Matching, Decoupling, and Threshold Evolution

Matching determines how short-distance physics enters a low-energy theory. The central operation is a controlled comparison: calculate the same low-energy object in a full theory and in its EFT, with the light states, infrared prescription, field normalization, basis, and renormalization scheme aligned. Their difference fixes Wilson coefficients; renormalization-group evolution then transports those coefficients between thresholds and the scale of observation.

This chapter develops that procedure in dependency order. It begins by integrating out a heavy field, compares on-shell, off-shell, background-field, and functional matching strategies, and verifies classical elimination at tree level. It then isolates hard information beyond tree level, checks infrared cancellation, states the hypotheses behind decoupling, evolves through multiple thresholds, and ends with nondecoupling and observable-level validation.

Enter through the matching problem

If the immediate question is…	Start with…	Leave with…
What does “integrating out” a field actually do to the action?	Integrating Out Heavy Fields	A nonlocal light-field action, its local heavy-mass expansion, and the determinant contribution
Should coefficients be matched with amplitudes, Green functions, background fields, or a functional method?	Matching with Amplitudes, Green Functions, and Background Fields	A strategy choice with its gauge, field-coordinate, equation-of-motion, basis, and infrared data declared
When does the classical heavy-field equation reproduce tree matching?	Tree-Level Matching and Classical Elimination	A sign-checked local action and an amplitude residual through the first omitted power
How is a loop-level hard coefficient separated from EFT loops?	Matching Conditions Beyond Tree Level	A renormalized full-minus-EFT coefficient in a stated scheme and basis
Why may infrared poles appear in both calculations but not in the Wilson coefficient?	Infrared Cancellation, Regulators, and Matching Consistency	An infrared record, including the ultraviolet and infrared content hidden by scaleless integrals
Under what assumptions does a heavy particle decouple?	Decoupling Theorems and Threshold Corrections	A theorem statement with its momentum, mass, coupling, operator, and scheme hypotheses
How should several separated heavy scales be crossed?	Running and Matching across Multiple Thresholds	A scale-ordered chain of evolution and threshold maps with a path-independence check
What can survive a formal heavy-mass limit?	Nondecoupling Effects and Matching Validation	A classification of symmetry-breaking, anomalous, logarithmic, and power-suppressed remnants tested in observables

The natural linear route is the order shown. A reader who needs only tree coefficients can stop after classical elimination. Loop matching requires the strategy and infrared-consistency pages even when a computer package performs the integrals: automation does not choose matching states, define redundant directions, or prove that a pole is ultraviolet rather than infrared.

The full-minus-effective statement

Let $M$ denote a heavy scale and let $p_i,m\ll M$ be external momenta and light masses. Choose a renormalized matching object $\Gamma$, expand the full theory in the low-energy variables, and require

$$\Gamma_{\mathrm{full}}^R (p_i,m,M;\mu_m) = \Gamma_{\mathrm{EFT}}^R (p_i,m;C(\mu_m),\mu_m) +O\!\left[\left(\frac{Q}{M}\right)^{N+1}\right]$$

through the target order $N$. The matching scale $\mu_m$ is normally chosen near $M$ to avoid large logarithms in the boundary coefficients. It is not a physical threshold by itself, and a correctly truncated prediction is insensitive to its variation up to higher-order terms.

At one loop, schematically,

$$\sum_i C_i^{(1)}(\mu_m) \langle\mathcal O_i\rangle^{(0)} = \Gamma_{\mathrm{full}}^{(1),R} -\Gamma_{\mathrm{EFT}}^{(1),R} \big|_{C^{(0)}}.$$

Both terms on the right can contain light-particle logarithms, infrared poles, gauge-dependent off-shell pieces, and scheme-dependent finite parts. Those structures cancel or translate only when the two sides use compatible conventions. What remains is local in the low-energy variables and belongs in the Wilson coefficients. Georgi emphasizes this separation between short-distance coefficient data and long-distance matrix elements in Georgi 1993, §§ 2–4, pp. 212–220.

Matching is not the act of setting a full-theory loop equal to a coefficient. EFT loops built from already-known lower-order interactions are part of the comparison. Omitting them contaminates the purported hard coefficient with infrared physics that the EFT itself reproduces.

One heavy scalar, three equivalent views

The chapter continues the light-scalar/heavy-mediator model,

$$\mathcal L = \frac12(\partial\phi)^2 +\frac12(\partial H)^2 -\frac12M^2H^2 -\frac g2H\phi^2.$$

Because the heavy field appears quadratically, three descriptions can be compared exactly at tree level.

First, the classical equation

$$(\Box+M^2)H=-\frac g2\phi^2$$

gives a nonlocal light-field interaction proportional to

$$\phi^2\frac{1}{M^2+\Box}\phi^2.$$

Second, Gaussian functional integration produces the same source–propagator–source term. If the heavy quadratic operator depends on the light background, its determinant also contributes at one loop; for the strictly linear model above, the field-independent determinant cancels from normalized light correlators.

Third, expanding the heavy propagator in an on-shell light-particle amplitude yields

$$\frac{1}{M^2-q^2} = \frac1{M^2} +\frac{q^2}{M^4} +\frac{q^4}{M^6} +\cdots.$$

The local action, functional integral, and amplitude therefore encode the same coefficients once signs, symmetry factors, field normalization, and redundant operators are treated consistently. Burgess derives the corresponding relation among heavy-field elimination, the Wilson action, and the inverse-mass expansion in Burgess 2021, §§ 2.2–2.4, pp. 26–44. Effective Field Theory as a Controlled Expansion derived the associated fixed-angle residual; this chapter turns that expansion into an explicit matching calculation.

The equivalence has a domain. The derivative expansion requires the external invariants to remain inside the nearest heavy singularity, while the exact nonlocal action can be meaningful beyond the convergence radius. A determinant can carry light-background dependence and anomalies that classical elimination misses. An off-shell Green function can contain equation-of-motion operators absent from an on-shell amplitude. These are differences of matching representation, not contradictory physics.

Choosing the matching object

No single strategy is best for every coefficient.

Matching object	Main advantage	Extra information that must be controlled
On-shell amplitudes	Directly removes field-redefinition and equation-of-motion redundancy from physical matrix elements	A complete set of external states and kinematics, infrared-safe comparison, and enough observables to separate coefficient combinations
Off-shell Green functions	Projects many tensor and derivative structures with convenient momenta	Gauge choice, field coordinates, equation-of-motion and BRST-exact sectors, contact terms, and basis reduction
Background-field Green functions	Preserves manifest background-gauge covariance in suitable schemes	Quantum gauge fixing, background-field normalization, and translation to the target operator basis
Functional matching	Generates whole classes of operators and determinant terms systematically	Expansion of differential operators, measure contributions, regularization, and a transparent map to a nonredundant basis

Two correct strategies can produce different intermediate coefficients if they use different redundant bases or finite field definitions. The comparison is meaningful only after transforming both results to the same operator normalization and recomputing an invariant amplitude or observable. The next chapter, Operator Bases and Field Redefinitions, develops the systematic quotient and basis-construction problem.

Running, matching, and decoupling

Below a heavy threshold, a mass-independent subtraction scheme does not automatically remove the heavy particle from beta functions. One explicitly matches onto a theory with different active content and then runs with the lower-energy anomalous dimensions. For one threshold,

$$C_{<}(\mu_m) = \mathcal M_M(\mu_m)C_{>}(\mu_m), \qquad C_{<}(\mu_l) = U_{<}(\mu_l,\mu_m)C_{<}(\mu_m).$$

For $M_1>M_2$, the ordered chain is

$$C(\mu_l) = U_2(\mu_l,\mu_2) \mathcal M_2(\mu_2) U_1(\mu_2,\mu_1) \mathcal M_1(\mu_1) C(\mu_h).$$

Each $U$ evolves within a fixed field content; each $\mathcal M$ changes the theory at a threshold. Varying the intermediate scales tests cancellation between the logarithms in $U$ and those in $\mathcal M$. One-step matching can be equivalent at a fixed order, but it leaves large $\ln(M_1/M_2)$ terms unresummed when the hierarchy is wide.

The Appelquist–Carazzone conclusion is correspondingly conditional. At external momenta much smaller than a large mass, under suitable coupling and renormalization assumptions, heavy-particle effects can be absorbed into renormalizations of low-dimension light-theory parameters plus local operators suppressed by powers of the heavy mass Appelquist and Carazzone 1975, pp. 2856–2861. Threshold shifts in marginal couplings and logarithms are part of this low-energy redefinition, not failures of decoupling.

Naive decoupling can fail when the heavy limit also increases a coupling, when the mass is generated by symmetry breaking, when an anomaly must remain matched, or when the proposed low-energy field content omits nonanalytic infrared physics. The final page tests those mechanisms rather than classifying every finite heavy threshold effect as “nondecoupling.”

Validation before interpretation

A matching result is accepted only after the following data are aligned or translated:

• the external states or off-shell fields and kinematic projections;
• the gauge fixing, field normalization, operator normalization, and basis;
• the ultraviolet subtraction scheme and matching scale;
• the infrared regulator, light masses, overlap or zero-bin prescription, and order of limits;
• the retained loop and inverse-mass orders on both sides;
• the threshold content and running between scales; and
• an observable or invariant amplitude that closes the comparison.

A reproducible calculation should retain separate hard and infrared pole records and a retained-order scale check for the scalar fixture. The chapter pages provide the analytic cancellation against which it can be validated.

A failed check has diagnostic value. Residual infrared poles indicate mismatched light physics or regulators. Residual gauge dependence indicates an incomplete projection, basis, or observable. Large matching-scale variation indicates missing running or threshold terms. A full-theory/EFT difference that remains nonanalytic at low momentum is not a Wilson coefficient; it signals that a light region has not canceled or that an active degree of freedom was removed incorrectly.

Scope and connections

This chapter develops the comparison that fixes Wilson coefficients, the cancellation of common infrared structure, threshold matching, decoupling hypotheses, and matching validation. Continue to Scattering Amplitudes and Collider Theory for general loop-integral reduction, expansion by regions, and process amplitudes, or to Gauge Theories and the Standard Model for named-model coefficient catalogues and Standard Model threshold applications.

Effective Field Theory: Construction and Power Counting supplies the domain, active fields, operator classes, and target order before matching begins. Operator Bases and Field Redefinitions reduces the resulting local structures to a chosen basis. Modes, Factorization, and Multiscale RG takes over when several homogeneous momentum regions require distinct fields or rapidity evolution rather than a single heavy-threshold expansion.

References

• Appelquist, Thomas, and J. Carazzone. “Infrared Singularities and Massive Fields.” Physical Review D 11 (1975): 2856–2861. DOI
• Burgess, C. P. Introduction to Effective Field Theory: Thinking Effectively about Hierarchies of Scale. Cambridge: Cambridge University Press, 2021. DOI
• Georgi, Howard. “Effective Field Theory.” Annual Review of Nuclear and Particle Science 43 (1993): 209–252. DOI