Introduction

Austenitic stainless steels have beenwidely used in nuclear power plants fordifferent applications such as superheatersand heater components, and they are alsoused in cryogenics and pressure vessels.These steels perform well at elevated tem-peratures and are applied extensively forsteam pipes and exhaust systems. How-ever, there are many difficulties in joiningthese steels using fusion welding methods.Despite the fact that there is some re-search on joining stainless steel using gastungsten arc welding (GTAW), laser beamwelding (LBW), and plasma arc welding(PAW), work continues on attaining a re-liable joint. A concern, when welding theaustenitic stainless steel by conventional

welding methods, is the susceptibility tosolidification and liquation cracking (Refs.1–4). In many instances, the formation ofbrittle intermetallic phases in the diffusionzone leads to unfavorable changes in themechanical and physical properties of themetallic bonds (Refs. 5–7). The researchwork presented in this study concerns thetransient liquid phase diffusion brazing ofstainless steel 304 using a pure copper in-terlayer. Copper as the interlayer does notform brittle intermetallic compounds withiron, and its melting point is lower than Feand Ni. Thus, the flowability increases athigher brazing temperatures and encour-ages a suitable contact between the fayingsurfaces. The brazing variables and theireffect on microstructural changes havebeen investigated using optical and scan-ning electron microscopy and energy dispersive spectrometry elemental analy-

ses. In addition, the brazing mechanismwas explained using numerical methods.The analysis was coupled with a corrosiontest, considering the effects of the brazingparameters.

Experimental Procedure

The stainless steel was received in theform of plate 2 mm thick, and the chemi-cal composition as shown in Table 1. Theplates were cut using an abrasive cuttingsaw to the dimensions of 2 × 10 × 20 mm.The mating surface of the stainless steelwas prepared by conventional grinding on1200-grade silicon carbide paper followedby polishing using diamond paste. Thespecimens then were cleaned in an ultra-sonic bath using acetone for 15 min anddried in air. A copper foil (50 μm thick,99.95% purity) was used as intermediatematerial, and the surface of the interlayerwas polished in the same fashion as it doesto the base metal. A stainless steel fixturewas designed to fix the specimen and holdthe sandwich assembly during the metallicbrazing process. The brazing process wasperformed in a vacuum furnace, and thesubstrate and interlayer contact area wasenhanced by a pressure of 0.5 MP at setbrazing temperatures. The brazing cycle isshown in Fig. 1. The test specimens wereheated to the brazing temperature, thenleft in the furnace for a variable holdingtime and heating rate, and finally cooledto room temperature. The brazing param-eters for the copper interlayer are shownin Table 2. The specimens were preparedby grinding on 240- to 2400-grade SiCpaper, then etched using 2 g FeCl3, 24 mLdistilled water, and 6 mL HCl. Mi-crostructural observations were conductedusing an optical microscope (Nikon mi-crophot-FXL), scanning electron micro-scope (SEM, PHILIPS XL 40) usingbackscattered mode, and energy-disper-sive spectrometry (EDS). Moreover, thecorrosion test was carried out in a 3.5%

SUPPLEMENT TO THE WELDING JOURNAL, MARCH 2013Sponsored by the American Welding Society and the Welding Research Council

Transient Liquid Phase Diffusion Brazingof Stainless Steel 304

The results of different brazing temperatures and holding times were compared to determine the best condition

BY M. MAZAR ATABAKI, J. NOOR WATI, AND J. IDRIS

KEYWORDS

Stainless SteelTransient Liquid Phase Diffusion BrazingMicrostructure

M. MAZAR ATABAKI (mmazaratabak@smu.edu, mamehdi2@live.utm.my) is withResearch Center for Advanced Manufacturing(RCAM), Department of Mechanical Engineer-ing, Southern Methodist University, Dallas, Tex.,and Department of Materials Engineering, Fac-ulty of Mechanical Engineering, UniversitiTeknologi Malaysia, Malaysia. J. NOOR WATIand J. IDRIS are with Department of MaterialsEngineering, Faculty of Mechanical Engineering,Universiti Teknologi Malaysia, Malaysia.

ABSTRACT

Transient liquid phase diffusion brazing was employed to join stainless steel 304using pure copper foil as the interlayer. The brazing process was carried out in a vac-uum furnace at various temperatures for a range of times. The joints were studiedwith optical and scanning electron microscopy, energy-dispersive spectrometry, andcorrosion testing. The diffusion of the main elements from the interlayer and basemetal into the braze line and brazing-affected areas was the main controlling factorpertaining to the microstractural evolution of the joint interface. The presence of eu-tectoid γFe+ eutectic Cu + Cr and γFe (Cr, Ni) intermetallic was detected at the in-terface of the joints brazed with copper interlayer. The average displacement of thesolid/liquid interface as a function of time was found to be about 0.36 of the brazingtime. The diffusivity of copper in the grain boundary of the stainless steel was foundto be around 56 × 105 times higher than the lattice diffusivity at the interface of thejoint, showing copper as a melting point depressant has the ability to produce grainboundary grooves facilitating the diffusion of the copper atoms.

57-sWELDING JOURNAL

WE

LD

ING

RE

SE

AR

CH

Atabaki 3-13_Layout 1 2/14/13 4:36 PM Page 57

NaCl solution in order to determine thecorrosion resistance of the bonds. To esti-mate the reliability of the joints, a sheartest was applied using an Instron tensiletest machine with a crosshead speed of 1mm/min.

Results and Discussion

Microstructure Analysis of 304SS-Cu-304SS

Figure 2A–C shows the optical mi-crostructure of the brazed specimens at900°, 950°, and 1000°C for 20 min holdingtime, respectively. It is observed that a cer-tain amount of diffusion occurs betweenthe interlayer and two substrates. A thindiffusion layer was revealed parallel to thejoint interface on the stainless steel side

for all joints. However, it is interesting tonote that the stainless steel after exposureto the brazing temperature exhibits differ-ent microstructure by emerging twins onthe surface of the steel — Fig. 2A. This ex-posure to the brazing temperature also re-sults in the formation of complex carbideswithin the austenite grains. This leads toan impoverishment of chromium in theaustenite solid solution.

The interface of the stainless steel andcopper consisted of a continuous reactionlayer free from voids on both sides. It wasobserved that the copper side showed theabsence of any recognizable diffusionzone. Images of the joints prepared at900°, 950°, and 1000°C for 20 min holdingtime, shown in Fig. 3A–C, were taken byscanning electron microscope in the back-scatter mode. It can be seen that grainboundary grooving increases the diffusion

of the main alloying elements due to thecurl shape liquid/solid interface and highergrain boundary diffusivity. The effect ofthe grain boundary grooving was indicated(Refs. 8, 9) considering a significant dif-ference in a predicted brazing time withdifferent grain sizes. Accordingly, if the in-terfacial curvature enhances the solute,solubility would be increased dependingon the stainless steel grain size.

Therefore, increasing the interfacial cur-vature at the higher temperatures is the rea-son for enhancing the solute influx into thesubstrate at both sides of the joint, especiallyin the diffusion zone. After the grain bound-ary grooving, the base metal dissolves intothe partial molten copper interlayer untilthe stainless steel in the molten partial meltreaches its equilibrium state at the brazingtemperature. While the partial melt touchesthe stainless steel, a high-carbon iron phaseformed and, as a result, the high-carbon ironphase in the molten copper interlayer ex-ceeded its equilibrium quantity. This phasethen deposits as a columnar dark structureof Fe-Cu-C phase. This phase grew increas-ingly with heating time and finally solidifiedat the interface of the base metal and interlayer.

Therefore, the process of dissolution

MARCH 2013, VOL. 9258-s

WE

LD

ING

RE

SE

AR

CH

Fig. 1 — Transient liquid phase diffusion brazing cycle and thedimensions of the brazing specimens.

Table 1 — Chemical Composition of the Materials (wt-%)

Material C Fe Mn Si P S Cr Ni N Cu

304 SS 0.08 Bal 2.0 0.75 0.045 0.03 20.0 10.5 0.1 —

Copper — — — — — — — — — 99.99Interlayer

Fig. 2 — The optical microstructure of the joints prepared at A — 900°C; B — 950°C; C — 1000°C for 20 min, using copper interlayer.

A

B C

Atabaki 3-13_Layout 1 2/14/13 4:36 PM Page 58

and solidifying the dark phase can be ex-plained by the distortion of the stainlesssteel structure due to the diffusion of Cuinto the steel. This causes the separation ofthe formed Fe-Cu solid solution clusters —Fig. 4. In order to determine the chemicalcomposition of the brazing interface, en-ergy-dispersive spectrometry (EDS) analysiswas performed using spot analysis takenfrom the stainless steel and copper inter-layer. From the observations, various diffu-sion zones were found in the joint area andbase metal. The brazing-affected zones canbe classified in terms of the shape and loca-tion of each phase. The area can be dividedinto three distinct zones, which are basemetal, diffusion zone, and interlayer, asshown in Fig. 5. Compositional profiles in-dicating the distribution of the alloying ele-ments from the base metal and interlayer atthe centerline are taken as shown in AB linein Fig. 5. The compositional profile for thespecimens brazed at 900°, 950°, and 1000°Care presented in Fig. 6.

Different rates of elemental distributionlike the distribution of Fe, Cr, and Cu alongthe joint is due to the different diffusion co-efficients. At the interlayer surface, the sidewhich is rich in Cu, however, only a smallamount Fe and Cr diffused from the basemetal. It can be seen that the weight-percentof Fe and Cr in the interlayer significantlydecreased while the weight-percent of Cuconsiderably increased in the base metal.The decreasing of Fe at the interface is dueto dissolution of the base metal. To furtheranalyze the distribution of Fe and Cr fromthe base metal to the interlayer and Cu fromthe interlayer to the base metal, Cu-Fe andFe-Cr-C phase diagrams were applied fordescribing the effect of the melting point de-pressant on the mechanism of the brazingprocess. A very low amount of Fe was de-tected in the copper interlayer due to thelimited solubility of Fe in copper at the tem-perature range of 900° to 1000°C. The cop-per started to dissolve a little amount of Feat the stainless steel/copper interlayer inter-face and continuously diffuses into the

interlayer. The diffusion of Cu into the

stainless steel produces a solidsolution in a very limited solu-bility. In addition, the solubilityof Cu in Fe was enhanced whenthe brazing temperature wasincreased. It can be assumedthat the diffusion layer at thebraze consists of γFe + (Cu).The element distribution analy-sis revealed that at any of thebrazing temperatures, Cutransfered a long distance inthe stainless steel side, whileFe, Cr, and Ni transverse a

59-sWELDING JOURNAL

WE

LD

ING

RE

SE

AR

CH

Table 2 — Bonding Process Parameters

Brazing Parameters

Low Brazing Temperature (°C) 850, 900, and 950Predry Time (min) 24

Rate (°C/min) 25

High Brazing Temperature (°C) 900, 950, and 1000Holding Time (min) 16, 20, 24, and 72Cooling Time (min) 10Vacuum (mm Hg) 740

Vacuum Start (mm Hg) 740Vacuum Release (mm Hg) 740

Table 3 — Average of the Diffusion-Affected Zone Thickness in the Joints Prepared at 900°,950°, and 1000°C at Various Holding Times

Temperature (°C) 900 950 1000

Holding Time 16 20 72 16 20 72 20 72(min)

X1 (μm) – 9.286 9.657 15.15 10.013 14.8 16.385 7.581 72Diffusion Zone

X2 (μm) – 47.129 46.933 46.05 45.113 43.357 43.24 42.355 7.854Interlayer

X3 (μm) – 7.996 8.664 11.91 6.257 9.748 13.895 5.957 8.432Diffusion Zone

Predicted XD.Z 8.245 8.996 13.451 8.452 13.765 14.673 6.941 41.121(μm)

Fig. 3 — SEM-BSE images of the joints prepared at A — 900°C; B — 950°C; C — 1000°C for 20 min holding time.

A B

C

Atabaki 3-13_Layout 1 2/14/13 4:36 PM Page 59

comparatively smaller distance in the Cu in-terlayer side.

Interface Characterization

For characterizing the diffusion zoneand diffusion-affected areas, images weretaken from the joint area at the stainlesssteel/copper interlayer interface as shown inFigs. 8 and 9. Moreover, a quantitativeoverview of the chemical composition of dif-ferent regions (region A to E) of the diffu-sion zone for the specimen brazed at 900°Cillustrated in Fig. 7A shows a dark-shadedregion (point B) and a light-shaded region(point C) at the interface. By applying EDSanalysis, it is shown that regions B and C areenriched with Fe and Cr with small quanti-ties of Ni, Mn, and Cu. The different con-trast occurs due to the dissimilarity inconcentration of Fe and Cr into the inter-layer and Cu into the base metal. The white

island (region A) found inthe copper interlayer is en-riched with Fe and Cr. Fig-ure 7B shows a significantdifference in the joint com-pared to the joints shown inFig. 7A, presenting a smalldiffusion layer. Copper isdissolved (regions D and E)due to infiltration of the liq-uidated copper at 1000°C.Figure 8 shows a schematicdiagram describing the pos-sible brazing mechanismand the changes of the phases from the be-ginning of the process until 20 min of hold-ing time.

At the initial stage, during the brazingprocess, the copper interlayer partially liq-uefied and then stainless steel dissolved bymigrating the liquid copper to the basemetal. As a result, some grain boundary

grooves appeared in the steel in the vicinityof the copper interlayer, and Fe-Cr from thebase metal flows into the diffusion zone. Atthe same time, Cu atoms of the melt diffuseinto the dissolved region. This is known as arelationship of the solid metal dissolutionand liquid metal migration.

While the liquid interlayer touches thesurface of the base metal, the migration of

MARCH 2013, VOL. 9260-s

WE

LD

ING

RE

SE

AR

CH

A B

C

Fig. 4 — Binary alloy phase diagrams. A — Cr-C; B — Fe-Cu; C — Cu-C. Fig. 5 — Three distinct zones in the joints, showing the solute infiltratedinto the grain boundaries of the base metal. This is more obvious in thediffusion zone.

Fig. 6 — Elemental distribution for the joints brazed at A — 900°C; B —950°C; C — 1000 °C for 20 min holding time.

B

A

C

Atabaki 3-13_Layout 1 2/14/13 4:37 PM Page 60

the melting point depressant to the basemetal occurs by diffusing into the grainboundary of the base metal (Ref. 10). Thegrain boundary grooving as a result ofheavy diffusion of the melting point de-pressant adjacent to the interfacial regioncaused great infiltration of copper into thestainless steel. The liquid copper changesinto Cu phase and reacts with Fe and Cralong the interface. As described, the dif-fusion coefficients of Fe, Cr, and Cu arequite different. For example, the diffusioncoefficient of Cu into Fe and Fe into Cu at920°C is 2.2 × 10–13 and 4.63 × 10–10 cm2/s,respectively (Ref. 11). By increasing thedissolution of the base metal with apply-ing a longer time (16 min), the phase is en-riched with Cu and Fe — Fig. 8B.

On the other hand, it was observed thatthe Fe and Cr atoms that flowed from thebase metal were transformed into a Fe-Crphase. By increasing the holding time to20 min, the amount of Fe-Cr phase showsa tendency to increase — Fig. 8C.

Widening of Diffusion Zone

The diffusion zone was widened duringthe brazing operation. The mechanism ofthe diffusion zone widening is studied bymeasuring the thickness of the diffusionarea on both sides of the interlayer with an

initial thickness of 50μm. The thickness ofthe diffusion zonesfor the joint preparedat 950°C for 16 min isshown in Fig. 9. Itcan be seen that thediffusion zone ad-vanced into the stain-less steel. Table 3shows the averagethickness of the dif-fusion zone for thejoints fabricated at900°, 950°, and1000°C for variousholding times.

As the morphol-ogy of the diffusionzone was not uniform, thickness of the dif-fusion-affected zones was taken at threedifferent locations. It was seen that thethickness of the diffusion zone increasedwith increased holding time. Generally,mass transfer has to be extended, depend-ing on the brazing temperature. By in-creasing the brazing temperature more,atoms migrated across the interface, hencethe diffusion zone widened. By approach-ing the melting temperature of the copperinterlayer (900°–950°C), Cu atoms werestimulated to move faster and in larger

quantity. Therefore, it is enough for Cuatoms to vibrate and give possibility to Feand Cr atoms to diffuse into the interlayer.If the growth of the diffusion zone as-sumed to be parabolic, the growth can bestated by a simple relation:

(1)

where XD.Z is the growth of the diffusionzone, DCu Fe is the diffusion coefficient ofCu atoms into the stainless steel structure,and tb is the brazing time. Calculating the

X D tD Z Cu Fe b. = →

61-sWELDING JOURNAL

WE

LD

ING

RE

SE

AR

CH

Fig. 7 — A — SEM-BSE image and EDS analysis of the joints prepared at 900°C; B — SEM-BSE image and EDS analysis of the joint prepared at 1000°C.

A B

Fig. 8 — Schematic of the brazing mechanism. Fig. 9 — The thickness of the diffusion area and its measurement.

Fig. 10 — The concentration profile of Cu, Cr, and Ni at the interface of thestainless steel and interlayer brazed at 900°C.

Atabaki 3-13_Layout 1 2/14/13 4:37 PM Page 61

growth of the diffusion zone implies that thedistance of the movement of Cu atoms intothe base metal can be predicted by the sim-ple parabolic law under the condition thatother parameters like self-diffusion andKirkendall effect are neglected from theevaluation (Table 3). To estimate the diffu-sion coefficient of Cu, Cr, and Ni, the Boltz-mann solution of Fick’s second law wasapplied based on the following definition:

(2)

where x is the position of the solute, tb is thebrazing time, and CS is the concentration ofthe solute in the stainless steel. It was re-ported that DCu →Fe at 900°C is 4.67 × 10–15

m2/s for a diffusion time of 50 min (Refs. 12,13), whereas DCu calculated in this research,despite applying lower diffusion time, is alittle higher, reaching to 1.06 × 10–13. Thereason behind this difference might be thepresence of Cu in the lattice and grainboundaries of the stainless steel. The pre-dicted value of the Cu diffusion in the grainboundaries of the stainless steel is 1.754 ×10–6. Another main element in the stainlesssteel is Ni, which for a stainless steel with 10wt-% Ni has a predicted diffusion coeffi-cient of 1.5 × 10–14 m2/s at 1280°C withholding specimens at 1280°C (Ref. 14), al-

though the diffusion coefficient of Ni at abrazing temperature of 900°C is estimatedto be 2.65 × 10–14 m2/s. The diffusion coef-ficient of Cr is also predicted to be 5.03 ×10−14 m2/s. Now, with having the diffusioncoefficients, the diffusion profiles of Cu, Cr,and Ni in and out of the stainless steel canbe estimated by solving the Fick’s secondlaw and applying an analytical model:

(3)

where C (t, x)Cu,Cr,Ni is the composition ofeach of Cu, Cr, and Ni individually in thejoint, D is the diffusion coefficient, t is thebrazing time, x is the distance from the in-terface of the substrate and interlayer, andC0 is the initial solubility of Cu, Cr, and Niin the stainless steel. The concentration ofCu, Cr, and Ni can be drawn via distancefrom the joint as shown in Fig. 10. It canbe seen that by increasing the brazing timethe concentration of the elements changesconsiderably leading to a change in the in-terlayer thickness. However, the thermalactivation of the solute atoms was higherat the higher brazing temperatures caus-ing a higher rate of mass transfer andtougher atomic brazing.

Corrosion Test Results

Figure 11A–C shows the optical mi-crostructure of the joints prepared at 900°,950°, and 1000°C after corrosion test in a3.5% NaCl solution for 12 h. Dark film de-posits were present along the copper surfacefor all joints. Chloride ions are very aggres-sive ions to the copper interlayer, due to thetendency of the chloride ion to form unsta-ble films (CuCl and CuCl3

2–) (Ref. 16).Therefore, even a little amount of chlorideions can cause severe corrosion attack. Fig-ure 12 shows polarization curves for thejoints as a function of the brazing time andtemperature. It can be declared that all thejoints show the same behavior in the polar-ization curve due to the uniform and con-tinuous penetration of the solute and otheralloying elements. However, optical exami-nation indicated that the corrosion mostlyattacks along the surface of the stainlesssteel and copper interlayer.

Obviously, the cathodic line for thebrazing temperature of 900°C exhibitshigh current density indicating a hydrio-genic reaction by reducing the holdingtime. Thus, the free corrosion potential(Ecorr) becomes more negative at thelower holding times. Moreover, by reduc-ing the brazing temperature from 1000° to900°C, cathodic hydrogen reactions are re-duced and Ecorr increases negatively. Theanodic polarization curve for all jointsshows that the current density significantly

( )

= −⎛

⎝⎜

⎞

⎠⎟

⎛

⎝⎜⎜

⎞

⎠⎟⎟

C t x

C erfx

D t

,

12

Cu Cr Ni

Cu Cr Ni b

, ,

0, ,

∫=∂

∂D

t

x

Cx dC

1

2Sb S

SC

C

S

S

1

MARCH 2013, VOL. 9262-s

WE

LD

ING

RE

SE

AR

CH

A CB

Fig. 11 — The optical microstructure of the joints prepared at A — 900°; B — 950°; C — 1000°C for a holding time of 20 min.

Fig. 12 — A — Linear polarization test results for the joints brazed at 900°C for three different holding times;B — linear polarization test results for the joints prepared at three temperatures for 20 min holding time.

Fig. 13 — The comparison of the shear strengths forthe joints prepared at different temperatures andholding times.

A B

Atabaki 3-13_Layout 1 2/14/13 4:37 PM Page 62

increases as a result of dissolution of theelectrode at the beginning of the anodicpolarization, and then the current densitydecreases because of the passivationprocess. In the passivation region, the jointsurface is covered by the passive oxidefilm. Then, the current density starts to in-crease suddenly due to the breakdown ofthe passive layer until the end of the po-larization curve.

As mentioned, the corrosion fre-quently occurred along the stainless steel-copper interface and inside the copperregion. At the anodic polarization curve,the current density was significantly in-creased for all the joints indicating disso-lution of some phases ( γFe + eutectic Cu+ Cr), which resulted in preferential lo-calized corrosion. As indicated in Fig. 12,severe corrosion attack with a wide open-ing (trench formation) occurred at the in-terface because of the dissolution of thebase metal. The corrosion develops in allthe joints and preferential dissolution ofthe copper interface was favored duringpolarization. In this case, stainless steel ismore positive than copper, so when stain-less steel is in contact with copper, thecopper corrodes first. The rate of corro-sion attack is governed by the size of thepotential difference. Referring to the joint,the surface area of the stainless steel is big-ger than copper.

As a result, stainless steel has a largesurface area in contact with the elec-trolyte, while the copper interlayer has avery small surface area in contact with3.5% NaCl solution; therefore, the stain-less steel generates a large corrosion cur-rent, concentrating on a small area of thecopper interlayer. However, intergranularcorrosion cracking was found at the cop-per surface. It is caused by interdiffusionof Fe and Cr along the copper grainboundary and also by Fe-Cr intermetalliccompound. The larger the area of thestainless steel, the greater is the accelera-tion of the copper corrosion.

On the copper side, chloride ions werevery aggressive, due to their tendency toform an unstable film (CuCl) and solublechloride complexes (CuCl2– and CuCl3

2–). Besides, intergranular corrosioncracking was also found at the copper sur-face. It is due to interdiffusion of Fe andCr elements along the copper grainboundary and also to the formation of Fe-Cr intermetallic phase. This phase dis-solves into (Fe + eutectic Cu + Cr) alongthe interface.

For the joints without an immersiontest, there was no pitting attack found onthe copper surface, as this can be seen inFig. 12A. Figure 12B shows a pitting at-tack on the surface of the copper. Gener-ally, grain boundaries; inclusions such assulphides, oxides, and nitrides; and localsegregation of the alloying elements canact as irregularities, initiating the pitting

corrosion. The main alloying elementssuch as Cr, Ni, and Mo were segregatedinto the copper interlayer. Pitting corro-sion attacks and propagates easily whereCr and Mo contents are locally depletedforming microsegregations, precipitationof carbides, and the formation of inter-metallic phases.

A comparison of shear strengths for thejoints is shown in Fig. 13. The shearstrength increased with the enhancing ofbrazing time for different temperatures.This increase in joint strength is related tothe diffusion of the main alloying elementsto the base metal during isothermal solid-ification, so an optimum brazing time is re-quired to improve the joint strength. Thehighest shear strength with a value of 180MPa was achieved for the joints preparedat 950°C for 72 min. It was noticed that thejoints prepared at 1000°C showed a reduc-tion in shear strength, confirming the li-quation at the grain boundaries of the basemetal in the vicinity of the interface.

Conclusions

Transient liquid phase diffusion braz-ing was applied to join stainless steel usinga copper interlayer. Brazing was carried at900°, 950°, and 1000°C for 16, 20, 24, and72 min in a vacuum furnace. The impor-tant findings are as follows:

1) The microstructure studies revealedthat γFe + eutectic Cu + Cr accumulatedalong the diffusion zone. The diffusionzone for the joints increases with increas-ing temperature and holding time exceptfor the case of the highest brazing tem-perature (1000°C), which did not producesignificant changes.

2) After brazing, the joint consisted ofthree distinct zones including base metal,diffusion zone, and diffusion-affectedzone. There were no eutectic structures inthe joint and a relatively uniform distribu-tion of the alloying elements across thejoints occurred, especially at the higherbrazing temperature.

3) It was shown that diffusion of Cu inthe lattice and grain boundary of the stain-less steel plays a significant role in alteringthe solute concentration in the joint regionduring the brazing process at the brazingtemperatures. Corrosion test resultsshowed that the interface is the weakestpart in the resistance against the corrosivesolution.

4) The joints developed crevice corro-sion due to a galvanic couple formed be-tween the stainless steel and copperinterlayer. It presented preferential disso-lution of the copper interlayer under an-odic polarization in 3.5% NaCl solution atroom temperature. Intergranular corro-sion was also found in the copper region.After immersing the joints for 12 h, pittingattack appeared at the copper surface. In

the shear test, the optimum shear strengthwas attained for the joints prepared at950°C for 72 min.

References

1. Orhan, N., Khan, T. I., and Eroglu, M.2001. Diffusion bounding of a microduplexstainless steel to Ti-6Al-4V. Scripta Mat. 45(4):441–446.

2. Atabaki, M. M., and Hanzaei, A. T. 2010.Partial transient liquid phase diffusion bondingof Zircaloy-4 to stabilized austenitic stainlesssteel 321. J. Mat. Character 61(10): 982–991.

3. Atabaki, M. M. 2010. Microstructuralevolution in the partial transient liquid phasediffusion bonding of Zircaloy-4 to stainless steel321 using active titanium filler metal. J. Nuc.Mater. 406(3): 330–344.

4. Atabaki, M. M. 2010. Recent progress injoining of ceramic powder metallurgy productsto metals. Metalurgija-J. Metallurgy (MJoM)16(4): 255–268.

5. Madaah, H. R. H., Atabaki, M. M., andTahaei, A. 2003. Brazing and Soldering, JahaneJame Jam Publisher, Tehran, Iran, ISBN 964-95030-5-6, in Persian, p. 127.

6. Atabaki, M. M., and Idris, J. 2012. Low-temperature partial transient liquid phase dif-fusion bonding of Al/Mg2Si metal matrixcomposite to AZ91D using Al-based interlayer.J. Mater. Des. 42: 172–183.

7. Mannan, S. K., Seetharaman, V., andRaghuthan, V. S. 1983. A study on interdiffu-sion between AISI type 316 stainless steel andaluminium. J. Mater. Sci. Eng. 60(1): 79–86.

8. Ikeuchi, K., Zhou, Y., Kokawa, H., andNorth, T. H. 1992. Liquid-solid interface migra-tion at grain boundary regions during transientliquid phase brazing. Metal. Trans. A 23A (10):2905–2915.

9. Ghoneim, A., and Ojo, O. A. 2011. Nu-merical modeling and simulation of a diffusion-controlled liquid-solid phase change inpolycrystalline solids. Comp. Mater. Sci. 50(3):1102–1113.

10. Cramer, S. D., and Covino, B. S. 2003.ASM International, Volume 13A: Corrosion:Fundamentals, Testing, and Protection, ISBN:978-0-87170-705-5.

11. Na, H. S., Kim, J. K., Jeong, B. Y., andKang, C. Y. 2007. Effect of brazing conditionson microstructure and mechanical properties ofduplex stainless steel to Cr-Cu alloy with Cu-base insert metal. Met. Mater. Int. 13(13):511–515.

12. Chang-Gyu, L., Yoshiyaki, I., Tatsuhiko,H., and Ken ichi, H. 1990. Diffusion chromiumin α-iron. Mater. Trans. 31(4): 255–61.

13. Padron, T., Khan, T. I., and Kabir, M. J.2004. Modelling the transient liquid phasebonding behaviour of a duplex stainless steelusing copper interlayers. Mater. Sci. Eng. A385(1-2): 220–228.

14. Yunker, L. M., and Van Orman, A. J.2007. Interdiffusion of solid iron and nickel athigh pressure. J. Earth Plan. Sci. Let 254(1-2):203–13.

15. E. A. Brandes, G. B. Brook, eds. 1992.Smithells Metals Reference Book, 7th ed., But-terworth–Heinemann, Oxford, p. 938.

16. Yeow, C. W., and Hibbert, D. B. 1983.Galvanostatic pulse plating of copper and cop-per (I) halides from acid copper (II) halide so-lutions. J. Electrochem. Soc. 130(4): 786–90.

63-sWELDING JOURNAL

WE

LD

ING

RE

SE

AR

CH

Atabaki 3-13_Layout 1 2/14/13 4:37 PM Page 63