Skip to content

Diamagnetism, Paramagnetism and Ferromagnetism

CSIR NET Part C magnetism: Bohr–van Leeuwen, Larmor and Landau diamagnetism, Curie and Brillouin, Pauli paramagnetism, Weiss mean field, Curie–Weiss law.

  • 20 min read
  • Updated
On this page · 16 sections
  1. Prerequisites
  2. The Hook: A Frog, a Fridge Magnet and a Theorem That Says Neither Should Work
  3. The Classification at a Glance
  4. The Bohr–van Leeuwen Theorem
  5. Diamagnetism
  6. Paramagnetism of Localised Moments
  7. Pauli Paramagnetism of Conduction Electrons
  8. Ferromagnetism: the Weiss Mean Field
  9. Antiferromagnetism and Ferrimagnetism
  10. Memory Tricks
  11. Don’t Confuse
  12. Common Mistakes
  13. Practice
  14. Quick Revision Table
  15. Key Takeaways
  16. What’s Next

Prerequisites

The Hook: A Frog, a Fridge Magnet and a Theorem That Says Neither Should Work

The Classification at a Glance

In SI, \( \mathbf M = \chi\mathbf H \) and \( \mathbf B = \mu_0(\mathbf H + \mathbf M) \); \( \chi \) is dimensionless. For weak magnetism \( \mathbf B \approx \mu_0\mathbf H \), so \( M \approx \chi B/\mu_0 \).

TypeSign and size of \( \chi \)Temperature dependenceOrigin
Diamagnetic (insulators)negative, about \( -10^{-6} \) to \( -10^{-5} \)independent of \( T \)orbital response of closed shells (Larmor)
Landau diamagnetismnegative, \( -\frac{1}{3}\chi_P \) for free electronsnearly independentorbital quantisation of conduction electrons
Superconductor\( -1 \) (perfect diamagnet)below \( T_c \)Meissner effect
Curie paramagnetpositive, about \( 10^{-5} \) to \( 10^{-3} \)\( \propto 1/T \)permanent moments of partly filled shells
Pauli paramagnetpositive, smallnearly independentspin of conduction electrons near \( E_F \)
Ferromagnethuge, field-history dependentspontaneous \( M \) below \( T_c \); \( C/(T - \theta) \) aboveexchange aligns moments
Antiferromagnetpositive, small\( C/(T + \theta) \) above \( T_N \); peak at \( T_N \)exchange anti-aligns sublattices
Ferrimagnetlargespontaneous \( M \) below \( T_c \)unequal antiparallel sublattices

The Bohr–van Leeuwen Theorem

For classical charges, a magnetic field enters the Hamiltonian only through \( \mathbf p \to \mathbf p - q\mathbf A(\mathbf r) \). In the classical partition function the momentum integral runs over all of \( \mathbb R^3 \), so the shift \( \mathbf p' = \mathbf p - q\mathbf A \) removes \( \mathbf A \) completely:

$$Z(B) = \int d^3r\,d^3p\;e^{-\beta[(\mathbf p - q\mathbf A)^2/2m + V]} = Z(0)$$

Hence \( F \) is independent of \( B \) and \( M = -\partial F/\partial B = 0 \). Diamagnetism and paramagnetism both require quantum mechanics. (The classical Langevin result for paramagnets assumes fixed-length permanent dipoles, which is already a quantum ingredient.)

Diamagnetism

Larmor (atomic) diamagnetism

A field adds \( (e^2B^2/8m)(x^2 + y^2) \) per electron to the Hamiltonian of an atom with closed shells. First-order perturbation theory gives an energy rising as \( B^2 \), so \( M = -\partial E/\partial B \) is opposite to \( B \). For \( N \) atoms per unit volume, each with \( Z \) electrons of mean-square radius \( \langle r^2\rangle \) (using \( \langle x^2 + y^2\rangle = \frac{2}{3}\langle r^2\rangle \)):

$$\boxed{\chi_{\text{dia}} = -\frac{\mu_0NZe^2\langle r^2\rangle}{6m}}$$

Temperature independent, proportional to the size of the electron clouds, and present in every material (it is masked when a larger positive term exists). Water (about \( -9\times10^{-6} \)) and bismuth (about \( -1.7\times10^{-4} \), one of the largest among ordinary solids) are standard examples. Physically it is Lenz’s law at the atomic scale: switching on \( B \) induces orbital currents that oppose it.

Landau diamagnetism of free electrons

Conduction electrons in a field occupy Landau levels \( (n + \frac{1}{2})\hbar\omega_c \), \( \omega_c = eB/m \). The resulting orbital susceptibility for a free-electron gas is

$$\boxed{\chi_L = -\tfrac{1}{3}\chi_P}$$

where \( \chi_P \) is the Pauli spin susceptibility below. In a real metal the band mass changes the ratio.

Paramagnetism of Localised Moments

From spin-1/2 to general J

An ion with total angular momentum \( J \) has moment \( \boldsymbol\mu = -g_J\mu_B\mathbf J/\hbar \) and \( 2J + 1 \) levels \( E = g_J\mu_Bm_JB \). With \( x = g_J\mu_BJB/k_BT \):

$$\boxed{M = Ng_J\mu_BJ\,B_J(x)}, \qquad B_J(x) = \frac{2J + 1}{2J}\coth\!\left(\frac{2J + 1}{2J}x\right) - \frac{1}{2J}\coth\!\left(\frac{x}{2J}\right)$$

Limits of the Brillouin function:

CaseResult
\( J = \frac{1}{2} \), \( g = 2 \)\( B_{1/2}(x) = \tanh x \): \( M = N\mu_B\tanh(\mu_BB/k_BT) \)
\( J \to \infty \) with \( g_J\mu_BJ \) fixedLangevin function \( L(x) = \coth x - 1/x \) (classical dipoles)
\( x \gg 1 \) (strong field, low \( T \))\( B_J \to 1 \): saturation \( M_s = Ng_J\mu_BJ \)
\( x \ll 1 \)\( B_J(x) \approx \dfrac{(J + 1)x}{3J} \)

The small-\( x \) limit gives the Curie law:

$$\boxed{\chi = \frac{\mu_0N\mu_{\text{eff}}^2}{3k_BT} = \frac{C}{T}, \qquad \mu_{\text{eff}} = p\,\mu_B, \quad p = g_J\sqrt{J(J + 1)}}$$

Note \( J(J + 1) \), not \( J^2 \): quantum fluctuations make the effective moment larger than the saturation moment \( g_JJ\mu_B \). For spin-1/2 (\( g = 2 \)), \( g^2J(J + 1)/3 = 1 \) and \( \chi = \mu_0N\mu_B^2/k_BT \), matching the tanh law; for classical dipoles \( \chi = \mu_0N\mu^2/3k_BT \).

Hund’s rules and the Landé g-factor

For the ground term of a partly filled shell:

  1. Maximise \( S \) (consistent with Pauli).
  2. Then maximise \( L \).
  3. \( J = \lvert L - S\rvert \) if the shell is less than half full; \( J = L + S \) if more than half full.
$$\boxed{g_J = \frac{3}{2} + \frac{S(S + 1) - L(L + 1)}{2J(J + 1)}}$$
Ion (configuration)\( S \)\( L \)\( J \)\( g_J \)\( p = g_J\sqrt{J(J+1)} \)
Ce\( ^{3+} \) (4f\( ^1 \))1/235/26/72.54
Nd\( ^{3+} \) (4f\( ^3 \))3/269/28/113.62
Gd\( ^{3+} \) (4f\( ^7 \))7/207/227.94

Quenching in the iron group

Rare-earth 4f shells lie deep inside the ion, so Hund’s-rule \( J \) values work. The 3d shell of iron-group ions is exposed to the crystal field, which quenches the orbital moment (\( \langle L\rangle \approx 0 \)). Then only spin counts:

$$\boxed{p_{\text{spin-only}} = 2\sqrt{S(S + 1)}}$$
Unpaired electrons12345
\( p \)1.732.833.874.905.92

So Mn\( ^{2+} \) (3d\( ^5 \), \( S = 5/2 \), \( L = 0 \)) has \( p = 5.92 \) either way, while for Cu\( ^{2+} \) (3d\( ^9 \)) the spin-only 1.73 is closer to experiment than the Hund value.

Adiabatic demagnetisation

For an ideal paramagnet the partition function depends on \( B \) and \( T \) only through \( B/T \), and so does the entropy. An adiabatic (isentropic) change therefore keeps \( B/T \) fixed:

$$\boxed{T_f = T_i\frac{B_f}{B_i}}$$

Magnetise the salt at fixed low \( T \) (heat flows out to a bath, entropy drops), isolate it, then reduce the field: the temperature falls in proportion. In practice small internal fields between the ions set the floor, so \( B_f \) is effectively never zero.

Pauli Paramagnetism of Conduction Electrons

A naive Curie law for \( N \) free electrons would give \( \chi \propto 1/T \), far larger than measured. The Pauli principle forbids most electrons from flipping: only those within about \( \mu_BB \) of \( E_F \) can. The spin-up and spin-down bands shift by \( \mp\mu_BB \), and the excess is \( g(E_F)\mu_BB \), where \( g(E_F) \) is the density of states (both spins) per volume at the Fermi level:

$$\boxed{\chi_P = \mu_0\mu_B^2g(E_F) = \frac{3N\mu_0\mu_B^2}{2E_F}}$$

the second form for a free-electron gas. Compared with the Curie value \( \mu_0N\mu_B^2/k_BT \), it is smaller by the factor \( 3k_BT/2E_F \), only about 1% at room temperature, and independent of \( T \) because \( E_F \gg k_BT \).

Free-electron total: \( \chi_P + \chi_L = \frac{2}{3}\chi_P > 0 \).

Ferromagnetism: the Weiss Mean Field

The Weiss model

Weiss assumed each moment feels the applied field plus a molecular field proportional to the magnetization, \( B_{\text{eff}} = B + \lambda M \). For spin-1/2 moments \( \mu = \mu_B \):

$$\boxed{M = N\mu\tanh\!\left[\frac{\mu(B + \lambda M)}{k_BT}\right]}$$

Spontaneous magnetization (\( B = 0 \)). With \( m = M/N\mu \) and \( T_c = N\mu^2\lambda/k_B \):

$$\boxed{m = \tanh\!\left(\frac{T_c}{T}m\right)}$$

Graphically, the line \( m \) meets the curve \( \tanh(T_cm/T) \) at a non-zero point only when the slope of the curve at the origin, \( T_c/T \), exceeds 1, i.e. for \( T < T_c \).

RegionBehaviour
Just below \( T_c \)expand \( \tanh y \approx y - y^3/3 \): \( m \approx \sqrt{3}\,(1 - T/T_c)^{1/2} \), so \( \beta = \frac{1}{2} \)
\( T \to 0 \)\( 1 - m \approx 2e^{-2T_c/T} \) (exponentially flat)
\( T = T_c \), small field\( m \propto B^{1/3} \): \( \delta = 3 \)
Heat capacityjump at \( T_c \), zero above (no short-range order in mean field)

Above \( T_c \), small \( B \). Linearise: \( M = (N\mu^2/k_BT)(B + \lambda M) \), so \( M = N\mu^2B/k_B(T - T_c) \):

$$\boxed{\chi = \frac{C}{T - T_c}} \quad \text{(Curie–Weiss law, } \gamma = 1\text{)}$$

with the same Curie constant \( C = \mu_0N\mu^2/k_B \) as the free paramagnet. For general \( J \), \( \theta = \lambda C/\mu_0 \). Experimentally, the intercept \( \theta \) of \( 1/\chi \) against \( T \) lies somewhat above the true \( T_c \), because mean-field theory ignores fluctuations.

How strong is the molecular field?

At \( T = 0 \), \( \lambda M_s = \lambda N\mu = k_BT_c/\mu \). For \( T_c = 1043 \) K (iron) and \( \mu = \mu_B \):

$$\lambda M_s = \frac{1.381\times10^{-23}\times1043}{9.274\times10^{-24}} \approx 1.6\times10^3\ \text{T}$$

No magnetic dipole–dipole field is remotely that large (dipolar fields between neighbouring moments are of order 0.1 to 1 T). The molecular field is really the exchange interaction, an electrostatic effect made spin-dependent by the Pauli principle. Heisenberg wrote it as

$$H = -J_{\text{ex}}\sum_{\langle ij\rangle}\mathbf S_i\cdot\mathbf S_j$$

with \( J_{\text{ex}} > 0 \) favouring parallel spins. Mean-field treatment with \( z \) nearest neighbours gives \( k_BT_c = zJ_{\text{ex}}S(S + 1)/3 \) (spins in units of \( \hbar \)). Replacing the vector spin by an Ising variable gives the model of the next page.

Approximate Curie temperatures: iron 1043 K, cobalt about 1390 K, nickel about 630 K, gadolinium about 293 K (just below room temperature).

Beyond mean field: domains and spin waves

  • Domains. A macroscopic sample below \( T_c \) usually shows little net moment: it splits into domains to reduce magnetostatic energy. An applied field moves domain walls and rotates domains, producing the hysteresis loop (remanence, coercive field). Soft magnets (transformer cores) have narrow loops; hard (permanent) magnets have wide ones.
  • Spin waves. The low-lying excitations of a Heisenberg ferromagnet are magnons with \( \hbar\omega \propto k^2 \) at small \( k \). Counting them like bosons gives the Bloch \( T^{3/2} \) law, \( M(0) - M(T) \propto T^{3/2} \), not the exponential of mean-field theory. Antiferromagnetic magnons have \( \omega \propto k \).

Antiferromagnetism and Ferrimagnetism

With \( J_{\text{ex}} < 0 \), neighbouring spins anti-align on two interpenetrating sublattices. Mean-field theory with opposite molecular fields gives

$$\boxed{\chi = \frac{C}{T + \theta} \quad (T > T_N)}$$

The negative intercept of \( 1/\chi \) against \( T \) at \( T = -\theta \) is the signature of antiferromagnetic coupling. In the simplest two-sublattice model \( \theta = T_N \); real materials usually have \( \theta > T_N \). Below the Néel temperature \( T_N \), \( \chi \) depends on direction: along the spin axis \( \chi_\parallel \to 0 \) as \( T \to 0 \); perpendicular to it \( \chi_\perp \) stays roughly at its \( T_N \) value. The susceptibility of a powder therefore peaks (with a cusp) at \( T_N \).

In a ferrimagnet the two sublattices carry unequal moments, so a net spontaneous moment survives. Magnetite (Fe\( _3 \)O\( _4 \)) and the ferrites are the standard examples.

graph TD
    A["Plot 1/chi against T above the ordering temperature"] --> B{"Intercept on the T axis"}
    B -- "at T = 0 (through origin)" --> C["Ideal paramagnet: Curie law"]
    B -- "positive, T = theta" --> D["Ferromagnetic coupling: orders near theta"]
    B -- "negative, T = -theta" --> E["Antiferromagnetic coupling: Neel order below T_N"]
    A --> F{"chi negative and T-independent?"}
    F -- "yes" --> G["Diamagnet"]

Memory Tricks

Don’t Confuse

ThisvsThat
Curie paramagnet: \( \chi \propto 1/T \), localised momentsPauli paramagnet: \( \chi \) nearly \( T \)-independent, conduction electrons
Larmor diamagnetism: bound electrons, \( \propto \langle r^2\rangle \)Landau diamagnetism: free electrons, \( -\frac{1}{3}\chi_P \)
Effective moment \( g\sqrt{J(J+1)}\,\mu_B \) (from \( \chi \))Saturation moment \( gJ\mu_B \) (from \( M \) at high \( B/T \))
Spin-1/2 Curie constant \( \mu_0N\mu_B^2/k_B \)Classical Langevin \( \mu_0N\mu^2/3k_B \)
Weiss molecular field: exchange, about \( 10^3 \) TDipolar field between moments: about 1 T or less
Mean-field low-\( T \) magnetization: exponential approach to saturationSpin waves: Bloch \( T^{3/2} \) law
Antiferromagnet \( \chi_\parallel \to 0 \) at \( T \to 0 \)\( \chi_\perp \) nearly constant below \( T_N \)

Common Mistakes

Practice

All questions below are original practice questions written for this page. They are not from any CSIR NET paper. Use \( \mu_B = 9.274\times10^{-24} \) J T\( ^{-1} \), \( k_B = 1.381\times10^{-23} \) J K\( ^{-1} \).

Level 1: Foundation

  1. What does the Bohr–van Leeuwen theorem state?
  2. How does the susceptibility of an ideal Curie paramagnet vary with temperature?
  3. Give the free-electron ratio \( \chi_L/\chi_P \).
Answers
  1. A classical system of charges in thermal equilibrium has zero magnetization in any applied field, because the vector potential can be shifted out of the momentum integral.
  2. \( \chi = C/T \), inversely proportional to absolute temperature.
  3. \( -\frac{1}{3} \): Landau diamagnetism is a third of the Pauli paramagnetism and opposite in sign.

Level 2: Part B Style

Practice Q1. A free ion has the ground term \( ^4I_{9/2} \) (\( S = 3/2 \), \( L = 6 \), \( J = 9/2 \)). Its effective moment \( p = g_J\sqrt{J(J+1)} \) in units of \( \mu_B \) is closest to:

(a) 3.62 (b) 3.87 (c) 4.97 (d) 6.63

Answer: (a) 3.62

\( g_J = \frac{3}{2} + \dfrac{S(S+1) - L(L+1)}{2J(J+1)} = 1.5 + \dfrac{3.75 - 42}{2\times24.75} = 1.5 - 0.773 = 0.727 \ (= 8/11) \).

\( p = 0.727\times\sqrt{24.75} = 0.727\times4.975 = 3.62 \).

(b) is the spin-only value for \( S = 3/2 \), which would apply only if the orbital moment were quenched.

Practice Q2. The susceptibility of a paramagnetic salt follows \( \chi = C/(T - \theta) \). It is \( 2.0\times10^{-3} \) at 400 K and \( 1.0\times10^{-3} \) at 600 K. Then \( \theta \) and \( C \) are:

(a) 200 K and 0.4 K (b) −200 K and 0.4 K (c) 200 K and 0.8 K (d) 0 and 0.8 K

Answer: (a) 200 K and 0.4 K

\( C = 2.0\times10^{-3}(400 - \theta) = 1.0\times10^{-3}(600 - \theta) \Rightarrow 800 - 2\theta = 600 - \theta \Rightarrow \theta = 200 \) K.

\( C = 2.0\times10^{-3}\times200 = 0.4 \) K. A positive \( \theta \) indicates ferromagnetic coupling.

Practice Q3. For a free-electron metal, the total spin plus orbital susceptibility of the conduction electrons is:

(a) \( \chi_P \) (b) \( \frac{2}{3}\chi_P \) (c) \( \frac{1}{3}\chi_P \) (d) \( -\frac{1}{3}\chi_P \)

Answer: (b) (2/3)χ_P

Pauli paramagnetism \( \chi_P \) plus Landau diamagnetism \( -\frac{1}{3}\chi_P \) gives \( \frac{2}{3}\chi_P \). (d) is the Landau term alone.

Practice Q4. An ideal paramagnetic salt is at 1.0 K in a field of 2.0 T. The field is reduced adiabatically and reversibly to 0.05 T. Ignoring internal fields, the final temperature is:

(a) 0.5 K (b) 0.1 K (c) 25 mK (d) 2.5 mK

Answer: (c) 25 mK

Entropy is a function of \( B/T \) only, so \( B/T \) stays fixed: \( T_f = 1.0\times0.05/2.0 = 0.025 \) K.

Level 3: Part C Style

Practice Q5. A dilute system of spin-1/2 ions (\( g = 2 \), moment \( \mu_B \)) is held at 1.0 K in a field of 1.0 T. Consider:

  1. The fractional magnetization \( M/M_s \) is about 0.59.
  2. The Curie law, applied naively at this field, overestimates \( M \) by about 15%.
  3. At 300 K in the same field, the Curie law is accurate to better than 0.1%.
  4. Doubling both \( B \) and \( T \) leaves \( M \) unchanged.

Which are correct?

(a) 1 and 4 only (b) 1, 2 and 4 only (c) 1, 2, 3 and 4 (d) 2 and 3 only

Answer: (c) 1, 2, 3 and 4

\( x = \mu_BB/k_BT = 9.274\times10^{-24}/1.381\times10^{-23} = 0.672 \).

  1. \( M/M_s = \tanh0.672 = 0.586 \). True.
  2. Curie law gives \( M/M_s = x = 0.672 \); ratio \( 0.672/0.586 = 1.146 \). True.
  3. At 300 K, \( x = 2.24\times10^{-3} \); \( \tanh x \approx x(1 - x^2/3) \), relative error \( x^2/3 \approx 1.7\times10^{-6} \). True.
  4. \( M \) depends only on \( B/T \). True.

Practice Q6. In the Weiss mean-field model of a spin-1/2 ferromagnet, \( m = \tanh(T_cm/T) \) with \( T_c = 600 \) K. Consider:

  1. At \( T = 594 \) K, the spontaneous \( m \) is close to 0.17.
  2. At \( T = 300 \) K, \( m \) is about 0.96.
  3. Above 600 K, the zero-field susceptibility varies as \( (T - 600)^{-1} \).
  4. In this model the magnetic heat capacity is zero just above \( T_c \) and jumps to a finite value just below.

Which are correct?

(a) 1 and 3 only (b) 1, 3 and 4 only (c) 1, 2, 3 and 4 (d) 2 and 4 only

Answer: (c) 1, 2, 3 and 4
  1. Near \( T_c \), \( m \approx \sqrt{3(1 - T/T_c)} = \sqrt{0.03} = 0.173 \); solving \( m = \tanh(m/0.99) \) numerically gives 0.1725. True.
  2. At \( T/T_c = 0.5 \), iterate \( m = \tanh 2m \): \( m = 0.9575 \). True.
  3. Curie–Weiss, \( \gamma = 1 \). True.
  4. Mean-field energy \( -\frac{1}{2}N\lambda M^2 \) vanishes above \( T_c \) and varies as \( (T_c - T) \) just below, so \( C \) jumps (\( \alpha = 0 \)). True.

Practice Q7. Susceptibility data above the ordering temperature of a magnetic insulator give a straight line \( 1/\chi = (T + 150\ \text{K})/C \). Below 80 K, single-crystal measurements show \( \chi \) along one axis falling towards zero as \( T \to 0 \), while \( \chi \) perpendicular to that axis stays nearly constant. Consider:

  1. The dominant exchange coupling is antiferromagnetic.
  2. The ordering (Néel) temperature is 150 K.
  3. The axis along which \( \chi \to 0 \) is the direction of the ordered spins.
  4. The material must have a spontaneous net magnetization below 80 K.

Which are correct?

(a) 1 and 3 only (b) 1, 2 and 3 only (c) 1 and 4 only (d) 2 and 4 only

Answer: (a) 1 and 3 only
  1. The intercept of \( 1/\chi \) is at \( T = -150 \) K: antiferromagnetic coupling. True.
  2. The data say ordering sets in near 80 K. The Curie–Weiss \( \theta \) (150 K) need not equal \( T_N \); here \( \theta/T_N \approx 1.9 \). False.
  3. \( \chi_\parallel \to 0 \) along the sublattice axis, because a field along the spins cannot tilt them and only thermal excitations respond. True.
  4. An antiferromagnet has zero net moment; that would describe a ferrimagnet. False.

Quick Revision Table

IdeaFormula or fact
Bohr–van Leeuwenclassical \( M = 0 \)
Larmor diamagnetism\( \chi = -\mu_0NZe^2\langle r^2\rangle/6m \), \( T \)-independent
Brillouin magnetization\( M = Ng_J\mu_BJB_J(x) \), \( x = g_J\mu_BJB/k_BT \)
Curie law\( \chi = \mu_0N\mu_{\text{eff}}^2/3k_BT \), \( p = g_J\sqrt{J(J+1)} \)
Landé \( g \)\( \frac{3}{2} + [S(S+1) - L(L+1)]/2J(J+1) \)
Spin-only\( p = 2\sqrt{S(S+1)} \): 1.73, 2.83, 3.87, 4.90, 5.92
Pauli\( \chi_P = \mu_0\mu_B^2g(E_F) = 3N\mu_0\mu_B^2/2E_F \)
Landau\( \chi_L = -\chi_P/3 \)
Weiss\( m = \tanh(T_cm/T) \); \( m \approx \sqrt{3(1 - T/T_c)} \)
Curie–Weissferro \( C/(T - \theta) \); antiferro \( C/(T + \theta) \)
Molecular field\( k_BT_c/\mu_B \approx 10^3 \) T: exchange, not dipolar
Demagnetisation\( T_f = T_iB_f/B_i \)
Magnonsferro \( \omega \propto k^2 \), Bloch \( T^{3/2} \)

Key Takeaways

What’s Next

Advertisement
Advertisement