Bose–Einstein Condensation
CSIR NET Part C: ideal Bose gas, critical temperature, condensate fraction, heat capacity cusp, pressure below Tc, harmonic traps, dimensionality, helium-4.
- 16 min
On this page · 14 sections
Prerequisites
- Canonical and Grand Canonical Ensembles: the grand partition function, fugacity \( z = e^{\beta\mu} \) and \( \langle N\rangle = k_BT\,\partial\ln\mathcal Z/\partial\mu \).
- Quantum Statistics: Ideal Bose and Fermi Gases: the Bose–Einstein distribution and Bose functions from the core unit.
- Identical Particles, Pauli Principle and Spin-Statistics: why bosons may pile into one state, and which composite particles are bosons.
- First- and Second-Order Phase Transitions: order parameters and the classification of transitions.
The Hook: A Phase Transition Without Forces
The Ideal Bose Gas
Spinless bosons of mass \( m \) in a box of volume \( V \), single-particle energies \( \varepsilon = p^2/2m \geq 0 \). The mean occupation of a state is the Bose–Einstein distribution
$$\boxed{\langle n_\varepsilon\rangle = \frac{1}{e^{\beta(\varepsilon - \mu)} - 1}}$$Occupations must be non-negative, so \( \mu < \varepsilon_{\min} = 0 \): the fugacity \( z = e^{\beta\mu} \) lies in \( 0 < z < 1 \).
Replace the sum over states by an integral with the 3D density of states \( g(\varepsilon) = \dfrac{V}{4\pi^2}\left(\dfrac{2m}{\hbar^2}\right)^{3/2}\varepsilon^{1/2} \). Because \( g(0) = 0 \), the integral drops the ground state, so add it back by hand:
$$\boxed{N = \frac{V}{\lambda^3}g_{3/2}(z) + \frac{z}{1 - z}}, \qquad \lambda = \frac{h}{\sqrt{2\pi mk_BT}}$$where \( \lambda \) is the thermal de Broglie wavelength and \( g_\nu(z) = \sum_{k=1}^{\infty}z^k/k^\nu \) is the Bose function. At \( z = 1 \), \( g_\nu(1) = \zeta(\nu) \), the Riemann zeta function:
| \( \zeta(3/2) \) | \( \zeta(2) \) | \( \zeta(5/2) \) | \( \zeta(3) \) |
|---|---|---|---|
| 2.612 | 1.645 | 1.341 | 1.202 |
Why Condensation Happens
The excited states can hold at most
$$N_{\text{ex}}^{\max} = \frac{V}{\lambda^3}g_{3/2}(1) = \frac{V}{\lambda^3}\zeta(3/2)$$particles, because \( g_{3/2}(z) \) is bounded as \( z \to 1 \). As \( T \) falls, \( \lambda \) grows and this capacity shrinks as \( T^{3/2} \). When it drops below \( N \), the remaining particles have nowhere to go but the ground state: \( z \) sticks just below 1 (\( z/(1 - z) \) becomes of order \( N \)) and \( N_0 \) becomes macroscopic.
The critical condition is
$$\boxed{n\lambda_c^3 = \zeta(3/2) = 2.612}$$i.e. the thermal wavelength becomes comparable to the interparticle spacing \( n^{-1/3} \). Solving for \( T \):
$$\boxed{T_c = \frac{2\pi\hbar^2}{mk_B}\left(\frac{n}{\zeta(3/2)}\right)^{2/3}}$$Scaling: \( T_c \propto n^{2/3}/m \). Doubling the density raises \( T_c \) by \( 2^{2/3} = 1.59 \); doubling the mass halves it.
Below \( T_c \) the excited states are full, \( N_{\text{ex}} = N(T/T_c)^{3/2} \), so
$$\boxed{\frac{N_0}{N} = 1 - \left(\frac{T}{T_c}\right)^{3/2}}$$graph TD
A["High T: z small, n lambda^3 << 1, classical gas"] --> B["Cooling: lambda grows as T^(-1/2)"]
B --> C{"n lambda^3 reaches 2.612?"}
C -- "no" --> B
C -- "yes, T = Tc" --> D["z sticks at 1, excited states saturate"]
D --> E["N0/N = 1 - (T/Tc)^(3/2)"]
E --> F["T = 0: all particles in the ground state"]Thermodynamics Below and Above Tc
Only excited particles carry energy and entropy. With \( z = 1 \) below \( T_c \):
| Quantity | \( T < T_c \) | \( T \gg T_c \) |
|---|---|---|
| Pressure | \( P = \dfrac{k_BT}{\lambda^3}\zeta(5/2) \propto T^{5/2} \), independent of \( n \) | \( P \to nk_BT \) |
| Energy | \( U = \frac{3}{2}PV = 0.770\,Nk_BT_c\left(\dfrac{T}{T_c}\right)^{5/2} \) | \( \frac{3}{2}Nk_BT \) |
| Heat capacity | \( C_V = \frac{15}{4}\dfrac{\zeta(5/2)}{\zeta(3/2)}Nk_B\left(\dfrac{T}{T_c}\right)^{3/2} \) | \( \frac{3}{2}Nk_B \) |
| Entropy | \( S = \frac{5}{2}k_B\dfrac{V}{\lambda^3}\zeta(5/2) \propto T^{3/2} \) | Sackur–Tetrode |
At \( T_c \):
$$\boxed{C_V(T_c) = \frac{15}{4}\frac{\zeta(5/2)}{\zeta(3/2)}Nk_B \approx 1.93\,Nk_B}$$which exceeds the classical \( 1.5\,Nk_B \). Above \( T_c \), \( C_V \) falls back towards \( 1.5\,Nk_B \). The curve is continuous at \( T_c \) but its slope jumps: a cusp, not a jump and not a divergence.
The leading quantum correction above Tc
For \( n\lambda^3 \ll 1 \), expanding in powers of \( z \):
$$\boxed{\frac{PV}{Nk_BT} = 1 - \frac{n\lambda^3}{2^{5/2}} + \cdots \ \ \text{(bosons)}, \qquad 1 + \frac{n\lambda^3}{2^{5/2}g_s} + \cdots \ \ \text{(fermions, spin degeneracy } g_s\text{)}}$$Bosons behave as if they attract (lower pressure); fermions as if they repel. This is a pure statistics effect, the thermodynamic face of the bunching and antibunching of identical particles.
What kind of transition is it?
The ideal-gas pressure below \( T_c \) does not depend on volume: isotherms in the \( P \)–\( v \) plane are flat for \( v < v_c = \lambda^3/\zeta(3/2) \), exactly like a liquid–vapour coexistence region, with the condensate playing the role of a phase of zero specific volume. The transition line \( P_c(T) \propto T^{5/2} \) then satisfies Clausius–Clapeyron with an effective latent heat per particle
$$L = Tv_c\frac{dP_c}{dT} = \frac{5}{2}\frac{\zeta(5/2)}{\zeta(3/2)}k_BT$$So viewed in the \( P \)–\( T \) plane the ideal-gas transition looks first order, while at fixed density \( C_V \) shows only a cusp. Interacting real systems (helium-4) instead show a continuous \( \lambda \) transition. Part C uses this as a “which statement is correct” trap; the safest exam statement is that \( C_V \) is continuous at \( T_c \) with a discontinuous derivative.
Density of States and Dimension
Whether BEC occurs depends on whether \( \int_0^\infty g(\varepsilon)\,d\varepsilon/(e^{\beta\varepsilon} - 1) \) converges at \( z = 1 \). Near \( \varepsilon = 0 \) the integrand behaves as \( g(\varepsilon)/\beta\varepsilon \). For \( g(\varepsilon) \propto \varepsilon^{\alpha - 1} \):
$$N_{\text{ex}}^{\max} \propto T^{\alpha}\,\Gamma(\alpha)\zeta(\alpha), \quad \text{finite only if } \alpha > 1$$and then
$$\boxed{\frac{N_0}{N} = 1 - \left(\frac{T}{T_c}\right)^{\alpha}}$$| System | \( g(\varepsilon) \) | \( \alpha \) | BEC at \( T > 0 \)? | Condensate fraction |
|---|---|---|---|---|
| 3D box | \( \propto\varepsilon^{1/2} \) | 3/2 | yes | \( 1 - (T/T_c)^{3/2} \) |
| 2D box | constant | 1 | no (\( \int d\varepsilon/\varepsilon \) diverges) | — |
| 1D box | \( \propto\varepsilon^{-1/2} \) | 1/2 | no | — |
| 3D harmonic trap | \( \varepsilon^2/2(\hbar\bar\omega)^3 \) | 3 | yes | \( 1 - (T/T_c)^3 \) |
| 2D harmonic trap | \( \varepsilon/(\hbar\omega)^2 \) | 2 | yes | \( 1 - (T/T_c)^2 \) |
| 1D harmonic trap | \( 1/\hbar\omega \) | 1 | no (in the thermodynamic limit) | — |
BEC in a harmonic trap
For an isotropic or anisotropic 3D trap with geometric-mean frequency \( \bar\omega = (\omega_x\omega_y\omega_z)^{1/3} \), \( N_{\text{ex}}^{\max} = \zeta(3)(k_BT/\hbar\bar\omega)^3 \), so
$$\boxed{k_BT_c = \hbar\bar\omega\left(\frac{N}{\zeta(3)}\right)^{1/3} \approx 0.94\,\hbar\bar\omega N^{1/3}}$$Example: \( N = 10^6 \) atoms in a trap with \( \bar\omega = 2\pi\times100 \) Hz gives \( T_c \approx 450 \) nK, the range in which the alkali-vapour experiments work. Note that \( T_c \) here depends on \( N \) and \( \bar\omega \), not on the mass.
Who Condenses and Who Does Not
Photons and phonons. Their number is not conserved, so \( \mu = 0 \) at all temperatures and there is no condition fixing \( N \). Blackbody radiation has no BEC: lowering \( T \) simply removes photons.
Helium-4 is a boson (six fermion constituents). Treating liquid helium as an ideal Bose gas with density about 145 kg m\( ^{-3} \) and \( m = 6.65\times10^{-27} \) kg gives
$$T_c = \frac{2\pi\hbar^2}{mk_B}\left(\frac{n}{2.612}\right)^{2/3} \approx 3.1\ \text{K}$$compared with the observed \( \lambda \) point of about 2.17 K. The agreement in order of magnitude is the classic argument (due to London) that superfluidity is related to BEC. Strong interactions shift \( T_c \), turn the cusp into a \( \lambda \)-shaped divergence, and deplete the condensate so that even at \( T = 0 \) only a minority of the atoms are in the zero-momentum state.
Helium-3 is a fermion and cannot condense this way; it becomes superfluid only at millikelvin temperatures after its atoms form pairs.
Dilute alkali gases (rubidium-87, sodium-23 and others) are weakly interacting, so ideal-gas formulas are a good first approximation, with interactions entering as a mean-field correction.
Memory Tricks
Don’t Confuse
| This | vs | That |
|---|---|---|
| BEC: macroscopic occupation of one state, needs conserved \( N \) | Blackbody photons: \( \mu = 0 \), no condensation | |
| Box: \( N_0/N = 1 - (T/T_c)^{3/2} \) | 3D harmonic trap: \( 1 - (T/T_c)^{3} \) | |
| 2D box: no BEC | 2D harmonic trap: BEC | |
| Ideal gas \( C_V \): cusp, \( 1.93\,Nk_B \) at \( T_c \) | Helium-4: \( \lambda \)-shaped divergence at 2.17 K | |
| Bosons: \( PV/Nk_BT = 1 - n\lambda^3/2^{5/2} \) | Fermions: \( 1 + n\lambda^3/2^{5/2}g_s \) | |
| Condensate fraction (particles in \( \mathbf p = 0 \)) | Superfluid fraction (two-fluid model): not the same quantity in interacting systems |
Common Mistakes
Practice
All questions below are original practice questions written for this page. They are not from any CSIR NET paper. Use \( \hbar = 1.055\times10^{-34} \) J s, \( k_B = 1.381\times10^{-23} \) J K\( ^{-1} \), \( 1\ \text{u} = 1.661\times10^{-27} \) kg, \( \zeta(3/2) = 2.612 \), \( \zeta(3) = 1.202 \).
Level 1: Foundation
- State the condition for the onset of BEC in a 3D ideal Bose gas in terms of \( n \) and \( \lambda \).
- Give the condensate fraction below \( T_c \) for a 3D box.
- Why do photons in a cavity not undergo BEC as the temperature is lowered?
Answers
- \( n\lambda^3 = \zeta(3/2) \approx 2.612 \) with \( \lambda = h/\sqrt{2\pi mk_BT} \).
- \( N_0/N = 1 - (T/T_c)^{3/2} \).
- Their number is not conserved, so \( \mu = 0 \) and there is no fixed \( N \) that the excited states must accommodate; photons are simply absorbed by the walls.
Level 2: Part B Style
Practice Q1. An ideal Bose gas in a 3D box is at \( T = T_c/2 \). The fraction of particles in the ground state is closest to:
(a) 0.35 (b) 0.50 (c) 0.65 (d) 0.88
Answer: (c) 0.65
\( 1 - (0.5)^{3/2} = 1 - 0.354 = 0.646 \). (a) is the excited fraction; (d) is the 3D harmonic-trap value \( 1 - 0.5^3 \).
Practice Q2. The number density of an ideal Bose gas in a box is doubled at fixed particle mass. The critical temperature changes by a factor of:
(a) \( 2^{1/3} \) (b) \( 2^{2/3} \) (c) \( 2^{3/2} \) (d) 2
Answer: (b) 2^(2/3) ≈ 1.59
\( T_c \propto n^{2/3} \), from \( n\lambda_c^3 = \) constant with \( \lambda \propto T^{-1/2} \).
Practice Q3. The heat capacity at constant volume of an ideal 3D Bose gas at \( T = T_c \) is closest to:
(a) \( 0.77\,Nk_B \) (b) \( 1.50\,Nk_B \) (c) \( 1.93\,Nk_B \) (d) \( 3.00\,Nk_B \)
Answer: (c) 1.93 Nk_B
\( C_V = \frac{15}{4}\times\frac{1.341}{2.612}Nk_B = 1.926\,Nk_B \). (a) is the coefficient of \( U/Nk_BT_c \) at \( T_c \); (b) is the classical limit.
Practice Q4. For non-interacting bosons, in which of the following does BEC occur at a finite temperature in the thermodynamic limit?
- A 3D box
- A 2D box
- A 2D isotropic harmonic trap
- A 1D box
(a) 1 only (b) 1 and 3 only (c) 1, 2 and 3 only (d) all four
Answer: (b) 1 and 3 only
BEC needs \( g(\varepsilon) \propto \varepsilon^{\alpha - 1} \) with \( \alpha > 1 \). 3D box \( \alpha = \frac{3}{2} \); 2D trap \( \alpha = 2 \): yes. 2D box \( \alpha = 1 \) and 1D box \( \alpha = \frac{1}{2} \): the excited-state integral diverges, no.
Level 3: Part C Style
Practice Q5. A uniform gas of bosonic atoms of mass 87 u has number density \( 1.0\times10^{20} \) m\( ^{-3} \). Treat it as ideal. Consider:
- Its BEC transition temperature is about 0.40 μK.
- The thermal de Broglie wavelength at \( T_c \) is about 0.30 μm.
- At \( T = T_c/2 \), about 65% of the atoms are in the condensate.
- A gas of atoms of twice the mass at the same density would condense at twice the temperature.
Which are correct?
(a) 1 and 2 only (b) 1, 2 and 3 only (c) 3 and 4 only (d) 1, 2, 3 and 4
Answer: (b) 1, 2 and 3 only
\( m = 87\times1.661\times10^{-27} = 1.445\times10^{-25} \) kg.
- \( T_c = \dfrac{2\pi\hbar^2}{mk_B}\left(\dfrac{n}{2.612}\right)^{2/3} \). \( (n/2.612)^{2/3} = (3.83\times10^{19})^{2/3} = 1.136\times10^{13} \) m\( ^{-2} \); prefactor \( 2\pi(1.055\times10^{-34})^2/(1.445\times10^{-25}\times1.381\times10^{-23}) = 3.50\times10^{-20} \) K m\( ^2 \). Product \( \approx 3.98\times10^{-7} \) K. True.
- \( \lambda_c = (2.612/n)^{1/3} = (2.612\times10^{-20})^{1/3} = 2.97\times10^{-7} \) m. True.
- \( 1 - 0.5^{3/2} = 0.646 \). True.
- \( T_c \propto 1/m \): it would halve. False.
Practice Q6. \( 2.0\times10^5 \) non-interacting bosonic atoms are held in a 3D isotropic harmonic trap of frequency \( \omega = 2\pi\times50 \) Hz. Consider:
- The condensation temperature is about 130 nK.
- At \( T = 0.8\,T_c \), about \( 9.8\times10^4 \) atoms are in the condensate.
- If \( N \) is increased eightfold at fixed \( \omega \), \( T_c \) doubles.
- \( T_c \) depends on the atomic mass even at fixed \( \omega \) and \( N \).
Which are correct?
(a) 1 and 2 only (b) 1, 2 and 3 only (c) 2 and 4 only (d) 1, 2, 3 and 4
Answer: (b) 1, 2 and 3 only
- \( k_BT_c = \hbar\omega(N/\zeta(3))^{1/3} \). \( (2\times10^5/1.202)^{1/3} = 55.0 \); \( \hbar\omega/k_B = 1.055\times10^{-34}\times314.2/1.381\times10^{-23} = 2.40\times10^{-9} \) K. \( T_c \approx 132 \) nK. True.
- \( N_0 = N[1 - (0.8)^3] = 2\times10^5\times0.488 = 9.76\times10^4 \). True.
- \( T_c \propto N^{1/3} \): \( 8^{1/3} = 2 \). True.
- At fixed \( \omega \), the formula contains no mass (mass enters only through how a given trap sets \( \omega \)). False.
Practice Q7. For an ideal 3D Bose gas in a box below \( T_c \), consider:
- The pressure depends on temperature but not on density.
- \( C_V \propto T^{3/2} \).
- The condensate contributes nothing to the energy or entropy.
- \( C_V \) is discontinuous at \( T_c \).
Which are correct?
(a) 1 and 2 only (b) 1, 2 and 3 only (c) 2, 3 and 4 only (d) 1, 2, 3 and 4
Answer: (b) 1, 2 and 3 only
- \( P = \zeta(5/2)k_BT/\lambda^3 \propto T^{5/2} \), independent of \( n \). True.
- \( U \propto T^{5/2} \Rightarrow C_V \propto T^{3/2} \). True.
- Condensed particles occupy the single \( \varepsilon = 0 \) state: zero energy, and a single state carries no entropy. True.
- \( C_V \) is continuous (both sides give \( 1.93\,Nk_B \)); only its slope jumps. False.
Quick Revision Table
| Idea | Formula or fact |
|---|---|
| Bose–Einstein | \( \langle n\rangle = 1/(e^{\beta(\varepsilon - \mu)} - 1) \), \( \mu \leq \varepsilon_0 \) |
| Thermal wavelength | \( \lambda = h/\sqrt{2\pi mk_BT} \) |
| Onset | \( n\lambda^3 = \zeta(3/2) = 2.612 \) |
| Box \( T_c \) | \( (2\pi\hbar^2/mk_B)(n/2.612)^{2/3} \) |
| Box fraction | \( N_0/N = 1 - (T/T_c)^{3/2} \) |
| Below \( T_c \) | \( P = \zeta(5/2)k_BT/\lambda^3 \); \( U = \frac{3}{2}PV \); \( C_V \propto T^{3/2} \) |
| \( C_V(T_c) \) | \( \frac{15}{4}\frac{\zeta(5/2)}{\zeta(3/2)}Nk_B = 1.93\,Nk_B \), cusp |
| Quantum correction | \( PV/Nk_BT = 1 \mp n\lambda^3/2^{5/2} \) (bosons minus) |
| Dimension rule | \( g \propto \varepsilon^{\alpha - 1} \): BEC iff \( \alpha > 1 \); fraction \( 1 - (T/T_c)^\alpha \) |
| 3D trap | \( k_BT_c = \hbar\bar\omega(N/1.202)^{1/3} = 0.94\hbar\bar\omega N^{1/3} \) |
| Helium-4 estimate | about 3.1 K versus \( \lambda \) point 2.17 K |
| Photons, phonons | \( \mu = 0 \), no BEC |
Key Takeaways
What’s Next
- Diffusion, Random Walk and Brownian Motion: leaving equilibrium, starting from the simplest stochastic process.
- The Ising Model: the interaction-driven counterpart to this statistics-driven transition.
- Phase Space, Microstates and the Microcanonical Ensemble: the classical Sackur–Tetrode gas that the Bose gas reduces to at high temperature.
- Boltzmann statistics and partition functions from the chemist’s side: Boltzmann Distribution.