Super-Ferritic Stainless Steels Rediscovered

184

INTRODUCTION

Recent worldwide demand has dri-ven nickel and molybdenum pricesto record high values. Alloys contai-ning significant amounts of nickeland molybdenum, such as the aus-tenitic and duplex grades, have ex-perienced significant price increasesand some spot shortages have resul-ted in some regions. Today's superaustenitic prices are more than twi-ce the value of late 2003. With lownickel content and reasonable mo-lybdenum content, super-ferriticstainless steels are now proving tobe the most cost effective.

Originally developed in 1970 by C.D. Schwartz, I.A.Franson, and R.J.Hodges of Allied Vacuum Metals, E-Brite 26-1 (S44627) was the firstcommercial super ferritic alloy1. Tominimize the detrimental effect ofcarbon and nitrogen, high puritymelting techniques were required.This was accomplished by combi-

S T A I N L E S S S T E E L W O R L D 2 0 0 5 © 2 0 0 5 K C I P U B L I S H I N G B V www.stainless-steel-world.net

Super-Ferritic Stainless SteelsRediscoveredAuthor: Daniel S. Janikowski,Plymouth Tube East Troy, USA

Keywords:High Performance StainlessSteel Tubing, Ferritic StainlessSteel, Super-ferritic stainlesssteel, Corrosion, Erosion,Vibration, stress corrosioncracking, Thermal conductivity

AbstractOriginally developed back in the early to mid-1970's, the current generation ofsuper-ferritic stainless steels have now returned to popularity. When they werefirst developed, the goal was to have an alternative to titanium grade 2 in appli-cations such as seawater and high chloride applications. At that time, titaniumwas in short supply, not unlike today. However, over the last 10 years, themajority of the seawater capable stainless steel literature has been focused onsuper-austenitic (6% and 7% Mo alloys) and super-duplex alloys. While theperformance of these alloys has been very good, today's material raw materialprices have driven the price of these alloys skyward. This has driven the redis-covery of the super-ferritic alloys. This paper traces usage in power plant con-densing applications and compares properties such as corrosion resistance,mechanical and physical properties for many of the seawater resistant grades.

P 5 1 1 4

ning vacuum induction meltingwith EBM or ESR. A few years later,M. A Streicher at DuPont developed29Cr-4Mo2 (S44700). Although the-se grades performed well in highchloride environments, the highcost of the double melting techni-que restricted the alloys to only afew applications.

The newer generation super-ferriticalloys were developed soon after. Toreduce the manufacturing cost, acombination AOD refining and Nband Ti stabilization eliminated thedetrimental effect of the residualcarbon and nitrogen content. R.Oppenheim and J. Lennartz atDeutsche Edelstahlwekes3 are belie-ved to have used this process with28Cr-2Mo in 1974. Monit®, 26Cr-4Mo-4Ni (S44635) was developedsoon afterward by Nyby-Uddeholm4, followed by AL29-4C®

(S44735) by Allegheny Ludlum. Themost commercially successful of thegroup, SEA-CURE® (S44660), wasdeveloped by K.E. Pinnow of Cruci-ble Research in 19775. Over 20,000,000 meters of tubinghas been shipped of this grade since1980. The chemistry of the earlyand current commercialized super-ferritic grades is summarized in Table 1.

One industry that has stronglyadopted high performance stainlesssteels is power production. Kovach6

has summarized the history andperformance of high performancestainless steel use in power plantcondensers through the late 1990's.The meters of condenser tubingshipped in each year is documentedin Figure 1 separated by stainlessgroup. Most of the early applica-tions were dominated by austenitics

Table 1: Typical Chemical Composition of Super-Ferritic Alloys

P5114.qxp 1-11-2005 13:18 Pagina 178

185

that included alloys such as AL6X®

and 254SMO®. Between 1980 and1985, applications of super-ferriticsmultiplied. Use in the United Sta-tes, Europe, and Japan was com-mon. The cumulative use of highperformance stainless steels for po-wer plant condensers is summarizedin Figure 2 by type: austenitic, ferri-tic, and duplex. The trends of highinitial austenitic use, followed thespurt of ferritic use. After the mid1980's growth rates of both austeni-tics and ferritics declined, probablybecause of the increased availabilityof titanium grade 2. However, theuse of ferritics declined significantlymore to the point where they wereonly being used in a few select loca-tions, predominately in the US.One additional limitation may havebeen the lack of availability of iden-tical tube sheet materials as the su-per-ferritic alloys have a thicknessrestriction due to low toughness inthick sections. In the late 1990's,the gradual price increases of thesuper-austenitic alloys started todrive the shift toward the super-fer-ritics. Since the year 2000, over 95%of the high performance stainlesssteel used in power plant conden-sing application has been super-fer-ritic based. This market alone hasaveraged over 500 metric tonnes ofsuper-ferritic alloy per year since2002. High performance duplexesnever grew in popularity for this ap-

plication. Until recently, technicaldifficulties prevented the cold rol-ling of these grades to the common0.5 to 0.7 mm thickness commonfor this application.

Since 2000, the use of super-ferriticstainless steel in other markets,such as the petrochemical industry,has also grown significantly. Twomajor projects exceeding 1,200,000meters selected S44660 to use forcooling gas and/or crude utilizingsea or brackish water. These includethe PDVSA collection towers in La-ke Maricaibo, Venezuela (one of the

most aggressive waters known), andthe U.S. government's Strategic Pe-troleum Reserve. In both cases, ex-tensive studies considered a numberof copper based, stainless steel ba-sed, nickel based, and titanium al-ternatives. Both studies determinedthat the super-ferritic alloy was themost cost-effective long-term choice.

PITTING AND CREVICE

CORROSION

The high performance stainlesssteels are commonly chosen for ap-plications where high chlorides,low pH, or high microbiological ac-tivity is present. Several alloyingelements, such as chromium, mo-lybdenum, and nitrogen, promotechloride resistance in this group ofalloys. Not all have the same effect.By investigating the impact of eachelement, Rockel7 developed a for-mula to determine the total stain-less steel resistance to chloride pit-ting as follows:

PREn = % Cr + 3.3 (% Mo) + 16 (N)

PREn represents the "Pitting Resi-stance Equivalent" number. Usingthis formula, various stainless steelscan be ranked based upon theirchemistry. In this formula, nitrogenis 16 times more effective and mo-lybdenum is 3.3 times more effecti-ve than chromium for chloride pit-

S T A I N L E S S S T E E L W O R L D 2 0 0 5 © 2 0 0 5 K C I P U B L I S H I N G B Vwww.stainless-steel-world.net

P 5 1 1 4

Figure 1: Installed High Performance Austenitic, Duplex, and Ferritic Condenser Tubing by Year

Figure 2: Cumulative High Performance Austenitic, Duplex, and Ferritic Stainless Steel

Installed in Condensing Applications.

P5114.qxp 1-11-2005 13:19 Pagina 179

186

ting resistance. The higher thePREn, the more chloride resistancean alloy will have. Additional workperformed using interlaboratory tes-ting reported in ASTM G 48-998

confirmed that the formula develo-ped by Rockel was realistic. In thistest, five alloys representing S31600 through nickel alloys wereexamined. These alloys showed thatthe multiplying effect of molybde-num is 3.04 and for nitrogen is12.67. It is interesting to note thatnickel, a very common stainlesssteel alloying element, has little orno effect on chloride pitting resi-stance. Kovach and Redmond9 refined thework of Rockel by evaluating a largedatabase of existing crevice corro-sion data and compared it to thePREn number. By plotting the rela-tionships between the PREn andthe G 48 method B critical crevicetemperature (CCT), they determi-ned that the relationship was also afunction of crystal structure. Thisrelationship is displayed in Figure 3.Three relatively parallel lines repre-sent each of the crystal structures.Ferritic stainless steels were foundto have the highest CCT for a parti-cular PREn, followed by the duplexgrade. The austenitic grades need

the greatest amount of chromium,molybdenum, and nitrogen to haveequivalent chloride resistance. One of the most common questionsasked is "What is the maximum ch-loride level that can be tolerated fora particular grade of stainless steel?"The answer varies considerably. Fac-tors include pH, temperature, pre-sence and type of crevices, and po-tential for active biological species.Tverberg and Blessman10, and Jani-kowski11 studied a number of am-bient temperature applications andfound that the relationship betweenchloride resistance and G-48 criticalpitting appears to be logarithmic.To easily use and understand the re-lationship of PREn, critical crevicetemperature and "safe" chloride levelas a function of stainless steel type,they added the maximum chloridelevels on the right side axis of theoriginal chart developed by Kovackand Redmond. This is presented onthe right hand axis of Figure 3. It isbased upon having a neutral pH,35° Centigrade flowing water (toprevent deposits from building andforming crevices) common in manyheat exchanging and condensingapplications. Once an alloy with aparticular chemistry is selected, thePREn can be calculated and then in-tersected with the appropriate slo-ped line. The suggested maximumchloride level can then be determi-ned by drawing a horizontal line tothe right axis. In general, if an alloyis being considered for brackish orseawater applications, it needs tohave a CCT above 25° Centigrade asmeasured by the G 48 method B test.When using this guide, additionalcaveats need to be considered. The-se are:1.The maximum acceptable chlori-

de level needs to be lowered if thetemperature is higher than 35°Centigrade.

2. If the pH is lower than 7, themaximum chloride level shouldbe lowered.

3.This guide is based upon having aclean surface. If deposits are allo-wed to form, the pH can be signi-ficantly lower under the deposits,and the chloride levels may bemuch higher than the bulk water.

This figure can be very useful forranking alloys. After a typical or mi-nimum chemistry is determined,the PREn can be calculated. Tocompare the corrosion resistance oftwo or more alloys, a line is drawnvertically from the calculated PREnfor each alloy to the appropriatesloped line for the structure. Thevertical line should stop at the bot-tom line for austenitics, such as TP304, TP 316, TP 317, 904L, S31254,and N08367. Duplex grades, such asS32304, S32003, S32205, andS32750, fall on the center line. Theferritics, such as S44660 andS44735, follow the top sloped line.From this intersection, a horizontalline should be drawn to the left axisto determine an estimated CCT. Ahigher CCT indicates more corro-sion resistance.

STRESS CORROSION CRACKING

Stainless steels are susceptible to afailure mechanism known as stresscorrosion cracking (SCC). For thisto occur, a combination of threefactors are needed: tensile stress, acorrodent known to depassivate thesurface, and a temperature above a"threshold" temperature. The stressis caused by a combination of fac-tors. These may include: residualstress, thermally induced stress, ser-


P 5 1 1 4

Figure 3: Relationship between G-48 crevice

corrosion, PREn, and acceptable chloride con-

tent of water. The right side axis is based

upon neutral pH, 35 degree C temperature, and

no films or crevices.

Figure 4: Relationship between breaking time of

nickel, chromium, and iron alloys wires stressed

and immersed in boiling magnesium chloride

solution.

P5114.qxp 1-11-2005 13:19 Pagina 180

187

vice applied stress (such as hoopstresses from the pressure inside thetube), and stress from other sources.Chlorides are the most common de-passivating corrodent for the stain-less steel alloys. Not all stainless steels are equallysusceptible to SCC. Copson12 deter-mined that a direct relationshipexists between the time to failureand the nickel content. As shownin Figure 4, a combination of timeand specific nickel concentrationsabove the curve failed, while thosebelow the curve did not. The stain-less steel nickel content with themost potential is 8%, which is thesame content of the workhorse ofthe industry, S30400. An alloy con-taining 11% nickel content, such asS31600, is still very susceptible ascan be seen by the slightly highertime to failure. Improvements in ti-me to failure come from selecting

an alloy with very low nickel, suchS43035, or significantly higher nic-kel, typically that above 25%. Con-trary to many beliefs, this curvedoes not appear to be affected by achange in the crystal structure!Crucible Research tested a group offerritic, duplex, and austenitic stain-less steels in a series of high tempe-rature, high pressure autoclave testsusing strip samples bent into a "U"shape placed in a solution contai-ning sodium chloride13. The resultsare presented in Table 2. The re-sults of this test mirrored the Cops-on results. The alloy containing 8%nickel failed in the least aggressiveenvironment. In this testing, onlyS43035, the alloy containing verylow nickel, escaped cracking.

MECHANICAL PROPERTIES

Mechanical properties of commonseawater heat exchanger candidates

are listed in Table 3. The copper al-loys generally have the loweststrength, hardness, and modulus ofelasticity. Because of this, these al-loys are normally used with thickerwalls than either the stainless steelsor titanium for most applications.The high performance stainlesssteels typically have higher mecha-nical properties than both the cop-per alloys and more conventionalstainlesses. They can be used inthinner walls than that traditional-ly considered. Many power plantcondensers are now being designedusing 0.5 and 0.55 mm thickness.Titanium tubing in this wall thick-ness range is also being used. Howe-ver, because of the very low modu-lus of elasticity, the designs may besignificantly different.

EROSION RESISTANCE

When fluid velocities exceed a levelthat the protective oxide can nolonger tolerate, then erosion-corro-sion results. In most cases, the ero-sion velocity is proportional to thehardness or tensile strength of thealloy. Maximum velocities that ha-ve been found to be limitations forthe various alloys are listed in Table 4. As can be seen, the super-ferritic stainless steels have excel-lent erosion resistance as comparedto many other candidates.

In some applications where highvelocity water droplet impact on tu-bing is possible, the erosion mecha-


P 5 1 1 4

Table 2: Cracking Results of Various Stainless Steels in High Temperature Solutions containing

Sodium Chloride

Table 3: Typical Mechanical Properties of Alloys Commonly Used in Seawater

P5114.qxp 1-11-2005 13:19 Pagina 181

188

nism may be somewhat different.In this case, the mechanism is rela-ted to resistance to minute impact.Eroded titanium grade 2 tubing fr-om water droplet impact driven byhigh velocity steam is shown in Figure 5. When the wet steam can-not be avoided, other alloys withmore erosion resistance need to beutililized. In North America andTaiwan, S44660 has been used, inJapan FS10 has solved the problem,and in Europe S44800, S31254, andS32654 have been utilized.

Tavist14 developed a test for compa-ring erosion resistance for this me-chanism using a variable speed pad-dle that is utilized for acceleratingthe water droplets. He confirmedthat the resistance is proportionalto the hardness of the alloy. Table 5 summarizes relative waterdroplet erosion resistance using tita-nium grade 2 as unity. High perfor-mance stainless steels show seventimes or greater droplet erosion resi-stance.

STIFFNESS & VIBRATION

RESISTANCE

Tubing vibration is a major concernin some applications. A number ofdifferent methods can be used todetermine safe spans for heatexchanger tubing materials. Eachmethod uses a number of assump-tions that may or may not be cor-rect for the specific application. Alt-hough the absolute value for safe

wall thickness or safe length may besignificantly different dependingupon the method selected, almostall methods generally concludewith a similar ranking when alloysare compared to each other.

One method that has been used asa basis for cross-flow steam drivenvibration in a condensing applica-tion is the one developed by Coit,et al.15, Using this, maximum sup-port plate spacing can be calculatedin a specific condenser comparingOD, wall, and grade of various al-loys. Coit developed the followingformulas:

L = 9.5 [( E I ) / p v2 D)]1/4

I = Pi / 64 ( D4 - ID4)

Where:E = Modulus of Elasticity (psi)I = Moment of Inertia (in4)p = Turbine Exhaust Density

(lb/ft3)v = Average Exhaust Steam

Velocity at Condenser Inlet

D = Tube Outside DiameterID = Tube Inside Diameter

It is clear from the formula, consi-dering the same OD and wall tube,the property that has the largest im-pact on vibration is the modulus ofelasticity. Higher modulus alloys arestiffer and have more vibration resi-stance. Using Coit's method, Table 6 dis-plays a calculated condenser mini-mum wall for different materialsusing the same steam flow, tubeOD, and 900 mm support spacing.For a given support spacing, alloyswith low modulus may require twi-ce the wall thickness as those with ahigher modulus to prevent the riskof vibration damage. Alternatively,if a heat exchanger is newly con-structed, the support plates need tobe significantly closer on the lowermodulus materials. Existingexchangers can be retubed with alower modulus material if staking isused. However, this can add signifi-cant additional cost and one shouldbe very careful of stake selection asthe reliability of stakes can vary sig-nificantly.

THERMAL CONDUCTIVITY

Overall heat transfer of a heatexchanger tube is a function notonly of the resistance to the tubewall material, but also of the ther-


P 5 1 1 4

Table 4: Maximum surface velocity of Seawater

Figure 5: Water droplet erosion on titanium grade 2 tubing caused by high velocity wet steam.

Table 5: Relative Water Droplet Erosion Resistance Based upon Tavist12 Test Data

P5114.qxp 1-11-2005 13:19 Pagina 182

189

mal barriers on both the OD and IDsurface. In support of the HeatExchanger Institute, Hefner16 as-sembled a heavily instrumentedcondensing heat exchanger so thatactual heat transfer rates that inclu-ded OD and ID surface resistancescould be accurately measured. Theresults of that study are presentedin Figure 6. The Admiralty brass tu-be exhibited the highest conducti-vity. Titanium grade 2 had the nextgreatest heat transfer, followed clo-sely by the super-ferritic stainlesssteel, S44660. S30400 thermal per-formance was approximately 5%below S44660, with the super-auste-nitic N08367 having the least ther-mal transfer of the materials testedin this study. The difference in thethermal transfer for each of the gra-des would be roughly equivalent tothe amount of additional surfacethat would be required to match agrade above it. Copper alloys form

significant patina on both OD andID surfaces. With time, this patinawill lower heat transfer. After thepatina develops, conductivity ofthis alloy would have been expec-ted to drop in the range of S30400.Only small changes occur with tita-nium and the stainless steels as theprotective oxides on these gradesare very thin and protective and donot change much with time.

LIMITATIONS OF SUPER-FERRITIC

STAINLESS STEELS

Although this group of materialshas a number of advantages, metal-lurgical restrictions prevent usage ofthese grades in some applications:

Toughness- The toughness of super-ferritic stainless steels drops signifi-cantly as the wall thickness increa-ses. S44735 is rarely used with wallthickness above 1.25 mm andS44660 is normally not used in sec-tions thicker than 2.11 mm. This li-mits the usage to heat exchangertubing and thin sheet applications.However, since the super ferriticstainless steels are galvanically simi-lar to the other high performancestainless steel, both super-austeniticand super-duplex tubesheets can beused with these alloys.

Hydrogen Embrittlement - Like tita-nium, super-ferritic stainless steelscan be embrittled when they en-counter nacent hydrogen. However,while titanium forms a stable inter-metallic compound, the hydrogendiffuses interstitially into the ferriticalloys. As the hydrogen does notform a second phase, the embrittle-ment is reversible once the source

of the monotomic hydrogen is re-moved.

High Temperature - Super-ferritics,like the duplex alloys, are also su-sceptible to a loss of ambient tem-perature ductility when exposed totemperatures between 315 and 600degrees Centigrade. The pheno-menom occurs most rapidly at 475degrees.

SUMMARY

The attractive mechanical proper-ties, high modulus of elasticity,high thermal conductivity, and mo-derate cost make super-ferritic stain-less steels cost effective alloys forheat exchanger and thin strip appli-cations where high chloride andacid resistance are needed. Thiscombination of properties has re-cently been recognized as the use ofthese alloys has grown dramaticallysince 1999.

REFERENCES

[1] C.D.Schwartz, I.A.Franson, R.J.Hodges,

Chemical Engineering, 77, April 20,

1970, Pages 164-167

[2] M.A.Streicher, Corrosion, 30, (3), 1974,

77-91

[3] R. Oppenheim, J. Lennartz, H.Laddach,

TEW-Techn. Ber., 2 (1), 1976, 3-13

[4] N. Pessall & J.I. Nurminen, Development of

Ferritic Stainless Steels for Use in Desali-

nation Plants, Corrosion, 30, (11), 1974,

381

[5] K.E. Pinnow, Progress in the Development

of High Chromium Ferritic Stainless

Steels Produced by AOD Refining, Stain-

less Steel 77, London, England, Septem-

ber 1977

[6] C. W. Kovach Report on Twenty-Five

Years Experience with High Performance

Stainless Steel Tubing in Power Plant

Condensers. International Joint Power

Conference, San Francisco, CA, July 25-

28, 1999

[7] M. Rockel, Use of Highly Alloyed Stainless

Steels and Nickel Alloys in The Chemical

Industry, ACHEMA Conf., Frankfurt,

Germany, 1978

[8] ASTM G48-99, Pitting and Crevice Corro-

sion Resistance of Stainless Steels and

Related Alloys by the Use of Ferric Chlo-

ride Solution. ASTM, West Conshohock-

en, PA, USA.

[9[ C.W. Kovach and J.D. Redmond, "Correla-


P 5 1 1 4

Table 6: Minimum wall thickness required to prevent vibration for a theoretical 900 mm span

support plate spacing using the Coit method vibration calculation method.

Figure 6: Overall Heat Transfer coefficient of vari-

ous materials in a heavily instrumented condenser.

P5114.qxp 1-11-2005 13:19 Pagina 183

190

tion Between the Critical Crevice Tem-

perature "Pre-Number", and Long-Term

Crevice Corrosion Data for Stainless

Steels," presented at the NACE Annual

Conference Corrosion 93, New Orleans,

LA (April 1993).

[10] Tverberg, J. and Blessman E., Superferritic

Stainless Steels For Steam Condensing

Service With High Chloride Cooling

Water, Properties and History,

IJPGC2002-2612, ASME conference,

Phoenix, AZ, USA, July 2002

[11] D. S. Janikowski, "Considerations in

Selecting Stainless Steel for Heat

Exchanger Applications in Power Gener-

ation", EPRI Conference, June 17-19,

2002.

[12] H.R. Copson. Physical Metallurgy of

Stress-Corrosion Fracture. New York:

Interscience, 1959, p. 247.

[13] Internal Research. Crucible Research Cen-

ter, Pittsburg, PA: 1987.

[14] Tavast, J.O. Steam Side Droplet Erosion in

Titanium Tubed Condensers - Experi-

ences and Remedies," ACOM. Schaum-

burg, IL: AvestaPolarit, Inc., April 1996.

[15] R.L. Coit, CC. Peake, and A. Lohmeier,

"Design and Manufacturing of Large

Surface Condensers - Problems and

Solutions," Volume XXVIII - Proceedings

of the American Power Conference.

1966, pp. 469-483.

[16] Hefner, R.J.. "Effect of Tube Material

SEACURE on Steam Condensation".

Rochester, NY: Rochester Institute of

Technology, July 1993.


P 5 1 1 4

P5114.qxp 1-11-2005 13:19 Pagina 184

Documents

Super-Ferritic Stainless Steels Rediscovered