19. VV SATYANARAYANA 2005 Dissimilar metal friction welding of austenitic–ferritic stainless steels

Journal of Materials Processing Technology 160 (2005) 128137

Dissimilar metal friction welding of austeniticferritic stainless steelsV.V. Satyanarayana a, G. Madhusudhan Reddy b, T. Mohandas b,

a Vasavi College of Engineering, Hyderabad 500 031, Indiab Defence Metallurgical Research Laboratory, Kanchanbagh, Hyderabad 500 058, India

Received 30 June 2003; received in revised form 20 May 2004; accepted 20 May 2004

Abstract

Continuous drive friction welding studies on austeniticferritic stainless steel combination has been attempted in this investigation.Parameter optimization, microstructuremechanical property correlation and fracture behaviour is a major contribution of the study. Soundwelds are obtained at certain weld parameter combinations only. The mechanical properties of dissimilar metal welds are comparable tothose of ferritic stainless steel welds. Evaluation of the joints for resistance to pitting corrosion revealed that the dissimilar welds exhibitlower resistance to pitting corrosion compared to the ferritic and austenitic stainless steel welds. Interface on the austenitic stainless steelside exhibited higher residual stress possibly due to its higher flow stress and higher coefficient of thermal expansion. 2004 Published by Elsevier B.V.

Keywords: Dissimilar metal welding; Friction welding; Austenitic stainless steel; Ferritic stainless steel; Microstructure; Notch tensile strength; Hardness;Impact toughness; Pitting corrosion; Residual stresses

1. Introduction

Several situations arise in industrial practice which callfor joining of dissimilar materials. The materials employedare location dependent in the same structure for effectiveand economical utilization of the special properties of eachmaterial. The joining of dissimilar metals is generally morechallenging than that of similar metals because of differencein the physical, mechanical and metallurgical properties ofthe parent metals to be joined. In order to take full advan-tage of the properties of different metals it is necessary toproduce high quality joints between them. Only in this waycan the designer use most suitable materials for each part ofa given structure. The growing availability of new materialsand higher requirements being placed on materials creates agreater need for joints of dissimilar metals.

Joining of ferritic stainless steels are faced with the prob-lem of coarse grains in the weld zone and heat affectedzone of fusion welds and consequent low toughness andductility due to the absence of phase transformation duringwhich grain refinement can occur [1,2]. In general austeniticstainless steels are easily weldable. When austenitic stain-less steel joints are employed in cryogenic and corrosiveenvironment the quantity of ferrite in the welds must beminimized/controlled to avoid property degradation during

Corresponding author. Tel.: +91 40 444 0051; fax: +91 40 444 0683.

service [3,4]. In addition these steels are prone to sensiti-zation of their fusion welds. These problems have been ad-dressed by solid state welding processes, such as frictionwelding [36].

Dissimilar metal combination between ferritic stainlesssteels and austenitic stainless steels is commonly employedin TiCl4 reduction retorts. This calls for welding of thecombination. Such transition joints are necessary becauseaustenitic stainless steels with superior creep strength andoxidation resistance are required in the higher temperatureregions, while ferritic stainless steels to avoid the problemof nickel leaching by molten magnesium. Welding of fer-ritic to austenitic stainless steels is considered to be a majorproblem due to difference in coefficient of thermal expan-sion, which may lead to crack formation at the interface,formation of hard zone close to the weld interface, relativelysoft regions adjacent to the hard zone; large hardness differ-ence between the hard and soft zones and expected differ-ences in microstructure may lead to failures in service [79].Solid state welding is a possible solution for these problems.This paper reports on a study that has been taken up to de-velop an understanding on the friction welding characteris-tics of austenitic stainless steelferritic stainless steel dissim-ilar metal welds. Detailed microstructural examination in thedifferent regions of the welds and a correlation between themicrostructure and mechanical properties and correspondingfracture behaviour forms the goal of the study and thereforeassumes special significance since such detailed studies are

0924-0136/$ see front matter 2004 Published by Elsevier B.V.doi:10.1016/j.jmatprotec.2004.05.017

V.V. Satyanarayana et al. / Journal of Materials Processing Technology 160 (2005) 128137 129

not hitherto reported. Parameter optimization also forms amajor contribution of the study.

2. Experimental

2.1. Parent metals

The parent metals employed in this study are AISI 304austenitic stainless steel and AISI 430 ferritic stainless steel.The composition and mechanical properties of the startingparent metals are given in Tables 1 and 2, respectively. Theaustenitic stainless steel contains equiaxed grains of austen-ite with occasional twinning, while the ferritic stainless steelconsists of coarse and elongated grains of ferrite (Fig. 1).For friction welding studies the material employed was in

Table 1Chemical composition of parent metals (wt.%)Composition C Si Mn Cr S P Ni

AISI 430 0.06 0.4 0.4 17.0 0.03 0.04 AISI 304 0.06 0.32 1.38 18.4 0.28 0.4 8.17

Table 2Mechanical properties of parent metals

Property Ultimate tensilestrength (MPa)

Yield strength(MPa)

Elongation (%)

AISI 430 488 380 28AISI 304 600 250 58

Fig. 1. Optical microstructure of parent metals: (a) austenitic stainlesssteel (AISI 304); (b) ferritic stainless steel (AISI 430).

Table 3Experimental design matrix

Trial run Friction force(kN) (X1)

Forge force(kN) (X2)

Burn-off(mm) (X3)

AF1 4 8 3AF2 6 8 3AF3 4 12 3AF4 4 8 5AF5 6 12 3AF6 6 8 5AF7 4 12 5AF8 6 12 5

the form of rods of 18 mm diameter obtained from the platematerial.

2.2. Friction welding

Welding was performed on a continuous drive frictionwelding machine at a speed of 1500 rpm in a continuouslyand step less variable speed machine of 15 kN capacity. Inthe continuous drive friction welding process a stationarymember is pressed against a rotating member with an axialpressure. The relative motion generates frictional heat whichcauses the material to soften and plastically deform. Af-ter a preset displacement (known as burn-off) has occurred,the machine is rapidly braked, and the pressure is increasedto generate a high quality solid state weld. During weld-ing the primary parameters (friction force, forge force, rota-tional speed and displacement) were continuously monitoredand recorded. To arrive at suitable welding parameters tri-als were carried out with the same parameters as employedfor ferriticferritic stainless steel and austeniticausteniticstainless steel [5,6]. Examination of the joints revealed de-fects like delaminations (Fig. 2). Few more trials were car-ried out with different parameters in order to get defect freewelds. Based on this a 23 factorial design of experiments wasadopted [10]. The experimental matrix is given in Table 3.The main parameters employed are friction force, forge forceand burn-off (length loss during friction/forge stage). Trialwelds were made by varying one parameter keeping otherparameters constant to find the limits.

2.3. Metallography

Low magnification stereo microscope of Leitz make wasemployed for observing the bead shape. To observe deforma-tion and microstructural features Leitz optical microscopewas employed. Fractographic examination was carried outunder a Leo scanning electron microscope.

2.4. Mechanical testing

The mechanical tests consist of Vickers hardness, notchtensile test and Charpy V notch impact testing. Hardnessmeasurements included hardness survey across the interface.In notch tensile and impact tests the notch was located at the

130 V.V. Satyanarayana et al. / Journal of Materials Processing Technology 160 (2005) 128137

Fig. 2. Typical micrograph showing delamination.

interface. The location of notch is positioned at the center ofthe weld as shown in Fig. 3. Standard specimen configura-tions were employed for notch tensile and impact toughnesstesting (Fig. 4). Plain tensile specimen was not included asfailures occurred outside the weld (Fig. 5) and data wouldyield only the parent metal properties.

2.5. Residual stress measurement

Residual stress measurements were carried out across theinterface employing X-ray stress measurement technique.The equipment employed for the purpose was model X 2002of American stress technologies. Before measurements theflash was machined and the surfaces were polished elec-trolytically. Residual stresses were measured in the weldbead center region using peak shift sin2 technique with CrK radiation.

2.6. Corrosion testing

The weld joints and the parent metals were tested for pit-ting corrosion in an electrolyte of 0.5 M H2SO4 + 0.5 M

HCl. The electrochemical measurements were made usinga potentiometer dedicated for the purpose. Steady state po-tential was recorded 10 min after immersion of the samplein to the electrolyte and the potential was raised anodicallyusing scanning potentiostat at a scan rate 2 mV/s. The poten-tial at which the current increases abruptly after the passiveregion was taken as pitting potential (Epit) [11]. Specimensthat exhibit more positive potential value were consideredto be those having better pitting resistance.

2.7. Statistical analysis of the data

The mechanical property data were subjected to statis-tical analysis to understand the influence of individual ef-fects of the parameters and their interactive effects on theproperties. ANOVA technique of Yates algorithm was em-ployed to study the significance of coefficients [10]. Re-gression equations were obtained from this analysis. Theregression equations enabled to understand the influence ofthe friction welding parameters on mechanical properties.Correlation coefficient was also obtained from the statisticalanalysis.


Fig. 3. Schematic diagram showing location and orientation of: (a) notchtensile sample; (b) Charpy V notch impact sample in bond zone.

Fig. 4. Configuration of: (a) notch tensile specimen; (b) Charpy V notch impact specimen.

Fig. 5. Appearance of friction welded plain tensile test specimen after test.

3. Results

3.1. Metallography and visual examination

A view of friction weld joints shown in Fig. 6 exhibitshigher flash with increase in burn-off and forge pressure.The flash was observed to be from ferritic stainless steel andaustenitic stainless steel did not participate in the flash for-mation suggesting deformation is mainly limited to ferriticstainless steel side. Typical cross-sectional view of the weld(AF7) and microstructure details at the center and peripheryare presented in Fig. 7 shows that the deformation is mainlyconfined to ferritic stainless steel. The central region con-sists of fine grains while peripheral region consists of coarsegrains.

The microstructure in the central region of the weldsshown in Fig. 8 indicate that all the welds consist of fineequiaxed grains of ferrite. The grain size is in general


Fig. 6. Visual view of friction welded ferriticaustenitic stainless steeljoints.

coarser in the low forge pressure welds as well as in highfriction pressure welds. It is to be noted that forge pres-sure aids in grain refinement while, friction pressure aidsgrain coarsening (Fig. 8). Typical microstructural featuresin various regions of the weld across the interface are

Fig. 7. Typical friction weld and its cross-sectional view (AF7).

shown in Fig. 9. The central region consists of equiaxedgrains and is confined to ferritic stainless steel. Adjacentto this region bent and elongated grains are observed onthe ferritic stainless steel side. The austenitic stainless steelside consists of parallel banded features adjacent to thecentral equiaxed grain structure at the interface. The inter-face is narrow and straight in high burn-off welds. At lowfriction and low burn-off parallel bands are observed inthe austenitic stainless steel side adjacent to the interface(Fig. 10).

3.2. Mechanical properties

Hardness, notch tensile strength (NTS) and Charpy Vnotch impact toughness data of the welds are presented inTable 4. The hardness of the weld region is in the range195 Hv (min) to 270 Hv (max). Highest hardness (265 Hv)and lowest hardness (195 Hv) were obtained at low burn-off


Table 4Mechanical properties of AISI 430AISI 304 friction welds

Run Parameter (X1X2X3) Notch tensile strength (MPa) Impact toughness (J) Hardness at centerTrial 1 Trial 2 Trial 1 Trial 2 Trial 1 Trial 2

AF1 483 668 630 15 17 247 236AF2 683 647 635 18 19 208 195AF3 4123 610 629 16 17 214 210AF4 485 663 607 18 20 222 230AF5 6123 640 669 21 18 270 265AF6 685 600 628 14 17 208 215AF7 4125 616 619 25 28 242 235AF8 6125 689 637 22 25 231 229

Fig. 8. Microstructure of friction welds at center: (a) AF1; (b) AF2; (c) AF3.

Fig. 9. Micrographs of friction welds at center (4125).


Fig. 10. Micrographs of the friction weld at the interface: (a) AF6 (685);(b) AF3 (4123).

and at high and low forge pressures, respectively. Impacttoughness was on the higher side at high burn-off whilenotch tensile strength followed a reverse trend. Notch ten-sile strength ranged from a minimum of 600 MPa to a max-imum of 689 MPa while impact toughness was in the range1528 J. Typical hardness distribution across the weld (AF7)shown in Fig. 11 reveals that hardness is higher on austeniticstainless steel side of the interface.

Regression analysis of the data in the form of regressionequations is presented in Table 5. It is observed that indi-vidual parameters of friction welding do not have an influ-ence on notch tensile strength and only friction and forgepressure exhibit an interactive effect. Burn-off has an influ-ence on impact toughness while forge and burn-off have aninteractive effect. Hardness is influenced by forge pressureand interactive effect of other parameters. The correlation

Table 5Regression equations for response function and their coefficient of correlation

S. no. Response Regression equation Coefficient of correlation

1 Notch tensile strength Y = 636.6 + 13.68X1X2 0.7072 Impact toughness Y = 18.75 + 2.38X3 + 2.388X2X3 0.843 Hardness Y = 228.56 + 8.44X2 + 12.69X1X2 4.81X1X3 11.89X1X2X3 0.98X1friction force; X2forge force; X3burn-off.

Fig. 11. Microhardness traverse along the bond line of a typical weld(4125).

coefficient is about 0.98 in respect of hardness, 0.84 in thecase of impact toughness and 0.7 in respect of notch tensilestrength.

3.3. Fractography

The fractographs for low and high notch tensile strengthare presented in Fig. 12. At low notch tensile strength thefracture features are of cleavage while for high NTS the fea-tures of fracture contain ductile micro-voids. Similar trendswere noted in the impact specimens.

3.4. Residual stresses

Residual stress data at interface on the austenitic stainlesssteel side are tabulated in Table 6. The magnitude of thestresses range from a minimum of 180 MPa to a maximumvalue of 260 MPa. On an average the stresses are maximumat high burn-off (241 MPa average), while they were low(221 MPa) at low burn-off. A consistent trend is noticedalthough the differences are marginal. Typical stress distri-bution across a weld (AF7) shown in Fig. 13 reveals thatthe stresses are higher on the austenitic stainless steel sideof the interface. Weld center also has a stress value almostequal to ferritic stainless steel side of the interface. Thestress distribution trends follow hardness distribution acrossthe weld (Fig. 11).


Fig. 12. Fractographs of notch tensile samples: (a) AF6 (685)low notch tensile strength; (b) and (c) AF8 (6125)high notch tensile strength.

Table 6Residual stress at the interface of austenitic stainless steel and weld

Run Parameter (X1X2X3) Residual stress (MPa)AF1 483 210AF2 683 260AF3 4123 180AF4 485 230AF5 6123 235AF6 685 250AF7 4125 225AF8 6125 260

3.5. Pitting corrosion

A typical polarization curve is shown in Fig. 14. Thepitting potential (Epit) was used as a measure of resistance topitting. Epit values for various types of welds are presented

Table 7Pitting potential of friction weld joints and parent metalsType of joint Pitting potential (mV)Ferriticferritic 1023Austeniticaustenitic 940Austeniticferritic 931Parent metal AISI 304 914Parent metal AISI 430 988

in Table 7. Pitting corrosion studies indicate that among thewelds ferriticferritic stainless steel joints exhibit highestpitting resistance (Epit 1023 mV) while the dissimilar jointsexhibit least resistance (Epit 931 mV). Similar metal weldsexhibit marginally superior performance compared to thecorresponding parent materials.

Fig. 13. Residual stress distribution traverse across the bond line of atypical weld (4125).


Fig. 14. Typical polarization behaviour dissimilar austeniticferritic stain-less steel weld.

4. Discussion

4.1. Deformation and microstructure

The different thermal and physical properties of the ma-terials to be welded in dissimilar metal welding (heat ca-pacity, thermal conductivity, relation between hardness andtemperature) generally results asymmetrical deformation.Austenitic stainless steel have lower thermal conductivityand greater hardness at higher temperatures compared to fer-ritic stainless steels. For this reason austenitic stainless steeldoes not undergo extensive deformation while ferritic stain-less steel specimen undergoes extensive deformation. Thesame phenomenon has been reported during friction weldingof dissimilar welds namely aluminium to copper, titaniumto steel, aluminium to steel, etc. [1216]. The formation ofupset collar (flash) on the ferritic stainless steel side onlyis due to low strength of the ferritic stainless steel. A nar-row zone of deformation bands (Fig. 7) on the austeniticstainless steel suggests that this region undergoes deforma-tion although it does not take part in the upset collar sug-gesting that the deformation is not extensive. The view thatonly ferritic stainless steel takes part in upset collar forma-tion is substantiated by the shortening of the ferritic stainlesssteel rod only. The zone of deformation band width is widerat low burn-off than at high burn-off could be attributedto lower heat content and high flow stress of the region atlow burn-off than when the burn-off is high that aids in thespread of heat resulting in increased heat content and con-

Table 8Comparison of notch tensile strength and impact toughness of similar and dissimilar combinations of austenitic and ferritic stainless steel

Material Austenitic stainless steel Ferritic stainless steel Austeniticferritic stainless steel

Notch tensilestrength (MPa)

Impacttoughness (J)


Impacttoughness (J)


Impacttoughness (J)

Parent metals 830896 213214 547590 67 Welds 693753 71148 660784 8.530 600697 1628

sequent lowering of flow stress in this region. The coarsergrain structure observed in low forge pressure combinationcan be attributed to lower degree of working of the materialthan at high forge pressure that results in higher degree ofworking. The central region consists of fine grains, while theperipheral region consists of coarse grains (Fig. 7). The finegrain size at the central region is due to dynamic recrystal-lization. The temperature of the peripheral region would behigher [17] and therefore exhibits coarse grain size. Despitethe coarse grain structure at the periphery, the overall bondregion remains stronger than the parent material (Fig. 7), asevidenced by the plain tensile test results shown in Fig. 5.

4.2. Mechanical properties and regression analysis

The mechanical and thermo-physical properties of dis-similar substrates will have a major influence on the proper-ties of the dissimilar joints because the temperature attainedby each substrate markedly depends on the thermo-physicalproperties of the two substrates and on the joining parame-ters selected. Consequently, the flow stresstemperature re-lations for each substrate will have an important influenceon the joint properties produced during friction welding.

In general high forge pressures resulted in high tough-ness and notch tensile strength. Fine grain structure exhib-ited high strength and low toughness while coarse grain mi-crostructure exhibited a reverse trend. Fracture features ofthe notch tensile and impact specimens are as per expectedtrends, in that at low strength and toughness the fractureis predominantly cleavage as against high energy ductilemicro-void fracture when the strength and toughness arehigh. It is opined that the grain size in the weld region dic-tates the mode of fracture in that fine grain gives raise toductile fracture while, coarse grain promotes cleavage frac-ture.

A comparison of properties of similar material combina-tion welds with the dissimilar combination welds (Table 8)shows that the properties of austenitic stainless steelferriticstainless steel compare well with the ferritic stainless steelwelds. This further substantiates that the deformation is con-fined to ferritic stainless steel and the interface propertiesare dictated by the properties of ferritic stainless steel alone.

Analysis of mechanical property data suggest that hard-ness is dictated by forge pressure and an interactive effectof friction and forge pressure while toughness is controlledby the interaction of friction pressure and burn-off. Thissuggests that toughness is dictated by the heat content


and spread of heat. Higher burn-off aids in spread of heatdue to the availability of friction time. These conditionsaid in stabilizing the microstructure as well and there-fore favour toughness improvement at higher burn-off.Hardness and notch tensile strength are mostly influencedby an interactive effect of friction and forge pressures,which are a combination of heat input and degree ofworking.

4.3. Pitting corrosion

The excellent improvement in the pitting corrosion re-sistance of the welds compared to parent metal couldbe due to the presence of dynamically recrystallized mi-crostructure and possible composition uniformity resultingfrom faster cooling rates that do not favour elementalsegregation.

4.4. Residual stresses

Higher residual stresses at the austenitic stainless steelside of the interface can be attributed to the higher flowstress of the austenitic stainless steel aided by higher co-efficient of thermal expansion of the austenitic stainlesssteel. Higher flow stress of austenitic stainless steel re-sulted in the deformation confined to a very narrow regionconsisting of deformation bands. These deformation bandsare likely to consist of high density of dislocations andhence the observed higher residual stress peak is con-fined to this region. Penetration depth of the X-rays is ofthe order of 20m and the stresses measured are within15 MPa.

5. Conclusions

(i) Continuous drive friction welding has been used to suc-cessfully join austeniticferritic stainless steel.

(ii) In friction welding of austeniticferritic stainless steel,deformation is confined to ferritic stainless steel only.

(iii) Higher forge pressure combinations exhibit fine grainsize and increased friction pressure aids in grain coars-ening.

(iv) The mechanical properties of austeniticferritic stain-less steel welds are similar to ferritic stainless steelwelds.

(v) The toughness and strength properties of dissimilarmetal welds are better than ferritic stainless steel parentmetal.

(vi) Notch tensile strength, hardness and impact toughnesscan be expressed in terms of the process parameters byregression equation obtained by statistical analysis.

Acknowledgements

The authors express their gratitude to Defence Researchand Development Organisation for the financial support tocarry out this programme. The authors are thankful to Dr.D. Banerjee Director, DMRL for his continued encourage-ment. One of the authors (V.V. Satyanarayana) is thankfulto the Principal and the management of Vasavi College ofEngineering, Hyderabad for their continued support duringthis work.

References

[1] F.B. Pickering, Physical metallurgy of stainless steel developments,Int. Met. Rev. 21 (1976) 227268.

[2] S. Kou, Welding Metallurgy, Wiley, 1987, pp. 383386.[3] W.A. Baeslack, D.J. Puquette, W.F. Savage, Effect of ferrite on stress

corrosion cracking in duplex stainless steel weld metals at roomtemperature, Corrosion 34 (1979) 46.

[4] E.R. Szumachowki, H.F. Reid, Cryogenic toughness of SMAaustenitic stainless steel weld metals part-1 role of ferrite, Weld. J.57 (1978) 325s333s.

[5] V.V. Satyanarayana, G. Madhusudhan Reddy, T. Mohandas, G.Venkata Rao, Continuous drive friction welding studies on AISI 430ferritic stainless steel welds, Sci. Technol. Weld. Join., in press.

[6] V.V. Satyanarayana, G. Madhusudhan Reddy, T. Mohandas, Contin-uous drive friction welding studies on AISI 304 austenitic stainlesssteel welds, Mater. Manuf. Process., communicated June 2003.

[7] S. Allabhakshi, G. Madhusudhan Reddy, V.V. Ramarao, C. PhaniBabu, C.S. Ramachandran, Studies on weld overlaying of austeniticstainless steel (AISI 304) with ferritic stainless steel (AISI 430),National Welding Conference, Chennai, India, January 2002, IndianInstitute of Welding, Paper 8

[8] A. Omar, Effects of welding parameters on hard zone formation atdissimilar metal welds, Weld. J. 77 (2) (1998) 86s93s.

[9] Welding Hand Book, 7th ed., vol. 4, American Welding Society,1984, pp. 93128.

[10] Kempthorne, Design and Analysis of Experiments, Wiley, New York,1966, pp. 234244.

[11] J.W. Oldfield, Test techniques for pitting and crevice corrosion resis-tance of stainless steels and nickel-base alloys in chloride-containingenvironments, Int. Mater. Rev. 32 (3) (1987) 162169.

[12] T.J. Jessop, Friction welding of dissimilar metal combinationsaluminium and stainless steel, Weld Institute Research Report,November 1995, 7375.

[13] V.V. Trutnev, The friction welding of steel to aluminium alloy VAD-1,Svar. Proi. 20 (37) (1973) 1819.

[14] S.Y. Bekir, Z.S. Ahmet, K. Nafiz, Z. Ahmed, Friction welding ofSt-Al and Al and Cu materials, J. Mater. Process. Technol. (49)(1995) 431443.

[15] D. Yashan, S. Tsang, W.L. Johns, M.W. Doughty, Inertia frictionwelding of 1100 aluminium to type 316 stainless steel, Weld. J. (8)(1987) 27s37s.

[16] K.K. Wang, Friction Welding Report WRC Bulletin 204, WeldingResearch Council, New York, 1972.

[17] S. Fukumoto, H. Tsubakino, M. Aritoshi, T. Tomita, K. Okita, Dy-namic recrystallisation phenomena of commercial purity aluminiumduring friction welding, Mater. Sci. Technol. (18) (2002) 219225.

Dissimilar metal friction welding of austenitic-ferritic stainless steelsIntroductionExperimentalParent metalsFriction weldingMetallographyMechanical testingResidual stress measurementCorrosion testingStatistical analysis of the data

ResultsMetallography and visual examinationMechanical propertiesFractographyResidual stressesPitting corrosion

DiscussionDeformation and microstructureMechanical properties and regression analysisPitting corrosionResidual stresses

ConclusionsAcknowledgementsReferences

19. VV SATYANARAYANA 2005 Dissimilar metal friction welding of austenitic–ferritic stainless steels

Documents