Initial Measurements On The Effect of Stress On P - and S-Wave Vel

UNLV Theses, Dissertations, Professional Papers, and Capstones
12-1-2020
Initial Measurements on the Effect of Stress on P- and S-wave

Velocities in Olivine
Taryn Traylor
Follow this and additional works at: https://digitalscholarship.unlv.edu/thesesdissertations
Part of the Geophysics and Seismology Commons
Repository Citation
Traylor, Taryn, "Initial Measurements on the Effect of Stress on P- and S-wave Velocities in Olivine" (2020).
UNLV Theses, Dissertations, Professional Papers, and Capstones. 4088.
http://dx.doi.org/10.34917/23469763
This Thesis is protected by copyright and/or related rights. It has been brought to you by Digital Scholarship@UNLV
with permission from the rights-holder(s). You are free to use this Thesis in any way that is permitted by the
copyright and related rights legislation that applies to your use. For other uses you need to obtain permission from
the rights-holder(s) directly, unless additional rights are indicated by a Creative Commons license in the record and/
or on the work itself.
This Thesis has been accepted for inclusion in UNLV Theses, Dissertations, Professional Papers, and Capstones by
an authorized administrator of Digital Scholarship@UNLV. For more information, please contact
[email protected].
INITIAL MEASUREMENTS ON THE EFFECT OF STRESS ON P- AND S-WAVE
VELOCITIES IN OLIVINE
By
Taryn Traylor
Bachelor of Science – Geology

Texas A&M University
2017
A thesis submitted in partial fulfillment

of the requirements for the
Master of Science – Geoscience
Department of Geoscience
College of Sciences
The Graduate College
University of Nevada, Las Vegas

December 2020
Thesis Approval
The Graduate College
The University of Nevada, Las Vegas
November 20, 2020
This thesis prepared by
Taryn Traylor
entitled
Initial Measurements on the Effect of Stress on P- and S-Wave Velocities in Olivine
is approved in partial fulfillment of the requirements for the degree of
Master of Science – Geoscience

Department of Geoscience
Pamela Burnley, Ph.D. Kathryn Hausbeck Korgan, Ph.D.

Examination Committee Chair Graduate College Dean
Michael Wells, Ph.D.

Examination Committee Member
Oliver Tschauner, Ph.D.

Examination Committee Member
Moses Karakouzian, Ph.D.

Graduate College Faculty Representative
ii
Abstract
It is well known that elasticity is a key physical property in the determination of the structure and
composition of the Earth and provides critical information for the interpretation of seismic data.
This study investigates the stress-induced variation in elastic wave velocities, known as the
acoustoelastic effect, in San Carlos olivine. A recently developed experimental ultrasonic
acoustic system, the Directly Integrated Acoustic System Combined with Pressure Experiments
(DIASCoPE), was used with the D-DIA multi-anvil apparatus to transmit ultrasonic sound waves
and collect the reflections. We use the DIASCoPE to obtain longitudinal (P) and shear (S) elastic
wave velocities from the sample which we compare to our known stress state in the D-DIA
derived from synchrotron X-ray diffraction. We use elastic-plastic self-consistent (EPSC)
numerical modeling to forward model X-ray diffraction data collected in D-DIA experiments to
obtain the macroscopic stress on our sample. We can observe the relationship between the
relative elastic wave velocity change (ΔV/V) and macroscopic stress to determine the
acoustoelastic constants, and interpret our observations using the linearized first order equation
based on the model proposed by Hughes and Kelly (1953). This work supports the presence of
the acoustoelastic effect in San Carlos olivine, which can be measured as a function of pressure
and temperature. This study will aid in our understanding of the acoustoelastic effect and provide
a new experimental technique to measure the stress state in elastically deformed geologic
materials at high pressure conditions.
iii
Acknowledgements
I would like to extend my deepest gratitude to my advisor and mentor Dr. Pamela
Burnley for providing the opportunity to pursue my passion for Geology and truly believing in
me as a scientist. I would also like to extend my gratitude to Dr. Matthew Whitaker, for his
collaboration in this research and instrumentation that made this project possible.
I would like to thank my advisory committee Dr. Michael Wells, Dr. Oliver Tschauner,
and Dr. Moses Karakouzian for their involvement and encouragement.
I would like to acknowledge Dr. Haiyan Chen, Dr. Shirin Kaboli, Dawn Reynoso, and
Jason Reek for their assistance in data collection.
Special thanks must be given to my office mate and friend, Genevieve Kidman. You
made me feel welcomed into the NeRD Lab group from day one, and our scientific discussions
have greatly aided in my development as a scientist.
Finally, I would like to thank my family and friends for their continued support. I would
especially like to thank my mom and Dalton McCaffrey for their endless motivation, love, and
support.
This research was supported by a grant from the National Science Foundation under
award NSF-EAR13613399, and by the National Nuclear Security Administration under the
Stewardship Science Academic Alliances program through DOE Cooperative Agreement #DE-
NA0001982. This research used resources of the Advanced Photon Source, a U.S. Department of
Energy (DOE) Oﬃce of Science User Facility operated for the DOE Oﬃce of Science by
Argonne National Laboratory under Contract No. DE-AC02-06CH11357. Use of the 6-BM-B
beamline was supported by COMPRES, the Consortium for Materials Properties Research in
iv
Earth Sciences, under NSF Cooperative Agreement No. EAR 16-61511 and by the Mineral
Physics Institute, Department of Geosciences, Stony Brook University.
v
Table of Contents
Abstract........................................................................................................................................ iii
Acknowledgements ..................................................................................................................... iv
List of Tables .............................................................................................................................. vii
List of Figures............................................................................................................................ viii
1: Initial Measurements on the Effect of Stress on P- and S-wave Velocities in Olivine .............1
Appendix 1: Tables......................................................................................................................38
Appendix 2: Figures ....................................................................................................................49
References ...................................................................................................................................63
Curriculum Vitae .........................................................................................................................67
vi
List of Tables
Table 1 Acoustoelastic constant values ...................................................................................... 38
Table S1 Unit cell parameters used in EPSC models ................................................................. 39
Table S2 Experimental conditions for each deformation sequence ............................................ 40
Table S3 Single crystal elastic constants (Cij) used in EPSC models ......................................... 41
Table S4 Slip Systems used in each EPSC model to fit the experimental diffraction data ........ 42
Table S5 Acoustoelastic Constants comparison table with literature ......................................... 43
Table S6 San Carlos Olivine velocity compilation of literature ................................................. 46
vii
List of Figures
Figure 1 Schematic diagram of ultrasonic wave propagation in an experimental set-up ........... 49
Figure 2 Schematic diagram illustrating how travel time measurements are determined ......... 50
Figure 3 Comparison of acoustoelastic slope at different pressure conditions ........................... 51
Figure 4 Comparison of acoustoelastic slope for differential ram advancement and retraction 52
Figure 5 Acoustoelastic constants as a function of pressure and temperature ............................ 53
Figure S1 Schematic cross-sectional view of the D-DIA apparatus ........................................... 54
Figure S2 Schematic of sample assembly for San_381and San_416 ......................................... 55
Figure S3 Thermocouple temperature vs. power calibration curve for San_381 ....................... 56
Figure S4 Schematic of the X-ray beam propagation during an experiment .............................. 57
Figure S5 Lattice strain vs. sample strain data for the 4 deformation sequences of San_381 .... 58
Figure S6 Lattice strain vs. sample strain data for the 3 deformation sequences of San_416 .... 59
Figure S7 Communication path for DIASCoPE ultrasonic interferometry measurements ........ 60
Figure S8 Comparison of acoustoelastic constants to Egle and Bray (1976) ............................. 61
Figure S9 Summary of San Carlos Olivine P- and S-wave velocity measurements .................. 62
viii
Initial Measurements on the Effect of Stress on P- and S-wave Velocities in Olivine
Background
Introduction
The measurement of mineral elastic properties through experimental studies is crucial for seismic
data interpretation and can aid in our understanding of the deformation processes that shape our
Earth. Measurement of acoustic velocities, absolute pressure determination, and derivation of
thermoelastic equations of state for various materials is traditionally done through experimental
collection of ultrasonic interferometry measurements in conjunction with synchrotron X-
radiation in a multi-anvil apparatus (Li et al., 2004). Lab-based elastic constants and velocity
measurements are used by seismologists to interpret the Earth’s interior, however, the constants
derived from the susceptibility of seismic velocity changes to stress perturbation has yet to be
measured to deep Earth conditions.
We evaluate the acoustoelasticity of polycrystalline San Carlos olivine at high pressure
and high temperature conditions analogous to the Earth's lithospheric mantle through the
measurement of ultrasonic wave velocities and the stress state during uniaxial deformation. The
current experimental set up allows us to investigate the acoustoelastic effect for two ultrasonic
waves; a longitudinal (P) wave with propagation and polarization parallel to the applied
compression direction (LW11), and a shear (S) wave with propagation parallel to the applied
stress and polarization normal to stress (SW12). Hughes and Kelly (1953) derived a series of
equations under the theory of second order elastic deformation of solids to obtain the third-order
elastic constants of isotropic materials from wave velocity measurements. A first-order
linearization of these equations provides a direct relationship between the relative ultrasonic
1
wave velocity change (ΔV/V) and the applied uniaxial stress for waves propagating parallel to
the applied stress (Chaki and Bourse, 2009, and references therein).
Nonlinear Elastic Behavior of Solids
The classical linear theory of elasticity encompasses the generalized Hooke's law, where for
infinitesimal deformation of an elastic material, the stress-strain relationship is expressed as a
first-order approximation. This first-order theory is derived from expressing the strain-energy
(W) function in terms of first- and second-degree strain (ε) products, with the corresponding first
and second-order elastic constants:
1
𝑊 = 𝐶𝑖𝑗 𝜀𝑖𝑗 + 𝐶𝑖𝑗𝑘𝑙 𝜀𝑖𝑗 𝜀𝑘𝑙 (1)
2
Differentiation of this expression with respect to strain, and the understanding that a material will
not store energy if it is not deformed, results in an expression for stress in terms of strain and the
second order elastic constants:
𝜎𝑖𝑗 = 𝐶𝑖𝑗𝑘𝑙 𝜀𝑘𝑙 (2)
Under this theory, P- and S-wave velocities remain constant when a stress is applied. However, it
is known that a change in the state of stress does have an effect on wave velocities in solids,
including rocks. Also, rocks may experience large deformations and behave plastically, resulting
in violation of the assumptions of the linear elastic theory.
An extension of the linear elastic theory was proposed by Murnaghan (1951) to include
finite deformation in elastically isotropic solids. This theory departs from the classical linear
elasticity theory by first including higher-order terms in the strain energy function (to the third-
2
order), and second, the resulting deformation is finite. For finite deformation, the initial and final
coordinates of a point are now defined, and either set of coordinates may be used as the
independent variables. Murnaghan’s (1951) theory uses the Lagrangian system exclusively, in
which the initial coordinates of a point are the independent variables. Murnaghan includes third-
order elastic constants defined as l, m, and n, in addition to the Lamé second-order elastic
constants, λ and μ. With terms in the strain energy function defined to the third-order,
experimental determination of the third-order elastic constants has shown that we can use the
acoustoelasticity theory to describe the dependence of ultrasonic wave velocities on the stress
state as the waves propagate through a solid (Smith, 1963).
Acoustoelasticity
The change in wave velocities that occurs when an elastic material is subjected to static stress,
deemed "acoustoelasticity" in 1959 by Benson and Raelson, has been studied for over half a
century. The idea of a stress-induced wave velocity change in a solid is derived from the
photoelastic method of stress analysis in which a change in the index of refraction occurs when a
polarized light beam propagates through a stressed optically transparent material (Benson and
Raelson, 1959). Previous studies have used acoustoelasticity to analyze the distribution of both
residual and applied stress, in various materials such as metal, concrete, wood, and rocks, as well
as to determine the third-order elastic constants. To study the acoustoelasticity of a material
many theories of acoustoelasticity have been proposed, including Hughes and Kelly (1953),
Toupin and Bernstein (1961), Thurston and Brugger (1964), Johnson (1981), Dey et al., (1984),
and Man and Lu (1987). For this study, we use the Hughes and Kelly (1953) theory on the
3
second-order elastic deformation of solids due to its applicability and ease of use, which we
discuss below.
Hughes and Kelly (1953) formulated a series of empirical relationships (equations 3a-3e)
to calculate elastic P- and S-wave velocities in stressed solids using the nonlinear behavior laws
derived by Murnaghan (1951). Hughes and Kelly evaluated P- and S-wave velocities as a
function of stress for the case of hydrostatic pressure and uniaxial compression for polystyrene,
Armco iron, and Pyrex glass. They used this data to determine the third-order elastic constants
for each material. To determine Murnaghan's third-order elastic constants l, m, and n for an
initially homogeneous isotropic material under uniaxial stress in principal direction one (Figure
1), the following equations are used:
2 𝜎11 𝜆+𝜇
𝜌0 𝑉11 = 𝜆 + 2µ − 2 [2𝑙 + 𝜆 + (4𝑚 + 4𝜆 + 10𝜇)] (3a)
3(𝜆+3𝜇) 𝜇
2 𝜎11 𝜆𝑛
𝜌0 𝑉12 =µ− 2 [𝑚 + + 4𝜆 + 4𝜇] (3b)
3(𝜆+3𝜇) 4𝜇
2 𝜎11 2𝜆
𝜌0 𝑉22 = 𝜆 + 2µ − 2 [2𝑙 − (𝑚 + 𝜆 + 2𝜇)] (3c)
3(𝜆+ 𝜇) 𝜇
3
2 𝜎11 𝜆𝑛
𝜌0 𝑉21 =µ− 2 [𝑚 + + 𝜆 + 2𝜇] (3d)
3(𝜆+3𝜇) 4𝜇
2 𝜎11 𝜆+𝜇
𝜌0 𝑉23 =µ− 2 [𝑚 + 𝑛 − 2𝜆] (3e)
3(𝜆+3𝜇) 2𝜇
where the elastic wave velocities (V) contain subscripts 1, 2, and 3, with the first index denoting
the direction of wave propagation and the second index denoting the direction of polarization, σ11
is the applied uniaxial compressive stress acting along the X1 axis, λ and μ are the Lamé second-
4
order elastic constants, and ρ0 is the material density in the unstressed state. Equations 3c-3e are
the only equations necessary to determine the third-order elastic constants, with equations 3a and
3b providing a check on these values (Hughes and Kelly, 1953).
The acoustoelastic effect, the change in wave velocities due to the imposition of stress, is
described by the acoustoelastic constants (Aij) which can be derived by a linearization of the
system of equations (3a-3e) above (Hughes and Kelly, 1953; Egle and Bray, 1976; Johnson et
al., 1986; Chaki and Bourse, 2009; Lillamand et al., 2010). Thus:
𝑉𝑖𝑗𝜎 −𝑉𝑖𝑗0
𝐴𝑖𝑗 𝜎11 = (4)
𝑉𝑖𝑗0
where σ11 is the applied uniaxial compressive stress acting along the X1 axis, Vijσ is the wave
velocity during deformation and Vij0 is the wave velocity in the hydrostatic state before
deformation, with i corresponding to the direction of propagation and j corresponding to the
direction of polarization. Since Vij is governed by the second-order and third-order elastic
constants, the acoustoelastic constants are therefore, also a function of the second-order and
third-order elastic constants. In this work, we will confine ourselves to the measurement of A11,
which can be determined from measurements of the P-wave (LW11) and A12, which can be
determined from measurements of the S-wave (SW12), as they both propagate parallel to the
applied stress.
Stress-induced Velocity Changes in Metallic Materials and Concrete
Ultrasonic pulse velocity (UPV) tests use an ultrasonic pulse passing through a material to study
the elastic properties and determine the level of heterogeneity and damage that affects the
5
strength in industrial materials, such as concrete and metallic materials. UPV tests can also
measure the stress state in a material. As previously mentioned, Hughes and Kelly (1953)
experimentally confirmed their theory of acoustoelasticity through the determination of third-
order elastic constants for various materials under uniaxial and hydrostatic compression.
Researchers have since investigated a range of metallic materials (e.g., Bergman and
Shahbender, 1958; Bateman et al., 1961; Rollins et al., 1963, Smith et al., 1966; Crecraft, 1967;
Egle and Bray, 1976; Nogueira, 2017) and concrete (e.g., Lillamand et al., 2010; Bompan and
Haach, 2018) using Hughes and Kelly’s theory as well as alternative theories of acoustoelasticity
mentioned above. Metallic material and concrete studies dominate a large portion of the
acoustoelastic literature, which is primarily due to their importance in industrial uses. Many
previous studies have used the theory of Hughes and Kelly (1953) to deduce the third-order
elastic constants and have thus only used the equations involving the P- and S-wave velocities in
the transverse direction to the applied stress (equations 3c-3e). The following summarized
studies have used an ultrasonic technique to evaluate the acoustoelasticity of a metallic material
or concrete under the theory of Hughes and Kelly, in addition, a few of the studies have
determined the acoustoelastic constants.
Metallic Materials. Crecraft (1967) showed that the acoustoelastic effect could be evaluated for
various structural materials including nickel-steel, copper, and aluminum. In 1976, Egle and
Bray experimentally derived the third-order elastic constants and the relationship between
relative wave velocity change and strain, rather than stress, which they termed "acoustoelastic
constants" for samples of steel; this is in contrast to the above authors who have defined
acoustoelastic constants as a function of stress. Egle and Bray measured all five possible wave
velocities and calculated the “acoustoelastic constants” in two ways: 1) using equations 4a-4e
6
stated in Egle and Bray (1976) which determined ΔV/V with axial strain using Poisson’s ratio
and second- and third-order elastic constants, and 2) determining the slope of the relationship
between ΔV/V and axial strain. The two calculations of the “acoustoelastic constants” agreed
within the experimental error, showing that the first-order relationship between the measured
ΔV/V and axial strain is adequate to describe the stress dependence of wave velocity changes. In
metallic materials, the third order elastic constants are generally negative and approximately an
order of magnitude greater than the second order elastic constants. Using the second and third
order elastic constants from the literature to calculate the acoustoelastic constants (see Table S5),
in general, these calculated acoustoelastic constants for metals yield ~3% wave velocity change
per GPa for P-waves and 0.6% wave velocity change per GPa for S-waves, and positive for both
P and S-waves propagating along the axis of compression.
Concrete. Concrete is a heterogeneous material fabricated with varying ratios of cement, sand,
and aggregate. Concrete can be considered a homogeneous material when the wavelength of an
ultrasonic wave is large compared to the heterogeneity size, and isotropic due to the random
distribution of the aggregate. Concrete shows a nonlinear behavior related to the presence of
microcracks and porosity in the sample even before damage. Lillamand et al. (2010) and
Bompan and Haach (2018) both conducted experiments on concrete samples under uniaxial
compressive loading using the first-order approximation (equation 4) to link the ultrasonic wave
velocity change to the mechanical stress state. Both studies showed that an increase in applied
stress due to compression resulted in increased wave velocities, observing the acoustoelastic
effect. These studies also observed a positive percentage of wave velocity change for P- and S-
waves propagating along the compression axis, with the greatest increase in velocity observed
for the P-wave. However, the microstructural state of the concrete has an important impact on
7
the magnitude of the acoustoelastic constants (Ankay and Zhang, 2019). Calculating the
acoustoelastic constants from the second and third order elastic constants, we observe that
concrete shows ~1-400% wave velocity change per GPa for P-waves and ~1-177% wave
velocity change per GPa for S-waves.
Stress-induced Velocity Changes in Geologic Materials
The nonlinear elastic behavior of rocks, as seen by departure from the generalized linear stress-
strain relationship of Hooke's law, is a well-observed occurrence (e.g., Birch, 1966; Johnson and
Rasolofosaon, 1996; Winkler and Liu, 1996). The opening and closing of pre-existing cracks and
damage at grain boundary contacts and crack tips are the typical causes of stress-induced
velocity changes published in the literature. Previous geologic studies have derived the third-
order elastic constants, using the theory of acoustoelasticity, for various sedimentary, igneous,
and metamorphic rocks. The experiments performed in these studies are typically at ambient to
low confining pressures (Zamora, 1990; Nur and Simmons, 1969; Lucet, 1989; Liu et al., 2007;
Winkler and Liu, 1996; Xie et al., 2018).
Johnson and Rasolofosaon (1996) used the finite elastic deformation theory of
Murnaghan (1951) to describe stress-induced velocity changes from published experimental data
for sandstone and marble (Zamora, 1990), Barre granite (Nur and Simmons (1969), and
Bauvilliers limestone (Lucet, 1989). Johnson and Rasolofosaon (1996) observed that the third-
order elastic constants are two to three orders of magnitude greater than the second-order elastic
constants, which indicates a significant deviation from the linear stress-strain relationships at
very small strain on the order of 10-3.
8
Wang et al. (2015) developed a method, in the framework of the theory of
acoustoelasticity, to determine the pressure in-situ during a high pressure-high temperature
experiment using synchrotron X-radiation and ultrasonic interferometry. This method had
success in determining the pressure, and differential stress due to the apparatus configuration, in
a Kawaii-type multi-anvil apparatus for a traditional ultrasonic experiment; these types of
experiments collect ultrasonic measurements during the increase and decrease in confining
pressure. However, Wang et al. did not quantify the acoustoelastic constants, nor did they
perform experiments under uniaxial load.
General Objective
To our knowledge, the acoustoelasticity of geologic materials during a uniaxial
deformation experiment at high pressure and high temperature conditions has not been evaluated.
In addition, previous studies investigating the stress-induced wave velocity change in rocks, with
the exception of Wang et al. (2015), have been conducted at low confining pressures where the
primary mechanisms contributing to a wave velocity change are microcracking and porosity
closure. There is a lack of studies published with experimental work done at high confining
pressures where these effects are less applicable. Whitaker et al. (2017) attributes this to the
inability of current ultrasonic interferometry technology to collect ultrasonic spectra fast enough
to observe the time-dependent phenomena that occur during a high pressure-high temperature
deformation experiment. In addition, the lack of integration of ultrasonic interferometry
technology into a deformation experiment has delayed advancement in the study of the
acoustoelastic effect. The current integration of the DIASCoPE experimental acoustic system
(Whitaker et al., 2017) into the D-DIA multi-anvil apparatus at the 6-BM-B synchrotron
9
beamline at Argonne National Laboratory, Chicago, Illinois, presents the opportunity to perform
steady-state deformation experiments at high pressure and high temperature in conjunction with
ultrasonic interferometry measurements. This capability allows us to study the stress dependence
of infinitesimal elastic wave velocities in a solid. We present an experimental study evaluating
the acoustoelasticity of San Carlos olivine at upper mantle pressure and temperature conditions.
With pressure conditions ranging from 3.2-10.5 GPa, the results of this study, including the
numerical value of the acoustoelastic constants for two samples of polycrystalline San Carlos
olivine, to our knowledge, will be the first to quantify the acoustoelasticity of olivine at upper
mantle conditions directly.
10
Methods
The ultrasonics-modified deformation-DIA (U-DIA) experiments were conducted using the D-
DIA multi-anvil apparatus (Durham et al., 2002) and DIASCoPE acoustic system (Whitaker et
al., 2017) located at the 6-BM-B beamline at the Advanced Photon Source at Argonne National
Laboratory, Chicago, Illinois. The D-DIA multi-anvil apparatus combined with a synchrotron
beamline provides the ability to measure sample stress and strain using in-situ X-ray techniques
during the deformation experiment. The incorporation of the DIASCoPE acoustic system into a
traditional D-DIA experiment allowed for simultaneous travel time measurements when
deforming. Using the sample length measurements from synchrotron X-radiographic imaging
and P- and S-wave travel times from the DIASCoPE, the elastic P- and S-wave velocities were
determined. Powder diffraction data collected from the sample during deformation was then
interpreted through elastic-plastic self-consistent (EPSC) modeling (Tomé and Oliver, 2002)
following the strategies devised by Burnley (2015) and Burnley and Kaboli (2019) to determine
the macroscopic stress on the sample. Correlating the elastic wave velocities with the
macroscopic stress, the acoustoelastic constants (Aij) were determined.
D-DIA Apparatus
The D-DIA multi-anvil apparatus is used to generate high pressure-high temperature conditions
(Durham et al., 2002). Three tungsten carbide (WC) anvils and one sintered diamond anvil are
horizontally distributed at 90 degrees, and two WC anvils are vertically opposed (Figure S1). To
achieve high pressure conditions, the six anvils are simultaneously advanced by the main ram.
The vertical anvils can then be independently displaced to uniaxially deform the sample, while
11
the horizontal anvils retract to maintain a constant force and sample cell volume (Durham et al.,
2002).
Sample Assembly Configuration. To integrate the DIASCoPE into a traditional D-DIA
experiment, a hybrid sample assembly was developed that included aspects of an ultrasonics
experiment and a deformation experiment (Figure S2). The sample assembly used for each
experiment is shown in the supplementary section. A cubic pressure medium of mullite is used
with a ~3 mm vertical hole drilled through the center. This hole accommodates a series of
concentric sleeves consisting of a crushable alumina support sleeve, a thin walled graphite
furnace and BN confining media. In the center of the BN confining media the sample, in series
with a fully dense sintered alumina (Al2O3) upper piston, is enclosed within a 25 µm thick nickel
sleeve. The sample consisted of pulverized San Carlos olivine that was isostatically hot-pressed
at 1150 °C and 296.5 MPa for 8 hours, cored to produce a right cylinder, and then polished to ¼
µm to produce parallel ends. The nickel sleeve is used to help prevent the iron in the olivine from
reducing. A cylinder of Coors AD-998 polycrystalline Al2O3 below the sample was used as a
waveguide to couple the WC anvil to the sample and will be referred to as the buffer rod from
now on. A crushable Al2O3 plug above the upper piston and BN sleeve was used to transmit the
load from the top anvil. All surfaces intersecting the ultrasonic wave path were polished to 1
micron to ensure the interfaces were flat and parallel within 0.05°; the interfaces of the sample
were polished to ¼ micron. 1 µm thick gold (Au) foils were placed at the sample interfaces, as
well as the bottom anvil-buffer rod interface, to improve the coupling of these surfaces and
minimize the loss of energy of the transmitted ultrasonic waves. In addition, the Au foils have a
high X-ray absorption compared to the other cell materials, so the foils at the sample interfaces
were used as strain markers in the X-radiographic images. A side-entry thermocouple was
12
inserted into the sample assembly to directly measure the sample temperature; the side entry
placement was necessary to ensure that the thermocouple did not interfere with the ultrasonic
wave travel path.
Experimental Procedure
The San_381 experiment was compressed to ~7 GPa, at room temperature, as estimated by
olivine diffraction. The temperature was then raised to 1000 °C and the sample was annealed for
~1 hour and 40 minutes to relax the stress in the grains developed during cold loading. Grain
growth was inferred in the sample by changes in diffraction peak intensity, so the temperature
was cooled to 850 °C for the initial advance of the differential rams. Once the rams were in
position to start the experiment, differential stress was observed again, so the temperature was
raised to 980 °C to relax the remaining differential stress. The total annealing process lasted 3
hours and 26 minutes. The sample temperature was then lowered to the first experimental
temperature of 450 °C.
The San_416 experiment was compressed to ~10 GPa, at room temperature, as estimated
by olivine diffraction. The temperature was then raised to 980 °C and the sample was annealed
for 43 minutes. Differential stress was still present, so the temperature was raised to 1130 °C for
37 minutes, and raised again to 1240 °C for 17 minutes. The total annealing process lasted ~1
hour and 37 minutes. The sample temperature was then lowered to the first experimental
temperature of 450 °C.
The initial annealing phase allows the cell materials to extrude through the gaps between
the anvils and relaxes the internal stresses within the cell assembly. This process results in a
significant pressure loss. When the temperature is decreased to the experimental condition, the
cell assembly pressure is further decreased due to thermal contraction. In our experiment, the
13
frictional behavior of the ceramic pressure media does not allow for precise pressure adjustment
after the initial compression and heating phase, therefore we chose not to modify the
experimental pressure between deformation sequences beyond the automatic feedback system
that maintains a constant oil pressure. By not modifying the pressure during the experiment, a
slight pressure increase occurred as the temperature was increased in each subsequent
deformation sequence.
For experiments San_381 and San_416, X-ray spectra were collected at the starting
condition, and then the D-DIA differential rams were advanced to deform the samples while
diffraction, radiographic, and ultrasonic measurements were made sequentially. In deformation
sequence 3 of San_381 and all deformation sequences of San_416, the sense of motion of the D-
DIA rams was immediately reversed upon reaching the desired amount of strain and diffraction,
radiographic, and ultrasonic measurements were made sequentially until the differential stress
was released.
In experiment San_381, the sample was uniaxially deformed at 3.2-4.4 GPa at a nominal
strain rate of ~3.5x10-6 sec-1 at 450°, 650°, 800°, and 950°C in 4 deformation sequences. In each
sequence the sample was deformed ~3-5%, and then with the exception of sequence 3 as
described above, the D-DIA motors for the differential rams were stopped. The temperature was
then raised to 900°C and the differential rams were retracted at a speed of 0.0006 mm/sec and
briefly at 0.002 mm/sec to release the differential stress. This was then followed by an additional
period of stress relation.
In experiment San_416, the sample was uniaxially deformed at 7.8-10.5 GPa and 450°,
650°, and 900°C in 3 deformation sequences. The sample was deformed ~2.5-4%, and then the
D-DIA motors for the differential rams were retracted until the differential stress was released.
14
Applying a uniaxial load resulted in a nominal strain rate of 3.2x10-6 sec-1 during differential ram
advancement, followed by a nominal strain rate of 0.9x10-6 sec-1 as the differential rams were
retracted.
In-situ Synchrotron X-ray Measurements
Measurements of Sample Strain. X-radiographic images of the sample were obtained at six-
minute intervals during each deformation sequence. The X-ray absorption contrast between the
sample and the metal foils placed at its interfaces allows the sample to be identified and the
length to be measured during the experiment. We then compare the intensity profiles of the first
photo in each deformation sequence with the subsequent photos to determine the amount of
sample strain. We determined the sample length, in pixels, for the X-radiographic image prior to
pressurization using the open-source software ImageJ (Abràmoff et al., 2004); this was done to
create a pixel to micron conversion using the known initial sample length. The pixel to micron
conversion allowed us to determine the starting length of the first photo in each deformation
sequence with ~0.3% precision. The radiographs were not taken simultaneously with diffraction
or ultrasonic measurements; therefore, sample strain as a function of time elapsed was fit with a
polynomial function to allow for the calculation of the sample strain at the time of each
diffraction and ultrasonic measurement. The sample length during deformation (𝑙) at the time of
the diffraction and ultrasonic wave measurements was then calculated using the sample strain
determined from the polynomial fit and the sample length, in microns, prior to deformation (𝑙0 )
in each deformation sequence:
𝑙 = ((𝜀 ∗ 𝑙0 ) + 𝑙0 ) (5)
15
A polynomial function was chosen to fit the sample strain vs. time data as opposed to a linear fit,
due to the observed sluggishness in the system at the start of deformation.
Measurement of Stress. Powder diffraction patterns were taken of the sample and alumina buffer
rod at alternating six-minute intervals using an array of 10 energy-dispersive detectors. The
primary detectors used in our data analysis procedure measured diffraction parallel and
perpendicular to the vertical compression axis. Two of these detectors are aligned along the
compression axis at ψ= 0°, 180° to obtain diffraction from lattice planes that are normal to
compression (Figure S4). A detector at ψ= 90° is also used to obtain diffraction from lattice
planes parallel to the compressive direction (Figure S4). The lattice strain (𝜀𝑙𝑎𝑡𝑡𝑖𝑐𝑒 ) was
calculated for each population of grains producing a diffraction peak as:
𝑑ℎ ℎ𝑘𝑙−𝑑𝑓 ℎ𝑘𝑙
𝜀𝑙𝑎𝑡𝑡𝑖𝑐𝑒 = (6)
𝑑ℎ ℎ𝑘𝑙
where 𝑑ℎ ℎ𝑘𝑙 is the d-spacing measured for a grain population immediately before deformation at
a given detector, and 𝑑𝑓 ℎ𝑘𝑙 is the d-spacing measured by a given detector during deformation.
The lattice strain is a measure of the average stress on the population of grains producing the
diffraction peak (Burnley and Zhang, 2008).
Elastic-Plastic Self-Consistent Modeling
Elastic-plastic self-consistent (EPSC) modeling (Turner and Tomé, 1994; Tomé and Oliver,
2002) is a numerical modeling technique that can be used to interpret synchrotron X-ray
diffraction measurements of lattice strain. Typically D-DIA studies have instead calculated the
differential lattice strain, which are interpreted using a method that assumes a Reuss state of
stress developed by Singh et al. (1998). However, other studies (Burnley and Zhang, 2008;
16
Hilairet et al., 2012; Raterron et al., 2013; Burnley, 2015; Burnley and Kaboli, 2019) have shown
that it does not suffice to look at this differential lattice strain to evaluate the macroscopic load
due to the heterogeneous distribution of stress among grain populations within a polycrystalline
aggregate. The EPSC model inputs include boundary conditions, crystal orientation, slip
systems, single crystal elastic constants, and unit cell dimensions. We model the eight commonly
observed slip systems in olivine, three unidirectional slip systems to simulate kink band
formation (Burnley, 2015; Kaboli et al., 2017; Burnley and Kaboli, 2019), and an isotropic
deformation mechanism composed of 30 slip planes (Burnley and Kaboli, 2019) to allow a small
amount of anelastic deformation during the initial elastic portion of the experiment. Lattice strain
vs. sample strain curves were compared with the EPSC models to determine the macroscopic
stress on the sample at any given time during the experiment.
DIASCoPE: Directly Integrated Acoustic System Combined with Pressure Experiments
The DIASCoPE is an experimental acoustic system that allows for the measurement of elastic
wave velocities during an in-situ high pressure experiment (Whitaker et al., 2017). The system
uses ultrasonic interferometry to measure acoustic wave velocities and is integrated into the D-
DIA and the 6-BM-B EPICS computing system which allows for automated collection of
experimental measurements. A dual-mode 10° Y-cut LiNbO3 piezoelectric transducer is attached
to the bottom of the lower tungsten carbide anvil (Figure 2). The transducer was used to transmit
frequencies and receive the reflections for collection of nearly synchronous P- and S-wave
measurements during the experiment. The transducer was set to a frequency of 35 MHz for S-
waves and 60 MHz for P-waves. The Pulse-Echo-Overlap method was implemented in Plot85, a
data analysis software developed for the 6-BM-B beamline, to analyze the data and obtain the P-
17
and S-wave travel times. The elastic wave velocities (V) were then calculated by dividing the
distance the acoustic wave traveled in the sample by the two-way travel time (∆𝑡):
2𝐿
𝑉= (5)
∆𝑡
where 𝐿 is the length of the sample. The uncertainty of the wave velocity measurements is a
function of the uncertainty in the absolute length of the sample and the uncertainty in the
measured travel time (~0.5 ns).
18
Experimental Results
ΔV/V for longitudinal and shear wave data is plotted as a function of macroscopic stress for
deformation sequences 1 and 2 of San_381 and San_416 during differential ram advancement
(Figure 3). The acoustoelastic constants are calculated from a linear regression of the data points
within the elastic limit, as determined by where the EPSC model needs to start including slip
activity in order to fit the diffraction data points (denoted by a black asterisk in the figure). We
observe that the ∆𝑉𝑝 ⁄𝑉𝑝0 consistently increases with increased differential stress as seen by the
positive acoustoelastic constant (A11). The ∆𝑉𝑠 ⁄𝑉𝑠0 generally decreases slightly with increased
differential stress, but is less sensitive to compression, with a slope trending near zero.
Figure 4(a-c) shows ∆𝑉𝑝 ⁄𝑉𝑝0 and ∆𝑉𝑠 ⁄𝑉𝑠0 as a function of macroscopic stress for deformation
sequences 1, 2, and 3 of experiment San_416 during differential ram advancement and retraction.
The acoustoelastic slopes during advancement to the elastic limit and retraction follow a similar
slope in the P-wave data, and in the S-wave data.
The acoustoelastic constants for P- and S-wave data are plotted as a function of pressure and
temperature in Figure 5. We do not observe temperature dependence, but minor pressure
dependence can be seen. When the temperature is comparable, it is observed that an increase in
pressure results in a decrease in the A11 acoustoelastic constant (Figure 3).
The experimental pressure and temperature conditions, wave velocity at hydrostatic stress, and
the acoustoelastic constants during differential ram advancement and retraction for each
deformation sequence are given in Table 1.
19
Discussion
Acoustoelastic Constants
Sensitivity of acoustoelastic constants. We observe that the acoustoelastic constant for the P-
wave propagating along the axis of compression (A11) has a greater sensitivity to compression;
this can be seen for other materials, including metals and concrete (Egle and Bray, 1976;
Bompan and Haach, 2018; Lillamand et al., 2010). The values of A11 for experiment San_381
are positive with an average uncertainty ± 0.0747; for San_416, the A11 values are positive with
an average uncertainty of ± 0.0150. The S-wave propagating along the axis of compression (A12)
shows less sensitivity to compression with values trending near zero. The values of A12 for
experiment San_381 data are both positive and negative with an average uncertainty ± 0.0742;
for San_416, the A12 values are negative with an average uncertainty of ± 0.0057.
Discrepancy in Differential Ram Advancement and Retraction Acoustoelastic Constants
We should expect that the ΔV/V and resulting acoustoelastic constants during differential ram
advancement and retraction would be within error of each other, as this has been observed in
other studies (Egle and Bray, 1976; Crecraft, 1967). However, we observe some differences.
These differences are likely caused by two different aspects of our experiments and analysis
procedure; potential pressure drift during the experiment and challenges with generating EPSC
models for the retraction phase of the experiments.
While the difference in the acoustoelastic constants are smaller as is reflected in similar
slopes between the advancement and retraction phases, the offset that is observed in the ΔV/V
between the acoustoelastic slopes during differential ram advancement and retraction is more
significant. In other words, with return to near hydrostatic state the wave velocity does not
20
always return to the same value. Looking at the ΔV/V value in Figure 4, at near hydrostatic
conditions for the differential ram retraction data, we can see that at 450°C the return wave
velocity at near hydrostatic is higher as indicated by the greater ΔV/V value, approximately the
same return velocity at 650°C, and at 900°C the return wave velocity is lower than the initial
hydrostatic wave velocity (as indicated by the negative ΔV/V value). It should be noted that we
assume that the pressure remains constant during each deformation cycle; however, a small
fluctuation in pressure may cause this observed offset, as the offset is not consistently positioned
above or below the differential ram advancement data for each deformation cycle. Further
analysis of the pressure evolution during each deformation cycle may provide clarity to this.
The macroscopic stress used in calculating the acoustoelastic constants is determined
through interpretation of our X-ray diffraction data using EPSC modeling. It is observed in the
EPSC models for the San_416 experiment (Figure S6) that as we increase the temperature for
each subsequent deformation sequence the model fit to the diffraction data during differential
ram retraction becomes increasingly difficult. This is partly due to the sample yielding sooner
with each subsequent temperature increase, resulting in a longer period of plastic deformation
during differential ram advancement that the model must fit. When running the EPSC model,
two mechanical processes, that include the stress or strain boundary conditions, are required to
produce a model that can interpret the diffraction data during the differential ram advancement
and during differential ram retraction. For our models, the start of iterations for the second
mechanical process begins based on the result of the first mechanical process. Therefore, a poor
fit of the diffraction data during differential ram advancement, towards the end of the plastic
deformation phase, will result in a poor fit during differential ram retraction. The macroscopic
stress output for the EPSC model will then be less precise, thus resulting in a discrepancy
21
between the acoustoelastic constant during differential ram advancement and retraction. Future
work needs to be done to understand how to best set up the second mechanical process in the
EPSC code so that modifications can be made to better fit the diffraction data during differential
ram retraction.
Acoustoelastic Constants Comparison with Literature.
As this is the first measurement of the acoustoelastic constants at high pressure, there are no
other comparable measurements with which to compare our results. However, it is informative to
make comparisons with measurements of metals at low pressure. Most measurements in the
metals literature report the third order elastic constants rather than the acoustoelastic constants,
thus, through reordering of equations 3a and 3b, the acoustoelastic constants, A11 and A12
respectively, can be determined as a function of the second and third-order elastic constants:
1 𝜆+µ
𝐴11 = − ((2(𝜆+2µ))∗(3𝜆+2µ)) ∗ (( ) ∗ (4𝜆 + 10µ + 4𝑚) + 𝜆 + 2𝑙) (7a)
µ
1 𝜆𝑛
𝐴12 = − ((2µ)∗(3𝜆+2µ)) ∗ (4𝜆 + 10µ + 𝑚 + ) (7b)
4µ
We include a negative sign in the acoustoelastic constant equations, as we define compression as
positive. Table S5 summarizes the second and third order elastic constants from various metals,
concrete, and rock from the literature and the acoustoelastic constants calculated from equations
7a-7b. We observe that the A11 and A12 values for polycrystalline San Carlos olivine at high
confining pressure follow more closely to the range of A11 and A12 values for metals at low
confining pressure. We should expect behavior more like metals at low confining pressure than
rocks at low confining pressure due to the fact that the acoustoelastic response for geologic
22
materials is attributed to pore closure and microcracks, which should not be present at high
confining pressure conditions.
Implications
A linear relationship is observed between the relative percentage of wave velocity change and
stress up to the elastic limit. Additionally, at the onset of plastic deformation the linearity
between ΔV/V and the macroscopic stress for the P-wave and S-wave data show a consistent
slope for a given percentage of strain. This relatively consistent slope suggests that acoustoelastic
measurements may provide an alternative method for deriving stress in offline experiments when
we have integrated ultrasonic data collection, but do not have access to a Synchrotron light
source.
High resolution P-wave tomography studies of the upper mantle have shown a ± 2%
wave velocity change in the cold, stressed region of a subducting slab (Zhao et al., 2017). We are
observing approximately 0.5-1% wave velocity change for the kinds of stress present in a
subducting slab. This illustrates the significance of the acoustoelastic effect and that integration
of acoustoelastic measurements may need to be considered in seismic interpretation of cold-
highly stressed regions of the Earth’s upper mantle, such as subduction zones.
23
Conclusions
The acoustoelastic effect allows us to observe the effect stress has on P- and S- wave velocities.
Using a linearization of the equations by Hughes and Kelly (1953) we can observe this effect in a
combined uniaxial deformation and ultrasonic experiment at a synchrotron beamline for P- and
S-wave velocities propagating along the axis of compression. We observe that P-wave velocities
increase with uniaxial compression, and S-wave velocities generally decrease with uniaxial
compression. Our results agree with the second order linear elasticity theory that the relative
wave velocity changes are a linear function of the stress. Greater acoustoelastic sensitivity is
observed for the P-wave propagating along the axis of compression, and the corresponding
acoustoelastic constant, A11, decreases slightly with pressure. These findings suggest that
acoustoelastic measurements may need to be considered in seismic data interpretation and
provide potential for the evaluation of stress through the relative wave velocity change in offline
experiments.
24
Supplementary Materials
Detailed Methods
Deformation DIA or D-DIA Multi-anvil Apparatus. The D-DIA multi-anvil apparatus (Durham
et al., 2002) uses six hard anvils, Tungsten Carbide (WC) or sintered diamond, in a cubic
configuration to generate high pressure-high temperature conditions. Four anvils are horizontally
distributed at 90 degrees, and two anvils are vertically opposed. To achieve high pressure
conditions, the six anvils are simultaneously advanced by the main ram. The vertical anvils can
then be independently displaced to uniaxially deform the sample (Figure S1). Hydraulic fluid is
pumped into the differential rams behind the vertical anvils allowing them to advance
independently to produce deviatoric stress, while simultaneously draining fluid from the main
ram to retract the four lateral anvils to maintain a constant force and volume of the cell (Durham
et al., 2002). A 0.0006 mm/sec motor speed of the D-DIA rams for San_381 was chosen to
produce a strain rate that would allow focus on the initial elastic low strain behavior. The sample
length for experiment San_416 was longer than the sample length for experiment San_381, so
the D-DIA motor speed was increased to 0.0008 mm/sec to obtain a similar strain rate.
Sample Assembly. The sample assembly consisted of a mullite cubic pressure medium with a 3
mm vertical hole that contained a cylindrical graphite furnace placed in an alumina support
sleeve (Figure S2). The sample of San Carlos olivine was cored from an isostatically hot pressed
block of polycrystalline San Carlos olivine. The San Carlos polycrystal was prepared by grinding
single crystals of olivine, ~5 mm in diameter, that had been tumbled to remove surface oxidation.
The fine fraction of the resulting powder was removed by settling in water and then the
remaining powder was dried under vacuum and then vacuum packed in a Ni-lined steel container
25
for hot pressing at 1150°C and 296.5 MPa to produce an approximately 5-20 µm average grain
size with some grains as large 23 µm x 70 µm. A single 2 mm diameter 1.13 mm long cylindrical
olivine sample was used for San_381. In San_416, two cylindrical olivine samples, 1.87 mm in
diameter with a combined length of 1.689 mm, were used; the use of two samples in San_416
was strictly to achieve the desired length, and therefore no foil was placed at the interface
between the samples. At the confining pressure used for San_416, the effect of two specimens
was not observed in the ultrasonic measurements. The sample in each experiment was in series
with a fully dense sintered alumina piston all enclosed within a 25 μm Ni sleeve in the center of
the cell assembly. The alumina piston was placed above the sample. The Ni sleeve acts as a
redox buffer to help stabilize the iron and is thin enough to minimize the absorption of the X-
rays. An alumina “buffer rod” was placed below the sample and a confining medium sleeve of
boron nitride (BN) enclosed the alumina piston, sample, and buffer rod within the graphite
furnace. A crushable alumina piston filled the remainder of the cylindrical hole and was used to
transmit the load from the above anvil. The interfaces of the buffer rod, sample, and the interface
of the upper alumina piston in contact with the sample were polished to ¼ µm to produce parallel
ends. One-micron thick gold (Au) foils were placed at the buffer rod-sample interface, alumina
piston-sample interface, and the buffer rod-WC anvil interface to improve the coupling of these
surfaces. The Au foils at the sample interfaces were also used as strain markers in the X-
radiographic images. A single side-entry W 3%-W 25% Re thermocouple within a fully dense
0.86 mm OD 4-hole Al2O3 insulating tube was inserted into the sample assembly to directly
measure the sample temperature. A thermocouple bead was formed by crossing the hooked ends
of the wires into the empty channels of the 4-hole insulator tube and applying Cotronics #940
Zirconia cement over the crossed wires.
26
Temperature Measurement. The U-DIA cell design incorporates a side entry thermocouple for
direct temperature measurements. However, the thermocouples are fragile and may break during
the experiment. In experiment San_381, the thermocouple was used for deformation sequence 1
but failed near the end of deformation sequence 2. Therefore, an extrapolation was made for the
temperature for the remainder of sequence 2 and sequence 3 by applying a polynomial fit to the
thermocouple data during the temperature increase (450-1000 °C) for the sample annealing
process prior to deformation sequence 2 (Figure S3). To compare the results of San_416 with
San_381, we set the sample power level for each deformation sequence in San_416 to
correspond with the power level and desired temperature for the San_381 deformation
sequences. However, with the same power level achieved, the San_416 thermocouple read a
temperature ~50-160 °C lower than the anticipated temperature based on the power-temperature
relationship from San_381 as well as the observation that this cell should generally produce
~3°C/watt. The lower temperatures read by the thermocouple in San_416 was likely due to the
thermocouple not being in direct contact with the sample. Thus, we judged this thermocouple to
be an unreliable temperature gauge.
In-situ X-ray measurements. The 6-BM-B synchrotron beamline uses bending magnets to
produce a white X-ray beam. The X-ray beam enters the assembly between gaps in the anvils
that are created by extrusion of the mullite confining media. Five of the six anvils are tungsten
carbide and the remaining anvil is sintered diamond, located downstream, to permit propagation
of the diffracted X-rays. This allows the differential stress and sample lengths to be monitored
in-situ by X-ray powder diffraction and imaging. For X-ray diffraction measurements the
incident X-ray beam is collimated to 100 µm x 100 µm by a pair of WC slits that sit upstream of
27
the D-DIA. Since APS 6-BM-B operates with a white beam, energy dispersive detectors are used
to measure diffraction. For X-ray diffraction to occur, Bragg’s condition must be satisfied:
𝑛𝜆 = 2𝑑𝑠𝑖𝑛𝜃 (S1)
where 𝑛 is the order of the reflection, 𝜆 is the wavelength of the incident X-rays, d is the
interplanar spacing (d-spacing), and θ is the angle of incidence; the wavelength of the incident
X-rays (𝜆) is related to the X-ray energy (E) by:
ℎ𝑐
𝐸= (S2)
𝜆
where h is Plank’s constant, and c is the speed of light. Therefore, the X-ray energy, the d-
spacing, and the diffraction angle are simply related. By constraining the 2θ angle using a
conical slit, the d-spacing can be uniquely determined by measuring the energy of each
diffraction peak. The conical slit system downstream of the D-DIA imposes a 6.50° diffraction
angle and defines the volume from which diffraction originates. We use a slit width such that the
diffraction volume is contained entirely within the sample, eliminating diffraction peaks from the
confining medium.
Diffraction from the sample is collected with a white X-ray source that allows each of the
10 energy dispersive detectors to collect a full powder diffraction pattern (Weidner et al., 2010).
Of the 10 detectors, we focus on diffraction collected from detectors at ψ= 0° and 180° in the
compression direction and ψ= 90° in the transverse direction (Figure S4). We cannot collect
diffraction for the other transverse detector at ψ= 270° because it is blocked by a WC anvil
during the experiment. D-spacing measurements reflect only the elastic strain in the sample as
the d-spacings are not sensitive to plastic strain (Wang et al., 2003).
28
Calibration spectra were collected at the start of the experiment; these spectra constrain
the relationship between the channel number and X-ray energy for the energy dispersive
detectors. Twenty 60-second diffraction spectra were collected and summed for the calibration
standard. Five 60-second diffraction spectra were collected and summed for both the sample and
the alumina buffer rod. Breaking the period of collection into 60 seconds per pattern allows for
spectra to be collected for different lengths of time while avoiding any intensity-based shifts that
may originate from the detector electronics. During each deformation sequence, the data
collection set includes a radiograph, P- and S-wave scans, five 60-second spectra of the sample,
radiograph, P- and S-wave scans, and five 60-second spectra of the alumina buffer rod.
We peak fit diffraction data from the compressional (ψ= 0°, 180°) and transverse
detectors (ψ= 90°) using Plot85. The olivine diffraction peaks measured to obtain d-spacing
values were (130), (131), (112), (122), (140), and (211). The point of hydrostatic stress for the
sample was determined by identifying the diffraction data set that had the smallest difference in
d-spacing value for each lattice reflection between the detectors in the compressional and
transverse direction. Using the d-spacing at the point of hydrostatic stress (𝑑ℎ ℎ𝑘𝑙) and the d-
spacings during deformation (𝑑𝑓 ℎ𝑘𝑙) the lattice strain (𝜀𝑙𝑎𝑡𝑡𝑖𝑐𝑒 ) evolution through the experiment
was determined:
𝑑ℎ ℎ𝑘𝑙−𝑑𝑓 ℎ𝑘𝑙
𝜀𝑙𝑎𝑡𝑡𝑖𝑐𝑒 = (S3)
𝑑ℎ ℎ𝑘𝑙
For X-radiographic imaging, the incident slit opening was enlarged to 3 mm horizontally
by 4 mm vertically to ensure that the entire sample area was exposed to the X-rays. The X-ray
beam travels through the cell assembly to a scintillator made of single crystal Yttrium-
Aluminum-Garnet (YAG). The YAG scintillator will fluoresce, converting the transmitted X-ray
29
beam into visible light which is reflected off an angled mirror for collection by a Prosilica CCD
camera. The X-ray absorption contrast between the sample and the metal foils (referred to as
“strain markers”) placed at its interfaces allows the sample to be identified and the length to be
measured during the experiment.
Radiographic images were measured using the open-source software ImageJ (Abràmoff
et al., 2004). For sample strain we only need to compare changes in sample length, but to
calculate travel time we need to also determine the relationship between pixels in the images and
distance in real space. This was done by determining the length, in pixels, for the sample in a X-
radiographic image prior to pressurization and using the sample length as measured using a
Mitutoyo digimatic micrometer to three decimal places. The pixel to micron conversion allowed
us to determine the starting length of the sample in the first radiograph in each deformation
sequence. To measure strain, we compared each radiograph in the deformation sequence to the
radiograph taken nearest to the point of hydrostatic stress for each deformation sequence. This
was accomplished using a Python code to calculate the standard deviation of the intensity for
each row of pixels in the vertically oriented radiograph. The metal foils on either end of the
sample produce a parabolic decrease in the standard deviation moving from the sample interface
to the end of the metal foil. This is caused by the similarity in the grayscale value created by the
shadow of the anvils and the metal foil. We then determine the amount of sample strain from
how much the hydrostatic image standard deviation profile needs to be stretched in order to
minimize the difference between it and each subsequent image. We chose to take the standard
deviation of the intensity for each row of the full radiographs, as opposed to the average
intensity, because the standard deviation profile eliminates the effect of changes in the grayscale
value from one end of the image to the other.
30
The radiographs were not taken simultaneously with the diffraction or ultrasonic
measurements, therefore, sample strain as a function of time elapsed data were fit with a
polynomial function to allow for calculation of the sample strain at the time the diffraction and
ultrasonic measurements were taken. The sample length, in microns, during deformation (𝑙) at
the time of ultrasonic wave measurements was then calculated using the sample strain (ε)
determined from the polynomial fit and the sample length, in microns, prior to deformation (𝑙0 )
in each deformation sequence:
𝑙 = ((𝜀 ∗ 𝑙0 ) + 𝑙0 ) (S4)
A polynomial function was chosen to fit the sample strain vs. time data as opposed to a linear fit,
due to the observed sluggishness in the system at the start of deformation. Strain rates during
differential ram advancement, given in Table S2, were based on the latter portion of the sequence
where the sample strain vs. time data becomes linear; measured strain rates during the
differential ram retraction (Table S2) included all sample strain vs. time data after the differential
rams were reversed as these data showed no transient behavior.
Pressure determination during the experiment. To determine the pressure for each deformation
sequence, we used Ross Angel’s EOSfit 7.0 PVT calculator (Gonzalez et al., 2016) and the
olivine sample as the pressure standard. To calculate the pressure for each deformation sequence
we used the olivine diffraction data collected at detectors ψ= 0°, 180°, and 90°, at hydrostatic
conditions. We calculated the unit cell volume at each detector by solving the lattice parameters
that minimized the difference between the calculated d-spacing and the observed d-spacing for
each lattice reflection. We then determined the average unit cell volume for the three detectors
and determined the pressure needed to produce this unit cell volume using the third-order Birch-
31
Murnaghan isothermal Equation of State (EoS), thermal expansion coefficients from Fei (1995),
bulk modulus from Knittle (1995), reference unit cell volume from Zha et al. (1998), and the
measured temperature conditions.
EPSC Models. Elastic-plastic self-consistent (EPSC) modeling is a numerical modeling
technique that can be used to aid in the interpretation of synchrotron X-ray diffraction data.
Lattice strain versus sample strain curves for the deformation sequences are compared with
simulated diffraction data using a modified version of the EPSC3 model code provided by C.N
Tomé (Tomé and Oliver, 2002; Burnley 2015; Burnley and Kaboli, 2019). EPSC modeling is
derived from Eshelby’s theory of inclusions, in which the model considers the deformation of an
individual crystal located in a homogeneous elastic medium (HEM), which is essentially an
average of all crystals in the material (Turner and Tomé 1994; Tomé and Oliver 2002). The
model will calculate the stress and strain of each crystal and then move to the next crystal while
also updating the stress and strain of the HEM. The model iterates until a self-consistent result is
obtained. EPSC modeling accounts for all crystals within a sample, including “silent grains”,
which are crystals that do not satisfy the Bragg condition. This model also allows us to take into
consideration both elastic and plastic anisotropy.
The EPSC model inputs include boundary conditions, single crystal elastic constants, unit
cell dimensions, crystal orientation, slip systems, and the geometry of diffraction planes.
Temperature and pressure enter the model through the effect on the unit cell dimensions and
elastic constants and do not change in the simulations. Single crystal elastic constants for the
EPSC models were calculated for the appropriate pressure and temperature conditions using
single crystal elastic constants and derivatives for our sample material from Isaak (1992),
Anderson and Isaak (1995), and Abramson et al. (1997) and are listed in Table S3. The crystal
32
orientation input file from Burnley (2015) was used, which incorporates 49,108 grains
distributed through Euler space in five degree increments. We modeled the eight commonly
observed slip systems in olivine, three unidirectional slip systems to simulate kink band
formation (Burnley, 2015; Kaboli et al., 2017; Burnley and Kaboli, 2019), and an isotropic
deformation mechanism composed of 30 slip planes (Burnley and Kaboli, 2019) to allow a small
amount of anelastic deformation during the initial elastic portion of the experiment. Table S4
gives the parameters used to create the EPSC model fits for each deformation sequence of
San_381 and San_416.
EPSC models run during differential ram advancement include one mechanical process
input file with a uniaxial strain boundary condition and strain as the control variable. Strain in
the transverse directions and all stresses were allowed to vary freely. Models were run to
approximately 2.6- 4.6% strain corresponding with the sample strain value reported for the final
ultrasonic scan before the D-DIA rams were reversed and differential stress was released. The
number of displacement increments was determined so as to produce a 0.0001 strain increment
in the output file.
For deformation sequences in our experiments where measurements were also taken
during the reversal of the differential rams, we ran an additional mechanical process following
the mechanical process discussed above. During differential stress release, EPSC models were
run with a uniaxial stress boundary condition and stress was used as the control variable. Strain
in the transverse directions was allowed to vary freely. EPSC models were run to the
approximate macroscopic stress value corresponding with the final set of lattice strain values
reported. The number of displacement increments was determined so as to produce a 0.0001
strain increment in the output file.
33
DIASCoPE: Directly Integrated Acoustic System for Pressure Experiments.
In the experimental setup, a dual-mode 10° Y-cut LiNbO3 piezoelectric transducer is attached to
the bottom of the tungsten carbide anvil in the vertical column. This type of transducer can
transmit and receive frequencies from ~20-70 MHz, which allows us to measure both
longitudinal (P) and shear (S) waves at optimal frequencies during the experiment. For our
experiment, the frequency was set at 60 MHz for P-waves and 35 MHz for S-waves. The P- and
S-waves propagate along the axis of compression, with polarization parallel and perpendicular to
the compression axis, respectively. Figure S7 shows the communication of the various parts of
the DIASCoPE ultrasonic system. The Wavetek pulse generator sets the timing to send the sine
wave pulses and receive the measured travel time data. To visualize the signal, the pulse
generator must communicate with the digital oscilloscope (Ztec ZT4611-E). The Ztec
oscilloscope triggers the Arbitrary Function Generator to send a multi-cycle sine wave to the
transducer; for our experiments, the P-wave input signal was 5 cycles of a 60 MHz sine wave,
and the S-wave input signal was 5 cycles of a 35 MHz sine wave. The transducer sends the
sound wave into the cell assembly and returns the acoustic response to the oscilloscope. The Ztec
digital oscilloscope and the Arbitrary Function Generator are built directly into the EPICS
beamline control system. The accessibility of the DIASCoPE through the beamline control
system allows us to switch between the P- and S-wave frequencies in ~6 milliseconds for our
experiments providing nearly synchronous measurements. Fast measurements are possible
because the acquisitions are collected and summed in the oscilloscope before output of the final
summed waveform. In experiment San_381 the system was set to collect and sum 1024
acquisitions, however, due to a firmware error in the oscilloscope processing, the system
34
presented the 1024th acquisition as the sum. This issue was corrected in experiment San_416,
resulting in an improvement of the signal to noise ratio on the order of 10.
We determined the two-way travel time for P- and S-waves propagating in the sample
assembly through the Pulse-Echo Overlap (PEO) method using Plot85. As the propagating wave
intersects the various interfaces of the cell assembly materials, the wave is partially reflected, and
the remaining energy is transmitted beyond the interface. In the PEO method, the two wave
reflections of interest are overlaid and matched to determine the time offset. For our experiments,
the alumina piston-sample reflection (R2) was overlain on the sample-buffer rod reflection (R1)
(Figure 2). The sintered alumina piston above the San Carlos olivine in the cell assembly has a
lower acoustic impedance compared to the olivine, which results in a negative reflection factor
that causes a 180° phase shift. Therefore, when performing the overlap, we must invert the R2
reflection to properly determine the two-way travel time.
Acoustoelastic Constants
To use the theory of Hughes and Kelly (1953), we assume that the initial material is isotropic,
homogeneous, and unstressed. Although the properties of an individual olivine grain are known
to be anisotropic, at a larger scale, the polycrystalline material behavior is the sum of the grains
which are at a random orientation. Thus, we can consider the sample material isotropic before
deformation. Residual stresses should not be present in our sample given the annealing of the
sample before each deformation sequence; however, we cannot rule out the presence of low level
stress during the hydrostatic state due to the complexity of the stress state in a polycrystalline
material and the stress resolution of the D-DIA multi-anvil apparatus.
35
The equations 3a-3e derived by Hughes and Kelly (1953) can be linearized to the first-
order to produce a simple relationship for a uniaxial stress state in which ΔV/V is a linear
function of the macroscopic stress (equation 4). The velocities were calculated from travel time
and sample length measurements, and the macroscopic stress was obtained from EPSC models.
Acoustoelastic constants are specific to the elastic regime; therefore, the point at which the
sample begins to yield and plastically deform must be identified. This elastic limit was
determined from the output of the EPSC models from the strain increment in which the first
grains began to slip. The uncertainty of the acoustoelastic constant values was determined from
the uncertainty in stress and the uncertainty in velocity which originates from the uncertainty in
sample length and travel time. The uncertainty in velocity is ± 0.011 and ± 0.015 km/s for
San_416 and San_381 respectively. The uncertainty in stress (± 0.01-0.05 GPa) is determined
by comparing the stress at the elastic limit for various EPSC model that produced a similar fit to
the diffraction data.
Acoustoelastic Constant Comparison with Egle and Bray (1976). One method for calculating the
“acoustoelastic constants” by Egle and Bray (1976) in their evaluation of steel was to determine
the slope of the linear relation between ΔV/V and the axial strain. Since they have directly
calculated the “acoustoelastic constants” through linearization of the equations of Hughes and
Kelly (1953) we converted our acoustoelastic constants to also determine the slope of the linear
relation between ΔV/V and the axial strain. Figure S8 shows the acoustoelastic constants
(denoted as the relative wave velocity change (ΔV/V) as a function of strain) as a function of
pressure to compare with the values of Egle and Bray (1976) for steel. The Egle and Bray (1976)
data follows the trend we observe in our data where the acoustoelastic constant (A11) decreases
36
with increasing pressure, with positive values for A11. For the shear velocity data, the Egle and
Bray (1976) steel values are positive, but overall, the trend nears zero.
San Carlos Olivine Wave Velocity Comparison with Literature
We observe that the wave velocities at the point of hydrostatic stress increase with increasing
pressure, as seen when comparing the wave velocity for deformation sequences with the same
starting temperature (Figure S9). To evaluate the validity of our San Carlos olivine velocity
measurements, we report our values in Table S6 along with previously published results of the
literature. It is observed that our wave velocities lie within the range of previous experimentally
determined wave velocity values for San Carlos olivine.
37
Tables
Table 1. Comparison of the acoustoelastic constants for P- and S-wave velocities, A11 and A12 respectively, determined for experiment San_381 and
San_416. The acoustoelastic constants are separated into “Aij Advancement” which represents the acoustoelastic constant during differential ram
advancement, and “Aij Retraction” which represents the acoustoelastic constant during differential ram retraction. *systematic error of 50 °C.
Experiment Seq. Pressure (GPa) Temperature* (°C) Wave Velocity (km/s) Aij Advancement Aij Retraction
1 3.2 ± 0.3 451 ± 1.75 7.99 ± 0.014 1.073 ± 0.0236 -

San_381
2 4.0 ± 0.3 650 ± 2.36 7.89 ± 0.014 1.384 ± 0.0496 -

P-WAVE (A11)
3 4.4 ± 0.3 804 ± 3.50 7.72 ± 0.015 0.5660 ± 0.1091 1.043 ± 0.0581
4 4.9 ± 0.3 946 ± 3.42 7.74 ± 0.015 2.075 ± 0.1832 -
1 7.8 ± 0.3 449 ± 2.26 8.80 ± 0.011 0.7631 ± 0.0172 0.5978 ± 0.0051
San_416
2 9.8 ± 0.3 656 ± 2.87 8.75 ± 0.011 0.5203 ± 0.0230 0.5172 ± 0.0071
3 10.5 ± 0.3 898 ± 4.52 8.57 ± 0.011 0.7657 ± 0.0186 0.6179 ± 0.0189
1 3.2 ± 0.3 451 ± 1.75 4.52 ± 0.006 0.3847 ± 0.0236 -
San_381
2 4.0 ± 0.3 650 ± 2.36 4.46 ± 0.007 -0.3456 ± 0.0582 -

S-WAVE (A12)
3 4.4 ± 0.3 804 ± 3.50 4.35 ± 0.007 -0.1238 ± 0.0633 0.1553 ± 0.0454
4 4.9 ± 0.3 946 ± 3.42 4.35 ± 0.007 0.1281 ± 0.2667 -
1 7.8 ± 0.3 449 ± 2.26 4.90 ± 0.005 -0.1059 ± 0.0038 -0.2079 ± 0.0043
San_416
2 9.8 ± 0.3 656 ± 2.87 4.84 ± 0.005 -0.2057 ± 0.0081 -0.2566 ± 0.0051
3 10.5 ± 0.3 898 ± 4.52 4.75 ± 0.005 -0.1827 ± 0.0079 -0.2863 ± 0.0049
38
Table S1. Unit cell parameters used in EPSC models.
Experiment Seq. Pressure (GPa) Temperature (°C) a (Å) b (Å) c (Å) V (Å3)
1 3.2 451 4.7560 10.1592 5.9743 288.7
2 4.0 650 4.7528 10.1716 5.9822 289.2
San_381
3 4.4 804 4.7544 10.1917 5.9850 290.0
4 4.9 946 4.7557 10.2023 5.9870 290.5
1 7.8 449 4.7194 10.0269 5.9127 279.8
San_416 2 9.8 656 4.7353 9.9782 5.8905 278.3
3 10.5 898 4.7175 10.0315 5.9112 279.7
39
Table S2. Pressure, temperature, and strain rate information for each deformation sequence for San_381 and San_416 experiments. Strain rate
advancement indicates the strain rate during differential ram advancement. Strain rate retraction indicates the strain rate during differential ram
retraction back to “hydrostatic” conditions. aUncertainty in temperature is based on the observed temperature variation during experiment.
b
Uncertainty in pressure includes both uncertainty in measured d-spacings and temperature uncertainty.
Strain Rate Advancement Strain Rate Retraction

Experiment Seq. Temperature a (°C) Pressureb (GPa)
(x10-6 sec-1) (x10-6 sec-1)
1 451 ± 1.75 3.2 ± 0.3 3.4 x 10-6 -
San_381 2 650 ± 2.36 4.0 ± 0.3 3.6 x 10-6 -
3 804 ± 3.50 4.4 ± 0.3 3.6 x 10-6 1.0 x 10-6
4 946 ± 3.42 4.9 ± 0.3 2.4 x 10-6 -
1 449 ± 2.26 7.8 ± 0.3 3.1 x 10-6 0.9 x 10-6
San_416 2 656 ± 2.87 9.8 ± 0.3 2.6 x 10-6 1.0 x 10-6
3 898 ± 4.52 10.5 ± 0.3 3.8 x 10-6 0.8 x 10-6
40
Table S3. Single crystal elastic constants (Cij) used in EPSC models (values in GPa).
Experiment Seq. Temperature (°C) Pressure (GPa) C11 C22 C33 C44 C55 C66 C12 C13 C23
1 451 3.2 314.4 194.4 228.9 62.0 74.2 75.8 73.3 76.5 81.9
2 650 4.0 308.5 190.6 224.0 60.3 72.0 73.6 73.3 76.6 82.3
San_381
3 804 4.4 302.5 186.5 219.1 58.5 69.9 71.4 72.5 76.0 81.8
4 946 4.9 297.8 183.4 215.2 57.2 68.2 69.7 72.2 75.7 81.8
1 449 7.8 343.7 218.5 253.5 69.5 78.9 84.5 90.5 92.5 97.0
San_416 2 656 9.8 344.5 220.3 254.4 69.5 76.6 84.2 94.7 96.4 101.0
3 898 10.5 335.0 213.7 246.5 66.7 72.8 80.7 93.4 95.3 100.2
41
Table S4. Summary of slip systems used in each EPSC model to fit the experimental diffraction data for San_381 and San_416. Kink system defines the
required slip systems used to simulate kink band formation: [𝟐 ̅𝟏𝟎] 𝐨𝐧 (120), [21𝟎] 𝐨𝐧 (𝟏𝟐 ̅𝟎𝟒] 𝐨𝐧 (405), [𝟓
̅𝟎), [𝟓 ̅𝟎𝟒 ̅). τ is the critically resolved
̅](40𝟓
shear stress and τ0, φ0, and φ1 are the hardening parameters in (GPa). The macroscopic stress (σ ) on the sample (in GPa) at 1.5% is shown.
a
Experiment Seq Slip Systems τ τ0 φ0 φ1 σa

Isotropic System 0.86 85 85 85
1.712
1 Group A: [001](100), [001]{110}, [001](010) 0.53 0.001 0.01 0.01
Group B: Kink system, [100]{011} 0.72 0.001 0.01 0.01
2 Group A: [001](100), [001]{110}, [001](010) 0.40 0.001 0.01 0.01 1.494
San_381
3 Group A: [001](100), [001]{110}, [100](001) 0.39 0.001 0.01 0.01 1.225
Group B: Kink system, [100]{011}, [001](010) 0.57 0.001 0.01 0.01

4 Group A: [001](100), [001]{110} 0.13 0.001 0.01 0.01 0.5528
Group B: Kink system, [100]{011}, [001](010) 0.21 0.001 0.01 0.01
Group A: [001](100),[001]{110},[001](010) 0.76 0.001 0.01 0.01 2.264

1
Group A: [001](100),[001]{110},[001](010) 0.56 0.001 0.01 0.01 1.833

San_416 2

3 Group A: [001]{110},[001](100) 0.56 0.001 0.01 0.01 1.484
42
Table S5. Second and third order elastic constants and acoustoelastic constants for various materials in literature.
Literature Material λ (GPa) μ (GPa) l (GPa) m (GPa) n (GPa) A11† A12†

San381 Seq 1 Polycrystalline San Carlos Olivine - - - - - 1.0730 0.3847
San381 Seq 2 Polycrystalline San Carlos Olivine - - - - - 1.3840 -0.3456
This Work

San381 Seq 4 Polycrystalline San Carlos Olivine - - - - - 1.9660 0.8343
Hughes and Kelly
Armco Iron 110 82 -348 -1030 -1100 2.6902 0.7787
(1953)
Egle and Bray
Rail Steel-1 115.8 79.9 -248 -623 -714 1.2134 0.1220
(1976)
Egle and Bray
Rail Steel-4 110.7 82.4 -302 -616 -724 1.2049 0.1060
(1976)
Ankay and Zhang
Steel 105.88 79.01 -332.15 -584.39 -717.75 1.2701 0.1135
(2019)
Metallic Materials
Crecraft (1962) REX 535 Nickel-steel 90.9 78 -46 -590 -730 1.2448 0.1900
Smith et al.
Rex 535 Nickel steel 109 81.8 -327.5 -578 -676 1.1268 0.0498
(1966)
Smith et al.
Steel-Hecla 37 (0.4% C) 111 82.1 -461 -636 -708 1.3956 0.1260
(1966)
Smith et al.
Steel-Hecla 17 (0.6% C) 110.5 82 -328 -595 -668 1.1653 0.0616
(1966)
Smith et al.
Steel-Hecla 138A 109 81.9 -426.5 -619 -708 1.3407 0.1132
(1966)
Smith et al.
Steel-Hecla ATV austenitic 87 71.6 -535 -752 -400 2.8422 0.4131
(1966)
Crecraft (1967) Ni-Steel S/NTV 109 81.7 -56 -671 -785 1.2513 0.2122
Johnson et al
Steel bolt - - - - - 0.0092 0.0004
(1986)+
43
Chaki and Bourse
Steel strand intertwined - - - - - 0.0048 -
(2009)+
Smith et al.
Alumium Alloy-2S 57 27.6 -311 -401 -408 7.7557 2.1884
(1966)
Smith et al.
Alumium Alloy-B53S M 58 26 -223.5 -237 -276 3.7454 0.4674
(1966)
Smith et al.
Alumium Alloy-B53S P 61.9 26.2 -201.5 -305 -300 5.0154 1.0403
(1966)
Smith et al.
Alumium Alloy-D54S 49.1 26 -387.5 -358 -320 8.7941 2.0136
(1966)
Smith et al.
Alumium Alloy-JH77S 57.5 26.8 -337 -395 -436 8.0016 2.4067
(1966)
Nogueria (2017) Aluminum 61.5671 26.9612 -449.8 -384.2 -229.8 7.6025 1.2534
Stobbe (2005) Aluminum 7075-T651 54.9 26.5 -252.2 -325 -351.2 6.2882 1.5713
Crecraft (1967) Aluminum (99%) 61 24.9 -47 -342 -248 5.9152 1.2963
Crecraft (1967) Copper (99.9%) 104 46 542 -372 -401 0.5100 -0.0036
Smith et al.
Magnesium Tooling Plate 25.9 16.6 -90.1 -141.6 -168.4 6.9740 1.0127
(1966)
Smith et al.
Molybdenum-Sintered 157 110 -308.5 -669 -772 0.5299 -0.0813
(1966)
Smith et al.
Molybdenum-Resintered 178 124 -301 -852 -908 0.5959 -0.0155
(1966)
Smith et al.
Tungsten-Sintered 75 73 -250.5 -391 -496 0.9200 -0.1359
(1966)
Smith et al.
Tungsten-Resintered 163 137 -472.5 -792 -1068 0.4936 -0.0432
(1966)
Rasolofosoan and
Quartz 38 48 -98 -89 -165 -0.5979 -1.1029
Yin (1996)
Rasolofosoan and
Rocks
Calcite 76 37 -77 -136 -141 -0.3521 -1.0900

Yin (1996)
Nur and Simmons
Barre Granite 13.8 18.2 -3371 -6742 -6600 687.8323 277.7303
(1969)*
Zamora (1990)* Dry F32 Foutainbleau Sandstone 15.3 11.7 -97800 -99400 -84900 20744.7664 7834.6203
44
Zamora (1990)* Dry F32, Thermally Cracked 7.8 5.7 -74000 -64500 -34900 56785.4958 19254.2533
Zamora (1990)* Wet F32, Thermally Cracked 30.2 17.5 -59100 -38200 -27500 3259.5241 1134.5197
Zamora (1990)* Dry D82 Marble 30 21.3 -40300 -35400 -20300 2185.5983 749.5926
Winkler and Liu
Berea SS-A 4.1 4.5 -1636 -3441.5 -451 5278.8075 1830.8961
(1996)
Winkler and Liu
Berea SS-B 3.7 3.7 -3470 -5541.5 -6233 12457.9255 5164.4631
(1996)
Winkler and Liu
Buff SS 9.9 9.5 -1204 -3489.5 -4215 1088.0448 487.4115
(1996)
Winkler and Liu
Hanson SS 10.4 10 -3425 -3806.5 -3371 1208.2319 449.3516
(1996)
Winkler and Liu
Limestone 1078 21 12.2 -3322 -1349.5 -1207 261.4952 81.4095
(1996)
Winkler and Liu
Limestone 1083 29.6 20.6 -8302 -11299 -9730 684.0596 272.4691
(1996)
Winkler and Liu
Massilon SS 6.1 6.3 -7900 -14435 -17530 11185.7171 4784.7057
(1996)
Winkler and Liu
Portland SS 9.7 7.3 -759 -980 -1122 488.5871 201.3603
(1996)
Winkler and Liu
Westerly Granite 29.9 23.6 -21377 -27262.5 -14071 1369.6308 487.5715
(1996)
Rasolofosoan and
Palatinat sandstone 7.3 6.3 -2530 -580 1140 718.2229 44.9418
Yin (1996)
Rasolofosoan and
Magnesian Marble 15 14 -25150 -7090 13590 1730.1120 163.1028
Yin (1996)
Rasolofosoan and
Colorado oil shale 21 18 -66600 7230 24420 619.7679 -407.0847
Yin (1996)
Bompan and
Concrete - - - - - 405.3 176.7
Haach (2018)+
Concrete
Lillamand et al.
Concrete - - - - - 1.3 5.0
(2010)+
Ankay and Zhang
High Performance Concrete 9.05 14.7 -613.36 -672.82 -457.57 121.1821 38.9906
(2019)
Ankay and Zhang Ultra High Performance Concrete 11.61 19.76 -60.8 -76.5 -104.5 2.7404 -1.1445
45
(2019)
Hughes and Kelly
Polystyrene 2.889 1.381 -18.9 -13.3 -10 93.6534 4.5931
(1953)
Hughes and Kelly
Pyrex Glass 13.53 27.5 14 92 420 -8.2558 -5.8542
Misc.
(1953)
Winkler and Liu
Beads 8 5.5 -716 -678 46 591.5174 157.7332
(1996)
Winkler and Liu
Lucite 5.6 2.3 -23 -21.5 -23 41.1976 3.9618
(1996)
†
Values calculated from the second and third order elastic constants (equations 7a-7b).
*
Values from Johnson and Rasolofosoan (1996).
+
A11 and A12 values were taken directly from literature.
46
Table S6. Summary of San Carlos Olivine P- and S-wave velocity measurement from the literature.
Reference Pressure (GPa) Temperature (K) Vp (km/s)† Vs (km/s)†

San_416 Seq1- This study 7.8 722.5 8.799 ± 0.011 4.905 ± 0.005
San_381 Seq2- This study 4 923.4 7.885 ± 0.014 4.461 ± 0.007
Zha et al. (1998)* Fo90Fa10 Ambient Ambient 8.39 4.84
Zha et al. (1998) Fo90Fa10 2.5 Ambient 8.55 ± 0.06 4.89 ± 0.04
Zha et al. (1998) Fo90Fa10 5 Ambient 8.79 ± 0.06 4.97 ± 0.04
Darling et al. (2004) Fo91.1Fa8.9 8.6 1273 8.31 4.81
Darling et al. (2004) Fo90.1Fa9.9 7 1223 8.36 4.86
Darling et al. (2004) 8.6 1272 8.38 4.86
Isaak (1992) Fo90.3Fa9.5 (Mn2SiO4)0.2 Ambient 300 8.346 ± 0.011 4.827 ± 0.009
Isaak (1992) Fo90.3Fa9.5 (Mn2SiO4)0.2 Ambient 500 8.238 4.752
Webb (1989)* Fo90.5Fa9.5 3 Ambient 8.36 4.82
Abramson et al. (1997)* 0 300 8.34 4.82
Mao et al. 2015† Fo90Fa10 0 300 8.35 4.82
Mao et al. 2015† Fo90Fa10 3.1 300 8.62 4.93
Mao et al. 2015† Fo90Fa10 5.3 300 8.79 4.98
Mao et al. 2015† Fo90Fa10 7.8 300 8.95 5.02
Mao et al. 2015† Fo90Fa10 9.9 300 9.08 5.06
47
Mao et al. 2015† Fo90Fa10 12.0 300 9.19 5.07
Mao et al. 2015† Fo90Fa10 14.0 300 9.29 5.10
Mao et al. 2015† Fo90Fa10 16.6 300 9.42 5.11
Mao et al. 2015† Fo90Fa10 19.2 300 9.51 5.12
Mao et al. 2015† Fo90Fa10 2.7 500 8.47 4.85
Mao et al. 2015† Fo90Fa10 6.8 500 8.78 4.97
Mao et al. 2015† Fo90Fa10 8.6 500 8.92 5.00
Mao et al. 2015† Fo90Fa10 10.9 500 9.04 5.03
Mao et al. 2015† Fo90Fa10 14.1 500 9.14 5.03
Mao et al. 2015† Fo90Fa10 4.0 750 8.47 4.83
Mao et al. 2015† Fo90Fa10 5.5 750 8.59 4.87
Mao et al. 2015† Fo90Fa10 9.0 750 8.86 4.96
Mao et al. 2015† Fo90Fa10 10.4 750 8.90 4.97
Mao et al. 2015† Fo90Fa10 12.5 750 8.97 4.98
Mao et al. 2015† Fo90Fa10 4.5 900 8.43 4.81
Mao et al. 2015† Fo90Fa10 8.4 900 8.71 4.90
Mao et al. 2015† Fo90Fa10 10.1 900 8.78 4.93
Mao et al. 2015† Fo90Fa10 13.3 900 8.96 4.97
Liu et al. 2005† Fo90Fa10 1.8 298 8.52 4.87
Liu et al. 2005† Fo90Fa10 2.8 298 8.60 4.91
Liu et al. 2005† Fo90Fa10 3.6 298 8.66 4.93
Liu et al. 2005† Fo90Fa10 4.7 298 8.74 4.94
Liu et al. 2005† Fo90Fa10 6.5 298 8.90 5.00
Liu et al. 2005† Fo90Fa10 2.2 473 8.45 4.80
Liu et al. 2005† Fo90Fa10 3.2 473 8.55 4.85
Liu et al. 2005† Fo90Fa10 4.0 473 8.62 4.88
Liu et al. 2005† Fo90Fa10 5.1 473 8.69 4.90
Liu et al. 2005† Fo90Fa10 6.8 473 8.85 4.96
48
Liu et al. 2005† Fo90Fa10 2.7 673 8.39 4.76
Liu et al. 2005† Fo90Fa10 3.6 673 8.49 4.80
Liu et al. 2005† Fo90Fa10 4.3 673 8.56 4.82
Liu et al. 2005† Fo90Fa10 5.4 673 8.64 4.85
Liu et al. 2005† Fo90Fa10 7.3 673 8.79 4.91
Liu et al. 2005† Fo90Fa10 4.0 873 8.44 4.74
Liu et al. 2005† Fo90Fa10 4.8 873 8.51 4.78
Liu et al. 2005† Fo90Fa10 5.9 873 8.58 4.80
Liu et al. 2005† Fo90Fa10 7.7 873 8.75 4.85
Liu et al. 2005† Fo90Fa10 5.4 1073 8.45 4.72
Liu et al. 2005† Fo90Fa10 6.4 1073 8.53 4.75
Liu et al. 2005† Fo90Fa10 6.7 1073 8.55 4.76
Liu et al. 2005† Fo90Fa10 8.2 1073 8.68 4.80
Zhang and Bass (2016) Fo90.5Fa9.5 10-5 300 8.33 ± 0.03 4.80 ± 0.03
Zhang and Bass (2016) Fo90.5Fa9.5 3.8 ± 0.1 300 8.65 ± 0.04 4.91 ± 0.03
Zhang and Bass (2016) Fo90.5Fa9.5 10-5 1300 7.89 ± 0.05 4.51 ± 0.04
†
The velocity values are extracted from figures in the original publications.
*Calculated from Bulk and Shear Moduli and density
49
Figures
Figure 1. A schematic illustrating the various ultrasonic waves which propagate from a single transducer
placed along the X1 axis of compression (left) and perpendicular to the X1 axis of compression (right)
(Inspired by Bompan and Haach, 2018).
50
Figure 2. (2a)-A transducer attached to the bottom anvil transmits an ultrasonic wave pulse (optimal for P- or
S-wave) and the reflection from the contact between different materials is recorded. (R0) is the reflection off
of the anvil-buffer rod interface, (R1) buffer rod-sample interface, (R2) sample-above sintered Al2O3 piston
interface, (R3) sintered Al2O3 piston-crushable Al2O3 piston. (2b)- The observed acoustic signal recorded for
file SAN_416_0033.udat in experiment San_416 for a 35 MHz wave pulse that generated S-waves. The
amplitude of the reflections from various material interfaces is shown as a function of time. (2c)- The pulse-
echo-overlap (PEO) method is used to determine the travel time in the sample by overlaying the (R2)
reflection on the (R1) reflection and measuring the time offset (Inspired by Whitaker et al., 2017).
51
Figure 3. Comparison of acoustoelastic slope shown as the relative wave velocity change (∆𝑽𝒊𝒋 ⁄𝑽𝟎𝒊𝒋 ) as a
function of the macroscopic stress (in GPa) for the first deformation sequence in San_381 (P=3.2 GPa) and
San_416 (P=7.8 GPa) at ~450 °C (3a, 3c), and second deformation sequence in San_381 (P=4.0 GPa) and
San_416 (P= 9.8 GPa) at ~650 °C (3b, 3d). Figures 3a, 3b show the P-wave (V11) data and figures 3c, 3d show
the S-wave (V12) data.
52
Figure 4. Comparison of acoustoelastic slope during differential ram advancement to the elastic limit and
during differential ram retraction for San_416 deformation sequence 1 (4a), 2 (4b), and 3 (4c). The shaded in
squares include data points used for calculating the A11 “advancement” acoustoelastic constant, and the
shaded in triangles include data points used for calculating the A12 “advancement” acoustoelastic constant.
The data used to determine the A12 “retraction” acoustoelastic constant for P- and S-wave data is indicated
by a trendline through the bolded data points. Pressure and temperature are denoted for each sequence.
53
Figure 5. Acoustoelastic constants as a function of pressure (5a, 5b) and temperature (5c, 5d) for San_381 and
San_416. The temperature and pressure conditions of each deformation sequence is listed next to the
corresponding data point. A11 and A12 are the acoustoelastic constants for the P- and S-wave, respectively.
Circle symbols denote San_381 data and square symbols denote San_416 data. Solid data points indicate Aij
advancement and bordered data points indicate Aij retraction.
54
Figure S1. A schematic cross-sectional view of the D-DIA apparatus (Wang et al., 2003).
55
Figure S2. Schematic sample assembly for San_381 (S2a) and San_416 (S2b).
56
Figure S3. Thermocouple temperature vs. Power calibration for San_381 used to extrapolate the temperature
after the thermocouple failed during deformation sequence 2. This calibration was also used to determine the
temperature in each deformation sequence for San_416.
57
Figure S4. A simplified schematic illustrating the propagation of the X-ray beam through the sample
assembly during an experiment, and the means of diffraction data collection from the array of 10 energy-
dispersive detectors (Modified from Burnley and Zhang, 2008).
58
Figure S5. Lattice strain vs. sample strain data (symbols) for the 4 deformation sequences of San_381.
Pressure and temperature conditions are labeled on each graph. The solid lines represent the EPSC model
simulations for each lattice reflection to closely match the corresponding diffraction data. Compression is (+)
with the average (AVE) lattice strain determined from detectors at ψ =0° and 180° in the positive quadrant.
The black star denotes the elastic limit. Data beyond the elastic limit is not used in calculation of the
acoustoelastic constants. The uncertainty in lattice strain is ±0.001, as shown by the error bar placed next to
each deformation sequence.
59
Figure S6. Lattice strain vs. sample strain data (symbols) for the 3 deformation sequences of San_416.
Pressure and temperature conditions are labeled on each graph. The solid lines represent the EPSC model
simulations for each lattice reflection to closely match the corresponding diffraction data. Compression is (+)
with the average (AVE) lattice strain determined from detectors at ψ =0° and 180° in the positive quadrant.
The black star denotes the elastic limit. Data beyond the elastic limit is not used in calculation of the
acoustoelastic constants. The uncertainty in lattice strain is ±0.001, as shown by the error bar placed next to
each deformation sequence.
60
Figure S7. The communication path for DIASCoPE ultrasonic interferometry measurements with arrows
indicating direction of communication flow (Modified from Whitaker et al., 2017).
61
Figure S8. Comparison of A11 (S8a) and A12 (S8b) as a function of pressure for San_381, San_416, and steel
values of Egle and Bray (1976). The acoustoelastic constants are defined here as the relative wave velocity
change (ΔV/V) as a function of strain to facilitate comparison with the values of Egle and Bray (1976).
62
Figure S9. Comparison of P-wave velocity (S9a) and S-wave velocity (S9b) measurements from the literature
for San Carlos olivine with the P- and S-wave velocities of this study.
63
References
Abràmoff, M.D., Paulo J.M., and Sunanda J.R., 2004, Image processing with ImageJ:
Biophotonics international, v.11, p. 36-42.
Abramson, E.H., Brown, J.M., Slutsky, L.J. and Zaug, J., 1997, The elastic constants of San
Carlos olivine to 17 GPa: Journal of Geophysical Research: Solid Earth, v. 102, p.12253-
12263.
Anderson, D.L., 2007, Chapter 18: Elasticity and solid-state Geophysics, in New theory of the
earth, Cambridge, Cambridge University Press, p. 233–245.
Ankay, B., and Zhang, C.. 2019, Acoustoelastic evaluation of ultra-high performance concretes:
AIP Conference Proceedings, v.2102, p. 110002.
Bateman, T., Mason, W.P. and McSkimin, H.J., 1961, Third‐order elastic moduli of germanium:
Journal of Applied Physics, v.32, pp.928-936.
Benson, R.W. and Raelson, V.J., 1959, From ultrasonics to a new stress-analysis
technique…Acoustoelasticity, Product Engineering, v.30, pp.56-59.
Bergman, R.H., and Shahbender, R.A., 1958, Effect of Statically Applied Stresses on the
Velocity of Propagation of Ultrasonic Waves: Journal of Applied Physics, v. 29, p. 1736–
1738, doi: 10.1063/1.1723035.
Birch, F., 1966, Compressibility; Elastic Constants, in Handbook of Physical Constants, New
York, Geological Society of America, p. 97–174, doi: 10.1130/mem97-p97.
Bompan, K.F., and Haach, V.G., 2018, Ultrasonic tests in the evaluation of the stress level in
concrete prisms based on the acoustoelasticity: Construction and Building Materials, v.
162, p. 740–750, doi: 10.1016/j.conbuildmat.2017.11.153.
Burnley, P.C., and Zhang, D., 2008, Interpreting in-situ X-ray diffraction data from high pressure
deformation experiments using elastic–plastic self-consistent models: an example using
quartz: Journal of Physics: Condensed Matter, v. 20, p. 285201, doi: 10.1088/0953-
8984/20/28/285201.
Burnley, P.C., 2015, Elastic plastic self-consistent (EPSC) modeling of plastic deformation in
fayalite olivine: American Mineralogist, v. 100, p. 1424–1433, doi:10.2138/am-2015-
5234ccbyncnd.
Burnley, P.C., and Kaboli, S., 2019, Elastic plastic self-consistent (EPSC) modeling of San
Carlos olivine deformed in a D-DIA apparatus: American Mineralogist, v. 104, p. 276–
281, doi: 10.2138/am-2019-6666.
Chaki, S., Corneloup, G., Lillamand, I., and Walaszek, H., 2007, Combination of Longitudinal
and Transverse Ultrasonic Waves for In Situ Control of the Tightening of Bolts: Journal
of Pressure Vessel Technology, v. 129, p. 383–390, doi: 10.1115/1.2748821.
Chaki, S., and Bourse, G., 2009, Stress Level Measurement in Prestressed Steel Strands Using
Acoustoelastic Effect: Experimental Mechanics, v. 49, p. 673–681, doi: 10.1007/s11340-
008-9174-9.
Crecraft, D.I., 1962, Ultrasonic Wave Velocities in Stressed Nickel Steel: Nature, v. 195, p.
1193–1194, doi: 10.1038/1951193a0.
Crecraft, D.I., 1967, The measurement of applied and residual stresses in metals using ultrasonic
waves: Journal of Sound and Vibration, v. 5, p. 173–192, doi: 10.1016/0022-
460X(67)90186-1.
Darling, K.L., Gwanmesia, G.D., Kung, J., Li, B. and Liebermann, R.C., 2004, Ultrasonic
measurements of the sound velocities in polycrystalline San Carlos olivine in multi-anvil,
high-pressure apparatus: Physics of the Earth and Planetary Interiors, v. 143, p.19-31.
64
Dey, S., Roy, N. and Dutta, A., 1984. P and S waves in a medium under initial stresses and under
gravity. Indian Journal of Pure and Applied Mathematics, v. 15, p.795-808.
Durham, W.B., Weidner, D.J., Karato, S., and Wang, Y., 2002, New Developments in
Deformation Experiments at High Pressure: Reviews in Mineralogy and Geochemistry, v.
51, p. 21–49, doi: 10.2138/gsrmg.51.1.21.
Egle, D. M., and Bray, D. E., 1976, Measurement of acoustoelastic and third‐order elastic
constants for rail steel: The journal of the Acoustical Society of America, v. 60, p. 741-
744, doi: 10.1121/1.381146
Fei, Y., 1995, Thermal expansion: Mineral physics and crystallography: a handbook of physical
constants, v. 2, p.29-44, doi: 10.1029/RF002p0029.
Gonzalez-Platas, J., Alvaro, M., Nestola, F., and Angel, R., 2016, EosFit7-GUI: a new graphical
user interface for equation of state calculations, analyses and teaching: Journal of Applied
Crystallography, v. 49, p. 1377–1382, doi: 10.1107/s1600576716008050.
Hilairet, N., Wang, Y., Sanehira, T., Merkel, S., and Mei, S., 2012, Deformation of olivine under
mantle conditions: An in situ high-pressure, high-temperature study using monochromatic
synchrotron radiation: Journal of Geophysical Research: Solid Earth, v. 117, doi:
10.1029/2011jb008498.
Hughes, D.S., and Kelly, J.L., 1953, Second-Order Elastic Deformation of Solids: Physical
Review, v. 92, p. 1145–1149, doi: 10.1103/physrev.92.1145.
Isaak, D.G., 1992, High‐temperature elasticity of iron‐bearing olivines: Journal of Geophysical
Research: Solid Earth, v. 97. p.1871-1885.
Johnson, G.C., 1981, Acoustoelastic theory for elastic–plastic materials: The Journal of the
Acoustical Society of America, v. 70, p. 591–595, doi: 10.1121/1.386748.
Johnson, G.C., Holt, A.C., and Cunningham, B., 1986, An Ultrasonic Method for Determining
Axial Stress in Bolts: Journal of Testing and Evaluation, v. 14, p. 253–259, doi:
10.1520/jte10337j.
Johnson, P.A., and Rasolofosaon, P.N.J., 1996, Nonlinear elasticity and stress-induced
anisotropy in rock: Journal of Geophysical Research: Solid Earth, v. 101, p. 3113–3124,
doi: 10.1029/95jb02880.
Kaboli, S., Burnley, P.C., Xia, G., and Green, H.W., 2017, Pressure Dependence of Creep in
Forsterite Olivine: Comparison of Measurements From the D‐DIA and Griggs Apparatus:
Geophysical Research Letters, v. 44, doi: 10.1002/2017gl075177.
Knittle, E., 1995, Static Compression Measurements of Equations of State: Mineral physics and
crystallography: a handbook of physical constants, v. 2, p. 98–142, doi:
10.1029/rf002p0098.
Lott, M., Remillieux, M. C., Garnier, V., Le Bas, P. Y., Ulrich, T. J., & Payan, C., 2017,
Nonlinear elasticity in rocks: A comprehensive three-dimensional description: Physical
Review Materials, v. 1, p. 023603, doi:10.1103/PhysRevMaterials.1.023603.
Li, B., 2004, Modern techniques in measuring elasticity of Earth materials at high pressure and
high temperature using ultrasonic interferometry in conjunction with synchrotron X-
radiation in multi-anvil apparatus: Physics of The Earth and Planetary Interiors, v. 143-
144, p. 559–574, doi: 10.1016/s0031-9201(04)00088-3.
Li, L., Weidner, D., Raterron, P., Chen, J., Vaughan, M., Mei, S., and Durham, B., 2006,
Deformation of olivine at mantle pressure using the D-DIA: European Journal of
Mineralogy, v. 18, p. 7–19, doi: 10.1127/0935-1221/2006/0018-0007.
Lillamand, I., Chaix, J.-F., Ploix, M.-A., and Garnier, V., 2010, Acoustoelastic effect in concrete
65
material under uni-axial compressive loading: NDT & E International, v. 43, p. 655–660,
doi: 10.1016/j.ndteint.2010.07.001.
Liu, J. X., Cui, Z. W., & Wang, K. X., 2007, Reflection and transmission of acoustic waves at
the interface between rocks in the presence of elastic–plastic deformations: Journal of
Geophysics and Engineering, v.4, p. 232, doi: 10.1088/1742-2132/4/2/012.
Liu, W., Kung, J. and Li, B., 2005, Elasticity of San Carlos olivine to 8 GPa and 1073 K:
Geophysical Research Letters, v. 32, p. L16301, doi:10.1029/2005GL023453.
Lucet, N., 1989, Velocity and attenuation of sonic and ultrasonic waves in rocks under confining
pressure (in French), Ph.D. thesis, Univ. of Paris VI, Paris.
Man, C.S. and Lu, W.Y., 1987, Towards an acoustoelastic theory for measurement of residual
stress: Journal of Elasticity, v.17, p. 159-182.
Mao, Z., Fan, D., Lin, J.F., Yang, J., Tkachev, S.N., Zhuravlev, K. and Prakapenka, V.B., 2015,
Elasticity of single-crystal olivine at high pressures and temperatures: Earth and
Planetary Science Letters, v. 426, p.204-215.
Murnaghan, F.D., 1951, Finite deformation of an elastic solid: Wiley.
Nogueira, C.L., 2017, Ultrasonic Evaluation of Acoustoelastic Parameters in Aluminum: Journal
of Materials in Civil Engineering, v. 29, p. 04017158, doi: 10.1061/(ASCE)MT.1943-
5533.0002009.
Nur, A., and Simmons, G., 1969, Stress-induced velocity anisotropy in rock: Journal of
Geophysical Research, v. 74, p. 6667-6674.
Raterron, P., Merkel, S. and Holyoke III, C.W., 2013, Axial temperature gradient and stress
measurements in the deformation-DIA cell using alumina pistons: Review of Scientific
Instruments, v. 84, p. 043906.
Rasolofosaon, P.N.J., and Yin, H., 1996, Simultaneous Characterization of Anisotropy and
Nonlinearity in Arbitrary Elastic Media - Reflections on Experimental Data: Seismic
Anisotropy, p. 141–179, doi: 10.1190/1.9781560802693.ch6.
Rollins Jr, F.R., Kobett, D.R. and Jones, J.L., 1963, Study of Ultrasonic Methods for
Nondestructive Measurement of Residual Stress, Part 2: Technical Documentary Report
No. WADD-TR-61-42, part 2.
Singh, A.K., Balasingh, C., Mao, H.-K., Hemley, R.J., and Shu, J., 1998, Analysis of lattice
strains measured under nonhydrostatic pressure: Journal of Applied Physics, v. 83, p.
7567–7575, doi: 10.1063/1.367872.
Smith, R., 1963, Stress-induced anisotropy in solids—the acousto-elastic effect: Ultrasonics, v.
1, p. 135–147, doi: 10.1016/0041-624x(63)90003-9.
Smith, R.T., Stern, R., and Stephens, R.W.B., 1966, Third‐Order Elastic Moduli of
Polycrystalline Metals from Ultrasonic Velocity Measurements: The Journal of the
Acoustical Society of America, v. 40, p. 1002–1008, doi: 10.1121/1.1910179.
Su, C., Liu, Y., Song, W., Fan, D., Wang, Z., and Tang, H., 2018, Thermodynamic properties of
San Carlos olivine at high temperature and high pressure: Acta Geochimica, v. 37, p.171-
179, doi: 10.1007/s11631-018-0261-z.
Thurston, R.N., and Brugger, K., 1964, Third-Order Elastic Constants and the Velocity of Small
Amplitude Elastic Waves in Homogeneously Stressed Media: Physical Review, v. 135,
doi: 10.1103/physrev.135.ab3.2.
Tome, C. N., and E. C. Oliver., 2002, Code Elasto-Plastic Self-Consistent (EPSC): Los Alamos
National Laboratory, USA.
66
Toupin, R.A., and Bernstein, B., 1961, Sound Waves in Deformed Perfectly Elastic Materials.
Acoustoelastic Effect: The Journal of the Acoustical Society of America, v. 33, p. 216–
225, doi: 10.1121/1.1908623.
Turner, P., and Tomé, C., 1994, A study of residual stresses in Zircaloy-2 with rod texture: Acta
Metallurgica et Materialia, v. 42, p. 4143–4153, doi: 10.1016/0956-7151(94)90191-0.
Wang, Y., Durham, W.B., Getting, I.C., and Weidner, D.J., 2003, The deformation-DIA: A new
apparatus for high temperature triaxial deformation to pressures up to 15 GPa: Review of
Scientific Instruments, v. 74, p. 3002–3011, doi: 10.1063/1.1570948.
Webb, S.L., 1989, The elasticity of the upper mantle orthosilicates olivine and garnet to 3 GPa:
Physics and Chemistry of Minerals, v. 16, p.684-692.
Weidner, D.J., Vaughan, M.T., Wang, L., Long, H., Li, L., Dixon, N.A., and Durham, W.B.,
2010, Precise stress measurements with white synchrotron x rays: Review of Scientific
Instruments, v. 81, p. 013903, doi: 10.1063/1.3263760.
Whitaker, M.L., Baldwin, K.J., and Huebsch, W.R., 2017, DIASCoPE: Directly integrated
acoustic system combined with pressure experiments—A new method for fast acoustic
velocity measurements at high pressure: Review of Scientific Instruments, v. 88, p.
034901, doi: 10.1063/1.4977596.
Winkler, K. W., and Liu, X., 1996, Measurements of third‐order elastic constants in rocks: The
Journal of the Acoustical Society of America, v. 100, p. 1392-1398, doi:
10.1121/1.415986.
Xie, F., Ren, Y., Zhou, Y., Larose, E., & Baillet, L., 2018, Monitoring local changes in granite
rock under biaxial test: A spatiotemporal imaging application with diffuse waves: Journal
of Geophysical Research: Solid Earth, v. 123, p. 2214-2227, doi:
10.1002/2017JB014940.
Yan, X., Dong, S., Xu, B., and Cao, Y., 2018, Progresses and Challenges of Ultrasonic Testing
for Stress in Remanufacturing Laser Cladding Coating: Materials, doi:
10.20944/preprints201801.0071.v1.
Zamora, M., 1990, Experimental study of the effect of the geometry of rock porosity on the
velocities of elastic waves (in French), Docteur bs Sciences thesis, Univ. of Paris VII,
Paris, 1990.
Zha, C.S., Duffy, T.S., Downs, R.T., Mao, H.-K., and Hemley, R.J., 1998, Brillouin scattering
and X-ray diffraction of San Carlos olivine: direct pressure determination to 32 GPa:
Earth and Planetary Science Letters, v. 159, p. 25–33, doi: 10.1016/s0012-
821x(98)00063-6.
Zhang, J.S. and Bass, J.D., 2016, Sound velocities of olivine at high pressures and temperatures
and the composition of Earth's upper mantle: Geophysical Research Letters, v. 43,
p.9611-9618.
67
Curriculum Vitae
Taryn Traylor
Email: [email protected]
EDUCATION
• B.S. Geology con. Petroleum Geology, Texas A&M University, College Station, TX,
2014-2017.
RESEARCH EXPERIENCE
• M.S. Thesis, University of Nevada, Las Vegas, 2018-present. Analyzing the
acoustoelasticity of polycrystalline olivine in an in-situ ultrasonic-modified D-DIA
experiment and its implications for the upper mantle.
• Research Assistant, Texas A&M University, July 2016-Dec 2016. Conducted sample
preparation and geochemical data interpretation on the Diagenesis of the Bataraja
Formation, Indonesia.
PROFESSIONAL EXPERIENCE
• Teaching Assistant for Geology 101 Lab, University of Nevada, Las Vegas, Spring 2018-
Fall 2018
o Responsibilities include instructing students on the basic application of geology
and geologic processes
ANALYTICAL EQUIPMENT
• D-DIA multi-anvil apparatus, DIASCoPE acoustic system, Synchrotron X-ray
Diffraction
o Conducted ultrasonic-modified deformation experiments at the 6-BM-B
Synchrotron beamline at Argonne National Laboratory
• Powder X-ray Diffractometer- UNLV
• Scanning Electron Microscope and Field Emission Scanning Electron Microscope-
UNLV
DATA PROCESSING TOOLS

• Experience with Adobe Illustrator & Photoshop, Elastic-Plastic Self-Consistent (EPSC)
numerical modeling, ImageJ, MATLAB, Microsoft Suite, Python, and Plot85
AWARDS/SCHOLARSHIPS
• UNLV Graduate Academic Achievement Award, Spring 2020
• COMPRES Student Travel Award, Summer 2019
• Access Grant, University of Nevada- Las Vegas, Fall 2019- Spring 2020
• Bernada French Scholarship, University of Nevada- Las Vegas, Fall 2018- Spring 2019
• Department of Geology Outstanding Senior Award, Texas A&M University, Summer
2017
68
• Geosciences Medallion Scholar, Texas A&M University, Summer 2017
• Marianna and William Green ’53 Scholarship, Texas A&M University, Spring 2017
• Dean’s List, Texas A&M University, Fall 2015
CONFERENCES/WORKSHOPS
• University of Nevada, Las Vegas Geosymposium, April 2020- Oral presentation
• 2020 Stewardship Science Academic Programs (SSAP) Symposium, Feb 2020- Poster
presentation
• American Geophysical Union Annual Meeting, Dec 2019- Poster presentation
• COMPRES 2019 Annual Meeting, July 2019- Poster presentation
• 9th International Symposium on Granite Pegmatites, June 2019
• University of Nevada, Las Vegas Geosymposium, April 2019- Poster presentation
• COMPRES Workshop: Envisioning the Next Generation of In-situ Synchrotron X-ray
Techniques in Large-Volume High Pressure Apparatus for Mineral and Rock Physics,
Sept 2018
• University of Nevada, Las Vegas Geosymposium, April 2018- Poster presentation
• Texas A&M University, Berg-Hughes Symposium, Oct 2016- Poster presentation
PROFESSIONAL ORGANIZATIONS
• American Geophysical Union, 2019-present
• Geological Society of Nevada (GSN), 2018-present
• Microanalysis Society, 2020-present
• Microscopy Society of America, 2020-present
• Society of Economic Geology (SEG), 2018-present
o UNLV Chapter Secretary- Spring 2019-Fall 2019
o UNLV Chapter Treasurer- Spring 2020-present
69

Initial Measurements On The Effect of Stress On P - and S-Wave Vel

Uploaded by

Copyright:

Available Formats

Initial Measurements On The Effect of Stress On P - and S-Wave Vel

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Initial Measurements On The Effect of Stress On P - and S-Wave Vel

Uploaded by

Copyright:

Available Formats

UNLV Theses, Dissertations, Professional Papers, and Capstones

Initial Measurements on the Effect of Stress on P- and S-wave

Follow this and additional works at: https://digitalscholarship.unlv.edu/thesesdissertations

Part of the Geophysics and Seismology Commons

Bachelor of Science – Geology

A thesis submitted in partial fulfillment

Master of Science – Geoscience

University of Nevada, Las Vegas

November 20, 2020

This thesis prepared by

Initial Measurements on the Effect of Stress on P- and S-Wave Velocities in Olivine

is approved in partial fulfillment of the requirements for the degree of

Master of Science – Geoscience

Pamela Burnley, Ph.D. Kathryn Hausbeck Korgan, Ph.D.

Michael Wells, Ph.D.

Oliver Tschauner, Ph.D.

Moses Karakouzian, Ph.D.

acoustoelastic effect, in San Carlos olivine. A recently developed experimental ultrasonic

derived from synchrotron X-ray diffraction. We use elastic-plastic self-consistent (EPSC)

materials at high pressure conditions.

and Dr. Moses Karakouzian for their involvement and encouragement.

Jason Reek for their assistance in data collection.

have greatly aided in my development as a scientist.

Physics Institute, Department of Geosciences, Stony Brook University.

List of Tables .............................................................................................................................. vii

List of Figures............................................................................................................................ viii

Appendix 2: Figures ....................................................................................................................49

Curriculum Vitae .........................................................................................................................67

Table 1 Acoustoelastic constant values ...................................................................................... 38

Table S1 Unit cell parameters used in EPSC models ................................................................. 39

Table S2 Experimental conditions for each deformation sequence ............................................ 40

Table S5 Acoustoelastic Constants comparison table with literature ......................................... 43

Table S6 San Carlos Olivine velocity compilation of literature ................................................. 46

Figure 1 Schematic diagram of ultrasonic wave propagation in an experimental set-up ........... 49

Figure 3 Comparison of acoustoelastic slope at different pressure conditions ........................... 51

Figure 5 Acoustoelastic constants as a function of pressure and temperature ............................ 53

Figure S1 Schematic cross-sectional view of the D-DIA apparatus ........................................... 54

Figure S2 Schematic of sample assembly for San_381and San_416 ......................................... 55

Figure S4 Schematic of the X-ray beam propagation during an experiment .............................. 57

Figure S7 Communication path for DIASCoPE ultrasonic interferometry measurements ........ 60

Figure S8 Comparison of acoustoelastic constants to Egle and Bray (1976) ............................. 61

Earth. Measurement of acoustic velocities, absolute pressure determination, and derivation of

collection of ultrasonic interferometry measurements in conjunction with synchrotron X-

measured to deep Earth conditions.

We evaluate the acoustoelasticity of polycrystalline San Carlos olivine at high pressure

elastic constants of isotropic materials from wave velocity measurements. A first-order

Nonlinear Elastic Behavior of Solids

infinitesimal deformation of an elastic material, the stress-strain relationship is expressed as a

and second-order elastic constants:

second order elastic constants:

𝜎𝑖𝑗 = 𝐶𝑖𝑗𝑘𝑙 𝜀𝑘𝑙 (2)

in violation of the assumptions of the linear elastic theory.

state as the waves propagate through a solid (Smith, 1963).

as to determine the third-order elastic constants. To study the acoustoelasticity of a material

1), the following equations are used:

3b providing a check on these values (Hughes and Kelly, 1953).

deformation, with i corresponding to the direction of propagation and j corresponding to the

Stress-induced Velocity Changes in Metallic Materials and Concrete

experimentally confirmed their theory of acoustoelasticity through the determination of third-