Study of 13Cr-4Ni-(Mo) (F6NM) Steel Grade Heat Treatment for Maximum Hardness Control in Industrial Heats

metals

Article Study of 13Cr-4Ni-(Mo) (F6NM) Steel Grade Heat Treatment for Maximum Hardness Control in Industrial Heats

Massimo De Sanctis 1,*, Gianfranco Lovicu 2, Massimiliano Buccioni 2, Angelo Donato 2, Maria Richetta 3 and Alessandra Varone 3

1 Dipartimento di Ingegneria Civile e Industriale, Università di Pisa, Largo Lucio Lazzarino, 2, 56122 Pisa, Italy 2 Nuovo Pignone Tecnologie Srl, BHGE, Via Felice Matteucci, 2, 50127 Firenze, Italy; [email protected] (G.L.); [email protected] (M.B.); [email protected] (A.D.) 3 Dipartimento di Ingegneria Industriale, Università di Tor Vergata, Via del Politecnico, 1, 00133 Roma, Italy; [email protected] (M.R.); [email protected] (A.V.) * Correspondence: [email protected]; Tel.: +39-050-2217827

Received: 16 June 2017; Accepted: 31 August 2017; Published: 6 September 2017

Abstract: The standard NACE MR0175 (ISO 15156) requires a maximum hardness value of 23 HRC for 13Cr-4Ni-(Mo) steel grade for sour service, requiring a double tempering heat treatment at temperature in the range 648–691 ◦C for the first tempering and 593–621 ◦C for the second tempering. Difficulties in limiting alloy hardness after the tempering of forged mechanical components (F6NM) are often faced. Variables affecting the thermal behavior of 13Cr-4Ni-(Mo) during single and double tempering treatments have been studied by means of transmission electron microscopy (TEM) observations, X-ray diffraction measurements, dilatometry, and thermo-mechanical simulations. It has been found that relatively low Ac1 temperatures in this alloy induce the formation of austenite phase above 600 ◦C during tempering, and that the formed, reverted austenite tends to be unstable upon cooling, thus contributing to the increase of final hardness via transformation to virgin martensite. Therefore, it is necessary to increase the Ac1 temperature as much as possible to allow the tempering of martensite at the temperature range required by NACE without the detrimental formation of virgin martensite upon final cooling. Attempts to do so have been carried out by reducing both carbon (<0.02% C) and nitrogen (<100 ppm) levels. Results obtained herein show final hardness below NACE limits without an unacceptable loss of mechanical strength.

Keywords: supermartensitic stainless steels; tempering treatments; austenite reversion

1. Introduction Low-carbon supermartensitic stainless steel 13Cr-4Ni-(Mo) has substituted 12% Cr for several decades, and more recently, CA15 steels in many applications in the oil/gas industry [1,2]. The combination of low carbon content with the addition of 3.5–4.5% Ni suppresses the formation of δ-ferrite and promotes the formation of a mixed structure of α’ (martensite) and γ ( austenite) at room temperature, which possesses superior mechanical properties and corrosion resistances [3–5]. This alloy is proposed in two delivery conditions: casted (CA6NM) or forged (F6NM), and it finds wide applications for handling fluids containing CO2 and H2S. These environments increase the danger of sulfide stress corrosion cracking (SSCC); the NACE MR0175 standard limits 13Cr-4Ni alloys to 23 HRC maximum for sour service. However, in industrial practice, considerable difficulties are faced in lowering the hardness of 13Cr-4Ni-(Mo) below 23 HRC. These difficulties are of crucial interest, since tempering

Metals 2017, 7, 351; doi:10.3390/met7090351 www.mdpi.com/journal/metals Metals 2017, 7, 351 2 of 14 treatments are carried out on expensive mechanical components (e.g., compressor rotors) in their final state. Tempering should maximize the recovery of the martensitic matrix and promote the retention of soft retained austenite at room temperature. Previous studies on the thermal behavior of low-carbon 13Cr-4Ni alloy [2] and 16Cr-5Ni-(Mo) [6] showed that the amount of retained austenite increases with increasing reversion treatment temperature, exhibiting a peak typically in the range 620–640 ◦C. Above this range, it decreases with increasing temperatures, leading to the formation of virgin martensite upon cooling, with increasing hardness. It has been suggested that the stability of reversed austenite initially increases because of the diffusion of nickel and other gamma-stabilizing elements from the martensitic matrix towards reverted austenite islands [7–9]. Above a critical tempering temperature, a gradual loss of stability would occur because of several factors, among which the most important seems to be the solute redistribution for increasing volume fractions of reversed austenite [10]. The NACE MR0175 standard requires a double tempering treatment after quenching for F6NM, with a tempering temperature within 650–690 ◦C for the first tempering and 595–620 ◦C for the second one. An initial high-temperature treatment which is even higher than the Ac1 temperature of the alloy will favor softening of as-quenched martensite, and at the same time, the formation of reverted austenite, which can at least partially transform into virgin martensite upon cooling. The aim of the second stage of tempering is the recovery of the newly formed virgin martensite. Experimental diagrams are available for 13Cr4Ni0.5Mo [11], giving the final hardness and content of retained austenite at room temperature as a function of temperature and time of tempering treatments using a “Larson–Miller”-type parameter (P = T(K)[20 + log t(h)] × 10−3). For 13Cr4Ni0.5Mo after single heat treatment, most reverted austenite formed up to P = 17.7 is expected to remain stable while cooling to room temperature, and according to these data sets, hardness would attain a minimum of 250 HB. Double treatments are indicated as effective in reducing hardness around 220 HB—at least when the second tempering is carried out below P = 18.2. Nevertheless, from industrial experience and previous indications, there is a considerable scatter of results with related economic losses. Gooch [12] studied the effects of heat treatments for welded 13Cr4Ni steels, and agreed on the difficulty in limiting the alloy hardness through post-weld heat treatments. He concluded that the major compositional factor determining hardness was carbon, which should be limited below 0.03%. The objective of this work is to carry out a detailed study on microstructural transformations occurring during single and double temper treatments of 13Cr4Ni(Mo) steel, with the aim of clarifying the compositional and microstructural variables influencing the final hardness of this alloy and to give useful guidelines for industrial practice.

2. Materials and Methods Two industrial heats of 13Cr4Ni0.5Mo have been considered in this work, with chemical compositions reported in Table1.

Table 1. Chemical composition of the studied alloys.

Heats %C %Si %Mn %P %Cr %Mo %S %Ni N (ppm) %Creq/%Nieq * A 0.027 0.35 0.90 0.019 12.89 0.52 0.001 3.84 180 2.45 B 0.030 0.44 0.73 0.018 12.72 0.55 0.001 3.85 280 2.35

* %Nieq and %Creq calculated following the DeLong formula.

The compositions are very close, with an almost coincidental Creq/Nieq ratio. The only signiﬁcant differences are the higher carbon (0.03% C) and nitrogen (280 ppm) contents in Heat B than in Heat A (0.027% C, 180 ppm N). These industrial heats were processed through the following treatments:

- Homogenizing at 1473 K (1200 ◦C) for 4 h, followed by preliminary hot forging (HR); - Intermediate stress relief at 700 ◦C; Metals 2017, 7, 351 3 of 14

These industrial heats were processed through the following treatments: Metals 2017, 7, 351 3 of 14 - Homogenizing at 1473 K (1200 °C) for 4 h, followed by preliminary hot forging (HR); - Intermediate stress relief at 700 °C; ◦ - - ForgingForging (F(F samples)samples) atat 14731473 KK (1200(1200 °C);C); - - AustenitizingAustenitizing 44 hh atat 10201020 ◦°CC followedfollowed byby oiloil quenchingquenching (OQ)(OQ) toto roomroom temperature.temperature.

TheThe Ac Ac11and and MsMs criticalcritical temperaturestemperatureswere weredetermined determined by by dilatometric dilatometric analysis analysis using using a a Gleeble Gleeble 38003800 thermo-mechanical thermo-mechanical simulatorsimulator (Dynamic(Dynamic SystemsSystems Inc.,Inc., Poestenkill,Poestenkill, NY,NY, USA),USA), equipped equipped withwith a a ISO-QISO-Q Quenching Quenching and and Deformation Deformation Dilatometer. Dilatometer. The The samples samples were were heated heated at at 1020 1020◦ C,°C, with with a a heating heating raterate of of 300 300◦ C/h,°C/h, heldheld forfor 11 hh atat thethe targettarget temperature,temperature, andand thenthencooled cooledto to room room temperature temperature using using differentdifferent cooling cooling rates rates of of 100 100 and and 3000 3000◦ C/h.°C/h. VolumeVolume variationsvariations ofof samplessamples wereweremeasured measured by by a a laser laser dilatometrydilatometry system system [ 13[13].]. Figure Figure1 shows1 shows the the geometry geomet andry and dimensions dimensions of theof the samples samples used. used. Details Details of theof the experimental experimental technique technique are are reported reported elsewhere elsewhere [14]. [14].

FigureFigure 1. Geometry 1. Geometry of the of sample the sample used for used dilatometric for dilatometric analysis (unit:analysis. mm).

® TheThe GleebleGleeble® 38003800 simulatorsimulator hashas alsoalso beenbeen usedused toto reproducereproduce singlesingle andand doubledouble temperingtempering treatmentstreatments in in the the temperature temperature range range 610–700 610–700◦ C°C using using heating heating rates rates of of 300 300◦ C/h°C/h withwith holdingholding times times of of 11 h h followed followed by airby cooling.air cooling. Treated Treated samples samples have been have used been to determineused to determine their mechanical their mechanical properties. properties. The volume fraction of retained austenite at room temperature (γr) was determined by The volume fraction of retained austenite at room temperature (γr) was determined by X-ray diffraction (Philips,X-ray diffraction Eindhoven, (Philips, The Netherlands). Eindhoven, XRD The spectraNetherlands). were collected XRD spectra in the 2θwereangular collected range in 5◦ theto 65 2◦θ byangular using range the Mo-K 5° toα 65°radiation by using (λ =the 0.71 Mo-K Å)α in radiation step scanning (λ = 0.71 mode Å) within step steps scanning of 0.05 mode◦ and with counting steps timeof 0.05° of 5 and s per counting step. The time relative of 5 s amountsper step. The of martensite relative amounts and retained of martensite austenite and were retained calculated austenite for eachwere sample calculated from for the each integrated sample intensities from the of integrated the reflections intensities of both phases,of the reflections following theof both ASTM phases, E975 standardfollowing [15 the]. ASTM E975 standard [15]. ® MechanicalMechanical properties properties were were measured measured by meansby means of tensile of tensile tests ontests a Galdabini on a Galdabini® SUN 1000 SUN tensile 1000 machinetensile machine (Cardano (Cardano al Campo al (VA), Campo Italy). (VA), Tensile Italy). specimens Tensile specimens which were which 1 mm were in thickness, 1 mm in 12.5 thickness, mm in width,12.5 mm and in with width, a gage and length with ofa gage 50 mm, length heat of treated 50 mm, by heat Gleeble treated 3800 by as describedGleeble 3800 above as described have been above used. Testshave were been performed used. Tests using were a performed crosshead displacementusing a crosshead rate ofdisplacement 1 mm/s, and rate the of gage 1 mm/s, length and elongation the gage waslength measured elongation using was an measured extensometer. using To an compare extensomet theer. quantity To compare of austenite the quantity in the microstructureof austenite in the of samplesmicrostructure subjected of to differentsamples quenchingsubjected andto temperingdifferent treatments,quenching theand instantaneous tempering strain-hardeningtreatments, the coefficientinstantaneous (n) was strain-hardening evaluated. It is coefficient defined as (n follows:) was evaluated. It is defined as follows: d ()lnσ n =d(ln σ) n = ()ε (1)(1) dd(lnlnε) wherewhereσ σand andε εare are true true stress stress and and true true strain, strain, respectively. respectively. TheThe presence presence of theof austenitic the austenitic phase—with phase—wi its associatedth its high associated work hardening high behavior—induceswork hardening highbehavior—inducesn values and pronounced high n values plateau andin pronounced the n vs. true plateau strain in curves, the n dependingvs. true strain on itscurves, volume depending fraction andon stability.its volume Moreover, fraction itand is also stability. responsible Moreover, for the it observed is also markedresponsible serration for the in theobserved instantaneous marked strainserration hardening in the instantaneous curves [16]. strain hardening curves [16]. MicrostructuralMicrostructural studies studies have have been been carried carried out out by by transmission transmission electron electron microscopy microscopy (TEM) (TEM) using using ® aa Philips Philips® CM12CM12 microscopemicroscope (Philips Electron Optics, Eindhoven, The The Netherla Netherlands)nds) equipped equipped with with a ® aBruker Bruker® QuantaxQuantax TEM TEM 200T 200T EDX EDX system system (Bruker, (Bruker, Berlin,Berlin, Germany). Thin foilsfoils werewere preparedprepared byby mechanical grinding followed by double-jet electropolishing in a solution of 10% perchloric acid and 90% butoxyethanol at 258 K (−15 ◦C) and 9 V. Metals 2017, 7, 351 4 of 14

Metals 2017, 7, 351 4 of 14 Alloy hardness was measured using Rockwell (Hardness tester, C.I.S.A.M. sas, Induno Olona, Italy) testsmechanical from industrialgrinding followed byproducts by double-jet and Vickers electropo (Micro-hardnesslishing in a solution tester, of 10% Shimadzu perchloric Corporation, acid and Kyoto,90% Japan) butoxyethanol tests from at 258 1 mm-thick K (−15 °C) and heat-treated 9 V. samples used in the Gleeble 3800 simulator. The VickersAlloy hardness hardness scale was is measured frequently using used Rockwell for these (Hardness heats for tester, weld C.I.S.A.M. procedure sas, qualiﬁcation Induno Olona, tests, and accordingItaly) tests to from the industrial BS860 1967 byproducts standard, and a Vickers 23 HRC (Micro-hardness limit wouldbe tester, equivalent Shimadzu to Corporation, 253 HV. It has Kyoto, Japan) tests from 1 mm-thick heat-treated samples used in the Gleeble 3800 simulator. The been shown that this conversion—as well as the ASTM E140-88 correlation—is not applicable to Vickers hardness scale is frequently used for these heats for weld procedure qualification tests, and 13Cr-4Ni-(Mo) materials, with several researchers indicating 23 HRC as equivalent to 275 HV [17,18]. according to the BS860 1967 standard, a 23 HRC limit would be equivalent to 253 HV. It has been Therefore,shown this that hardness this conversion—as equivalence willwell be as used the hereinafter.ASTM E140-88 correlation—is not applicable to 13Cr-4Ni-(Mo) materials, with several researchers indicating 23 HRC as equivalent to 275 HV 3. Results [17,18]. Therefore, this hardness equivalence will be used hereinafter. Table2 compares the hardness values of preliminary hot forged (HR) and as-quenched (after 3. Results forging) (OQ) samples from different heats. In spite of their similar chemical compositions, hot rolled samples areTable very 2 differentcompares in the hardness, hardness withvalues Heat of preliminary A (31.1 ± 1.0hot HRC)forged being (HR) signiﬁcantlyand as-quenched harder (after than Heat Bforging) (25.1 ± (OQ)0.8 HRC). samples from different heats. In spite of their similar chemical compositions, hot rolled samples are very different in hardness, with Heat A (31.1 ± 1.0 HRC) being significantly harderTable than 2. HeatHardness B (25.1 of ± Heats 0.8 HRC). A and B in hot rolled (HR) and oil quenched (OQ) conditions.

Table 2. Hardness of Heats A and B in hot rolledHardness (HR) and (HRC) oil quenched (OQ) conditions. Heat Hardness (HRC) Heat HR-Samples As-Quenched Samples HR-Samples As-Quenched Samples A 31.1 ± 1.0 35.3 ± 1.3 BA 31.1 25.1 ±± 1.00.8 35.3 ± 1.3 34.2 ± 1.2 B 25.1 ± 0.8 34.2 ± 1.2

The hardnessThe hardness difference difference was was partially partially recovered recovered in in OQ OQsamples samples where Heat Heat B B(34.2 (34.2 ± 1.2± 1.2HRC) HRC) showedshowed about about the same the same hardness hardness as Heat as Heat A (35.3A (35.3± ±1.3 1.3HRC). HRC). Considering that that the the hardness hardness value value is is also affectedalso affected by the by amount the amount of retained of retained austenite, austenit thee, Ms–Mf the Ms–Mf range range for Heatfor Heat A is A expected is expected to be to higherbe in temperaturehigher in temperature than for Heat than B. for Heat B. The Ms and Ac1 critical temperatures have been determined through dilatometric analysis, and The Ms and Ac1 critical temperatures have been determined through dilatometric analysis, and relevant experimental curves for Heat A and Heat B are reported in Figure 2. relevant experimental curves for Heat A and Heat B are reported in Figure2. The experimental Ac1 temperatures for Heats A and B were calculated from the dilatometric Thecurves experimental and by analyzing Ac1 temperatures their derivatives. for Heats Figure A 3 andshows B werethe derivative calculated curves from in the the dilatometric range of curvestemperature and by analyzing close to theirthe transformation derivatives. Figurepoint. 3Three shows different the derivative curves have curves been in considered, the range of temperatureperformed close at to different the transformation cooling rates point. (two samples Three different at 100 °C/h curves and haveone at been 3000 considered, °C/h), but with performed the ◦ ◦ at differentsame coolingheating ratesrate (300 (two °C/h). samples For at the 100 sameC/h Heat, and onethere at is 3000 a goodC/h), agreement but with among the same the heatingAc1 ◦ rate (300temperaturesC/h). For determined the same Heat, from theredifferent is a sa goodmples. agreement The uncertainty among of the calculated Ac1 temperatures Ac1 was determined determined from differentusing the samples.temperature The range uncertainty in which it of is calculatedpossible to verify Ac1 was the slope determined change of using dilatometric the temperature curves range inand which their derivatives. it is possible to verify the slope change of dilatometric curves and their derivatives.

(a)

Figure 2. Cont. Metals 2017, 7, 351 5 of 14

Metals 2017, 7, 351 5 of 14 Metals 2017, 7, 351 5 of 14

(b)) Figure 2. Dilatometric curves of (a) Heat A and (b) Heat B. FigureFigure 2. 2.Dilatometric Dilatometric curves curves of of ( a(a)) Heat Heat A A and and ( b(b)) Heat Heat B. B.

Figure 3. Derivatives of dilatometric curves of Heat A and Heat B samples. Different curves have been shifted to avoid overlapping and to allow a better comparison among them.

Ac1 temperatures for Heat A and B were around 620 ± 10 and 600 ± 10 °C, respectively. These Figure 3. Derivatives of dilatometric curves of Heat A and Heat B samples. Different curves have Figureresults 3. Derivatives could be compared of dilatometric to the curves calculated of Heat Ac A1 temperature and Heat B samples.for 13% DifferentCr steels curveshaving have carbon been shiftedbeencontents shifted to avoid lower to overlappingavoid than overlapping0.05% and[19], toaccording and allow to aallow betterto the a following comparisonbetter comparison formula: among among them. them. ° = − + − − + + + − Ac 1( C) 850 1500(C N) 50Ni 25Mn 25Si 25Mo 20(Cr 10) (2) Ac1 temperatures for Heat A and B were around 620 ± 10 and 600 ± 10◦ °C, respectively. These Ac1 temperatures for Heat A and B were around 620 ± 10 and 600 ± 10 C, respectively. These results couldThe be predicted compared Ac1 temperature to the calculated for Heats AcA and1 temperature B were 647 andfor 63313% °C, Cr respectively. steels having These carbon results could be compared to the calculated Ac1 temperature for 13% Cr steels having carbon contents contentsvalues lower are than higher 0.05% than [19], the according experimental to thetemper followingature, although formula: it is confirmed that the Ac1 lower thantemperature 0.05% [19 of], Heat according A is higher to thewith following respect to Heat formula: B. ° = − + − − + + + − The MsAc temperature1( C) 850 for1500(C Heat B (250N) ± 1050Ni °C) is lower25Mn than25Si Heat A25Mo (270 ± 1520(Cr °C), thus10) supporting (2) ◦ the propensityAc1( C) =for850 this −alloy1500 to (retainC + N higher) − 50Ni volume− 25Mnfractions+ of25Si austenite+ 25Mo in HR+ 20 and(Cr OQ− products.10) (2) The predictedFigure 4 Acshows1 temperature electron micrographs for Heats fromA and OQ B sampleswere 647 of andHeat 633A and°C, respectively.Heat B. The These ◦ valuesThe microstructuresare predicted higher Ac than1 consisttemperature the of experimentalan α for’ martensite Heats Atemper matrix, and B ature,in were which 647although fine and and 633 highlyit C,is respectively. confirmeddislocated laths Thesethat are the values Ac1 distributed parallel to one another. The presence of retained austenite is not evident in OQ state, aretemperature higher than of theHeat experimental A is higher with temperature, respect to although Heat B. it is confirmed that the Ac1 temperature of being scarcely visible only in Heat B in the form of thin interlath films (Figure 4b). Heat AThe is higherMs temperature with respect for toHeat Heat B (250 B. ± 10 °C) is lower than Heat A (270 ± 15 °C), thus supporting Figure 5 shows stress–strain curves and Figure 6 yield strength (YS) and ultimate tensile the propensity for this alloy to retain higher volu◦ me fractions of austenite in HR◦ and OQ products. Thestrength Ms temperature (UTS) values for obtained Heat B on (250 OQ± samples10 C) heat is lower treated than (tempered) Heat A by (270 Gleeble± 15 3800C), for thus one supportinghour the propensityFigureat different 4 forshows temperatures this alloyelectron to in retain the micrographs temperature higher volume range from 610–700 fractions OQ °C.samples of austenite of Heat in HR A and and OQ Heat products. B. The microstructuresFigure4 shows consist electron of micrographsan α’ martensite from OQmatrix, samples in which of Heat fine A andand Heat highly B. The dislocated microstructures laths are consistdistributed of an αparallel’ martensite to one matrix, another. in whichThe presence fine and of highly retained dislocated austenite laths is are not distributed evident in parallel OQ state, to onebeing another. scarcely The visible presence only of in retained Heat B in austenite the form is of not thin evident interlath in OQ films state, (Figure being 4b). scarcely visible only in HeatFigure B in the5 shows form of stress–strain thin interlath curves films (Figureand Figure4b). 6 yield strength (YS) and ultimate tensile strengthFigure (UTS)5 shows values stress–strain obtained curveson OQ andsamples Figure heat6 yield treated strength (tempered) (YS) and by ultimateGleeble 3800 tensile for strength one hour (UTS)at different values temperatures obtained on OQ in samplesthe temperature heat treated range (tempered) 610–700 °C. by Gleeble 3800 for one hour at different temperatures in the temperature range 610–700 ◦C.

Metals 2017, 7, 351 6 of 14 Metals 2017, 7, 351 6 of 14 Metals 2017, 7, 351 6 of 14

Figure 4. TEM electron micrographs of (a) Heat A and (b) Heat B after oil quenching (OQ samples). Figure 4. TEM electron micrographs ofof ((aa)) HeatHeat AA andand ((bb)) HeatHeat BB afterafter oiloil quenchingquenching (OQ(OQ samples).samples).

(a) (a)

(b) (b) Figure 5. Experimental stress–strain curves of (a) Heat A and (b) Heat B after single tempering at FigureFigure 5.5. ExperimentalExperimental stress–strain curves of (a) HeatHeat AA andand ((bb)) HeatHeat BB afterafter singlesingle temperingtempering atat different temperatures. differentdifferent temperatures.temperatures.

Metals 2017 7 Metals, , 2017 351 , 7, 351 7 of 714 of 14

Metals 2017, 7, 351 7 of 14

FigureFigure 6. Yield 6. (YS)Yield and (YS) tensileand tensile (ultimate (ultimate tensile tensile strength, strength UTS), UTS)strengths strengths of of Heat Heat A A (red) (red) and and Heat Heat B B (black)(black) after single after single tempering tempering at different at different temperatures temperatures (610–700 (610–700◦ °C).C).

FigureThe Vickers6. Yield (YS) hardness and tensile (HV10) (ultimate measurements tensile strength of thes, UTS)e samples strengths are of Heatshown A (red) in Figure and Heat 7. It B can be Theobserved Vickers(black) afterthat hardness singleHeat temperingB remained (HV10) at measurements differentconstantly temperatures harder of than these (610–700 the samples limit °C). of are275 shownHV (23 inHRC), Figure whereas7. It can the be observedhardness that Heat reduction B remained of Heat A constantly allowed values harder close than to the the limit. limit Moreover, of 275 HV for both (23 HRC), heats there whereas was a the The Vickers hardness (HV10) measurements of these samples are shown in Figure 7. It can be hardnessslight reduction decrease ofof Heatstrengths A allowed and hardness values when close increasing to the limit. the tempering Moreover, temperature for both heats from there 610 to was observed that Heat B remained constantly harder than the limit of 275 HV (23 HRC), whereas the a slightaround decrease 650 of°C, strengths followed by and a step hardness increase when for higher increasing temperatures. the tempering temperature from 610 to hardness reduction of Heat A allowed values close to the limit. Moreover, for both heats there was a around 650Figure◦C, followed 8 shows by volume a step fractions increase of for retained higher austenite temperatures. measured in Heat A and Heat B after slight decrease of strengths and hardness when increasing the tempering temperature from 610 to Figuresingle8 tempering.shows volume The austenite fractions content of retained of both austenite heats was measured relatively inlow, Heat with A the and values Heat consistently B after single aroundbelow 650 10%. °C, For followed Heat B bythe a peak step wasincrease reached for higher at 630 temperatures.°C, whereas for Heat A it was around 640 °C. tempering.Figure The austenite8 shows volume content fractions of both of heats retained was austenite relatively measured low, with in the Heat values A and consistently Heat B after below Strength and hardness were reduced ◦in the presence of retained austenite, but they◦ showed a 10%.single For Heat tempering. B the peak The austenite was reached content at of 630 bothC, heat whereass was relatively for Heat low, A it with was the around values 640 consistentlyC. further increase when austenite disappeared, as shown in Figure 6. This was also confirmed by the below 10%. For Heat B the peak was reached at 630 °C, whereas for Heat A it was around 640 °C. Strengthinstantaneous and hardness strain hardening were reduced coefficient, in the n, presencemeasuredof during retained tensile austenite, tests of butsingle they tempered showed a Strength and hardness were reduced in the presence of retained austenite, but they showed a furthersamples, increase as when shown austenite in Figure disappeared, 9. This coefficient as shown is sensitive in Figure to6 all. This factors was influencing also confirmed the strain by the instantaneousfurtherhardening increase strain behavior when hardening austeniteof the coefficient, steel. disappeared, For steelsn, measured aswhose show nmicrostructure during in Figure tensile 6. This is tests a was mixture of also single confirmedof temperedmartensite by samples,the and as showninstantaneousaustenite, in Figure both 9strain. the This meanhardening coefficient value andcoefficient, is sensitive the punctual n, to measured all variability factors during influencing of n were tensile affected the tests strain byof relativesingle hardening temperedamounts behavior of samples, as shown in Figure 9. This coefficient is sensitive to all factors influencing the strain of the steel.soft retained For steels austenite. whose As microstructure shown in Figure is a9, mixturethe gradual of martensiteloss of stability and of austenite, reverted austenite both the for mean hardening behavior of the steel. For steels whose microstructure is a mixture of martensite and value andtempering the punctual temperatures variability above of640n °Cwere in Heat affected A is byapparent, relative with amounts a practical of soft absence retained of retained austenite. austenite, both the mean value and the punctual variability of n were affected by relative amounts of austenite at room temperature for tempering treatments above 670 °C. As shownsoft retained in Figure austenite.9, the gradualAs shown loss in Figure of stability 9, the of gradual reverted loss austenite of stability for of temperingreverted austenite temperatures for ◦ abovetempering 640 C in temperatures Heat A is apparent, above 640 with °C ain practical Heat A is absence apparent, of retainedwith a practical austenite absence at room of retained temperature ◦ for temperingaustenite at treatments room temperature above 670 for temperingC. treatments above 670 °C.

Figure 7. Hardness values for Heat A and Heat B as a function of the single tempering treatment.

FigureFigure 7. Hardness 7. Hardness values values for for Heat Heat A A and and HeatHeat B as a a function function of of the the single single tempering tempering treatment. treatment.

Metals 2017Metals, 7 ,20172017 351,, 77,, 351351 88 ofof 14148 of 14

FigureFigure 8. Volume 8. VolumeVolume fraction fractionfraction of retainedofof retainedretained austenite austeniteaustenite in ininfunction functionfunction ofof temperingtempering tempering temperatures.temperatures. temperatures.

FigureFigure 9. Instantaneous 9. InstantaneousInstantaneous strain strainstrain hardening hardeninghardening coefﬁcient coefficientcoefficient of ofof Heat HeatHeat AA after after differentdifferent different temperingtempering tempering treatments.treatments. treatments.

From these results, the final hardness after single temperingtempering appearsappears toto bebe thethe resultresult ofof bothboth thethe Fromstructuralstructural these recoveryrecovery results, ofof the untempereduntempered ﬁnal hardness martensitemartensite after singleandand thethe tempering relativerelative stabilitystability appears ofof revertedreverted to be the austeniteaustenite result of duringduring both the structuralprevious recovery heating. of Strength untempered and hardness martensite initially and decreased the relative after stability single tempering of reverted above austenite 600 °C due during previoustoto increasedincreased heating. recoveryrecovery Strength ofof andas-quenchedas-quenched hardness martensitemartensite initially anan decreasedd the retention after of single stable temperingretained austenite. above For 600 ◦C due toincreasingincreasing increased temperatures,temperatures, recovery of as-quenchedrevertedreverted austeniteaustenite martensite becabecame and unstable, the retention forming of virgin stable martensite retained austenite.upon For increasingcooling,cooling, thusthus temperatures, determiningdetermining aa reverted furtherfurther inin austenitecreasecrease ofof bothboth became hahardnessrdness unstable, andand strength.strength. forming virgin martensite upon cooling, thusFigure determining 10 shows the a further microstructure increase ofof of HeatHeat both AA hardness afterafter singlesingle and temperingtempering strength. atat 620620 °C°C (Figure(Figure 10a)10a) andand 680 °C (Figure 10b), respectively, below and above the critical temperature for austenite Figure 10 shows the microstructure of Heat A after single tempering at 620 ◦C (Figure 10a) and destabilization upon cooling. The microstructure of samples tempered at 620 °C consisted of 680 ◦C (Figure 10b), respectively, below and above the critical temperature for austenite destabilization temperedtempered martensitemartensite withwith elongatedelongated islandsislands ofof reretainedtained austeniteaustenite◦ formedformed inin thethe interlathinterlath regionsregions upon cooling.with no evidence The microstructure of untempered of samplesmartensite. tempered Afterr temperingtempering at 620 Catat consisted680680 °C,°C, manymany of tempered islandsislands ofof martensite highlyhighly with elongateddislocated virgin islands martensite of retained were austeniteobserved. formed in the interlath regions with no evidence of ◦ untemperedFinally, martensite. Figure After11 shows tempering the alloy at microstructure 680 C, many after islands tempering of highly at 650 dislocated °C, where virgin all possible martensite were observed.structuralstructural constituentsconstituents maymay coexist—specifically,coexist—specifically, recoveredrecovered originaloriginal martensite,martensite, virginvirgin martensite,martensite, Finally,and retained Figure stable 11 shows austenite. the alloy microstructure after tempering at 650 ◦C, where all possible structural constituentsAs shown in mayFigure coexist—speciﬁcally, 11 and Table 3, image recovered contrast original coupled martensite, with energy-dispersive virgin martensite, X-ray and retainedspectroscopyspectroscopy stable austenite. (EDS)(EDS) measurementsmeasurements cancan bebe convenconvenientlyiently usedused toto distinguishdistinguish betweenbetween differentdifferent constituents.constituents. RegionsRegions wherewhere α’’ toto γ reversionreversion occurredoccurred areare characterizedcharacterized byby anan increasedincreased amountamount ofof As shown in Figure 11 and Table3, image contrast coupled with energy-dispersive X-ray spectroscopy (EDS) measurements can be conveniently used to distinguish between different constituents. Regions where α’ to γ reversion occurred are characterized by an increased amount of both Ni and Mn compared to the nominal heat composition, as a consequence of solute redistribution between phase constituents during tempering. Metals 2017, 7, 351 9 of 14 both Ni and Mn compared to the nominal heat composition, as a consequence of solute Metalsboth 2017Ni , 7and, 351 Mn compared to the nominal heat composition, as a consequence of solute9 of 14 redistribution between phase constituents during tempering.

Figure 10.10. TEM electron micrographs of Heat A afterafter single tempering treatmenttreatment at: (a) 620620 ◦°CC and ((b)) 680680 ◦°C.C.

Figure 11. TEM micrograph of Heat A tempered at 650 °C. Numbers in the figure identify the regions Figure 11. TEM micrograph of Heat A tempered at 650 °C◦C.. Numbers in the figure ﬁgure identify the regions where EnergyEnergy dispersive dispersive X-ray X-ray EDX EDX microanalysis microanalysi wass was performed. performed. Results Results for Cr, for Mn, Cr, and Mn, Ni and weight Ni percentagesweight percentages are reported are reported in Table 3in. Table 3.

Table 3. Content of Cr, Mn, and Ni measured by EDX microanalysis from the different regions Table 3.3.Content Content of of Cr, Cr, Mn, Mn, and and Ni measuredNi measured by EDX by microanalysisEDX microanalysis from thefrom different the different regions markedregions marked in Figure 11. inmarked Figure in 11 Figure. 11. Region Marked in Figure 11 wt % Cr wt % Mn wt % Ni Region Marked in Figure 11 wt % Cr wt % Mn wt % Ni 1 (retained austenite) 12.47 ± 0.5 2.40 ± 0.2 7.32 ± 0.3 ± ± ± 21 (retained(tempered austenite) martensite) 12.47 13.57 ±0.5 0.4 0.76 2.40 ± 0.10.2 3.30 7.32 ± 0.20.3 ± ± ± 2 (tempered3 (virgin martensite) martensite) 13.57 13.16 ±0.4 0.7 1.99 0.76 ± 0.30.1 5.07 3.30 ± 0.20.2 33 (virgin (virgin martensite) martensite) 13.16 13.16± ±0.7 0.7 1.99 1.99 ±± 0.30.3 5.07 5.07 ±± 0.20.2

Metals 2017, 7, 351 10 of 14 Metals 2017,, 7,, 351351 10 of 14 Among these areas, retained austenite islands can be further distinguished by a low density of Among these areas, retained austenite islands can be further distinguished by a low density of dislocationsAmong (Figure these areas, 11, Region retained 1), austenitein contrast islands to virgin can martensite be further with distinguished a high density by a lowof dislocations density of dislocations (Figure 11, Region 1), in contrast to virgin martensite with a high density of dislocations dislocations(Figure 11, Region (Figure 3). 11 Finally,, Region regions 1), in contrast of tempered to virgin martensite martensite are with characterized a high density by decreased of dislocations nickel (Figure 11, Region 3). Finally, regions of tempered martensite are characterized by decreased nickel (Figurecontent 11compared, Region 3).to the Finally, nominal regions heat of composition, tempered martensite as well as are by characterizedpartial recovery by of decreased the martensitic nickel content compared to the nominal heat composition, as well as by partial recovery of the martensitic contentmatrix (Figure compared 11, Region to the nominal 2). heat composition, as well as by partial recovery of the martensitic matrix (Figure 11, Region 2). matrixThe (Figure effect 11 of, Regiona second 2). tempering treatment was investigated, taking into consideration NACE The effect of a second tempering treatment was investigated, taking into consideration NACE recommendationsThe effect of a for second the temperingfirst cycle treatment (for which was NACE investigated, requires taking a temperature into consideration in the NACErange recommendations for the first cycle (for which NACE requires a temperature in the range recommendations648–690 °C); two fortemperatures the ﬁrst cycle corresponding (for which NACE to th requirese presence a temperature or total absence in the of range stable648–690 austenite◦C); 648–690 °C); two temperatures corresponding to the presence or total absence of stable austenite twowere temperatures chosen: 650 °C corresponding (P = 18.4) and to680 the °C presence (P = 19.0), or respectively. total absence For of th stablee second austenite cycle, the were minimum chosen: were chosen: 650 °C (P = 18.4) and 680 °C (P = 19.0), respectively. For the second cycle, the minimum 650and◦ maximumC(P = 18.4) NACE-recommendedand 680 ◦C(P = 19.0), temperatures—595 respectively. For the °C second (P = cycle,17.3) and the minimum620 °C (P and= 17.9)—were maximum and maximum NACE-recommended temperatures—595 °C (P = 17.3) and 620 °C (P = 17.9)—were NACE-recommendedused. Figure 12 shows temperatures—595YS and UTS, and Figure◦C(P 13= 17.3)shows and the 620 hardness◦C(P = after 17.9)—were double tempering. used. Figure 12 used. Figure 12 shows YS and UTS, and Figure 13 shows the hardness after double tempering. shows YS and UTS, and Figure 13 shows the hardness after double tempering.

Figure 12. YS and UTS of Heat A and Heat B following double temper treatments. Figure 12. YS and UTS of Heat A and Heat B following double temper treatments. Figure 12. YS and UTS of Heat A and Heat B following double temper treatments.

Figure 13. Hardness values for Heat A and Heat B in functionfunction ofof doubledouble temperingtempering treatments.treatments. Figure 13. Hardness values for Heat A and Heat B in function of double tempering treatments. BothBoth tensiletensile strength strength and and hardness hardness levels levels were signiﬁcantlywere significantly lower comparedlower compared to the ﬁrst to tempering the first cycle,temperingBoth with tensilecycle, hardness withstrength reduced hardness and to reducedhardness 266 HV to after levels266 1HV hwe atafterre 650 significantly1◦ hC, at and 650 1°C, h lower atand 620 1 hcompared◦ C,at 620 whereas °C, towhereas nothe useful first no reductionstemperinguseful reductions cycle, were obtainedwith were hardness obtained for Heat reduced for B, Heat with to B,266 tensile wi HVth strengthsaftertensile 1 hstrengths at above 650 °C, 800above and MPa, 1800 h withat MPa, 620 hardness °C, with whereas hardness values no remainingusefulvalues reductionsremaining around aroundwere 300 HV10. obtained 300 HV10. for Heat B, with tensile strengths above 800 MPa, with hardness values remaining around 300 HV10.

Metals 2017, 7, 351 11 of 14

Metals 2017, 7, 351 11 of 14 The effects of intermediate and final sub-zero cooling (−30 ◦C) for double tempering treatment were studiedThe effects using of 650 intermediate◦C for the and first final tempering sub-zero temperaturecooling (−30 °C) and for varying double tempering the temperature treatment of the secondwere cycle studied from using 595 to 650 620 °C◦C. for The the resultsfirst temperin reportedg temperature in Figure 13 and indicate varying that the an temperature intermediate of the cooling second cycle from 595 to 620 °C. The results reported in Figure 13 indicate that an intermediate between the first and the second temper was beneficial for the reduction of hardness, especially for cooling between the first and the second temper was beneficial for the reduction of hardness, Heat B, thus confirming Gooch’s suggestions [12]. especially for Heat B, thus confirming Gooch’s suggestions [12]. Moreover,Moreover, a final a final sub-zero sub-zero treatment—which treatment—which is expected to to partially partially destabilize destabilize retained retained austenite—inducedaustenite—induced an increasean increase in in hardness, hardness, especially especially for Heat B B (as (as shown shown in inFigure Figure 14). 14 ).

FigureFigure 14. Hardness14. Hardness values values for for Heat Heat A A and and HeatHeat B followingfollowing double double tempering tempering treatments treatments with with intermediate and final sub-zero (−30 °C) cooling. intermediate and final sub-zero (−30 ◦C) cooling. 4. Discussion 4. Discussion Despite similar chemical compositions and almost similar Creq/Nieq ratios of Heats A and B, Despitetheir responses similar to chemical heat treatments compositions were rather and differ almostent. It similar is apparent Creq that/Ni aeq significantratios of reduction Heats A of and B, their responseshardness below to heat the treatmentsNACE limit were(23 HRC) rather could different. be accomplished It is apparent only for that Heat a significant A, whereas reduction Heat B of hardnessremained below hard the NACEwhile limitalso (23adopting HRC) couldthe double be accomplished tempering onlystrategy for Heat recommended A, whereas by Heat B remainedNACE hard standards. while also adopting the double tempering strategy recommended by NACE standards. In orderIn order to obtain to obtain significant significant hardness hardness reductions reductions after tempering, tempering, the the as-quenched as-quenched martensite martensite matrix should recover at high temperatures, allowing the retention of soft austenite islands without matrix should recover at high temperatures, allowing the retention of soft austenite islands without destabilization to virgin martensite. For the industrial heats studied, two major difficulties are destabilizationpresented. The to virginfirst difficulty martensite. is to obtain For signific the industrialant volume heats fractions studied, of stable two retained major austenite difficulties at are presented.room temperature, The first difficulty potentially is toas a obtain result of significant the relatively volume high Ms fractions values ofmeasured stable retained(Figure 2). austenite The at roomsecond temperature, difficulty is potentiallyan intrinsic instability as a result of ofreversed the relatively austenite highformed Ms at valuesa temperature measured that favors (Figure 2). The secondthe transformation difficulty is to an virgin intrinsic martensite instability during of cooling reversed and austenite limits the formed hardness at areduction temperature after that favorstempering the transformation in the final product. to virgin martensite during cooling and limits the hardness reduction after temperingThe in thestability final product.of reversed austenite decreased during tempering as consequence of solute Thepartitioning stability of γ of-stabilizer reversed elements austenite (especia decreasedlly Ni and during C). As the tempering tempering as temperature consequence increases, of solute the volume fraction of reversed austenite increases and the consequent solute redistribution is no partitioning of γ-stabilizer elements (especially Ni and C). As the tempering temperature increases, longer able to stabilize the γ phase on cooling. This aspect is believed to be critical for 13Cr-4Ni-(Mo) the volume fraction of reversed austenite increases and the consequent solute redistribution is no industrial heats, which are characterized by relatively low carbon and nickel contents as compared γ longerto ableother to supermartensitic stabilize the phasealloys on[6].cooling. Moreover, This alloying aspect with is believed molybdenum to be critical(0.5%) reduces for 13Cr-4Ni-(Mo) carbon industrialdiffusion heats, towards which reverted are characterized austenite islands, by relatively thus further low limiting carbon stabilization and nickel of contents reverted asaustenite compared to otherand supermartensiticlowering kinetics for alloys martensite [6]. Moreover, recovery. Moly alloyingbdenum with is effective molybdenum in decreasing (0.5%) the reduces diffusion carbon diffusioncoefficient towards of elements reverted such austenite as carbon islands, and sulfur thus within further ferritic/mart limiting stabilizationensitic structures, of reverted because of austenite its and loweringstrong affinity kinetics for those for martensite elements [20,21]. recovery. Loveless Molybdenum et al. [22] worked is effective within 13Cr4Ni decreasing alloys, the finding diffusion coefficientthat variants of elements containing such molybdenum as carbon and presented sulfur withinincreasing ferritic/martensitic difficulties for hardness structures, reduction because of itsthrough strong tempering. affinity for those elements [20,21]. Loveless et al. [22] worked with 13Cr4Ni alloys,

ﬁnding that variants containing molybdenum presented increasing difﬁculties for hardness reduction through tempering. Metals 2017, 7, 351 12 of 14

Additionally, it was observed that the retained austenite in these alloys was rather unstable. As demonstrated by the results reported in Figure 14, there is an apparent destabilization effect of sub-zero cooling for Heat B. This aspect is believed to be important for mechanical components that work at low service temperatures and/or under manufacturing and service strains, since a loss in toughness and in corrosion resistance may be envisaged. Under these circumstances, limiting the formation of reversed austenite is believed to be useful during tempering, and can be accomplished by increasing the Ac1 critical temperature of the alloy as much as possible. This consequently allows for the recovery of the as-quenched microstructure at higher temperatures without the detrimental effect of austenite destabilization and the formation of virgin martensite upon cooling. For example, Heats A and B appear different for experimental Ac1 and Ms temperatures, with ◦ ◦ ◦ both temperatures being higher for Heat A (Ac1 = 620 C, Ms = 270 C) than for Heat B (Ac1 = 600 C, Ms = 250 ◦C). These differences might be the consequence of higher carbon and nitrogen contents for Alloy B, with nickel and manganese contents of alloys being almost identical (Table1). Therefore, more stringent control of carbon and nitrogen may be the key to controlling the ﬁnal hardness of 13Cr-4Ni-(Mo) industrial heats. A ﬁrst attempt to verify these conclusions has been done with Heat C, reported in Table4.

Table 4. Chemical composition of the alloy C.

Heat %C %Si %Mn %P %Cr %Mo %S %Ni N (ppm) %Creq/%Nieq C 0.018 0.420 0.68 0.023 12.53 0.53 0.001 4.00 68 2.70

The carbon content was reduced to 0.018% C, and the nitrogen content was reduced to 68 ppm. ◦ ◦ The predicted Ac1 temperature for Heat C is 670 C, which is about 15 C higher than Heat A. From this it would be possible to increase tempering temperature without the formation of large volume fractions of unstable reverted austenite. Heat C was treated through double tempering at 665 and 615 ◦C. The ﬁnal mechanical properties—averaged from seven different samples—gave the following results: YS = 583 MPa, UTS = 751 MPa, with hardness of 21.4 HRC. These results allow for the supposition that the reduction of both carbon level below 0.02% C and nitrogen below 100 ppm are effective in controlling the ﬁnal hardness of 13Cr-4Ni-(Mo) with industrial heats below NACE recommendations (23 HRC), without an unacceptable loss of mechanical strength.

5. Conclusions Control of the final mechanical properties of industrial 13Cr-4Ni-(Mo) heats has proven to be difficult when using double tempering treatments in accordance with NACE specifications. From the results obtained here, this difficulty is mainly related to the quantity of austenite reversion that can form during tempering that—if it is not sufficiently stable—transforms into virgin martensite during cooling, thus limiting the effect of the tempering process. This, in turn, is due to the relatively low Ac1 temperatures of industrial heats, with the associated formation of unstable reverted austenite at temperatures above 600 ◦C. Therefore, limiting the amount of austenite reversion is believed to be necessary while favoring the recovery of untempered martensite at tempering temperatures as high as possible, without allowing the formation of virgin martensite upon final cooling. Reduced carbon (<0.02% C) and nitrogen (<100 ppm) levels have proven effective in controlling the final hardness of 13Cr-4Ni-(Mo) industrial heats below NACE recommendations (23 HRC), without an unacceptable loss of mechanical strength. Moreover, the retention of significant amounts of soft retained austenite at room temperature for 13Cr-4Ni-(Mo) alloys should be carefully considered, as this phase appeared to transform into martensite following a sub-zero treatment at −30 ◦C. This phenomenon could be detrimental for mechanical components working at low service temperatures and/or under service strains. Metals 2017, 7, 351 13 of 14

Author Contributions: Massimiliano Buccioni and Angelo Donato performed all heat treatments; Gianfranco Lovicu carried out hardness tests, tensile tests, Gleeble tests and performed TEM studies together with Massimo De Sanctis; Alessandra Varone and Maria Richetta performed XRD; All the authors discussed the results and contributed in writing the paper. Conﬂicts of Interest: The authors declare no conﬂict of interest.

Nomenclature

Ac1 Austenite transformation temperature Creq Chromium equivalent (percentage of ferrite stabilizing elements) δ, α’, γ phases designation HB Brinell hardness HV Vickers hardness HR Preliminary hot forging HRC Rockwell hardness, C scale Ms Martensite start temperature Mf martensite ﬁnish temperature n Instantaneous strain-hardening coefﬁcient Nieq Nickel equivalent (weight percentage of austenite stabilizing elements) OQ Oil Quenched UTS Ultimate Tensile Strength YS Yield Strength

References

1. Kondo, K.; Ueda, M.; Ogawa, K.; Amaya, H.; Hirata, H.; Takabe, H.; Miyazaki, Y. Alloy design of super 13Cr martensitic stainless steel (Development of uper 13Cr martensitic stainless steel for line pipe-1). In Proceedings of the Supermartensitic Stainless Steels 99 Conference, Belgian, Brussels, 27–28 May 1999; pp. 11–18. 2. Niederau, H.J. A New Low-Carbon 16Cr-5Ni Stainless Martensitic Cast Steel. In ASTM STP 756 Stainless Steel Castings; Behal, V.G., Melilli, A.S., Eds.; American Society for Testing and Materials: West Conshohocken, PA, USA, 1982; pp. 382–393. 3. Nakada, N.; Tsuchiyama, T.; Takaki, S.; Miyano, N. Temperature Dependence of Austenite Nucleation Site on Reversion of Lath Martensite. Mater. Sci. Forum 2010, 638–642, 3424–3429. [CrossRef] 4. Liu, Z.B.; Yang, Z.Y.; Liang, J.X.; Zhang, L.N.; Zhang, X.L. Precipitation behavior and transformation kinetics of reverted austenite in ultra-high strength stainless steel. Trans. Mater. Heat Treat. 2010, 31, 39–44. 5. Park, E.S.; Yoo, D.K.; Sung, J.H.; Kang, C.Y.; Lee, J.H.; Sung, J.H. Formation of reversed austenite during tempering of 14Cr-7Ni-0.3Nb-0.7Mo-0.03C super martensitic stainless steel. Met. Mater. Int. 2004, 10, 521–525. [CrossRef] 6. De Sanctis, M.; Lovicu, G.; Valentini, R.; Dimatteo, A.; Ishak, R.; Migliaccio, U.; Montanari, R.; Pietrangeli, E. Microstructural Features Affecting Tempering Behavior of 16Cr-5Ni Supermartensitic Steel. Metall. Mater. Trans. A 2015, 46, 1878–1887. [CrossRef] 7. Song, Y.Y.; Li, X.Y.; Yin, F.X.; Ping, D.H.; Li, Y.Y. Variation of the Reversed Austenite Amount with the Tempering Temperature in a Fe-13%Cr-4%Ni-Mo Martensitic Stainless Steel. Mater. Sci. Forum 2010, 650, 193–198. [CrossRef] 8. Song, Y.Y.; Li, X.Y.; Rong, L.J.; Ping, D.H.; Yin, F.X.; Li, Y.Y. Formation of the reversed austenite during intercritical tempering in a Fe-13%Cr-4%Ni-Mo martensitic stainless steel. Mater. Lett. 2010, 64, 1411–1414. [CrossRef] 9. Bilmes, P.D.; Solari, M.; Llorente, C.L. Characteristics and effects of austenite resulting from tempering of 13Cr–NiMo martensitic steel weld metals. Mater. Charact. 2001, 46, 285–296. [CrossRef] 10. Lovicu, G.; De Sanctis, M.; Valentini, R.; Dimatteo, A.; Ishak, R.; Migliaccio, U.; Montanari, R.; Pietrangeli, E. Inﬂuence of presence and stability of reverted austenite on mechanical properties of 16Cr-5Ni supermartensitic stainless steel after tempering treatment. Metall. Ital. 2013, 105, 29–35. Metals 2017, 7, 351 14 of 14

11. Haynes, A.G. Some factors governing the metallurgy and voidability of 13%Cr and newer Cr-Ni martensitic stainless steels. In Proceedings of the Supermartensitic Stainless Steels ‘99, Brussels, Belgium, 27–28 May 1999. 12. Gooch, T.G. Heat treatment of welded 13% Cr-4% Ni martensitic stainless steels for sour service. Weld. J. 1995, 74, 213s–223s. 13. APN005 Gleeble System Application Note; Dynamic System Inc.: Poestenkill, NY, USA, 2001. 14. Colla, V.; De Sanctis, M.; Dimatteo, A.; Lovicu, G.; Valentini, R. Prediction of Continuous Cooling Transformation Diagrams for Dual-Phase Steels from the Intercritical Region. Metall. Mater. Trans. A 2011, 42, 2781–2793. [CrossRef] 15. American Society for Testing and Materials. ASTM E975 Standard Practice for X-ray Determination of Retained Austenite in Steel with Near Random Cryslalographic Orientation; ASTM: Philadelphia, PA, USA, 2003. 16. De Moor, E.; Föjer, C.; Clarke, A.J.; Penning, J.; Speer, J.G. Quench & partitioning response of a Mo-alloyed CMnSi steel. In Proceedings of the New Developments on Metallurgy and Applications of High Strength Steels Conference, Buenos Aires, Argentina, 26–28 May 2008; The Minerals, Metals & Materials Society: Warrendale, PA, USA, 2008; pp. 721–729. 17. Fisher, R.B.; Larson, J.A. Experience and some pitfalls in processing CA6NM. In Proceedings of the Transactions of the American Foundrymen’s Society, AFS Annual Meeting, Chicago, IL, USA, 19–23 April 1982; pp. 103–113. 18. Sherman, E.T.; Nalbone, C.S.; Scott, W.D. Processing low-carbon CA6NM (steel) to meet NACE (National Association of Corrosion Engineers) hardness requirements. In Proceedings of the 28th Annual Conference, Technical Advances in Steel Castings, Coventry, UK, 24–25 May 1983; Steel Castings Research and Trade Associations: Shefﬁeld, UK, 1983. 19. Gooch, T.G.; Woollin, P.; Haynes, A.G. Welding Metallurgy of Low Carbon 13% Chromium Martensitic Steels. In Proceedings of the Supermartensitic Stainless Steels ‘99, Brussels, Belgium, 27–28 May 1999; pp. 188–195. 20. Wada, H.; Pehlke, R.D. Nitrogen solubility and nitride formation in austenitic Fe-Ti alloys. Metall. Trans. B 1985, 16, 815–822. [CrossRef] 21. Akben, M.G.; Bacroix, B.; Jonas, J.J. Effect of vanadium and molybdenum addition on high temperature recovery, recrystallization and precipitation behavior of niobium-based microalloyed steels. Acta Metall. 1983, 31, 161–174. [CrossRef] 22. Loveless, R.W.; Smith, W.C.; Templeton, N.C. Mechanical Testing for Deformation Model Development, ASTM STP 765; ASTM: Philadephia, PA, USA, 1982; pp. 394–402.

© 2017 by the authors. Licensee MDPI, Basel, Switzerland. This article is an open access article distributed under the terms and conditions of the Creative Commons Attribution (CC BY) license (http://creativecommons.org/licenses/by/4.0/).