Classical Limit of Quantum Statistics

Every gas is, underneath, a quantum gas. Its particles are identical bosons or fermions, and the many-body wavefunction must be symmetric or antisymmetric under exchange. Yet the air in this room, the steam in a kettle, and the argon in a light bulb all obey the humble Maxwell-Boltzmann distribution and the ideal-gas law with astonishing accuracy. Why does quantum indistinguishability leave no fingerprint? The answer is the classical limit of quantum statistics: the regime in which the Bose-Einstein and Fermi-Dirac occupation numbers, which differ only by a $\pm1$ in their denominators, both collapse onto the same exponential $e^{-\beta(\epsilon-\mu)}$ . The whole story is controlled by one dimensionless number, the degeneracy parameter $n\lambda^3$ , where $\lambda=h/\sqrt{2\pi m k_BT}$ is the thermal de Broglie wavelength — the fuzziness of each particle's wavepacket. When $n\lambda^3\ll1$ the packets are far smaller than the spacing between particles, they never overlap, double occupancy is vanishingly rare, and the $\pm1$ that encodes exchange is negligible. Both quantum distributions reduce to Maxwell-Boltzmann, and quantum statistical mechanics becomes classical kinetic theory. As $n\lambda^3$ climbs toward one, the gas becomes degenerate — bosons head for condensation, fermions for the filled Fermi sea — and the classical description fails. The reduction is exact only asymptotically, and the first correction is the most instructive part. Expanding the equation of state in powers of $n\lambda^3$ produces a second virial coefficient of purely statistical origin, with no underlying force: fermions feel an effective short-range repulsion (Pauli exclusion) and bosons an effective attraction (bunching). Exchange is a genuine, measurable interaction that switches off only as the gas is diluted or heated. And one quantum remnant never leaves: the Gibbs factor $1/N!$ survives to keep the entropy extensive and resolve the Gibbs paradox — the classical limit is not $\hbar\to0$ .

Key takeaways

The mean occupation of a single-particle state is $\langle n_\epsilon\rangle=1/(e^{\beta(\epsilon-\mu)}\pm1)$ — upper sign fermions, lower sign bosons — and both reduce to the Maxwell-Boltzmann form $e^{-\beta(\epsilon-\mu)}$ when the fugacity $z=e^{\beta\mu}\ll1$ .
The single control parameter is the degeneracy parameter $n\lambda^3$ with $\lambda=h/\sqrt{2\pi m k_BT}$ ; classical statistics holds when $n\lambda^3\ll1$ , i.e. the interparticle spacing $n^{-1/3}$ far exceeds $\lambda$ so wavepackets do not overlap.
The leading quantum correction is a statistical virial term: $\dfrac{PV}{Nk_BT}=1\pm\dfrac{n\lambda^3}{2^{5/2}}$ — $+$ for fermions (effective repulsion), $-$ for bosons (effective attraction).
The correction to the occupation is $\langle n_\epsilon\rangle\approx e^{-\beta(\epsilon-\mu)}\big(1\pm e^{-\beta(\epsilon-\mu)}\big)$ : bosons bunch ( $+$ , enhanced), fermions anti-bunch ( $-$ , suppressed).
Air at STP has $n\lambda^3\sim10^{-7}$ (deeply classical); conduction electrons in a metal have $n\lambda^3\sim10^3$ (fully degenerate) — the same criterion decides both.
The classical limit is high $T$ or low $n$ , never literally $\hbar\to0$ : $\hbar$ survives as the phase-space cell $h^3$ and the Gibbs factor $1/N!$ that keep entropy meaningful and extensive.
Key idea: both quantum distributions $1/(e^{\beta(\epsilon-\mu)}\pm1)$ collapse onto Maxwell-Boltzmann $z\,e^{-\beta\epsilon}$ exactly when the fugacity $z=n\lambda^3\ll1$ — the thermal wavelength is small compared to the interparticle spacing, so exchange is negligible and the sign of the statistics survives only as a tiny $\pm$ correction.

The core insight

Both quantum distributions $\langle n_\epsilon\rangle=1/(e^{\beta(\epsilon-\mu)}\pm1)$ collapse onto the Maxwell-Boltzmann form $z\,e^{-\beta\epsilon}$ precisely when the fugacity $z=e^{\beta\mu}\ll1$ , equivalently $n\lambda^3\ll1$ : the thermal wavelength is far smaller than the interparticle spacing, wavepackets do not overlap, and exchange survives only as a $\pm$ statistical correction of order $n\lambda^3$ .

Where this lives today

The degeneracy parameter $n\lambda^3$ is the experimentalist's daily compass. In ultracold-atom laboratories it is literally the dial: a cloud of rubidium or sodium is laser- and evaporatively cooled to push $n\lambda^3$ from the classical regime ( $\ll1$ , a "thermal cloud") up through the condensation threshold $\zeta(3/2)\approx2.612$ , where a Bose-Einstein condensate appears. The same number, for fermionic atoms like lithium-6, marks the onset of a degenerate Fermi gas used to simulate electrons in solids. In condensed-matter physics, $n\lambda^3\gg1$ for conduction electrons is why metals require the full Fermi-Dirac treatment — the Sommerfeld theory of the electronic heat capacity, the Pauli paramagnetism, and the very stability of white dwarfs against gravity all rest on the failure of the classical limit. The statistical virial correction $\pm n\lambda^3/2^{5/2}$ , with its characteristic $\lambda^3\propto T^{-3/2}$ dependence, has been measured directly in weakly interacting quantum gases as the leading exchange correction to the equation of state in the near-classical regime.

The physical picture

Picture each particle as a thermal wavepacket of characteristic size $\lambda$ , the thermal de Broglie wavelength. If the typical spacing $n^{-1/3}$ is much larger than $\lambda$ , neighboring wavepackets overlap only negligibly, so exchange effects are negligible and Bose-Einstein and Fermi-Dirac statistics become experimentally indistinguishable from Maxwell-Boltzmann statistics. Exchange effects — the Pauli veto that keeps two fermions apart, the bunching that draws bosons together — only bite when packets overlap, which happens only when $n\lambda^3\gtrsim1$ . In the dilute, hot gas, the quantum statistics is a whisper. In the grand-canonical picture the same thing reads off the occupation numbers. When nearly every single-particle state has occupation $n_\varepsilon\ll1$ , the $\pm1$ in the Bose--Einstein and Fermi--Dirac denominators becomes a small correction, since the fugacity satisfies $z=e^{\beta\mu}\ll1$ . Heating the gas spreads its particles over exponentially many states, diluting occupancy and driving $z\to0$ ; lowering the density does the same. The classical ideal gas is the universal high-entropy attractor that any ideal quantum gas flows toward when it is warmed or thinned.

Figure

Three particles as fuzzy wavepackets of size λ. Left: dilute (spacing n^(-1/3) ≫ λ), packets never overlap, statistics irrelevant, classical. Right: dense (spacing ≈ λ, nλ³ ≈ 1), packets overlap, exchange matters, quantum-degenerate.

The second figure is the punchline: plot all three distributions — Bose-Einstein, Maxwell-Boltzmann, Fermi-Dirac — against $(\epsilon-\mu)/k_BT$ . Near $\mu$ they split sharply: the Fermi-Dirac occupation is capped at one (anti-bunching), the Bose-Einstein occupation soars (bunching), and the classical Maxwell-Boltzmann sits between them. But for $(\epsilon-\mu)/k_BT\gtrsim1$ – $2$ all three merge onto the same exponential tail. That convergence is the classical limit: once occupations drop below one, the sign of the statistics no longer matters, and the three curves become indistinguishable.

Figure

Punchline: Bose-Einstein (top), Maxwell-Boltzmann (middle), Fermi-Dirac (bottom, capped at 1) occupation vs (ε−μ)/kT. They split near μ but converge onto the same exponential tail for (ε−μ)/kT ≳ 1-2 — that merge is the classical limit.

The core equations

The starting point is the grand-canonical occupation number of an ideal quantum gas — the mean number of particles in a single-particle state of energy $\epsilon$ . Fermions and bosons differ only by a sign in the denominator, and everything about the classical limit follows from what happens to that sign.

\langle n_\epsilon\rangle=\frac{1}{e^{\beta(\epsilon-\mu)}\pm1} \qquad (1)

where: $\langle n_\epsilon\rangle$ — mean occupation of the state of energy $\epsilon$ ; $\beta=1/k_BT$ ; $\mu$ — chemical potential; the upper sign ( $+$ ) is Fermi-Dirac (occupation capped at one), the lower sign ( $-$ ) is Bose-Einstein (unbounded). Reading it: when $e^{\beta(\epsilon-\mu)}\gg1$ the $\pm1$ is negligible and $\langle n_\epsilon\rangle\to e^{-\beta(\epsilon-\mu)}$ — the Maxwell-Boltzmann occupation, identical for both statistics. Keeping the next term reveals the sign of the statistics. Writing $w=z\,e^{-\beta\epsilon}$ with the fugacity $z=e^{\beta\mu}$ , the geometric expansion gives the leading correction.

\langle n_\epsilon\rangle=\frac{w}{1\pm w}=w\big(1\mp w+\cdots\big),\qquad w=z\,e^{-\beta\epsilon} \qquad (2)

where: the leading term $w$ is Maxwell-Boltzmann; the correction $\mp w^2$ carries the statistics — for bosons ( $-$ in the denominator) the correction is $+w^2$ (occupation enhanced, bunching), while for fermions ( $+$ ) it is $-w^2$ (occupation suppressed, anti-bunching). The correction is exponentially small whenever $w=z\,e^{-\beta\epsilon}\ll1$ . The fugacity is tied to the density through the thermal wavelength, which is what makes $n\lambda^3$ the master parameter. Summing the occupations over all states and converting to a momentum integral gives the density as a Bose/Fermi function of $z$ .

\frac{n\lambda^3}{g}=f_{3/2}(z)=z\mp\frac{z^2}{2^{3/2}}+\frac{z^3}{3^{3/2}}\mp\cdots,\qquad \lambda=\frac{h}{\sqrt{2\pi m k_BT}} \qquad (3)

where: $n=N/V$ is the number density, $g$ is the spin degeneracy, and $\lambda$ is the thermal de Broglie wavelength; the upper sign is for Fermi statistics and the lower sign for Bose statistics. To leading order, $n\lambda^3/g=z$ , so the dilute condition $z\ll1$ is equivalent to $n\lambda^3\ll1$ . This is the criterion for the classical limit. Dividing the pressure ( $f_{5/2}$ ) by the density ( $f_{3/2}$ ) and re-expanding gives the equation of state — with a virial coefficient that has no force behind it.

\frac{PV}{Nk_BT}=\frac{f_{5/2}(z)}{f_{3/2}(z)}=1\pm\frac{n\lambda^3}{2^{5/2}}+\mathcal O\big((n\lambda^3)^2\big) \qquad (4)

where: the upper sign ( $+$ ) is Fermi — the statistical repulsion of Pauli exclusion raises the pressure; the lower sign ( $-$ ) is Bose — statistical attraction (bunching) lowers it; $2^{5/2}=4\sqrt2\approx5.657$ . As $n\lambda^3\to0$ this returns $PV=Nk_BT$ , the ideal-gas law, independent of statistics — the classical limit proper.

Derivation ABoth quantum distributions collapse onto Maxwell-Boltzmann

The goal is to show explicitly that both the Fermi-Dirac and Bose-Einstein occupation numbers reduce to the same Maxwell-Boltzmann form when the fugacity is small, and to extract the leading correction that carries the sign of the statistics. The strategy is to rewrite the occupation in terms of the small variable $w=z\,e^{-\beta\epsilon}$ and expand geometrically, watching how the $\pm1$ in the denominator produces a $\mp$ in the correction — the whole physical content of quantum statistics in the near-classical regime.

Step 1. Multiply numerator and denominator of equation (1) by $z\,e^{-\beta\epsilon}=e^{-\beta(\epsilon-\mu)}$ to get $\langle n_\epsilon\rangle=\dfrac{z\,e^{-\beta\epsilon}}{1\pm z\,e^{-\beta\epsilon}}=\dfrac{w}{1\pm w}$ with $w=z\,e^{-\beta\epsilon}$ .

Step 2. Expand geometrically for $w<1$ : $\dfrac{w}{1\pm w}=w\big(1\mp w+w^2\mp\cdots\big)=w\mp w^2+w^3\mp\cdots$ .

Step 3. The leading term $w=z\,e^{-\beta\epsilon}$ is the Maxwell-Boltzmann occupation, identical for both statistics. As $z\to0$ (equivalently $n\lambda^3\to0$ ) only this term survives: $\langle n_\epsilon\rangle\to z\,e^{-\beta\epsilon}$ . Why the sign is the whole story. The correction is $\mp w^2$ : for a Bose gas the denominator carries $-1$ , so the correction is $+w^2$ — the occupation is enhanced relative to classical (bosons bunch). For a Fermi gas the denominator carries $+1$ , so the correction is $-w^2$ — the occupation is suppressed (fermions anti-bunch, a foreshadow of the Pauli veto). The magnitude vanishes as $w\to0$ ; only the sign remembers what the particles are.

Step 4. Both series begin with the same term and differ only in the sign of every even-order correction, so at leading order the statistics is invisible — the classical limit. The corrections are of relative order $w=z\,e^{-\beta\epsilon}\le z=n\lambda^3\ll1$ . The expansions $1/(e^x+1)=e^{-x}-e^{-2x}+\cdots$ (FD, negative correction) and $1/(e^x-1)=e^{-x}+e^{-2x}+\cdots$ (BE, positive correction) confirm the unified $w/(1\pm w)$ series with its $\mp$ correction.

Go deeper · Derivation B · 1 minThe statistical virial coefficient: repulsion for fermions, attraction for bosons

Derivation BThe statistical virial coefficient: repulsion for fermions, attraction for bosons

The goal is to turn the fugacity expansion into a virial expansion of the equation of state in powers of $n\lambda^3$ , producing a second virial coefficient of purely statistical origin. The strategy is to invert the density relation $n\lambda^3/g=f_{3/2}(z)$ to get $z$ in terms of $n\lambda^3$ , form the pressure ratio $f_{5/2}(z)/f_{3/2}(z)$ , and re-expand — the sign of the leading correction is the sign of the effective statistical interaction.

Step 1. Set $g=1$ and write $x=n\lambda^3$ . The density and pressure functions are $f_{3/2}(z)=z\mp z^2/2^{3/2}+\cdots$ and $f_{5/2}(z)=z\mp z^2/2^{5/2}+\cdots$ (upper Fermi, lower Bose).

Step 2. Invert the density relation $x=z\mp z^2/2^{3/2}+\cdots$ by series reversion: $z=x\pm x^2/2^{3/2}+\cdots$ .

Step 3. Form the compressibility factor $\dfrac{PV}{Nk_BT}=\dfrac{f_{5/2}(z)}{f_{3/2}(z)}=\dfrac{z\mp z^2/2^{5/2}}{z\mp z^2/2^{3/2}}=1\pm z\Big(\dfrac{1}{2^{3/2}}-\dfrac{1}{2^{5/2}}\Big)+\cdots$ . Why the coefficient is $1/2^{5/2}$ . The difference $\dfrac{1}{2^{3/2}}-\dfrac{1}{2^{5/2}}=\dfrac{1}{2^{5/2}}(2-1)=\dfrac{1}{2^{5/2}}$ , so the two polylog series combine into a single clean coefficient.

Step 4. Substitute the leading-order relation $z\simeq n\lambda^3$ : $\dfrac{PV}{Nk_BT}=1\pm\dfrac{n\lambda^3}{2^{5/2}}$ . Fermions ( $+$ ) raise the pressure — an effective Pauli repulsion; bosons ( $-$ ) lower it — an effective statistical attraction (bunching). This is a genuine second virial coefficient, $B_2^{\text{stat}}=\mp\lambda^3/2^{5/2}$ , arising with no interparticle force whatsoever. The series reversion yields the pressure ratio $1\pm z/2^{5/2}$ (with $+$ Fermi, $-$ Bose), where $2^{5/2}=4\sqrt2\approx5.657$ , and the $z\to0$ limit correctly recovers $PV=Nk_BT$ .

Limiting cases

The regimes are organized by the degeneracy parameter $n\lambda^3$ ; each row marks a specialization or the boundary where the classical description breaks.

Case	Result	Meaning
Classical, $n\lambda^3\ll1$	$\langle n_\epsilon\rangle\approx z\,e^{-\beta\epsilon}$ , $PV=Nk_BT$	Maxwell-Boltzmann statistics; ideal-gas law and Sackur-Tetrode entropy hold.
First correction	$\dfrac{PV}{Nk_BT}=1\pm\dfrac{n\lambda^3}{2^{5/2}}$	$+$ fermions (statistical repulsion), $-$ bosons (statistical attraction).
Degenerate, $n\lambda^3\gtrsim1$	$z\sim\mathcal O(1)$ or $z>1$	Fermi: filled Fermi sea, $\mu\to E_F>0$ . Bose: condensation at $n\lambda^3=\zeta(3/2)\approx2.612$ .
Formal $\hbar\to0$	$\lambda\to0$ , $n\lambda^3\to0$ always	Non-degenerate for all $n,T$ , but $h^3$ in the entropy diverges — absolute entropy loses meaning.

Common trapThe classical limit means taking Planck's constant to zero

The classical limit is not obtained by taking $\hbar\to0$ . Planck's constant plays two indispensable roles: it sets the phase-space cell $h^3$ , giving entropy an absolute scale, and it enforces the indistinguishability factor $1/N!$ that keeps entropy extensive. Formally setting $\hbar\to0$ does send $\lambda\to0$ and $n\lambda^3\to0$ (always non-degenerate), but it also sends the absolute entropy to $+\infty$ because the phase-space cell vanishes. Drop the $1/N!$ and you resurrect the Gibbs paradox; drop $h^3$ and there is no third law. Also beware that "high temperature" means high relative to the degeneracy temperature $T_d\sim\hbar^2 n^{2/3}/mk_B$ , not high in absolute terms — helium at $4\,\mathrm K$ can be more classical than electrons at $10^4\,\mathrm K$ , because the tiny electron mass makes $\lambda$ enormous.

Worked examples

Example 1 — Is ordinary air classical? Compute the degeneracy parameter and the quantum pressure correction (enumerate, apply, boundary number). Take nitrogen at $T=300\,\mathrm K$ and $P=1\,\mathrm{atm}$ , so the number density is $n=P/k_BT\approx2.5\times10^{25}\,\mathrm m^{-3}$ .

Step 1. Compute the thermal wavelength: $\lambda=h/\sqrt{2\pi m k_BT}$ with $m=28\times1.66\times10^{-27}\,\mathrm{kg}=4.65\times10^{-26}\,\mathrm{kg}$ . This gives $\lambda=1.9\times10^{-11}\,\mathrm m$ — about $0.019\,\mathrm{nm}$ , far smaller than an atom.

Step 2. Apply the degeneracy criterion: the interparticle spacing is $n^{-1/3}\approx3.4\,\mathrm{nm}$ , so $n\lambda^3=(2.5\times10^{25})(1.9\times10^{-11})^3\approx1.7\times10^{-7}$ . The packets are $\sim180$ times smaller than the spacing.

Step 3. Boundary number: the quantum correction to the pressure is $|\Delta P/P|=n\lambda^3/2^{5/2}\approx1.7\times10^{-7}/5.66\approx3\times10^{-8}$ . The deviation from the ideal-gas law due to quantum statistics is about three parts in a hundred million — utterly negligible, which is exactly why classical thermodynamics works so well for everyday gases.

Example 2 — Why do metals need Fermi-Dirac statistics? The conduction-electron degeneracy parameter. Conduction electrons in copper have $n\approx8.5\times10^{28}\,\mathrm m^{-3}$ and the tiny electron mass $m_e=9.1\times10^{-31}\,\mathrm{kg}$ , still at $T=300\,\mathrm K$ .

Step 1. The thermal wavelength scales as $\lambda\propto m^{-1/2}$ , so the small electron mass makes it large: $\lambda_e=h/\sqrt{2\pi m_ek_BT}\approx4.3\times10^{-9}\,\mathrm m=4.3\,\mathrm{nm}$ .

Step 2. Apply the criterion: $n\lambda^3=(8.5\times10^{28})(4.3\times10^{-9})^3\approx6800\gg1$ . The spacing $n^{-1/3}\approx0.23\,\mathrm{nm}$ is twenty times smaller than the thermal wavelength — the wavepackets overlap massively.

Step 3. Boundary number: with $n\lambda^3\approx7\times10^3$ , the classical expansion in $n\lambda^3$ is completely invalid—the electron gas is deeply degenerate, with a Fermi temperature $T_F\sim10^5\,\mathrm{K}\gg T$ . Maxwell--Boltzmann statistics would predict the classical equipartition result for the electronic heat capacity, whereas the observed electronic heat capacity is proportional to $T$ because only electrons within about $k_BT$ of the Fermi surface can be thermally excited. This is the textbook demonstration that Maxwell--Boltzmann statistics is inapplicable to conduction electrons in metals.

Origins · 1 minclassical limit and how this was discovered

classical limit

quantum-statistics concept · Bose 1924, Einstein 1924-25, Fermi & Dirac 1926

The classical limit is the regime in which the observable consequences of quantum statistics become negligible. Satyendra Nath Bose (1924) derived Planck's radiation law by a new counting of photon states, and Einstein (1924–25) extended it to massive bosons, predicting condensation. Enrico Fermi and Paul Dirac independently (1926) found the complementary statistics for particles obeying the Pauli principle. In every case the new distributions reduce to Maxwell-Boltzmann when occupations fall below one — a fact the founders emphasized as the bridge back to classical kinetic theory. The criterion $n\lambda^3\ll1$ , with $\lambda$ the thermal de Broglie wavelength, became the standard measure of whether a gas is classical or quantum degenerate. The same parameter now distinguishes a dilute thermal cloud from a Bose-Einstein condensate.

History

1900

— Planck introduces the quantum of action, implicitly the phase-space cell that survives the classical limit.

1924

— Bose derives Planck's law by counting photon states; Einstein extends it to massive bosons.

1926

— Fermi and Dirac independently formulate the statistics of particles obeying the exclusion principle.

1927

— Sommerfeld applies Fermi-Dirac statistics to metal electrons, resolving the electronic heat-capacity puzzle.

1995

— Cornell, Wieman, and Ketterle realize Bose-Einstein condensation in dilute atomic gases at nλ³ ≈ 2.6.

Going further

References. Pathria & Beale, Statistical Mechanics, 3rd/4th ed., Ch. 6–7 · Kardar, Statistical Physics of Particles, Ch. 7 · Reif, Fundamentals of Statistical and Thermal Physics, Ch. 9 · Sethna, Statistical Mechanics: Entropy, Order Parameters, and Complexity, Ch. 7 · Landau & Lifshitz, Statistical Physics Part 1, §45, §53–56. The Bose/Fermi polylog function series is $f_\nu(z)=\sum_k(\mp1)^{k+1}z^k/k^\nu$ . Key results: the statistical virial coefficient is $\pm1/2^{5/2}$ ( $2^{5/2}=4\sqrt2\approx5.657$ ); the BEC threshold is $\zeta(3/2)\approx2.612$ ; the FD and BE expansions give $1/(e^x+1)=e^{-x}-e^{-2x}+\cdots$ and $1/(e^x-1)=e^{-x}+e^{-2x}+\cdots$ (the $\mp$ correction); the density series is $f_{3/2}(z)=z\mp z^2/2^{3/2}+\cdots$ ; the pressure ratio is $f_{5/2}/f_{3/2}=1\pm z/2^{5/2}$ ( $+$ Fermi, $-$ Bose), recovering $PV=Nk_BT$ as $z\to0$ ; and $n\lambda^3\sim10^{-7}$ for air and $\sim10^3$ for metal electrons. Constants used: $h=6.626\times10^{-34}\,\mathrm{J\,s}$ , $k_B=1.380649\times10^{-23}\,\mathrm{J/K}$ .

Builds on: Thermal de Broglie Wavelength · Correct Boltzmann Counting · Classical Ideal Gas · Equipartition Theorem Leads to: Sackur-Tetrode Equation · Gibbs Paradox · Virial Equation of State · Maxwell-Boltzmann Velocity Distribution

Why do both Bose and Fermi statistics give the same classical limit?

Because the difference between them is a single $\pm1$ in the denominator of the occupation number, and that $\pm1$ becomes negligible precisely when occupations fall below one. Write the occupation as $w/(1\pm w)$ with $w=z\,e^{-\beta\epsilon}$ ; the leading term $w$ is Maxwell-Boltzmann and is identical for both, while the statistics only enters the correction $\mp w^2$ . When $w\ll1$ — the dilute, hot regime $n\lambda^3\ll1$ — the correction is exponentially small, so both distributions collapse onto the same exponential. The particles are still bosons or fermions; it is just that the gas is so sparse that they almost never meet in the same state, so exchange has nothing to act on.

What is the physical meaning of the degeneracy parameter nλ³?

It is the number of particles within a thermal de Broglie volume $\lambda^3$ — equivalently, the ratio $(\lambda/n^{-1/3})^3$ of the wavepacket size to the interparticle spacing, cubed. When $n\lambda^3\ll1$ each particle's wavepacket occupies a region far smaller than the space around it, so wavepackets do not overlap and the particles behave like distinguishable classical objects. When $n\lambda^3\sim1$ the packets begin to overlap and quantum statistics takes over. It is the single dimensionless combination that decides classical-versus-quantum, which is why the same criterion works for air ( $10^{-7}$ ), helium near condensation ( $\sim2.6$ ), and metal electrons ( $\sim10^3$ ).

Is the classical limit the same as taking ℏ to zero?

No, and this is a common trap. Setting $\hbar\to0$ does drive $\lambda\to0$ and $n\lambda^3\to0$ , making the gas formally non-degenerate — but it also makes the phase-space cell $h^3$ vanish, so the absolute entropy diverges to $+\infty$ and the third law is lost. The physical classical limit keeps $\hbar$ finite and is reached by raising the temperature or lowering the density. Two pieces of $\hbar$ survive into the classical description: the cell $h^3$ that sets the zero of entropy, and the Gibbs factor $1/N!$ that keeps entropy extensive. The classical gas is a limit of the quantum gas, not a world where $\hbar$ has been switched off.

Why does the Gibbs 1/N! survive if it is quantum in origin?

Because indistinguishability is exact at all temperatures — permuting two identical particles never produces a new microstate, no matter how dilute the gas. What the classical limit switches off is the exchange-energy effect, which requires wavepacket overlap and vanishes as $n\lambda^3\to0$ . The combinatorial fact that identical particles cannot be labeled does not depend on overlap, so the factor $1/N!$ that corrects the overcounting remains. This is why the classical ideal-gas entropy is the Sackur-Tetrode formula with a $V/N$ (not $V$ ) inside the logarithm: the $1/N!$ makes the entropy extensive and resolves the Gibbs paradox, a purely quantum remnant in an otherwise classical result.

Do fermions really "repel" and bosons "attract" without any force?

Yes — this is one of the most striking features of the near-classical limit. The leading correction to the equation of state, $PV/Nk_BT=1\pm n\lambda^3/2^{5/2}$ , acts exactly like a second virial coefficient, even though the gas is ideal with no interparticle potential. Fermions ( $+$ ) push the pressure up as if they repelled one another: the Pauli principle keeps them out of the same state, an effective short-range repulsion. Bosons ( $-$ ) pull the pressure down as if they attracted: they preferentially occupy the same states (bunching), an effective attraction. These "statistical interactions" are genuine and measurable, distinguished from real forces by their characteristic $\lambda^3\propto T^{-3/2}$ temperature dependence.

How large can nλ³ get for a Bose gas before the classical picture is hopeless?

For an ideal Bose gas the excited states can only hold $n\lambda^3=\zeta(3/2)\approx2.612$ particles per thermal volume (with spin degeneracy $g=1$ ). Beyond this the surplus particles pile into the single-particle ground state, forming a Bose-Einstein condensate. Long before that point — already at $n\lambda^3\sim1$ — the expansion in powers of $n\lambda^3$ has broken down, so the classical description and its first virial correction are useless there. The value $2.612$ is the sharp threshold where a macroscopic condensate appears; it is the number ultracold-atom experiments aim for when they cool a thermal cloud into a condensate.

At the same temperature, which becomes classical first as the density drops — a Bose or a Fermi gas?

Both become classical simultaneously. The leading criterion $n\lambda^3\ll1$ is identical and statistics-independent, because at leading order $n\lambda^3=z$ for both. As you dilute either gas at fixed temperature, the fugacity $z$ falls and both distributions approach Maxwell-Boltzmann in lockstep. The only difference is the sign of the first correction: the Fermi gas keeps a slightly higher pressure (residual statistical repulsion) and the Bose gas a slightly lower one (residual attraction) at the same $n\lambda^3$ . So they reach the classical limit together, but they approach it from opposite sides.