Dissimilar Non-Ferrous Metal Welding: An Insight on Experimental and Numerical Analysis

Abstract

In recent years Gas Metal Arc Welding (GMAW) technology has expanded its functionalities in various areas which have further motivated its usage in several emerging manufacturing industries. There are several issues and challenges associated with this technology, especially in dissimilar metal welding (DMW). One of the predominant challenges is selecting appropriate welding parameters which influence the efficiency of this technology. To explore several modern advancements in this expertise, this paper has done an exclusive survey on various standards of GMAW and its variants for selecting suitable parameters for welding dissimilar nonferrous metals. This review summarizes various experimental and numerical results along with related illustrations to highlight the feasibility of welding dissimilar nonferrous metals using traditional GMAW and investigations on advanced GMAW processes such as cold metal transfer (CMT) and pulsed GMAW (P-GMAW). Simulation and modeling of nonferrous DMW have identified several research gaps and modeling problems. Researchers and manufacturers can use this review as a guideline to choose appropriate welding parameters to implement GMAW and its variants for non-ferrous dissimilar welding. It found that by controlling the heat input and effective post-heat treatments, adequate joint properties can be achieved. Automated large -scale manufacturing will widen the utilization scope of GMAW and avoid some costly methods such as laser welding, ultrasonic welding, and friction stir welding etc.

1. Introduction

In the early 1990s, electronic power control, especially the use of specific electric power converter sources, allowed better efficient control of output signal, contributing to the emergence of several unique metal transfer control methods. Such advancements further led to substantial consumer advantages and a wider variety of GMAW technologies. GMAW is a widely used process for welding both similar and dissimilar metals. Due to its versatility, speed, adaptability to robotic automation, efficiency, and economy, GMAW is preferred over other standard joining methods such as bolting, riveting, and mechanical interlocking [1]. Brazing stands in a queue followed by the welding process for DMW especially for non-ferrous metals. Though it holds an advantage over welding process in case of minimum distortion, the major fallback is the strength of the joint made through brazing will not be adequate for some industrial application [2]. Welded joints are generally stronger than bolted joints, due to the absence of perforations that reduces the load-carrying capacity of a joint. The application of GMAW are manufacturing units of the automobile industry, aerospace industry, and various manufacturing domains [3,4,5]. The feasibility of a product to be utilized in any practical welding application is connoted by its weldability character.

Generally, the process of preventing metal cracking during welding is termed as weldability [6]. Due to differences in mechanical, microstructural, and thermal characteristics, weldability varies by material. These material properties may cause cracking in the weld joint region, and also contribute to the creation of high-hardness weld zones due to a fast rate of solidification [7], and may release of environmental gases such as Argon, nitrogen, carbon dioxide, carbon monoxide [8]. Therefore, a material with high weldability quality should be selected for welding to avoid different types of cracking, thereby avoiding repair or revamping. During the welding process, the heating phase considerably affects welded joints and their heat-affected zones (HAZ) by modifying their chemical configurations, micro and macro-structural characteristics, and physical properties. To avoid significant contrasts in the base metal, an appropriate welding technique should be chosen. Among the fusion joining processes, the most common process is electric arc welding, which is widely used in the welding of metals such as steel [9,10], aluminum [11], titanium [12], and magnesium alloys [13]. Unfortunately, welding these metals together in a dissimilar manner is highly complicated and challenging.

The dissimilar metals have typically lower weldability than similar metals because of improper joint design, formation of different intermetallic compounds (IMC), differences in metallic compositions, mechanical and thermal properties. Corrosion issues, such as galvanic corrosion, oxidation, and hydrogen-induced fracture, are some of the key factors that reduce DMW efficiency [14]. To achieve high-quality dissimilar welds, a proper filler material is vital [15,16,17].

Dissimilar metals can be joined by conventional welding methods such as shielded metal arc welding, GMAW, submerged arc welding, flux-cored arc welding, electron beam welding (EBW) [18], ultrasonic welding, friction stir welding [19], laser beam welding [20], and gas tungsten arc welding (GTAW). Although welding methods such as EBW and laser beam welding are suitable for joining dissimilar materials, they are inadmissible for small production industries because of their high cost and maintenance. Also, conventional GMAW causes considerable dimensional deformation and has difficulties in customizing and controlling heat input. Therefore, this study focuses on analyzing various aspects of advanced GMAW. In DMW, an advanced GMAW method provides a massive benefit. It is apparent that the different GMAW variants have a major impact on the characteristics of the joint. It also controls the width of the HAZ, the content of intermetallic layer, the microstructure, and therefore the technical effectiveness of the joint. Furthermore, when combined with an appropriate metal, improved controllability of advanced GMAW can improve ductility of IMC [21]. Despite the challenges, practical fusion welding methods for DMW have been ceaselessly explored [22]. Yet, the actual experimental welding process and physical testing will be extensive, laborious as well as horrendously overpriced. Because of this problem, the welding method necessitates the employment of advanced computational and modeling tools.

Several welding methods can be modeled using simulation study to determine various system performances. For a better understanding of the extensive thermo-mechanical processes related to GMAW, many simpler numerical models were created. Mathematical modeling of GMAW offers to work on a wide range of applications. The detailed analysis of either the GMAW process or the characteristics of the welded joints can produce results that have advanced significantly. Moreover, no comprehensive review specifically in arc welding numerical analysis has ever been reported so far. This paper examines recent advances in numerical techniques for welding non-ferrous DMW by arc welding, covering the welding processes and the microstructure as well as properties of the resultant joints. Material flow modeling, meshing, heat source modeling, and the latest software selection method are all presented in detail. Despite the fact that welding is widely simulated, the analyses do not reflect actual weld application standards. New-era materials and a deeper knowledge of the process are still required. Nevertheless, using improved numerical simulation analysis modules, we can determine the result of a process at the beginning of its working phase.

This literature review attempts to convey the magnitude of the problems associated with dissimilar joints. It also discusses the successful DMW of non-ferrous metals using conventional GMAW and advanced GMAW-adapted methods such as globular, pulsed arc welding, spray type, and CMT. This review further investigates the mechanical and microstructural properties of weld zones and provides detailed analyses of fusion zone porosities and the evolution of intermetallic compounds at interfacial regions. It concludes with opportunities for future research.

The novelty of the current review is the cumulative survey on experimental and numerical analyses in the DMW of non-ferrous metal welding using GMAW has not been reported extensively. The contents herein provide a resource using which young researchers can structure their studies and research in the welding field. Welding research is divided into specialized areas, and global researchers have created models and simulations for a better understanding of the welding processes. Further, this review answers the following questions in detail: (1) which factors affect the weldability of dissimilar non-ferrous metals using GMAW? (2) Which types of software are available for modeling GMAW processes and DMW? (3) Which variants of GMAW are suitable for the effective welding of non-ferrous metals?

The research is organized into four stages. The challenges in welding non-ferrous DMW are recognized following a comprehensive literature review. The GMAW method is selected, and their variations can be utilized to overcome the challenges. Since the goal of most numerical simulation studies is to develop finite element models (FEMs) to solve real-world problems, adopting the latest software to carry out numerical analysis is a crucial part of every discipline. In the second phase, detailed availability and comparison of different simulation software, used specifically for simulating welding process, are discussed (Figure 1).

In the other two consecutive phases, a deep dive insight on the experimental analysis for DMW is discussed. The process parameters and the filler wire selected in the respective researches along with their results and microstructural study are detailed.

Architecture of the Review

The earlier GMAW methods lacked accurate control of the heat input, which is necessary for successful welding of different materials, whereas advanced GMAW technology development in recent times has opened up new opportunities. So, the numerical and experimental feasibility of nonferrous welding by GMAW and its variants such as CMT, P-GMAW, and other transfer modes are collated from the various scientific literature. Figure 2 shows the architectural layout of the process layout carried out in the current survey. The layout infers the flow of review depicting the selection of welding process followed by the FEM simulation and experimental feasibility.

2. Finite Element Analysis (FEA)

Computational models that replicate the welding process and optimize the operating conditions have been extensively utilized over the past two decades. FEA is an analysis of the thermal and mechanical behavior of FEM that is usually modeled in design software and assigned with specific material property (Figure 3). FEM in fusion welding, modeling the metal droplets and formation during the joining process, has improved weld quality and process stability [23,24].

Normally, the current, speed, wire-feed rate, gas flow rate, and voltage inputs of GMAW are the parameters that most significantly affect the geometry, quality, and cost of the weld. The effect of welding current on the penetration using GMAW was examined [25]. The penetration depth was more severely affected by current increments than voltage increments.

Presently, characteristics of welds are evaluated using fatigue behavior forecasting approaches such as the equivalent structural stress technique, structural stress (σ_HS) method, notch stress method, and local strain method. However, predicting weld defects using FEM is more important in industrial sectors. To obtain the correct output, FEM requires a properly selected mesh, model, and heat source.

2.1. Mesh

The mesh is an essential design factor and must be carefully selected and designed to achieve high model performance. One of the most important considerations to achieve simulation accuracy is the creation of a high-quality meshing. In general, increasing the number of elements in the mesh (i.e., refining the mesh) increases the reliability of the simulation result. When the outputs of the mesh refinement converge, further refining will no longer improve the results. Additionally, poor geometry meshing provides inaccurate simulated results, but it may also lead the solver to generate an error owing to instability. Some of the additional features to be taken into consideration are using a well-defined, simple, and clear geometry, mesh refinement in weld joints, and boundary layers.

The computational meshes are made up of several sub-meshes, which are created through linearly extending unit space. By carefully selecting the necessary variables of the mesh element, adjustable unit spacing can be accomplished. In plastic deformation simulations, hexahedral elements are superior to linear and quadratic tetrahedral elements [26,27].

2.2. Models

Complex models such as metallurgical and thermo-mechanical models [28,29] consider many factors and yield high-precision results, but are slow. In contrast, simple models are imprecise but significantly quicker. Friedman expressed the heat flux at a heating spot as a Gaussian distributed heat source, a popular choice among second-generation models [30].

Whenever the moving heat source is employed in computations, the results are calculated using the “transient” approach, also known as the “step-by-step” method. The time taken to perform analysis can be default or modified by the operator, based on the mesh size and refinement of the design. Semi-ellipsoidal and Double–ellipsoidal heat source models were proposed by Goldak et al., 1984 [31] as well as rectangular heat sources [32] which are used in most current FEM models [33] (Figure 4). The type of FEM selected determines the accuracy and time consumption of the analysis (Figure 5).

2.2.1. Isotropic Hardening

Isotropic (IS) hardening increases the yield surface under sustained plastic strain. Preliminary isotropy introduces and maintains a constant fluid density, achieving IS hardening and satisfying the normality principle [34]. Equation (1) determines the plastic flow in a FEM analysis.

f = σ_eq − k − σ_y = 0

(1)

where, f = plastic flow, σ_eq is the Von Mises equivalent stress, σ_y is the initial yield stress, and k is the degree of strain hardening. The types of hardening models are listed in

Table 1. Types of hardening Models for arc welding FEM.

Hardening Model	Nomenclature	Usage
Isotropic	IS	Calculate plastic strain at reduced computational cost
Kinematic	Km	Obtain the mechanical behavior during heating and cooling cycles but is costly
Mixed isotropic & kinematic	Mix	Perform expansion and translation of the yield surface

.

2.2.2. Kinematic Hardening

Kinematic hardening includes a quantity called back stress. In this case, the yield surface converts as the plastic strain accumulates. The yield surface in this model is defined by Equation (2).

f = f (σ − X) − σ_y

(2)

where X is the back stress tensor, whose rate of change, equivalent plastic strain rate, is defined as Xʹ = C ε^Pl·1/σ_y (σ − X), where 0 < ε^Pl < ε0, and C is a material parameter; ε^Pl—plastic strain tensor.

The heating–cooling cycles in the welding process affect the mechanical behavior of the material, but these effects are represented only in the kinematics hardening model. Therefore, the application of less complex models such as elastic–plastic deformation without hardening or IS hardening is justified by the reduced computational expense [35].

3. Simulation Software

Numerical simulations improve the efficiency and effectiveness of welding by avoiding the need for rectification in the welding workshop, which requires expensive materials and a long working time. To reduce the lead time and cost of numerous prototypes and the design of complex structures, researchers are developing many computation welding mechanics tools for modeling joining processes. Recently updated computational welding mechanics (CWM) principles can well predict the residual stresses and distortion in welded materials [36,37].

The analysis method depends on the study objective, either thermal or structural and the welding process. When simulating welding and surface treatments, we must compute the temperature, phase proportions, and stresses. This can be achieved by solving the problem in a moving frame associated with the heat source. For example, when investigating the microstructural changes in a material that is susceptible to the molten state, melt pool analysis is necessary.

However, before selecting FEM software for numerical simulations, researchers should have some knowledge of the types of hardening models (Table 1), heat source models [33,38,39,40] (

Table 2. Types of heat source models; cumulative data obtained from [33,38,39,40] form open access journals.

Heat Source Model	Nomenclature	Simulation Time	Accuracy
Uniform Surface Heat Flux	USHF	≈15–35 h	Moderate
Goldak’s Heat Source	GHS	≈10–36 h	High
Gaussian Surface Heat Flux	GSHF	-	Precise
Uniform Volume Heat Flux	UVHF	-	Moderate
Conical Heat Source Model	CHS	-	Moderate
Rectangular Heat Source	RHS	≈20–35 h	High

), and element types [41,42] (

Table 3. Types of elements meshes in FEM; data obtained from [41,42] form open access journals.

Element Mesh 1,*		Nomenclature	Accuracy 2,*
2D	Triangle	tri	High
	Quadrilateral	quad	Moderate
3D	Hexahedral	hex	High
	Tetrahedral	ted	Moderate
	Tri-Prism	tip	Moderate

* ¹-mesh includes all possible variants such as HEX8, HEX20, TET4, and TET10. * ²-accuracy depends on the number of Gauss points.

) available in the software. As previously mentioned, plastic deformation should be simulated on hexahedral elements rather than on linear or quadratic tetrahedral elements [27]. Though the research application is different in biomechanics and orthopedics, Ruggerio et al. [41] clearly identified and compared different types of meshes. The currently available and most commonly used software packages for welding simulations by industries and researchers were listed and described as follows (

Table 4. Descriptions about simulation software (Nomenclature is described in Table 1, Table 2 and Table 3).

SI.No	Software	Hardening Model	Heat Source Model	Element Type	Simulation Time *	Approach Method
1	ANSYS®	Is, Km, and Mix	All	All	Moderate	Subroutine programming and adaptive meshing
2	SIMUFACT®	Is and Km	No RHS	No tip	Moderate	Adaptive meshing
3	ESI SYSWELD®	Is, Km, and Mix	All	All	Moderate	Adaptive meshing
5	NX NASTRAN®	Is, Km, and Mix	All	All	Moderate	Subroutine programming and adaptive meshing
6	ABAQUS®	Is, Km, and Mix	All	All	Moderate	Subroutine programming and adaptive meshing
7	FLOW-3D WELD®	Is and Km	No RHS	All	Thermal analysis only	Adaptive meshing
8	TRANSWELD®	Is and Km	No CHS and RHS	All	Moderate	Adaptive meshing

* Note: if the number of system variables (elements and nodes) is high, then the processing time considerably increases with an increase in mesh size (possible duration may range from minutes to hours).

):

Welding design can be modeled in ANSYS Fluent^® (version 2020, ANSYS Inc.; Canonsburg, PA, USA), but the modeling of weld design is not straightforward. Residual stresses can be determined using ANSYS APDL or Workbench for thermomechanical analysis using element births and deaths procedure. Design and analysis can be completed on the same platform.

MSC NASTRAN^® (Version 2020, Hexagon Corporate Services Ltd.; Surrey, UK) is advanced high-end Computer-aided design (CAD)/Computer-aided Manufacturing (CAM)/Computer-aided Engineering (CAE) software providing additional structural and thermal solutions that are unavailable in other software packages. Design and analysis can be implemented on the same NX platform. When the model is massively meshed with solid 3-D C-TETRA and/or C-HEXA elements, an iterative solver (element iterative solver) is recommended, but models with two-dimensional shells and one-dimensional elements should always be solved using the default solver named, direct sparse solver [43].

ABAQUS^® (Version 2018, ABAQUS Inc.; Providence, RI, USA) with DFLUX, GAPCON, and FILM user subroutines are available for thermal investigations. DFLUX describes the welding heat input. The heat source moves along the welding direction to represent the motion of the electrode. GAPCON triggers the heat transfer between the accumulated filler and the base metals until the electrode has reached the preferred position. FILM resembles GAPCON in operation but enables film coefficients to model convective atmospheric cooling until the electrode crosses a specified position. The thermal evaluation measures the heat transfer from the thermal load of the traveling electrode. However, ABAQUS cannot mesh a region with hexahedral elements if one or more edges lie along the axis of revolution.

Meanwhile, SYSWELD^®, MSC.MARC Simufact welding^® are FEA software packages dedicated to welding. Simufact Welding has a user-friendly interface and achieves high accuracy through volume elements and implicit time integration. Users can simulate either thermal or thermomechanical processes. MSC SIMUFACT^® (version 8, Hexagon Corporate Services Ltd.; Surrey, UK) enables data exchange between Simufact Forming and Simufact Welding. Therefore, welding can be integrated with preprocessing and post-processing within a common numerical framework. However, only basic shapes can be directly designed in the software; the design of complex shapes is precluded. Simufact welding analysis can be performed in both thermal and thermomechanical modes. There are additional options that include phase transformation of the metals that can be used to relate the properties such as tensile strength and hardness with the results of experimental analyses.

ESI SYSWELD^® (version 2019.0, ESI Group PLM; Aix-en-Provence, France) integrates the effects linked to metallurgical transformations in the thermal, mechanical, and hydrogen diffusion analysis process. Simulation is broken down into a number of successive steps due to the modular aspect of the product (Figure 6). ABAQUS and ANSYS require additional user subroutines for the welding simulation, such as heat source geometry definitions, moving heat source functions, heat treatment examinations, and thermochemical treatments, all of which are readily accessible in SYSWELD. The existing two types of analysis options such as local and global analysis can be used based on the end results expected from the simulation. Local model analyzes temperature field, metallurgy phase, residual stress, and lateral displacement. Global analysis, on the other hand, examines just the dimensional change (displacements) of welded components as well as stress patterns [32,40]. SYSWELD appears to consume less lead time than ABAQUS in process modeling and is more attractive than ABAQUS for complex welding models [44].

FLOW-3D WELD (V12.0, Flow Science Inc.; Santa Fe, NM, USA) simulates a wide range of mass transfer, fluid flow, heat transfer, and welding problems. This software accounts for the laser power, heat flux profile distribution, laser pulse duration, and scan paths. Temperature-dependent material properties can be assigned to both the welded materials, which is important for tracking the evolution and mixing of the two materials in the melt pool. FLOW-3D WELD simulations provide valuable insights into flow process parameters. After solidification, the deformation and residual stress evolution during the weld cooling can be determined by transferring the three-dimensional flow data obtained by FLOW-3D WELD into FEA software such as ABAQUS.

TRANSWELD^® (Transvalor Software Company; Sophia Antipolis, Biot, France) offers an industrial solution for predicting the thermomechanical behavior of an assembly during a welding process. TRANSWELD^® is available for arc and laser welding and allows microstructural studies of the solid-state assembly. Welding processes can be numerically observed under real conditions in this software. For example, users can visualize the heat source movement during the running time of the simulation.

SORPAS^® (version 10, SWANTEC Software and Engineering Company; Lyngby, Denmark) is professional FEM/FEA software dedicated to welding simulations and optimization of resistance welding and mechanical joining processes. It has been developed for engineers by engineers aiming to solve industrial welding problems.

During the evaluation of residual stresses, for establishing a stress relief strategy, or evaluating distortion of welded component, for controlling dimensional changes, the weld pool geometry as well as phase transformation can be neglected. And the temperature distribution with subsequent elastic-plastic analysis should be properly calculated without any error. This analysis becomes unmanageable in complex structures and has been attempted only in approximate mixed methods (an intuitive simple analysis in the weld bead and FEA in the remaining structure), with varying degrees of success.

A study to understand the selection of simulation parameters such as boundary conditions, heat input, material modeling, mesh size, etc. was proposed by Caprace J.D. et al., focused on the impact of design work and FEM codes on weld simulations [38]. They numerically analyzed the residual stress, heat inputs, boundary condition, and material models of T-joint metals used in ships. A comparison between the numerical and experimental results confirmed that the design procedure can significantly affect the welding residual stresses but has little effect on welding displacement evaluation. The residual stress estimation was mainly influenced by minor alterations of the mechanical properties coupled with simple variations in identical radiant heat. Simulation matrices and their meshing parameters are stated in

Table 5. Simulation matrices and their meshing parameters, data obtained from [38] with permission from Elsevier.

Simulation ID	Software	Hardening Model	EHS Model	Element Type	Node Number	Elements Activation
S1	ANSYS	IS	USHF	8N-Parallelepiped	33,337	Yes
S2	ANSYS	Kinematic	USHF	8N-Parallelepiped	33,337	Yes
S3	SYSWELD	IS	Goldak’s	8N-Parallelepiped	194,568	Yes
S4	SYSWELD	IS	Goldak’s	8N-Parallelepiped	194,568	Yes
S5	Virfac	IS	Goldak’s	4N-Tetrahedral	127,155	Yes
S6	ABAQUS	IS	Goldak’s	8N-Parallelepiped	82,614	Yes

.

FEM and welding simulation of a dissimilar joint from Stainless steel SS304 and copper using laser beam welding was studied by modeling in Ansys software using Gaussian heat flux distribution [45]. The welding load on the formed joint was modeled as a static heat load, and the deformation and residual stress distributions were studied using the temperature-dependent thermal and structural properties. A comparison of the thermal distribution, distortion, and residual stress profiles of the SS, Cu, and SS–Cu welded joints revealed that the SS–Cu joint exhibited a temperature of 1307 °C in the fusion zone, four times that of the base metals); distortion of 4.4 × 10⁻⁵ m, double that of the jointed copper plates); and residual stress of 74MPa, which was less than the yield stress.3.1. Study on nonferrous welding simulations

A study by Kulkarni, et al., modeled and evaluated CMT welding of lap joint formed from two dissimilar metals: aluminum alloy (AA 6061 T6) and zinc-coated steel. The tensile strength of the lap joint was measured using a universal testing machine (UTM). The FEA analysis results were correlated with the experimental results when iterating with the yield of the weld material as 130 MPa, while the weld process simulation gave a bond length of 4.8 mm. By iterating parameters such as geometry (profile) of the weld zone, yield of the weld material and bond length of the weld, it was found that the yield of the material and bond length has a high impact on the test result [46].

In all numerical models, the element type used in the analysis is very important for achieving results that are comparable to the experiment. The effects of GTAW on the deformation properties of dissimilar welded structure from carbon steel CK4 and stainless steel AISI409 (Figure 7) were studied by Farajpour and Ranjbarodeh. They also modeled the deformation using three-dimensional solid and shell elements in ANSYS 11.0 FEA software (Ansys, Inc., Canonsburg, PA, USA). The results of the inherent deformation method utilizing 4-node shell elements are compared with those obtained by a traditional model using 9-node strong components. Although the former method minimized the necessary storage (and hence the time and expense of the analysis), comparisons with experimental observations confirmed that both methods can be used to model DMW with similar efficacy [47].

A study investigated the residual stress in the dissimilar joint from AISI 304 and Monel 400. The predicted residual stress distribution and thermal field (Figure 8) in a FEM model developed was performed in ANSYS. Solid 90 elements and a sequentially coupled thermomechanical transient method were used. They reported that the finite element three-dimensional heat source model effectively estimates the precision residual stresses in the weld zone. The predicted findings correlated with the experimental results, with only 5% defects [48]. Mostly, FEA analysis of welded components is typically performed to investigate the distortion and deformation of the component during the welding process.

Welding is the heating of two or more materials to fuse them together. The uneven heating and cooling phase during the welding process causes expansion and contraction, resulting in distortion. Residual stress induced in the component is the primary cause of distortion. The deformation increases as the total heat input increases. This is frequently caused by differences in electrode/filler size and power output. Residual stresses and welding temperature of a dissimilar butt joint welded by GTAW method were studied by Zhao and his research team. In particular, the researchers investigated the impact of heat input, layer number, and groove form on the residual stress distribution in a pipe joint welded with dissimilar steels (T92 and S30432). Wide gradients were found in the HAZ of the T92 steel side. Simulations were performed in the general-purpose finite element software code of ABAQUS (Figure 9). Further analysis revealed hoop and axial stresses in the weld zone. Meanwhile, the residual stress on the S30432 pipe was minimized because the heat input was lower at this side, changing the thermal profile [49].

The residual stress and distribution temperature field along a welded transverse control arm was investigated in a model with linear multinode finite elements [50]. In the welding sequence and clamping, the authors compared the weld distortions in a transverse control arm welded with GMAW in various simplified models (

Table 6. Comparison of the results of SYSWELD and other models [51] (from open access journal).

Criteria	Thermal Shrinkage Model	Pre-Stressed Truss Element Model	Initial Stress and Strain Model	SYSWELD
Number of material parameters	Small	Small	Small	Large
Preprocessing time required	Small	Small	Small	Large
Relative distortion with respect to the experiment (%)	93	103	96	92
Total computation time	41 s	26 s	13 s	38 h

[51]) in ABAQUS 6.5 and SYSWELD.

Selection of type of heat sources various with change in the type of the welding process in the numerical simulation. Finite element code COMSOL Multiphysics was used for an investigation on a titanium-aluminum dissimilar butt joint [52]. They investigated the welding process under two- and three-dimensional Gaussian heat sources and compared the simulation results with the experimentally observed fiber-laser-welded fusion zone of the joint. The size of the fusion zone was better estimated in the two-dimensional model than in the three-dimensional model, but the three-dimensional model well estimated the field of the weld puddle and the cooling rate. The simulations captured the actual thermal field (recorded using temperature sensors) at the time of the joining process. The primary finding of this research was the high simulation accuracy after refining the mesh size near the fusion zone.

Heat transfer, fluid flow, and species distribution in the welding process of dissimilar aluminum alloys have been studied in the numerical finite volume model (FVM). FVM speed and thermal analyses were performed on two aluminum alloys with different Si contents (5% and 10%), using Scheil’s model for the electrode melting process and cavity models for metal addition to the weld zone [53]. The results reported that convection plays a major role in thermal analysis and that the simulated temperature and velocity fields are similar.

4. Nonferrous Dissimilar Welding

The practice of satisfying weld quality requirements is essential for any product; particularly for non-ferrous materials and it should not be compromised. Before welding any metals, welders should prepare the base metal and remove any debris, oxides, and hydrocarbon impurities introduced by leftover contaminants or lubricants from earlier processes such as machining. Removing the oxides is very important, as the melting temperature of the oxide layer on the metal surface highly differs from that of non-oxidized aluminum.

Mostly, oxides are removed using steel bristle hairbrushes or cleansing solvents along with mark-removing agents. To avoid contamination of the scrubbed layer from oxides, an appropriate method of scrubbing the brush should be employed. The brush motions should be gentle and unidirectional, with no transverse or backward scraping. Additionally, traverse contamination of different materials must be prevented. When using etching solvents, the traces of the solvent must be thoroughly extracted from the metal before introducing weld heat. To reduce the risk of hydrocarbon contamination from the coolant liquid or cutting solvents during the welding, the work material should be rinsed with a grease-removing chemical product. As weldments are prone to fracture under various loading conditions, automobile engineers must also consider the fatigue properties of weld metals despite compromising the actual welds.

The formation of brittle IMCs limits the welding of non-ferrous DMW such as aluminum to copper and titanium to aluminum. Moreover, titanium and copper have different physical properties and low solubility. When titanium is fusion-welded to any nickel-based alloy, the IMC (composed of elements such as nickel, iron, silicone, chromium, and manganese) increases the brittleness of the joint. By contrast, copper and nickel (and their alloys) are easily welded as these two metals are compatible and mutually soluble.

4.1. Welding of Titanium to Aluminum

Titanium is ubiquitously employed in the aviation industry. The final components of aviation require expensive subtractive procedures such as processing or turning, which remove almost 90% of the material as fine chips from certain parts [54]. Titanium-based alloys are broadly applied in the metallurgical, papermaking, medical, aerospace, and chemical industries. Such applications require astounding mechanical properties and high-temperature resistance [55].

GMAW is not usually applied to titanium owing to high spatter generation and wandering of the welding arc (with consequent waviness of the weld bead). Meanwhile, the properties of titanium alloys such as low heat transfer, low density, and high interfacial tension at the weld pool, positively affect the fusion welding process [56].

Load tolerance and long weld life are regarded as crucial factors of the mechanical properties of joints formed from DMW. Satisfactory welded joints were obtained while the welding-brazing overlap of titanium on the top of an aluminum sheet having a dimension of 200 mm × 50 mm × 1 mm by CMT using ER4043 as filler wire. The microstructures of the joints were analyzed using an optical microscope and a scanning electron microscope SEM 6700F revealed that several IMC elements such as Ti3Al, TiAl, TiAl2, Ti2Al5, and TiAl3 are formed in the Ti/Al fusion process. In a tensile strength test, two types of fracture modes occurred at the interface at the heat affected zone of Aluminum [57].

Ti-2Al-Mn alloy and aluminum 1060 plate was welded by Wei and his fellow researchers, using an AlSi5 filler wire by P-GMAW indicating a good weld. Under a scanning electron microscope, the joint formed by the controlled welding process was found to be free of pores and cracks. When aluminum and titanium are fusion-welded, IMCs such as TiAl₃ and Ti₂Al render the weld zone exceedingly brittle, thereby decreasing the mechanical properties of the joint. Despite its low density, TiAl₃ exhibits high microhardness and high temperature and oxidation resistance. Wei et al. reported TiAl₃, Ti₇Al₅Si₁₂, and α-Ti in the HAZ [58].

The microhardness values of the transition region, grain region, fusion zone, and HAZ were 280, 220–270, 35–40, and 25–30 HV0.05, respectively. The parent metal Ti–2Al– Mn was especially hard (200–210 HV0.05). Furthermore, the amount of TiAl₃ precipitation depended on the difference in the welding heat input. Defect-free joints were achieved by maintaining the heat input at 1.85–2.10 kJ/cm. Fracture occurred in the HAZ of the fusion zone of the weld formed from these two metals. Microstructures of weld metal near the Ti-Al interface under different heat inputs: (a) 1.82–1.91 kJ/cm, (b) 1.90–1.99 kJ/cm, and (c) 2.05–2.14 kJ/cm has been illustrated in Figure 10.

Titanium Ti6Al-4V and aluminum 5A05Al metals are butt welded by CMT using Al–Mg5 filler wire. As revealed in tensile tests and fractography analysis, fully penetrated crack-free joints were obtained at a high heat input (Figure 11). Ti_3.3Al and TiAl₃ layers were formed in an orderly sequence [59].

4.2. Welding of Titanium to Steel

Selecting the welding parameters is considered a critical task in DMW welding processes. Pardal et.al stated that a high heat input maximizes the tensile strength and microhardness [60]. The research investigated the joint strength by varying the filler offset distance (0.5–1.2 mm) and welding speed (5–9 mm/min) in the GMAW of stainless steel (AISI 316L) and titanium (AMS 4911L; Ti–6Al–4V) using copper (CuSi-3) as the filler wire (Figure 12). At the end of the tensile strength test, most of the fractures appeared on the stainless steel side because of its less strength comparatively. Due to the great bio-compatible properties, specific strength, and superior corrosive tolerance, Ti as well as their alloys are the best choice in biomedical field [61,62].

As titanium and steel dissimilar joints are lightweight, they are suitable for applicable in spacecraft and rockets technology [63]. A macroscopic analysis showed that the welding wire positioning did not influence the weld bead geometry (Figure 13). Samples selected for tensile testing failed at the Fe–Cu and Cu–Ti interface at maximum strength of 200 Mpa. Nonetheless, the adaptability of this joining process is easier and more cost-effective than infrared brazing and EBW in the industry.

Also, Information on joints formed from different metals such as Ti, steel, and different compounds is limited, and the related mechanical properties of these joints are poorly understood. Accordingly, the embrittlement impact of carbides that may form when Ti is joined to Fe, Ni, or Cu has not been elucidated.

4.3. Welding of Titanium to Nickel

When adequate techniques are practiced, Ni-based metals can be successfully welded with no complications. To achieve high-quality welds, the welding design should consider the base and filler metals, joint geometry, preheating and input temperatures, and the post-weld heat treatment specifications. The power sources used in traditional welding processes deliver sufficient performance and controls, but the weld heat should usually be regulated from low to moderate levels. In stringer bead welding methods, the electrode and torch should be modified to avoid wide beads. The welding of titanium and nickel alloys has relevant application to the field of light-weight and high-temperature purposes like aircraft propulsion [64].

Due to the intricacy and amendments in the operating environment of all welded structures, prior concern on weld deposition has to be taken. P-GMAW in the synergistic phase was used to weld titanium with nickel (ErNi-1) electrode and examined the variational effect of the wire feed rate, weld current, voltage, and weld speed of one-way deposition. Microhardness study (Figure 14) and Macroscopic images confirmed strong bonding of the alloyed metal layer to the unalloyed layer, and no fractures or permeability were observed at three wire-feed speeds (2.0, 2.2, and 2.5 m/min) [65].

Meanwhile, microstructure images confirmed a Ni3-Ti phase and another eutectic mixture. The surface hardness in the fusion zone ranged from 650 to 700 HV. These differences caused diverse IMC-based microstructures on the surfaces of joints welded by this method. Ti & Ni IMC layer was generated by the surface alloying of titanium with a pure nickel electrode, thereby improving the hardness of the material.

Dissimilar joints of Ti and Ni have not been reported in the literature, but titanium plates have been welded with nickel filler wire using P-GMAW. Yet welding of a dissimilar joint from titanium and nickel was not accomplished through conventional GMAW. Correspondingly this is inciting on the future research of joining Ti and Ni by utilizing advanced or hybrid welding processes.

4.4. Welding of Titanium to Copper

The weld interface in Ti–Cu metal welding is large because copper is strongly heat-conductive and there is low atomic diffusion between the dissimilar Ti and Cu metals in the welding process [66]. Moreover, the thermal diffusivities differ between the two materials. Copper has a lower melting point than titanium [67], and the mechanical properties widely differ between titanium and copper. In the sectors of aircraft, instruments, electronics, and petrochemical industries, Ti/Cu dissimilar joints have good thermal, electrical, corrosive properties, and abrasion resistance, as well as meeting the criteria of good mechanical properties under less weight circumstances [68].

The presence of an inappropriate IMC layer causes deterioration of the mechanical properties of dissimilar combinations of titanium and copper. According to certain researchers, this problem can be minimized using copper as the filler wire in the CMT method [69]. Two types of lap joints were made, one formed from pure titanium (TA2) and the other from pure copper (T2) at a torch-weld distance of 10 mm, using ERCu-NiAl (AWS A5.7/A5.7M) copper wire with a diameter of 1.2 mm (Figure 15). The first lap joint, comprising a copper sheet placed on a titanium sheet, was welded at 6 mm/s with a wire feed speed of 5–6 m/min; welding voltages of 11.0, 11.6, and 12.6 V; and welding currents of 112, 125, and 136 A. The torch deviated by 1.0 mm. The second joint, comprising a titanium plate overlapping a copper plate, was welded at 6 mm/s with a wire feed rate of 6–7.5 m/min, and the torch did not deviate. The weld geometry and weld stability of the process were investigated in a microstructural and tensile strength analysis and were desirably high.

A clear fusion line was formed on the weld interface of the Cu-weld bead (Figure 16-Region A). This indicates a prominent possibility of welding these two dissimilar metals with the improved microstructural properties.

In region B of Figure 16-the microstructures of weld metal clearly have dark Ti–Cu–Al–Ni–Fe multi-phase (denoted by arrow (1) and bright and white Cu solid phase (denoted by arrow (2). In region C of Figure 16 the weld interface zone showed complex chemical and metallurgical reactions. It consists of a gray Ti₂Cu phase (denoted by arrow 1), light gray TiCu phase (denoted by arrow 2).

CMT can provide low heat input, minimize crack formation and thereby considerably enhance the mechanical characteristics. An energy-dispersive X-ray line analysis of two joints showed that a large amount of copper was diffused into the weld zone and IMCs layers of Ti–Cu, Ti2–Cu, and AlCu2Ti were formed in the joints. The IMC layers were 140–160-μm thick. After lap-welding Cu–T2 to Cu–T2 using CMT, the tensile shear strength of both joints was 194 N/mm. The fracture locations in the HAZ of the copper sheet were the same in both joints and exhibited a plastic fracture mode. Though the metals were welded successfully, the joints obtained had tensile strength less than that of cu base metal (Figure 17).

Although Cu–Ti welds are strongly crack and corrosion-resistant and can withstand high loads, their weldability is challenging owing to the chemical composition difference between the two metals. Zhao et al. welded titanium TA18 and pure copper T2 sheets of 5-mm-thickness using an YLS-6000-S2-TR fiber laser. The 0.45mm offset of the laser beam to the copper side resulted in high tensile strength, and the high heat input allowed the diffusion of Ti elements into Cu, thus improving the plasticity and toughness of the welded metal. However, ICMs such as TiCu, TiCu₂, Ti₂Cu, and Ti₃Cu₄ reduced the plasticity and strength in the regions of their formations [70].

An analysis of Liu and his colleagues examined the weld reliability of TC4 (Ti6Al4V) and 304SS (1Cr18Ni9Ti) sheets using three independent electrodes: ERCu (Cu), ERCuSi-A (Cu–Si), and ERCuNi30 (Cu–Ni) with diameters of 0.8, 1.0, and 1.2 mm, respectively [71]. They applied a unique welding process called MIG–TIG double-sided arc brazing (DSAB), in which the TIG and MIG electrodes were placed opposite each other and the TIG electrode was offset by 1.5 mm from the center of the weld line.

Macroscopic findings showed no macro flaws (such as cracks and pores) in the weld regions. Both base metals barely dissolved under the low heat input and retained their initial solid-state interface. The MIG–TIG DSAB method joined the titanium and stainless steel with good weld formation. Depending on the input factors, the weld zone and HAZ primarily exhibited a dark-gray Ti₂Cu zone, a light-gray TiCu zone with (Ti₂Cu + TiCu) and TiCu intergranular layers, or a (TiCu + Ti₃Cu₄) zone. Phases of Ti–Cu or Ti–Cu–Ni in the IMC increased the microhardness values of the different joints, which peaked at 573 Hv on the Ti side. During a tensile strength test, the thick Ti side of the joint was fractured because of its high brittleness, showing a maximum tensile strength of 276 Mpa. The feasibility of welding non-ferrous metals are discussed as follows and the 645 same is tabulated in

Table 7. Possible DMW combinations of nonferrous metals by GMAW and its variants.

Welding Process	DMW	Results	Difficulties & Possible Solution
P-GMAW [58]	Ti–Al	1. Joint made was free of pores and cracks.	1. Formation of brittle IMC (Ti3Al, TiAl3, and TiAl). 2. Not successfully welded by conventional arc welding processes.
P-GMAW [65]	Ti–Ni	1. Not successful by conventional arc welding processes. 2. Good bonding of metals was achieved at wire-feed rates of 2.2 and 2.5 m/min on welding Ti with Ni filler wire.	1. Formation of brittle IMC. 2. Using copper or columbium as the filler may form joints without harmful IMC.
CMT [69]	Ti & Cu	1. Copper was highly diffused into titanium, forming a firm joint.	1. Differences in thermal diffusivity 2. Low solubility of Cu in Ti. 3. Joints of Ti–30Cb and Ti–3Al–6.5Mo–11Cr exhibited the highest tensile strength and ductility.
CMT [72]	Ti & Mg	1. Good weld quality and high tensile load capability of 2 kN was achieved. 2. A filler wire with aluminum content should be selected for the arc welding process.	1. Challenges are the low solubility and lack of inter-reaction between the metals. 2. Not successfully achieved by conventional arc welding processes.
GMAW-TIG Hybrid [73]	Cu & Al	1. Heat input had direct relation with the strength of the joint.	1. Eutectic Si particle caused the crack propagation
External magnet assisted CMT [74]	Cu & Al	1. A copper–aluminum transition filler should be used. 2. The Al2Cu layer thickness evidently decreased under a high EMF.	1. Brittle IMC formation. 2. A buttering layer of silver or silver alloy will help the welding of these materials using GMAW.
GMAW [75]	Cu & Ni	1. The base metal and filler wire were properly mixed. 2. No serious problems arose when welding these two metals and their alloys.	2. Any filler wire containing nickel and copper can be used for welding.
GMAW [76] & CMT [77]	Mg & Al	1 The molten zinc compound prohibited the combination of the weld zone and the Mg base metal at the joint interface. 2. Zinc layer is used in the intermediate region of the weld plates.	1. Brittle IMC and differences in melting points inhibit the welding. Precise control of the heat input is necessary. 2. A multilayered microstructure of eutectic structure, Mg17AL12, and Mg2Al3 layer forms in the fusion zone.
CMT [78]	Mg & Al	1. Zinc-based Inconel 625 filler wire is used.	1. Formation of Mg2Al3 had limitation on the strength of the joint.
TIG [79]	Co & Ni	1. Performance of the joint was satisfactory under continuous high-temperature conditions, and it resisted oxidation and corrosion. 2. Heat input for welding cobalt should be kept at low-to-moderate level for better results.	1. Not successfully achieved by conventional arc welding processes. 2. Good cobalt-welding practices require adequate joint preparation and thorough cleaning before welding to ensure sound joints.
Not welded by GMAW	Cu & Mg	-	Extremely difficult to join Mg alloys to Cu through GMAW.
Not welded by GMAW	Co & Cu; Co & Ni	-	Cracking can easily occur when welding cobalt and copper.

and

Table 8. Weldability of dissimilar metals using GMAW & its variants.

Metals	Al	Ni	Fe	Cu	Ti	Mg	Co
Al	-	W	W	W	W	W	N
Ni	W	-	W *	W	W	W	W
Fe	W	W *	-	W	W	W	W *
Cu	W	W	W	-	W	C	Cr
Ti	W	W	W	W	-	W	N
Mg	W	N	W	C	W	-	C
Co	C	N	W *	Cr	C	N	-

Metals (Al: aluminum; Ni: nickel; Fe: iron; Cu: copper; Ti: titanium; Mg: magnesium) and their results on dissimilar joints (W: weldable by the respective welding process; C: complex structures may exist; N: not weldable; Cr: cracking occurs). * Note—readily weldable with ferrous alloy metals.

.

4.5. Welding of Copper to Aluminum

Aluminum and its alloys have been widely used in manufacturing applications such as food manufacturing, marine, etc. because of properties such as low specific gravity, high electrical conductivity, and strong corrosion resistance. Multi-material items may improve the economic and technological advantages of industries. Owing to their admirable metallurgical properties, thermal conductivity, and electrical resistance, the aluminum-copper combination is widely used in various applications. Cylindrical sheets, capacitors, transport bars, heat-exchanger tubes, and electrical connectors are the basic applications [80]. According to some researchers, the Wright brothers benefited from the strengthening effects of age-hardened Al-8%Cu during their first flight in 1903, as parts of the aircraft engine were cast using this material [81].

To lower the weight and raise the ductility of parts, many industries use copper and aluminum. In pressure vessel construction, shipbuilding, and renewable energy, these metals are favored for their good formability and long life. However, they are very difficult to weld without modified heat input. Using the hybrid weld method, a lap joint of 1.0-mm-thick 5052 aluminum alloy and 2.0-mm-thick pure copper was made by double-electrode GMAW and GTAW with an ER4047 filler wire [73].

By precisely controlling the heat input into the base metals, they also controlled the formation of IMC layers and the microstructure of the weld. The brazed zone was mainly composed of Al2Cu IMCs and eutectic Al–Cu. Under tensile loading, fracture occurred at three positions in the Al & Cu lap joints such as HAZ (Figure 18), fusion zone (Figure 19), and interface of the weld with different heat input. The thicknesses of the IMC layer and fracture zone increased when the heat input was increased. Eutectic Si particles and IMCs break the bond in between base metals, so crack propagation more possible.

Some researchers have also studied the effect of pulse current and pulse frequency on dissimilar welded joints. Liu and his research team outlined the application of a robot in a mass manufacturing environment with the help of an axisymmetric applied magnetic field in the weld zone of the lap joint with an overlap distance of 10 mm between a red copper T2 plate and an aluminum alloy 6061-T6 plate [74].

The welding process was CMT using an AlSi5 aluminum alloy wire with a diameter of 1.2 mm. The welding electrode moved by a six-axis ABB robot was maintained parallel to the usual position of the work material. The shielding gas was argon flowing at 15 L min⁻¹. When an electromagnetic field (EMF) was applied, both the welding arc and weld beads were shifted by the magnetic field and a major interaction region between the metal and Cu surface was induced by the effective saturated conduct. The weld interface consists of an Al-Cu eutectic and Al₂Cu layer which caused infiltration of Cu elements to the Aluminum side (Figure 20).

To mitigate the failure of welded joints under high load, several researchers have proposed a high laser power during the welding process. A study discussed the effect of laser power on the mechanical properties and microstructure for lap joining 1.6 mm thick plates of 6061 aluminum and 110 copper [82]. The shear strength of the joint gradually rose and then declined as the laser power increased. The maximum shear strength was 99.8 MPa when the laser power reached 2.45 kW. The IMC widths also depend on laser power, but the IMC thickness differed from the depth of the hardness indentation.

The thickness of the Al₂Cu layer was evidently reduced as the EMF increased. The interfacial microstructure was fundamentally unchanged by raising the EMF frequency, but the structure and development path of the Al₂Cu substrate at the weld interface distinctly differed from the crystal structure after welding without the magnetic field.

4.6. Welding of Copper to Nickel

Yang and Kou investigated the proper mixing of base metals (Ni 200 and Cu 101 of thickness 6.4 mm) welded by GMAW with three different filler wires (pure Ni of diameter 1.1 mm, pure Cu of diameter 1.3 mm, and Cu–30.40Ni of diameter 1.1 mm) [75]. The liquidus temperatures of the bulk weld metal TLW, base metal T_LB, filler metal T_LF, and partially mixed filler metal near the pool bottom T_LF were measured for thermal analysis. Macrosegregation occurred in the weld metal and was anticipated to occur with high probability at the bottom of any weld metal. A filler-rich zone was observed near the bottom of a Cu weld with Ni filler metal (T_LW = 1366 °C and T_LB = 1085 °C) and a similar weld (T_LW = 1338 °C and T_LB = 1085 °C), where T_LW > T_LB. Conclusively, Copper and nickel DMW can be achieved using any copper based filler wire. Copper-nickel alloys can be used for a long period in submerged applications, such as seawater pipework, heat exchangers, battery applications, unloading boats, and commercial boats [83]. They are also used in desalination and power plants.

4.7. Welding of Copper to Brass

Copper has unique characteristics and provides excellent heat conductivity, electrical conductivity, corrosion resistance, and machinability. Copper can be alloyed with zinc. By varying the percentages of zinc and copper in brass alloys, the mechanical properties can be optimized.

Dissimilar copper–brass joints formed under a low heat flow have relatively low mechanical strength. This problem can be rectified using activated TIG (ATIG), which directly applies a DC as a straight polarity (

Table 9. Effects of frequency on the grain types observed in the microstructure of brass–copper welds [84].

Hardness	Type of Grains	High Frequency—100 Hz	Low Frequency—5 Hz
High	Coarse	DC	DC
Intermediate	Intermediate coarse dendritic	AC, mixed	Mixed
Low	Fine dendritic	-	AC

). The resulting DC-ATIG and its AC counterpart (AC-ATIG) may improve the weld penetration. These type of dissimilar joints are widely used in automobile industries for manufacturing copper-based radiator systems [2].

Flux powders such as SiO₂, Al₂O₃, and TiO₂ are used for welding brass to Cu in both AC and DC modes. In the surface morphology study, AC welding yielded a fine dendritic grain under a TiO₂ flux and coarse dendritic structures under a SiO₂ flux. Meanwhile, DC welding created a coarse columnar structure under an Al₂O₃ flux. As a fine dendritic grain structure enhances the strength, AC-ATIG is more effective than DC-ATIG [84].

4.8. Welding of Cobalt to Nickel

Comparison study of hardness and weld bead geometry of dissimilar joints made from AMS 5536 nickel and AMS 5608 cobalt alloys using GTAW and EBW clearly shows that during the EBW hardness value of the joint got reduced (Figure 21) [79]. This unique heterogeneous metal joint confers the combined characters of its base metals: oxidation resistance, corrosion resistance, high-temperature strength, and good metallurgical stability. The hardness of the welded dissimilar metal was measured at different points. The hardness values in the weld zones and HAZ of the joint formed by TIG reached 250 and 276 HV0.5, respectively. The weld zone and HAZ of the joint formed by EBW shared the same hardness value (244 HV0.5).

When joining dissimilar alloys, a “global” filler material is preferable over a more specialized one. Global filler can potentially combine several distinct alloys when the highest weld quality is not mandatory and achieves acceptable material characteristics and resistance to corrosion. The HAYNES 556 alloy (AWS A5.9 ER3556) has exceeded its expectations, providing a flexible filler metal for a range of dissimilar high-temperature welding applications including Ni-, Fe-, and Co-base metals, which are chosen for their good welding properties, resistance to corrosion, and exceptional weld strength.

A standardized filler wire should also cope with the corrosion conditions of different alloys. This need is particularly valid when “upset” situations are inevitable or in multiapplication environments in which the alloys encounter a broad range of chemicals.

4.9. Welding of Magnesium to Aluminum

CMT with AlSi5/ER4043 and pure copper (HS201) filler wires have been used for welding AZ31B magnesium and 6061 aluminum alloys [77] (Figure 22). A microstructural study of the joint welded with AlSi5 revealed three types of zones. The third zone was characterized by occasional porosity shrinkage caused by a wide range of solidifying temperatures. The excessively low heat input and the Si additive in the filler wire hindered the creation and development of brittle IMCs in the weld metal, thus improving the quality of the joint (Figure 23).

Meanwhile, in the joint welded with pure copper wire, the penetration increased with current, voltage, and wire feed rate. An undeniable multilayer microstructure of eutectic structures such as AlCu, CuAl₂, Cu₉Al₄, Cu₂Mg and Mg₁₇AL₁₂, and Mg₂Al₃ were found in the fusion zone of joint welded by pure copper and ER4043 respectively. The microhardness distribution was determined by a microhardness tester on the fusion zones. In both joints, the microhardness increased more sharply at the Mg side than at the Al side where the IMCs were reduced (Figure 24). The brittle fracture occurred in the IMC layer of the fusion zone at the Mg side because a large particle phase was formed in that zone. Joining Al and Mg in a hybrid structure can allow those metals to be used in automobile sector [85], lightweight joints in manufacturing industries [86].

High-penetration melting of the base metal and macroscopic brittle cracking in aluminum-magnesium welds can be mitigated by inserting a zinc interlayer between the two metals. These outcomes have inspired new techniques for joining two metals using traditional fusion welding processes. The joint made using nickel-based Inconel 625 as filler wire, had a significant effect on the tensile-shear load of the lap joint [79].

Zinc filler wires are usually used in arc and laser fusion welding. Microstructural growth was investigated by Zhang and Song to weld dissimilar Al-Mg lap joint using the GMAW technique with a zinc interlayer [76]. An aluminum–silicone hypoeutectic arrangement and an aluminum-zinc eutectic arrangement were observed in the weld zone. Diverse aluminum-zinc intermetallic elements were produced at the boundary between the weld region and the undissolved magnesium metal.

In pulsed GTAW, the weld heat input mainly depends on the current and speed of the welding. When the heat input variation reaches its peak value, it is directly related to the microstructure of the weldment and the penetration and fusion of filler metal into the base metal. An Al-5052-H34 aluminum alloy sheet lap was successfully joined to alloy steel sheet st-12 using P-GTAW [87]. Scanning electron microscopy of the joint formed under a low heat input revealed α-Al dendrites and Al-Si eutectics in the regions between the joint interface and the fracture locations. When the heat input increased, the temperature obviously increased and the rate of cooling reduced. In this case, Al3Fe and Al5Fe2 phases are the intermetallic compounds formed. Notably, as the heat input increased from 100 to 250 J·mm⁻¹ and above 380 J·mm⁻¹, the mechanical shear tensile strength rose and declined. The current frequency achieved sufficient penetration under a low heat input.

4.10. Welding of Magnesium to Titanium

The microstructure and mechanical strength of Mg–Ti welding depends on the eutectic phase and the Ti-Al IMCs [88] formed near the interface of the Ti/Mg dissimilar material. The successful dissimilar joint by arc welding techniques to produce Mg & steel welds using CMT was studied [89]. The joint was made from pure Ti (TA2) and AZ31B (3 wt. % Al) in different configurations by the CMT technique with an AZ61 filler (6 wt. % Al). An analysis of the interface characteristics showed that in both joint configurations, the brazing interfaces determining the joint quality were composed of Ti₃Al and Mg₁₇Al₁₂ phases resulting from diffusion of Al elements from the molten magnesium base metal. Meanwhile, the filler wire was aggregated at the liquid/solid interface and reacted with the titanium substrate [72].

Higher hardness was also found in the interface of titanium and magnesium base metal (Figure 25). The brazing of these dissimilar metals produced an interface mainly composed of Ti₃Al, Mg₁₇Al₁₂, and Mg_0.97Zn_0.03 intermetallic in investigation carried out previously [69]. The selection of filler wire made of magnesium consists of aluminum and Zinc in fair content is very essential to join Mg and Ti base metals successfully. The use of magnesium and titanium alloys has increased dramatically in recent years, notably in the transport and aviation industries [90].

Meanwhile, TIG welding is favored over MIG welding because the TIG arc is stable even at low currents; therefore, TIG is promising for joining thin parts. TIG welding offers additional benefits such as flexibility, high efficiency, and good joint quality and can potentially join dissimilar metals such as Mg and Ti.

Excessive grain coarsening occurred at the fusion zone of titanium TA2 and magnesium AZ31B lap-welded by TIG with AZ31B Mg-based filler [91]. This phenomenon was caused by differences in the thermal properties of the welded metals. A welded–brazed joint was formed and metallurgical bonding at the joint interface was achieved by the coarse-grained fusion zone with a precipitated Mg17Al12 phase and a distributed Mg/Ti solid solution. Under the optimum processing parameters, the maximum shear strength of the joint reached 193.5 N·mm⁻².

5. Conclusions

This review provides insights on the current status of welding research, focusing on the feasibility of non-ferrous dissimilar welding and the shortcomings of existing GMAW methods are highlighted. Some conclusions made are as follows:

• When welding dissimilar nonferrous metals, various specifications must be considered, including the base metal characteristics, electrode selection, and welding phase selection. The traditional GMAW process is not recommended for welding non-ferrous metals because the heat input control and arc stability are inadequate for this purpose.
• During welding of dissimilar non-ferrous and ferrous metals, deciding the welding cycle is as important as selecting the appropriate electrode that to be used in the welding process. Specifically while welding different non-ferrous metals, the welding method should be deliberately selected as heat input affects shrinkage and element relocation. With its high stability and tolerance, adaptive GMAW can minimize the shrinkage and residual stress induced by thermal coefficient variations.
• AZ31 magnesium and aluminum were effectively welded using an adaptive GMAW. The hardness in the intermetallic layer of the welded metals was increased by increasing the temperature and duration of the heat input. Similarly, Cu–Ni joints can be welded by GMAW using any nickel-based filler wires.
• CMT welding enables dissimilar metal joining of Ti & Cu; Ti & Mg; Cu & Al, and Mg & Al respectively.
• The research aspects that are not yet covered thoroughly are modeling of certain non-ferrous dissimilar metal joints like Ti& Al, Ti & Mg, and Al & Mg, which are based on the findings of this review.
• Researchers and manufacturers can use this review as a guideline to choose appropriate welding parameters to implement GMAW and its variants for non-ferrous dissimilar welding.

6. Future Research Perspectives

(a)

Investigate new welding techniques for preventing undesirable intermetallic compound forming during the melting process.

(b)

Instead of welding, a new way of combining dissimilar metals could be more appropriate. An alloying approach, for manufacturing a unique material as per the requirement, maybe a way to create a component with a homogeneous and well-bonded microstructure.

(c)

Further study into different defects in welding dissimilar metals is needed. Improved weld zone hardness, joint strength, oxide formation, and reduction or removal of intermetallic compounds are indeed a few examples.

(d)

Examination of the impact of combining mechanical interlocking with welding for joint strength can be done.

(e)

Introducing composite filler wire used for ferrous and nonferrous metal welding is worthy of further research associated with DMW. It could lead to improvement in mechanical and metallurgical properties of the joint.