Bloch–Nordsieck and KLN Cancellation

Individual real-emission and virtual contributions can diverge even though a measurable probability is finite. The cancellation occurs only after the probability sums over states that the experiment treats as degenerate. Bloch–Nordsieck cancellation supplies the classic massive-QED soft example; the Kinoshita–Lee–Nauenberg statement is broader and makes the required sums over degenerate initial as well as final states explicit.

Required background. Soft and Collinear Singularities identifies the unresolved regions. The Optical Theorem and Cut Interpretation relates virtual discontinuities to sums over on-shell intermediate states.

A regulated soft cancellation

Consider a massive charged process, and let $\omega_{\mathrm{IR}}$ be an auxiliary soft-energy regulator. Suppose the leading unresolved virtual correction relative to the Born rate contains

$$\delta_V=-A\int_{\omega_{\mathrm{IR}}}^Q\frac{\mathrm d\omega}{\omega},$$

where $A>0$ includes the angular integral and charge correlations. If the detector does not distinguish any real photon below energy $\Delta E$, the corresponding real correction is

$$\delta_R=A\int_{\omega_{\mathrm{IR}}}^{\Delta E}\frac{\mathrm d\omega}{\omega}.$$

Their sum is

$$\delta_V+\delta_R=-A\ln\frac{Q}{\Delta E}.$$

The unphysical $\omega_{\mathrm{IR}}$ cancels, but the physical resolution $\Delta E$ does not. This simple integral captures the essential logic: the virtual loop samples unresolved momenta, so the rate must include the real states that are experimentally indistinguishable from the nominal final state. Weinberg derives this cancellation with the full eikonal charge correlations in Weinberg 1995, §§ 13.2–13.3, printed pp. 539–548.

The exclusive probability for emitting exactly no soft photons tends to zero as the regulator is removed. Schematically, exponentiating independent leading emissions gives

$$P_0(\omega_{\mathrm{IR}})\sim \exp\!\left[-A\ln\frac{Q}{\omega_{\mathrm{IR}}}\right] \longrightarrow0.$$

That is not a failure of probability conservation. Probability has migrated into states containing arbitrarily soft quanta, none of which a finite- resolution detector can separately count.

Bloch–Nordsieck versus KLN

The two names should not be used as synonyms for an automatic pole cancellation.

Bloch–Nordsieck. For massive QED, fixing the incoming charged particles and summing over unresolved final soft photons cancels the soft divergence in inclusive probabilities. The particle mass prevents a genuine collinear divergence. Bloch and Nordsieck established the multiple-soft-photon resolution of the apparent catastrophe in Bloch and Nordsieck 1937, pp. 54–59.

Kinoshita–Lee–Nauenberg. Order by order in perturbation theory, mass singularities cancel after summing transition probabilities over complete sets of degenerate initial and final states, subject to the theorem’s unitarity and degeneracy hypotheses. The need for an initial-state sum is important: in a prepared beam one often does not average over every collinear degeneracy, so the remaining universal singularity is instead factorized into an initial-state distribution or retained as a physical mass logarithm. See Kinoshita 1962, §§ 2–4, pp. 652–668 and Lee and Nauenberg 1964, §§ II–IV, pp. B1551–B1560.

Unitarity underlies the matching between real cuts and virtual absorptive parts, but the optical theorem alone does not specify what a detector calls degenerate. That information belongs to the observable.

The measurement-function test

A weighted cross section can be written schematically as

$$\begin{aligned} \sigma[F] ={}&\frac{1}{\mathcal F}\sum_n\frac{1}{S_n} \int \mathrm d\Phi_n\,\overline{|\mathcal M_n|^2}\\ &\times F_n(p_1,\ldots,p_n). \end{aligned}$$

Here $\mathcal F=4\sqrt{(p_1\!\cdot p_2)^2-m_1^2m_2^2}$ is the invariant incoming flux, $S_n$ removes identical-state overcounting, and the bar denotes the declared initial averages and final sums. These normalization factors do not drive infrared cancellation, but retaining them keeps $\sigma[F]$ an actual cross section rather than an unnormalized transition weight.

For a massless final-state observable to be infrared safe, its measurement functions must satisfy the limiting relations. Writing the resolved hard momenta compactly as $\{p\}_n$,

$$\lim_{k\to0}F_{n+1}(\{p\}_n,k)=F_n(\{p\}_n),$$

and

$$\begin{aligned} &\lim_{a\parallel b} F_{n+1}(\ldots,zp,(1-z)p,\ldots)\\ &\qquad=F_n(\ldots,p,\ldots). \end{aligned}$$

These conditions make the real-emission weight approach the virtual configuration’s weight point by point in each unresolved limit. They are necessary local tests, not an all-orders proof: overlapping limits, initial-state factorization, phase-space boundaries, and non-global restrictions may require further analysis. Schwartz works through the real–virtual structure and infrared-safe observables in Schwartz 2014, §§ 20.1–20.3, printed pp. 356–373.

A veto that forbids every photon, or an observable that changes by order one when a particle splits into two parallel particles, fails these tests. A finite detector threshold can still make a non-infrared-safe observable numerically finite, but its prediction then carries regulator-, mass-, or fragmentation-sensitive logarithms rather than a universal massless limit.

What remains after the poles cancel

Cancellation removes auxiliary infrared poles; it does not guarantee a small perturbative correction. The finite logarithm $\ln(Q/\Delta E)$ can be large and require resummation. Nor does KLN remove long-distance hadronization, nonperturbative initial-state structure, or confinement. Local real–virtual subtraction is a computational method for implementing the cancellation and is developed separately from the theorem.

Likewise, “inclusive” must name an actual sum or measurement. Summing over soft radiation in only part of angle space, tagging one member of a collinear pair, or fixing an initial state that participates in a degeneracy changes the hypotheses and can leave logarithms or poles.

Check your understanding

Apply the two measurement-function limits to an event shape that sums energy flow and to the bare multiplicity of massless particles. Explain why the first can be unchanged by an unresolved splitting while the second jumps by one.

References

• Bloch, Felix, and Arnold Nordsieck. “Note on the Radiation Field of the Electron.” Physical Review 52 (1937): 54–59. doi:10.1103/PhysRev.52.54.
• Kinoshita, Toichiro. “Mass Singularities of Feynman Amplitudes.” Journal of Mathematical Physics 3 (1962): 650–677. doi:10.1063/1.1724268.
• Lee, Tsung-Dao, and Michael Nauenberg. “Degenerate Systems and Mass Singularities.” Physical Review 133 (1964): B1549–B1562. doi:10.1103/PhysRev.133.B1549.
• Schwartz, Matthew D. Quantum Field Theory and the Standard Model. Cambridge University Press, 2014, §§ 20.1–20.3, printed pp. 356–373. doi:10.1017/9781139540940.
• Weinberg, Steven. The Quantum Theory of Fields, Volume I: Foundations. Cambridge University Press, 1995, §§ 13.2–13.3, printed pp. 539–548. doi:10.1017/CBO9781139644167.