Comparison of corrosion-resistance and hydrogen permeation properties of Zn–Ni, Zn–Ni–Cd and Cd coatings on low-carbon steel

Corrosion Science 45 (2003) 1505–1521 www.elsevier.com/locate/corsci

Comparison of corrosion-resistance and hydrogen permeation properties of Zn–Ni, Zn–Ni–Cd and Cd coatings on low-carbon steel Hansung Kim a, Branko N. Popov

a,*

, Ken S. Chen

b

a

b

Center for Electrochemical Engineering, Department of Chemical Engineering, University of South Carolina, Columbia, SC 29208, USA Engineering Science Center, Sandia National Laboratories, Albuquerque, NM 87185, USA Received 5 April 2002; accepted 20 November 2002

Abstract Zn–Ni–Cd alloy was electroplated from an alkaline sulfate bath under potentiostatic conditions. The corrosion and hydrogen permeation characteristics of Zn–Ni–Cd alloy coatings electrodeposited from alkaline bath were studied and compared with those of Cd and Zn–Ni coatings obtained using commercial baths. Zn–Ni–Cd alloy was electroplated from an alkaline sulfate bath under potentiostatic conditions. The corrosion potential of this Zn– Ni–Cd coating was )0.62 V vs. SCE, which is still negative potential compared to iron. The corrosion rate of Zn–Ni–Cd coated steel was 0.073 mm y
1 , which is estimated in a 0:5 M Na2 SO4 þ 0:5 M H3 BO3 solution at a pH of 7. This value is much lower than the corrosion rate of Zn–Ni alloy (0.502 mm y
1 ) and Cd (0.306 mm y
1 ) coatings deposited from commercial baths. Zn–Ni–Cd alloys are also demonstrated to have superior hydrogen permeation inhibition properties compared to Cd and Zn–Ni coatings. Kinetic parameters of hydrogen permeation such as the transfer coefficient, a, the modified exchange current density, i00 , thickness dependent adsorption–absorption rate constant, k 00 , recombination rate constant, k3 , surface hydrogen coverage, hs , were evaluated by applying a mathematical model to analyze experimental results. Ó 2003 Elsevier Science Ltd. All rights reserved.

*

Corresponding author. Tel.: +1-803-777-7314; fax: +1-803-777-8265. E-mail address: [email protected] (B.N. Popov).

0010-938X/03/$ - see front matter Ó 2003 Elsevier Science Ltd. All rights reserved. doi:10.1016/S0010-938X(02)00228-7

1506

H. Kim et al. / Corrosion Science 45 (2003) 1505–1521

Nomenclature a b Cs D F f ic ja ir i00 k1 k3 kabs kads k 00 L R T

a constant, F =RT , 1/V a constant, L=FD, mol/A cm absorbed hydrogen concentration, mol/cm3 hydrogen diffusion coefficient, cm2 /s FaradayÕs constant, 96 487 C/mol a constant ¼ c=RT , dimensionless charging current density, A/cm2 steady state permeation flux, A/cm2 steady state recombination flux, A/cm2 modified exchange current density, A/cm2 discharge reaction rate coefficient, mol/cm2 s recombination rate constant, mol/cm2 s absorption rate constant, mol/cm2 s adsorption rate constant, cm/s thickness dependent absorption–adsorption rate constant, mol/cm3 membrane thickness, cm gas constant, 8.314 J/mol K temperature, K

Greek symbols a transfer coefficient, dimensionless g overvoltage, V hs surface hydrogen coverage, dimensionless c gradient of the apparent standard free energy of adsorption with hydrogen coverage, J/mol

1. Introduction Hydrogen is evolved during electrochemical processes such as electroplating, corrosion and cathodic protection. A portion of the evolved hydrogen is adsorbed on the metallic surface, the extent of which depends on the adsorption kinetics of the surface. A part of the adsorbed hydrogen absorbs and subsequently diffuses into the crystalline lattice of the substrate, where it reacts with the metal atoms to form brittle metal hydrides, causing the structure to fail even at applied stresses far below the yield strength [1]. The hydride formation is general mechanism of hydrogen embrittlement and applies only to those metal that can form hydrides. The hydrogen atoms may also combine into hydrogen molecules in the microvoids of the material. The pressure exerted by molecular hydrogen and atomic hydrogen is high enough in most cases to cause blisters and ultimately failure of the material. Various methods have been proposed in the literature to decrease the hydrogen embrittlement [2–5]. These methods include the post-heat treatment, alloying, laser

H. Kim et al. / Corrosion Science 45 (2003) 1505–1521

1507

surface modification and shot preening. These methods however, do not reduce the hydrogen entry below the threshold level that is safe of cracking hazards. On the other hand, coating with a suitable metal/alloy can successfully reduce the hydrogen ingress into the metal [6]. The coatings can either be metallic or may consist of metallic particles bonded with an inorganic matrix, often phosphates and/or chromates. An important requirement, however, is that the coating should provide sacrificial protection if the substrate becomes exposed. Extensive research has been done on reducing the corrosion and hydrogen permeation of different metals in corroding environment [7–17]. Chen and Wu [7] reported the effect of copper, tin, silver and nickel-plating on hydrogen permeation inhibition in AISI 4140 steel. It was found that copper and tin effectively reduced the hydrogen permeation by over 80%. However, the corrosion resistance behavior was not presented for these electrodeposited metals. Cadmium plating has been extensively used as a corrosion resistant coating on hard steel for various applications [8]. Zamanzadeh et al. [9] found that deposits of cadmium reduced the hydrogen permeation on iron. Previously our group [12–17] has developed a variety of electrodeposition schemes to provide corrosion and hydrogen permeation properties equal to or better than that of cadmium. The acid based Zn–Ni–Cd deposition process resulted in deposits with better corrosion and hydrogen permeation resistant characteristics. However, it was observed that the deposit from the acid bath showed poor adherence [18]. Recently we developed an alkaline electrodeposition process to deposit good quality Zn–Ni–Cd deposits with high Ni and Zn contents [19]. The resultant Zn–Ni–Cd deposits have excellent corrosion and barrier properties. In this work, we compare the corrosion and hydrogen permeation inhibition properties of Zn–Ni–Cd deposits obtained using the alkaline process with Zn–Ni and Cd deposits obtained using commercial electrolytic baths. The permeation and kinetic parameters for hydrogen permeation such as the absorption–adsorption constant, the recombination reaction constant, the modified exchange current density and hydrogen surface coverage were determined using the Iyer et al. [20] model or its modified model. The mechanism of preventing hydrogen permeation for each coating will be compared and explained using the kinetic parameters obtained by fitting data to the models.

2. Experimental Zn–Ni–Cd, Zn–Ni and Cd thin films were electrodeposited on low-carbon steel foils with thickness of 0.1 mm and area of 3 cm2 . Prior to the electrodeposition, the steel membrane was mechanically polished with 600 grade sand paper and 0.5 micron high purity alumina powder to a mirror finish. It was then cleaned in an ultrasonic cleaning bath. The cleaned steel membrane was mounted on the special sample holder exposing only one side of the membrane to the plating solution. Depositions were carried out in a three electrode setup using an EG&G (type 273A) potentiostat/galvanostat. Standard calomel electrode (SCE) was used as the reference electrode and a platinum mesh served as the counter electrode. Zn–Ni–Cd was electrodeposited from the alkaline sulfate bath developed by us at USC [19]. Deposits obtained from

1508

H. Kim et al. / Corrosion Science 45 (2003) 1505–1521

electrolytic baths containing 60 g/l of ZnSO4 , 40 g/l of NiSO4 , 1 g/l CdSO4 and 80 g/l of (NH)4 SO4 in the presence of additives at pH ¼ 9:3 have a composition of 50 wt% of Zn, 28 wt% of Ni and 22 wt% of Cd. Zn–Ni and Cd depositions were carried out using a commercial bath obtained from SIFCOâ . The other side of the steel membrane was electrodeposited with the thin layer (0:15–0:2 lm) of palladium to avoid corrosion reactions on steel that might perturb the permeation analysis. Pd deposition was carried out using 1  10
4 M potassium tetranitropalladate (K2 Pd(NO2 )4 ) at a current density of 200 lA/cm2 for 2 h. The prepared membrane samples were finally washed thoroughly with deionized water, dried in air and fitted into the permeation cell. Electrochemical characterization was carried out in a solution containing 0:5 M Na2 SO4 þ 0:5 M H3 BO3 at a pH of 6.5 using the same three-electrode setup that was used for electrodeposition. The rate of hydrogen permeation through the membranes was measured continuously as a function of time using the Devanathan– Stachurski permeation cell [22]. This cell-setup consists of two working compartments, cathodic and anodic chamber, and the membrane is placed between two chambers. The palladium-plated side faces the anodic side and Zn–Ni–Cd deposited side faces the cathodic side. Permeation studies were done using an EG&G PAR (model 273A) potentiostat connected to each side. Platinum mesh was used as a counter electrode and the membrane was used as a bipolar working electrode. SCE and mercuric oxide (Hg–HgO) reference electrodes were used for the cathodic and anodic sides, respectively. The anodic compartment was filled with 0.2 M of NaOH solution and potential was kept at )0.3 V vs. Hg/HgO reference electrode until the background current was reduced to below 1 lA/cm2 . This potential corresponds to a practically zero concentration of atomic absorbed hydrogen on the surface. Then, the cathodic compartment was filled with a supporting electrolyte containing 0.5 M Na2 SO4 and 0.5 M H3 BO3 buffer solution with pH ¼ 7. Nitrogen was purged on both sides through out the experiment in order to remove dissolved oxygen from the electrolytes. The membrane on the cathodic side of the cell was polarized potentiostatically, creating conditions for hydrogen evolution. Hydrogen generated on the cathodic side permeates through the membrane and gets oxidized on the anodic surface of the membrane. The steady state currents associated with anodic (permeation current) and cathodic (charging current) reactions were monitored continuously with changing the overpotential of hydrogen evolution reaction (HER) at the cathode side.

3. Results and discussion 3.1. Corrosion characterization The corrosion characteristics of Zn–Ni–Cd ternary alloy deposit from alkaline bath were evaluated using Tafel and linear polarization and compared with Cd and Zn–Ni films deposited from a commercial bath. Tafel polarization behavior shown in Fig. 1 indicates that the corrosion current was much smaller for Zn–Ni–Cd than those of Cd and Zn–Ni coatings. Zn–Ni–Cd alloy deposited potentiostatically at

H. Kim et al. / Corrosion Science 45 (2003) 1505–1521

1509

-0.5

-0.6

Potential (V vs. SCE)

-0.7

-0.8

Zn-Ni-Cd from alkaline bath E corr : -0.62 V I : 4.82 × 10-6 A/cm2 corr

Commercial Cd E corr : -0.79 V

-5

I corr : 1.43 × 10

A/cm2

-0.9

-1.0

-1.1

Commercial Zn-Ni (86/14) E corr : -1.16 V I : 3.5 × 10-5 A/cm2 corr

-1.2

-1.3 10 -8

10-7

10-6

10-5

10-4

10-3

10-2

Current density (A/cm2) Fig. 1. Tafel polarization plots of Zn–Ni–Cd deposited from the alkaline sulfate solution and Zn–Ni, Cd deposits obtained from the commercial baths.

)1.3 V vs. SCE from alkaline bath of pH 9.3 containing 1 g/l of CdSO4 , 40 g/l of NiSO4 and 60 g/l ZnSO4 showed the composition ratio of 50/28/22 for Zn, Ni and Cd, respectively. By addition of Cd to the Zn–Ni system, the Zn content in the deposit decreased, causing a shift of the open circuit potential of the alloy to )0.62 V vs. SCE. Co-deposition of Cd mainly replaces the portion of Zn and also results in an increase of the nickel content in the deposit. The linear polarization plots shown in Fig. 2 indicate that Zn–Ni–Cd has higher polarization resistance when compared to the other coatings. The polarization resistance obtained using the linear and Tafel polarization studies were used along with the compositional analysis to estimate the corrosion rates of different coatings. Table 1 summarizes the corrosion resistance of different coatings tested in 0:5 M H3 BO3 þ 0:2 M Na2 SO4 solution at a pH of 6.5. The estimated corrosion rate of Zn–Ni–Cd alloy from the Tafel polarization is smaller than that of Zn–Ni and Cd by a factor of 7 and 4, respectively. This ensures the longevity of the coating for industrial applications. A much thinner coating can be used to achieve the same protection that would have been obtained using a thicker cadmium coating. The results indicated that Zn–Ni–Cd deposit has excellent barrier properties in terms of corrosion resistance. 3.2. Hydrogen permeation test Fig. 3 shows the hydrogen permeation current of the Cd deposited steel membrane as a function of the applied cathodic potential. During the start-up, the

1510

H. Kim et al. / Corrosion Science 45 (2003) 1505–1521

6.0

Zn-Ni-Cd (ratio: 50/28/22) P

4.5

Overpotential (mV vs. SCE)

Commercial Cd R = 403 Ω P E corr = -0.79 V

R = 2942 Ω E

corr

= -0.62 V

3.0 Commercial Zn-Ni (86/14)

1.5

R = 98 Ω P E corr = -1.16 V

0.0 -1.5 -3.0 -4.5 -6.0

-40

-30

-20

-10

0

10

20

30

40

Current Density (µ Α/cm ) 2

Fig. 2. Linear polarization plots for Zn–Ni–Cd, Zn–Ni and Cd coatings in a 0:5 M Na2 SO4 þ 0:5 M H3 BO3 solution at a pH of 7.0.

Table 1 Composition and corrosion rates of different coatings Deposit

Ni

Cd

Corrosion potential (V vs. SCE)

Corrosion rate (mm y
1 )

Tafel constant (V)

Zn

ba

bc

Zn [28] 100 Ni [28] 0 Zn–Ni (commercial bath) 86 Cd (commercial bath) 0 Zn–Ni–Cd (alkaline bath) 50

0 100 14 0 28

0 0 0 100 22

)1.14 )0.358 )1.16 )0.79 )0.62

1.239 0.0218 0.502a /0.909b 0.306a /0.467b 0.073a /0.062b

– – 0.020 0.027 0.041

– – )0.069 )0.081 )0.091

a b

Composition (wt%)

Calculated from Tafel polarization. Calculated from linear polarization.

background current on the anodic side was allowed to stabilize below 1 lA/cm2 . Following this, the cathodic side of the membrane was polarized in the negative direction to provide conditions for hydrogen evolution. Hydrogen that comes out of the anodic side is oxidized potentiostatically in order to obtain the permeation current. The steady state permeation current for each value of applied cathodic potential was noted down. After stepping the potential a few times, the applied potential was switched off and the decay curve was recorded. Using the same process, the permeation current and hydrogen evolution current of the Zn–Ni and Zn–Ni–Cd coated membranes were measured for each values of the applied overpotential.

H. Kim et al. / Corrosion Science 45 (2003) 1505–1521

1511

Permeation current density (µΑ/cm2)

10 OCV : -0.78 V vs. SCE

-1.33 V

8

-1.28 V

-1.23 V 6

-1.18 V 4

-1.08 V

2

0 0

1000

2000

3000

4000

5000

6000

7000

Time (s) Fig. 3. The hydrogen permeation transients through Cd deposited iron membrane as a function of time for different applied cathodic potentials in an electrolyte containing 0:5 M H3 BO3 þ 0:5 M Na2 SO4 (pH ¼ 7:0) solution.

Figs. 4 and 5 show the dependence of the HER current (ic ) and the hydrogen permeation current (ia ) for the three different coatings (Cd, Zn–Ni and Zn–Ni–Cd). As seen from Fig. 4, the cathodic current density is less for Cd deposits at any particular overpotential indicating a slower kinetics for HER. Zn–Ni coatings exhibited high hydrogen discharge currents due to the presence of high amount of Zn [21]. Zn–Ni–Cd coatings, owing to the presence of Cd had lower current for hydrogen evolution indicating slower kinetics. For a typical applied overpotential of )400 mV, the hydrogen evolution current density is around 51 lA/cm2 for Cd deposited steel. On the other hand, Zn–Ni coated steel has a very high cathodic current density of 5000 lA/cm2 at the even low overpotential of )240 mV. The hydrogen evolution current density in case of Zn–Ni–Cd was around 270 lA/cm2 at the overpotential of )310 mV. Fig. 5 compares the permeation current densities for the Cd, Zn–Ni and Zn–Ni–Cd. A very large decrease in the permeation current density is seen for Zn–Ni–Cd deposits obtained from alkaline electrolytes. The permeation current values were nearly zero under steady state conditions. For applied overpotential of about )310 mV, the permeation current increased to a value of about 0.23 lA/cm2 in the case of Zn–Ni–Cd deposited steel. The increase in permeation current density was about 10 times higher in the case of Cd deposited steel and about at least 20 times higher in the case of Zn–Ni deposited steel. Cd deposited steel, in spite of possessing low currents for hydrogen evolution, exhibits higher hydrogen permeation. This suggests that Cd could possess lower kinetics for HER, and unfortunately

1512

H. Kim et al. / Corrosion Science 45 (2003) 1505–1521 0.0

Zn-Ni

Overpotential (V)

-0.1

-0.2

Zn-Ni-Cd Cd

-0.3

-0.4

-0.5

-0.6 1

10

100

1000

Cathodic current density (µΑ

10000

/cm2)

Fig. 4. The dependence of cathodic current densities (ic ) on the overpotential (g) for Zn–Ni, Zn–Ni–Cd and Cd deposited iron membranes.

0.0

Zn-Ni

Overpotential (V)

-0.1

-0.2

Zn-Ni-Cd -0.3

Cd -0.4

-0.5

-0.6 0.1

1

10

Permeation current density (µΑ/cm2) Fig. 5. Plot of permeation current densities (ia ) as a function of applied overpotential (g) for Zn–Ni, Zn– Ni–Cd and Cd deposited iron membranes.

H. Kim et al. / Corrosion Science 45 (2003) 1505–1521

1513

a higher adsorption and absorption kinetics, which may cause majority of the hydrogen adatoms to diffuse into the substrate. In order to gain better understanding about the permeation inhibition properties of different coatings, it is important to evaluate the various kinetics parameters governing hydrogen permeation. Several models [20,23–25] have attempted to explain the hydrogen entry into metals and to evaluate the kinetic and permeation parameters required to explain the reaction process. The Tafel slope and dg=dlogðia Þ obtained for any particular system decide the selection of an appropriate model that can be used to evaluate the kinetics parameters for hydrogen permeation. Since the Tafel slope has a range of 124–130 mV/decade and dg=dlogðia Þ shows 225–235 mV/ decade for Zn–Ni and Zn–Ni–Cd, suggesting a presence of a coupled discharge– chemical recombination mechanism, the Iyer–Pickering–Zamanzadeh model (I–P–Z model) was chosen to fit the data [20]. The model analysis carried out for Cd, Zn–Ni and Zn–Ni–Cd is presented below along with a brief description of the details of the model. The basic I–P–Z model assumes that: (i) the HER is a coupled discharge–recombination process; (ii) g  RT =F so that the backward reactions are negligible; (iii) the surface coverage is low enough so that a Langmuir isotherm is used for the hydrogen coverage; (iv) the intermediate hydrogen adsorption–absorption reaction is in local equilibrium; (v) steady state condition is assumed and once the steady state is attained the permeation current density is described by a simple diffusion of hydrogen through the metal membrane without any trapping. With these assumptions, the following relationships were obtained: ic ¼ i00 ð1
hs Þe
aag

ð1Þ

ir ¼ Fk3 h2s ¼ ic
ia

ð2Þ

ia ¼ F ðkabs hs
kads Cs Þ ¼ F ðD=LÞCs

ð3Þ




 kads þ DL Cs bia Cs ¼ 00 ¼ 00 hs ¼ kabs k k

ð4Þ

k 00 pffiffiffi ia ¼ pffiffiffiffiffiffiffi ir b Fk3

ð5Þ

F ag

ic e RT ¼


bi00 ia þ i00 k 00

ð6Þ

where ic is the charging current; ia is the permeation current; ir is the recombination current; i00 ¼ Fk1 ; with k1 being the discharge rate coefficient; F is Faraday constant; hs is the surface coverage by hydrogen; a ¼ F =RT ; a is the transfer coefficient; g is the overpotential; k3 is the recombination rate constant; Cs is the surface hydrogen concentration; k 00 is called the equilibrium absorption–adsorption constant; kabs is the absorption rate constant; kads is the adsorption rate constant; D is the hydrogen diffusion coefficient and L is the thickness of the membrane.

1514

H. Kim et al. / Corrosion Science 45 (2003) 1505–1521

pffiffiffi The square root of the recombination current density ( ir ) vs. the permeation current density (ia ) for the Zn–Ni, Cd and Zn–Ni–Cd deposited steel are plotted in Figs. 6 and 7, respectively. For the case of Zn–Ni and Zn–Ni–Cd, the plots are linear and pass through or close to the origin, which is in agreement with the assumptions and subsequent deductions achieved in Eq. (5). To further ascertain the validity of the model and for parameter estimation, the linear dependence between ia and ic eaag , a consequence of Eq. (6), is verified. According to Figs. 8 and 9, the dependences in the case of Zn–Ni and Zn–Ni–Cd show linear slopes. The slopes and intercepts in Figs. 8 and 9 are used to calculate i00 and k 00 for the Zn–Ni and Zn–Ni–Cd coated steel samples. The transfer coefficient, a, which was required for the parameter estimation was derived using the following quadratic equation [20]: a2 þ ½ð2Sc
Sa Þ=aa þ ½Sc ðSc
Sa Þ=a2 Þ ¼ 0

ð7Þ

where Sc ¼ dlnðic Þ=dg and Sa ¼ dlnðia Þ=dg. Sc and Sa , the cathodic and anodic slopes required for the transfer coefficient estimation are obtained from Figs. 4 and 5. Using this equation, the transfer coefficient, a, was determined to be 0.47 and 0.46 for Zn– Ni and Zn–Ni–Cd, respectively. Two roots were obtained from the calculation and the one that results in negative values of i00 and k 00 was discarded. pffiffiffi In the case of Cd coated steel, we did not obtain linear dependences for ir vs. ia . This non-linearity arises due to the activation of HER [20,26]. Under these conditions, the Frumkin–Temkin corrections have to be applied to the discharge and 10

2 ia (µΑ/cm )

8

Cd

6

Zn-Ni

4

2

0 0

10

20

30

40 0.5

(ir)

50

60

70

80

2 0.5

(µΑ/cm )

pffiffiffi Fig. 6. The hydrogen permeation current (ia ) vs. square root of the recombination current ( ir ) for Zn–Ni and Cd coatings.

H. Kim et al. / Corrosion Science 45 (2003) 1505–1521

1515

0.40 0.35

Zn-Ni-Cd

2 ia (µΑ/cm )

0.30 0.25 0.20 0.15 0.10 0.05 0.00 0

4

8

12

16 0.5

(ir)

20

24

28

32

(µΑ/cm2)0.5

pffiffiffi Fig. 7. The hydrogen permeation current (ia ) vs. square root of the recombination current ( ir ) for Zn– Ni–Cd alloy. 63

Zn-Ni

ic exp(aαη) (µΑ /cm2)

62

61

60

59

58

57 0

1

2

3

4

5

6

ia (µΑ/cm2) Fig. 8. Plot of the hydrogen charging function (ic eaag ) vs. steady state hydrogen permeation current (ia ) for Zn–Ni coating.

1516

H. Kim et al. / Corrosion Science 45 (2003) 1505–1521 0.96

ic exp(aαη) (µΑ/cm2)

0.94

Zn-Ni-Cd

0.92

0.90

0.88

0.86 0.10

0.15

0.20

0.25

0.30

0.35

0.40

0.45

ia (µΑ/cm2) Fig. 9. The hydrogen charging function (ic eaag ) vs. steady state hydrogen permeation current (ia ) for Zn– Ni–Cd coating.

recombination currents [27]. Eqs. (8) and (9) give the modified set of charging and recombination currents, ic ¼ i00 ð1
hs Þe
aag e
af hs

ð8Þ

ir ¼ Fk3 h2s e2af hs

ð9Þ

where f 0 ¼ c=RT and c is the gradient of the apparent standard free energy of adsorption with coverage. The value of f 0 is taken to be equal to 4.5 [27]. After suitable modifications, Eqs. (8) and (9) can be written as ! afbia
 ic e k00 ð10Þ ln ¼
aag þ ln i00 bia 1
k00 ln

 pffiffiffi   pffiffiffiffiffiffiffi afb b Fk3 ir þ ln ¼ i a k 00 ia k 00

ð11Þ

pffiffiffi The non-linear trend obtained in case of Cd coated steel for the. ir vs. ia dependence arises as a result of the exponent term added to the characterizing currents to compensate for the higher surface coverage. pffiffiffi As presented in Fig. 10, a plot of lnð ir =ia Þ vs. ia shown for Cd coated steel is linear. The parameters of interest were evaluated using the slopes and intercepts of

H. Kim et al. / Corrosion Science 45 (2003) 1505–1521

1517

1.5

0.5

ln( ir /ia )

1.0

0.5

0.0

-0.5 2

4

6

8

10

ia (µΑ/cm2) pffiffiffi Fig. 10. Relationship between lnð ir =ia Þ and the hydrogen permeation current (ia ) for Cd coating.

plotting equations (10) and (11) (Figs. 10 and 11) by using an iterative procedure. The value of a was assumed to be 0.5 for the first guess. pffiffiffiWith this value for a, the constant k 00 can be obtained from the slope of plot ln( ir =ia ) vs. ia (Fig. 10). The recombination rate constant k3 can be determined from the intercept of the same plot. The modified exchange current density i00 is determined from the intercept of the lnfic eðaf hs Þ =ð1
hs Þg vs. g plot. Knowing i00 from the intercept of the plot, we proceeded to verify the a value assumed. Using the a value obtained from the slope of Eq. (10), a new value for k 00 was found by a regression analysis of Eq. (11). The procedure was repeated until a converged to a fixed value. The surface coverage of hydrogen was calculated using Eq. (4) and is plotted as a function of applied overpotential. Fig. 12 shows the hydrogen surface coverage for Zn–Ni, Zn–Ni–Cd and Cd deposited steel. At the applied range of overpotential, the hydrogen surface coverage on Cd coated steel showed higher value than 0.1 which allows to the Tempkin isotherm. The surface coverage remained low in the case of Zn–Ni–Cd and Zn–Ni coated steel at all applied potentials validating the use of Langmuir isotherm. Table 2 summarizes different constants that characterize hydrogen permeation in the Zn–Ni–Cd, Zn–Ni and Cd coated steel samples. A diffusion coefficient of 2:67  10
7 cm2 /s was used for these model analyses [18]. The thickness of the coating was assumed to be negligible when compared to the thickness of the steel membrane. Thus, any effect of the coating on the diffusion of hydrogen is neglected.

1518

H. Kim et al. / Corrosion Science 45 (2003) 1505–1521

ln{icexp(αf'bia/k'')/(1-bia/k'')} (µΑ/cm2)

10 9 8 7 6 5 4 3 2 -0.60

-0.55

-0.50

-0.45

-0.40

-0.35

-0.30

-0.25

Overpotential (V) Fig. 11. Plot of hydrogen coverage corrected ic vs. hydrogen overpotential (g) for Cd coating.

Hydrogen surface coverage (θs)

0.5

0.4

Cd 0.3

0.2

Zn-Ni-Cd

0.1

Zn-Ni

0.0 0.0

-0.1

-0.2

-0.3

-0.4

-0.5

-0.6

Overpotential (V) Fig. 12. Plot of surface hydrogen coverage (hs ) vs. hydrogen overpotential (g) for Zn–Ni, Zn–Ni–Cd and Cd coatings.

H. Kim et al. / Corrosion Science 45 (2003) 1505–1521

1519

Table 2 The kinetic parameters characterizing the hydrogen evolution and permeation through various coatings Deposit (wt%) Zn–Ni–Cd (50/28/22) Zn–Ni (86/14) Cd

i00 (A/cm2 )
7

9.17  10 6.30  10
5 6.22  10
9

k 00 (mol/cm3 )
6

1.11  10 2.10  10
5 7.32  10
5

k3 (mol/cm2 s) 0.38  10
6 5.73  10
6 1.66  10
9

As can be seen in the table, Zn–Ni coated steel shows the highest recombination constant. High recombination constant is beneficial in preventing the hydrogen permeation. However, it has a high value of i00 , explaining the preference of HER when compared to those of Zn–Ni–Cd and Cd coated steel. It also has a more negative corrosion potential ()1.16 V vs. SCE) due to the high concentration of Zn in the Zn–Ni alloy [19]. The adsorption–absorption coefficient, k 00 , was the largest for Cd coated steel followed by Zn–Ni and Zn–Ni–Cd. Therefore, the positive effect of the decrease in the HER of Cd is diminished by the increased adsorption–absorption rate constant and the decreased recombination rate constant. With addition of Cd into the Zn–Ni, the Zn content in the alloy is suppressed and the corrosion potential shifts to positive values ()0.62 V vs. SCE) [19]. This causes the observed decrease in the exchange current density resulting in a slower discharge kinetics of hydrogen evolution. Also, there is a large increase in the value of recombination rate constant, k3 , for Zn–Ni–Cd alloy when compared to Cd and plain steel. Thus, Zn–Ni–Cd possesses excellent sacrificial protection properties and also inhibits the hydrogen permeation through the substrate.

4. Conclusion Corrosion resistant Zn–Ni–Cd ternary alloy with higher nickel content than conventional Zn–Ni bath was synthesized using an alkaline deposition process. Electrochemical techniques were used to characterize the hydrogen permeation properties of Zn–Ni–Cd coated steel and to compare it with those of Zn–Ni and Cd coated steel in aqueous solutions. The corrosion rate of Zn–Ni–Cd coated steel estimated in a 0:5 M Na2 SO4 þ 0:5 M H3 BO3 solution at a pH of 7 was 0.073 mm y
1 . This value is much lower than the corrosion rate of Zn–Ni alloy (0.502 mm y
1 ) and Cd (0.306 mm y
1 ) coatings deposited from commercial baths. Zn–Ni–Cd alloy offers maximum hydrogen permeation resistance. The hydrogen permeation experimental results were analyzed using I–P–Z model. The results of this study can be summarized as follows: The polarization and permeation studies indicated that the mechanism for HER on Zn–Ni and Zn–Ni–Cd (low-Cd composition) deposits on steel is coupled discharge–recombination with Langmuir isotherm for hydrogen coverage, while in the case of Cd deposited steel, HER mechanism followed the Tempkin isotherm.

1520

H. Kim et al. / Corrosion Science 45 (2003) 1505–1521

Hydrogen permeation current was found to be nearly zero under normal corroding conditions for Zn–Ni–Cd coated steel and it increased to 0.23 lA/cm2 at a cathodic overpotential of 310 mV. The hydrogen permeation current density for Zn– Ni and Cd coated steel under similar conditions were 5 and 2 lA/cm2 , respectively. It was observed that the HER on the surface was suppressed with increasing the Cd ratio in deposits. The presence of Cd in Zn–Ni alloy reduces the hydrogen evolution and also increases the recombination constant. This ensures most of the hydrogen that is adsorbed on the surface to recombine chemically, thus reducing in large extent the hydrogen ingress into the alloy contributing to smaller permeation currents. Corrosion studies were done on the deposits and the results indicated that Zn–Ni–Cd possesses superior corrosion resistance in the solution of 0:5 M Na2 SO4 þ 0:5 M H3 BO3 at a pH of 6.5 when compared to Zn–Ni and Cd deposits obtained from commercial baths.

Acknowledgements Financial Support by Dr. Vinod Agarvala, the Office of Naval Research under Grant no. N00014-00-1-0053 and AESF Research Contract, Project 107 are gratefully acknowledged. This work was also partially supported by Sandia National Laboratories. Sandia is a multi-program laboratory operated by Sandia Corporation, a Lockheed Martin Company, for the United States Department of Energy under contract DE-AC04-94-AL85000.

References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] [16] [17] [18] [19] [20]

S. Jin, A. Atrens, Appl. Phys. A 42 (1987) 149. B.E. Wilde, T. Shimada, Scripta Metall. 22 (1988) 552. P.J. Grobner, D.L. Sponseller, D.E. Diesburg, Corrosion 35 (1979) 240. M. Manohar, Ph.D. Thesis, The Ohio State University, Columbus, OH, 1990. B.E. Wilde, L. Chattoraj, Scripta Metall. 26 (1992) 627. K.R. Baldwin, C.J.E. Smith, Trans. Inst. Met. Finish. 74 (1996) 202. J.M. Chen, J.K. Wu, Plating Surf. Finish. 33 (1992) 657. K.R. Baldwin, C.J.E. Smith, Trans. Inst. Met. Finish. 74 (1996) 202. M. Zamanzadeh, A. Allam, C. Kato, B. Ateya, H.W. Pickering, J. Electrochem. Soc. 129 (1982) 284. D. McLandolt, Electrochim. Acta 39 (1994) 1075. J.W. Dini, H.R. Johnson, Met. Finish. 77 (1979) 31. B.N. Popov, G. Zheng, R.E. White, Corrosion 50 (1994) 613. B.N. Popov, G. Zheng, R.E. White, Corrosion 51 (1995) 429. G. Zheng, B.N. Popov, R.E. White, J. Electrochem. Soc. 140 (1993) 3153. G. Zheng, B.N. Popov, R.E. White, J. Electrochem. Soc. 141 (1994) 1220. G. Zheng, B.N. Popov, R.E. White, J. Electrochem. Soc. 141 (1994) 1526. B.N. Popov, G. Zheng, R.E. White, Corros. Sci. 36 (1994) 2139. A. Durairajan, Ph.D. Thesis, University of South Carolina, 2001. H. Kim, B.N. Popov, K.S. Chen, J. Electrochem. Soc. 150 (2003) in press. R.N. Iyer, H.W. Pickering, M. Zamanzadeh, J. Electrochem. Soc. 136 (1990) 2463.

H. Kim et al. / Corrosion Science 45 (2003) 1505–1521 [21] [22] [23] [24] [25] [26]

1521

D.H. Coleman, G. Zheng, B.N. Popov, R.E. White, J. Electrochem. Soc. 143 (1996) 1871. M.A.V. Devanathan, Z. Stachurski, J. Electrochem. Soc. 110 (1963) 886. J. McBreen, L. Nanis, W. Beck, J. Electrochem. Soc. 113 (1966) 1218. B.S. Chaudhari, T.P. Radhakrishnan, Surf. Technol. 22 (1984) 353. M. Ramasubramanian, B.N. Popov, R.E. White, J. Electrochem. Soc. 145 (1998) 1907. E. Gileadi, B.E. Conway, in: J.O.M. Bockris, B.E. Conway (Eds.), Modern Aspects of Electrochemistry, vol. 3, Butterworths, Washington, DC, 1964, pp. 347–442. [27] R.N. Iyer, I. Takeuchi, M. Zamanzadeh, H.W. Pickering, Corrosion 46 (1990) 460. [28] A. Durairajan, B.S. Haran, R.E. White, B.N. Popov, J. Electrochem. Soc. 147 (2000) 1781.