Oxidation of Metals, Vol. 40, Nos. 1/2, 1993

The Effect of AI on the Corrosion Behavior of N i - M o in a H 2 / H 2 0 / H 2 S Gas Mixture Y. R. He* and D. L. Douglass* Received January 15, 1993; revised March 10, 1993

The corrosion behavior of seven N i - M o - A I alloys was investigated over the temperature range of 600-950~ in a mixed-gas atmosphere of 1-12/11:0/ H2S. The parabolic law was followed at low temperatures, while linear kinetics were generally observed at higher temperatures. At a fixed Mo content, the transition from parabolic to linear kinetics shifted to higher temperature with increasing AI concentration. Double-layered scales generally formed on alloys having a low AI content, consisting of an outer layer of nickel sulfide and a complex inner scale. The thickness of the outer scale and the inner scale decreased as the A1 content increased. The outer scale became porous and discontinuous with increasing AI content and temperature. Al203 was detected in the scales of all alloys corroded at higher temperatures (> 800~ even though the amount of A1203 was very small in some cases. The decrease in corrosion rate with increasing AI content may be attributed to the formation of A1203, Alo 55Mo2S4, and Al2S3 in the inner scale. KEY WORDS: mixed gas; Ni-Mo-A1 alloys; A1203; MoS2; AIo.ssM02S4; Chevrel phase.

INTRODUCTION A research program aimed at improving the sulfidation behavior of Ni-, Co-, and Fe-base alloys has been underway in our laboratory for several years. It was found that Mo and Nb, as alloying additions, reduced the sulfidation rate considerably, with Mo being more effective than Nb. ~-6 However, the formation of base-metal sulfide could not be totally suppressed even at an *Materials Science and Engineering Department, University of California at Los Angeles, Los Angeles, California 90024-1595. 119 0030-770x/93/0800-0119507.00/09 1993PlenumPublishingCorporation

120

He and Douglass Table I. N i - M o - A 1 Alloy Composition wt.%

at.%

N i - 10Mo-7.5A1 Ni-10Mo-10A1 N i - 10Mo-15A1 Ni-20Mo-7.5A1 N i - 2 0 M o - 15A1 Ni-30Mo-3.8A1 Ni-30Mo-7.5AI

Ni-5.8Mo-15.6A1 Ni-5.7Mo-20.2A1 Ni-5.4Mo-28.7A1 Ni-12.1Mo-16.1A1 N i - 11.1Mo-29.7A1 N i - 19.8Mo-8.9A1 Ni-18.9Mo-16.8A1

alloying level of 40wt.%. The formation of base-metal sulfides has been attributed to the intercalation of base-metal ions into the octahedral interstitial sites between weakly bonded layers of MoS 2 and NbS2, causing rapid diffusion of these ions to the surface where they react to form their sulfides. The effect of Cr, Ti, Mn, V, and A1 additions on the sulfidation behavior of N i - M o , C o - M o , and F e - M o alloys has also been studied. 7-1~ A1 was found to be the most effective addition. The sulfidation of some Ni-A1 alloys has also been analyzed by some authors. The sulfidation of Ni-A1 alloys containing 3 - 1 3 a t . % A1 in H z S - H 2 atmospheres has been investigated by Bhide and Smeltzer. ~1 AIzS 2 and a liquid phase were found in the scales. Godlewska e t al. 12 also studied the sulfidation behavior of/3-NiA1 at sulfur pressures from 5 - 2 x 103 Pa. The scales were found to be multiphase and multilayered. AlzS 3 was also found in the inner scale, and Ni3S2 was detected on the outer surface. In practice, it is of importance to study the corrosion behavior of a metal and its alloys in environments that contain both oxygen and sulfur. Recently, the corrosion behavior of six binary alloys in a H 2 - H 2 0 - H 2 S gas mixture has been investigated by Douglass' group. 13-1s The corrosion rate decreased with increasing Mo content. In addition to MoS2, double sulfides also formed in the scales. Base-metal-sulfide formation was not suppressed even for alloys containing 40 wt. % Mo. Oxidation could serve as one approach to improve the corrosion resistance of an alloy in environments containing both oxygen and sulfur. This is because the oxidation rate of common metals is generally much lower than Table II. Calculated Partial Pressures of Po2 and Ps2 at Different Temperatures Temp. (~ Po2 Ps2

900 1.0 • 10-19 2.5 • 10-5

800

700

600

8.9 • 10-22 4.5 • 10-6

2.9 • 10-24 5.6 • 10-7

2.7 • 10-27 4.5 x 10-8

121

Corrosion Behavior of N i - M o

40/

~60

927c .\.

Ni

/

90

'c

\

"~- ~ .

80

70

60

50

40

30

20

10

No

Fig. 1. Ternary isotherm of the Ni-Mo-A1 system at 927~ (a) Ni-10Mo-7.5A1, (b) Ni-10Mo-10A1, (c) Ni-10Mo-15A1, (d) Ni-20Mo-7.5A1, (e) Ni-20Mo-15A1, (f) Ni-30Mo-3.8AI, and (g) Ni-30Mo-7.5A1.

their sulfidation rate due to the lower defect concentration in oxides.19 Therefore, if an adherent and continuous oxide film is developed and maintained, the protective oxide scale can act as a barrier between the substrate and the gaseous environment. Among the numerous components of scaling-resistant alloys, A1 plays an important role in atmospheres containing oxygen. Giggins and Pettit2~ studied the corrosion of Ni, Co, Fe, and alloys of these metals containing Cr and A1 in atmospheres containing multi-oxidant species. It was found that continuous layers of oxides formed on the alloys initially, that sulfidation occurred eventually, depending on the gas composition and alloy composition, and that oxide layers could affect the sulfidation behavior. The research conducted by Ohla e t a / . 21-22 pointed out that thin alumina scales formed on Ni-20Cr-A1 alloys improved the sulfidation ye~istance significantly. So far, no report on the corrosion behavior of Ni-Mo-A1 alloys in H2-HzO-H2S atmospheres has been published. The purpose of this research was to study the effect of A1 additions on the corrosion behavior of Ni-Mo alloys in Hz-HzO-H2S atmospheres. EXPERIMENTAL PROCEDURES

Seven ternary Ni-Mo-A1 alloys were prepared by arc-melting in a Tigettered argon atmosphere from elemental metals (> 99.8%). Buttons were cut into halves, sealed in evacuated quartz tubes, and annealed at 900~ for 36 hours. The compositions of the alloys are listed in Table I in both at.% and wt.% (alloy composition will be given in weight percentage in this paper). Samples were prepared in the same way as described previously. 17

122

He and Douglass

Fig. 2. (a) Microstructure of Ni- 10Mo-7.5A1, (b) microstructure of Ni- 10Mo- 10A1, (c) microstructure of Ni-10Mo-15A1, (d) microstructure of Ni-20Mo-7.5A1, (e) microstructure of Ni-20Mo-15A1, (f) microstructure of Ni-30Mo-3.8A1, and (g) microstructure of Ni-30Mo-7.5AI.

Corrosion Behavior of Ni-Mo

123

Fig. 2. Continued.

Corrosion experiments were carried out in a system described by Shing gas mixture containing 1.92H2S, 13.40H2, and 84.68Ar (vol.%) was first passed through a HEO-filled beaker at room temperature and then through a gas washer containing distilled water at 0~ in order to obtain e t al. 15 A

124

He and Douglass

Fig. 2. Continued.

Corrosion Behavior of Ni-Mo

125

Fig. 2. Continued.

saturated water vapor. Based on thermodynamic data, 23 the equilibrium partial pressures were calculated and are shown in Table II. Flow rate of the mixed gas was 150 ml/min. Characterization of the scales was performed using X-ray diffraction, scanning electron microscopy (SEM), and electron-probe microanalysis (EPMA) with a wavelength dispersive X-ray spectrometer (WDS).

RESULTS AHoy Constitution

The compositions of all seven alloys are shown on the partial isotherm on the Ni-Mo-A1 system at 927~ 24 (Fig. 1). Due to the absence of a standard X-ray diffraction pattern of ~(NiMo), it is impossible to compare the X-ray diffraction results of Ni-20Mo-7.5A1 and Ni-30Mo-3.8A1 with the phase diagram, which shows that Ni-20Mo-7.5A1 lies in the -y'(Ni3A1)+ ~(Mo) + ~ region, and Ni-30Mo-3.SA1 consists of "~'+ "~(Ni)+ ~ (three phases). Besides these two alloys, X-ray diffraction results of the other five alloys were in good agreement .with Fig. 1. Microstructures of the seven alloys are shown in Fig. 2. Ni-10Mo-7.5A1 consisted of "y'+ 7(Ni). The Microstructure of Ni-10Mo-15A1 is different from Ni-20Mo-15A1, even

126

He and Douglass

10

a

[]

9

~

800oc

Ni-10Mo-7.5AI

8

7 6

5 4 3 2 1 0

,s-r-.l~ 50

b

100 150 t~ (see1,2)

800~

'7

250

20O

Ni-10Mo-10AI

6 "O

5 4

7 ~

3 2

600"

1 0

i

50

i

100

150

200

250

i

i

300

350

400

tl/2 (~t/2) Fig. 3. Corrosion kinetics: parabolic plots of (a)Ni-10Mo-7.SAI, (b) Ni-10Mo-IOAI, (c) Ni-10Mo-15AI, (d) Ni-20Mo-7.5A1, (e) Ni-20Mo-15AI, (f) Ni-30Mo-3.8AI, and (g) Ni-30Mo-7.5Al.

Corrosion Behavior of Ni-Mo

127

12

C Ni-10Mo-15AI

10 950~ o I:=J0

E

4

r

900oc

~

2

o

o

C 0

50

100

150

200

250

300

350

400

450

t in (secIn) 14

d

O

Ni-20Mo-7.5AI

12 900oCff ~ r

E

8

~0

6

A o 800oc

"4

,

i

50

100

i

150

Fig. 3. Continued.

200

i

250

300

350

128

He and Douglass 1.8

e Ni-20Mo-15AI

1.6

1,4 E{ O

1.2

"~

0.B

~

0.6

900~

800oc+

7OO~c ~

'

~

0.4 A

A

0.2 ~

0

50

100

J

i

150

200

250

J

i

i

300

350

400

450

t 1/2 (sex~la)

35

f

Ni-30Mo-3.8AI

3O o

800~

~

900~

25

E 20

10 600~

50

100

tv2 ( S C C

150 1/2)

Fig. 3. Continued.

200

Corrosion Behavior of Ni-Mo

129

g 7

800.~/00.C

Ni-30Mo-7.5AI

6 5 E

rA

•

4 3 2 1 0 50

100

150

200

250

300

350

t ~ (secl~) Fig. 3. Continued.

though both lie in the same phase region of "y' + a +/3(NiA1) as shown in Fig. 1. A similar phenomenon was also observed for Ni-10Mo-10A1 and Ni-30Mo-7.5A1, which consisted of two phases, 7' + a.

Kinetics

Corrosion of the seven ternary alloys was performed over the temperature range of 600-950~ Parabolic plots of weight gain vs. square root of time are shown in Fig. 3. Most of the kinetics at lower temperatures followed the parabolic rate law, although some showed two parabolic stages. Samples corroded at higher temperatures tended to obey linear kinetics. Increasing A1 content tended to shift linear kinetics to higher temperatures. The effect of A1 composition on the kinetics at each temperature is shown in Fig. 4. An Arrhenius plot of the parabolic rate constants is shown in Fig. 5. "Apparent" activation energies for the alloys corroded in the mixed gas in this as well as the previous experiments 13:7 were calculated and are listed in Table III. The dependence of log kp on aluminum content (Fig. 6) shows that corrosion rates decreased significantly with increasing A1 content.

4 a

~

o

Ni-10Mo-xA! at 7000C 7.5AI

3.5

+

3

.~

§

10AI

n

2.5

+

q.+

[] +

0

2

.~1.5 O0

15AI

1

O O

0.5 I

I

!

5

lO

15

0 o

'

I

'

I

2O

25

30

Time (hour)

b

Ni-20Mo-xAl at 700"C

7 7.5A1

da

6 O

5

O D

r

4

t~

3

o o

2

[]

0 ~x_.~xar:~ I 5 0

X

15AI

x

X

X

I

I

I

I

I

10

15

20

25

30

35

Time (hour)

c

Ni-30Mo-xAI at 700"C

8

~

o 3.8A1

7 n 12 []

A

A

t, O

4

7.5A1

0 o O

2

3

0

AA

I

I

i

I

I

I

5

10

15

20

25

30

35

Time (hour) Fig. 4. Effect of A1 content on corrosion kinetics at 700~ (a) Ni-10Mo-xA1, (b) Ni-20Mo-xA1, and (c) Ni-30Mo-xA1.

Corrosion Behavior of Ni-Mo

131

Scale Morphology and Constitution The scale morphology and constitution depended on alloy composition for a fixed experimental condition. A double-layered scale generally formed on alloys with low A1 content, i.e., an outer layer of nickel sulfide and a complex inner scale. The thickness of both the outer scale and the inner scale decreased with increasing A1 content. Furthermore, the outer scale becomes porous and discontinuous with increasing A1. In addition to temperature, corrosion time is another important parameter affecting the scale morphology. Figure 7a displays the morphology of the scale formed on Ni-10Mo-7.5A1 coded at 600~ for 5 hours. Two regions (bright and dark) were observed on the scale. However, a micrograph of the same alloy corroded for 12 hours (Fig. 7b) shows that the dark region already covers the entire surface. According to EDX and X-ray diffraction analysis, the dark region consisted of Ni3S 2, while the bright region consisted of the elements Ni, A1, and S with or without Mo. In the next section, the morphology and constitution of the scales formed on the different alloys will be presented for two corrosion temperature ranges.

Scales Formed at Lower Temperatures (600-700~ Figure 8 shows the scales formed on Ni-10Mo-10A1 after corrosion at 700~ The irregularly shaped phase with multiple facets was found to be Ni3S2 in a plan view of the surface (Fig. 8a). A whisker-like phase was also found, in which Ni, S, and A1 were detected by EDX. A cross-section of the scale is shown in Fig. 8b. Figure 8c is the concentration profile across the scale. The outer scale was Ni3S2, while MoS2, A12S3, A10.55Mo2S4, and A120 3 were found in the inner scale by X-ray diffraction. The duplex scale formed on Ni-20Mo-7.5A1 corroded at 600~ for 31 hours is seen in Fig. 9a. A continuous layer of Ni3S2 formed on the top of a multiple-phase inner scale, which consisted of Ni2.sMo6S6.7, MoS2, A12S3, A10.55MoES4, uncorroded Mo, and a trace amount of A120 3. The same whiskers, as found in Fig. 8a were also observed growing from blocks of NiaS 2 on the surface of Ni-20Mo-7.5A1 after 23-hour corrosion at 600~ as shown in Fig. 9b, although the diameter of the whiskers is smaller. Figure 10a depicts the surface morphology of the scale formed on Ni-20Mo-15A1 corroded at 700~ Figure 10b and c are backscattered electron image (BEI) and corresponding composition profile across the scale, respectively. The corrosion scale is very thin, and no continuous layer of Ni3S2 formed. X-ray diffraction showed that the outermost part consisted of Ni3S2, MoS2, and A10.55MoES4, while MoS2, A12S3, A10.ssMo2S4, and uncorroded Mo existed in the innermost part of the scale.

132

He and Douglass -6.5 -7,0

..... Ni- IOMo-7.5AI ..... N i - l O M o - IOAI * * * * ' * N i - I O M o - 15AI ,,,,,__A pure Mo , "-." ".--" pure Ni

-7.5 i---,

7O3

9~t

-8.0

\

a \ \

\ ,.

--8.5

--9.0

0 ~

-9.5

0.-10.0 "x" -10.5 _,_0 - 1 1.0

- 1 1.5 -12.0

7 .........

8' . . . . . . . . . . . . 9

I/T

""' ....

i~

.......

. . . 12 ..........

~'1' . . . . .

' 13

x 10000 ( l / K )

-6.5 -7.0 -7.5

=-.--9 ~,-,

.....

Pure Mo pure Ni Ni-20Mo

b

\ \

Ni- 20Mo-7.SAI

\\

9 * * * ~ , N i - 2 O M o - 15AI

-8.0 7 .r I

t'N

-8.5 E

-9.0

0 ~

-9.5 -10.0 -10.5

0

-11.0 -11.5 -12.0

7" ....... ~ ......... ~ . . . . . . . .

~

.......

~'; . . . . . . .

~'~ . . . . . . .

~'3

I/T x I0000 (I/K) Fig. 5. (a) Arrhenius plot comparing the corrosion rate of N i - 1 0 M o xA1 with pure Ni, Mo, (b) Arrhenius plot comparing the corrosion rate of Ni-20Mo-xAl with pure Ni, Mo, and (c) Arrehnius plot comparing the corrosion rate of Ni-30Mo-xA1 with pure Ni, Mo.

Corrosion Behavior of Ni-Mo

133

-6.0

-7.0 It

-8.0

I I E

-9.0

0 ,,,~.,~-- 10.0 CL

- 1 1.0 ; , 9 ,..__, Pure Mo 9__, =._, pure Ni

E~ 0 -12.0

-13.0

..... 9 ****

Ni-3OMo Ni-3OMo-3.8AI Ni-3OMo-7.SAI

iiiiwlllllillltllll~l

7

Ii,

8

....... ~'6 .....

1,IT x 10000

,,lllllllllllll

1';'

~2

(I/K)

'i'3

Pig. 5. Continued.

A cross-section of the scale formed on Ni-30Mo-3.8A1 corroded at 600~ is shown in Fig. 1la. X-ray diffraction results showed that the continuous outer scale consisted of Ni3S2 plus Ni7S6, while the inner scale contained primarily Ni2.5Mo656. 7 and uncorroded Mo plus small amounts of MoS2 and A10.ssMoaS4. Further, A1 was found dissolved in the outer scale as seen in the composition profile across the scale in Fig. 1lb. Table 1II. Activation Energy for Corrosion in Mixed Gases Alloy Pure Ni Pure Mo Ni-20Mo Ni-30Mo Ni-10Mo-7.5Al Ni-10Mo-10A1 Ni-10Mo-15A1 Ni-20Mo-7.5AI N i - 2 0 M o - 15Al Ni-30Mo-3.8A1 Ni-30Mo-7.5Al

Q (kcal/rnol) 68.5 19.2 49.4 77.8 15.9 42.9 42.5 31.6 17.9 25.1 50.4

134

He and Douglass -8.5

a N i - 1 0 M o - xAI

-9.0

_.--.. I (/)

-9.5

-al-

l

E -IO.O O

r

C~ -10.5 ,,/s O~ 0 - 1 1.0 .=..

-11.5

,~,~, 700*C sAaaa 800~ 9 9

-12.0

i l l l l l l l l l l l l l l i l l l | l l l l m " l l l l l l l ' l " l

6

10

14

wt.~ -7.0

b

L,,

\

-8.0

:)ure Mo

18

AI

Ni-20Mo-xAI %. \

A

T

-o.o O3

I

~E - 1 0 . 0 O

-11.0 O~ O -

-

.**** 600*C o o ~ 1 7 6700~

-12.0

-13.0

0 .....

3 . . . . . '6 . . . . .

9 . . . . . I'2. . . . . I'5. . . . . I'8. . pure .. Mo

wt.~6 AI Fig. 6. (a) Effect of A1 content on logkp of Ni-10Mo-xA1, and (b) effect of A1 content on log kp of Ni-20Mo-xA1.

Corrosion Behavior of N i - M o

Fig. 7. Surface morphology of the scale formed on Ni-10Mo-7.5A1 at 600~ for (a) 5 hours, and (b) 12 hours.

135

136

He and Douglass

Fig. 8. (a) Surface morphology of the scale formed on N i - 1 0 M o - 10A1 corroded at 700~ for 22 hours, (b) cross-section of 8a, and (c) composition profile of 8b.

Corrosion Behavior of N i - M o

C ~

basemetal

137

,..I inner L.,

outerscale

J'[ scale F'w .:!

/! I" {.,"

i

/

/

1 0

20

40

60

,

1

80

1O0

distance (fun) Fig. 8. Continued.

Figure 12 shows the scale formed on Ni-30Mo-7.5A1 after corrosion at 700~ The large dark faceted blocks are Ni3S2. The inner scale is composed mainly of MoS2 and uncorroded Mo plus small amounts of A10.55Mo2S4, A1203, and Ni2.sMo6S6.7.

Scales Formed at Higher Temperatures (800-900~ Large spheres or ellipsoids, 300-400 #m in size, which are visible to the naked eye, were found on the surface of Ni-10Mo-7.5A1 after 800~ corrosion for 5 hours, as shown in Fig. 13a. Some small islands were observed on the spheres, as in Fig. 13b. The EDX analysis and X-ray diffraction showed that the large spheres consist of Ni3 $2 and that the small islands might be AlzS 3. A cross-section of the scale and its composition profile are given in Figs. 13c and d, respectively. The outer layer consisted mainly of Ni3S2. Phases of Ni3S2, MoS2, Ni25Mo6S6.7, A12S3, and AlzO 3 were found in the inner scale by X-ray diffraction analysis. Although A1203 was not detected by X-ray diffraction in the outer scale, Fig. 13d shows that some A1 exists on the surface. The preferential corrosion of grain boundaries can also be seen in Fig. 13c. Figure 14a shows a plan view of the scale formed on Ni-10Mo-10A1 corroded at 800~ for 5 hours. The size of the spheres is smaller than that

138

He and Douglass

Fig. 9. (a) Cross-section of scale formed on N i - 2 0 M o - 7 . 5 A 1 corroded at 600~ for 31 hours, and (b) whisker-like phase on the scale formed on N i - 2 0 M o - 7 . 5 A 1 corroded at 600~ for 23 hours.

Corrosion Behavior of N i - M o

139

formed on the surface of Ni-10Mo-7.5A1 under the same experimental condition (Fig. 13a). In addition, A1 was detected on the surface of the spheres in addition to Ni and S. Figures 14b and c represent the image and composition profile of the scale cross-section, respectively. Only some discrete Ni3Szspots can be seen on the outermost part of the scale. The phases found in the scale were Ni3S2, MoS2, A12S3, and A1203. Localized attack also occurred in Ni3A1 in this sample. Figures 15a and b show the surface and cross-section of the scale formed on Ni-20Mo-7.5A1 after corrosion at 800~ By comparing the size of the spheres and the scale thickness shown in Fig. 15 with those in Figs. 13 and 14, it can be deduced that A1 is more efficient than Mo in improving corrosion resistance. In addition to Ni3S2, a trace amount of A1203 was found on the surface of the scale by X-ray diffraction. MoS2, A12S3, A1203, A10.55Mo2S4, Ni2.5Mo686.7, and Mo were found in the inner scale. The scale formed on Ni-20Mo-15A1 corroded at 800~ is shown in Fig. 16. Along with some ellipsoids described above, crooked sticks with spherical heads were also observed on the surface in Fig. 16a. According to the EDX analysis, Ni and S were detected on the spherical heads. However, A1 was also found on the surface of the sticks besides Ni and S. The outer scale became discontinuous. X-ray diffraction analysis shows that the outermost part of the scale was composed of Ni3S2, A12S3, A10.55Mo2S4, and trace amounts of A1203 and MoS2. The inner part of the scale consisted mostly of A12S3 with some A10.55Mo2S4, A1203, and Mo particles. The presence of A12S3 was also proved in Fig. 16c. In contrast to the thin scale and wavy inner-scale/base-metal interface formed on Ni-20Mo-15A1 corroded at the same temperature described above, the scale formed on Ni-30Mo-3.8A1 after 1 &our corrosion at 800~ is very thick, as shown in Fig. 17a. The bright phase in the discontinuous outermost scale is pure Ni, and the dark phase was found to be Ni3S2. The inner scale consisted of pure Ni (brightest phase), A10.55Mo2S4 (gray), Ni3S2 (darkest), as well as trace amounts of A12S3 and A1203. When approaching the interface of the inner scale and the substrate, the amount of Ni3S2 and Ni decreased while the amount of A10.55Mo2S4 increased. A line scan across the scale in Fig. 17b shows the composition distribution. J

DISCUSSION From the results described above, the primary features can be summarized as follows: (1) the corrosion rate decreased with increasing A1 content, (2) most of the kinetics followed the parabolic law at low temperatures,

140

He and Douglass

Fig. 10. Surface morphology of the scale formed on N i - 2 0 M o - 15A1 corroded at 700~ for 31 hours, (b) cross-section of 10a, and (c) composition profile of 10b.

Corrosion Behavior of N i - M o

C

~,= -~

141 I

scale

~ jr[1 ~-~

base metal

AI," "

. . - - h =~

,.

. . . . . . . . . .

~ aI

II

Ni

0

5

10

15

20

25

distance (pal) Fig. 10. Continued.

while linear kinetics were observed primarily at higher temperatures, (3) at a fixed Mo content, the transition from parabolic to linear kinetics shifted to higher temperatures with increasing A1 concentration in the alloys, (4) double-layered scales generally formed on the alloys with low A1 content, with an outer layer of nickel sulfide and a complex inner scale, (5) the thickness of the outer scale and the inner scale decreased as the A1 content increased, the outer scale becoming porous and discontinuous for alloys containing high A1 content, (6) spherical Ni3S2 formed on the surface of the alloy after corrosion at temperature > 800~ (7) according to the X-ray diffraction analysis, trace amounts of A1203 were detected in the inner scales of all alloys corroded at higher temperatures (> 800~ and it was also found on the surface of some alloys with at least 7.5% A1 content. Stability phase diagrams of A1 at different temperatures have been calculated based on thermodynamic data 25 and are shown in Fig. 18. It is found that the gas composition lies in the A1203-stable region at all temperatures studied. The stability phase diagrams for Ni and Mo show that Ni3S2 and MoS2 should be the stable phases, respectively, according to previous studies. ~7'18 Although it is the alloy and not the pure element which is exposed to the gases, the calculated stability diagrams are still useful in predicting the approximate phase equilibria. It is reasonable to expect that Ni3S2 exists in the outermost scale due to its faster growth rate and the large content of Ni in the alloys. Most of

142

He and Douglass

b

~"

base metal

inner scale

% fj el i

,

J i

S

I\\" . ~ '

,t ",,, \,,d*

i

0

r

i

20

40

1

60 d i s t a n c e (Ima)

\

t

.k. ~,

AI

t

80

. IO0

Fig. 11. (a) Cross-section of the scale formed on Ni-30Mo-3.8A1 corroded at 600~ for 7 hours, and (b) composition profile of 1 la.

Corrosion Behavior of N i - M o

Fig. 12. (a) Surface morphology of the scale formed on Ni-30Mo-7.5A1 corroded at 700~ for 31 hours, and (b) cross-section of 12a.

143

144

He and Douglass

Fig. 13. (a) Surface morphology of the scale formed on N i - 10Mo-7.5A1 corroded at 800~ for 5 hours, (b) small islands on large sphere of 13a, (c) cross-section of 13a, and (d) composition profile of 13c.

Corrosion Behavior of Ni-Mo

145

J

base metal

d

inner scale ~

outer scale

_~

,i t,

'AI

!

,,',~

grain boudary

- ;

Ni

I

30

I

60

90 distance (pro)

Fig. 13. Continued.

120

150

180

146

He and Douglass

Fig. 14. (a) Surface morphology of the scale formed on Ni-10Mo-10A1 corroded at 800~ for 5 hours, (b) cross-section of 14a, and (c) composition profile of 14b.

Corrosion Behavior of N i . M o

C

~,

"I

147

scale

--

~<1----

b a s e rnetai

l~

~j

l.'

:

i : :

:,A

AI

i ls: ":::. . . . . i

,.,_IV

,

./\

, i

9

Ni

j

X

-

,._ ,~" " "'~""~ 0

10

,

20

30

40

~

50

~

'

,

60

70

distance (pro) Fig. 14. Continued.

the scale structures formed in this experiment are indeed consistent with predictions. However, Ni3S2 and some pure Ni were observed in the inner scales formed on Ni-10Mo-7.5A1 and Ni-30Mo-3.SA1 after corrosion at 800~ As known, a eutectic reaction exists between Ni and Ni3S2 at 645~ in the Ni-S binary system. It appears that there must also be a eutectic reaction in various ternary systems. The eutectic-liquid phase is expected to exist during the corrosion and decompose into Ni, Ni3S2, and other phases, which could be MoS2, A1203, A12S3, or AI0.55Mo2S4, depending on the composition of the alloys. Therefore, Ni3S2 and Ni in the inner scale can be regarded as the products of the decomposition of the eutectic liquid phase. It is also reasonable that A1203 remains in the inner layer if it can nucleate before the Ni3S2 screens A1 from the gas, or if O can diffuse through the Ni3S2 outer layer after the screening. Some A1203 has been detected in scales formed after corrosion at higher temperatures. In addition, the amount of A1203 increases toward the interface between the inner scale and substrate. As observed, the Ni3S2 outer scale became more porous and discontinuous with increasing temperature. Thus, it is possible for oxygen to diffuse in and to react with A1 due to the lack of screening from a continuous Ni3S2 layer. The slow growth rate of AlzO3 results in its being buried in the inner scale. No continuous layer of A1203 has been observed in this study, which behavior can be attributed to the multi-phase

148

He and Douglass

Fig. 15. (a) Surface morphology of the scale formed on N i - 2 0 M o , 7 . 5 A 1 corroded at 800~ for 5 hours, and (b) cross-section of 15a.

Corrosion Behavior of N i - M o

149

nature of the alloys and to the slow growth rate of A1203. The reason why A1203 was detected only in the scales formed after corrosion at higher temperatures might be as follows: (1) the higher oxygen partial pressure at high temperatures in the mixed gas was favorable to A1203 formation; and (2) the amount of A1203 formed at low temperatures was too small to be detected by X-ray diffraction. According to X-ray diffraction analysis, A12S3 was found in the scales. Other evidence is the production of some H2S due to A12S3 hydrolysis, during the sequential grinding of the scales for X-ray diffraction analysis. Although the formation of A12S3 is less favorable in terms of thermodynamics, compared to A1203, the kinetics factor must be considered also to comprehend the corrosion-scale structures. The sulfur partial pressure is much higher than that of oxygen in the mixed gases, as shown in Table II. For example, the difference between the free energies of A12S3 and A1203 formation is slightly more than a factor of two (-574.4 kJ/mol for A12S3 and - 1370.2 kJ/mol for A1203). However, the sulfur partial pressure is seventeen orders of magnitude higher than that of oxygen at 700~ Thus, there is much more sulfur than oxygen to be in contact with A1. Furthermore, once nucleated, A12S3 will grow faster than A1203. The formation of an A12S3 inner layer in Ni-A1 alloys was believed to contribute to the reduction of the sulfidation rate of N i , as studied by Chen et al. 8 In addition, some A12S3 was observed also on the outermost scales after higher-temperature corrosion. This might be the result of the decomposition of the liquid phase, as described above. As in the sulfidation of Ni-Mo-A1 in pure sulfur, the spinel phase, Alo.55M02S4, was also detected in the inner layer of scales formed in the mixed-gas corrosion. The presence of the spinel phase was always associated with the considerable reduction in the sulfidation rate in previous research. 7-1~ Therefore, it is fair to say that the formation of Alo.55M02S4 should contribute also to the decrease in the corrosion rate with increasing A1 content. Apart from Alo.55MozS4, another double sulfide, Ni2.sM06S6.7, known as a Chevrel phase (with a formula MxM06Ss) , was observed also inside some inner scales of the alloys with relatively low Mo or low A1 contents. This phase was found responsible for the significant decrease of the corrosion rate observed in a previous study. 17 Increasing the Mo and A1 contents resulted in decreasing corrosion rates of the Ni-Mo-A1 alloys in the mixed gas. Of the seven alloys studied, the lowest corrosion rate was found for Ni-20Mo-15A1, the weight gain of which was about 0.4mg/cm2 after corrosion at 600~ for 34.4hours. The corrosion rate of this alloy is of the same order as that of pure Mo, as shown in Fig. 6. The reduction of corrosion rate with increasing A1 content

150

He and Douglass

Fig. 16. (a) Surface morphology of the scale formed on Ni-20Mo-15A1 corroded at 800~ for 28 hours, (b) cross-section of 16a, and (c) composition profile of 16b.

Corrosion Behavior

of Ni-Mo

151

~1...=

base metal

.J ,q

scale

"T"

AI

.=~

S

Mo

9 ~

9

"

~

"-

.

.

.

.

.

.

.

.

.

"

9,

0

5

10

15

20

/

.

25

\

30

distance (lun)

Fig. 16. Continued. of the alloys can be ascribed mostly to the formation of A1203, A12S3, and Alo.55MoES4 in the inner scales. Another interesting aspect is that the Ni3S2 phase formed on all the alloys, although the amount of which was very small in some cases, i.e., Ni-20Mo-15A1. This might be due chiefly to the multiphase nature of the alloys. Various phases in the substrate react with the gas species simultaneously at different rates. The phase having the fastest growth rate tends to cover the entire surface and form the outermost part of the scale, while the phase with the lowest growth rate is located in the inner part of the scale. NiaS2, not A1203, happens to be the phase with the fastest growth rate in present study. Therefore, it is difficult to form a continuous and protective A1203 layer, especially in the present mixed gas with a high $2 partial pressure and low O2 partial pressure. CONCLUSIONS Ni-Mo-A1 alloys corroded in the mixed gas followed mostty the parabolic law at low temperatures, while linear kinetics were observed primarily at higher temperatures. At a fixed Mo content, the transition from parabolic to linear kinetics shifted to higher temperature with increasing A1 concentration in the alloys. Double-layered scales generally formed on alloys with low A1 contents, with an outer layer of nickel sulfide and a

152

He and Douglass

base metal

b

kl.d

inner scale

"l"

/ ~.

f'~ I %

I

k

I

I ~

~ I~

~

A 1

!1

40

~

I/

,-,,

.,/,,,,..-.: .:.,... ,"-20

~,: I~

,

~

60

80

100

scale-~~

;~,1 ,,

S

I I

t" . . . . o

~ ~--

120

'1~1' t~, , '(

"'" I

I

,,:1',1 ~'~ ; ,: tlt

l

A.;

[

140

It

160

180

distance (tun) Fig. 17. (a) Cross-section of the scale formed on Ni-30Mo-3.8A1 corroded at 800~ for 1 hour, and (b) composition profile of 17a.

S3

-~o

-26-~o

.

-40

-is -40 togPo2

3

logPo~

-45

.

_

-3~

-35

3

.

.

-3o

AtzO3

-30

-~5

-25

x

-2o

-20

m

-55

-26,

2

-18

-to

-i

-e~60

-22

-t@

0)-14

E

-I0

-6

-2

-50

-50

At

-,45

900"C AI2S3

-55

AL

AL~S3

700~

-45

-40

-40

-35

-30

At~O~

LooPoz

-~

togPo~

-25

-30

Fig. 18. Stability phase diagrams of A1-O-S. X indicates the gas composition used.

At

-s5

/

AI~ S

-50

800~

-55

I

AI~

-2a

-18

D_ O) -14 o

-10

-2 -6

-60

-26

-22

-18

E 0)-14

-10

-6

600"C /

•

-20

-e5

x

-16

-eO

T"

&

O

154

He and Douglass

c o m p l e x inner scale. T h e thickness o f the o u t e r scale a n d the inner scale d e c r e a s e d as the A1 c o n t e n t increased. The o u t e r scale becomes p o r o u s a n d d i s c o n t i n u o u s with increasing A1 content, a n d increasing t e m p e r a t u r e . A1203 was d e t e c t e d in scales o f all alloys c o r r o d e d at higher t e m p e r a t u r e ( > 800~ even t h o u g h the a m o u n t o f A1203 is very small in some cases. T h e decrease in c o r r o s i o n rate with increasing A1 c o n t e n t m a y be a t t r i b u t e d to the presence o f the f o r m a t i o n o f A1203, Alo.55M02S 4, a n d A12S3 in the inner scale. ACKNOWLEDGMENTS T h e a u t h o r s are grateful to the Electrical P o w e r R e s e a r c h Institute, P a l o A l t o , C a l i f o r n i a , for financial s u p p o r t for this work. T h e e n c o u r a g e m e n t o f J. Stringer is greatly a p p r e c i a t e d . REFERENCES

1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25.

M. F. Chert and D. L. Douglass, Oxid. Met. 32, 185 (1989). M. F. Chen, D. L. Douglass, and F. Gesmundo, Oxid. Met. 31, 237 (1989). B. Gleeson, D. L. Douglass, and F. Gesmundo, Oxid. Met. 31, 209 (1989). B. Gleeson, D. L. Douglass, and F. Gesmundo, Oxid. Met. 33, 425 (1990). R. V. Carter, D. L. Douglass, and F. Gesmundo, Oxid. Met. 31, 341 (1989). G. Wang, R. V. Carter, and D. L. Douglass, Oxid. Met. 32, 273 (1989). M. F. Chen and D. L. Douglass, Oxid. Met. 33, 103 (1990). M. F. Chert, D. L. Douglass, and F. Gesmundo, Oxid. Met. 33, 399 (1990). B. Gleeson, D. L. Douglass, and F. Gesmundo, Oxid. Met. 34, 123 (1990). G. Wang, D. L. Douglass, and FI Gesmundo, Oxid. Met. 35, 349 (1991). V. S. Bhide and W. W. Smeltzer, J. Eleetrochem. Soc, 128, 902 (1981). E. Godlewska, K. Godlewski, and S. Mrowec, Mater. Sci. Eng. 87, 183 (1987). W. Kai, D. L. Douglass, and F. Gesmundo, Oxid. Met. 37, 389 (1992). W. Kai, D. L. Douglass, and F. Gesmundo, Oxid. Met. 37, 189 (1992). C. C. Shing, D. L. Douglass, and F. Gesmundo, Oxid. Met. 37, 441 (1992). C. C. Shing, D. L. Douglass, and F. Gesmundo, Oxid. Met. 37, 167 (1992). Y. R. He, D. L. Douglass, and F. Gesmundo, Oxid. Met. 37, 413 (1992). Y. R. He, D. L. Douglass, and F. Gesmundo, Oxid. Met. 37, 217 (1992). S. Mrowec and K. Przyblyski, High-Temp. Mater. Proc. 6, 1 (1984). C. S. Giggins and F. S. Pettit, Oxid. Met. 14, 363 (1980). K. Ohla, S. W. Kim, H. Fischmeister, and E. Fromm, Oxid. Met. 36, 379 (1991). S. W. Kim, K. Ohla, H. Fischmeister, and E. Fromm, Oxid. Met. 36, 395 (1991). J. Phys. Chem. Ref. Data 14 (Suppl.), 1516-1526 (1985). D. B. Miracle, K. A. Lark, and V. Srinivasan, Met. Trans. 15A, 481 (1984). J. Phys. Chem. Ref Data 14 (Suppl.), 156-165 (1985).