las difusi Al

Page 1

DA

ALE

a

ARRAA

KDCMSMJ

1

auosortt1ih

plewaau

0h

Journal of Materials Processing Technology 213 (2013) 543– 552

Contents lists available at SciVerse ScienceDirect

Journal of Materials Processing Technology

jou rna l h om epa g e: www.elsev ier .com/ locate / jmatprotec

iffusion welding of aluminium alloy strengthened by Al2O3 particles through anl/Cu multilayer foil

natolii Ustinov ∗, Yury Falchenko, Tatyana Melnichenko, Andrey Shishkin, Gennady Kharchenko,idia Petrushinets

.O. Paton Electric Welding Institute of NAS of Ukraine, 11 Bozhenko Street, Kiev 03680, Ukraine

r t i c l e i n f o

rticle history:eceived 15 March 2012eceived in revised form 9 November 2012ccepted 14 November 2012vailable online 23 November 2012

a b s t r a c t

In the diffusion welding (DW) of difficult-to-deform materials (such as composites and intermetallics),the application of intermediate multilayer foils (MF), which have alternating layers of elements thatform intermetallics, allows for production of a permanent joint under milder conditions. In this paper,the processes occurring in the joint zone (JZ) during DW of Al–5 wt.%Mg+27 wt.%Al2O3 composite mate-rial through the Al/Cu interlayer were studied. It was shown that, while heating of such a foil, phasetransformations that are due to the reaction diffusion of elements, run in it. At MF heating under a con-

eywords:iffusion weldingompositeultilayer foil

uperplasticity

tinuously applied external load, the materials are plastically deformed. It is established that the intensityof foil plastic deformation at a specified load non-monotonically depends on temperature. It is shownthat welding temperature is determined by the temperature at which MF can undergo superplastic flowunder the impact of applied pressure. A mechanism of formation for a solid-phase joint of high-strength

yers b

icrostructureoint zone

materials through interla

. Introduction

Producing permanent joints of materials based on compositesnd intermetallics, for example, by melting in the joint zone (JZ), issually accompanied by the degradation of their properties. The usef diffusion welding (DW) prevents any substantial changes in thetructure of such materials in the JZ. However, practical applicationf this method is complicated by the need to preheat these mate-ials to high temperatures above 0.7 Tm, where Tm is the meltingemperature of the based material; applying high pressure ensureshe plastic deformation of the material being welded (Kazakov,985). DW involves certain difficulties because materials based on

ntermetallics and composites are difficult-to-deform, even aftereating to high temperatures (Lee et al., 1999).

To reduce the heating temperature, the magnitude of the appliedressure and time of DW for difficult-to-deform materials, inter-

ayers based on ductile metals were used. Previously, Nesmikht al. (1987) used aluminium foil as an interlayer in the diffusionelding of titanium to UF46 ceramics containing 72 wt.%Al2O3,

nd in the welding of a composite material based on 2124 Al

lloy strengthened by 12% SiC. Urena and Gomez de Salazar (1993)sed interlayers of superductile materials, Al–Li 8090 and Al–Cu

∗ Corresponding author. Tel.: +380 44 200 6180; fax: +380 44 205 2277.E-mail address: [email protected] (A. Ustinov).

924-0136/$ – see front matter © 2012 Elsevier B.V. All rights reserved.ttp://dx.doi.org/10.1016/j.jmatprotec.2012.11.012

ased on the MF of intermetallic-forming elements is proposed.© 2012 Elsevier B.V. All rights reserved.

“Supral-100”; in both cases the materials were welded at lowertemperatures than those welded without an interlayer.

It is hypothesised that during welding, these interlayers undergoplastic flow as a result of applied heat and pressure, which causedeformations in of the subsurface layers of the based materials thatare in contact with them. Such plastic deformations in the interlay-ers provide physical contact between surfaces to be weld (e.g., theoxide layer is removed from the welding surface and discontinu-ities formed at the intersection of rough surfaces are eliminated,etc.). This process leads to an increase in the defect density in thesubsurface layers, thereby promoting acceleration of the diffusionprocesses in the JZ of the based materials. This acceleration resultsin the intergrowth of grains in the JZ between surfaces being weldedat lower temperatures and applied pressures, compared to DW thesame materials without an interlayer.

Applying Ni and Cu interlayers 2–5 �m thick produced by vac-uum deposition when welding 2024 Al alloy strengthened by SiCparticles ensures a welded joint strength of approximately 52%of the inherent strength of these composite materials (Ryabovand Cherepivskaya, 2002), using 150 �m-thick Al interlayers whenwelding of Al + 4 wt.%C composite, ensures a welded joint strengthof approximately 70% (Ryabov et al., 1996).

Nami et al. (2010) believe that chemical inhomogeneity in the JZdue to these interlayers lowers the strength and service properties

of joints produced with ductile interlayers. Reducing the degree ofchemical inhomogeneity in the JZ by reducing the thickness of theductile interlayer requires a considerable increase in the weldingpressure (Musin et al., 1979).

Page 2

5 rocessing Technology 213 (2013) 543– 552

acTtotthntupTut

dtOtdei

(ttM

2

bsewWtoMioiaeM5cMdUppfoistatc

t

welding temperature ranged from 350 to 500 C at constant valuesof pressure (44 MPa) and time (20 min).

Scanning electron microscopy (SEM, CamScan 4, operating volt-age of 20 kV) with energy dispersive spectrometer (EDS, Link

44 A. Ustinov et al. / Journal of Materials P

Reduction of chemical inhomogeneity in the JZ in DW throughn interlayer can be achieved by using an interlayer of the sameomposition, as materials being welded. To weld alloys based oniAl intermetallics, Duarte et al. (2006) used interlayers with a mul-ilayer structure of titanium and aluminium that were depositedn surfaces being welded by magnetron sputtering in a propor-ion close to the composition of the based materials. The coatingotal thickness was 2–2.5 �m, and the period was 4 nm. Subsequenteating above 600 ◦C for 1 h allows for the transformation of theanolayer into an equilibrium TiAl structure with a finer grain thanhat of the based material (Ramos et al., 2006). Ustinov et al. (2008a)sed a 20–30 �m-thick interlayer foil with a multilayer structure,roduced by electron beam physical vapour deposition (EBPVD).he formation of full-strength joints in the solid phase is ensurednder milder conditions (Ustinov et al., 2009a), which indicateshat the interlayers accelerate diffusion processes in the JZ.

Because alloys based on TiAl intermetallics are difficult-to-eform, acceleration of the diffusion processes in the JZ occurs dueo processes within the MF while heating under stationary loading.ne can assume that the approach to welding TiAl intermetallics

hrough a multilayer foil can also be applied to welding otherifficult-to-deform materials, such as composite materials. How-ver an important first step is to investigate the processes occurringn the JZ and to correlate them with the welding conditions.

In this work, Al–5 wt.%Mg + 27 wt.%Al2O3 composite materialhere, referred to as ‘composite’) was used to study the proper-ies of permanent joint formation through a MF (Al/Cu) by DW ando correlate them to the phase and microstructural changes in the

F while heating under continuously applied loads.

. Materials and methods

A composite of Al + 5 wt.%Mg aluminium alloy1 strengthenedy 27 wt.%Al2O3 was used as the model material in this work totudy DW conditions. According to previous research (Falchenkot al., 2010; Koryagin et al., 2000), such a composite has a Rock-ell hardness of 99–100 MPa and a tensile strength of 340 MPa.elding was conducted on 6 mm × 15 mm × 15 mm samples. Prior

o welding, the base material was treated on abrasive wheels tobtain a smooth surface (roughness Rz of approximately 6 �m).echanically prepared sample surfaces and foils applied as an

ntermediate layer were degreased using acetone. MF consistingf layers of aluminium and copper were used as an interlayer dur-ng composite welding. The thicknesses of the aluminium layerslternated between 600 nm and 60 nm, and that of the copper lay-rs alternated between 100 nm and 10 nm. The overall thickness ofF with these submicron and nanosized alternating of layers was

0 �m. The stoichiometry of the Al/Cu MF was close to the eutecticomposition: 69.1 wt.%Al, 30.9 wt.%Cu (84 at.%Al, 16 at.%Cu). TheseF were produced using a layer-by-layer electron beam vacuum

eposition method of aluminium and copper, as described in detailstinov et al. (2009b). Briefly, aluminium and copper ingots werelaced into in a vacuum chamber fitted with two water-cooled cop-er crucibles separated by a screen (Fig. 1), and the steel substrateor metal vapour phase deposition was placed above the cruciblesn a vertical shaft aligned with the plane of the screen separat-ng the crucibles. Due to the position of the crucibles and theireparating screen during electron beam evaporation of the ingots,he metal vapour is only deposited on the substrate surface that is

bove the crucibles. Rotation of the substrate on a vertical shaft overhe evaporators allows for deposition of alternating aluminium andopper layers on multiple sections of the substrate surface. Before

1 Impurity content in the alloy is 5–8 wt.%, of which Fe has the highest concen-ration (>0.5 wt.%).

Fig. 1. Schematic of the electron beam unit used to produce the multilayer foil: (1)preheating guns, (2) rotating substrate, (3) evaporation guns, (4) targets, (5) screen.

elemental deposition, the substrate was coated with a thin layer ofsalt, which allowed for easy separation of MF from the substrateafter the process was completed. Thus, MF were not subjected toheat treatment before welding.

Welding was conducted in a unit vacuum chamber fitted with aradiation heating system (Fig. 2). To be weld samples were mountedin a fixture between the upper and lower rods. An intermedi-ate layer of multilayer foil was placed between the aluminiumcomposite surfaces being joined. Heating of the JZ was done withmolybdenum heaters located around the sample. The temperaturewas controlled by a chromel–alumel thermocouple fastened to thefixture. Pressure was applied to the samples from a press throughthe lower rod and a wedge; the magnitude of the pressure wascontrolled by a dynamometer.

A sequence diagram of the DW process is shown in Fig. 3. Accord-ing to the sequence diagram, samples are heated, and, after a certainwelding temperature (Tw) has been reached, a pressure (Pw) isapplied to them. The welding time is tw (at temperature Tw andapplied pressure Pw). After welding, the samples are cooled andthen unloaded. The main parameters of the welding process arethe welding temperature, welding pressure and welding time. The

◦

Fig. 2. Schematic of the welding unit: (1) vacuum chamber of the unit; (2) molybde-num heater; (3) upper rod; (4) base materials with interlayer; (5) fixture; (6) lowerrod; (7) wedge; (8) press.

Page 3

A. Ustinov et al. / Journal of Materials Process

Fig. 3. Sequence diagram of the DW process: the time dependence of temperatureiiw

Swsiisi

tat2

naptwtc

tf

Fop

n the JZ is shown on the upper graph, and the time dependence of applied pressuren the JZ is shown on the lower graph. Tw is the welding temperature, Pw is the

elding pressure, and tw is the welding time under specified conditions.

ystems Oxford Instruments ENERGY 200, 20 kV operating voltage)as conducted on transverse sections to investigate the compo-

ition and microstructure of the welded joint and those of thentermediate layer in its as-deposited condition, after heating ints free state and under a continuously applied load. Microscopyamples were prepared using a standard procedure with grind-ng/polishing equipment from Struers Company.

The deformation behaviour of MF under a continuously appliedensile load with heating was studied in a dilatometer fitted with

device that allows application of a constant uniaxial tensile loado the foil strip during heating as shown in Fig. 4 (Ustinov et al.,009b).

Springs were used to create a continuously applied load, con-ecting a mobile table to one end of the foil strip (5 mm × 20 mm)nd a linear displacement sensor. The other end of the foil strip sam-le was fastened to a stationary base that allowed for determininghe tensile load on the sample and recording the change in lengthith a highly sensitive linear displacement sensor. The sample was

hen placed in a vacuum pipe furnace. The sample temperature was

ontrolled by a chromel–alumel thermocouple attached to it.

Differential thermal analysis (DTA) was used to determine theemperature range of the phase transformations in the multilayeroil during heating. DTA curves were obtained in a VDTA8M unit

ig. 4. Schematic of dilatometer used to measure the MF elongation under continu-usly applied force with heating: (1) foil sample, (2) spring, (3) mobile, (4) stationaryarts of the measurement block.

ing Technology 213 (2013) 543– 552 545

using a so-called diluent method (Shishkin et al., 2007), in which100 �m-thick copper foil was used as the diluent. Specimens of10 mm × 10 mm were cut out of the MF foil and were placedbetween the copper plates. The samples were heated at a constantrate of 50 ◦C/min in helium gas in the DTA unit up to the requiredtemperature.

3. Results and discussion

Producing a permanent joint during DW process implies forma-tion in the joint zone of grains, which interpenetrate the surfacesbeing joined (Kazakov, 1985). In order to achieve such a structuralcondition in the joint zone, DW conditions (temperature, pressureand welding time) should ensure establishing a physical contactbetween the surfaces being welded and intensive running of the dif-fusion processes. It is known that these conditions are determined,mainly, by physical and mechanical properties of the material beingwelded. When an interlayer is used, the properties of the inter-layer material can significantly influence the DW conditions. It canbe assumed that the phase and microstructural changes that willproceed in the MF while heating under an external mechanical loadmay affect the processes occurring in the JZ. To establish a corre-lation between these processes and the phase changes, we studiedthe temperature intervals of the phase transformations and thedeformational behaviour of MF foils while heating under externalloading, similar to the conditions implemented during DW.

The cross-sectional microstructure of multilayer Al/Cu foil inits the as-deposited condition with submicron-period (700 nm) ofaluminium and copper is shown in Fig. 5a. In the diffractrograph ofAl/Cu foil in as-welded condition shown in Fig. 5b, just the diffrac-tion peaks corresponding to copper and aluminium are observed.This leads to the conclusion that during the deposition process, dif-fusion between the elements does not lead to new phase formation.

Therefore, the light and dark bands observed in foil microstruc-ture corresponded to alternating layers of copper and aluminiumrespectively.

From the equilibrium diagram of the Al–Cu system (Massalskiet al., 1986), it follows that, at the selected composition, theequilibrium foil structure should consist of a mixture of phases(Al + Al2Cu). To study the transition from a metastable state to anequilibrium state, the foil was heated at a constant rate in a high-temperature X-ray diffractometer fitted with a vacuum chamber;and diffractometry measurements were conducted in situ using aposition-sensitive detector. As observed in the diffractograms inFig. 6, under continuous heating of the MF, the formation of an equi-librium structure that consists of a mixture of phases (Al + Al2Cu)proceeds through a number of intermediate states that can be con-ditionally divided into two stages (Ustinov et al., 2008a,b, 2009a,b):

Al + Cu → Al + AlCu → Al + Al2Cu

At the first stage, the dominating process is the formation of anAlCu intermetallic, and at the second stage, the dominating processis the formation of an Al2Cu compound.

As observed in the DTA curves (Fig. 7a), phase transformationsare accompanied by intense heat evolution, occurring in two stages:the first stage occurs at 80–150 ◦C, and the second stage occurs at150–350 ◦C. Comparison of the data obtained by XRD (Fig. 6) andDTA (Fig. 7a) leads to the hypothesis that synthesis of the AlCu inter-metallic predominates during the first stage, whereas formation ofthe Al2Cu compound predominates during the second stage. Theabsence of any thermal effects at an additional increase in the MF

temperature allows for the reasonable assumption that the MF isin relative equilibrium at heating above 350 ◦C.

It is known that roughness of surfaces being welded is one ofthe obstacles for establishing the physical contact between them.

Page 4

546 A. Ustinov et al. / Journal of Materials Processing Technology 213 (2013) 543– 552

F n 700

d er foi

Cta

hticca

Ff

ig. 5. Cross-sectional microstructure of multilayer Cu/Al foil (period of modulatioark bands correspond to aluminium layers; X-ray diffractograms of Al/Cu multilay

onditions of plastic deformation of the interlayer are importanto establish the physical contact between the surface being weldednd the interlayer.

To investigate the deformational behaviour of the MF, it waseated under continuously applied tensile forces, similar to thosehat arise in the foil at the welding pressure MF strips were placed

nto the dilatometer and heated (Fig. 4). Fig. 7b and c shows theurves of foil elongation and deformation rate, respectively, withontinuous heating under a continuously applied tensile forcesnd a pressure of 4 MPa. It is seen that with temperature increase

ig. 6. X-ray diffractograms were taken at continuous heating of Al/Cu multilayeroil (period of modulation 700 nm) with the rate of 10 ◦C/min (b).

nm) in its as-deposited condition (a). Light bands correspond to copper layers, andl in its the as-deposited condition (b).

the degree of foil deformation rises right up to 250 ◦C. The defor-mation rate also rises. However, at temperatures above 250 ◦Cfoils deformation abruptly slows down, and its rate decreases.Above the temperature of 300 ◦C foil deformation starts increas-ing smoothly, and at the temperature above 450 ◦C it acquires anexponential dependence on temperature. Thus, two intervals canbe singled out on the temperature dependence of MF deforma-tion behaviour, where its intensive plastic deformation is observed:low-temperature interval of 150–250 ◦C and high-temperatureinterval above 450 ◦C. By comparing the deformation curves to theDTA curves (Fig. 7), the low-temperature interval can be observedto coincide with the temperature range in which phase transfor-mations proceed.

The temperature range in which formation of the AlCu com-pound predominates is characterised by an increase in the rateof sample elongation with increasing temperature, whereas at thesecond stage, the temperature range in which the AlCu phase trans-forms into the Al2Cu phase is characterised by a decrease in the rateof sample deformation.

To clarify the nature of non-monotonic change of foil deforma-tion rate at its heating in the temperature range of 150–300 ◦C,calculation of the change of sample length due to phase transfor-mations was performed. Considering that the volume per one atomof elements entering into the synthesis reaction, varies, calcula-tion of the change of sample volume for various stages of phasetransformations was performed, proceeding from crystalline latticeparameters taken at room temperature. As is seen from Table 1, syn-thesis of AlCu intermetallic should be accompanied by an increasein sample volume, whereas transition of AlCu compound into Al2Cushould lead to a reduction in sample volume. Considering that�l/l ≈ (1/3)(�V/V), calculation of the change of sample length wasperformed based on the calculated data. One can see that at thefirst stage of transformation sample elongation should have beenequal to about 0.68%, and at the second stage sample length should

have decreased by 0.77%. Comparison of calculated values of sam-ple elongation, due to volume effects of phase transformations andthose measured experimentally, shows that experimental valuesof sample elongation are greater than the calculated values both in

Table 1Calculated changes in volume and elongation of Cu/Al MF due to intermetallicformation.

Type of reaction Changes of volume,�V/V (%)

Elongation, �l/l≈ (1/3)(�V/V) (%)

2Al + Cu = AlCu + Al 2.03 0.68AlCu + Al = Al2Cu −2.32 −0.772Al + Cu = Al2Cu −0.33 −0.11

Page 5

A. Ustinov et al. / Journal of Materials Process

For

itipoas

mptacst

tion upon subsequent heating in the low-temperature region.Because the high-temperature range of MF plastic defor-

mation is not accompanied by phase transformations, it is dueto microstructural changes. These diffusion processes were

ig. 7. The DTA curve for this foil (period of modulation 700 nm) with a heating ratef 20 ◦C/min (a); temperature dependence of the elongation rate (b) and deformationate (c) on Cu/Al foil under a continuously applied tensile force.

ndividual stages of transformation, and after their completion. Onhis basis one can assume that sample elongation is due mainly tots plastic deformation, and the contribution of the change of sam-le length due to volume effects of phase transformations, leads tonly non-monotonic change of the deformation rate: increasing itt the stage of AlCu compound formation and decreasing it at thetage of Al2Cu formation.

The onset of plastic deformation at loads much lower than theaterial yield point at room temperature can be attributed to the

hase transformations occurring in this temperature region. Phaseransformations can promote the plastic deformation of materials

t loads below their yield point, such as the plastic deformationaused by the initiation of a martensitic transformation in TRIPteels with an external load (Greenwood and Jonson, 1965). In addi-ion, diffusion processes are activated by phase transformations in

ing Technology 213 (2013) 543– 552 547

the MF that can facilitate dislocation displacement and plastic flowlocally where forces are applied. Therefore, plastic deformation of asample in the low-temperature region results from phase transfor-mations, i.e., due to the plasticity of these phase transformations.Thus, completion of phase transformations in the layered foil isaccompanied by an abrupt decrease in the rate of sample deforma-

Fig. 8. Cross-sectional microstructure of the multilayer foil (period of modulation700 nm) after heating up to 300 ◦C (a) and 450 ◦C (b) under constantly applied tensileforce, and after heating up to 450 ◦C (c) at a pressure of 44 MPa.

Page 6

548 A. Ustinov et al. / Journal of Materials Processing Technology 213 (2013) 543– 552

F d after(

sttfNrf

ibrTpoaie

ieTpiagam

aibf

ig. 9. Cross-sectional microstructure of multilayer Al/Cu foil with nanosized perioc).

tudied by analysing the microstructure of MF samples heatedo 300–500 ◦C. As observed in Fig. 8a, after completing phaseransformations due to diffusion-induced elemental mixing, theoil forms a multi-layered structure of Al2Cu intermetallic and Al.o phase transformations were observed in the high temperature

ange of MF plastic flow; the phase composition (Al2Cu + Al) thatormed at low temperatures remained stable.

However, based on images of the MF microstructure after heat-ng under tensile (Fig. 8b) and compressive (Fig. 8c) forces, it cane observed that foil plastic deformation in the high-temperatureegion is accompanied by breaking up of the foil lamellar structure.hen, the separation of intermetallic layers into fragments takeslace, and bridges between MF components are formed. The degreef layer fragmentation increases with decreasing thickness (Fig. 9and b). Notably, the fragmentation of the intermetallic layers intondividual grains is also observed at MF while heating withoutxternal impact (Fig. 9c).

Thus, the high-temperature region of plastic flow in the foils accompanied by fragmentation of the intermetallic layers intoquiaxed grains aligned in the direction of the applied tensile force.he shape of the equiaxed grains indicates that the deformationrocesses in the heterophasic structure of the foil occurred predom-

nantly via a grain-boundary slipping mechanism, leading to grainlignment in the direction of the applied force, not through intra-ranular dislocation processes, which lead to grain elongation. Such

plastic deformation mechanism is characteristic of superplasticaterials (Zhilayev and Pshenichyuk, 2008).Therefore, applying pressure to a multilayer foil with aluminium

nd copper layers will change its deformational behaviour, depend-ng on the temperature at which the pressure is applied. It maye assumed that in the temperature ranges of the phase trans-ormations (the low-temperature region) and the microstructural

heating up to 450 ◦C under uniaxial tension (a and b) and without external impact

changes (the high-temperature region) the foil will be prone toplastic flow; however, between these temperature intervals (from300 ◦C to 450 ◦C), plastic deformation of foil may be difficult.

To verify this assumption, MF macrostructure was studied afterheating it up to the specified temperature and then applying pres-sure at this temperature. For this purpose MF were placed betweencopper surfaces, and this assembly was heated to the specified tem-perature and pressure was applied. To separate the surfaces beingwelded, one of them was coated with a layer of boron nitride to pre-vent its diffusion interaction with MF. The time of assembly soakingafter its loading was not less than 1 min. Then the assembly wascooled, and a copper plate with boron nitride layer applied on it,was separated from MF.

Fig. 10 shows micrographs of the foil after the thermome-chanical treatment described above. At 350 ◦C, the application ofpressure leads to the brittle fracturing of the foil, whereas at 450 ◦Cunder applied pressure, the foil maintains the integrity of its struc-ture with traces of plastic flow visible on its surface.

The data show that the deformational behaviour of the foilbelow 450 ◦C with applied pressure is characterised by low duc-tility. Based on these results, DW of high-strength aluminiumalloys through a MF should be conducted at temperatures above450 ◦C, i.e., under the conditions in which the onset of plasticdeformation due to super-plasticity is possible. To study the cor-relation between DW conditions and the temperature range of MFsuperplastic behaviour, pressure welding of aluminium compositethrough Al/Cu MF was performed at welding temperatures (Tw) of350, 400, 450, and 500 ◦C; at a constant pressure (Pw) of 44 MPa;

and at a welding time (ts) of 20 min. As observed in Fig. 11, thealuminium composite microstructure is characterised by thepresence of fragmented Al2O3 particles in the aluminium matrixthat were randomly distributed throughout the volume fraction.

Page 7

A. Ustinov et al. / Journal of Materials Processing Technology 213 (2013) 543– 552 549

Fig. 10. Appearance of MF surface produced while heating up to temperatures of 350 ◦C (awelded and at a compressed force of 44 MPa.

Fig. 11. Microstructure of a Al + 5 wt.%Mg aluminium alloy strengthened by2t

Pdr

boacwobJiM

TC

7 wt.%Al2O3: dark particles are Al2O3 inclusions and the light particles correspondo Fe-rich precipitates.

resence of Al2O3 particles in the metal essentially lowers itsuctility and prevents joint formation in the sites of the particleseaching the surfaces being welded.

When the DW of these materials using MF at temperatureselow 450 ◦C, no adhesion of the surfaces being welded to eachther was observed; the surfaces being welded were separatedfter pressure elimination. The joints of these materials with a suffi-ient strength level started forming only after the parts being joinedere heated up to 450 ◦C. The microstructure of the welded joints

btained at such welding temperatures is shown in Fig. 12a. It cane observed that two microstructural regions are observed in the

Z: the MF region, which preserves its lamellar structure after heat-ng up to 450 ◦C, and the composite material region adjacent to the

F, in which copper-enriched precipitates are observed (Table 2).

able 2hemical composition of different sections of the welded JZ.

Spot Chemical composition (wt.%), EDS spectra, K series

O Mg Al Mn Ti Fe Cu

1 2.25 2.82 74.4 0.29 17.332 29.57 2.82 16.71 0.35 50.553 8.03 65.16 26.814 1.02 57.5 0.56 12.03 28.89

) and 450 ◦C (b) in the form of an intermediate layer between copper surfaces being

Upon increasing the welding temperature, the welding zonestructure underwent essential changes (Fig. 12b and c). The degreeof MF dispersion and thickness of the JZ at this point is notable. Thehighest degree of dispersion of the intermediate layer and smallestthickness of the JZ were achieved at 500 ◦C.

Under these conditions a composite structure forms in the JZ,which consists of an aluminium matrix and a copper-enrichedstrengthening particles (supposedly, formed on the basis of Al2Cu-phase). At this point, the density and distribution of the Al2O3particles in the composite material regions adjacent to the weldingzone remained practically unchanged.

It is believed that plastic flow of the interlayer under the impactof applied pressure may lead to establishing of a physical contactbetween the surfaces being welded and acceleration of diffusionprocesses in them. However, an important factor that ensures theprocesses required for the formation of a high-strength joint is theelimination of discontinuities, which can form on the boundarybetween the high-strength materials because of their roughness.In this case, the super-plastic behaviour of the interlayer can alsopromote the elimination of discontinuities upon contact of real sur-faces of the parts being welded. Formation of a permanent jointbetween high-strength materials via DW through an intermedi-ate layer consisting of layers of intermetallic-forming elementsis presented as a schematic in Fig. 13, which details the follow-ing: during the first stage, heating of the JZ to a temperature highenough to ensure phase transformations in the multilayer foil thatare associated with the synthesis of an intermetallic compound(i.e., transition of the weld structure shown in Fig. 13a to the struc-ture shown in Fig. 13b); during the second stage, heating of the JZto the heterophasic foil transition temperature to induce a super-plastic state and applying a pressure that provides the conditionsnecessary for plastic flow (Fig. 13c).

Processes occurring during the second stage promote contactdeformation of the subsurface layers for the parts being welded,i.e., the “flow” of the foil material into the discontinuities, exist-ing on the contact boundary of these surfaces, which modify themicrostructural state of the foil itself (i.e., the fragmentation ofintermetallic interlayers, which form in MF during the heating pro-cess). “Flowing” of foil into discontinuities on the interface of thematerial being welded and the interlayer ensures establishment ofphysical contact between the surfaces being welded and the inter-layer and is the necessary condition for intensive mass transfer

between them. This will be also promoted by longer welding time.In addition, during the welding process chemical inhomogeneities,which are due to the presence of an interlayer in the welding zone,will be reduced owing to running of the diffusion processes.

Page 8

550 A. Ustinov et al. / Journal of Materials Processing Technology 213 (2013) 543– 552

F produ2 70 ◦C,

bpfatptl

a

ig. 12. Cross-sectional microstructure of a welded joint of aluminium composite

0 min and at a pressure of 44 MPa for the following temperatures: (a) 450 ◦C, (b) 4

Clearly, the degree of realisation of the processes describedy this schematic essentially depends on the rate of foil super-lastic flow under the impact of applied pressure. As is observedrom Fig. 7, foil deforms most intensively when the temperaturespproach 500 ◦C. From Fig. 12, it can be observed that increasinghe welding temperature from 450 ◦C up to 500 ◦C in the aboverocesses realises the most complete weld. Therefore, the factors

hat lower the temperature of MF superplastic flow also ensure theowering of the DW temperature.

As noted earlier, after the MF was heated to temperaturesbove 300 ◦C, a transition occurred from a lamellar structure

ced by the DW method though intermediate Al/Cu multilayer at a welding time of (c) 500 ◦C.

based on pure elements to one consisting of layers of alu-minium and Al2Cu intermetallic, with subsequent fragmentationof intermetallic layers upon heating. Comparing the cross-sectionalmicrostructures of foil with different period that were subjected toheating at continuously applied tensile forces (Figs. 8c and 9a, b)one can see that fragmentation of intermetallic interlayers alongtheir length is observed more often in foil with a smaller

period.

Therefore, at a heating temperature of 450 ◦C, the intermetalliclayers in the foil with nanosized period were fully fragmented(Fig. 9b). Thus, it is hypothesised that a reduction in the thickness

Page 9

A. Ustinov et al. / Journal of Materials Processing Technology 213 (2013) 543– 552 551

Fig. 13. Schematic of the microstructural changes in JZ of a high-strength material being welded with the interlayer of a multilayer structure placed between them surfaces(a) while being heated: at the phase transition temperature in the foil intermetallic layers are formed (b); while applying pressure at the temperature of superplastic flowleads to making of physical contact between surface being welded and activates the diffusion processes in the JZ (c); long-term soaking of such a structure at temperaturetw promotes leads to the “dissolving” of interlayer (d).

F MF with period of equal to 70 nm at heating under the conditions of continuously appliedt

os

umttimpaip

wsnfb

ig. 14. Temperature elongation dependencies (a) and deformation rate (b) of Cu/Alensile forces.

f the period reduces the transition temperature of the foil into auperplastic state.

Fig. 14 gives the deformation curves obtained while heatingnder tensile forces applied to MF with 70 nm period. The defor-ational behaviour of the foil remains unchanged from that of

he MF with submicron (700 nm) period. However, the tempera-ure at the onset of superplastic flow in the nanolayer foil shiftsnto the lower-temperature region, unlike that of the MF with sub-

icron period. The temperature at which the maximum rate oflastic flow occurs also decreases. Based on these results, it can bessumed that using MF with nanosized period promotes a decreasen the DW temperature, as opposed to using MF with submicroneriod.

The microstructure of the DW joints that used nanolayer foilith 70-nm period is shown in Fig. 15. Comparing the micro-

tructures of DW joints using MF with either submicron oranosized period, one can see that for nanolayered MF a JZ can be

ormed, in which concentrated inhomogeneities due to interlayerehaviour are practically absent.

Fig. 15. Cross-sectional microstructure of a welded joint of aluminium compos-ite produced by the DW method through intermediate Al/Cu multilayer foil withnanosized period (70 nm) at a welding time of 20 min, a pressure of 44 MPa, and atemperature of 500 ◦C.

Page 10

5 rocess

4

1

2

3

R

D

F

G

K

52 A. Ustinov et al. / Journal of Materials P

. Conclusions

. It is established that deformational behaviour of Al/Cu MFdepends on temperature; there exist two temperature ranges,a low-temperature range from 150 to 250 ◦C and a high-temperature range above 400–450 ◦C, in which the foilundergoes plastic flow under applied external forces. Plastic flowin the low-temperature range is associated with phase trans-formations (plasticity transformation); in the high-temperaturerange, plastic flow is associated with microstructure changes(super-plasticity).

. It is shown that the lower temperature limit of pressure weld-ing a high-strength aluminium composite using an intermediatelayer of Al/Cu MF is determined by the temperature at the onsetof the plastic flow due to super-plasticity in the multilayer foil.

. It is shown that reducing the modulation period of the multilayerAl/Cu foil from a submicron to nanosized scale lowers the super-plasticity temperature interval by 50 ◦C. Applying nanolayer foilas an interlayer decreases the temperature at which the forma-tion of a permanent joint is possible by 50 ◦C, compared withthe welding temperature for a multilayer foil with submicronlayers.

eferences

uarte, L.I., Ramos, A.S., Vieira, M.F., Viana, F., Vieira, M.T., Kocak, M., 2006. Solid-state diffusion bonding of gamma-TiAl alloys using Ti/Al thin films as interlayers.Intermetallics 14, 1151–1156.

alchenko, Yu.V., Muravejnik, A.N., Kharchenko, G.K., Fedorchuk, V.F., Gor-dan, G.N., 2010. Pressure welding of micro-dispersed composite materialAMg5 + 27%Al2O3 with application of rapidly solidified interlayer of eutectic

alloy Al + 33%Cu. The Paton Welding Journal 2, 7–10.

reenwood, G.W., Jonson, R.H., 1965. The deformation of metals under small stressduring phase transformation. Proceedings of Royal Society A 283, 403–423.

azakov, N.F., 1985. Diffusion Bonding of Materials. English version. Pergamon PressInc., New York, USA, p. 304.

ing Technology 213 (2013) 543– 552

Koryagin, S.I., Pimenov, I.V., Khudyakov, V.K., 2000. Processing Techniques of Mate-rials. Kaliningrad University, Kiliningrad, Russia, p. 448.

Lee, C.S., Li, H., Chandel, R.S., 1999. Simulation model of the vacuum-free diffusionbonding of aluminum metal-matrix composite. Journal of Materials ProcessingTechnology 89–90, 344–349.

Massalski, T.B., Okamoto, H., Subramanian, P.R., Kacprzak, L., 1986. Binary AlloyPhase Diagrams. ASM International, Materials Park, OH, p. 526.

Musin, R.A., Antsiferov, V.N., Kvasnitskii, V.F., 1979. Diffusion Welding of High-temperature Alloys. Metallurgia, Moscow, Russia, pp. 207 (in Russian).

Nami, H., Halvaee, A., Adgi, H., Hadian, A., 2010. Investigation on microstructureand mechanical properties of diffusion bonded Al/Mg2Si metal matrix com-posite using copper interlayer. Journal of Materials Processing Technology 210,1282–1289.

Nesmikh, V.S., Malevskii, Yu.B., Kushnareva, T.N., Urodkov, V.M., Peregud, V.I., 1987.Diffusion welding of titanium with ceramics through an aluminium interlayer.Automatic Welding 5, 69–72 (in Russian).

Ramos, A.S., Vieira, M.T., Duarte, L.I., Vieira, M.F., Viana, F., Calinas, R., 2006. Nanjmet-ric multilayers: a new approach for joining TiAl. Intermetallics 14, 1157–1162.

Ryabov, V.R., Cherepivskaya, E.V., 2002. Diffusion welding of composite of aluminumalloy + 20% SiC. Technological Systems 1, 40–47 (in Russian).

Ryabov, V.R., Kharchenko, G.K., Markashova, L.I., Muravejnik, A.N.,1996. On theweldability of dispersion strengthened composite (Al + 4%C). In: InternationalWelding Conference. Technical University of Kosice, Kosice, pp. 271–273.

Shishkin, A.E., Shishkin, E.A., Ustinov, A.I., 2007. Thermal analysis of microlaminatefillers based on intermetallic forming elements. The Paton Welding Journal 12,23–27.

Urena, J.M., Gomez de Salazar, J.M., 1993. Bonding of aluminium matrix compositesfor application in the transport industry. Jornal de Physique 4, 1037–1042.

Ustinov, A.I., Falchenko, Yu.V., Ishchenko, A.Ya., Kharchenko, G.K., Melnichenko, T.V.,Muraveynik, A.N., 2008a. Diffusion welding of �-TiAl based alloys through nano-layered foil of Ti/Al system. Intermetallics 16, 1043–1045.

Ustinov, A.I., Olikhovskaya, L.A., Melnichenko, T.V., Shishkin, A.E., Matvienko, Ya.I.,2008b. Solid-phase reaction in heating of multilayer Al/Ti foils produced byelectron beam deposition method. Advances in Electrometallurgy 2, 19–27.

Ustinov, A.I., Falchenko, Yu.V., Ishchenko, A.Ya., Kharchenko, G.K., Melnichenko, T.V.,Muraveynik, A.N., 2009a. Producing permanent joints of �-TiAl based alloysusing nanolayered Ti/Al interlayer by vacuum diffusion welding. The PatonWelding Journal 1, 12–15.

Ustinov, A.I., Matvienko, Ya.I., Polishchuk, S.S., Shishkin, A.E., 2009b. Investigation ofphase transformations and plastic deformation at continuous heating of Al/Cumultilayer foil. The Paton Welding Journal 10, 23–27.

Zhilayev, A.P., Pshenichyuk, A.I., 2008. Superplasticity and Grain Boundaries inUltrafine-grained Materials. Fizmatlit, Moscow, Russia, p. 315 (in Russian).