Skip to content

Fidelity and quantum geometry approach to Dirac exceptional points in diamond nitrogen-vacancy centers

FieldValue
JournalKent Academic Repository (University of Kent)
Published2026-01-31
AuthorsChia-Yi Ju, Gunnar Möller, Yu-Chin Tzeng
DOI10.48550/arxiv.2602.00666

The real part of the fidelity susceptibility diverges to negative infinity along the non-reciprocal coupling axis of a Dirac exceptional point in a diamond nitrogen-vacancy center. The point lies at (q1, q2) = (0, 1), where two eigenvalues coalesce at E0 = 3. Along the effective momentum axis, the susceptibility vanishes exactly in the studied model. The authors derive these results numerically and analytically for an effective Hamiltonian. This directional response identifies a geometric singularity without a parity-time symmetry-breaking transition.

Directional response at the Dirac exceptional point

Section titled “Directional response at the Dirac exceptional point”

The Dirac point lies in the parity-time unbroken region, where the eigenvalues stay purely real. Two branches meet with linear dispersion around (0, 1). At the point, the Hamiltonian is defective and supports a rank-2 Jordan block. Conventional exceptional points usually mark the boundary of parity-time symmetry breaking. Increasing q2 eventually crosses a critical threshold and produces conventional exceptional points and complex conjugate eigenvalues. The authors therefore focus on the real-spectrum regime near the Dirac point.

Effective Hamiltonian and synthetic parameter space

Section titled “Effective Hamiltonian and synthetic parameter space”

The modeled object is one nitrogen-vacancy center, treated as a spin-1 electronic qutrit. Its effective non-Hermitian Hamiltonian has two synthetic control parameters. The parameter q1 acts as effective momentum, while q2 measures the degree of non-reciprocal coupling. The paper treats q1 and q2 as synthetic parameters and reports no physical units for them. The non-Hermitian term is sqrt(2)(Sx - i q2 Sy), and q1 enters only through the diagonal term 2q1 Sz.

The 3 × 3 effective Hamiltonian is obtained by quantum dilation from a 6 × 6 Hermitian Hamiltonian. The electronic qutrit couples to an ancillary nuclear spin. Conditional measurements and post-selection on the ancilla produce the effective non-Hermitian evolution. This implementation connects the analysis to the NV-center platform used in prior experimental work.

Fidelity is defined as the complex-valued product of biorthogonal overlaps between states at neighboring parameter values. Its expansion uses a unit displacement vector n, and the fidelity susceptibility is the second-order coefficient along that direction. The measure is therefore directional by construction. Within the parity-time unbroken phase, the fidelity is purely real, so the susceptibility is also real. The calculations examine the n = 0 eigenstate.

A parameter-space map identifies the Dirac singularity at (0, 1) and a separate exceptional line from conventional exceptional points. Along q1 = 0, with displacement n = (0, 1), the real susceptibility diverges to negative infinity at both types of singularity. The authors interpret this as extending the cited fidelity-susceptibility signature beyond the parity-time transition.

A finite-step test used delta q = 10^-2. When q2 < qEP < q2 + delta q across a conventional exceptional line, the real fidelity approaches 1/2. The value matches the reference limit reported for a transition between unbroken and broken phases.

Angular tests separate the two response directions. The authors set q1 = r cos phi and q2 = 1 + r sin phi, then compared states at r and r minus delta r. The calculations used r = 0.1, 0.2, and 0.3. Strong negative divergences occurred at phi = pi/2 and 3pi/2, which follow the q2 axis. The susceptibility was numerically zero at phi = 0 and pi, which follow the q1 axis. The divergent term is associated with an r^-2 contribution, while the dark-axis response is suppressed to numerical precision.

The anisotropy follows from the defective eigenstate structure rather than a vanishing energy gap. Near the Dirac point, one eigenstate branch remains geometrically inert to linear order. The other branch acquires a linear admixture of the generalized Jordan vector |chi>. Its coefficient A(phi) measures the projection of the perturbation onto that defective direction. Since the eigenstate splitting remains linear and nonzero in every real direction, the directional response comes from the eigenstate deformation.

The perturbation is delta H = r(cos phi partial q1 H + sin phi partial q2 H). Because q1 enters only through a diagonal operator, partial q1 H is diagonal. The appendix uses evolution generators to prove exact susceptibility vanishing along the line q2 = 1. Thus the numerical zero along q1 is not only a leading-order effect.

The proof gives a broader condition. If the parameter derivatives collectively span the off-diagonal operator space, at least one dark direction must exist. Opposite vectors define the same line through the Dirac point. The number and orientation of dark lines remain model dependent. The paper treats this dark-line result as generic under the stated spanning condition.

Quantum geometry and conventional exceptional points

Section titled “Quantum geometry and conventional exceptional points”

The fidelity susceptibility is a directional projection of the quantum geometric tensor’s quadratic form. In the parity-time unbroken phase, the eigenvectors are real. The symmetric, real, and real-and-symmetric definitions of the quantum metric therefore coincide. The antisymmetric component vanishes throughout this phase, including near the Dirac point. The singular response is consequently carried by the quantum metric rather than Berry curvature.

This behavior separates Dirac and conventional exceptional points. A conventional exceptional point has square-root energy behavior and an effectively isotropic susceptibility divergence. The Dirac point has linear energy dispersion and a rank-one leading geometry associated with one Jordan direction. Perturbations perpendicular to that direction do not cause first-order eigenstate deformation.

The paper also checks the cited conventional-point result within the same model. The real fidelity tends to 1/2 across the exceptional line. Earlier lemmas tied that limit and the negative susceptibility divergence to a broken-unbroken transition. The Dirac point satisfies neither condition, yet it still produces the negative divergence.

The analysis is theoretical and numerical. It does not report a new measurement of fidelity susceptibility. The authors state that the diagnostics are accessible in principle through conditional dynamics. For NV centers, conditional measurements and post-selection can provide biorthogonal overlaps. Repeated state preparation and measurement can infer the required overlaps on other post-selected platforms.

Trapped-ion simulators are proposed as another route. A three-level internal manifold could encode the qutrit, and tunable coherent couplings could reproduce the spin-1 structure. Ancilla-assisted dilation or state-selective dissipation could engineer the non-Hermitian dynamics. The paper states that a direct trapped-ion realization of this Dirac exceptional point remains to be demonstrated.

Several limits remain. At the exact exceptional point, coalesced states make the biorthogonal basis incomplete, so the study evaluates the approach from a real, nondegenerate regime. A metric operator would be required for a rigorous treatment at the defective point. Fidelity also has several non-Hermitian definitions. The selected biorthogonal-overlap definition resolves metric ambiguity within the unbroken phase, but it remains a specific choice.

The engineering implication is direct. Parameter variations should be aligned with q2 when seeking the largest geometric response. A sweep along q1 can miss the Dirac signature in the fidelity diagnostic despite approaching its energy coalescence. The work does not quantify a sensing enhancement or noise floor. It also leaves the broader quantum-metric effect across non-Hermitian phase diagrams for future study.

The modeled device requires a diamond host containing a single nitrogen-vacancy center. 6CCVD’s diamond materials catalog includes single-crystal and polycrystalline diamond grown by MPCVD, but NV-center creation is not among the stated product facts. The linked engineering support can be consulted for custom fabrication requirements around the required host material.

View Original Abstract

Dirac exceptional points (EPs) represent a novel class of non-Hermitian singularities that, unlike conventional EPs, reside entirely within the parity-time unbroken phase and exhibit linear energy dispersion. Here, we theoretically investigate the quantum geometry of Dirac EPs realized in nitrogen-vacancy centers in diamond, utilizing fidelity susceptibility as a probe. We demonstrate that despite the absence of a symmetry-breaking phase transition, the Dirac EP induces a pronounced geometric singularity, confirming the validity of the fidelity in characterizing non-Hermitian EPs. Specifically, the real part of the fidelity susceptibility diverges to negative infinity, which serves as a signature of non-Hermitian criticality. Crucially, however, we reveal that this divergence exhibits a distinct anisotropy, diverging along the non-reciprocal coupling direction while remaining finite along the detuning axis. Furthermore, we establish that this anisotropy, characterized by at least one exact dark direction coexisting with divergent directions, is a generic consequence of the Dirac EP structure whenever the parameter derivatives collectively span the off-diagonal operator space at the Dirac EP. This behavior stands in stark contrast to the omnidirectional divergence observed in conventional EPs. Our findings provide a comprehensive picture of the fidelity probe near the Dirac EP, highlighting the critical role of parameter directionality in exploiting Dirac EPs for quantum control and sensing applications.