. H.-S. Lin, S.C. Dexter, Corrosion 44 (1988): p. 615.
7. J.E. Finnegan, K.P. Fischer, Calcareous Deposits: Calcium and
Magnesium Ion Concentrations, CORROSION/89, paper no.
581 (Houston, TX: NACE, 1989).
8. K.P. Fischer, J.E. Finnegan, Cathodic Protection Behavior of
Steel in Seawater and the Protective Properties of the Calcare-
ous Deposits, CORROSION/89, paper no. 582 (Houston, TX:
NACE, 1989).
9. J.S. Luo, R.U. Lee, T.Y. Chen, W.H. Hartt, S.W. Smith, Corro-
sion 47 (1991): p. 189.
10. K.E. Mantel, W.H. Hartt, T.Y. Chen, Corrosion 48 (1992): p. 489.
11. W. Wang, W.H. Hartt, S. Chen, Corrosion 52 (1996): p. 419.
12. W.H. Hartt, S. Chen, D.W. Townley, Corrosion 54 (1998), p. 317.
13. D.W. Townley, Unified Design Equation for Offshore Cathodic
Protection, CORROSION/97, paper no. 473 (Houston, TX:
NACE, 1997).
14. NACE Standard RP0176-76, Corrosion Control of Steel-Fixed
Offshore Platforms Associated with Petroleum Production
15. NACE Standard RP0176-94, Corrosion Control of Steel-Fixed
Offshore Platforms Associated with Petroleum Production
16. Recommended Practice RP B401, Cathodic Protection Design
(Hovilc, Norway: Det Norske Veritas Industri NorRe AS, 1993).
17. M.W. Mateer, K.J. Kennelley, Design of Platform Anode Retro-
fits Using Measured Structure Current Density, CORROSION/
93, paper no 526 (Houston, TX: NACE, 1993).
18. S.L. Wolfson, W.H. Hartt, Corrosion 37 (1981): p. 70.
19. S.W. Smith, K.M. McCabe, D.W. Black, Corrosion 45 (1989):
p. 790.
20. P.O. Gartland, E. Bardal, R.E. Andersen, R. Johnsen, Corrosion
40 (1984): p. 127.
21. T.K. Ross, D.H. Jones, J. Appl. Chem. 12 (1962): p. 314.
22. W.H. Hartt, Fatigue of Welded Structural Steel in Seawater,
paper no. 3982, Proc. Offshore Technol. Conf., Houston, May,
1981 (Dallas, TX: Society of Petroleum Engineers, 1981).
23. R.A. Humble, Corrosion 4 (1948): p. 358.

99020115

Transféré par

Informations du document

Copyright

Formats disponibles

Partager ce document

Partager ou intégrer le document

Options de partage

Avez-vous trouvé ce document utile ?

Ce contenu est-il inapproprié ?

Droits d'auteur :

Formats disponibles

99020115

Transféré par

Droits d'auteur :

Formats disponibles

CORROSIONVol. 55, No.

) casting. The cathodes were

Vous aimerez peut-être aussi