Find a frequency by counting light
An NV center is a small defect in the diamond lattice: a nitrogen atom replaces a carbon atom beside a vacancy. Its electron spin responds to a magnetic field. Green excitation produces red fluorescence, and collection optics send part of that light to a photon detector. Each detection contributes a count.
Green light also favors the more fluorescent spin state. When a microwave frequency matches a spin transition, some population moves into a less fluorescent state and the count falls. A sweep turns this change into a dip. Light supplies a practical way to locate a spin energy gap.
Separation and common motion carry different information
Near room temperature, the ground-state spin levels already have a separation of about 2.87 gigahertz at zero applied field. An axial magnetic field separates the two nearby microwave transitions: one moves down in frequency and the other moves up. Their separation measures an axial field projection in the small-field limit.
Temperature moves their shared center. Near room temperature, a one-kelvin increase reduces the zero-field splitting by roughly 74 kilohertz. Following only one resonance can mix a thermal shift with a magnetic signal. Measuring both branches supplies a separation and a midpoint, which answer different questions.
The spectrum has finer structure as well. The common nitrogen-14 nucleus has spin, so an electronic transition can have three hyperfine lines. Interpreting several neighboring dips requires a model of those nuclear populations and line shapes before counting them as distinct electronic transitions.
Four crystal axes add directional information
One NV axis measures a projection, much as a shadow records only part of a spatial arrow. Different three-dimensional fields can give that same projection. Diamond permits four tetrahedral NV orientations, so one field gives several different projections within one crystal.
Known crystal geometry links these projections. Yet measured frequencies do not arrive with an axis label, and an unsigned splitting does not directly reveal a projection’s sign. Sorting eight resonance groups alone does not always select a unique field vector.
An independently known bias field provides a useful starting point. Its measured spectrum establishes which resolved groups to follow. A bounded change around that field can stay on the same branch if the groups remain separate. Transverse fields also shift the levels, so accurate reconstruction compares all the resonances with eigenvalues of the complete spin matrix.
Frequency and optical references serve different purposes
Dip depth also depends on the collection angle, laser and microwave conditions. The objective collects a fraction of the emitted light, followed by filter losses and detector efficiency. More detected photons usually make a dip easier to resolve. Brightness alone does not supply its resonance frequency.
The sweep’s frequency scale needs a reference of its own. Separate frequency-counter readings test the synthesizer, a dark exposure measures detector background, and a known-field diamond measurement establishes the resonance pattern. A further known magnetic increment checks whether the chain works beyond the measurements used to set it up.
Strain can share the thermal shift
Crystal strain can also move or separate spin levels. If axial strain and temperature both move a resonance pair together, the same midpoint shift has two physical explanations. Stable-strain evidence, an independent temperature measurement or additional observables are needed to distinguish them. A magnetic result carries the physical conditions that made its interpretation possible.