Exploratory Search for Anomalies in Coherence Mediated Entanglement Generation
Paper details
Section titled “Paper details”| Field | Value |
|---|---|
| Journal | Zenodo (CERN European Organization for Nuclear Research) |
| Published | 2026-06-01 |
| Authors | Walter Ayala |
| DOI | 10.5281/zenodo.20491651 |
A numerical simulation detected the proposed anomaly of $\eta = 0.05$ in 33 of 35 candidates. The maximum significance was $Z = 16.55$. The standard quantum mechanics scenario produced zero false positives. These results validate the statistical protocol, not a physical anomaly. The paper reports no laboratory measurement of entanglement without direct electromagnetic interaction.
Search target and null prediction
Section titled “Search target and null prediction”The protocol asks whether the quantum coherence of a messenger NV center, M, can induce entanglement between two target spins, A and B, without direct electromagnetic interaction. It starts from a product state, with A and B in $|0\rangle$, and applies microwave pulses and dynamical decoupling only to M. The standard prediction is $W_{\min} \geq 0$ for the minimum of four Bell-state witnesses.
The anomaly hypothesis assigns $\eta > 0$ to an unspecified mechanism. It separates temporal, distance, and angular behavior into $g(\tau)$, $h(d)$, and $k(\theta)$. Standard physics sets $\eta = 0$. No coupling mechanism is assumed, so a positive result would identify a deviation before it supplied a physical explanation.
The primary observable uses correlators $\langle XX\rangle$, $\langle YY\rangle$, and $\langle ZZ\rangle$ for each witness. A negative minimum witness value establishes entanglement of A and B. Each correlator is measured through basis rotation and optical readout of A and B. Optional tomography after a significant Stage 2 result uses nine measurement bases and 27 correlators at $10^6$ repetitions. Maximum likelihood then reconstructs the density matrix.
Diamond platform and interaction bounds
Section titled “Diamond platform and interaction bounds”The experimental platform uses three $S = 1$ nitrogen-vacancy centers. A and B are fabricated $100\ \mu\text{m}$ apart on the same diamond substrate. M sits on an independent piezoelectric mount at 1 mm, 1 cm, or 10 cm from each target. The design operates at 1.5 K in a continuous-flow superfluid helium cryostat under $B_0 = 0.5$ T. The field is aligned with the NV axis. Diamond is enriched in $^{12}$C to $[^{13}\text{C}] < 0.01%$.
At 1.5 K, the zero-field splitting is about 2.877 GHz, and the measured transition is about 16.877 GHz. The 7 MHz shift from the room-temperature value of 2.870 GHz makes in situ photoluminescence calibration necessary before setting pulse frequencies. XY-8 decoupling supports $T_2 \sim 1$ ms for up to 1000 pulses, while $T_1$ exceeds 100 ms.
The protocol bounds conventional channels before testing the anomaly. At the $100\ \mu\text{m}$ A-B separation, a dipolar phase of $\pi$ would require about 111 days at $\tau \sim 1$ ms. At 1 mm, the M-target dipolar phase is $3.27 \times 10^{-13}$ rad during 1 ms. This corresponds to about 300 years for a $\pi$ phase.
A single $^{13}$C nucleus cannot bridge the targets. Its effective 1 Hz range is only 15 to 20 nm, versus $10^5$ nm for the A-B separation. Triple-layer shielding and active magnet stabilization suppress common field fluctuations. RC filters and ferrites suppress power line noise. Residual fluctuations are expected below $\Delta B_0/B_0 < 10^{-8}$.
Sequence and artifact controls
Section titled “Sequence and artifact controls”Phase 2 polarizes all three centers with a 532 nm laser, then applies a $\pi/2$ pulse of about 10 ns to M. The resulting messenger coherence is verified as $C_M = 0.5 \pm 0.02$, while A and B remain in $|0\rangle$. XY-8 applies eight $\pi$ pulses with alternating phases and repeats them to set $\tau$. The delay is swept through 1, 3, 10, 30, 100, 300, and 1000 $\mu$s with 100 ns pulse spacing.
Each point measures $\langle XX\rangle$, $\langle YY\rangle$, and $\langle ZZ\rangle$ on the targets and calculates $W_{\min}$. The positive control uses a demonstrated gate between NVs at about 10 to 30 nm, or one mediated by a $^{13}$C nucleus. That gate uses a separation far below the $100\ \mu\text{m}$ A-B spacing. The negative control expects $Z < 2$ when $C_M = 0$.
| Control | Comparison | Standard result | Anomaly or artifact signal |
|---|---|---|---|
| Messenger coherence | $C_M = 0.5$ and $C_M = 0$ | $W_{\min} \geq 0$ | Negative only with $C_M > 0$; a negative result at $C_M = 0$ suggests an artifact |
| Distance | 1 mm, 1 cm, and 10 cm | $W_{\min} \geq 0$ | Weak or constant $h(d)$; a dipolar or radiative decay suggests an electromagnetic artifact |
| Orientation | Five angles from 0° to 135° | $W_{\min} \geq 0$ | No dipolar angular pattern suggests an anomaly; a dipolar pattern suggests an artifact |
| Mixed messenger state | $C_M < 0.01$ | $W_{\min} \geq 0$ | A negative result suggests a pulse artifact |
Two-stage decision rule
Section titled “Two-stage decision rule”The search grid contains $7 \times 3 \times 5 \times 3 = 315$ combinations of $\tau$, distance, orientation, and coherence state. Stage 1 uses $10^4$ repetitions, $\delta W \approx 1.4 \times 10^{-2}$, $Z > 3$, and a 5% false discovery rate correction. Stage 2 uses $10^5$ repetitions, $\delta W \approx 4.5 \times 10^{-3}$, $Z > 5$, and Bonferroni correction. The one-tailed $p$ threshold associated with $Z > 5$ is $2.87 \times 10^{-7}$.
| Stage | Repetitions | Decision condition | Error control | Stated duration |
|---|---|---|---|---|
| Stage 1 | $10^4$ | $Z > 3$ | FDR at 5% | About 73 days serial or 7 days parallel |
| Stage 2 | $10^5$ | $Z > 5$ | Bonferroni | Not stated |
| Tomography | $10^6$ | Follows a confirmed anomaly | Nine bases and 27 correlators | Not stated |
The parallel estimate assumes ten pairs measured concurrently. The proposed parallelization uses a $10 \times 10$ matrix of A-B pairs on a $25\ \text{mm}^2$ substrate.
Numerical validation
Section titled “Numerical validation”The synthetic validation contains 669 records across two scenarios. The anomaly case sets $\eta = 0.05$, $g(\tau) = \exp(-\tau/300\ \mu\text{s})$, $h(d) = 1$, and $k(\theta) = 1$. The standard scenario sets $\eta = 0$ and includes only Gaussian noise.
At Stage 1, the anomaly case reports $W_{\min} = -0.464$ for the positive control. The value is $-0.019$ for phases 2 to 3, and $-0.002$ for phase 4 and phase 6. The Stage 2 value is $Z_{\min} = 13.5$ for phases 2 to 3. Across the full simulation, 33 of 35 candidates are confirmed in Stage 2, with $Z_{\max} = 16.55$.
The standard quantum mechanics scenario returns zero false positives. The injected $\eta = 0.05$ sits above the Stage 2 floor of 0.045 and the tomography floor of 0.015. The simulation supports sensitivity to the injected anomaly and control of statistical errors. It does not establish that the phenomenological profile occurs in NV centers. The constant distance and angular profiles were inputs to the anomaly scenario, so the simulation could not independently establish them.
Calibration and limitations
Section titled “Calibration and limitations”Ramsey calibration requires contrast $C > 0.95$ and $|\Delta\omega| < 1$ kHz. Rabi and Hahn tests require pulse fidelity $F_\pi > 0.995$. Data are discarded and recalibrated if $C < 0.90$ or $|\Delta\omega| > 5$ kHz.
Local magnetic gradients separate NV transition frequencies by at least 50 MHz. About 1 MHz pulse bandwidth provides more than 2500 suppression, or 34 dB, before local shielding. Local shielding adds more than 20 dB, giving total suppression above 54 dB.
The stated detection floor is $\eta \gtrsim 0.045$ at $10^5$ repetitions. Tomography lowers it to $\eta \gtrsim 0.015$ at $10^6$ repetitions. The explored window stops at $\tau = 1$ ms and $d = 10$ cm. A null result applies only to this NV platform and the searched conditions. It cannot exclude mechanisms that require other timescales, distances, or platforms.
The phase numbering is inconsistent. One table labels the mixed control as Phase 6, while the protocol assigns Phase 6 to distance and Phase 8 to the mixed control. The paper also does not reconcile 669 synthetic records with 315 grid points and 35 candidates.
Implications for diamond device engineering
Section titled “Implications for diamond device engineering”The material must support $S = 1$ NV centers within a $^{12}$C-enriched lattice, long dynamical-decoupling coherence, and reproducible microwave control. A and B need a fixed $100\ \mu\text{m}$ pattern within one substrate. M must remain independently movable at tested distances of 1 mm, 1 cm, and 10 cm.
The proposed $10 \times 10$ matrix shows why repeatable fabrication and parallel optical readout matter for device selection. Calibration limits also set material acceptance criteria. Pulse fidelity above 0.995 and Ramsey contrast above 0.95 are required before data collection. Selection must account for operation at 1.5 K under $B_0 = 0.5$ T and for microwave channel separation of at least 50 MHz.
A null outcome would bound this protocol’s sensitivity, not all coherence-mediated entanglement. Extending that bound would require extending the tested delay, distance, coherence, and NV platform conditions.
Materials and support
Section titled “Materials and support”The paper’s quantum-device application needs a $^{12}$C-enriched diamond substrate with NV centers, a $100\ \mu\text{m}$ A-B spacing, and a patterned $25\ \text{mm}^2$ area. 6CCVD’s diamond materials catalog lists MPCVD-grown single-crystal and polycrystalline diamond plates and wafers up to 125 mm. This matches the plate and wafer form for review, but the catalog does not specify NV-center realization or $^{12}$C enrichment. Engineering support can confirm the isotope, defect, and geometry requirements before material is assigned to the protocol.
View Original Abstract
A complete experimental protocol is presented for the exploratory search ofanomalies in entanglement generation in three-center NV diamond systems. Theexperiment investigates whether the quantum coherence of a messenger spin M caninduce entangled correlations between two target spins A and B without directelectromagnetic interaction. The design is strictly model-independent: it assumesno specific mechanism, explicitly states the null prediction of standard quantummechanics (Wmin ≥ 0), and incorporates systematic controls. Four complementaryentanglement witnesses are used, with optional tomography for anomalous states(9 bases, 27 correlator measurements). The statistical protocol is two-staged withmultiple comparison correction (FDR + Bonferroni) and realistic time estimates.Quantitative controls for microwave crosstalk, referenced magnetic shielding, andcalibration of π and π/2 pulses are included.Through numerical simulation with realistic noise (N = 669 points, 104-105rep-etitions per point), the protocol is validated: a scenario with a phenomenologicalanomaly (η = 0.05) is detected with Z > 5 in Stage 2 (33/35 candidates con-firmed, Zmax = 16.55), while the standard quantum mechanics scenario yields 0false positives. The observed distinctive signatures (weak distance dependence, ab-sence of angular dipolar pattern, contrast between control phases) are inconsistentwith known electromagnetic artifacts. A detection limit η ≳ 0.045 is established forN = 105 and η ≳ 0.015 for tomography with N = 106. The protocol is ready forpeer review and experimental implementation.
Electronic Grade Intrinsic Diamond Epitaxial Wafer H Terminated Superior 100 Mosfet Ready Substrate
View Specs →
Single Crystal Cvd Diamond Plate
View Specs →