C. e': e '.\
The Application of Noble Metals in Light-Water Reactors Young-Jin Kim, Leonard W. Niedrach, Maurice E. Indig, and Peter L. Andresen Corrosion potential is a primary determinant of the stress-corrosion cracking susceptibility of structural materials in high-temperature water. Efforts to minimize stresscorrosion cracking in light-water reactors include adding hydrogen. In someplants'outof-core regions, the hydrogen required to achieve the desired corrosion potential is relatively high. In-core, more hydrogen is needed for an equivalent reduction in corrosion potential. Additionally, side effects of high hydrogen-addition rates, including increased 16N turbine shine and 6OCO deposition, have also been observed in some cases. An approach involving noble-metal coatings on and alloying additions to engineering materials dramatically improves the efficiency with which the corrosion potential is decreased asa function ofhydrogen addition, such that very low potentials are obtained once a stoichiometric concentration of hydrogen (versus oxygen) is achieved.
INTRODUCTION The important influence of the corrosion potential on stress-corrosion cracking (seC) (Figure 1),1 irradiationassisted sec (IASCC) (Figure 2),2 flow304 Stainless Steel
10-6
25 mm CT Specimen
Constant Load
288· Water Test Conditions: .15 C1cm2 EPR .27.5 MPaYm 0.1 - 0.3 ,.,slem
Predicted Curves From PLEDGE Code
0.1
10-9
L...........................................'""---".....&.-...........................................~
.(l.6
.(l..
.(l2
0
02
0.•
Corrosion Potential. V..
Figure 1. Observed and predicted crack growth rate as a function of corrosion potential for furnace-sensitized type 304 stainless steel at 27.5-30 MPav'ffi in 288°C water over the range of solutions conductivities of 0.1 to 0.5 ~S/cm. Data points at elevated corrosion potentials and growth rates correspond to irradiated water chemistry conditions in test or commercial reactor's.',2 14
assisted corrosion of carbon steel (FACCS),3 and other facets of boilingwater reactor (BWR) operation is well established. Early attention was focused on the relationship between the corrosion potential and sec, which resulted in the discovery that sec can be markedly suppressed if the potential is maintained below about-O.23 Vshe' thereby reducing the thermodynamic driving force for crack advance. 4,5 While many other factors (e.g., metal properties, temperature, irradiation, flow characteristics, aqueous impurities, etc.) are involved in establishing the corrosion potential, the redox properties of the coolant water are usually of primary importance. Under normal water chemistry (NWC) conditions, oxidants such as dissolved oxygen, hydrogen peroxide, and chromate ions are in stoichiometric excess of reductants (primarily hydrogen). This results in relatively high corrosion potentials, often ~.10 V.he. Hydrogenwaterchemistry(HWC),inwhich hydrogen is injected into the feedwater, has subsequently been adopted at numerous BWRs to shift the redox balance in favor of lower potentials. Under unirradiated conditions, some plants were found to require comparatively high hydrogen-addition rates to shift the corrosion potential into the desired range (less than-O.23 V,he)' In-core, more hydrogen is needed for an equivalent shift in corrosion potential. Apart from cost, undesirable side effects of high hydrogen-addition levels have also been observed in some cases, including turbine shine from 16N carry-over and WCo deposition, both of which increase the radiation level in parts of the plant. Noble metals have long been recognized as recombination catalysts for oxygen and hydrogen dissolved in water. However, it was recently found that the corrosion potential of a surface containing noble metals dropped to very low values when hydrogen was present in stoichiometric amounts (or greater), even in the absence of complete volume recombination of O 2 and H2 in the water.&-
that the effectiveness of hydrogen was limited by the irreversibility of the hydrogen/water half-reaction on oxidized stainless-steel surfaces, and that it should be possible to achieve lower corrosion potentials more readily if the catalytic properties of the stainless-steel surface were enhanced for this reaction. 6 Details of this hypothesis were reported and confirmatory data obtained using electrochemical studies and constant extension rate tests (CERTs) in Data of Ja::c:tls (e) & Asaro (•• T) SSAT2·3xl0-7 s·' ~ • CP 304SS 3 x , 021 nJcm2 .CP304&S4.6 x l 021 ,. . CP 316SS6.9 x 102' . . .... CP316SS 9.2 x 102' ..
-0_2
-0.4
0.2
Corrosion Potential. Vshe
Figure 2. The dependence of IASCC on corrosion potential as measured in slow strain tests at 3.7 x 10-7 S-1 in 288°C water on types 304 and 316 stainless steel pre-irradiated to several specific fluences. RE'of:RSII.E
Oz I HzO
iL ~ Diffusion Umited Currents lotlll-
Figure 3. Schematic E (potential) vs. log i (absolute value of current density) curves showing the interaction of H2 and O2 on a catalytically active surface such as platinum or palladium. The quantity io is the exchange current density, which is a measure of the reversibility of the the reaction. Above io. activation polarization (Tafel behavior) is shown in the sloped. linear regions. The quantity iL is the limiting current densities for oxygen diffusion, which vary with mass transport rate (e.g .. oxygen concentration, temperature, and convection). The corrosion potential in high-temperature water containing oxygen is usually controlled by the intersection of the O2 reduction curve (02 +2Hp+4e--740H-)withthe H2 0xidation curve (H 2 -7 2H+ + 2e-).
JaM. April 1992
.. ...J
;:::
...z ...o
>-
Loq Iii -
Figure 4. Schematic E (potentia/) vs. log i (absolute value of current density) curves showing the interaction of H2 and 02 on a catalytically inactive surface such as stainless steel. 200 100
• PI ELECTRODE
o 55 ELECTRODE A
l!. SS ELECTRODE B·Pd (ON METAL)
o S8 ELECTRODE A·Pd (ON OXIDE)
o 5S ELECTRODE C
--.JOU>.------O;--"--
100
200 ppb O2 IN WATER
300
400
Figure 5. The corrosion potentials of platinum, stainless steel, and stainless steel with palladium coatings on cleaned or preoxidized stainless steel vs. dissolved oxygen content in 285°C water containing 150 ppb H2 • The palladium-coated stainless steels exhibit catalytically active behavior just like platinum, and maintain a low corrosion potential despite elevated dissolved oxygen concentrations, provided stoichiometric excess hydrogen is maintained. which stainless-steel samples with surface coatings of palladium were employed. 7 Subsequently, the incorporation of small amounts of palladium or platinum (:;;0.1 wt. %) into stainless steels and related materials were also found to be effective. B In addition to improving the performance of the various steels under HWC conditions, work has now been extended to similar alloys and coatings to promote the volume recombina tion of H2 and O 2in the core water to enhance the effectiveness of HWC. Lab and field experience shows the benefits of HWC to extend to mitigation of IASCC. 2The mechanism of IASCC is similar to that of intergranular SCC (IGSCC), although detailed dependencies are shifted because of the influence of radiation damage to the alloy (e.g., hardening, creep, and segregation) and the presence of numerous radiolysis products in the water such as hydrogen peroxide (H20 2), hydroxyl and other radicals (e.g., OHo, H02°, HO), hydrogen, etc. (Figures 1 and 2.)2 Ongoing incore corrosion potential measurements contribute to a better understanding. 9 1992 April. JaM
Conversely, FACCS is markedly enhanced when the corrosion potential falls below -0.6 V,he as hydrogen is added to the feed water. This type of attack, associated with regions of high and/ or disturbed flow, stems in part from the less protecti ve corrosion film on carbon steel versus stainless steel. 3 In deaerated hightemperature water, carbon steel tends to depassivate more readily than stainless steel, and hence develops corrosion potentials in the active corrosion range (e.g., -O.8V,he versus-O.5 V,heforstainlesssteel in deaerated pure water). Once in this region, corrosion of carbon steel can occur, balanced by water reduction and oxygen reduction reactions. Because of the high overvoltage (poorly catalyzed) hydrogen-evolution reaction on the oxidized-steel surface, corrosion potentials can fall to -0.80 V,he' iO At such low potentials, the solubility of magnetite increases as "reductive dissolution"]] occurs and enhances corrosion. The presence of a catalytically active palladium or platinum coating can raise the potential to that of the reversible hydrogen/ water reaction (e.g., about -0.55 V,he) where a protective passive film may form or, at least, "reductive dissolution" would be lessened. In addition, palladium or platinum coatings may form a mechanically protective layer that also aids in mitigating corrosion. These important, interrelated behaviors are now the focus of in-depth studies designed to achieve a better understanding and thereby better control of the processes. These studies include the investigation of possible side effects of the noble-metal coatings on turbine shine from IbN carry-over and bOCO deposition.
drogen reaction, and little deviation in corrosion potential from the reversible hydrogen/water potential occurs; this potential is -0.25 V below the accepted range of the "critical" potential for protection from SCc. Figure 4 shows the more complex situation for stainless steel, wherein the exchange current density for both the hydrogen/water and oxygen/water reactions is smaller than platinum or palladium. Also indicated on this figure is a polarization curve for oxidation of stainless steel in high-temperature water. This has been drawn with only a small maximum in the active region before the current decreases to a somewhat lower passive value that is shown as essentially constant over a wide potential range. If hydrogen is in stoichiometric excess with respect to oxygen and the mass transport conditions are similar to those in Figure 3 for platinum, stainless steel exhibits much different behavior. The exchange current density (io) for the hydrogen/water reaction is much lower on stainless steel; thus, the Tafel regime for this reaction begins at much lower current densities. The intersection of this curve with the limiting current densities for the oxygen/water reaction (iL) begins to deviate from the reversible hydrogen potential even at relatively low 6 SS elECTRODE B·PeI
METAL)
055 ELECTRODEC
~ -100
I I
I ~ -200
•
•
• I c I
-300
I
\~=====<'i~
BEHAVIOR OF STAINLESS STEEL It is useful to consider the behavior of the corrosion potential (mixed potential) with the aid of schematic Evans diagrams. Figure 3 represents the behavior of hydrogen and oxygen on catalytically active metals such as platinum. A high exchange-current density (io) is shown for the hydrogen/water reaction and a somewhat lower one for the oxygen/ water reaction. In both cases, Tafel (activation polarization) behavior has been assumed with a transition to diffusion control (i L) at higher current densities for the oxygen reduction reaction. When hydrogen is in stoichiometric excess (HWC conditions), several limiting oxygen reduction-rate curves (iL ) can be shown that correspond to different oxygen concentrations and, hence, different rates of mass transport of oxygen to the electrode. In this system, over a wide range of oxygen concentrations, the intersection of the oxygen reduction curves with the hydrogen oxidation curve occurs below the Tafel region for the hy-
(O~
o SS ElECfRODE A-Pd (ON OXIDEI
ppb H2 IN WATER
Figure 6. The corrosion potentials of type 316 stainless steel and palladium-coated type 316
stainless steel In 285°C water containing 300 ppb O2 and various amounts of H2 • At about the stoichiometric H2 concentration, the corrosion potentials of the catalytically active specimens drop to approximately -0.5 V.he • iii ::
'" > e
o :304SS ... : PurePt • : 304SS+1.O"Pd
...i ...,
a : 304SS+3.0
i= z
'"E-o 0.
z
S '"0
''0""
·200
-400
u
·600
0
20
40
60
80
HYDROGEN IN WATER, ppb
Figure 7. The corrosion potentials of platinum, type 304 stainless steel, and type 304 stainless steel alloyed with 1 wt.% Pd and 3 wt.% Pd in 285°C water containing 350 ppb O2 and various amounts of H2 • 15
forSCC protection. While only platinum and palladium were evaluated, all noble metals are known to be catalytically active toward the hydrogen/water reaction. Unpublished attempts to achieve such catalytic activity from the oxide itseIfby modifying its properties through incorporation of additional nickel, cobalt, and other ions from the water were unsuccessful.
oxyg~n concentrations (e.g., i L, and i L3 ). At hIgh oxygen mass-transport rates, the limiting current for oxygen reduction is sufficiently high (e.g., iL ) that the hydrogen oxidation curve inte'rsects the linear Tafel region for oxygen reduction. This represents a maximum value for the corrosion potential for these temperature and concentration conditions, independent of any further increase in mass transport rate. Thus, improving the catalytic properties for the hydrogen/water reaction on stainless steel surfaces should markedly improve the efficiency by which hydrogen additions reduce the corrosion potential below the "critical" value
w: g:j
;;
!
0.1% PI
!
0.35% PI
: 1.0% Pt : 1.0% Pd
E
......s 1= z
'"b c.. z
o ;;; o ..: ..: o u
20
40 60 80 IMMERSION TIME, day
100
IJOO ppb 02.
120
Figure 8. The effect of palladium and platinum alloying additions to type 304 stainless steel on corrosion potentials in 285°C water containing H2 and 02'
w:
=
'" ;; E
...s ... 1= z
...'0"
z
The behavior of stainless-steel electrodes with and without palladium coatings is shown in Figure S for 28S o C water containing 150ppb hydrogen with various substoichiometric oxygen concentrations. 7 One palladinized electrode had the noble metal applied to a fresh stainless steel surface, while the other was applied to a preoxidized surface. In both cases the corrosion potentials are essentially identical to the platinum electrode over the entire range of dissolved oxygen. The two untreated electrodes polarized to much higher corrosion potentials, even at very low oxygen concentrations. The performance of both types of palladinized stainless steel electrodes was maintained for the entire -13 month duration of the tests. This behavior as a function of oxygen could also be simulated electrochemically by applying anodic currents to the electrode in deaerated water contained ISO ppb hydrogen. 7 Figure 6 shows the effect of palladium in28SoCwater.Oxygenwasmaintained at 300 ppb 02; hydrogen concentration was varied? In contrast to the unpalladinized stainless steel, where the change hibited an abrupt, large shift from -0.1 V,he to about -0.5 V,he above -24 ppb hydrogen. This hydrogen concentration is less than the stoichiometric amount required for 300 ppb oxygen (Le., 37.S ppb hydrogen). This is attributable to the higher diffusivity of hydrogen versus oxygen in the water boundary layer, since the recombination reaction occurs atthe metal surface. (A somewhat smaller diffusivity ratio of 1.83 was obtained from recent measurements of the diffusion coefficients of hydrogen and oxy-
0
<;; 0 .400
'"0
:II U
50 DISSOLVED HYDROGEN IN WATER, ppb
Figure 9. Corrosion potential measurements on platinum, type 304 stainless steel, and type 304 stainless steel thermally sprayed by the hypervelocity oxy-fuel (HVOF) technique with a powder of type 308l stainless steel containing 0.1 wt.% Pd. Data were obtained in 285°C water containing 200 ppb oxygen and varying amounts of hydrogen.
,No PtiPd ;o.t%Pd : 0.311. Pd
~:~~~~~~;~i;::'+"'.....
\
:::::::::
q
~:~
;........,,:.~ ........................... , \
\.
\.,
\:::...
in corrosion potential was small (Figure 6), the palladinized stainless steel ex-
c..
".9
••••••••,.
Corrosion Potentials on Noble Metal Coatings
200~~~~~1'~~==~==~ 150
~Ir.~~~~~~~==~
!"'~:~; ; ; ; ; ; ; ; ;,
20 40 60 80 DISSOLVED HYDROGEN IN WATER, ppb
Figure 10. The effect of palladium and plat· inum alloying additions to Stellite 6b on its corrosion potential in 285°C water containing 300 ppb 02 and varying amounts of H2. Duane Arnold BWR
0.2
!
0.1
..:
0
~
-0.1
§
-0.2
>
'"
~ .;;;
_____ .304 S8, Top of Core ..---- _____ 0 Pt, Top Qf Core
•
~
55. Bottom of Core
e -0.3 <;
C,)
-0.4 -0.5
Feedwater Hydrogen Addition. SCFM
Figure 11. The effect of feedwater hydrogen addition on the corrosion potential of stainless steel and platinum at several locations in the Duane Arnold BWR.9
gen in water over 10-S0°CY) Clearly, on a catalyzed surface, the corrosion potential can be reduced to the desired level with far less hydrogen than is required in the absence of a catalyst. This should be beneficial with regard to
the control of J6N "shine" in the turbine
building, since there is considerable evidence that low hydrogen levels in the water result in less production of volatile nitrogen species.
Effect of Noble Metal Alloying on Corrosion Potential In many instances, it is more feasible to create a catalytic surface by directly alloying with noble metals, or by thermal spraying with metallic powders containing the noble elements. Figure 7 shows the effect of palladium additions to stainless steel on the corrosion po-
Table I. Results of Constant Extension Rate Tests·
CERT No.
-I
2 3 4 5 6
Pd Thick. (11m)
0+ 0.77 0.77 0.77 0.07 0.03
ppbH, 161 161 16 9 19 20
ppbO, 95 104 196 196 251 263
Molar Ratio of H,:O. in Water at Surface (calc,§) (calc})
27.1 24.8 1.3 0.7 1.2 1.2
49.6 45.3 2.4
1.3 2.2 2.2
Potential vs. SHE Autoclave Sample (mV) (mV) -102±l2 31±11 -535±45 -110±20 -IOO±30 -515±25 S0±30 -102±32 -490±30 -150±20 -400±30 -110±10
Time to Failure (h) 70 124 125 76 118 126
Max. Stress (MPa) 407 462 476 407 469 482
Strain to Failure (%)
25 45 45 27 42 45
IGSCC (%)
--26 0 0 33 0 0
• Conditions: 287°C 0.3 M H,SO" conductivity = 0.3 )15/cm, strain rate = 1 x l()-< 5",
t Control
t
Molar ratio in water = 16 x ppb H,/ppb 0,
§ Molar ratio at surface = 1.83 x molar ratio in water, where 1.83 is the ratio of the diffusion coefficients for
16
~and
02 in water.
JOM • April 1992
tential in 285°C water containing 350 ppb O 2 and varying amounts of ~ .8 Similar data are obtained for platinum additions. These data show that the palladium-alloyed electrode drops to low corrosion potentials and exhibits behavior identical to the platinum electrode. Thus, even with small amounts of palladium or platinum addition to the stainless steel, corrosion potentials below the "critical" value « -0.23 V,he) for protection from sec can be achieved with relatively small amounts of hydrogen. Figure 8 shows that with increasing immersion time, alloys with lower levels of platinum or palladium (0.1 % Pt, 0.35% Pt, 1% Pt, and 1% Pd) respond fully catalytically. This may be attributable to the enrichment of platinum/ palladium on the surface as corrosion occurs, or to the fact that only a comparatively small number of catalytically active sites are required on the surface to achieve an adequate surface recombination rate of~ and O2 (relative to diffusion through the boundary layer). Preliminary data obtained by Auger electron spectroscopy show that some, but not extensive, surface enrichment of palladium may occur with time. 100
+
•
1.3
-100
0 41.1
+
+
+
+ +
o NO Pd • 0.77 I'm Pd 8 0.03.0.07 ~ Pd + AUTOCLAVE (NO Pdl
-400
•
•
-500
•
22
Demonstration of Reduction In
2.2
2.'
-~L---~--~~3~~4--~--~--~
eERT NO.
Figure 12. A summary of corrosion potential and CEAT data showing the relationship to the "critical" potential below which IGSCC susceptibility is low. The numbers under the points represent the H2:02 ratios from Table I. ;;;
..
== ~ E
·300
.J
...
...
1= z
·500
~ z ~
'"0OC
.700
: carbon steel : carbon sted-Pd
OC 0 <.i
:PurtPl
·900
0
100
200
300
400
DISSOLVED OXYGEN IN WATER, ppb
Figure 13. The corrosion potentials of platinum, carbon steel, and a palladium-coated ·carbon steel in 285°C water containing 150 ppb H2 and various amounts of 02'
1992 April. JaM
Similar observations of catalytic behavior have been obtained on stainlesssteel specimens thermally sprayed (using, for example, the hypervelocity oxyfuel technique) with a powder of type 308L stainless steel containing 0.1 % Pd (Figure 9). Likewise, catalytic response has also been observed for a range of engineering materials alloyed with 0.1 to 1 % Pd or Pt, including alloy 600, low alloy steel, and Stellite 6b (Figure 10). Evaluation of the effects on processing showed that alloying with noble metal had no ad verse effect on tensile yield stress, ultimate tensile strength, elongation to failure, thermal spray processing parameters, or mechanical characteristics of the coating. These quantitative measures were consistent with the qualitative observation that there was no negative impact on the ability to forge or roll these alloys. Initially, there was concern for the higher corrosion potentials obtained on catalytic versus noncatalytic surfaces in substoichiometric hydrogen conditions (e.g., the left side of Figures 6, 7, 9, and 10). This would indicate that the noble metal approach is only beneficial as long as excess hydrogen is present, and disadvantageous otherwise. However, careful evaluation of extensive in-reactor measurements (e.g., Figure 11)9 show that the potentials on platinum are below stainless steel under all conditions. While not yet confirmed in controlled laboratory experiments, the difference is most likely the result of the effect of hydrogen peroxide (which is always present in the reactor water) on the reversibility of the oxygen reduction reaction on platinum, a phenomenon known to occur in, for example, fuel cells.
see
Constant extension rate tests to directly evaluate the effects of noble metal coatings on sec were performed using previously described techniques. 13 Type 304 stainless steel (containing 0.051 % C) specimens were machined from welded 254 nun-diameter pipe that was further sensitized at 500°C for 24 hours. Results are summarized in Table J.7 The measured corrosion potentials were similar to the type 316 stainless steel samples used in the electrochemical measurements described previously; their relationship to the "critical" potential for sec is shown in Figure 12. In all cases, the corrosion potential of the type 316 stainless steel autoclave was above the critical potential since it was not palladinized. Inall CERT experiments, the dissolved oxygen concentration was maintained at a much higher level than is consistent with HWC specifications in BWRs. Also, the first two tests, which included the unpalladinized control specimen, were performed at high hydrogen-to-oxygen
. .
200
ISO
.. N
<
...
E
0
150
0
E
.,;
....'" 0
100
.
50
::== ~
H2.
usc
: C.s.·Pd, 16cc/.,I.,
: C.s.. Pd,
llOccl~n .
: c.s., 2Occlmin. : c.s.. 200cdmin .
O~~~==~~==~~ o
50
100
150
200
250
300
DISSOLVED OXYGEN IN WATER. ppb
Figure 14. The weight loss of carbon steel and palladium-coated carbon steel in 285°C water containing 150 ppb H2 as a function of O2 concentration. 200 0 0
0
150
E
"t
~
.
0
~
100
.
..J
~
~
50
0 -800
o :C.S.-Pd, 20cc/min. • :C.S.-Pd, 20Occ/min. ~ :C.S. 2Occ/min. D:C.S. 200cc/mln.
~ -600
:.
150 ppb H, + 0. 285"C -400
-200
ELEClROOE POTENTIAL, mV(SHE)
Figure 15. The effect of palladium coating on the weight loss of carbon steel as a function of corrosion potential in 285°C water containing 150 ppb H2 and various amounts of 02'
ratios. For the remainder of the tests, the molar ratio of hydrogen to oxygen at the sample surface was held close to the stoichiometric value (2:1) for the formation of water at the surface, accounting for the ratio of 1.83 for the diffusion coefficients of hydrogen and oxygen, as discussed above. When the molar ratio was greater than 2.0, the corrosion potential of the palladinized samples was well below the "critical" value (approximately-O.23 V'he)' even with only a 0.03j.lm thick palladium coating. With a ratio of less than 2.0, the corrosion potential of all specimens was above the critical value . Scanning electron microscopy analysis of the fracture surfaces reveals that the only specimens exhibiting IGseC (which was extensive) were the unpalladinized control specimen and the palladinized specimen from test 4 (Table 1), which was performed under substoichiometric hydrogen conditions (high corrosion potential). A number of intergranular cracks were also evident on the free surface near the fracture surface in both cases. In all other tests, the fracture surface was predominantly ductile morphology, with sometransgranularcrack17
ing on the fracture and free surfaces, as has been widely observed in other CERT experiments at low potentials and relatively high strain rates. 14•15 Thus, in spite of the relatively high oxygen concentrations and low hydrogen concentrations in the water, the palladium coatings successfully catalyzed corrosion potentials lower than those of the control specimens and, indeed, below the "critical" value for IGSCC. Further, this behavior was achieved and sustained with a palladium coating as thin as 0.03 Ilm. In agreement with the measured corrosion potentials, the control sample and the palladinized sample deliberately held at a high potential (with a hydrogen-tooxygen ratio of less than 2.0) manifested IGSCC, while the remainder of the palladinized samples did not. This clearly demonstrates that the corrosion potential, rather than the palladium coating thickness, per se, is responsible for the improved behavior. BEHAVIOR OF CARBON STEEL
While sec can also occur in carbon steels at elevated potentials, this section focuses on the benefits of noble-metal coatings and alloying additions in reducing flow-assisted corrosion. Since they essentially lack chromium and nickel, carbon steels tend to depassivate more readily than stainless steel and hence develop corrosion potentials in the active corrosion range (Figure 4). Once in this region, accelerated corrosion can occur, with the metal dissolution balanced by water reduction as well as oxygen reduction. In the absence of dissolved oxygen, dissolution of carbon steel is balanced by reduction of water (hydrogen/water reaction), and thus potentials below the reversible hydrogen/watervalue (Le., less than -0.55 V,he in pure water) can occur. Indeed, because of the high overvoltage (poorly catalyzed) hydrogen-evolution reaction on the oxidized steel surface, the corrosion potential can fall well below -0.55 Vshe' to approximately -0.8 Vshe ' At such low potentials, the solubility of magnetite increases as "reductive dissolution" occurs and enhances corrosion. By catalyzing the hydrogen/water reduction reaction, the presence of a catalytically active palladium coating can raise the potential to that of the reversible hydrogen/water reaction where a protective passive film may form, or, at least, "reductive dissolution" would be lessened. Also, thicker films form a mechanically protective layer that aids in mitigating corrosion. The complementary benefits of noble metals result from catalyzing the hydrogen/water reaction. At high dissolvedoxygen concentrations (with excess hydrogen), the corrosion potential is controlled by reducing oxygen and oxidiz1.8
ing hydrogen. By catalyzing the hydrogen/water oxidation reaction, the corrosion potential is shifted toward that reaction, to lower values. In contrast, with carbon steel at very low oxygen concentrations, the corrosion potential is controlled by the hydrogen/water reduction reaction and the carbon steel dissolution reaction (Figure 4). Catalyzing the hydrogen/water reduction reaction again shifts the corrosion potential toward that reaction, but now to higher values. Observed Potentials
Figure 13 shows the corrosion potential of carbon steel and palladium-coated carbon steel in 285°C water containing 150 ppb dissolved hydrogen as a function of oxygen concentration. 10 The fully palladium-coated carbon steel electrode possessed a potential similar to that of the platinum electrode in water containing 150 ppb hydrogen and was not affected by oxygen concentration. In contrast, the behavior of the unpalladinized carbon steel electrode did not parallel that of platinum; it exhibited a much lower corrosion potential at low oxygen concentrations. This is attributable to a high overvoltage for hydrogen evolution and to the formation of a less protective oxide, permitting high corrosion rates in the low-oxygen regime. Weight Loss Data
Figures 14 and 15 show the effect of palladium coating on the weight loss of carbon steel in water containing 150 ppb hydrogen at 285°C as a function of oxygen concentration and as a function of electrode potential, respectively.lO The data represent the average value of three specimens. The weight loss of carbon steel can be significantly reduced by palladium coatings O.3-0Allm thick that serve as a protective barrier layer. At low oxygen concentrations, the unpalladinized carbon steel electrode suffered relatively high corrosion. However, there was no Significant effect of the two different flow rates used in this experiment. CONCLUSIONS
These results clearly demonstrate some of the benefits of noble metal coatings and alloying additions on stainless steel in efficiently reducing the corrosion potential below the "critical" value for SCC, even at relatively low hydrogen and high oxygen concentrations. The corrosion potential benefits described for stainless steel also apply to other iron-, nickel-, and cobalt-base alloys, such as low-alloy steel, alloy 600, and Stellite. Advantages of this approach include the ability to achieve low corrosion potentials with relatively high oxygen concentrations in the water; a reduction in the amount of hydrogen required to attain corrosion potentials below the
"critical" value for SCC; and the likelihood that the increase in radiation levels in parts of the plant from turbine shine from 16N carry-over and 6OCO deposition can be minimized. For carbon steels, palladium (and other noble-metal) coatings are successful in minimizing flow-assisted corrosion by preventing the corrosion potential from decreasing from about-O.55 Vshe to about -0.8 Vshe in reducing environments (e.g., where oxygen is absent). As with stainless steel, small hydrogen additions coupled with noble metal coatings on carbon steel will also aid in achieving corrosion potentials below the "critical" potential for sec in oxidizing environments. ACKNOWLEDGEMENT
The authors acknowledge General Electric Nuclear Energy who funded this work and provided advice and encouragement. References I. R.L. Jones et aI .• Corr05ion/84 (Paper no. 167 presented at the National Association of Corrosion Engineers, Houston, TX,1984). 2. P.L. Andresenet aI., Proc. FourthInternatioMI Symposium on
Environmental Degradation of MlIterials in Nuclear Power Sys-
tems (Houston, TX: NACE, 1990), pp. 1-83 to 1-121. 3. G.J. Bignold et ai., Nucl. Eng. Inter. (1981), p. 37. 4. M.E. Indig et aI., Proc. 2nd Int. Conf. on Environmental
Degradation of MlIterials in Nuclear Power Systems-Water
Reactors (La Grande Park, IL: ANS, 1986), p. 411. 5. L.G. Ljungberg. D. Cubicciotti, and M. Trolle, Corrosion, 42 (1986), p. 263. 6. L.W. Niedrach and W.H. Stoddard, Corrosion, 42 (1986), p. 696. 7. L.W. Niedrach, Corrosion, 47 (1991), p. 162. 8. P.L. Andresen, Y.J. Kim, and L.W. Niedrach, unpublished work. 9. M.E. Indig and J.L. Nelson, Corrosion, 47 (1991), p. 202. 10. Y.J. Kim and L.W. Niedrach, Extended Abstracts, 91-2 (Pennington, NJ: Electrochem. Soc., 1991), abstract no. 174. 11. Ph. Berge, C. Ribon, and P. SI. Paul. Corrosion, 32 (1976), p.223. 12. P.T.H.M. Verhallen el.aI., Chem. Eng. Science, 39 (1984), p. 1535. 13. P.L. Andresen, Environment-Sensitive Fracture: Evaluation and Comparison of rest Methods, SIP 821, ed. S.W. Dean et aI. (Philadelphia. PA: ASTM, 1984), p. 271. 14. F.P. Ford and M.J. Povich, Corrosion, 35 (1979), p. 569. 15. W.E. Ruther et ai., Corrosion, 40 (1984). p. 518.
ABOUT THE AUTHORS _ _ _ __ Young-Jln Kim earned his Ph.D. in metallurgy/corrosion science at the University of Minnesota in 1985. He is currently a member ~f the R&D staff at the General Electric Corporate R&D Center in Schenectady, New York. Leonard W. Niedrach earned his Ph.D. in physical chemistry at Harvard University in 1948. He is currently a consultant at the General Electric Corporate R&D Center in Schenectady, New York. Maurice E. Indig earned his Ph.D. in materials science at Rensselaer Polytechnic Institute in 1974. He is currently a lead engineer at General Electric Nuclear Energy in Pleasanton, California. Peter L. Andresen earned his Ph.D. in materials science at Rensselaer Polytechnic Institute in 1978. He is currently a staff scientist at General Electric Corporate R&D Center in Schenectady. New York. Dr. Andresen is also a member of TMS. H you want more information on this subject, please circle reader service card number 53.
JOM • April1992
