Parr Stress Crack Corrosion Info

7/22/2019 Parr Stress Crack Corrosion Info

1/8

ST IN LESS STEEL 408ALKAllES

Hyd. oridehydjoxlde solutions at eompsratively low temperatures andbra quite noncorrosive to both the chromium and chromlum-

loya. Higher and inconsistent corrosion rates ocour in more con-solutions and are somewhat accelerated under applied pressure2.corrosion rates for several stainless steels are shown in Table

Potassium hydroxide eolutiona would be expected to show similarstainless steela.osion graph in Figure 16,Q summarizes the performance of

austenitic chromium-nickel atsinless ateels in aodium hydroxide. The in-retie hold true prirqerily for Typeci 304 and 316. A atreaa-corrosion

zone baaed on faihrea reported in the literature is also shown inboundary of the cracking cone is shown aa a broken line,

ay not be completely defined. Moat of the crackingrred at temperaturea near the boiling point. Where

have been reported below the boiling point, there is a possibilitychloride may have been piesent in the aolutiona and con-

to the failures. Aeration and corrosion rates may also be factorsthe cracking.

TABLE 16.32. CORROSION r STAINLEM RTEELBBY Hmtou Hvnnoxwm SOLUTIONS~~

Type CW&C;lll . Tempcrdurc, FI Tcylw;hm, Corradon e,wv-_410 20 122 to 140 134 0.1430 20 122 to 140 134 ( 0.1309 20 122 to 140 134


2/8

MATERIALS SELECTION & DESIGN

Corrosion of stainless steel by hot causticResearch using solutions of

chemically pure caustic (sodiumhydroxide [NaOH]) led to thedevelopment of a diagram thatattempts to delineate the param-eters of concentration and tem-perature governing stresscorrosion cracking (SCC) of type300 series austenitic stainless steels(SS), such as types 304 (UNSS30400), 316 (S31600), and theirlow-carbon forms, types 304L(S30403) and 316L (S31603). Thisdiagram (Figure 1) is analogous toa similar diagram for causticembrittlement of carbon steel un-der stress from welding or cold-forming. The term causticembrittlement is a misnomer be-cause the phenomenon is simplySCC of steel in the alkalinesolution.

The 1 mpy (0.0254 mm/y)isocorrosion line in Figure 1 isnearly constant at 100C (212F)from -20 to 50 caustic. Thedashed line delineating SCC (la-beled Apparent SCC boundary) isU-shaped, with the minimum at-40 caustic and -240F (115C).

In actuality, there is a realpossibility that type 300 series SSmay lose passivity and undergorapid general corrosion in hot40 to 50 caustic. Probable safelimits are well below those indi-cated by the diagram, perhaps70C (158F) for 50 caustic and80C (177F) for 40 solution.However, oxidizing contami-nants can maintain passivity.Extra-low interstitial ferritic SSgrades, such as alloy 26-l(S44627), have been used insteadof Ni (UNS N02200) in causticevaporators provided the chlor-ate content is sufficiently high.

600 8 1 1 I &

500 - Stress cracking zone -

Apparent SCC boundary

0 20 40NaOH (wt )

60 80

FIGURE 1lsocorroslon diagram for type 300 series austenitlc SS in NaOH (caustic).

Above 300C (570F), the dan-ger of caustic SCC is very great.Bellows-type piping expansionjoints made of type 321 (S32100, Fe-18 Cr-10 Ni-Ti stabilized) in300-lb to 400-lb steam (-215C to230C [QOF to 445F]) are prone torapid SCC if there is entrainment ofcaustic from boiler treatment. Whenhigh-temperature caustic SCC of SS

is encountered, there is a charac-teristic gunmetal blueing of thesurface.

The role of chlorides in causticcracking is often misunderstood.Chlorides, if present, are not afactor. There have been failures ofalloys 800 (NOSSOO, Fe-30 Ni-20

64 MP/January 199


3/8

i

MATERIALS SELECTION & DESIGN

Phorgotten Phenomena


4/8

STRESS CORROSION CR CKING

Stress corrosion cracking is something which the de-engineer should always keep in mind in specifying

particularly in the case of pressure vessels.presence of internal stress should always be takenaccount when deciding the magnitude of exter-

tresses to which the equipment can besubjected. Unfortunately there are no guiding

be followed. About all that can bethat stress corrosion cracking is specific both formetal and for the environment. In certain special

the presence of tensile stresses may leadthe cracking of certain metals. With other environ-

or with other metals or alloys no difficulty isReliance must be placed on practical

or on laboratory or field tests.The term stress corrosion cracking is used to indicate

combined action of static tensile stress and corro-which leads to cracking. The principal factors are

magnitude of the stress, the nature of the environ-the length of time involved and the internal

of the alloy. These factors are not independ-but interact, one accelerating the action of another.relative importance varies with conditions.

If stress corrosion cracking is to occur there must besses at the surface. The stresses may be in-applied, the two types being additive. Ex-

of internal stresses are those produced by de-during cold work, by unequal cooling from

temperature and by internal structural rearrange-involving volume changes. Stresses inducedpiece is deformed, those induced by press and

fits and those near welds, rivets and bolts mayclassed as internal stresses.In many cases these concealed stresses are of greater

than actual operating stresses. This isof pressure vessels, except perhaps for thoseat loads which are high in relation to the

of the material. When the factor of safetyin design is considered, operating stresses

generally low enough to be of comparatively littlece, except as they add to the internal stresses.

The actual stresses may vary greatly from point towithin the metal, and in some locations are much

han the average value. A nonuniform stressis expected, nevertheless a high localized

considered more damaging than a uniformGenerally tensile stresses in the neighborhood ofyield strength are present in stress corrosion crack-failures, but failures are known which have occurred

much lower stresses. In any case the stress levelslow enough so that normally a great deal of general

could be tolerated. The interaction of thestess and corrosion produces cracking where it would

otherwise be expected.Stress corrosion cracking has been observed in almostmetal systems. Yet for each metal specific environ-

required to produce it. No stress corrosion

cracking has occurred in a vacuum. The environmentthat induces cracking frequently attacks the metal onlysuperficially if stresses are absent. Many of the en-vironments that cause cracking tend to produce a pit-ting type of corrosion.

One of the curious aspects of stress corrosion crackingis the wide difference in time required for failure, whichvaries from a matter of minutes to many years. Asso-ciated with this is the probability of cracking. Speci-mens which are apparently similar may not behave alike,with perhaps 40 cracking in a short time, and the restremaining untracked for a much longer time. Labora-tory tests require severe conditions to produce crack-ing in reasonable time, whereas in service much milderconditions may cause cracking in the longer time avail-able.

Considerable time may be required before corrosionproceeds to the extent that it begins to be acceleratedby the tensile stresses present. The more severe thecorrosive conditions and the higher the stress level thesooner this will happen. With some alloys there is anincubation period, during which precipitation or otherstructural changes may be occurring. For example,aluminum-magnesium alloys (over 6 magnesium),immediately after heat treatment, may not show anysusceptibility to cracking in accelerated laboratorytests, but after aging at room temperature for 6 months,stress corrosion cracks may form rapidly in the sametest.

As just indicated the internal structure of the metalor alloy can be of considerable importance. The in-ternal structure is dependent upon composition, uponthe method of fabrication and whether the metal is as-cast, hot worked or cold worked. It is also dependenton thermal treatments and the extent of natural aging.

There have been numerous reviews and books onstressc orrosion cracking.gv 20-24 There is also a volu-minous literature. This should be consulted for de-tailed information. The more important instances ofstress corrosion cracking are discussed below.

Caus t i c Emb r i t t l emen tA well-known example of st.ress corrosion cracking is

the caustic embrittlement of steel in steam boilers.25The cracking is associated with the presence of sodiumhydroxide in the boiler and hence the name caustic em-brittlement. The metal away from the cracks, how-ever, is ductile and not brittle. When a boiler lets gobecause of caustic embrittlement, results can be dis-astrous as illustrated in Fig. 1.26 Other examples ofsevere explosions have been cited by Zapffe.27

The cracking is said usually to be predominantlyintercrystalline, and Fig. 2 is an illustration of this.This photomicrograph was prepared from the steel ofan autoclave exposed to 507a caustic soda at 250and 400 psi. This is perhaps a more concentratedsolution than normally encountered in steam boilers,

.but it is an excellent example of caustic embrittlementjust the same. Oxides are present in the cracks, which


5/8

F i g . 1 Powe r p l a n t f ter exp l o s i on due to caus t i c em-b r i t t l emen t o f s t e am bo i l e +

F i g . 2 Ca u s t i c emb r i t t l emen t o f s t e el e x p os ed t o 5 0c a u s t i c s od a a t 2 5 0 F . 2 5 0 X

is typical. In other cases the cracks may be partlytranscrystalline, or even, as illustrated in Fig. 5, pre-dominantly transcrystalline. It used to be thoughtthat stress corrosion cracks were characteristicall--intergranular, but many cases are now known whei&the cracking follows a path across the grains.

In steam boilers the caustic present concentrates atsmall leaks or capillary spaces. The caustic concen-tration builds up to high values at such places. Hightensile stresses must be present, of course, and usuallythe cracking takes place along rows of rivets where theremay be slight leaks. Salt deposits have been observedin some cracked rivet seams.

In laboratory tests a U-bend specimen with a boltthrough the legs is convenient for studying caustic em-brittlement. Specimens of this nature are shown inFig. 3. The U-bend specimen has tensile stresses in theouter fibers in the neighborhood of the yield point, andhigh stress gradients such as frequently occur in prac-tice. Any cracking is located on the outside of thebend as shown in the figure. The white material inthe cracks is residual caustic. These steel specimenswere exposed to boiling 33oj, sodium hydroxide con-taining 0.1% lead oxide for 14 days.

U-bend specimens of this type were used by Berkand Waldeckz8 to outline dangerous concentrations andtemperatures of caustic. They obtained no cracking in30 days or longer at concentrations below 15y0 or above43% or at temperatures below 180 F. However, und Cilong-time service conditions, cracking has been ob-served well outside these limits29 Also it is known thatconstituents present in small amounts may act as ac-celerators or as inhibitors of the cracking.

In laboratory tests it is difficult to obtain consistentbehavior in pure caustic. Cracking is readily producedby adding certain oxidizing agents or accelerators.Thus cracking in boiling 33% sodium hydroxide can

F i g . 3 S t e el U -b en d s p ec ime n s a j -t e r e x p o su r e f o r 1 4 d a y s t o b o i l i n g 3 3 s o d i um h y d r o x i d e c o n t a i n i n g 0 . 1 l e a do x i d e . N a t u r a l si z e


6/8

4 I n t e r c r y s t a l l i n e c ra ck in g of s teel U-bend spec imens1 5 d a y s i n boiling 33q0 s o d i um h y d r o x i d e c o n t a i n i n g0 . 1 l e a d o x i d e . 5 0 0 X

consistently obtained by the addition of 0.1 leadide. Figure 4 il lustrates the intergranular nature of

racking in steel U-bend specimens exposed to thisdia. Cracking can also be consistently obtainedsimilar conditions by the addition of 0.3 so-il icate. Figure 5 shows that under these slightly

onditions the cracking was transcrystalline.es 4 and 5 are for the same steel and practicallycal conditions except for the change in minor

nstituents in the caustic. Other stress corrosionsystems are known where slight changes in

nditions have changed the path of the cracking.s fact is not explained by some of the theories ofess corrosion cracking.Under service conditions caustic embrittlement is

etimes avoided by adding inhibitors of the crackingthe water. Thus tannins, lignins, quebracho ex-ther additives have been beneficial in some

5 T r a n s c r y s t a l l i n e c r a c k i n g of s tee l U-bend spec i -a f t e r 1 2 d a y s in 33% s o d i um h y d r o x i d e containingsod i um s i l i c a te a t 300 F . Same s teel as F i g . 4 .

5 0 0 x

An embrittlement detector has been devised whichcan be attached directly to a pressure vessel to deter-mine whether or not the water is capable of producingcaustic embrittlement.30 The detector has a basewhich consists of a rectangular block with a hole throughwhich the water circulates. The test specimen isclamped to this base. I t is bent and maintained understress by means of an adjusting screw which passesthrough the specimen and presses against this baseblock. A small hole conducts the water from the baseblock to the contact surface between the base and thespecimen. By correct setting of the clamping nuts andadjusting screw a very slow leak of steam is estab-lished. Thus a concentrated solution forms under thebent area of the specimen. I f the water is embrittl ingand sufficient time is allowed, such as 30 days or more,the specimen wil l crack. I f the water is not embrittlingno cracking will occur.

Al l steels are not equally susceptible to caustic em-brittlement. Deoxidation practice and the residualelements present have some effect. However, the im-provements have been minor and no steel has been de-vised which is completely resistant. Low-alloy steelscrack as readily as plain mild stee1.z7s29

A welded construction is sometimes recommended asbeing superior to a riveted construction, the argumentbeing that this should prevent the concentration ofcaustic at leaks and capillary spaces. Welds are apt tointroduce high internal stresses, however, and weldedpressure vessels have failed by caustic embrittlement.31As-welded steel tanks are recommended for causticservice up to 140 F , but for higher temperatureswelded tanks must be stress relieved.*9 A low-tem-perature stress relief of welds has been proposed.32 Thisconsists of heating a narrow band each side of the weldto 350-400 F and quenching. This produces localizedstretching with a resultant decrease in stress. Ofcourse, where practical, a full stress relief anneal of theentire vessel is far safer.

Nickel cladding has been used successfully to preventcaustic embrittlement. Where the clad areas arejoined by welding care must be taken to insure soundwelds. The cracking illustrated in Fig. 2 was in anautoclave lined with nickel. The welds were defectiveand allowed the caustic to contact the steel with the re-sult that cracking of the steel occurred.

Lowering the tensile stresses present when possible isa standard means of preventing caustic embrittlement.This and other means of preventing stress corrosioncracking are considered further below.

S t r e s s Cor r o s i o n C ra ck i n g o f I r o nIn addition to caustic embrittlement, iron and steel

alloys are subject to rapid stress corrosion cracking insome nitrate solutions.33 Cracking has occurred inconcentrated calcium nitrate and concentrated am-monium nitrate. Highly stressed bridge cable wirecracked in dilute ammonium nitrate and in dilute so-

7


7/8

F i g . 6 T r a n s c r y st a l l i n e c r a c k i n g o tee l U-bend spec i -me n s e x p o se d f o r 6 d a y s t o w a t e r s a t u r a t e d w i t h a 5 0 1 5 0m i x t u r e o f h y d r o g en s u l f i d e a n d ca r b o n d i o x i d e a t 1 0 0 F .

5 0 0 xF i g . 8 S t r e ss co r r o s i o n c r a c k i n g o f s t a i n l e s s s t e el k e t t l eh a n d l i n g a b a k e d b ea n s a u c e h i g h i n c h l o r i d e s . N a t u r a l

s ize

S ta i n l e s s S tee lFigure 7 shows an instance of stress

corrosion cracking of Type 302 stainlesssteel. Disk-shaped specimens were im-mersed for 7 days in a solution contain-ing 12 hydrofluoric acid and 0.2fluosilicic acid at 182 F. Cracksformed around stenciled identificationmarks causing these sections of tb-specimens to fall away completer&Cracks can also be seen at the machinededges.

F i g . 7 S t r e s s c o r r o s i o n c r a c k i n g o f s t a i n l e s s s t e el a r o u n d s t e n c i l e d i d e n t i j i c a -t i o n ma r k s a n d a t ma c h i n e d e d g es . T h e s p ec ime n s w e r e e x p o se d i n 1 2 h y d r o -f l u o r i c a c i d c o n t a i n i n g O .2 ~ e jl u o s i l i c i c a c i d . N a t u r a l s i r e

Figure 8 shows another instance ofstress-corrosion cracking, this time ofa Type 304 stainless steel kettle han-dling a baked bean sauce which washigh in chlorides. The cracks hada radiating pattern and tended

dium nitrate, but not in distilled water, dilute am-monium sulfate, ammonium nitrite or sodium hy-droxide.34 Cracking has occurred in concentrated ni-tric acid and also in dilute nitric acid containing manga-nese dichloride as an accelerator.35 All these failureswere largely intergranular.

Transcrystalline cracking has been observed in tanksholding certain gases under pressure. This has beenattributed to moisture and traces of hydrogen cya-nide,3a and to moisture and traces of hydrogen sulfide.3Figure 6 shows the transcrystalline cracking of steelU-bend specimens exposed for 6 days to water at 100F saturated with a SO:50 mixture of hydrogen sulfideand carbon dioxide.38 Embrittlement by hydrogenmay be involved in this cracking. Cracking of similarnature has occurred in high-pressure gas condensatewells containing hydrogen sulfide and carbon dioxide.3gAmmonium thiocyanate seemed important in the stresscorrosion cracking of steel gas mains.*O Undoubtedlyadditional corrosives which cause stress corrosioncracking of steel will come to light. This should be con-sidered in exposures involving new chemicals.

to line up with small sharp pits.Stress corrosion cracks in stainless steel are usually

transcrystalline. A typical example is shown in Fig. 9.These cracks occurred in a specimen exposed to a boil-ing calcium-magnesium chloride brine. Intergranularcracking has been observed, but only when the heat

F i g . 9 T r a n s c r y st a l l i n e n a t u r e o f t h e s t r e ss co r r o s i o nc r a c k i n g o f s t a i n l e s s s t e el . 1 0 0 X

8


8/8

treatment has been such as to make the stainless steelsusceptible to general intergranular corrosion.

The stresses required to cause cracking in the stain-less steels sometimes seem comparatively low, a stressof less than 20,000 psi being reported sufficient by Scheilin one instance.23 Some Type 347 tubing was foundto be susceptible to cracking in boiling calcium mag-nesium chloride brine after stress-relief annealing from1350 F, but resistant to cracking after a slow cool from1600 F. Slow cooling is essential as quenchingstresses may be sufficient to cause cracking.

Ferritic stainless steels are less susceptible to stresscorrosion cracking than the austenitic alloys. Gener-ally, cracking cannot be avoided by shifting from oneaustenitic grade to another, although there may bereal differences in behavior, with increased corrosionresistance and increased austenitic stability beinghelpfu14i

A thorough review of environments causing stresscorrosion cracking of stainless steels was prepared re-cently by Nathorst.42 The number of environmentsseems large, but in spite of this the austenitic chromiumnickel stainless steels perform satisfactorily under manyservice conditions.

Most cases of cracking involve the presence of chlo-ride ions, particularly if the solution is acid. Hotconcentrated solutions of chlorides of magnesium, cal-cium, barium, cobalt, zinc, lithium, ammonium and

all cause rapid cracking.43 Cooler or moredilute solutions may be satisfactory. Organic chlo-rides, such as ethyl chloride, which decompose in thepresence of moisture to form hydrochloric acid maycause cracking.44

In many cases where some other environment wasthought to cause cracking, closer investigation hasshown that chlorides were actually present, even if onlyas an impurity. Thus cracking has been reported insulfite waste liquors of the cellulose industry, but inalmost every case brackish waters were used. Likewisecracking has been reported in steam and hot water, butin such cases it appears that the design is such as to allowlocal concentrations of chlorides.

Stainless steels are susceptible to stress corrosioncracking in hot concentrated caustic solutions. Thepresence of sodium sulfide and reducing substances issaid to increase the danger of cracking in caustic. Reeshas reported cracking in moist hydrogen sulfide.24Cracking has also been reported in some other environ-ments.42

cracking, but oxygen and carbon dioxide have a con-tributing effect. Cracking is fast in contaminatedatmospheres, but has occurred under apparently normaloutdoor and indoor conditions. Evans has speculatedon the role of ammonia in promoting the cracking.24

Susceptibility to cracking increases with tensile stress.Stresses of 12,000 to 20,000 psi readily cause cracking,but cracking is rare with stresses below 12,000 psi.Susceptibility to cracking increases greatly with zinccontent. Alloys with 85 to 90 copper are practicallyimmune, and with 90 copper they are fairly free fromcracking. Two-phase brass compositions, such as 60copper, 40 zinc, are more susceptible than alloys withless zinc. Special brasses which contain other elementsbehave similarly to the straight zinc brasses. Thecracking is usually intergranular, but transcrystallinecracking has been reported, particularly in the betaphase of high zinc brasses.

As compared with the brasses, other commercialcopper alloys, as aluminum bronze, tin bronze, sili-con bronze and cupronickel show comparatively littletendency to season crack, although failures are known,and sometimes the failures may occur in other thanammoniacal atmospheres. For example, the ASMEBoiler Code cautions on the use of silicon bronze insteam above 212 F. Cook24 reported pure copper tobe immune to cracking for all practical purposes.Thompson and Tracy4* found most additions to copperto cause a rapid increase in susceptibility to cracking,but larger additions of the same elements caused thesusceptibility to decrease again.

An acid mercury salt solution is sometimes used asan inspection test to determine the susceptibility ofcopper alloys to stress corrosion cracking.2g, 4g--62Metallic mercury is liberated on the surface and pene-trates stressed metal intergranularly. This is a differ-ent type of attack than stress corrosion cracking, butresults are roughly comparable. However, crackingin service has been known to occur in material whichhas passed the mercury test.

A better but more difficult test involves exposure to agas phase containing ammonia, air, water vapor andcarbon dioxide.48+ 4g This relates directly to serviceconditions. For reproducible results the temperaturemust be controlled and also the composition of the gasphase. With this test there seems to be no thresholdstress below which cracking will not occur in time.This probably relates to the fact that some intergranu-lar attack occurred in the absence of stress. This testis particularly suited for experimental studies.

CoPerhaps the best known example of stress corrosion Pure aluminum is quite resistant to stress corrosion

cracking in copper alloys is the season cracking of cracking. On the other hand aluminum alloys con-brass,45-47 so-called because the cracks resemble those taining more than 12 zinc or more than 6 magne-in seasoned wood. Exposure to moist ammoniacal sium have cracked in such mild environments as theatmospheres is believed to be necessary to produce the atmosphere and tap water. 63

9

Documents

Parr Stress Crack Corrosion Info