CIRP Annals - Manufacturing Technology 59 (2010) 271–274

Contents lists available at ScienceDirect

CIRP Annals - Manufacturing Technology

jou rnal homep age : ht t p: // ees .e lse vi er. com/ci rp/ def a ult . asp

A new fixture for FSW processes of titanium alloys

L. Fratini (2)*, F. Micari (1), G. Buffa, V.F. Ruisi
Dipartimento di Tecnologia Meccanica, Produzione e Ing. Gestionale, Università di Palermo, Viale delle Scienze, 90128 Palermo, Italy

A R T I C L E I N F O A B S T R A C T

Keywords: FSW of titanium alloys is nowadays one of the most challenging welding operations, even with a solid
Friction Stir Welding
state process, due to the thermo-mechanical and thermo-chemical characteristics of such materials. Due
Titanium
to the relevant application of titanium alloys in the aeronautic and aerospace industries, in the recent
Microstructure
years few attempts were carried out to develop FSW processes aimed to maximize the mechanical
performances of the welded parts. In the paper a new fixture is presented allowing obtaining effective
FSW joints of titanium blanks, which were investigated through mechanical and metallurgical tests
highlighting the peculiarities of FSW of titanium alloys.
ß 2010 CIRP.

1. Introduction papers can be found in literature on FSW of titanium alloys. Lee et al.
[13] and Zhang et al. [14], for instance, studied the mechanical and
Nowadays Friction Stir Welding (FSW) [1–3] can be considered metallurgical properties of FSW joints obtained from commercially
mature for butt joints configurations and aluminum alloys as a large pure titanium sheets of 5.6 and 3 mm, respectively, finding the
number of publications can be found in literature focusing on feasibility of the process for such materials and pointing out the
different aspects of the process itself such as microstructural issues, main differences with the already known microstructure evolution
process parameters influence, joints fatigue life and material flow phenomena occurring in FSW of aluminum alloys. Pasta and
analysis. Currently two main front end topics are capturing the Reynolds [11] investigated the residual stress effects on fatigue
interest of researchers around the world, namely the development of crack growth also using numerical simulation. The paper is focused
complex geometry joints [4,5] and welding of high strength on Ti–6Al–4V titanium alloy, which is by far the most commonly
materials as titanium alloys [6,7]. Regarding the latter issue, the used Ti alloy for industrial applications; the numerical model,
choice is primarily triggered by the needs of industry [8], demanding limited to the prediction of the crack growth rate, showed a good
very high strength, low weight and corrosion resistant materials for matching with experimental results.
aeronautical, aerospace, nautical and nuclear applications. In the present paper a dedicated experimental fixture able to
Overall welding of titanium alloys is a challenging process due to overcome the shortcomings of FSW of titanium alloys is designed
the chemical, mechanical and thermal characteristics of such and proposed with particular attention to the choice of the
materials. First of all the latter are subjected to atmosphere materials and to the cooling systems, both under the backplate and
contamination resulting in joint hydrogen, oxygen and nitrogen around the tool. FSW joints were then developed out of Ti–6Al–4V
embrittlement; furthermore, due to the high melting temperature, sheets and tested in order to verify the mechanical performances.
large distortion and residual stress are found in joints obtained by Finally micro- and macro-analyses permitted to highlight the final
traditional fusion welding processes as gas metal arc welding [9], microstructure and the metallurgical phenomena occurring during
electron beam welding and laser welding [10]. In this way a solid the process.
state process, as FSW, represents a valid choice in order to overcome
problems related to the material melting. It should be noticed that 2. Proposed fixture an experimental details
FSW of titanium alloys is definitely more complex than the same
process referred to aluminum alloys. In fact, as a consequence of the As briefly outlined in the previous paragraph, FSW of titanium
large reached temperatures and reacting forces on the pin, the choice alloys requires a more careful design of both the clamping fixture
of tool material is limited to extremely high strength refractory and the tooling with respect to FSW of aluminum alloys. As far as
materials as WC alloys, Re and Mo based alloys and PcBNs [6,11,12]. the clamping fixture design is regarded a first problem to be
What is more, a proper cooling system must be used in order to overcome is due to the high temperatures reached during the
prevent failures both in the tool and in the anvil of the utilized process; under such extreme conditions, the welded blanks are
fixture. Finally a gas shield must be used due to the reactivity of the likely to remain stuck to the backplate compromising both the
considered alloys with air at the process temperatures. Just few soundness of the joint and the integrity of the fixture itself. A
cooling circuit made by three circular channels – 16 mm in
diameter each – drilled in the backplate all along the welding
* Corresponding author. direction resulted as an effective solution in order to avoid such

0007-8506/$ – see front matter ß 2010 CIRP.

doi:10.1016/j.cirp.2010.03.003
272 L. Fratini et al. / CIRP Annals - Manufacturing Technology 59 (2010) 271–274

temperatures that can be reached especially at the base of the tool

pin, at the tool shoulder–workpiece interface. In the present
application a tungsten carbide tool with a 16 mm shoulder and a
308 conical pin, 2.6 mm in height and 5 mm in major diameter, was
utilized. Furthermore, a cooling system able to subtract the excess
of heat from the welding was designed and developed. Next Fig. 2
shows a sketch of the developed tool set made by a fixed external
collar featuring water admission and discharge holes, together
with a sealing O-ring in order to avoid leakage on the weld. The
latter event would be highly detrimental for the weld quality due
to the elevated reactivity of titanium with oxygen and hydrogen,
especially at the process temperatures. Finally inside the collar the
tool, cooled by the water flow, is free to rotate in order to generate
the frictional heat needed for the process.
It should be noticed that the above described features of the
fixture derive from progressive design refinement carried out on
the basis of the ongoing experimental activity. As far as the
developed experiments are regarded, 3 mm thick Ti–6Al–4V
titanium alloy sheets 200 mm  200 mm in dimensions were
welded together under different process conditions. In particular
rotating speeds of 300, 500, 700 and 1000 rpm were selected. Fixed
advancing speed equal to 50 mm/min, nuting angle equal to 28 and
tool shoulder sinking of 0.2 mm were considered for all the welds.
Fig. 1. Sketch of the developed fixture for FSW of titanium alloys. A tungsten carbide tool with a 16 mm shoulder and a 308 conical
pin, 2.6 mm in height and 5 mm in major diameter, was utilized.
Both the backplate and the tool were cooled by a 2 l/min flow of
water. Temperatures were measured during the welding first of all
side effects (see Fig. 1 also for the position of the channels). In this by a thermocouple placed between the two sheets, at 1.5 mm from
way the downward thermal flow through the titanium sheets the bottom of the joints, at a distance of 30 mm from the plunging
thickness was also improved with further beneficial effects for the of the tool. In this way temperatures can be measured in the center
joint quality. Nevertheless, detrimental effects on the overall weld of the weld till the tool shoulder contacts and destroys the
integrity may come from carbon contamination of titanium. A thermocouple itself. What is more two more thermocouples were
30 mm wide tungsten insert was then introduced in a pocket placed below the tungsten insert and between the insert and the
milled in the center of the backplate in order to assure the contact blanks to be welded, respectively, just at the middle of the joint, i.e.
between tungsten and titanium at least close to the welding line, half away from the welding start. All the welds were protected
i.e. where temperature levels may lead to contamination problems. from atmospheric contaminants by a shield of argon inert gas. Each
In Fig. 1 a sketch of the developed fixture, showing the relative experiment was repeated 5 times and specimens were cut by EDM
positioning between sheets to be welded, tungsten insert and from the obtained joints for tensile tests and macro- and micro-
cooling system in the backplate, is presented. observations. In particular the specimens were embedded by hot
The tool represents a further decisive key factor for the process; compression mounting, polished and finally etched with Keller
shoulder and pin wear and breakage set a severe limitation for the reagent before being observed under a light microscope for optical
tool material choice. As a matter of fact the tool must be able to characterization.
maintain high mechanical resistance as well as sufficient wear and
oxidation resistance at the temperatures reached during the 3. Obtained results
process. It should be observed that in FSW processes temperature
typically reaches about 80% of the processed material melting First of all, tensile tests were carried out on the cut specimen. In
temperature; however, temperature in the tool is always larger the following Fig. 3 the average results in terms of percentage of
than maximum values reached in the sheets because of the less the parent material ultimate tensile stress (UTS) are reported. As it
favorable thermal exchange conditions. can be seen from the figure, all the investigated rotational speed
Based on the above considerations WC alloys, Re and Mo based values allow obtaining a resistance higher than 70% of the base
alloys and PcBNs appears at the moment as the only possible material. However, when utilizing a 300 rpm speed, an incomplete
choice although not even such high refractory materials can assure filling at the base of the weld, close to the pin root, is found
the required mechanical properties over the extremely high resulting in the so-called ‘‘root defect’’. Such defect is typically due

Fig. 3. Tensile tests results in terms of percentage with respect to the parent
Fig. 2. Sketch of the utilized tool set. material UTS.
 L. Fratini et al. / CIRP Annals - Manufacturing Technology 59 (2010) 271–274 273

Fig. 4. Transverse section of a welded joint – 700 rpm case study.

to an insufficient thermal flux through the joint. At the increase

of the rotational speed, i.e. at the increase of the specific thermal
contribution (STC) [3] conferred to the joint, UTS% increases till a
quite satisfying average value of 87% obtained with 700 rpm. It
has to be noticed that with a further STC increase the joint
resistance decreases due to the occurrence of local micro-fusion
phenomena. Fig. 6. Temperature histories in the middle of the joint at the top and the bottom of
A macro-image of the transverse section of the joint obtained the tungsten insert.
with rotational speed of 700 rpm is shown in Fig. 4. An interesting
observation can immediately be made: FSW of titanium alloys
results in a microstructure that is significantly different form the
one observed for aluminum alloys. In particular, no thermo- mances. In Fig. 5 micro-images of the SZ, together with maximum
mechanically affected zone (TMAZ) [6] and no nugget is observed temperature measured along the welding line and average a + b
[1–3], but just two different areas can be identified (apart from the grain equivalent diameter (DAVG), are shown for the developed case
parent material), namely a stirred zone (SZ) and a heat affected studies.
zone (HAZ). A sharp transition is observed between the two latter In particular, a duplex structure is obtained for the 300 rpm case
zones as the deformed grains, typically found in the TMAZ of study (Fig. 5a), indicating that temperature maintained below the
aluminum alloys FSW joints, are replaced by new transformed b-transus temperature. Nevertheless temperature was high
grains, either of the SZ or of the HAZ, due to the thermal cycle the enough to permit a dynamic recrystallization phenomenon, due
material undergoes during the process. Based on the above to the tool stirring action, resulting in a smaller grain dimension
observations, it immediately descends that the stir zone and in with respect to the parent material. When rotational speed is equal
particular its metallurgical properties have a dramatic influence on to 500 rpm (Fig. 5b) an increase in the average grain dimension is
the soundness of the obtained joints. As known the utilized alloy, observed and few b grains with a + b lamellar structure are found.
i.e. Ti–6Al–4V, is a dual phase a + b alloy in which b-transus is A fully lamellar microstructure is visible in Fig. 5c, corresponding
about 1000 8C [7]. to a rotational velocity of 700 rpm. In this case the b-transus
Depending on the temperature levels reached in the SZ a temperature was reached and a further increase in the average
different final microstructure can be observed. In particular, if the grain dimension is observed. Finally temperature values signifi-
selected process parameters result in temperatures in excess of the cantly larger than the ones corresponding to the b-transus
b-transus temperature, a lamellar a + b microstructure is found; it temperature are reached in the 1000 rpm case study (Fig. 5d). In
should be observed that the dimension of the b grains containing such conditions fully lamellar and larger grain dimension are
the lamellae increases at the increasing of the conferred STC. On detected implying that the entire deformation cycle induced by the
the other hand, if temperature levels are below the b-transus mechanical action of the tool took place above the b-transus
temperature, a duplex microstructure is obtained characterized by temperature; in other words, at the end of the stirring action, the
small equiaxed a grains and a + b lamellae inside b grains. The material begins the cool down phase experiencing the b ! a + b
wide range of rotational velocities utilized in this work (300– phase transformation. This eventually results in fully lamellar
1000 rpm) permitted to obtain significantly different macrostruc- structure and grain growth.
tures within the SZ, both in terms of grains dimensions and Further information can be derived from Fig. 6 where the
morphology, dramatically affecting the joints mechanical perfor- temperature histories measured by the thermocouples placed at
the middle of the joint between the blanks and tungsten insert (i.e.
at the top of the insert) and at the interface between the latter and
the steel backplate (i.e. at the bottom of the insert), are shown. It
should be observed the effect of the thermal barrier given by the
tungsten insert limiting the temperature in the backplate. Fig. 7, in

Fig. 5. 250 magnification of the stir zone for the developed weds: (a) 300 rpm, (b)
500 rpm, (c) 700 rpm and (d) 1000 rpm. Fig. 7. Microhardness profiles at 1.5 mm form the bottom of the joints.
274 L. Fratini et al. / CIRP Annals - Manufacturing Technology 59 (2010) 271–274

turn, shows the microhardness profiles measured at 1.5 mm from  As in the FSW of aluminum alloys, the mechanical performances
the bottom of the joints, i.e. at the middle of the sheet thickness, for of the titanium joints strictly depend on the metallurgical
all the considered case studies and the parent material. evolutions induced by the process. Actually, the b ! a + b phase
Overall microhardness increases at the decreasing of the transition of the considered Ti–6Al–4V alloy strongly determines
average grain size; additionally, larger values are found with the final material microstructure at least in the SZ and, as a
lamellar microstructure. The latter considerations explain why consequence, the joint mechanical strength.
larger values with respect of the base material are found at the
center of all the considered welds while the minimum values, even Finally, the already carried out research activities definitely
smaller than the base materials, are found in the HAZ. The highlight very interesting perspectives of further research and
maximum hardness values decrease at the decreasing of the industrial applications regarding FSW operations of titanium
average grain size and at the increasing of the rotational speed, alloys and other high strength materials.
which, as previously observed, strictly determines the conferred
STC. It should be noticed that the extension of the high
microhardness areas increases at the increasing of the rotational References
speed and that can explain the peak of mechanical resistance
observed for the 700 rpm case study, where an effective [1] Mishra RS, Ma ZY (2005) Friction Stir Welding and Processing. Materials Science
compromise is found between microhardness peak value and and Engineering R Reports 50(1–2):1–78.
[2] Nandan R, DebRoy T, Bhadeshia HKDH (2008) Recent Advances in Friction-stir
high quality area extension. Finally it should be observed that Welding—Process, Weldment Structure and Properties. Progress in Materials
microhardness profiles observed in Fig. 6 are extremely different Science 53(6):980–1023.
from the ones known for aluminum alloys, where, as far as [3] Buffa G, Fratini L, Pasta S, Shivpuri R (2008) On the Thermo-mechanical Loads
and the Resultant Residual Stresses in Friction Stir Processing Operations. CIRP
precipitation hardening alloys are regarded, a strong material
Annals 57(1):287–290.
softening is observed all along the joint transverse section due to [4] Buffa G, Fratini L, Hua J, Shivpuri R (2006) Friction Stir Welding of Tailored
the density reduction of the precipitates. The latter phenomenon is Blanks: Investigation on Process Feasibility. CIRP 55(1):279–282.
[5] Fratini L, Buffa G, Shivpuri R (2009) Influence of Material Characteristics on
determined by the heat flux generated during the process resulting
Plastomechanics of the FSW Process for T-joints. Materials & Design
in local temperature levels larger than the solubilization thresh- 30(7):2435–2445.
olds of the precipitates. Such effect is partially limited in the nugget [6] Zhang Y, Sato YS, Kokawa H, Park SHC, Hirano S (2008) Microstructural
due to the reduced grain size influence [3]. Characteristics and Mechanical Properties of Ti–6Al–4V Friction Stir Welds.
Materials Science and Engineering A 485(1–2):448–455.
[7] Pilchak AL, Juhas MC, Williams JC (2007) Microstructural Changes Due to
4. Conclusions Friction Stir Processing of Investment-cast Ti–6Al–4V. Metallurgical and Mate-
rials Transactions A Physical Metallurgy and Materials Science 38(2):401–408.
[8] Alberti N, Forcellese A, Fratini L, Gabrielli F (1998) Sheet Metal Forming of
In the paper a new fixture for FSW of titanium alloys was Titanium Blanks using Flexible Media. CIRP Annals-Manufacturing Technology
proposed and the results of an experimental campaign were 47(1):217–244.
presented. On the basis of the obtained results the following [9] Short AB (2009) Gas Tungsten Arc Welding of a + b Titanium Alloys: A Review.
Materials Science and Technology 25(3):309–324.
advances in the specific knowledge of FSW of titanium alloys can [10] Yunlian Q, Ju D, Quan H, Liying Z (2000) Electron Beam Welding, Laser Beam
be pointed out: Welding and Gas Tungsten Arc Welding of Titanium Sheet. Materials Science
and Engineering A 280(1):177–181.
[11] Pasta S, Reynolds AP (2008) Residual Stress Effects on Fatigue Crack Growth in
 Due to the thermo-mechanical peculiarities of the titanium a Ti–6Al–4V Friction Stir Weld. Fatigue and Fracture of Engineering Materials
alloys the FSW tooling (both the clamping fixture and the and Structures 31(7):569–580.
rotating tool) must present specific features: proper refrigerating [12] Mironov S, Zhang Y, Sato YS, Kokawa H (2008) Development of Grain Structure
in b-Phase Field During Friction Stir Welding of Ti–6Al–4V Alloy. Scripta
systems must be used both for the tool and the anvil, the tool
Materialia 59(1):27–30.
itself must be obtained from a qualified high temperature [13] Lee W-B, Lee C-Y, Chang W-S, Yeon Y-M, Jung S-B (2005) Microstructural
resistant material, a non reactive refractory material must be Investigation of Friction Stir Welded Pure Titanium. Materials Letters
included in the backplate at least close to the weld center and, 59(26):3315–3318.
[14] Zhang Y, Sato YS, Kokawa H, Park SHC, Hirano S (2008) Stir Zone Micro-
finally, a shield of inert gas must be used to prevent oxidation structure of Commercial Purity Titanium Friction Stir Welded Using pcBN Tool.
phenomena; Materials Science and Engineering A 488(1–2):25–30.