Di-nucleons do not form bound states at heavy pion mass

This high-statistics lattice QCD study at a heavy pion mass ($m_\pi \simeq 714$ MeV) demonstrates that di-nucleons do not form bound states, attributing previous claims of deeply bound states to misidentifications of the spectrum arising from off-diagonal correlation function elements rather than physical hexaquark states.

The Gist

Imagine the universe is built out of tiny, invisible Lego bricks called quarks. When you snap three of these bricks together, you get a proton or a neutron (collectively called nucleons). When you snap two nucleons together, you get the nucleus of an atom.

For decades, physicists have been trying to figure out exactly how these Lego bricks snap together. Specifically, they wanted to know: Do two neutrons stick together on their own? Do a proton and a neutron stick together to form a deuteron (the heart of heavy hydrogen)?

This paper is like a high-stakes detective story where a team of scientists finally solves a mystery that has confused the physics community for over ten years.

The Mystery: The "Ghost" Bond

In the past, some scientists used a specific method to simulate these particles on supercomputers. They claimed to find that two neutrons could stick together tightly, forming a "bound state" (like a very strong magnet). Other scientists, using a different method, said, "No way, they don't stick at all."

It was like two groups of people looking at the same blurry photo. One group said, "That's a dog!" and the other said, "No, it's a cat!"

The problem was that the "dog" (the bound state) might have been a trick of the light—a hallucination caused by the way the photo was taken.

The Investigation: A High-Definition Camera

The authors of this paper decided to settle the argument once and for all. They didn't just pick one camera; they brought out the entire camera shop. They used the same supercomputer data (the same "Lego set") but applied every major method used in the field to analyze it.

Here is how they did it, using some everyday analogies:

1. The Signal-to-Noise Problem (The Static on the Radio)

Imagine you are trying to listen to a very quiet whisper (the interaction between two neutrons) while standing next to a roaring jet engine (the massive energy of the particles themselves).

• The Challenge: In computer simulations, the "whisper" is so faint that the "jet engine" noise drowns it out almost instantly.
• The Solution: The team used a technique called sLapH. Think of this as a super-smart noise-canceling headphone that filters out the jet engine so you can hear the whisper clearly. They also used a "Conspiracy Model," which is like assuming that the background noise in the whisper and the background noise in the roar are actually related, allowing them to cancel each other out mathematically.

2. The "Hexaquark" Trap (The Magic Trick)

Some previous studies used a special tool called a Hexaquark operator. Imagine you are trying to find a hidden treasure.

• The Old Way: You used a metal detector (momentum operators) that scans the whole beach.
• The New Way: Some scientists said, "No, use a magic wand (Hexaquark operator) that only beeps if the treasure is buried exactly here."
• The Discovery: This paper tested the magic wand. They found that the wand was actually beeping at the wrong things. It was picking up "ghosts" (excited states) and confusing them with the real treasure. When they added the magic wand to their analysis, it didn't find any new treasure; it just confirmed that the old "treasure" was a mirage.

3. The "False Plateau" (The Flat Spot on a Hill)

This is the most critical part of the discovery.
Imagine you are hiking up a mountain (representing time in the simulation). You want to find the peak (the true energy of the particles).

• The Mistake: In the old studies, the path looked like it flattened out (a "plateau") at a certain height. The hikers thought, "Great, we reached the top!" and stopped.
• The Reality: The paper shows that this flat spot was an optical illusion caused by the "noise" of the mountain. If you kept hiking, you would see the path actually went down or stayed higher than they thought. The "bound state" was just a false plateau created by mathematical errors in how they read the map.

The Verdict: No Magic Glue

After running thousands of simulations and checking every angle, the team reached a clear conclusion:

At the heavy mass they tested (which is like testing the Lego bricks with a slightly heavier plastic), two neutrons do NOT stick together.

• The Deuteron (Proton + Neutron): They found it is almost a bound state, but not quite. It's like two magnets that are very close to sticking but are just a tiny bit too far apart.
• The Di-neutron (Neutron + Neutron): They definitely do not stick. They are just two lonely particles passing each other.

Why Does This Matter?

You might ask, "So what? They don't stick at this heavy mass."

This is huge because:

1. It Clears the Air: It proves that the "deeply bound" states seen in older papers were mistakes, not real physics. It stops the field from building theories on a foundation of sand.
2. It Validates the Tools: It shows that the new, high-tech "noise-canceling" methods (like the one used here) are reliable.
3. The Road to Reality: Now that they know their tools work and the "ghosts" are gone, they can move on to simulating particles with their real physical weights (lighter mass). This is the first step toward understanding how the stars burn and how the universe is built, all from the bottom up, using only the fundamental laws of nature.

In short: The scientists looked at the Lego bricks with a microscope, a telescope, and a magic wand. They realized the "glue" they thought they saw was just a trick of the light. The bricks don't stick together the way some thought, and now we can finally build a correct model of the universe.

Technical Summary

1. Problem Statement

The paper addresses a long-standing and significant discrepancy in the lattice Quantum Chromodynamics (LQCD) literature regarding the existence of bound di-nucleon states (specifically the deuteron in the isospin-0 channel and the di-neutron in the isospin-1 channel) at heavy pion masses.

• The Conflict: Early LQCD calculations (e.g., NPLQCD collaborations) using local hexa-quark (HX) creation operators reported the existence of deeply bound di-nucleon states at pion masses around $m_\pi \approx 800$ MeV. Conversely, calculations using the HAL QCD potential method and later studies utilizing momentum-space operators found no evidence of such bound states, suggesting only scattering states or virtual states.
• The Challenge: Determining the interaction energy of two nucleons is extremely difficult because it is an $O(0.1\%)$ shift relative to the total energy of the system. This is compounded by the exponential degradation of the signal-to-noise ratio in LQCD and the difficulty of isolating the ground state from excited state contamination.
• The Hypothesis: The authors hypothesize that the previous identification of deeply bound states was an artifact caused by the misidentification of plateaus in effective energy plots derived from off-diagonal correlation functions involving HX operators, where negative spectral weights and slow-decaying excited state contamination create "false plateaus."

2. Methodology

The authors performed a high-statistics LQCD calculation on the same set of gauge configurations to isolate methodological systematic uncertainties.

• Lattice Setup:
  • Ensemble: CLS $N_f=2+1$ isotropic clover-Wilson action with non-perturbative $O(a)$ improvement.
  • Parameters: Pion mass $m_\pi \approx 714$ MeV (where $m_\pi = m_K$, the SU(3) flavor symmetric limit). Lattice spacing $a \approx 0.086$ fm, volume $48^3 \times 96$.
  • Statistics: 1,490 configurations from four replicas.
• Operator Construction:
  • Primary Method: Used the stochastic Laplacian Heaviside (sLapH) method to construct nucleon operators projected to definite momentum. This allows for the construction of multi-hadron operators that are linear combinations of products of individually momentum-projected hadrons.
  • Basis: A large basis of interpolating fields was used, including momentum-space creation operators for various irreducible representations (irreps) of the little group.
  • HX Test: To directly address the controversy, the authors added local hexa-quark (HX) operators to their correlation matrix basis to test if they couple to a deeply bound state missed by momentum-space operators.
• Analysis Techniques:
  • Correlation Matrix: Used the Hermitian correlator matrix method with the Generalized Eigenvalue Problem (GEVP) to extract the energy spectrum.
  • "Conspiracy" Model: Introduced a novel fitting model for the excited state contamination. Instead of treating NN excited states agnostically, this model assumes the excited states in the NN correlator are inelastic excitations of the single nucleons used to build the operators. This allows the excited states to "conspire" to cancel in the ratio correlator, a phenomenon observed in previous data.
  • HAL QCD Potential: Performed a parallel calculation of the HAL QCD potential on the same ensembles to cross-check results using a different formalism (solving the Nambu-Bethe-Salpeter equation).
  • Toy Models: Constructed toy models to demonstrate how off-diagonal correlators with negative weights can produce false plateaus in effective mass plots, mimicking bound states.

3. Key Contributions

1. Resolution of the Discrepancy: By applying multiple methods (momentum-space operators, HX operators, and HAL QCD potential) to the same gauge ensemble, the authors isolated the source of the discrepancy. They demonstrated that the inclusion of HX operators does not alter the low-lying spectrum, proving that previous claims of deep binding were not due to a lack of operator overlap.
2. Diagnosis of "False Plateaus": The paper provides a rigorous explanation for the previous false positives. It shows that off-diagonal correlators (HX source to NN sink) can have negative spectral weights. When combined with slow-decaying excited state contamination, these can create a "false plateau" in the effective mass at intermediate times, leading to the misidentification of a deeply bound state.
3. The "Conspiracy" Model: The authors developed and validated a robust fitting strategy ("conspiracy model") that accounts for the specific structure of excited state contamination in two-nucleon systems, yielding more stable and reliable extractions of the ground state energy compared to agnostic multi-exponential fits.
4. Cross-Validation: The study provides the first high-statistics comparison where the Lüscher finite-volume method (QC2) and the HAL QCD potential method yield qualitatively consistent results on the same ensemble.

4. Key Results

• No Bound States Found:
  • Deuteron ($I=0$): The analysis rules out a bound state with a confidence level of $>5\sigma$. The scattering length is determined to be negative (indicating no bound state), and a virtual bound state is identified.
  • Di-neutron ($I=1$): A bound state is excluded with a confidence level of $>2.5\sigma$.
• Ineffectiveness of HX Operators: Adding HX operators to the correlation matrix basis did not shift the low-lying energy levels. The overlap factors showed that HX operators couple strongly only to high-energy states and negligibly to the low-lying scattering states of interest.
• HAL QCD Agreement: The HAL QCD potential analysis, after controlling for systematic uncertainties (excited state contamination, discretization, finite volume), yielded phase shifts consistent with the QC2 results, confirming the absence of bound states.
• Toy Model Validation: The toy models successfully reproduced the "false plateau" behavior seen in earlier literature, confirming that the discrepancy arises from the misinterpretation of off-diagonal correlators rather than a failure of the momentum-space method.

5. Significance

• Settling the NN Controversy: This work effectively resolves a decade-long controversy in nuclear LQCD. It establishes that the previously reported deeply bound di-nucleons at heavy pion masses were artifacts of analysis methods (specifically the use of off-diagonal correlators with negative weights) rather than physical reality.
• Methodological Rigor: The paper sets a new standard for two-baryon calculations by demonstrating the necessity of using momentum-space operators and large operator bases, and by introducing the "conspiracy model" for handling excited state contamination.
• Path to Physical Point: By validating the reliability of these methods at heavy pion masses, the authors pave the way for future calculations at lighter pion masses ($m_\pi \lesssim 200$ MeV) and ultimately the physical point. This is a critical step toward deriving nuclear physics (such as the properties of the deuteron and nuclear forces) directly from the fundamental theory of QCD.
• Implications for Previous Work: The findings invalidate previous LQCD matrix element calculations for few-body systems that relied on the assumption of deeply bound di-nucleons, highlighting the need to re-evaluate those results.

In conclusion, the paper provides compelling evidence that di-nucleons do not form bound states at heavy pion masses, attributing previous contradictory results to systematic errors in the extraction of energy levels from off-diagonal correlation functions.