08.15.03 · stat-mech / topological-defects

Topological defects in ordered media

shipped3 tiersLean: none

Anchor (Master): Toulouse-Kleman 1976 (J. Phys. Lett. 37:L149); Mermin 1979; Nelson 'Defects and Geometry in Condensed Matter Physics' Ch. 1; KTHNY two-stage melting

Intuition Beginner

Take an ordered material: a stack of compass needles all leaning the same way, the atoms of a crystal sitting on a neat grid, or the rod-shaped molecules of a liquid-crystal display all lying parallel. Order means there is a preferred local arrangement that neighbors agree on. A defect is a spot where that agreement cannot be patched up: no matter how gently you nudge the surrounding region, one stubborn point or line is left where the order breaks down.

The companion vortex of the XY magnet 08.15.01 is the first example. Walk a loop around the vortex core and the needle direction turns through one full circle. You cannot iron this out by small adjustments, because the number of full turns you collect on the loop is a whole number, and a whole number cannot drift to zero by gentle changes. That integer is the defect's charge.

The same idea reappears across very different materials. A crystal has dislocations, where a row of atoms ends in the middle of the lattice. A liquid crystal has disclinations, where the molecular axis swings through a half-turn around a line. Each is a place the order cannot be smoothed away, and each carries a stubborn integer-like label that survives any gentle reshaping of the material.

The deep surprise is that one piece of bookkeeping predicts which defects a given material can host, using only the shape of its space of ground states. Counting the kinds of loops in that space tells you the kinds of defects, before you know anything about the forces.

Visual Beginner

The three panels show the three canonical defects. On the left, a vortex in a planar magnet: small arrows wind once around the core through a full circle. In the middle, a hedgehog in a three-dimensional magnet: arrows point radially outward from a single center like spines on a sea urchin, a defect at a point rather than along a line. On the right, a nematic disclination: the rod-shaped molecules swing through only a half-turn around the core line, because a rod pointing up looks the same as a rod pointing down.

The picture carries the main lesson. The vortex and the hedgehog both come from arrows on a sphere or circle, where a full turn is the smallest stubborn unit. The nematic is different: because its rods have no head or tail, a half-turn already brings the configuration back to itself, so the smallest stubborn unit is half as large. The shape of the local order, not the strength of the forces, decides what the smallest indestructible defect is.

Worked example Beginner

Count the charge of a defect by walking a loop and adding up how far the order rotates, then compare a magnet to a nematic.

Step 1. Start with a planar magnet, where the order is an arrow pointing somewhere in a full circle of directions. Walk a closed loop around a suspected defect and record the direction of the arrow as you go. When you return to the start, the arrow must point the same way it did when you left, so the total rotation must be a whole number of full turns: $0$ , one turn, two turns, and so on.

Step 2. Say you measure exactly one full turn, which is $360$ degrees. That is a charge-one vortex. A loop that collects zero turns has no defect inside it. The charge is the number of full turns, and it can only be a whole number, so it cannot slip to a different value when you jiggle the arrows.

Step 3. Now do the same in a nematic, where the order is a headless rod. A rod pointing at $0$ degrees is the same physical state as a rod pointing at $180$ degrees, since the rod has no front or back. Walk a loop and watch the rod axis. This time the rod can come back to itself after only a half-turn of $180$ degrees, because a half-turn maps the rod onto its indistinguishable flipped self.

Step 4. Put in numbers. In the magnet the allowed loop rotations are $0, 360, 720, \dots$ degrees, so charges are $0, 1, 2, \dots$ . In the nematic the allowed loop rotations are $0, 180, 360, 540, \dots$ degrees, so charges come in halves: $0, \frac{1}{2}, 1, \frac{3}{2}, \dots$ . The nematic has a defect of charge one-half that the magnet simply cannot have.

What this tells us: the menu of possible defects is set by what counts as "the same state" for the local order. Headless rods allow a half-charge defect; headed arrows do not. The geometry of the order parameter, decided before any energy is computed, fixes which defects the material can carry.

Check your understanding Beginner

Exercise (easy, multiple choice).

What makes a defect in an ordered medium impossible to remove by small local adjustments?

A. The material is too cold to rearrange B. It carries a whole-number (or fixed-fraction) charge measured by walking a loop around it C. There are no atoms at the defect core D. The defect is glued to the boundary of the sample

Hint

Think about why the count of full turns of the order around a loop cannot drift continuously to zero.

Answer

B. It carries a whole-number (or fixed-fraction) charge measured by walking a loop around it. The charge is the number of turns the local order makes around a loop, and that count is restricted to integers (or a fixed set of fractions for a nematic), so it cannot move to zero under small smooth changes. Feedback-correct: yes, the discreteness of the charge is what protects the defect. Feedback-wrong: temperature, missing atoms, and boundaries are not what makes the charge stubborn; the discreteness of the loop count is.

Exercise (easy, multiple choice).

A hedgehog in a three-dimensional magnet, with arrows pointing radially outward from a center, is best described as which kind of defect?

A. A line defect, stable because of loops around it B. A point defect, where the arrows cover all directions of a sphere C. A surface where two crystals meet D. A defect that can always be smoothed away

Hint

The arrows of the hedgehog sweep out every direction on a sphere, and the defect sits at one point.

Answer

B. A point defect, where the arrows cover all directions of a sphere. The hedgehog lives at a single point in three dimensions, and the surrounding arrows wrap once around the sphere of directions, which is its stubborn integer charge. Feedback-correct: yes, the hedgehog is a point defect counted by how many times the arrows cover the sphere. Feedback-wrong: it is not a line defect or a surface, and its wrapping number cannot be smoothed to zero.

Formal definition Intermediate+

Let a medium have order described by a field $n (x)$ taking values in the order-parameter manifold $M$ , the space of degenerate ground states. When a symmetry group $G$ breaks to a residual subgroup $H$ , this manifold is the coset space $M = G / H$ . For the planar (XY) magnet the order is a direction in a plane and $M = S^{1}$ ; for the Heisenberg magnet it is a direction in space and $M = S^{2}$ ; for the uniaxial nematic it is a headless axis, a point of the sphere with antipodes identified, so $M = RP^{2} = S^{2} / Z_{2}$ .

Away from its core, a defect is a smooth field. Surround a defect by a sphere $S^{k}$ in the sample that does not touch the core; restricting $n$ to that sphere gives a map $S^{k} \to M$ . Two such maps describe the same defect when one deforms continuously into the other, so a defect is a homotopy class of maps $S^{k} \to M$ , an element of the homotopy group $π_{k} (M)$ . The dimension $k$ is fixed by the geometry: enclosing a defect by a sphere $S^{k}$ inside a $d$ -dimensional sample with a $p$ -dimensional defect requires $k = d - p - 1$ .

The basic dictionary follows ^{[Mermin 1979]}:

Line defects in $3$ D and point defects in $2$ D are enclosed by a loop $S^{1}$ , hence classified by the fundamental group $π_{1} (M)$ 03.12.00.
Point defects in $3$ D are enclosed by a sphere $S^{2}$ , hence classified by $π_{2} (M)$ .
Domain walls (codimension-one) are enclosed by a pair of points $S^{0}$ , classified by $π_{0} (M)$ , the set of connected components.

The worked instances are these. The XY model has $M = S^{1}$ and $π_{1} (S^{1}) = Z$ : vortices labeled by an integer winding number $q$ , exactly the defect of 08.15.01. The Heisenberg magnet has $M = S^{2}$ ; since $π_{1} (S^{2}) = 0$ there are no stable line defects (any vortex line can be unwound by tilting the spin into the third dimension, the "escape into the third dimension"), while $π_{2} (S^{2}) = Z$ gives stable point defects, the hedgehogs, labeled by an integer wrapping number. The nematic has $M = RP^{2}$ , whose fundamental group is

π_{1} (RP^{2}) = Z_{2} = {0, 1},

so it has exactly one species of stable disclination line, of half-integer character, and two such lines can annihilate. Its point defects obey $π_{2} (RP^{2}) = π_{2} (S^{2}) = Z$ , the nematic hedgehogs.

Counterexamples to common slips

The defect charge is not a property of the field everywhere; it is the homotopy class of the field restricted to a surrounding sphere, evaluated in $M$ , not in the physical sample. Confusing the two leads to counting defects with the wrong group.
$π_{1} (S^{2}) = 0$ does not say the Heisenberg magnet has no defects; it says it has no stable line defects. The stable defects there are points, governed by $π_{2}$ , a different group entirely.
The nematic disclination has charge in $Z_{2}$ , not in $Z$ . Writing its strength as a half-integer is the physicist's bookkeeping for the same fact; the homotopy-invariant content is only "present or absent", since $1 + 1 = 0$ in $Z_{2}$ .

Key theorem with proof Intermediate+

The organizing statement is the homotopy classification itself, which lifts the ad-hoc winding number of 08.15.01 to a uniform rule.

Theorem (homotopy classification of defects). Let an ordered medium have order-parameter manifold $M$ . Stable defects of codimension $k + 1$ are in one-to-one correspondence with elements of the homotopy group $π_{k} (M)$ , with the class of the constant map corresponding to the absence of a defect. In particular, line defects in three dimensions and point defects in two dimensions are classified by $π_{1} (M)$ , and point defects in three dimensions by $π_{2} (M)$ .

Proof. Fix a defect of dimension $p$ in a $d$ -dimensional sample and choose a sphere $Σ ≅ S^{k}$ with $k = d - p - 1$ that links the defect once and avoids its core. On $Σ$ the field $n$ is smooth and defines a continuous map $f : S^{k} \to M$ . Two configurations of the medium that can be deformed into one another without passing the core through $Σ$ restrict to homotopic maps on $Σ$ , and conversely a homotopy of $f$ extends to a deformation of the field in a collar neighborhood of $Σ$ . Hence the defect determines, and is determined up to such deformation by, the homotopy class $[f] \in π_{k} (M)$ once a basepoint and orientation of $Σ$ are fixed.

The class is independent of the choice of linking sphere: any two spheres linking the defect once are themselves homotopic in the complement of the core, and the field carries that homotopy to a homotopy of the restricted maps. The constant map, where $n$ is uniform on $Σ$ , extends to a defect-free field on the disk bounded by $Σ$ , so the identity element of $π_{k} (M)$ is exactly the removable (absent) defect. Group multiplication in $π_{k} (M)$ corresponds to merging two defects enclosed by a common larger sphere, which establishes the correspondence as one of pointed sets, and of groups when $π_{k}$ is a group. Specializing $k = 1$ gives line defects in $3$ D and point defects in $2$ D classified by $π_{1} (M)$ , and $k = 2$ gives point defects in $3$ D classified by $π_{2} (M)$ . $□$

Bridge. This classification builds toward the energetic theory of the Kosterlitz-Thouless transition 08.15.01 and appears again in the homotopy backbone of every later defect computation; the foundational reason a single group controls a material's whole defect spectrum is that the protected information in a defect is purely the homotopy class of a map into $M$ , nothing more. This is exactly the structure that the XY winding number $q \in π_{1} (S^{1}) = Z$ exhibited in the special case, and the present theorem generalises it to arbitrary $M$ . The map from physical defect to homotopy class is dual to the construction that reads a defect off a given class by prescribing the field on a linking sphere and filling in. The central insight is that the linking sphere $S^{k}$ converts a $d$ -dimensional embedding problem into the pure homotopy of $π_{k} (M)$ , and putting these together with the coset description $M = G / H$ yields the defect spectrum directly from the symmetry-breaking pattern, which is the bridge from group theory to the observed catalogue of vortices, hedgehogs, and disclinations.

Exercises Intermediate+

Exercise 2 (easy, multiple choice).

The Heisenberg magnet has order-parameter manifold $M = S^{2}$ . Which statement about its defects is correct?

A. It has stable line defects because $π_{1} (S^{2}) = Z$ B. It has no stable line defects because $π_{1} (S^{2}) = 0$ , but stable point defects because $π_{2} (S^{2}) = Z$ C. It has only domain walls D. It has the same defects as the XY magnet

Hint

The sphere is simply connected but not $2$ -connected.

Answer

B. Since $π_{1} (S^{2}) = 0$ , any line defect (vortex line) can be smoothed away by tilting spins into the third direction; since $π_{2} (S^{2}) = Z$ , there are stable point defects, the hedgehogs, with integer charge. Feedback-correct: yes, $S^{2}$ kills line defects but supports point defects. Feedback-wrong: $π_{1} (S^{2}) = 0$ , not $Z$ ; the XY case is $S^{1}$ , a different manifold. Rubric: full credit for B.

Exercise 3 (medium, short-answer).

Explain why the uniaxial nematic supports a stable disclination of strength one-half whereas the XY magnet does not, in terms of the order-parameter manifolds and their fundamental groups.

Hint

Compare $M = RP^{2}$ with $M = S^{1}$ and identify the antipodal identification.

Answer

The XY magnet has a headed direction, $M = S^{1}$ , with $π_{1} (S^{1}) = Z$ ; the smallest stable defect is a full winding, charge $1$ . The nematic has a headless axis, so antipodal directions are identified and $M = RP^{2} = S^{2} / Z_{2}$ . A path on $S^{2}$ from a point to its antipode projects to a closed loop in $RP^{2}$ that is not contractible, generating $π_{1} (RP^{2}) = Z_{2}$ . Physically this loop rotates the molecular axis by a half-turn, the strength- $\frac{1}{2}$ disclination. It is stable because $1 \neq = 0$ in $Z_{2}$ , but two of them annihilate since $1 + 1 = 0$ . The XY manifold has no antipodal identification, so no half-strength loop exists. Rubric: full credit for naming $M = RP^{2}$ , the antipodal lift, $π_{1} = Z_{2}$ , and the half-turn interpretation.

Exercise 4 (medium, symbolic).

Using the covering $S^{2} \to RP^{2}$ with deck group $Z_{2}$ and the fact that $S^{2}$ is simply connected, compute $π_{1} (RP^{2})$ from the covering-space exact sequence.

Hint

For a covering $\tilde{X} \to X$ with $\tilde{X}$ simply connected, $π_{1} (X)$ equals the deck transformation group.

Answer

The double cover $p : S^{2} \to RP^{2}$ identifies antipodes, with deck group $Z_{2}$ . Covering-space theory gives the exact sequence $π_{1} (S^{2}) \to π_{1} (RP^{2}) \to Z_{2} \to π_{0} (S^{2})$ . With $π_{1} (S^{2}) = 0$ (simply connected) and $π_{0} (S^{2}) = 0$ (connected), the middle map is an isomorphism, so $π_{1} (RP^{2}) ≅ Z_{2}$ . Equivalently, a universal cover that is simply connected has $π_{1} (base) =$ the deck group $= Z_{2}$ . The generator is the image of any path joining antipodal points of $S^{2}$ . Rubric: full credit for using simple-connectedness of $S^{2}$ and the deck group $Z_{2}$ to conclude $π_{1} (RP^{2}) = Z_{2}$ .

Exercise 5 (medium, short-answer).

A biaxial nematic breaks rotational symmetry down to a discrete group, with order-parameter manifold $M = S O (3) / D_{2}$ whose fundamental group is the quaternion group $Q_{8}$ , which is non-abelian. What new physical phenomenon, absent for abelian $π_{1}$ , does non-commutativity of the defect charges produce?

Hint

When $π_{1}$ is non-abelian, the combination of two defects can depend on the order and the path, and charges are classified by conjugacy classes.

Answer

When $π_{1} (M)$ is non-abelian, defect charges no longer simply add. Two defect lines can become topologically entangled so that they cannot pass through one another: the obstruction is the commutator $[a, b] = ab a^{- 1} b^{- 1}$ of their charges, which differs from the identity when they fail to commute. The physically meaningful invariant of a single line becomes its conjugacy class, not its individual group element, because dragging one defect around another conjugates its charge $a \mapsto ba b^{- 1}$ . This gives entanglement and non-commutative combination rules with no analogue in the abelian XY or nematic-uniaxial cases. Rubric: full credit for naming entanglement/non-commuting combination and the conjugacy-class classification of charges.

Exercise 6 (hard, symbolic).

Show that the hedgehog point defect of the Heisenberg magnet ( $M = S^{2}$ ) has an integer charge given by the degree of the map $n : S^{2} \to S^{2}$ on an enclosing sphere, and write that degree as a surface integral.

Hint

$π_{2} (S^{2}) = Z$ is detected by the Brouwer degree, computable as the normalized signed area swept on the target sphere.

Answer

A point defect in three dimensions is enclosed by a sphere $Σ = S^{2}$ , and the order field restricts to $n : S^{2} \to S^{2}$ . Its homotopy class in $π_{2} (S^{2}) = Z$ is the Brouwer degree, the number of times $n$ wraps the target sphere. The degree is the normalized signed area swept on the target,

q = \frac{1}{4 π} \int_{Σ} n \cdot (\frac{\partial n}{\partial u} \times \frac{\partial n}{\partial v}) d u d v,

where $(u, v)$ parametrize $Σ$ and $4 π$ is the area of the unit target sphere. For the radial hedgehog $n = \hat{r}$ the map is the identity, the integrand integrates to the full sphere area $4 π$ , and $q = 1$ . The integral is an integer for any smooth $n$ because it counts signed coverings of the target, which is the defining property of the degree. Rubric: full credit for identifying $π_{2} (S^{2}) = Z$ with the degree, the surface-integral formula, and $q = 1$ for the radial hedgehog.

Exercise 7 (hard, short-answer).

In two-dimensional melting (KTHNY), the solid first loses translational order by unbinding dislocations and only later loses orientational order by unbinding disclinations. Using the order-parameter structure, explain why dislocations and disclinations are distinct defects and why their unbinding happens at two separate temperatures rather than one.

Hint

A two-dimensional crystal has both translational order (broken translations, defects = dislocations) and bond-orientational order (broken rotations, defects = disclinations); a disclination is energetically much costlier than a dislocation.

Answer

A two-dimensional crystal breaks both translational and rotational symmetry, so it carries two order parameters. Dislocations are the defects of the translational order (a Burgers vector, the broken-translation charge), while disclinations are the defects of the bond-orientational order (a rotation by a lattice angle). A disclination has an energy that grows with system size more steeply than a dislocation, since a dislocation is a tightly bound disclination pair. As temperature rises, dislocations unbind first, destroying translational order but leaving bond-orientational order intact: this is the hexatic phase, with short-range positional order and quasi-long-range orientational order. At a higher temperature the disclinations themselves unbind, destroying orientational order and giving the isotropic liquid. Two distinct defect species with two distinct binding energies produce two distinct Kosterlitz-Thouless-type transitions. Rubric: full credit for distinguishing dislocation (translational) from disclination (orientational) defects, the hexatic intermediate, and the two-stage unbinding.

Advanced results Master

The classification by $π_{k} (M)$ is the start; the algebraic structure of these groups encodes how defects combine, merge, and become trapped. The systematic theory is due to Toulouse and Kléman ^{[Toulouse-Kleman 1976]} and Mermin ^{[Mermin 1979]}.

Theorem (combination rule from group multiplication). Let two defects of the same codimension, classified by $α, β \in π_{k} (M)$ , be enclosed by a common sphere. The composite carries the product $α β \in π_{k} (M)$ . For $k \geq 2$ the group $π_{k} (M)$ is abelian, so composite charges add commutatively; for $k = 1$ the group $π_{1} (M)$ may be non-abelian, and the conjugacy class of a defect, not its individual element, is the path-independent invariant.

The non-abelian case has no analogue in the XY or uniaxial-nematic worlds, where $π_{1}$ is $Z$ or $Z_{2}$ . For the biaxial nematic $M = S O (3) / D_{2}$ with $π_{1} (M) = Q_{8}$ the quaternion group, two disclination lines with non-commuting charges $a, b$ cannot cross: the obstruction is the commutator $ab a^{- 1} b^{- 1} \neq = e$ , and dragging one line around another conjugates its charge. The physical defect species are therefore the five conjugacy classes of $Q_{8}$ , and the entanglement of defect lines is a direct readout of group non-commutativity.

Theorem (action of $π_{1}$ on higher defects). The fundamental group $π_{1} (M)$ acts on $π_{2} (M)$ . A point defect of class $γ \in π_{2} (M)$ transported around a line defect of class $σ \in π_{1} (M)$ returns with charge $σ \cdot γ$ . When this action differs from the identity the integer point-defect charge is only defined up to the $π_{1}$ -orbit.

For the uniaxial nematic, $π_{1} (RP^{2}) = Z_{2}$ acts on $π_{2} (RP^{2}) = Z$ by sign reversal: carrying a hedgehog of charge $+ 1$ around a disclination returns it as charge $- 1$ . The hedgehog charge is therefore well-defined only modulo this sign, so in the presence of disclinations the nematic point-defect charge lives in $Z / (\pm) = {0, 1, 2, \dots}$ as an unsigned integer.

Theorem (KTHNY two-stage melting; Halperin-Nelson-Young). A two-dimensional crystal melts through two successive Kosterlitz-Thouless transitions. At $T_{m}$ the dislocations (defects of the translational order, Burgers-vector charges) unbind, destroying quasi-long-range translational order and producing the hexatic phase, which retains quasi-long-range bond-orientational order. At a higher $T_{i} > T_{m}$ the disclinations (defects of the orientational order) unbind, destroying orientational order and giving the isotropic liquid.

The hexatic phase is the signature prediction: an intermediate phase with no positional order but exponentially-correlated sixfold bond-orientational order, separated from both solid and liquid by continuous transitions of the Kosterlitz-Thouless universality class 08.15.01. The translational stiffness undergoes a universal jump at $T_{m}$ and the orientational (Frank) stiffness a universal jump at $T_{i}$ , each set by the renormalization-group fixed point of its own vortex gas, the dislocation gas below and the disclination gas above. The theory was developed by Halperin and Nelson ^{[Halperin-Nelson 1978]} and by Young ^{[Young 1979]}, whence the acronym KTHNY.

Synthesis. The homotopy theory of defects is the foundational reason a single discrete invariant survives in a continuously deformable medium, and it is exactly the structure that the Kosterlitz-Thouless vortex made concrete in $π_{1} (S^{1}) = Z$ 08.15.01. The central insight is that the order-parameter manifold $M = G / H$ carries the entire defect spectrum in its homotopy groups: $π_{0}$ for walls, $π_{1}$ for lines, $π_{2}$ for points, with the linking-sphere dimension $k = d - p - 1$ selecting which group acts. Putting these together, the abelian higher groups give additive point charges while the possibly non-abelian $π_{1}$ gives entanglement and conjugacy-class charges, and the action of $π_{1}$ on $π_{2}$ generalises the bare integer hedgehog charge to a charge defined only up to a group orbit. This pattern recurs and is dual to the symmetry-breaking description: every entry in the catalogue of vortices, hedgehogs, and uniaxial and biaxial disclinations is read directly off $π_{k} (G / H)$ , and the bridge to thermodynamics is that the unbinding of each defect species, weighted by its homotopy charge, drives a transition of Kosterlitz-Thouless type, culminating in the two-stage KTHNY melting where dislocation and disclination unbinding are two separate homotopy sectors flowing under two separate renormalization groups.

Full proof set Master

Proposition (vortex charge is the winding number, $π_{1} (S^{1}) = Z$ ). For $M = S^{1}$ the line/point defects are classified by an integer, the winding number, recovering the XY vortex of 08.15.01.

Proof. The order field on a linking circle is a map $f : S^{1} \to S^{1}$ . Lift $f$ along the universal cover $R \to S^{1}$ , $t \mapsto e^{2 π i t}$ , to a path $\tilde{f} : [0, 1] \to R$ with $f = e^{2 π i \tilde{f}}$ . Since $f$ is a loop, $\tilde{f} (1) - \tilde{f} (0) = q \in Z$ , and $q$ is invariant under homotopy of $f$ because the lift varies continuously and $q$ is integer-valued. The assignment $[f] \mapsto q$ is a homomorphism onto $Z$ (concatenation adds winding) and injective (equal winding numbers give homotopic lifts rel endpoints, hence homotopic loops), so $π_{1} (S^{1}) = Z$ . The integer $q$ is the circulation $\frac{1}{2 π} \oint \nabla θ \cdot d ℓ$ of 08.15.01. $□$

Proposition (no stable line defects in the Heisenberg magnet). For $M = S^{2}$ , $π_{1} (S^{2}) = 0$ , so every line defect is removable.

Proof. A line defect restricts to a map $f : S^{1} \to S^{2}$ . The image of a circle is a closed curve on the sphere; since $S^{2}$ is simply connected, any loop bounds a disk on the sphere, and contracting the loop along that disk is a homotopy of $f$ to a constant map. A constant map on the linking circle extends to a uniform field on the bounding disk in the sample, i.e. a defect-free configuration. Hence the line defect carries the identity class and can be smoothed away; physically this is the escape of the spin into the third dimension. $□$

Proposition (nematic disclination, $π_{1} (RP^{2}) = Z_{2}$ ). The uniaxial nematic has exactly one stable line-defect species beyond the removable one, of order two.

Proof. The double cover $p : S^{2} \to RP^{2}$ identifies antipodes and has deck group $Z_{2}$ . A loop $γ$ in $RP^{2}$ lifts to a path $\tilde{γ}$ in $S^{2}$ ; the lift either closes ( $\tilde{γ} (1) = \tilde{γ} (0)$ ) or ends at the antipode ( $\tilde{γ} (1) = - \tilde{γ} (0)$ ). The class in $π_{1} (RP^{2})$ is detected by which of the two occurs, since $S^{2}$ is simply connected and a closed lift is contractible while an antipodal lift is not. Composition multiplies these $Z_{2}$ -valued endpoints, so $π_{1} (RP^{2}) = Z_{2}$ . The generating element is the strength- $\frac{1}{2}$ disclination; two of them compose to a closing lift, the removable class, so $1 + 1 = 0$ . $□$

Proposition (hedgehog charge is the degree, $π_{2} (S^{2}) = Z$ ). Point defects of the Heisenberg magnet carry an integer charge equal to the Brouwer degree of the enclosing map.

Proof. A point defect restricts to $n : S^{2} \to S^{2}$ . By the Hopf degree theorem, two maps $S^{2} \to S^{2}$ are homotopic if and only if they have equal Brouwer degree, and every integer is realized, so $π_{2} (S^{2}) = Z$ with the degree as the isomorphism. The degree equals the signed count of preimages of a regular value, equivalently the normalized integral $\frac{1}{4 π} \int_{S^{2}} n \cdot (\partial_{u} n \times \partial_{v} n) d u d v$ . For the radial field $n = \hat{r}$ this is the identity map of degree $1$ , the elementary hedgehog. $□$

Proposition ( $π_{1}$ acts on $π_{2}$ ; nematic hedgehog sign). In the uniaxial nematic the disclination $σ \in π_{1} (RP^{2}) = Z_{2}$ acts on the hedgehog charge $γ \in π_{2} = Z$ by $σ \cdot γ = - γ$ for the generating $σ$ .

Proof. Transporting a point defect around a loop $σ$ acts on $π_{2} (M)$ through the standard $π_{1}$ -action on higher homotopy groups, realized here by the deck transformation. The generating element of $π_{1} (RP^{2})$ is the antipodal map $a : S^{2} \to S^{2}$ , $n \mapsto - n$ , which has degree $(- 1)^{2 + 1} = - 1$ . Composing the enclosing map $n$ with $a$ multiplies its degree by $- 1$ , so $σ \cdot γ = - γ$ . Hence a hedgehog carried once around a disclination reverses sign, and the hedgehog charge is well-defined only up to sign in the presence of disclinations. $□$

The KTHNY two-stage melting statement is established in Halperin-Nelson ^{[Halperin-Nelson 1978]} and Young ^{[Young 1979]} through the renormalization-group flow of the dislocation and disclination Coulomb gases; the present unit states the result and locates each stage in its homotopy sector, with the quantitative recursion relations carried over from 08.15.01.

Connections Master

The Kosterlitz-Thouless transition 08.15.01. The XY vortex of that unit is the case $M = S^{1}$ , $π_{1} (S^{1}) = Z$ , of the present classification, and its energy-entropy and renormalization-group analysis is the thermodynamic content that homotopy theory leaves open. The homotopy charge says which defects exist; the KT analysis says when they unbind. The two are complementary, and the KTHNY two-stage melting discussed here is the direct multi-defect generalization of the single-vortex KT flow.
Disclinations and two-dimensional melting 08.15.02. The dislocation and disclination defects whose homotopy classes are catalogued here are the dynamical actors of two-dimensional melting; that unit carries out the quantitative renormalization-group treatment of their successive unbinding, the hexatic phase, and the universal stiffness jumps, with the defect identities supplied by the classification proved here.
The fundamental group 03.12.00. The line-defect charge is literally an element of $π_{1} (M)$ , and every structural fact used here, conjugacy-class invariance of charges, the deck-group computation of $π_{1} (RP^{2}) = Z_{2}$ , and the action of $π_{1}$ on higher groups, is a theorem of that unit's homotopy theory applied to the order-parameter manifold. The Seifert-van Kampen theorem 03.12.09 of that block is the tool that computes $π_{1} (G / H)$ for the coset manifolds appearing as order-parameter spaces.

Historical & philosophical context Master

The recognition that defects in ordered media are classified by homotopy groups of the order-parameter space emerged in the mid-1970s from condensed-matter and field theory simultaneously. Gérard Toulouse and Maurice Kléman gave the classification scheme in 1976 in Journal de Physique Lettres 37 (1976) L149, "Principles of a classification of defects in ordered media" ^{[Toulouse-Kleman 1976]}, identifying the order-parameter manifold $V = G / H$ and the role of $π_{k} (V)$ for defects of each codimension, and applying it to superfluids, magnets, and liquid crystals. Independent and closely related formulations were given by G. E. Volovik and V. P. Mineev and by Louis Michel. N. David Mermin's 1979 review in Reviews of Modern Physics 51 (1979) 591, "The topological theory of defects in ordered media" ^{[Mermin 1979]}, consolidated the theory, worked the examples of the XY model, Heisenberg magnet, uniaxial and biaxial nematics, and superfluid helium-3, and treated the non-abelian and higher-homotopy refinements including the action of $π_{1}$ on $π_{2}$ .

The two-dimensional melting application was developed by Bert Halperin and David Nelson in 1978 in Physical Review Letters 41 (1978) 121 ^{[Halperin-Nelson 1978]} and by A. Peter Young in 1979 in Physical Review B 19 (1979) 1855 ^{[Young 1979]}, extending the Kosterlitz-Thouless dislocation theory of Kosterlitz and Thouless to a two-stage scenario with an intermediate hexatic phase, the KTHNY theory. The hexatic phase was subsequently observed in two-dimensional colloidal crystals and in liquid-crystal films, and the homotopy classification became the standard organizing framework for defects across superfluids, superconductors, liquid crystals, and cosmological field theories, where the same $π_{k} (G / H)$ bookkeeping classifies cosmic strings, monopoles, and domain walls.

Bibliography Master

@article{ToulouseKleman1976,
  author  = {Toulouse, G{\'e}rard and Kl{\'e}man, Maurice},
  title   = {Principles of a Classification of Defects in Ordered Media},
  journal = {J. Phys. Lett. (Paris)},
  volume  = {37},
  year    = {1976},
  pages   = {L149--L151}
}

@article{Mermin1979,
  author  = {Mermin, N. David},
  title   = {The Topological Theory of Defects in Ordered Media},
  journal = {Rev. Mod. Phys.},
  volume  = {51},
  year    = {1979},
  pages   = {591--648}
}

@article{HalperinNelson1978,
  author  = {Halperin, B. I. and Nelson, David R.},
  title   = {Theory of Two-Dimensional Melting},
  journal = {Phys. Rev. Lett.},
  volume  = {41},
  year    = {1978},
  pages   = {121--124}
}

@article{Young1979,
  author  = {Young, A. P.},
  title   = {Melting and the Vector Coulomb Gas in Two Dimensions},
  journal = {Phys. Rev. B},
  volume  = {19},
  year    = {1979},
  pages   = {1855--1866}
}

@article{VolovikMineev1977,
  author  = {Volovik, G. E. and Mineev, V. P.},
  title   = {Investigation of Singularities in Superfluid {He}$^3$ and Liquid Crystals by Homotopic Topology Methods},
  journal = {Sov. Phys. JETP},
  volume  = {45},
  year    = {1977},
  pages   = {1186--1196}
}

@book{NelsonDefects,
  author    = {Nelson, David R.},
  title     = {Defects and Geometry in Condensed Matter Physics},
  publisher = {Cambridge University Press},
  year      = {2002}
}

@book{KardarFields,
  author    = {Kardar, Mehran},
  title     = {Statistical Physics of Fields},
  publisher = {Cambridge University Press},
  year      = {2007}
}

@book{Kleman2003,
  author    = {Kl{\'e}man, Maurice and Lavrentovich, Oleg D.},
  title     = {Soft Matter Physics: An Introduction},
  publisher = {Springer},
  year      = {2003}
}

Prerequisites

08.15.01

Tier anchors

beginner: Kardar Ch. 8 vortex / defect picture; Mermin 1979 RMP figures of vortices, hedgehogs, disclinations
intermediate: Mermin 1979 (Rev. Mod. Phys. 51:591) homotopy classification; Kardar Ch. 8
master: Toulouse-Kleman 1976 (J. Phys. Lett. 37:L149); Mermin 1979; Nelson 'Defects and Geometry in Condensed Matter Physics' Ch. 1; KTHNY two-stage melting

References

raw/pdfs/kardar-1-3-laws-entropy-potentials.pdf · Kardar statistical physics lecture notes (anchor for conventions)
quantum-well · statistical physics index
Mermin — The topological theory of defects in ordered media · Rev. Mod. Phys. 51 (1979) 591
Toulouse and Kleman — Principles of a classification of defects in ordered media · J. Phys. Lett. (Paris) 37 (1976) L149
Halperin and Nelson — Theory of two-dimensional melting · Phys. Rev. Lett. 41 (1978) 121
Young — Melting and the vector Coulomb gas in two dimensions · Phys. Rev. B 19 (1979) 1855

Estimated time

beginner: 18m
intermediate: 42m
master: 80m