Corrosion behaviour of the extruded and unidirectionally solidified Al-Al4Ca eutectic alloy in a neutral sodium chloride solution

Corrosion behaviour of the extruded and unidirectionally solidified Al-Al4Ca eutectic alloy in a neutral sodium chloride solution

Corrosion Science, Vol. 19, pp. 553 to 571 Pergamon Press Ltd. 1979. Printed in Great Britain. CORROSION BEHAVIOUR OF THE EXTRUDED AND UNIDIRECTIONAL...

4MB Sizes 0 Downloads 15 Views

Corrosion Science, Vol. 19, pp. 553 to 571 Pergamon Press Ltd. 1979. Printed in Great Britain.

CORROSION BEHAVIOUR OF THE EXTRUDED AND UNIDIRECTIONALLY S O L I D I F I E D A1-A14Ca E U T E C T I C ALLOY IN A NEUTRAL SODIUM CHLORIDE SOLUTION* L. PERALDO BICELLI, C. ROMAGNANI a n d D. SINIGAGLIA Istituto di Chimica Fisica, Elettrochimica e Metallurgia del Politecnico di Milano, Centro di studio sui processi elettrodici del C.N.R., Piazza Leonardo da Vinci, 32-20133 Milano, Italy Abstract--The corrosion behaviour of extruded and unidirectionally solidified AI-AI~Ca eutectic alloy in a neutral sodium chloride solution has been compared with the behaviour of commercially pure AI and AI-Cu alloy. From both anodic and cathodic cyclic linear potentiodynamic measurements and from S.E.M. observations it was possible to point out that the corrosion behaviour of unidirectionally solidified alloy is substantially equal to that of the extruded alloy, slightly different from that of commercially pure AI and better than AI-Cu alloy. Therefore, the use of this new lamellaereinforced A1-AI4Ca composite is promising, even though from the corrosion point of view further studies involving different techniques and conditions (solutions, temperature, surface state, stress, etc.) should be undertaken. INTRODUCTION

As PART OF the widely debated problem of saving geochemically scarce metals, the use of calcium in alloys looks promising, especially because this element is relatively abundant in the earth's crustY Therefore, the fairly recent technique of unidirectional solidification has been employed to prepare the A1-AI4Ca eutectic. This consists of parallel lamellae (ca. 30 ~o of the total volume) of a relatively hard material (A14Ca, an intermetallic compound) incorporated in a ductile matrix of practically pure aluminium.~ The A1-A14Ca eutectic can be rolled at temperatures between 400-500°C and its mechanical properties are typical of fibre composites, i.e. yield strength almost equal to tensile strength, low percentage elongation, a In particular, two of the most notable properties are the low specific gravity and superplastic behaviour, a In order to complete the understanding of the properties of this new material, it was necessary to determine the corrosion resistance--especially to localized corrosion - - o f the extruded and unidirectionally solidified A1-AI4Ca eutectic. A systematic study of the anodic and cathodic behaviour of this material extending the study to the commercially pure aluminium and age-hardened A1-Cu alloy for the purpose of comparison and completeness was undertaken. There is a large amount of data in the scientific literature on the corrosion behaviour of these latter materials. 4-6 The conditions for the reliable use of these materials in various fields were defined some years ago. 7 The same is not true for the unidirectionally solidified AI-Ca alloy and the fibre composites in general s Since aluminium and aluminium alloys are usually passive in the most common aggressive environments and are subject to pitting or crevice corrosion in the presence of an aggressive anion (such as the chloride ion), it was decided to use a neutral *Manuscript reeeivexl 15 March 1978; in amertd~d form 19 May 1978. $53

554

L. PERALDO BICELLI, C. ROMAGNANIand D. SINIGAGLIA

chloride solution (3.5 ~o NaCI) which is often used to simulate seawater corrosion. Furthermore, since a comparison between the behaviour of different active-passive materials was wanted, it was decided to use the cyclic linear polarization method, which has already been defined by Wilde a n d Williams for the stainless F e - C r - N i alloys 9 and applied to some a l u m i n i u m alloys, for example by Nilsen a n d Barda110 and Broli and Holtan. 11 T h u s there is a reasonably fast a n d simple method for measuring both the pitting a n d protection potentials, and the area of the anodic hysteresis loop, i.e. a method for comparing pitting corrosion resistance of the various materials. F u r t h e r study will then be made of their corrosion behaviour in chemical tests u n d e r resting conditions (these tests may last some time), following the suggestions a n d the aim of Fouroulis32 E X P E R I M E N T A L METHOD Materials

The following metallic materials were studied: AI--Ca alloy (Ca = 7.4~) in extruded and unidirectionally solidified rods (diameter 10 mm); this material was compared with commercially pure aluminium (> 99.9~: the main impurities were: Cu: n.d.,* Fe: 0.06, Mg: n.d., Mn: n.d., Zn: n.d., Ti" n.d.), in extruded rods (diameter 10 ram) and the AI-Cu alloy (Cu: 5.4, Fe: 0.23, Si: 0.13, Mg: n.d., Mn: n.d., Zn: 0.02, Ti: 0.06), in rods (diameter 10 mm) hardened in water (from ca. 530°C) and aged at room temperature (natural ageing). The AI-Ca alloyt was made in two steps. First of all, the master alloy, containing "30~Ca, was made from Raffinal aluminiumand calcium (> 99.99 ~ pure) by melting the two metals in a graphite crucible in an induction furnace with an argon atmopshere. The master alloy was analysed to determine its calcium content, then melted again. At this point, Raffinal aluminium was added to bring the percentage of calcium to 8~o (this is a little higher than the percentage of calcium in the eutectic: 7.6 .~/o). Since calcium has a high vapour pressure, the final percentage of calcium in the alloy was 7.4~.~ The alloy was extruded at 450°C into rods (diameter 10 mm) (Fig. 1 (a)). Lengths of extruded rods were unidirectionally solidified in a controlled temperature gradient furnace. The specimens could be moved vertically through the furnace and their velocity could be adjusted, a The alloy that was formed had a typically lamellar structure (Fig. l(b)). X-ray analysis showed that these lamellae consisted of AI4Ca intermetallic compound. Specimen surface preparation

The cylindrical specimens (diameter ca. 8.5 mm after turning) were sealed in Perspex or in cold cured epoxide resin (of the Presi Mecaprex-Km type) so that the surface perpendicular to the hot forming or unidirectional solidification was exposed to the aggressive environment. The surfaces were prepared as follows: (a) wet abrasion with emery paper (grain size decreasing down to 15 ~tm); (b) polishing with cloths and diamond paste (grain size decreasing down to 1 ~tm). Each time the cloth was changed, the specimens were rinsed in distilled water and cleaned ultrasonically. After the last polishing operation, they were rinsed in methanol and then in distilled water and cleaned ultrasonically; (c) drying for 2 h at 60°C in a water pump vacuum. This treatment is especially necessary for the AI-Ca alloy because Evans and Braddick B showed that all traces of liquid must be removed from the crevices at the matdx-intermetallie compound interface. (d) oxidation in dry air or anodic oxidation (in 15~H2SO4; time = 0.2 h; current density = 1.6 A dm-~). According to Hunter and Fowle, a3 the first type of oxidation gives a barrier-type film ca. 2 nm thick; the second gives a duplex film (barrier + porous type) ca. 5 v.m thick, a4 The pores of the duplex film were sealed by specimen immersion in boiling distilled water (pH = 6-7) for at least 600 s.1~ *n.d. = none detected. 1"Thealloy was made by the Materials Science Division of the C.C.R. Euratom, at the Ispra Joint Center. ~:It was not necessary to bring the percentage of calcium to the exact eutectie value because the mechanical properties of the alloy are substantially the same as those of the eutectic.

Corrosion behaviour of the AI-AI,Ca eutectic alloy

555

Test solution

A 3.5 % sodium chloride solution (0.5M NaCl), simulating seawater aggressivity was used. The pH was adjusted to 7 by adding 0.1N NaOH solution. Reagent grade chemical and distilled water were employed. The solution was stirred by bubbling a nitrogen* or chromatographic air flow ( c a . 5 1h-I) through the solution. The solution and the cell were de-aerated for ca. 6 h. The ratio between the solution volume and the exposed surface area was ca. 950-1000 rnl cm-2. A thermostat was used to maintain a temperature of 20°C _-4_-2°C. Cell and electrochemical

measurements

A glass polarization cell with two separate compartments was used. The cell description is reported elsewhere.16 For the electrode assembly a cylindrical Perslmx holder was used in which a Perspex cylinder was screwed, so that the specimen was pressed against a Vyton gasket. This gasket precisely defined the exposed surface and reduced the likelihood of localized attacks along the three-phase boundary (metal-gasket-solution). 16 The distance between the electrode surface and the Haber-Luggin capillary was set at about the optimum value of I mm to minimize experimental error (especially ohmic drop and screen effects). The cyclic linear potentiodynamic method was employed. The specimen was left in the solution for ca. 0.25 h and then polarized at the rate of 30 mV min-L During the anodic polarization, the polarization direction was inverted at a potential ca. 100-150 mV greater than the pitting potential. Then, when the potential had decreased to a value corresponding to zero anodic current density (c.d.), the polarization direction was inverted again and the potential was raised beyond the pitting potential (repeated cyclic linear anodic polarization). In the cathodic polarization the polarization direction was inverted at a value of -- 2000 mV(SCE) (this value was used during all the experiments) and the potential was increased until the cathodic c.d. reached zero. An AMEL Mod. 551/SU potentiostat was used, together with a saturated calomel electrode (SCE) ( + 242 mV vs NHE). The potential values were not corrected for ohmic drop.'[" A Siemens Mod. Autoscan scanning electron microscope with a 25 kV filament was used to characterize the localized attack morphology. The specimens were not covered with evaporated metal prior to examination because it was noted that the different tendencies of the various areas to emit secondary electrons and to accumulate charge created sufficient contrast, t7 Tilting was applied arbitrarily, so the tilt angles are not quoted.

E X P E R I M E N T A L R E S U L T AND D I S C U S S I O N Corrosion

potential

The corrosion potential (scanned at the beginning of the measurements for 15 min) for the different materials is only slightly reproducible ( + 100 mV for Al; :k 200 mV for A1-Ca alloy); the A1-Cu alloy is an exception to this rule ( ± 20 mV). However, the a l u m i n i u m corrosion potential (average value c a . - - 850 mV(SCE)) is less noble t h a n A1-Cu alloy potential (average value c a . - - 680 mV(SCE)). This decrease of nobility is especially marked for the extruded a n d the unidirectionally solidified A I - C a alloys (average value c a . - - 1250 mV(SCE)). The calcium present in the intermetallic c o m p o u n d AI4CA xs can play a n i m p o r t a n t role in m a k i n g the corrosion potential less noble. M u c h of the data i n the available literature 1~ indicates that the a l u m i n i u m corrosion rate in chloride or non-chloride solutions is i n d e p e n d e n t of the potential when the potential is between - - 800 and - - 1600 mV(SCE), since in such conditions a l u m i n i u m is in the passive range. F o r the A I - C u alloy, the corrosion potential differs only slightly from the pitting potential. The corrosion potential for the four materials does not appear to be influenced by anodic oxidation or the presence of oxygen. *Nitrogen (supplied by SIAD) contained 2 ppm of oxygen and a few ppm of CO + CO2. tThe error caused by the inter-liquidjunction is negligible because an agar-agar bridge saturated with potassium chloride was used.

556

L. PERALDO BICELLI, C. ROMAGNANIand D. SINIGAGLIA

In the case of aluminium, the results do not agree with those obtained by Di Bar and Read. 19 They found that the corrosion potential was slightly more noble (ca. 70 mV) in the presence of oxygen and generally more reproducible ( + 30 mV in an aerated environment; -4- 6 mV in a de-aerated environment). This disagreement may only be apparent because their data ~refer to a solution with a different pH value (pH = 4) and temperature (30°C) and a different surface preparation of the specimens (electrochemical polishing).

Anodic measurements Aluminium. Figure 2 shows the potentiodynamic anodic curves of non-anodized and anodized aluminium; these data refer to de-aerated and aerated environments respectively. The shape of the curves, with the characteristic hysteresis loop, is typical of an initially passive material subject to pitting corrosion. The pitting potential is between -- 650 and -- 750 mV(SCE). In the aerated environment the pitting potential of the anodized specimens was only poorly reproducible. This is due to the characteristics of the system (oxidized metal/aggressive medium) and is not caused by localized attack along the three-phase boundary (metalgasket-solution).

-500

I

I

I

i

I

I

I

I

I

I

I

I

[

I

I

I

-600

-700 -800 - 900

-IO00 -I100 UJ -1200 0 '..n -1500 :~

"500

-(c)

(d)

-600 -700

-800 -900

- iooo -1100

02

-1200

-1300

I [0 -4

I0"3

I JO"2

r

t

I0 "i

I

I 10.3

i

i

I

ZO-~

I0"i

!

i, mA cm'2

FIG. 2. Anodic polarization curves of non-anodized (a, b) and anodized (c, d) commercially pure aluminium in neutral chloride solution (0.5M NaCI, p H = 7,

20°C).

FIG. 1. Microstructure of extruded (a) and unidirectionally solidified (b) A1-AI~Ca eutectic alloy (transverse section; the AI~Ca lamellae are the dark ones).

4~

FIG. 3.

<

,-"

#-,-4,aPg,~,,t

S.E.M. micrographs showing a typical pit on the anodized commercially pure aluminium.

~e

t'

,~,~!.

Fr6. 5.

S.E.M. micrographs showing a typical pit on the anodized AI-Cu alloy.

(a)

(b)

FIG. 8. S.E.M. micrographs showing a typical localized attack on anodized: (a) extruded alloy; (b) unidirectionally solidified alloy (the AItCa lamellae are preferentially attacked).

Ft6. 14. S.E.M. micrographs showing a typical cathodic structural attack on the non-anodized extruded (a) and unidirectionally solidified (b) AI-AI.~Ca eutectic alloy (the AI~Ca lamellae are preferentially attacked).

Corrosion behaviour of the AI-AI4Ca eutoctic alloy

563

Practically no influence due to the anodizing treatment and air presence was observed for the pitting potential and the hysteresis loop. This independence was observed by Diggle et al. 2° and Downie et a l ) z respectively. The return curves of the hysteresis loops are characterized by a rapid drop in the anodic c.d. (corresponding to the potential range where the localized attack occurs during the ascending curve). Then the c.d. decreases slowly until it reaches the final corrosion potential, which is much less noble that the initial value. When the electrode is repolarized in the anodic direction, the pitting potential is not subject to any significant change and it maintains the values originally measured (this was also observed by Di Bari and Read). z9 After each anodic measurement, the aluminium showed localized crystallographic attack with evident cubic planes (Fig. 3). This attack morphology is analogous to that found by Richardson and Wood for anodized and non-anodized aluminium exposed to a potassium chloride solution, t7 A1-Cu alloy. Figure 4 shows the anodic potentiodynamic curves for the AI-Cu alloy. These curves are reproducible and have a marked hysteresis (as was the case with aluminium). The final corrosion potential is less noble than the initial value. The pitting potential lies between -- 600 and -- 650 mV(SCE) which is slightly more noble than the value for aluminium. This alloy had practically the same behaviour in all four experimental conditions. -sool - ( a ) -600'

I

I

I

I

-I

I

I

I

I

I

I

(b)

-700

-800 -900 -

I000

-II00

_

Oz

N2

c~ -1200 03

-1300 >

E

-

500

tJ

-600

I

I

I

I

-(c)

l

I

I

(.d)

-700 -800

-900 -

I000

- I IO0 -

Nz

02

-1200 -1300

I

l

10"3

10"z

l

I

I

I

I

I

l

10-j

I

10"3

10"2

10"t

[

I0

i, mA cm-2

FIG. 4.

Anodic polarization curves of non-anodized (a, b) and anodized (c, d) AI-Cu alloy in neutral chloride solution (0.SM NaCI, pH = 7; 20°C).

_

L. PERALDOBICELLI,C. ROMAONANIand D. SINIOAOLIA

564

Metallographic examination showed local corrosion (Fig. 5)* which consists of a very large number of pits; these pits were already present in the sample whichwas simply immersed in the solution. The local attack was intensified by slight anodic polarization. The pits form and grow at iron-bearing constituents, e.g. FeAIa and ~-AIFeSi particles 22and perhaps also around the intermetallic compound CuAI2 with a mechanism analogous to the one suggested by Sugimoto and Sawada3 a

Extruded and unidirectionally solidified eutectic alloy. Figures 6 and 7 show typical curves which represent the average anodic behaviour of the extruded and unidirectionally solidified alloy respectively. Both alloys cover a wide range of potentials before the film becomes pitted. Analysis of their hysteresis loops leads to the following conclusions: (i) The potential measured at zero c.d. at the end of the first hysteresis loop is more noble than the initial corrosion potential. This result is particularly evident in Fig. 6(b). (ii) The pitting potential of the anodized materials becomes less noble during subsequent polarizations. It becomes equal to the value for the non-anodized material which on the contrary is not affected by the polarization cycles. (iii) The descending region of the curve shows that the pits tend to' de:activate rapidly; this is followed by a slow de-activation. The pits are reactivated at the same potential during the successive polarization cycles. ]

-700

l

[

I

I

-800

I

I

I

I I0"

I I

(b)

(o) -

-900 -I000 - I [00 -1200 02

-1300

-1400 > E

tJ

-1500

I

I

[

t

O]c) -200

-

-400

-

"(d) -~

- oo -800 -I000

-1200 -1400 -1600 I0 -4

i IO.3

I I0 .2

[ I 0 -I

I I0 -3

1

I i,

l I0-2

m A crn 2

Fzo. 6. Anodic polarization curves of non-anodized (a, b) and anodized (c, d) extruded A]-A]4Ca eutcctic ahoy in neutral chloride solution (0.5M NaC], pH = 7; 20°C).

*In some areas a slight intergranular corrosion was observed,

.

Corrosion behaviour of the AI-AI4Ca eutectic alloy (a)

I

~

[

i

(b)

l

565

t

{

t

I

[

I

- 600 -700

/

-800 -900 -I000 - I IOC) -1200 u~

> E

-1300

~w

-1400

(C)

w I

I

I

I

I

I

J

(d)

I

0 -200 -400

N2

-60o -800 -I000 - 1200 -1400

I

10-4

I0 -:~

l

I0 -z

I0"

I

I~ 4

I

!

I

[

Iff a

I~ a

I 0 "t

[

mA cm 2

Fzo. 7. Anodic polarization curves of non-anodized (a, b) and anodized (c, d) unidirectionally solidified AI-AI4Ca eutectic alloy in neutral chloride solution (0.SM NaCI, pH = 7; 20°C).

(iv) The hysteresis loop area is practically the same for the extruded and the unidirectionally solidified eutectic alloys. This would indicate that the pitting on the Al4Ca lamellae or on the lamellae-matrix interface does not produce at the microstructural level the conditions characterizing crevice corrosion development. The pitting potential of the non-anodized alloys is always less noble than the pitting potential of aluminium and the AI-Cu alloy and it improves after anodization. Indeed, the pitting potential practically falls in the range -- 650 to -- 750 mV (SCE)* (the same was observed for anodized and non-anodized aluminium) and it is slightly less noble than that of the A1-Cu alloy. Therefore the anodizing treatment improves the localized corrosion resistance of these two alloys and it becomes comparable to the aluminium corrosion resistance. Furthermore, no difference in the behaviour of these two alloys was observed. If the limited reproducibility of the measurements is taken into account, it appears that the presence of oxygen in the aggressive medium has no influence on anodic behaviour and on the pitting potential in particular. Non-crystallographic (in some cases spherical in shape) attack was observed in the extruded alloy (Fig. 8(a)) and geometrical and preferential attack of the lamellae in the unidirectionally solidified alloy (Fig. 8(b)). *In some cases the film pitted even at -- 200 mV (SCE).

566

L. PERALDO BICELLI, C. ROMAGNANI a n d D . SINIGAGLIA

Cathodic measurements A l u m i n i u m . Figure 9 shows some typical curves which represent the average

cathodic behaviour of aluminium. The shape of the curves is about the same in all four experimented conditions. The first region is almost horizontal, i.e. characterized by a c.d. increasing rapidly with small variations in the applied potential. This is followed by an almost vertical region where the c.d. is practically independent of the potential, as also reported by Diggle et a l ? o for acid (pH = 1) chloride (CI- = IN) solutions. This is followed by a linear Tafel range with the slope varying between 150 and 220 mV/dec. Sometimes the curve becomes non-linear for high c.d. owing to ohmic drop and concentration polarization. For the curves obtained in an aerated environment, the vertical region cannot be attributed to oxygen reduction under diffusion conditions. This vertical region is also observed in the absence of oxygen, as already found by Di Bari and Read. 19 Moreover the c.d. magnitude is much lower than the value obtained during cathodic measurements on a platinum electrode. For platinum the value of the diffusion limiting c.d. is about 10 times greater than the value of the vertical region in Fig. 9. Consequently in the case of aluminium, this vertical region is not due to a diffusion limited oxygen reduction, in agreement with the observation of Pryor and Keir 24 and of Kaesche? '~ According to these authors, oxygen reduction on filmed aluminium is a l m o s t completely inhibited due to the very low electronic conductivity of the film. Therefore, unlike other metals (Cu, Fe, etc.), it is impossible to reach conditions where oxygen

- 600

~~~-(a) I

- 8OO

I

I

--(b)

I

I

I

I

I

-IOOO -1200 -1400 -1600 -1800 -2000

-

-2200

I I I I

i I i [

-;oo t J - Iooo --12oo -1400 -1600 - i800 - 2000 -220(3

~ 10 -4

I I 0 -s

I

I

I

I0z

I0 "r

I

10.4 i,

rnA

I

I

I

1

10"~

I 0 -a

10 "1

I

c r n- 2

Fio. 9. Cathodic polarization curves of non-anodized (a, b) and anodized (c, d) commercially pure aluminium in neutral chloride solution (0.5M NaCI, pH = 7; 20°C).

-

Corrosion behaviour of the AI-AI4Ca eutectic alloy

567

reduction is controlled by diffusion only, unless high cathodic potentials are applied; but in this last case, the cathodic c.d. is mainly due to the hydrogen evolution reaction. In spite of the irreproducibility of the curves (especially the portion of the cathodic curve preceding the linear Tafel region), the following interpretation is suggested. The schematic partial curves and the resulting curve, analogous with the Di Bari and Read paper, x9 are shown in Fig. 10. As stated, aluminium was initially covered by a stable oxide film. Applying an increasing cathodic polarization creates conditions which favour gradual film modification from stable to less stable. When the potential values fall below the vertical region of the cathodic curve the film is practically dissolved. Since the aluminium surface state is not very reproducible, the initial passivating film may be more or less stable. This is reflected by the size of the anodic loop and therefore by the shape of the region of the resulting cathodic curve which precedes the Tafel line. When the first part of the cathodic curve has a low (high) resulting c.d. (the type shown in Fig. 9a and c respectively) this corresponds to a high (low) anodic partial c.d. This indicates that the aluminium is initially covered with a less (more) stable oxide.* The shape of the return curve, where the vertical region disappears, confirms that a change in the surface state occurred during the initial cathodic polarization, making the aluminium active. In fact, the final value of the corrosion potential is much less noble than the initial value and it stays that way for a certain time (for c a . 1 h at least). It may be concluded that aluminium remains passive during cathodic polarization, even though hydrogen is evolved, as long as the value of the applied potential

I~--"'~-

lWffhout c r

Ep= . . . . . . . . . EHz"H' ~ , .

~' /~.

"*External onodic curve

=Y.i

Eco=

-

-

-:

With CI" _

_

;~L~--'~.x / C a t h o d i c self polorizotion curve ~"L'\ (2H'+ Ze-H 2)

]

i....

cathodic

-

~.

Anoch'cself'" . . . . . . . polarization curve~

~

~'\

"-.~,,,

-

Log i

FIG. 10. Schematic self-polarization and external anodic and cathodic curves of the commercially pure aluminium in neutral chloride solution (0.5M NaCI, pH = 7; 20°C). *This interpretation appears even more valid if the fact that in general an increase in film stability implies a general de-activation ('lower exchange c.d.) of the hydrogen evolution process ~6 is considered.

568

L . PERALDO B I C E L L I , C . R O M A G N A N I a n d D . S I N I G A G L I A

does not fall below the primary passivation potential and/or an increase in the pH of the solution does not lead to chemical film dissolution. At the end of the cathodic polarization cycles, the aluminium shows general attack. Slight localized attack was also observed (this was also observed by Di Bari and Read19). Anodizing tends to shift the vertical region of the cathodic curves towards higher c.d. values, in agreement with the greater stability of the artificial compared with the natural oxide. The oxygen reduction overvoltage is so high that in practice it does not contribute to the resulting cathodic process. Furthermore, the presence of oxygen in the solution does not appear to modify film stability. AI-Cu alloy. The curves for this alloy (Fig. 11) are reproducible and have the same shape as the ones for aluminium. In one case, the vertical region (Fig. l l(b)) was clearly shaped like a reversed "S", in accordance with the prediction based on the resulting cathodic curve in Fig. 10. During the reverse polarization, the Tafel line shows a marked hysteresis correlated with the formation of a fine black powder. The presence of this powder (presumably copper or the intermetallic compound CuAI2z3) lowers the hydrogen overvoltage values since the hydrogen Ùvervoltage for copper and the intermetallic compound 18 is much lower than for aluminium. Furthermore, this powder can develop electrocatalytic activity. In this case, tbo, the final corrosion potential is less noble than the initial value; this might indicate that the material changed from the passive to the active state. When compared with aluminium, the A1-Cu alloy has a shorter vertical region and a much lower hydrogen overvoltage (compare Figs. 9 and I 1); the slope is between 110 and 180 mV/dec. Furthermore, this alloy shows much more general corrosion. Anodizing tends to improve the surface oxide stability, whereas the effect of oxygen present in the aggressive medium is negligible. -6OO -800 -1000 -1200 -1400 -1600 -Ie00 ~o

-2000 -2200

I

~

I

I

10"z

10"~

I

10

I

I

I

J

1

I

l0z m A cme

i

[

1

10-z

10"~

I

I 10

[ l0 t

-800 UJ -1000

-

-

-1200 -1400

-16(30

-,eo.o -2000 -2200 10-.~

i,

FIo. 11. Cathodic polarization curves of non-anodized (a, b) and anodized (c, d) AI--Cu alloy in neutral chloride solution (0.5M NaCI, pH = 7; 20°C).

Corrosion behaviour of the AI-A14Ca eutectic alloy

569

Extruded and unidirectionally solidified eutectic alloy. In comparison with the curves for aluminium, the vertical region is generally missing. The example in Fig. 12 shows the most frequently observed curve-type. This behaviour is not influenced by anodizing nor by the presence of oxygen in the aggressive medium. -

I000

o)

-1200 -1400

F

N2

-1800 - 2000 b.J f,,.,)

cO - 2 2 0 0

-

>

(b)

~:

I

i

I

I

-1200 -

-1400 -

N2

-

-1800 -2000 - 2200 10-4

I

f

lO"~l

i, FIG. 12.

I

IG z FnA

I 0- I

I I

crri z

C a t h o d i c p o l a r i z a t i o n curves of" n o n - a n o d i z e d e x t r u d e d (a) a n d u n i d i r e c t i o n -

ally solidified (b) AI-AI4Ca eutectic alloy in neutral chloride solution (0.5M NaCI, pH = 7; 20°C).

The lack of the vertical region in the resulting cathodic curves leads to the conclusion that for these two alloys, if there is an-oxide film, it is very unstable. In some cases, when the initial corrosion potential of these alloys falls in the --800 to 1000 mV (SCE) range (i.e. when it nears the initial corrosion potential values which are most frequently observed for aluminium and the AI-Cu alloy) the cathodic curves do have a vertical region and the final corrosion potential is much less noble than the initial corrosion potential (see Fig. 13, for example). The behaviour of the alloys in the Tafel line is sufficiently reproducible (the slope is ca. 130-170 mV/dec), the corresponding hydrogen overvoltage is high and closer to the hydrogen overvoltage for aluminium than for the A1-Cu alloy. After the cathodic measurements, the alloys show strong attack; this attack is generalized and uniform and it emphasizes the structure (Fig. 14). -

CONCLUSIONS

The following conclusions may be drawn from the cyclic linear potentiodynamic measurements (both anodic and cathodic) made on the AI-AI4Ca extruded and unidirectionally solidified eutectic alloy in a neutral chloride solution, compared with the measurements made on aluminium and the AI-Cu alloy.

L. PERALDOBICELLI) C. ROMAGNANIand D. SINIGAGLIA

570

-800 - 1000 w

02

-1200 -1400

> E

-1600

"J

- 1800

"

2000

-2200 II] 3

i I() z i,

I I0-' m A crn 2

p t

I0

FIo. 13.

Cathodic polarization curve of non-anodized unidirectionally solidified AI-AI~Ca eutectic alloy in neutral chloride solution (0.5M NaCI, pH = 7; 20°C).

(a) The corrosion potential of the extruded and unidirectionally solidified eutectic alloy is much less noble than for aluminium. The latter has a slightly less noble corrosion potential than the A1-Cu alloy. Both the corrosion potential and the corrosion rate (weight losses o f ca. mg cm -~in a 32 h immersion test) are similar to the values for metals which are covered with a more or less stable film of oxide in resting conditions. (b) The pitting potential for the non-anodized eutectic alloys is less noble than for commercially pure aluminium and the AI-Cu alloy. When these four materials are anodized, the pitting potential is practically the same for all. Oxygen does not influence the results. (c) The hysteresis loop (on the anodic side) shows a rapid and then slower de-activation of the pits formed when the potential is greater than the pitting potential. In particular, repeating the anodic cycle brings the pitting potential of the anodized eutectic alloys back to the values for the non-anodized alloys. This cancels the positive effects of anodic oxidation (see item b). (d) For aluminium and the A1-Cu alloy, the cathodic behaviour is characterized by a vertical region which cannot be attributed to the cathodic oxygen reduction process (even in an aerated environment) under diffusion c.d. conditions. It is caused by the presence of an oxide film which is dissolving. This vertical region is generally not found in the eutectic alloys. It is not affected by oxygen, but by specimen oxidation, which tends to shift this region towards higher cathodic c.d. All four materials showed some generalized attack after cathodic polarization. (e) The behaviour of the eutectic alloys in the Tafel region is more similar to the behaviour of aluminium than to the behaviour of the A1-Cu alloy. The latter has a smaller hysteresis loop (on the cathodic side) and a lower hydrogen evolution overvoltage. (f) The electrochemical (and therefore the corrosion) behaviour of the unidirectionaUy solidified A1-AI4Ca eutectic alloy is substantially equal to that of the extruded alloy. The behaviour of both anodized alloys appears to be only slightly different from that of aluminium and better than that of A1-Cu alloy.

Corrosion behaviour of the AI-AIaCa eutectic alloy

571

T h e use o f this n e w c o m p o s i t e m a t e r i a l l o o k s p r o m i s i n g e v e n t h o u g h f r o m the c o r r o s i o n p o i n t o f v i e w f u r t h e r studies i n v o l v i n g different t e c h n i q u e s a n d c o n d i t i o n s (solutions, t e m p e r a t u r e , surface state, stress, etc.) s h o u l d be u n d e r t a k e n . REFERENCES 1. G. PELLEGRINIand G. PIATTI, Eurospectra 13, 112 (1974). 2. K. N. STREET, G. F. St. JOHN and G. PIA'VEI,J. Inst. Metals 96, 326 (1967). 3. R. MATERA,G. PIATTIand K. N. STREET,Aluminhtm 49, 335 (1973); G. PIATTI,G. PELLEGRINIand R. TRIPPODO, J. lVlater. Sci. 11, 186 (1976). 4. NACE-3, Localized Corrosion (Edited by R. W. STAEHLE, B. E. BROWN, J. KRUGER and A. AGRAWAL), pp. 515-622. Nace, Houston (1974). 5. E. MA'VrSON, Eurocorr '77, 6th European Congress on Metallic Corrosion, London, 19-23 September 1977. 6. I. L. MULLERand J. R. GALVELE, Corros. Sci. 17, 179 (1977). 7. H. P. GODDARD, The Corrosion o f Light Metals, pp. 45-69. John Wiley, New York (1965). 8. J. M. EVANSand D. M. BRADDICK,Corros. Sci. II, 611 (1971). 9. B. E. WILDE and E. WILLIAMS,J. electrochem. Soc. 117, 755 (1970). 10. N. NILSEN and E. BARDAL,Corro$. Sci. 17, 635 (1977). 1 I. A. BROLI and H. HOLTAN, Corros. Sci. 13, 237 (1973). 12. Z. A. FOUROULISand M. J. THUBRIKAR,Electrochim. Acta 21,225 (1976). 13. M. S. HUNTER and P. FOWLE, J. electrochem. Soc. 103, 482 0956). 14. S. TAJIMA, Anodic oxidation of aluminium, in Advances in Corrosion Science and Technology (Edited by M. G. FONTANAand R. W. STAEHLE),pp. 230-362. Plenum Press, New York (1970). 15. D. A. JONES, Corrosion 25, 187 (1969). 16. B. MAZZA, P. PEDEFERRI, D. SINIGAGLIA,A. CIGADA, L. LAZZARI, G. RE and D. WENGER, J. elect,'ochem. Soc. 123, 1157 (1976). 17. J. A. RICHARDSONand G. C. WOOD, Corros. Sci. 10, 313 0970). 18. A. I. GOLUBEVand M. N. RONZHIN, Protection of Metals 1, 169 (1965). 19. G. A. DI BARI and H. J. READ, Corrosion 27, 483 (1971). 20. J. W. DIGGLE, T. C. DOWNIE and C. W. GOULDING, Corros. Sci. 8, 907 (1968). 21. T. C. DOWNIE and C. W. GOULDING, Metallurgia 73, 93 (1966). 22. C. PANSERI,Manuale di lecnologia delle leghe leggere da lavorazione plastica, pp. 225-227. Hoepli, Milano (1957). 23. K. SUGIMO'rOand Y. SAWADA,Nippoll Kink. Gakk. 3, 148 0973). 24. M. J. PP,YOR and D. S. KEIR, J. electrochem. Soc. 102, 605 (1955). 25. H. KAESCHEin Localized Corrosion (Edited by R. W. STAEHLE,B. F. BROWN,J. KRUGERand A. AGRAWAL), NACE-3, pp. 516-525. Houston (1974). 26. A. K. Vl.m, Electrochemistry and Semiconductors, pp. 173-177. Marcel Dekker, New York (1973).