Chromatic Perturbation in Quantum Tomography

• Chromatic perturbation module is a framework for modeling chromatic aberration in polarization qubit tomography by explicitly accounting for wavelength-dependent dispersion in wave plates.
• It constructs spectral-averaged measurement operators (POVMs) that mitigate systematic errors through integration over finite spectral bandwidth and parasitic effects.
• The method enhances quantum state estimation by restoring the correct 1/N fidelity scaling and enabling robust, high-order tomography under broadband illumination.

The chromatic perturbation module is a quantitative and algorithmic framework for modeling the effects of chromatic aberration in optical polarization qubit tomography when using birefringent wave plates, such as half-wave plates (HWPs) and quarter-wave plates (QWPs), in the presence of finite spectral bandwidth. By explicitly accounting for wavelength-dependent dispersion and parasitic effects in wave plates, this approach constructs a physically adequate model of quantum state measurement, yielding more accurate reconstructions than standard projective models that neglect such imperfections. The chromatic perturbation module can suppress systematic errors and enables high-fidelity, unbiased tomography even in setups utilizing high-order plates or broadband illumination (Bantysh et al., 2020).

1. Theoretical Model of Chromatic Aberration in Wave-Plate Tomography

Chromatic aberration in quantum state measurement arises from the dispersive properties of birefringent crystals and the resultant wavelength dependence of wave-plate retardance. For a given wavelength $\lambda$, a sequence of HWP and QWP at angles %%%%1%%%% implements the transformation

$$U(\alpha, \beta; \lambda) = U_{WP}(\delta_{QWP}(\lambda), \beta)\;U_{WP}(\delta_{HWP}(\lambda), \alpha),$$

where $U_{WP}(\delta, \alpha)$ denotes the Jones matrix for a wave plate of phase delay $\delta$ at angle $\alpha$,

$$U_{WP}(\delta, \alpha)=\begin{pmatrix} \cos\delta - i\sin\delta\cos2\alpha & -i\sin\delta\sin2\alpha \ -i\sin\delta\sin2\alpha & \cos\delta + i\sin\delta\cos2\alpha \end{pmatrix},$$

with the phase delay

$$\delta_X(\lambda)=\frac{\pi h_X[n_o(\lambda)-n_e(\lambda)]}{\lambda},$$

where $h_X$ is the plate thickness, and $n_o(\lambda)$, $n_e(\lambda)$ the ordinary and extraordinary indices, respectively, for $X\in\{\mathrm{HWP},\mathrm{QWP}\}$.

In polychromatic or broadband scenarios, each wavelength component undergoes a slightly different transformation. The effective quantum channel becomes a spectral average:

$$\rho \mapsto \int d\lambda\,P(\lambda)\;U(\alpha,\beta;\lambda)\,\rho\,U(\alpha,\beta;\lambda)^{\dagger},$$

with $P(\lambda)$ the normalized spectral distribution.

2. Spectral-Averaged POVMs and the Fuzzy Quantum Measurement Model

Chromatic perturbations require recalculating the effective measurement operators. In the monochromatic limit, each setting corresponds to projectors $P_j=|j\rangle\!\langle j|$ ($j=0,1$) following the transformation. Under chromatic spread, the POVM elements become:

$$\Lambda_j(\alpha, \beta) = \int d\lambda\, P(\lambda)\, U(\alpha, \beta;\lambda)^{\dagger} P_j U(\alpha, \beta;\lambda),$$

for $j=0,1$. Each $\Lambda_j$ is full rank and collectively they satisfy $\Lambda_0 + \Lambda_1 = \mathbb{I}$. The rank and structure of $\Lambda_j$ encode the degree of chromatic "fuzziness" in the measurement—ideal projectors become spectrally-averaged effects whose detailed form is determined by crystal dispersion and spectral bandwidth.

3. Information-Theoretic Analysis: Fisher Matrix and State Estimation Fidelity

The inferential power of chromatic-perturbed tomography is captured via the classical Fisher information matrix, with density matrix $\rho$ parametrized as $\theta_i$:

$$[I_F]_{ik}(\rho) = N\sum_{j=0}^1 \frac{1}{p_j(\rho)}\frac{\partial p_j(\rho)}{\partial\theta_i}\frac{\partial p_j(\rho)}{\partial\theta_k},$$

where $$p_j(\rho) = \operatorname{Tr}[\rho\,\Lambda_j(\alpha, \beta)]$$ and $N$ is the total sample size.

An alternative, "root" approach forms the real-symmetric information matrix

$$H=2 \sum_{j,m} n_{jm} \frac{(\Lambda_{jm} c)(\Lambda_{jm} c)^{T}}{p_{jm}},$$

where $n_{jm}$ is the count at measurement setting $(j,m)$, $c$ the vectorization of the purified amplitudes, and $p_{jm} = \operatorname{Tr}[\rho\,\Lambda_{jm}]$. The nonzero eigenvalues $h_k$ quantify Fisher information directions, scaling linearly with $N$ and encoding bandwidth dependence via $\Lambda_{jm}(\Delta\lambda)$.

Asymptotically, the purified-state infidelity,

$$1 - F = \sum_{k=1}^{\nu_P} d_k\,\xi_k^2, \quad \xi_k \sim \mathcal{N}(0,1), \quad d_k = \frac{1}{2h_k},$$

yields the mean infidelity

$$\langle 1 - F \rangle = \sum_k d_k = \frac{1}{2} \sum_k \frac{1}{h_k(\Delta\lambda)} \sim \frac{L(\Delta\lambda)}{N},$$

and average fidelity

$$F(N, \Delta\lambda) = 1 - \frac{L(\Delta\lambda)}{N},$$

with $L(\Delta\lambda)$ the loss function encoding bandwidth-induced information loss.

4. Chromatic-Perturbation Tomography Algorithm

The chromatic-perturbation module comprises a four-step algorithmic pipeline:

1. Calibration of Dispersion: Determine refractive indices $n_o(\lambda), n_e(\lambda)$ from empirical data or literature (e.g., for quartz) across the spectral interval. Compute $\delta_{HWP}(\lambda)$ and $\delta_{QWP}(\lambda)$ for given physical thicknesses and plate orders.
2. Spectral-Averaged POVM Construction: For each setting $(\alpha_m, \beta_m)$, discretize $\lambda$ over $$[\lambda_0-\Delta\lambda/2,\lambda_0+\Delta\lambda/2]$$ (central wavelength $\lambda_0$, bandwidth $\Delta\lambda$, e.g., 650 nm and up to 0.02 μm, with $M\approx 50$–200 spectral points). Compute $U(\alpha_m, \beta_m, \lambda_\ell)$ for each $\lambda_\ell$, and assemble

$$\Lambda_{jm} = \sum_{\ell=1}^M P(\lambda_\ell) U(\alpha_m, \beta_m, \lambda_\ell)^{\dagger} P_j U(\alpha_m, \beta_m, \lambda_\ell) \frac{\Delta\lambda}{M}.$$

1. Measurement and Data Acquisition: For each setting $m$, direct $N_m$ photons through the apparatus, recording counts $k_{0m}, k_{1m}$, with total $N = \sum_m N_m$.
2. State Reconstruction:
   • Linear Inversion (if measurement matrix $B$ is full rank): Solve $\vec{p} = B\,\text{vec}(\rho)$, yielding $\hat{\rho} = B^+\,\vec{p}$, with $\vec{p}_{jm} = k_{jm}/N_m$.
   • Maximum-Likelihood (Root Approach):
     1. Initialize a purified state $\psi_0$.
     2. Iterate: $I\,\psi_{n+1} = J(\psi_n)\,\psi_n$, where $I = \sum_{jm} N_m \Lambda_{jm}$, $$J(\psi) = \sum_{jm} \frac{k_{jm}}{p_{jm}(\psi)} \Lambda_{jm}$$.
     3. After convergence ($\sim$10–20 iterations), reconstruct $\hat{\rho} = \psi_\infty \psi_\infty^{\dagger}$.

By following all steps, chromatic aberration is rigorously modeled in quantum state reconstructions (Bantysh et al., 2020).

5. Quantitative Impact Versus Standard Projective Tomography

The standard projective measurement model, which neglects chromatic averaging,

$$\Lambda_{jm}^{\text{ideal}} = U(\alpha_m, \beta_m)^{\dagger} P_j U(\alpha_m, \beta_m),$$

exhibits significant systematic bias under realistic conditions. In the cube protocol (three settings), the ideal (zero bandwidth) theoretical loss $L$ lies in $[1.0, 1.125]$, while at $\Delta\lambda = 0.01\,\mu m$ the chromatic-perturbed (“fuzzy’’) model predicts $L \in [1.216, 1.857]$. The chromatic model preserves the correct asymptotic $1/N$ scaling of mean infidelity and yields a proper $\chi^2$ distribution (1 d.o.f.), whereas the standard projective case saturates—often achieving only $\approx$99.58% fidelity as $N \to \infty$ for $\Delta\lambda=0.01 \mu m$—and manifests unmodeled systematics in the fit quality.

6. Experimental Parameters, Resource Requirements, and Protocols

The module's implementation depends on precise empirical and design parameters:

• Crystal Dispersion: Empirical values per Ghosh (Opt. Commun. 163, 95, 1999).
• Wavelength and Bandwidth: Central $\lambda_0 = 650\,\mathrm{nm}$; uniform $\Delta\lambda$ up to $0.02\,\mu m$.
• Wave Plate Specifications: Orders $k_{HWP}=k_{QWP}=10$; thicknesses $h_{HWP}=756\,\mu m$, $h_{QWP}=738\,\mu m$ (typical for high-order plates).
• Discretization: Spectral resolution $M \approx 50$–200.
• Protocols: Cube ($l=3$) and octahedron ($l=4$) configurations for tomography.
• Sample Size: $N$ per experiment $=10^2$–$10^5$.
• Computational Cost: Forming $\Lambda_{jm}$ is $O(M)$ per setting; assembling information matrix $H$ is $O((2s)^3 l)$; ML root-iterations converge within $\sim$10–20 steps.

All components are thus directly informed by and parameterized through the physical and experimental conditions of the optical tomographic setup.

7. Significance and Practical Implications

Employing the chromatic perturbation module allows substantial mitigation of systematic reconstruction errors in polarization qubit tomography, particularly relevant for experiments utilizing high-order wave plates and sources with non-negligible spectral bandwidth. By modeling the effect of parasitic dispersion and chromatic spread in the basis-change transformation, the module restores informative Fisher metrics and provides unbiased fidelity estimates, supporting robust quantum state estimation in practical, non-ideal optical systems. The framework is extensible to various tomographic protocols and is implementable with moderate computational overhead for resource ranges commonly encountered in quantum optics experiments (Bantysh et al., 2020).