Perturbative construction of amplitudes from on-shell trees with vacuum pairs: the all-plus four-gluon amplitude through order $\boldsymbol{g}^{\boldsymbol{6}}$

This paper proposes a fixed-order perturbative on-shell construction of scattering amplitudes using BCFW-generated trees and integrated vacuum pairs, successfully reproducing the known one- and two-loop all-plus four-gluon amplitudes up to order $g^6$ through a polygon-organized inclusion-exclusion framework.

The Gist

Imagine you are trying to figure out how four tiny, invisible marbles (gluons) bounce off each other. In the world of quantum physics, calculating exactly how they interact is like trying to solve a massive, 3D puzzle where the pieces keep changing shape.

Usually, physicists solve this by drawing "Feynman diagrams." Think of these diagrams as blueprints that show every possible path the marbles could take, including paths that go through "ghost" states—things that exist mathematically but can't actually be seen. These blueprints are accurate, but they are messy, full of redundant steps, and often require canceling out huge numbers just to get a simple answer.

This paper proposes a cleaner way to build the solution, called the "Vacuum-Pair Construction." Here is how it works, using simple analogies:

1. The Building Blocks: On-Shell Trees

Instead of using the messy blueprints with ghost states, the authors start with the simplest, most solid building blocks: three-point interactions. Imagine these as the basic "handshakes" between three particles.

• The Rule: If you know how three particles can shake hands, you can build a whole tree of interactions by gluing these handshakes together.
• The Problem: This only works for "tree-level" interactions (simple bounces). It doesn't account for the complex loops and delays that happen in real, high-energy collisions (like the "one-loop" or "two-loop" effects).

2. The Secret Ingredient: "Vacuum Pairs"

To fix the missing complexity, the authors introduce a trick. They imagine inserting invisible pairs of particles into the mix.

• The Analogy: Think of a vacuum pair like a ghostly echo. You have a particle moving forward and its "conjugate" (a mirror image) moving backward. Together, they carry zero net energy and zero net momentum. You can't see them, and they don't change the final outcome, but they act like a temporary scaffold.
• The Process: The authors take their "tree" of handshakes and insert these invisible vacuum pairs into the gaps. They then "integrate" (sum up) over all the possible ways these pairs could exist. It's like shaking a box of invisible marbles and seeing how they rearrange the visible ones.

3. The Accounting Trick: Inclusion-Exclusion

Here is the clever part. If you just add up all these vacuum pair scenarios, you might count the same physical situation twice.

• The Analogy: Imagine you are counting people in a room. If you count everyone wearing a red hat, then everyone wearing a blue hat, you might double-count the person wearing both.
• The Solution: The authors use an "Inclusion-Exclusion" sign rule.
  • Add the scenarios with one invisible pair (+).
  • Subtract the scenarios with two invisible pairs (–) because they overlap too much.
  • Add the scenarios with three pairs (+) to fix the subtraction.
  • This ensures every unique physical possibility is counted exactly once, no more, no less.

4. The Polygon Game

To keep track of all these combinations, the authors use a visual method involving polygons (shapes with many sides).

• The Analogy: Imagine the particles are vertices on a polygon.
  • A Hexagon (6 sides) represents a specific type of interaction with one vacuum pair.
  • Two Quadrilaterals (4 sides each) represent a split interaction with two vacuum pairs.
  • An Octagon (8 sides) represents a more complex interaction with two vacuum pairs.
• The paper systematically lists every possible polygon shape that fits the rules for a specific level of complexity (called "order $g^4$" and "order $g^6$").

5. The Results: Rebuilding the Puzzle

The authors tested this method on a specific, difficult problem: the "all-plus four-gluon amplitude." This is a scenario where four gluons interact, and they all have the same "spin" direction (like four spinning tops all spinning clockwise).

• The Test at Order $g^4$ (One-Loop): They built the solution using their vacuum pairs and polygons. The result perfectly matched the known, standard answer for a one-loop interaction. It was like rebuilding a known house using only bricks and mortar, without the original blueprints, and getting the exact same structure.
• The Test at Order $g^6$ (Two-Loop): This is the big test. They went deeper, looking at more complex interactions involving octagons, hexagons, and pentagons.
  • They found that the "vacuum pair" method naturally produced the exact same mathematical expressions as the standard, messy Feynman diagrams.
  • They identified specific "sectors" (like the Octagon, the Hexagon-Quadrilateral, and the Bow-Tie shapes) that correspond to the complex "planar" and "non-planar" loops found in traditional physics.

The Bottom Line

The paper claims that you don't need to rely on "off-shell" (unobservable, gauge-dependent) fields to calculate these complex particle interactions. Instead, you can:

1. Start with simple, observable three-particle handshakes.
2. Glue them together into trees.
3. Insert invisible "vacuum pairs" to simulate loops.
4. Use a specific "plus-minus" counting rule to avoid double-counting.
5. Organize everything into polygon shapes.

By doing this, they successfully reconstructed the known, complex two-loop results for four-gluon scattering. It's a new, cleaner way to build the same physical reality, proving that you can get the full picture just by gluing together the simplest, most solid pieces of the puzzle.

Technical Summary

Technical Summary: Perturbative Construction of Amplitudes from On-Shell Trees with Vacuum Pairs

Problem Statement
The paper addresses the organization of perturbative scattering amplitudes in gauge theories without relying on off-shell fields or gauge-fixed Feynman diagrams. While modern amplitude programs have successfully reconstructed tree-level amplitudes using on-shell three-point building blocks and BCFW recursion, extending this on-shell philosophy to loop orders remains a challenge. Standard loop calculations involve integrations over virtual momenta and off-shell propagators, introducing unphysical intermediate quantities and gauge dependence. The authors ask whether fixed-order perturbative amplitudes beyond tree level can be constructed entirely from on-shell tree amplitudes by supplementing them with specific auxiliary structures, thereby avoiding off-shell intermediate states.

Methodology
The authors propose a "vacuum-pair construction" formulated at fixed perturbative orders. The core methodology involves the following steps:

1. Input Data: The construction relies on the particle spectrum, allowed on-shell three-point amplitudes (fixed by Lorentz invariance and little-group scaling), and color ordering.
2. Vacuum Pairs: To generate loop-level contributions, the authors insert $r \ge 0$ unobservable, state-conjugate on-shell pairs, termed "vacuum pairs," with momenta $(\ell_a, -\ell_a)$. These pairs are integrated over their Lorentz-invariant phase space.
3. Tree Gluing: The amplitude is constructed by gluing ordinary connected on-shell tree amplitudes. These trees are generated recursively from the three-point data using BCFW recursion. The vacuum pairs act as internal legs connecting these tree components.
4. Inclusion-Exclusion Sign Prescription: To prevent double-counting of overlapping phase-space regions, the authors assign alternating signs to contributions based on the number of simultaneous vacuum-pair insertions ($r$) within a connected phase-space component. A component with $r$ insertions carries a sign of $(-1)^{r-1}$. For factorized contributions with $c$ independent components, the total sign is $(-1)^{r-c}$.
5. Polygon Bookkeeping: To organize the fixed-order bookkeeping, $n$-point tree amplitudes with $r$ vacuum pairs are represented as polygons with $n+2r$ sides. The coupling order $g^{N_3}$ is determined by the number of cubic vertices ($N_3$) in the polygon decomposition. Sectors are labeled by the side counts of the polygons involved (e.g., $\{6\}$, $\{4,4\}$).
6. Dimensional Regularization: The calculation is performed in $D = 4-2\epsilon$ dimensions. The internal gluon state sums include $D_s - 2$ physical states, where the transverse momentum components ($\lambda$) are crucial for generating non-vanishing rational terms in all-plus amplitudes.

Key Contributions and Results
The paper tests this framework on the color-ordered four-gluon all-plus amplitude ($A_4(1^+, 2^+, 3^+, 4^+)$) through orders $g^4$ and $g^6$.

• Order $g^4$ (One-Loop):

  • The construction identifies two sectors: the one-pair hexagon sector $\{6\}$ and the two-pair product of quadrilaterals $\{4, 4\}$.
  • The strictly 4D helicity part of the hexagon vanishes, but the transverse ($\lambda$-dependent) part of the $D_s$-dimensional state sum yields a non-zero rational numerator.
  • The $\{6\}$ sector provides single-cut contributions, while the $\{4, 4\}$ sector provides double-cut contributions.
  • Summing these signed contributions reproduces the standard Feynman-tree theorem opening of a scalar box integral. The result matches the known finite rational one-loop all-plus amplitude.

• Order $g^6$ (Two-Loop):

  • The fixed-order sectors identified are the octagon $\{8\}$, the hexagon-quadrilateral product $\{6, 4\}$, the two-pentagon product $\{5, 5\}$, and the three-quadrilateral product $\{4, 4, 4\}$.
  • Sector Analysis:
    • The $\{8\}$ sector (one octagon, two vacuum pairs) generates contributions to the planar double-box and crossed double-box families.
    • The $\{6, 4\}$ sector (one hexagon, one quadrilateral, three vacuum pairs) produces contributions to both the seven-propagator double-box family and the six-propagator bow-tie family.
    • The $\{5, 5\}$ sector (two pentagons, three vacuum pairs) contributes to the planar double-box family.
    • The $\{4, 4, 4\}$ sector (three quadrilaterals, four vacuum pairs) splits into contributions for the double-box and bow-tie families.
  • Vanishing Configurations: The authors explicitly demonstrate that configurations leading to soft boundaries, collinear boundaries with no scale, or isolated all-plus tree factors vanish in dimensional regularization.
  • Reproduction of Standard Results: By summing the signed phase-space integrals over these sectors, the construction reproduces the known planar, non-planar, and bow-tie expressions for the two-loop all-plus amplitude. The numerators derived from the vacuum-pair state sums match the standard double-box ($N_{DB}$) and bow-tie ($N_{BT}$) numerators found in the literature.

Significance and Claims
The paper claims to demonstrate that fixed-order perturbative amplitudes can be systematically constructed using only on-shell tree amplitudes and vacuum-pair insertions, without invoking off-shell propagators or gauge-fixing as intermediate building blocks.

• Scope: The authors emphasize that the particle spectrum, three-point couplings, and color structure are inputs, not derived from the method. The contribution lies in showing how these inputs generate higher-loop amplitudes via on-shell data alone.
• Validation: The method is validated by its ability to reproduce the known one-loop and two-loop all-plus four-gluon amplitudes, including the specific rational numerators and the correct denominator families (planar, non-planar, and bow-tie).
• Framework: The work extends the ideas proposed in previous literature (cited as [15–18]) by providing the first explicit two-loop analysis. It establishes a "fixed-order bookkeeping" system using polygons to organize the combinatorics of vacuum-pair insertions and phase-space integrations.

The paper concludes that this construction offers a viable alternative organization for gauge-theory amplitudes, where the physical on-shell result is obtained by summing signed on-shell tree products over vacuum-pair phase spaces, effectively bypassing the need for unphysical off-shell intermediate states.