Constructive Scattering, Particle Structure, and Mass Gaps

A constructed Euclidean measure does not automatically contain a stable particle or an $S$-matrix. One first reconstructs the relativistic Hilbert space, then proves a spectral gap and an isolated mass hyperboloid, and only then applies Haag–Ruelle estimates. Asymptotic completeness remains a further, usually harder, conclusion.

Required background. Constructing P(φ)₂ and φ⁴₃ models supplies the Euclidean model; Haag–Ruelle scattering-state construction supplies the asymptotic theorem; clustering, vacuum uniqueness, and mass-gap implications links Euclidean decay to spectrum.

Helpful background. Asymptotic completeness: definitions and known models separates existence of scattering states from completeness; the OS reconstruction theorem supplies the Hilbert-space bridge.

From Euclidean decay to an isolated shell

Let a limiting Schwinger hierarchy satisfy the OS hypotheses and reconstruct $(\mathcal H,U,\Omega,\phi)$. Exponential decay of connected Euclidean correlations at rate $m_\mathrm{gap}>0$ implies that the joint energy–momentum spectrum has vacuum at zero separated from the rest by a positive mass gap, under the usual regularity assumptions. A stable particle requires more: an isolated hyperboloid

$$H_m^+=\{p:p^2=m^2,\ p^0>0\}$$

in the spectrum, separated locally from multiparticle continuum, and a local operator with nonzero projection onto its one-particle subspace.

For weakly coupled massive $P(\phi)_2$, constructive cluster estimates yield a spectrum contained in $\{0,m\}\cup[m_1,\infty)$ with $m_1$ tending to $2m_0$ as the coupling vanishes. Hence the one-particle shell is isolated. This is stronger than “there is a mass gap”: a theory can be gapped while its first spectral band has no isolated one-particle pole. Summers 2016, §3.1, pp. 12–13 states the spectral separation and the resulting Haag–Ruelle application; the primary construction and particle analysis is Glimm, Jaffe, and Spencer 1974, pp. 585–632.

Haag–Ruelle construction in the worked model

Choose local operators $A_1,A_2$ whose energy–momentum transfers meet the isolated shell and smooth positive-energy Klein–Gordon wave packets $f_{i,t}$ with disjoint velocity supports. Define

$$A_i(f_{i,t})=\int d^dx,f_{i,t}(x)\,A_i(x).$$

Locality makes commutators between the two time-dependent operators small because their wave packets become spacelike separated linearly in $|t|$. Spectral isolation removes off-shell contributions, and Cook-type estimates prove

$$\Psi^{\mathrm{in/out}}=lim_{t\to\mp\infty}A_1(f_{1,t})A_2(f_{2,t})\Omega.$$

The limits define isometric wave operators on the symmetric Fock space over the one-particle subspace. Their inner products factor into one-particle inner products. The in/out maps and their relation are developed physically at in and out states.

For $P(\phi)_2$, this gives a well-defined, nontrivial scattering theory and particle production in appropriate models. It does not show that the ranges of the wave operators exhaust all of $\mathcal H$. Even two-particle asymptotic completeness, where known for special interactions, requires extra spectral analysis of bound states and thresholds.

An independent check takes zero coupling. The connected scattering amplitude must vanish, the wave operators reduce to the free identification, and $S=1$. The Haag–Ruelle normalization must also reproduce the one-particle inner product; failure indicates a missing mass-shell residue.

Adversarial threshold test

Suppose the proposed particle mass lies at or inside a continuum threshold, so no open spectral neighborhood separates $H_m^+$ from other spectrum. The time averaging used to isolate a one-particle component cannot suppress the continuum uniformly, and the Haag–Ruelle hypothesis fails. A peak in a finite-volume spectral density or exponential two-point decay does not restore isolation.

Likewise, a mass gap alone allows a lowest excitation band without a sharp particle. The strongest surviving claim is a gapped reconstructed theory. Scattering-state existence, an $S$-matrix, nontrivial scattering, and asymptotic completeness must not be inferred.

Bound states require equal care. A second isolated hyperboloid below a two-particle threshold represents another stable species and should be added to the asymptotic particle space. By contrast, a resonance pole on a continued sheet is not a normalizable one-particle vector and does not satisfy the Haag–Ruelle spectral hypothesis. The distinction can be tested directly with the spectral measure of a local interpolating operator: an isolated atom supplies a one-particle projection, whereas an absolutely continuous bump does not. This check prevents a perturbative resonance from being promoted to a constructively proved asymptotic particle.

Exercises

1. Velocity separation. Why do disjoint velocity supports make two wave packets spacelike separated at large equal times?

Solution

Their centers separate as $t(v_1-v_2)$ while each packet remains concentrated in a narrower cone around its velocity support. A positive separation between the supports therefore produces spatial separation proportional to $|t|$, exceeding bounded localization radii.

2. Gap versus particle. Give a spectral set with a mass gap but no isolated mass shell.

Solution

The set $\{0\}\cup\{p:p^0\ge\sqrt{|\mathbf p|^2+m^2}\}$ is separated from the vacuum but fills the entire region above threshold. It contains no isolated hyperboloid distinguished from continuum.

References

• Glimm, James, Arthur Jaffe, and Thomas Spencer. “The Wightman Axioms and Particle Structure in the $P(\phi)_2$ Quantum Field Model.” Annals of Mathematics 100 (1974): 585–632. DOI.
• Haag, Rudolf. “Quantum Field Theories with Composite Particles and Asymptotic Conditions.” Physical Review 112 (1958): 669–673. DOI.
• Ruelle, David. “On the Asymptotic Condition in Quantum Field Theory.” Helvetica Physica Acta 35 (1962): 147–163. E-Periodica record.
• Summers, Stephen J. “A Perspective on Constructive Quantum Field Theory.” arXiv:1203.3991, revised 2016. Open PDF.