Modes, Factorization, and Multiscale RG

Multiscale effective field theory is needed when a low-energy problem contains several parametrically distinct momentum regions that remain dynamically coupled. Instead of representing all infrared physics by one local field with momenta of a single size, the EFT introduces homogeneous mode fields, expands their interactions consistently, subtracts their overlaps, and evolves each factor from its natural scale.

This chapter develops that construction in dependency order. It distinguishes integration regions from dynamical modes, performs the multipole and overlap expansion, matches sector operators, organizes virtuality and rapidity evolution, and then applies the framework to SCET and to unstable resonances. Observable-specific factorization theorems and precision phenomenology remain with their application volumes.

Enter through the scale hierarchy

If the immediate question is…	Start with…	Leave with…
Which momentum regions should become EFT fields?	Modes, Virtualities, and EFT Scale Separation	Homogeneous component scalings, virtualities, rapidities, and a mode-relevance test
How are interactions expanded without mixing powers or double counting?	Multipole Expansion and Homogeneous Mode Power Counting	A coordinate expansion, homogeneous operators, and an overlap or zero-bin check
What turns a hard coefficient and sector matrix elements into a factorized operator statement?	Matching onto Factorized Operator Structures	Sector building blocks, Wilson lines, measurement insertions, and stated obstructions
How are logarithms moved between natural scales and resummed?	Evolution Kernels, Consistency Relations, and Resummation Architecture	Boundary matching, path-ordered kernels, consistency sums, and a declared logarithmic accuracy
Why do equal-virtuality modes sometimes require a second scale?	Rapidity Renormalization and Two-Scale Evolution	A regulator-specific rapidity counterterm and a checked $\mu$–$\nu$ evolution path
Which SCET variant and building blocks fit a hard, collinear, and soft problem?	Soft-Collinear Effective Theory: Architecture and Validity	Light-cone fields, sector gauge symmetries, Wilson-line operators, and validity limits
How is a narrow resonance expanded without making it an external asymptotic state?	Unstable-Particle Effective Theory and the Width Expansion	Complex-pole input, resonant and nonresonant modes, width counting, and an energy-window test

The first five pages form the general method. The SCET page collects those ingredients into one named relativistic framework. The unstable-particle page uses the same logic of modes and matching but a different small parameter, $\Gamma/M$, and should not be read as a collinear specialization.

Momentum components, virtuality, and rapidity

Choose null reference vectors $n^2=\bar n^2=0$ and $n\!\cdot\bar n=2$. Any momentum decomposes as

$$p^\mu =\frac{\bar n\!\cdot p}{2}n^\mu +p_\perp^\mu +\frac{n\!\cdot p}{2}\bar n^\mu.$$

For the ordered components $(n\!\cdot p,\bar n\!\cdot p,p_\perp)$ and a hard scale $Q$, representative scalings are

Mode	Momentum scaling	Virtuality	What separates it
Hard	$Q(1,1,1)$	$Q^2$	Integrated out at hard matching
$n$-collinear	$Q(\lambda^2,1,\lambda)$	$Q^2\lambda^2$	Large $\bar n\!\cdot p$ and large rapidity
$\bar n$-collinear	$Q(1,\lambda^2,\lambda)$	$Q^2\lambda^2$	Large $n\!\cdot p$ and opposite rapidity
Soft in an equal-virtuality theory	$Q(\lambda,\lambda,\lambda)$	$Q^2\lambda^2$	Central rapidity rather than a new virtuality
Ultrasoft in a separated-virtuality theory	$Q(\lambda^2,\lambda^2,\lambda^2)$	$Q^2\lambda^4$	Lower virtuality in every component

This is a menu, not a universal field list. A SCET-I-like problem contains collinear and ultrasoft modes at different virtualities. A SCET-II-like problem contains collinear and soft modes with comparable virtuality but parametrically different rapidity. The observable, measurement, and pinch structure decide which version applies; naming both soft and ultrasoft without a scaling argument double counts or invents degrees of freedom.

Rapidity may be represented schematically by

$$y_p=\frac12\ln\left|\frac{\bar n\!\cdot p}{n\!\cdot p}\right|.$$

Virtuality RG changes $p^2$ scales and is conventionally parameterized by $\mu$. When two sectors have the same $p^2$ but widely separated $y_p$, dimensional regularization and $\mu$ evolution alone may not separate them; a rapidity regulator and scale $\nu$ can then be required.

Becher, Broggio, and Ferroglia derive the region scalings, scalar EFT, QCD power counting, Wilson-line building blocks, and RG organization using the Sudakov form factor as a common example Becher, Broggio, and Ferroglia 2015, §§ 2–5, preprint pp. 4–53, Open PDF.

From a region to a field mode

Expansion by regions is a rule for asymptotically expanding an integral. A region becomes an EFT mode only when the low-energy theory needs a field with that homogeneous scaling to reproduce long-distance propagation and interactions across the class of observables under consideration.

A useful decision sequence is:

1. Identify a physical hierarchy and the low-energy pinch surfaces or near-on-shell configurations that can produce nonanalytic dependence on small scales.
2. Assign every candidate region homogeneous momentum components and determine its measure, propagator, and interaction scaling.
3. Test whether existing EFT fields reproduce that region after their loop momenta and multipole-expanded interactions are included.
4. If a new field is required, specify its gauge transformations, operator building blocks, and overlap with every existing mode.
5. Match and renormalize the enlarged theory, then verify an amplitude or factorized quantity through the retained power.

A saddle or integration domain that contributes to one diagram is not automatically a new particle species. Conversely, two modes may represent the same microscopic field near different momentum configurations and therefore require independent EFT fields with an explicit exclusion of their common limit.

The zero-bin construction implements that exclusion. If a collinear integrand $I_c(k)$ contains a limit $I_{c\to s}(k)$ already represented by the soft sector, the subtracted contribution has the schematic form

$$I_c^{\rm sub}=I_c-I_{c\to s}.$$

The subtraction depends on the regulator and on the complete set of nested overlaps, but its physical role is fixed: each momentum configuration is counted once. Manohar and Stewart show how such subtractions reproduce the infrared structure while separating mode contributions in Manohar and Stewart 2007, §§ I and IV, preprint pp. 3–20, Open PDF.

Factorization and its consistency equations

After hard fluctuations are integrated out, a leading-power observable often has the schematic form

$$\sigma =H(Q,\mu) \otimes \left[\prod_i J_i(\mu,\nu)\right] \otimes S(\mu,\nu),$$

where the products and convolutions depend on the measurement. This expression is not a factorization theorem merely because its factors have been named. One must derive the sector operator, decouple or account for leading interactions, insert the measurement consistently, subtract overlaps, and exclude or include Glauber exchange according to the observable.

Renormalization gives a stringent check. For a multiplicatively represented quantity with no residual $\mu$ dependence,

$$\gamma_H^\mu +\sum_i\gamma_{J_i}^\mu +\gamma_S^\mu=0.$$

If rapidity renormalization is present, the physical product also requires cancellation of $\nu$ dependence, while each factor obeys compatible two-scale evolution. In differential form, path independence requires the $\mu$ and $\nu$ evolution operators to commute through the computed order. Chiu, Jain, Neill, and Rothstein introduce the rapidity RG precisely to resum logarithms associated with large rapidity ratios in Chiu et al. 2012, pp. 1–4, Open PDF.

Canonical scales minimize logarithms in each boundary function: a hard scale for $H$, invariant-mass scales for jets, and a soft or measurement scale for $S$. Evolution brings the factors to common scales. Matching-scale independence and anomalous-dimension consistency test the construction; they do not by themselves prove that no endpoint, Glauber, non-global, or power-suppressed coupling has been omitted.

Width as a multiscale expansion

For an unstable particle with pole mass $M$ and width $\Gamma\ll M$, the near-resonance hierarchy is

$$\delta \sim\frac{s-M^2}{M^2} \sim\frac{\Gamma}{M} \ll1.$$

Hard fluctuations have momenta of order $M$, while a resonant field carries a large mechanical momentum plus residual momentum of order $M\delta$. Its inverse propagator is of order $M^2\delta$, so self-energy effects that would be perturbative away from resonance become leading and must be incorporated through complex-pole matching. Nonresonant local operators contribute at definite orders in the same expansion.

The unstable excitation is not an external LSZ state. Production and decay are combined into amplitudes for stable external particles, and resonant plus nonresonant contributions are matched so that the result is gauge independent. Beneke, Chapovsky, Signer, and Zanderighi formulate this simultaneous expansion in the coupling and $\Gamma/M$ in Beneke et al. 2004, pp. 1–4, Open PDF.

Scope and connections

This chapter develops the translation from physical momentum hierarchies to EFT modes, the homogeneous interaction expansion, overlap removal, sector-operator matching, and multiscale evolution. Matching, Decoupling, and Threshold Evolution supplies ordinary hard matching and infrared consistency; Operator Bases and Field Redefinitions supplies quotient and translation machinery.

Scattering Amplitudes and Collider Theory develops expansion by regions as an integral technique, observable factorization theorems, Glauber and endpoint failure analysis, and precision resummed predictions. Gauge Theories and the Standard Model treats QCD processes and named particle applications. Landau–Ginzburg–Wilson Criticality treats momentum-shell mode elimination near criticality, which is conceptually related but not a light-cone factorization problem.

The chapter’s checks are therefore structural: a complete mode list for the stated hierarchy, homogeneous power counting, overlap cancellation, consistent anomalous dimensions, scale-path independence, and a declared validity window. Passing those checks establishes a controlled EFT construction; downstream observables remain outside this volume’s scope.

Review the chapter

1. Why is a loop-momentum region not automatically an EFT mode?

   Solution

   A mode must represent independent long-distance propagation and admit a consistent field-theory realization across the stated observable class. Some regions are hard contributions to matching, while others are overlaps already reproduced by the limiting expansion of retained modes.

2. When is rapidity evolution needed in addition to ordinary virtuality evolution?

   Solution

   It is needed when factorized modes have comparable invariant mass, so $\mu$ evolution cannot separate them, but parametrically different light-cone component ratios generate large rapidity logarithms.

References

• Becher, Thomas, Alessandro Broggio, and Andrea Ferroglia. 2015. Introduction to Soft-Collinear Effective Theory. Lecture Notes in Physics 896. Cham: Springer. DOI. Open PDF.

• Beneke, Martin, A. P. Chapovsky, Adrian Signer, and Giulia Zanderighi. 2004. “Effective Theory Approach to Unstable Particle Production.” Physical Review Letters 93 (1): 011602. DOI. Open PDF.

• Chiu, Jui-yu, Ambar Jain, Duff Neill, and Ira Z. Rothstein. 2012. “The Rapidity Renormalization Group.” Physical Review Letters 108 (15): 151601. DOI. Open PDF.

• Manohar, Aneesh V., and Iain W. Stewart. 2007. “The Zero-Bin and Mode Factorization in Quantum Field Theory.” Physical Review D 76 (7): 074002. DOI. Open PDF.