Skip to content

Multi Channel Detection Framework for Ultralight Axion like Fields Complete Integration Statistical Validation and Degeneracy Breaking with Heterogeneous Quantum Sensors, Atomic Clocks and Gravitational Wave Detectors

FieldValue
JournalZenodo (CERN European Organization for Nuclear Research)
Published2026-05-03
AuthorsWalter Ayala
DOI10.5281/zenodo.20007043

The six-channel framework projects a one-year, 5σ reach of $d\alpha/\Lambda_\alpha < 2.4 \times 10^{-14}\ \mathrm{eV}^{-1}$ at $m_\Phi \sim 10^{-14}\ \mathrm{eV}$ using six terrestrial optical clocks. Its simulated detector has an ROC area under the curve of $0.91 \pm 0.02$. These are projections and Monte Carlo results, not a reported field detection.

The framework treats Φ as a generic ultralight field within an effective field theory. Its observables constrain couplings through $g_e$, $g_N$, $d\alpha/\Lambda_\alpha$, and $\xi/M_{\rm Pl}$. If Φ forms part of the local dark matter, its frequency is $f_\Phi=mc^2/(2\pi\hbar)$. The model uses a quality factor of about $10^6$ and a virial velocity of about $10^{-3}c$.

The field is treated as a classical coherent wave. Correlations between detectors therefore arise from field coherence. The author finds no sensitivity benefit from distributing quantum entanglement between sensors.

The six inputs are terrestrial optical clocks, GNSS atomic clocks, NV centers in diamond, hyperpolarized xenon-129 with optical magnetometry, SQUID, and pulsar timing arrays. Fiber links provide a common time reference with $10^{-19}$ precision, but the author assigns them only a $\times 1.0$ SNR improvement. The terrestrial clock channel dominates the combined SNR.

Each time series contains stationary Gaussian noise and a sinusoidal signal set by the EFT parameters. A matched filter is applied within each coherent segment. Results are then combined across segments and channels.

The atomic sensitivity factor K is about 1 for Sr, Yb, and Al$^+$ optical transitions. It is about 0.1 for Rb, Cs, and H-maser hyperfine transitions. The H-maser value is a conservative estimate because the required atomic physics is not fully calculated.

The nominal coherence time at $m_\Phi \sim 10^{-14}\ \mathrm{eV}$ is $2.6 \times 10^6\ \mathrm{s}$, or about 30 days. Projections instead use $10^6\ \mathrm{s}$, or about 12 days, which gives 32 coherent segments in one year. After incoherent segment combination, SNR scales as $\sqrt{T}$. The simulations found deviations below 3% over 30 to 365 days.

The generalized likelihood ratio has five degrees of freedom for $g_e$, $g_N$, $d\alpha/\Lambda_\alpha$, $m_\Phi$, and one phase parameter. Because zero amplitude lies on the parameter boundary, the correct null distribution is a chi-square mixture. Its 5σ threshold is 37.7, while the paper uses the conservative value 52.1 from uncorrected Wilks’ theorem. The adopted threshold corresponds to $p=2.87\times10^{-7}$.

The simulation used $T=3.15\times10^7\ \mathrm{s}$, $\tau_c=10^6\ \mathrm{s}$, and 32 coherent segments. It injected $A_\Phi=2.4\times10^{-14}$ at the one-year 5σ threshold. Signal frequencies were 2.4 Hz for terrestrial clocks and $2.31\times10^{-5}$ Hz for GNSS. Clock noise used an Allan variance power spectral density, while quantum-sensor noise was white Gaussian.

The study used $10^3$ realizations for SNR estimation and $10^2$ for the ROC. Under H1, the mean likelihood ratio was $58.3\pm2.1$ versus 57.1 predicted. Detection probability was $0.69\pm0.03$ versus 0.72 predicted. Simulated false alarm probability was $(2.1\pm1.5)\times10^{-7}$ versus $2.87\times10^{-7}$ predicted.

The cross-channel consistency score reached $0.94\pm0.03$ for injected signals. Noise-only trials scored below 0.5 in 99% of cases. The 100 ROC realizations cannot directly validate a $10^{-7}$ false alarm rate. The paper estimates that about $10^8$ realizations would be required, so Wilks’ theorem was checked only at accessible significance levels near $10^{-3}$.

The Fisher information matrix describes the information that the combined data provide about the EFT parameters. Its Jacobian must have full column rank for the relevant couplings to be identifiable. The complete six-channel Jacobian has this property only if all six channels operate with sufficient sensitivity.

Different channel subsets provide the required parameter separation. Terrestrial and GNSS clocks constrain $d\alpha/\Lambda_\alpha$. NV centers and SQUID provide electron-coupling information. Xe-129 with optical magnetometry and SQUID provide nucleon-coupling information.

The current terrestrial, GNSS, and pulsar timing system constrains only $d\alpha/\Lambda_\alpha$. The $g_e$ and $g_N$ parameters remain degenerate. For the complete system at $m_\Phi=10^{-14}\ \mathrm{eV}$, the Fisher matrix condition number is about $8.3\times10^3$. Projected one-year precisions are 5%, 30%, and 40% for $d\alpha/\Lambda_\alpha$, $\Lambda_e^{-1}$, and $\Lambda_N^{-1}$. Relatively weak NV and Xe-129 signals limit the $g_e$ and $g_N$ precision.

Sensitivity and comparison with prior work

Section titled “Sensitivity and comparison with prior work”

For $d\alpha/\Lambda_\alpha=10^{-13}\ \mathrm{eV}^{-1}$, the paper gives the following projected times and one-year reaches.

Configuration5σ detection timeOne-year $d\alpha/\Lambda_\alpha$ reach
Single terrestrial clock pairAbout 300 days$9\times10^{-14}\ \mathrm{eV}^{-1}$
Six-clock network, baselineAbout 20 days$2.4\times10^{-14}\ \mathrm{eV}^{-1}$
Full system integrationAbout 48 days, conservative$2.4\times10^{-14}\ \mathrm{eV}^{-1}$, validated

The 48-day projection is conditional on continuous clock comparison data and the validated statistical pipeline. This row includes a conservative allowance for clock systematics.

The paper describes the six-clock reach as competitive with existing laboratory constraints while covering a complementary mass range. Its projected range extends between $10^{-14}$ and $10^{-11}\ \mathrm{eV}$, whereas terrestrial clock comparisons discussed in the paper probe masses much below $10^{-22}\ \mathrm{eV}$. At $m_\Phi\sim10^{-14}\ \mathrm{eV}$, the projected fractional variation is $|\delta\alpha/\alpha|\sim2.4\times10^{-20}$. This is about $10^3$ times weaker than the cited laboratory limit of $10^{-17}$.

The corrected improvement budget gives a total sensitivity improvement of $\times2.5$, rather than the earlier $\times18.6$. Synchronization contributes $\times1.0$ rather than $\times3.2$, and multi-channel fusion contributes $\times1.0$ rather than $\times1.8$.

The estimated one-year clock systematic floor is about $10^{-19}$. Temperature-induced drift contributes about $10^{-19}$, seismic correlations $10^{-20}$, solar activity $10^{-21}$, fiber-link phase noise $10^{-22}$, and residual relativistic corrections $10^{-23}$. The floor is below the projected statistical sensitivity for six clocks over one year, but the paper requires it to be monitored and subtracted in any detection claim.

Stationary noise is another limitation. Real satellite clocks can exhibit frequency jumps and thermal transients, so the Allan variance model requires robust outlier rejection. Environmental correlations from temperature, seismic activity, and solar cycles also require modeling beyond the simulation.

Phase 1 uses existing PTB-SYRTE-NPL fiber comparisons, IGS and CDDIS GNSS products, and the 15-year NANOGrav pulsar timing dataset. Continuous year-long clock data may require metrology institute collaborations. Reaching $10^{-15}$ GNSS stability requires dual-frequency carrier phase processing beyond standard IGS products. Phase 2 runs from 2028 to 2035 for Xe-129 with optical magnetometry and SQUID arrays. Phase 3 runs from 2035 to 2040 for NV-diamond sensors and LISA, when the paper places full degeneracy breaking.

The NV-diamond channel supplies electron-coupling information that the clock channels cannot provide alone. Its contribution still depends on adequate sensitivity because the formal rank argument assumes all six channels are operational. The Fisher condition number shows that weak NV and Xe-129 signals limit the recovered electron and nucleon couplings.

The paper does not specify NV concentration, crystal dimensions, sensor geometry, resonance linewidth, operating temperature, or optical readout. Those missing specifications prevent selection of a diamond grade based on this work alone. The material and device work must define them before the projected Phase 3 performance can be treated as an engineering target.

The paper’s future channel requires nitrogen-vacancy centers in diamond for local electron-coupling sensing. 6CCVD’s diamond materials catalog lists MPCVD-grown single-crystal and polycrystalline diamond, which provides a base material family, but the stated facts do not specify NV defect density or a ready sensor grade. Use the linked engineering support service to confirm defect processing, geometry, and readout requirements before treating the channel as available.

View Original Abstract

I present the complete design, rigorous statistical validation, and quantitativeanalysis of a fully interconnected multi-channel detection framework forultralight axion-like fields. The system integrates six complementary sensor techno-logies — terrestrial optical atomic clocks, GNSS satellite atomic clocks, nitrogen-vacancy (NV) centers in diamond, hyperpolarized xenon-129 with optical magneto-metry (OPM), superconducting quantum interference devices (SQUID), and pulsartiming arrays — into a unified instrument synchronized by the existing global fiber-optic metrology network. A statistically complete detection pipeline is pre-sented, including likelihood-based inference, generalized likelihood ratio tests withWilks’ theorem incorporating boundary corrections, finite coherence time correc-tions, and Monte Carlo validation. The Fisher information matrix is constructedto formally prove that the heterogeneous architecture breaks degeneracies betweenEFT parameters (ge, gN , dα/Λα, ξ/MPl) inaccessible to any single-channel experi-ment, contingent on the deployment of future quantum sensors. The terrestrial op-tical clock network provides a dual function: (i) as the most sensitive Φ detectionchannel, probing dα/Λα ∼ 2,4×10−14 eV−1 at mΦ ∼ 10−14 eV with 6 clocks in 1 year,and (ii) as the global frequency reference distributing absolute time via operationalfiber-optic links with 10−19 precision. I rigorously validate the detection frameworkthrough Monte Carlo simulation, achieving area under the ROC curve of 0,91±0,02,and confirm that the false alarm probability at the 5σ threshold matches the theoretical prediction from Wilks’ theorem within statistical uncertainties. The projected5σ detection time is approximately 48 days for dα/Λα ∼ 10−13 eV−1, conditionalon the validated statistical pipeline and assuming continuous clock comparison dataavailability. I emphasize that the projected sensitivity requires addressing data accesschallenges and systematic environmental effects beyond the idealized models presen-ted here. The framework leverages existing infrastructure with minimal incrementalhardware requirements for Phase 1 (terrestrial clocks + GNSS + PTA), while Phase2 (quantum sensors) and Phase 3 (full integration) require significant technologicaldevelopment.