Materials and Design 31 (2010) 1981–1992

Contents lists available at ScienceDirect

Materials and Design

journal homepage: www.elsevier .com/locate /matdes

Experimental investigation of laser beam welding of explosion-weldedsteel/aluminum structural transition joints

L. Tricarico, R. Spina *

Dept. of Mechanical and Management Engineering (DIMeG), Politecnico di Bari, Viale Japigia 182, 70126 Bari, Italy

a r t i c l e i n f o a b s t r a c t

Article history:Received 15 September 2009Accepted 19 October 2009Available online 1 November 2009

Keywords:Structural transition jointLaser material processingMechanical characterization

0261-3069/$ - see front matter � 2009 Elsevier Ltd. Adoi:10.1016/j.matdes.2009.10.032

* Corresponding author. Tel./fax: +39 0805962768.E-mail address: r.spina@poliba.it (R. Spina).

The steel/aluminum structural transition joints are widely used in shipbuilding industry due to theadvantages of joining these two materials with important weight savings while exploiting their bestproperties. The use of laser welding to strongly connect components made of Fe and Al alloys as basematerials with Fe/Al structural transition joints is very attractive. The authors report results achievedduring the laser welding of these particular joints with the scope to evaluate effects of the laser-inducedthermal loads on the integrity of the Fe/Al bond interface, from metallurgical and mechanical points ofview. The increase of both inter-metallic film thickness and extension were detected as a result of thelaser beam induced heat on the Fe/Al bond interface. These increases did not cause severe reductionsof the mechanical resistance of the investigated structural transition joint.

� 2009 Elsevier Ltd. All rights reserved.

1. Introduction

A great interest in joining aluminum, magnesium, titanium orcopper to steel currently exists in transportation industry. Alumi-num superstructures to steel hulls for ships, aluminum or magne-sium to steel tailor-welded blanks for automotive components,titanium to steel sheets for aeronautic and aerospace parts andmany other emerging applications are dependent on dissimilarmaterial joints for reducing the total component weight and/or im-prove performances. Dissimilar material joining presents chal-lenges significantly different from those explored during similarmaterial joining, due to the difference in chemical, mechanicaland thermal properties of involved metals/alloys [1]. Over theyears, similar materials have been coupled by using mechanicalinterlocking and connections, gluing or chemical bonding, welding,brazing and soldering. Some of these processes may become lim-ited if they are applied to dissimilar materials. For example, fusionwelding of clad materials results in too extended undesirable brit-tle inter-metallics if the process is not accurately control, limitingthe resulting weld quality. In particular defect-free welds can beobtained when limiting penetration to below 500 lm for alumi-num–steel joints [2], the appropriate filler to was selected [3] orthe diffusion was accurately controlled [4]. Consequently, there isa driving force to improve existing techniques and develop newmethods for joining such dissimilar lightweight materials.

Focusing the attention on the construction of yachts and fastvessels, the application of steel/aluminum transition joints is sig-

ll rights reserved.

nificant because the design criteria require the reduction of weightin several regions of the vessel. Bars and plates made of steel/alu-minum alloys with thicknesses greater than 20 mm are of relevantinterest. At present, this different material combination can bemainly joined via explosion welding. No alternative commercialtechnology is available to realize a direct bonding of these dissim-ilar materials with these great thicknesses [5]. In explosion bond-ing, the explosive and prime metal are placed together andspaced slightly away from the backing metal. When the explosivesare detonated, the prime metal collides with the backer metal. Jet-ting, which takes place ahead of the collision, acts to clean the jointzone and the clean surfaces are subjected to high pressures in thecollision region, causing the plastic deformation and materialjoining.

The direct assembly in service of the structural transition jointswith welding processes is became a crucial topic in manufacturing.With more and more special alloys being developed, choosing theright joining methods to employ with these explosion weldedstructural transition joints is very important [6]. The heat gener-ated during welding of steel/aluminum materials can result inmetallurgical structure changes that can alter the material corro-sion characteristics. The welded structure design and the tech-niques used to weld the structures can impact the useful life ofthe component, as it happens during heat treatment [7–9]. Theobjective of the present research is to investigate modification in-duced by laser beam welding onto steel/aluminum bonding inter-face. Bead-on-plate tests were performed on explosion welded barsand plates with different thicknesses, keeping constant processingparameters such as the laser power and travel speed and varyingthe distance between the fusion area and bonding interface. By

1982 L. Tricarico, R. Spina / Materials and Design 31 (2010) 1981–1992

the localized energy input of the laser beam and a controlled heatdistribution, a minimized interaction of the joined materials wasrealized, thus avoiding embrittlement of the joint.

2. Experimental activity

A tri-metallic structural transition joint (STJ) was chosen for thisstudy due to its industrial importance for the fast vessel construc-tion. The rough material was the Triclad

�STJ, a trade name of Mer-

rem & la Porte, produced by with open-air explosion welding. Inparticular, the selected rough material consisted of a ASTM A516grade 55 steel backer plate clad to a AA5083 flyer plate, with com-mercial purity aluminum (AA1050) interlayer plate placed be-tween the former two. The presence of the AA1050 interlayerwas necessary to improve STJ diffusion resistance with both ironand aluminum. The chemical compositions of the base materialsof the STJ are reported in Table 1.

The investigated STJ, realized by the supplier in compliancewith specification ASTM B898, was analyzed with ultrasonicinspection from the manufacturer to confirm the whole weld inter-face integrity.

Laser welded bead-on-plate were performed on AA5083 andASTM A516 steel plates as well as as-clad STJs. Tests on STJs wereperformed on both aluminum and steel side. The experimentalactivity was divided into:

– Preliminary analysis: Welding tests were performed on alumi-num and steel plates to identify the processing conditions touse during the following tests on STJs. The most appropriate pro-cessing conditions were selected by analyzing shape and dimen-sions of the weld cross sections.

– STJ analysis: Bead-on-plate welding were carried-out directly onSTJs with processing parameters previously identified. Weldingtests were performed on aluminum and steel sides, using spec-imens the thickness of which was modified by machining.

The trials kept the average power at the work surfaceconstant at 4.3 kW in continuous wave regime – approximately85% of full power. Additional process parameters kept constantduring all tests were the nozzle stand-off distance, beam focusposition and gas flow-rate, the values of which were300.0 mm, 0.0 mm (on the surface) and 35.0 Nl/min respectively.Helium was used as shielding gas. The laser head was angled of5.0� respect to the Z axis to avoid beam reflections. All surfacesof plates and specimens, interacting with laser beam, wereprepared by mechanical removing oxides and cleaning withacetone.

In the preliminary analysis, AA5083 plates (12.0 ± 0.3 mmthick) and ASTM A516 steel plates (14.0 ± 0.3 mm thick) were used.These thicknesses were quite equal to those existing in of STJs cladsections, allowing the reproduction of the same heat transfer con-ditions during welding. Several welding paths were realized oneach plate, waiting a time sufficient to take back the temperatureof the plate to ambient before performing the successive weld.The travel speed of the laser head varied in specific ranges in orderto achieve, at the same time, the partial deep penetration and

Table 1Chemical composition of base materials.

Material Elements (%)

AA1050 Si 0.25, Fe 0.40, Cu 0.05, Mn 0.05, Mg 0.05, Zn 0.07, Ti 0.05AA5083 Si 0.40, Fe 0.40, Cu 0.10, Mn 0.40–1.0, Mg 4.0–4.9, Cr 0.05–

0.25, Zn 0.25, Ti 0.15ASTMA516

grade 55C 0.20, Mn 0.60–1.20, P 0.035, S 0.035, Si 0.55

stable keyhole regime. Fig. 1 shows the weld top surface of sometrials made on AA5083 plates with travel speeds equal to 1.0, 1.5and 2.0 m/min.

Table 2 reports the maximum penetration depth and width ob-tained for these travel speeds.

The analysis of the weld profiles pointed-out that the best re-sults was achieved for travel speed equal to 1.5 m/min. Weldingtests on ASTM A516 steel plates were performed by operating inthe same manner of AA5083 plates. The scope was, in this case,the identification of the travel speed allowing welding profiles tobe comparable with those of AA5083 welds. A travel speed equalto 1.0 m/min was recognized as the target speed, resulting in amaximum penetration depth of 5.0 mm. Additional weld profilesparameters for this speed are reported in Table 2.

In the STJ analysis, welding tests on STJs were performed onspecimens machined from the same bar, which section and lengthwere 26.3�26.0 mm2 and 1000.0 mm, respectively. The materials ofthe STJ bar were steel – ASTM A516 grade 55 – and aluminum al-loys – AA1050 and AA5083, the thicknesses of which were13.7 ± 0.4 mm for steel and 11.7 ± 0.3 mm for aluminum alloys.Bead-on-plate were carried-out by using process parameters iden-tified in the previous preliminary analysis and varying only thejoint geometry. In this way, the laser-induced thermal load waskept constant on the top surface while Fe/Al interface was subjectto different thermal stresses due to the reduction of the specimenthickness. To achieve this result, upper section interacting with thelaser beam was machined to reduce the distance d between thefused zone and bond interface (Fig. 2).

For this reason specimens with different thicknesses were em-ployed. The specimen length, equal to 80.0 mm, was chosen toachieve stationary conditions during welding. Some trial tests onSTJ specimens were performed to verify if the main geometricalparameters of weld sections were the same of those achieved inthe preliminary analysis on the base materials. Weld geometrywas characterized by the penetration depth and width equal to5.4 and 5.0 mm, quite equal to those measured on the base mate-rials. The small variations between results were inputted to the la-ser beam power fluctuation, specimen geometry (plate vs. bar) aswell as materials compositions (mono-material vs. tri-materials).

The following tests on STJs were performed with the laser beampower equal to 4.3 kW and travel speed set to 1.5 or 1.0 m/minrespectively for aluminum or steel top surface. The specimenLPBAL1, LPBAL2 and LPBAL3 were realized by machining the alu-minum side (Table 3).

The distance d between the melt zone and the Fe/Al wasplanned to 3.0, 1.5 and 0.0 mm, assuming that the penetrationdepth remained constant to 5.5 mm. Table 3 also reports cut sec-tion, details of the weld fused area and geometrical parameters

Fig. 1. Welding trails on AA5083 plates.

Fig. 2. Principal weld geometrical dimensions.

Table 2Tests on AA5083 and ASTM A516 steel plates.

Weld Speed, v (mm) Power, (kW) Depth, h (mm) Width, r (mm) P/(v � h) (J/mm2)

AA5083 1.0 4.3 5.6 6.0 46.1

1.5 4.3 5.4 4.5 31.62.0 4.3 4.5 4.0 28.7

ASTM A516 1.0 4.3 5.0 4.5 51.6

L. Tricarico, R. Spina / Materials and Design 31 (2010) 1981–1992 1983

such as the penetration depth h and width r. In particular, thesetwo parameters decreased with the reduction of the distance d,in spite of the thermal load on the top surface was the same forall specimens. Possible justifications of this trend were the lowerconductivity of steel, that reduced the heat transmission at bondinterface as well as in the steel region, and the smaller aluminumthickness, that lowered the possibility of the laser-induced heatto quickly went away from the weld fused area. The decrease ofthe main weld parameters become more evident by comparingthe as-clad specimen with LPBAL3. The difference of the penetra-tion depth between these two specimen was equal to 0.9 mmwhile the difference of the penetration width was equal to0.3 mm. The same experimental framework was applied for speci-men welded from the steel side. The specimen LPBST1, LPBST2 andLPBST3, realized by the steel side, were machined from the samebar of the specimen processed from the aluminum side to elimi-nate differences due to manufacturing batch. The distance d be-tween the melt zone and the Fe/Al was planned to 3.0, 1.5 and0.0 mm, assuming that the penetration depth remained constantto 5.0 mm. Table 4 reports specimen dimensions, cut section, de-tails of the weld fused area and geometrical parameters such asthe penetration depth h and width r. The reduction of penetrationdepth h and width r with the decrease of the distance d were alsodetected for these tests but this reduction was considered asnegligible.

3. Micro-structural measurements at the steel/aluminum bondinterface

The study of micro-structure of the bond interface was manlyaddressed to the qualitative and quantitative analysis of the Fe/Alinter-metallic film in terms of variation of its extension and thick-ness. Sections of as-clad and welded specimen were cut 15.0 mmfar from specimen edges by using an abrasive wheel cut-off ma-chine (Discotom-2 of STRUERS) in transverse direction to thelength of the rough bar, taking care of minimizing the mechanicaland thermal distortions of the Fe/Al interface. These sections were

then prepared by grinding with 200–1000-grit silicon carbide pa-pers, followed by mechanical polishing from 6 lm to 1 lm dia-mond abrasive on short nap clothes. Etching was then performedon the steel side of specimens with Nital solution (2 ml HNO3

and 98 ml of C2H5OH) in distilled water for 15 s in order to high-light grain structures as well as inter-metallic phases. Keller’s re-agent (5 ml HNO3 and 190 ml of H2O) was applied for 15 s toaluminum side to point macro-structures.

The visual inspection of the STJ specimens by using the metal-lographic microscope was very useful to investigate modificationsof Fe/Al interface caused by laser-induced thermal loads. The as-clad specimen sections were initially analyzed and different areaswere detected, as Fig. 3 shows.

Ripples with different morphological characteristics were lo-cated at the interface. These ripples, formed from the rapidquenching of melt regions caused by explosion, consisted of a mix-ture of different inter-metallic phases, as the grey scale variationsuggests. Areas surrounding these ripples, and sometimes locatedinside them, exhibited the typical dendrite morphology of a slowcooling process after melting. Small-sized clusters of inter-metalliccompounds, formed in not equilibrium cooling conditions, werealso observed along the Fe/Al interface, pointing out the interfacediscontinuity. The cluster thickness ranged between 50 and160 lm. Along the Fe/Al interface, the inter-metallic phases weredetected as a discontinuous narrow film, less than 10 lm wide.This film, caused by the agglomeration of Fe and Al, was thick inareas submitted to high thermal gradient while it was very thinor absent in areas subjected to very low thermal gradient. The verybrittle inter-metallic phases identified in this band at room tem-perature in the as-clad STJ were the FeAl3 and Fe2Al5 on the alumi-num side and steel side respectively. Further metallographicfeatures were noted for the STJ base materials. The micro-structureof the ASTM A516 steel consisted of ferrite (lighter constituent)with pearlite (darker constituent). Small-sized elongated grains,characteristic of the cold-working conditions, were observed nearthe interface while medium-sized regular ones were identified inareas immediately after the Fe/Al interface until to the specimenboundaries. As concern the AA1050 side, the micro-structure con-sisted of insoluble FeAl3 particles (dark constituent) dispersed inthe aluminum matrix (lighter constituent). The morphology ofthese particles seemed to be not influenced by explosion welding[10]. Different welding interfaces (straight, wavy and continuoussolidified-melted) may be obtained by changing explosive weldingparameters such as stand-off distance, explosive loading and an-vils, as [11] reports.

The micro-structural measurements involved the use an opticalmicroscope connected to a high resolution digital camera and com-puterized image tool software. During the acquisition step, the en-tire Fe/Al interface of the specimen was captured by shootingmultiple images at different locations, performing the brightness/contrast adjustment, joining them in a single frame and finally

Table 3Tests on STJs – aluminum side.

1984 L. Tricarico, R. Spina / Materials and Design 31 (2010) 1981–1992

over-laying a 100 lm grid. The complete images of the Fe/Al inter-face of the different specimens were thus achieved (Table 5).

In the measurement step, the presence of inter-metallic phaseswas evaluated for each sector of 100 m length. These phases, dar-ker than aluminum and lighter than ferrite, were searched at inter-face. In case of a not very clear distinction between light and darkzones, the inter-metallic phases were considered as not present. Atthe same time, several random areas were selected from the entireimage of the Fe/Al interface and inter-metallic thickness measured.The above procedure was repeated for all specimens.

The results of the metallographic examinations are reported inTable 6 in terms of the reference length LREF and the inter-metallicextension LINT as well as the ratio between these twomeasurements.

The reference length LREF of 15.0 mm along the Fe/Al interfacewas chosen for all specimens. The results pointed-out that the in-ter-metallic extension LINT was greater than 50% of the total lengthLREF of the cut section in the as-clad specimen. The length LINT in-creased with the decrease of the distance d between weld fusedarea and bond interface, as Fig. 4 shows.

This increase was more rapid for specimens welded from thesteel side than those welded from the aluminum side. Anotherimportant aspect to underline was that the two main factors linkedto ripples and film growth contributed to the inter-metallic exten-sion value. Ripples existed in the as-clad specimen, as the mainfeature of the explosion welding process. The laser-induced heatloads influenced inter-metallic ripples, promoting their growthby inter-diffusion, but no new ripples were created during welding.The inter-metallic film was mainly promoted by laser-inducedthermal loads because it also aroused in areas in which it did notexist at all. The contribution of the inter-metallic film was by mademore evident by the LFILM/LREF ratio in Table 7. The enlargement ofthe inter-metallic film was also in this case higher for specimenswelded from the steel side. The inter-metallic film thickness wasevaluated in terms of average value of several random measuresnear the fused area. The maximum and the minimum thicknesseswere also evaluated, considering that these values were represen-tative of local conditions. The evaluation of the average value ofthe film growth made observations independent of the previousstate of the as-clad material, linking results to the effects of the

Fig. 3. Fe/Al interface (as-clad condiBon).

Table 4Tests on STJs – steel side.

L. Tricarico, R. Spina / Materials and Design 31 (2010) 1981–1992 1985

laser-induced heat. The average film thickness increased with thereduction of the distance d, as Fig. 5 shows, with the same trendindependently from the material of the welding side. The increaseof film thickness was greater for specimen welded from the alumi-num side than those welded from the steel side. This behavior,opposite to that recorded for the film extension growth, pointingout that laser-induced heat remained on aluminum side of thetri-material specimen because steel created a thermal barrier with

its lower thermal conductivity. This behavior was coherent to theresults achieved in the analysis of the post-weld heat treatmenton the interface micro-structure of explosively welded titanium–stainless steel composite in which the increase in the processingtemperature enhanced diffusion distance of Ti in the AISI 304stainless steel matrix [12].

The SEM analyses were then performed from the qualitativepoint of view by visually inspecting morphology of the Fe/Al

Table 5Fe/Al interfaces.

1986 L. Tricarico, R. Spina / Materials and Design 31 (2010) 1981–1992

interface. The back-scattered electron (BSE) images near the spec-imen Fe/Al interface showed the AA1050 in dark grey, the ASTMA516 steel in the light grey the ‘‘wavy” interfacial area with dif-

ferent grays, in function existing FexAly inter-metallics. Fig. 6 isone of the acquired BSE images for the specimen LPBAL1, weldedfrom the aluminum side. The analysis in three different positions

Fig. 4. RaBo LINT/LMEA vs. distance d.

Fig. 5. Average inter-metallic film thickness vs. distance d.

Table 6Inter-metallic extension.

Condition ID Distance, d (mm) Length, LREF (mm) Length, LINT (mm) LINT/LREF (%) LFILM/LREF (%)

As-clad – – 15.0 8.4 56.0 8.0

Welded from aluminum side LPBAL1 3.0 15.0 8.4 56.0 8.4LPBAL2 1.5 15.0 8.4 56.0 8.6LPBAL3 0.0 15.0 8.9 59.3 10.9

Welded from steel side LPBST1 3.0 15.0 8.4 56.0 8.4LPBST2 1.5 15.0 9.5 63.3 14.7LPBST3 0.0 15.0 10.4 69.3 16.1

L. Tricarico, R. Spina / Materials and Design 31 (2010) 1981–1992 1987

allowed to show different weight content of Al and Fe, pointingout the presence of several inter-metallic compounds. Howeverthe number was limited, as the gray variations highlighted. Thesame conditions was pointed out by analyzing the BSE image ofthe specimen LPBAL2, also welded from the aluminum side(Fig. 7). Different conditions were detected when specimens werewelded from the steel side (Figs. 8 and 9). The number of inter-

Table 7Inter-metallic film thickness.

Condition ID Distance, d (mm) Max. HF

As-clad – – 10.29

Welded from aluminum side LPBAL1 3.0 17.45LPBAL2 1.5 14.07LPBAL3 0.0 17.70

Welded from steel side LPBST1 3.0 11.53LPBST2 1.5 10.65LPBST3 0.0 11.05

metallic compounds were more numerous and fragile, as figuresshow. This experimental evidence was probably due by the lowerconduction coefficient of steel than that of aluminum, that causedheat to be slowly removed after welding. Figs. 6 and 7 also reportthe results of the micro-analysis of the area indicated from thethree shapes (circle, rectangle and triangle). With energy disper-sive X-ray spectrometers (EDS), chemical compositions was deter-mined quickly. Despite the ease in acquiring X-ray spectra andchemical compositions, the potentially major sources of errorwere minimized by optimizing the operative conditions necessaryto improve the statistical meaning of the electron counter. In par-ticular, the scanning area was equal to 1 lm2, the incident energywas 25 keV on the specimen surface with a working distance of10 mm (in this way the X-ray take-off distance was equal to35�), the electronic current was tuned in order to generate a X-ray counter rate of 2000 pulse per second and the effective coun-ter time was equal to 100 s [13].

4. Mechanical strengths of laser welded specimens

The mechanical characterization of the welded specimens al-lowed the modifications to the mechanical properties (shear andtensile strengths) caused by laser beam interaction to be evaluated.

Shear test was very important to assess transition joint techni-cal conformity by measuring the maximum shear strength of thebond interface and comparing it to the maximum shear strengthof the weakest material of the STJ [14]. Prismatic samples were fab-ricated from Fe/Al STJ (Fig. 10) in compliance with ASTM A264-03standard by removing the majority of the weakest clad materials(AA5083 and AA1050) and leaving a small nub with length wand thickness a. The ratio between w and a was 1½ with the max-imum dimension of w equal to 3.18 mm (1/80 0). Other importantdimensional features were the specimen width equal to25.40 mm (10 0), length equal to 63.50 mm (2½0 0), distance betweenside surfaces equal to 19.05 mm (3/40 0) and minimum steel thick-ness t greater than 2w. For thickness t lower than that specifiedby standard, the specimen dimensions were scaled down, keepingthe distance between side surfaces and ratio w/a unchanged.

The shear strength of the test piece was defined as the peakshearing load divided by the sheared area. Interfacial bonding ofthe clad specimen was evaluated by shear stress measured with

ILM (lm) Avg. HFILM (lm) Min. HFILM (lm) FILM growth (%)

6.51 3.10 –

8.11 3.51 258.27 4.58 278.95 3.01 38

7.30 3.92 127.40 3.43 148.21 5.53 26

Fig. 6. SEM analysis – specimen LPBAL1.

Fig. 7. SEM analysis – specimen LPBAL3.

Fig. 8. SEM analysis – specimen LPBST1. Fig. 9. SEM analysis – specimen LPBST3.

1988 L. Tricarico, R. Spina / Materials and Design 31 (2010) 1981–1992

a universal test machine INSTRON 4485 with a 200 KN loading celland a home designed testing equipment in which the test pieceand the jig were fitted (Fig. 11). The equipment was designed byrespecting dimensional features of the universal test machine, fastassembly and disassembly, easy specimen blocking and high posi-tioning accuracy as well as flexibility of use for other requiredtests.

The tensile strength of the bond zone was evaluated by RamTensile Test. A specimen with a particular parallelepiped/cylindri-cal shape was used in order to concentrate stresses in the sectionjust above the transition area corresponding to the bond interface,as Fig. 12 shows.

The force was applied to the direction perpendicular to theplanes delimiting different material interface. The MIL-J-24445A

standard specifies maximum dimensions of the bi-material speci-men but its principles can be easily extended to tri-material joints.In this case, the strength of the steel/AA1050 bond interface wasestimated rather than that of the AA5083/AA1050 one. The speci-men, punch and matrix prescribed by the above cited standardare reported in Fig. 13.

The matrix was fixed while the punch moved in contact withthe bond interface, compressing it until rupture. The MIL-J-24445A standard does not give values or suggestions linked totesting parameters (e.g. punch speed) and/or specimen clamping.The specimen was simply rested on the matrix top surface, allow-ing its rotation around the cylindrical axis. In the hypothesis thatthe pressure between punch and specimen during contact wasuniform, the applied load could be considered as axial-symmetric.

Fig. 11. Shear test – equipment.

Fig. 10. Shear test – specimen.

Fig. 12. Ram tensile test – specimen.

L. Tricarico, R. Spina / Materials and Design 31 (2010) 1981–1992 1989

The thickness D of the bond zone is always equal to1.72 ± 0.1 mm, independently from the values of steel and alumi-num thicknesses. The thickness D was achieved by simply varyingthe hole depth.

Shear and ram tensile samples were achieved from the sameplate with the sampling scheme shows in Fig. 14 in order to avoiddifference in STJ lot characteristics and based on literature results.In fact a previous work showed that little difference in the shear

Fig. 13. Ram tensile test – equipment.

Fig. 14. Specimen sampling.

1990 L. Tricarico, R. Spina / Materials and Design 31 (2010) 1981–1992

strength values of samples taken parallel and transverse to the det-onation existed [14]. In addition, the shear strength perpendicular

to the detonation appeared to be inversely proportional to theexplosive thickness whereas the shear strength parallel to thedirection of detonation remained almost constant.

The laser beam passed at the center of the small nub of theshear test specimens and sufficient far from the ram tensilespecimens. The area of the small nub of the shear specimenswere consequently subjected to the highest thermal stresseswhile the bonding area AA1050/steel of the ram tensile speci-mens were uniformly thermally loaded. The process parametersused for bead-on-plate welds were the same of those employedfor STJ bars in the previous experimental step. Increasing ther-mal loads at the bond interface were achieved by removingmaterial from the surface interacting with the laser beam. Thereduction of the plate thickness required the scaling down ofspecimen dimensions for some samples. Tables 8 and 9 reportthe dimensions of shear and ram tensile test specimens. Anadditional test specimen was cut at the center of the plate toevaluate the maximum welding penetration depths in compari-son with those of the welded bars as well as hardness values[15].

The shear test were performed with the mobile crosshead mov-ing at 3.0 mm/min. The acquisition of several high resolution digi-tal images during tests was useful to visually understand themechanisms of deformation of the small nub (Fig. 15), comparedwith numerical data.

The two repetitions for each welding conditions were character-ized by load–displacement curves wholly overlaid, as Fig. 16 showsfor samples B1 and B2.

The evolution of stress–displacement curve initially presented arapid increase of the stress value, its stabilization and finally its ra-pid reduction. The rupture was localized in the AA1050 and not atinterface AA1050/steel, justifying the trend of this loading curve.Table 8 reports the final results of all tests in term of maximumshear load and stress.

Table 8Shear test – sample dimensions and results.

Condition ID Sample Al thick (mm) Fe thick (mm) a (mm) w (mm) t (mm) T (kN) s (MPa)

As-clad – A1–A2 11.7 13.7 3.00 4.50 9.00 10.0 87.4

Welded from aluminum side LPBAL1 B1–B2 8.5 13.7 3.00 4.50 9.00 10.1 88.3LPBAL2 C1–C2 7.0 13.7 3.00 4.50 9.00 9.3 81.7LPBAL3 D1–D2 5.5 13.7 3.00 4.50 9.00 9.0 78.7

Welded from steel side LPBST1 E1–E2 11.7 9.0 3.00 4.50 9.00 10.0 87.4LPBST2 F1–F2 11.7 7.5 2.50 3.75 7.50 8.8 86.6LPBST3 G1–G2 11.7 6.0 2.00 3.00 6.00 6.3 83.0

Table 9Ram tensile test – sample dimensions and results.

Condition ID Sample Al thick (mm) Fe thick (mm) Diameter (mm) F (kN) r (MPa)

As-clad – RA1–RA2 11.0 13.0 12.7 29.6 235.3

Welded from aluminum side LPBAL1 RB1–RB2 8.0 10.0 9.7 28.5 226.4LPBAL2 RC1–RC2 6.5 10.0 8.2 26.8 213.0LPBAL3 RD1–RD2 5.0 10.0 6.7 23.9 199.9

Welded from steel side LPBST1 RE1–RE2 11.0 8.5 12.7 28.1 223.7LPBST2 RF1–RF2 11.0 7.0 12.7 26.6 211.2LPBST3 RG1–RG2 11.0 5.5 12.7 26.1 206.8

Fig. 16. Shear stress vs. punch stroke – samples B1 and B2. Fig. 17. Shear stress vs. punch stroke – samples B1, C1 and D1.

Fig. 18. Shear stress vs. punch stroke – samples E1, F1 and G1.

Fig. 15. DeformaBon Bmes of sample B1.

L. Tricarico, R. Spina / Materials and Design 31 (2010) 1981–1992 1991

All stress values recorded during tests were decidedly higherthan 50–60 MPa prescribed from Lloyd’s Register of Shipping,revealing the good fabrication quality of the observed STJ. Resultsalso pointed-out that the reduction of the specimen thicknessand the consequent reduction of the distance between weld fusedarea and bond interface caused the decrease of the maximum shearstrength. This decrease was more evident for specimen weldedfrom the aluminum side than those welded from the steel side,as Figs. 17 and 18 show. The ram tensile test were then carried-out. Two repetitions for each welding condition was useful to as-sess test repeatability. Fig. 19 reports results of the samples RA1and RA2, in terms of stress–displacement in which the maximumtensile stress, equal to 235.3 MPa and corresponding to a maxi-mum load of 29.6 KN, was equal for the two samples.

Fig. 19. Tensile stress vs. punch stroke – samples RA1 and RA2.

1992 L. Tricarico, R. Spina / Materials and Design 31 (2010) 1981–1992

The rupture was always localized at the Fe/Al interface due tothe specimen shape. Table 9 reports the final results of all testsin term of maximum tensile load and stress. The results showedthe same trend detected during the shear test linked to the moreevident reduction of the final strength of specimens welded fromthe aluminum side than those welded from the steel side. Thereduction of the distance between the fused area and bond inter-face was less important because specimens were realized in areafar from the laser beam interaction and consequently they weresubjected to mild laser-induced thermal loads.

The comparison between mechanical results and inter-metallicfilm thickness was very interesting. The reduction of the maximumtensile and shear stresses could be inputted to the increase of theinter-metallic film thickness. In fact lower values of the mechanicalstrength was detected for higher values of the film thickness. Thishypothesis also confirmed that specimen welded from the steelside were more critical than those welded from the aluminum side.However, the mechanical strength of the welded specimens wereonly blindly affected by the laser beam interaction because themeasured strengths were much more higher than those normallyrequired.

5. Conclusions

In this work the influence of different thermal loads caused bylaser beam on mechanical strength of the Fe/Al explosion weldedjoints has been investigated. The present research has investigatedthe application of laser beam to weld components made of struc-tural transition joints (STJs) between steel and aluminum basematerials. The metallographic examinations and the tensile testshave pointed-out that laser-induced thermal loads strongly pro-moted inter-metallic growth at steel/aluminum interface.However, these loads have limited influence on integrity of bead-on-plate welds because rupture mainly occurred in the weakest

material (aluminum alloy). The condition has been also verifiedby reducing the thickness of the STJ.

Acknowledgements

The present research is only a part of activities performed byConsortium for Laser and Electron Beam Applications (CALEF Con-sortium), the main members of which are the CRF, ENEA, Polytech-nic of Bari and Rodriquez Cantieri Navali SPA. In particular,Rodriquez Cantieri Navali SPA directly funded the present researchthrough the ENVIROALISWATH project.

The authors wish to thank Dr. Gianframco Palumbo and Dr.Donato Sorgente of DIMeG – Politecnico di Bari, Dr. Marco Brandiz-zi of the Fiat Research Center (CRF), Enzo PUTIGNANO and RobertoDebonis of CALEF Consortium, Laura CAPODIECI of ENEA – Dept.FIM – Composites & Nanostructured materials Section for theirprecious collaboration.

Reference

[1] Chen L et al. Clad metals, roll bonding and their applications for SOFCinterconnects. J Power Sources 2005;152:40–5.

[2] Sierra G et al. Steel to aluminium key-hole laser welding. Mater Sci Eng A2007;447:197–208.

[3] Missori S, Murdolo F, Sili A. Single-pass laser beam welding of clad steel plate.Weld J 2004;02:65–71.

[4] Kobayashi S, Yakou T. Control of intermetallic compound layers at interfacebetween steel and aluminum by diffusion-treatment. Mater Sci Eng A2002;338:44–53.

[5] Mousavi SAAA, Barrett LM, Al-Hassani STS. Explosive welding of metal plates. JMater Process Technol 2008;1–3:224–39.

[6] Durgutlu A, Gülenç B, Findik F. Examination of copper/stainless steel jointsformed by explosive welding. Mater Design 2005;26:497–507.

[7] Lee JE et al. Effects of annealing on the mechanical and interface properties ofstainless steel/aluminum/copper clad-metal sheets. J Mater Process Technol2007;187–188:546–9.

[8] Naoi D, Kajihara M. Growth behavior of Fe2Al5 during reactive diffusionbetween Fe and Al at solid-state temperatures. Mater Sci Eng A2007;459:375–82.

[9] Li S-X et al. Fatigue damage of stainless steel diffusion-bonded joints. Mater SciEng A 2008;480:125–9.

[10] Tricarico L, Spina R, Sorgente D, Brandizzi M. Effects of heat treatments onmechanical properties of Fe/Al explosion-welded structural transition joints.Mater Design 2009;30:2693–700.

[11] Acarer M, Gülenç B, Findik F. The influence of some factors on steel/steelbonding quality on there characteristics of explosive welding joints. J Mater Sci2004;39:6457–66.

[12] Mousavi SAAA, Sartangi PF. Effect of post-weld heat treatment on the interfacemicrostructure of explosively welded titanium–stainless steel composite.Mater Sci Eng A 2008;494:329–36.

[13] Capodiceci L. Analysis of explosion welding steel/aluminum transition jointwith scanning electronic microscope and X-ray microanalysis. ENEA InternalTechnical Report, RT/ENEA/ 06/11 of FIM COMP Dept.

[14] Mousavi SAAA, Al-Hassani STS, Atkins AG. Bond strength of explosively weldedspecimens. Mater Design 2008;29:1334–52.

[15] Tricarico L et al. Effects of laser welding on properties of Fe/Al explosionwelded structural transition joints. In: ATA proceedings of 2nd internationalconference on advanced materials and technologies for transportationindustry. Italy: ELASIS in Pomigliano D’Arco; 2007.