Vous êtes sur la page 1sur 26

Materials and Design 141 (2018) 230–255

Contents lists available at ScienceDirect

Materials and Design

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

Friction stir welding of titanium alloys: A review


Kapil Gangwar, M. Ramulu ⁎
University of Washington, 98195 Seattle, USA

H I G H L I G H T S G R A P H I C A L A B S T R A C T

• Friction stir welding of titanium alloys


critical review from year 2007 onward
is presented.
• Different titanium sheet and plate alloys,
similar and dissimilar titanium's FSW
process conditions were highlighted.
• Microstructural (including texture evo-
lution), and mechanical properties are
discussed.
• Current challenges and suggestion for
future research.

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

Article history: The candidacy of ninth earth abundant material, titanium is supreme when it comes to the higher strength to
Received 28 July 2017 weight ratio, and higher corrosive resistance. Next generation of jet engines for aerospace industry clearly de-
Received in revised form 1 December 2017 pends on the manufacturability and improved ability of titanium alloys that can withstand the high tempera-
Accepted 16 December 2017
tures. However, the primary sheets, plates, billets, ingots, or rods are of limited sizes that need to be either
Available online 21 December 2017
machined or welded in order to produce a desired structure with optimal the buy-to-fly ratio. This article primar-
Keywords:
ily summarizes the research in the field of joining of titanium sheets with a direct focus on friction stir welding
Friction stir welding (FSW). The industrial requirements, whether to weld similar or dissimilar titanium alloys, or of different thick-
Titanium alloys nesses, or even with composites, are often challenged by the selection of the tool, position of the alloying sheets,
Microstructural and mechanical properties cooling of the weld nugget, material composition, complex material flow in the wake of the rotating, and travers-
Texture ing tool etc. This article provides a review of FSW of similar and dissimilar titanium alloys focusing on surface, and
subsurface properties, such as microstructural, and mechanical properties, texture evolution, current challenges
summarizing a possible remedy, encompassing the recent development and research in the field.
© 2017 Elsevier Ltd. All rights reserved.

1. Introduction continuous focus on reducing the buy-to-fly ratio, and on inventing


the new lighter weight material is an example of such diligence. Con-
Amidst the presence of a constant conflict between the cost and the ventional titanium alloys are continuously challenged by the operation
application of the titanium and its alloys, there exists an equally assidu- at temperatures higher than 550 °C due to poor oxidation resistance,
ous pursuit to use the material efficiently. Titanium alloys, over the de- followed by poor creep properties, undesirable grain growth in single
cades have drawn attraction due to their high strength, excellent phase titanium alloys, and consequent degradation of mechanical prop-
corrosion resistance, and biocompatibility. Aerospace industry's erties. On the contrary, biomedical industries are also in relentless quest
of developing alloys that exhibit high strength, and low elastic modulus
⁎ Corresponding author. matching the mechanical properties of human bone. A metastable β
E-mail address: ramulum@u.washington.edu (M. Ramulu). type titanium alloys Ti-33Nb-4Sn (wt%) developed by Shun et al. [1]

https://doi.org/10.1016/j.matdes.2017.12.033
0264-1275/© 2017 Elsevier Ltd. All rights reserved.
K. Gangwar, M. Ramulu / Materials and Design 141 (2018) 230–255 231

Neo. Compatibility of titanium with composites has given rise to the


use of titanium in aerospace industry, as shown in the Fig. 1 [4].
In addition to existing titanium alloys, third generation forgeable γ-
TiAl for low pressure turbine blades in the Pratt &Whitney PW1000G
engines have made their impact in the aerospace industry. Simulta-
neously, engine frames, and nacelle components are a result of new al-
loys, owing to their forgeability, superplastic formability. A recent
development of new α + β high temperature titanium alloys have
been marked by Kobe Steel Ltd., Japan, Allegheny Technology Inc.
(ATI), Pittsburghg, and Timet Metal Corp., Exton, PA with improved me-
chanical properties [4]. Table 1 shows the recent developments in the
field of titanium alloys.
An increasing use, cost associated with the titanium alloys, and how-
ever, difficult machinability of titanium has overshadowed conventional
manufacturing processes. Long before friction stir welding came into
existence, fusion welding had, and still does, an innate ability to join
the structures with complicated geometries. What seemed impossible,
thanks to fusion welding, appeared achievable with a gamut of advan-
tages that fusion based additive manufacturing processes render. How-
ever, similar to any happenstance, the pros of the fusion welding
Fig. 1. Increasing use of titanium in commercial aircrafts. Bubble size denotes the techniques set up an overdue tryst with inevitable cons as the research
passenger capacity [4]. Images not to scale but to show an increasing passages capacity in the realms of thermodynamics, and phase transformation progresses.
as the time progresses.
Dendritic microstructures, and the associated phase transformation due
to melting of titanium followed by subsequent cooling, threatened the
structural integrity of the welded joints that presented themselves as
has shown elastic modulus as low as 36GPa, and tensile strength over a site susceptible to crack initiation during either cyclic, or monotonic
853 MPa by retaining low content of β stabilizers at room temperature loading in fusion welding techniques. The formation of intermettalic
by high density of dislocations and grain boundaries introduced by cold phases in the weld, and the contaminations imparted by the filler mate-
rolling (CR) and annealing treatment (CRA). In a study of multi-pass rials are the primary downsides of fusion welding. Going through in-
friction stir processed CP-Ti, Arash et al. [2] have concluded no effect tense research and pushing through the boundaries of unimaginable,
of FSP on semiconducting behavior. However, the presence of passive fusion based additive manufacturing technologies have emerged as
film due to multi-pass FSP followed by emergence of diverse structural having the ability to join the metal sheets to form near net shapes. Ap-
defects and grain refinement resulted in an improved passive behavior plications of high intensity processes such as Electron Beam Welding
and higher corrosion resistance in strongly acidic solution of 0.5 M (EBW), Laser Beam Welding (LBW), Plasma Arc Welding (PAW) range
H2SO4. Mott–Schottky analysis, in this study shows that multi-pass from low cost automotive in-line part unit, to high cost, and extremely
FSP has great potential for increasing the corrosion resistance of pure reliable units in aircraft industry [5]. EBW, however, a superior example
Ti, mainly due to the formation of more resistive, thicker, and yet less of fusion welding technique, the higher heat intensity associated with
defective passive films (with lowered donor densities on all passes). the EBW, creates a much smaller fusion zone (FZ), and a heat affected
Titanium alloys are primarily classified as α, α + β, and β. Temper- zone (HAZ), and can be adapted for thick components in aerospace,
ature above which only β phase exists is known as β transus tempera- and automotive industry. In addition to narrower FZ, and HAZ, high
ture. Upon cooling from above β transus temperature, the titanium depth-to-width ratio, and excellent dimensional tolerance makes
alloys can be manipulated to obtain a variety of microstructure with tai- EBW, a prevalent manufacturing process for joining of thick titanium
lored mechanical properties. α, and near α alloys have iron (Fe), and ox- sheets. Deep penetration, and absence of material contamination, due
ygen (O) as an alloying elements. Mo, and V are β stabilizers, hence aid to high vacuum, makes EBW, a suitable candidate for joining titanium
in lowering the β transus temperature. An intentional addition of α (Al, alloys [6]. Nonetheless, the macrostructural changes, concomitant of
or Sn), or β stabilizer (Mo, or V) can have a significant effect on the mi- high energy density, and rapid cooling rate are primarily marked by
crostructure. A detailed information about the prevalent titanium alloys the introduction of porosity in the weld nugget, and the formation of
can be found here in reference [3]. The present and future of the current, brittle phase, martensite. For Ti-6Al-4V, the mechanical properties
and next generation air frames and engines is closely related with the such as yield strength (YS), ultimate tensile strength (UTS), and micro-
precocious research and development in order to meet the stringent hardness increase significantly after one pass EBW, owing to finer la-
fuel efficiency standards, and cost pressures associated with the increas- mella spacing observed in the Widmanstatten microstructures due to
ing orders of highly advanced aircrafts such as Boeing 787, 777X, A320 extensive crisscrossing (or nucleation of multiple variants of α phase

Table 1
Advancements in titanium alloys [4].

Titanium alloys Quality Company

Ti-0.4Ni-0.015Pd-0.025Ru-0.14Cr (AKOT) Corrosion resistant Kobe Steel


Ti-0.5Al-0.35Si (Ti-0.9SA) High temperature oxidation resistance to 765 °C. Silicide inhibits grain growth
Ti-0.5Al-0.45Si-0.2Nb(Ti-1.2ASN)
Ti-2Al-1Sn-1Fe-1Cu-0.5Cr-0.3Si (Ti-2111S) Lower Levels of anisotropy, and improved superplasticity
Ti-5Al-4V-0.75Mo-0.5Fe (Ti-54M) Superplastic forming (SPF) capability at temperatures as low as 775 °C; TIMET
Ti-5.5Al-5V-5Mo-2.4Cr-0.75Fe-0.15O (TIMETAL® 18) Good combination of strength and ductility
Ti-4Al-2.5V-1.5Fe-0.25O (ATI 425) Excellent cold and hot workability ATI
Ti-5.5Al-4.3Zr-5.7V-1.3Mo-0.10O-0.06Pd (RTI-XPT™) Highly stress tubular application RTI
232 K. Gangwar, M. Ramulu / Materials and Design 141 (2018) 230–255

on the prior β grain boundary), leading to considerable amount of diffi- and evaluated in greater details for a better understanding of the
culty for intergranular slip [7]. process.
For structures as big as engine nacelle, and as intricate as bio im- List of abbreviations used in Fig. 2 is shown in Table 2. Complexity
plants, even a slightest window of leeway is blasphemous. With the lim- and correlation among multiple parameters involved in FSW is shown
ited, and pre-defined shapes and sizes titanium sheets comes in, it is in Fig. 2, developed using Circos [22]. Connecting ribbons indicate inter-
inevitable to either machine, or join the sheets, and if necessary dependence on each parameter. In this representation, it has been as-
superplastically form it as well sans compromising the integrity of the sumed that the thickness of alloying sheets (similar titanium alloys
structure as far as the microstructure and other inherent properties with similar β transus temperature) is known and kept constant for a
are concerned. To the point where fruition of the fusion welding tech- set of experiments.
niques can longer be realized, revolutionary paradigm shift in friction Instructions for figure b):
based manufacturing processes emerged as a state of the art solid
state joining technique where phase transformation, and evolution of 1. The outer arcs and corresponding color bars show primary character-
microstructure in the weld nugget is induced by high temperature se- istics of the FSW process.
vere plastic deformation. Hence a parallel research, in the field of an al-
2. Inner arcs represent multiple parameters that are involved in the
ternate additive manufacturing has been aced by friction based additive
process.
manufacturing (FBAM) processes, such as rotary friction welding
3. Interconnecting ribbons depicts inter-dependence on each parame-
(RFW), friction deposition, friction surfacing, linear friction welding,
ter. Width of ribbons shows the dominance of interdependence.
friction stir welding etc.
4. For example, FSWed titanium alloys (T1-ADV, and T2-RET) show a
So far, a vast literature on friction stir welding and its adaptability in
variety of metallurgical (Metall.) and mechanical (Mech.) properties
a variety of materials, including titanium, is available. A detailed review
under varying parameters.
on friction stir welding of metal matric composites (MMC), aluminum
alloys can be found here [8–10]. General consents about friction stir
welding and friction stir processing have been well documented in liter- The surface roughness of the joining sheets is important in
ature so far [11,12]. Numerical and analytical models associated with governing the coefficient of the friction in FSW. The tool geometry,
temperatures and material flow during friction stir welding have been governing friction, material flow, and corresponding cooling rate deter-
discussed in details [13–15]. Furthermore, a thorough analysis for tool mines the temperature profile as the tool progresses along the welding
geometry and tool materials have been carried out by Zhang et al. [16] line. The resulting flow of the material, disparities associated with the
and Rai et al. [17]. A brief review about mechanical and metallurgical initial microstructure, the dependence of the multiphase materials on
properties of friction stir welded titanium alloys have been provided the temperature (texture), thermal conductivity of the materials, local-
by Lienert et al. [18] for articled published until year 2007. As seen in ized heating or cooling (promoting one phase over another) govern the
Fig. 1, with an increasing use of titanium in aerospace industry it is of ut- final microstructure and mechanical properties of FSWed titanium
most importance to use the material efficiently. Over the last ten years alloys.
due to rapid adaptation of this welding technique in aerospace industry,
a variety of titanium alloy sheets of different thicknesses have been 2. Titanium alloys
welded by FSW. We believe that such database along with significant
information about characterization schemes of FSWed near α, α + β, Over the years, since invention of FSW in 1991, a variety of FSWed
and β alloys should be readily available to process engineers utilizing titanium alloys have been studied for their mechanical and metallurgi-
FSW and to the metallurgist of titanium alloys. In the last decade there cal properties. Commercially pure titanium (CP-Ti; α), Ti-6242 (near
have been several advancements and improvements; not only in the α), Ti-6Al-4V (α + β), Ti-54 M (α + β), and Ti-5111 (near α) are to
field of automation of FSW [19] but also in the process design, tool de- name a few covered in this study. Near α, and α + β alloys behave qual-
velopment and microstructure controls to improve the mechanical itatively similarly but the transformation temperature from one constit-
properties of the joints. With an advent of new titanium alloys, it is uent to another, and the growth kinetics differ from one to another due
even more important to extend our prevalent knowledge of FSW for to the difference in the composition.
an efficient applicability of titanium alloys. However, to the best of our Depending on the alloying elements present in the binary titanium
knowledge, a detailed review on the friction stir welding of titanium, alloy, the titanium alloys can be classified as fully α, near α, binary
and it's alloys that comprises the information on tool selection, welding phase α + β, and fully- β (metastable β) [23]. The β transus tempera-
parameters, material flow, microstructural evolution, temperature dis- ture is an indication of the alloying element, i.e. alloying elements that
tribution, mechanical properties (fracture, fatigue, residual stress), and raise the β transus temperature, are known as α stabilizers while others
texture evolution does not exist after year 2007. Therefore, the scope will suppress the β transus temperature and are referred as β stabilizers.
of this review is limited to the friction stir welding of titanium alloys Al, O, N, and C are common α stabilizers. Aluminum, however, is the
for articles published over the last decade in order to encourage engi- most sought alloying element in titanium alloys due to its good solid
neers to use this technique more effectively along with suggested future solubility in both α, and β, is a potent solid solution strengthener, and
work to contribute more to scholastic community. increases the β transus, in addition to lowering the density of titanium
The key questions that are being raised in the scientific and technical alloys.
activities in the field of friction stir welding (in general) can be summa- β stabilizers can be further categorized as β isomorphs (V, Nb, Mo),
rized based on the Fig. 2 as shown below. Fig. 2 shows the viewpoint of and β eutectoid forming (Cr, Fe) elements depending on the type of the
friction stir welding for multiple parameters that are of importance in binary phase diagram the elements forms with titanium [24]. Sn, and Zr
the process [19–21]. While considering an impact of these parameters have no effect on the phase stability of the titanium alloys [23,25]. Small
it has been assumed that the automation of the FSW machine, tool tilt addition (1–2%) of β phase stabilizer in fully α titanium alloys results in
angle, tool insert hole and location of alloying sheets (in consideration the formation of near α microstructure, consisting of primary α and
with the need of pre-heating and presence of backing anvil) are opti- retained β (b5%) [26]. Further addition of β phase stabilizers (4–6%)
mized. In addition, it has also been assumed that the associated vibra- in the α + β alloys results in the β volume fraction of (5–40%). An in-
tion dynamics of entire system are capable of producing defect free creased amount of β stabilizers (10–15%) results in metastable/near β
FSWed joints of titanium alloys. Although, these assumptions are quite alloys, and single phase β alloys (N 20%β stabilizers) [27]. The low diffu-
hypothetical yet, after decades of research in FSW of titanium alloys, sivity of the β stabilizers facilitate full retention of the β phase to the
we can safely assume that these parameters are thoroughly studied room temperature at moderate cooling rate.
K. Gangwar, M. Ramulu / Materials and Design 141 (2018) 230–255 233

Fig. 2. A viewpoint set for friction stir welding.

Ti-5Al-2Sn (Ti-5-2.5), Ti-3Al-2.5V (Ti-325), Ti-2.25Al-11Sn- 5Zr- increasing the Fe concentration. Otherwise these alloys cannot be
1Mo-0.2Si, and Ti-8Al-1Mo-1V and CP-Ti are the typical example of α strengthened by heat treatment.
alloy where α phase is predominant in the microstructure. These alloys Near α, and α + β phase alloys behave qualitatively similarly but the
have good strength and high workability which can further be improved transformation temperature from one constituent to another, and the
by increasing the O content. Due to absence of ductile-to-brittle transi- grown kinetics differ from one another due to the difference in the com-
tion in α alloys makes them a suitable candidate for high temperature as position. Ti-6Al-4V, and Ti-6Al-2Sn-4Zr-2Mo-0.1Si (Ti-6242) are the
well as cryogenic application. These alloys have a small fraction of β typical examples of α + β and near α alloy respectively. The β transus
phase present at the room temperature which can be increased by temperature of these alloys depend on the O and Al content. These al-
loys exhibits a good combination of strength and ductility, however,
the operating temperatures for these alloys are limited by the creep re-
sistance, and the oxidation [24].
Table 2
List of abbreviations for terms used in Fig. 2.
Furthermore, the microstructure, and associated mechanical proper-
ties of these alloys can be manipulated by the optimized heat treatment.
Anvil Backing anvil Many of the affected properties due to heat treatment are inversely re-
Ar-Flow Argon flow
Cyl. Cylindrical
lated. For an instance, for Ti-6Al-4V the fatigue crack initiation resis-
Design Tool design tance increases as the fatigue crack growth resistance decreases. Ti-
Displac. Displacement control 6Al-4V, is by far the most widely used titanium alloy with a wide
FSW-A FSW-automation range of application that runs from dental crowns to aircraft engines,
FSWed Friction stir welded
from orthopedic implants to critically loaded structures in aircraft oper-
Load Load controlled
Mate. Tool material ating under different thermal conditions.
Mech. Mechanical properties Titanium alloys with high percentage of β stabilizers, commonly
Metall. Metallurgical properties known as β alloys- Ti-15Mo-3Nb-3Al-0.2Si (Timetal 21S, or β21S), Ti-
Power FSW-A power 5Al-5V-5Mo-3Cr (Ti-5553), and Ti-11Mo-6Zr-4.5Sn. The formation of
rpm Rotation per minute
Stiffness Stiffness of FSW-A
β′, and omega phases has been described earlier. Furthermore, low elas-
Strain-R. Strain rate tic modulus, and the biocompatibility of the β alloys makes them an
T1-ADV Titanium alloy 1 kept on advancing side ideal candidate for the medical implants because these alloys are in
T2-RET Titanium alloy 2 kept on retreating side the range of the elastic modulus of human bones, hence, lowering the
Taper Tapered
interfacial stresses on the implants. In addition, the β alloys find their
Temp. Temperature
Thread Threaded application in spring, fasteners, and landing gear due to the low elastic
Tilt Tilt angle modulus.
Tool Tool Fig. 3 summarizes the properties and characteristics of different tita-
Traverse Feed rate (mm-min−1) nium alloys. α alloys exhibits higher β transus, higher flow stress, great-
Wear Wear
er ability for welding, and high strength at increasing temperature. β
234 K. Gangwar, M. Ramulu / Materials and Design 141 (2018) 230–255

Fig. 3. Characteristics of titanium alloys.

alloys exhibit higher forming capacity, high strain rate sensitivity, steel and titanium if threaded tools are used. Thus, heat and deforma-
higher ability for heat treatment, and higher strength at the room tem- tion induced by the tool are an important part of the complexities that
perature. However, service temperatures for β alloys have limitations are followed by the consequent properties once the tool passes from
due to lack of creep resistance. Further information about the effect of one location to another.
alloying elements on titanium alloys can be found here [24,27–29]. A schematic diagram of the friction stir welding, tool shape and size,
and the material flow in the wake of tool is shown in the Fig. 4. The most
3. FSW process trivial welding configuration is the square groove butt joint, however, a
variety of configurations such as corner joint, T-butt joint, lap joint have
Friction stir welding (FSW), developed in 1991 at The Welding Insti- been produced [11]. Once the tool design, and the material is agreed
tute (TWI), is performed by a non-consumable rotating and traversing, upon, tool rotation speed and the traverse speed come into play to tailor
and slightly tilted (2°–3°) tool in the softer material's side in case of dis- the microstructural and mechanical properties followed by underlying
similar materials) inserted into the interface at the contiguous line of material flow. The rotation of the tool facilitate the mixing of the two ad-
the metal plates. By means of automated FSW (either displacement con- jacent material and translation of the tool deposit the softened material
trolled or load controlled), the tilted tool is then inserted to create a start in the wake of the tool from retreating side to advancing side. Higher ro-
hole in order to reduce the forces for tool propagation. Accordingly, a tation speed causes, higher temperature and subsequently higher stir-
backing anvil (either preheated or at room temperature) also needs to ring which as a result cases an excessive amount of flash generation
be clamped under the welding sheets. As the tool rotates and traverses, giving rise to defects, such as porosity, and inadequate material flow
severely plasticized material migrates from retreating, to advancing in the weld nugget. Hence, identification of the optimized parameters
side, and deposits in the wake of the tool. The metal plates are thus corresponding to the appropriate tool design is necessary to produce a
joined by the friction between the tool shoulder, tool probes, and sound weld with desired mechanical properties [30]. Fig. 4a is the typi-
alloying sheets. The combination of frictional and adiabatic localized cal schematic of the friction stir welding depicting the fundamental
heating in the weld nugget softens the material promoting the material terms use in this study. Fig. 4b shows the variety of tools (images not
flow around the rotating tool. The traverse speed determines the expo- to scale) that have been used by Edwards et al. [31] to join (3-, 6-, 9-,
sure time, i.e. the softened material is continuously transferred around and 12-mm thick) Ti-6Al-4V sheets. Fig. 4c and Fig. 4d are the top
the tool and then deposited on the trailing side forming a solid-phase view of Fig. 4a and optimized processing parameters (for 3-, 6-, 9-,
bond between the two sheets. In friction stir welding the tool geometry and 12-mm thick Ti-6Al-4V sheets) respectively.
typically consists of round shoulder, and a cylindrical pin. Based on the During friction stir welding tool plunge depth, and the axial forces
requirements, and the material to be joined, the pin of the tool can ei- applied on the tool, can be categorized as two processes: (a) position
ther be threaded, or tapered, or threaded with flutes or no feature at control; where the plunging depth is constant, and controlled regardless
all. The shape and size of the tool are the significant factors in heat gen- of the welding force, (b) load control; a fixed load is applied and the tool
eration during friction stir welding. Threads and scrolls are generally in is able to adjust its position into the butting line communicating with
the opposite direction of the tool rotation such that the material is being the controlled load. Furthermore, the applied load is dependent on the
pulled into the center of the weld instead of casting it outside the weld flow stress of the workpiece, which is related with the temperature
nugget. Threads on the other hand are designed to push the material that is controlled by the tool rotation speed. Variation in the plunging
into the weld zone, but can also be engineered to pull the material up- depth can also affect the temperature and loads. An increased plunging
wards. Tapered tools simply facilitates the flow of the material and are depth will increase the process temperature, decrease the flow stress,
generally used for high strength aluminum alloys, high strength steels, and may cause the thinning of the weld zone due to a significant
and titanium alloys. High strength materials, with high work hardening amount of the material being displaced at the joining line. Conversely,
coefficients are likely to undergo an incomplete penetration, or plasti- an adverse situation may appear due to less frictional heat, higher work-
cizing hence the presence of either threads or flutes in the tool can result piece flow stress, and increased forge loads if the plunging depth is not
in the formation of cavities, or other undesirable properties in the weld. adequate.
Furthermore, the poor conductivity of titanium and its alloys poses an-
other challenge for process and properties optimization due to continu- 3.1. Tool selection, and thickness variation of alloying titanium sheets
ous thermal and microstructural gradient in and around weld nugget.
High strain rates involved in the deformation by FSW can have a signif- Unarguably, the FSW of low melting temperature alloys of Al and Mg
icant effect on the resulting properties of high strength alloys such as do not suffer much from the challenges of tool wear. Although matured,
K. Gangwar, M. Ramulu / Materials and Design 141 (2018) 230–255 235

Fig. 4. (a) Schematic of friction stir welding; isometric view, (b) Front view of (a), (c) schematic of transverse cross section of FSWed dissimilar titanium alloys and (d) Identification of the
processing window for FSW of (3-, 6-, 9-, and 12-mm thick) Ti-6Al-4V sheets [32].

FSW of titanium due to lower thermal conductivity (7 W/mK) of titani- where τ is the shear yield strength as a function of the temperature and
um and high melting temperature, results in an uneven thermal gradi- strain rate, κ is the contract condition coefficient between tool and the
ent from base material on the advancing side to base material on the workpiece, ω is the rotation speed, and ν is the traversing speed of the
retreating side. Tool selection and parallel consideration of the sheet tool. C1, and C2 are the constants derived from the tool geometry. A va-
thickness are of equal importance to produce a defect free joints. Com- riety of titanium alloys have been welded by friction stir welding by
mercially pure tungsten tools upon cooling from 1473 K cause recrystal- adopting a variety of processing parameters. Edwards et al. [30], for
lization and embrittlement. However, an addition of rhenium (Re) identification of the suitable processing conditions for 3-,6-,9-, and 12-
reduces the ductile to brittle transition temperature (DBTT) by influenc- mm thick Ti-6Al-4V (as shown in Fig. 4d), have provided a detailed in-
ing the Peierls stress for the motion of the dislocations. Similarly, tung- formation about the processing parameters and thickness of alloying
sten carbide (WC) based tool due to their excellent toughness and sheets along with tool dimensions. Hereafter, as the tool material, tool
hardness, and insensitivity to sudden changes in temperature and design, and processing conditions are identified; primary microstruc-
loads, have also been used for the FSW of titanium alloys. The contami- ture, and β transus temperature of the titanium alloys become of impor-
nation of workpiece by Si and N from the tool was prevented by TiC/TiN tance in understanding the evolution of the microstructure in the WN,
coating. Sintered TiC welding tool, with a water cooling arrangement to TMAZ/HAZ. To produce a defect free weld and to achieve a full penetra-
extract excessive heat from the tool, has been used for successful FSW of tion depth in thicker material, a rather longer tool geometry is sug-
titanium. A detailed review about tool dimensions and processing con- gested. A study conducted by Shude et al. [21] for friction stir lap
ditions for FSW of titanium alloys can be found in a study conducted by welded Ti-6Al-4V has suggested that a slight penetration into lower
Rai et al. [17]. sheet can significantly reduce the hook effect. Their study, conducted
in the light of temperature gradient as it forms in the weld nugget, illus-
3.2. Parameters selection trates that higher temperature (assisted with higher rotation speed) re-
sults in formation of void-like defect.
Microstructural evolution, mechanical properties, residual stress, Higher energy inputs, and the optimized cooling rate (a flow of inert
hardness, fracture toughness, fatigue crack propagation rate, accumula- gas (Ar) on the top of the welded surface, and back anvil) can influence
tion of high angle grain boundary, grain refinement, strain rate, material the evolution of the microstructure through thickness of the weld nug-
flow, and super plastic forming ability of the welded joint are primarily get. Furthermore, tool degradation due to hot adhesion can also take
affected by the adopted processing parameters. An optimization of tool place under the influence of high temperature and vertical pressure at
rotation speed, and tool traverse speed is much sought for in order to the end of the tool which was in the rotating and advancing process
avoid any welding defect congruent of high temperature (due to high [33]. Jinwen et al. [34] proposed a model for optimizing the rotation
rotation speed), and less exposure time (due to high traverse speed). and transverse speed for FSW of AA6061 can be extended for titanium
Energy input due to friction is a function of tool geometry, hence, the en- alloys by employing suitable thermodynamic, and mechanical proper-
ergy input of thicker sheets is ought to be higher due to longer length of ties of titanium alloys [24,27–29].
the welding tool. The energy input per unit thickness of the sheet can be
obtained by following relationship. 3.3. Material flow

Once the tool design, and processing parameters are optimized, un-
Q ¼ τκ ðωC 1 þ νC 2 Þ derlying material flow in the wake of the tool throughout the weld
236 K. Gangwar, M. Ramulu / Materials and Design 141 (2018) 230–255

nugget rises to wily power in defining the course of macrostructure and In α + β titanium alloys, due to large differences in the flow stress of
microstructure as the plasticized materials cools, and deposits from the existing phases, and the thermal conductivity, it is formidable to
retreating side to advancing side. In FSW, due to complexity of the de- compare the flow patterns as they emerged in the aluminum alloys.
formation, the shear directions, and hence the resulting texture are con- However, as challenging as it appeared, in a study performed by Ed-
stantly evolving as the material flows around the tool. Studies by Seidel wards et al. [41] a tracer technique has been used to visualize the mate-
et al. [35], Colligan et al. [36] and Schmidt et al. [37] and Morisada et al. rial flow in FSWed Ti-6Al-4V by X-ray and CT-scan. The tracer material
[38] demonstrated a detailed overview about the material flow in Al al- used was a tungsten (W), 25% rhenium (Re) blend powder 20 μm in
loys. It has been suggested that a material in the form of embedded size. W and Re have densities roughly 5 times that of Ti-6Al-4V, which
marker, or steel shot beads, or strips of metal, or wire markers, or tracers is one of the main reasons they were selected as the powders to be
should not be reacting with the base material, and must possess the embed in the titanium workpiece. Fig. 6 shows X-ray, CT-scan, and ra-
ability to sustain temperatures as high as being produced during friction diographic visualization of the material flow during friction stir welding
stir welding. of Ti-6Al-4V [41].
X-ray, tomography, CT-scan, and radiography can be further used to
understand the material flow. Seidel et al. [35], Colligan et al. [36],
Guerra et al. [39] have used the embedded markers, steel shot beads, 3.4. Temperature distribution
and strips of copper respectively to visualize the material flow by
adapting the X-ray technique. Not only similar, but also dissimilar mate- The friction stir welding of titanium alloys, especially of α and near α
rial flow has also been observed for underlying material flow. Schmidt alloys is often met with challenges posed by tool limitation, and thick-
et al. [40] have performed CT-scan to visualize the material flow. A nu- ness of the alloying sheets. A simple, either tapered or cylindrical, design
merical models have also been proposed in consideration of welding pa- of the tool is responsible for the preempted vertical flow of the material
rameters [37]. Fig. 5 is a three dimensional illustration of the material unlike aluminum. The high temperatures originating in the root of the
flow by radiography. The material flow and associated velocities have tool are responsible for the material softening and the consequent ma-
also be exercised by Morisada et al. [38] to understand the effect of terial flow. However, in case of α or near α titanium alloys the higher
strain and strain rate in determining the recrystallization of the grains flow stress of α phase, and lower thermal conductivity, and higher
in aluminum alloys during friction stir welding. Material flow, especially heat capacity of the titanium makes it difficult for the selection of the
in case of FSW of dissimilar alloys, is extremely complicated. Not only tool materials for titanium alloys with high β transus temperatures. β al-
welding parameters play an important role, material composition loys, or α + β alloys are prone to undergo par β transus temperature
poses another challenge in determining the migration of the alloying during friction welding depending on the welding parameters and the
elements. thermal distribution during FSW.

Fig. 5. Locus of the tracer observed by the X-ray transmission real-time imaging systems. (a) three-dimensional image, (b) two-dimensional image on TD-WD plane (c) schematic drawing
of calculation method for strain, (d) schematic drawing of calculation method for strain rate [38].
K. Gangwar, M. Ramulu / Materials and Design 141 (2018) 230–255 237

Fig. 6. Material flow patterns as a function of welding parameters shown by (a) X-ray and (b) CT scan [41].

Spatial and temporal dependence of evolving microstructure is a re- converted into heat giving rise to adiabatic shear bands with localized
sult of temperature profiles, and associated phase transformation in the heating and the irregular microstructure.
titanium alloys. However, temperature measurement during friction
stir welding is a herculean task due to intense plastic deformation in- 4. Microstructural evolution during FSW of titanium alloys
volved in the process. Edwards et al. [32] have attempted to measure
the temperature distribution by embedding the thermocouples either Ever since FSW of high temperature alloys came into realms of
outside or in the weld nugget. Their experimental measurements achievable paradigm; titanium, and steel have been of prime focus
were made to determine the peak temperatures during friction stir among researchers [43–48]. A ubiquitous fact about FSW of aluminum
welding of Ti–6Al–4V alloy as a function of the processing conditions alloys, that it produces three distinct zones; weld nugget (WN), thermo
such as tool rotation speed and traverse speed. It was found that the mechanically affected zone (TMAZ) heat affected zone (HAZ), holds its
spindle speed has a dominant effect on peak temperatures, while tra- nerves when it comes to the unanimity of the FSW zones in titanium al-
verse speed controls exposure time. Low spindle speed conditions loys. Some researchers claim no existence of the TMAZ [49,50], while
lead to peak temperatures near, or below, the β transus temperature some bolster their findings about narrow TMAZ by showing a distinct
of the material, 1000 °C (1800 °F), while high spindle speed welds result microstructural observation between HAZ, and TMAZ [51,52]. Accord-
in peak temperatures above 1200 °C (2100 °F). Fig. 7 demonstrate the ing to Pilchak et al. [51] and Knipling et al. [52] the weld nugget, appar-
temperature profile as measured by Edwards et al. [32]. ently acting as a heat source, transfers a small amount of heat in its
A fully-coupled thermal–mechanical finite element model by using surrounding, therefore, plastic deformation in these regions was limited
commercially available softwares, such as Ansys, Abaqus, Deform can and a negligible thermos mechanically affected zone has been formed.
be used to numerically simulate the joining process with corresponding The microstructure evolving in the TMAZ, HAZ, is primarily affected by
heat generation, in order to reduce the number of experiments. While the microstructure of the base material (BM). However, the microstruc-
there is no suitable study demonstrating the temperature profile around ture of the weld nugget is a result of processing conditions, and material
the weld nugget, an analogy can be made from the studies of Cho et al. composition [51,52].
[42] based on the modelling results of friction stir welding of 304 stain-
less steel using finite elements method. Advancing side, with rotation 4.1. Microstructure in WN, and TMAZ
and traversing speed in the same direction, is likely to experience higher
temperatures in comparison with the retreating side as shown in Fig. 8 Due to severe plastic thermo mechanical deformation the grains are
[18]. The top surface on the either side of the weld cools by two mech- refined in the WN. In the TMAZ/HAZ, deformation, and continuous dy-
anism. 1) Due to flow of the argon gas, 2) thermal diffusivity of the tita- namic recrystallization (CDRX) associated with higher temperature
nium. Due to which a thermal gradient develops. Owing to its high heat are in competition to define the microstructure in this region. Due to
capacity and poor thermal conductivity, a large portion of plastic work is complex deformation, and the irregularity observed in the temperature

Fig. 7. a) Temperature versus time profiles for a varying traverse speed at constant spindle speed of 300 rpm; and b) varying spindle speeds for constant traverse speed of 100 mm/min
[32].
238 K. Gangwar, M. Ramulu / Materials and Design 141 (2018) 230–255

Fig. 8. a) Schematic of typical thermal profile generating during friction stir welding of TIMET 54 M and Ti-6242 (a combination of dissimilar titanium alloys; notice the emergence of
streaking pattern on the ADV side). Note: Image and colors not to scale; b) Typical heat affected zone thermal cycles for friction stir weld of Ti-6Al-4V. A, advancing side; R, retreating
side [18].

profiles of the FSW of titanium alloys, the evolving microstructure, and size of the α colonies, and prior β grains indicated the increase in the
the textures are persisted as the material flows around the tool and de- temperature as the rotation speed increased, Fig. 9b.
posits in its wake. The presence of simple shear crystallographic texture One of the most agreed mechanisms for the grain refinement in the
as it evolves in the direction of tool rotation can be observed due to TMAZ is considered to be CDRX related to shear-induced lattice rotation.
bending, and breaking of the lamellar microstructure of Ti-6Al-4V [53]. According to the literatures, the root of the CDRX in FSW can be found in
The transformation, nucleation, and growth of the respective phases, the very early studies on FSW of aluminum alloy [56]. However, some
also depends on the amount of α and β phases present in the base ma- recent works mentioned the discontinuous DRX [57]. In the first place,
terial. The grain refinement of near α, and α + β alloys have been inves- although enough high strain rate can be predicted, it is difficult to
tigated by several researchers. Ti-6Al-4V, by far, is the most studied prove DRX in the FSW process where the plastic deformation stage
titanium alloy (commonly known as the workhorse of the industry) and post heat affected stage (cooling stage) cannot be separated
for friction stir welding [54]. Edwards et al. [32] have studied FSWed completely. Therefore, further research in a controlled environment
12 mm thick Ti-6Al-4V sheets with a water cooled W\\La alloy with that accounts for stage by stage observation of microstructure, needs
25 mm shoulder diameter and a tapered pin, 6 mm in length, adopting to be carried out in order to understand the true mechanism behind.
different rotation, and traverse speed to study the peak temperatures, In case of FSW of titanium alloys where material in the WN (and
and the resulting microstructure in the weld nugget. Fig. 9a, shows around) is constantly compressed (from tool shoulder and backing
that a minor influence of rotation speed, and traverse speed has been anvil), and sheared (due to tool rotation and traverse), imparted plastic
observed in the microstructure of the WN for 12 mm thick Ti-6Al-4V. deformation and resultant temperatures play a dramatic role in devel-
Nonetheless, higher rotation speed and lower traverse speed resulted opment of microstructure. On one hand, assumption of simple mecha-
in a coarser grained microstructure due to increased temperatures and nism either CDRX or DRX seems reasonable if heating and cooling
extended exposure times. Zhang et al. [55] have demonstrated the effect were the only two factors governing the evolution of microstructure.
of varying rotation speeds on FSWed 3 mm thick Ti-6Al-3V. The base In addition to these two factors, relaxation of material under compres-
material with an α/β lamellar microstructure evolved into a fully lamel- sive and shear forces as the tool departs from one location to another,
lar microstructure with refined prior β grains in the WN. The coarser followed by the impact of heating and cooling rate on β transus

Fig. 9. a) Macrographs and micrographs of welds as function of rotation speed and traverse speed [32], b) SEM images of the WN produced at different rotational speeds. Lamellar α/β,
prior β and α colonies are shown by arrows in the 400 rpm weld [55].
K. Gangwar, M. Ramulu / Materials and Design 141 (2018) 230–255 239

Fig. 10. SEM images of the fracture locations (HAZ) in the 300, 500 and 600 rpm welds [55].

temperature of titanium alloys, are few other dominating parameters in Fig. 12a, of BM shows elongated α with subgrains, equiaxed α and a
thorough understanding of microstructural evolution in FSW of titani- lower volume fraction of residual β (black region marked with white
um alloys. As shown in Fig. 2, introduction of other parameters (i.e. β arrow), which were distributed intermittently between α grains. In
transus temperature and the extent of phase transformation mecha- Fig. 12b, the border of TMAZ II and WN is defined according to whether
nism β → β + α) will add more to the complexity of FSW. or not the material exceeded β transus. The presence of primary α in the
Furthermore, as the two phases α (HCP), and β (BCC) evolve, it is TMAZ, and the fully lamellar microstructure in the WN suggests that the
most likely that the deformation of these two phases, whether by temperature in the TMAZ, and WN was below and above β transus re-
grain boundary sliding (GBS), or by twinning is affected by the presence spectively. In TMAZ I (Fig. 12c) due to low angle grain boundaries
of each other. Processing par β transus temperature results in the com- (LAGBs) introduced initially, the recrystallized grains subdivided into
plete transformation of the α phase into β phase giving rise to the fur- smaller subgrains. As the strain increased from TMAZ I to TMAZ II, the
ther nucleation of the twelve different variants of α phase. The subdivided α grains were refined, and slightly elongated along the bor-
resulting microstructure, and concomitant mechanical properties are a der of the TMAZ II and WN. A similar patterns has also been observed by
results of that transformation. The effect of rotation speed, as observed Pilchak et al. [53] in an investment casted, FSWed Ti-6Al-4V sheet.
by Zhang et al. [55] on the microstructural evolution in the HAZ is As a result of adopted welding parameters the temperature in the
shown in the Fig. 10. The HAZ contained a bimodal microstructure TMAZ and WN was below and above β transus temperature respective-
consisting of primary α and α/β lamellar microstructure within prior ly. In TMAZ I, α to β phase transformation during heating, and β to α +
β grains. The prior grain size increases with rotational speed. As seen β transformation during cooling, occurred between two original elon-
in Fig. 10, it can be observed that as the rotational speed increased, gated α grains. Groove, a result of diffusional flux from β phase into
higher temperatures associated with the processing resulted in the α/α boundaries, were formed at the α/α boundaries. Dislocations intro-
CDRX, with evolution of rather bigger prior β grains in the HAZ. duced as a result of thermomechanical process arrange themselves in
the form of subgrain boundaries by DRV. In TMAZ II due to higher tem-
4.2. TMAZ and HAZ (deformation mechanism) perature than TMAZ I, a larger degree of β phase penetration occurred
along the α/α boundaries and it broke up large α grains into refined
Thermo-mechanically affected zone due to lowest hardness (as de- subgrains facilitated by the grain boundary sliding (GBS) in the direc-
scribed later), if measured on the transverse cross section of the FSW tion of shear stress. Furthermore, for a two phase Ti-6Al-4V titanium al-
joint, is most prone to failure under transverse tensile tests. The evi- loys, α/β interphase boundary also plays an important role during high
dence of weaker nature of the thermo-mechanically affected zone can temperature deformation. According to this study it has been postulated
also be seen in case of the electron beam welding where the digital that the α/β phase boundary sliding (PBS) occurs more rapidly than α/
image correlation based strain measurement showed the accumulation α; grain boundary sliding (GBS) [60]. Therefore, for the deformation
of maximum strain in the TMAZ [58]. Therefore, a close inspection of mechanism in TMAZ resulting in α/β phase transformation, and groov-
TMAZ is necessary to understand the deformation mechanism. Wu et ing along the α/α boundary favors PBS. The presence of high angle grain
al. [59] have studied the microstructural evolution in the thermo-me- boundaries, and sheared refined shear grains in the direction of applied
chanically affected zone in a 2.4 mm thick Ti-6Al-4V sheet friction stir stress by rotating tool suggest that GBS, PBS, and grooving are interrelat-
welded (800 rpm, and 200 mm/min) with a polycrystalline BN tool ed and facilitate each other. The EBSD, and TEM images of the TMAZ and
with a concave shoulder 15 mm in diameter and 2.2 mm in length in WN are shown in the Fig. 13. For a detailed information, the reader is
an argon shielded environment. The typical macrostructure of the highly recommended to refer to [59,60].
FSWed Ti-6Al-4V sheets is shown in Fig. 11 [59]. Corresponding texture
and microstructural evolution in TMAZ has been discussed in the fol- 4.3. Microstructural evolution in the WN, TMAZ, and HAZ
lowing section.
According to Wu et al. [59] the deformation mechanism in the Although the resulting microstructure in the WN, and TMAZ/HAZ is
thermomechanically affected zone (if divided into TMAZ 1, beside BM, a culmination of initial microstructure of the BM, tool geometry, pro-
and TMAZ II beside WN) is governed by dynamic recovery (DRV) in cessing condition, cooling rates, chemical composition, yet assuming
TMAZ I and by continuous dynamic recrystallization (CDRX) in TMAZ that the temperature goes par β transus in the WN, the mechanism
II as shown in the Fig. 12. for the microstructure evolution in the WN can be described in the fol-
lowing section. On the contrary (assuming that) in TMAZ, the tempera-
tures are below β transus, the microstructure evolution is a function of
initial BM microstructure, and is described below in Fig. 14.
As the tool approaches the point of interest along the joining line, (or
a particular element in the workpiece), the temperature begins to rise,
and as a results the β phase located at the grain boundaries of α begin
to grow in expanse of α. As the temperature rises above the β transus
temperature there exists only one phase in prior β grains. The size of
the prior β grains is primarily dependent on the initial microstructure,
Fig. 11. Friction stir welded 2.4 mm thick Ti-6Al-4V sheets. Instead of HAZ, TMAZ has been and temperature. Due to severe plastic deformation induced by the
formed with deformation characteristics similar to TMAZ [59]. tool, the β phase deforms with shear, and compressive components.
240 K. Gangwar, M. Ramulu / Materials and Design 141 (2018) 230–255

Fig. 12. EBSD maps of (a) BM, (b) TMAZ, with selected areas shown at higher magnification in (c) and (d). SZ = Stir zone (equivalent to weld nugget (WN)) [59].

After the deformation, as the tool departs from the particular position, can see that, below β transus temperature, the pre-existing phase is
the material begins to cool down from the peak temperature in the combination of α and β phases. Upon cooling, corresponding transfor-
weld nugget ensued by the frictional heat, and deformation. As a results mations in the β phase begins. Amount of transformed β is likely to be
of dynamic recrystallization and dynamic recovery the growth of β higher in the top of weld in comparison with HAZ. Width of α lath is
phase (as shown in the Fig. 14. However, not necessarily of the same higher (in comparison with what is observed for the case where tem-
size as beginning) grows until the nucleation of α phase begins at the peratures are par β transus) marked by slower cooling rate in this
prior β grain boundary. On the contrary the temperature in the TMAZ/ case. In the center and bottom of the weld, only refinement of α grain
HAZ the temperatures are well below the β transus temperature so has been observed due to plastic deformation imparted by tool pin. Ad-
not all the primary α is converted into β. Upon cooling the β phase starts ditionally, heat dissipation is relatively higher in the bottom of the weld
transforming into acicular α surrounding the primary α. A schematic di- in comparison with the center of the weld. As a result, dynamic recovery
agram of microstructure evolution in the titanium friction stir weld is of refined grains in the bottom of the weld is more rapid than in the cen-
shown in Fig. 14. ter of the weld.
The microstructural evolutions have been demonstrated with aid of A research study conducted by Shude et al. [21] suggests that using a
above and below β transus temperature. Main focus of evolution of mi- back heated assisted friction stir welding is a feasible solution to reduce
crostructure is on the top and bottom of the weld nugget. Correspond- the tearing effect in 2 mm thick Ti-6Al-4V sheets. Their study suggests
ing cooling scheme has also been highlighted during evolution. In that for processing parameters 350 rpm and 50 mm/min, back heating
summary, it can be deduced that initial microstructure, during heating at temperature 480 °C results into defect free weld with minimum
starts consuming primary α. As temperature reaches above β transus, tool wear in addition to a significant drop in temperature gradient
all β phase exists and goes through shear and compressive deformation through the thickness. Furthermore, the microstructure in the weld
followed by dynamic recovery of β grains. Growth of β grains is nugget is characterized as α/β lamella microstructure correlating with
inhibited by the nucleating α. Nucleation of α at the prior β grain the higher values of hardness [61]. In a study of similar kind, thermal
boundaries, and at the triple junctions of β grains results in final micro- stir weld (TSW) have been analyzed by Fonda et al. [62] for evolving mi-
structure that consists of plates of α in a retained β (acicular or basket- crostructures and textures. Authors have shown that sheets of α and
weave) morphology. Microstructural evolution in the TMAZ/HAZ of ti- near-α titanium alloy of thickness 12.3 mm can be welded by TSW, a
tanium friction stir weld during heating β consumes α but tempera- variant of FSW, due to presence of an external heat source to preheat
tures are below β transus, so primary α remains untransformed. the alloying sheets. As a result, the weld microstructures at the root re-
Furthermore, the prior β grains β after heating, promotes nucleation flects the influence of lower temperatures from the reduced induction
of α at the grain boundaries of prior β. The acicular morphology sur- heating, smaller tool cross section, and the heat sink effect from the un-
rounds the primary α [18]. derlying anvil. These factors cause a substantial refinement in the grain
Evolution of microstructure below β transus temperature follows size at the base of the weld and limit the width of the weld nugget. How-
quite a similar trend except the pre-existing phases, and corresponding- ever, based on this study it has been concluded that tensile welding
ly transformed phases are significantly different. As shown in Fig. 14, we stresses associated with the severe deformation, and stirring of the

Fig. 13. TEM micrographs showing grain structures of (a and b) BM, (c and d) TMAZ I and (e and f) TMAZ II: (a) elongated α grains, (b) coarse recrystallized α grains, (c) grooving across α/
α boundaries, (d) dislocations in initial recrystallized α grains, (e) fragmented α grains and (f) refined equiaxed α grains [59].
K. Gangwar, M. Ramulu / Materials and Design 141 (2018) 230–255 241

Fig. 14. Microstructural evolution in the friction stir welded titanium alloys. Are encircled by red and blue depicts evolution of microstructures above and below β transus temperatures
respectively.

material are not dependent on the peak temperatures, but also on the 1) Absence of second phase to initiate the fracture.
temperature gradient along the thickness of the sheets [61]. 2) The plastic anisotropy of α phase limits the availability of the slip
(and or twin) systems that can operate during loading.
3) Presence of two ductile phases in near α, and α + β alloys can result
5. Mechanical properties of FSW joints of titanium alloys in a complicated fractured morphologies.

Mechanical properties of the friction stir welded titanium alloys are 5.1. Tensile properties
evidently dependent on the evolved microstructure. The macrostruc-
ture as observed in the titanium alloys, is of parabolic [63] shape in com- The peculiarity in the global stress strain curves of the FSWed titani-
parison with banded (or onion rings) elliptical structure as observed in um alloys is the lack of work hardening. In fact, the FSWed titanium al-
aluminum alloys [11] where band spacing and tool traverse per rotation loys exhibit a small amount of work softening. In addition, the percent
are in accordance. In aluminum alloys, as observed by Fonda et al. [64], elongation, as observed in the global stress strain curves of FSWed tita-
the banding can be attributed to the grain size effect and fluctuations in nium alloys is relatively lower in comparison with the base material
the density of the second phase particles or changes in the crystallo- owing to the non-uniform elongation throughout the guage length of
graphic texture. In titanium alloys, on the other hand, lack of any hard the transverse tensile specimen stemming from the microstructural
second phase or inclusion so the only formation of banding in the gradients created by the welding process. For FSWed specimens, the
weld nugget can be a result of difference in the texture. Furthermore, strains are usually carried by the narrow regions of TMAZ. A full field
Lienert et al. [65] have suggested that upon cooling, once the tool has strain mapping of the electron beam welded Ti-6Al-4V has shown the
passed, β to α + β transformation can potentially mask the formation dissimilarity in the strain fields. Based on the study conducted by
of bands in the weld nugget. In a study conducted by Reynolds et al. Saranath et al. [58] for EBWed Ti-6Al-4V it has come to our illustration
[66] for the FSWed β-21S it has been found that if the cooling rates that region around the weld pool is most susceptible to fracture. The
are sufficiently high, β phase (with lower flow stress) can be retained heterogeneity in the elastic, and plastic properties as calculated by the
at room temperature. Because of a wide range of microstructural gradi- uniform stress method (USM), and virtual field method (VFM), is an in-
ent developing in the weld nugget a variety of mechanical properties dication of the complexity of the microstructural gradient that are asso-
can germinate. Based on the temperature profiles as shown in Fig. 7 ciated in the joining of the titanium alloys.
the microstructures are slightly different on both sides of the weld nug- Lee et al. [67], based on the FSW of CP-Ti has drawn some insightful
get. As a result, the fracture properties, and morphology of the mechan- conclusions about the deformation mechanism involved during the
ically tested surface during fatigue, and fracture can be different welding. Materials with HCP structure, such as α-Ti (case of CP-Ti) are
depending on the nature of loading. The fundamental fracture mecha- easily crystallized under severe plastic deformation, and heat, thus,
nisms for titanium alloys can be interpreted as following. the overall grain structure in the WN of the CP-Ti is marked by the
242 K. Gangwar, M. Ramulu / Materials and Design 141 (2018) 230–255

temperature, dynamic recrystallization of β grains followed by nucle-


ation, and growth of α variants inside the prior β grains that are delin-
eated by the grain boundary α. High rotational speed, resulting in a
higher temperature accounts of larger proper β grains, hence lower
values of hardness have been recorded. On the contrary some re-
searchers have shown that hardness values in the weld nugget have de-
creased in comparison with the base material, shown in Fig. 15 [70].
From Fig. 16a, it can be concluded that for 400 rpm the temperatures
in the WN have not exceeded β transus resulting in a rather refined
equiaxed α and transformed β. As the rpm increased upto 500 and
600, the prior β grains with grain boundary α and lamellar α + β colo-
nies formed in the prior β grains are marked by temperature reaching
par β transus in the WN. The subsequent decrease in the hardness
values is shown in the Fig. 15. Based on the temperature profile as
shown in Fig. 8, the advancing side of the weld shows higher hardness
due to slightly higher temperature in comparison to retreating side
followed by DRX.
The decreased hardness values in the WN of the FSWed Ti-6Al-4V
are explained by the annealing effect caused by the frictional heat gen-
erated between tool (tungsten–rhenium alloy pin tool) and the work-
Fig. 15. Hardness variation as recorded on the center of the cross section of the FSWed Ti- piece (2 mm thick mill annealed Ti–6Al–4V plates) [70,71]. The
6Al-4V [55], effect of rotation [70] and traverse speed [71]. reduction in the dislocation density for the welds processed at
500 rpm, and 600 rpm, followed by DRX is responsible for the hardness
decrease in the weld nugget. Similarly, the effect of traverse speed, if
recrystallized grains. Juhas et al. [68] has proposed that the sizeable higher, appears to be detrimental for the hardness values as shown in
strains associated with the deformation imparted in the WN, as a result Fig. 15. At higher traverse speed, due to lesser exposure time, the DRX
of β to α + β transformation, were accommodated and distributed in the WN is not fully developed resulting in lower values of hardness.
through the surrounding materials, resulting in a narrow TMAZ. Zhang As the traverse speed increases, the dislocation density also increases
et al. [69] have also shown that that all the welds of Ti-6Al-4V (trans- resulting in higher values of hardness in comparison to lower traverse
verse tensile tests) have lower strength, and lower elongation in com- speed [71].
parison with BM. In the same study, all the tensile coupons were The hardness values in the HAZ are not affected by either rotational
fractured in the HAZ so the mechanical properties, such as yield speed, or traverse speed. The microstructures on the AS, and RS are
strength (YS), ultimate tensile strength (UTS), and percentage elonga- quite similar as observed in [70,71]. This could be attributed to the
tion are the evaluation of the properties of the HAZ. The increased rota- fact that sheets of Ti-6Al-4V used in their study were of 2 mm thickness,
tional speed has resulted in larger prior β grains, and hence lower hence the cooling rates by the continuous flow of argon, and back anvil
ductility in titanium alloys with a bimodal microstructure consisting of steel sheet were more dominant in comparison with the generating fric-
equiaxed primary α grains and α β lamellae [69]. Effect of processing tional heat. Based on their study it can be inferred that if a sheet of
parameters and rotational speed has further been studied by Zhou et higher thickness processed with similar welding conditions will result
al. [49,70]. It would not be entirely futile to assume that FSWed tensile in a different hardness values on the AS, and RS of the WN. Edwards
specimens in transverse tests are likely to fail in a region where the et al. and Sanders et al. [72,73] have successfully joined the sheets of
hardness is lower; especially TMAZ (in general) in titanium alloys. How- thickness 3 mm, 6 mm, 9 mm, and 12 mm with varying processing con-
ever, in a study by Zhou et al. [70] all the transverse tensile specimen ditions, and with varying tool dimensions and higher values of hardness
fractured in the WN which also had showed an interesting hardness pro- in the WN have been observed.
file with lower hardness recorded in the WN for the adopted processing
conditions as shown in the following section. 5.3. Fatigue, and fracture toughness

5.2. Hardness In a study conducted by Edwards et al. [74] for the fracture and fa-
tigue crack propagation of friction stir welded 6 mm sheets of Ti-6Al-
The hardness profiles in the center of the cross section of the trans- 4V, after post weld heat treatment, it has been concluded that the ap-
verse weld are been shown in the Fig. 15. In general the WN of the parent fracture toughness was lower than the wrought base material,
FSWed has shown higher values of the hardness due to grain refine- 7–25% depending on the crack orientation relative to the weld, but the
ment and high angle grain boundaries, as shown in the Fig. 15. During crack growth performance (ΔK vs. da/dN) of the weld in the absence
FSW processing if the temperatures are rendered above β transus of weld-induced residual stresses was identical to the base material. In

Fig. 16. SEM micrographs of the WN produced at: (a) 400 rpm, (b) 500 rpm and (c) 600 rpm [70].
Table 3
Fracture and fatigue data for friction stir welded titanium alloys.

Material and thickness Processing conditions (rpm/mm-min−1) Heat treatment Fracture/fatigue specimen characteristics, and Conclusions
properties evaluated

Ti-6Al-4V, Mill annealed, 240/100 NO Effect of residual stress on near-threshold crack growth Evidence of dominant influence of residual stress M(T)
6.35 mm thick [63] in the welded region using edge crack compact tension geometry is less susceptible to residual stress
C(T), eccentrically loaded single edge ESE(T) and
center crack configuration M(T).
Effect of R (0.05–0.8) on crack growth in the center of Parametric analysis, using uniform residual stress
the weld using M(T). distribution along the crack plane, suggests that the
Geometry effect on ΔK-Kmax. change in Kres is much higher for C(T) than M(T).
Comparison of predicted (AFGROW) and measured

K. Gangwar, M. Ramulu / Materials and Design 141 (2018) 230–255


fatigue crack growth.
Ti-6Al-4V, grade 5 mill 1) 250–320/45–100 Post weld heat treatment. Held at 760 °C for 30 min in a 1) Butt, 2) corner, and 3) T joints Fatigue performance of FSW corner joint is better than
annealed plate 6 mm thick vacuum furnace, and air-cooled identical structural elements fabricated from wrought
[75] Ti-6Al-4V.
2) 330–450/65–100 S-N curves were produced Using appropriate Kt for geometry and loading type can
3) 250–450/65–100 In addition machined bar and extruded machined bar give good estimate (within 10%).
were tested for uni-axial fatigue testing
R = 0.1 or −1 and Kt = 1.0 or 1.3
Ti-6Al-4V, 6 mm thick [76] 280/100 Post weld heat treatment. Held at 760 °C, 816 °C, 871 Low Kt = 1.2, and Dog-bone shape geometry As welded fatigue life was 50% lower than base
°C, 927 °C for 45 min in a vacuum furnace then material
air-cooled Constant applied cyclic stress 800 MPa at 20 Hz, Two lowest heat treatment temperature further
and R = −0.2 until failure reduced the performance by 70% in comparison with
base material
S-N curves were calculated Higher heat treatment brought the fatigue
performance within 20% of the base material
Ti-6Al-4V, 3 mm, and 6 mm For 3 mm, 300/75 Stress relieved at 730 °C for 30 min Maximum applied stress 830 MPa at 20 Hz, For low Kt (1.0) the weld showed similar endurance
thick [32] and R = 0.1, and 0.6 until failure limit as base material (Compared with MIL handbook).
For 6 mm, 280/100 Kt = 1 to 3, dog-bone shape geometry
S-N curves were calculated
Ti-5111, 12.7 mm and 6.3 225 rpm constant Half of the as-welded and base material C (T) Effect of welding speed on fatigue crack growth rate For base materials, W1, W2, and W3 Kth is 6, 8, 10.2,
mm thick [77] specimens were stress relieved in vacuum at 1100 °C kinetics in a C(T) at a frequency of 10 Hz, and R = 0.1. and 6 MPa√m.
For 12.7 mm welding speed was for 2 h. Followed by rapid cooling in flowing argon at Notch direction parallel to the welding direction, For stress relieved base material, W1, W2, and W3 is
51 mm/min (W1) ambient temperature fatigue crack growth were conducted 3.5, 2, 2, 2 MPa√m respectively.
And for 6.3 mm two welding speeds were
compared; 25.4 (W2), and 102 mm/min (W3)
Ti-6Al-4V, 2.5 mm, 3 mm 300/75 Most of the test panels were stress relieved using a Dog-bone transverse to weld and longitudinal fatigue Base materials fatigue endurance limit that was about
thick [79] vacuum furnace with a temperature of 730 °C and a coupons, S\
\N curves were measured at a frequency of 620 MPa.
soak time of 30 min, which was followed by furnace 15 Hz, and R = 0.1 Transverse FSW coupons (tool marks removed by
cooling machining and also had LPB to induce fatigue-resistant
compressive stress) is 565 MPa.
The transverse FSW coupons (tool marks removed by
machining) showed endurance limit of 496 MPa
Ti-6Al-4V, 2.54 mm thick 300/75 One set was stress relieved at 774 °C for 45 min Fracture toughness measured using R-curve method In the weld center for the SR condition, an increase in
[72] C(T) specimen. fracture toughness due to the stress relieving and
machining has been observed.

(continued on next page)

243
244
Table 3 (continued)

Material and thickness Processing conditions (rpm/mm-min−1) Heat treatment Fracture/fatigue specimen characteristics, and Conclusions
properties evaluated

As welded (AW), stress relieved (SR), SR-machined Higher fracture toughness in the weld zone interface
(M). than the WN center for all processing conditions.
Crack at weld center, crack perpendicular to the weld, Weld interface results on the advancing side give a
crack parallel to the weld slight increase in fracture toughness as compared with
the weld center in the SR condition
Ti-6Al-4V, 6 mm thick [74] 280/100 Thermal stress relief cycle of 760 °C for 45 min in a Fracture toughness measured using R-curve method Fracture toughness of base metal, longitudinal weld,
vacuum and then furnace-cooled C(T) specimen (longitudinal to weld) with crack and transverse weld is 90.9, 84.9, 69.9 MPa√m
perpendicular to the weld, and in a C(T) (transverse to
the weld) with crack in the center of the weld parallel
to the welding direction.

K. Gangwar, M. Ramulu / Materials and Design 141 (2018) 230–255


Same geometry for fatigue crack growth. 10 Hz, and R ΔKth for base metal, longitudinal weld, and transverse
= 0.1 weld is 8.2,8.6,8.2 MPa√m
Ti-6Al-4V, 24 mm thick [78] 150/75 Thermal stress relief cycle of 730 °C for 60 min in a Fracture toughness and fatigue crack growth for C(T) Fracture toughness for base metal, FSW longitudinal,
vacuum furnace and then furnace cooled specimen in both longitudinal and transverse and transverse are 58.4, 78.1, and 81.8 MPa√m.
orientation.
R = 0.1, and 10 Hz, S\\N curves are measured. For BM-longitudinal, and transverse, and FSW-L, and
FSW-T the ΔKth is nearly same for all cases; 9 MPa√m.
Ti-6Al-4V, 2 mm thick [80] 150/100 NO C(T) specimens for decreasing and constant dKI-app (30 Residual stress measured by cut and compliance
MPa√m) were measured at R = 0.1, and 10 Hz. method was analyzed and considered as a factor
affecting the FCG.
Constant load tests for comparison with AFGROW and Compressive residual stress outside the weld and
measured FCG curves tensile stresses inside the weld were found.
Crack growth rate at constant dKI-app was measured FCG rate perpendicular to the FSW joint, showed a
across the weld line longer life sustainability in comparison with residual
stress free base material.
A correspondence between AFGROW and measured
data validates the adoptability of AFGROW while
residual stresses are in consideration.
AFGROW does not incorporate the microstructural
effect hence, numerical analysis does not provide a
thorough analysis.
Ti-6Al-4V, 7 mm thick [81] 120/15 Effect of residual stress and post weld heat treatment C(T) specimens at R = 0.6, and 20 Hz with initial FCG at low ΔK levels, in the stirred zone and the
(PWHT) at 650 °C for 2 h in argon gas environment notches in the BM and the weld affected regions: stir interfacial zones are relatively higher in comparison
followed by furnace cooling zone, inaterfacial zone, heat affected zone. with base metal and heat affected zone.
Fatigure crack growth rate vs ΔK with in-situ digital In case of as welded specimens, microstructure and
microscopy residual collectively account for the FCP, however, in
case of PWHT specimens, microstructure if the
determining factor for fatigue crack growth rate
calculation.
FCP, for microstructurally influenced profiles is a result
of consideration of ΔKeq. The results show that
nominally linear cracks in the weld nugget and in the
interfacial regions facilitate higher growth rate in
comparison with base metal and heat affected zone.
Since temperature in the WN are higher than β transus
resulting in an increment in the β phase fraction
causing undesirable FCG rates.
Table 4
Summary of friction stir welded titanium alloys, processing conditions and metallurgical characteristics.

Thickness (in mm) RPM mm/min Tool design BM ADV CEN CEN RET Texture Ref.
top bot

1.8 180 and 150 30 W-Re concave shoulder right-hand thread YES YES YES YES YES NO [21]

K. Gangwar, M. Ramulu / Materials and Design 141 (2018) 230–255


2 350 50 W-Re alloy tool with a concave shoulder and a YES YES YES YES YES NO [20]
threaded pin
3.17 500 40 and 120 and 200 W-La truncated YES YES YES YES YES NO [145]
2 500 and 600 and 700 and 1000 and 1250 and 1500 100 W-C, conical shape pin with a flat shoulder YES YES YES YES YES NO [146]
2 500 150 W-5Re, concave shoulder tapered pin YES YES YES YES YES NO [147]
2 150 30 W-Re, Shoulder diameter 15 mm YES YES YES YES YES NO [148]
12.7 200 88.8 W based tapered pin tool YES YES YES YES YES YES [149]
Ti-5111
7 120 15 Co-based tool (no geometry provided) YES YES YES YES YES NO [81]
5 50 and 150 10 Co-based tapered tool YES YES YES YES YES YES [98]
5 150 10 Co-based tapered tool YES YES YES YES YES YES [150]
3 and 6 and 9 and 12 300, and 280, and 270, and 170 75, and 100, and 65, and 65 W-La tapered (different for different thickness) YES YES YES YES YES NO [31]
2.5 300 76.2 W-La tapered (different for different thickness) YES YES YES YES YES NO [72]
2 225 and 250 and 300 50 WC tool with a shoulder and probe YES YES YES YES YES YES [151]
2 1100 and 1000 and 1200 and 1200 and 1000 40 and 50 and 50 and 60 and 50 WC tool with a tapered geometry YES YES YES YES YES NO [33]
2 1200 60 A H13 steel holder with fix screws, Ti-Al interfaces were NO [152]
a nickel-base super- characteristics after
alloy (Inconel 718) shoulder and a multi-pass FSP
WC–13 wt% Co pin
2 100 and 1000 and 1000 and 900 and 900 25.4 and 50.8 and 100.6 and 25.4 W–1% La2O3 tool with shoulder and pin YES YES YES YES YES YES [54]
and 900 and 800 and 800 and 50.8 and 100.6 and 25.4 and 50.8
6 Range of 150 to 450 RPM Ranged from 0.85 to 5.1 mm/s A truncated W-1%La2O3 tool YES YES YES YES YES NO [153]
4 450 and 650 and 750 and 850 65 W–25%Re tapered circular YES YES YES YES YES NO [154]
1.25 300 and 500 and 700 25 and 50 and 75 W25Re conical pin YES YES YES YES YES NO [155]
2.5 900 and 1000 and 1100 25 and 50 W–1.1%La2O3 and WC–Co based materials YES YES YES YES YES NO [156]
of two grades
2 Variation between 300 and 1000 Variation between 25 and 400 WC based tool with shoulder and pin YES YES YES YES YES NO [157]
2.5 1000 (for both tools) 25.8 (W-La) and 51 (WC-Co) W-La (shoulder and cylindrical probe) YES YES YES YES YES NO [158]

(continued on next page)

245
246
Table 4 (continued)

Thickness (in mm) RPM mm/min Tool design BM ADV CEN CEN RET Texture Ref.
top bot

WC-Co (shoulder and stepped spiral)


2.4 800 200 Polycrystalline cubic BN tool with a concave YES YES YES YES YES YES [59]
shoulder and a triangular prism pin
2 400 51 W-Re pin tool YES YES YES YES YES YES [96]
6 Combinations of 200 and 300 and 400 Combinations of 50, and 100 and 150 W-La tapered tool Tracer technique has been NO [41]

K. Gangwar, M. Ramulu / Materials and Design 141 (2018) 230–255


adapted to characterize the
material flow.
6 280 100 W-La tapered tool YES YES YES YES YES NO [74]
7 280 100 W-La tapered tool YES YES YES YES YES NO [76]
2 300 100 WC-based tool (shoulder and cylindrical pin) Microstructures and textures were [141] CP-Ti
observed for three cases. 1) high
pressure air cooling, 2) ice–water
mixture cooling and 3) liquid nitrogen
cooling
3 630 4–22 mm/min WC-Co based frustum tool YES YES YES YES YES NO [159]
12.3 300 76.2 NA YES YES YES YES YES YES [62]
CP-Ti
24 150 75 W-La tapered YES YES YES YES YES NO [78]
4 300 100 W-La tapered YES YES YES YES YES NO [137] Dissimilar,
Ti-54 M and ATI-425
7 120 15 Co-based tool YES YES YES YES YES NO [81]
3 275 and 300 50.8 Densimet-176 (stepped pin profile) YES YES YES YES YES NO [160]
3 500 600 Mo based tool YES YES YES YES YES YES [161] VT6
7 800 150 Shoulder; Si3N4 and probe (W-25% Re) YES YES YES YES YES YES [162]
2 50 to 400 50 to 150 WC-Co tool with a shoulder and probe YES YES YES YES YES YES [99] CP-Ti
6 250–450 45–100 W–La based pin tool YES NO NO YES YES NO [75]
3 350–650 40–120 W-1%La2O3 shoulder and conical pin NO NO NO NO NO NO [163]
2.5 100 and 120 and 150 and 200 30 W–Re based tool with right hand threaded pin YES YES YES YES YES NO [61]
3 300 and 500 and 700 35 and 50 W-25Re based tool with a shoulder and conical pin YES YES YES YES YES NO [164]
12.7 200 88.8 W based tapered pin tool NO NO NO NO NO NO [165] Ti-5111
1.2 for CP-Ti, and 1.25 for Ti-6Al-4V CP-Ti (500) and Ti-6Al-4V (700) CP-Ti (55) and Ti-6Al-4V (80) W-La based tool with a shoulder and conical pin YES YES YES YES YES YES [166] similar, CP-Ti and
Ti-6Al-4V
1.8 550 127 W based tool with conical pin NO NO YES NO NO NO [167]
1.6 640; steady after 4 s 620 thereafter 102 W-25%Re based truncated conical pin YES YES NO NO NO NO [168]
3 300–500 50 WC tool with a shoulder and a conical pin YES YES YES YES YES NO [169]
2 250 50 and 75 and 100 W-Re tool with a shoulder and pin YES YES YES YES YES NO [170]
2 1400 40 WC and HSS tool with shoulder and hexagonal pin YES YES YES YES YES NO [2] CP-Ti
K. Gangwar, M. Ramulu / Materials and Design 141 (2018) 230–255 247

Table 5
Summary of mechanical properties in friction stir welded titanium alloys.

YS UTS %Elong Hardness Heat treatment Fractography Fracture Fatigue Wear Residual Remarks Ref.
toughness stress
i
NO NO NO YES NO Yes NO NO YES NO [21]
ii
NO NO NO YES YES NO NO NO YES NO [20]
iii
NO YES YES YES NO NO YES YES NO YES [145]
iv
NO NO NO NO NO NO NO NO Yes NO [146]
v
Superplasticity NO Tensile tests NO NO NO NO NO [147]
achieved by performed at different temperature
lamellar
microstructure.
Data provided
with
temperature
and strain rates
vi
Superplasticity NO Tensile tests NO NO NO NO NO [148]
achieved by performed at different temperature
lamellar
microstructure.
Data provided
with
temperature
and strain rates
vii
Tensile data is YES NO YES NO NO NO NO [149]
provided for Ti-5111
transverse
specimens from
various part of
the weld
(multiple
specimens)
viii
Local tensile YES YES YES YES YES YES YES [81]
properties of
specimens
extracted from
different
locations in the
weld nugget are
provided
ix
NO NO NO YES NO NO NO NO NO NO [98]
x
No data provided YES YES NO NO NO YES [31]
except base
material
xi
Data is provided YES Stress relieving has been done YES YES NO NO NO [72]
for as welded,
stress relieved
and machined,
and stress
relieved
xii
NO NO NO NO NO NO NO NO NO NO [151]
xiii
NO NO NO YES NO NO NO NO YES NO [33]
xiv
NO NO NO NO Heat-inputs during FSLW-processing the upper NO NO NO Interfacial layer NO [152]
pure Al plate analysis by EDS
and XRD
xv
YS, UTS, %Elong NO NO NO NO NO Microstructure NO [54]
data is provided is affected
in the article for by tool wear
all 8 conditions (no data
provided)
xvi
NO NO NO NO NO NO NO NO [153]
xvii
NO YES NO NO NO NO NO NO [154]
xviii
NO YES NO Fracture NO NO NO NO [155]
locations
considered
xix
NO NO NO NO NO NO YES, BSE NO [156]
images
with W debris
xx
YES YES NO NO YES NO NO NO [157]
xxi
YES (high NO NO YES NO NO Debris of tools NO [158]
temperature were
deformation observed in the
behavior is microstructures
discussed)
xxii
NO Yes NO NO NO NO NO NO [59]
xxiii
NO NO NO NO [96]
xxiv
NO NO NO NO [41]

(continued on next page)


248 K. Gangwar, M. Ramulu / Materials and Design 141 (2018) 230–255

Table 5 (continued)

YS UTS %Elong Hardness Heat treatment Fractography Fracture Fatigue Wear Residual Remarks Ref.
toughness stress
xxv
YES (for as welded YES YES YES YES YES NO YES [74]
and stress
relieved
conditions)
xxvi
YES (for as welded Yes YES Yes NO Yes (SN NO NO [76]
and for four heat curves
treated with a site
conditions) for
crack
initiation
and
propagation)
xxvii
YES YES Three cases. High pressure cooling, ice-water Fractured NO NO NO NO [141]
cooling, and liquid nitrogen cooling, were locations
considered are shown
xxviii
YES YES NO YES No NO YES (SEM, EDS, NO [159]
and XRD data
have been
presented)
xxix
NO NO An external heat source is used to preheat the NO NO NO NO NO [62]
workpiece in order to reconstruct
high temperature microstructure and texture
xxx
YES YES Stress relieving has been done YES YES YES NO Stress [78]
relived
xxxi
Yes YES NO NO NO NO NO NO [137]
xxxii
Local tensile YES YES YES YES YES No YES [81]
strengths and
elongations of
the miniature
specimens is
provided
xxxiii
NO NO NO Yes NO NO NO YES YES NO [160]
xxxiv
NO NO NO NO NO NO NO NO NO NO [161]
xxxv
NO NO NO NO NO NO NO NO NO NO [162]
xxxvi
YES YES YES YES NO Fractured No No NO No [99]
locations
are shown
xxxvii
NO NO NO NO YES YES NO YES NO NO [75]
xxxviii
YES YES YES NO NO Crack NO YES NO NO [145]
initiation
sites
are
characterized
xxxix
YES YES YES YES YES YES NO NO YES NO [61]
xl
NO YES NO YES NO NO NO NO NO NO [164]
xli
Tensile data is NO NO YES NO NO NO NO [165]
provided for
transverse
specimens from
various part of
the weld
(multiple
specimens)
xlii
YES YES YES YES NO NO NO NO NO NO [166]
xliii
Global and local NO NO NO NO NO NO NO [167]
stress strain
curves are
embellished. YS,
UTS, and %
elongation is
provided for
corresponding
specimens and
locations
xliv
Load displacement YES YES NO NO S-N curves NO NO [168]
curves for three were
specimens presented
tested under the for as
same conditions welded and
finished
surface
xlv
NO YES YES NO NO NO NO NO [169]
xlvi
Mechanical NO NO YES NO NO NO NO [170]
properties for
three different
K. Gangwar, M. Ramulu / Materials and Design 141 (2018) 230–255 249

Table 5 (continued)

YS UTS %Elong Hardness Heat treatment Fractography Fracture Fatigue Wear Residual Remarks Ref.
toughness stress

conditions are
provided
NO This study presented the effects of FSP on the microstructure and electrochemical behaviors of pure Ti in the phosphate buffer solution. [2]
FSP reduced the grain size significantly down to submicron range as a result of dynamic recrystallization phenomena.
1
[171]
2
[142]
1
Microstructure evolution of dynamic recrystallization (DRX) was simulated by cellular automaton method based on metal plastic flow analysis. The predicted flow stress curves are
similar to experimental measurements. Influences of strain rate and temperature on the evolution of the average dislocation density (equivalent to flow stress) as well as the final grain
size are analyzed.
2
This paper focuses on the advances of FSW in relation to common titanium alloys and explores their potential within the space sector. The study concentrates on research that has been
performed in the past 20 years. The outputs identify achieved successes, encountered limitations and unexplored opportunities. The competitiveness of the more standard welding pro-
cesses mentioned has also been considered to better understand the limited use of FSW for titanium alloys. The requirements of the European Space Agency (ESA) remain primary focus in
consideration along with other government agencies (NASA).
i
Ti-6Al-4V alloy has been friction stir lap welded by slightly penetrating into the lower sheet for the purpose of reducing the hook defect.
ii
To eliminate the tearing defect and to widen the welding parameters for welding Ti–6Al–4V alloy, the current research puts forward a new back heating assisted FSW into use.
iii
Extremely high cooling rates resulted in high Vickers hardness and fine acicular α (needle-like α) in prior β grain boundaries in the weld nugget compared to low Vickers hardness
and Basket weave α/β lamellar colonies for FSW in the weld nugget. High tensile residual stresses were measured in the weld nugget for LBW in comparison to FSW which had high tensile
residual stresses in the HAZ.
iv
Tool wear is strongly affected by the tool rotational speed. The highest tool wear was obtained under low rotational speeds.
v
Fine lamellar microstructure in the nugget of FSW Ti-6Al-4V alloy joint exhibited a low temperature superior plasticity (LTSP) of 442% at 650 °C.
vi
The maximum elongation of FSW Ti-6Al-4V alloy was obtained at 890°C and 2.6 × 10–3 s−1 was 617%.
vii
In terms of transverse direction properties, the advancing interface was found to result in strength values comparable to the BM. It also displayed ductile fracture but exhibited a low
strain to failure. The retreating interface showed highly brittle fracture and values below that of the BM.
viii
Microstructural and local mechanical properties of a FSWed Ti–6Al–4V alloy were evaluated with main focus on the fatigue crack propagation behavior in the different micro-
structural zones.
ix
The temperature at the top surface was lower than the β-transus temperature. The microstructure and the texture distribution in the thickness and the horizontal of the SZ were
investigated.
x
In general, tensile residual stresses are present in longitudinal to the welding direction and compressive stresses are present transverse to the welding direction in as-welded condition
(or prior to a post-weld thermal stress relief cycle).
xi
1) Weld center has higher fracture toughness in comparison to weld interface, 2) impact energy data has been provided.
xii
Analysis of fraction of α and β phase has been.
xiii
Severe wear (hot adhesion) caused inclusions of fragments inside the samples and the loss of surface quality.
xiv
Surface aluminizing on Ti–6Al–4V alloy was successfully performed via a novel solid-state method of multi-pass friction-stir lap welding (FSLW).
xv
Compared with the base material, the SZs of FSP samples exhibit higher tensile strength and comparable ductility. The tensile strength was influenced by a microstructure related to
the prior β grain size and α colony size, which are controlled by processing parameters.
xvi
Dimensionless heat input, suggests that stir zone microstructure is more dependent on RPM than travel speed and that the heat input parameter is not a good indicator of the
processing temperature.
xvii
Comparison between GTAW and FSW is presented by phases detected by SEM and EDS.
xviii
Numerical simulation of temperature, and effective strain distribution.
xix
Tool degradation mechanisms examined in terms of nature of interaction between tool and workpiece.
xx
The effect of the thermal cycle during friction stir welding (FSW) on the microstructure.
xxi
Comparison of room temperature and high temperature deformations was made.
xxii
Grain refinement and deformation process of TMAZ during FSW is discussed.
xxiii
Texture in stir zone is discussed in details.
xxiv
Tracer material used was a tungsten (W), 25% rhenium (Re) blend powder 20 μm in size.
xxv
As welded and stress relieved conditions were characterized.
xxvi
Residual stresses were removed by post weld heat treatments. Additionally, processing conditions were characterized.
xxvii
Finest grain size of 2.0 μm with abundant twin boundaries of 26.8% was obtained by FSW with liquid nitrogen cooling.
xxviii
The presence of WC-Co particles in the microstructure of weld bead results in lowering strength and ductility.
xxix
Thermal stir welding uses an external heat source to preheat the workpiece ahead of the rotating tool.
xxx
Comparison of mechanical and metallurgical properties have been made for FSW and EBW 24 mm thick Ti-6Al-4V sheets.
xxxi
First study to report FSW of dissimilar titanium alloys.
xxxii
Fatigue crack propagation rates obtained through miniature CT specimens with initial notches in the base metal (BM), stirred zone (SZ), interfacial zone (IZ) and heat affected zone
(HAZ) were higher in the SZ and the IZ.
xxxiii
Microstructure characterization of parts produced using a hybrid manufacturing process consisting of laser metal deposition (LMD) and friction stir processing (FSP).
xxxiv
The β → α phase transformation that takes place after FSW is found to obey the Burgers orientation relationship.
xxxv
A unique method was developed in this work to differentiate the HAZ and TMAZ. Crystallographic texture evolution in FSW of Ti–6Al–4V has been shown to follow simple shear of
bcc materials, and textures are presented by the pole figures.
xxxvi
A distinctive plastic deformation extensively occurred within the weld region during the initial tensile deformation stage because the Schmid factor for prismatic slip in the weld
region is much higher than that in the BM.
xxxvii
Friction stir welding testing of titanium corner & T-joints for stress relieved specimens.
xxxviii
Comparison between as welded and polished specimens for their fatigue life (S\ \N curve) has been carried out.
xxxix
Effect of back heating has been characterized.
xl
FEM model was produced from tool plunge to the final in air cool down of the joint, by means of a model that is implicit, Lagrangian, thermo-mechanically coupled with visco-plastic
material characterization.
xli
This work is focused on micro sample characterization of the mechanical properties of a Ti-5111 FSW in the longitudinal direction.
xlii
Maximum hardness values have been achieved in the shoulder contact area. With strain controlled tensile tests of the base material and the FSW welding zone it could be proved that
for the weld the tensile strength decreases 3.1%, but the yield point increases by 11.5% respectively.
xliii
Full field strain maps on the longitudinal cross section of FSWed Ti-6Al-4V are provided by digital image correlation (DIC).
xliv
Fatigue tests performed on titanium alloy FSW joints showed that the fatigue strength of finished joint was greater than the as welded joints. The increase of the fatigue life is up to
1300%.
xlv
Due to both dissolution of smashed β phase, refined microstructure and strain hardening induced by the process, the average values of microhardness is higher in weld nugget (in
comparison with base material) and decrease with increasing rotation speed.
xlvi
Grain size of shoulder-affected zone is bigger than that of weld nugget. At the rotational velocity of 250 rpm, the grain size of weld nugget is decreased by increasing the welding
speed. By decreasing the welding speed, the tensile strength and the ductility of weld are increased.
250 K. Gangwar, M. Ramulu / Materials and Design 141 (2018) 230–255

a follow up study by Edwards et al. [75] to evaluate the fatigue perfor- surface residual stresses developed during welding are independent of
mance of structural joints (T-Joints, and corner), they have performed thickness. However, 12 mm weld did not agree with the concurrent re-
friction stir welding of 6 mm thick Ti-6Al-4V. The fatigue performance sults, with a possibility to either validate or intrigue the scientific com-
of the structural joints was either equivalent or better than the identical munity for future research.
structural element manufactured from the wrought Ti-6Al-4V. It has
been observed that a suitable heat treatment of the FSWed Ti-6Al-4V 5.5. Superplastic deformation
can significantly be improved due to residual stresses being released
[76]. However, with the post weld heat treatment one might encounter The agglomeration of friction stir welding, and super plastic forming
the deterioration in the hardness, and strength. Hence, it is recommend- is capable of producing large, single, and near net shape structure. By
ed to optimize the heat treatment temperature high enough without additional manufacturing processes such as joining, either by fasteners,
compromising the strength, ductility, and hardness of the welded joints. or rivets, or by welding, followed by superplastic forming could be
Pao et al. [77] have been found that fatigue crack growth rates, and fa- employed to pre-obtained welded and super plastically formed sheets
tigue crack growth threshold are significantly lower, and significantly in order to produce a bigger, and rather monolithic part [85,86] in
higher for friction stir welded Ti-5111 plates. From their study, it can order to meet the buy-to-fly ratio requirements in aerospace industry.
be concluded that with increasing traverse speed, FCP rates are progres- In [87] it has been shown that FSW has produced a lamellar microstruc-
sively higher and FCP threshold values are lower through the weld. Al- ture with small thickness and aspect ratio in Ti-6Al-4V. The lamellar mi-
though, stress relieving omits these differences in the FCP kinetics, yet, crostructure evolved in the weld nugget of Ti-6Al-4V is of 172 nm and
the FCP rates are still higher in comparison with the base materials. Fur- an aspect ratio of 8:1 exhibiting a superior super plasticity at 925 °C
thermore, the fatigue and fracture toughness of 24 mm Ti-6Al-4V has and strain rate 3 × 10−3 s−1. The deformation in the weld nugget is a
been compared with the electron beam welded sheets of same thick- result of grain boundary sliding, and phase boundary sliding between
ness [78]. It has been found that due to slower cooling rates as observed α and β. Further research on work hardening type super plastic defor-
in FSW, a relatively coarser lamellar α transforming from β, has better mation behavior of FSW of Ti-6Al-4V has shown that the super plastic
ductility, fatigue life, fracture toughness, and crack growth resistance tensile elongation of weld nugget is unaffected by the temperature.
in comparison with base metal and electron beam welded joints However, deformation rate has contrary effects on super plasticity of
where rapid cooling results in the evolution of fine acicular α from weld nugget. It has been demonstrated that for lower strain rates the su-
prior β grains [78]. The advances carried out in the field of fracture perplastic elongation was unaffected by the temperature conditions.
and fatigue of FSWed titanium alloys are summarized in Table 3 However, higher superplastic elongation (617%) were obtained at 890
shown below. °C with a higher strain rate of 2.6 × 10−3 s−1.

5.4. Residual stress 6. Texture evolution in FSWed titanium alloys

Residual stresses, depending on the nature; either compressive or The mechanism of superplastic deformation in friction stir welded
tensile, can be detrimental to the fatigue crack growth. A limited study titanium alloys can be extremely intricate. The microstructural gradient,
has been done on the residual stress measurements, and their effect as it develops, from advancing side to retreating side, has a significant
on the fatigue crack growth in friction stir welded titanium alloys [31, impact on the superplastic forming ability of the titanium alloys. The de-
63,80,82–84]. In a study conducted by John et al. [63] for FSWed 6 mm formation modes vary as a function of material composition, deforma-
thick Ti-6Al-4V sheets, compressive residual stresses (500 MPa), and tion, resulting texture, and in application loading direction. Presence
tensile residual stresses (approximately 100 MPa) were observed in of the α, or β stabilizer in a titanium alloys can significantly affect the
the root, and in the center of the weld nugget. Prime et al. [82] have pre- mode of deformation; either prismatic, or basal. For instance, the
sented the contours of residual stresses on the transverse cross section lower aluminum containing materials, prismatic slip dominated the
of the FSWed Ti-6Al-4V sheets. The stresses profiles were similar to tensile deformation. However, the basal slip activity increased with in-
the stress profile as observed in case of the fusion welding than to the creased aluminum content in the alloys. For the Ti-5Al-2.5Sn and Ti-
typical two-peak stress distribution as it has been reported for the 6Al-4V alloys, basal slip was almost as equally active as prismatic slip.
FSW in aluminum alloys [83]. The stir zone or nugget exhibited residual In fact, for some testing conditions of these alloys, such as the 728 K ten-
tensile stress. A single peak of tensile stresses exceeding 400 MPa was sion tests of Ti- 6Al-4V, basal slip was more active than prismatic slip.
observed slightly below the surface corresponding to the shoulder of Basal slip activity was also enhanced by increasing temperature [88].
the tool. The stresses were asymmetric with higher stresses on the ad- Friction stir welding produces a predominant shear texture with the
vancing side of the weld. The high compressive stresses seen at the close-packed directions aligned with the shear direction (SD) and the
edge of the plate on the advancing side were likely anomalous and prob- close-packed plane normal perpendicular to both the SD and the shear
ably arose from machining of the original plates [82]. In a research plane normal [89]. Traces of sheared plastic flow associated with severe
study, conducted by Pasta et al. [80] where they have measured the re- plastic deformation caused by rotating tool in FSW, create crystallo-
sidual stresses in 2 mm thick Ti-6Al-4V sheets by cut compliance meth- graphic arrangements for preferential slip plane (basal (0002) plane)
od, it has been observed that the high tensile stresses (200 MPa) were in such a way that these basal planes surround the tool pin surface in
present in the WN that were balanced by compressive stress in the ad- the WN [90–92]. In the center of the weld, while passing from one loca-
jacent BM. A numerical validation of the residual stress, and consequent tion to another, the pin column surface of the tool is perpendicular to
FCP study has also been performed with the help of AFGROW software. welding direction. Shear plastic flow along the pin column surface
Fatigue tests results showed that for cracks growing perpendicular to breaks the texture of the base material region and produces a texture
the FSW joint, life was longer than for a stress free base metal specimen with the (0001) plane accumulated on the surface perpendicular to
[80]. Recently, Steuwer et al. [84] evaluated the residual stresses in 3- the welding direction in this region [90]. In regions near top and bottom
mm Ti-6Al-4V FSWs as a function of tool travel speed. Increasing weld of the surfaces, however, material is pushed by the compressive forces
traverse speed resulted in narrower and higher residual stress profile. provided by tool shoulder and backing anvil, so that the shear planes
Edwards et al. [31] have studied the surface residual stresses by Hole- in this region are likely to incline from the plane along the surface of
Drilling Strain-Gage Method in 6, 9, and 12 mm thick FSWed Ti-6Al- the pin column to that parallel to the backing plate. Introduction of
4V sheets. They have found that tensile residual stresses were present such shear stress states may produce a texture component with the ac-
in longitudinal direction, and compressive in transverse. Residual stress cumulation in the 〈0002〉 directions near the normal direction in a re-
profiles were similar for 3 mm, 6 mm, and 9 mm welds, exhibiting that gion that in inclined − 65° from the welding direction to the normal
K. Gangwar, M. Ramulu / Materials and Design 141 (2018) 230–255 251

direction [92]. Another trace of shear can be observed in evolving micro- materials such as aluminum and magnesium [100–105], aluminum and
structures. In a study conducted by Prangnell et al. [93] the material in copper [106–110] have been studies for their mechanical and metallur-
the WN was found to have a strong ideal {112} b 110N texture or strong gical properties, but also combinations of harder and softer material are
B=B simple shear component, with the shear plane normal and shear di- also well documented in literature for aluminum and steel [111–116],
rection aligned approximately perpendicular to and tangential with the magnesium and steel [117], aluminum and titanium [118–120], and
flow lines in the nugget. In addition to shearing and corresponding crys- magnesium and titanium [121–123]. Numerous combinations of differ-
tallographic orientation, imparting strains (due to severe plastic defor- ent alloys have been studies for weld soundness and material flow in
mation) and strain rates are far more complicated. The material flow in butt and lap joints [110]. Depending on the materials present on ad-
FSW is continuously accelerated (from RET side), steadied (in the WN), vancing and retreating side, a variety of flow patterns have been ob-
and decelerated (on the ADV side) resulting in a non-uniform velocity served. Vortex-like material flow, sharp interfaces at the weld
field and resultant strain histories. Consequently, the illustration of boundary, hook like flow are to name a few. Presence of micro-cracks,
shear, either by texture development or by microstructural observation, and the formation of intermetallic compounds (IMCs) at the interfaces
is a challenging task. Therefore, in wake of such complex mechanics re- are expected during cooling stages due to variation in the work harden-
searchers and engineers are continuously encouraged to contribute ing exponents of dissimilar alloys [124,125]. A variety of processing con-
more to the field of FSW of materials with higher strain hardening, and ditions with respect to the formation of IMCs and material flow have
with rather complex slip systems as observed in case of hcp materials. been studied and correlated in literature. Mechanical properties of
In α, or near α titanium alloys, slip occurs most readily on prismatic these joints have been characterized in terms of yield strength, ultimate
planes and to a lesser extent on pyramidal or basal planes. In a study con- tensile strength, location of fracture, and percentage elongation. Fur-
ducted by Mironov et al. [94] (CP-Ti, α titanium alloy), and Knipling et al. thermore, effects of IMCs on joint properties have also been discussed
[95] (near α, Ti-5111) on the transverse cross section of the weld, electron in details.
back scattered diffraction (EBSD) revealed the existence of P1(1100)[112 In light of interdependence of multiple parameters involved in fric-
0] hcp ideal shear texture. However, Fonda and Knipling) have observed tion stir welding, several modifications have been suggested to produce
FSWed joints with improved mechanical and metallurgical properties
D1 (112) [111] bcc simple shear texture as well. The presence of bcc tex-
for dissimilar combinations of alloys [113,126–128]. For instance, mod-
ture indicates the material has reached above β transus temperature dur-
ifying heat input by pre-heating the harder plate could improve the duc-
ing welding and that the observed texture was inherited directly from
tility of welded joint. However, thickening of IMCs layer is an impending
that high temperature bcc shear texture as the material cooled through
drawback of pre-heating the harder plate. Therefore, an optimization,
the allotropic transformation. The strength of either of these two textures
among pre-heat temperature, tool rotation speed and corresponding
depends on the processing conditions. High temperature D1 (bcc shear
cooling scheme, is required. Furthermore, it has also been suggested
texture) is dominant at fast feed rates (lower temperature), however
that modifying the distribution of IMCs could also avoid the possibility
lower temperature P1 (hcp shear texture) is strongest at slower feed
of localized fracture in the vicinity of thick IMC layer.
rated (high temperature). They interpreted this apparent contradiction
Modifications in work piece geometry (grooving the harder plate)
to reflect a variation in the stirring efficiency in the wake of the tool, pro-
could potentially enhance the flow of softer material in the weld nugget
posing that fast welding speeds can allow the high temperature deforma-
and improve interlocking abilities of the welded joint [129]. In another
tion structure to be retained in the wake of the tool while slower welding
effort to modify IMC layer, a new technology called friction melt bond-
speeds can enable a more efficient transfer of the stirring deformation and
ing (FMB) has been introduced by researchers. FMB involves melting of
thereby continue to deform the deposited weld as it cools through the β
softer welds where a pinless tool generates enough heat to melt the
transus temperature [95]. It should be noted that while the presence of
softer plate and generate an IMC sealing at the lap joint with harder
P1, hcp shear texture components in a titanium alloy could indicate that
plate on the top [130–133]. The mechanical properties of these welds
the shear deformation was introduced into the material at temperatures
are similar to what can be reached by classical dissimilar FSW. In all
below the β transus, it could also result from an a variant selection as
these modified processes, the welding tool is reduced to a simple cylin-
the material cools through the β transus in the presence of a shear defor-
der facilitating tool manufacturing. Now these processes cannot be
mation field, since the hcp P1 shear texture is identical to one of the Bur-
strictly classified as solid-state processes as here melting of one of the
gers variants of the bcc D1 shear texture [89]. A systematic study of the
two materials to be joined is reached. A study conducted by Aude
texture for β-21 s is performed by Reynolds et al. [66], for Ti-6Al-4V by
Simar et al. [134] presents a detailed overview of challenges and solu-
Pilchak et al., Zhou et al., and Semiatin et al. [53,96,97]. In a recent study
tions for friction stir welding of dissimilar alloys.
conducted by Yoon et al. [98], for the texture development on friction
Despite advancements in FSW of similar and dissimilar alloys, there
stir welded 5 mm thick Ti-6Al-4V, fully equiaxed structure was observed
is a limited literature available for the FSW of dissimilar titanium alloys.
even near the top surface in the 50 rpm joint, while a mixture of the la-
One of the many challenges that are being encountered in FSW of dis-
mellar structure with equiaxed grains was found near the top surface in
similar titanium alloys, is the minimal difference in melting tempera-
the 150 rpm joint. The existence of the lamellar structure resulted in a
ture of titanium alloys. Furthermore, the presence of various alloying
strong [0001] α concentration in this area due to the variant selection
elements and their nature of either being either α, or β stabilizer has a
during the phase transformation [98]. The developing texture has signifi-
major impact on the emerging phases and IMCs in and around the
cant effect on the mechanical properties as well. As observed in the typical
weld nugget. Next section, primarily discusses the challenges associated
stress strain curve, the friction stir welded specimens of CP-Ti fractured in
with friction stir welding of similar titanium alloys with a suggest strat-
the BM [99]. In their studies, it has been found that for tensile strain b0.1%,
egy to tackle emerging problems associated with the friction stir
an extensive distinctive plastic deformation occurred in the WN because
welding of dissimilar titanium alloys.
Schmid factor for prismatic slip in WN is higher than the BM. However,
for strain values N0.1% this difference in the Schmid factor was overcome
7. Challenges and suggested remedies
by the grain size effect, therefore, the local deformation in the WN that
had a finer grain structure restrained, and the fracture occurred in the BM.
By far the scholastic community has been convinced with the notice-
able achievements that friction stir welding can render. With the en-
6.1. FSW of dissimilar alloys hanced technology, automation, suitable tool design, and material
selection, FSW of higher work hardening materials has also been possi-
Over the years, varieties of dissimilar alloys have been joined by fric- ble. Advantages, such as defect free weld, and almost no porosity (due to
tion stir welding under optimized processing conditions. Not only softer solid state joining nature of the FSW) are often challenged by the
252 K. Gangwar, M. Ramulu / Materials and Design 141 (2018) 230–255

flexibility of materials, especially materials with different melting point. For three cases (case 1-high pressure air cooling; case 2-ice-water mix-
In aerospace application, several parts operate at different temperature, ture cooling; case 3-liquid nitrogen cooling) studied by Nan et al. [141]
and under different loading condition. Similarly, in the light of recent for friction stir welded 2 mm CP-Ti sheets, have shown that case 3 result-
development at several automobile industries joining of two different ed in finest grains with abundant twin boundaries (26.8%). An interest-
materials (especially steel and aluminum) can retract a huge payload ing fact to notice in their study is that temperatures generated by
of aluminum. Another challenge that friction stir welding faces, is the adapted processing conditions were below β transus temperature. For
development of highly sensitive microstructure due to complicate ma- case 3, where the primary mechanism of grain refinement during
terial flow, if dissimilar materials are considered. The compatibility of ti- welding is geometric dynamic recrystallization (GDRX), f1012gb1011N
tanium with composite materials is not only alluring but at the same twin also partially contributes to the grain refinement. Mechanical prop-
time joining of composite with titanium by welding appear to be a her- erties and texture evolution are also discussed in that regard. While alu-
culean task. Recently Curtis et al. [135] have developed friction scribe minum climbs on the top of the list to be joined by FSW, high strength
technology to join material with different melting temperature. materials such as steel, titanium, and nickel based alloys have made an
There has not been any published literature describing the FSW in a impact in the transportation, and space industry as well [142].
nonlinear manner. Therefore, the only window that remains operable is In this article, we have presented a brief review of friction stir
the joining in a straight line. In addition to directional limitability, mate- welded titanium alloy with a focus on its escalating use in aerospace
rial composition, presence of α and β stabilizer, β transus temperature, over the last decade (2007–2017). While tool design, tool material,
and the preprocessing conditions, and the position of the alloying sheets and sheet thickness are subjected to change from one application to an-
(either AS, or RS) are few of the many variables that need to be consid- other, processing parameters also need to be optimized in order to
ered before joining of dissimilar materials. In a study performed by Jata achieve a full penetration depth with either load controlled, or position
et al. [136] where they have joined Ti-17, and Ti-6Al-4V, a necklace like controlled FSW process. Material flow, depending on the position of the
structure has emerged in the weld nugget due to different material alloying materials (either AS, or RS) can be significantly different for the
composition. A post annealing of FSWed specimens resulted in a signif- FSW of dissimilar materials. Concurrently, the β transus temperature of
icant improvement in the FCG rates. Another study of FSWed dissimilar the individual titanium alloy ought never be overlooked for tempera-
titanium alloys (Timet-54M, and ATI-425) summarizes the evolution of ture profiles on AS, or RS are quite different. Later on, we have presented
microstructural and mechanical properties in details [137]. Gangwar et the microstructural development in the WN along with the deformation
al. [137] have shown that the presence of alloying sheet (whether on mechanisms in the TMAZ. Based on the processing conditions adopted,
advancing, or retreating side) has significant effect on metallurgical the deformation mechanism can either be sliding (either GBS, or PBS),
mixing. A systematic study conducted by Gonser et al. [138] has docu- or twinning. As a result, grains of different textures are formed with a fa-
mented the mixing pattern, microstructural and mechanical properties vorable slip planes in the loading (and or shearing) direction.
evolution in FSWed Ti-6Al-4V, and Timetal-21S sheets. Mechanical properties of the welded joints have been summarized
These cost sensitive tempests can be subdued by the adoption of nu- in terms of tensile strength, hardness, fatigue, fracture, and residual
merical or analytical approach. Temperature measurement at the root of stress. In general weld nugget showed relatively higher strength. The
the weld is extremely difficult, hence these numerical models can prove global stress strain curves as obtained in the referred studies are a result
themselves to be insightful in determining the true nature of position of deformation of the heterogeneous gauge length containing BM, HAZ,
controlled, and load controlled welding processes. Nonetheless, the TMAZ, WN, TMAZ, HAZ, and BM from AS to RS. Zone wise deformational
evaluation of intrinsic mixing patterns in the weld nugget is still histories recorded by digital image correlation for EBW Ti-6Al-4V have
thwarted by the material composition, and initial microstructure of exhibited an unsymmetrical strain profile with least deformation re-
the BM. One approach that could tackle this issue, would be to couple corded in the weld nugget [58]. Fracture and fatigue crack growth
computational fluid dynamics (CFD), crystal plasticity finite element rates are dependent on the microstructure as developed in the WN. It
method (CP-FEM) along with appropriate temporal conditions that has been shown that widmanstattan morphology developed in the
are responsible for either CDRX, or DRV in the WN, and in the TMAZ weld nugget provides good resistance to FCW rate in comparison to
respectively. equiaxed grains. Residual stresses are tensile in the WN and are com-
Furthermore, an introduction of the vector valued phase field model pressive in the BM. However, post heat treatment can significantly re-
in which phase field model is combined with an orientation field for duce the level of stresses. Unlike FSWed aluminum alloys, only single
representing different crystals or grain orientations, can simulate the peak of tensile stress has been observed.
microstructures at a mesoscale by considering the different thermody- Evolving texture in the weld nugget governs the nature of super
namic driving forces such as bulk and interfacial energy, elastic energy, plastic forming ability of the welded sheets. In presence of two phase
and electric or magnetic energy, and the effect of different transport (α or β) texture can either be bcc or hcp shear dominant or both. Mate-
processes [139]. rial composition has a significant role in determining the nature of slip
and emerging strain rates during friction stir welding. Depending on
8. Summary the values of Schmid factor for corresponding slip planes, assisted by ei-
ther GBS, or PBS the plastic deformation can take place differently. How-
Since 1991, friction stir welding has changed the course of additive ever, Hall-Petch relationship determining the dependence of loading on
manufacturing. Prior to its invention, fusion based additive manufactur- the grain size, can be competitive with the emerging texture.
ing processes, championed by EBW, and laser welding (LW), had the At the end of the article we have suggested that a conglomeration of
ability to produce almost any structural joint required. Unlike fusion several techniques such as CFD (for material flow), CP-FEM (for under-
welding techniques, FSW has flourished as a most viable solid state join- standing the effect of grain boundaries and particle reinforced metallic
ing process that has given a miraculous hope to join materials that con- alloys), and numerical analysis can be useful in extracting valuable in-
tain second phase particles and are highly susceptible to intermetallic formation from the FSW process. The concept of material composition
formation. Strategies like heat treatment, solution hardening, and can be utilized by carefully adjusting the phase field model with aid of
aging have been proposed to improve the work hardening behavior of vector-valued of phase field model, and CALPHAD to predict the grain
titanium alloys. With advancements in additive manufacturing, it is nec- orientation with corresponding elements' migration underlying in the
essary to adopt strategies that can tailor the microstructure from β (at process of FSW. Lastly, as researchers are pushing through various fron-
par β transus temperature) to either β + α (furnace cooling), or α + tiers it becomes of paramount to develop new alloys with long term
α′ (rapid cooling; water quenching) [140]. Another proposed strategy goals. For instance, development of nanostructured β titanium alloys
for modification of mechanical properties, is to induce twin boundaries. with high strength [143], metastable β-type titanium alloys with
K. Gangwar, M. Ramulu / Materials and Design 141 (2018) 230–255 253

ultralow elastic modulus and high strength [1], provide ideas and infor- [15] D.M. Neto, P. Neto, Numerical modeling of friction stir welding process: a literature
review, Int. J. Adv. Manuf. Technol. 65 (1) (2013) 115–126.
mation about how microstructural control can render optimized me- [16] Y.N. Zhang, X. Cao, S. Larose, P. Wanjara, Review of tools for friction stir welding
chanical properties. A noble example of such pursuit is development and processing, Can. Metall. Q. 51 (3) (2012) 250–261.
of metastable nanolaminate steels with bone-like crack resistance [144]. [17] R. Rai, A. De, H.K.D.H. Bhadeshia, T. DebRoy, Review: friction stir welding tools, Sci.
Technol. Weld. Join. 16 (4) (2011) 325–342.
Table 4, represents summary of frictions stir welded titanium alloys. [18] T. Lienert, Microstructure and mechanical properties of friction stir welded titani-
References highlighted with text in the column correspond with re- um alloys, Friction Stir Welding and Processing (2007) 123–154.
search articles that discuss titanium alloys other than Ti-6Al-4V. Due [19] N. Mendes, P. Neto, A. Loureiro, A.P. Moreira, Machines and control systems for
friction stir welding: a review, Mater. Des. 90 (2016) 256–265.
to complexity of microstructures that are subjected to initial base mate-
[20] S. Ji, Z. Li, L. Zhang, Y. Wang, Eliminating the tearing defect in Ti-6Al-4V alloy joint
rial (BM) microstructure, sheet thickness, welding parameters, tool ge- by back heating assisted friction stir welding, Mater. Lett. 188 (2017) 21–24.
ometry; a wide range of microstructures have been obtained at multiple [21] S. Ji, Z. Li, Reducing the hook defect of friction stir lap welded Ti-6Al-4V alloy by
slightly penetrating into the lower sheet, J. Mater. Eng. Perform. 26 (2) (2017)
locations. Therefore, to summarize metallurgical properties we have il-
921–930.
lustrated whether (YES) or not (NO) metallurgical characterization has [22] M.I. Krzywinski, J.E. Schein, I. Birol, J. Connors, R. Gascoyne, D. Horsman, S.J. Jones,
been performed at prescribed locations. ADV, CEN, and RET represent M.A. Marra, Circos: an information aesthetic for comparative genomics, Genome
advancing side, center of the weld, and retreating side respectively. Con- Res. 19 (9) (2009) 1639–1645.
[23] G. Welsch, R. Boyer, E. Collings, Materials Properties Handbook: Titanium Alloys,
clusive remarks have been highlighted for few cases. ASM international, 1993.
Table 5 represents mechanical properties of FSWed titanium alloys. [24] G. Lütjering, J.C. Williams, Titanium, Springer, 2007.
Articles discussing mechanical properties of FSWed titanium alloys, [25] M.J. Donachie, Titanium: A Technical Guide, ASM International, 2000.
[26] C. Leyens, M. Peters, Titanium and Titanium Alloys: Fundamentals and Applica-
other than Ti-6Al-4V, are highlighted along with corresponding refer- tions, John Wiley & Sons, 2003.
ences. Conclusive remarks corresponding to each reference are present- [27] V.A. Joshi, Titanium Alloys: An Atlas of Structures and Fracture Features, Crc Press,
ed at the end of the article. 2006.
[28] T. Massalski, H. Okamoto, P. Subramanian, L. Kacprzak, Binary Alloy Phase
Diagrams, ASM International, Materials Park, OH, 1990.
[29] H. Okamoto, Phase Diagrams for Binary Alloys, 366, ASM International, Materials
Acknowledgement Park, OH, 2000 553.
[30] P. Edwards, M. Ramulu, Identification of process parameters for friction stir
welding Ti–6Al–4V, J. Eng. Mater. Technol. 132 (3) (2010), 031006.
The authors acknowledge the support of many of the former doctor- [31] P. Edwards, M. Ramulu, Surface residual stresses in Ti-6Al-4V friction stir welds:
al, master's degree students at UW and research collaborators in devel- pre-and post-thermal stress relief, J. Mater. Eng. Perform. 24 (9) (2015)
3263–3270.
oping the friction stir welding and superplastic forming of titanium
[32] P. Edwards, M. Ramulu, Peak temperatures during friction stir welding of Ti–6Al–
alloys. Professor Ramulu would like extend special thanks to Dr. Dan 4V, Sci. Technol. Weld. Join. 15 (6) (2010) 468–472.
Sanders and Dr. R. Boyer from Boeing Company; Dr. Paul Edwards from [33] A. Farias, G.F. Batalha, E.F. Prados, R. Magnabosco, S. Delijaicov, Tool wear evalua-
Tesla; Dr. N. Kulkarni from Honeywell; Dr. T. Morton from Blue Origin; tions in friction stir processing of commercial titanium Ti–6Al–4V, Wear 302 (1)
(2013) 1327–1333.
Mr. Kevin Colligan; Mr. A. Cantrell. We would like to thank Professor [34] J. Qian, J. Li, F. Sun, J. Xiong, F. Zhang, X. Lin, An analytical model to optimize rota-
A.P. Reynolds of USC; G. Grant of PNNL and Prof Brian Flinn from depart- tion speed and travel speed of friction stir welding for defect-free joints, Scr. Mater.
ment of materials science and engineering at University of Washington 68 (3) (2013) 175–178.
[35] T. Seidel, A.P. Reynolds, Visualization of the material flow in AA2195 friction-stir
(UW) for their contribution to Solheim Manufacturing Science and welds using a marker insert technique, Metall. Mater. Trans. A 32 (11) (2001)
Technology Laboratory at UW. We, also, sincerely acknowledge the 2879–2884.
Boeing-Pennell Professorship fund and The Boeing Company for [36] K. Colligan, Material flow behavior during friction welding of aluminum, Weld. J.
75 (7) (1999) 229s–237s.
supporting titanium structures manufacturing research. [37] H.B. Schmidt, J.H. Hattel, Thermal modelling of friction stir welding, Scr. Mater. 58
(5) (2008) 332–337.
References [38] Y. Morisada, T. Imaizumi, H. Fujii, Determination of strain rate in friction stir
welding by three-dimensional visualization of material flow using X-ray radiogra-
[1] S. Guo, Q. Meng, X. Zhao, Q. Wei, H. Xu, Design and fabrication of a metastable β- phy, Scr. Mater. 106 (2015) 57–60.
type titanium alloy with ultralow elastic modulus and high strength, Sci. Rep. 5 [39] M. Guerra, C. Schmidt, J.C. McClure, L.E. Murr, A.C. Nunes, Flow patterns during fric-
(2015) 14688. tion stir welding, Mater. Charact. 49 (2) (2002) 95–101.
[2] A. Fattah-alhosseini, M. Vakili-Azghandi, M. Sheikhi, M.K. Keshavarz, Passive and [40] H.N.B. Schmidt, T.L. Dickerson, J.H. Hattel, Material flow in butt friction stir welds in
electrochemical response of friction stir processed pure titanium, J. Alloys AA2024-T3, Acta Mater. 54 (4) (2006) 1199–1209.
Compd. 704 (2017) 499–508. [41] P.D. Edwards, M. Ramulu, Material flow during friction stir welding of Ti-6Al-4V, J.
[3] R.R. Boyer, An overview on the use of titanium in the aerospace industry, Mater. Mater. Process. Technol. 218 (2015) 107–115.
Sci. Eng. A 213 (1) (1996) 103–114. [42] J.-H. Cho, P.R. Dawson, Investigation on texture evolution during friction stir
[4] V. Venkatesh, A.L. Pilchak, J.E. Allison, S. Ankem, R.R. Boyer, J. Christodoulou, H.L. welding of stainless steel, Metall. Mater. Trans. A 37 (4) (2006) 1147–1164.
Fraser, M.A. Imam, Y. Kosaka, H.J. Rack, Proceedings of the 13th World Conference [43] R. Nandan, G. Roy, T. Lienert, T. Debroy, Three-dimensional heat and material flow
on Titanium, John Wiley & Sons, 2016. during friction stir welding of mild steel, Acta Mater. 55 (3) (2007) 883–895.
[5] D. Powers, G. Schubert, Electron beam welding: a useful tool for the automotive in- [44] K. Ishida, Y. Gao, K. Nagatsuka, M. Takahashi, K. Nakata, Microstructures and me-
dustry, Weld. J. 79 (2) (2000) 35–38. chanical properties of friction stir welded lap joints of commercially pure titanium
[6] C. Liu, B. Wu, J.X. Zhang, Numerical investigation of residual stress in thick titanium and 304 stainless steel, J. Alloys Compd. 630 (2015) 172–177.
alloy plate joined with electron beam welding, Metall. Mater. Trans. B Process [45] J. Liao, N. Yamamoto, H. Liu, K. Nakata, Microstructure at friction stir lap joint inter-
Metall. Mater. Process. Sci. 41 (5) (2010) 1129–1138. face of pure titanium and steel, Mater. Lett. 64 (21) (2010) 2317–2320.
[7] P. Wanjara, M. Brochu, M. Jahazi, Thin gauge titanium manufacturing using multi- [46] M. Fazel-Najafabadi, S. Kashani-Bozorg, A. Zarei-Hanzaki, Joining of CP-Ti to 304
ple-pass electron beam welding, Mater. Manuf. Process. 21 (5) (2006) 439–451. stainless steel using friction stir welding technique, Mater. Des. 31 (10) (2010)
[8] O.S. Salih, H. Ou, W. Sun, D.G. McCartney, A review of friction stir welding of alu- 4800–4807.
minium matrix composites, Mater. Des. 86 (2015) 61–71. [47] M.M. Husain, R. Sarkar, T. Pal, N. Prabhu, M. Ghosh, Friction stir welding of steel:
[9] M.-N. Avettand-Fènoël, A. Simar, A review about friction stir welding of metal ma- heat input, microstructure, and mechanical property co-relation, J. Mater. Eng. Per-
trix composites, Mater. Charact. 120 (2016) 1–17. form. 24 (9) (2015) 3673–3683.
[10] M. Sivashanmugam, T. Kumar, S. Ravikumar, V.S. Rao, D. Muruganandam, A review [48] M. Fazel-Najafabadi, S. Kashani-Bozorg, A. Zarei-Hanzaki, Dissimilar lap joining of
on friction stir welding for aluminium alloys, Frontiers in Automobile and Mechan- 304 stainless steel to CP-Ti employing friction stir welding, Mater. Des. 32 (4)
ical Engineering -2010 (2010) 216–221. (2011) 1824–1832.
[11] R.S. Mishra, Z.Y. Ma, Friction stir welding and processing, Mater. Sci. Eng. R. Rep. 50 [49] L. Zhou, H. Liu, Q. Liu, Effect of process parameters on stir zone microstructure in
(1) (2005) 1–78. Ti–6Al–4V friction stir welds, J. Mater. Sci. 45 (1) (2010) 39.
[12] R. Nandan, T. DebRoy, H.K.D.H. Bhadeshia, Recent advances in friction-stir welding [50] Y. Zhang, Y.S. Sato, H. Kokawa, S.H.C. Park, S. Hirano, Stir zone microstructure of
— process, weldment structure and properties, Prog. Mater. Sci. 53 (6) (2008) commercial purity titanium friction stir welded using pcBN tool, Mater. Sci. Eng.
980–1023. A 488 (1) (2008) 25–30.
[13] X. He, F. Gu, A. Ball, A review of numerical analysis of friction stir welding, Prog. [51] A. Pilchak, M. Juhas, J. Williams, Microstructural changes due to friction stir process-
Mater. Sci. 65 (2014) 1–66. ing of investment-cast Ti-6Al-4V, Metall. Mater. Trans. A 38 (2) (2007) 401–408.
[14] H. Schmidt, J. Hattel, J. Wert, An analytical model for the heat generation in friction [52] K. Knipling, R. Fonda, Microstructural evolution in Ti-5111 friction stir welds,
stir welding, Model. Simul. Mater. Sci. Eng. 12 (1) (2004) 143. Metall. Mater. Trans. A 42 (8) (2011) 2312–2322.
254 K. Gangwar, M. Ramulu / Materials and Design 141 (2018) 230–255

[53] A. Pilchak, J. Williams, Microstructure and texture evolution during friction stir [84] A. Steuwer, D. Hattingh, M. James, U. Singh, T. Buslaps, Residual stresses, micro-
processing of fully lamellar Ti-6Al-4V, Metall. Mater. Trans. A 42 (3) (2011) structure and tensile properties in Ti–6Al–4V friction stir welds, Sci. Technol.
773–794. Weld. Join. 17 (7) (2012) 525–533.
[54] J. Su, J. Wang, R.S. Mishra, R. Xu, J.A. Baumann, Microstructure and mechanical [85] D. Sanders, M. Ramulu, E. Klock-McCook, P. Edwards, A.P. Reynolds, T. Trapp, Char-
properties of a friction stir processed Ti–6Al–4V alloy, Mater. Sci. Eng. A 573 acterization of superplastically formed friction stir weld in titanium 6Al-4V: pre-
(2013) 67–74. liminary results, J. Mater. Eng. Perform. 17 (2) (2008) 187–192.
[55] Y. Zhang, Y.S. Sato, H. Kokawa, S.H.C. Park, S. Hirano, Microstructural characteristics [86] D. Sanders, M. Ramulu, P. Edwards, Superplastic forming of friction stir welds in ti-
and mechanical properties of Ti–6Al–4V friction stir welds, Mater. Sci. Eng. A 485 tanium alloy 6Al-4V: preliminary results, Mater. Werkst. 39 (4–5) (2008)
(1) (2008) 448–455. 353–357.
[56] K. Jata, S. Semiatin, Continuous Dynamic Recrystallization during Friction Stir [87] L. Wu, B. Xiao, D. Ni, Z. Ma, X. Li, M. Fu, Y. Zeng, Achieving superior superplasticity
Welding of High Strength Aluminum Alloys, Air Force Research Lab Wright- from lamellar microstructure of a nugget in a friction-stir-welded Ti–6Al–4V joint,
Patterson AFB OH Materials and Manufacturing Directorate, 2000. Scr. Mater. 98 (2015) 44–47.
[57] S. Mironov, K. Inagaki, Y.S. Sato, H. Kokawa, Development of grain structure during [88] H. Li, Analysis of the Deformation Behavior of the Hexagonal Close-Packed Alpha
friction-stir welding of Cu–30Zn brass, Philos. Mag. 94 (27) (2014) 3137–3148. Phase in Titanium and Titanium Alloys, Michigan State University, 2013.
[58] K. Saranath, M. Ramji, Local zone wise elastic and plastic properties of electron [89] R. Fonda, K. Knipling, Texture development in friction stir welds, Sci. Technol.
beam welded Ti–6Al–4V alloy using digital image correlation technique: a com- Weld. Join. 16 (4) (2011) 288–294.
parative study between uniform stress and virtual fields method, Opt. Lasers [90] S.H.C. Park, Y.S. Sato, H. Kokawa, Effect of micro-texture on fracture location in fric-
Eng. 68 (2015) 222–234. tion stir weld of Mg alloy AZ61 during tensile test, Scr. Mater. 49 (2) (2003)
[59] L. Wu, D. Wang, B. Xiao, Z. Ma, Microstructural evolution of the 161–166.
thermomechanically affected zone in a Ti–6Al–4V friction stir welded joint, Scr. [91] W. Woo, H. Choo, D.W. Brown, P.K. Liaw, Z. Feng, Texture variation and its influ-
Mater. 78 (2014) 17–20. ence on the tensile behavior of a friction-stir processed magnesium alloy, Scr.
[60] J. Kim, J. Kim, Y. Lee, C. Park, C. Lee, Microstructural analysis on boundary sliding Mater. 54 (11) (2006) 1859–1864.
and its accommodation mode during superplastic deformation of Ti–6Al–4V [92] S.H.C. Park, Y.S. Sato, H. Kokawa, Basal plane texture and flow pattern in friction stir
alloy, Mater. Sci. Eng. A 263 (2) (1999) 272–280. weld of a magnesium alloy, Metall. Mater. Trans. A 34 (4) (2003) 987–994.
[61] S. Ji, Z. Li, Y. Wang, L. Ma, Joint formation and mechanical properties of back [93] P.B. Prangnell, C.P. Heason, Grain structure formation during friction stir welding
heating assisted friction stir welded Ti–6Al–4V alloy, Mater. Des. 113 (2017) observed by the ‘stop action technique, Acta Mater. 53 (11) (2005) 3179–3192.
37–46. [94] S. Mironov, Y. Sato, H. Kokawa, Development of grain structure during friction stir
[62] R.W. Fonda, K.E. Knipling, A.L. Pilchak, Thermal stir welds in titanium, Metall. welding of pure titanium, Acta Mater. 57 (15) (2009) 4519–4528.
Mater. Trans. A 47 (1) (2016) 360–367. [95] K. Knipling, R. Fonda, Texture development in the stir zone of near-α titanium fric-
[63] R. John, K. Jata, K. Sadananda, Residual stress effects on near-threshold fatigue tion stir welds, Scr. Mater. 60 (12) (2009) 1097–1100.
crack growth in friction stir welds in aerospace alloys, Int. J. Fatigue 25 (9) [96] Z. Li, H.-J. Liu, L.-Z. Wu, Texture of friction stir welded Ti–6Al–4V alloy, Trans. Non-
(2003) 939–948. ferrous Metals Soc. China 24 (2) (2014) 368–372.
[64] R. Fonda, J. Bingert, Texture variations in an aluminum friction stir weld, Scr. Mater. [97] S. Semiatin, K. Kinsel, A. Pilchak, G. Sargent, Effect of process variables on transfor-
57 (11) (2007) 1052–1055. mation-texture development in Ti-6Al-4V sheet following beta heat treatment,
[65] T. Lienert, K. Jata, R. Wheeler, V. Seetharaman, Proc. Joining of Advanced and Spe- Metall. Mater. Trans. A 44 (8) (2013) 3852–3865.
cialty Materials III, 2001. [98] S. Yoon, R. Ueji, H. Fujii, Microstructure and texture distribution of Ti–6Al–4V alloy
[66] A.P. Reynolds, E. Hood, W. Tang, Texture in friction stir welds of Timetal 21S, Scr. joints friction stir welded below β-transus temperature, J. Mater. Process. Technol.
Mater. 52 (6) (2005) 491–494. 229 (2016) 390–397.
[67] W.-B. Lee, C.-Y. Lee, W.-S. Chang, Y.-M. Yeon, S.-B. Jung, Microstructural inves- [99] F. Liu, J. Liao, Y. Gao, K. Nakata, Influence of texture on strain localization in stir
tigation of friction stir welded pure titanium, Mater. Lett. 59 (26) (2005) zone of friction stir welded titanium, J. Alloys Compd. 626 (2015) 304–308.
3315–3318. [100] A. Ikuta, Y. Yin, T. North, Influence of tool thread on mechanical properties of dis-
[68] M.C. Juhas, G.B. Viswanathan, H.L. Fraser, Characterization of Microstructural Evo- similar Al alloy friction stir spot welds, Sci. Technol. Weld. Join. 17 (8) (2012)
lution in a Ti-6Al-4V Friction Stir Weld, Lightweight Alloys for Aerospace Applica- 622–629.
tion, John Wiley & Sons, Inc., 2001 209–217. [101] A. Somasekharan, L. Murr, Microstructures in friction-stir welded dissimilar mag-
[69] Z. Zhang, B. Li, X. Feng, Y. Shen, W. Hu, Friction-stir welding of titanium/aluminum nesium alloys and magnesium alloys to 6061-T6 aluminum alloy, Mater. Charact.
dissimilar alloys: joint configuration design, as-welded interface characteristics 52 (1) (2004) 49–64.
and tensile properties, Proc. Inst. Mech. Eng. B J. Eng. Manuf. 228 (11) (2014) [102] J. Yan, Z. Xu, Z. Li, L. Li, S. Yang, Microstructure characteristics and performance of
1469–1480. dissimilar welds between magnesium alloy and aluminum formed by friction stir-
[70] L. Zhou, H. Liu, Q. Liu, Effect of rotation speed on microstructure and mechanical ring, Scr. Mater. 53 (5) (2005) 585–589.
properties of Ti–6Al–4V friction stir welded joints, Mater. Des. 31 (5) (2010) [103] V. Firouzdor, S. Kou, Al-to-Mg friction stir welding: effect of material position, trav-
2631–2636 (1980–2015). el speed, and rotation speed, Metall. Mater. Trans. A 41 (11) (2010) 2914–2935.
[71] H. Liu, L. Zhou, Q. Liu, Microstructural characteristics and mechanical properties of [104] P. Venkateswaran, A.P. Reynolds, Factors affecting the properties of friction stir
friction stir welded joints of Ti–6Al–4V titanium alloy, Mater. Des. 31 (3) (2010) welds between aluminum and magnesium alloys, Mater. Sci. Eng. A 545 (2012)
1650–1655. 26–37.
[72] D. Sanders, P. Edwards, A. Cantrell, K. Gangwar, M. Ramulu, Friction stir-welded ti- [105] B. Fu, G. Qin, F. Li, X. Meng, J. Zhang, C. Wu, Friction stir welding process of dissim-
tanium alloy Ti-6Al-4V: microstructure, mechanical and fracture properties, JOM ilar metals of 6061-T6 aluminum alloy to AZ31B magnesium alloy, J. Mater. Pro-
67 (5) (2015) 1054–1063. cess. Technol. 218 (2015) 38–47.
[73] P. Edwards, M. Ramulu, Investigation of microstructure, surface and subsurface [106] L. Murr, Y. Li, R. Flores, E.A. Trillo, J. McClure, Intercalation vortices and related mi-
characteristics in titanium alloy friction stir welds of varied thicknesses, Sci. crostructural features in the friction-stir welding of dissimilar metals, Mater. Res.
Technol. Weld. Join. 14 (5) (2009) 476–483. Innov. 2 (3) (1998) 150–163.
[74] P. Edwards, M. Ramulu, Fracture toughness and fatigue crack growth in Ti-6Al-4V [107] P. Xue, D. Ni, D. Wang, B. Xiao, Z. Ma, Effect of friction stir welding parameters on
friction stir welds, Fatigue Fract. Eng. Mater. Struct. 38 (8) (2015) 970–982. the microstructure and mechanical properties of the dissimilar Al–Cu joints, Mater.
[75] P. Edwards, M. Ramulu, Fatigue performance of friction stir welded titanium struc- Sci. Eng. A 528 (13) (2011) 4683–4689.
tural joints, Int. J. Fatigue 70 (2015) 171–177. [108] I. Galvao, J. Oliveira, A. Loureiro, D. Rodrigues, Formation and distribution of brittle
[76] P. Edwards, M. Ramulu, Fatigue performance of friction stir welded Ti–6Al–4V subject- structures in friction stir welding of aluminium and copper: influence of process
ed to various post weld heat treatment temperatures, Int. J. Fatigue 75 (2015) 19–27. parameters, Sci. Technol. Weld. Join. 16 (8) (2011) 681–689.
[77] P. Pao, R. Fonda, H. Jones, C. Feng, D. Moon, Fatigue Crack Growth in Friction Stir [109] J. Ouyang, E. Yarrapareddy, R. Kovacevic, Microstructural evolution in the friction
Welded Ti-5111, Naval Research Lab Washington DC, 2009. stir welded 6061 aluminum alloy (T6-temper condition) to copper, J. Mater. Pro-
[78] P.D. Edwards, M. Ramulu, Comparative study of fatigue and fracture in friction stir cess. Technol. 172 (1) (2006) 110–122.
and electron beam welds of 24 mm thick titanium alloy Ti-6Al-4V, Fatigue Fract. [110] P. Liu, Q. Shi, W. Wang, X. Wang, Z. Zhang, Microstructure and XRD analysis of FSW
Eng. Mater. Struct. 39 (10) (2016) 1226–1240. joints for copper T2/aluminium 5A06 dissimilar materials, Mater. Lett. 62 (25)
[79] D. Sanders, M. Ramulu, P. Edwards, A. Cantrell, Effects on the surface texture, su- (2008) 4106–4108.
perplastic forming, and fatigue performance of titanium 6AL-4V friction stir [111] R.S. Coelho, A. Kostka, S. Sheikhi, J. dos Santos, A.R. Pyzalla, microstructure and
welds, J. Mater. Eng. Perform. 19 (4) (2010) 503–509. mechanical properties of an AA6181-T4 aluminium alloy to HC340LA high
[80] S. Pasta, A.P. Reynolds, Residual stress effects on fatigue crack growth in a Ti-6Al- strength steel friction stir overlap weld, Adv. Eng. Mater. 10 (10) (2008)
4V friction stir weld, Fatigue Fract. Eng. Mater. Struct. 31 (7) (2008) 569–580. 961–972.
[81] M. Muzvidziwa, M. Okazaki, K. Suzuki, S. Hirano, Role of microstructure on the fa- [112] R. Coelho, A. Kostka, J. Dos Santos, A. Kaysser-Pyzalla, Friction-stir dissimilar
tigue crack propagation behavior of a friction stir welded Ti–6Al–4V, Mater. Sci. welding of aluminium alloy to high strength steels: mechanical properties and
Eng. A 652 (2016) 59–68. their relation to microstructure, Mater. Sci. Eng. A 556 (2012) 175–183.
[82] M. Prime, T. Lienert, W. Stellwag Jr., R. Casey, A. Residual, Stress measurement in [113] H. Bang, H. Bang, G. Jeon, I. Oh, C. Ro, Gas tungsten arc welding assisted hybrid fric-
friction stir welded Ti-6Al-4V, Fabtech International and AWS Welding Show tion stir welding of dissimilar materials Al6061-T6 aluminum alloy and STS304
2005, November 13 (16) (2005). stainless steel, Mater. Des. 37 (2012) 48–55.
[83] E. Lima, J. Wegener, C.D. Donne, G. Goerigk, T. Wroblewski, T. Buslaps, A.R. Pyzalla, [114] M. Dehghani, A. Amadeh, S.A. Mousavi, Investigations on the effects of friction stir
W. Reimers, Dependence of the microstructure, residual stresses and texture of AA welding parameters on intermetallic and defect formation in joining aluminum
6013 friction stir welds on the welding process, Z. Met. 94 (8) (2003) 908–915. alloy to mild steel, Mater. Des. 49 (2013) 433–441.
K. Gangwar, M. Ramulu / Materials and Design 141 (2018) 230–255 255

[115] H.A. Derazkola, H.J. Aval, M. Elyasi, Analysis of process parameters effects on dis- [144] M. Koyama, Z. Zhang, M. Wang, D. Ponge, D. Raabe, K. Tsuzaki, H. Noguchi, C.C.
similar friction stir welding of AA1100 and A441 AISI steel, Sci. Technol. Weld. Tasan, Bone-like crack resistance in hierarchical metastable nanolaminate steels,
Join. 20 (7) (2015) 553–562. Science 355 (6329) (2017) 1055–1057.
[116] K. Ramachandran, N. Murugan, S.S. Kumar, Effect of tool axis offset and geometry [145] P.M. Mashinini, D.G. Hattingh, H. Lombard, Mechanical properties and microstruc-
of tool pin profile on the characteristics of friction stir welded dissimilar joints of ture of friction stir and laser beam welded 3 mm Ti6Al4V alloy, Proceedings of the
aluminum alloy AA5052 and HSLA steel, Mater. Sci. Eng. A 639 (2015) 219–233. World Congress on Engineering 2 (2016) 1140–1145.
[117] H. Kasai, Y. Morisada, H. Fujii, Dissimilar FSW of immiscible materials: steel/mag- [146] A. Fall, M.H. Fesharaki, A.R. Khodabandeh, M. Jahazi, Tool wear characteristics and
nesium, Mater. Sci. Eng. A 624 (2015) 250–255. effect on microstructure in Ti-6Al-4V friction stir welded joints, Metals 6 (11)
[118] U. Dressler, G. Biallas, U.A. Mercado, Friction stir welding of titanium alloy TiAl6V4 (2016) 275.
to aluminium alloy AA2024-T3, Mater. Sci. Eng. A 526 (1) (2009) 113–117. [147] L. Wu, P. Xue, B. Xiao, Z. Ma, Achieving superior low-temperature superplasticity
[119] B. Li, Z. Zhang, Y. Shen, W. Hu, L. Luo, Dissimilar friction stir welding of Ti–6Al–4V for lamellar microstructure in nugget of a friction stir welded Ti-6Al-4V joint, Scr.
alloy and aluminum alloy employing a modified butt joint configuration: influ- Mater. 122 (2016) 26–30.
ences of process variables on the weld interfaces and tensile properties, Mater. [148] M.J. Fu, X.H. Li, X.Q. Han, H.Y. Xu, Research on Work Hardening Type Superplastic
Des. 53 (2014) 838–848. Deformation Behavior of FSW Titanium Alloy, Materials Science Forum, Trans Tech
[120] Z. Song, K. Nakata, A. Wu, J. Liao, L. Zhou, Influence of probe offset distance on in- Publ, 2016 506–511.
terfacial microstructure and mechanical properties of friction stir butt welded joint [149] S. Nimer, J. Wolk, M. Zupan, Local property characterization of friction stir welded
of Ti6Al4V and A6061 dissimilar alloys, Mater. Des. 57 (2014) 269–278. Ti-5111: transverse orientation measurements, Acta Mater. 61 (8) (2013)
[121] M. Aonuma, K. Nakata, Dissimilar metal joining of ZK60 magnesium alloy and tita- 3050–3059.
nium by friction stir welding, Mater. Sci. Eng. B 177 (7) (2012) 543–548. [150] S. Yoon, R. Ueji, H. Fujii, Effect of initial microstructure on Ti–6Al–4V joint by fric-
[122] M. Aonuma, K. Nakata, Effect of alloying elements on interface microstructure of tion stir welding, Mater. Des. 88 (2015) 1269–1276.
Mg–Al–Zn magnesium alloys and titanium joint by friction stir welding, Mater. [151] S. Yoon, R. Ueji, H. Fujii, Effect of rotation rate on microstructure and texture evo-
Sci. Eng. B 161 (1) (2009) 46–49. lution during friction stir welding of Ti–6Al–4V plates, Mater. Charact. 106 (2015)
[123] M. Aonuma, K. Nakata, Effect of calcium on intermetallic compound layer at inter- 352–358.
face of calcium added magnesium–aluminum alloy and titanium joint by friction [152] B. Li, Y. Shen, L. Luo, W. Hu, Z. Zhang, Surface aluminizing on Ti–6Al–4V alloy via a
stir welding, Mater. Sci. Eng. B 173 (1) (2010) 135–138. novel multi-pass friction-stir lap welding method: preparation process, oxidation
[124] A. Esmaeili, M.B. Givi, H.Z. Rajani, A metallurgical and mechanical study on dissim- behavior and interlayer evolution, Mater. Des. 49 (2013) 647–656.
ilar friction stir welding of aluminum 1050 to brass (CuZn30), Mater. Sci. Eng. A [153] J.C. Lippold, J.J. Livingston, Microstructure evolution during friction stir processing
528 (22) (2011) 7093–7102. and hot torsion simulation of Ti-6Al-4V, Metall. Mater. Trans. A 44 (8) (2013)
[125] H. Uzun, C. Dalle Donne, A. Argagnotto, T. Ghidini, C. Gambaro, Friction stir welding 3815–3825.
of dissimilar Al 6013-T4 to X5CrNi18-10 stainless steel, Mater. Des. 26 (1) (2005) [154] M. Esmaily, S.N. Mortazavi, P. Todehfalah, M. Rashidi, Microstructural characteriza-
41–46. tion and formation of α′ martensite phase in Ti–6Al–4V alloy butt joints produced
[126] X. Fei, X. Jin, Y. Ye, T. Xiu, H. Yang, Effect of pre-hole offset on the property of the by friction stir and gas tungsten arc welding processes, Mater. Des. 47 (2013)
joint during laser-assisted friction stir welding of dissimilar metals steel and alumi- 143–150.
num alloys, Mater. Sci. Eng. A 653 (2016) 43–52. [155] G. Buffa, L. Fratini, M. Schneider, M. Merklein, Micro and macro mechanical charac-
[127] D. Yaduwanshi, S. Bag, S. Pal, Numerical modeling and experimental investigation terization of friction stir welded Ti–6Al–4V lap joints through experiments and nu-
on plasma-assisted hybrid friction stir welding of dissimilar materials, Mater. Des. merical simulation, J. Mater. Process. Technol. 213 (12) (2013) 2312–2322.
92 (2016) 166–183. [156] J. Wang, J. Su, R.S. Mishra, R. Xu, J.A. Baumann, Tool wear mechanisms in friction
[128] X. Liu, S. Lan, J. Ni, Electrically assisted friction stir welding for joining Al 6061 to stir welding of Ti–6Al–4V alloy, Wear 321 (2014) 25–32.
TRIP 780 steel, J. Mater. Process. Technol. 219 (2015) 112–123. [157] K. Kitamura, H. Fujii, Y. Iwata, Y. Sun, Y. Morisada, Flexible control of the micro-
[129] T. Ogura, T. Nishida, Y. Tanaka, H. Nishida, S. Yoshikawa, M. Fujimoto, A. Hirose, Mi- structure and mechanical properties of friction stir welded Ti–6Al–4V joints,
croscale evaluation of mechanical properties of friction stir welded A6061 alumin- Mater. Des. 46 (2013) 348–354.
ium alloy/304 stainless steel dissimilar lap joint, Sci. Technol. Weld. Join. 18 (2) [158] J. Wang, J. Su, R.S. Mishra, R. Xu, J.A. Baumann, A preliminary study of deformation
(2013) 108–113. behavior of friction stir welded Ti-6Al-4V, J. Mater. Eng. Perform. 23 (8) (2014)
[130] B. Kuang, Y. Shen, W. Chen, X. Yao, H. Xu, J. Gao, J. Zhang, The dissimilar friction stir 3027–3033.
lap welding of 1A99 Al to pure Cu using Zn as filler metal with “pinless” tool con- [159] A. Nasresfahani, A. Soltanipur, K. Farmanesh, A. Ghasemi, Effects of tool wear on
figuration, Mater. Des. 68 (2015) 54–62. friction stir welded joints of Ti–6Al–4V alloy, Mater. Sci. Technol. 33 (5) (2017)
[131] C. van der Rest, P.J. Jacques, A. Simar, On the joining of steel and aluminium by 583–591.
means of a new friction melt bonding process, Scr. Mater. 77 (2014) 25–28. [160] R. Francis, R. Francis, J. Newkirk, J. Newkirk, F. Liou, F. Liou, Investigation of forged-
[132] S. Crucifix, C.V.D. Rest, N. Jimenez-Mena, P. Jacques, A. Simar, Modelling thermal like microstructure produced by a hybrid manufacturing process, Rapid Prototyp. J.
cycles and intermetallic growth during friction melt bonding of ULC steel to alu- 22 (4) (2016) 717–726.
minium alloy 2024-T3, Sci. Technol. Weld. Join. 20 (4) (2015) 319–324. [161] S.Y. Mironov, Evolution of the microstructure of a VT6 alloy during friction stir
[133] G. Zhang, W. Su, J. Zhang, Z. Wei, Friction stir brazing: a novel process for fabricat- welding, Russian Metallurgy (Metally) 2015 (4) (2015) 301–308.
ing Al/steel layered composite and for dissimilar joining of Al to steel, Metall. [162] X. Jiang, B. Wynne, J. Martin, Microstructure and texture evolution of stationary
Mater. Trans. A 42 (9) (2011) 2850–2861. shoulder friction stir welded Ti6Al4V alloy, Sci. Technol. Weld. Join. 20 (7)
[134] A. Simar, M.-N. Avettand-Fènoël, State of the art about dissimilar metal friction stir (2015) 594–600.
welding, Sci. Technol. Weld. Join. 22 (5) (2017) 389–403. [163] P.M. Mashinini, D. Hattingh, H. Lombard, Influence of process heat input on static
[135] T. Curtis, C. Widener, M. West, B. Jasthi, Y. Hovanski, B. Carlson, R. Szymanski, W. and dynamic properties of friction stir welded 3 mm Ti6Al4V alloy, Adv. Mater.
Bane, Friction Stir Scribe Welding of Dissimilar Aluminum to Steel Lap Joints, Fric- Res. 1019 (2014).
tion Stir Welding and Processing VIII, Springer, 2015 163–169. [164] G. Buffa, A. Ducato, L. Fratini, FEM based prediction of phase transformations dur-
[136] K. Jata, P. Subramanian, A. Reynolds, T. Trapp, E. Helder, Friction Stir Welding of Ti- ing friction stir welding of Ti6Al4V titanium alloy, Mater. Sci. Eng. A 581 (2013)
tanium Alloys for Aerospace Applications: Microstructure and Mechanical Behav- 56–65.
ior, The Fourteenth International Offshore and Polar Engineering Conference, [165] S. Nimer, J. Wolk, M. Zupan, Location and orientation specific material property
International Society of Offshore and Polar Engineers, 2004. evaluation of friction stir welded Ti-5111: a microsample approach, Adv. Eng.
[137] K. Gangwar, M. Ramulu, A. Cantrell, D.G. Sanders, Microstructure and mechanical Mater. 16 (4) (2014) 452–458.
properties of friction stir welded dissimilar titanium alloys: TIMET-54M and ATI- [166] N. Buhl, G. Wagner, D. Eifler, M. Gutensohn, F. Zillekens, Microstructural and Me-
425, Metals 6 (10) (2016) 252. chanical Investigations of Friction Stir Welded TI/TI- and TI-Alloy/TI-Alloy-Joints,
[138] M.J. Gonser, Microstructure Evolution and Material Flow Behavior in Friction-Stir Friction Stir Welding and Processing VII, Springer, 2013 141–149.
Welded Dissimilar Titanium Alloys, The Ohio State University, 2010. [167] M. Ramulu, T. Greenwell, P. Labossiere, Full-Field Strain Behavior of Friction Stir-
[139] N. Moelans, B. Blanpain, P. Wollants, An introduction to phase-field modeling of Welded Titanium Alloy by Digital Image Correlation, Applied Mechanics and Mate-
microstructure evolution, Calphad 32 (2) (2008) 268–294. rials, Trans Tech Publ, 2014 490–496.
[140] C. de Formanoir, A. Brulard, S. Vivès, G. Martin, F. Prima, S. Michotte, E. Rivière, A. [168] E. Lertora, C. Mandolfino, C. Gámbaro, Ti 6AI-4V FSW weldability: mechanical char-
Dolimont, S. Godet, A strategy to improve the work-hardening behavior of Ti– acterization and fatigue life analysis, Key Eng. Mater. 611-612 (2014) 1476–1483.
6Al–4V parts produced by additive manufacturing, Math. Res. Lett. 5 (3) (2017) [169] L. Paola, C. Emanuela, Friction stir welding of Ti-6AI-4V alloy, Mater. Sci. Forum 783
201–208. (2014) 574–579.
[141] N. Xu, Q. Song, Y. Bao, Y. Jiang, J. Shen, X. Cao, Twinning-induced mechanical prop- [170] S. Ji, B. Zhuo, S. Gao, Y. Huang, L. Zhang, J. Li, Y. Ma, The microstructure and me-
erties' modification of CP-Ti by friction stir welding associated with simultaneous chanical properties of friction stir welded Ti6Al4V titanium alloy under β transus
backward cooling, Sci. Technol. Weld. Join. (2017) 1–7. temperature, Eng. Rev. 35 (1) (2015) 27–31.
[142] W. Brassington, P.A. Colegrove, Alternative friction stir welding technology for tita- [171] K. Song, Z. Dong, K. Fang, X. Zhan, Y. Wei, Cellular automaton modelling of dynamic
nium–6Al–4V propellant tanks within the space industry, Sci. Technol. Weld. Join. recrystallisation microstructure evolution during friction stir welding of titanium
22 (4) (2017) 300–318. alloy, Mater. Sci. Technol. 30 (6) (2014) 700–711.
[143] A. Devaraj, V.V. Joshi, A. Srivastava, S. Manandhar, V. Moxson, V.A. Duz, C. Lavender,
A low-cost hierarchical nanostructured beta-titanium alloy with high strength, 7
(2016), 11176.

Vous aimerez peut-être aussi