⚛️ phenomenology

Light top-flavoured dark matter in dark SMEFT: one-loop matching and phenomenology

This paper establishes a unified one-loop DSMEFT framework for light top-flavoured dark matter to analyze its phenomenology across rare meson decays, direct and indirect detection, and thermal relic abundance, finding that while tree-level operators can explain the Belle II $B^+\to K^+\nu\bar\nu$ excess, significant portions of the favored parameter space are already constrained by $B^0\to K^{*0}+\mathrm{inv}$ and direct detection experiments.

Original authors: Meng-Chao Gao, Xin-Qiang Li, Ya-Dong Yang, Xing-Bo Yuan, Xin Zhang

Published 2026-10-27

📖 5 min read🧠 Deep dive

CC BY 4.0

Original authors: Meng-Chao Gao, Xin-Qiang Li, Ya-Dong Yang, Xing-Bo Yuan, Xin Zhang

Original paper licensed under CC BY 4.0 (http://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

For decades, the most popular idea about the invisible substance holding galaxies together was that it consists of heavy, slow-moving particles, perhaps a thousand times heavier than a proton. These hypothetical particles, known as weakly interacting massive particles, were expected to be heavy enough to bump into atomic nuclei and leave a detectable trace in deep underground tanks. However, as scientists built more sensitive detectors, they found nothing. The silence from these experiments has forced a shift in thinking. Perhaps the invisible matter is much lighter, too light to trigger the heavy sensors used for decades. If these particles are light, they might be too faint to see in a direct collision, but they could still reveal themselves by slipping away from decaying atoms, carrying energy with them that never returns.

This is the territory explored by a team of physicists who investigated a specific scenario: dark matter that interacts with the heaviest known quark, the top quark, but is light enough to be produced in the decay of B-mesons. These mesons are unstable particles containing a bottom quark that occasionally decay into a lighter kaon and something invisible. Recently, the Belle II experiment in Japan observed more of these decays than standard physics predicts, a discrepancy that could be explained if the missing energy was actually a stream of dark matter particles. The researchers set out to see if this explanation holds up when tested against every other known constraint, from the stability of the universe to the behavior of light.

To do this, the team constructed a detailed theoretical bridge connecting the high-energy world where new physics might exist to the low-energy world of our current experiments. They started with a framework that describes how new particles might interact with the standard model at very high energies, specifically focusing on interactions involving the top quark. They then mathematically translated these high-energy rules down to the lower energies where B-mesons and K-mesons live. This translation was not simple; it required calculating complex quantum loops where top quarks and other heavy particles briefly appear and disappear, effectively changing the rules of interaction for the lighter particles. This process allowed them to predict exactly how this light dark matter would behave in rare particle decays, how it would scatter off atomic nuclei in direct detection experiments, and how it would have behaved in the early universe.

The results of this comprehensive analysis paint a picture of a scenario that is tightly constrained. The researchers found that for the dark matter to explain the excess of decays seen by Belle II, the interaction strength must be relatively strong, but only if the dark matter is produced through a specific type of interaction that happens directly, without the need for complex intermediate steps. If the interaction requires those complex loops, the theory breaks down or requires unrealistic values. More importantly, the study shows that this explanation is in constant tension with other observations. While the dark matter could fit the Belle II data, it is heavily restricted by the non-observation of similar decays in other particles, such as the decay of a neutral B-meson into a K-star meson and invisible particles. This specific decay mode acts as a powerful filter, ruling out large portions of the parameter space that would otherwise look promising.

Furthermore, the team discovered that for certain types of dark matter, specifically those that interact via scalar currents, the constraints from direct detection experiments are surprisingly strong. Experiments like CRESST-III and DarkSide-50, which look for the tiny recoil of nuclei when hit by dark matter, are competitive with the rare decay searches. For dark matter masses above a certain threshold, roughly between one and two GeV, these direct detection experiments effectively rule out the very region of parameter space that the Belle II excess seems to favor. This means that if the dark matter is heavy enough to be seen by these detectors, it cannot be the one causing the Belle II anomaly. For lighter masses, the rare decay constraints remain the primary limit.

The study also looked at the cosmic history of this dark matter. If these particles were created in the early universe and survived to the present day, their abundance would be determined by how often they annihilated with each other. The calculations show that for the interactions described, the dark matter would annihilate too slowly to produce the correct amount of dark matter observed in the universe today. In most of the favored region, the universe would be filled with far too much of this substance. The only narrow window where the correct amount could be produced is at the very edge of the allowed mass range, where new decay channels open up. This suggests that if the Belle II excess is indeed caused by dark matter, there must be additional mechanisms at play, perhaps involving other particles or forces, to ensure the universe has the right amount of dark matter.

Ultimately, the paper concludes that while a light, top-flavored dark matter particle is a theoretically consistent way to explain the Belle II anomaly, it is a very fragile explanation. It survives only in a narrow strip of possibilities, constantly squeezed by the lack of signals in other decay channels and the sensitivity of direct detection experiments. The work demonstrates that the flavor of the interaction—specifically its connection to the top quark—creates a unique fingerprint that links rare particle decays, direct detection, and cosmological history together. By following this fingerprint, the researchers have shown that the simplest versions of this idea are likely incorrect, pointing toward a more complex reality where additional physics is needed to reconcile the experimental excess with the rest of the universe.

Technical Summary: Light Top-Flavoured Dark Matter in Dark SMEFT

Problem Statement
The particle nature of dark matter (DM) remains unknown. While Weakly Interacting Massive Particles (WIMPs) have been the leading paradigm, direct detection experiments (e.g., XENONnT, LZ) have increasingly constrained the parameter space for DM masses from a few GeV to several TeV. This has shifted focus toward lighter DM candidates (O(GeV)), where nuclear recoil energies become too small for traditional detectors. Concurrently, the Belle II collaboration reported an excess in the rare decay $B^+ \to K^+ \nu\bar{\nu}$ , with a branching ratio approximately $2.7\sigma$ above the Standard Model (SM) prediction. This excess could be interpreted as a decay into light invisible particles, such as DM. However, explaining this excess requires a consistent theoretical framework that connects high-energy new physics (NP) to low-energy flavor observables, direct detection, and cosmological constraints. Existing analyses often rely on simplified models or tree-level effective field theory (EFT) matching, potentially missing crucial loop-induced effects, particularly those involving the top quark.

Methodology
The authors employ the Dark Standard Model Effective Field Theory (DSMEFT) to study light top-flavoured DM (interactions restricted to third-generation quarks) below the electroweak (EW) scale. The analysis covers scalar ( $\phi$ ), fermionic ( $\chi$ ), and vector ( $X_\mu$ ) DM, as well as axion-like particles (ALPs). The methodology follows a rigorous EFT chain:

Framework Definition: The DSMEFT is defined at the EW scale ( $\mu_{EW}$ ) with dimension-6 operators (and one dimension-7 operator for fermionic DM) coupling DM fields to SM gauge-singlet quark currents. The analysis assumes a "top-flavoured" scenario where Wilson coefficients are non-zero only for the third generation ( $C_{33}$ ).
One-Loop Matching: A primary technical contribution is the performance of the one-loop matching of the DSMEFT onto the Dark Low-Energy Effective Field Theory (DLEFT) at the EW scale. This involves integrating out the top quark, Higgs boson, and $W/Z$ bosons. The calculation utilizes the standard off-shell matching approach, employing packages like FeynRules, FeynArts, and FeynCalc. Special care is taken with axial-vector currents and $\gamma_5$ traces using the "reading point" prescription to preserve Ward identities.
Renormalization Group Evolution (RGE): The resulting DLEFT Wilson coefficients are evolved from $\mu_{EW}$ down to relevant low-energy scales ( $\mu_B \approx 4.8$ GeV, $\mu_K \approx 2.0$ GeV, and $\mu_H \approx 1.0$ GeV) using QCD renormalization group equations. This accounts for the mixing of operators and the evolution of quark currents.
Matching to Non-Relativistic EFT (NREFT): For direct detection, the DLEFT is matched onto a non-relativistic EFT describing DM-nucleus scattering. The authors extend existing tools (DirectDM and DDCalc) to include vector DM and derive the necessary response functions.
Phenomenological Analysis: The framework is applied to:
- Flavor Observables: Rare $B$ and $K$ meson decays ( $B \to K^{(*)} + \text{inv}$ , $K \to \pi + \text{inv}$ , etc.).
- Direct Detection: Constraints from CRESST-III and DarkSide-50.
- Relic Density: Thermal freeze-out calculations using micrOMEGAs and DRAKE, considering all 32 $Z_2$ -symmetric DLEFT operators.
- Indirect Detection: Constraints from CMB (Planck) and Fermi-LAT on DM annihilation into photons.

Key Contributions

One-Loop Matching: The paper provides the first complete one-loop matching of DSMEFT operators (coupling exclusively to top quarks) onto the DLEFT. This captures top-quark loop effects that generate flavor-changing neutral currents (FCNCs) and flavor-diagonal couplings to light quarks and gluons, which are absent at tree level in the low-energy theory.
Unified EFT Chain: The work establishes a unified framework linking high-scale DSMEFT operators to low-energy phenomenology via the chain: DSMEFT $\to$ DLEFT (with QCD running) $\to$ NREFT.
Vector DM Extension: The authors extend the NREFT formalism and response functions to include vector DM, implementing these into public codes.
Comprehensive Constraints: A simultaneous analysis of flavor, direct detection, relic density, and indirect detection constraints for light top-flavoured DM is presented, identifying the specific operator structures that can or cannot explain the Belle II excess.

Results

Belle II Excess Interpretation: The analysis finds that only DSMEFT operators contributing at the tree level to the $b \to s$ transition can accommodate the Belle II $B^+ \to K^+ \nu\bar{\nu}$ excess with $\mathcal{O}(1)$ couplings for a new physics scale of $\Lambda = 1$ TeV. Loop-induced operators require couplings of $\mathcal{O}(10\text{--}100)$ , pushing the EFT to the boundary of its validity.
Flavor Constraints:
- For tree-level operators, the decay $B^0 \to K^{*0} + \text{inv}$ provides the strongest constraint over large portions of the parameter space.
- For scalar currents, $K^+ \to \pi^+ + \text{inv}$ dominates at low DM masses, while $B_s \to \text{inv}$ , $B^+ \to \pi^+ + \text{inv}$ , and $\Upsilon(1S) \to \text{inv}$ become important at higher masses.
- The viable parameter space for the Belle II excess is significantly restricted by these flavor bounds.
Direct Detection:
- For operators involving scalar quark currents (e.g., $Q_{u\phi^2}$ , $Q_{u\chi^2}$ , $Q_{uX^2}$ ), direct detection limits (CRESST-III and DarkSide-50) are competitive with flavor constraints. These operators generate unsuppressed gluonic interactions via top loops.
- Direct detection excludes the entire Belle II-favoured band for $m_{DM} \gtrsim 1.35$ GeV (scalar), $2.3$ GeV (fermion), and $1.8$ GeV (vector).
- For vector-current operators (e.g., $Q_{q\chi}$ ), direct detection bounds are orders of magnitude weaker than flavor constraints due to CKM or loop suppression.
Relic Density:
- Standard thermal freeze-out via top-flavoured interactions alone generally overproduces DM in the region favoured by the Belle II excess, as the annihilation cross-section is suppressed by CKM factors or loop effects.
- A simultaneous explanation of the Belle II excess and the observed relic density is only possible in a narrow mass window ( $m_{DM} \simeq 2.3\text{--}2.4$ GeV) where new annihilation channels (e.g., $s\bar{b}$ ) open up, or if additional annihilation mechanisms (e.g., into dark sector states) are invoked.
Indirect Detection: CMB and Fermi-LAT constraints on diphoton annihilation do not exclude the parameter space favoured by the Belle II excess for the operators considered.

Significance
The paper claims that the one-loop matching is essential for a predictive description of light top-flavoured DM, as it generates the necessary flavor-diagonal and gluonic interactions that control direct detection and relic density. The study demonstrates that while the top-flavoured scenario offers a minimal and predictive link between flavor anomalies and DM, it faces tight constraints. Specifically, the tension between explaining the Belle II excess and satisfying direct detection limits (for scalar currents) or relic density requirements (for most operators) suggests that a simple top-flavoured EFT explanation may be insufficient without additional model-building features, such as light mediators or additional dark sector states. The work provides the necessary theoretical tools (matching coefficients, RGE evolution, and NREFT extensions) for future precision tests of light DM scenarios at Belle II and direct detection experiments.

More like this