Indentation-Based Analysis of Interfacial Properties of The Film-Substrate Structure and Instrument Development

Download as pdf or txt
Download as pdf or txt
You are on page 1of 17

CHAPTER

Indentation-based
analysis of interfacial
31
properties of the
film-substrate structure
and instrument
development
Xu, Long, PhD, Northwestern Polytechnical
University, Professor, School of Mechanics,
Civil Engineering and Architecture, Xi'an,
China, Yangtze Delta Region Institute of
Tsinghua University, Scientist, Jiaxing, China

Indentation-based analysis of interfacial


properties of the film-substrate structure and
instrument development
Ziyi Shen1, Xu Long1,2, Yutai Su1, Xin Wan2, Fengrui Jia2
1
School of Mechanics, Civil Engineering and Architecture, Northwestern
Polytechnical University, Xi’an, China
2
Yangtze Delta Region Institute of Tsinghua University, Zhejiang,
Jiaxing, China
*Corresponding author. Email address: xulong@​nwpu​.edu​.cn

I Introduction
As multi-​material science research progresses, traditional mechanical
material characterization methods for substrate materials increasingly
fall short of current testing requirements. Nanoindentation technique,
known for its use of its ultra-​low loads and monitoring sensors with dis-
placement resolution better than 1 nm, has been widely used in measuring
the mechanical properties of substrate materials at micro and nanoscale
(Lin et al., 2009; Nagamura and Tanaka, 2009; Gizatulin and Sultanov, 2018;
Long, Li, et al., 2022). Numerical simulations have been used to study the
intrinsic relationship between the morphology, mechanical properties
of elastoplastic materials, and the applied load–​penetration depth curve
(P–​h curve) has been studied through dimensionless analysis (Long, Hu,
1
2 Small-​Scale Mechanical Testing

et al., 2019; Long, Du, et al., 2018; Long, Zhang, et al., 2019). The plastic zone
radius proposed in dimensionless analysis helps overcome the challenging
uniqueness issue (Long, Shen, et al., 2022). Given the difficulties of tradi-
tional numerical simulations, researchers have proposed deep learning
methods to learn the time series of P–​h curves, thereby mapping the P–​h
curves to the stress-​strain response of elastoplastic materials (Long et al.,
2023). Sintered silver nanoparticles (AgNPs) can form thin films with suf-
ficiently high thermal conductivity, which can be widely applied in high-​
power fields. Long et al. (Long et al., 2021; Long, Tang, et al., 2018) evaluated
the stress-​strain relationship of AgNPs materials under different strain
rates using Berkovich nanoindentation method, revealing the advantage of
evaluating the sensitivity of AgNPs to strain rates. Furthermore, due to the
intrinsic differences between layered materials, mechanical and thermal
effects can lead to mismatches in stress and strain between two materials,
resulting in functional failure of thin film materials (Dean, Aldrich-​Smith
and Clyne, 2011).
Due to the advantages of simplicity in adhesive use, lightweight, corro-
sion resistance, and good fatigue performance, the film-​substrate structure
is widely used in industries such as automotive, aerospace, and micro-
electronics (He, 2011). For more reliable application of the film-​substrate
structures, precise analysis and prediction of the influence of the cohesive
force of the film on the substrate are essential. Cohesive force models are
used to simulate and predict the cracking and crack propagation processes
of adhesive structures. This model was first proposed by Barenblatt (Baren-
blatt, 1959) and Dugdale (Dugdale, 1960). When the tensile force reaches a
critical value, cracks begin to initiate, then the tensile force decreases, and
the cracks propagate forward until complete detachment. Cohesive force
models are also widely used in analyzing the fracture failure of metals,
ceramics, polymer materials, and composite materials. They can accurately
analyze phenomena such as the plastic zone at the crack tip, cracking, and
creep behavior (Feraren and Jensen, 2004; Park, Paulino and Roesler, 2008).
Campilho et al. (Campilho et al., 2013) studied the influence of cohesive
force model shape (triangular, exponential, or trapezoidal) on modeling
of single-​lap adhesive joint with thin adhesive layer. The study showed
that the shape of the cohesive force model has a significant effect on the
overlap structure bonded with ductile adhesives but has little effect on
brittle adhesives. Currently, widely adopted cohesive force models include
the parabolic model in exponential form, the elastic perfectly plastic tri-
linear model (Feraren and Jensen, 2004), and the elastic bilinear model
(Park, Paulino and Roesler, 2008). Tvergaard and Hutchinson (Tvergaard
and Hutchinson, 1992) proposed elastic and ideal plastic trilinear cohesive
force models, while Camacho and Oritiz proposed a bilinear cohesive force
model for brittle fracture. Evers and Parks (Evers et al., 2002) established
Chapter 31 | Indentation-​based analysis of i 3

a crystal plasticity hardening model based on geometrically necessary


dislocation-​related factors. Bittencourt and Needleman (Bittencourt et al.,
2003) explained the scale effect from the dislocation theory and compared
the results with the model of Gurtin (Gurtin, 2000).
The film-​substrate system, by covering the carrier with one or more
layers of films with special properties, endows the entire system with
additional functionalities. These surface-​functional films, although their
thickness may be only a fraction of the overall structural size, ranging
from a few tenths to a few hundredths, can confer higher wear resistance,
corrosion resistance, and high-​temperature resistance to the components.
This enables the components to operate in harsh working environments,
significantly enhancing their performance and service life. These excellent
properties make thin film materials highly promising in surface protection
technologies for microelectronics and microelectromechanical systems
(MEMS), such as wear resistance, corrosion resistance, and oxidation resis-
tance. However, the use of film-​substrate technology also presents many
challenges, such as the tendency of films to fracture, buckle, and delaminate
under external loads. These issues severely hinder the further promotion
and application of thin film materials.
This chapter introduces the damage mechanisms and evolution laws
of cohesive layers. It conducts nanoindentation experiments using the
developed indentation instrument by the research team, tests the surface
of encapsulation materials, and obtains the P–​h curve. Through reverse
analysis, important information regarding the mechanical properties of
materials can be obtained. The analysis of the mechanical properties of
the film-​substrate interface is of significant importance in various appli-
cation fields, including microelectronics, optical coatings, and biomedical
devices. In microelectronics, understanding the mechanical properties of
the film-​substrate interface can help optimize the design and performance
of integrated circuits. In optical coatings, a deep understanding of the film-​
substrate interface can improve the wear resistance and stability of optical
components. This indentation-​based analysis method not only provides
quantitative data on the mechanical properties of film-​substrate interface
materials but also offers crucial guidance for optimizing material design
and engineering applications.

Ⅱ Cohesive interface
Various parameters within the cohesive layer describe the mechanical state
of the cohesive zone. The traction–​separation relationship within the cohe-
sive zone mainly includes bilinear, exponential, and trapezoidal relation-
ships. Figure 31.1 shows the typical traction–​separation curve, illustrating
the relationship between the normal stress on the crack surface and the
4 Small-​Scale Mechanical Testing

Fig. 31.1 Typical traction–​


separation curve Typical traction–​
separation curve No Permission
Required

displacement on the crack surface within the cohesive zone. In 1998, Mi et


al. (Mi et al., 1998) proposed a method to model the progressive mixed-​mode
layering in fiber composite materials. This approach was incorporated into
nonlinear finite element methods, based on the use of interface elements
and the softening relationship between stress and relative displacement.
This model, known as the cohesive zone model, relates the area under the
stress-​strain curve to the critical fracture energy. Subsequently, Camanho
(Camanho, Davila and de Moura, 2003) and Turon (Turon et al., 2006) devel-
oped a new equation based on continuum damage mechanics to simulate
the initiation and propagation of layering, proposing their own bilinear
cohesive zone model. This model was implemented in finite element for-
mulations, and numerical predictions were compared with experimental
results from composite material specimens and structural components. To
meet the needs of different materials, many scholars have further improved
these models. However, the fundamental characteristics of these models are
consistent: the interface stress within the cohesive zone initially increases
with increasing interface displacement until it reaches a peak, after which
it decreases to zero with further interface displacement, resulting in inter-
face separation and crack formation. The decreasing trend in stress after
reaching the peak indicates material softening during the fracture process.

Starting with the premise that stress at the crack tip cannot be infinitely
large, Dugdale (Dugdale, 1960) introduced the material's yield stress as the
cohesive stress but stipulated that the stress required to open the crack must
be significantly greater than the equivalent stress under multiaxial stress
states. Delamination between thin films and substrates is a critical factor
in device failure, making in-​depth study of delamination phenomena in
thin films highly significant. In order to better simulate this delamination
behavior between thin films and substrates, a layer of cohesive elements
was added between the thin film and the substrate in the finite element
model of this study.
Chapter 31 | Indentation-​based analysis of i 5

In the study of the cohesive zone, it was found that since the attraction
between atoms is a function of the distance between atoms being pulled
apart, a certain crack tensile displacement δ less than a certain critical
value can exist on the crack surface in this tiny region. The stress σ on the
crack surface is thus a function of this opening displacement. Therefore,
within the cohesive zone, the normal stress on the crack surface is typically
defined as a function of the displacement on the crack surface, known as
the Traction–​Separate Law (TSL) on the crack interface, or the constitu-
tive relationship of the cohesive zone model. The functional expression
is as follows:
​σ ​= f​(δ)​​ (31.)
From an energy perspective, the energy released during the formation of
new crack surfaces in the cracking process is defined as its fracture energy
φ. The general formula is:
​ϕ ​= ∫ σdδ ​= ∫ f​(δ)​dδ​ (31.)

2.1 Traction–separation relationship


The TSL constitutive relationship describes the relationship between stress
and opening displacement in the cohesive zone, but there is no unified
fixed pattern, and different scholars adopt different constitutive relation-
ships. The general form of the TSL constitutive relationship is as follows:
when the cohesive zone is subjected to loading, stress on the cohesive crack
surface increases with increasing opening displacement until it reaches a
limit value. The stress reaching the limit value means that the material's
bearing capacity reaches its maximum value, stiffness begins to degrade,
and internal damage occurs. The generation of damage causes stress to
decrease with increasing opening displacement, decreasing to zero when
the material is completely damaged and fails, and the cohesive zone com-
pletely fractures and becomes a free surface at that point. At this point,
the fracture energy at the material point also reaches its maximum value,
which is the area under the curve of the TSL constitutive relationship. In
summary, the cohesive force model has three main parameters: maximum
stress, fracture energy, and maximum interface opening displacement.
Because there is a functional relationship between stress and opening
displacement, once the shape rules of the TSL constitutive relationship
curve are given, only two of the three parameters are independent. Many
researchers (Volokh, 2004) believe that two independent parameters in the
cohesive force model, fracture energy combined with either maximum
stress (fracture strength) or maximum interface opening displacement,
are sufficient to characterize the mechanical behavior of the cohesive zone.
On the interface, the shape of the TSL constitutive relationship curve is
6 Small-​Scale Mechanical Testing

not as important. However, recent studies have shown that the shape of
the TSL constitutive relationship curve in the cohesive zone has a signifi-
cant impact on the macroscopic mechanical behavior of crack surfaces in
material structures (Alfano, 2006).
In the continuous application and research of cohesive force models,
various TSL constitutive relationships have been developed. Among them,
the most commonly used ones include bilinear, trapezoidal (tri-​linear),
exponential, and polynomial TSL constitutive relationships. In the material
property module of the finite element software ABAQUS, only bilinear and
exponential TSL constitutive relationships are included. If one wishes to
utilize other forms of TSL constitutive relationships for numerical simu-
lations in ABAQUS, they must make use of the subroutine interface pro-
vided by ABAQUS and write the necessary subroutine for the desired TSL
constitutive relationship implementation.

2.2 Damage initiation criterion of cohesive force


model
Eqs. (31.3) and (31.4) represent the governing equations for the traction-​
separation relationship in the classical bilinear adhesive force model.
⎧​σ
_
​  ​​
​ max ​  δ​δ ​≤ ​δ​  0t​  ​)​
⎪ ​δ​  0n ​​  (
​​σ​  n​​ ​= ⎨​ ​   ​ ​​​ (31.)
⎪​σ​  ​​ ​ _ ​δ​  fn​ ​  − δ
​​(δ ​> ​δ​  t​  ​)​
0
⎩ max ​δ​  fn​ − ​ ​  δ​  0n ​​ 
⎧ ​τ_ ​  max​​


σ ​  ​​ ​  ​  δ​(δ ​≤ ​δ​  0t​  ​)​
max
​δ​  0n ​​ 
​​τ​  n​​ ​= ⎨ ​ ​   ​ ​​​ (31.)
⎪​τ​  ​​ ​ _ ​δ​  fn​ ​  − δ
​​(δ ​> ​δ​  t​  ​)​
0
⎩ max ​δ​  fn​ − ​ ​  δ​  0n ​​ 

where σn represents the normal stress value, and τn represents the tan-
gential stress value. σ max and τ max are the maximum normal and tangential
stress values, respectively, corresponding to the crack interface opening
displacement values δ​​ ​  0n ​​a​  nd δ​​ ​  0t​  ,​​ which are characteristic displacement values
of the cracking process. When reaching their maximum values, the stress
begins to decrease to zero, indicating the completion of crack initiation,
with corresponding displacement values of the final crack initiation dis-
placements ​​δ​  fn​​a​  nd ​​δ​  ft​​.​  The critical fracture energy thresholds in all direc-
tions, ​​ϕ​  cn​ ​​and ​​ϕ​  ct​  ​​, are calculated as:
​​ϕ​  cn​ ​ ​= _ ​  1 ​ ​σ​  max​​  · ​δ​  fn​​​  (31.)
_ 2
1
​​ϕ​  t c
​  ​ =
​ ​  ​ ​ τ ​  ​​  · ​δ​  f
​​ ​  (31.)
2 max t

When simulating material damage using cohesive force models, the


material failure is determined based on the damage initiation criteria,
Chapter 31 | Indentation-​based analysis of i 7

indicating the onset of initial damage, which implies the degradation of


material stiffness. The conditions for degradation initiation are when the
values of stress and strain reach predefined thresholds for damage occur-
rence. Six types of damage initiation criteria are provided by ABAQUS in
the damage for TSLs, namely: maximum nominal stress criterion, qua-
dratic nominal stress criterion, maximum nominal strain criterion, qua-
dratic nominal strain criterion, maximum principal stress criterion, and
maximum principal strain criterion. Among these, the first four criteria are
generally used for simulating composite material delamination, while the
latter two are used for simulating discontinuous problems such as crack
propagation using extended finite element method. For models based on
cohesive elements, the following four types of damage initiation criteria
are utilized: maxs damage is the quadratic nominal strain criterion, where
damage begins when the sum of squares of the nominal strain ratios equals
1. In this study, the initial damage criterion adopted for the elements is
the first stress criterion, specifically:
​σ​  n​​ _ ​σ​  s​​ _ ​σ​  t​​
{ n
​max​ _​  max

σ ​  ​ 
​  , ​  max
​ ​
σ ​  s
​  ​ ​
σ ​  t​  ​}
​  , ​  max ​ ​=1​ (31.)

2.3 Degradation evolution law of cohesive force


models
The crack propagation in cohesive force models is determined based on
damage evolution criteria. Damage evolution refers to the degradation
process of material mechanical properties after damage initiation, typically
described by stiffness degradation. Introducing a stiffness degradation
coefficient D, which ranges from 0 to 1, where D = 0 represents the mate-
rial in an intact state, and D = 1 represents complete material failure. The
stiffness of the damaged material is expressed by the following equation:
​​Ki​  ​​ ​= ​Ki​  0​  ​​(1 − D)​​ (31.)
where ​​Ki​  0​  ​​represents the stiffness value when the material is intact. The
crucial aspect of describing material damage evolution lies in the calcula-
tion of D. For cohesive force models, the stiffness degradation coefficient
can be calculated through various constitutive models, with the most com-
monly used being the bilinear cohesive force model.
The damage evolution law of cohesive force models is divided into two
types: based on equivalent displacement and equivalent energy. For the
evolution law based on equivalent displacement, damage evolution refers
to the decay of stiffness, typically represented by the damage variable D.
When the material reaches the fracture damage state, the damage evolution
stage begins, with an initial D = 0 indicating intact material.
8 Small-​Scale Mechanical Testing

​δ​  fm(​ ​​ ​δ​  max


​  ​ − ​δ​  0m ​)​  ​
​D ​= ___________
​ max m f ​​ (31.)
​δ​  m​  (​​ ​δ​  m​ ​ − ​δ​  0m ​)​  ​

where δ​​ ​  fmrepresents


​​​  the relative displacement occurring at the structure
or interface failure, δ​​ ​  0m ​​r​  epresents the relative displacement at the initiation
of damage in the structure or interface, and ​​δ​  mmax ​  ​​represents the relative
displacement during the damage evolution process in the structure or
interface. The equivalent displacement ​​δ​  m​​​is commonly used to describe
damage evolution during normal and tangential separation on the inter-
face. It can be expressed as
_

​​δ​  m​​ = ​ ​δ​  2n ​ ​  + ​δ​  2s​  ​ + ​δ​  2t​  ​ ​​,
indicating that when the combined separation displacements in three
directions reach a certain effective displacement, the damage reaches its
maximum, leading to structural fracture or crack propagation.

III Finite element interface mechanics simulation


Finite element cohesive interface mechanics simulation is used to simu-
late the mechanical behavior of interfaces within materials through finite
element analysis. This simulation method is commonly employed to study
interface effects in complex multiphase materials or composite materials,
enabling a better understanding of material responses and performance
under stress conditions. When performing finite element modeling of
nanoindentation using ABAQUS, the process of a conical indenter press-
ing into the material is a typical axisymmetric problem. Due to the faster
algorithm speed and satisfactory accuracy of axisymmetric problem solving,
an axisymmetric model can be chosen to simulate the nanoindentation
process. According to Saint Venant's principle, the deformation zone of the
tested material is concentrated near the indenter, while stress and strain
in the distant material tend to zero. Therefore, local material analysis can
be conducted. In this case, the substrate is a cylinder with a height of 3000
μm and a radius of 3000 μm, consisting of 49,349 axisymmetric elements.
Since the maximum penetration depth is 6 μm, the mesh size of the sub-
strate is sufficiently small to simulate semi-​infinite space. Campilho et al.
(Campilho et al., 2013; Dattaguru et al., 1994; Krueger and O’Brien, 2001)
pointed out that to ensure numerical convergence and accuracy, the mesh
density should be sufficiently high. The indenter used in the finite element
model is a Berkovich indenter. To simplify the modeling complexity, it
is approximated as a conical indenter with an angle of 70.3° based on the
equivalent area criterion. That is, when a non-​conical indenter and a conical
indenter act on the same material and penetrate to the same depth, if the
vertical projected area of contact between the two indenters and the tested
Chapter 31 | Indentation-​based analysis of i 9

Fig. 31.2 Schematic diagram of


film–​
s ubstrate cohesive model
Schematic diagram of film–​substrate
cohesive model No Permission
Required

material is the same, then the two indenters are considered to follow the
principle of equivalent area. This criterion is particularly suitable for the
contact problem of nanoindentation in this study. The entire model con-
sists of 94,054 nodes and 93,916 elements, with a maximum mesh size of
0.05 mm and a minimum mesh size of 0.001 mm. As shown in Figure 31.2,
the thin film is made of linear elastic metal material with a thickness of
10 μm, a Young's modulus range of 10 to 50 GPa and a Poisson's ratio range
of 0.1 to 0.5; the range for the metal substrate is from 80 to 120 GPa. For
the Berkovich indenter, an elastic model with a Young's modulus of 1060
GPa and a Poisson's ratio of 0.07 is used to describe the diamond material.
The cohesive zone model parameters are taken from reference (Li and Suo,
2007) as τmax = 100 MPa,δ = 0.1 μm,𝜙 = 5 J/m2.
The cohesive zone model is commonly used to simulate the bonding
behavior of interfaces. Its main feature is that it can not only predict the
fracture behavior of structures but also be used to study crack initiation
and propagation. Traditional cohesive zone models are independent of
loading history when studying delamination crack propagation, cannot
record the damage behavior of interfaces, and are reversible. Moreover,
they cannot handle the coupling behavior between normal and tangential
directions well when the normal opening displacement is negative (due
to mutual penetration of the thin layer and substrate under compressive
loading). Therefore, the cohesive zone model used in this study employs
the bilinear cohesive zone model proposed by Camanho (Camanho, Davila
and de Moura, 2003) for interface failure analysis. In the two-​dimensional
10 Small-​Scale Mechanical Testing

Fig. 31.3 Traction–​separation curves


of film–​s ubstrate cohesive model
under different cohesive parameters:
(a) Damage initial displacement; (b)
Strength; (c) Fracture energy Traction–​
separation curves of film–​s ubstrate
cohesive model under different cohesive
parameters: (a) Damage initial displace-
ment; (b) Strength; (c) Fracture energy
No Permission Required

finite element model, artificial viscosity effects are introduced to enhance


numerical convergence and explore the effects of different cohesive zone
parameters, such as the strength of cohesion, on numerical simulation con-
vergence and computational efficiency. As shown in Figure 31.3, interface
failure accompanied by obvious failure evolution occurs when the cohesive
zone strength T = 100 MPa. Under the same cohesive zone strength and
stiffness, different stiffness determines the occurrence of initial displace-
ment loss. The earlier the initial displacement loss occurs, the earlier the
interface failure occurs.
Comparing the P–​h curves of the bilinear cohesive zone model under
various damage initiation displacements, it is observed that as the inden-
tation depth increases, the required load also increases, a phenomenon
also noted by Alfano et al. (Mi et al., 1998; Turon et al., 2007; Liu and Islam,
2013). As illustrated in Figure 31.4, comparing the P–h​ curves of the bilinear
cohesive zone model under different cohesive zone strengths, although the
strength of cohesion is positively correlated with the required load, this
influence is not very pronounced in the curves. The setting of different
fracture energies directly affects the ease of later damage evolution. If the
fracture energy is too large, there is a significant reserve of energy within
the cohesive zone to counteract the damage effects generated during the
indentation process.
Chapter 31 | Indentation-​based analysis of i 11

Fig. 31.4 P–​ h Response of film–​


substrate cohesive model with differ-
ent cohesive parameters: (a) Damage
initial displacement; (b) Strength; (c)
Fracture energy P–​h Response of film–​
substrate cohesive model with different
cohesive parameters: (a) Damage initial
displacement; (b) Strength; (c) Fracture
energy No Permission Required

IV Development of nanoindentation
instrumentation
Based on the mechanical performance characteristics of electronic pack-
aging materials, this study developed a set of dimensionless functions
comprising parameters such as temperature, strain rate, elastic modulus,
yield stress, void volume fraction, characteristic stress, characteristic strain,
and plastic zone radius, along with corresponding indentation analysis
reversion algorithms. These were used to evaluate the in-​situ interface
properties of electronic packaging materials based on TSLs. By considering
the plastic zone radius under the indenter as a key variable, the constitutive
parameter values corresponding to different microstructural morphol-
ogies of packaging materials were obtained from the P–​h curves of the
indentation, thus avoiding the uniqueness issue in traditional indentation
reversion analysis. Furthermore, statistical machine learning and analysis
were performed on a large amount of experimental and simulation data
to further refine the quantitative deformation laws of the plastic zone size
under the indenter, depicting the unique characteristics of indentation
conditions in three-​dimensional phase diagrams and proposing a funda-
mental solution to the uniqueness problem of constitutive relationships
and parameters of packaging materials. An efficient testing instrument,
the electronic packaging material mechanical property nanoindentation
12 Small-​Scale Mechanical Testing

Fig. 31.5 Machine learning-​d riven


package material constitutive param-
eter indentation instrument Machine
learning-​driven package material consti-
tutive parameter indentation instrument
No Permission Required

tester, and its supporting testing instruments were developed for in-​situ
measurement of local material mechanical properties.
Furthermore, based on the proposed reversion algorithm for electronic
packaging material mechanical properties, this study aims to develop an
efficient testing instrument capable of in-​situ measurement of local mate-
rial mechanical properties–​the electronic packaging material mechanical
property nanoindentation tester. As shown in Figure 31.5, the instrument
mainly consists of components such as the indenter, connecting rod, spec-
imen stage, upper computer, lower computer, force sensor, displacement
sensor, fixation device, and display screen. Leveraging extensive finite
element simulations and optimized machine learning algorithms con-
ducted in this project, a complete set of inversion algorithms has been
embedded into the control system of the instrument's upper computer.
When users operate the instrument, the surface of the specimen to be
tested is fixed facing upward on the specimen stage, and the indentation
conditions are set via the display screen. The upper computer controls
the force sensor and displacement sensor, while the lower computer con-
ducts the indentation test on the specimen. Based on the P–​h curve of the
indentation process, the upper computer can instantaneously calculate
the elastic and plastic mechanical parameters of the material, obtain the
material's complete constitutive characteristics, and display and save the
data files on the display screen.

V Material parameter determination based on


nanoindentation method
In the manufacturing of printed circuit boards (PCBs), underfill adhesive
is a material used to fill the gaps between layers of the circuit board. PCBs
Chapter 31 | Indentation-​based analysis of i 13

Fig. 31.6 P–​h curve corresponding


to five types of underfill adhesives
Schematic diagram of the classical P–​h
curve No Permission Required

are substrates used to support and connect electronic components, typically


made of insulating materials, while underfill adhesive is used to fill the
gaps between different layers of the circuit board to enhance mechanical
strength, prevent moisture ingress, and resist chemical intrusion. Under-
fill adhesive is commonly employed to enhance the mechanical strength
of the PCB, hence, measuring its mechanical properties is crucial. This
includes testing parameters such as tensile strength, flexural strength,
hardness, etc., to ensure that the underfill adhesive effectively reinforces
the mechanical stability of the PCB. In this study, five types of underfill
adhesives commonly used in surface-​mount technology for enhancing
the mechanical resistance of solder joints on PCBs were tested using the
indentation instrument and its associated testing equipment, and the P–​h
curves obtained were shown in Figure 31.6. Parameter reversion was per-
formed using the mean values to obtain their elastic parameters.
As shown in Figure 31.7, Sc represents the contact stiffness, which is con-
sidered as the contact area between the indenter and the sample, defined
by Eq. (31.10).
_
​​Sc​  ​​ ​= _​  2_​ ​Er​  ​​ ​√​Ac​  ​​ ​​ (31.)
​√π  ​

The elastic modulus E can be expressed as follows:


14 Small-​Scale Mechanical Testing

Fig. 31.7 Schematic diagram of


the classical P–​h curve Schematic
diagram of the classical P–​h curve No
Permission Required

Table 31.1 Reversion results of the five underfill adhesives parameters


No. hc (nm) Er (MPa) E (MPa)
1# 0.00578 7297.18 6346.42
2# 0.00491 8946.54 7445.33
3# 0.00630 6183.58 5657.57
4# 0.00912 4027.37 3677.82
5# 0.00423 6524.71 5971.46
No Permission Required

(​Er​  ​​ ​Ed​  ​​ )
1 − ​υ​  2d​ ​
​  1 ​ − ​_
​E ​= (​ 1 − ​υ​​  2)​​ / ​ _  ​​​ (31.)

where Ed and υd represent the elastic modulus and Poisson's ratio of


the indenter, respectively. The indenter used in this study is the diamond
Berkovich indenter, with a value of 1, 141 GPa for Ed and 0.07 for υd . The
(Long et al., 2021; Long, Li, et al., 2022; Long, Shen, et al., 2022; Long, Tang, et
al., 2018; Long, Zhang, et al., 2019)reverse analysis results of the five underfill
adhesives are shown in the Table 31.1.

VI Conclusions
Film-​substrate systems are significant in many critical engineering applica-
tions. Accurate analysis of the cohesive effects on the mechanical response
of the film-​substrate structure is challenging yet critical for reliable appli-
cations. In this study, the damage mechanism and evolution behavior
of the cohesive layer in the film-​substrate structure are presented. As a
Chapter 31 | Indentation-​based analysis of i 15

benchmark simulation, a finite element model of the film-​substrate model


with a perfect interface between film and the substrate is established as an
axisymmetric problem of a rigid Berkovich indenter on a semi-​infinite sub-
strate. Subsequently, the comparison with the simulations considering the
cohesive layer between film and the substrate underlines the importance of
considering the cohesive force when evaluating the indentation response
of a typical soft film/rigid substrate in the electronic packaging industry.
To the end of estimating the in-​situ interfacial properties following the
TSL, an efficient test instrument is developed based on the characteristics
of mechanical properties of electronic packaging materials. By utilizing the
indentation instrument and its supporting test equipment, five types of
underfill adhesives are tested, which are commonly adopted to strengthen
the mechanical resistance of solder joints by the surface-​mount technology
to assemble the electronic components on printed circuit boards. Based
on the loading and unloading curves in the P–​h response, the elastic and
plastic properties of the five underfill adhesives are estimated by the finite
element simulations. The obtained elastic and plastic properties of under-
fill materials can be further used to establish a reliable numerical model to
predict the mechanical response in the film-​underfill-​substrate structure.
This makes it possible to theoretically unveil the underlying mechanism of
cohesive layer to influence the film-​substrate structure with different film
and substrate properties. The proposed approach of this indentation-​based
analysis not only provides quantitative data on the mechanical properties
of the film-​substrate interface, but also provides important guidance for
optimizing material design and engineering applications.

REFERENCES
Alfano, Giulio (2006) “On the influence of the shape of the interface law on the application of
cohesive-​zone models,” Composites Science and Technology. Elsevier BV, 66(6), pp. 723–​730.
doi: 10.1016/j.compscitech.2004.12.024.
Barenblatt, G.I (1959) “The formation of equilibrium cracks during brittle fracture. General
ideas and hypotheses. Axially-​symmetric cracks,” Journal of Applied Mathematics and
Mechanics. Elsevier BV, 23(3), pp. 622–​636. doi: 10.1016/0021-​8928(59)90157-​1.
Bittencourt, E. et al. (2003) “A comparison of nonlocal continuum and discrete dislocation
plasticity predictions,” Journal of the Mechanics and Physics of Solids. Elsevier BV, 51(2), pp.
281–​310. doi: 10.1016/s0022-​5096(02)00081-​9.
Camanho, P. P., Davila, C. G. and de Moura, M. F. (2003) “Numerical Simulation of Mixed-​
Mode Progressive Delamination in Composite Materials,” Journal of Composite Materials.
SAGE Publications, 37(16), pp. 1415–​1438. doi: 10.1177/0021998303034505.
Campilho, R.D.S.G. et al. (2013) “Modelling adhesive joints with cohesive zone models:
effect of the cohesive law shape of the adhesive layer,” International Journal of Adhesion
and Adhesives. Elsevier BV, 44, pp. 48–​56. doi: 10.1016/j.ijadhadh.2013.02.006.
Dattaguru, B. et al. (1994) “Finite element estimates of strain energy release rate components
at the tip of an interface crack under mode I loading,” Engineering Fracture Mechanics.
Elsevier BV, 49(3), pp. 451–​463. doi: 10.1016/0013-​7944(94)90273-​9.
16 Small-​Scale Mechanical Testing

Dean, J., Aldrich-​Smith, G. and Clyne, T.W. (2011) “Use of nanoindentation to measure
residual stresses in surface layers,” Acta Materialia. Elsevier BV, 59(7), pp. 2749–​2761. doi:
10.1016/j.actamat.2011.01.014.
Dugdale, D.S. (1960) “Yielding of steel sheets containing slits,” Journal of the Mechanics and
Physics of Solids. Elsevier BV, 8(2), pp. 100–​104. doi: 10.1016/0022-​5096(60)90013-​2.
Evers, L.P. et al. (2002) “Crystal plasticity model with enhanced hardening by geometrically
necessary dislocation accumulation,” Journal of the Mechanics and Physics of Solids. Elsevier
BV, 50(11), pp. 2403–​2424. doi: 10.1016/s0022-​5096(02)00032-​7.
Feraren, P. and Jensen, H.M. (2004) “Cohesive zone modelling of interface fracture near
flaws in adhesive joints,” Engineering Fracture Mechanics. Denmark, 71(15), pp. 2125–​2142.
doi: 10.1016/j.engfracmech.2003.12.003.
Gizatulin, A.R. and Sultanov, A.K. (2018) “Application of whispering gallery modes (WGM)
in optical communications,” Proceedings of SPIE - The International Society for Optical
Engineering. Russian Federation: SPIE. doi: 10.1117/12.2317739.
Gurtin, M.E. (2000) “On the plasticity of single crystals: Free energy, microforces, plastic-​
strain gradients,” Journal of the Mechanics and Physics of Solids. United States: Elsevier
Science Ltd, 48(5), pp. 989–​1036. doi: 10.1016/S0022-​5096(99)00059-​9.
He, X. (2011) “A review of finite element analysis of adhesively bonded joints,” Interna-
tional Journal of Adhesion and Adhesives. China, 31(4), pp. 248–​264. doi: 10.1016/j.ijad-
hadh.2011.01.006.
Krueger, R. and O’Brien, T.K. (2001) “Shell/3D modeling technique for the analysis of
delaminated composite laminates,” Composites Part A: Applied Science and Manufacturing.
United States: Elsevier Science Ltd, 32(1), pp. 25–​44. doi: 10.1016/S1359-​835X(00)00133-​0.
Li, T. and Suo, Z. (2007) “Ductility of thin metal films on polymer substrates modulated by
interfacial adhesion,” International Journal of Solids and Structures. United States, 44(6), pp.
1696–​1705. doi: 10.1016/j.ijsolstr.2006.07.022.
Lin, K.L. et al. (2009) “Recycling thin film transistor liquid crystal display (TFT-​LCD) waste
glass produced as glass-​ceramics,” Journal of Cleaner Production. Taiwan, 17(16), pp. 1499–​
1503. doi: 10.1016/j.jclepro.2009.05.012.
Liu, P.F. and Islam, M.M. (2013) “A nonlinear cohesive model for mixed-​mode delamination
of composite laminates,” Composite Structures. Elsevier BV, 106, pp. 47–​56. doi: 10.1016/j.
compstruct.2013.05.049.
Long, X., Hu, B., et al. (2019) “Correlation of microstructure and constitutive behaviour of
sintered silver particles via nanoindentation,” International Journal of Mechanical Sciences.
China: Elsevier Ltd, 161–​162. doi: 10.1016/j.ijmecsci.2019.105020.
Long, X., Shen, Z., et al. (2022) “Determine the unique constitutive properties of elastoplastic
materials from their plastic zone evolution under nanoindentation,” Mechanics of Mate-
rials. China: Elsevier B.V., 175. doi: 10.1016/j.mechmat.2022.104485.
Long, X., Li, J., et al. (2022) “Dimensionless Analysis to Determine Elastoplastic Properties of
Thin Films by Indentation,” Coatings. China: MDPI, 12(11). doi: 10.3390/coatings12111768.
Long, X., Du, C., et al. (2018) “Finite Element Analysis to the Constitutive Behavior of Sintered
Silver Nanoparticles Under Nanoindentation,” International Journal of Applied Mechanics.
China: World Scientific Publishing Co. Pte Ltd, 10(10). doi: 10.1142/S1758825118501107.
Long, X. et al. (2023) “Indentation Reverse Algorithm of Mechanical Response for Elasto-
plastic Coatings Based on LSTM Deep Learning,” Materials. China: MDPI, 16(7). doi:
10.3390/ma16072617.
Long, X., Zhang, X.-​D., et al. (2019) Micromachines calibration of a constitutive model from tension
and nanoindentation for lead-​free solder.
Chapter 31 | Indentation-​based analysis of i 17

Long, X., Tang, W., et al. (2018) “Strain rate sensitivity of sintered silver nanoparticles using
rate-​jump indentation,” International Journal of Mechanical Sciences. China: Elsevier Ltd,
140, pp. 60–​67. doi: 10.1016/j.ijmecsci.2018.02.035.
Long, X. et al. (2021) “Strain rate shift for constitutive behaviour of sintered silver nanopar-
ticles under nanoindentation,” Mechanics of Materials. China: Elsevier B.V., 158. doi:
10.1016/j.mechmat.2021.103881.
Mi, Y. et al. (1998) “Progressive delamination using interface elements,” Journal of Com-
posite Materials. United Kingdom: SAGE Publications Ltd, 32(14), pp. 1246–​1272. doi:
10.1177/002199839803201401.
Nagamura, T. and Tanaka, K. (2009) “Physical properties of polymer thin Films for device
application,” Iete Tech Rev, 109, pp. 1–​2.
Park, K., Paulino, G.H. and Roesler, J.R. (2008) “Determination of the kink point in the
bilinear softening model for concrete,” Engineering Fracture Mechanics. United States,
75(13), pp. 3806–​3818. doi: 10.1016/j.engfracmech.2008.02.002.
Turon, A. et al. (2006) “A damage model for the simulation of delamination in advanced
composites under variable-​mode loading,” Mechanics of Materials. Elsevier BV, 38(11), pp.
1072–​1089. doi: 10.1016/j.mechmat.2005.10.003.
Turon, A. et al. (2007) “An engineering solution for mesh size effects in the simulation of
delamination using cohesive zone models,” Engineering Fracture Mechanics. Elsevier BV,
74(10), pp. 1665–​1682. doi: 10.1016/j.engfracmech.2006.08.025.
Tvergaard, V. and Hutchinson, J.W. (1992) “The relation between crack growth resistance and
fracture process parameters in elastic-​plastic solids,” Journal of the Mechanics and Physics
of Solids. Denmark, 40(6), pp. 1377–​1397. doi: 10.1016/0022-​5096(92)90020-​3.
Volokh, K. Y. (2004) “Comparison between cohesive zone models,” Communications in Numer-
ical Methods in Engineering. Wiley, 20(11), pp. 845–​856. doi: 10.1002/cnm.717.

You might also like