The potentials plotted in Fig. 2 are the peak- 123docz.net

16 SLOW STRAIN RATE TESTING

2. The potentials plotted in Fig. 2 are the peak (maximum) free corrosion potentials measured

Shibata [9] also concluded that the SSR test technique was not effective at reproducing the cracking observed in constant load tests in lower concentration (CaC12 concentrations below 25%) solutions. However, comparison of his test results with those of Daniels [5]

suggests that the strain rate used by Shibata was too fast for the low concentration solutions.

The study by Shibata demonstrates the two most important parameters affecting cracking, namely, potential and strain rate.

Potential and strain rate effects may not account for all of the discrepancies in the data between the SSR test technique and other SCC test techniques. Yang [14] evaluated the effects of temperature, chloride content, and dissolved oxygen content on the susceptibility to SCC of slightly sensitized alloy 304 in high-temperature water. U-bend and SSR test

BEAVERS AND KOCH ON SSRT LIMITATIONS 29

- 0 . 3 0 0

w - 0 . 3 2 0

0 I

I -0.340

~ C

~ - 0 . 3 6 0 0

m - 0 . 3 8 0

0 C o n s t a n t Load Test 9 SSRT

55% ; 20%

~ 2 0 %

5 0 ~ % ~ 30%

- 0 . 4 0 0 . . . . i . . . . i . . . . t . . . .

3 4 5 6 7

FIG. 2--Relation between peak potential and p H for alloy 304 in CaCI~ solutions of different concen- trations (after Shibata [9]).

techniques were used and the testing was performed under freely corroding conditions.

Oxide film thickness and composition were measured using Auger Electron Spectroscopy (AES). In water containing 100 ppm CI- and 8 ppm 02 over the temperature range of 200 to 300~ the U-bend specimens showed a decreasing susceptibility to cracking with increasing temperature while the SSR specimens exhibited a maximum in susceptibility to cracking at 250~ (see Fig. 3).

25 100

0 -1-

- 15

D -

O 10

F= E o 5

0 SSR o U - B e n d

0 i i I

150 3 0 0

o / • o

0 0

[ ]

2 0 0 2 5 0

T e m p e r a t u r e , ~

90 80 70

60 r o

3 o ~ 20 []

, o

3 5 0

FIG. 3--Time to failure qf alloy 304 in high-temperature water for U-bend and SSR tests as a function of temperature (after Yang [14]).

University of Washington (University of Washington) pursuant to License Agreement. No further reproductions authorized.

30 SLOW STRAIN RATE TESTING 1.0

0.8

o_ 0 . 6

0 c

0 . 4

o 0 . 2

9 n

n . , - - n I ,

2OO

9 SSR o U - b e n d

0.0 l , i ,

0 400 600 800

Oxide T h i c k n e s s , nm

FIG. 4--Resistance to SCC of alloy 304 in high-temperature water as a function of oxide thickness for U-bend and SSR tests. The resistance to cracking for U-bend is expressed as the f?action of specimens remaining uncracked after the test period. Cracking resistance for the SSR tests is equated to time-to- failure in water/time-to-failure in argon (after Yang [14]).

The difference in the behavior for the two test techniques was attributed to a difference in response to the thicker films formed at higher temperatures. Resistance to SCC was found to increase with increasing film thickness for the U-bend tests while the opposite effect was observed for the SSR tests; see Fig. 4. In Fig. 4, the cracking resistance p a r a m e t e r for U- bends is expressed as the fraction of specimens remaining uncracked after the test period.

The cracking resistance parameter for the SSR tests is equated to time-to-failure in water/

time-to-failure in argon. While these cracking resistance parameters are calculated differently for the two test techniques, they provide similar conclusions at the limits of behavior; a cracking resistance p a r a m e t e r of 1 indicates negligible cracking.

It was also noted in the study by Yang [14] that the composition of the films formed on the SSR test specimens was enhanced with respect to nickel while no such enhancement was observed for the films on the U-bend specimens. Only limited free corrosion potential data were given for the SSR specimens and no potential data were given for U-bend specimens. It is tempting to attribute the difference in composition of the films on the two types of specimens to differences in potential during testing. However, this interpretation does not account for the effects of film thickness on cracking behavior for the two test techniques.

From a mechanistic standpoint, the film thickness effects also are reasonable. U n d e r SSR test conditions, a thick film may fracture more readily than a thin film, while a thick film may be more protective for a specimen tested under constant strain.

Nickel-Base A l l o y s

Four responses to the questionnaire survey indicated anomalous behavior for the nickel- base alloys. Sridhar [15] and Kolts [16] reported on the anomalous behavior of alloy G-3 and alloy 825 in H2S environments. In the research by Sridhar [15], the behavior clearly can

BEAVERS AND KOCH ON SSRT LIMITATIONS 31 be attributed to solution chemistry effects. He reported that alloy G-3 and alloy 825 cracked in C-ring tests in a NaC1-H_,S-S ~ environment in several hundred hours while, under the same conditions, no cracking was observed in the SSR tests. SSR tests performed by Wilhelm

[17] under similar conditions produced cracking of these alloys. A close comparison of the test procedures at the two laboratories indicated that Wilhelm agitated the autoclave fol- lowing H2S addition to more rapidly reach equilibrium. When Sridhar followed this test procedure, he also observed cracking in SSR tests in these alloy-environment systems.

Wilhelm attributed the anomalous SSR behavior to the difficulty in achieving the proper H2S concentration in the solution in the short time period of the SSR test.

It also has been reported that the deaeration procedure used in SSR testing of nickel- base alloys in H~S environments can significantly affect the results. Ikeda [18] found that SCC susceptibility of several nickel-base alloys was much greater when only N2 sparging, as opposed to vacuum deaeration, was used to deaerate the autoclave prior to SSR testing in H2S-CO2-CI environments at 150 to 175~

Kolts [16] reported anomalous behavior in SSR tests on alloy 825 and alloy G-3 in high- temperature CO2-H2S environments. In long-term (over one year) tests, C-ring specimens of alloy 825 cracked but no cracking was observed in SSR tests at a strain rate of 2 • 10-"

s L. The C-ring and SSR tests also gave reverse rankings for the two alloys. The C-ring specimens were prepared from sheet and had a total applied strain of about 6%. The tests were performed on the two alloys at the same strength level, which resulted in a higher degree of cold work for the alloy 825 specimens. Typically the alloy 825 specimens had an elongation to failure in standard mechanical tests in air of only 7% whereas the alloy G-3 specimens had an elongation of 25%. As a result of this difference in ductility, the SSR test times for the alloy 825 specimens were considerably shorter than those for the alloy G-3 specimens. Kolts speculated that this difference in time in the test solution may account for the anomalous behavior observed.

Asphahani [19] reported similar anomalous behavior for the SSR test technique in SCC studies of alloy C-276 and alloy MP35N in 50% N a O H at 147~ Constant strain (C-ring) tests of these alloys indicated that the 50% cold-worked specimens were more susceptible to SCC than solution annealed specimens, based on the maximum depth of cracking. On the other hand, SSR tests of the same materials indicated that the annealed material was more susceptible to SCC. Asphahani [19] attributed this behavior to the fact that the cold- worked C-ring specimens experienced much higher stresses than the annealed specimens.

In these tests, the total strain was about 6%, which corresponded to a stress of 200 ksi (1380 MPa) for the cold-worked specimens and 70 ksi (483 MPa) for the annealed specimens of alloy C-276. Strain rate and test time effects were considered but rejected as an explanation for the behavior. In the SSR tests, the failure times for the cold-worked material were shorter than the annealed material because of the lower ductility of the former. However, the cold-worked specimens exhibited shallower cracks than the annealed specimens even when the strain rate of the tests on the cold-worked material was reduced in order to achieve comparable failure times for the two metallurgical conditions.

In both studies, the cold-worked specimens exhibited higher susceptibility in the C-ring tests than in the SSR tests. In the former, the behavior was attributed to a test-time effect (cold-worked specimens failed in a shorter test time than annealed specimens because of the reduced ductility) while in the latter, the behavior was attributed to a strength level effect. One could argue that the test-time effect is a deficiency of the SSR test technique.

On the other hand, one could equally argue that strength-level effects on failure times are an inherent deficiency of constant-strain techniques such as the U-bend technique. When several alloys of different strength levels are plastically strained to the same value, the higher strength materials will experience a higher stress assuming similar rates of work hardening.

University of Washington (University of Washington) pursuant to License Agreement. No further reproductions authorized.

32 S L O W STRAIN RATE TESTING

This problem can be avoided in C-ring tests by controlling the stress level rather than the strain level. The most appropriate test technique depends on the loading conditions in service.

Was [20] reported anomalous behavior in SSR tests on alloys 600, 690, and X750 in deaerated high-purity water at elevated temperatures. Constant load and U-bend tests exhibited SCC while SSR tests did not show significant cracking. The questionnaire response by Payer [21] may provide an explanation for the anomalous behavior reported by Was [20].

Payer reported that alloy 600 has failed by SCC in primary-side water in pressurized water reactors while it is difficult to reproduce the cracking in SSR tests in the laboratory unless very slow strain rates (10 -7 to 10 ~ s ~) are used. The strain rates used by Was were not indicated in the response but it is likely that they were faster than those indicated by Payer.

Bandy [22] also reported, in SSR studies of nickel-base alloys in deaerated high-temperature water, that it was found to be necessary to decrease the strain rate in the range of 1 • 10 .-6 to 3 • 10 s s 1 in order to see SCC of more resistant alloys; details of the testing or actual data were not given in the reference.

Studies by Page [23] are contradictory to the behavior reported by Was [20]. He investigated the use of creviced and uncreviced SSR specimens for the study of the SCC of alloy 600 in partially deoxygenated (200 ppb) and oxygenated (8 to 16 ppm) pure water at 288~

Smooth constant-load test specimens were evaluated for comparison purposes. It was found that the specimens cracked, regardless of heat treatment, in the SSR tests under oxygenated conditions only when crevices were present. No cracking was observed under partially deoxygenated conditions or in the constant load tests under any of the test conditions. The reasons for the discrepancy between the researchers is not readily apparent but may be attributable to environmental factors such as oxygen content or water purity.

Reports of anomalous behavior from the open literature were limited and, in most cases, the root cause was confirmed or speculated by the authors. Examples include the cold work effect described by Asphahani [19] and previously discussed and the effects of multiple cracking on the specimen potential, as reported by Newman [24]. SSR and interrupted SSR testing were performed on alloy 600 in 0.21 M boric acid containing lithium hydroxide, sodium thiosulfate, and sodium tetrathionate at 40~ It was found that the corrosion potential of the specimen was depressed by the initiation of the stress-corrosion cracks in the gage section, which reduced the cracking velocity. The addition of the lithium ion to the test solution greatly reduced the number of cracks in the specimen, which promoted an increase in the potential and the cracking velocity. A t still higher lithium ion concentrations, the inhibiting effect of the ion dominated and the cracking velocity decreased; see Figs. 5 and 6. One might speculate that these deleterious effects could be avoided by performing SSR tests under potential control, but, as described by Newman [24], this approach may not be effective in highly resistive solutions. A similar effect of SSR testing on corrosion potentials was previously reported for stainless steels.

Asphahani [19] also reported that the SSR technique could not distinguish between intergranular attack ( I G A ) and intergranular SCC of nickel-base alloys. In SSR tests on sensitized alloy C-276, intergranular cracking was observed on anodically polarized specimens in a chloride environment. U n d e r similar test conditions, I G A was found on both the tension and compression faces of C-ring specimens. As described by Asphahani [19], both phenomena are capable of causing service failures and the SSR technique is capable of revealing each.

In other work related to the SSR test technique, Asphahani [19] investigated the parameters used to assess SCC susceptibility in a SSR test. It was found that the mechanical properties parameters often used to measure susceptibility, such as percent elongation, percent reduction in area, load at failure, and the total time to failure, were not always

BEAVERS AND KOCH ON SSRT LIMITATIONS 33 2 0 I t

a Interrupted SSR Tests (1 Crock)

18 [] 0 Standard SSR Test

T 16 ~

E \

c 1 4 []

. - - [ ]

~ \

10 []

o 8 S ~

o o

6 / []\

~ 4. q

0 , I ~ I , [ ] ,

0 1 2 3 4 5 6

Lithium Ion Concentration, ppm

FIG. 5 - - E f f e c t o f L i O H addition on the mean crack velocity o f solution-annealed and sensitized lncone1600 in air-saturated 1.3% H3B03 + O. 7 p p m sulfur as Na2S203. The mean crack velocity is shown f o r SSR-tests at 3 • 10 6 s-1 strain rate. For the interrupted SSR test, the crack velocity was estimated

visually and f r o m the load decay at constant deflection (after Newman [24]).

100

,, 1 0

- - Interrupted SSR T e s t s 0 Standard SSR Test

I I I I I

1 2 3 4 5 6

Lithium Ion Concentration, ppm

FIG. 6 - - E f f e c t o f L i O H addition on the number o f cracks in solution-annealed and sensitized lnconel 600 in air-saturated 1.3 H3BOs + O. 7 p p m sulfur as Na2S203. The strain rate was 3 • 10 -6 s ~ (after Newman [24]).

University of Washington (University of Washington) pursuant to License Agreement. No further reproductions authorized.

34 S L O W STRAIN RATE TESTING

consistent indicators of SCC of nickel-base alloys. It was concluded that the metallographic measurement of Secondary crack depth is the most reliable technique for the determination of SCC of nickel-base alloys. Payer [25] reached a similar conclusion for the general application of the SSR technique "Metallography or fractography is essential to confirm the presence or absence of SCC."

Copper Alloys

No responses to the questionnaire indicating anomalous behavior of SSR tests on copper alloys were received. However, effects of strain rate, potential, and metallurgy, which could produce anomalous results, were reported in the literature. Bradford [26], Birley [27I, and Yu [28] found that, in different SCC inducing environments, cracking occurs only below a certain strain rate (10 4 s-~). It should be noted that a minimum strain rate below which SCC would not occur was not detected. Scully [29] further showed that a narrow range of potentials exist around the free corrosion potential for brass in a 15 N ammoniacal solution, where there was maximum susceptibility to SCC.

Finally, various metallurgical parameters such as grain size and degree of work-hardening were reported to affect the results of SSR testing. Yu [28] demonstrated that grain size could affect the various SSR parameters such as crack velocity, frequency of cracking, and percent reduction in area. The effect of work-hardening of copper alloys on SSR testing was demonstrated by Scully [29]. It can be hypothesized that metallurgical factors influence the measured SSR parameters by affecting the strain or stressing rate at the crack tip.

Carbon Steels

There were three responses to the questionnaire indicating anomalous or potentially anomalous behavior with SSR testing of carbon steels. The effect of electrochemical potential on SCC reported by Clayton [30] clearly indicated that if close attention is not paid to the potential at which the SSR test is conducted, the SSR test may exhibit anomalous behavior.

He showed that SCC in continuous digesters occurs in a very narrow potential range and that, for SSR testing in this environment to be meaningful, the tests must be performed within this range. This strong potential dependence was supported by Singbeil [31,32] in tests on carbon steel in typical kraft digester liquors (see Fig. 7), by Sriram [33] who conducted potentiostatically controlled SSR tests in various caustic environments, and by Parkins [34] who studied potential dependence of cracking of carbon steel in carbonate- bicarbonate solutions. In these environments, carbon steel exhibits active-passive behavior in the potentiodynamic polarization curves and the potential range for SCC is associated with the potential range in which the active-passive transitions occur.

Speidel [35] reported on anomalies in the fracture mode of a 3.5% nickel turbine rotor steel in high-temperature pure water. The cracking mode during SSR testing was found to be transgranular while service failures were intergranular. He also reported that cracking was intergranular in fracture mechanics tests performed under constant stress intensity conditions. This apparent anomaly may be attributed to strain-rate effects. Several authors are in agreement that the SSR test is generally more severe than actual field conditions because of the high stresses generated at the crack tip. Also different fracture mechanisms may predominate at different strain rates. For example, Kim [36] showed that, while on one hand intergranular SCC in sodium carbonate/bicarbonate solutions occurs in a very narrow range of strain rates, transgranular hydrogen induced cracking is less strain rate dependent and will thus occur at a much broader range of strain rates.

BEAVERS AND KOCH ON SSRT LIMITATIONS 35

O Q)

.o_

" 6 o w b 3

.~_ ~

1.1 1.0 0.9 0.8 0.7 0.6 0.5 0.4

J o o~

O ~ o / ~

Strain Rate 2.5 X 1 0 - e s - 1

0.3 9 9 Strain Rate 1.0 X 1 0 - s s - 1

0.2 , i , i , i , 1 , i , i ,

- ' 200 - 1 1 0 0 - 1 0 0 0 - 9 0 0 -BOO - 7 0 0 - 6 0 0 - 5 0 0 Potential, m V - v s - S C E

FIG. 7--Effects of potential and strain rate on the reduction in area ratio for SSR tests performed on A516 Grade 70 steel in an impregnation zone liquor containing 20-70 g/l NaOH and 20-25 g/l Na2S at I IO~ (after Singbeil [31]).

Murata [37] reported difficulty in reproducing SCC of a carbon steel in an organic liquid using a conventional SSR test on a specimen with a smooth gage section. However, he could reproduce SCC when using a slow bending rate test technique on prenotched specimens.

In a conventional SSR test, multiple stress-corrosion cracks generally initiate in the gage section of the specimen. Thus, the crack tip strain rate varies during the test and is ill- defined. On the other hand, the strain rate can be more accurately controlled in a precracked or prenotched specimen tested at a constant deflection rate under bending or at a constant cross head speed under tension. Thus, the observation by Murata [37] suggests that stresses at the crack tip and the resulting strain rate played an important role in this material- environment system. The differences in cracking susceptibilities that were reported in the literature for the various test techniques and other environments may also relate to differences in strain rate at the crack tip.

A l u m i n u m Alloys

Experience with the SSR test technique for studying SCC of aluminum alloys has been good. Negative responses, indicating problems with the technique, for this alloy system, were not received from the questionnaire survey. Moreover, several researchers have reported favorable comparisons between the SSR technique and other SCC test techniques for aluminum alloys. As has been reported for other alloy systems, the optimum strain rate for SSR testing was found to be a function of the specific alloy. Thus one must exercise care in comparing SCC susceptibility for different alloys at a given strain rate.

Two articles by Buhl [38,39] were the only references in which the validity of the SSR technique was questioned for application to aluminum alloys. He performed SSR tests on three aluminum alloys in aqueous NaC1 under potentiostatic control and found a ranking of SCC resistance of 7075 > 2014 > 2024 where alloy 7075 was the most resistant. On the other hand, Brown [40] performed constant load and constant strain SCC tests and found Copyright by ASTM Int'l (all rights reserved); Sat Dec 19 20:05:50 EST 2015

Downloaded/printed by

University of Washington (University of Washington) pursuant to License Agreement. No further reproductions authorized.

The potentials plotted in Fig. 2 are the peak (maximum) free corrosion potentials measured

DEWES ET AL. ON MEASUREMENT OF IN-REACTOR DEFORMABILITY 93

DIETZEL AND SCHWALBE ON RISING DISPLACEMENT TEST 145