MTE 4050 Physical Metallurgy – Mechanical Properties

Fall 2024
Reading 07 Dislocation Physics

This Lecture:
• Dislocation geometry
• Forces on dislocations
• Dislocation/defect interactions
• Dislocation stress fields
• Dislocation energy
• Dislocation structure & density

Basic Dislocation Types

Dislocations come in two types: edge and screw. These two different types have different
topology, but the practical effect on a crystal by their passage is the same in the end. The two
types are shown below. An edge dislocation transmits shear deformation or “slip” in the direction
of its Burgers vector b in the slip plane. The dislocation line ℓ lies in the slip plane perpendicular
to b and the direction of travel d. This is the feature that defines purely edge dislocations: b | ℓ.
For screw dislocations, the situation is different; b||ℓ, and both are perpendicular to d. (The
screw dislocation is sometimes imagined to have a “parking-lot structure,” in which one complete
cycle around the dislocation line moves one up or down by a distance b.) In general, the angle
between b and ℓ can be any value between 0 and 90°; these dislocations are termed mixed
dislocations.

J. B. Puthoff Cal Poly Pomona MTE 4050

Furthermore, dislocations can have a “positive” (edge/screw) or “negative” (antiedge/antiscrew)
sense. The passage of a positive dislocation to the right produces slip equivalent to the passage
of a negative dislocation to the left. The situation is shown above for both the edge and the screw
cases. Dislocations can also form another important structure called a dislocation loop. A
dislocation loop is produced when a single dislocation line turns back upon itself and closes, as
shown above. The loop is of mixed character, except at the 4 “corners,” where the dislocation is
variously edge, antiedge, screw, and antiscrew. A shear stress in the plane of the dislocation
orientated along b causes the dislocation loop to expand, producing its increment of slip.

Dislocation/Defect Interactions
Dislocations can also contain “defects” themselves in the form of kinks and jogs. These are
produced when dislocations pass through one another. Some kinked/jogged edge dislocations
are below. The deformation imparted by a dislocation with a kink in its slip plane is unaffected,

J. B. Puthoff Cal Poly Pomona MTE 4050

but the presence of a jog will modify the shape of the slipped region. Kinks and jogs are formed
where dislocation lines cross each other during slip (see below).

Dislocations will also interact with point defects, such as vacancies, and each other. A number of
cases is shown below: (i) edge dislocations travelling in the same plane, (ii) edge dislocations
travelling in adjacent planes, (iii) edge dislocations travelling in overlapping adjacent planes, (iv)
an edge dislocation encountering a vacancy, and (v) an edge dislocation emitting a vacancy.

(i)

(ii)

(iii)

J. B. Puthoff Cal Poly Pomona MTE 4050

 (iv)

(v)

The intersection of propagating dislocations, one edge and one screw, is shown below. When the
dislocations cross, jogs or kinks are introduced since certain regions are slipped with respect to
one another.

J. B. Puthoff Cal Poly Pomona MTE 4050

Forces on Dislocations
Dislocation motions results from internal and externally applied forces. The effect of these forces
on a volume of material can be described using the stress state σ. In the image below, a
dislocation (with Burgers vector b) passes through a volume of material dV = dAxdAydAz. The

element is effectively displaced by the passage of the dislocation, so we can calculate the plastic
work dW done on the material by the stress σ as

𝑑𝑊 = 𝑑𝐀 ∙ 𝛔 ∙ 𝐛
σ𝑥𝑥 τ𝑥𝑦 τ𝑥𝑧 𝑏𝑥
= [𝑑𝐴𝑥 𝑑𝐴𝑦 𝑑𝐴𝑧 ] [ τ𝑥𝑦 σ𝑦𝑦 τ𝑦𝑧 ] [𝑏𝑦 ]
τ𝑥𝑧 τ𝑦𝑧 σ𝑧𝑧 𝑏𝑧
= (σ𝑥𝑥 𝑑𝐴𝑥 + τ𝑥𝑦 𝑑𝐴𝑦 + τ𝑥𝑧 𝑑𝐴𝑧 )𝑏𝑥 + (τ𝑥𝑦 𝑑𝐴𝑥 + σ𝑦𝑦 𝑑𝐴𝑦 + τ𝑦𝑧 𝑑𝐴𝑧 )𝑏𝑦
+ (τ𝑥𝑧 𝑑𝐴𝑥 + τ𝑦𝑧 𝑑𝐴𝑦 + σ𝑧𝑧 𝑑𝐴𝑧 )𝑏𝑦 .

As expected, the work dW is a scalar. This result is not particularly interesting, but we can rewrite
the expression above to extract the force on the dislocation produced by the stress field. If we
consider a volume element lying in the dislocation’s slip plane with one edge r, another edge s,
and volume dV = (dr × ds) · b, then we obtain

𝑑𝑊 = (𝑑𝐀 ∙ 𝛔) ∙ 𝐛 = [(𝑑𝐫 × 𝑑𝐬) ∙ 𝛔] ∙ 𝐛 = 𝑑𝐫 ∙ [𝑑𝐬 × (𝛔 ∙ 𝐛)] = [𝑑𝐬 × (𝛔 ∙ 𝐛)] ∙ 𝑑𝐫 = 𝑑𝐅 ∙ 𝑑𝐫 ,

where dF = ds × (σ · b). (F is sometimes called the Peach-Kohler force.) Consider now the case of
a dislocation in the xy plane traveling in the x direction (shown below). The dislocation line is in
the y direction and the relevant length ds = dℓ. This gives

0 σ𝑥𝑥 τ𝑥𝑦 τ𝑥𝑧 𝑏 0 σ𝑥𝑥 𝑏

𝑑𝐅edge = 𝑑𝐬 × (𝛔 ∙ 𝐛) = [𝑑ℓ] × [ τ𝑥𝑦 σ𝑦𝑦 τ𝑦𝑧 ] [0] = 𝑑ℓ [1] × [ τ𝑥𝑦 𝑏 ] .
0 τ𝑥𝑧 τ𝑦𝑧 σ𝑧𝑧 0 0 τ𝑥𝑧 𝑏

J. B. Puthoff Cal Poly Pomona MTE 4050

The force per unit dislocation length is then

𝑑𝐅edge τ𝑥𝑧 𝑏
=[ 0 ] .
𝑑ℓ −σ𝑥𝑥 𝑏

Notice that this force has two components: τxzb is called the glide force and –σxxb is called the
climb force. The glide force acts in the slip plane of the dislocation in the direction of the Burgers
vector and drives the dislocation in this direction, producing slip. The climb force, related to the
compressive stress –σxx, pushes the dislocation upward, causing it to climb through the crystal to
the next plane above. (Since the climb process absorbs vacancies, an offsetting chemical
balancing force arises to counteract unrestricted climb.) For the screw case, take b = [b 0 0] and
ds = [–dℓ 0 0] (i.e., b is parallel to ℓ) to get

−1 σ𝑥𝑥 τ𝑥𝑦 τ𝑥𝑧 𝑏 −1 σ𝑥𝑥 𝑏 0

𝑑𝐅screw
= [ 0 ] × [ τ𝑥𝑦 σ𝑦𝑦 τ𝑦𝑧 ] [0] = [ 0 ] × [ τ𝑥𝑦 𝑏 ] = [ τ𝑥𝑧 𝑏 ] .
𝑑ℓ τ𝑥𝑧 τ𝑦𝑧 σ𝑧𝑧 0 −τ𝑥𝑦 𝑏
0 0 τ𝑥𝑧 𝑏

This means that the screw dislocation experiences the same force (= τ xzb) perpendicular to its
dislocation line as the edge dislocation: another way that the behavior of dislocations of different
types is analogous. The –τxy component also drives the dislocation’s motion; screw dislocations
can glide in multiple planes, unlike edge dislocations that are confined to a single one!

In general, a dislocation segment of length ℓ experiences a force of magnitude

𝐹 = τ𝑏 ,

where τ is the shear stress in the plane of the dislocation in the direction of the Burgers vector.

J. B. Puthoff Cal Poly Pomona MTE 4050

Dislocation Stress Fields
Dislocations induce elastic stress fields in a crystal they inhabit, and these stresses reflect the
distortion of the lattice in the vicinity of the dislocation. We can calculate these stress fields using
what we know about the geometry of dislocations and elasticity. Consider the screw dislocation
shown below. Because of the topology of the defect (i.e., a parking lot), by traversing a pathway
s = s(x, y) around the core of the dislocation, you accumulate displacement uz in the z direction.
The rate at which you travel upward is given by duz/ds. After a complete trip back to your original
position (x, y), you have moved a distance uz = b in the z direction.

This situation can be represented using the contour integral

𝑑𝑢𝑧
∮ 𝑑𝑠 = 𝑏 .
𝑑𝑠

To evaluate this integral, take a simple circular path of increasing angular displacement θ. Far
away from the core of the dislocation, the rate of upward travel will be uniform along this path,
i.e. duz/dθ = constant = C. This gives

2π 2π
𝑑𝑢𝑧 𝑑𝑢𝑧
𝑏=∮ 𝑑𝑠 = ∫ 𝑑θ = 𝐶 ∫ 𝑑θ = 2π𝐶 ,
𝑑𝑠 0 𝑑θ 0

or C = b/2π. This means that uz = bθ/2π (neglecting a constant of integration corresponding to

rigid-body motion).

J. B. Puthoff Cal Poly Pomona MTE 4050

We can now find the stress that corresponds to this amount of deformation using Hooke’s Law
and the strain-displacement relations:

𝜕𝑢𝑥 𝜕𝑢𝑧 𝜕𝑢𝑧 𝜕𝑢𝑧

τ𝑥𝑧 = 𝐺γ𝑥𝑧 = 𝐺 ( + ) = 𝐺 (0 + )=𝐺
𝜕𝑧 𝜕𝑥 𝜕𝑥 𝜕𝑥

(there is no displacement ux in the x direction). Since tan θ = y/x and (∂/∂u)arctan u = 1/(1 + u2),
we get

𝜕 𝑏 𝑦 𝐺𝑏 𝜕 𝑦 𝜕 𝐺𝑏 𝑦 1
τ𝑥𝑧 (𝑥, 𝑦) = 𝐺 ( arctan ) = ( ) (arctan 𝑢) = (− 2 ) [ ]
𝜕𝑥 2π 𝑥 2π 𝜕𝑥 𝑥 𝜕𝑢 2π 𝑥 1 + (𝑦/𝑥)2

or

𝐺𝑏 𝑦
τ𝑥𝑧 (𝑥, 𝑦) = − .
2π 𝑥 2 + 𝑦 2

Correspondingly

𝜕𝑢𝑦 𝜕𝑢𝑧 𝜕𝑢𝑧 𝐺𝑏 𝑥

τ𝑦𝑧 (𝑥, 𝑦) = 𝐺 ( + ) = 𝐺 (0 + ) = .
𝜕𝑧 𝜕𝑦 𝜕𝑦 2π 𝑥 + 𝑦 2
2

These are the only two components of the stress tensor (i.e., the stress field around a screw
dislocation is one of pure shear), and it does not depend on the z coordinate (i.e., it is symmetric
along the z axis).

0 0 τ𝑥𝑧 (𝑥, 𝑦)
𝛔screw (𝑥, 𝑦) = [ 0 0 τ𝑦𝑧 (𝑥, 𝑦)] .
τ𝑥𝑧 (𝑥, 𝑦) τ𝑦𝑧 (𝑥, 𝑦) 0

The edge dislocation is a bit different, with components

𝐺𝑏 𝑦(3𝑥 2 + 𝑦 2 ) 𝐺𝑏 𝑦(𝑥 2 − 𝑦 2 )
σ𝑥𝑥 (𝑥, 𝑦) = − , σ𝑦𝑦 (𝑥, 𝑦) = , and
2π(1 − ν) (𝑥 2 + 𝑦 2 )2 2π(1 − ν) (𝑥 2 + 𝑦 2 )2

𝐺𝑏 𝑥(𝑥 2 − 𝑦 2 )
τ𝑥𝑧 (𝑥, 𝑦) = .
2π(1 − ν) (𝑥 2 + 𝑦 2 )2

The stress component σzz = –ν(σxx +σyy) exists but is frequently ignored. So

σ𝑥𝑥 (𝑥, 𝑦) 0 τ𝑥𝑧 (𝑥, 𝑦)

𝛔edge (𝑥, 𝑦) = [ 0 σ𝑦𝑦 (𝑥, 𝑦) 0 ].
(𝑥,
τ𝑥𝑧 𝑦) 0 (𝑥,
σ𝑧𝑧 𝑦)

J. B. Puthoff Cal Poly Pomona MTE 4050

Energy of Dislocations
Because dislocations introduce strain into a lattice (as compared to a perfect, unstrained crystal),
they introduce strain energy into the system. Recall that the elastic energy contained within a
volume V of material can be represented as

⬚
𝑈el = ∭ 𝑑𝑈el .
𝑉

Expressed in term of stresses, the elastic energy increment is

1 1
𝑑𝑈el = [ (σ2 + σ2𝑦𝑦 + σ2𝑧𝑧 ) + τ2𝑥𝑦 + τ2𝑥𝑧 + τ2𝑦𝑧
2𝐺 2(1 + ν) 𝑥𝑥
ν
− (σ σ + σ𝑥𝑥 σ𝑧𝑧 + σ𝑦𝑦 σ𝑧𝑧 )] 𝑑𝑉 .
1 + ν 𝑥𝑥 𝑦𝑦

Since the screw dislocation has a stress field

𝐺𝑏 𝑦 𝐺𝑏 𝑥
τ𝑥𝑧 (𝑥, 𝑦) = − and τ𝑦𝑧 (𝑥, 𝑦) = ,
2π 𝑥 2 + 𝑦 2 2π 𝑥 2 + 𝑦 2

we can find

1 1 𝜈
𝑑𝑈el = [ (0 + 0 + 0) + 0 + τ2𝑥𝑧 + τ2𝑦𝑧 − (0 + 0 + 0)] 𝑑𝑉
2𝐺 2(1 + 𝜈) 1+𝜈
2 2
1 𝐺𝑏 𝑦 𝐺𝑏 𝑥
= [(− ) +( ) ] 𝑑𝑉
2𝐺 2π 𝑥 2 + 𝑦 2 2π 𝑥 2 + 𝑦 2
𝐺𝑏 2
= 2 2 𝑑𝑉 .
8π (𝑥 + 𝑦 2 )2

In a rectangular coordinate system, the volume element is dV = dxdydz. We can recalculate the
strain energy above using a cylindrical system, instead. This gives us a variable change x2 + y2 →
r2 and dxdydz → rdrdθdz, as shown below. After this change, we have

J. B. Puthoff Cal Poly Pomona MTE 4050

 𝐺𝑏 2 𝐺𝑏 2
𝑑𝑈el = 𝑟𝑑𝑟𝑑θ𝑑𝑧 = 𝑑𝑟𝑑θ𝑑𝑧 .
8π2 𝑟 2 8π2 𝑟

We can now integrate over this cylindrical region (see below) to find the total energy of the
dislocation. Since the stresses (and hence energy) do not vary in the θ or z directions

⬚ ℓ 2π ∞
𝐺𝑏 2 𝐺𝑏 2 ℓ ∞ 𝑑𝑟
𝑈el = ∭ 𝑑𝑈el = ∫ ∫ ∫ 𝑑𝑟 𝑑θ 𝑑𝑧 = ∫ ,
𝑉 0 0 0 8π2 𝑟 4π 0 𝑟

where ℓ is an arbitrary distance along the dislocation line/z direction. This permits us to recast
the energy of the dislocation U′screw as an energy per dislocation line length:

′
𝑑𝑈el 𝐺𝑏 2 ∞ 𝑑𝑟
𝑈screw = = ∫ .
𝑑ℓ 4π 0 𝑟

This final integration presents some problems, since the region of integration includes a
singularity at r = 0 and the integral diverges as the upper limit → ∞. There is no way to deal with
the singularity or divergence mathematically, but we can still perform the calculation if we invoke
our knowledge of the physics involved. The stress field was derived for a material that is
completely continuous, but we know that a continuum picture of matter breaks down at atomic
scales. I.e., the singularity is just a mathematical artifact. We can remove it by taking

𝑟0 ∞
′
𝐺𝑏 2 𝑑𝑟 𝑑𝑟 ′
𝐺𝑏 2 ∞ 𝑑𝑟
𝑈screw = (∫ +∫ ) = 𝑈core + ∫ ,
4π 0 𝑟 𝑟0 𝑟 4π 𝑟0 𝑟

J. B. Puthoff Cal Poly Pomona MTE 4050

where U′core has been estimated to be ~ Gb2/10 and r0 ~ 5b. Furthermore, we know that the upper
limit at ∞ is also physically unrealistic; rather, there is some physical cut-off distance R beyond
which the integral is effectively 0. This gives

∞ 𝑅 ∞ 𝑅
𝑑𝑟 𝑑𝑟 𝑑𝑟 𝑑𝑟 𝑅
∫ =∫ +∫ =∫ + 0 = ln .
𝑟0 𝑟 𝑟0 𝑟 𝑅 𝑟 𝑟0 𝑟 𝑟0

The final energy for the screw dislocation is

′ ′
𝐺𝑏 2 𝑅 𝐺𝑏 2 𝐺𝑏 2 𝑅
𝑈screw = 𝑈core + ln = + ln .
4π 𝑟0 10 4π 5𝑏

The energy for the edge dislocation is similar:

′ 𝐺𝑏 2 𝐺𝑏 2 𝑅
𝑈edge = + ln .
10 4π(1 − ν) 5𝑏

Note that this result also has the implication that dislocation energies U ∝ b2. This means that
dislocations tend to form in and adopt configurations that minimize b2. Consider the “reaction”
between two dislocations with Burgers vectors b1 and b2 to form a two-dislocation complex b12:

𝐛1 + 𝐛2 → 𝐛12 .

Initially, the system had energy U1 + U2 = b12 + b22. After the association, the system has energy
U12 = b122. The difference ΔU is

∆𝑈 = 𝑈12 − 𝑈1 − 𝑈2 = (𝐛12 + 𝐛1 ∙ 𝐛2 + 𝐛22 ) − 𝐛12 − 𝐛22 = 𝐛1 ∙ 𝐛2 .

Recognize that ΔU will be positive when b1 • b2 > 0 (indicating repulsion between dislocations
with like polarity) and will be negative when b1 • b2 < 0 (indicating attraction between
dislocations with opposite polarity).

Dislocation Structure and Density

Dislocations typically do not exist in a material in isolation. Rather, they form a network that
rearranges itself under the influence of an applied stress. An idealized picture of what this
network might look like is shown below. The network is comprised of a uniform array of
crisscrossing edge dislocation lines of alternating polarity. The spacing between dislocations is L.
This partitions the material into “subcrystals” of volume V = L3.

Each of these cells contains a certain amount of dislocation line. A cubic subcrystal has 12 edges,
each with dislocation line length L. Since each edge is shared among 4 neighboring subcrystals,
the total dislocation line for each cell is Ltot = 12L/4. We can now calculate a dislocation density ρ
as

J. B. Puthoff Cal Poly Pomona MTE 4050

 total dislocation line length 𝐿total 12𝐿/4 3
ρ= = = = 2.
subcrystal volume 𝑉 𝐿3 𝐿

Since the stress field of a dislocation exists essentially up to the point where it overlaps with
another, negating stress field, we can interpret the upper limit of integration R in the dislocation
energy using this picture. This gives

𝐿
𝑅= .
2

Accordingly

3
𝑅=√ = 0.866ρ−1/2 ≈ ρ−1/2 .
4ρ

This permits us to write

′
𝐺𝑏 2 𝐺𝑏 2 √3
𝑈screw = + ln ,
10 4π 5𝑏√4ρ

and

′ 𝐺𝑏 2 𝐺𝑏 2 √3
𝑈edge = + ln .
10 4π(1 − ν) 5𝑏√4ρ

A unified way to write the dislocation energy might be something like

J. B. Puthoff Cal Poly Pomona MTE 4050

 𝐺𝑏 2 𝐺𝑏 2 √3
𝑈′(α) = + (1 − ν cos 2 α) ln ,
10 4π(1 − 𝜈) 5𝑏√4ρ

where 0 ≤ α ≤ π/2. The α = 0 case corresponds to pure screw character, while α = π/2 corresponds
to pure edge character.

As an example, consider a specimen of pure, annealed copper (b = 0.25 nm, G = 48.3 GPa). In the
annealed state, the dislocation content is low: ρ = 108 cm–2 = 1012 m–2. We can now determine
(for dislocations of purely screw character)

′
𝐺𝑏 2 𝐺𝑏 2 √3
𝑈screw = + (1 − ν cos2 α) ln
10 4π(1 − ν) 5𝑏√4ρ
𝐺𝑏 2 𝐺𝑏 2 √3
= + (1 − ν cos2 0) ln
10 4π(1 − ν) 5(0.25 × 10−9 )√4 × 1012
2
= (0.1 + 0.042 ln 693)𝐺𝑏
= 0.62𝐺𝑏 2
= 1.9 × 10−9 J/m .

The intermediate result U′screw = 0.62Gb2 furnishes us with another estimate of the dislocation
energy. Since the ln function varies slowly over the range of dislocation densities that materials
exhibit and the “mixing parameter” α ≈ 50%, sometimes the first-order approximation

𝐺𝑏 2
𝑈′~
2

is used. This example calculation also leads us to an associated calculation: the total dislocation
energy Utotal contained within the specimen of volume V. Since U′screw is the dislocation energy
per unit dislocation line length

′
J m
𝑈total = 𝑈screw × ρ × 𝑉 = 1.5 × 10−9 × 1012 3 × 𝑉 = (1.5 × 103 J/m3 )𝑉 .
m m

I.e., if we know the dislocation density, we can determined how much elastic energy associated
with dislocations is stored with a piece of material. Roughly

𝐺𝑏 2
𝑈total ~ρ 𝑉 .
2

J. B. Puthoff Cal Poly Pomona MTE 4050