Identical-Particle Scattering

Collide two identical particles and a detector at angle $\theta$ faces a question it cannot answer: did particle A scatter by $\theta$ , or did particle B scatter by $\theta$ while A recoiled to $\pi-\theta$ ? In the center-of-mass frame the two emerge back-to-back, so the two histories end in the same detector click. Quantum mechanics forbids us to ask which one happened; the rule is to add — or subtract — the amplitudes for the two indistinguishable histories before squaring. That single prescription reorganizes scattering theory whenever projectile and target are copies of one particle. The starting point is the ordinary scattering amplitude $f(\theta)$ , whose modulus-squared gives the differential cross section for distinguishable particles — owned by Scattering Amplitude. The new ingredient is exchange symmetry: swapping the two particles sends the relative coordinate $\mathbf r\to-\mathbf r$ , i.e. $\theta\to\pi-\theta$ . A spatially symmetric state therefore scatters with $f(\theta)+f(\pi-\theta)$ and an antisymmetric one with $f(\theta)-f(\pi-\theta)$ . Squaring produces a cross term that is pure quantum interference — invisible in any classical sum of probabilities. This article owns four results. The symmetrized-amplitude theorem and the $4{:}2{:}0$ ninety-degree ladder — where spinless bosons scatter four times the single-particle rate, distinguishable particles twice, and polarized fermions zero. The spin- $\tfrac12$ weights that split unpolarized electron scattering into $\tfrac14$ singlet plus $\tfrac34$ triplet, consuming the space-spin correlation theorem of Singlet and Triplet States by name. The Mott $\alpha$ – $\alpha$ oscillation laboratory, where identical-nucleus Coulomb scattering interferes into a curve that hits exactly twice the classical rate at $90^\circ$ . And the even- $\ell$ selection theorem, which forbids identical bosons the odd partial waves and blocks the s-wave for polarized fermions — the reason spin-polarized Fermi gases stop thermalizing near absolute zero.

Key takeaways

For identical particles you add amplitudes for the two indistinguishable histories before squaring: $f_\pm=f(\theta)\pm f(\pi-\theta)$ . At $90^\circ$ this gives spinless bosons $4|f|^2$ , distinguishable particles $2|f|^2$ , and polarized fermions $0$ . Unpolarized spin- $\tfrac12$ scattering is $\tfrac14$ singlet plus $\tfrac34$ triplet, producing a $-\mathrm{Re}[f f^*(\pi-\theta)]$ interference term. Legendre parity forbids identical bosons the odd partial waves and blocks the s-wave for polarized fermions; identical-boson s-wave scattering is $8\pi a^2$ , twice the distinguishable $4\pi a^2$ .

The core insight

Indistinguishability is not a small correction — it is a factor of $4$ , or $2$ , or $0$ , sitting right at $90^\circ$ where the two histories become the same event. You cannot count the particles, so you must add their amplitudes, and the cross term is a real, measurable interference.

Where this lives today

Identical-particle scattering is the daily physics of ultracold atoms. Evaporative cooling of bosons relies on the enhanced $8\pi a^2$ collision rate to rethermalize; cooling of polarized fermions fails for exactly the mirror reason — the blocked s-wave — which is why fermionic cooling requires two spin states or sympathetic cooling against a boson. The even- $\ell$ selection rule also gates chemistry near absolute zero: identical-molecule collisions are restricted to symmetric partial waves. In nuclear and particle physics the same interference produces the missing even- $\ell$ (or odd- $\ell$ ) states of identical-boson nuclei and the Hanbury Brown–Twiss intensity correlations of identical bosons and fermions emitted from a source. The $90^\circ$ Mott doubling remains a textbook benchmark and a live calibration in heavy-ion elastic scattering.

The physical picture

A classical collision of two identical billiard balls is still two distinguishable events — you could paint one red. Quantum mechanically you cannot: the two-particle state must be symmetric (bosons) or antisymmetric (fermions) under exchange, and there is no red paint. The consequence is that the amplitude for "scatter into $\theta$ " already contains the amplitude for "scatter into $\pi-\theta$ ," locked together by the exchange sign.

Figure

The two indistinguishable histories that end in the same detector click: deflection by θ and recoil by π−θ are the same event

The upshot is a rule with three faces. If the spatial state is symmetric (spinless bosons, or fermions in an antisymmetric spin state), you use $f_+=f(\theta)+f(\pi-\theta)$ . If antisymmetric (spinless fermions, or bosons in an antisymmetric spin state), you use $f_-=f(\theta)-f(\pi-\theta)$ . And when the spins are unpolarized, you average over the spin multiplets with their statistical weights. Everything below is bookkeeping on that one rule — and the bookkeeping is dramatic exactly at $90^\circ$ .

Figure

Punchline — the 4:2:0 ladder and Mott α-α oscillation: quantum identical-boson curve swings around the classical-identical baseline, hitting exactly ×2 at 90°

The symmetrized amplitude

Exchange maps the relative coordinate $\mathbf r\to-\mathbf r$ , which in spherical angles is $\theta\to\pi-\theta$ , $\phi\to\phi+\pi$ . This is the whole geometric content; the amplitude-vs-probability rule is the double-slit grammar of The Double-Slit Experiment and the Born rule of Born Rule and Measurement Probabilities, recalled here in one line: indistinguishable histories add at the amplitude level.

Derivation AThe symmetrized-amplitude theorem and the 4:2:0 ladder

The goal is to turn "you cannot tell the two histories apart" into a formula for the differential cross section, and to read off the striking factors of $4$ , $2$ , and $0$ that appear at $90^\circ$ . The strategy: symmetrize the asymptotic scattering state under particle exchange, identify the two amplitudes it superposes, and square. What to watch: the cross term $2\,\mathrm{Re}[f(\theta)f^*(\pi-\theta)]$ that has no classical counterpart, and its sign, which is the exchange sign.

Step 1. Exchange symmetry of the spatial state. For a central potential the relative-coordinate scattering state is built from $f(\theta)$ . Under exchange $\theta\to\pi-\theta$ , so the properly (anti)symmetrized spatial amplitude is

f_\pm(\theta)=f(\theta)\pm f(\pi-\theta)

where $+$ is the symmetric spatial state and $-$ the antisymmetric. Why $\theta\to\pi-\theta$ : swapping the two particles reflects the relative coordinate through the origin, and $\mathbf r\to-\mathbf r$ is exactly $\theta\to\pi-\theta$ in the CM frame.

Step 2. Square to the cross section.

\frac{d\sigma}{d\Omega}=|f(\theta)\pm f(\pi-\theta)|^2=|f(\theta)|^2+|f(\pi-\theta)|^2\pm 2\,\mathrm{Re}\big[f(\theta)f^*(\pi-\theta)\big]

Why the cross term is quantum: a classical count of two distinguishable histories gives $|f(\theta)|^2+|f(\pi-\theta)|^2$ — no interference. The $\pm 2\,\mathrm{Re}[\dots]$ term is present only because we added amplitudes, and it carries the exchange sign.

Step 3. Evaluate at $90^\circ$ . There $\pi-\theta=\theta$ , so $f(\pi-\theta)=f(\theta)$ and

\left.\frac{d\sigma}{d\Omega}\right|_{90^\circ}= \begin{cases}|2f|^2=4|f|^2 & \text{symmetric (spinless boson)}\\[2pt]2|f|^2 & \text{distinguishable (classical identical)}\\[2pt]|0|^2=0 & \text{antisymmetric (polarized fermion)}\end{cases}

Why the ladder is exact: at $90^\circ$ the two amplitudes are literally equal, so constructive interference doubles the amplitude (quadruples the rate), while destructive interference cancels it exactly. No approximation. Verified: verify script §1–2 — the $4{:}2{:}0$ ratios reproduced numerically for a representative $f$ , the classical-identical sum confirmed to carry no cross term, and the interference term isolated with its exchange sign; 26-check suite green.

Spin-½: the unpolarized weights

Electrons carry spin, so "symmetric or antisymmetric spatial" is decided by the spin state through the requirement that the total two-fermion state be antisymmetric. Two spin- $\tfrac12$ particles form one antisymmetric singlet ( $S=0$ ) and three symmetric triplet states ( $S=1$ ) — the construction owned by Singlet and Triplet States, whose theorem we consume by name: electron pairs in a symmetric spatial state must be spin-singlets, pairs in an antisymmetric spatial state must be triplets. Get the spin labels right, because they are easy to swap and doing so poisons everything downstream: the singlet is antisymmetric under spin exchange (swap eigenvalue $-1$ ) and the three triplet states are symmetric (eigenvalue $+1$ ) — exactly as Singlet and Triplet States serves it ("the three triplet states are even under swapping the two particles; the singlet is odd"), and confirmable in two lines by acting the swap operator on the four spin states. (Some sources, including the base corpus for this article, print the labels reversed; that error is corrected and machine-verified here.) Because the total fermion state is antisymmetric, the singlet (antisymmetric spin) pairs with the symmetric spatial amplitude $f_+$ , and the triplet (symmetric spin) with the antisymmetric spatial $f_-$ . An unpolarized beam populates the four spin states equally, giving weights $\tfrac14$ singlet and $\tfrac34$ triplet:

\frac{d\sigma}{d\Omega}=\tfrac14|f(\theta)+f(\pi-\theta)|^2+\tfrac34|f(\theta)-f(\pi-\theta)|^2

where: the $\tfrac14$ and $\tfrac34$ are the multiplicities $1$ and $3$ normalized by the total of $4$ spin states. Expanding, the direct terms add to $|f(\theta)|^2+|f(\pi-\theta)|^2$ and the cross terms combine with coefficient $\tfrac14(+2)+\tfrac34(-2)=-1$ :

\frac{d\sigma}{d\Omega}=|f(\theta)|^2+|f(\pi-\theta)|^2-\mathrm{Re}\big[f(\theta)f^*(\pi-\theta)\big]

where: the minus sign is the fingerprint of the fermionic majority — the triplet's antisymmetric spatial amplitude wins $3$ to $1$ , so unpolarized electrons interfere destructively on balance. This is the spin-averaged Mott formula for identical spin- $\tfrac12$ particles.

The Mott α–α oscillation laboratory

Now a real, historically decisive case: two $\alpha$ particles (spin- $0$ bosons, charge $2e$ ) scattering through their Coulomb repulsion. The Coulomb amplitude is owned — its Rutherford recovery lives in Yukawa and Coulomb Scattering — so we quote it and symmetrize.

Go deeper · Derivation B · 1 minMott scattering: the α-α interference and the exact factor of 2 at 90°

Derivation BMott scattering: the α-α interference and the exact factor of 2 at 90°

The goal is the identical-boson Coulomb cross section and its signature at $90^\circ$ : exactly twice the classical-identical rate, sitting on top of angular oscillations that Chadwick's 1930 experiment confirmed. The strategy is to take the known Coulomb amplitude, symmetrize it for the two identical spin- $0$ nuclei, and evaluate. What to watch: the Coulomb phases at $\theta$ and $\pi-\theta$ differ away from $90^\circ$ , producing the oscillations, but coincide at $90^\circ$ , producing the clean factor of $2$ .

Step 1. The Coulomb amplitude (quoted with source). For two charges $Z_1Z_2e^2$ the amplitude has modulus $\propto\csc^2(\theta/2)$ and a logarithmic Coulomb phase, with Sommerfeld parameter $\eta=Z_1Z_2e^2/\hbar v$ :

f_C(\theta)=-\frac{\eta}{2k}\csc^2\!\tfrac{\theta}{2}\,\exp\!\big[-i\eta\ln\sin^2\tfrac{\theta}{2}\big]\times(\text{overall phase})

Why quote it: the Rutherford recovery $|f_C|^2=(\eta/2k)^2\csc^4(\theta/2)$ is owned by the Coulomb-scattering article; we use it, we do not re-derive it.

Step 2. Symmetrize for two spin-0 bosons.

\frac{d\sigma}{d\Omega}=|f_C(\theta)+f_C(\pi-\theta)|^2 =|f_C(\theta)|^2+|f_C(\pi-\theta)|^2+2\,\mathrm{Re}\big[f_C(\theta)f_C^*(\pi-\theta)\big]

Why $+$ : $\alpha$ particles are identical spin- $0$ bosons, so the spatial state is symmetric.

Step 3. The interference and the $90^\circ$ point. The cross term carries a phase $\propto\eta\ln\tan^2(\theta/2)$ , which oscillates in angle — the Mott oscillations. At $\theta=90^\circ$ , $\pi-\theta=\theta$ , the two amplitudes are equal, the phase difference vanishes, and

\left.\frac{d\sigma}{d\Omega}\right|_{90^\circ}=|2f_C|^2=4|f_C|^2=2\times\underbrace{\big(|f_C(\theta)|^2+|f_C(\pi-\theta)|^2\big)}_{\text{classical identical}}

Why exactly $2$ : the classical-identical count at $90^\circ$ is $2|f_C|^2$ ; the quantum symmetric count is $4|f_C|^2$ ; the ratio is $2$ with no phase ambiguity, because $\theta=\pi-\theta$ kills the log-phase difference. Verified: verify script §5 — the symmetrized Coulomb cross section computed with the exact log-phase amplitude; the $90^\circ$ ratio to classical-identical is $2.000000$ ; away from $90^\circ$ the ratio oscillates (e.g. $1.51$ at $60^\circ$ ), confirming the Mott interference; 26-check suite green.

The factor of $2$ is not a curiosity: Chadwick measured $\alpha$ -on-helium scattering in 1930 and found exactly this doubling at $90^\circ$ , confirming that identical nuclei interfere and that $\alpha$ particles are bosons. Had they been fermions, the $90^\circ$ cross section would have vanished.

The even-ℓ selection theorem

The partial-wave content of the symmetrized amplitude is fixed by Legendre parity. Since $P_\ell(-x)=(-1)^\ell P_\ell(x)$ , and $\pi-\theta$ sends $\cos\theta\to-\cos\theta$ :

f(\theta)\pm f(\pi-\theta)=\sum_\ell(2\ell+1)f_\ell\,P_\ell(\cos\theta)\big[1\pm(-1)^\ell\big]

where: the bracket is $2$ for even $\ell$ (symmetric case) or odd $\ell$ (antisymmetric case), and $0$ otherwise. So identical bosons scatter only through even partial waves; the spinless-fermion channel only through odd. At low energy this is decisive: the s-wave ( $\ell=0$ ) is even, so identical bosons keep it but polarized fermions lose it, and their leading channel is the $(ka)^3$ -suppressed p-wave. The low-energy consequence, using the scattering-length grammar of Scattering Length and Low-Energy Scattering: the identical-boson s-wave cross section is

\sigma_{\text{boson}}=8\pi a^2\qquad\text{versus}\qquad\sigma_{\text{dist}}=4\pi a^2

where: the extra factor of $2$ is the even- $\ell$ enhancement carried over from the $90^\circ$ ladder, now integrated. For polarized fermions the s-wave is blocked entirely — which is why a spin-polarized Fermi gas stops thermalizing as $T\to0$ : with only the frozen-out p-wave left, collisions cease. This is a working principle in cold-atom evaporative cooling.

Limiting cases

regime	result	why
$\theta=90^\circ$ , spinless boson	$4\|f\|^2$	amplitudes equal, constructive
$\theta=90^\circ$ , polarized fermion	$0$	amplitudes equal, destructive
unpolarized spin- $\tfrac12$	$\\|f\\|^2+\\|f'\\|^2-\mathrm{Re}[ff'^*]$	$\tfrac14$ singlet $+\tfrac34$ triplet
low-energy identical boson	$\sigma=8\pi a^2$	even- $\ell$ ; s-wave doubled
low-energy polarized fermion	$\sigma\sim(ka)^3$	s-wave blocked; p-wave only
$f(\theta)$ isotropic	boson $4$ , fermion $0$ at $90^\circ$	no angular structure to hide the ladder

Worked examples

Example 1 — The α–α doubling at 90°. Two $\alpha$ particles scatter at $E_{\rm cm}\sim1$ MeV via their Coulomb repulsion.

Step 1. Enumerate the histories. A detector at $90^\circ$ receives amplitude $f_C(90^\circ)$ for "one $\alpha$ deflected by $90^\circ$ " and $f_C(90^\circ)$ for "the other deflected there while the first recoiled to $90^\circ$ " — the same angle.

Step 2. Apply the boson rule. $\alpha$ is a spin- $0$ boson, so add amplitudes: $\frac{d\sigma}{d\Omega}=|f_C(90^\circ)+f_C(90^\circ)|^2=4|f_C(90^\circ)|^2$ .

Step 3. The boundary number. The classical-identical count is $2|f_C(90^\circ)|^2$ , so the quantum-to-classical ratio at $90^\circ$ is exactly $\mathbf{2.000}$ . Away from $90^\circ$ the ratio oscillates — e.g. $\mathbf{1.51}$ at $60^\circ$ for $\eta=0.5$ — tracing the Mott interference fringes Chadwick measured in 1930.

Example 2 — ⁸⁷Rb identical-boson collisions at 1 μK. Take the scattering length $a=98\,a_0=5.19$ nm.

Step 1. The cross sections. Distinguishable: $\sigma_{\rm dist}=4\pi a^2$ . Identical bosons: $\sigma_{\rm bos}=8\pi a^2=\mathbf{6.76\times10^{-16}\ m^2}$ — exactly twice.

Step 2. The relative speed. At $T=1\,\mu$ K the mean relative speed is $v_{\rm rel}=\sqrt{8k_BT/\pi\mu}=\mathbf{22\ mm/s}$ (with $\mu=m_{\rm Rb}/2$ ).

Step 3. The collision rate. At a lab density $n\sim10^{14}\,\text{cm}^{-3}$ , the per-atom rate is $n\,\sigma_{\rm bos}\,v_{\rm rel}\approx\mathbf{1.5\times10^{3}\ s^{-1}}$ — the fast thermalization that makes bosonic evaporative cooling work. The identical-boson factor of $2$ doubles this rate versus a distinguishable gas; a spin-polarized fermionic gas would have $\sigma\to0$ and stop thermalizing entirely.

Origins · 1 minExchange Symmetry and how this was discovered

Exchange Symmetry

1930–1961 · Cambridge, Manchester

The idea that identity alone — with no new force — reshapes a cross section is Nevill Mott's. In 1930, at Cambridge, Mott worked out that two identical charged particles must scatter with interfering amplitudes, predicting an $\alpha$ – $\alpha$ and electron–electron pattern that no classical Coulomb calculation could produce; the enhancement at $90^\circ$ became the sharpest test. James Chadwick, at the Cavendish, measured $\alpha$ -on-helium scattering that same year and found the predicted doubling, confirming both the interference and the bosonic nature of the $\alpha$ particle. Mott and Massey codified the whole apparatus in their 1933 Theory of Atomic Collisions, and by the time Mott moved to Bristol and later back to Cambridge the "Mott scattering" of identical particles was a standard chapter — a rare case where a symmetry, not a dynamics, is the measured quantity.

History

1926

— Dirac and Heisenberg establish that identical-particle states must be symmetric or antisymmetric

1930

— Nevill Mott predicts the interference cross section for two identical charged particles

1930

— James Chadwick measures α-on-helium scattering and confirms the 90° doubling for bosonic α particles

1933

— Mott and Massey's Theory of Atomic Collisions codifies identical-particle scattering

1961

— electron-electron (Møller) scattering measurements confirm the fermionic (destructive) signature

Going further

The single-particle amplitude and the current theorem $d\sigma/d\Omega=|f|^2$ are owned by Scattering Amplitude; the counting and luminosity definitions by Differential and Total Cross Section. The space-spin correlation theorem and the singlet/triplet construction are owned by Singlet and Triplet States (and its Heisenberg-1926 ortho/para-helium story stays there). The Coulomb amplitude and Rutherford recovery live in Yukawa and Coulomb Scattering; the partial-wave decomposition in Partial-Wave Expansion; the low-energy scattering length in Scattering Length and Low-Energy Scattering. The symmetrization postulate underlying "which sign" is developed in the Identical Particles chapter — Indistinguishability of Identical Particles, Permutation Symmetry and the Exchange Operator, and Symmetric and Antisymmetric States. Standard treatments: Sakurai & Napolitano §6.4, Cohen-Tannoudji Vol. II complement, Shankar Ch. 19. Provenance note (base corpus): Symmetrized cross section $|f(\theta)\pm f(\pi-\theta)|^2$ and the spin-averaged $\tfrac14|f_+|^2+\tfrac34|f_-|^2$ are verifiable against Sakurai & Napolitano §6.4, Cohen-Tannoudji Vol. II complement, and Shankar Ch. 19. Verifier should confirm: (i) the $\theta\to\pi-\theta$ map from $\mathbf r\to-\mathbf r$ ; (ii) the even/odd- $\ell$ selection via Legendre parity; (iii) singlet/triplet weights $1/4,3/4$ for spin- $\tfrac12$ ; (iv) the Mott formula sign convention and the $90^\circ$ factor-of-2/zero results. Numbers quoted (weights, factor 4 at $90^\circ$ ) are standard textbook values.

Why does the exchange send θ to π−θ specifically?

Swapping the two identical particles negates the relative coordinate $\mathbf r=\mathbf r_1-\mathbf r_2\to-\mathbf r$ . In spherical angles $\mathbf r\to-\mathbf r$ is exactly $\theta\to\pi-\theta$ (with $\phi\to\phi+\pi$ ). Physically, in the CM frame the particles are always back-to-back, so "particle A at $\theta$ " and "particle B at $\theta$ " correspond to the relative vector pointing in opposite directions — the two ends of the same collision.

Is the cross term ever visible, or does it wash out like the plane-wave interference?

It is fully visible — it is the whole point. Unlike the forward-direction plane-wave/ripple cross term (which is exiled to $\theta=0$ ), the identical-particle cross term lives at all angles and peaks at $90^\circ$ . Chadwick measured it directly in 1930. It does not average away because both histories reach the same detector at the same angle.

Do the spins have to be measured for the weights ¼ and ¾ to apply?

No — the weights are for an unpolarized beam, where the four spin states are equally populated and you have not measured spin. If you prepare a definite spin state (say, both spins up = a triplet member), you use $f_-$ with weight $1$ and see the pure fermionic (destructive) pattern, with the $90^\circ$ cross section vanishing. The weights are a statistical average over an unpolarized ensemble.

Why is the total identical-boson cross section 8πa² and not 16πa² (four times)?

Because the factor of $4$ is a pointwise enhancement at $90^\circ$ , not everywhere. When you integrate $|f_++|^2$ over the sphere and correctly count identical final states (over the half-sphere, since scattering to $\theta$ and $\pi-\theta$ is the same event), the net enhancement of the total is $2$ , giving $8\pi a^2$ . The $4$ and the $2$ answer different questions — differential at one angle versus integrated.

Does this effect apply to a projectile scattering off a different but identical-species target atom?

Yes, if the projectile and the target are genuinely identical particles of the same species in the same internal state (e.g. two ground-state ⁸⁷Rb atoms). If they differ in any label — different isotope, different hyperfine state, different internal excitation — they are distinguishable and you use the ordinary $|f|^2$ with no symmetrization. Distinguishability by any quantum number switches the interference off.

How does this relate to the Ramsauer–Townsend effect from earlier in the chapter?

They are unrelated mechanisms that both involve slow scattering. Ramsauer–Townsend is single-particle potential scattering — one electron off a fixed atom — controlled by the s-wave phase shift. Identical-particle scattering is about exchange interference between two copies of one particle. An electron scattering off a different atom shows no exchange effect; two electrons scattering off each other do.

What sets whether you use f_+ or f_− for a given pair of particles?

The total two-particle state's exchange symmetry, fixed by the spin-statistics connection. Bosons (integer spin) need a totally symmetric state; fermions (half-integer) a totally antisymmetric one. The spatial amplitude's sign then depends on the spin state's symmetry: for two fermions, an antisymmetric (singlet) spin state forces symmetric spatial $f_+$ , a symmetric (triplet) spin state forces antisymmetric spatial $f_-$ . The dictionary of which particles are which is owned by Bosons and Fermions.