Skip to content

Quantum Fluctuation Spectroscopy of Chemical Bonds Using NV Centers and Scanning Tunneling Microscopy

FieldValue
JournalZenodo (CERN European Organization for Nuclear Research)
Published2026-05-01
AuthorsWalter Ayala
DOI10.5281/zenodo.19959323

The proposed framework predicts a local fluctuation-spectrum contrast of (C(\mathrm{C=C}):C(\mathrm{C{-}C}\ \mathrm{arom.}):C(\mathrm{C{-}C}) \approx 2.3:1.6:1). Order-of-magnitude estimates put the nitrogen-vacancy (NV) decoherence contrast at (10^{-3}) to (10^{-1}) and the scanning tunneling microscopy (STM) current-noise contrast at (10^{-3}) to (10^{-2}). The paper presents calculations and a proposed protocol, not measured bond contrasts.

The scaling relation is (C \propto b^2/E_{\mathrm{bond}}), where (b) is bond order and (E_{\mathrm{bond}}) is bond energy. The (b^2) term comes from the bonding-antibonding matrix element, while the energy denominator contributes (1/E_{\mathrm{bond}}). Ethane supplies (b=1) and 3.6 eV, benzene supplies (b=1.5) and 4.5 eV, and ethylene supplies (b=2) and 6.3 eV. These inputs first give (2.28:1.80:1). A 10% delocalization correction for the aromatic bond gives (2.28:1.60:1), which rounds to the proposed (2.3:1.6:1). The author presents this result as parameter-free because the bond orders and energies come from molecular structure.

The recommended calculation uses PBE0, the def2-TZVP basis, and Casida linear response for the 50 lowest singlet excitations. The proposed frequency grid covers 0 to 10 eV at 0.05 eV resolution. The bond volume is a 1.4 Å cube at the bond midpoint, while the reference volume is centered 3 Å above the molecular plane. The paper states that PBE0, B3LYP, and LDA change the absolute contrast while preserving the ratios within about 5%. No experimental ratio is reported for comparison.

The GFET calculation serves as a negative platform comparison. Its predicted normalized conductance change is about (10^{-12}), seven to nine orders below the quoted experimental sensitivity. Table III gives a broader (10^{-12}) to (10^{-10}) range, but the paper rejects capacitive detection in both cases.

The TDDFT treatment starts from the Kohn-Sham response. A Dyson equation adds the Coulomb interaction and exchange-correlation kernel to obtain the interacting response. The local density fluctuation spectrum is written as (S_n(\mathbf r,\omega)=S_0(\omega)+\beta(\omega)n(\mathbf r)). The bond-region contribution comes mainly from (\sigma\rightarrow\sigma^*) transitions localized on the molecule. A local expansion is used when one transition length scale dominates and the electron density varies slowly over that scale.

The fluctuation-dissipation theorem converts the bond polarizability into a charge fluctuation spectrum. The (f)-sum rule and a single pole at (\omega_0=E_{\mathrm{bond}}/\hbar) determine the estimate. At 4.2 K and (\hbar\omega_0 \approx 5) eV, the paper sets (\coth(\hbar\omega_0/2k_BT) \approx 1). A single-pole model with (N_{\mathrm{eff}} \approx 2) for a two-electron bond gives (\delta q \approx 10^{-2}e). Across bond energies of 3.6 to 6.3 eV, the estimate varies by less than 30% because (\delta q \propto E_{\mathrm{bond}}^{-1/2}). The static-polarizability check gives about 1 ų, consistent with the paper’s reference to known bond polarizabilities. This charge fluctuation enters both sensor estimates.

Shallow nitrogen-vacancy centers at 1 to 5 nm depth in (^{12}\mathrm{C})-enriched diamond provide the first sensing route. The protocol specifies enrichment above 99.99%, optical initialization, and Hahn echo or CPMG sequences at 4.2 K. The spin is read out through optically detected magnetic resonance (ODMR). The NV ground state has spin 1 and a zero-field splitting of 2.87 GHz. The decoherence rate (\Gamma=1/T_2) responds to electric-field noise through a Stark response of about 35 Hz/(V/cm). The modeled electronic frequency near (10^{15}) Hz reaches the readout through nonlinear downconversion to kHz-MHz bands.

At distances above 2 nm, the leading noise term is dipolar: (S_E(\omega) \propto \langle\delta q^2\rangle a_0^2/r^6). The first quadrupolar correction is suppressed by ((a_0/r)^2 \approx 5\times10^{-3}). The paper therefore states that the dipolar approximation is accurate to better than 0.5% for (r \gtrsim 2) nm. The depth test uses (\Delta\Gamma \propto 1/(d+h)^6), with an adsorption height (h=0.35) nm. Centers at 2, 3, and 5 nm should reproduce that law.

ODMR reaches a relative sensitivity of about (10^{-4}) with 100 s averaging. The predicted contrast is (10^{-3}) to (10^{-1}), corresponding to a stated signal-to-noise range of 10 to (10^3). The parameter table gives an intrinsic decoherence reference of (\Gamma_0 \approx 1) kHz.

STM noise spectroscopy accesses a different observable from conventional STM imaging. Conventional STM measures mean local density of states (LDOS) through average tunneling current, while the proposal measures its quantum variance through (S_I(f)). By the Wiener-Khinchin theorem, (S_I(f)) is the Fourier transform of the current autocorrelation and probes (\langle\delta\rho,\delta\rho\rangle). The proposed signal is therefore a two-point electronic correlator rather than another mean LDOS map.

The protocol uses an STM under ultra-high vacuum at 4 K, with molecules on graphene or HOPG, or directly on diamond. It samples noise from 100 Hz to 100 kHz. The estimate assumes (z \approx 5) Å, a 100 kHz bandwidth, and 100 s averaging. The quoted STM noise resolution is (10^{-4}) to (10^{-3}), within the predicted (10^{-3}) to (10^{-2}) noise excess. A two-dimensional noise map would test whether excess current noise follows bond positions.

The paper labels the table’s normalized values as phenomenological signal estimates.

MoleculeBond and energyEstimated (\Delta\Gamma/\Gamma_0)Estimated (\Delta S_I/S_{I,0})
EthaneC-C, 3.6 eV(5\times10^{-3})(2\times10^{-3})
BenzeneC-C arom., 4.5 eV(1\times10^{-2})(5\times10^{-3})
EthyleneC=C, 6.3 eV(3\times10^{-2})(8\times10^{-3})

The TDDFT comparison instead gives relative contrasts of 1.0 for ethane, 1.6 for benzene, and 2.3 for ethylene.

The measurement geometry includes a bond center, a ring center or other non-bonding region, and a bare substrate. Several controls separate the proposed signal from known backgrounds.

  • Isotopic substitution compares (\mathrm{C_2H_6}) with (\mathrm{C_2D_6}). The vibrational frequency ratio is about 1.36, but the electronic signal should remain invariant.
  • Low-temperature tests require a constant signal below (\hbar\omega_0/k_B), while thermal noise should scale with (T).
  • Spatial mapping requires the excess to follow a translated molecule rather than a substrate defect.
  • NV depth measurements test the (1/(d+h)^6) law, and cross-platform measurements must agree.
  • Spectral tests require bond-linked structure instead of featureless (1/f^\alpha) surface noise, with (\alpha \approx 0.8) to 1.2.
  • TDDFT calculations must reproduce the measured contrast within the stated experimental and theoretical uncertainties.

The null criterion requires both (\Delta\Gamma/\Gamma_0) and (\Delta S_I/S_{I,0}) below (10^{-4}) across all molecules. The ratio test rejects the model if departures exceed combined experimental uncertainty of 15% to 20% and theoretical uncertainty of 5%.

The paper reports no measured bond contrasts, noise spectra, or NV depth series. Every platform signal is a model estimate. The local approximation requires (r \gg a_0) and depends on the assumed transition length scale. Vertex corrections may change absolute contrast values, although the paper assumes that they preserve the predicted hierarchy. Weak physisorption is also assumed, so strongly chemisorbed molecules require separate TDDFT treatment.

Surface charge noise remains the main experimental confound for shallow NVs. The proposed spectral, spatial, and isotopic filters reduce this risk but do not remove it. The GFET value ranges from (10^{-12}) in the text to (10^{-12}) to (10^{-10}) in Table III. This range does not affect the paper’s selection of NV and STM measurement platforms.

NV sensing requires a diamond host with shallow centers, high isotopic enrichment, controlled surfaces, and operation at 4.2 K. The paper’s surface-noise concern requires controls for charge traps and two-level systems. NV centers offer the higher sensitivity, while STM provides atomic-scale spatial localization. They therefore serve complementary roles in the proposed test. The depth series gives a direct material and implantation check because a valid dipolar signal must follow (1/(d+h)^6). Until the predicted ratios are measured, these results should guide test-sample design rather than define a production material specification.

For 6CCVD, the diamond materials catalog lists single-crystal and polycrystalline diamond grown by MPCVD, including plates and wafers up to 125 mm. The NV application calls for a (^{12}\mathrm{C})-enriched host with shallow centers at 1 to 5 nm, which are not among the listed specifications. Those device-specific requirements need separate definition through engineering support.

View Original Abstract

I propose a dual-platform experimental approach to probe quantum electromagnetic fluctuationsat the scale of individual chemical bonds. Instead of invoking unobservable virtual particles, I for-mulate the problem in terms of measurable correlation functions of charge and current densities.I show that covalent bonds are expected to exhibit enhanced local quantum fluctuations relativeto non-bonding regions due to increased electronic overlap. I demonstrate that this effect is unde-tectable using conventional graphene-based capacitive sensors (∆SG/SG,0 ∼ 10−12), but becomesobservable using two existing high-sensitivity platforms: nitrogen-vacancy (NV) centers in diamondand scanning tunneling microscopy (STM) noise spectroscopy. Order-of-magnitude estimates yield∆Γ/Γ0 ∼ 10−3-10−1for NV decoherence and ∆SI /SI,0 ∼ 10−3-10−2for STM current noise,both within current experimental reach. A first-principles connection to time-dependent densityfunctional theory (TDDFT) is established via the interacting density response function, enablingdirect computation of the observable contrast C(ω). I derive a universal, parameter-free scalingprediction—the primary falsifiable result of this work—for the contrast ratios across bond types:C(C=C) : C(C-C arom.) : C(C-C) ≈ 2.3 : 1.6 : 1. I provide falsifiable predictions including isotopicinvariance (H/D substitution), temperature independence at low T, 1/r6 depth scaling for NV cen-ters, cross-platform consistency, and quantitative TDDFT benchmarking. A positive result wouldestablish a new spectroscopic modality based on quantum noise and provide direct experimentalaccess to spatially resolved electronic correlations in chemical bonds.