Born Approximation

Green's Function for Scattering ended on a loaded object: the exact formula $f=-\frac{1}{4\pi}\int e^{-i\mathbf k'\cdot\mathbf r'}U\psi\,\mathrm d^3r'$ , "the launch pad for the Born approximation, which evaluates it." This article fires the engine. One substitution — the unknown $\psi$ inside the potential replaced by the incident plane wave — and the amplitude collapses into the Fourier transform of the potential at the momentum transfer $\mathbf q=\mathbf k'-\mathbf k$ . A scattering experiment becomes a microscope: the angular distribution reads $|\tilde V(q)|^2$ component by component, the scattering angle its magnification dial. We earn that claim honestly. We derive the Born amplitude twice — from the integral equation, and from Fermi's golden rule, cashing the matrix element Differential and Total Cross Section deliberately left abstract — and prove the routes identical. We derive the validity criterion from the distortion the wave actually suffers, then put it on trial by machine against exact cross sections: the error tracks the criterion, dips once in accidental agreement, and collapses catastrophically once the well binds states. And we prove an uncomfortable structural fact: the first-order amplitude is real — zero of the forward imaginary part probability conservation demands — a deficit repaired order by order in Born Series and Higher-Order Corrections and resolved exactly in Optical Theorem. The screened-Coulomb showpiece, Rutherford included, is Scattering by a Screened Coulomb Potential's.

Key takeaways

Insert $\psi\to e^{i\mathbf k\cdot\mathbf r}$ into the exact amplitude integral of Green's Function for Scattering: $f_B(\mathbf q)=-\frac{m}{2\pi\hbar^2}\tilde V(\mathbf q)$ , $q=2k\sin(\theta/2)$ — scattering measures the potential's Fourier content.
The same formula drops out of the golden-rule dictionary of Differential and Total Cross Section once the abstract matrix element is evaluated for plane waves — a two-route consistency we verify symbolically.
Central potentials: $f_B(q)=-\frac{2m}{\hbar^2q}\int_0^\infty rV(r)\sin(qr)\,\mathrm dr$ ; square well and Gaussian solved in closed form; forward amplitude $=$ the volume integral of $V$ .
Validity, derived not decreed: distortion at the origin small ⟺ $m|V_0|a^2/\hbar^2\ll1$ (low energy), $m|V_0|a/\hbar^2k\ll1$ (high energy) — machine-tested: the error tracks the criterion until bound states appear, then explodes (factor 31 at $C=8$ ).
$f_B$ is real for a real central potential, so $\mathrm{Im}\,f_B(0)=0$ while $\sigma_B>0$ : first order cannot balance probability. The repair is second order's job; the exact statement is the optical theorem's.
Key idea: replace the exact wave function inside the potential by the unperturbed incident plane wave; the scattering amplitude then becomes the Fourier transform of the potential evaluated at the momentum transfer $\hbar\mathbf q=\hbar(\mathbf k'-\mathbf k)$ .

Where this lives today

Every technique that reports a "form factor" runs this article's dictionary in production. Cryo-EM image formation models the specimen as a weak phase object — exactly the $C\ll1$ regime — and the 2017 chemistry Nobel sits on that linearity. X-ray crystallography's kinematic approximation is Born on a periodic potential (Bragg peaks are $\tilde V$ sampled on the reciprocal lattice); "dynamical diffraction" corrections are precisely the multiple-scattering terms Born drops. Elastic electron-nucleus data are published as $|F(q)|^2$ — measured over point-target cross section — because Born cleanly strips kinematics from structure; proton-radius extractions start from this factorization. And plasma and warm-dense-matter codes drive transport coefficients with screened-Coulomb Born cross sections, switching to full phase-shift machinery exactly where Derivation C's audit says they must.

The physical picture

Illuminate the potential with the plane wave and let every volume element answer: each cell re-radiates a spherical wavelet with strength $\propto V(\mathbf r)$ — what the Lippmann–Schwinger kernel says, term by term. In the single-scattering picture wavelets never talk to each other; they add downstream with inherited phases. A wavelet born at $\mathbf r$ carries excitation phase $e^{i\mathbf k\cdot\mathbf r}$ and path phase $e^{-i\mathbf k'\cdot\mathbf r}$ ; the product is $e^{-i\mathbf q\cdot\mathbf r}$ . Summing wavelets weighted by $V$ with exactly this phase is a Fourier transform: first-order quantum scattering is Fraunhofer diffraction with the potential as the aperture.

Figure

Single scattering as Fraunhofer diffraction: every volume element re-radiates, and the phases assemble the Fourier transform of V at the momentum transfer

The momentum transfer is the resolution dial. Small angles mean small $q$ : the phase barely turns across the target, wavelets add coherently, and you learn only the potential's volume integral. Large angles mean $q$ up to $2k$ : only structure on the scale $1/q$ survives the winding phase. That is why a potential of range $a$ scatters almost everything into a cone $\theta\lesssim1/(ka)$ at high energy — beyond $qa\sim1$ the aperture has no Fourier content left — and why cross sections plotted against $q$ rather than $\theta$ superpose across beam energies. When does the cartoon lie? When the wave inside the target is not the plane wave we assumed — potential too strong, or passage too slow. The criterion, derived below and put on trial numerically, compares the accumulated distortion at the worst point (the center) against the undisturbed amplitude. One flaw the cartoon hides entirely: wavelets excited by a real potential sum to a real forward amplitude, and probability bookkeeping cannot live with that — a beam's shadow requires an imaginary forward part. First order cannot cast a shadow. We quantify the deficit at the end and hand its repair to the next article.

The inherited machinery, and what this article owns

Two serials feed this one. Green's Function for Scattering derived, exactly,

f(\mathbf k',\mathbf k)=-\frac{1}{4\pi}\int e^{-i\mathbf k'\cdot\mathbf r'}\,U(\mathbf r')\,\psi(\mathbf r')\,\mathrm d^3r', \qquad U=\frac{2m}{\hbar^2}V \qquad (1)

where: $\psi$ is the exact scattering state — a launch pad, not an answer, the unknown still under the integral. And Lippmann-Schwinger Equation, packaging all scattering into $\hat T=\hat V+\hat V\hat G_0^{(+)}\hat T$ , noted the perturbative reading: "keeping $\hat T\approx\hat V$ is the Born approximation, evaluated in its own article." This is that article. It owns the evaluation: the plane-wave insertion, the Fourier reading, the closed central-potential forms, the golden-rule consistency, the validity criterion with its machine trial, the reality deficit. What it does not own it names: screened-Coulomb/Rutherford → Scattering by a Screened Coulomb Potential; the iteration and its convergence → Born Series and Higher-Order Corrections; exact forward consistency → Optical Theorem.

Derivation AThe plane-wave insertion: scattering as a Fourier transform

The goal is the first Born amplitude, extracted from the exact integral (1) by the single approximation that defines this article, then reduced to one radial integral for central potentials and evaluated in closed form for the square well and the Gaussian. The strategy is deliberate minimalism: approximate only the wave function inside the potential — the one place we cannot see — and track what the substitution discards, because that piece is the next article's entire subject. Watch the momentum transfer emerge as the only variable the answer knows.

Step 1. Inside the integral (1), replace the exact state by its zeroth iterate, the incident plane wave: $\psi(\mathbf r')\to e^{i\mathbf k\cdot\mathbf r'}$ . This is legitimate when the true state barely differs from the beam within the range of $V$ — the quantitative criterion is Derivation C's. The two exponentials merge:

e^{-i\mathbf k'\cdot\mathbf r'}e^{i\mathbf k\cdot\mathbf r'}=e^{-i(\mathbf k'-\mathbf k)\cdot\mathbf r'}\equiv e^{-i\mathbf q\cdot\mathbf r'}

Why this variable: $\hbar\mathbf q$ is the momentum the particle gained — Scattering Amplitude's kinematic dictionary, whose identity $q=2k\sin(\theta/2)$ (elastic) we reuse as is. The sign of $\mathbf q$ is a convention; for the real, even integrands below it never surfaces.

Step 2. The unknown is gone, and the integral is a pure functional of the potential:

f_B(\mathbf q)=-\frac{m}{2\pi\hbar^2}\int e^{-i\mathbf q\cdot\mathbf r}\,V(\mathbf r)\,\mathrm d^3r=-\frac{m}{2\pi\hbar^2}\,\tilde V(\mathbf q)

Why the prefactor is safe: $-\frac{1}{4\pi}\cdot\frac{2m}{\hbar^2}=-\frac{m}{2\pi\hbar^2}$ — pinned symbolically in the verify script against (1)'s normalization, so no factor of $2\pi$ has drifted between serials.

Step 3. Central potentials collapse further. Align the polar axis with $\mathbf q$ ; the azimuth gives $2\pi$ and the polar integral is the angular shell average $\int_0^\pi e^{-iqr\cos\alpha}\sin\alpha\,\mathrm d\alpha=2\sin(qr)/qr$ (verified symbolically). Then

f_B(q)=-\frac{2m}{\hbar^2 q}\int_0^\infty r\,V(r)\,\sin(qr)\,\mathrm dr

— one radial quadrature, with all angular dependence hiding inside $q(\theta)$ .

Step 4. Two closed forms. Attractive square well, $V=-V_0$ for $r<a$ :

f_B(q)=\frac{2mV_0}{\hbar^2}\,\frac{\sin qa-qa\cos qa}{q^3}\;\xrightarrow{\;q\to0\;}\;\frac{2mV_0a^3}{3\hbar^2}

Gaussian well, $V=-V_0e^{-r^2/a^2}$ : $\tilde V=-V_0\pi^{3/2}a^3e^{-q^2a^2/4}$ — a pure Gaussian in $q$ , the cleanest forward-cone laboratory. Both forward limits are $-\frac{m}{2\pi\hbar^2}\int V\,\mathrm d^3r$ : the potential's volume integral, manifestly real — remember that. Verified: the angular shell integral, the prefactor match to Green's Function for Scattering, the square-well and Gaussian transforms, both forward limits, and $q^2=4k^2\sin^2(\theta/2)$ are all symbolic checks in this article's verify script (33 checks).

Reading the result is half its value:

you measure	Born reads it as	so you learn
$\mathrm d\sigma/\mathrm d\Omega$ at $\theta$	$\left(\frac{m}{2\pi\hbar^2}\right)^2\lvert\tilde V(q)\rvert^2$	one Fourier component of $V$ per angle
forward peak height	volume integral of $V$ , squared	the potential's total strength
angular width $\theta_{1/2}\simeq1.18/ka$	Fourier support $qa\lesssim1$	the range $a$ of the potential
large- $q$ falloff	short-distance structure of $V$	edges, cores, singularities

This dictionary, with $V$ read as a charge density, is the form factor language of electron and X-ray scattering — small- $q$ expansions deliver mean-square radii; nuclear charge distributions were first mapped this way.

The same amplitude from the golden rule

Differential and Total Cross Section built the golden-rule counting dictionary with the box volume cancelling, then stopped cold, on purpose: "evaluating it for plane waves is the Born approximation's job, next door." Constant Perturbation Switched On made the same promise from the time-dependent side — potential scattering is "at bottom, a constant-perturbation golden rule, the first Born approximation, which the scattering chapter derives by name." Both IOUs are cashed here, and the payoff is a consistency theorem: a time-independent integral equation and a time-dependent rate calculation, with different logic, must and do produce the same amplitude.

Go deeper · Derivation B · 1 minThe golden-rule route gives the identical formula

Derivation BThe golden-rule route gives the identical formula

The goal is $\mathrm d\sigma/\mathrm d\Omega=|f_B|^2$ from Fermi's golden rule alone, every box-normalization factor tracked to its cancellation — the two-route consistency anchoring Born in both halves of the course. Strategy: rate $=$ golden rule with the angular density of states, cross section $=$ rate over flux, compare with Derivation A. The only new work is the plane-wave matrix element the opener left abstract; the rest is quotation.

Step 1. In a quantization box of side $L$ , plane waves are $L^{-3/2}e^{i\mathbf k\cdot\mathbf r}$ , so the static potential's beam-to-detector matrix element is

\langle\mathbf k'|V|\mathbf k\rangle=\frac{1}{L^3}\int e^{-i\mathbf q\cdot\mathbf r}V(\mathbf r)\,\mathrm d^3r=\frac{\tilde V(\mathbf q)}{L^3}

— the same Fourier transform, in box normalization.

Step 2. Golden rule with the angular density of states $\rho(E,\Omega)=\left(\frac{L}{2\pi}\right)^3\frac{mk}{\hbar^2}\,\mathrm d\Omega$ (Density of Final States' object, aimed into a solid angle by the chapter opener):

\mathrm d\Gamma=\frac{2\pi}{\hbar}\,\frac{|\tilde V(\mathbf q)|^2}{L^6}\,\left(\frac{L}{2\pi}\right)^3\frac{mk}{\hbar^2}\,\mathrm d\Omega

Why elastic: the golden rule's energy delta was spent building $\rho$ — it enforces $|\mathbf k'|=|\mathbf k|$ , leaving $q=2k\sin(\theta/2)$ as the only kinematic variable.

Step 3. Divide by the incident flux $j=\hbar k/(mL^3)$ of one box-normalized particle:

\frac{\mathrm d\sigma}{\mathrm d\Omega}=\frac{\mathrm d\Gamma/\mathrm d\Omega}{j}=\frac{2\pi}{\hbar}\cdot\frac{|\tilde V|^2}{L^6}\cdot\frac{L^3mk}{8\pi^3\hbar^2}\cdot\frac{mL^3}{\hbar k}=\left(\frac{m}{2\pi\hbar^2}\right)^2|\tilde V(\mathbf q)|^2

The $L^6$ from the matrix element is annihilated by $L^3$ from the density of states and $L^3$ from the flux — the opener's abstract cancellation, now closed. Comparing with Derivation A: $\mathrm d\sigma/\mathrm d\Omega=|f_B(q)|^2$ , identically. Why this matters: the rate route generalizes instantly — inelastic channels, relativistic kinematics, many-body targets. Agreement at first order is the license for that generalization. Verified: the full symbolic cancellation, including both $2\pi$ towers, is check 7 of the verify script.

When the plane wave is a lie

The Born approximation stands or falls on one question: how badly does the true wave differ from the beam inside the potential? The honest way to answer is to compute the discarded piece and compare.

Go deeper · Derivation C · 1 minThe validity criterion, derived from the distortion it neglects

Derivation CThe validity criterion, derived from the distortion it neglects

The goal is the standard pair of Born validity conditions — weak potential at low energy, fast passage at high energy — derived from the first-order distortion at the center of the potential, with closed-form limits for the square well. The strategy: the quantity Derivation A threw away is $\psi^{(1)}=\int G_0\,U\,e^{i\mathbf k\cdot\mathbf r'}\mathrm d^3r'$ ; self-consistency demands $|\psi^{(1)}|\ll1$ at the worst point, the origin. The subtlety worth watching: the incident wave must be angle-averaged honestly — the lazy bound $e^{i\mathbf k\cdot\mathbf r}\to e^{ikr}$ gets the scalings but hides where the oscillation suppression comes from.

Step 1. At the origin, with the kernel of Green's Function for Scattering,

\psi^{(1)}(0)=-\frac{2m}{4\pi\hbar^2}\int\frac{e^{ikr'}}{r'}\,V(r')\,e^{i\mathbf k\cdot\mathbf r'}\,\mathrm d^3r'

Step 2. For central $V$ , the angular average of the incident plane wave is $\sin(kr')/kr'$ (the same shell integral as Derivation A, at $q=k$ ). The distortion becomes a single radial integral:

C(k)\equiv|\psi^{(1)}(0)|=\frac{2m}{\hbar^2k}\left|\int_0^\infty e^{ikr}\sin(kr)\,V(r)\,\mathrm dr\right|

Why the combination $e^{ikr}\sin(kr)$ : one factor is the outgoing kernel, one is the averaged beam — their product $\frac{e^{2ikr}-1}{2i}$ is what will oscillate at high energy and saturate at low energy, generating both regimes from one expression.

Step 3. Low energy ( $ka\ll1$ ): the product $\to kr$ , so $C\to\frac{2m}{\hbar^2}\int_0^\infty r|V|\,\mathrm dr$ ; for depth $V_0$ , range $a$ : $\boxed{C_{\rm low}=mV_0a^2/\hbar^2\ll1}$ (exact square-well limit, verified; order-unity prefactors are conventional). A well Born-treatable at low energy is comparable to one too shallow to bind.

Step 4. High energy ( $ka\gg1$ ): the square-well closed form $\int_0^ae^{ikr}\sin(kr)\,\mathrm dr=\frac{ia}{2}+\frac{1-e^{2ika}}{4k}$ (verified) has modulus $\to a/2$ , so $\boxed{C_{\rm high}=mV_0a/(\hbar^2k)=V_0a/(\hbar v)\ll1}$ : the criterion becomes the phase imprinted during the crossing — a strong well turns Born-treatable if crossed fast enough. This $1/k$ rescue is why the method owns the high-energy world. Verified: both limits symbolic; the closed form exact; and the machine trial below (checks 11–16) measures how the true error tracks $C$ .

Criteria are cheap; trials are not. We computed exact cross sections for the attractive square well — partial-wave machinery run as arithmetic; its anatomy belongs to Phase Shifts — and compared Born across a strength ladder:

Figure

Punchline — the criterion on trial: Born error vs the derived criterion C, with the accidental-agreement dip and the bound-state catastrophe measured

Three lessons, all measured. The error is of order $C$ while $C\lesssim1$ — at $ka=10$ the weak well errs by $3\times10^{-4}$ against a criterion of 0.05; the $C=1$ well errs by 22%. The approach is not monotone: our $C=4$ well beats the $C=1$ well, an accidental cancellation that should cure anyone of treating $C$ as an error bar. And once the well holds bound states the failure is not gradual — the factor-31 collapse at $C=8$ is the wave rattling inside the well, scattering many times, physics the single-scattering ansatz cannot see. Born Series and Higher-Order Corrections turns exactly this collapse into a convergence theorem.

The amplitude that cannot cast a shadow

For a real central potential $\tilde V(q)$ is real at every $q$ , so the first Born amplitude satisfies

\mathrm{Im}\,f_B(0)=0 \qquad\text{while}\qquad \sigma_B=\int|f_B|^2\,\mathrm d\Omega>0 \qquad (2)

where: the tension is with probability conservation — Scattering Amplitude showed the beam-ripple interference survives averaging only forward, its value tied to $\sigma_{\rm tot}$ (the theorem is Optical Theorem's: $\sigma_{\rm tot}=\frac{4\pi}{k}\mathrm{Im}f(0)$ ). A real forward amplitude removes no flux: first order scatters particles without dimming the beam that carries them. For our $ka=1$ , $C=1$ well the ledger demands $\mathrm{Im}\,f(0)=k\sigma_B/4\pi=0.30$ (well-radius units); Born supplies zero. This is order bookkeeping, not scandal. $\sigma_B$ is second order in $V$ , so the balancing imaginary part must be second order too — it lives in the second Born term, generated by the $i\epsilon$ of the free Green's function, and Born Series and Higher-Order Corrections proves $\mathrm{Im}\,f^{(2)}(0)$ delivers precisely $\frac{k}{4\pi}\sigma_B^{(1)}$ . Truncate consistently and no paradox survives.

Limiting cases

case	result	meaning
forward, $q\to0$	$f_B(0)=-\frac{m}{2\pi\hbar^2}\int V\,\mathrm d^3r$	volume integral; real — Eq. (2)'s deficit
high energy, fixed $\theta\neq0$	$f_B\to$ tail of $\tilde V$	fixed-angle cross sections collapse; data superpose in $q$
$ka\gg1$	cone $\theta\lesssim1.18/(ka)$	Fourier support $qa\lesssim1$ (measured: checks 18–19)
$ka\ll1$	$f_B\to$ constant, isotropic	one number survives — the world of Low-Energy Scattering and the Scattering Length
criterion $C\to0$	error $\to0$ proportionally	measured at $ka{=}10$ : err $3{\times}10^{-4}$ at $C{=}0.05$
well binds states	error catastrophic (×31.6 measured)	multiple scattering; the convergence story of Born Series and Higher-Order Corrections

Worked examples

Example 1 — a 10 MeV neutron on a nuclear well: run the criterion before trusting the answer. An attractive square well, $V_0=40$ MeV, $a=3$ fm — a caricature of a light nucleus. Compute the Born observables, then audit.

Step 1. $k=\sqrt{2mE}/\hbar=\sqrt{2\cdot939.57\cdot10}/197.33=0.695\ \mathrm{fm^{-1}}$ , so $ka=2.08$ — neither limit is clean.

Step 2. $f_B(0)=\frac{2mV_0a^3}{3\hbar^2}=\frac{2\cdot939.57\cdot40\cdot27}{3\cdot(197.33)^2}=17.4$ fm, so $\mathrm d\sigma/\mathrm d\Omega(0)=3.02$ b/sr.

Step 3. From the closed form at $q=2k\sin(\theta/2)$ : 1.19 b/sr at 60°, 0.094 b/sr at 120° — forward-heavy, spanning a factor 32; $\sigma_B=9.4$ b by quadrature.

Step 4. The audit: $C_{\rm low}=8.7$ , $C_{\rm high}=4.2$ — both violated severalfold; this well binds states. Verdict: shapes, not predictions — at these strengths the punchline figure documents order-of-magnitude failure. Born always produces a number; the criterion decides whether to publish it. (Low-energy nuclear scattering belongs to Phase Shifts and Low-Energy Scattering and the Scattering Length.)

Example 2 — a 1 keV electron on a screened atomic site: Born in its element. A Gaussian well, $V_0=10$ eV, $a=1$ Å.

Step 1. $k=\sqrt{2m_ec^2E}/\hbar c=\sqrt{2\cdot511000\cdot1000}/1973.3=16.2\ \mathrm{Å^{-1}}$ , so $ka=16.2\gg1$ .

Step 2. Audit first: $C_{\rm high}=0.081\ll1$ — trustworthy. Note $C_{\rm low}=1.3$ : the low-energy form alone would fail this potential; the fast passage rescues it — Derivation C's $1/k$ at work.

Step 3. From the Gaussian transform: $f_B(0)=\frac{mV_0a^3\sqrt\pi}{2\hbar^2}=1.16$ Å; $\mathrm d\sigma/\mathrm d\Omega(0)=1.35$ Å²/sr.

Step 4. The forward cone: the distribution is $e^{-q^2a^2/2}$ , half maximum at $\theta_{1/2}=1.18/ka=0.073$ rad $=$ 4.2° — essentially everything within a few degrees of the beam. Total: $\sigma_B=f_B(0)^2\cdot2\pi/(k^2a^2)=0.032$ Å², the cone's solid angle shrinking as $1/k^2$ . Answer: $\sigma_B=0.032$ Å², delivered with its validity certificate. All numbers machine-computed (verify checks 20–33).

Origins · 1 minBorn Approximation and how this was discovered

Born Approximation

1926–1954 · Göttingen

The Born approximation was born in the same four pages as quantum probability itself. In June 1926 Max Born, treating collisions with Schrödinger's brand-new wave equation because matrix mechanics could not handle unbound states, computed the scattered wave to first order — and had to say what the amplitudes meant. The preliminary note's famous footnote, added in proof, declared probability proportional to the amplitude squared; the full July paper built the perturbative collision theory whose leading term this article derives. The scattering problem and the statistical interpretation are, historically, one calculation. Wentzel evaluated the Coulomb case that year, finding Rutherford's formula. The 1954 Nobel citation honored the footnote; working physicists honor the integral, still the first thing anyone computes about a new interaction. (Born the person appears across this course; this card is the approximation's biography, not his.)

History

1911

— Rutherford's scattering analysis gives the angular distribution the quantum theory will one day have to reproduce

1926

— Born's June collision note computes first-order scattering and, in a footnote added in proof, states the probability interpretation; the full theory follows in July

1926

— Wentzel obtains the Rutherford cross section from the first-order quantum amplitude for a Coulomb field

1932

— Feenberg's analysis of forward consistency exposes what first order omits — the thread the optical theorem completes

1951

— Jost and Pais give the convergence criterion that tells you when the series behind the approximation actually sums

1954

— Born's Nobel Prize cites the statistical interpretation born inside the collision calculation

Going further

References. Griffiths 3rd ed. §11.4 · Sakurai & Napolitano 2nd ed. §6.3 · Shankar 2nd ed. Ch. 19 · Cohen-Tannoudji, Diu & Laloë Vol. II Ch. VIII · Newton, Scattering Theory of Waves and Particles, Ch. 8.

Provenance. All formulas are standard textbook results. A verifier should confirm: (1) the prefactor $m/(2\pi\hbar^2)$ and the Fourier-transform structure of the first Born amplitude (Griffiths §11.4; Sakurai §6.3; Shankar Ch. 19); (2) the angular reduction giving $f(\theta)=-\frac{2m}{\hbar^2 q}\int_0^\infty rV\sin(qr)\,\mathrm{d}r$ with $q=2k\sin(\theta/2)$ ; (3) the Yukawa amplitude $-2m\beta/[\hbar^2(\mu^2+q^2)]$ and the Rutherford limit $(Z_1Z_2e^2/4E\sin^2(\theta/2))^2$ (Griffiths, Sakurai); (4) the validity criteria $2m|V_0|a^2/\hbar^2\ll1$ (low energy) and $m|V_0|a/(\hbar^2k)\ll1$ (high energy), Cohen-Tannoudji Ch. VIII. No numerical constants beyond standard ones are introduced. (Lane note: items (1), (2) and (4) are re-derived and machine-verified in this article — the validity pair from the distortion integral, with the exact limits confirming the quoted scalings. Item (3), the base's Yukawa/Rutherford derivation, is chapter property of Scattering by a Screened Coulomb Potential and is relocated there in full; this article evaluates no specific long-range potential. No physics error found in the base's displayed formulas.)

Builds on: Lippmann-Schwinger Equation · Green's Function for Scattering · Scattering Amplitude · Asymptotic Form of the Scattering Wave Function Leads to: Born Series and Higher-Order Corrections · Scattering by a Screened Coulomb Potential · Optical Theorem · Differential and Total Cross Section · Partial-Wave Expansion

Is the Born approximation the same as first-order perturbation theory?

Same expansion parameter, different observable. Bound-state perturbation theory corrects discrete energies; Born is first-order perturbation theory for the continuum, where the output is an amplitude. The golden-rule derivation makes the kinship exact. It is also why Born fails where bound-state physics contaminates the continuum — resonances, binding wells — where " $V$ is a small correction to free motion" is false at every order.

Why evaluate the distortion at the origin rather than averaging it over the potential?

Validity needs the replacement to hold everywhere Derivation A's integrand is nonzero, and for an attractive well the distortion peaks at the center — every point of the well radiates into it coherently. The worst point gives the most conservative simple criterion. Norm-based criteria (on $G_0V$ ) are stricter and become the Born-series convergence statement; all versions agree on the scalings, differing in order-unity prefactors — why textbooks disagree cheerfully about factors of 2 in $C_{\rm low}$ .

The criterion at high energy involves the velocity. Is Born secretly a classical short-time approximation?

The high-energy form $V_0a/\hbar v\ll1$ is a time-energy statement: the potential acts for the crossing time $a/v$ , and $V_0\,(a/v)/\hbar$ is the phase it deposits. Born holds when that phase is small; the eikonal approximation (the chapter's closer) is the complementary regime — large phase, straight-line accumulation. The output stays quantum: the forward cone $\theta\sim1/ka$ is a diffraction width no trajectory picture produces.

If scattering measures only |Ṽ(q)|², can two different potentials have identical Born cross sections?

Yes — the phase of $\tilde V(q)$ is invisible at first order, so potentials whose transforms share a modulus are Born-indistinguishable: crystallography's phase problem, in scattering dress. Beyond first order the degeneracy breaks — the second Born term correlates Fourier components pairwise, and Phase Shifts' analysis reconstructs more than $|\tilde V|$ . At first order the microscope reads intensities, not phases.

Why does the momentum transfer q, not the energy, control what the scattering resolves?

Elastically the target absorbs momentum but no energy, so the only scale imprinted on the potential is $1/q$ : the kernel averages $V$ over cells that size. Energy enters only through the reach $q\leq2k$ — a faster beam extends resolution but does not change what a given $q$ sees. Hence plotting data against $q$ : different beam energies land on one curve, which is $|\tilde V(q)|^2$ in the Born regime.

Does the Born approximation conserve probability at all?

At its own order, yes; as a truncated answer, no. The real Born amplitude contributes zero forward imaginary part — but the deficit is $O(V^2)$ , the order of the neglected second term, so nothing kept contradicts anything kept. Needing unitarity at finite order? Include $\mathrm{Im}f^{(2)}$ (next article) or repackage through phase shifts, where it is built in channel by channel (The S-Matrix and Unitarity in Scattering).

Born Approximation

Where this lives today

The physical picture

The inherited machinery, and what this article owns

The same amplitude from the golden rule

When the plane wave is a lie

The amplitude that cannot cast a shadow

Limiting cases

Worked examples

Going further

Before this

Where this leads