Anda di halaman 1dari 66

THERMAL PROPERTIES OF YTTRIUM ALUMINUM GARNETT FROM MOLECULAR DYNAMICS SIMULATIONS

by Majid S. alDosari

Thesis Submitted to the Faculty of the Graduate School of Vanderbilt University in fulllment of the requirements for the degree of

MASTER OF SCIENCE in Mechanical Engineering

May, 2011

Nashville, Tennessee

Approved:

Date:

MECHANICAL ENGINEERING

THERMAL PROPERTIES OF YTTRIUM ALUMINUM GARNETT FROM MOLECULAR DYNAMICS SIMULATIONS

MAJID S. ALDOSARI

Thesis under the direction of Professor Greg Walker Yttrium Aluminum Garnet (YAG, Y3 Al5 O12 ) and its varieties have applications in thermographic phosphors, lasing mediums, and thermal barriers. In this work, thermal properties of crystalline YAG where aluminum atoms are substituted with gallium atoms (Y3 (Al1-x Gax )5 O12 ) are explored with molecular dynamics simulations. For YAG at 300K, the simulations gave values close to experimental values for constant-pressure specic heat and thermal expansion. For various values of x, the simulations predicted no change in thermal expansion, and a decrease in specic heat with increasing x. While non-equilibrium simulations predicted bulk thermal conductivity approaching experimental bulk values, equilibrium (Green-Kubo) simulations predicted conductivities 20% to 50% lower. Furthermore, equilibrium simulations showed no increase with system length in contrast to non-equilibrium simulations. Nonetheless, both simulation types gave conductivity for x=0 higher than that of x=1, which was in turn higher than that of x=0.5. The difference in the vibrational frequency content of the simulations is believed to be responsible for the disparity. Finally, a procedure is introduced to objectively evaluate Green-Kubo integrals.

Approved:

Date:

AKNOWLEDGEMENTS

Acknowledgment is mainly to Saudi ARAMCO and its Advanced Degree Program which sponsored my degree. Also, every Thermal Engineering Lab member has contributed in some way to this work. But, it is difcult to order their contribution so they are listed here according to their seating in counter-clockwise order starting from the labs only window: Nicholas Roberts, Terry Musho, Ebonee Walker, Mike Myers, and Rachel Hansel. Last but not least, I appreciate Dr. Walkers magic in directing this work which I have found to be interesting, enjoyable, and satisfying.

ii

TABLE OF CONTENTS

Page LIST OF FIGURES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . v

LIST OF TABLES . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vi

NOMENCLATURE . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . vii

Chapter I. INTRODUCTION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1

II. YAG MODEL . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

III. LATTICE CONSTANT, SPECIFIC HEAT CAPACITY, AND THERMAL EXPANSION . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1. Simulation Setup. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 2. Results and Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

7 7 8

IV. THERMAL CONDUCTIVITY FROM NON-EQUILIBRIUM SIMULATIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 1. Simulation Setup. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 11 2. Results and Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13 V. THERMAL CONDUCTIVITY FROM EQUILIBRIUM SIMULATIONS. . . 16 1. Simulation Setup. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 2. Results and Discussion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 17 VI. CONCLUSIONS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25

APPENDIX. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 26

iii

BIBLIOGRAPHY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57

iv

LIST OF FIGURES

Figure 1 Figure 2

CONVENTIONAL UNIT CELL YAG SYMMETRY ON [1 0 0] FACE. Red:O, Green:Al, Blue:Y. (Image generated by Jmol [1]). . . . . . POLYHEDRA IN YAG [2] (with permission from the American Chemical Society) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . TEMPERATURE VS. TIME FOR 25% Al SUBSTITUTION . . . . . . . . c p VS. SUBSTITUTION FRACTION, x . . . . . . . . . . . . . . . . . . . . . . . .

4 5 8 9

Figure 3 Figure 4 Figure 5 Figure 6

NON-EQUILIBRIUM MD CONFIGURATION . . . . . . . . . . . . . . . . . . 12 AVERAGE TEMPERATURE PROFILE FOR A SAMPLE SIZE 8 UNIT CELLS WIDE, WITH 50% Al SUBSTITUTION, AVERAGED OVER .618 ns . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 14 NON-EQUILIBRIUM THERMAL CONDUCTIVITY TRENDS FOR x = 0, 50, AND 100% . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 15 HEAT CURRENT AUTOCORRELATION (V.8a 13 & V.8b) AND ITS INTEGRAL (V.8c & V.8d) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18 TIME DEPENDENCE OF THE HCACF [(A) & (B)] AND ITS INTEGRAL [(C) & (D)] (WHOSE CONVERGED VALUE IS THE THERMAL CONDUCTIVITY) FOR QUARTZ(A) AT T=250 K AND QUARTZ(C) AT T=200 K (q=S. with permission from Elsevier) . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 19 ENVELOPE OF SINUSOIDAL EXPONENTIAL DECAY IS EXPONENTIAL DECAY . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 ILLUSTRATION OF PROCEDURE FINDING RELAXATION TIME . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 21 EQUILIBRIUM THERMAL CONDUCTIVITY TRENDS FOR x = 0, 50, AND 100%. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22 HEAT CURRENT POWER SPECTRAL DENSITY . . . . . . . . . . . . . . . 24

Figure 7

Figure 8 Figure 9

Figure 10 Figure 11 Figure 12 Figure 13

LIST OF TABLES

Table 1 Table 2 Table 3 Table 4 Table 5

THERMAL PROPERTIES CRYSTALLINE OF YAG . . . . . . . . . . . . . . TWO-BODY PARAMETERS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . THREE-BODY PARAMETERS . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

2 3 6

THERMAL CONDUCTIVITY SIMULATION SETTINGS . . . . . . . . . 13 EQUILIBRIUM SIMULATION SETS . . . . . . . . . . . . . . . . . . . . . . . . . . 17

vi

NOMENCLATURE

A Ac B E Ets avg Ets HC HCAC MD NPT NVE NVT S T U V V0 a cp cv fi j k kB q q ri j r0,i j t u.c. v x jik 0

two-body potential parameter cross-sectional area two-body potential parameter total energy change in system energy due to a thermostat average of the absolute values of Ets heat current heat current autocorrelation molecular dynamics isothermal and isobaric ensemble: conserves moles (N), pressure (P) , and temperature (T) microcanonical ensemble: conserves moles, volume (V), and energy (E) canonical ensemble: conserves moles, volume, and temperature heat current temperature potential energy volume initial volume lattice spacing constant-pressure specic heat constant-volume specic heat force of atom i on atom j thermal conductivity Boltzmanns constant elementary charge heat ux distance between atoms i and j atomic bonding cutoff between atoms i and j time unit cell velocity fraction of Al randomly substituted with Ga: Y3 (Al1-x Gax )5 O12 mean free path three-body angular expression thermal expansion three-body potential parameter energy three-body potential parameter frequency angle jik three-body potential parameter relaxation time constant vii

CHAPTER I

INTRODUCTION

Yttrium Aluminum Garnet (YAG, Y3 Al5 O12 ) is a synthetic crystal with a complicated atomic structure. Doped varieties are used for solid state lighting [3], as thermographic phosphors [4], in thermal barriers [5], and as lasing mediums [6]. The inherent thermal design issues of these applications motivate the study of the thermal transport in YAG. Since YAG has a relatively low thermal conductivity, its thermal transport is primarily governed by phonons, which are inferred from atomic motion. Therefore, classical molecular dynamics (MD) simulations can offer insight into the thermal properties of YAG of varying system parameters as long as temperatures are above 50 K, approximately 10% of the Debye temperature. In the work introduced by Hansel [4], gallium is substituted for aluminum in various percentages to study its effect on luminescence. In this work, aluminum substitution with gallium will be a variable to study its effect on constant-pressure specic heat, lattice constant, thermal expansion, and thermal conductivity. In addition, system size will be introduced as a variable for thermal conductivity. Thermal properties of pure, crystalline YAG are reported in Table 1 for comparison to results from the simulations.

Table 1 THERMAL PROPERTIES CRYSTALLINE OF YAG Melting Temperature, Tm Thermal Expansion, Specic Heat, cp Thermal Conductivity, k Debye Temperature, TD 2213 K [7] 7.010-6 K-1 @300 K [8] 0.6 J g-1 K-1 @300 K [9] [10] 1014 W m-1 K-1 [11] 750 K [12]

CHAPTER II

YAG MODEL

YAG crystallizes in a highly symmetric cubic structure (detailed in Reference [2]). The conventional unit cell of YAG contains 160 atoms comprising of 24 yttrium, 40 alu minum, and 96 oxygen atoms; with a lattice parameter of 12.01 A [13]. The relative arrangement of Y, O, and Al can be described with polyhedra as shown in Figure 2, while the symmetry can be seen in Figure 1. Jun, et al. [14] have modeled the complex structure of pure YAG for classical molecular dynamics simulation by tting elasticity and lattice constant data to the potential energy functions reproduced here. However, for the purposes of this work, gallium atoms are considered substitutable for aluminum atoms, representing only a change in mass. The potential energy for any pair of atoms (ions), i and j, are given by the sum of a Coulombic and exponential term given by ri j qi q j + Ai j exp ri j Bi j

Ui j = c

(1)

where c is an energy conversion constant equal to 14.4 eVA e1 .

Table 2 TWO-BODY PARAMETERS ij Ai j (eV) B1 (A-1 ) qi (e) ij O-O 2449.44 3.44 2 Al-O 1740.31 3.44 +3 Y-O 1250.85 2.86 +3 Al-Al 312.11 14.06 Y-Y 245.14 14.06 Y-Al 256.55 14.06

Figure 1 CONVENTIONAL UNIT CELL YAG SYMMETRY ON [1 0 0] FACE. Red:O, Green:Al, Blue:Y. (Image generated by Jmol [1])

In addition to pair potentials, three-body potentials were employed in order to enforce angular bonding between certain triplets of atoms as seen in Figure 2. For a triplet of atoms; i, j, and k; the three-body interaction, U jik , is expressed as

U jik =

jik exp

i j ri j ri j,0

ik + rik rik,0 jik

ri j < ri j,0 and rik < rik,0 ri j ri j,0 or rik rik,0

(2)

jik is an angular factor whose functional form depends on which triplet of atoms is being considered. For bonds where j and k represent oxygen atoms, and i represents an aluminum or yttrium atom (OAl/YO), jik is expressed as
0 jik = [(cos jik cos jik ) sin jik cos jik ]2 .

(3)

Figure 2 POLYHEDRA IN YAG [2] (with permission from the American Chemical Society)

Table 3 THREE-BODY PARAMETERS jik jik (eV) (A) r0,i j (A) 0, jik ( ) Al/YOAl/Y 6.242 2.0 2.6 109.5 OAlO 149.324 2.8 3.0 109.5 OYO 168.250 2.8 3.0 109.5

For triplets Al/YOAl/Y, jik is


0 jik = (cos jik cos jik )2 .

(4)

The parameters of Equations 2, 3, and 4 are in Table 3. All simulations were performed using the June 5, 2010 version of the Large-scale Atomic/Molecular Massively Parallel Simulator (LAMMPS) package [15]. Furthermore, all simulations used the Verlet algorithm [16] for time integration. Periodic boundary conditions were imposed in all directions on all simulations to balance the repulsive forces from similarly charged two-body interactions. The two-body interactions expressed by Equation 1 are implemented using the included Born-Mayer-Huggins potential [17] where Coulombic interactions greater than a specied cutoff were computed in k-space using a particle-particle particle-mesh (PPPM) solver. The LAMMPS was modied to include the potential energies expressed by Equations 2, 3, and 4 (see Appendix).

CHAPTER III

LATTICE CONSTANT, SPECIFIC HEAT CAPACITY, AND THERMAL EXPANSION

This section describes how the constant-pressure heat capacity, thermal expansion, and the lattice constant are found from simulated heating. These quantities are checked for dependence on temperature and random aluminum substitution fraction, x. Obtaining the lattice constant is necessary to perform the thermal conductivity simulations.

3.1

Simulation Setup

a, c p , and are calculated from ve simulations to represent a random aluminum substitution percentage of 0, 25, 50, 75, and 100 percent. The system size was 222 unit cells (1280 atoms) for all simulations. Each simulation was initialized with a Gaussian velocity distribution giving a temperature of 10 K. Then, using NPT time integration, the system is thermostatted at 10 K and barostatted at 10 bars for 80,000 timesteps. Subsequently, the temperature is ramped up by 100 K. After that, the system is again barostatted at 10 bars, thermostatted at the new temperature of 110 K, and time-integrated for another 80,000 timesteps. This process is repeated until the system reaches a temperature of 3000 K. Except for aluminum substitution, this simulation setup closely follows that of Jun [14] to allow for comparison. The simulations ran with a 0.002 ps timestep, 10-4 PPPM precision, and a 7 A cutoff (for two-body interactions). These parameters were found to balance energy conservation with execution speed. The damping parameter on the thermostat was 1 ps while the damping parameter on the barostat was 10 ps. An additional drag factor of 1 was imposed to dampen oscillations. 7

3000 2500 temperature (K) 2000 1500 1000 500 00 500 1220 1200 1180 1160 1140 1120 1100 950 960 970 980 990 1000 1010 940 1000 1500 2000 2500 timestep 10-3 (.002 ps)

Figure 3 TEMPERATURE VS. TIME FOR 25% Al SUBSTITUTION

A time series for the total energy, temperature, pressure, and the lattice constant was collected from the simulation by an averaging procedure: the quantities are sampled every 10 timesteps 100 times, and subsequently averaged. This average is then recorded.

3.2

Results and Discussion

The stability of kinetic energy in the simulation can be seen in the time series for temperature in Figure 3. Variations in each temperature step were within 10 K while exhibiting almost no transients. Similarly, the lattice parameter, a, exhibited variations of 0.1% at most. Furthermore, there is a linear relationship between temperature and total energy for all simulations which is expected from classical simulations. c p for all temperatures can then be computed using the slope of this relationship 8

0.85 cp (J g-1 K-1 ) 0.80 0.75 0.70 0.65 0.0 0.2 0.4 0.6 Substitution Fraction 0.8 1.0

Figure 4 c p VS. SUBSTITUTION FRACTION, x

cp =

E . mT

(5)

Figure 4 shows c p from all simulations. Given that E/T were close to equal for all simulations, c p is primarily a function of mass. However, the calculated c p of pure YAG is about 40% greater than the experimental value of 0.6 J g-1 K-1 . This discrepancy may be due to the interatomic potentials since they were specied with no thermal considerations. Similarly, a linear relationship was also found between temperature and lattice constant for all simulations. can be found using the slope of this relationship V . V0 3T

(6)

However, was 7.810-6 K-1 for all simulations showing no dependence on aluminum

substitution. This is expected given that the potential energy functions were the same for the substituted atoms. Nonetheless, this compares well with the established gure of 7.010-6 K-1 for pure YAG. As a result, the lattice constant, a, was taken to be only a function of temperature expressed in angstroms as a = 9.67 105 T A + 12.4 A . K

(7)

So at 300 K, the lattice constant is 12.43 A which is 3.6% over published gures [13] but only 0.6% over the result from MD simulations of Reference [14]. But, the validity of and a for gallium-substituted YAG can be questioned since they have not been veried by experiment.

10

CHAPTER IV

THERMAL CONDUCTIVITY FROM NON-EQUILIBRIUM SIMULATIONS

The simulations described in this section measure thermal conductivity in the [1 0 0] direction. Since YAG is arranged in a cubic structure, its thermal conductivity is given by a diagonal tensor

k 0 0 ki j = 0 k 0 . 0 0 k Therefore, the thermal conductivity tensor is fully described by this one measure. An effective thermal conductivity, k is obtained by imposing a temperature gradient so that Fouriers law of conduction in one dimension may be evaluated directly. qx dx Ac dT (8)

k=

(9)

4.1

Simulation Setup

Two temperature gradients were imposed in a symmetric conguration, as represented in Figure 5, since the simulations feature periodic boundary conditions. This conguration ensures that the hot slab is not exposed to the cold slab at the simulation boundary. For all simulations, the lattice spacing was 12.43 A. Furthermore, the hot slab was 811 (LHD) unit cells, while each cold slab was 411 unit cells. As such, the atoms in each 411 region were assigned to four thermostats. The thermostats were applied to the hot slab and cold slab such that the average temperature of the temperature gradient was 300 K. Additionally, the temperature difference between the thermostats was 11

Periodic Boundary Periodic Boundary Periodic Boundary

Cold Slab

Sample Region q

Hot Slab

Sample Region q

Cold Slab

Periodic Boundary
Figure 5 NON-EQUILIBRIUM MD CONFIGURATION

xed at 100 K for the largest simulations, and 40 K for the others. However, the simulations were varied in the size of the sample region and in random aluminum substitution fraction, x. The sample sizes were 4, 8, 16, and 32; while x was 0, 50, and 100 percent, giving a total of 12 simulations. An initial linear temperature gradient was imposed over the sample region with a random gaussian velocity distribution. The temperature gradient is then maintained by the thermostats for 1000 timesteps times the system size in unit cells to achieve steady state. The region controlled by the thermostats are time-integrated with the NVT integrator, while the sample region was time-integrated with the NVE integrator. Subsequently, the simulations ran for at least 0.2 fs to at most 2.5 fs, depending on the amount of time necessary to minimize short term energy uctuations. During the run, the average temperature in each unit cell is calculated every 10 timesteps. The time average is then recorded for 100 of these calculations. Meanwhile, the data from the thermostats are recorded every 1000 timesteps. The data from the thermostats represent the cumulative energy change of the system due to their action. In particular, here it represents an energy addition or subtraction due to the ow of energy from the hot slab to the cold slab. Table 4 lists the simulation settings. The cutoffs apply to the two-body potential

12

Table 4 THERMAL CONDUCTIVITY SIMULATION SETTINGS timestep .001 ps exp. term cutoff 4.0 A Coulombic term cutoff 7.0 A PPPM precision 10-5 thermostat damping 1 ps

(Equation 1).

4.2

Results and Discussion

Figure 6 typies the temperature prole found in all simulations when averaged over their run time. To calculate thermal conductivity according to Fouriers law (Equation 9), the slope from a line t to the sample region from each side of the temperature prole is averaged to obtain x/(Ac T ). In addition, qx is E/t where E is represented by the energy
avg difference of Ets between the start and end of the simulation in steady state and Ac is the

sample cross-sectional area. Figure 7 shows the results of the thermal conductivity calculation. For pure YAG, the trend is approaching (the lower end of) its accepted bulk value of 10 Wm-1 K-1 [11]. All trends of a certain substitution fraction (varying system size) can be explained by simplied kinetic theory: the relationship between thermal conductivity, specic heat capacity, group velocity, and mean free path is given by 1 k = vcv . 3

(10)

Since system size is the only variable, the decrease in thermal conductivity with decreasing system size is due to decreased mean free path. Meanwhile, group velocity is needed to fully explain the trends for a certain sys13

330 320 Temperature (K) 310 300 290 280 2700 50 100 150 200 250 Distance (A) 300 350 400

Figure 6 AVERAGE TEMPERATURE PROFILE FOR A SAMPLE SIZE 8 UNIT CELLS WIDE, WITH 50% Al SUBSTITUTION, AVERAGED OVER .618 ns

14

9 8 Conductivity (W m-1 K-1 ) 7 6 5 4 3 0.000 0.005 0.010 0.015 -1 (A-1 ) length 0.020

100 50 0

0.025

Figure 7 NON-EQUILIBRIUM THERMAL CONDUCTIVITY TRENDS FOR x = 0, 50, AND 100%

tem size (varying substitution fraction). Nonetheless, assuming acoustic phonons are the primary thermal carriers, the lower thermal conductivity of YAG fully substituted with gallium compared to pure YAG can be explained by the difference in their phonon dispersions: as gallium is a heavier element than aluminum, the group velocity of fully substituted YAG will be lower, resulting in lower thermal conductivity. Still, decreased mean free path explains why the thermal conductivity for x=50% is less than x=0% and x=100%. The presence of gallium atoms distributed non-homogeneously in the structure represents massdifference impurity phonon scattering centers that reduce the mean free path of phonons.

15

CHAPTER V

THERMAL CONDUCTIVITY FROM EQUILIBRIUM SIMULATIONS

Equilibrium simulations provide an alternative method of obtaining the bulk thermal conductivity of YAG. However, practical limits on computational resources restrict simulating YAG systems to less than 30 nanometers in length. Nonetheless, this more closely represents bulk transport in the sense that temperature is virtually constant on the nanoscale. The Green-Kubo expression is used to relate the heat current, S, to the thermal conductivity through the integral of its autocorrelation [18], 1 kBV T 2
0

k=

Sx (t) Sx (0) .

(11)

The heat current is dened in this work as [19] 1 fi j vi + v j ri j 2 i< j

S = Ei vi +
i

(12)

where i and j are indices to all atoms. The variables in the expression are all scalars representing quantities along an axial direction as only scalar values are needed to represent the thermal conductivity as explained in Section IV. The physical interpretation of Equation 12 is that it represents the direction of energy transfer at a point in time [20]; while Equation 11 implies that the amount of time S stays correlated with itself is proportional to thermal conductivity.

16

Table 5 EQUILIBRIUM SIMULATION SETS L (u.c.) HD (u.c.) x (%) No. of init. vel. 4, 8, 16 22 0, 50, 100 8 8 22 50 3 32, 64 22 0 3 8 44 0, 50, 100 3 4, 8, 16 11 0, 50, 100 1

5.1

Simulation Setup

Simulation parameters resemble the non-equilibrium simulations to allow comparison. As such, the simulations can be divided into three sets based on their size: The main set had a depth and height of 22 unit cells with lengths of 4, 8, and 16 unit cells for substitution fractions 0, 50, and 100%. For each size and substitution fraction, 8 independent simulations were started with different random initial velocities. This set, along with other sets, are summarized in Table 5. Each entry in the table that contains a substitution fraction of 50% represents a particular random distribution of a certain simulation size. As in the non-equilibrium simulations, all simulations were thermostatted by a NVT integrator at 300 K with the lattice constant equal to 12.43 A. Other simulation settings, listed in Table 4, are also the same. Also, in a manner similar to the non-equilibrium simulations, steady state was achieved by running the simulations for 1000 timesteps times the number of unit cells being simulated. After that, the heat current is recorded at every timestep for 106 timesteps, 1 ns.

5.2

Results and Discussion

Finding the thermal conductivity of each simulation involves the numeric evaluation of Equation 11 from simulation data (of nite time). The heat current autocorrelation functions (HCAC) for all simulations were oscillatory. However, for simulations with x=50%,

17

less oscillations were observed in a given time. These qualitative differences are shown in Figure 8 for two samples of dimensions 1622 unit cells (2002525 A).

1.0 Sx(t) Sx(0) / Sx(0) Sx(0) 0.5 0.0 0.5 0


Sx(t) Sx(0) / Sx(0) Sx(0)

a x=50%

10 time (ps)

15

20

1.0 0.8 0.6 0.4 0.2 0.0 0.2 0.4 0.6 0.80

b x=0%

10 time (ps)

15

20

14 Conductivity (W m-1 K-1 ) 10 8 6 4 2 00 5 10 time (ps) c x=50% 15 20


Conductivity (W m-1 K-1 )

14 12 10 8 6 4 2 00 5 10 time (ps) d x=0% 15 20

12

Figure 8 HEAT CURRENT AUTOCORRELATION (V.8a 13 & V.8b) AND ITS INTEGRAL (V.8c & V.8d)

Subjectively, the integrals show convergence around the time when the oscillations minimize. McGaughey evaluated similar-behaving integrals in the analysis of complex silica structures [21], shown in Figure 9. A convergence region is subjectively identied over which an average of the integral is taken as the thermal conductivity. So, it follows that the integrals objective evaluation requires two steps. The rst is to determine at what time to terminate the integral as the convergent value does not

18

Figure 9 TIME DEPENDENCE OF THE HCACF [(A) & (B)] AND ITS INTEGRAL [(C) & (D)] (WHOSE CONVERGED VALUE IS THE THERMAL CONDUCTIVITY) FOR QUARTZ(A) AT T=250 K AND QUARTZ(C) AT T=200 K (q=S. with permission from Elsevier)

19

always hold indenitely. The second step is to take an average over the time in which the integral is convergent. More specically, the rst step involves nding some relaxation time, , representing the exponentially decaying nature of the autocorrelation function. So, an exponential decay function, constantexp(t/), is t to the peak points [22] of the absolute value of the autocorrelation function. The t corresponds to the exponential function in the sinusoidal decay, constantexp(t/)cos(t), since

| exp(t/) cos(t)| = exp(t/)

when

d | exp(t/) cos(t)| = 0 dt . This is shown graphically in Figure 10. But, due to noise in the autocorrelation, peaks below 5% of the maximum peak value are not taken in the t. 5% is chosen because it select points mostly in the transient part of the autocorrelation. Otherwise this tolerance does not affect the t much as long as it is reasonably low. This procedure is illustrated in Figure 11 for the autocorrelation in Figure V.8d. The peaks are shown as dots on the absolute value of the autocorrelation with an exponential decay t (of relaxation time equal to 0.62 ps) going through most of these dots. Now, the terminal time for the autocorrelation integral is taken to be 30 times as the subjective evaluation of the integral agrees with result from the (objective) evaluation procedure in the second. In the second step, a histogram is made for the (discrete) values of the terminated HCAC integral with 100 equally spaced bins ranging from zero to the maximum value of the integral. The convergent value is taken as the midpoint of the bin range with the highest frequency. The highest frequency bin is expected to represent the converged value because 20

Figure 10 ENVELOPE OF SINUSOIDAL EXPONENTIAL DECAY IS EXPONENTIAL DECAY

1.0 | Sx(t) Sx(0) / Sx(0) Sx(0) | 0.8 0.6 0.4 0.2 0.0 0 1 2 time (ps)
21

Figure 11 ILLUSTRATION OF PROCEDURE FINDING RELAXATION TIME

8 7 Conductivity (W m-1 K-1 ) 6 5 4 3 2 1 0.000

50.0

100

0.005

0.010 0.015 -1 (A-1 ) length

0.020

Figure 12 EQUILIBRIUM THERMAL CONDUCTIVITY TRENDS FOR x = 0, 50, AND 100%

other values are transient in comparison. So, for the HCAC integrals shown in Figure V.8c and V.8d, the convergent values are 3.0 W m-1 K-1 and 4.6 W m-1 K-1 respectively. In these cases, the apparent increase in the integral value after the oscillations minimize is part of low frequency uctuations about the converged value. Figure 12 summarizes the results from thermal conductivity evaluations for the 22 sets. The error bars represent the standard deviation of thermal conductivity evaluations from different initial velocities. Clearly, no dependence on length is observed and the conductivity for x=100% is lower than x=0%, while the conductivity for x=50% is lower than 100% and 50%. Also 22

(not shown), there is no signicant difference in conductivity when compared to the 4x4 set. Similarly, for the same x, different random atom substitution distributions had no effect. In contrast, the 11 set HCAC integrals were not stable and showed no convergence and thus could not be evaluated. To explain these trends (except the 11 cases), the power spectral density (PSD) of the heat current was examined for each simulation. Denite peaks were found for all simulations as in Figure 13 which shows the heat current PSD for the two simulations whose HCACs and HCAC integrals were shown in Figure 8. In general, the peaks for x=0% or 100% were more dened and an order of magnitude higher than peaks for x=50%. McGaughey [23] associates peaks in the HC PSD of complex silica structures with optical phonon modes and suggests that the electrostatic term in their interatomic potential as their cause. A similar argument can be made for YAG given its complex structure and the electrostatic term in its interatomic potential. From here, it is suggested that acoustic phonons do not contribute signicantly to the thermal conductivity. This is supported by observing that no signicant sub-THz frequency components are present in the HC PSD of all simulations, where acoustic phonons are expected to manifest.

23

10-6 10-7 10-8 10-9 5 10

a.u./frequency

x=0%
15 20 frequency (THz)

x=50%
25 30

Figure 13 HEAT CURRENT POWER SPECTRAL DENSITY

24

CHAPTER VI

CONCLUSIONS

The model for pure YAG shows good agreement with experimental values for constant-pressure specic heat, thermal expansion, and non-equilibrium thermal conductivity at 300 K. The relationship between thermal conductivity and aluminum substitution is not monotonic. Consideration of this relationship is important in high temperature-transient applications. In absolute terms, optical phonon contribution to thermal conductivity is relatively constant and represents a lower limit. As a fraction, optical phonons contribute about 20% to 50% depending on system size and substitution fraction. The remaining acoustic contribution is only present in non-equilibrium. However, more investigation is needed to explain and quantify the discrepancy in thermal conductivities between non-equilibrium and equilibrium methods. Furthermore, experimental thermal properties for gallium-substituted YAG are needed to conrm the YAG model used in this work.

25

APPENDIX

LAMMPS SOURCE MODIFICATION In order to compute three-body energies and their forces for this work, code for two potentials was added based on the included Stillinger-Weber potentials (composed of two and three-body interactions). One implemented the three-body interaction formed by Equation 2 and Equation 3 while zeroing the original two-body interaction. The other involved only zeroing the two-body part of the Stillinger-Weber potentials since the threebody interaction formed by Equation 2 and Equation 4 is compatible with the original code. The following code shows the implementation of Equations 2 and 3 (mbmh potential in LAMMPS script).
/ LAMMPS Larges c a l e A t o m i c / M o l e c u l a r M a s s i v e l y P a r a l l e l S i m u l a t o r h t t p : / / lammps . s a n d i a . gov , S a n d i a N a t i o n a l L a b o r a t o r i e s S t e v e P l i m p t o n , s j p l i m p @ s a n d i a . gov C o p y r i g h t ( 2 0 0 3 ) S a n d i a C o r p o r a t i o n . Under t h e t e r m s o f C o n t r a c t DEAC0494AL85000 w i t h S a n d i a C o r p o r a t i o n , t h e U . S . Government retains certain rights in t h i s software . This s o f t w a r e i s d i s t r i b u t e d under t h e GNU G e n e r a l P u b l i c L i c e n s e . S e e t h e README f i l e i n t h e t o p l e v e l LAMMPS d i r e c t o r y . / / C o n t r i b u t i n g a u t h o r : A i d a n Thompson ( SNL ) / # include # include # include # include # include # include # include # include # include # include # include # include # include # include # include math . h s t d i o . h s t d l i b . h s t r i n g . h pair mbmh . h atom . h neighbor . h n ei gh re qu es t . h force . h comm . h memory . h neighbor . h n e i g h l i s t . h memory . h e r r o r . h

26

u s i n g n a m e s p a c e LAMMPS NS ; # d e f i n e MAXLINE 1024 # d e f i n e DELTA 4 / / PairMBMH : : PairMBMH (LAMMPS lmp ) : P a i r ( lmp ) { single enable = 0; one coeff = 1; nelements = 0; e l e m e n t s = NULL ; nparams = 0 ; maxparam = 0 ; p a r a m s = NULL ; elem2param = NULL ; } / c h e c k i f a l l o c a t e d , s i n c e c l a s s can be d e s t r u c t e d when i n c o m p l e t e / PairMBMH : : PairMBMH ( ) { i f ( elements ) f o r ( i n t i = 0 ; i < n e l e m e n t s ; i ++) d e l e t e [ ] e l e m e n t s [ i ] ; delete [] elements ; memory>s f r e e ( p a r a m s ) ; memory>d e s t r o y 3 d i n t a r r a y ( elem2param ) ; if ( allocated ) { memory>d e s t r o y 2 d i n t a r r a y ( s e t f l a g ) ; memory>d e s t r o y 2 d d o u b l e a r r a y ( c u t s q ) ; d e l e t e [ ] map ; } } / / v o i d PairMBMH : : compute ( i n t e f l a g , i n t v f l a g ) { i n t i , j , k , i i , j j , kk , inum , jnum , jnumm1 , i t a g , j t a g ; i n t i t y p e , j t y p e , ktype , ijparam , ikparam , i j k p a r a m ; d o u b l e xtmp , ytmp , ztmp , d e l x , d e l y , d e l z , evdwl , f p a i r ; double rsq , rsq1 , r s q 2 ; double d e l r 1 [ 3 ] , d e l r 2 [ 3 ] , f j [ 3 ] , fk [ 3 ] ; i n t i l i s t , j l i s t , numneigh , f i r s t n e i g h ; evdwl = 0 . 0 ; i f ( eflag | | vflag ) ev setup ( eflag , vflag ) ; else evflag = v f l a g f d o t r = 0;

27

d o u b l e x = atom>x ; d o u b l e f = atom>f ; i n t t a g = atom>t a g ; i n t t y p e = atom>t y p e ; i n t n l o c a l = atom>n l o c a l ; i n t n e w t o n p a i r = f o r c e >n e w t o n p a i r ; inum = l i s t >inum ; i l i s t = l i s t > i l i s t ; numneigh = l i s t >numneigh ; f i r s t n e i g h = l i s t > f i r s t n e i g h ; / / l o o p o v e r f u l l n e i g h b o r l i s t o f my a t o m s f o r ( i i = 0 ; i i < inum ; i i ++) { i = ilist [ ii ]; itag = tag [ i ] ; i t y p e = map [ t y p e [ i ] ] ; xtmp = x [ i ] [ 0 ] ; ytmp = x [ i ] [ 1 ] ; ztmp = x [ i ] [ 2 ] ; / / twobody i n t e r a c t i o n s , s k i p h a l f o f them jlist = firstneigh [ i ]; jnum = numneigh [ i ] ; f o r ( j j = 0 ; j j < jnum ; j j ++) { j = jlist [ jj ]; jtag = tag [ j ] ; if ( itag > jtag ) { i f ( ( i t a g + j t a g ) % 2 == 0 ) c o n t i n u e ; } else if ( itag < jtag ) { i f ( ( i t a g + j t a g ) % 2 == 1 ) c o n t i n u e ; } else { i f ( x [ j ] [ 2 ] < ztmp ) c o n t i n u e ; e l s e i f ( x [ j ] [ 2 ] == ztmp && x [ j ] [ 1 ] < ytmp ) c o n t i n u e ; e l s e i f ( x [ j ] [ 2 ] == ztmp && x [ j ] [ 1 ] == ytmp && x [ j ] [ 0 ] < xtmp ) continue ; } j t y p e = map [ t y p e [ j ] ] ; d e l x = xtmp x [ j ] [ 0 ] ; d e l y = ytmp x [ j ] [ 1 ] ; d e l z = ztmp x [ j ] [ 2 ] ; rsq = delx delx + dely dely + delz delz ; i j p a r a m = elem2param [ i t y p e ] [ j t y p e ] [ j t y p e ] ; i f ( r s q > params [ i j p a r a m ] . c u t s q ) continue ; twobody (& p a r a m s [ i j p a r a m ] , r s q , f p a i r , e f l a g , evdwl ) ;

28

f[i f[i f[i f[j f[j f[j

][0] ][1] ][2] ][0] ][1] ][2]

+= += += = = =

delx f p a i r dely f p a i r delz fpair delx f p a i r dely f p a i r delz fpair

; ; ; ; ; ;

i f ( evflag ) e v t a l l y ( i , j , nlocal , newton pair , evdwl , 0 . 0 , f p a i r , d e l x , d e l y , d e l z ) ; } jnumm1 = jnum 1 ; f o r ( j j = 0 ; j j < jnumm1 ; j j ++) { j = jlist [ jj ]; j t y p e = map [ t y p e [ j ] ] ; i j p a r a m = elem2param [ i t y p e ] [ j t y p e ] [ j t y p e ] ; d e l r 1 [ 0 ] = x [ j ] [ 0 ] xtmp ; d e l r 1 [ 1 ] = x [ j ] [ 1 ] ytmp ; d e l r 1 [ 2 ] = x [ j ] [ 2 ] ztmp ; rsq1 = delr1 [0] delr1 [0] + delr1 [1] delr1 [1] + delr1 [2] delr1 [ 2 ] ; i f ( r s q 1 > params [ i j p a r a m ] . c u t s q ) continue ; f o r ( kk = j j + 1 ; kk < jnum ; kk ++) { k = j l i s t [ kk ] ; k t y p e = map [ t y p e [ k ] ] ; i k p a r a m = elem2param [ i t y p e ] [ k t y p e ] [ k t y p e ] ; i j k p a r a m = elem2param [ i t y p e ] [ j t y p e ] [ k t y p e ] ; d e l r 2 [ 0 ] = x [ k ] [ 0 ] xtmp ; d e l r 2 [ 1 ] = x [ k ] [ 1 ] ytmp ; d e l r 2 [ 2 ] = x [ k ] [ 2 ] ztmp ; rsq2 = delr2 [0] delr2 [0] + delr2 [1] delr2 [1] + delr2 [2] delr2 [2]; i f ( r s q 2 > params [ ikparam ] . c u t s q ) continue ; t h r e e b o d y (& p a r a m s [ i j p a r a m ] ,& p a r a m s [ i k p a r a m ] ,& p a r a m s [ i j k p a r a m ] , r s q 1 , r s q 2 , d e l r 1 , d e l r 2 , f j , fk , e f l a g , evdwl ) ; f [ i ][0] f [ i ][1] f [ i ][2] f [ j ][0] f [ j ][1] f [ j ][2] f [k][0] f [k][1] f [k][2] = = = += += += += += += f j [ 0] + fk [ 0 ] ; f j [ 1] + fk [ 1 ] ; f j [ 2] + fk [ 2 ] ; fj [0]; fj [1]; fj [2]; fk [ 0 ] ; fk [ 1 ] ; fk [ 2 ] ;

i f ( e v f l a g ) e v t a l l y 3 ( i , j , k , evdwl , 0 . 0 , f j , fk , d e l r 1 , d e l r 2 ) ; } } }

29

if ( vflag fdotr ) virial compute () ; } / / v o i d PairMBMH : : a l l o c a t e ( ) { allocated = 1; i n t n = atom>n t y p e s ; s e t f l a g = memory> c r e a t e 2 d i n t a r r a y ( n +1 , n +1 , p a i r : s e t f l a g ) ; c u t s q = memory>c r e a t e 2 d d o u b l e a r r a y ( n +1 , n +1 , p a i r : c u t s q ) ; map = new i n t [ n + 1 ] ; } / global se t t i ng s / v o i d PairMBMH : : s e t t i n g s ( i n t n a r g , char a r g ) { i f ( n a r g ! = 0 ) e r r o r > a l l ( I l l e g a l p a i r s t y l e command ) ; } / s e t c o e f f s f o r one o r more t y p e p a i r s / v o i d PairMBMH : : c o e f f ( i n t n a r g , char a r g ) { int i , j , n ; if (! allocated ) allocate () ; i f ( n a r g ! = 3 + atom>n t y p e s ) e r r o r > a l l ( I n c o r r e c t a r g s f o r p a i r c o e f f i c i e n t s ) ; / / i n s u r e I , J s are i f ( strcmp ( arg [ 0 ] , ) != 0 | | strcmp ( arg [ 1 ] , ) != 0) e r r o r > a l l ( I n c o r r e c t a r g s f o r p a i r c o e f f i c i e n t s ) ; // // // // r e a d a r g s t h a t map atom t y p e s t o e l e m e n t s i n p o t e n t i a l f i l e map [ i ] = w h i c h e l e m e n t t h e I t h atom t y p e i s , 1 i f NULL nelements = # of unique elements e l e m e n t s = l i s t o f e l e m e n t names

i f ( elements ) { f o r ( i = 0 ; i < n e l e m e n t s ; i ++) d e l e t e [ ] e l e m e n t s [ i ] ; delete [] elements ; } e l e m e n t s = new char [ atom>n t y p e s ] ;

30

f o r ( i = 0 ; i < atom>n t y p e s ; i ++) e l e m e n t s [ i ] = NULL ; nelements = 0; f o r ( i = 3 ; i < n a r g ; i ++) { i f ( s t r c m p ( a r g [ i ] , NULL ) == 0 ) { map [ i 2] = 1; continue ; } f o r ( j = 0 ; j < n e l e m e n t s ; j ++) i f ( s t r c m p ( a r g [ i ] , e l e m e n t s [ j ] ) == 0 ) break ; map [ i 2] = j ; i f ( j == n e l e m e n t s ) { n = s t r l e n ( arg [ i ] ) + 1; e l e m e n t s [ j ] = new char [ n ] ; strcpy ( elements [ j ] , arg [ i ] ) ; n e l e m e n t s ++; } } / / read p o t e n t i a l r e a d f i l e ( arg [ 2 ] ) ; setup () ; / / c l e a r s e t f l a g s i n c e c o e f f ( ) c a l l e d once w i t h I , J = n = atom>n t y p e s ; f o r ( i n t i = 1 ; i <= n ; i ++) f o r ( i n t j = i ; j <= n ; j ++) s e t f l a g [ i ][ j ] = 0; / / s e t s e t f l a g i , j f o r t y p e p a i r s where b o t h a r e mapped t o e l e m e n t s int count = 0; f o r ( i n t i = 1 ; i <= n ; i ++) f o r ( i n t j = i ; j <= n ; j ++) i f ( map [ i ] >= 0 && map [ j ] >= 0 ) { s e t f l a g [ i ][ j ] = 1; c o u n t ++; } i f ( c o u n t == 0 ) e r r o r > a l l ( I n c o r r e c t a r g s f o r p a i r c o e f f i c i e n t s ) ; } / i n i t s p e c i f i c to t h i s pair s t y l e / v o i d PairMBMH : : i n i t s t y l e ( ) { i f ( atom>t a g e n a b l e == 0 ) e r r o r > a l l ( P a i r s t y l e S t i l l i n g e r Weber r e q u i r e s atom IDs ) ; i f ( f o r c e >n e w t o n p a i r == 0 ) e r r o r > a l l ( P a i r s t y l e S t i l l i n g e r Weber r e q u i r e s newton p a i r on ) ; f i l e and i n i t i a l i z e p o t e n t i a l parameters

31

/ / need a f u l l neighbor l i s t i n t i r e q u e s t = n e i g h b o r >r e q u e s t ( t h i s ) ; n e i g h b o r >r e q u e s t s [ i r e q u e s t ]> h a l f = 0 ; n e i g h b o r >r e q u e s t s [ i r e q u e s t ]> f u l l = 1 ; } / i n i t f o r one t y p e p a i r i , j and c o r r e s p o n d i n g j , i / d o u b l e PairMBMH : : i n i t o n e ( i n t i , i n t j ) { i f ( s e t f l a g [ i ] [ j ] == 0 ) e r r o r > a l l ( A l l p a i r c o e f f s a r e n o t s e t ) ; r e t u r n cutmax ; } / / v o i d PairMBMH : : r e a d f i l e ( char f i l e ) { int params per line = 14; char words = new char [ p a r a m s p e r l i n e + 1 ] ; memory>s f r e e ( p a r a m s ) ; p a r a m s = NULL ; nparams = 0 ; / / open f i l e on p r o c 0 FILE f p ; i f ( comm>me == 0 ) { fp = fopen ( f i l e , r ) ; i f ( f p == NULL) { char s t r [ 1 2 8 ] ; s p r i n t f ( s t r , C a n n o t open S t i l l i n g e r Weber p o t e n t i a l ; e r r o r >one ( s t r ) ; } }

f i l e %s , f i l e )

/ / r e a d e a c h s e t o f params f r o m p o t e n t i a l f i l e / / one s e t o f params can s p a n m u l t i p l e l i n e s / / s t o r e params i f a l l 3 e l e m e n t t a g s a r e i n e l e m e n t l i s t i n t n , nwords , i e l e m e n t , j e l e m e n t , k e l e m e n t ; char l i n e [MAXLINE] , p t r ; int eof = 0; while ( 1 ) { i f ( comm>me == 0 ) {

32

p t r = f g e t s ( l i n e , MAXLINE, f p ) ; i f ( p t r == NULL) { eof = 1; f c l o s e ( fp ) ; } else n = s t r l e n ( line ) + 1; } M P I B c a s t (& e o f , 1 , MPI INT , 0 , w o r l d ) ; i f ( e o f ) break ; M P I B c a s t (&n , 1 , MPI INT , 0 , w o r l d ) ; M P I B c a s t ( l i n e , n , MPI CHAR , 0 , w o r l d ) ; // s t r i p comment , s k i p l i n e i f b l a n k

i f ( p t r = s t r c h r ( l i n e , # ) ) p t r = \0 ; nwords = atom>c o u n t w o r d s ( l i n e ) ; i f ( nwords == 0 ) c o n t i n u e ; / / c o n c a t e n a t e a d d i t i o n a l l i n e s u n t i l h a v e p a r a m s p e r l i n e words w h i l e ( nwords < p a r a m s p e r l i n e ) { n = strlen ( line ) ; i f ( comm>me == 0 ) { p t r = f g e t s (& l i n e [ n ] , MAXLINEn , f p ) ; i f ( p t r == NULL) { eof = 1; f c l o s e ( fp ) ; } else n = s t r l e n ( line ) + 1; } M P I B c a s t (& e o f , 1 , MPI INT , 0 , w o r l d ) ; i f ( e o f ) break ; M P I B c a s t (&n , 1 , MPI INT , 0 , w o r l d ) ; M P I B c a s t ( l i n e , n , MPI CHAR , 0 , w o r l d ) ; i f ( p t r = s t r c h r ( l i n e , # ) ) p t r = \0 ; nwords = atom>c o u n t w o r d s ( l i n e ) ; } i f ( nwords ! = p a r a m s p e r l i n e ) e r r o r > a l l ( I n c o r r e c t f o r m a t i n S t i l l i n g e r Weber p o t e n t i a l / / words = p t r s t o a l l words i n l i n e nwords = 0 ; words [ nwords ++] = s t r t o k ( l i n e , \ t \n \ r \ f ) ; w h i l e ( words [ nwords ++] = s t r t o k (NULL, \ t \ n\ r \ f ) ) c o n t i n u e ; / / ielement , jelement , kelement = 1 s t args / / i f a l l 3 args are in element l i s t , then parse t h i s l i n e / / else skip to next entry in f i l e for ( ielement = 0; ielement < nelements ; i f ( s t r c m p ( words [ 0 ] , e l e m e n t s [ i e l e m e n t ] ) i f ( i e l e m e n t == n e l e m e n t s ) c o n t i n u e ; for ( jelement = 0; jelement < nelements ; i f ( s t r c m p ( words [ 1 ] , e l e m e n t s [ j e l e m e n t ] ) i e l e m e n t ++) == 0 ) break ; j e l e m e n t ++) == 0 ) break ;

file ) ;

33

i f ( j e l e m e n t == n e l e m e n t s ) c o n t i n u e ; f o r ( k e l e m e n t = 0 ; k e l e m e n t < n e l e m e n t s ; k e l e m e n t ++) i f ( s t r c m p ( words [ 2 ] , e l e m e n t s [ k e l e m e n t ] ) == 0 ) break ; i f ( k e l e m e n t == n e l e m e n t s ) c o n t i n u e ; / / l o a d up p a r a m e t e r s e t t i n g s and e r r o r c h e c k t h e i r v a l u e s i f ( n p a r a m s == maxparam ) { maxparam += DELTA ; p a r a m s = ( Param ) memory>s r e a l l o c ( params , maxparam s i z e o f ( Param ) , p a i r : params ) ; } params [ nparams ] . i e l e m e n t = i e l e m e n t ; params [ nparams ] . j e l e m e n t = j e l e m e n t ; params [ nparams ] . kelement = kelement ; p a r a m s [ n p a r a m s ] . e p s i l o n = a t o f ( words [ 3 ] ) ; p a r a m s [ n p a r a m s ] . s i g m a = a t o f ( words [ 4 ] ) ; p a r a m s [ n p a r a m s ] . l i t t l e a = a t o f ( words [ 5 ] ) ; p a r a m s [ n p a r a m s ] . lambda = a t o f ( words [ 6 ] ) ; p a r a m s [ n p a r a m s ] . gamma = a t o f ( words [ 7 ] ) ; p a r a m s [ n p a r a m s ] . c o s t h e t a = a t o f ( words [ 8 ] ) ; p a r a m s [ n p a r a m s ] . b i g a = a t o f ( words [ 9 ] ) ; p a r a m s [ n p a r a m s ] . b i g b = a t o f ( words [ 1 0 ] ) ; / / params [ nparams ] . powerp = a t o f ( words [ 1 1 ] ) ; p a r a m s [ n p a r a m s ] . q i = a t o f ( words [ 1 1 ] ) ; / / params [ nparams ] . powerq = a t o f ( words [ 1 2 ] ) ; p a r a m s [ n p a r a m s ] . q j = a t o f ( words [ 1 2 ] ) ; p a r a m s [ n p a r a m s ] . t o l = a t o f ( words [ 1 3 ] ) ; i f ( params [ nparams ] . e p s i l o n < 0 . 0 | | params [ nparams ] . sigma < 0 . 0 | | p a r a m s [ n p a r a m s ] . l i t t l e a < 0 . 0 | | p a r a m s [ n p a r a m s ] . lambda < 0 . 0 | | p a r a m s [ n p a r a m s ] . gamma < 0 . 0 | | p a r a m s [ n p a r a m s ] . b i g a < 0 . 0 | | / / params [ nparams ] . b i g b < 0 . 0 | | params [ nparams ] . powerp < 0 . 0 | | params [ nparams ] . bigb < 0 . 0 | | / / params [ nparams ] . q i < 0 . 0 | | i m i g h t want a neg v a l u e / / params [ nparams ] . powerq < 0 . 0 | | params [ nparams ] . t o l < 0 . 0 ) / / params [ nparams ] . q j < 0 . 0 | | i m i g h t want a neg v a l u e params [ nparams ] . t o l < 0 . 0 ) e r r o r > a l l ( I l l e g a l S t i l l i n g e r Weber p a r a m e t e r ) ; n p a r a m s ++; } d e l e t e [ ] words ; } / / v o i d PairMBMH : : s e t u p ( ) { i n t i , j , k , m, n ; double rtmp ;

34

/ / s e t elem2param f o r a l l t r i p l e t c o m b i n a t i o n s / / m u s t be a s i n g l e e x a c t match t o l i n e s r e a d f r o m f i l e / / do n o t a l l o w f o r ACB i n p l a c e o f ABC i f ( elem2param ) memory>d e s t r o y 3 d i n t a r r a y ( elem2param ) ; elem2param = memory> c r e a t e 3 d i n t a r r a y ( n e l e m e n t s , n e l e m e n t s , n e l e m e n t s , p a i r : elem2param ) ; f o r ( i = 0 ; i < n e l e m e n t s ; i ++) f o r ( j = 0 ; j < n e l e m e n t s ; j ++) f o r ( k = 0 ; k < n e l e m e n t s ; k ++) { n = 1; f o r (m = 0 ; m < n p a r a m s ; m++) { i f ( i == p a r a m s [m ] . i e l e m e n t && j == p a r a m s [m ] . j e l e m e n t && k == p a r a m s [m ] . k e l e m e n t ) { i f ( n >= 0 ) e r r o r > a l l ( P o t e n t i a l f i l e h a s d u p l i c a t e e n t r y ) ; n = m; } } i f ( n < 0 ) e r r o r > a l l ( P o t e n t i a l f i l e i s m i s s i n g an e n t r y ) ; elem2param [ i ] [ j ] [ k ] = n ; }

/ / compute parameter v a l u e s d e r i v e d from i n p u t s / / s e t cutsq using s h o r t c u t to reduce neighbor l i s t for accelerated / / c a l c u l a t i o n s . c u t must remain unchanged as i t i s a p o t e n t i a l parameter / / ( c u t = a sigma ) f o r (m = 0 ; m < n p a r a m s ; m++) { p a r a m s [m ] . c u t = p a r a m s [m ] . s i g m a p a r a m s [m ] . l i t t l e a ; r t m p = p a r a m s [m ] . c u t ; i f ( p a r a m s [m ] . t o l > 0 . 0 ) { i f ( p a r a m s [m ] . t o l > 0 . 0 1 ) p a r a m s [m ] . t o l = 0 . 0 1 ; i f ( p a r a m s [m ] . gamma < 1 . 0 ) rtmp = rtmp + p a r a m s [m ] . gamma p a r a m s [m ] . s i g m a / l o g ( p a r a m s [m ] . t o l ) ; e l s e rtmp = rtmp + p a r a m s [m ] . s i g m a / l o g ( p a r a m s [m ] . t o l ) ; } p a r a m s [m ] . c u t s q = r t m p r t m p ; p a r a m s [m ] . sigma gamma = p a r a m s [m ] . s i g m a p a r a m s [m ] . gamma ; p a r a m s [m ] . l a m b d a e p s i l o n = p a r a m s [m ] . lambda p a r a m s [m ] . e p s i l o n ; / / params [m ] . l a m b d a e p s i l o n 2 = 2 . 0 params [m ] . lambda params [m ] . e p s i l o n ; p a r a m s [m ] . e r f c = 1 4 . 3 9 9 6 5 1 7 3 p a r a m s [m ] . q i p a r a m s [m ] . q j ; p a r a m s [m ] . d e l e r f c = 1 6 . 2 4 8 2 6 7 p a r a m s [m ] . q i p a r a m s [m ] . q j / p a r a m s [m ] . b i g b

35

; p a r a m s [m ] . d e l e x p = 3 . 4 4 8 2 7 5 8 6 2 p a r a m s [m ] . b i g a ; / params [m ] . c1 = params [m ] . b i g a params [m ] . e p s i l o n params [m ] . powerp params [m ] . b i g b pow ( params [m ] . sigma , params [m ] . powerp ) ; params [m ] . c2 = params [m ] . b i g a params [m ] . e p s i l o n params [m ] . powerq pow ( params [m ] . sigma , params [m ] . powerq ) ; params [m ] . c3 = params [m ] . b i g a params [m ] . e p s i l o n params [m ] . b i g b pow ( params [m ] . sigma , params [m ] . powerp + 1 . 0 ) ; params [m ] . c4 = params [m ] . b i g a params [m ] . e p s i l o n wont n e e d s o many c s pow ( params [m ] . sigma , params [m ] . powerq + 1 . 0 ) ; params [m ] . c5 = params [m ] . b i g a params [m ] . e p s i l o n params [m ] . b i g b pow ( params [m ] . sigma , params [m ] . powerp ) ; params [m ] . c6 = params [m ] . b i g a params [m ] . e p s i l o n pow ( params [m ] . sigma , params [m ] . powerq ) ; / } / / s e t c u t m a x t o max o f a l l params cutmax = 0 . 0 ; f o r (m = 0 ; m < n p a r a m s ; m++) { r t m p = s q r t ( p a r a m s [m ] . c u t s q ) ; i f ( r t m p > cutmax ) cutmax = r t m p ; } } / / v o i d PairMBMH : : twobody ( Param param , d o u b l e r s q , d o u b l e &f f o r c e , i n t e f l a g , d o u b l e &eng ) { / / d o u b l e r , r i n v s q , rp , rq , r a i n v , r a i n v s q , e x p s r a i n v ; / o r i g i n a l code r = sqrt ( rsq ) ; rinvsq = 1 . 0 / rsq ; r p = pow ( r ,param>powerp ) ; r q = pow ( r ,param>powerq ) ; r a i n v = 1 . 0 / ( r param>c u t ) ; r a i n v s q = r a i n v r a i n v r ; / / why t i m e s r ? e x p s r a i n v = e x p ( param>s i g m a r a i n v ) ; f f o r c e = ( param>c1 r p param>c2 r q + ( param>c3 r p param>c4 r q ) r a i n v s q ) e x p s r a i n v r i n v s q ; i f ( e f l a g ) eng = ( param>c5 r p param>c6 r q ) e x p s r a i n v ; / / / V = ( 1 4 . 3 9 9 6 5 1 7 3 q1 q2 e r f c ( r / B ) / r+A e x p ( r / ( . 2 9 ) ) where A and B a r e f a c t o r s ; and q i s a c h a r g e / /dV / d r =14.39965173 q1 q2 e r f c ( r / B ) / r 2 + 2 8 . 7 9 9 3 0 3 4 6 q1 q2 e x p ( r 2 / B 2 ) / ( r s q r t ( P i ) B ) +3.448275862A e x p ( 3.448275862 r ) / below i s d i s a b l e d m o d i f i e d code f o r s e p a r a t i o n double r ; double paramerfcrbyB , expnegrbypt29 ;

36

r = sqrt ( rsq ) ; p a r a m e r f c r b y B = param>e r f c e r f c ( r / param>b i g b ) ; e x p n e g r b y p t 2 9=e x p ( 3.448275862 r ) ; / f f o r c e = 0 ; / / ( 1 . 0 / r ) ( p a r a m e r f c r b y B / r s q+param>d e l e r f c ( e x p ( r s q / ( param >b i g b param>b i g b ) ) ) / r+param>d e l e x p e x p n e g r b y p t 2 9 ) ; / / e x t r a r b / c i t s x i e d by l e n g t h v e c t o r i n c o m p u t e ( ) i f ( e f l a g ) eng = 0 ; / / ( p a r a m e r f c r b y B ) / r + param>b i g a e x p n e g r b y p t 2 9 ; } / / v o i d PairMBMH : : t h r e e b o d y ( Param p a r a m i j , Param p a r a m i k , Param p a r a m i j k , double rsq1 , double rsq2 , double d e l r 1 , double d e l r 2 , d o u b l e f j , d o u b l e fk , i n t e f l a g , d o u b l e &eng ) { d o u b l e r1 , r i n v s q 1 , r a i n v 1 , g s r a i n v 1 , g s r a i n v s q 1 ; / / , e x p g s r a i n v 1 ; d o u b l e r2 , r i n v s q 2 , r a i n v 2 , g s r a i n v 2 , g s r a i n v s q 2 ; / / , e x p g s r a i n v 2 ; d o u b l e r i n v 1 2 , cs , d e l c s , d e l c s s q , expexp , l a m b d a e p s i l o n e x p e x p ; d o u b l e c s s q , s n s q , c s c b , cspwr4 , s n c s s q , d e l s q x s n c s s q ; / / added double gsterm1 , gsterm2 , cs3term , cs3term1 , cs3term2 , cs4term , cs4term1 , cs4term2 ; r1 = s q r t ( rsq1 ) ; rinvsq1 = 1.0/ rsq1 ; r a i n v 1 = 1 . 0 / ( r 1 p a r a m i j >c u t ) ; g s r a i n v 1 = p a r a m i j >sigma gamma r a i n v 1 ; g s r a i n v s q 1 = g s r a i n v 1 r a i n v 1 ; / / / r1 ; / / e x p g s r a i n v 1 = exp ( g s r a i n v 1 ) ; r2 = s q r t ( rsq2 ) ; rinvsq2 = 1.0/ rsq2 ; r a i n v 2 = 1 . 0 / ( r 2 p a r a m i k >c u t ) ; g s r a i n v 2 = p a r a m i k >sigma gamma r a i n v 2 ; g s r a i n v s q 2 = g s r a i n v 2 r a i n v 2 ; / / / r2 ; / / e x p g s r a i n v 2 = exp ( g s r a i n v 2 ) ; rinv12 = 1 . 0 / ( r1 r2 ) ; cs = ( delr1 [0] delr2 [ 0] + delr1 [1] delr2 [ 1] + delr1 [2] delr2 [ 2 ] ) rinv12 ; cssq= cs cs ; s n s q =1.0 c s s q ; cscb=cssq cs ; cspwr4= cscb cs ; sncssq= cssq snsq ; d e l c s = c s p a r a m i j k >c o s t h e t a ; delcssq = delcs delcs ; delsqxsncssq=delcssq sncssq ; l a m b d a e p s i l o n e x p e x p = p a r a m i j k >l a m b d a e p s i l o n exp ( g s r a i n v 1 + g s r a i n v 2 ) ; / / e x p g s r a i n v 1 e x p g s r a i n v 2 ; / / m a j i d : why n o t e x p ( a+b ) ? gsterm1 = delsqxsncssq gsrainvsq1 ; gsterm2= delsqxsncssq gsrainvsq2 ;

37

c s 4 t e r m = 2 . 0 ( ( d e l c s c s c b + d e l c s s q c s s q ) s n s q d e l c s s q c s p w r 4 ) ; c s 4 t e r m 1 =cs4term / r1 ; cs4term2=cs4term / r2 ; c s 3 t e r m = 2 . 0 ( ( d e l c s c s s q d e l c s s q c s ) s n s q + d e l c s s q c s c b ) lambdaepsilonexpexp ; cs3term1=cs3term / r1 ; cs3term2=cs3term / r2 ; // f (v) function . ( d e l c s 2 c s ( v ) 2 s n 2 g s / ( r 1 ( v )a s ) 2 + ( ( 2 d e l c s c s ( v ) 3+2 d e l c s 2 c s ( v ) 2 ) s n 2 2 d e l c s 2 c s ( v ) 4 ) / r 1 ) LE e x p 2 V r 1 h a t +(( 2 d e l c s c s ( v ) 2 2 d e l c s 2 c s ( v ) ) s n 2+2 d e l c s 2 c s ( v ) 3 ) LE e x p 2 Vr2hat / r1 f j [ 0] = ( gsterm1+cs4term1 ) lambdaepsilonexpexp delr1 [ 0 ] / r1+cs3term1 delr2 [ 0 ] / r2 ; f j [ 1] = ( gsterm1+cs4term1 ) lambdaepsilonexpexp delr1 [ 1 ] / r1+cs3term1 delr2 [ 1 ] / r2 ; f j [ 2] = ( gsterm1+cs4term1 ) lambdaepsilonexpexp delr1 [ 2 ] / r1+cs3term1 delr2 [ 2 ] / r2 ; fk [ 0] = ( gsterm2+cs4term2 ) lambdaepsilonexpexp delr2 [ 0 ] / r2+cs3term2 delr1 [ 0 ] / r1 ; fk [ 1] = ( gsterm2+cs4term2 ) lambdaepsilonexpexp delr2 [ 1 ] / r2+cs3term2 delr1 [ 1 ] / r1 ; fk [ 2] = ( gsterm2+cs4term2 ) lambdaepsilonexpexp delr2 [ 2 ] / r2+cs3term2 delr1 [ 2 ] / r1 ; i f ( e f l a g ) eng = l a m b d a e p s i l o n e x p e x p d e l s q x s n c s s q ; }

38

LAMMPS NON-EQUILIBRIUM THERMAL CONDUCTIVITY CODE Simulation Setup Script This script sets up the simulation.
label inits dimension 3 u n i t s metal boundary p p p newton on atom style charge v a r i a b l e a e q u a l ( ${Tk } 9 . 6 7 1 1 4 3 4 9 e 05) + 1 2 . 3 9 1 8 # geometry # p r i m i t i v e v e c s a1 .5 . 5 . 5 a2 . 5 .5 . 5 a3 . 5 . 5 .5 & # c o n v e n t i o n a l a1 1 0 0 a2 0 1 0 a3 0 0 1 & # b a s i s v e c s b e l o w , A1 ( 1 6 ) , A2 ( 2 4 ) ,O( 9 6 ) , Y ( 2 4 ) i n o r d e r , f o r c o n v e n t i o n a l #ERROR I n p u t l i n e t o o l o n g ! ! c h a n g e d i n p u t . cpp MAXLINES l a t t i c e c u s t o m $a & basis 0.0 0.0 0.0 & basis 0.5 0 0.5 & basis 0 0.5 0.5 & basis 0.5 0.5 0 & basis 0.75 0.25 0.25 & basis 0.75 0.75 0.75 & basis 0.25 0.25 0.75 & basis 0.25 0.75 0.25 & basis 0.25 0.75 0.75 & basis 0.25 0.25 0.25 & basis 0.75 0.75 0.25 & basis 0.75 0.25 0.75 & basis 0.5 0.5 0.5 & basis 0 0.5 0 & basis 0.5 0 0 & basis 0 0 0.5 & basis 0.375 0 0.25 & basis 0.125 0 0.75 & basis 0.625 0.5 0.25 & basis 0.875 0.5 0.75 & basis 0.25 0.375 0 & basis 0.75 0.125 0 & basis 0.25 0.625 0.5 & basis 0.75 0.875 0.5 & basis 0 0.25 0.375 & basis 0 0.75 0.125 & basis 0.5 0.25 0.625 & basis 0.5 0.75 0.875 & basis 0.75 0.625 0 & basis 0.75 0.375 0.5 & basis 0.25 0.875 0 & basis 0.25 0.125 0.5 & basis 0.125 0.5 0.25 & basis 0.875 0 0.25 & basis 0.375 0.5 0.75 & basis 0.625 0 0.75 &

39

basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis

0 0.5 0 0.5 0.28023 0.21977 0.71977 0.78023 0.19922 0.69922 0.30078 0.80078 0.1011 0.8989 0.6011 0.3989 0.8511 0.6489 0.3511 0.1489 0.03023 0.96977 0.46977 0.53023 0.94922 0.44922 0.05078 0.55078 0.71977 0.78023 0.28023 0.21977 0.80078 0.30078 0.69922 0.19922 0.8989 0.1011 0.3989 0.6011 0.1489 0.3511 0.6489 0.8511 0.96977 0.03023 0.53023 0.46977 0.05078 0.55078 0.94922 0.44922 0.78023 0.71977

0.25 0.25 0.75 0.75 0.1011 0.8989 0.6011 0.3989 0.28023 0.21977 0.71977 0.78023 0.19922 0.69922 0.30078 0.80078 0.53023 0.46977 0.96977 0.03023 0.44922 0.94922 0.55078 0.05078 0.3511 0.1489 0.8511 0.6489 0.8989 0.1011 0.3989 0.6011 0.71977 0.78023 0.28023 0.21977 0.80078 0.30078 0.69922 0.19922 0.46977 0.53023 0.03023 0.96977 0.55078 0.05078 0.44922 0.94922 0.6489 0.8511 0.1489 0.3511 0.6011 0.3989

0.875 0.125 0.625 0.375 0.19922 0.69922 0.30078 0.80078 0.1011 0.8989 0.6011 0.3989 0.28023 0.21977 0.71977 0.78023 0.05078 0.55078 0.94922 0.44922 0.1489 0.3511 0.6489 0.8511 0.96977 0.03023 0.53023 0.46977 0.80078 0.30078 0.69922 0.19922 0.8989 0.1011 0.3989 0.6011 0.71977 0.78023 0.28023 0.21977 0.94922 0.44922 0.05078 0.55078 0.8511 0.6489 0.3511 0.1489 0.03023 0.96977 0.46977 0.53023 0.69922 0.19922

& & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & &

40

basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis

0.21977 0.28023 0.69922 0.19922 0.80078 0.30078 0.6011 0.3989 0.1011 0.8989 0.3511 0.1489 0.8511 0.6489 0.53023 0.46977 0.96977 0.03023 0.44922 0.94922 0.55078 0.05078 0.21977 0.28023 0.78023 0.71977 0.30078 0.80078 0.19922 0.69922 0.3989 0.6011 0.8989 0.1011 0.6489 0.8511 0.1489 0.3511 0.46977 0.53023 0.03023 0.96977 0.55078 0.05078 0.44922 0.94922 0.125 0.375 0.875 0.625 0.25 0.75 0.25 0.75

0.1011 0.8989 0.78023 0.71977 0.21977 0.28023 0.69922 0.19922 0.80078 0.30078 0.03023 0.96977 0.46977 0.53023 0.94922 0.44922 0.05078 0.55078 0.8511 0.6489 0.3511 0.1489 0.3989 0.6011 0.8989 0.1011 0.21977 0.28023 0.78023 0.71977 0.30078 0.80078 0.19922 0.69922 0.96977 0.03023 0.53023 0.46977 0.05078 0.55078 0.94922 0.44922 0.1489 0.3511 0.6489 0.8511 0 0 0.5 0.5 0.125 0.375 0.875 0.625

0.80078 0.30078 0.6011 0.3989 0.1011 0.8989 0.78023 0.71977 0.21977 0.28023 0.55078 0.05078 0.44922 0.94922 0.6489 0.8511 0.1489 0.3511 0.46977 0.53023 0.03023 0.96977 0.30078 0.80078 0.19922 0.69922 0.3989 0.6011 0.8989 0.1011 0.21977 0.28023 0.78023 0.71977 0.44922 0.94922 0.55078 0.05078 0.3511 0.1489 0.8511 0.6489 0.53023 0.46977 0.96977 0.03023 0.25 0.75 0.25 0.75 0 0 0.5 0.5

& & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & &

41

basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis

0 0 0.5 0.5 0.875 0.625 0.125 0.375 0.75 0.25 0.75 0.25 0 0 0.5 0.5

0.25 0.75 0.25 0.75 0 0 0.5 0.5 0.875 0.625 0.125 0.375 0.75 0.25 0.75 0.25

0.125 0.375 0.875 0.625 0.75 0.25 0.75 0.25 0 0 0.5 0.5 0.875 0.625 0.125 0.375

& & & & & & & & & & & & & & & &

region wholething block 0 1 0 1 0 1 u n i t s l a t t i c e c r e a t e b o x 4 w h o l e t h i n g # no . i s n t y p e s # a f t e r b a s i s kw 1 s t no . i s b a s i s atm , 2 nd i s t y p e create atoms 1 box basis 1 1 # c r e a t e s 160 a t o m s a t once # . . . w i l l need t o spec t y p e s g r o u p Al1 i d <> 1 16 g r o u p Al2 i d <> 17 40 g r o u p Al u n i o n Al1 Al2 g r o u p O i d <> 41 136 g r o u p Y i d <> 137 160 s e t g r o u p Al t y p e 1 s e t group O type 2 s e t group Y type 3

# Atom p r o p e r t i e s mass 1 2 6 . 9 8 # A l mass 2 1 5 . 9 9 9 #O mass 3 8 8 . 9 0 6 #Y mass 4 6 9 . 7 # Dopant Ga f o r A l # can add a d o p a n t a s mass 4 s e t g r o u p O c h a r g e 2.0 #O ( 2) s e t g r o u p Al c h a r g e 3 . 0 # A l ( + 3 ) . . . t h e dopedf o r atom s e t g r o u p Y c h a r g e 3 . 0 #Y ( + 3 ) i n c l u d e r u n r e q s . lmpin # looking variable variable variable variable variable variable a t one s i d e o f s i m box dx e q u a l $dx hsx e q u a l $hsx # h e a t e r or s i n k width handsw e q u a l ${ h s x }2 ly equal $ly lz equal $lz h a l f s i m b o x w i d t h e q u a l ${ dx }+$ { handsw }

42

v a r i a b l e s i m b o x w i d t h e q u a l 2 $ { h a l f s i m b o x w i d t h } # c r e a t e s l a b t o be e x t r u d e d r e p l i c a t e 1 $ly $lz r e p l i c a t e ${ simboxwidth } 1 1

# random s u b s t i t u t i o n s e t g r o u p Al t y p e / f r a c t i o n 4 $dp $ d s e e d g r o u p Ga t y p e 4 s e t g r o u p Ga c h a r g e 3 . 0 #Ga

# |0 s i n k n1dxn2h e a t e r n3h e a t e r n4dxn5s i n k n6 | v a r i a b l e p a r t g index l s i n k ldx l h e a t e r r h e a t e r rdx r s i n k v a r i a b l e p a r t l i n d e x $ h s x $dx $ h s x $ h s x $dx $ h s x v a r i a b l e s l a b i e q u a l 160 $ l y $ l z variable s t a r t i d equal 1 label extrude v a r i a b l e e n d i d e q u a l ${ s t a r t i d }+$ { s l a b i } ${ p a r t l }1 g r o u p ${ p a r t g } i d <> $ { s t a r t i d } $ { e n d i d } v a r i a b l e s t a r t i d e q u a l ${ e n d i d }+1 next partg next p a r t l jump s e t u p . l m p i n e x t r u d e group s i n k s union l s i n k r s i n k group h e a t e r s union l h e a t e r r h e a t e r group grads union ldx rdx

variable variable velocity variable variable variable velocity velocity

T e q u a l $Tk TC e q u a l $Tk$dT / 2 a l l c r e a t e ${TC} $ v s e e d d i s t g a u s s i a n TH e q u a l $Tk+$dT / 2 d T o v e r d x e q u a l $dT / $dx dT e q u a l $dT s i n k s s c a l e ${TC} h e a t e r s s c a l e ${TH}

v a r i a b l e nGa e q u a l c o u n t ( Ga ) # l e f t side v a r i a b l e xmn e q u a l bound ( l d x , xmin ) v a r i a b l e xmx e q u a l bound ( l d x , xmax ) i f $ {nGa} == 0 t h e n & v a r i a b l e g r o u p i i n d e x Al O Y & v a r i a b l e t p g i n d e x l A l lO lY e l s e & v a r i a b l e g r o u p i i n d e x Al O Y Ga & v a r i a b l e t p g i n d e x l A l lO lY lGa label tpl g r o u p ${ t p g } i n t e r s e c t l d x ${ g r o u p i } v a r i a b l e m index 26.98 15.99 88.906 69.7 v a r i a b l e VEhi e q u a l 1 . 5 7 9 3 4 7 7 7 4 ($m$ {dT } ) ( 1 / 2 ) / $m # 1.579347774($m

43

3 0 0 ) ( 1 / 2 ) / $m v e l o c i t y ${ t p g } ramp vy 0 $ { VEhi } x $ {xmn} $ {xmx} sum y e s u n i t s box next tpg next groupi next m jump s e t u p . l m p i n t p l v a r i a b l e tpg d e l e t e variable groupi delete variable m delete # right side v a r i a b l e xmn e q u a l bound ( rdx , xmin ) v a r i a b l e xmx e q u a l bound ( rdx , xmax ) i f $ {nGa} == 0 t h e n & v a r i a b l e g r o u p i i n d e x Al O Y & v a r i a b l e t p g i n d e x r A l rO rY e l s e & v a r i a b l e g r o u p i i n d e x Al O Y Ga & v a r i a b l e t p g i n d e x r A l rO rY rGa label tpr g r o u p ${ t p g } i n t e r s e c t r d x ${ g r o u p i } v a r i a b l e m index 26.98 15.99 88.906 69.7 v a r i a b l e VEhi e q u a l 1 . 5 7 9 3 4 7 7 7 4 ($m$ {dT } ) ( 1 / 2 ) / $m# 1.579347774($m 3 0 0 ) ( 1 / 2 ) / $m v e l o c i t y ${ t p g } ramp vy 0 $ { VEhi } x $ {xmx} $ {xmn} sum y e s u n i t s box next tpg next groupi next m jump s e t u p . l m p i n t p r v a r i a b l e tpg d e l e t e variable groupi delete variable m delete reset timestep 0 write restart r . restart

Interatomic Potentials runreqs.lmpin


timestep .001 k s p a c e s t y l e pppm . 0 0 0 0 1 # u s u a l l y a t . 0 0 0 1 b u t n e e d some s p e e d , . 0 0 0 0 1 conserves but slow # p o t e n t i a l s block p a i r s t y l e h y b r i d / o v e r l a y b o r n / c o u l / l o n g 4 . 0 7 . 0 mbmh mbmh2 # #2 body params #w / o Ga p a i r c o e f f 2 2 born / coul / long 2449.44 0.2907 0 0 0 p a i r c o e f f 1 2 born / coul / long 1740.31 0.2907 0 0 0 p a i r c o e f f 2 3 born / coul / long 1250.85 0.3497 0 0 0 p a i r c o e f f 1 1 born / coul / long 312.11 0.071124 0 0 0 p a i r c o e f f 3 3 born / coul / long 245.14 0.071124 0 0 0 p a i r c o e f f 1 3 born / coul / long 256.55 0.071124 0 0 0 #w Ga , r e p l a c e d t h e 1 s w 4 s ( b u t s m a l l e r no . h a s t o be 1 s t

44

p a i r c o e f f 2 4 born / coul / long 1740.31 0.2907 0 0 0 p a i r c o e f f 4 4 born / coul / long 312.11 0.071124 0 0 0 p a i r c o e f f 1 4 born / coul / long 312.11 0.071124 0 0 0 p a i r c o e f f 3 4 born / coul / long 256.55 0.071124 0 0 0 # #3 body params p a i r c o e f f mbmh / yag / commoninput / p o t e n t i a l s / 0 / yag . mbmh A O Y A # maps 4 th type to A p a i r c o e f f mbmh2 / yag / commoninput / p o t e n t i a l s / 0 / yag . mbmh2 A O Y A # r e c o v e r y f r o m a b i n a r y w / d i f f e r e n t no o f p r o c s neigh modify delay 0 every 1 check yes

3-Body parameters (yag.mbmh)


# 3 body o n l y f o r c e r t a i n k i n d s and o n l y f o r XYY # ( a ( s i g m a =1) ) i s r i j 0 , lambda ( e p s i l o n =1) i s j u s t lambda # KEEP I T 1 FOR ALL EVEN IF NO THREE BODY INTERACTION ! # b / c same param u s e d i n 2 body and we d o n t want t o z e r o t h a t A A A 1.0 A A O A A Y A O A A O O A O Y A Y A A Y O A Y Y O A A O A O O A Y O O A O O O O O Y O Y A O Y O O Y Y 1.0 5.0 0.0 0.0 0.333333333333 # cut@5 312.11 14.06 3.0 3.0 0.0 1.0 1.0 0.0 0.0 0.0 0.333333333333 # no 3b , no 2 b 0.0 0.0 0.0 0.0 0.0 1.0 1.0 0.0 0.0 0.0 0.333333333333 # no 3b , no 2 b 0.0 0.0 0.0 0.0 0.0 1.0 1.0 0.0 0.0 0.0 0.333333333333 # no 3b , no 2 b 0.0 0.0 0.0 0.0 0.0 1.0 1.0 3.0 149.324 2.8 0.333333333333 # OAlO 1 7 4 0 . 3 1 3 . 4 4 2.0 3 . 0 0 . 0 1.0 1.0 0.0 0.0 0.0 0.333333333333 # no 3b , no 2 b 0.0 0.0 0.0 0.0 0.0 1.0 1.0 0.0 0.0 0.0 0.333333333333 # no 3b , no 2 b 0.0 0.0 0.0 0.0 0.0 1.0 1.0 0.0 0.0 0.0 0.333333333333 # no 3b , no 2 b 0.0 0.0 0.0 0.0 0.0 1.0 1.0 5.0 0.0 0.0 0.333333333333 # no 3b , cut@5 256.55 14.06 3.0 3.0 0.0 1.0 1.0 3.0 0.0 0.0 0.333333333333 # 3 b i n 2 nd f i l e 1 7 4 0 . 3 1 3 . 4 4 2.0 3 . 0 0 . 0 1.0 1.0 0.0 0.0 0.0 0.333333333333 # no 3b , no 2 b 0.0 0.0 0.0 0.0 0.0 1.0 1.0 0.0 0.0 0.0 0.333333333333 # 3 b i n 2 nd f i l e , no 2 b 0.0 0.0 0.0 0.0 0.0 1.0 1.0 0.0 0.0 0.0 0.333333333333 # no 3b , no 2 b 0.0 0.0 0.0 0.0 0.0 1.0 1.0 3.0 0.0 0.0 0.333333333333 # cut@3 2 4 4 9 . 4 4 3 . 4 4 2.0 2.0 0 . 0 1.0 1.0 0.0 0.0 0.0 0.333333333333 # no 3b , no 2 b 0.0 0.0 0.0 0.0 0.0 1.0 1.0 0.0 0.0 0.0 0.333333333333 # 3 b i n 2 nd f i l e , no 2 b 0.0 0.0 0.0 0.0 0.0 1.0 1.0 0.0 0.0 0.0 0.333333333333 # no 3b , no 2 b 0.0 0.0 0.0 0.0 0.0 1.0 1.0 3.0 0.0 0.0 0.333333333333 # 3 b i n 2 nd f i l e 1 2 5 0 . 8 5 2 . 8 6 2.0 3 . 0 0 . 0

45

Y A A 1.0 Y A O Y A Y Y O A Y O O Y O Y Y Y A Y Y O Y Y Y

1.0 5.0 0.0 0.0 0.333333333333 # no 3b , 256.55 14.06 3.0 3.0 0.0 1.0 1.0 0.0 0.0 0.0 0.333333333333 # no 3b , 0.0 0.0 0.0 0.0 0.0 1.0 1.0 0.0 0.0 0.0 0.333333333333 # no 3b , 0.0 0.0 0.0 0.0 0.0 1.0 1.0 0.0 0.0 0.0 0.333333333333 # no 3b , 0.0 0.0 0.0 0.0 0.0 1.0 1.0 3.0 168.25 2.8 0.333333333333 # 1 2 5 0 . 8 5 2 . 8 6 3 . 0 2.0 0 . 0 1.0 1.0 0.0 0.0 0.0 0.333333333333 # no 3b , 0.0 0.0 0.0 0.0 0.0 1.0 1.0 0.0 0.0 0.0 0.333333333333 # no 3b , 0.0 0.0 0.0 0.0 0.0 1.0 1.0 0.0 0.0 0.0 0.333333333333 # no 3b , 0.0 0.0 0.0 0.0 0.0 1.0 1.0 3.0 0.0 0.0 0.333333333333 # no 3b , 245.14 14.06 3.0 3.0 0.0

cut@5 no 2 b no 2 b no 2 b

no 2 b no 2 b no 2 b cut@3

3-Body parameters (yag.mbmh2)


# 3 body o n l y f o r c e r t a i n k i n d s and o n l y f o r XYY # ( a ( s i g m a =1) ) i s r i j 0 , lambda ( e p s i l o n =1) i s j u s t lambda # KEEP I T 1 FOR ALL EVEN IF NO THREE BODY INTERACTION ! # a c t u a l l y z e r o t h e twobody i n t e r a c t i o n h e r e A A A 1.0 A A O A A Y A O A A O O A O Y A Y A A Y O A Y Y O A A O A O O A Y ?) O O A O O O 1.0 0.0 0.0 0.0 0.333333333333 # no 3b , no 2 b 0.0 0.0 0.0 0.0 0.0 1.0 1.0 0.0 0.0 0.0 0.333333333333 # no 3b , no 2 b 0.0 0.0 0.0 0.0 0.0 1.0 1.0 0.0 0.0 0.0 0.333333333333 # no 3b , no 2 b 0.0 0.0 0.0 0.0 0.0 1.0 1.0 0.0 0.0 0.0 0.333333333333 # no 3b , no 2 b 0.0 0.0 0.0 0.0 0.0 1.0 1.0 0.0 0.0 0.0 0.333333333333 # no 3b , no 2 b 0.0 0.0 0.0 0.0 0.0 1.0 1.0 0.0 0.0 0.0 0.333333333333 # no 3b , no 2 b 0.0 0.0 0.0 0.0 0.0 1.0 1.0 0.0 0.0 0.0 0.333333333333 # no 3b , no 2 b 0.0 0.0 0.0 0.0 0.0 1.0 1.0 0.0 0.0 0.0 0.333333333333 # no 3b , no 2 b 0.0 0.0 0.0 0.0 0.0 1.0 1.0 5.0 0.0 0.0 0.333333333333 # no 3b , cut@5 256.55 14.06 3.0 3.0 0.0 1.0 1.0 2.6 6.242 2.0 0.333333333333 # 3 b i n t h i s f i l e 0.0 0.0 0.0 0.0 0.0 1.0 1.0 0.0 0.0 0.0 0.333333333333 # no 3b , no 2 b 0.0 0.0 0.0 0.0 0.0 1.0 1.0 2.6 6.242 2.0 0.333333333333 # 3 b i n t h i s f i l e ( symm , no 2 b 0.0 0.0 0.0 0.0 0.0 1.0 1.0 0.0 0.0 0.0 0.333333333333 # no 3b , no 2 b 0.0 0.0 0.0 0.0 0.0 1.0 1.0 0.0 0.0 0.0 0.333333333333 # no 3b , no 2 b

46

O O Y 1.0 O Y A ?) O Y O O Y Y Y A A Y A O Y A Y Y O A Y O O Y O Y Y Y A Y Y O Y Y Y

0.0 0.0 0.0 0.0 0.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 1.0 2.6 6.242 2.0 , no 2 b 0.0 0.0 0.0 0.0 0.0 1.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 1.0 2.6 6.242 2.0 0.0 0.0 0.0 0.0 0.0 1.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 1.0 1.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0 0.0

0.333333333333 # no 3b , no 2 b 0.333333333333 # 3 b i n t h i s f i l e ( symm

0.333333333333 # no 3b , no 2 b 0.333333333333 # 3 b i n 2 nd f i l e , no 2 b 0.333333333333 # no 3b , no 2 b 0.333333333333 # no 3b , no 2 b 0.333333333333 # no 3b , no 2 b 0.333333333333 # no 3b , no 2 b 0.333333333333 # no 3b , no 2 b 0.333333333333 # no 3b , no 2 b 0.333333333333 # no 3b , no 2 b 0.333333333333 # no 3b , no 2 b 0.333333333333 # no 3b , no 2 b

Simulation Execution
label start ## i n p u t read restart r . restart # s t u f f not in r e s t a r t f i l e i n c l u d e r u n r e q s . lmpin set set set set group group group group Al t y p e 1 O type 2 Y type 3 Ga t y p e 4

v a r i a b l e a e q u a l ( ${Tk } 9 . 6 7 1 1 4 3 4 9 e 05) + 1 2 . 3 9 1 8 l a t t i c e s c 1 s p a c i n g $a $a $a v a r i a b l e s i m s i z e e q u a l ( 2 $ h s x +$dx ) 2 $ l z $ l y v a r i a b l e s X t s e q u a l 1 0 0 0 ( ${ s i m s i z e } ) # c o o k t i m e variable U equal e t o t a l # v a r i a b l e Ta e q u a l temp variable ts equal step v a r i a b l e dx e q u a l $dx v a r i a b l e hsx e q u a l $hsx # h e a t e r or s i n k width v a r i a b l e handsw e q u a l ${ h s x }2

47

v a r i a b l e ly equal $ly variable lz equal $lz v a r i a b l e h a l f s i m b o x w i d t h e q u a l ${ dx }+$ { handsw } v a r i a b l e Tk e q u a l $Tk v a r i a b l e TH e q u a l $Tk+$dT / 2 v a r i a b l e TC e q u a l $Tk$dT / 2 v a r i a b l e d T o v e r d x e q u a l $dT / $dx v a r i a b l e dT e q u a l $dT v a r i a b l e dTH e q u a l ${dT } / 2 v a r i a b l e dTC e q u a l $ {dT } / 2 compute k e a a l l ke / atom # c o n v e r t s KE t o temp v a r i a b l e Ta atom 7 7 3 6 . 3 4 c k e a #=kb / 1 . 5 = 1 1 6 0 4 / 1 . 5 # a s s i g n m e n t o f groups o f atoms # |0 s i n k n1dxn2h e a t e r n3h e a t e r n4dxn5s i n k n6 | v a r i a b l e p a r t g index l s i n k ldx l h e a t e r r h e a t e r rdx r s i n k v a r i a b l e p a r t l i n d e x $ h s x $dx $ h s x $ h s x $dx $ h s x v a r i a b l e s l a b i e q u a l 160 $ l y $ l z variable s t a r t i d equal 1 label extrude v a r i a b l e e n d i d e q u a l ${ s t a r t i d }+$ { s l a b i } ${ p a r t l }1 g r o u p ${ p a r t g } i d <> $ { s t a r t i d } $ { e n d i d } v a r i a b l e s t a r t i d e q u a l ${ e n d i d }+1 v a r i a b l e xmn e q u a l bound ( ${ p a r t g } , xmin ) v a r i a b l e xmx e q u a l bound ( ${ p a r t g } , xmax ) p r i n t b o u n d s f o r $ { p a r t g } : min : ${xmn } , max : ${xmx} next partg next p a r t l jump a c t i o n s . l m p i n e x t r u d e group s i n k s union l s i n k r s i n k group h e a t e r s union l h e a t e r r h e a t e r group grads union ldx rdx

f i x l g f l d x nve f i x r g f r d x nve f i x l h e a t e r f l h e a t e r n v t temp $ {TH} $ {TH} 1 f i x r h e a t e r f r h e a t e r n v t temp $ {TH} $ {TH} 1 #why h a v e t o s e p a r a t e ? ! f i x l s i n k f l s i n k n v t temp $ {TC} ${TC} 1 f i x r s i n k f r s i n k n v t temp $ {TC} ${TC} 1 fix modify l h e a t e r f energy yes fix modify r h e a t e r f energy yes fix modify l s i n k f energy yes fix modify r s i n k f energy yes

# two s t a g e s : 1 . t h e r m a l i z e 2 . s s w h i l e r e c o r d i n g # what s t a g e i s t h e s i m a t ? i f $ { t s } < $ { s X t s } t h e n jump a c t i o n s . l m p i n t r a n s i e n t & e l s e jump a c t i o n s . l m p i n s s

48

label transient r u n $ { s X t s } u p t o e v e r y 1000 w r i t e r e s t a r t r . r e s t a r t # n e e d t o e r a s e f i x d a t a b / c i don t want t o r e c o r d i t unfix lgf unfix rgf unfix l h e a t e r f unfix rheaterf unfix lsinkf unfix rsinkf w r i t e r e s t a r t r . r e s t a r t # s o i don t s a v e p r e v i o u s f i x i n f o label ss f i x l g f l d x nve f i x r g f r d x nve f i x l h e a t e r f l h e a t e r n v t temp $ {TH} $ {TH} 1 f i x r h e a t e r f r h e a t e r n v t temp $ {TH} $ {TH} 1 f i x l s i n k f l s i n k n v t temp $ {TC} ${TC} 1 f i x r s i n k f r s i n k n v t temp $ {TC} ${TC} 1 fix modify l h e a t e r f energy yes fix modify r h e a t e r f energy yes fix modify l s i n k f energy yes fix modify r s i n k f energy yes # e n e r g y t h a t h a s gone t h r u s y s ( e a c h s i d e ) . . / 4 i s f o r avg . / 2 t o s p l i t sys E in 2 run 0 # v a r i a b l e Epc e q u a l 1 0 0 ( s q r t ( f l s i n k f 2 ) + s q r t ( f l h e a t e r f 2 ) + s q r t ( f r h e a t e r f 2 )+s q r t ( f r s i n k f 2 ) ) / ( 4 s q r t ( e t o t a l 2 ) / 2 ) # j u s t s o i can h a v e f i l e names w / t i m e s t e p s f i x a v g f i x a l l a v e / s p a t i a l 10 100 1000 x l o w e r 1 v Ta f i l e $ { t s } . tempprofile . avespacial f i x q f a l l a v e / t i m e 1 1 1000 f l s i n k f f l h e a t e r f f r h e a t e r f f r s i n k f f i l e ${ t s } . cumdE . a v e t i m e i f $ { t s } > 1000000 t h e n jump a c t i o n s . l m p i n end l a b e l keepgoing r u n 1000 write restart r . restart r u n 0 # i d k why i n e e d t o i n c l u d e t h i s # i f ${ Epc } > . 1 t h e n jump a c t i o n s . l m p i n end e l s e jump a c t i o n s . l m p i n keepgoing i f $ { t s } > 1000000 t h e n jump a c t i o n s . l m p i n end e l s e jump a c t i o n s . lmpin keepgoing l a b e l end

49

LAMMPS EQUILIBRIUM THERMAL CONDUCTIVITY CODE Simulation Setup


label inits dimension 3 u n i t s metal boundary p p p newton on atom style charge

v a r i a b l e a e q u a l ( ${Tk } 9 . 6 7 1 1 4 3 4 9 e 05) + 1 2 . 3 9 1 8 # geometry # p r i m i t i v e v e c s a1 .5 . 5 . 5 a2 . 5 .5 . 5 a3 . 5 . 5 .5 & # c o n v e n t i o n a l a1 1 0 0 a2 0 1 0 a3 0 0 1 & # b a s i s v e c s b e l o w , A1 ( 1 6 ) , A2 ( 2 4 ) ,O( 9 6 ) , Y ( 2 4 ) i n o r d e r , f o r c o n v e n t i o n a l # half of that for primitive #ERROR I n p u t l i n e t o o l o n g ! ! c h a n g e d i n p u t . cpp MAXLINES # u s i n g $a a s i n i t i a l l a t t i c e c u s t o m $a & basis 0.0 0.0 0.0 & basis 0.5 0 0.5 & basis 0 0.5 0.5 & basis 0.5 0.5 0 & basis 0.75 0.25 0.25 & basis 0.75 0.75 0.75 & basis 0.25 0.25 0.75 & basis 0.25 0.75 0.25 & basis 0.25 0.75 0.75 & basis 0.25 0.25 0.25 & basis 0.75 0.75 0.25 & basis 0.75 0.25 0.75 & basis 0.5 0.5 0.5 & basis 0 0.5 0 & basis 0.5 0 0 & basis 0 0 0.5 & basis 0.375 0 0.25 & basis 0.125 0 0.75 & basis 0.625 0.5 0.25 & basis 0.875 0.5 0.75 & basis 0.25 0.375 0 & basis 0.75 0.125 0 & basis 0.25 0.625 0.5 & basis 0.75 0.875 0.5 & basis 0 0.25 0.375 & basis 0 0.75 0.125 & basis 0.5 0.25 0.625 & basis 0.5 0.75 0.875 & basis 0.75 0.625 0 & basis 0.75 0.375 0.5 & basis 0.25 0.875 0 & basis 0.25 0.125 0.5 & basis 0.125 0.5 0.25 & basis 0.875 0 0.25 &

50

basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis

0.375 0.625 0 0.5 0 0.5 0.28023 0.21977 0.71977 0.78023 0.19922 0.69922 0.30078 0.80078 0.1011 0.8989 0.6011 0.3989 0.8511 0.6489 0.3511 0.1489 0.03023 0.96977 0.46977 0.53023 0.94922 0.44922 0.05078 0.55078 0.71977 0.78023 0.28023 0.21977 0.80078 0.30078 0.69922 0.19922 0.8989 0.1011 0.3989 0.6011 0.1489 0.3511 0.6489 0.8511 0.96977 0.03023 0.53023 0.46977 0.05078 0.55078 0.94922 0.44922

0.5 0 0.25 0.25 0.75 0.75 0.1011 0.8989 0.6011 0.3989 0.28023 0.21977 0.71977 0.78023 0.19922 0.69922 0.30078 0.80078 0.53023 0.46977 0.96977 0.03023 0.44922 0.94922 0.55078 0.05078 0.3511 0.1489 0.8511 0.6489 0.8989 0.1011 0.3989 0.6011 0.71977 0.78023 0.28023 0.21977 0.80078 0.30078 0.69922 0.19922 0.46977 0.53023 0.03023 0.96977 0.55078 0.05078 0.44922 0.94922 0.6489 0.8511 0.1489 0.3511

0.75 0.75 0.875 0.125 0.625 0.375 0.19922 0.69922 0.30078 0.80078 0.1011 0.8989 0.6011 0.3989 0.28023 0.21977 0.71977 0.78023 0.05078 0.55078 0.94922 0.44922 0.1489 0.3511 0.6489 0.8511 0.96977 0.03023 0.53023 0.46977 0.80078 0.30078 0.69922 0.19922 0.8989 0.1011 0.3989 0.6011 0.71977 0.78023 0.28023 0.21977 0.94922 0.44922 0.05078 0.55078 0.8511 0.6489 0.3511 0.1489 0.03023 0.96977 0.46977 0.53023

& & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & &

51

basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis

0.78023 0.71977 0.21977 0.28023 0.69922 0.19922 0.80078 0.30078 0.6011 0.3989 0.1011 0.8989 0.3511 0.1489 0.8511 0.6489 0.53023 0.46977 0.96977 0.03023 0.44922 0.94922 0.55078 0.05078 0.21977 0.28023 0.78023 0.71977 0.30078 0.80078 0.19922 0.69922 0.3989 0.6011 0.8989 0.1011 0.6489 0.8511 0.1489 0.3511 0.46977 0.53023 0.03023 0.96977 0.55078 0.05078 0.44922 0.94922 0.125 0.375 0.875 0.625 0.25 0.75

0.6011 0.3989 0.1011 0.8989 0.78023 0.71977 0.21977 0.28023 0.69922 0.19922 0.80078 0.30078 0.03023 0.96977 0.46977 0.53023 0.94922 0.44922 0.05078 0.55078 0.8511 0.6489 0.3511 0.1489 0.3989 0.6011 0.8989 0.1011 0.21977 0.28023 0.78023 0.71977 0.30078 0.80078 0.19922 0.69922 0.96977 0.03023 0.53023 0.46977 0.05078 0.55078 0.94922 0.44922 0.1489 0.3511 0.6489 0.8511 0 0 0.5 0.5 0.125 0.375

0.69922 0.19922 0.80078 0.30078 0.6011 0.3989 0.1011 0.8989 0.78023 0.71977 0.21977 0.28023 0.55078 0.05078 0.44922 0.94922 0.6489 0.8511 0.1489 0.3511 0.46977 0.53023 0.03023 0.96977 0.30078 0.80078 0.19922 0.69922 0.3989 0.6011 0.8989 0.1011 0.21977 0.28023 0.78023 0.71977 0.44922 0.94922 0.55078 0.05078 0.3511 0.1489 0.8511 0.6489 0.53023 0.46977 0.96977 0.03023 0.25 0.75 0.25 0.75 0 0

& & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & & &

52

basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis basis

0.25 0.75 0 0 0.5 0.5 0.875 0.625 0.125 0.375 0.75 0.25 0.75 0.25 0 0 0.5 0.5

0.875 0.625 0.25 0.75 0.25 0.75 0 0 0.5 0.5 0.875 0.625 0.125 0.375 0.75 0.25 0.75 0.25

0.5 0.5 0.125 0.375 0.875 0.625 0.75 0.25 0.75 0.25 0 0 0.5 0.5 0.875 0.625 0.125 0.375

& & & & & & & & & & & & & & & & & &

region wholething block 0 1 0 1 0 1 u n i t s l a t t i c e c r e a t e b o x 4 w h o l e t h i n g # no . i s n t y p e s # a f t e r b a s i s kw 1 s t no . i s b a s i s atm , 2 nd i s t y p e create atoms 1 box basis 1 1 # c r e a t e s 160 a t o m s a t once # . . . w i l l need t o spec t y p e s g r o u p Al1 i d <> 1 16 g r o u p Al2 i d <> 17 40 g r o u p Al u n i o n Al1 Al2 g r o u p O i d <> 41 136 g r o u p Y i d <> 137 160 s e t g r o u p Al t y p e 1 s e t group O type 2 s e t group Y type 3

# Atom p r o p e r t i e s mass 1 2 6 . 9 8 # A l mass 2 1 5 . 9 9 9 #O mass 3 8 8 . 9 0 6 #Y mass 4 6 9 . 7 # Dopant Ga f o r A l # can add a d o p a n t a s mass 4 s e t g r o u p O c h a r g e 2.0 #O ( 2) s e t g r o u p Al c h a r g e 3 . 0 # A l ( + 3 ) . . . t h e dopedf o r atom s e t g r o u p Y c h a r g e 3 . 0 #Y ( + 3 ) i n c l u d e r u n r e q s . lmpin # l o o k i n g a t one s i d e o f s i m box v a r i a b l e dx e q u a l $dx v a r i a b l e ly equal $ly variable lz equal $lz # c r e a t e s l a b t o be e x t r u d e d

53

r e p l i c a t e 1 $ly $lz r e p l i c a t e $ { dx } 1 1 # random s u b s t i t u t i o n s e t g r o u p Al t y p e / f r a c t i o n 4 $dp $ d s e e d g r o u p Ga t y p e 4 s e t g r o u p Ga c h a r g e 3 . 0 #Ga

v a r i a b l e T e q u a l $Tk v e l o c i t y a l l c r e a t e ${Tk} $ v s e e d d i s t g a u s s i a n reset timestep 0 write restart r . restart

Interatomic Potentials runreqs.lmpin


timestep .001 k s p a c e s t y l e pppm . 0 0 0 0 1 # u s u a l l y a t . 0 0 0 1 b u t n e e d some s p e e d , . 0 0 0 0 1 conserves but slow p a i r s t y l e h y b r i d / o v e r l a y b o r n / c o u l / l o n g 4 . 0 7 . 0 mbmh mbmh2 # #2 body params #w / o Ga p a i r c o e f f 2 2 born / coul / long 2449.44 0.2907 0 0 0 p a i r c o e f f 1 2 born / coul / long 1740.31 0.2907 0 0 0 p a i r c o e f f 2 3 born / coul / long 1250.85 0.3497 0 0 0 p a i r c o e f f 1 1 born / coul / long 312.11 0.071124 0 0 0 p a i r c o e f f 3 3 born / coul / long 245.14 0.071124 0 0 0 p a i r c o e f f 1 3 born / coul / long 256.55 0.071124 0 0 0 #w Ga , r e p l a c e d t h e 1 s w 4 s ( b u t s m a l l e r no . h a s t o be 1 s t p a i r c o e f f 2 4 born / coul / long 1740.31 0.2907 0 0 0 p a i r c o e f f 4 4 born / coul / long 312.11 0.071124 0 0 0 p a i r c o e f f 1 4 born / coul / long 312.11 0.071124 0 0 0 p a i r c o e f f 3 4 born / coul / long 256.55 0.071124 0 0 0 # #3 body params p a i r c o e f f mbmh / yag / commoninput / p o t e n t i a l s / 0 / yag . mbmh A O Y A # maps 4 th type to A p a i r c o e f f mbmh2 / yag / commoninput / p o t e n t i a l s / 0 / yag . mbmh2 A O Y A # r e c o v e r y f r o m a b i n a r y w / d i f f e n t no o f p r o c s neigh modify delay 0 every 1 check yes

3-Body parameters are the same as the non-equilibrium script. Simulation Execution
label start ## i n p u t read restart r . restart # s t u f f not in r e s t a r t f i l e i n c l u d e r u n r e q s . lmpin

54

s e t g r o u p Al t y p e 1 s e t group O type 2 s e t group Y type 3 s e t g r o u p Ga t y p e 4 g r o u p AlGa u n i o n Al Ga v a r i a b l e a e q u a l ( ${Tk } 9 . 6 7 1 1 4 3 4 9 e 05) + 1 2 . 3 9 1 8 l a t t i c e s c 1 s p a c i n g $a $a $a v a r i a b l e s i m s i z e e q u a l ( $dx ) $ l z $ l y v a r i a b l e s X t s e q u a l 1 0 0 0 ( ${ s i m s i z e } ) # c o o k t i m e variable U equal e t o t a l variable ts equal step v a r i a b l e t s m i n u s s X t s e q u a l s t e p ${ s X t s } # s t u p i d lammps s y n t a x . . no s p a c e a r o u n d t h e m i n u s s i g n ! ! ! v a r i a b l e dx e q u a l $dx v a r i a b l e ly equal $ly variable lz equal $lz v a r i a b l e Tk e q u a l $Tk compute k e a a l l ke / atom f i x n v t f a l l n v t temp $Tk $Tk 1 fix modify n v t f energy yes # two s t a g e s : 1 . t h e r m a l i z e 2 . s s w h i l e r e c o r d i n g # what s t a g e i s t h e s i m a t ? i f $ { t s } < $ { s X t s } t h e n jump a c t i o n s . l m p i n t r a n s i e n t & e l s e jump a c t i o n s . l m p i n s s label transient r u n $ { s X t s } u p t o e v e r y 1000 w r i t e r e s t a r t r . r e s t a r t # n e e d t o e r a s e f i x d a t a b / c i don t want t o r e c o r d i t unfix nvtf #reset timestep 0 w r i t e r e s t a r t r . r e s t a r t # s o i don t s a v e p r e v i o u s f i x i n f o label ss f i x n v t f a l l n v t temp $Tk $Tk 1 fix modify n v t f energy yes run 0 compute AlGaKE AlGa ke / atom compute AlGaPE AlGa pe / atom compute A l G a S t r e s s AlGa s t r e s s / atom v i r i a l # v i r i a l means KE c o n t r i b not included compute f l u x A l G a AlGa h e a t / f l u x AlGaKE AlGaPE A l G a S t r e s s # a l l i g n o r e d ? a l l same g r p v a r i a b l e JxAlGa e q u a l c f l u x A l G a [ 1 ] # / v o l compute compute compute compute OKE O ke / atom OPE O pe / atom O S t r e s s O s t r e s s / atom v i r i a l f l u x O O h e a t / f l u x OKE OPE O S t r e s s # a l l i g n o r e d ? a l l same g r p

55

variable compute compute compute compute variable

JxO e q u a l c f l u x O [ 1 ] # / v o l YKE Y ke / atom YPE Y pe / atom Y S t r e s s Y s t r e s s / atom v i r i a l f l u x Y Y h e a t / f l u x YKE YPE Y S t r e s s # a l l i g n o r e d ? a l l same g r p JxY e q u a l c f l u x Y [ 1 ] # / v o l

# j u s t s o i can h a v e f i l e names w / t i m e s t e p s f i x q f a l l a v e / t i m e 1 1 1 v JxAlGa v JxO v JxY f i l e ${ t s } . cumdE . a v e t i m e i f $ { t s m i n u s s X t s } > 1000000 t h e n jump a c t i o n s . l m p i n end l a b e l keepgoing r u n 1000 write restart r . restart r u n 0 # i d k why i n e e d t o i n c l u d e t h i s i f $ { t s m i n u s s X t s } > 1000000 t h e n jump a c t i o n s . l m p i n end e l s e jump a c t i o n s . lmpin keepgoing l a b e l end

56

BIBLIOGRAPHY [1] Jmol: an open-source java viewer for chemical structures in 3d, On the WWW, URL http://www.jmol.org/. [2] L. Dobrzycki, E. Bulska, D. A. Pawlak, Z. Frukacz, and K. Wozniak, Structure of YAG crystals doped/substituted with erbium and ytterbium, Inorganic Chemistry, vol. 43, no. 24, pp. 76567664, 2004, pMID: 15554630. [3] A. Potdevin, G. Chadeyron, and R. Mahiou, Tb3+ -doped yttrium garnets: Promising tunable green phosphors for solid-state lighting, Chemical Physics Letters, vol. 490, no. 1-3, pp. 50 53, 2010. [4] R. Hansel, S. Allison, and G. Walker, Temperature-dependent luminescence of gallium-substituted YAG:Ce, Journal of Materials Science, vol. 45, pp. 146150, 2010, 10.1007/s10853-009-3906-9. [5] X. Q. Cao, R. Vassen, and D. Stoever, Ceramic materials for thermal barrier coatings, Journal of the European Ceramic Society, vol. 24, no. 1, pp. 1 10, 2004. [6] A. Ikesue, Y. L. Aung, T. Taira, T. Kamimura, K. Yoshida, and G. L. Messing, Progress in ceramic lasers, Annual Review of Materials Research, vol. 36, no. 1, pp. 397429, 2006. [7] J. L. Caslavsky and D. J. Viechnicki, Melting behaviour and metastability of yttrium aluminium garnet (YAG) and YAlO3 determined by optical differential thermal analysis, Journal of Materials Science, vol. 15, pp. 17091718, 1980, 10.1007/BF00550589. [8] R. Wynne, J. L. Daneu, and T. Y. Fan, Thermal coefcients of the expansion and refractive index in YAG, Appl. Opt., vol. 38, no. 15, pp. 32823284, 1999. [9] Red Optronics, Undoped yag crystal (Y3 Al5 O12 ), On the WWW, URL http://www.redoptronics.com/undoped-YAG-crystal.html. [10] H. Qiu, P. Yang, J. Dong, P. Deng, J. Xu, and W. Chen, The inuence of Yb concentration on laser crystal Yb:YAG, Materials Letters, vol. 55, no. 1-2, pp. 1 7, 2002. [11] J. Morikawa, C. Leong, T. Hashimoto, T. Ogawa, Y. Urata, S. Wada, M. Higuchi, and J. ichi Takahashi, Thermal conductivity/diffusivity of Nd3+ doped GdVO4 , YVO4 , LuVO4 , and Y3 Al5 O12 by temperature wave analysis, Journal of Applied Physics, vol. 103, no. 6, p. 063522, 2008. [12] J. F. W. Gunsser, U. Wolfmeier, 1.3.8.4 Debye Temperature, A. M. H. K.-H. Hellwege, Ed. Springer-Verlag, 1978, vol. 12a.

57

[13] J. Bodzenta, A. Ka mierczak-Baata, K. B. Wokulska, J. Kucytowski, T. ukasiewicz, z and W. Hofman, Inuence of doping on thermal diffusivity of single crystals used in photonics: measurements based on thermal wave methods, Appl. Opt., vol. 48, no. 7, pp. C46C54, 2009. [14] C. Jun, C. Dong-Quan, and Z. Jing-Lin, Molecular dynamics simulation of thermodynamic properties of YAG, Chinese Physics, vol. 16, no. 9, p. 2779, 2007. [15] S. Plimpton, Fast parallel algorithms for short-range molecular-dynamics, Journal of Computational Physics, vol. 117, no. 1, pp. 119, March 1 1995. [16] L. Verlet, Computer Experiments on Classical Fluids. II. Equilibrium Correlation Functions, Phys. Rev., vol. 165, no. 1, pp. 201214, Jan 1968. [17] M. P. Tosi and F. G. Fumi, Ionic sizes and born repulsive parameters in the NaCltype alkali halidesII : The generalized Huggins-Mayer form, Journal of Physics and Chemistry of Solids, vol. 25, no. 1, pp. 45 52, 1964. [18] D. A. McQuarrie, Statistical Mechanics. University Science, 2000.

[19] J. Li, Modeling microstructural effects on deformation resistence and thermal conductivity, Ph.D. dissertation, Massachusetts Institute of Technology, 2000. [20] A. McGaughey and M. Kaviany, Phonon transport in molecular dynamics simulations: Formulation and thermal conductivity prediction, Advances in Heat Transfer, vol. 39, pp. 169256, 2006. [21] A. J. H. McGaughey and M. Kaviany, Thermal conductivity decomposition and analysis using molecular dynamics simulations: Part ii. complex silica structures, International Journal of Heat and Mass Transfer, vol. 47, no. 8-9, pp. 1799 1816, 2004. [22] Billauer, Eli, peakdet: Peak detection using matlab (non-derivative local extremum), On the WWW, URL http://www.billauer.co.il/peakdet.html. [23] A. J. McGaughey, Phonon transport in molecular dynamics simulations: Formulation and thermal conductivity prediction, Ph.D. dissertation, University of Michigan, 2004.

58