research30 min studyLesson 674 in Physics

Dimensional loop integration

Derive a Euclidean master integral and expand its ultraviolet pole with an explicit scale and subtraction convention.

Before you begin

Feynman parameter combination of denominators

On this page

The question for today

What you’ll understand

Derive the Gamma-function master integral using Schwinger parameters.
Expand a logarithmically divergent integral in d = 4 − 2ε.
Distinguish a dimensionally regularized zero from separate ultraviolet and infrared finiteness.

Dimensional regularization defines loop integrals as analytic functions of a dimension parameter and continues them to the physical dimension. Its power lies in preserving many useful symmetries while isolating divergent behavior as poles. It is not a claim that a particular real integral literally converges in four dimensions, nor that every pole has an ultraviolet origin. We work in a positive-mass Euclidean domain first, where a Schwinger-parameter derivation makes every normalization and energy dimension visible.

Build a master integral in its convergence domain

Let $d=4-2\epsilon$ and include one factor $\mu^{2\epsilon}$ per loop in the scalar integral defined here. The scale mu has energy units. For a denominator power a and positive mass-squared combination Delta, introduce its Schwinger representation. The Gaussian momentum integral supplies $(4\pi s)^{-d/2}$ with the measure $d^d\ell_E/(2\pi)^d$ . Integrating the positive parameter s produces a Gamma function and a power of Delta. No Minkowski factor of i is hidden in this Euclidean definition.

The direct derivation converges at large momentum when the real part of a exceeds half the real part of d, with positive Delta controlling the infrared region. The resulting expression can then be analytically continued beyond that domain. A pole encountered in that continuation records the regulated divergence. The arbitrary scale compensates the change of dimension so that the logarithmic a equal to two integral is dimensionless near four dimensions. Other authors absorb constants into a modified scale, which changes the displayed finite terms but not a matched observable.

J_a(\Delta)=\mu^{2\epsilon}\int\frac{d^d\ell_E}{(2\pi)^d}\frac1{(\ell_E^2+\Delta)^a}=\frac{\mu^{2\epsilon}}{(4\pi)^{d/2}}\frac{\Gamma(a-d/2)}{\Gamma(a)}\Delta^{d/2-a}

Delta has energy-squared units, mu has energy units, and the stated measure is Euclidean. With d = 4−2 epsilon, J_a has energy dimension 4−2a after including the scale factor.↔ Scroll wide equations horizontally.

Expand the logarithmic case consistently

For a equal to two, the Gamma function becomes Gamma of epsilon. Expand that function, the power of mu, the power of Delta, and the factor of four pi together. A finite logarithm is produced by a small power multiplying the pole. Keeping only the Gamma pole while replacing all other dimension-dependent factors by their four-dimensional values would lose necessary finite terms. The logarithm must involve a dimensionless ratio of energy-squared scales.

Define the modified minimal-subtraction pole to include the Euler constant and logarithm of four pi. Subtracting that combination gives a finite term proportional to minus the logarithm of Delta over mu squared. Ordinary minimal subtraction removes only the pure epsilon pole in this scale convention, so it retains different constants. A finite counterterm or a finite redefinition of a coupling relates such schemes. A single subtracted master integral is scheme dependent and should not be called an observable by itself.

J_2(\Delta)=\frac1{16\pi^2}\left[\frac1{\bar\epsilon}-\ln\frac\Delta{\mu^2}\right]+O(\epsilon),\qquad \frac1{\bar\epsilon}=\frac1\epsilon-\gamma_E+\ln(4\pi)

This is a positive-Delta Euclidean integral continued from its convergence domain. The pole is ultraviolet because Delta > 0 regulates the infrared. gamma_E is Euler’s constant; the scale convention is the one in J_a above.↔ Scroll wide equations horizontally.

Continue analytically without erasing infrared information

After a Minkowski loop is safely Wick rotated, its overall i factors must be restored separately according to propagator powers and the chosen amplitude definition. For timelike invariants, Delta acquires the inherited negative infinitesimal inside logarithms and powers. Replacing a negative argument by its magnitude loses the absorptive part. Analytic continuation in dimension regularizes integration, whereas continuation in an external invariant selects a physical sheet; these are related operations with distinct roles.

Scaleless integrals vanish in dimensional regularization by analytic continuation, but that vanishing can combine ultraviolet and infrared singular contributions. A massless zero-momentum bubble is a standard warning: its zero does not prove that each end of momentum integration is finite. Keep the origins of poles labeled when combining virtual and real contributions. The heat-mode instrument can illustrate exponential suppression as a mathematical analogue of a Schwinger factor, but it neither integrates over momentum nor continues dimension or external energy.

\frac1{(\ell_E^2+\Delta)^a}=\frac1{\Gamma(a)}\int_0^\infty ds\,s^{a-1}e^{-s(\ell_E^2+\Delta)}

The Schwinger parameter s has inverse-energy-squared units. This direct representation assumes positive real part of a and a denominator in a convergence sector; causal Minkowski continuation must be specified separately.↔ Scroll wide equations horizontally.

Work it through

A finite change of mass scale

In the stated modified minimal-subtraction convention, compare $J_{2,R}(4\mu^2)$ with $J_{2,R}(\mu^2)$ . Give both values and explain why the difference is ultraviolet finite.

At $\Delta=\mu^2$ , the finite logarithm vanishes, so $J_{2,R}(\mu^2)=0$ in this particular subtraction convention. The unsubtracted integral still contains its ultraviolet pole.
At $\Delta=4\mu^2$ , the finite result is $-\ln4/(16\pi^2)\simeq-0.008779$ . A negative subtracted value does not contradict positivity of the regulated original Euclidean integrand.
Subtracting the two unsubtracted expressions cancels their identical pole and common subtraction constants. Their difference is the same finite logarithm, since positive mass scales keep infrared behavior regulated.

The result

The two subtracted values are zero and −ln 4/(16π²); their difference is finite even though each original integral is ultraviolet divergent.

The living notebook

Explore the model

Interactive lab

Interactive lab · heat

An exponential factor without loop integration

Reset and set mode $n=1$ , diffusivity $\kappa=1$ ; after one Step predict amplitude $e^{-0.5}$ . Repeat from Reset with both controls $n=1$ , $\kappa=2$ ; after one Step predict $e^{-1}$ . Renaming the decay rate gives the algebraic form of a positive Schwinger exponential. This periodic heat mode does not perform a loop integral, change dimension, or choose an analytic branch.

Current temperature modeInitial sineCurrent amplitude

Current amplitude: 1
Spatial L² norm: 1.772
Decay rate · κn²: 3.6 /s

Spatial mode n3Diffusivity κ0.4

t = 0 s

The exact periodic heat-equation mode is u(x,t) = exp(−κn²t) sin(nx), on [0,2π]. Playback advances physical time. The initial sine is shown separately, and the unnormalized L² norm is √π exp(−κn²t). This analytic instrument does not run a finite-difference solver or test discretization stability. Parameter changes pause playback and recompute the model at the current elapsed time. Reset restores both parameters and returns time to zero. Playback starts paused. Adjust either parameter, then choose Play or Step.

Look forExponential suppression makes a parameter representation useful, while the integration measure and continuation determine its full value.

Pause & reason

Check your understanding

Interactive checks need JavaScript. You can still read the explanation.

Read the checkpoint explanation

The mass-squared parameter keeps the denominator nonzero in the infrared. Power counting then identifies the logarithmic divergence at large Euclidean momentum.

Your turn

Try it yourself

Use the master formula at $d=4$ , $a=3$ , and $\Delta=2E_*^2$ . Find $J_3$ and state whether an ultraviolet pole occurs.

Take a moment to make a prediction or work it out on paper. Open a hint when you need one.

Give me a hint

The relevant Gamma factors are Γ(1) and Γ(3); retain their ratio and the inverse power of Delta.

Compare with the solution

The finite integral is $J_3=\Gamma(1)\Delta^{-1}/[(4\pi)^2\Gamma(3)]=1/(32\pi^2\Delta)=1/(64\pi^2E_*^2)$ . It has inverse-energy-squared units. Large-momentum power counting is convergent, and positive Delta also regulates small momentum, so no ultraviolet or infrared pole occurs in this example. The absence of a pole follows from these conditions rather than from a subtraction convention.

Keep the idea

Take it with you

The loop measure and scale convention fix Gamma functions, constants, and dimensions.
Small dimension-dependent powers multiply poles to produce finite logarithms.
A scaleless zero or a canceled pole must not be mistaken for separate ultraviolet and infrared finiteness.

Read further

Original sources and deeper explanations to follow at your own pace.

Gudrun Heinrich, Introduction to Loop Calculations, §§1.1–1.4: dimensional integration, Gamma functions, and singularities

Optional progress, saved in this browser when available. Revisit any lesson at any time.

Lesson not marked complete.