Problem Set 3

Due Oct 16, 2024, 11:59pm. Upload onto Canvas

206 Points Total; Scaled to 100 pt.
Problem 1: Kroger-Vink notation for ZrO2 (14 pt)
The fluorite compound ZrO2 has many diverse applications. The formal oxidation state
for these elements are Zr: 4+ and O2-. In parts a-c, write the symbols for the following
defects in Kroger-Vink notation.
a. Oxygen vacancy (3 pt)
b. Substitution of Zr4+ with Y3+ (3 pt)
c. Zr vacancy (3 pt)

d. The introduction of Y3+ into Zr4+ sites will create an effective negative charge
on the Zr sites. This charge is balanced by oxygen vacancies, which is
positively charged when referenced to oxygen anions. What is the ratio of
Y3+ to oxygen vacancies (5 pt)?

Problem 2: Perovskite oxides (15 pt)

Perovskite oxides, with the general composition ABO3, are used extensively in energy
storage and conversion. Take the perfect crystalline solid Sr2+Ti4+O32- as an example. For
electronic defects, assume that they are localized on the Ti atom. Using Kroger-Vink
notation, write the symbols for the following defects (3 pt each):
a) Oxygen vacancy
b) Anti-site defect (there are 2 symbols here)
c) Localized hole (positively charged electronic defect)
d) Substituting Ti4+ with Fe3+
e) Substituting Ti4+ with Nb5+
 Problem 3: Configurational entropy and dilute approximation (30 pt)
Consider a crystal with N lattice sites. Assume that only vacancy defects form
a. Derive the equation for the defect configurational entropy, Δ Scfg, for a large
number of defects n and lattices sites N. Hint: start with the expression for the
number of combinations given N sites and n defects, and then use Stirling’s
approximation: 𝐥𝐧 𝑵! ~ 𝑵 𝐥𝐧 𝑵 − 𝑵 (5 pt)
𝑛 𝑁
Δ𝑆!"# = −𝑛𝑘$ ln + 𝑁𝑘$ ln
𝑁−𝑛 𝑁−𝑛
b. The entropic component to the chemical potential is also known as the partial
%&'
molar entropy, and is given by Δ𝑠!"# = %(!"# . Derive the equation for the partial
molar entropy in the non-dilute limit. Note this is lowercase s (5 pt)
𝑛
Δ𝑠!"# = −𝑘$ ln
𝑁−𝑛
c. Next, write down Δ𝑠!"# in the dilute limit where n ≪ N (5 pt)
d. Sketch Δ𝑠!"# vs n/N, or [P], in both the non-dilute limit and the dilute limit. The
dilute limit assumes that 1-[P] = 1 because [P]≪1, or N-n = N because n≪N (5 pt)
(
e. Calculate the critical value of ) where the dilute limit breaks down, which we
define as when the error between dilute and nondilute becomes greater than 1%
of the value of the partial molar entropy (5 pt)
f. Compute the vacancy concentration in the dilute and non-dilute limit for Au at
500K and 1000K (use the values in the table for problem 4) (10 pt).
 Problem 4: Melting temperature (35 pt)
The vacancy formation enthalpy and vibrational entropy for five metals is tabulated in
this table below:
a. The melting temperature can be approximated by when the equilibrium vacancy
concentration reaches a critical value 𝑐*+ , a value typically between 0.1% and
0.01%. Using the values on this table, predict the melting temperature (in K) for
these five metals. The vibration enthalpy is given in units of the Boltzmann
constant (lowercase k in this table). (25 pt)

b. Find the coefficient of determination, or R2, between the predicted melting

temperature (Tpredicted) and the actual tabulated melting temperature (Tactual), for
these five metals (10 pt). Make sure you use the same units for temperature, and
do not mix up Kelvin, Celsius, and Fahrenheit. R2 is a statistical quantity to
define how good a prediction is. It is given by the following, where var is the
variance
𝑆𝑆-./
𝑅, = 1 −
𝑆𝑆010
𝑆𝑆010 = 𝑣𝑎𝑟(𝑇+!02+3 )
𝑆𝑆-./ = 𝑣𝑎𝑟(𝑇4-.56!0.5 − 𝑇+!02+3 )
 Question 5: Interconnects in Integrated Circuits (35 pt)
In integrated circuits, Silicon is used as the active material for transistors, while
metal interconnects are used to electrically connect the Si transistors (see image below).
However, even small amounts (<<1 part per million) of metal impurities in silicon
strongly degrades the performance of the transistors.

Metal Interconnects
SiO2

Silicon Transistors

Source: J. Walvern et al. Sandia National Labs, SAND2017-9661C

One process in the fabrication of Si microelectronics is post-metallization anneal,
where the chip is annealed at 400C (673K) for 30 minutes in H2/N2 gas mixture. Metal
interconnects not only be stable at this temperature, but also not diffuse into Si to damage
the transistors. Materials that meet these criteria (along others) are termed CMOS-
compatible materials; in many cases, CMOS-incompatible materials like Gold are not
allowed in the same clean room as CMOS materials.
Appendix 1 lists some measured diffusion parameters for different metals in
silicon; EA is equivalent to the vacancy migration enthalpy HM. The D0 pre-factor
combines all of the temperature-independent terms for diffusion (vibration frequency,
atomic jump distance, etc). Consider four metals for use as interconnects: copper (Cu),
gold (Au), silver (Ag), and aluminum (Al).
a) Compute the diffusivity of Cu, Au, Ag, and Al in Si at 400C (673K). Assume that
the vacancy migration enthalpy does not depend on temperature. If there are
multiple diffusion mechanisms (vacancy, interstitial) listed for a material, pick the
one with the highest computed diffusivity (20 pt). Remember that the computed
diffusivity depends both on the D0 pre-factor AND the activation energy.
b) Compute the characteristic diffusion length for 30 minutes for each of these elements
(10 pt)
c) A benchmark is that the metal must not diffuse more than 10 nm in Si during
postmetallization anneal, or it will start entering the transistor channel and change
the transistor doping levels and its electronic properties. Explain why integrated
circuits would use Al instead of Cu*, Ag, or Au (5 pt)

*in practice, integrated circuits started to use Cu interconnects in the late 1990s, as
copper diffusion barriers based on tantalum nitride and other materials were
developed.
Problem 6: Defect concentration (25 pt)
Inert pure metals X and Y form either vacancies or interstitials. An interstitial
forms when an atom enters a lattice site, resulting in a loss of an atom from the surface; a
vacancy forms when an atom moves from the bulk to the surface, leaving behind an
empty lattice site. The initial volume of the material is computed by slowly cooling the
material down from high temperature such that the defects are in equilibrium. To
compute the defect concentration as a function of temperature, we can heat up the
material to a high temperature and enable its defects to reach equilibrium, then rapidly
quench it such that the defects cannot equilibrate at lower temperatures. This would
enable the material to retain the same defect concentration as at the high temperature.
The volume of the material after quenching is recorded.
The volumes of materials X and Y are recorded. The masses after each heat and
quench experiment do not change. Assume that the lattice does not become distorted in
any way despite the addition of interstitials or vacancies; in other words, the lattice
constants are independent of the defect fraction.

1.00016
Final volume / initial volume

1.00012 Material X
Material Y
1.00008

1.00004

1.00000

0.99996

400 600 800 1000 1200 1400

Temperature (K)

a) Is the primary defect in X vacancies or interstitials (3 pt)?

b) What is the point defect fraction [P] at 1200K for material X (2 pt)?
c) What is the defect formation enthalpy for material X? Assume no vibrational
entropy component (10 pt).
d) Does material Y contain primarily vacancies or interstitials (3 pt)?
e) What is the defect formation enthalpy for material Y? Assume no vibrational
entropy component (7 pt).
 Problem 7: Types of dislocations (36 pt)
The following diagram shows the types of dislocations in a dislocation loop. Note: we
define the Burgers vector as Right Hand, Finish to Start. Other texts define Burgers as Start to
Finish.

+Y

-X +X

-Y

Source: MIT OCW, C. Schuh

a. For the four types of dislocations shown above, state whether the dot and cross
products are positive, negative, or 0 (16 pt)

Dot Product (𝑏=⃗ ⋅ 𝑡⃗) Cross Product (𝑏=⃗ × 𝑡⃗)

Right-handed screw
Left-handed screw
Negative Edge
Positive Edge
b. An edge dislocation is drawn below. The sense vector is in the [010] direction.
We define the positive Y direction as the direction given by the Y arrow below
(into the cube). Draw the Burgers circuit and the Burgers vector. Find (𝑏=⃗ ⋅ 𝑡⃗) and
(𝑏=⃗ × 𝑡⃗), and identify this as a positive or negative edge dislocation using your
chart in part (a) (10 pt)
 z
y

c. Now consider renaming the directions such that the sense vector is in the [01C0]
direction. Draw the Burgers circuit and the Burgers vector. Find (𝑏=⃗ ⋅ 𝑡⃗) and
(𝑏=⃗ × 𝑡⃗), and identify this as a positive or negative edge dislocation (5 pt)

z
y

d. Explain why it did not matter if we drew the sense vector in the positive or
negative direction in terms of identifying the type of dislocation (5 pt)
 Problem 8: Edge and Screw Dislocations (16 pt)
Pure edge and screw dislocations are rare in real materials; most dislocations contain
edge and screw character, and are called mixed dislocations. The dislocation below
changes direction while it is inside the crystal. At point B, the sense vector is in the [210]
direction; the vectors are given as [xyz]. The xyz coordinates axes (cyan color) denote
the positive direction for each axis.

B
A
C
Z
𝒕⃗ (sense vector)
Y
x

Source: thesis of Francesca Boioli, University of Milan-Biococa

a. Write the coordinates of the Burgers vector at A and C in terms of [hkl] or [xyz]
(6 pt)
b. Identify whether the dislocation at points A and C are left handed screw, right
handed screw, positive edge, or negative edge dislocations. (6 pt)
c. A mixed dislocation is one where the sense and burgers vector are neither
parallel nor perpendicular to each other; it is characterized by the angle between
the Burgers and Sense vector. Find the angle between the sense and Burgers
vector at point B (4 pt)
 Appendix 1

Source: Scotten Jones, Diffusion in Silicon,

[Link]