Thermomechanical Simulation of Heat-Affected Zones in Nickel-Free High Nitrogen Stainless Steel: Microstructural Evolution and Mechanical Property Studies

Abstract Three different Heat Affected Zones (HAZ) in hot rolled Nickel Free High Nitrogen Stainless Steels (NFHNSS) based on three different peak temperatures were physically simulated using Gleeble Simulator to investigate microstructural evolution and structure-property correlation. Optical microscopy revealed that the austenite grains are recrystallized in the simulated heat affected zone in the peak temperature range of 750 oC to 1050 oC. Extent of recrystallization of grains and nucleation of precipitates varied with peak temperatures. TEM characterization showed the presence of Cr2N precipitate having an average particle size in the range of 300 nm to 395 nm in the simulated HAZ were confirmed by Selected Area Electron Diffraction (SAED) analysis. Precipitation kinetics of Cr2N were simulated using Thermo-Calc were found to correlate well with experimental values. Mechanical properties of specimens taken from three different HAZ were evaluated for tensile strength and hardness. Variation in strength of the different specimens has been discussed using various strengthening models. Fractography analysis was also carried out to understand the effect of peak temperature on fracture behaviour. Transition in fracture patterns in NFHNSS from ductile to mixed mode was observed for different specimens.


Introduction
Nickel-Free High Nitrogen Stainless Steel (NFHNSS) containing 0.4% N are being considered for applications in defence, power plant and aerospace applications.In particular, ballistic tank shields fabricated from High Nitrogen Stainless Steel (HNSS) can be a viable replacement for conventional armour steels.HNSS possess excellent yield strength (>1000 MPa) along with good toughness, ductility and pitting resistance 1,2 .Nickel is replaced by nitrogen in austenitic stainless steel for both superior mechanical properties as well as cost benefits.In NFHNSS, addition of manganese aids in reducing the stacking fault energy which facilitates increased solubility of nitrogen in steel.However, manganese has a detrimental effect on localized corrosion resistance and can be compensated by increasing the nitrogen content of the steel 3 .
Welding of NFHNSS is an important area of research since a quality weld with adequate mechanical properties along with good corrosion resistance is required for many applications.Nitrogen loss in the weld zone, nitrogen induced porosity, solidification cracking, liquation cracking, δ-ferrite formation during different thermal cycling as well as precipitation of carbides and nitrides in the Heat Affected Zone (HAZ) are significant metallurgical issues involved in the welding of NFHNSS [4][5][6] .Proper selection of welding process as well as welding parameters, choice of electrodes and careful evaluation of the solidification behaviour helps in achieving a good weldment property.Mohammed et al. 7 reported the influence of microstructure on HAZ properties in NFHNSS formed under various thermal cycling.Mukherjee and Pal 8 reported that the heat input played a major role in the formation of M 23 C 6 and M 2 N precipitates along the grain boundaries during high-temperature thermal cycling.Rod like Cr 2 N precipitates were found in simulated HAZ specimens of weld thermal cycle with heat input of 5 kJ/cm on duplex steel exposed at 700 o C was observed by Chen and Yang 9 .Wang et al. 10 reported that increasing the heat input from 8 kJ/cm to 36 kJ/cm with different t 8/5 values increases the volume fraction of precipitates from 10.07% to 16.85% respectively.Similarly, with the increase of cooling rate, the fraction of Cr 2 N precipitates decreases from 13.8% to 2.9% in HNSS is observed by Li et al. 11 .Nitrogen enhances the properties when it is retained in solid solution, but nitrides were formed at higher peak temperatures has a detrimental effect on ductility and mechanical strength.Precipitation kinetics of Cr 2 N were reported in Fe-20Mn-12Cr-0.24C-0.32Nalloy that precipitation of Cr 2 N is favoured at 900 °C, as the driving force is higher at this temperature 12 .Krishna Kumar et al. 13 reported the kinetics of nitride precipitation at 700°C to 900°C aging temperatures using CALPHAD simulation.Simulations studies were also carried out by Lang et al. 14 to investigate the effect of various alloying elements (C, Cr, Mn and Ni) on nitrogen solubility at 1000°C to 1200°C and found that stable austenite phase were simulated within a given range of alloying elements using Thermo-Calc.Behjati et al. 15 investigated the effect of nitrogen content on grain refinement and strengthening mechanisms in Ni-free 18Cr-12Mn austenitic stainless steel.Various strengthening models influencing the mechanical strength of Fe-13Mn-1.3Caustenitic steels under different deformation conditions have been reported by Maier and Astafurova 16 .Similarly, quantitative models have been developed by Farooq 17 for predicting the strength contribution in austenitic stainless steels at elevated temperatures.
Limited literature were reported on simulation studies on precipitation kinetics and the effect of peak temperature and cooling rate on the microstructural and mechanical properties of NFHNSS weldments.Hence, the main objective of the present study was to simulate three different HAZ's with different peak temperature using Gleeble thermo-mechanical simulator and correlate the evolved microstructure with the mechanical properties of NFHNSS.Further, Precipitation kinetics of Cr 2 N particles in the various HAZ zones of NFHNSS have also been investigated using Thermo-Calc.

Materials and Methods
Nickel-Free High Nitrogen Stainless Steel (NFHNSS) having a chemical composition of 0.06C, 18Cr, 22Mn and 0.55N (wt%) was chosen for thermo-mechanical simulation.Welding parameters for thermo-mechanical simulation of heat affected zone were obtained by conducting SMAW trial experiments on 10 mm thick hot rolled NFHNSS plates.From the trial study data, three different HAZ were identified from the weld centre-line towards base material based on different microstructural morphology and were designated as HAZ-1 HAZ-2 and HAZ-3 respectively (Figure 1).Three specimens having dimension of 86 mm (length) x 10 mm (width) x10 mm (thickness) were used in the HAZ simulation for three different zones.Physical simulation of the samples was carried out using Gleeble-3500 under argon atmosphere with welding heat input of 6 kJ/cm and welding speed of 4 mm/s.3D-Rykalin mathematical model was used for determining the thermal cycle curve of NFHNSS specimens [18][19][20] .Simulated specimens were subsequently investigated for microstructural characterization and mechanical property evaluation.
As received (hot rolled) and HAZ simulated NFHNSS specimens were polished and electrolytically etched using an aqueous solution containing a mixture of 20 g oxalic acid and 200 ml distilled water.A Leica (DMi8C model) light polarized microscope was used to characterize the microstructure of the specimens.Grain size measurement was carried out using a Leica grain size analyzer as per ASTM E-112 standard 21 .An average of 10 readings was taken as the average grain size with an accuracy level of 0.097 microns/pixels.Specimens for Transmission Electron Microscope (TEM) investigations were prepared by mechanical polishing and further thinned down by ion milling.TEM analysis, using both imaging and Selected Area Diffraction modes were carried out on a JEOL-JEM-2100 instrument at an operating voltage of 200 kV.Volume fraction of secondary phases and dislocation density were estimated from TEM micrographs using Image-J software.Lattice points with zone axis of the primary phase and secondary phases in NFHNSS specimens were also investigated by SAED imaging.
Simulation of phase diagram of NFHNSS based on CALPHAD approach was carried out using Thermo-Calc software (TCFE8 database).Temperature-time data obtained from physical simulation of three different HAZ specimens were used for simulating the precipitation kinetics based on Langer-Schwartz theory and Kampmann-Wagner numerical approach using TC-PRISMA (Thermo-Calc Precipitation) module.Nucleation rate (m -3 s -1 ), mean particle size (nm), driving force for precipitation and volume fraction of precipitates were determined using TC-PRISMA mobility database under arbitrarily selected non-equilibrium conditions based on experimental thermal profiles.These simulated data were used to determine the kinetics of particle coarsening of Cr 2 N precipitates 22 .Room temperature tensile testing of the simulated specimens was carried out using an INSTRON micro-tensile testing machine at a constant strain rate of 0.001 s -1 .Tensile specimens having a gauge length of 16 mm were prepared by Electric Discharge Machining (EDM) and an average of three measurements was taken as the tensile strength.Fractography of tensile tested specimen was carried out using JEOL Scanning Electron Microscope.Subsequently, hardness measurements were done using a Zwick Vickers micro hardness tester with a 10 kgf load for a dwell time of 15 seconds and an average of five trials were taken as the hardness.NFHNSS specimen (Figure 2a) showed deformed austenite grains containing annealing twins with an average grain size of about 40 ± 2 μm.Few grains showed abnormal growth due to strain-induced grain boundary migration during hot rolling of NFHNSS.Similar observations have been reported by Rios et.al during their investigations on austenitic stainless steels 23 .

Microstructural characterization
Microstructure of the HAZ-1 specimen (Figure 2b) showed predominantly equiaxed austenite grains, with an average grain size of about 33 ± 1.65 μm.The formation of such grains is favoured by static recrystallization owing to a reduction in the stored energy of deformed grains and twins during thermal cycling.The degree of recrystallization depends on local differences in dislocation density which provides the driving force for nucleation of grain at the grain boundaries as reported by Sellars and Whiteman 24 .Grain coarsening was facilitated by high peak temperatures with slow cooling rate.For HAZ-2 specimen, microstructure comprised of fine recrystallized austenite grains along with a few deformed grains having an average grain size of 19 ± 0.95 μm (Figure 2c).This can be associated with a relatively low peak temperature and shorter time for recrystallization.In the case of HAZ-3 specimen, optical micrograph (Figure 2d) showed bimodal grain morphology of partially recrystallized austenite grains and deformed grains having an average grain size of 26 μm ± 0. 5 μm is due to lower peak temperature during the thermal cycle.Primary reason for the partial recrystallization and bimodal grain morphology is due to lower peak temperature and some of the deformed grains having higher stored energy recrystallize effectively compared to the grains with relatively lower stored energy.Misra have reported a similar result during his studies on hot rolled micro alloyed steels 25 .
The TEM micrograph of hot rolled NFHNSS specimen (Figure 3a) exhibited the presence of dense dislocation and deformation twins in the austenite matrix.The dislocation density for the specimen was calculated using intercept method (ρ= 2nM/Lt ) by drawing a circle with circumference (L) 200 mm over the micrograph of 100000X magnification (M) using Image-J software.The numbers of dislocation intercept (n) were counted as 70 with foil thickness (t) of 120 nm.The dislocation density of hot rolled specimen was 5.87 x 10 14 m -2 26 .Austenite phase was identified in the SAED pattern with a zone axis of [0 1 1].TEM analysis of base metal show absence of Cr 2 N precipitate.Figure 3b shows the bright field image of the HAZ-1 specimen.It can be observed that the formation of Cr 2 N precipitates occurred at the grain boundaries of the austenitic matrix.It has been reported that higher nitrogen levels in the alloy reduced the driving force for nucleation and growth of M 23 C 6 and sigma phase compared to Cr 2 N precipitates in NFHNSS 27 .
Figure 4 shows the peak temperature of physically simulated NFHNSS specimen.The peak temperature and cooling rate of HAZ-1 specimen nearer to the fusion zone are 1052 o C and 20 °C/s respectively.Since HAZ-1 exhibits peak temperature above 925 °C and cooling rate is 20 °C/s , nucleation of Cr 2 N is favoured at the grain boundaries during thermal cycling.This can be justifiable since nitrogen has more affinity towards chromium than carbon 28 .Woo and Kikuchi 2 also reported that Cr 2 N precipitates at a very slow cooling rate below 20 °C/s in the HAZ containing 0.75 wt.%N in HNSS since there is sufficient time for precipitation during cooling.In the present study, it can be observed that the Cr 2 N precipitate was also surrounded by a few dislocation tangles.Thus, it can be concluded that stored energy of dislocation tangles provides a sufficient driving force for the formation of Cr 2 N precipitates.The dislocation density was found to reduce to 2.2 x 10 14 m -2 compared to that of the hot rolled specimen.Further, the morphology of Cr 2 N precipitates was found to be globular in nature (395 nm), facilitated by high peak temperature.The presence of Cr 2 N precipitates was confirmed by the indexed SAED pattern obtained from the [1 0 1] zone axis representing a HCP lattice with a = 0.2748 nm and c = 0.4438 nm respectively.This data closely matches with the standard diffraction pattern of Cr 2 N precipitate 29 .
Similarly, the peak temperature and cooling rate of the HAZ-2 specimen was found to be 910 o C and 17 °C/s respectively also favoured the formation of Cr 2 N precipitates.As seen from Figure 3c, the HAZ-2 specimen showed the presence of catenary type Cr 2 N precipitate with a particle size of 260 nm at the grain boundary.It was also found that a dislocation density (2.93 x 10 14 m -2 ) in the HAZ-2 specimen was relatively higher compared to that of the HAZ-1 specimen due to a low peak temperature.Significant difference in the morphology of Cr 2 N precipitate between the HAZ-1 and HAZ-2 specimen arises due to variation in surface energy of precipitates imposed by the different peak temperature and cooling rate during thermal cycle.
HAZ-3 specimen (Figure 3d) showed the presence of dense dislocation tangles with the dislocation density of 5.13 x 10 14 m -2 and very few Cr 2 N precipitates owing to low peak temperature .Peak temperature of the HAZ-3 specimen was 785 o C and cooling rate is 16 °C/s .At this temperature, the extent of formation of Cr 2 N precipitates was found to reduce significantly due to low driving force.It can be also inferred that the driving force for nucleation of Cr 2 N precipitates tends to vary with cooling rate and exposure time at higher temperatures 30 .Thus, precipitate morphology and nucleation time for the formation of Cr 2 N precipitates is determined by cooling rate and reaction kinetics between nitrogen with chromium.

Precipitation kinetics of Cr 2 N
Simulated property diagram (Figure 5) showed that formation of delta ferrite begins at a temperature of 1375 o C and subsequently gets transformed to an austenite phase at a temperature of 1185 o C during solidification.A complete

Mechanical property studies
Average hardness of the hot rolled plate and simulated HAZ specimen are shown in Figure 7.It can be observed that hot rolled base plate display maximum hardness of was 340 HV10.This hardness is due to deformed microstructure containing higher dislocation density.It can be inferred that the hardness values of simulated HAZ vary significantly with distance towards the base metal.While the HAZ-1 specimen displayed lowest hardness 280 HV10, the HAZ-3 specimen exhibited highest hardness of 328 HV10.Hardness of the HAZ-3 specimen was increased by 14% compared to that of the HAZ-1 specimen.This hardness variation can be related to the variation in peak temperature experienced by different HAZ leading to the differences in grain size.In addition, HAZ -1 specimen exhibited a relatively lower dislocation density compared to that of HAZ-3 specimen as seen from TEM investigations.This difference in dislocation densities is due to variation in the extent of dislocation annihilation in HAZ-1 and HAZ-3, since the specimens experienced different peak temperatures.HAZ-2 specimen showed an intermediate hardness of 294 HV10 owing to intermediate dislocation density that of HAZ-1 and HAZ-3.
Figure 8a shows the engineering stress-strain curve of hot rolled and HAZ specimens.Hot rolled specimen shows average yield strength of 945 ± 6 MPa whereas, the HAZ-1 specimen exhibited the average yield strength of 734 ± 20 MPa, which has been discussed in detail based on strengthening mechanism models.While the ultimate tensile strength of the hot rolled specimen was found to be 1127 MPa ± 9 MPa, the ultimate tensile strength of HAZ-1, HAZ-2 and HAZ-3 simulated specimen were 954 ± 3 MPa, 1096 ± 8MPa, 1113 ± 6 MPa respectively.The percentage of elongation to failure of the hot rolled plate was observed to be 16%, and elongation of HAZ-1, HAZ-2 and HAZ-3 specimens were found to be 13%, 12% and 12.5% respectively.In general, the hot rolled NFHNSS specimen experiences maximum yield strength in the range of (900 to 1100 MPa) and reasonable ductility in the range of 10 to 18%.It can be seen that simulated HAZ specimens experience a slight decrease in ductility compared to that of the hot rolled specimen.The difference in ductility of the HAZ specimen arises because of variations in grain sizes, formation of Cr 2 N and dislocation density imposed by thermal cycles.Figure 8b shows the variation in grain size with the yield strength of the material.Yield strength of simulated HAZ samples is contributed by grain boundary strengthening and other strengthening mechanism.Thus, contribution of various strengthening on yield strength of the material is discussed.
Figure 9 micrographs shows the fracture surface morphologies of hot rolled and HAZ specimens after tensile testing.The micrographs of fracture surface in Figure 9a, exhibited dimple morphology featuring micro void formation in the hot rolled specimen, which experienced significant plastic deformation during tensile testing.Tensile test results also confirmed that hot rolled specimens exhibited a maximum ductility of 16%.Fractography analysis showed the transition from a ductile to a mixed mode of fracture surfaces in HAZ specimens because of variation in the extent of plastic deformation.HAZ-1 specimen failed by transgranular fracture owing to the nucleation and growth of cavities due to particle pull-out as indicated in Figure 9b.Both the HAZ-2 and HAZ-3 specimens exhibit a mixedmode of fractures as shown in Figure 9c, d.As indicated in Figure 9d, some degree of particle decohesion occurred between the particle-matrix interfaces.This means that the matrix deformation was certainly limited by the presence of precipitates which tends to reduce the ductility of the HAZ-2 and HAZ-3 specimen as seen from the tensile results.Moreover, the fracture patterned surfaces in HAZ-2 and HAZ-3 shows visible tearing ridges degrade the ductility of the specimen.

Strengthening mechanisms in NFHNSS
The theoretical yield strength of the NFHNSS can be estimated by considering four different strengthening contribution mechanisms namely solid solution strengthening (σ SS ), grain boundary strengthening (σ GB ), dislocationassisted strengthening (σ DS ) and precipitation or Orowan strengthening (σ PS ) The solid solution strengthening of NFHNSS could be calculated using Equation 2. It is understood that alloying of interstitial elements (N ,C) greatly impacts the yield strength of Fe-Mn-Cr-N austenite 28 The strengthening effect of grain boundaries is estimated by Hall-Petch relation 28

 
Where, σ f is the frictional stress, K y is the Hall-Petch slope is taken as 200 MPa and 16.5 31   (5)   Where α is a constant taken as 0.21, ρ is length of dislocation per unit volume, n is the number of intercepts with the circle, M is the magnification, L is the circumference of the circle in the micrograph and t is the foil thickness.The dislocation density were calculated using by intercept method The resistance offered by precipitates to the passage of dislocations can be estimated using Orowan-Ashby-equation 32,33 0.13 ln /  µ PS b σ r b λ (6)   where µ the shear modulus taken as 78 GPa, b is the Burgers vector taken as 0.25 nm, r is the particle radius, r = d p /2, and λ is the interparticle spacing, expressed as , Where Vp is volume fraction of precipitates and particle diameter (d p ) calculated from TEM results.Contribution of different strengthening mechanisms in base metal and simulated HAZ of NFHNSS is illustrated in Figure 10.An analysis of different strengthening mechanisms showed that solid solution strengthening appears to be the most dominant mechanism in hot rolled NFHNSS steels.Gavriljuk and Berns 28 discussed similar solid solution strengthening model in high nitrogen steel.The Equation 2suggests that nitrogen plays a major role in determining the solid solution strengthening owing to the large mismatch between nitrogen and Fe atoms in steels.The next dominant strengthening mechanism is grain boundary strengthening which contributes nearly 32% to yield strength.This grain refinement may be because of thermo mechanical processing

Figure 2 Figure 1 .
Figure 2 shows the optical micrographs of base metal and simulated HAZ of NFHNSS specimens.Hot rolled

Figure 7 .
Figure 7.Comparison of Vickers hardness of hot rolled and simulated HAZ specimens with Error bars representing standard deviations.

Figure 8 .
Figure 8.(a) Stress-strain curve of Hot rolled and HAZ specimens, (b) Relationship between yield strength and grain size of simulated NFHNSS.
and d g is the average austenite grain size.The Taylor equation is used to calculate dislocation strengthening is expressed as26 ,