The International Journal of Advanced Manufacturing Technology (2023) 129:3485–3496

https://doi.org/10.1007/s00170-023-12524-5

ORIGINAL ARTICLE

Dissimilar interface and joint strength of SS 304 and titanium friction

stir spot welds: a numerical and experimental analysis
Neeraj K. Mishra1 · Yogesh K. Churasiya1 · Amber Shrivastava1

Received: 3 April 2023 / Accepted: 17 October 2023 / Published online: 31 October 2023
© The Author(s), under exclusive licence to Springer-Verlag London Ltd., part of Springer Nature 2023

Abstract
The joining of Ti to stainless steel would allow cost-effective and light weight designs for many applications. For friction
stir spot welding of Ti to steel with lap configuration, dwell time is an important process parameter. The higher dwell time
promotes higher heat generation, which helps material movement and mixing for joint creation. However, higher dwell time
also leads to higher peak temperature and incubation time, which encourage undesirable IMC formation. This research
focuses on the impact of dwell duration on the elemental diffusion and IMC growth during FSSW of CP-Ti and SS 304.
This research numerically and experimentally analyzes the diffusion of elements, intermetallic compound formation, and
shear strength of stainless steel 304 and titanium dissimilar friction stir spot welds. The temperature profiles from numerical
thermal model show that the peak temperatures at the dissimilar joint interface increases as dwell time increases. The dif-
fusion model and characterization results reveal that Ni, Fe, and Cr diffuse across the dissimilar joint interface from steel to
titanium side. Since Ni, Fe, and Cr are β-Ti stabilizers, the β-Ti stabilizes next to the dissimilar interface. The α-β Ti needles
are observed between the stabilized β-Ti and α-Ti region towards base material. The numerical predictions of β-Ti and Fe-Ti
IMC compare well against the experimental results. From shear strength (tensile) tests, the maximum shear strength (tensile)
of ~ 8 kN is achieved from the SS 304-CP-Ti joint performed with 10 s dwell time. At the joint interface, the presence of
β-Ti strengthens the joint and the presence of Fe-Ti IMCs encourages abrupt brittle failure and reduces strength. With dwell
times less than 10 s, the Fe-Ti IMC thickness is less than 1 μm, which leads to higher shear strength (tensile) with higher
β-Ti thickness as dwell time increases. With 20 s dwell time, the Fe-Ti IMC thickness of about 1 μm leads to reduced shear
strength (tensile). The EDS analysis of the fracture surfaces reveal that dissimilar SS 304-CP-Ti joints failed from the dis-
similar interface containing Fe-Ti IMC.

Keywords FSSW · SS 304 · CP-Ti · Numerical modelling · IMC prediction

1 Introduction Interlayers were introduced between base materials to limit

the formation of IMCs with conventional joining methods,
The joints of stainless steels and titanium would enable for example, vanadium and tantalum interlayers between
petrochemical, aerospace, and nuclear sectors to achieve stainless steel (SS 304) and commercially pure titanium
sophisticated designs with less weight and expense [1, 2]. (CP-Ti) during laser welding [5]. However, any significant
The laser and tungsten inert gas welding are traditional reduction in IMC thickness could not be achieved. The IMCs
techniques for joining these two materials. At the dissimilar seen while welding CP-Ti and SS 304 were Fe-Ti, Cr-Ti, and
interface, the significant heat deposition involved in these Ni–Ti IMCs [6]. More recently, to address IMC formation at
processes leads to thick intermetallic compounds (IMCs) the dissimilar interface, interlayers of Cu [7], Nb [8], Ag [9],
[3]. These thick intermetallic compounds impart brittleness and Ni [10] and a combination of Ni + Nb [11] are utilized
and reduce ductility, deteriorating joint performance [4]. to join titanium alloys with stainless steels. The interlayer
reduces the thickness of the IMC by limiting direct elemen-
* Amber Shrivastava tal diffusion between the substrates [12]. Although the use of
ashrivastava.me@iitb.ac.in interlayer may inhibit/reduce the Fe-Ti IMCs, another IMC
of respective interlayer (Cu-Ti with Cu interlayer, Ni–Ti
1
Department of Mechanical Engineering, Indian Institute with Ni interlayer, etc.) replaces the Fe-Ti IMCs, which are
of Technology, Bombay, Mumbai, MH 400076, India

13
Vol.:(0123456789)
3486 The International Journal of Advanced Manufacturing Technology (2023) 129:3485–3496

comparatively less brittle than Fe-Ti. The complete elimi- more deformation of Al plate and weld defects were formed
nation of the IMC formation is extremely difficult during at the joining interface. Alaeibehmand et al. performed the
dissimilar joining. However, improved joint performance FSSW of aluminium (AA6061) and dual phase (DP600)
can be achieved by reducing the IMC growth. The fusion steel with a pinless tool [21]. At the dissimilar interface, a
welding processes involve melting and solidification during gray band of F­ e2Al5Zn0.4 IMC was reported. Furthermore,
joint formation, leading to significant intermixing and high the band towards the SS side had a smooth interface, in com-
temperatures for a relatively longer period. These factors parison to discontinuous and rough interface on Al side.
prevent any significant reduction of IMCs growth. The effect of tool penetration, pin length, and interlayer
The materials do not melt and solidify with solid-state materials has been studied on the IMCs in the dissimilar
joining techniques, which can reduce IMC formation. Fric- FSSW joints. The dwell time is another critical process input
tion stir welding (FSW) is one such technique explored for which affects the joint properties and IMC formation. More
dissimilar joining [13]. A few studies have explored the dwell time promotes higher heat generation, which helps
welding of steels and Ti with FSW. Various phases and IMCs material movement and mixing for joint creation. However,
like sigma, F­ e2Ti, ­Cr2Ti, ­Fe2Ti4O, and α-Ti were noticed in more dwell time also leads to a higher peak temperature and
the FSW of CP-Ti and AISI 316L, which adversely affected incubation time for IMC formation. This research focuses on
the joint strength [14]. Zhang et al. performed FSW in lap the impact of dwell duration on the elemental diffusion and
configuration for Ti6Al4V plates with colonies of β grains IMC growth during the FSSW of CP-Ti and SS 304. Further-
and α grains [15]. The stir zone had fine lamellar α + β more, the effectiveness of dwell time is evaluated towards
microstructure, which led to improved tensile strength and the inhibition/reduction of IMC formation and growth dur-
elongation. The underwater FSW and FSW with liquid nitro- ing dissimilar FSSW. For this purpose, the microstructure
gen were explored for welding Al and Mg alloys to control and phases, including IMCs and mechanical properties, are
the grain size [16]. Due to lower peak temperatures dur- investigated for the SS 304 and CP-Ti joints prepared with
ing the processes, the brittle IMCs were reported to be sup- different dwell times.
pressed. The joints from underwater FSW of Al–Mg alloy
and low carbon steel exhibited improved mechanical proper-
ties due to the stir zone with refined grains [17].
2 Numerical approach
Friction stir spot welding (FSSW) is similar to FSW and
used for fabricating spot welds. The dissimilar FSSW faces
A three-dimensional thermal model is created for the FSSW
similar challenges as FSW. Boucherit et. al. noticed Al-Cu
of SS 304 and CP-Ti, which estimates the temperature pro-
IMCs in the FSSWed joints of Al and Cu [18]. A Zn inter-
file at the joint interface. Figure 1 presents a schematic of
layer between Al and Cu during FSSW inhibited the forma-
the FSSW process. The thermal and diffusion models are
tion of A
­ l4Cu9 IMC; however, the shear tensile characteris-
discussed in the following sub-sections. Table 1 lists the
tics remained unaffected, with respect to the joints without
properties of both tool and workpiece materials [22].
Zn interlayer. For low carbon steel and Al 6063, the impact
of pin length on dissimilar FSSWed joints was studied by
Piccini et al. [19]. An increase in the tool penetration and
reduction in the pin length led to improved load bearing 3 Thermal model
capacity of the dissimilar joints. Similarly, with more tool
penetration, higher tensile shear loads were achieved for dis- The thermal model predicts the workpieces’ temperature
similar FSSWed joints of Al–Mg alloy and low-carbon steel during FSSW. The model consists of a three-dimensional
[20]. However, excessive penetration led to deterioration of transient conduction heat transfer equation, as shown below
joint properties. The excessive tool penetration resulted in [23]:

Fig. 1  Schematic of FSSW of

CP-Ti and SS 304

13
The International Journal of Advanced Manufacturing Technology (2023) 129:3485–3496 3487

Table 1  Material properties of workpiece and tool materials Table 2  Diffusion constant and Phases Do ­(mm2/s) A (kJ/mol)
activation energy for β-Ti and
Material Density Thermal Thermal diffusivity Fe-Ti phases [34] β-Ti 5.37 × ­10−2 272.9
(kg/m3) conductivity (m2/s)
(W/m–K) Fe-Ti 2.52 × ­10−5 228.2

SS 304 8000 16 4.2 × ­10−6

CP-Ti 4500 22.54 6.9 × ­10−6
Tungsten carbid tool 19,400 110 %1.%2 × ­10−5
3.2 Diffusion model and IMC thickness estimation

Equation 7 is the temperature-dependent diffusion model,

which captures the elemental diffusion according to Fick’s
second law of diffusion [32]:
( 2 )
𝜕T 𝜕 T 𝜕2T 𝜕2 T Q
=𝛼 + 2 + 2 + (1)
𝜕t 𝜕x2 𝜕y 𝜕z K
𝜕C 𝜕2C
=D 2 (7)
where T, t, α, Q, and K represent the temperature, time, ther- 𝜕t 𝜕X
mal diffusivity, internal heat generation rate, and thermal
where C, D, X, and t are the concentration, diffusion coef-
conductivity of workpiece materials, respectively. The heat
ficient, position, and time, respectively. The temperature-
produced at tool shoulder and workpiece contact, q, is given
dependent diffusion coefficient is defined as,
below [24]:
A
( )
= 𝜂m 𝛿𝜔r𝜇f P (2) D = Do exp − (8)
RT
where, P, r, and ω are axial pressure, the radial distance where Do is the diffusion constant (­ m2/s), R is the gas con-
from tool axis, and tool angular velocity, respectively. ηm is stant, and A is activation energy (J/mol). An empirical rela-
the fraction of heat at the tool-workpiece interface, which tionship is used to calculate IMC thickness as follows [33]:
transfers to the workpiece and is taken to be 0.7 [25] in this
study. An average axial pressure of 12 kN is used for the dx
= Dt−0.5 (9)
given workpieces, their dimensions and process parameters, dt
as per [26]. The local slip δ [27] and coefficient of friction μf where x is IMC thickness, and D is diffusion coefficient,
[28] vary along radial direction, as below: which is calculated by Eq. 8. Do and A for β-Ti and Fe-Ti are
(
r𝜔
) listed in Table 2.
𝛿 = 0.31exp − 0.026 (3) The thermal model employs finite differencing, with a
1.87
central difference scheme for space and the forward dif-
ference strategy for time. The domain is divided into
𝜇f = 0.5exp(−δrω) (4)
360 × 40 × 6 elements differential volumes along x, y, and
z directions, respectively. This resulted in a uniform grid
size of 0.5 mm in each direction. A time step of 0.005 s is
3.1 Thermal boundary and initial conditions used. For the diffusion model, 1D simulations over a domain
length of 40 μm were performed, with a grid size of 0.1 μm,
For tool and workpieces, 298 K (25 °C) is the initial tem- for 5 s with a time step of 0.003 s.
perature. Newton’s law defines the convection boundary
condition at all the exposed workpiece surfaces, as shown
below [29]: 4 Experimental methods
dT h(
(5) Friction stir spot welding of 100 mm × 20 mm × 1.5 mm aus-
)
= T − Tamb
dn K tenitic stainless steel (SS 304) and commercially pure tita-
where n and h are surface normal and convective heat trans- nium strips was performed with lap joint configuration. The
fer coefficients, respectively. h is 10 W/m2 K [30]. The heat SS 304 strips were on top of CP-Ti strips with an overlap of
transfer at the workpieces-backing plate interface is mod- 20 mm (Fig. 1). The 16.5-mm-diameter tool was made of
elled as per [31] with following heat transfer coefficient: tungsten carbide. Table 3 displays the workpiece materials’
chemical compositions [35].
A three-axis milling machine was used to perform
)2
(6)
(
h = 0.008 T − Tamb
FSSW. The FSSW was carried with 1800 rpm tool rota-
where Tamb is the ambient atmospheric temperature. tion speed, 0.5 mm plunge depth, and 0° tool tilt. The

13
3488 The International Journal of Advanced Manufacturing Technology (2023) 129:3485–3496

Table 3  Chemical composition Material Ti Fe C H Si N Cr Ni Mn P

of SS 304 and CP-Ti (wt%)
SS 304 - Bal 0.08 - 0.42 0.05 18.40 8.15 1.60 0.03
CP-Ti 99 - < 0.1 < 0.015 < 0.25 < 0.03 - - - -

mild steel backing plates were used to support the work- 5 Results and discussion
pieces on the work table. The welds were prepared with
a dwell time of 0 s, 2.5 s, 5 s, 10 s, and 20 s. For micro- 5.1 Temperature distribution
structural investigation, the 8-mm-long samples were
prepared from the middle of the joints, revealing the The numerical model is used to capture the joint interface
transverse cross-section of the weld zone. These samples temperature during FSSW. Figure 2 a and b present the tem-
were mechanically polished, then electropolished for 8 s perature distributions at vertical cross-section (xz plane;
at 20 V and − 10 °C using an electrolyte consisting of Fig. 1) at mid-width (10 mm from workpiece edges) and xy
perchloric acid, ethanol, and butanol. To analyze the weld plane containing the dissimilar interface (Fig. 1), respec-
cross-section, the Zeiss optical microscope was used. The tively, for the dwell time of 20 s. Figure 3 a–c show the
energy-dispersive spectroscopy (EDS) captured the ele- temperature distributions in the xy and xz planes containing
mental distribution on a scanning electron microscope the dissimilar interface for the dwell times of 2.5 s, 5 s, and
(Gemini SEM 300). Four joints were prepared for each 10 s, respectively.
weld condition (dwell time), to achieve one sample for The temperature of the workpieces is the highest at the
microstructural investigations (SEM and EDS) and three top and reduces towards the bottom. The highest tempera-
for tensile strength testing. Electron back scattered dif- tures at the top correspond with the maximum heat genera-
fraction (EBSD) analysis was performed on Zeiss Gemini tion at the tool shoulder-SS 304 contact. Figure 2 b shows
Crossbeam 350 SEM at 10 k × magnification over 22 × 17 the temperature variation at the xy plane containing the dis-
μm 2 with a step size of 0.05 μm. The X-ray diffraction similar interface for 20 s dwell time. Figure 3 also shows the
(XRD) scans were performed with PANalytical Empyrean temperature contours for dwell time of 2.5 s, 5 s, and 10 s.
equipment (2θ in the range of 20–100°, dwell time of 1 s, From Figs. 2 and 3, it can be seen that the maximum tem-
step size of 0.05°), and 2θ data were analyzed with the peratures are noticed beneath the tool-SS 304 contact with
high score plus software. The MTS landmark machine the plate. The maximum temperature is predicted at a small
was used for tensile strength testing, and the fracture sur- distance from the center and then decreases as the distance
face was scanned with SEM and EDS. increases. Such a temperature distribution is due to (i) non-
uniform heat generation from center to periphery, which can
be noticed from Eq. 2, where the heat generation increases

Fig. 2  Predicted temperature

profiles at a vertical cross-
section (xz plane, 10 mm from
workpiece edges) and b xy
plane containing the dissimilar
interface, for 20 s dwell time

13
The International Journal of Advanced Manufacturing Technology (2023) 129:3485–3496 3489

Fig. 3  Predicted temperature profiles containing dissimilar interfaces for a 2.5 s, b 5 s, and c 10 s dwell times

as radius increases, and (ii) the maximum heat loss from 5.2 Diffusion estimation
conduction is at the outer periphery, right below the tool-
SS 304 contact. The dwell time also affects the temperature As the material heats up during the FSSW process, Fe, Cr,
field, and the same can be noticed in Fig. 3 a, b and c. With and Ni diffuse from SS 304 into CP-Ti. Similarly, Ti dif-
increasing dwell time, the peak temperature rises. The peak fuses from CP-Ti into SS 304. For the numerical estimation
temperatures for the dwell times of 20 s, 10 s, 5 s, and 2.5 s of elemental diffusion, the activation energies and diffusion
are 1321 K, 1276 K, 1134 K, and 792 K, respectively. At the constants are presented in Table 4. Table 4 also shows the
tool-SS 304 contact, the maximum temperature of 1321 K diffusion coefficients for Cr, Fe, Ni into Ti, and Ti into Fe
is predicted for a dwell time of 20 s (Fig. 2a), which is less for the peak temperature of 1321 K, corresponding to the
than the melting point of SS 304 (1450 °C). The temperature representative joint prepared with 20 s dwell time [34].
decreases along the depth (-z direction). Fick’s second law of diffusion was solved with these diffu-
Furthermore, for the CP-Ti-SS 304 interface, the maxi- sion coefficients to estimate the concentration of Cr, Fe, Ni
mum temperature of 1192 K is predicted, which is less than in Ti, and Ti in Fe.
the melting of CP-Ti (1668 °C). The numerically calculated Figure 4 a shows the SEM image of the dissimilar inter-
peak temperatures for each weld dwell were used to estimate face from the CP-Ti-SS 304 joint, prepared with 20 s dwell
the elemental diffusion coefficients (Cr, Ni, Fe in Ti, and Ti time. From the EDS line scan (Fig. 4a), the concentration
in Fe). These were further used to predict the β-Ti and Fe-Ti profile of elements is shown in Fig. 4b. Figure 4 c shows the
IMC thicknesses for different dwell times. numerically predicted concentration profile of elements. The

Table 4  Activation energies, diffusion constants, and diffusion coefficients (for 1321 K peak temperature with a representative dwell time of
20 s) for the elemental diffusion [34]
Inter-diffusion Cr into Ti Fe into Ti Ni into Ti Ti into Fe

Activation energy (kJ/mol) 158 134 123.9 293.2

Diffusion constant ­(m2/s) 1 × ­10−6 8.5 × ­10−7 9.2 × ­10−7 0.21
Diffusion coefficient (­ m2/s) 1.202 × ­10−13 1.238 × ­10−12 3.404 × ­10−12 2.215 × ­10−14

13
3490 The International Journal of Advanced Manufacturing Technology (2023) 129:3485–3496

Fig. 4  a SEM image, b concentration profile of elements obtained from EDS line scan, and c concentration profile of elements from numerical
prediction, for the representation CP-Ti-SS 304 joint prepared with 20 s dwell time

predicted elemental concentration profile well captures the β-Ti phase and Fe-Ti IMC near the interface. Figure 6 a to
concentration profile of elements observed from the EDS d show the micrographs for the joints prepared with dwell
scan. From Fig. 4b, Fe diffused in Ti to a longer distance times of 0 s, 2.5 s, 5 s, and 10 s, respectively. Primarily, SS
from dissimilar interface, in comparison to the diffusion of 304 consists of austenitic phase (γ-Fe) and CP-Ti consists
Ti into Fe. Some Cr and Ni from the SS side also diffused of the α-Ti phase.
towards the Ti side. The predicted concentration profiles During FSSW, the temperature rise led to the phase trans-
compare well with the experimentally observed profiles. formation from α-Ti to β-Ti near CP-Ti side of the dissimilar
interface, which further transforms back majorly to α-Ti.
5.3 Microstructural analysis Any phase transformation with SS 304 was not noticed, as
the peak temperatures were not high enough to trigger the
During FSSW process, as the rotating tool contacts the γ-Fe to δ-Fe transformation (~ 1500 K [36]). The Fe-Ti IMCs
work piece, heat generates between tool and SS304. The were noticed at the dissimilar SS 304-CP-Ti interface, which
heat generated at the interface is conducted to CP-Ti plate are discussed in next sub-section. The β-transus temperature
beneath SS304 and leads to a non-uniform temperature field for CP-Ti is 882 ± 2 °C [37] and during FSSW transforma-
in the joining materials, as discussed in Sect. 5.1. At higher tion of α-Ti to β-Ti takes place. During cooling, the β-Ti
temperature, diffusion of elements from one side to another transforms back into α-Ti. However, Cr Fe, and Ni diffuse
side takes place and affects the formation of different phases from SS 304 side to CP-Ti side, which are the β-Ti stabiliz-
during cooling. Figure 5 shows the schematic of the phase ers at lower temperatures [38]. This leads to β-Ti next to the
transformation mechanism at the dissimilar interface. From dissimilar interface in the Fe/Cr/Ni rich region and α-β Ti
Fig. 5, the temperature rise accelerates the diffusion of ele- needles between β-Ti and α-Ti regions, as β-Ti transforms
ments across the boundary, which influences the stability of to α-Ti during cooling, which can be seen in Fig. 5c. Same

13
The International Journal of Advanced Manufacturing Technology (2023) 129:3485–3496 3491

Fig. 5  Evolution of phase transformation during FSSW process

Fig. 6  Dissimilar joint interface for dwell time of a 0 s, b 2.5 s, c 5 s, and d 10 s

kind of α-β Ti needles was also seen in different joints made interface. Figure 7 a shows the EBSD phase map near the
between SS and Ti [39–41]. joining interface. The β-Ti formed next to α-β Ti needles.
From Fig. 6, the α-β needles band is thickest for 5 s The phase fraction of β-Ti is ~ 20%. Near the interface, nee-
weld dwell and its thickness reduces as the weld dwell time dle structures are rich in β-Ti and, away from the interface,
increases. The peak temperature increases with dwell time, the major phase is α-Ti. On the SS304 side, no phase trans-
which encourages the higher/more diffusion of β-Ti stabiliz- formation is observed and primarily γ-austenite phase is pre-
ing elements (Fe, Cr, Ni) towards the CP-Ti side. More β-Ti sent. X-ray diffraction (XRD) is also performed to further
(from stabilizing effect of Fe, Cr, Ni) reduces the thickness confirm the phases. In addition to γ-austenite, α-Ti, and β-Ti,
of α-β needles. This promotes higher thickness of β-Ti and the Fe-Ti IMCs are also noticed from XRD 2θ plot (Fig. 7b).
lower thickness of α-β needles band as dwell time increases Similar peaks of β-Ti and Fe-Ti IMCs were also reported by
(Fig. 6a–d). previous studies [39, 40, 42, 43] on the joining of SS to Ti.

5.4 EBSD phase map and XRD 5.5 β‑Ti and Fe‑Ti IMC estimation

Electron back scattered diffraction (EBSD) analysis was per- Figure 8 shows experimentally measured and numerically
formed at higher magnification on a representative sample predicted β-Ti and Fe-Ti IMC thicknesses. The IMC thick-
(20 s dwell), to confirm the phases present at the joining nesses were measured from the SEM images (Figs. 4a

13
3492 The International Journal of Advanced Manufacturing Technology (2023) 129:3485–3496

Fig. 7  A EBSD phase map and b XRD 2θ plot of the joining interface

Fig. 8  Experimental and

predicted thicknesses of a β-Ti
and b Fe-Ti IMC, for different
dwell times

and 6). From Fig. 8, both β-Ti and Fe-Ti IMC thicknesses 5.6 Shear strength (tensile) results
increase as dwell time increases. The higher dwell time pro-
motes more heat generation, which leads to higher peak tem- For each FSSW condition, three shear strength (tensile)
perature (Eq. 1) and encourages more IMC formation (Eqs. 8 tests were conducted. Figure 9 displays the tensile shear
and 9). The β-Ti thicknesses for 0 s, 2.5 s, 5 s, 10 s, and 20 s test results for various dwell times. Shear strength (tensile)
weld dwell are 1.5 μm, 2.6 μm, 5.8 μm, and 11.1 μm, respec- goes up from ~ 4.5 to ~ 8 kN as dwell time goes up from 2.5
tively. The numerically predicted β-Ti thicknesses for 2.5 s, to 10 s. Furthermore, an increase in dwell time to 20 s led to
5 s, 10 s, and 20 s dwell time are 0.5 μm, 1 μm, 4.2 μm, and reduced shear strength (tensile) of ~ 7 kN. From Fig. 9, the
9.1 μm, respectively. Similarly, the Fe-Ti IMC thicknesses β-Ti thickness also increases as the dwell time increases. The
for 2.5 s, 5 s, 10 s, and 20 s dwell time from SEM images are β-Ti is relatively ductile, and the α-β alloys of Ti are known
0.2 μm, 0.3 μm, 0.7 μm, and 1 μm, respectively. Also, the to exhibit higher strengths as compared to pure (α or β) Ti.
predicted IMC thicknesses for corresponding dwell times are The presence of β-Ti at dissimilar interfaces tends
0.1 μm, 0.2 μm, 0.3 μm, and 2 μm, respectively. Numerically to improve joint strength. However, as the temperature
predicted thicknesses are in accordance with experimental increases with higher dwell time, the Fe-Ti IMC also
observations. The β-Ti and Fe-Ti IMC thickness were not increases at the dissimilar interface. The Fe-Ti IMCs are rel-
predicted for 0 s dwell time as the plunging phase was not atively more brittle. The excessive presence of Fe-Ti IMCs
modelled. At high dwell times, the model over predicts the at dissimilar interfaces would promote abrupt brittle failure
Fe-Ti IMC thickness. This could be due to inconsistent dif- and lower the strength. This suggests that with lower dwell
fusion constant and activation energy for Fe-Ti (Table 2) at times (up to 10 s), the presence of β-Ti and relatively less
high temperature, from high dwell time. Fe-Ti IMC (thickness < 1 μm (Sect. 5.5)) leads to improved/

13
The International Journal of Advanced Manufacturing Technology (2023) 129:3485–3496 3493

transforms to the α-Ti phase, which has hexagonal close-

packed (HCP) structure. For temperatures above β-transus,
the change of crystal structure from HCP to BCC enables
more open crystallography for easy diffusion of elements.
From SS304 side, Fe, Cr, and Ni diffused towards CP-Ti side
and Ti from CP-Ti side, diffused towards SS304 side. Fe, Cr,
and Ni (β-Ti stabilizers) stabilized the β-Ti near the interface
and inhibited the complete transformation of β-Ti into α-Ti
during cooling. The diffusion of these elements in titanium
lattice lowers the eutectoid transformation temperature and
encourages the β-Ti and α-β needle structure during cool-
ing [44]. Also, the presence of Fe, Cr, and Ni in Ti lattice
reduces the mobility of titanium atoms (The atomic size fac-
tors of Fe, Cr, and Ni are compatible with the β-Ti phase of
titanium.), making it more difficult for the crystal lattice to
rearrange and transform to the α-Ti phase.
As dwell time increases, the phase fraction of β-Ti and
Fig. 9  Tensile shear strength for the joints prepared with different Fe-Ti IMC increases, which affects the mechanical proper-
dwell times ties of the joint. The higher phase fraction of β-Ti improves
the strength. Previously, it is reported that as the phase
higher shear strength (tensile). However, for the joint pre- fraction of β-Ti increases, initially, the strength increases;
pared with the highest dwell time of 20 s, the Fe-Ti IMC however, further increase in β-Ti phase fraction decreases
thickness of about 1 μm reduces shear strength (tensile). The the strength [45, 46]. Similarly in the current study, as the
fracture surfaces from the representative shear test (tensile) dwell time increases up to 10 s, the presence of β-Ti phase
specimens were examined with SEM. The features on the and Fe-Ti IMC increases the tensile shear strength of the
fracture surfaces indicated primarily brittle failure. Results joint. However, higher dwell time (more than 10 s) leads to
of EDS analysis on the fracture surface of the joint prepared more β-Ti and Fe-Ti IMC, which makes the joint brittle and
at 20 s dwell time are shown in Fig. 10. Point EDS was deteriorates the join performance.
performed at three locations, and corresponding values are
listed alongside. The considerable Fe and Ti presence at all
locations strongly indicates Fe-Ti IMCs. Furthermore, the 6 Conclusion
dissimilar interface with Fe-Ti IMC is the location of the
fracture upon the shear strength (tensile) test. For friction stir spot welding of Ti to steel with lap config-
The numerical simulation shows that the dwell time uration, dwell time is an important process parameter. This
affects the heat generation at the dissimilar interface, such study investigates the dissimilar FSSW of CP-Ti and SS
that the interface temperature increases as dwell time 304, experimentally and numerically. The dwell time was
increases. The temperature at the Ti side increases beyond varied across the dissimilar FSSWed joints of CP-Ti and
the β-transus temperature, which leads to α-Ti to β-Ti phase SS 304. The tensile testing and characterization of FSSW
transformation during heating. The β-Ti phase has body- joints revealed the joint shear strength and microstructure
centered cubic (BCC) crystal structure, which is stable at of the dissimilar interface. The temperature profile during
elevated temperatures and as temperature decreases, the β-Ti FSSW, elemental distribution, and phase (β-Ti and Fe-Ti

Fig. 10  SEM image and

elemental distribution from Element Location
EDS for fractured surface from (wt. %) 1 2 3
the joint prepared with 20 s
weld dwell Fe 37.6 36 41.4

Ti 52.2 54.6 45

Ni 5.2 5 4.8

Cr 5 4.4 6

13
3494 The International Journal of Advanced Manufacturing Technology (2023) 129:3485–3496

IMC) thicknesses at the dissimilar interface are estimated funding acquisition, project administration, writing—reviewing and
numerically. The presence of these phases was further con- editing.
firmed by the EBSD phase map and XRD 2θ analysis. The
Declarations
following inference can be made from the analysis:
Competing interests The authors declare no competing interests.
(1) Dwell time is an important process parameter which
affects the heat generation during the process. The tem-
perature at the joning interface increases as dwell time
increases, which enables the interdiffusion of elements References
across the interface.
(2) The numerical and experimental results demonstrate 1. Cooke KO, Atieh AM (2020) Current trends in dissimilar diffu-
sion bonding of titanium alloys to stainless steels, aluminium and
that as the dwell time increases, the Fe from the SS side magnesium. J Manufact Materials Processing 4(2):39. https://​doi.​
diffuses to larger depths into the CP-Ti side. Some Cr org/​10.​3390/​jmmp4​020039
and Ni from the SS side also diffused towards the CP-Ti 2. Chen HC, Bi G, Lee BY, Cheng CK (2016) Laser welding of
side. CP-Ti to stainless steel with different temporal pulse shapes. J
Mater Process Technol 231:58–65. https://d​ oi.o​ rg/1​ 0.1​ 016/j.j​ matp​
(3) The microstructural analysis shows no significant rotec.​2015.​12.​016
change on SS 304 side, for all the FSSW joints. On 3. Li J, Schneiderman B, Song S, Gilbert SM, Vivek A, Yu Z, ... &
Ti side, α-Ti transforms to β-Ti during heating. The Daehn GS (2020). High strength welding of Ti to stainless steel
diffusion of Fe, Cr, and Ni into CP-Ti side stabilizes by spot impact: microstructure and weld performance. The Inter-
national Journal of Advanced Manufacturing Technology, 108,
β-Ti phase near the interface and inhibited the com- 1447–1461. https://​doi.​org/​10.​1007/​s00170-​020-​05506-4
plete transformation of β-Ti into α-Ti upon cooling. 4. Kale GB, Patil RV, Gawade PS (1998) Interdiffusion studies in
Furthermore, the significant presence of Fe and Ti led titanium–304 stainless steel system. J Nucl Mater 257(1):44–50.
to the formation of Fe-Ti intermetallics, which were https://​doi.​org/​10.​1016/​S0022-​3115(98)​00179-2
5. Shanmugarajan B, Padmanabham G (2012) Fusion welding stud-
confirmed with EBSD phase map and XRD. ies using laser on Ti–SS dissimilar combination. Opt Lasers Eng
(4) From shear strength (tensile) tests, as dwell time 50(11):1621–1627. https://​doi.​org/​10.​1016/j.​optla​seng.​2012.​05.​
increases the shear strength (tensile) initially increases, 008
however further increase in dwell time reduces the joint 6. Tomashchuk I, Sallamand P, Belyavina N, Pilloz M (2013) Evo-
lution of microstructures and mechanical properties during dis-
strength. The maximum shear strength (tensile) of ~ 8 similar electron beam welding of titanium alloy to stainless steel
kN is achieved corresponding to10 s dwell time. The via copper interlayer. Mater Sci Eng, A 585:114–122. https://​doi.​
joint is strengthened by the presence of β-Ti at the org/​10.​1016/j.​msea.​2013.​07.​050
dissimilar interface, and the presence of Fe-Ti IMCs 7. Pardal G, Ganguly S, Williams S, Vaja J (2016) Dissimilar metal
joining of stainless steel and titanium using copper as transition
encourages abrupt brittle failure and reduces strength. metal. Int J Adv Manufact Technol 86:1139–1150. https://d​ oi.o​ rg/​
For the joints prepared with dwell times less than 10 s, 10.​1007/​s00170-​015-​8110-2
the Fe-Ti IMC thickness is less than 1 μm, which leads 8. Zhang Y, Sun D, Gu X, Li H (2016) A hybrid joint based on two
to higher shear strength (tensile) with higher β-Ti thick- kinds of bonding mechanisms for titanium alloy and stainless steel
by pulsed laser welding. Mater Lett 185:152–155. https://​doi.​org/​
ness as dwell time increases. For the joint prepared with 10.​1016/j.​matlet.​2016.​08.​138
dwell time more than 10 s, the relatively higher Fe-Ti 9. Balasubramanian M (2016) Characterization of diffusion-bonded
IMC thickness reduces shear strength (tensile). The titanium alloy and 304 stainless steel with Ag as an interlayer. The
EDS analysis of the fracture surfaces from the shear Int J Adv Manufact Technol 82(1–4):153–162. https://​doi.​org/​10.​
1007/​s00170-​015-​7376-8
strength (tensile) test reveals that dissimilar joints failed 10. Cheepu M, Ashfaq M, Muthupandi V (2017) A new approach for
from the dissimilar interface containing Fe-Ti IMC. using interlayer and analysis of the friction welding of titanium to
stainless steel. Trans Indian Inst Met 70:2591–2600. https://​doi.​
org/​10.​1007/​s12666-​017-​1114-x
Acknowledgements The authors gratefully acknowledge Dr. Nikhil 11. Li JL, Huo LP, Zhang FS, Xiong JT, & Li WY (2009). Frac-
Gotawala’s support with initial experiments and discussions regard- ture characteristics of vacuum diffusion bonded TA2 titanium
ing numerical model. We would also like to acknowledge the partial to 1Cr18Ni9Ti stainless steel joint with Nb+ Ni interlayers.
support of this work by the Advanced Manufacturing Processes Lab, In Materials Science Forum (Vol. 620, pp. 399–402). Trans Tech
Microstructural Mechanics and Micro forming Lab and Machine Tools Publications Ltd.https://d​ oi.o​ rg/1​ 0.4​ 028/w
​ ww.s​ cient​ ific.n​ et/M
​ SF.​
Lab at the Department of Mechanical Engineering, Indian Institute of 620-​622.​399.
Technology Bombay, Mumbai. 12. He P, Zhang J, Zhou R, Li X (1999) Diffusion bonding technology
of a titanium alloy to a stainless steel web with an Ni interlayer.
Author contribution Neeraj K. Mishra: investigation, conceptualiza- Mater Charact 43(5):287–292. https://​doi.​org/​10.​1016/​S1044-​
tion, methodology, formal analysis, data curation, validation, software, 5803(99)​00008-X
writing-editing. Yogesh Churasiya: investigation, conceptualization, 13. Mishra RS (2005) MA ZY. Friction stir welding and processing.
methodology, software, writing-original draft. Amber Shrivastava: Mater Sci Eng R 50:1–78. https://​doi.​org/​10.​1016/j.​mser.​2005.​07.​001

13
The International Journal of Advanced Manufacturing Technology (2023) 129:3485–3496 3495

14. Akbarimousavi SAA, GohariKia M (2011) Investigations on the welding process for AA5083-H18 sheets. Europ J Mechanics-A/
mechanical properties and microstructure of dissimilar cp-tita- Solids 29(2):204–215. https://​doi.​org/​10.​1016/j.​eurom​echsol.​
nium and AISI 316L austenitic stainless steel continuous friction 2009.​10.​006
welds. Mater Des 32(5):3066–3075. https://​doi.​org/​10.​1016/j.​ 30. Sahlot P, Singh AK, Badheka VJ, Arora A (2019) Friction
matdes.​2010.​12.​057 stir welding of copper: numerical modeling and validation.
15. Zhang Y, Sato YS, Kokawa H, Park SHC, Hirano S (2008) Micro- Trans Indian Inst Met 72:1339–1347. https://​doi.​org/​10.​1007/​
structural characteristics and mechanical properties of Ti–6Al–4V s12666-​019-​01629-9
friction stir welds. Mater Sci Eng, A 485(1–2):448–455. https://​ 31. Soundararajan V, Zekovic S, Kovacevic R (2005) Thermo-
doi.​org/​10.​1016/j.​msea.​2007.​08.​051 mechanical model with adaptive boundary conditions for
16. Mofid MA, Abdollah-Zadeh A, Ghaini FM, Gür CH (2012) Sub- friction stir welding of Al 6061. Int J Mach Tools Manuf
merged friction-stir welding (SFSW) underwater and under liq- 45(14):1577–1587. https://​doi.​org/​10.​1016/j.​ijmac​htools.​2005.​
uid nitrogen: an improved method to join Al alloys to Mg alloys. 02.​008
Metall and Mater Trans A 43:5106–5114. https://d​ oi.o​ rg/1​ 0.1​ 007/​ 32. Momeni F, Ni J (2020) A new model for predicting the thickness
s11661-​012-​1314-2 of intermetallic compounds in friction stir welding. J Manuf Pro-
17. Derazkola HA, Eyvazian A, Simchi A (2020) Submerged fric- cess 56:1331–1339. https://d​ oi.o​ rg/1​ 0.1​ 016/j.j​ mapro.2​ 020.0​ 4.0​ 28
tion stir welding of dissimilar joints between an Al-Mg alloy and 33. Jedrasiak P, Shercliff HR, Reilly A, McShane GJ, Chen YC, Wang
low carbon steel: thermo-mechanical modeling, microstructural L, ... & Prangnell P (2016). Thermal modeling of Al-Al and Al-
features, and mechanical properties. J Manuf Process 50:68–79. steel friction stir spot welding. Journal of Materials Engineer-
https://​doi.​org/​10.​1016/j.​jmapro.​2019.​12.​035 ing and Performance, 25, 4089–4098. https://​doi.​org/​10.​1007/​
18. Boucherit A, Avettand-Fènoël MN, Taillard R (2017) Effect of s11665-​016-​2225-y.
a Zn interlayer on dissimilar FSSW of Al and Cu. Mater Des 34. Neumann G, Tuijn C (2011) Self-diffusion and impurity diffusion
124:87–99. https://​doi.​org/​10.​1016/j.​matdes.​2017.​03.​063 in pure metals: handbook of experimental data. Elsevier
19. Piccini JM, Svoboda HG (2015) Effect of pin length on friction 35. Campo KN, Campanelli LC, Bergmann L, dos Santos JF, Bol-
stir spot welding (FSSW) of dissimilar aluminum-steel joints. farini C (2014) Microstructure and interface characterization of
Procedia Materials Science 9:504–513. https://​doi.​org/​10.​1016/j.​ dissimilar friction stir welded lap joints between Ti–6Al–4V and
mspro.​2015.​05.​023 AISI 304. Mater Des 1980–2015(56):139–145. https://d​ oi.o​ rg/1​ 0.​
20. Lee CY, Choi DH, Yeon YM, Jung SB (2009) Dissimilar friction 1016/j.​matdes.​2013.​11.​002
stir spot welding of low carbon steel and Al–Mg alloy by forma- 36. Tseng CC, Shen Y, Thompson SW, Mataya MC, Krauss G (1994)
tion of IMCs. Sci Technol Weld Joining 14(3):216–220. https://​ Fracture and the formation of sigma phase, M 23 C 6, and austenite
doi.​org/​10.​1179/​13621​7109X​400439 from delta-ferrite in an AlSl 304L stainless steel. Metall and Mater
21. Alaeibehmand S, Mirsalehi SE, Ranjbarnodeh E (2021) Pinless Trans A 25:1147–1158. https://​doi.​org/​10.​1007/​BF026​52290
FSSW of DP600/Zn/AA6061 dissimilar joints. J Market Res 37. Ślęzak T, Zmywaczyk J, & Koniorczyk P (2019). Thermal diffu-
15:996–1006. https://​doi.​org/​10.​1016/j.​jmrt.​2021.​08.​071 sivity investigations of the titanium grade 1 in wide temperature
22. Pasang T, Pramana SS, Kracum M, Misiolek WZ, Aziziderouei M, range. In AIP Conference Proceedings (2170(1):020019). AIP
Mizutani M, Kamiya O (2018) Characterization of intermetallic Publishing LLC. https://​doi.​org/​10.​1063/1.​51327​38
phases in fusion welded commercially pure titanium and stainless 38. Peart RF, Tomlin DH (1962) Diffusion of solute elements in beta-
steel 304. Metals 8(11):863. https://​doi.​org/​10.​3390/​met81​10863 titanium. Acta Metall 10(2):123–134. https://​doi.​org/​10.​1016/​
23. Vicharapu B, Lemos GVB, Bergmann L, dos Santos JF, De A, & 0001-​6160(62)​90057-3
Clarke T (2021). Probing underlying mechanisms for PCBN tool 39. Kundu S, Ghosh M, Laik A, Bhanumurthy K, Kale GB, Chat-
decay during friction stir welding of nickel-based alloys. Tecno- terjee S (2005) Diffusion bonding of commercially pure titanium
logia em Metalurgia, Materiais e Mineração. 18:0–0. https://​doi.​ to 304 stainless steel using copper interlayer. Mater Sci Eng, A
org/​10.​4322/​2176-​1523.​20202​455. 407(1–2):154–160. https://​doi.​org/​10.​1016/j.​msea.​2005.​07.​010
24. Yang X, Feng W, Li W, Xu Y, Chu Q, Ma T, Wang W (2018) 40. Li J, Vivek A, Daehn G (2021) Improved properties and thermal
Numerical modelling and experimental investigation of thermal stability of a titanium-stainless steel solid-state weld with a nio-
and material flow in probeless friction stir spot welding process bium interlayer. J Mater Sci Technol 79:191–204. https://​doi.​org/​
of Al 2198–T8. Sci Technol Weld Joining 23(8):704–714. https://​ 10.​1016/j.​jmst.​2020.​11.​050
doi.​org/​10.​1080/​13621​718.​2018.​14698​32 41. Kundu S, Chatterjee S (2010) Evolution of interface microstruc-
25. Bang HS, Hong SM, Das A, Bang HS (2020) A prediction of ture and mechanical properties of titanium/304 stainless steel
Fe-Al IMC layer thickness in TIG-assisted hybrid friction stir diffusion bonded joint using Nb interlayer. ISIJ Int 50(10):1460–
welded Al/steel dissimilar joints by numerical analysis. Int J 1465. https://​doi.​org/​10.​2355/​isiji​ntern​ation​al.​50.​1460
Adv Manufact Technol 106:765–778. https://​doi.​org/​10.​1007/​ 42. Gotawala N, Shrivastava A (2021) Thermodynamics-based analy-
s00170-​019-​04560-x sis of formation and growth of FeTi and β-Ti during friction stir
26. Su P, Gerlich A, Yamamoto M, North TH (2007) Formation welding of SS304 and pure titanium. J Mater Sci 56(34):19180–
and retention of local melted films in AZ91 friction stir spot 19198. https://​doi.​org/​10.​1016/j.​msea.​2020.​139658
welds. J Mater Sci 42:9954–9965. https://​d oi.​o rg/​1 0.​1 007/​ 43. Hacisalihoglu I, Samancioglu A, Yildiz F, Purcek G, Alsaran A
s10853-​007-​2061-4 (2015) Tribocorrosion properties of different type titanium alloys
27. Mehta M, Chatterjee K, De A (2013) Monitoring torque and trav- in simulated body fluid. Wear 332:679–686. https://​doi.​org/​10.​
erse force in friction stir welding from input electrical signatures 1016/j.​wear.​2014.​12.​017
of driving motors. Sci Technol Weld Joining 18(3):191–197. 44. Ghosh M, Chatterjee S (2003) Diffusion bonded transition joints
https://​doi.​org/​10.​1179/​13621​71812Y.​00000​00084 of titanium to stainless steel with improved properties. Mater
28. Das H, Jana SS, Pal TK, De A (2014) Numerical and experimental Sci Eng, A 358(1–2):152–158. https://​doi.​org/​10.​1016/​S0921-​
investigation on friction stir lap welding of aluminium to steel. Sci 5093(03)​00298-3
Technol Weld Joining 19(1):69–75. https://d​ oi.o​ rg/1​ 0.1​ 179/1​ 3621​ 45. Filip R, Kubiak K, Ziaja W, Sieniawski J (2003) The effect of
71813Y.​00000​00166 microstructure on the mechanical properties of two-phase titanium
29. Kim D, Badarinarayan H, Kim JH, Kim C, Okamoto K, Wagoner alloys. J Mater Process Technol 133(1–2):84–89. https://​doi.​org/​
RH, Chung K (2010) Numerical simulation of friction stir butt 10.​1016/​S0924-​0136(02)​00248-0

13
3496 The International Journal of Advanced Manufacturing Technology (2023) 129:3485–3496

46. Lütjering GERD (1998) Influence of processing on microstructure Springer Nature or its licensor (e.g. a society or other partner) holds
and mechanical properties of (α+ β) titanium alloys. Mater Sci exclusive rights to this article under a publishing agreement with the
Eng, A 243(1–2):32–45. https://​doi.​org/​10.1​ 016/​S0921-​5093(97)​ author(s) or other rightsholder(s); author self-archiving of the accepted
00778-8 manuscript version of this article is solely governed by the terms of
such publishing agreement and applicable law.
Publisher's Note Springer Nature remains neutral with regard to
jurisdictional claims in published maps and institutional affiliations.

13