You are on page 1of 11

Optics & Laser Technology 165 (2023) 109629

Contents lists available at ScienceDirect

Optics and Laser Technology


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

Multi-scale simulation study on the evolution of stress waves and


dislocations in Ti alloy during laser shock peening processing
Cheng Gu a, *, Minghua Su a, Zenghui Tian a, Jianhua Zhao a, b, Yajun Wang a
a
College of Materials Science and Engineering, Chongqing University, Chongqing 400045, China
b
State Key Laboratory of Mechanical Transmission, Chongqing University, Chongqing 400044, China

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

Keywords: This study investigates the effect of laser shock peening (LSP) on the evolution of stress waves and dislocations in
Laser shock peening Ti-6Al-4V alloy by multi-scale simulation. The finite element model on the macroscale and molecular dynamics
Ti-6Al-4V alloy model on the microscale were built, and multi-scale simulations based on the finite element and molecular
Multi-scale simulation
dynamics were performed. Under the high pressure on material induced by LSP, elastic deformation and plastic
Stress wave
Dislocation
deformation are generated. With the increase of stress wave traveling distance, its intensity gradually decays. The
stress wave in the material completely vanishes eventually. The macroscale simulation shows that the average
velocity of the stress wave equals 6.0 km/s and does not change much with the loading pressure. The microscale
simulation shows that the average velocity of the stress wave in single crystal α-Ti is in the range of 7.1~7.8 km/
s. The average velocity of stress wave in polycrystals is always lower than that in the single crystal and has a
trend to decrease with the increase of grain number until a stable value of around 6.0 km/s. Considering the
induced shock pressure, the piston velocity should be below 0.4 km/s to recreate the LSP process. This paper
provides a comprehensive understanding of the effect of LSP on the evolution of stress waves and dislocations
and also provides a new numerical method to further study laser shock peening processing.

1. Introduction generated plane compression waves by an impact of a flyer plate. Jones


et al. [16] presented a series of experiments probing the martensitic
Titanium and its alloys have the characteristics of high specific hexagonal close-packed to simple hexagonal transition in titanium
strength, excellent corrosion resistance, and good biocompatibility, under shock-loading to peak stresses around 15 GPa. Gurusami et al.
which makes them suitable for a wide range of applications in aviation, [17] investigated the effect of LSP on the surface integrity of commer­
automotive, and biological field [1-3]. However, most of the titanium cially pure titanium and found that laser shock makes the surface of the
components tend to be subjected to extreme working conditions, which sample with an increase in tensile strength, microhardness, surface
may cause deformation and even fracture of the products [4,5]. There­ roughness, and compressive residual stresses. Jin et al. [18] investigated
fore, several surface treatment techniques have been conducted to the influence of LSP on the microstructure and fatigue behavior of
modify the surface-related properties and enhance the fatigue perfor­ additive-manufactured Ti-6Al-4V alloy. They found that the run-out
mance of products such as laser shock peening (LSP) [6,7], physical fatigue strength was significantly increased in the LSP-treated sample.
vapor deposition (PVD) [8], chemical vapor deposition (CVD) [9], etc. Sun et al. [19] experimentally and numerically investigated the residual
Compared with other technologies, LSP is able to penetrate the stresses induced by LSP in Ti-17 titanium specimens. Pavan et al. [20]
compressive residual stress into the material at the cost of fewer defects studied fatigue crack growth in a LSP-treated residual stress field and
on the treated material surface [10-12]. The compressive residual stress developed a linear-elastic finite-element crack growth prediction model.
is effective in improving fatigue properties by delaying the crack initi­ Petronić et al. [5] applied LSP on Ti alloy surface, with and without a
ation and by decelerating the crack propagation rate [13,14]. Kanel protective and transparent layer, and found that laser shock can improve
et al. [15] studied the effect of temperature on the mechanical property the surface topology and microhardness. In most of the publications, the
of high-purity and commercial-purity titanium and Ti-6Al-2Sn-2Zr-2Cr- effects of LSP on the properties are based on the measurements of the
2Mo-Si alloy upon sub-microsecond-scale shock-wave loading, and they residual stress and the crack length, and the evolutions of the stress wave

* Corresponding author.
E-mail address: gucheng.90@cqu.edu.cn (C. Gu).

https://doi.org/10.1016/j.optlastec.2023.109629
Received 10 December 2022; Received in revised form 22 April 2023; Accepted 17 May 2023
Available online 23 May 2023
0030-3992/© 2023 Elsevier Ltd. All rights reserved.
C. Gu et al. Optics and Laser Technology 165 (2023) 109629

Fig. 1. (a) The schematic diagram of the LSP process, (b) the finite element model, (c) the loading profile, and (d) the molecular dynamics model of monocrystalline
titanium processed by the shock wave.

and microstructure during LSP are less studied. waves and dislocations in Ti-6Al-4V alloy. The finite element model on
Several numerical studies were performed to investigate residual the macroscale and molecular dynamics model on the microscale are
stress and plastic strain by using the finite element method [21,22]. built, and multi-scale simulations based on the finite element model and
Peyre et al. [23] developed an axisymmetric FEM model for calculating molecular dynamics are performed. Based on the established model, we
the residual stresses induced by laser shock and found that it was a have investigated the effect of LSP on stress wave propagation and
rather efficient method to predict the surface modifications induced in dislocation evolution.
materials. Peyre et al. [24] also studied the influence of the coating on
pressure loading and found that the simulations gave satisfactory 2. Numerical model of laser shock peening
agreement with the experimental residual stress data. Voothaluru et al.
[25,26] developed a 3D finite element model for simulating the LSP The schematic diagram of the LSP process is shown in Fig. 1(a).
process and performed a study on the effect of multiple laser shocks and Before the LSP process, an ablative layer is added on the surface of the
their extent of overlap on the affected depths and the tensile and target material as the protection layer to avoid the material being ab­
compressive residual stresses. Zhou et al. [27] found that the residual lated by laser. The ablative layer is then covered with a transparent layer
stress and equivalent plastic strain induced by laser shock are sensitive as the confined overlay to raise the peak pressure of the shock wave and
to mesh size and can be well predicted. Pei et al. [28] applied numerical prolong its action time. During the LSP process, a high-power and short-
simulation to analyze the stress-wave propagation rules in Ti-17 tita­ duration laser pulse travels through the transparent overlay and irra­
nium alloy plates with different thicknesses. They found that the thinner diates onto the surface of the ablative layer. The ablative material is
specimen shows a higher reflection frequency of the stress wave and vaporized, and the high-pressure plasma is generated in an ultra-short
greater residual stress weakening. Bhamare et al. [21] employed a nu­ time. According to Fabbro et al. [33], for a given laser intensity,
merical approach to explore the relation between the processing pa­ higher peak pressures are obtained with longer pulse durations in the
rameters and the residual stress distribution. Luo et al. [29] developed a range of 0.6–30 ns experimentally. In addition, the action time of high-
finite element model to simulate the effects of overlapping rate, laser pressure plasma lasts two to three times longer than the duration of the
spot diameter, and laser power density on the residual stress of Ti-6Al- laser pulse. In this study, we have performed LSP experiments with a
4V alloy. Zhang et al. [30] investigated the residual stress field laser pulse width of 20 ns. Therefore, we set the duration of the laser
induced in a thin Ti-6Al-4V alloy plate through numerical simulation pulse as 20 ns. The duration of amplitude higher than 60 % is around 20
and experiments. It was found that the increasing laser shock pressure ns and the whole duration of the pressure pulse is around 60 ns as shown
had significant effects on the residual stress field. On the other hand, to in Fig. 1(c). The extremely rapid expansion of the plasma exerts pressure
study the dislocation of atoms induced by LSP, Meng et al. [31] used a on the surface of the target material under the constraint of the confining
molecular dynamics simulation to investigate the propagation of laser layer, resulting in its compression. As a result, the stress wave is induced
shock waves. Zhou et al. [32] used dislocation dynamics simulation to and propagates in the target material.
investigate laser shock-induced ultra-high strain rate plastic deforma­
tion and employed molecular dynamics to calculate dislocation
2.1. Finite element model in macroscale
mobility. Although studies have been performed on the residual stress
on the macroscale or dislocation on the microscale separately, some
In this study, the material response to laser shock was simulated by
parameters, such as particle velocity, are set based on the assumptions
using commercial software Abaqus, which has been used to simulate LSP
and it is difficult to investigate the evolutions of stress waves and dis­
[23,24,29,30]. The procedure included the dynamic analysis step and
locations according to the corresponding parameters of LSP.
the following static analysis step. In the dynamic analysis step, Abaqus/
This article is to study the effect of LSP on the evolution of stress
Explicit was utilized, and the dynamic analysis was implemented to

2
C. Gu et al. Optics and Laser Technology 165 (2023) 109629

Table 1 carried out by LAMMPS [35]. The molecular dynamic model is shown in
Physical and mechanical properties of Ti-6Al-4V alloy [30]. Fig. 1(d). In this study, the assumption was adopted that Ti alloy was
Materials ρ (kg/ E ν A B n C considered as a single crystal α-Ti and a polycrystal α-Ti, respectively.
m3) (GPa) (MPa) (MPa) The single crystal α-Ti had hexagonal close-packed (HCP) lattices and its
Ti-6Al- 4500 110 0.342 1098 1092 0.93 0.014 lattice constant was about 2.95 Å. The polycrystal α-Ti was established
4V by using Atomsk software and the Voronoi method [36]. The grain
number was set as 15 with an average grain size of about 6.5 nm, and the
thickness of the grain boundary was in the range of 1~1.5 nm. In both
capture the rapid response to the laser shock of the material. In the static cases, the dimensions of the initial simulation region were 118 Å × 204
analysis step, the transient results from the dynamic calculation were Å × 723 Å, and the total number of atoms was 1,500,000. Non-periodic
input into Abaqus/Standard to perform static calculation and to obtain boundaries were used in the shock direction (Z direction), while periodic
the final residual stress distribution in the static equilibrium state. boundaries were utilized in other directions (X and Y directions). The
Finally, the corresponding residual stress induced by several successive interatomic interactions in Ti were defined by an EAM potential [37].
impacts can be obtained. A piston with a particle velocity Up was used to achieve laser shock
A 3-D finite element model was established as shown in Fig. 1(b). on the microscale. The particle velocity should be obtained before
The target size was 15 mm × 15 mm × 5 mm and the symmetrical simulation. Based on the propagation process of the shock wave, the
boundary conditions were used in the model. The element type was total strain ε can be related to the particle velocity:
C3D8R. It is known that the finer mesh leads to more accurate results but
needs higher computational costs. Therefore, finer elements of 0.1 mm ε=
Up dt
(5)
at the center and coarser elements of 1 mm at the edge were set as the L
mesh. The element length along the z direction was 0.1 mm. The number
where dt is the time interval, and L is the length in shock direction. The
of total elements in the model was 288,800. The displacements and
total strain ε induced by laser shock wave can be written as [31]:
rotations of nodes at the edge of the sample were fully constrained as
( )
shown in Fig. 1(b). The laser local spot was 2.5 mm in diameter. The 2HEL Pmax
ε= − 1 (6)
shock wave pressure is related to laser power density. According to 3λ + 2G HEL
Fabbro et al. [33], the peak plasma pressure, P0 , induced by laser can be
calculated by: where HEL is the Hugoniot elastic limit, λ is the Lame constant, G is the
√̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅ shear modulus, and Pmax is the peak pressure induced by the laser shock
α wave. The peak pressure can be calculated by [33]:
P0 = 0.01 ZI0 (1)
2α + 3
( α )12 ( 4E )12
(7)
1
where α is the fraction of the internal energy devoted to the thermal Pmax = 0.01Z2
2α + 3 π R2 τ
energy, Z is the reduced shock impedance, and I0 is the laser power
density. The duration of the loaded pressure is shown in Fig. 1(c). where Z is the reduced shock impedance between the target material
The target material was assumed to be elastic–plastic homogeneous and the confining medium, α is the portion of absorbed energy
isotropic. Since the strain rate during LSP is very high (106 s− 1), the contributing to the thermal energy of the plasma, E is the laser energy, R
Johnson-Cook model is considered in the simulation which can be is the radius of the laser spot, and τ is the laser pulse width.
expressed as [26,30,34]: Since the pressure induced by laser shock is related to the particle
velocity, a relatively large range of particle velocity of 0.1~1.5 km/s
σ = (A + Bεn )(1 + Clnε̇* )[1 − (T * )m ] (2)
was used to recreate the LSP process in this study. In the LSP process, the
shock pressure is around ~GPa level. The relationship between the
ε̇* = ε̇/ε̇0 (3)
shock pressure and the particle velocity will be discussed and the
T − T0 appropriate particle velocity used for the LSP process will be given. The
T* = (4) procedures of molecular dynamics simulation can be described as fol­
Tm − T0
lows: (1) The whole model was relaxed for 10 ps to make the lattices
where ε is the equivalent plastic strain, A is the yield stress at the stable. (2) A particle velocity was loaded on the piston in the direction of
reference strain rate ε̇0 (106 s− 1), B and exponent n are the strain [0001] orientation for 10 ps. Since the boundaries can lead to rebound
hardening effect, C is the strain rate hardening effect, ε̇* is the normal­ shock waves, the loading time of 10 ps was selected to study the evo­
lution of lattice and dislocation before the wavefront spreading to the
ized effective plastic strain rate, ε̇ is the imposed strain rate, T * is the
boundaries.
homologous temperature, T m is the melting point of the material, and T 0
The common neighbor analysis (CNA) was used to measure the lat­
is the room temperature. It should be noted that the temperature effect
tice distortion and the characteristics of defects induced by laser shock
in the Johnson-Cook model in this study was neglected since the ablative
peening, and the cut-off radius can be calculated by [35]:
layer was sacrificed to protect the target from laser irradiation and the
⎛ √̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅̅ ⎞
target could be shielded from laser thermal effects. The parameters of Ti-
1 4 + 2B2 ⎠
6Al-4V alloy are listed in Table. 1 [30]. Due to the reflection and rhcp
c = ⎝1 + a, B = (c/a)/1.633 (8)
2 3
interaction of the multiple stress wave propagation in the material, the
dynamic solution time in the dynamic step takes much longer than the
action time of the laser shock wave. The dynamic solution time and where a and c are lattice parameters. Since the lattice distortion
static solution time were set to 5000 ns and 1 ms, respectively. After the generated by different lattice defects is significantly different, the
static analysis was accomplished, the treated material reached the final parameter can be used to distinguish different types of lattice defects.
static equilibrium state. The dislocation extraction algorithm (DXA) [38] based on the funda­
mental concept of the Burgers circuit was used to recognize dislocations.
The open-source software Ovito [39] was adopted to display the atom
2.2. Molecular dynamics model in microscale
configurations.
The molecular dynamics simulations were performed to analyze the
evolution of dislocation atoms and dislocation structure, which were

3
C. Gu et al. Optics and Laser Technology 165 (2023) 109629

Fig. 2. Simulation results of the evolution of stress wave propagation in the 5 mm thick target with time (a) 40 ns, (b) 90 ns, (c) 210 ns, (d) 400 ns, (e) 800 ns, (f) 900
ns, (g) 1100 ns, (h) 1400 ns, (i) 1800 ns, (j) 2000 ns, (k) 5000, and (l) 1.005 ms.

Fig. 3. (a) The profiles of stress waves induced by the pressure of 3 GPa, and (b) comparison of the stress wave profiles induced by the pressure of 3 GPa and 6 GPa.

3. Results and discussion 800 ns induced by shock waves with pressures of 3 GPa was recorded as
shown in Fig. 3(a). It is clear that the intensity of the stress wave
3.1. Stress wave on the macroscale gradually decays with the increase of traveling distance which agrees
with the results in Fig. 2. It is because (1) the energy is dispersed by the
When the target material is impacted by laser shock peening, it will lateral release waves, and (2) the elastic–plastic deformation created by
induce a compressive stress wave propagating into the solid target. The the stress wave consumes the energy. Due to the speed of the elastic
axial stress waves propagate along the direction of the target thickness. wave being faster than that of the plastic wave, the profile of stress
Fig. 2 shows the evolution of stress wave propagation in the 5 mm thick waves is characterized by a two-wave structure. The results agree well
target induced by the pressure of 6 GPa. The dynamic characteristics of with the research by Asay et al. [40,41]. Fig. 3(b) compares the stress
stress waves at different times can be observed. At the initial stage, the wave profiles induced by the pressure of 3 GPa and 6 GPa. It shows that a
precursor wave is a planar compressive wave within the whole laser spot higher loading pressure results in higher dynamic stress. At 400 ns, a
loaded by pressure loading. When t = 200 ns, a release wave is generated shock front appears at the depth of 2.2 mm in the target. At 800 ns, its
due to the unloading. The release wave and the compressive stress wave penetrated depth reaches 4.6 mm. Therefore, it can be calculated that
are moving towards the opposite side. At 900 ns, the compressive stress the average velocity of the stress wave is about 6.0 km/s. It can also be
waves encounter the other surface of the target. And then, the obtained that the velocity of the stress wave does not change much with
compressive stress wave is reflected by the free surface and propagates the loading pressure. Meanwhile, it is known that the high-pressure
in the opposite direction. At 1800 ns, the stress wave travels through the shock leads to elastic and plastic compressive wave propagation in the
whole section again and goes back to its starting surface. In the mean­ material [30]. When the high-pressure shock is unloaded, a release wave
time, with the increase of stress wave traveling distance, its intensity is generated, which travels along the same direction as the previous
gradually decays. Therefore, the stress wave in the material completely compressive wave. And then, the elastic release wave catches the plastic
vanishes eventually. compressive wave and reduces the amplitude of the plastic wave.
Under the high pressure on material induced by laser shock, elastic Fig. 4 compares the stress results induced by the pressure of 3 GPa
deformation and plastic deformation are generated. Therefore, elas­ and 6 GPa. The stress induced by the pressure of 6 GPa is higher than
tic–plastic waves occur in the target. The profiles of stress waves within that of 3 GPa. When the peak pressure induced by the stress wave goes

4
C. Gu et al. Optics and Laser Technology 165 (2023) 109629

Fig. 4. Comparison of the stress wave contours induced by the pressure of 3 GPa and 6 GPa: (a) 500 ns, (b) 1000 ns, (c) 2000 ns, (d) 5000 ns, and (e) 1.005 ms.

Fig. 5. Velocity profile of different piston particles in single crystal α-Ti when the laser shock loading time is 3 ps, 5 ps, and 7 ps: (a) 0.5 km/s, (b) 1.0 km/s, (c) 1.5
km/s, and (d) velocity profile of different piston particles when the laser shock loading time is 7 ps.

5
C. Gu et al. Optics and Laser Technology 165 (2023) 109629

Fig. 6. Development of the crystal structure in single crystal α-Ti induced by LSP with different piston particle velocities: (a) 0.5 km/s, (b) 1.0 km/s, and (c) 1.5 km/s.

beyond the dynamic yield strength of the material, the plastic defor­ Pressure is one of the most important shock wave parameters. The
mation starts to occur at the local impacted zone, and its surrounding increase in pressure will increase the density of dislocations. According
zone is elastically compressed simultaneously. After the pressure to the wavefront at different times, it can be calculated that the average
unloading, the surrounding elastically condensed zone tries to get back velocity of the stress wave equals 7.4 km/s, which is higher than 6.0 km/
to its original state but is held back by the plastic deformation. There­ s obtained in the macroscale simulation while agrees well with the ref­
fore, the plastic strain layer of the target is subjected to compressive erences [42,43]. Hao et al. [43] studied the elastic properties of titanium
residual stresses, and the region next to this layer is subjected to tensile by the First-principle method and found the elastic sound velocity was
residual stresses to maintain equilibrium in the target without external between 6.0 and 7.5 km/s due to the difference in pressure. Mack­
force. It can be seen in Fig. 4(d) that the residual stress induced by the enchery and Dongare [42] performed atomistic shock simulations for
pressure of 6 GPa is higher than that of 3 GPa. single-crystal titanium and proposed the elastic sound velocity was be­
tween 7.1 and 7.7 km/s. The reason that the average velocity of the
3.2. Stress wave and dislocations on the microscale stress wave in the microscale is higher than that obtained in the
macroscale may be that the microscale simulation is performed in single
Substructures such as crystal structure and dislocations induced by crystal α-Ti and the stress wave move along [0001] direction. Besides,
laser shock cannot be obtained from the finite element simulation on the according to Hugoniot relations, the loading pressure induced by the
macroscale. Therefore, molecular dynamics simulations were performed particle velocity can be calculated by:
to analyze the effect of LSP. Molecular dynamics simulation results of Pmax = ρUs Up (9)
laser shock peening on α-Ti were visualized and analyzed by using Ovito
software. Fig. 5(a-c) show the particle velocity profiles along the Z where ρ is the density of the target material and Us is the velocity of the
[0001] direction in single crystal α-Ti when piston particles are stress wave. Therefore, the shock pressures induced by laser shock at Up
impacted with speeds of 0.5 km/s, 1.0 km/s, and 1.5 km/s, respectively. of 0.5, 1.0, and 1.5 km/s are 16.7, 33.3, and 50.0 GPa, respectively. The
With the extension of shock loading time, the front position of the laser result shows that increasing shock pressure will also increase the particle
shock wave moves forward along the Z [0001] direction, indicating that velocity. In this case, the loading pressure can be obtained which can be
the laser shock wave moves forward along the loading direction. When used in macroscale simulations conversely. However, it should be noted
the particle velocity is 0.5 km/s, only elastic waves appear in the profile, that these shock pressures are relatively higher than the shock pressure
and the piston particle velocity fluctuates around 0.5 km/s as shown in induced in the actual LSP process. To explain the reason for the higher
Fig. 5(a). It is because the impact pressure is lower than the elastic limit. stress wave velocity, we performed similar MD simulations with shock
A similar phenomenon can be observed in Fig. 5(b) when the particle directions in the other two directions of the HCP structure. The results
velocity is 1.0 km/s. However, with a relatively higher velocity, the show that the average velocity of the stress wave is in the range of
piston particle velocity fluctuates more seriously. In the meantime, the 7.1~7.8 km/s. However, these results are still higher than 6.0 km/s. As
fluctuated velocity may cause the local impact pressure to be higher than mentioned above, the shock pressure in the LSP process is around ~GPa
the Hugoniot elastic limit (HEL) of the material. When the particle ve­ level. However, the shock pressure induced by laser shock at Up of 1.5
locity is 1.5 km/s in Fig. 5(c), the impact pressure exceeds the HEL of the km/s is 50.0 GPa, which is unreachable in the LSP process. Therefore,
material, and a double wave structure with the separation of elastic considering the induced shock pressure, the piston velocity should be
wave and plastic wave appears in the particle velocity profile. It can be below 0.4 km/s to recreate the LSP process. The reason that dislocation
observed that the moving of the laser shock wave leads to a relatively atoms start to form when the particle velocity is no less than 1.0 km/s
higher velocity at the front of the wave.

6
C. Gu et al. Optics and Laser Technology 165 (2023) 109629

Fig. 7. Velocity profile of different piston particles in polycrystal α-Ti when the laser shock loading time is 3 ps, 5 ps, and 7 ps: (a) 0.1 km/s, (b) 0.5 km/s, and (c) 1.5
km/s.

(the shock pressure is 33.3 GPa) might be that the microscale simulation continue to expand forward with the increase of impact loading time.
is performed in single crystal α-Ti. Therefore, the stress wave velocity Dislocation atom density increases with the increase of time. It can be
and the Hugoniot elastic limit are relatively higher than that in the seen that there are more dislocation atoms in α-Ti when the piston ve­
macroscale simulation. locity is higher. During the propagation process of the laser shock wave
To observe and compare the particle velocity profile clearly, we plot in α-Ti, dislocations are mainly formed due to the collapse of vacancy
the velocity profile of different piston particle velocities in single crystal clusters. When the shock pressure exceeds the amplitude of the elastic
α-Ti when the laser shock loading time is 7 ps as shown in Fig. 5(d). It wave, the shock wave will act on the surface of α-Ti repeatedly. Su­
can be seen that when the local impact pressure is lower than the HEL, persaturated vacancies gather into vacancy clusters under the plastic
the velocity fluctuates but there is only one elastic wave. However, when deformation caused by laser shock peening, and gradually evolve into
the local impact pressure is higher than the HEL (such as the curves of sheet-like vacancy clusters. Due to the instability of sheet-like vacancy
1.1, 1.3, and 1.5 km/s), two waves of elastic wave and plastic wave start clusters, they are easy to collapse, providing a position for dislocation
to appear. It should be noted that the higher the particle velocity, the nucleation. Atoms are staggered and polarized under the effect of the
shorter the distance between the elastic front and the plastic front. When shock wave, forming a large number of dislocation nuclei. In addition,
the piston velocity is 0.1 km/s, the shock pressure induced by laser shock after the nucleation of the dislocation, the dislocation will continue to
is calculated to be 3.3 GPa. expand in the slip plane and develop into a more complex dislocation
Fig. 6 shows the crystal structure evolution in single crystal α-Ti structure. Under the continuous effect of the laser shock wave, the shock
under the effect of LSP. Green atoms represent FCC crystal structure, red wave is repeatedly reflected on the interface of different structures,
atoms represent HCP crystal structure, and white atoms represent other resulting in the further development of complex dislocation structure,
crystals with irregular structure. When the particle velocity is 0.5 km/s, and finally, FCC phase and stacking faults (SF) are formed, as shown in
elastic deformation occurs in the crystal structure. The elastic wave front Fig. 6(c). The formation and propagation of dislocations lead to the
continues to expand forward with the increase of time. There are also transformation of the HCP structure into the FCC structure.
limited structure transformations that occurred as circled out in Fig. 6(a) To explain the difference in the stress wave velocity between
which show the formation of the white atoms with irregular structure. It macroscale simulation and microscale simulation, we performed MD
indicates that at low particle velocity, vacancy can also be formed. When simulations in polycrystal α-Ti and the particle velocity profiles along
the particle velocity is 1.0 km/s, the phase transformation from HCP to the Z [0001] direction are shown in Fig. 7. 15 grains were randomly set
FCC does not occur at the beginning. When the impact loading time is 5 in the calculation domain. Similar to the results in Fig. 5, the laser shock
ps, the FCC structure starts to appear at the front of the plastic wave. The wave moves forward along the loading direction. However, the particle
reason is that the fluctuated velocity may cause the local impact pressure velocities in polycrystals are much more unstable than that in a single
to be higher than the HEL of the material during the expansion of the crystal. When the particle velocity is 0.1 km/s in Fig. 7(a), the particle
laser shock waves. It also should be noted that the phase transformation velocity fluctuates seriously. When the particle velocity is 0.5 km/s and
not only forms the FCC phase but also causes the formation of the white 1.5 km/s, a double wave structure with the separation of elastic wave
atoms with irregular structure as shown in Fig. 6(b1) and (b2). When the and plastic wave appears in the particle velocity profile. It can be
particle velocity is 1.5 km/s in Fig. 6(c), atoms with FCC structure calculated that the average velocity of the stress wave is about 6.7 km/s,

7
C. Gu et al. Optics and Laser Technology 165 (2023) 109629

Fig. 8. Development of the crystal structure in polycrystal α-Ti induced by LSP with piston particle velocities of (a) 0.5 km/s and (b) 1.5 km/s.

Fig. 9. Dislocation evolution induced by LSP (a) at the original state of 0 ps, and at different times with piston particle velocities of (b) 0.1 km/s, (c) 0.5 km/s, and (d)
1.5 km/s.

which is still higher than 6.0 km/s obtained in the macroscale simula­ induced by laser shock with piston particle velocities of 0.5 km/s and
tion. However, the average velocity of the stress wave in polycrystals is 1.5 km/s. Green atoms represent FCC crystal structure, red atoms
lower than that in a single crystal. At this time, the shock pressures represent HCP crystal structure, blue atoms represent BCC crystal
induced by laser shock at Up of 0.1, 0.5, and 1.5 km/s are 3.0, 15.1, and structure, and white atoms represent other crystals with irregular
45.0 GPa, respectively. It can be obtained that the grain boundary has a structure. Compared with the single crystal structure evolution in Fig. 6
great influence on the propagation of the stress wave. (a), polycrystal structure evolution presents a different process when the
Fig. 8 shows the crystal structure evolution in polycrystal α-Ti piston particle velocity is 0.5 km/s. It can be seen that the phase

8
C. Gu et al. Optics and Laser Technology 165 (2023) 109629

Fig. 10. The variation of (a) the number of dislocation segments and (b) total dislocation length caused by impact load at different piston particle velocities with
impact time.

shows that the number of dislocation segments and total dislocation


length produced by piston particle velocity of 0.5 km/s are larger than
those of 0.1 km/s. It is interesting to point out that the strain rate can
affect the changing trend of the number of dislocation atoms induced by
plastic deformation. At 9 ps, when the particle velocity of the piston is
1.5 km/s, the dislocation length caused by laser shock is 83,062 Ȧ. The
dislocation length induced by particle velocity of 1.5 km/s is 87% higher
than that of 0.5 km/s.
It is found in the above study that the average velocity of the stress
wave in a polycrystal, when the grain number is 15, is lower than that in
a single crystal. Therefore, we performed MD simulations with different
grain numbers in the same calculation domain and calculated the cor­
responding average velocity of the stress wave as shown in Fig. 11. It can
be seen that the velocity in a polycrystal is always lower than that in a
single crystal. The average velocity of the stress wave has a trend to
decrease with the increase of grain number until a stable value. When
the grain number is above 30, the velocity of the stress wave reaches the
stable value of around 6.0 km/s which is also the velocity in macroscale
simulation. Based on the above results, it can be obtained that there is a
Fig. 11. The relationship between the average velocity of the stress wave and
the grain number in microscale simulation.
certain threshold for the strengthening effect of LSP. The average ve­
locity of the stress wave is related to the grain size. The initial state of the
Ti alloy affects the evolution of stress waves and dislocations and also
transformation from HCP to FCC occurs as shown in Fig. 8(a). With the
affects the final results after the LSP process.
increase of time, the stress wave front continues to expand forward and
the number of the dislocation atoms continues to increase. It also should
4. Conclusions
be noted that in this case, the nucleation of the dislocation or the slip
starts from the grain boundary. The phase transformation not only forms
In this article, the finite element model on the macroscale and mo­
the FCC phase and BCC phase but also causes the formation of white
lecular dynamics model on the microscale are built to study the effect of
atoms with irregular structures. When the particle velocity is 1.5 km/s in
LSP on the evolution of stress waves and dislocations in Ti-6Al-4V alloy.
Fig. 8(b), the grains are finer compared with the untreated structure.
The main conclusions are as follows:
When the piston velocity is higher, there are more dislocation atoms
formed in the simulation domain.
(1) Under the high pressure on material induced by laser shock,
To observe the evolution of dislocation, we used the dislocation
elastic deformation and plastic deformation are generated. The
extraction algorithm in Ovito software to plot the evolution of disloca­
intensity of the stress wave gradually decays with the increase in
tion structure in polycrystal α-Ti under the effect of LSP as shown in
traveling distance. The macroscale simulation shows that the
Fig. 9. It can be seen that dislocations exist at the grain boundary at the
average velocity of the stress wave equals 6.0 km/s and does not
beginning. With the increase of laser shock loading time, the shock wave
change much with the loading pressure. The stress induced by the
front gradually moves along the impact direction, and the dislocation
pressure of 6 GPa is higher than that of 3 GPa.
structure induced by shock loading continues to expand forward.
(2) In the molecular dynamic simulation results, the moving of the
However, when the particle velocity is 0.1 km/s, the dislocation struc­
laser shock wave leads to a relatively higher velocity at the front
ture does not change much. With the increase of the laser shock load
of the wave. The elastic wave front continues to expand forward
when the particle velocity is 0.5 km/s and 1.5 km/s, the dislocation
with the increase of time. When the piston velocity is higher,
structure density increases.
there are more formed dislocation atoms.
In addition, the dislocation information can be reflected by calcu­
(3) The microscale simulation shows that the average velocity of the
lating the number of dislocation segments and total dislocation length.
stress wave is related to the grain number. The average velocity of
Fig. 10(a) and 10(b) show the variation of the number of dislocation
the stress wave in single crystal α-Ti is in the range of 7.1 ~ 7.8
segments and total dislocation length caused by impact load at different
km/s. The average velocity of stress wave in polycrystals is al­
piston particle velocities with impact time. At the same piston particle
ways lower than that in a single crystal and has a trend to
velocity, both the number of dislocation segments and total dislocation
decrease with the increase of grain number until a stable value.
length induced by laser shock have an increasing trend with time. It

9
C. Gu et al. Optics and Laser Technology 165 (2023) 109629

When the grain number is above 30, the velocity of the stress wave deforming, Phys. Solid State 45 (2003) 656–661, https://doi.org/10.1134/
1.1569001.
wave reaches the stable value of around 6.0 km/s.
[16] D.R. Jones, B.M. Morrow, C.P. Trujillo, G.T. Gray, E.K. Cerreta, The α-ω phase
transition in shock-loaded titanium, J. Appl. Phys. 122 (2017) 1–7, https://doi.
CRediT authorship contribution statement org/10.1063/1.4987146.
[17] K. Gurusami, D. Chandramohan, S. Dinesh Kumar, M. Dhanashekar, T. Sathish,
Strengthening mechanism of Nd: Yag laser shock peening for commercially pure
Cheng Gu: Investigation, Data curation, Writing – original draft. titanium (CP-TI) on surface integrity and residual stresses, Mater. Today Proc. 21
Minghua Su: Investigation, Data curation, Writing – review & editing. (2020) 981–987, https://doi.org/10.1016/j.matpr.2019.09.141.
Zenghui Tian: Investigation, Writing – review & editing. Jianhua [18] X. Jin, L. Lan, S. Gao, B.o. He, Y. Rong, Effects of laser shock peening on
microstructure and fatigue behavior of Ti–6Al–4V alloy fabricated via electron
Zhao: Conceptualization, Investigation, Supervision. Yajun Wang: beam melting, Mater. Sci. Eng. A 780 (2020) 139199.
Writing – review & editing. [19] R. Sun, S. Keller, Y. Zhu, W. Guo, N. Kashaev, B. Klusemann, Experimental-
numerical study of laser-shock-peening-induced retardation of fatigue crack
propagation in Ti-17 titanium alloy, Int. J. Fatigue 145 (2021) 106081.
Declaration of Competing Interest [20] M. Pavan, D. Furfari, B. Ahmad, M.A. Gharghouri, M.E. Fitzpatrick, Fatigue crack
growth in a laser shock peened residual stress field, Int. J. Fatigue 123 (2019)
157–167, https://doi.org/10.1016/j.ijfatigue.2019.01.020.
The authors declare that they have no known competing financial
[21] S. Bhamare, G. Ramakrishnan, S.R. Mannava, K. Langer, V.K. Vasudevan, D. Qian,
interests or personal relationships that could have appeared to influence Simulation-based optimization of laser shock peening process for improved
the work reported in this paper. bending fatigue life of Ti-6Al-2Sn-4Zr-2Mo alloy, Surf. Coatings Technol. 232
(2013) 464–474, https://doi.org/10.1016/j.surfcoat.2013.06.003.
[22] A.W. Warren, Y.B. Guo, S.C. Chen, Massive parallel laser shock peening:
Data availability simulation, analysis, and validation, Int. J. Fatigue. 30 (2008) 188–197, https://
doi.org/10.1016/j.ijfatigue.2007.01.033.
Data will be made available on request. [23] P. Peyre, A. Sollier, I. Chaieb, L. Berthe, E. Bartnicki, C. Braham, R. Fabbro, FEM
simulation of residual stresses induced by laser Peening, Eur. Phys. J. Appl. Phys.
23 (2003) 83–88, https://doi.org/10.1051/epjap:2003037.
Acknowledgments [24] P. Peyre, I. Chaieb, C. Braham, FEM calculation of residual stresses induced by
laser shock processing in stainless steels, Model. Simul. Mater. Sci. Eng. 15 (2007)
205–221, https://doi.org/10.1088/0965-0393/15/3/002.
The authors would like to acknowledge the financial supports from [25] R. Voothaluru, C.R. Liu, Finite element analysis of the effect of overlapping impacts
the National Natural Science Foundation of China (NO. 51875062, NO. of laser shock peening within annealed AISI 1053 steel, ASME 2010 Int. Manuf. Sci.
52205336) and the China Postdoctoral Science Foundation (NO. Eng. Conf. MSEC 2010 (2) (2010) 221–228, https://doi.org/10.1115/MSEC2010-
34164.
2021M700567). [26] R. Voothaluru, C. Richard Liu, G.J. Cheng, Finite element analysis of the variation
in residual stress distribution in laser shock peening of steels, J. Manuf. Sci. Eng.
References 134 (2012) 1–8, https://doi.org/10.1115/1.4007780.
[27] W. Zhou, X. Ren, Y. Yang, Z. Tong, E. Asuako Larson, Finite element analysis of
laser shock peening induced near-surface deformation in engineering metals, Opt.
[1] X. Zhang, S. Pfeiffer, P. Rutkowski, M. Makowska, D. Kata, J. Yang, T. Graule, Laser
Laser Technol. 119 (2019), 105608, https://doi.org/10.1016/j.
cladding of manganese oxide doped aluminum oxide granules on titanium alloy for
optlastec.2019.105608.
biomedical applications, Appl. Surf. Sci. 520 (2020), 146304, https://doi.org/
[28] Y. Pei, C. Duan, Study on stress-wave propagation and residual stress distribution
10.1016/j.apsusc.2020.146304.
of Ti-17 titanium alloy by laser shock peening, J. Appl. Phys. 122 (19) (2017)
[2] M.D. Hayat, H. Singh, Z. He, P. Cao, Titanium metal matrix composites: an
193102.
overview, Compos. Part A Appl. Sci. Manuf. 121 (2019) 418–438, https://doi.org/
[29] K.Y. Luo, J.Z. Lu, Q.W. Wang, M. Luo, H. Qi, J.Z. Zhou, Residual stress distribution
10.1016/j.compositesa.2019.04.005.
of Ti-6Al-4V alloy under different ns-LSP processing parameters, Appl. Surf. Sci.
[3] S. Liu, Y.C. Shin, Additive manufacturing of Ti6Al4V alloy: a review, Mater. Des.
285 (2013) 607–615, https://doi.org/10.1016/j.apsusc.2013.08.100.
164 (2019), 107552, https://doi.org/10.1016/j.matdes.2018.107552.
[30] X. Zhang, H. Li, S. Duan, X. Yu, J. Feng, B. Wang, Z. Huang, Modeling of residual
[4] X. Cao, W. He, B. Liao, G. He, Y. Jiao, D. Huang, S. Wang, Effect of TiN/Ti coating
stress field induced in Ti-6Al-4V alloy plate by two sided laser shock processing,
combined with laser shock peening pre-treatment on the fatigue strength of Ti-6Al-
Surf. Coatings Technol. 280 (2015) 163–173, https://doi.org/10.1016/j.
4V titanium alloy, Surf. Coatings Technol. 403 (2020), 126393, https://doi.org/
surfcoat.2015.09.004.
10.1016/j.surfcoat.2020.126393.
[31] X. Meng, J. Zhou, S. Huang, C. Su, J. Sheng, Properties of a laser shock wave in Al-
[5] S. Petronić, K. Čolić, B. Đorđević, D. Milovanović, M. Burzić, F. Vučetić, Effect of
Cu alloy under elevated temperatures: a molecular dynamics simulation study,
laser shock peening with and without protective coating on the microstructure and
Materials (Basel) 10 (1) (2017) 73.
mechanical properties of Ti-alloy, Opt. Lasers Eng. 129 (2020) 106052.
[32] W. Zhou, X. Ren, Y. Yang, Z. Tong, L. Chen, Dislocation behavior in nickel and iron
[6] R. Sun, L. Li, W. Guo, P. Peng, T. Zhai, Z. Che, B. Li, C. Guo, Y. Zhu, Laser shock
during laser shock-induced plastic deformation, Int. J. Adv. Manuf. Technol. 108
peening induced fatigue crack retardation in Ti-17 titanium alloy, Mater. Sci. Eng.
(2020) 1073–1083, https://doi.org/10.1007/s00170-019-04822-8.
A. 737 (2018) 94–104, https://doi.org/10.1016/j.msea.2018.09.016.
[33] R. Fabbro, J. Fournier, P. Ballard, D. Devaux, J. Virmont, Physical study of laser-
[7] S.J. Lainé, K.M. Knowles, P.J. Doorbar, R.D. Cutts, D. Rugg, Microstructural
produced plasma in confined geometry, J. Appl. Phys. 68 (1990) 775–784, https://
characterisation of metallic shot peened and laser shock peened Ti–6Al–4V, Acta
doi.org/10.1063/1.346783.
Mater. 123 (2017) 350–361, https://doi.org/10.1016/j.actamat.2016.10.044.
[34] G.R. Johnson, W.H. Cook, Fracture characteristics of three metals subjected to
[8] M.Y.P. Costa, M.L.R. Venditti, M.O.H. Cioffi, H.J.C. Voorwald, V.A. Guimarães,
various strains, strain rates, temperatures and pressures, Eng. Fract. Mech. 21
R. Ruas, Fatigue behavior of PVD coated Ti-6Al-4V alloy, Int. J. Fatigue. 33 (2011)
(1985) 31–48, https://doi.org/10.1016/0013-7944(85)90052-9.
759–765, https://doi.org/10.1016/j.ijfatigue.2010.11.007.
[35] A.P. Thompson, H.M. Aktulga, R. Berger, D.S. Bolintineanu, W.M. Brown, P.
[9] J. Zhai, X. Yao, Z. Xu, H. Chen, Electric fatigue properties of Pb(Zr, Ti)O3 thin films
S. Crozier, P.J. in ’t Veld, A. Kohlmeyer, S.G. Moore, T.D. Nguyen, R. Shan, M.
grown on LaNiO3 buffer Pt/Ti/SiO2/Si substrate by metalorganic chemical vapor
J. Stevens, J. Tranchida, C. Trott, S.J. Plimpton, LAMMPS - a flexible simulation
deposition, Integr. Ferroelectr. 75 (2005) 47–54, https://doi.org/10.1080/
tool for particle-based materials modeling at the atomic, meso, and continuum
10584580500413251.
scales, Comput. Phys. Commun. 271 (2022) 108171.
[10] Y. Liao, C. Ye, G.J. Cheng, [INVITED] A review: warm laser shock peening and
[36] P. Hirel, Atomsk: a tool for manipulating and converting atomic data files, Comput.
related laser processing technique, Opt. Laser Technol. 78 (2016) 15–24, https://
Phys. Commun. 197 (2015) 212–219, https://doi.org/10.1016/j.cpc.2015.07.012.
doi.org/10.1016/j.optlastec.2015.09.014.
[37] G.J. Ackland, Theoretical study of titanium surfaces and defects with a new many-
[11] P. Peyre, C. Carboni, A. Sollier, L. Berthe, C. Richard, E. de Los Rios, R. Fabbro,
body potential, Philos. Mag. A Phys. Condens. Matter, Struct. Defects Mech. Prop.
New trends in laser shock wave physics and applications, High-Power Laser
66 (1992) 917–932, https://doi.org/10.1080/01418619208247999.
Ablation IV. 4760 (2002) 654. 10.1117/12.482138.
[38] A. Stukowski, V.V. Bulatov, A. Arsenlis, Automated identification and indexing of
[12] G.I. Kanel, G.V. Garkushin, A.S. Savinykh, S.V. Razorenov, T. de Resseguier, W.
dislocations in crystal interfaces, Model. Simul. Mater. Sci. Eng. 20 (8) (2012)
G. Proud, M.R. Tyutin, Shock response of magnesium single crystals at normal and
085007.
elevated temperatures, J. Appl. Phys. 116 (14) (2014) 143504.
[39] A. Stukowski, Visualization and analysis of atomistic simulation data with OVITO-
[13] L. Lan, X. Jin, S. Gao, B. He, Y. Rong, Microstructural evolution and stress state
the Open Visualization Tool, Model. Simul. Mater. Sci. Eng. 18 (1) (2010) 015012.
related to mechanical properties of electron beam melted Ti-6Al-4V alloy modified
[40] J. Lipkin, J.R. Asay, Reshock and release of shock-compressed 6061–T6 aluminum,
by laser shock peening, J. Mater. Sci. Technol. 50 (2020) 153–161, https://doi.
J. Appl. Phys. 48 (1977) 182–189, https://doi.org/10.1063/1.323306.
org/10.1016/j.jmst.2019.11.039.
[14] X.C. Zhang, Y.K. Zhang, J.Z. Lu, F.Z. Xuan, Z.D. Wang, S.T. Tu, Improvement of
fatigue life of Ti-6Al-4V alloy by laser shock peening, Mater. Sci. Eng. A. 527
(2010) 3411–3415, https://doi.org/10.1016/j.msea.2010.01.076.
[15] G.I. Kanel, S.V. Razorenov, E.B. Zaretsky, B. Herrman, L. Meyer, Thermal
“softening” and “hardening” of titanium and its alloy at high strain rates of shock-

10
C. Gu et al. Optics and Laser Technology 165 (2023) 109629

[41] H. Huang, J.R. Asay, Reshock and release response of aluminum single crystal, [43] Y. Hao, J. Zhu, L. Zhang, J. Qu, H. Ren, First-principles study of high pressure
J. Appl. Phys. 101 (6) (2007) 063550. structure phase transition and elastic properties of titanium, Solid State Sci. 12
[42] K. Mackenchery, A. Dongare, Shock hugoniot behavior of single crystal titanium (2010) 1473–1479, https://doi.org/10.1016/j.solidstatesciences.2010.06.010.
using atomistic simulations, AIP Conf. Proc. 1793 (2017), https://doi.org/
10.1063/1.4971589.

11

You might also like