Ti joints with AZ91 Mg based filler

    Microstructure and mechanical properties of laser welded-brazed Mg/Ti joints with AZ91 Mg based filler Caiwang Tan, Bo Chen, Shenghao Meng, Kaiping Zhang, Xiaoguo Song, Li Zhou, Jicai Feng PII: DOI: Reference:

S0264-1275(16)30355-0 doi: 10.1016/j.matdes.2016.03.073 JMADE 1551

To appear in: Received date: Revised date: Accepted date:

29 January 2016 13 March 2016 14 March 2016

Please cite this article as: Caiwang Tan, Bo Chen, Shenghao Meng, Kaiping Zhang, Xiaoguo Song, Li Zhou, Jicai Feng, Microstructure and mechanical properties of laser welded-brazed Mg/Ti joints with AZ91 Mg based filler, (2016), doi: 10.1016/j.matdes.2016.03.073

This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

ACCEPTED MANUSCRIPT Microstructure and mechanical properties of laser welded-brazed Mg/Ti joints with AZ91 Mg based filler

IP

Xiaoguo Songa, Li Zhoua, Jicai Fenga,b

T

Caiwang Tana,b*, Bo Chena, Shenghao Menga, Kaiping Zhanga,

SC R

a. Shandong Provincial Key Laboratory of Special Welding Technology, Harbin Institute of Technology at Weihai, Weihai 264209, China

Technology, Harbin 150001, China

NU

b. State Key Laboratory of Advanced Welding and Joining, Harbin Institute of

MA

* Corresponding author: Caiwang Tan. Tel./Fax: +86 631 5678211 E-mail address: ([email protected])

D

Abstract

TE

AZ31B Magnesium (Mg) alloys and Ti-6Al-4V titanium (Ti) alloys were joined by laser welding-brazing process with AZ91Mg based filler. Uniform and continuous

CE P

weld surfaces without obvious defects were produced in a relatively large processing window. In the process Al element diffused from the filler and reacted with Ti

AC

resulting in metallurgical bonding of Mg/Ti joint. An ultra-thin reaction layer with serrate-shaped morphology was evidently observed at the interface of AZ91 fusion zone/Ti. The thickness was varied slowly with the change of the heat input. Newly formed interfacial compound was then identified as Ti3Al phase by transmission electron microscopy (TEM) analysis. The maximum tensile-shear strength reached 2057 N, representing 50% joint efficiency relative to Mg base metal. Two kinds of fracture modes were noticed during the tensile-shear test. Observation of fracture surfaces suggested some reaction products were attached to Ti substrate in both cases, which was found to prevent crack propagation effectively and thus improved the joint strength. 1

ACCEPTED MANUSCRIPT Key words: laser welding-brazing; dissimilar metals; microstructure; mechanical properties

T

1. Introduction

IP

In advanced manufacturing industry, dissimilar materials combination not only

SC R

makes the best use of the properties of each material, but also provides the great flexibility using different materials in one product [1, 2]. The hybrid structure of Mg-Ti dissimilar metals has been of particular interest since both metals have some

NU

excellent features, such as low density, high specific strength and good formability. It

MA

can offer significant advantages over weight reduction in automobile and aerospace industries, thereby improving fuel efficiency and load capacity. Therefore, reliable

D

joining of Mg to Ti must be addressed, which will in turn expand the engineer

TE

application of Mg alloy and Ti alloy in various fields. However, the great differences in physical and metallurgical properties between

CE P

Mg and Ti pose a huge challenge when joining the two metals. The melting points of Mg and Ti are 649oC and 1678oC, respectively. The evaporation point of Mg alloy is

AC

only 1091oC. It suggests that catastrophic evaporation of Mg element is inevitable when melting Mg and Ti simultaneously using conventional fusion welding process, which is thus not applicable for reliable joining of Mg to Ti. Furthermore, Mg and Ti are immiscible according to Mg-Ti binary diagram, indicating that no intermediate reaction layer or atomic diffusion occurs after solidification. Therefore, an intermediate element which can react with or possess substantial solid solubility in Mg and Ti must be added to realize metallurgical bonding of the two metals. Various methods such as friction stir welding (FSW) [3, 4], transient liquid phase (TLP) welding [5-8], cold metal transfer (CMT) welding [9], tungsten inert gas (TIG) welding [10, 11], laser welding [12, 13] have been adopted to join Mg alloys and Ti 2

ACCEPTED MANUSCRIPT alloys in previous works. In the case of the FSW process, Al element from Mg base metal was employed to bond with Ti giving rise to metallurgical bonding at the

T

interface. The results indicated that Ti-Al intermetallic compound layers formed at the

IP

interface due to external force and stirring effect, while magnesium alloy with higher

SC R

Al content could result in the formation of thick reaction layer deteriorating the tensile strength [3, 4]. In the TLP welding process, Ni coating was first electrodeposited on the Ti surface. Eutectic structure was observed to form at the Mg/Ni alloy interface

NU

and a solid-state diffusion bonding was produced at the Ni/Ti alloy interface. The

MA

maximum shear strength was 34 MPa. After study the thickness of Ni coating was proved to affect the joint strength [6]. Additionally, nano-scale Ni or Cu dispersion in

D

Ni coating was reported to accelerate the TLP bonding process because of the shorter

TE

diffusion distances compared to traditional coated foil. Cao et al. [9] reported that formation of Ti3Al at the interface could be attributed to element Al diffusing from the

CE P

molten AZ61 filler wire and reacting with Ti base metal. The metallurgical bonding at the interface produced peak load of 2.1 kN. In addition, laser keyhole welding was

AC

also utilized to butt join Mg and Ti [12, 13]. Laser beam offset was found to play an important role in the bonding mechanism and properties of the joint. The investigation of bonding mechanism suggested that intermixing of molten Ti-6Al-4V with the liquid AZ31B caused the formation of lamellar and granular mixtures in the fusion zone, which yielded acceptable joints with highest tensile strength of 266 MPa. Laser welding-brazing technique has the great potential for increased flexibility and adaptability when joining of dissimilar metals such as Al/steel [14], Al/Ti [15, 16] and Mg/steel [17, 18]. It used welding for materials with low melting point, and brazing for materials with high melting point. Joining of Mg to steel has the same challenge as joining of Mg to Ti because of their similar immiscibility characteristics. 3

ACCEPTED MANUSCRIPT In our previous studies [17, 18], metallurgical bonding of Mg/steel was achieved by Zn coating or interdiffusion of alloying element from base metal. Al-Fe intermediate

T

phase was produced by dissolution and diffusion from filler metal and steel substrate

IP

when locally experienced high temperature.

SC R

To improve the flexibility of controlling the interfacial reaction, different kinds of alloying elements were usually added into the filler wire in dissimilar materials joining [19, 20]. In this work, Al element was selected as the intermediate element to

NU

bond Mg and Ti based on the Mg-Al and Al-Ti binary diagrams. AZ91 filler wire was

MA

employed which could contain maximum content of 9 wt. % Al without any crack after wire drawing. The objective of the current work is to study the characteristics of

D

laser welding-brazing of Mg alloys to Ti alloys with AZ91 filler wire. Weld

TE

appearances at different welding parameters were observed. The microstructure at the interface was characterized and mechanical properties were evaluated. Based on these

CE P

analyses, the bonding mechanism of Mg and Ti was expected to be clarified. 2. Experimental details

AC

The materials used in this work were AZ31B magnesium alloys and Ti-6Al-4V alloys, both with a thickness of 1.5 mm. Both sheets were machined to rectangular strips of 100 mm×30 mm. The filler wire adopted was AZ91 filler with diameter of 1.2 mm. The chemical compositions of two base metals and filler metal are listed in Table 1 and Table 2. The experiments were performed with a 10 kW fiber laser (IPG YLR-10000). The laser beam had a wavelength of 1070 nm and a beam parameter product of 7.2 mm mrad. The focused spot size was 0.2 mm. The schematic of laser welding-brazing Mg to Ti is illustrated in Fig. 1. The assembly was fixed with Mg sheet placing on top of Ti sheet in a lap configuration. Both sheets were ground and degreased prior to 4

ACCEPTED MANUSCRIPT welding. Laser beam was irradiated on the surface of Ti vertically. Filler wire was fed in front of the laser beam. The angle of the filler wire and the workpiece were

T

adjusted for smooth wire feed. To completely irradiate the filler metal and promote

IP

brazing between molten filler metal and titanium, the laser beam was defocused. The

SC R

defocus distance from steel surface was +10 mm. Argon gas was provided to prevent oxidation of the molten filler metal with the flow rate of 20 L/min. The commercial flux used in the experiment was Superior No. 21 manufactured by Superior Flux and

NU

Manufacturing Company. The powder flux consisted of LiCl (35-40 wt.%), KCl

MA

(30-35 wt.%), NaF (10-25 wt.%), NaCl (8-13 wt.%) and ZnCl2 (6-10 wt.%). The process parameters adopted in the present work referred to the results achieved laser

D

welding-brazing of Mg to steel in our previous study [17], since they had the similar

listed in Table 3.

TE

welding characteristics. The optimized process parameters used in the experiment are

CE P

After laser welding-brazing process, test specimens were cut transverse to the weldment. Standard metallographic preparation procedures were utilized. The

AC

reaction layer along the interface was observed by scanning electron microscope (SEM). Transmission electron microscopy (TEM) with a Tecnai-G2 F30, operating at a nominal voltage of 300 kV, was used to characterize the microstructure in detail. Z-contrast images were acquired using a high angle annular dark field (HAADF) detector in scanning transmission electron microscopy (STEM) mode. Phase identification was performed by selected-area electron diffraction (SAED) accompanied with energy dispersive spectroscopy (EDS). Vickers hardness measurement was performed across the Mg fusion zone/Ti interface under a test load of 100 g and a dwell time of 10 s. The 10-mm-wide, rectangular-shaped specimens were cut and subjected to tensile-shear test which was evaluated by a testing machine 5

ACCEPTED MANUSCRIPT (INSTRON-5569) at a cross-head speed of 1 mm/min at room temperature, according to GB/T 2651-2008 standard (equivalent to ISO 9016: 2001) [21]. Fracture load was

T

calculated via the tensile testing of at least three specimens.

IP

3. Results and discussion

SC R

3.1 Weld appearances

Fig. 2 shows typical appearances of laser welded-brazed Mg/Ti joints made at different welding parameters. For joint MT1, the rough surface and non-continuous

NU

weld was observed at low heat input as shown in Fig. 2a). Incomplete fusion welding

MA

of base metal and molten filler occurred resulting in a void arrowed in Fig. 2a). It was because most of energy from the laser beam was used to melt the filler wire and

D

pre-coated flux causing insufficient wetting of molten Mg filler on the Ti surface.

TE

With the increase of laser power to 2000 W, uniform and visually acceptable joints without obvious defects were obtained, indicating a stable process as shown in Fig.

CE P

2b). Small voids were also found at the surface of Mg base metal near the fusion zone, which was caused by slag erased after the process. Stable appearances were visible in

AC

a relatively wide processing window from joints MT-3 to MT-6. However, some evaporation of filler metal was observed when using high laser power of 3200 W at relatively high welding speed of 1 m/min, in the case of joint MT-7 as indicated in Fig. 2f). Fig. 3 presents cross sections of laser welded-brazed Mg/Ti joints. Fusion joint formed by mixing molten Mg base metal and AZ91 Mg based filler wire, while brazed joint was produced by molten filler and Ti substrate. The main differences in these joints achieved in the present work were distinguished from their wettability. The more surface wetted out, the smaller the contact angle. With the increase of heat input, the contact angle became lower as expected, compared with Fig. 3a) and Fig. 3b). 6

ACCEPTED MANUSCRIPT Note that the contact angle of joint produced with high laser power was a little lower than that with low laser power despite the same heat input (Fig. 3a and Fig. 3d). It

T

could be attributed to the higher local energy density under the condition of high laser

IP

power. This suggested that high power at high welding speed was beneficial to the

SC R

wettability of molten filler on the Ti and thereafter mechanical properties. 3.2 Interfacial microstructure

Fig. 4 shows the interfacial microstructure of laser welded-brazed Mg/Ti joints

NU

achieved using parameters listed in Table 3. The location where the interface was

MA

observed is indicated in the inset of Fig. 4a). All the micrographs were taken at the back-scattered electron (BSE) mode. No obvious interfacial layer was observed when

D

using AZ31 Mg based filler wire (3wt.% Al in the filler) with the same parameter as

TE

the joint MT-1. Mechanical bonding was obtained when direct joining Mg to Ti or using Mg based filler with low Al content, which was reported in our previous study

CE P

[22]. In contrast, an ultra-thin interfacial reaction layer was evidenced along the Mg/Ti interface at high magnification shown in Fig. 4b). The feature was found to

AC

exhibit continuous and serrate-shaped morphology. The thickness of the layer was less than 1μm. That was to say, metallurgical bonding rather than mechanical bonding was achieved at the fusion zone (FZ)/Ti interface when using filler with more Al elements. The feasibility of improving bonding between immiscible Mg and Ti by adding more Al content (9 wt.%) into the Mg based filler was thus confirmed. With the increase of heat input, the reaction layer was characterized by block-shaped morphology (Fig. 4d). Some grew into the fusion zone adjacent to the interface as shown in Fig. 4e), producing island-like reaction products. The thickness of reaction layer at the FZ/Ti interface was varied slowly, with maximum thickness of 1.5 μm as indicated in Fig. 4d). At the high welding speed of 1m/min, the reaction layer in the joint MT-5 was the 7

ACCEPTED MANUSCRIPT thinnest among all the joints due to the lowest heat input in the study. With further increasing heat input, the reaction layer grew slightly. The thickness of reaction layer

T

of joint MT-7 was similar to that of joint MT-1 at the same heat input.

IP

Concentration profiles of the main alloying elements across the fusion zone/Ti

SC R

interface were obtained from line scanning analysis (shown in Fig. 4a) and the results were presented in Fig. 5. The results showed that Mg content decreased and Ti increased gradually from the Mg fusion zone side to the Ti side. No enrichment of Al

NU

element was found across the interface of fusion zone/Ti joint produced with AZ31

MA

filler as shown in Fig. 5a). However, an apparent Al segregation was observed at the interface of AZ91 fusion zone/Ti shown in Fig. 5b), indicating the occurrence of

D

atomic diffusion or dissolution of Al atoms from filler. It then enriched at the interface

TE

and induced interfacial reaction resulting in the Ti-Al intermetallic compounds. Relatively high concentration of Al was noticed at the FZ/Ti interface in all applied

CE P

heat inputs. Besides, the segregation zone of Al atoms was observed to become wider as the heat input increased as shown in Fig. 5c). Similar phenomenon was observed

AC

when using arc welding-brazing method [9]. However, an obvious and ultra-thin reaction layer formed at the interface was not observed in the study. The formed layer was so ultra-thin in the present work that the above EDS result was not accurate enough to identify its phase structure. Therefore, further investigation via TEM analysis was required. The results in Figs. 4-5 suggested small change in thickness of Ti-Al reaction layer with the variation of heat input. Similar phenomenon was observed when laser welding-brazing of Mg to steel [23]. The main reason for it could be the diffusion-controlled characteristics of interfacial reaction and fast laser heating and cooling rate. It was believed that the precipitation and growth of Ti-Al phase was 8

ACCEPTED MANUSCRIPT controlled by atomic diffusion which was highly dependent on the amount of activated Al atoms diffusing from fusion zone to the interface nearby. However, the Al

T

content was only 9 wt.% in our work. Therefore, the diffusion-controlled growth of

IP

Ti-Al phase was limited leading to the formation of ultra-thin reaction layer as shown

SC R

in Fig. 4. Meanwhile, the characteristics of fast heating and cooling rate also restricted the interfacial reaction during the laser welding-brazing process. The interface was heated to a high temperature in a short time which could enhance mutual solubility of

NU

Al and Ti. The liquid/solid interface was then cooled down without too much

MA

precipitation and growth of reaction layer. As a result, the metallurgical bonding was achieved while the thickness of the reaction layer was kept far below 10 µm, which

D

was beneficial to the joint strength.

TE

TEM analysis was performed to further identify the composition and structure of the reaction layer formed at the interface of FZ/Ti. Fig. 6 shows TEM bright field

CE P

image and HAADF micrograph with corresponding SADP taken at the interface. An ultra-thin interfacial reaction layer differing from Ti substrate was clearly observed

AC

from the TEM image shown in Fig. 6a). Block and rod shaped reaction layer was found at higher magnification (Fig. 6b). Interfacial microstructure characteristics of Mg/Ti joint were evidenced from HAADF micrograph as shown in Fig. 6c). Compositional analysis by TEM-EDS suggested that the reaction layer at location P1 was mainly composed of 70.06 at.% Ti, 26.78 at.% Al and 3.15 at.% Mg. Based on SADP calibration result, the intermetallic compound was indexed as Ti-rich Ti3Al phase. 3.3 Mechanical properties Fig. 7 shows hardness distribution profile across the fusion zone/Ti interface of joint MT-3. The average hardness values of AZ91 fusion zone and Ti were 62 HV and 9

ACCEPTED MANUSCRIPT 315 HV, respectively. It was worth noticing that the hardness of AZ91 fusion zone/Ti adjacent to the interface increased sharply, which was higher than the weld and close

T

to the neighboring Ti substrate. The abrupt change in hardness value was closely

IP

associated with the formation of an intermetallic compound and diffusion of Al atoms

SC R

to the Ti side, which was in good consistence with the observation in Fig. 4 and Fig. 6. The size of the microhardness indenter was too large to measure precisely the hardness of the thin IMC layers formed at the interface. However, higher hardness

NU

values were expected for the IMC layer, since the reported average hardness of the

MA

Ti3Al phase was 414 HV which was much higher than that of base metals [24]. Similar distinct rise in hardness at the interface was also observed when CMT

D

welding-brazing of Mg and Ti using AZ61 filler wire [9]. In contrast, no distinct

TE

change was observed when direct joining or using filler which has the same chemical composition as the AZ31 Mg base metal. It was attributed to no obvious atomic

CE P

diffusion at the interface. Detailed description about difference in the hardness value has been reported elsewhere [22].

AC

Tensile-shear test was performed to evaluate the joint strength and the result is shown in Fig. 8. The fracture load increased first from joint MT-1 to MT-3 with the increase of laser power at low welding speed of 0.5 m/min. The maximum value attained 2057 N in the case of joint MT-3, representing 50 % joint efficiency with respect to Mg base metal. The joint strength then decreased with further increase in laser power at high welding speed. Two main fracture modes were found from all the joints: interfacial failure and fusion zone fracture which was divided in the Fig. 8. At low welding speed, the increase of laser power or heat input promoted the wetting and spreading of molten filler metal on the Ti substrate. The mutual diffusion of Al and Ti atoms were accelerated giving rise to the increasing fracture load. However, further 10

ACCEPTED MANUSCRIPT increase of laser power caused excessive evaporation of Mg based filler metal, which resulted in weak interfacial bonding and thereafter the reduced joint strength. Note

T

that the tensile-shear strength of joint obtained at high welding speed was higher than

IP

that of joint made at low welding speed at the same heat input. The main reason for it

SC R

could be the difference in formation process of reaction layer. At rapid laser heating and cooling characteristics, the interface experienced at high temperature accelerated the diffusion process resulting in the high bonding force.

NU

Fig. 9 presents SEM morphologies of fracture surfaces of Mg/Ti joints with

MA

different fracture behaviors after tensile-shear testing. For the Mg/Ti joints fractured with interfacial failure mode as shown in Fig. 9a), the feature of its fracture surface

D

exhibited mixed rupture characteristics of tear ridge and dimples at Mg side indicated

TE

in Fig. 9b). At the Ti side, the fracture surface was characterized by fine particles mingled with residual Mg indicated in Fig. 9c). These dispersed particles on the Ti

CE P

side were mainly composed of 19.6 at.% Al, 42.3 at.% Ti, according to the EDS result of P1 at higher magnification shown in the inset of Fig. 9c). It corresponded to the

AC

observation of interfacial reaction layer in Fig. 4. The serrate-shaped microstructure reaction layer formed at the interface was attached to the Ti substrate during tensile-shear testing, which was mainly responsible for high resistance to the crack propagation. However, in the case of joint MT-3 fractured at the fusion zone side, the crack propagated along the FZ/Ti interface at the seam root and then moved to the fusion zone due to high interfacial bonding (Fig. 9d). Magnified SEM image indicated in the inset of Fig. 9e) showed cracking with small particles attached to Mg fusion zone side, which was different from the morphology observed in Fig. 9a). As shown in Fig. 9f), more tearing of Ti substrate was noticed at the Ti side, indicating higher resistance to 11

ACCEPTED MANUSCRIPT crack propagation compared to interfacial failure. According to EDS result of rectangular area, these particles contained 23.2 at.% Al, 57.7 at.% Ti , which was also

T

originated from the Ti3Al phase formed at the interface observed in Fig. 4. In

IP

comparison, smooth fracture surface was observed when using AZ31 filler without

SC R

any particles [22]. 3.4 Bonding mechanism of interfacial reaction layer

Based on the analyses above, the bonding mechanism of Mg to Ti by laser

NU

welding-brazing process was clarified and schematic diagram is shown in Fig. 10.

MA

Firstly, laser beam directly irradiated the filler wire and the Ti substrate. Under high temperature, the filler wire melted and the Ti substrate was in dissolution mode. The

D

Ti atom was thus activated and dissolved continuously into the molten filler adjacent

TE

to the liquid/solid interface, as shown in Fig. 10a). Since Mg-Ti system was immiscible, they did not react with each other. Next, the Al atoms tended to diffuse

CE P

from Mg based filler to the liquid/solid interface and was mixed with Ti atoms diffusing from Ti substrate and segregated at the front of liquid/solid interface caused

AC

by the effect of chemical potential, indicated in Fig. 10c). The chemical potential of Mg-Ti system reduced because of the involvement of Al atoms, which was reported in previous study [22]. The decrease of Al chemical potential near the interface could promote the diffusion of Al from a little far away from the liquid to the liquid/solid interface. Meanwhile, the increase of Ti molar fraction adjacent to the interface induced by the dissolution of Ti into the liquid filler would further decrease the Al chemical potential, which in turn accelerated the diffusion of Al into the liquid/solid interface [25]. When the temperature decreased to about 1100oC, the Al atom in the liquid at the front of solid/liquid interface was saturated inducing the precipitation of Ti3Al phase, as shown in Fig. 10d). As the temperature further decreased to 650 oC or 12

ACCEPTED MANUSCRIPT below, the α-Mg began to precipitate from the AZ91 liquid. When temperature decreased to 325 oC, eutectic reaction occurred to the remaining liquid with the newly

T

formed structure (α-Mg+Mg17Al12) indicated in Fig. 10e). As a result, the interfacial

SC R

bonding, which enhanced joint strength as expected.

IP

reaction of Mg and Ti with AZ91 Mg based filler was produced causing metallurgical

4. Conclusions

Laser welding-brazing of Mg to Ti with AZ91 filler was carried out. The weld

NU

appearances and cross sections of laser welded-brazed Mg/Ti joints were observed.

MA

The interfacial microstructures were characterized and identified. The mechanical properties of Mg/Ti joint were evaluated. The bonding mechanism of Mg to Ti was

D

clarified based on the above results. The following conclusions were drawn:

TE

(1) Mg alloys and Ti alloys were successfully joined by laser welding-brazing process using AZ91 Mg based filler wire. Continous and uniform joints were obtained

CE P

in a relatively large processing window. (2) An ultra-thin reaction layer was observed along the fusion zone/Ti interface,

AC

indicating the occurrence of metallurgical bonding of Mg and Ti. Its thickness was varied slowly with the function of heat input, with the maximum value of 1.5 μm. The newly formed reaction layer was thereafter identified as Ti3Al phase by TEM analysis. (3) The maximum value of tensile-shear strength reached 2057 N, representing 50% joint efficiency with respect to Mg base metal. Two main fracture modes were found from all the joints: interfacial failure and fusion zone fracture. (4) Newly formed reaction products were attached to the Ti substrate after testing, which was mainly responsible for high resistance to the crack propagation and thereafter the improvement of mechanical properties. Acknowledgments 13

ACCEPTED MANUSCRIPT The research was financially supported by National Natural Science Foundation of China (Grant No.51504074), Shandong Provincial Young and Middle Aged Scientists

T

Research Awards Fund (Grant No. BS2015ZZ008), the Fundamental Research Funds

SC R

and Technology Major Project (No.2014ZX04001131).

IP

for the Central Universities (Grant No. HIT.NSRIF.2016094), and National Science

References

[1] L. Liu, D. Ren, F. Liu, A Review of Dissimilar Welding Techniques for Magnesium Alloys to

NU

Aluminum Alloys. Mater. 7 (2014) 3735-3757.

[2] Z. Sun, J. Ion, Laser welding of dissimilar metal combinations, J. Mater. Sci. 30 (1995) 4205-4214.

MA

[3] M. Aonuma, K. Nakata, Effect of alloying elements on interface microstructure of Mg–Al–Zn magnesium alloys and titanium joint by friction stir welding, Mater. Sci. Eng. B 161 (2009) 46-49. [4] M. Aonuma, K. Nakata, Effect of calcium on intermetallic compound layer at interface of calcium

D

added magnesium–aluminum alloy and titanium joint by friction stir welding. Mater. Sci. Eng. B 173

TE

(2010) 135-138.

CE P

[5] A. M. Atieh, T. I. Khan, Effect of process parameters on semi-solid TLP bonding of Ti–6Al–4V to Mg–AZ31, J. Mater. Sci. 48 (2013) 6737-6745. [6] A. M. Atieh, T. I. Khan, Application of Ni and Cu nanoparticles in transient liquid phase (TLP)

AC

bonding of Ti-6Al-4V and Mg-AZ31 alloys, J. Mater. Sci. 49 (2014) 7648-7658. [7] A. M. Atieh, T. I. Khan, Effect of Interlayer Thickness on Joint Formation Between Ti-6Al-4V and Mg-AZ31 Alloys, J. Mater. Eng. Perform. 23 (2014) 4042-4054. [8] A. M. Atieh, T. I. Khan, TLP bonding of Ti-6Al-4V and Mg-AZ31 alloys using pure Ni electro-deposited coats. J. Mater. Process. Technol. 214 (2014) 3158-3168. [9] R. Cao, T. Wang, C. Wang, Z. Feng, Q. Lin, J. H. Chen, Cold metal transfer welding–brazing of pure titanium TA2 to magnesium alloy AZ31B, J. Alloy. Compd. 605 (2014) 12-20. [10] C. Xu, G. Sheng, H. Wang, X. Yuan, Reinforcement of Mg/Ti joints using ultrasonic assisted tungsten inert gas welding–brazing technology, Sci. Technol. Weld. Joi. 19 (2014) 703-707. [11] C. Xu, G. M. Sheng, Y. Q. Deng, X. J. Yuan, K. L. Tang, Microstructure and mechanical properties of tungsten inert gas welded–brazed Mg/Ti lap joints, Sci. Technol. Weld. Joi. 19 (2014) 443-450. [12] M. Gao, Z. M. Wang, X. Y. Li, X. Y. Zeng, Laser Keyhole Welding of Dissimilar Ti-6Al-4V 14

ACCEPTED MANUSCRIPT Titanium Alloy to AZ31B Magnesium Alloy, Metall. Mater. Trans. A 43 (2011) 163-172. [13] M. Gao, Z. M. Wang, J. Yan, X. Y. Zeng, Dissimilar Ti/Mg alloy butt welding by fibre laser with Mg filler wire – preliminary study, Sci. Technol. Weld. Joi. 16 (2011) 488-496.

T

[14] C. Dharmendra, K.P. Rao, J. Wilden, S. Reich, Study on laser welding-brazing of zinc coated steel

IP

to aluminum alloy with a zinc based filler, Mater. Sci. Eng. A 3 (2011) 1497-1503.

SC R

[15] Y. B. Chen, S. H. Chen, L. Q. Li, Effects of heat input on microstructure and mechanical property of Al/Ti joints by rectangular spot laser welding-brazing method, Int. J. Adv. Manuf. Tech. 3-4 (2009) 265-272.

NU

[16] S. Chen, L. Li, Y. Chen, J. Huang. Joining mechanism of Ti/Al dissimilar alloys during laser welding-brazing process, J. Alloy. Compd. 509 (2011) 891-898.

MA

[17] L. Li, C. Tan, Y. Chen, W. Guo, X. Hu, Influence of Zn Coating on Interfacial Reactions and Mechanical Properties During Laser Welding-Brazing of Mg to Steel, Mater. Des. 43 (2012) 4740-4754.

D

[18] L. Li, C. Tan, Y. Chen, W. Guo, F. Song, Comparative study on microstructure and mechanical

TE

properties of laser welded–brazed Mg/mild steel and Mg/stainless steel joints, Mater. Des. 43 (2013) 59-65.

CE P

[19] H. Dong, W. Hu, Y. Duan, X. Wang, C. Dong, Dissimilar metal joining of aluminum alloy to galvanized steel with Al–Si, Al–Cu, Al–Si–Cu and Zn–Al filler wires. J. Mater. Process. Technol. 212

AC

(2012) 458-464.

[20] J. L. Song, S. B. Lin, C. L. Yang, C. L. Fan. Effects of Si additions on intermetallic compound layer of aluminum–steel TIG welding–brazing joint, J. Alloy. Compd. 488 (2009) 217-222. [21] GB/T 2650-2008/ISO 6-9016: 2001. Tensile test method on welded joints, Standardization Administration of the People’s Republic of China; 2008. [22] C. Tan, X. Song, B. Chen, L. Li, J. Feng. Enhanced interfacial reaction and mechanical properties of laser welded-brazed Mg/Ti joints with Al element from filler, Mater. Lett. 167 (2016) 38-42. [23] C. W. Tan, L. Q. Li, Y. B. Chen, A. M. Nasiri, Y. Zhou, Microstructural characteristics and mechanical properties of fiber laser welded-brazed Mg alloy stainless steel joint, Weld. J. 93 (2014) 399s-409s. [24] D. M. Fronczek, J. Wojewoda-Budka, R. Chulist, A. Sypien, A. Korneva, Z. Szulc, N. Schell, P. Zieba, Structural properties of Ti/Al clads manufactured by explosive welding and annealing. Mater. 15

ACCEPTED MANUSCRIPT Des. 91 (2016) 80-89. [25] S. H. Chen, L. Q. Li, Y. B. Chen, D. J. Liu, Si diffusion behavior during laser welding-brazing of

AC

CE P

TE

D

MA

NU

SC R

IP

T

Al alloy and Ti alloy with Al-12Si filler wire. T. Nonferr. Metal. Soc. 20 (2010) 64-70.

16

ACCEPTED MANUSCRIPT

Figure captions

T

Fig. 1 Schematic of laser welding-brazing process and specimen for tensile-shear

IP

testing : (a) welding-brazing process and (b) testing specimen.

MT-3, (d) MT-4, (e) MT-6 and (f) MT-7.

SC R

Fig. 2 Weld appearances of laser welded-brazed Mg/Ti joints: (a) MT-1, (b) MT-2, (c)

Fig. 3 Cross sections of laser welded-brazed Mg/Ti joints: (a) MT-1, (b) MT-2, (c)

NU

MT-3 and (d) MT-7.

MA

Fig. 4 Interfacial microstructure of brazed FZ/Ti side: (a) the same parameter as MT-1 with AZ31 filler, (b) MT-1, (c) MT-2, (d) MT-3, (e) MT-4, (f) MT-5, (g) MT-6

D

and (h) MT-7.

TE

Fig. 5 Line scanning results of brazed interface for Mg/Ti joint: (a) the same parameter as MT-1 with AZ31 filler, (b) MT-2 and (c) MT-3.

CE P

Fig. 6 TEM investigation of the Mg/Ti interface: (a) bright field image taken from the interface, (b) higher magnification of (a), (c) HAADF image of (a), and (d)

AC

SAED of the interfacial reaction layer. Fig. 7 Hardness distribution across the brazed Mg/Ti interface of joint MT-3. Fig. 8 Tensile-shear strength of laser welded-brazed Mg/Ti joints. Fig. 9 Fracture surface morphologies of the Mg/Ti joints at different fracture modes: (a) fracture location with interfacial failure, (b) FZ side of (a), (c) Ti side of (a), (d) fracture location with fusion zone fracture, (e) FZ side of (d) and (f) Ti side of (d). Fig. 10 Bonding mechanism of Mg/Ti joint: (a), (b) melting of filler and flux, c) dissolution and diffusion of Al atoms and Ti atoms at the interface, (d), (e) solidification of interfacial zone at different temperature ranges. 17

ACCEPTED MANUSCRIPT

Table captions

T

Table 1 Chemical compositions of AZ31B Mg base metal and AZ91 Mg based filler.

SC R

Table 2 Chemical compositions of Ti-6Al-4V. (wt.%)

IP

(wt.%)

AC

CE P

TE

D

MA

NU

Table 3 Experimental parameters used in the process.

18

TE

D

MA

NU

SC R

IP

T

ACCEPTED MANUSCRIPT

CE P

Fig. 1 Schematic of laser welding-brazing process and specimen for tensile-shear

AC

testing : (a) welding-brazing process and (b) testing specimen.

19

MA

NU

SC R

IP

T

ACCEPTED MANUSCRIPT

D

Fig. 2 Weld appearances of laser welded-brazed Mg/Ti joints: (a) MT-1, (b) MT-2, (c)

AC

CE P

TE

MT-3, (d) MT-4, (e) MT-6 and (f) MT-7.

Fig. 3 Cross sections of laser welded-brazed Mg/Ti joints: (a) MT-1, (b) MT-2, (c) MT-3 and (d) MT-7.

20

CE P

TE

D

MA

NU

SC R

IP

T

ACCEPTED MANUSCRIPT

Fig. 4 Interfacial microstructure of brazed FZ/Ti side: (a) the same parameter as MT-1

MT-7.

AC

with AZ31 filler, (b) MT-1, (c) MT-2, (d) MT-3, (e) MT-4, (f) MT-5, (g) MT-6 and (h)

21

AC

CE P

TE

D

MA

NU

SC R

IP

T

ACCEPTED MANUSCRIPT

Fig. 5 Line scanning results of brazed interface for Mg/Ti joint: (a) the same parameter as MT-1 with AZ31 filler, (b) MT-2 and (c) MT-3.

22

CE P

TE

D

MA

NU

SC R

IP

T

ACCEPTED MANUSCRIPT

AC

Fig. 6 TEM investigation of the Mg/Ti interface: (a) bright field image taken from the interface, (b) higher magnification of (a), (c) HAADF image of (a), and (d) SAED of the interfacial reaction layer.

23

MA

NU

SC R

IP

T

ACCEPTED MANUSCRIPT

AC

CE P

TE

D

Fig. 7 Hardness distribution across the brazed Mg/Ti interface of joint MT-3.

Fig. 8 Tensile-shear strength of laser welded-brazed Mg/Ti joints.

24

AC

CE P

TE

D

MA

NU

SC R

IP

T

ACCEPTED MANUSCRIPT

Fig. 9 Fracture surface morphologies of the Mg/Ti joints at different fracture modes: (a) fracture location with interfacial failure, (b) FZ side of (a), (c) Ti side of (a), (d) fracture location with fusion zone fracture, (e) FZ side of (d) and (f) Ti side of (d).

25

MA

NU

SC R

IP

T

ACCEPTED MANUSCRIPT

D

Fig. 10 Bonding mechanism of Mg/Ti joint: (a), (b) melting of filler and flux, c)

TE

dissolution and diffusion of Al atoms and Ti atoms at the interface, (d), (e)

AC

CE P

solidification of interfacial zone at different temperature ranges.

26

ACCEPTED MANUSCRIPT

Table 1 Chemical compositions of AZ31B Mg base metal and AZ91 Mg based filler.

Zn

Mn

AZ31B

3

1

0.3

AZ91 filler

9

1

0.24

Si

Mg

0.1

Bal.

0.007

Bal.

IP

Al

NU

SC R

Elements

T

(wt.%)

V

Fe

6

4

0.3

C

N

H

Ti

0.1

0.05

0.015

Bal.

AC

CE P

TE

D

Al

MA

Table 2 Chemical compositions of Ti-6Al-4V. (wt.%)

27

ACCEPTED MANUSCRIPT

Table 3 Experimental parameters used in the process.

T

Welding parameter

IP

Wire feed

Joint

Laser power

Welding speed

(P, kW)

(V, m/min)

3.5

MT-2

2000

0.5

3.5

MT-3

2400

0.5

3.5

MT-4

2400

0.8

4.5

MT-5

2400

1

5

MT-6

1

5

3200

1

5.5

TE

2800

AC

CE P

MT-7

NU

0.5

MA

1600

(Vw, m/min)

D

MT-1

SC R

speed

28

ACCEPTED MANUSCRIPT

AC

CE P

TE

D

MA

NU

SC R

IP

T

Graphical abstract

29

ACCEPTED MANUSCRIPT Highlights Laser welding-brazing of Mg to Ti with Mg based filler was studied.



Metallurgical bonding was achieved at the Mg/Ti interface.



Al diffusing from filler reacted with Ti resulting in interfacial reaction layer.



Newly formed Ti3Al phase improved the interfacial bonding and joint strength.



The joining mechanism of Mg to Ti with AZ91 filler was clarified.

AC

CE P

TE

D

MA

NU

SC R

IP

T



30