Skip to main content

Measuring Populations and Coherence Phases

Choose a seven-level quantum system, a reproducible preparation and a basis set by the apparatus. The state is a trace-one Hermitian matrix Γ\Gamma. Its seven populations sum to one, and each of the 21 off-diagonal pairs contains one complex number and its conjugate. A full-rank state has 48 independent real coordinates. Which of them can be recovered depends on the physical tests available on repeated preparations. Coherence matrix

What an entry and a phase measure​

γij=∣γij∣eiθij,γji=γij‾,Gap(i,j)=∣Im⁡γij∣∣γij∣=∣sin⁡θij∣(γij≠0).\gamma_{ij}=|\gamma_{ij}|e^{i\theta_{ij}},\qquad \gamma_{ji}=\overline{\gamma_{ij}},\qquad \mathrm{Gap}(i,j)=\frac{|\operatorname{Im}\gamma_{ij}|}{|\gamma_{ij}|} =|\sin\theta_{ij}|\quad(\gamma_{ij}\ne0).

At zero coherence the phase is undefined. Under ∣i⟩↦eiαi∣i⟩|i\rangle\mapsto e^{i\alpha_i}|i\rangle the angle changes by αj−αi\alpha_j-\alpha_i. One nonzero pair can therefore acquire any Gap value through a change of phase reference. With a calibrated interference probe one can measure a quadrature relative to that reference. Around a closed cycle the phase changes cancel, so the argument of the product of its nonzero coherences is independent of those local choices. For the complete graph of seven axes there are 21−7+1=1521-7+1=15 independent cycle angles. Seven Fano triangles measure seven of them. At a vanished edge its phase drops out. Cycle observations · Reconstruction G5b

Probability current​

For unitary dynamics generated by a specified Hamiltonian, the diagonal of Γ˙=−i[H,Γ]/ℏ\dot\Gamma=-i[H,\Gamma]/\hbar gives

Ji←j=2ℏIm⁡(Hijγji),γ˙ii=∑j≠iJi←j,Ji←j=−Jj←i.J_{i\leftarrow j}=\frac{2}{\hbar}\operatorname{Im}(H_{ij}\gamma_{ji}), \qquad \dot\gamma_{ii}=\sum_{j\ne i}J_{i\leftarrow j}, \qquad J_{i\leftarrow j}=-J_{j\leftarrow i}.

The current is invariant under rephasing of the basis. For diagonal HH it vanishes even when individual coherence phases rotate as θij(t)=θij(0)−(Ei−Ej)t/ℏ\theta_{ij}(t)=\theta_{ij}(0)-(E_i-E_j)t/\hbar. A diagonal dephasing channel can attenuate a coherence while keeping its phase fixed.

What the Fano instrument can distinguish​

Label the axes by the nonzero binary triples i∈F23i\in\mathbb F_2^3. A Fano line consists of {i,j,i+j}\{i,j,i+j\} for distinct i,ji,j. The operator Πℓ=∑i∈ℓ∣i⟩⟨i∣\Pi_\ell=\sum_{i\in\ell}|i\rangle\langle i| projects onto those three axes. Every axis belongs to three lines, so ∑ℓΠℓ=3I\sum_\ell\Pi_\ell=3I. These incidence data specify a seven-outcome instrument at the level of its effects.

Let BℓiB_{\ell i} be the incidence matrix and take the quantum effects Eℓ=Πℓ/3E_\ell=\Pi_\ell/3. The seven Born probabilities are q=Bp/3q=Bp/3 where pi=γiip_i=\gamma_{ii}. From B⊤B=2I+11⊤B^\top B=2I+\mathbf1\mathbf1^\top one recovers p=(3B⊤q−1)/2p=(3B^\top q-\mathbf1)/2. All diagonal populations are identifiable with this instrument. Fixing an interior population vector leaves 42 real off-diagonal degrees of freedom near a full-rank state.

G5b tests a sharper distinction. Two positive definite conjugate states have equal populations and coherence magnitudes. All real instruments have identical record distributions on them, including along adaptive sequences of real operations. Adding a specified complex rank-one projector yields probabilities 13/8413/84 and 11/8411/84. This is an operational privacy statement relative to the allowed instruments. Relating the hidden difference to a felt experience would require a separately testable report or intervention. Exact states and probe · Practice and report

Which phase values are possible?​

For a selected pair, (∣i⟩+i∣j⟩)/2(|i\rangle+i|j\rangle)/\sqrt2 gives Gap(i,j)=1\mathrm{Gap}(i,j)=1. The uniform real superposition 7−1/2∑i∣i⟩7^{-1/2}\sum_i|i\rangle gives zero on every pair in the selected basis. Fano incidence alone therefore permits both extreme values. A tighter bound requires a restriction on preparations or a calibrated dynamics that enforces it.

Reconstructing the entries experimentally​

First measure in the coordinate basis to obtain pi=Γiip_i=\Gamma_{ii}. For each pair (i,j)(i,j), use two further settings on independently prepared copies. Each setting has two outcomes in the pair subspace and an additional outcome on its orthogonal complement.

SettingStates defining the two projectorsDifference of outcome probabilities
Real quadrature(∣i⟩±∣j⟩)/2(\lvert i\rangle\pm\lvert j\rangle)/\sqrt2p+−p−=2Re⁡γijp_+-p_-=2\operatorname{Re}\gamma_{ij}
Imaginary quadrature(∣i⟩∓i∣j⟩)/2(\lvert i\rangle\mp i\lvert j\rangle)/\sqrt2p+−p−=2Im⁡γijp_+-p_-=2\operatorname{Im}\gamma_{ij}

Together the six independent populations and 42 quadratures specify Γ\Gamma within the chosen representation. Each column of a data table needs its actual preparation, instrument setting and outcome counts. Predicting responses to a new setting tests the reconstruction. A subject's report, when available, is a further record with its own elicitation procedure. Its relation to any matrix coordinate requires a separately testable prediction. Observation models and identifiability · Practice and report

Cross-References​