Mech Time-Depend Mater (2022) 26:1–20

https://doi.org/10.1007/s11043-020-09472-y

Protocols for studying the time-dependent mechanical

response of viscoelastic materials using spherical
indentation stress-strain curves

M.T. Abba1 · S.R. Kalidindi1

Received: 2 June 2019 / Accepted: 1 October 2020 / Published online: 23 October 2020
© Springer Nature B.V. 2020

Abstract Spherical nanoindentation has been used successfully to extract meaningful in-
dentation stress-strain curves in hard materials such as metals and ceramics. These methods
have not yet been applied on viscoelastic-viscoplastic polymer samples. This study explores
the potential of the current spherical nanoindentation analysis protocols in extracting inden-
tation stress-strain curves and viscoelastic properties on samples exhibiting time-dependent
material response at room temperature. These new protocols were tested on polymethyl
methacrylate, polycarbonate, and low-density polyethylene. The properties extracted under
different loading rates and indenter tip sizes conditions were observed to be consistent. It
is further demonstrated that it is possible to recover the compression stress-strain curves
for polymethyl methacrylate and low-density polyethylene from the measured indentation
stress-strain curves. This study establishes some of the foundations needed for the devel-
opment of protocols needed to reliably investigate the local time-dependent mechanical re-
sponse of materials using spherical nanoindentation.

Keywords Nanoindentation · Viscoelasticity · Stress-strain · Microscale · Finite element

analysis

1 Introduction

Optimizing the overall mechanical properties of most advanced materials requires a good
understanding of the mechanical properties of its microscale constituents. This task requires
the development and validation of novel multi-resolution mechanical characterization assays
(Hemker and Sharpe 2007; Jaya and Alam 2013; Srikar and Spearing 2003). In addition to
being able to measure the mechanical properties in very small volumes typical of the sizes
of the microscale constituents in advanced materials, it is desirable to develop methods ca-
pable of high-throughput testing. This is essential to obtain critical information on the very

B S.R. Kalidindi
surya.kalidindi@me.gatech.edu
1 George W. Woodruff School of Mechanical Engineering, Georgia Institute of Technology, Atlanta,
GA 30332-0405, USA
2 Mech Time-Depend Mater (2022) 26:1–20

large variety of potential microscale constituents that could be found in advanced material
systems (Kalidindi et al. 2017). Nanoindentation offers tremendous promise in addressing
this critical need, and has been successfully employed to study the microscale properties
of metal and ceramic constituents in advanced material systems (Khosravani et al. 2020;
Weaver and Kalidindi 2016; Fischer-Cripps 2002). There is a critical need to extend these
protocols to measuring the time-dependent mechanical responses of the viscoelastic con-
stituents in complex material systems.
More specifically, modern nanoindenters possess high load resolution and depth sens-
ing capabilities, and can be used to characterize the local mechanical behavior in materials
at very small spatial resolutions ranging from ∼100 nm to ∼500 µm. Moreover, recently
developed instrumented spherical indentation data analysis methods have demonstrated a
tremendous potential to transform the raw load-displacement data into very meaningful
indentation stress-strain curves (Kalidindi and Pathak 2008; Pathak and Kalidindi 2015;
Weaver and Kalidindi 2016; Vachhani et al. 2016; Patel and Kalidindi 2017). These pro-
tocols have been able to capture the local loading and unloading elastic moduli, the lo-
cal indentation yield strengths, and some characteristics of the post-yield behavior in
several samples (Pathak et al. 2008; Pathak et al. 2009a). The validity of these pro-
tocols has been critically evaluated using finite element models (Donohue et al. 2012;
Patel and Kalidindi 2016). By recovering indentation stress-strain curves from the mea-
sured indentation load-displacement data, we can get a much more reliable and objective
comparison of material behavior between different samples or between different locations
in the sample, compared to the conventional hardness measurements. These new protocols
have thus far been explored mostly on materials that do not experience significant time-
dependent deformation.
A salient feature of the methods described above is that they isolate the initial elastic
loading segment (in the initial contact of the indenter with the sample), and analyze it rigor-
ously using Hertz’s theory (Kalidindi and Pathak 2008). It is extremely critical to establish
the elastic stiffness of the virgin sample, as any amount of plastic deformation changes the
unloaded sample surface geometry and adds significant error to the proper analyses of the in-
dentation data using Hertz’s theory (Pathak and Kalidindi 2015). In the indentation analyses
protocols, continuous and accurate measurement of the elastic stiffness (as will be discussed
later) is central to the extraction of the indentation stress-strain curves, even into the inelas-
tic regime well beyond the elastic regime. This is because only the elastic stiffness gives
an estimate of the indenter-sample contact geometry that is fully consistent with Hertz’s
theory (Donohue et al. 2012). Any other direct or indirect estimate of the contact geometry
results in an erroneous application of Hertz’s theory for indentation analyses. Indeed, this
requirement has posed major challenges to the conventional protocols for the extraction of
indentation stress-strain curves from the raw load-displacement measurements, and has only
been circumvented with the high resolution measurement capabilities of modern nanoinden-
ters. In this regard, it should be noted that the initial elastic segment has been successfully
isolated and analyzed with only spherical indenters, as the sharp indenters usually produce
inelastic deformation as soon as they touch the sample surface. For the present study, our
goal is to extend the previously established spherical nanoindentation stress-strain protocols
to materials that exhibit viscoelasticity at room temperature. This work lays a foundation
for future work aimed at studies of viscoelastic and viscoplastic properties of individual
mesoscale constituents of composite material systems, including interphases.
The protocols developed in this work are demonstrated on three commercially avail-
able polymers (polymethyl methacrylate (PMMA), polycarbonate (PC), and low-density
polyethylene (LDPE)) using three different indenter tip radii (16 µm, 100 µm, and 1500 µm).
Mech Time-Depend Mater (2022) 26:1–20 3

The properties and stress-strain curves obtained from nanoindentation are compared against
results obtained from conventional simple (uniaxial) compression tests performed on the
same polymers at a similar strain rate. Finite element (FE) simulations of spherical nanoin-
dentation and uniaxial compression are also performed to understand the differences in the
indentation and simple compression responses in these materials.

2 Background: viscoelastic indentations and analyses

Most prior studies on nanoindentation of polymers have focused on extracting the creep
compliance or relaxation modulus, where these quantities are expressed either using a me-
chanical model or a Prony series representation (Huang and Lu 2006; Cheng et al. 2005;
Huang and Lu 2007). These studies have generally used sharp indenter tips (e.g., Berkovich
or conical) (Briscoe et al. 1998; Jakes et al. 2012; Cheng and Cheng 2005b). In some studies,
spherical tips were used to investigate the viscoelastic behavior of different polymers (Zhou
and Lu 2010; Knauss et al. 2008; Martynova 2016). Because of viscoelasticity, the loading
conditions imposed in the indentation play an important role. The most common approach
has been to ramp the load or displacement to a preset value, and holding it for some time.
The relaxation modulus or creep compliance is then extracted from the hold regime using
established viscoelastic indentation models. It is important to recognize that a step loading
condition is impossible to implement practically, and therefore it is important to account
for the ramp load before the hold in the analyses of the raw data. Oyen (2005) has demon-
strated a “ramp correction factor” for the exponential decay terms to address this challenge.
As mentioned earlier, there may be substantial amount of plastic deformation imposed on
the sample in the initial ramp, especially with the use of sharp indenters. As already noted,
any amount of plastic deformation adds significant errors to the analyses based on Hertz’s
theory because of the changes induced in the unloaded sample surface geometry (its radius
is no longer known a priori). Because of such errors, the properties extracted from the hold
regime in these experiments may not represent faithfully the properties of the as-received
sample. A number of indentation studies have also explored the response of a viscoelas-
tic material to dynamic (oscillatory) loading at different frequencies (Huang et al. 2004;
Odegard et al. 2005; Jäger and Lackner 2008) in order to estimate the storage and loss mod-
uli.
Theories of linear viscoelastic indentation started in the mid-1950s with the work of Lee
(1956), Radok (Lee and Radok 1960), Hunter (1960), Graham (1965), and Ting (1966).
They developed an approach for finding the viscoelastic solution in cases where the corre-
sponding solution for the purely elastic case is already known (referred as the viscoelastic
correspondence principle). Using these theories, the load-displacement relationship for the
frictionless indentation of an isotropic, linear viscoelastic, solid using a rigid indenter ex-
hibiting a quadratic axisymmetric surface geometry in the contact region can be expressed
in two equivalent forms as

3(1 − ν 2 ) t
dP (τ )
h3/2 (t) =  J (t − τ ) dτ (1)
4 Reff 0 dτ

  t 3
4 Reff dh 2 (τ )
P (t) = E(t − τ ) dτ (2)
3(1 − ν 2 ) 0 dτ
4 Mech Time-Depend Mater (2022) 26:1–20

where P (t) and h(t) denote the load and displacement, respectively, as functions of time. In
Eq. (1) and Eq. (2), J (t), E(t) and ν denote the isotropic creep compliance, isotropic relax-
ation modulus and the time-independent Poisson’s ratio, respectively, of the material being
indented. Reff denotes an effective curvature of the indenter-sample contact defined through
1/Reff = 1/Ri + 1/Rs , where Ri and Rs denote the radii of curvature of the indenter and the
sample, respectively, in the undeformed condition. For the initial elastic indentation of a flat
sample, Reff = Ri . However, if one loads the sample beyond its elastic limit, Reff continu-
ously changes with the imposed deformation. The central challenge in analyzing indentation
datasets stem from the difficulty of measuring or estimating the change in Reff beyond the
very small initial elastic loading segment (Pathak and Kalidindi 2015). An important result
from the linear viscoelastic indentation theory is the relationship between the indentation
depth and the contact area (a) expressed as (Cheng and Cheng 2005a)

a(t) = h(t)Reff (3)

Note that this relationship is the same as the original Hertz model for the case of the
elastic indentation.
Most modern indenters also provide a continuous stiffness measurement (CSM). This
signal provides a reliable estimate of the elastic stiffness (S) by imposing a small, sinu-
soidally varying load on top of the DC signal that drives the indenter, allowing for a con-
tinuous measurement of the resulting amplitude and phase of the displacement oscillation
(Li and Bhushan 2002). In elastic indentation, this measurement of the stiffness allows
for a reliable estimate of the contact radius using the relationship (Donohue et al. 2012;
Kalidindi and Pathak 2008)
S
a= (4)
2Eeff
1 1 − νs2 1 − νi2
= + (5)
Eeff Es Ei

where Eeff is called the effective elastic modulus and ν is the Poisson’s ratio, while sub-
scripts s and i refer to the specimen and indenter, respectively. It is not yet clear if this
equation needs to be modified in any way for the viscoelastic indentation. We will explore
this question critically in this study through suitably designed FEM simulations.
Creep compliance used in Eq. (1) is usually expressed as a time series called the Prony
series (Lakes 2009):

N
J (t) = J∞ − Ji e−t/τci (6)
i=1

where J∞ is the steady state creep compliance of the system and Ji and τci are the compli-
ances and time constants associated with the underlying viscoelastic mechanisms of defor-
mation in the material. Note that the instantaneous creep compliance, J0 , can be computed
from Eq. (6) by setting t = 0. In an analogous manner, one can express the relaxation mod-
ulus as

N
E(t) = E∞ + Ei e−t/τri (7)
i=1

where E∞ is the steady state stiffness of the system and Ei are the stiffnesses capturing the
viscoelastic behavior of the material. The creep compliance and the relaxation modulus are
Mech Time-Depend Mater (2022) 26:1–20 5

related to each other through

 t  t
J (t − τ )E(τ )dτ = E(t − τ )J (τ )dτ = t (8)
0 0

It is noted that the Prony series is a common method of describing viscoelastic be-
havior that is available in most commercial finite element software. In the present work,
we implemented this description of the viscoelastic material in a finite element model of
spherical indentation using the commercial software ABAQUS (Dassault Systemes Simulia
Corp. 2014). In this implementation, Prony series was defined for both the shear and the
bulk relaxation moduli by assuming that the Poisson’s ratio is a constant and that G(t) =
E(t)/2(1 + ν) and K(t) = E(t)/3(1 − 2ν) (Brinson and Brinson 2008; Gauthier 1995;
Lakes and Wineman 2006).

3 Material and methods

3.1 Materials

All the polymers used in this study were obtained from McMaster-Carr (Elmhurst, IL).
For nanoindentation, the specimen were obtained from a 1.25 inch diameter extruded
rods of PMMA (density 0.043 lbs/in3 ; glass transition temperature 105 °C), PC (density
0.045 lbs/in3 ; glass transition temperature 145 °C), and LDPE (density 0.033 lbs/in3 ; glass
transition temperature −125 °C). For compression tests, the PMMA and PC specimen were
taken from extruded rods of approximately 0.625 in diameter and 1.1 in length and the LDPE
specimen from extruded rods of approximately 1 in diameter and 2 in length. PMMA spec-
imens were annealed at 110 °C and PC specimens were annealed at 150 °C. All the speci-
mens were annealed in a Thermo Scientific Lindberg/Blue MTM MoldathermTM box furnace
(Waltham, MA) for two hours and then slowly cooled down to room temperature at a rate
of 5 °C/hr. The nanoindentation specimen were cut perpendicular to the extruded direction
using an Allied TechCut 5TM (Rancho Dominguez, CA) precision sectioning machine and
then polished using silicon carbide papers of decreasing grit sizes (320, 800, 1200, 2400,
and 4000 grit) using a StruersTegramin-30 (Cleveland, OH). Each polishing step except the
4000 grit was performed for 2 minutes, followed by washing to remove debris. The 4000
grit polishing was performed for 6 minutes. This was followed by polishing with a 1 µm
alcohol based diamond suspension (Struers DP-Suspension) for 20 minutes and a 0.05 µm
colloidal silica suspension (Buehler MasterMet) for 20 minutes.

3.2 Uniaxial compression tests

Compression tests were performed on an MTI Phoenix Universal Testing Machine with a
20,000 lb load cell. Tests were performed according to ASTM standard D695 for testing
plastics (D695-10 2010) at a speed of 0.05 in/min. Displacement was measured using a
capacitance gage and all load-displacement data was converted to true stress-strain curves.
At least five samples were tested for each material with the data reported as mean ± standard
deviation.
6 Mech Time-Depend Mater (2022) 26:1–20

3.3 Indentation tests

The indentation tests were performed on an Agilent G200 (Keysight Technologies Inc.,
Santa Rosa, CA) Nano Indenter with an XP head. The indenter has a maximum load of
500 mN with a high load option of 10 N and a load resolution of 50 nN. The maximum
indentation depth achievable is greater than 500 µm with a resolution of less than 0.01 nm.
The indents are performed on the polished surfaces of the polymers and the spherical di-
amond tips used have radii of 16 µm, 100 µm, and 1500 µm. Prior to use, the tips were
tested on Silica standards to confirm that the measured mechanical properties are close to
the expected values. The tips were also visually inspected in the SEM for defects or dirt.
Since changes in temperature can cause expansion or shrinkage of materials leading to er-
rors in measurement, the indenter measures a thermal drift before each test by holding the
indenter on the surface of the material and measuring any changes in displacement. All
tests were performed after the measured indenter drift rate reached and maintained a value
less than 0.05 nm/s. The machine applies a drift correction as part of the standard operat-
ing procedure. CSM corrections were applied to the displacement, load, and stiffness of all
the tests using protocols described in previously published papers (Vachhani et al. 2013;
Pharr et al. 2009). The stress-strain curves extracted in this study are from indentations per-
formed at rates of 0.05 s−1 and 0.5 s−1 , defined as the loading rate, dP /dt , divided by the
measured load, P . At least five curves were used to extract any mechanical properties in this
study.

3.4 Finite element model of spherical indentation

In prior studies (Donohue et al. 2012; Patel and Kalidindi 2016; Bouzakis et al. 2002), finite
element models were found invaluable in improving the understanding and analyses of the
measurements obtained from indentation experiments. This is mainly because they allow us
to evaluate critically the consequences of many assumptions and idealizations made in the
analyses of the measured datasets, such as the definitions of indentation stress and indenta-
tion strain, the estimation of the contact radius, and the relationships between indentations
stress-strain curves and simple compression stress-strain curves.
A two-dimensional axisymmetric finite element model of the spherical indentation shown
in Fig. 1 was produced using the commercial finite element code ABAQUS (Dassault Sys-
temes Simulia Corp. 2014). The indenter is assumed to be rigid with a radius of 10 µm, and
the sample is assumed to exhibit an isotropic constitutive response. The sample was meshed
using four-noded axisymmetric (CAX4) elements. The sample is constrained from moving
in the 1-direction (x-axis) along the axis of symmetry and in the 2-direction (y-axis) along
the bottom surface (see Fig. 1). The highest mesh density was achieved in the region where
highest stresses or strains were expected. A mesh density higher than 64 elements/µm (along
the 1-direction) is used for all simulations in this zone. A downward vertical displacement
condition was imposed on the node at the center of the indenter, which was tied to the surface
of the indenter. Contact between the sample and indenter was defined as a hard surface-to-
surface contact with the indenter as the master surface and the sample as the slave surface.
The FE model was validated by comparing the predicted results in elastic and viscoelastic
indentations against the known theoretical solutions (see also Eq. (2)). The FE result showed
good agreement with the analytical solutions, thereby validating the FE model developed in
this study for simulating indentation.
Mech Time-Depend Mater (2022) 26:1–20 7

Fig. 1 a) Schematic of indenter, sample, and boundary conditions used for finite element simulations. b) An
example of a ramp and hold nanoindentation simulation compared to the theoretical solution obtained from
Eq. (2)

3.5 Nanoindentation protocols for viscoelastic materials

New indentation protocols have been developed in this work for studying viscoelastic re-
sponse of materials. These will be presented next, and comprise five distinct steps. The
first step is designed to determine if the material has a sufficiently dominant viscous com-
ponent to affect its deformation behavior. This is particularly useful for new materials
for which prior information is not adequate. In some materials, the time constants of the
viscous response are very small, allowing us to treat the material simply as an elastic-
plastic material. The first step allows us to make a decision in this regard. The second
step is designed to determine which region of the indentation load-displacement curve is
viscoelastic. Any properties extracted assuming viscoelastic behavior should come from
this region. The third step aims to extract the creep compliance after an indentation has
been performed. The fourth step ensures that the initial portion of the load-displacement
curve is corrected for the zero-point. A zero-point identifies the point of effective initial
contact between the indenter and the sample. In prior work (Kalidindi and Pathak 2008;
Donohue et al. 2012), it has been clarified that the point of effective contact is the point
on the measured load-displacement curve that produces the best fit of the initial elastic (ex-
tended here to viscoelastic) portion to the theoretically expected response. This concept
is useful to circumvent unavoidable consequences from sample preparation artifacts such as
surface roughness or very small disturbed surface layers. The identification of the zero-point
has a major influence on the extracted indentation stress-strain curves. The final step is to
calculate the stress-strain curves from the nanoindentation data collected. All of the steps
are described in detail next.

3.5.1 Step 1: determine material viscoelasticity

Material viscoelasticity can be studied simply by indenting the sample at different displace-
ment/loading rates and examining the differences in the measured load-displacement curves
in the viscoelastic regime. The expected trends can also be studied using the FE model. This
8 Mech Time-Depend Mater (2022) 26:1–20

Fig. 2 FE simulation of
indentation loading on LDPE at
different displacement rates

is illustrated in Fig. 2 using the viscoelastic material properties obtained from tests on LDPE
(presented in the next section), expressed as a two-term Prony series

E(t) = 0.15 + 0.057e−0.091t + 0.043e−0.0094t GPa (9)

The material properties used for the fully elastic simulations were the steady state stiff-
ness, 0.25 GPa, and LDPE’s Poisson’s ratio, 0.45. In these simulations, the maximum in-
dentation depth was selected as 50 nm, and the indentation rates were selected as 50 nm/s,
5 nm/s, 0.5 nm/s, and 0.05 nm/s. As seen from Fig. 2, the material responds differently at the
different indentation rates. At the high displacement rates, the response is almost identical to
the elastic case. This is because the time it takes to reach the specified displacement is lower
than the material’s time constants, thus not giving the material enough time to relax. At a
lower displacement rate, the material has enough time to relax thus the response is lower.

3.5.2 Step 2: determine viscoelastic region

This can be determined by loading and unloading the material at increasing loads using the
same loading rate. At the end of each unload a small load is held on the material and the
displacement is allowed to recover. The goal is to observe where the displacement stops
decreasing and how this residual displacement changes at different loads. Ideally, if the
indentation is still within the viscoelastic regime, the residual displacement will be close to
zero at the different maximum loads. For highly viscoelastic materials, it may take much
longer for the material to return to its initial condition. Thus it is very difficult in practice to
quantify the exact point of yield on the load-displacement curve. Our goal here is to make
an educated guess. Since this is not an exact value, it is generally better to be conservative in
identifying the load or displacement levels where the viscoelastic behavior of the material
ends. It is important to avoid plasticity because it results in dramatic changes to Reff (see
Eq. (1)). As soon as plasticity initiates, one needs a different set of protocols to estimate
the contact radius and its evolution through the entire indentation test (Pathak and Kalidindi
2015; Pathak et al. 2008; Pathak et al. 2009b).
Each sample studied in this work was loaded and unloaded at the same rate with three
or four different maximum loads. Before completely unloading the material, a small pre-
set load was left on the material for an extended period of time. Using the 100 µm radius
indenter tip PMMA and PC were loaded at a loading rate of 2.5 mN/s with maximum loads
of 40 mN, 80 mN, 160 mN, and 320 mN for PMMA and 20 mN, 40 mN, 80 mN, and
Mech Time-Depend Mater (2022) 26:1–20 9

160 mN for PC. For LDPE a loading rate of 2 mN/s and maximum loads of 8 mN, 16 mN,
and 32 mN were used. The loads held at unload were 5 mN for PMMA and PC, and 2 mN
for LDPE.

3.5.3 Step 3: extract viscoelastic response

The material’s viscoelastic response in this study is extracted from a ramp and hold test
using the method presented by Oyen (2005). The loading conditions can be written as

P (t) = kt 0 ≤ t ≤ th
Pmax = kth th < t ≤ tu (10)

where th is the time taken to reach the maximum load and k is the loading rate. This means
that the viscoelastic equation for creep (Eq. (1)) must be solved for the ramp and the hold.
These solutions can be expressed as

3 3(1 − ν) min(t,th )
h 2 (t) = √ J (t − τ )kdτ (11)
8 Reff 0

Equation (6) can then be substituted into Eq. (11) and fit to the measurements to extract
the creep compliance function. This method explicitly accounts for the ramp before the
hold, rather than assuming it is fast enough to be considered a step load. The relaxation
modulus can then be extracted by using Eq. (8) using deconvolution techniques (Cirnu 2010;
Kreyszig 2010).
In this work, the viscoelastic indentation responses were measured using ramp and hold
tests on all three materials using three different indenter tip radii. For each material, the load
was increased to three different loads at a constant loading rate, held for 300 seconds, and
then unloaded at a constant unloading rate. The viscoelastic properties were extracted from
the hold portion.

3.5.4 Step 4: perform zero point correction

As mentioned earlier, the identification of the effective zero-point is central to extract-

ing a meaningful and reliable indentation stress-strain curve (Kalidindi and Pathak 2008;
Donohue et al. 2012; Fischer-Cripps 2000; Moseson et al. 2008; Moseson et al. 2009). The
central challenge comes from the fact that the machine-identified zero-point (using built-in
software provided by the manufacturer) is usually incorrect. This is particularly important if
our goal is to obtain reliable indentation stress-strain curves. This difference is illustrated in
Fig. 3. Combining Eqs. (1) and (3), and introducing a correction to the zero-point displace-
ment (h∗ ) results in

∗ 3(1 − ν) t dP (τ )
h(t) = h̃(t) − h = J (t − τ ) dτ (12)
8a(t) 0 dτ

where h̃(t) is the indentation depth measured by the machine, and h(t) is the corrected
indentation depth. The zero-point correction can be easily estimated from Eq. (12) through
linear regression, if we have information on the creep compliance (from step 3) and the
contact radius.
In the elastic indentation, the contact radius is estimated using CSM and Eq. (4). In these
protocols, stiffness is estimated through the analyses of the superimposed oscillations by the
10 Mech Time-Depend Mater (2022) 26:1–20

Fig. 3 a) Load-displacement curve from indentation test on LDPE. Points labeled a and b on the curve denote
the machine identified zero-point and the effective zero-point identified using the protocols presented in this
work, respectively. b) Indentation stress-strain curves extracted using the two different zero-points identified
in (a)

CSM module. In the application of CSM to the viscoelastic response, one needs to verify
if the measurements are sensitive to the applied frequency of the superimposed oscillations.
For the present study, the oscillations applied in CSM were simulated using the FE indenta-
tion model (with the viscoelastic properties measured in this study for LDPE as these sam-
ples produced the largest viscoelastic responses as shown later) at a range of frequencies.
Selected results from these simulations are presented in Fig. 4. The phase angle between
the applied displacement oscillations and the resulting load oscillations were extracted from
these simulations. At high frequencies, the phase angle between the oscillations approached
zero, indicating that one can idealize these as elastic response (Gutierrez-Lemini 2014;
Mase 1970). Therefore, we expect the relationship between the effective elastic modulus,
contact area, and stiffness shown in Eq. (4) to hold at high frequencies. In the work pre-
sented here, the CSM oscillations were imposed at a frequency of 45 Hz and amplitude of
2 nm. These specific values were identified in prior work (Vachhani et al. 2013) to provide
the most reliable results for the CSM.
Another important relationship that needs to be evaluated critically using the FE simu-
lations is the relationship between contact radius, displacement, and effective radius given
by Eq. (3). This is because the contact radius can be extracted directly in the FE simulation
through the contact definition. The results of such an investigation are presented in Fig. 5 for
both the elastic and the viscoelastic cases; four different displacement rates were examined
for the viscoelastic case. These results show that, as the indenter is pushed into the material,
the contact radius does not depend on the rate of indentation and is the same as the elastic
case. In other words, the contact radius depends only on the indentation depth as expressed
in Eq. (3). It should be noted that Eq. (3) is a geometric relation, and in the viscoelastic
correspondence principle, such relationships are generally unaffected when ones goes from
the elastic solution to the viscoelastic solution. In the present study, we have confirmed this
with the FE simulations.

3.5.5 Step 5: extract stress-strain curves

The indentation stress-strain curves can be recovered from the load-displacement data using
suitable definitions of indentation stress and indentation strain (Kalidindi and Pathak 2008;
Mech Time-Depend Mater (2022) 26:1–20 11

Fig. 4 Finite element

simulations for indentations on
LDPE at displacement oscillation
frequencies of 1 Hz, 10 Hz, and
100 Hz. The phase angle, , is
between the applied displacement
oscillations and the resulting load
oscillations

Fig. 5 Contact area versus

displacement from FE
simulations of indentations on
LDPE at different displacement
rates

Donohue et al. 2012). The indentation stress is defined classically as the load applied over
the projected contact area,

P (t)
σind (t) = (13)
πa(t)2
12 Mech Time-Depend Mater (2022) 26:1–20

Indentation strain is defined as (Kalidindi and Pathak 2008)

4 h(t) h(t)
εind (t) = ≈ (14)
3π a(t) 2.4a(t)
This definition of strain can be interpreted as change in length over the original length by
idealizing the primary zone of indentation as a cylinder of radius, a, and length, 2.4a, which
is then compressed by the indentation depth h. This definition has been validated using
both numerical simulations and experimental measurements (Kalidindi and Pathak 2008;
Pathak et al. 2008; Donohue et al. 2012; Pathak et al. 2009b). As mentioned earlier, the
CSM oscillations can be considered elastic. Therefore, it is reasonable to estimate the contact
radius throughout the test using Eq. (4), where the instantaneous relaxation modulus is used
as the elastic modulus (Huang and Lu 2006). With these definitions of indentation stress
and indentation strain, the load-displacement curves obtained by nanoindentation can be
converted to indentation stress-strain curves.

4 Results and discussion

4.1 Uniaxial compression tests

The results from the compression tests are shown in Fig. 6. The strain rates obtained in the
tests were 8 × 10−4 s−1 for PMMA and PC, and 5 × 10−4 s−1 for LDPE. It was observed
that the PMMA and PC samples failed by shearing with polycarbonate shearing more than
PMMA which may have caused some of the drop in stress with increasing strain. This drop
in stress has been reported in literature and is usually attributed to a combined effect of
strain softening and thermal softening (Arruda et al. 1995). Care was taken to ensure that
the sample geometry was always a right cylinder and the axis of loading was aligned, as
any misalignment would increase the shear on the material. The moduli, measured from the
initial linear portion of the true stress-strain curves, were 1.9±0.031 GPa, 1.42±0.031 GPa,
and 0.154 ± 0.0014 GPa for PMMA, PC, and LDPE, respectively. These results will be
compared to the indentation stress-strain curves collected at similar strain rates.

4.2 Spherical nanoindentation

4.2.1 Material viscoelasticity

The results from the indentations at two different rates (0.05 s−1 and 0.5 s−1 ) using the
16 µm radius indenter on PMMA, PC, and LDPE are shown in Fig. 7. It is clearly seen that
LDPE shows the largest viscoelastic component among the three materials, while PC is the
least viscoelastic.

4.2.2 Viscoelastic region

The results from the tests performed to identify the regime of viscoelasticity for PMMA
(without initiating significant plastic deformation) using the 100 µm radius indenter tip are
shown in Fig. 8. The figure shows all curves for the four different maximum loads with at
least five tests for each. The standard deviation for all cases was less than 5%. It is seen that
above a certain load, there is a clearly observable difference in residual displacements. In
other words, the material has experienced discernable plastic deformation above those loads.
These results are summarized in Table 1, and are used to estimate the viscoelastic regime of
the measured load-displacement curve.
Mech Time-Depend Mater (2022) 26:1–20 13

Fig. 6 True stress-strain curves

for a) PMMA, b) PC, and c)
LDPE measured from simple
compression tests on five samples
each

Table 1 Estimated viscoelastic

load limits for the three different Indenter size PMMA PC LDPE
indenter tip radii for the samples
tested in this work 16 µm 40 mN 30 mN 3 mN
100 µm 160 mN 80 mN 16 mN
1500 µm 500 mN 350 mN 50 mN

4.2.3 Viscoelastic properties

The creep compliance and corresponding relaxation modulus were extracted from the ramp
and hold tests and using the methods explained in the previous section and best fits based on
linear regression with five tests for each measurement. The extracted viscoelastic properties
are represented in terms of a two-term Prony series, and tabulated in Table 2. The standard
deviation for all cases was less than 5%. It is noted that the viscoelastic properties obtained
from the tests performed with the distinctly different indenter tip radii showed reasonable
agreement with each other. This agreement validates the protocols developed in this work.

4.2.4 Stress-strain curves

Once the viscoelastic properties have been extracted, the stress-strain curves can then be
produced. As explained earlier, a zero-point correction for the displacement needs to be
14 Mech Time-Depend Mater (2022) 26:1–20

Fig. 7 Measured indentation

load-displacement curves for a)
PMMA, b) PC, and c) LDPE at
two different loading rates with
five tests performed at each rate

established by fitting the measurements within the viscoelastic regime to Eq. (12). The zero-
point correction for the load is then taken simply as the load at the corrected displacement.
Since these values depend on the location of the indent, surface roughness, or other surface
artifacts that may remain after polishing they should be established separately for each in-
dentation test (even when performed on the same sample). The measured load-displacement
curves are converted to stress-strain curves using the protocols described earlier. Represen-
tative results at different strain rates and different indenter tips for PMMA, PC, and LDPE
are plotted in Fig. 9. The strain levels attained with the 1500 µm indenter were fairly low
due to the load limit on the nanoindenter. Consequently, they covered only the very early
portions of the indentation stress-strain curves shown in Fig. 9, and were seen to be in ex-
Mech Time-Depend Mater (2022) 26:1–20 15

Fig. 8 Measured indentation

load-displacement curves for
PMMA in experiments designed
to identify the viscoelastic
regime. The sample is loaded to
different maximum loads and
unloaded to 5 mN and held while
the time-dependent residual
displacement is measured. The
inset shows that after being
loaded to 320 mN, the residual
displacement is higher than the
others meaning that discernable
plastic deformation has occurred

Table 2 Extracted creep compliance and relaxation modulus for the three materials studied in this work in
measurements performed using three different indenter tip radii

Material Indenter size Creep compliance (1/GPa) Relaxation modulus (GPa)

PMMA 16 µm 0.43 − 0.032e−t/18 − 0.060e−t/216 2.32 + 0.26e−t/16 + 0.37e−t/186

100 µm 0.45 − 0.044e−t/15 − 0.076e−t/225 2.23 + 0.37e−t/13 + 0.44e−t/187
1500 µm 0.41 − 0.025e−t/25 − 0.034e−t/220 2.46 + 0.20e−t/23 + 0.22e−t/202
PC 16 µm 0.47 − 0.012e−t/15 − 0.010e−t/136 2.11 + 0.059e−t/15 + 0.047e−t/133
100 µm 0.49 − 0.013e−t/9 − 0.014e−t/151 2.03 + 0.059e−t/9 + 0.058e−t/146
1500 µm 0.52 − 0.021e−t/11 − 0.016e−t/162 1.92 + 0.084e−t/11 + 0.061e−t/157
LDPE 16 µm 6.48 − 1.22e−t/17 − 1.58e−t/154 0.15 + 0.072e−t/13 + 0.046e−t/118
100 µm 6.76 − 1.12e−t/14 − 1.61e−t/138 0.15 + 0.057e−t/11 + 0.043e−t/106
1500 µm 6.54 − 1.18e−t/15 − 1.73e−t/146 0.15 + 0.072e−t/11 + 0.051e−t/108

cellent agreement with the measurements obtained with the smaller indenter tips. In order
to maintain the clarity of the figures, these were not included in these plots. Two particular
trends are clearly seen from the results presented in Fig. 9: (i) there is remarkable agreement
between the indentation stress-strain curves measured using indenter tips of vastly different
radii, and (ii) the indentation stress-strain curves exhibit the expected strain rate sensitivity
inferred from Step 1 (see Fig. 7). When the ratios of the stresses of the viscoelastic region
for the 16 µm at the two strain rates are compared we get an average of 1.14 for PMMA,
2.22 for PC, and 1.31 for LDPE. The ratios obtained using FE simulations were 1.25 for
PMMA, 2.26 for PC, and 1.31 for LDPE. These results validate the protocols presented in
this work, including the normalizations implied in the indentation stress and strain measures
employed, and point to their tremendous potential in studies of multiscale material response.
Comparing the indentation stress-strain curves in Fig. 9 with the simple compression
stress-strain curves in Fig. 6, we observe that the indentation stress-strain curves are higher
for all three materials studied in this work. This, of course, is expected because of at least
two factors: (i) the higher levels of hydrostatic pressure in indentation cause the effective
indentation moduli to be higher than in simple compression, and (ii) most polymers exhibit
a pressure-dependent yield that further increases the indentation yield strengths as compared
to the simple compression yield strengths (Spitzig and Richmond 1979). It is therefore in-
structive to compare these values quantitatively to derive new insights into the mechanical
response of the polymers studied. In the case of rate-independent isotropic elastic-plastic
16 Mech Time-Depend Mater (2022) 26:1–20

Fig. 9 Comparison of the

measured indentation
stress-strain curves at different
loading rates with different
indenter tip radii for a) PMMA,
b) PC, and c) LDPE

deformation in metals, it was seen that the conversion of indentation stress-strain (ISS)
curves to the simple compression stress-strain curves could be accomplished using three
simple factors (Patel and Kalidindi 2016). In this prior work, the ratio between the indenta-
tion stress and uniaxial stress was found to be 2.2, the ratio between the indentation elastic
strain and the uniaxial elastic strain was found to be 2, and the ratio between the indenta-
tion plastic strain and uniaxial plastic strain was found to be 1.3. We used the same factors
to scale our indentation stress-strain curves and compared them to the compression stress-
strain curves for PMMA, PC, and LDPE in Fig. 10. The curves obtained using the 16 µm
indenter with strain rates of 9 × 10−4 s−1 , 1.1 × 10−3 s−1 , and 8 × 10−4 s−1 for PMMA,
PC, and LDPE, respectively, were used for the scaling. It can be seen that there is good
agreement between the curves for PMMA and LDPE, but not for PC. These results are
Mech Time-Depend Mater (2022) 26:1–20 17

Fig. 10 A comparison of the

scaled indentation stress-strain
curve with the compression
stress-strain curve for a) PMMA,
b) PC, and c) LDPE

somewhat surprising because one would have expected the scaling factors to change for the
different materials depending on the degree of pressure-dependence exhibited by them. In-
deed, PC is likely to exhibit a high level of pressure-dependent yield (Caddell et al. 1974;
Bahr et al. 1998), even with low levels of viscoelasticity. Clearly, more measurements on
additional material systems are needed to clarify the scaling factors.

5 Conclusions

A set of protocols were presented to reliably extract viscoelastic properties and indenta-
tion stress-strain curves from spherical nanoindentation, and validated using three different
indenter tip sizes on three different commercially available polymers – PMMA, PC, and
LDPE. The viscoelastic indentation properties obtained with the different indenter tips were
found to be remarkably consistent across all three indenter sizes for each polymer mate-
rial. The indentation stress-strain curves were also compared at two different strain rates
18 Mech Time-Depend Mater (2022) 26:1–20

and the measured strain rate dependence agreed well with the FE results. It was also noted
that LDPE was the most viscoelastic and PC was the least viscoelastic. All of these results
validate the protocols presented in this work for viscoelastic materials.
The indentation stress-strain curves at the lower strain rate were also compared to the
compression stress-strain curves. It was also shown that the uniaxial compression stress-
strain curve can be recovered from the indentation stress-strain curve for PMMA and LDPE
by using the three scaling factors developed in prior work for elastic-plastic pressure-
independent material behavior. For PC, a different set of scaling factors are needed. Al-
though it is suspected that this is a consequence of pressure-dependent inelastic deformation
in PC, it is clear that further detailed studies involving a number of other polymers are
needed to fully explain this observation.
The overall goal of this paper was to provide a comprehensive workflow for studying
constitutive response in viscoelastic-viscoplastic materials. We therefore defined a workflow
comprising a total of five steps that end in meaningful indentation stress-strain curves that
can be compared to uniaxial stress-strain curves. To the best of our knowledge, no one has
actually put together the protocol described in this paper with comparable end results. We
believe that comprehensive workflows are useful to move towards standardization of char-
acterization tasks, and eventually to potential automation. Therefore, there is tremendous
value to the research community is formalizing all the steps involved in the workflow.

Acknowledgements The authors gratefully acknowledge support from the National Science Foundation
(Grant# NSF 1761406).

Publisher’s Note Springer Nature remains neutral with regard to jurisdictional claims in published maps
and institutional affiliations.

References
Arruda, E.M., Boyce, M.C., Jayachandran, R.: Effects of strain rate, temperature and thermomechanical cou-
pling on the finite strain deformation of glassy polymers. Mech. Mater. 19(2), 193–212 (1995)
ASTM D695-10: Standard Test Method for Compressive Properties of Rigid Plastics. ASTM International
(2010)
Bahr, D.F., Watkins, C.M., Kramer, D.E., Gerberich, W.W.: Yield point phenomena during indentation. In:
Fundamentals of Nanoindentation and Nanotribology. Symposium, 13-17, vol. 522, pp. 83–88. Mater.
Res. Soc., Warrendale, PA, USA (1998). 1998
Bouzakis, K.D., Michailidis, N., Hadjiyiannis, S., Skordaris, G., Erkens, G.: Continuous FEM simulation of
the nanoindentation: actual indenter tip geometries, material elastoplastic deformation laws and univer-
sal hardness. Z. Met.kd. 93(9), 862–869 (2002)
Brinson, H., Brinson, C.: Polymer Engineering Science and Viscoelasticity: An Introduction. Springer, New
York (2008)
Briscoe, B.J., Fiori, L., Pelillo, E.: Nano-indentation of polymeric surfaces. J. Phys. D, Appl. Phys. 31(19),
2395 (1998)
Caddell, R.M., Raghava, R.S., Atkins, A.G.: Pressure dependent yield criteria for polymers. Mater. Sci. Eng.
13(2), 113–120 (1974)
Cheng, Y.T., Cheng, C.M.: Relationships between initial unloading slope, contact depth, and mechanical
properties for spherical indentation in linear viscoelastic solids. Mater. Sci. Eng.: A 409(1–2), 93–99
(2005a)
Cheng, Y.T., Cheng, C.M.: Relationships between initial unloading slope, contact depth, and mechanical
properties for conical indentation in linear viscoelastic solids. J. Mater. Res. 20(4), 1046–1053 (2005b)
Cheng, L., Xia, X., Scriven, L.E., Gerberich, W.W.: Spherical-tip indentation of viscoelastic material. Mech.
Mater. 37(1), 213–226 (2005)
Cirnu, M.I.: Linear discrete convolution and its inverse. Part 2. Deconvolution. J. Inf. Syst. Oper. Manag.
4(2), 43–55 (2010)
Dassault Systemes Simulia Corp.: ABAQUS. In: Providence, RI, USA, (2014)
Mech Time-Depend Mater (2022) 26:1–20 19

Donohue, B.R., Ambrus, A., Kalidindi, S.R.: Critical evaluation of the indentation data analyses methods
for the extraction of isotropic uniaxial mechanical properties using finite element models. Acta Mater.
60(9), 3943–3952 (2012)
Fischer-Cripps, A.C.: A review of analysis methods for sub-micron indentation testing. Vacuum 58, 569–585
(2000)
Fischer-Cripps, A.C.: Nanoindentation. Mechanical Engineering Series. Springer, New York (2002)
Gauthier, M.M.: Engineered Materials Handbook. ASM International, Russel, Township, OH (1995)
Graham, G.A.C.: The contact problem in the linear theory of viscoelasticity. Int. J. Eng. Sci. 3(1), 27–46
(1965)
Gutierrez-Lemini, D.: Engineering Viscoelasticity. Springer, New York (2014)
Hemker, K.J., Sharpe, W.N.: Microscale characterization of mechanical properties. In: Annual Review of
Materials Research, vol. 37, pp. 93–126. Annual Reviews, Palo Alto (2007)
Huang, G., Lu, H.: Measurement of Young’s relaxation modulus using nanoindentation. Mech. Time-Depend.
Mater. 10(3), 229–243 (2006)
Huang, G., Lu, H.: Measurements of two independent viscoelastic functions by nanoindentation. Exp. Mech.
47(1), 87–98 (2007)
Huang, G., Wang, B., Lu, H.: Measurements of viscoelastic functions of polymers in the frequency-domain
using nanoindentation. Mech. Time-Depend. Mater. 8(4), 345–364 (2004)
Hunter, S.C.: The Hertz problem for a rigid spherical indenter and a viscoelastic half-space. J. Mech. Phys.
Solids 8(4), 219–234 (1960)
Jäger, A., Lackner, R.: Identification of viscoelastic model parameters by means of cyclic nanoindentation
testing. Int. J. Mater. Res. 99(8), 829–835 (2008)
Jakes, J.E., Lakes, R.S., Stone, D.S.: Broadband nanoindentation of glassy polymers: Part I. Viscoelasticity.
J. Mater. Res. 27(2), 463–474 (2012)
Jaya, N.B., Alam, M.Z.: Small-scale mechanical testing of materials Curr. Sci. 105(8), 1073–1099 (2013)
Kalidindi, S.R., Pathak, S.: Determination of the effective zero-point and the extraction of spherical nanoin-
dentation stress–strain curves. Acta Mater. 56(14), 3523–3532 (2008)
Kalidindi, S.R., Mohan, S., Rossi, A.: Mechanical characterization of mesoscale interfaces using indentation
techniques. JOM 69(1), 22–29 (2017)
Khosravani, A., Caliendo, C.M., Kalidindi, S.R.: New insights into the microstructural changes during the
processing of dual-phase steels from multiresolution spherical indentation stress-strain protocols. Metals
10(1), 18 (2020)
Knauss, W.G., Emri, I., Lu, H.: Mechanics of polymers: viscoelasticity. In: Sharpe, W.N. (ed.) Springer
Handbook of Experimental Solid Mechanics, pp. 49–95. Springer, Boston (2008)
Kreyszig, E.: Advanced Engineering Mathematics. Wiley, Hoboken (2010)
Lakes, R.: Viscoelastic Materials. Cambridge University Press, Cambridge (2009)
Lakes, R.S., Wineman, A.: On Poisson’s ratio in linearly viscoelastic solids. J. Elast. 85(1), 45–63 (2006)
Lee, E.H.: Stress analysis in viscoelastic materials. J. Appl. Phys. 27(7), 665–672 (1956)
Lee, E.H., Radok, J.R.M.: The contact problem for viscoelastic bodies. J. Appl. Mech. 27(3), 438–444 (1960)
Li, X., Bhushan, B.: A review of nanoindentation continuous stiffness measurement technique and its appli-
cations. Mater. Charact. 48(1), 11–36 (2002)
Martynova, E.: Determination of the properties of viscoelastic materials using spherical nanoindentation.
Mech. Time-Depend. Mater. 20(1), 85–93 (2016)
Mase, G.: Schaum’s Outline of Continuum Mechanics. McGraw-Hill Education, New York (1970)
Moseson, A.J., Basu, S., Barsoum, M.W.: Determination of the effective zero point of contact for spherical
nanoindentation. J. Mater. Res. 23(1), 204–209 (2008)
Moseson, A.J., Basu, S., Barsoum, M.W.: (2009). U.S. Patent Application No. 12/184,711
Odegard, G., Gates, T., Herring, H.: Characterization of viscoelastic properties of polymeric materials through
nanoindentation. Exp. Mech. 45(2), 130–136 (2005)
Oyen, M.L.: Spherical indentation creep following ramp loading. J. Mater. Res. 20(8), 2094–2100 (2005)
Patel, D.K., Kalidindi, S.R.: Correlation of spherical nanoindentation stress-strain curves to simple compres-
sion stress-strain curves for elastic-plastic isotropic materials using finite element models. Acta Mater.
112, 295–302 (2016)
Patel, D.K., Kalidindi, S.R.: Estimating the slip resistance from spherical nanoindentation and orientation
measurements in polycrystalline samples of cubic metals. Int. J. Plast. 92, 19–30 (2017)
Pathak, S., Kalidindi, S.R.: Spherical nanoindentation stress–strain curves. Mater. Sci. Eng. R 91, 1–36
(2015)
Pathak, S., Kalidindi, S.R., Klemenz, C., Orlovskaya, N.: Analyzing indentation stress–strain response of
LaGaO3 single crystals using spherical indenters. J. Eur. Ceram. Soc. 28(11), 2213–2220 (2008)
Pathak, S., Stojakovic, D., Doherty, R., Kalidindi, S.R.: Importance of surface preparation on the nano-
indentation stress-strain curves measured in metals. J. Mater. Res. 24(3), 1142–1155 (2009a)
20 Mech Time-Depend Mater (2022) 26:1–20

Pathak, S., Stojakovic, D., Kalidindi, S.R.: Measurement of the local mechanical properties in polycrystalline
samples using spherical nanoindentation and orientation imaging microscopy. Acta Mater. 57(10),
3020–3028 (2009b)
Pharr, G.M., Strader, J.H., Oliver, W.C.: Critical issues in making small-depth mechanical property mea-
surements by nanoindentation with continuous stiffness measurement. J. Mater. Res. 24(3), 653–666
(2009)
Spitzig, W., Richmond, O.: Effect of hydrostatic pressure on the deformation behavior of polyethylene and
polycarbonate in tension and in compression. Polym. Eng. Sci. 19(16), 1129–1139 (1979)
Srikar, V.T., Spearing, S.M.: A critical review of microscale mechanical testing methods used in the design
of microelectromechanical systems. Exp. Mech. 43(3), 238–247 (2003)
Ting, T.C.T.: The contact stresses between a rigid indenter and a viscoelastic half-space. J. Appl. Mech. 33(4),
845–854 (1966)
Vachhani, S.J., Doherty, R.D., Kalidindi, S.R.: Effect of the continuous stiffness measurement on the me-
chanical properties extracted using spherical nanoindentation. Acta Mater. 61(10), 3744–3751 (2013)
Vachhani, S.J., Doherty, R.D., Kalidindi, S.R.: Studies of grain boundary regions in deformed polycrystalline
aluminum using spherical nanoindentation. Int. J. Plast. 81, 87–101 (2016)
Weaver, J.S., Kalidindi, S.R.: Mechanical characterization of Ti-6Al-4V titanium alloy at multiple length
scales using spherical indentation stress-strain measurements. Mater. Des. 111, 463–472 (2016)
Zhou, Z., Lu, H.: On the measurements of viscoelastic functions of a sphere by nanoindentation. Mech.
Time-Depend. Mater. 14(1), 1–24 (2010)