Bidirectional Nonlinear Optical Tomography Characterizes Coupling Efficiencies, Separating and for Unbiased On-Chip Performance Evaluation

Evaluating the performance of advanced photonic circuits presents a significant challenge, as accurately determining how efficiently light enters and exits these devices is crucial for reliable operation and scaling. Bo-Han Wu, Mahmoud Jalali Mehrabad, and Dirk Englund from the Massachusetts Institute of Technology address this problem with a new technique called bidirectional nonlinear optical tomography. This method overcomes the limitations of existing approaches by independently measuring both input and output coupling efficiencies, eliminating a common source of error that can lead to inaccurate assessments of on-chip performance. By linking off-chip measurements to internal circuit properties, and accounting for real-world noise, the team demonstrates unbiased and reproducible characterisation of coupling efficiencies, paving the way for more reliable benchmarking and improved design of scalable photonic systems for applications ranging from optical communications to precision sensing.

Chip-Scale Nonlinear Optics and Quantum Photonics

Research in integrated photonics is rapidly advancing, with a focus on controlling and optimizing nonlinear optical processes, such as Second Harmonic Generation and squeezed light generation, within miniaturized devices. This work is crucial for developing new technologies in quantum communication, optical sensing, and advanced optical systems. The core aim is to create high-performance, stable, and scalable photonic integrated circuits (PICs) that harness the power of light. Photonic Integrated Circuits (PICs) are central to this field, analogous to electronic circuits but utilizing light instead of electrons.

They offer the potential for miniaturization, integration, and mass production. Nonlinear optics forms the foundation for many advanced photonic applications, enabling frequency conversion and the creation of non-classical states of light. Squeezed light, a state with reduced noise in specific properties, is particularly important for quantum technologies. Researchers are developing methods to map the efficiency of nonlinear processes, measure squeezing levels, and stabilize PIC performance. Quantum technologies are a major driving force behind this research, as squeezed light and other quantum states of light are essential for quantum communication, computing, and sensing.

This research has potential applications in a wide range of fields, including secure quantum communication, building quantum computers, developing highly sensitive sensors, and improving optical metrology. It also promises advancements in biophotonics for biomedical imaging and sensing, and increasing the capacity and efficiency of optical communication networks. The field of integrated photonics is evolving rapidly, with immense potential for future innovation.

Bidirectional Tomography Separates Chip Coupling Efficiencies

Scientists have developed bidirectional nonlinear optical tomography (BNOT), a new metrology technique for accurately assessing nonlinear photonic integrated circuits. Conventional methods only measure the combined effect of input and output coupling efficiencies, obscuring individual losses and introducing bias when evaluating on-chip performance. BNOT overcomes this limitation by employing both forward and backward probing with complementary nonlinear signals, enabling the separation of these efficiencies with high confidence. The technique links off-chip measurements to on-chip quantities through a compact observation model that explicitly accounts for fluctuations in the laser and noise within the detectors.

This model frames the extraction of efficiency values as a constrained optimization problem, allowing for a precise determination of individual coupling efficiencies. Monte Carlo simulations demonstrate that BNOT converges to accurate efficiency values across realistic operating conditions. Researchers used these efficiency estimates to reconstruct on-chip nonlinear figures of merit, achieving distributions centered on the true values with reduced variance, a significant improvement over the biased results obtained from conventional methods. BNOT is compatible with existing hardware and is platform-agnostic, providing unbiased characterization of coupling efficiencies for a range of nonlinear processes, enabling reproducible benchmarking for scalable systems in quantum optics, frequency conversion, and precision metrology.

Bidirectional Tomography Maps Circuit Coupling Efficiencies

Scientists have developed bidirectional nonlinear optical tomography (BNOT), a new method for accurately characterizing photonic integrated circuits. Conventional calibration methods struggle to independently determine input and output coupling efficiencies, leading to systematic errors in performance assessment. BNOT overcomes this limitation by utilizing forward and backward probing with complementary nonlinear signals, enabling the direct estimation of individual interface efficiencies with significantly improved confidence. The research demonstrates that conventional linear transmission calibration only recovers the product of input and output efficiencies, while BNOT breaks this limitation by linking off-chip measurements to on-chip quantities through a compact observation model.

This model explicitly accounts for fluctuations in laser power and detector noise, framing efficiency extraction as a constrained optimization problem. Monte Carlo studies confirm that BNOT converges to accurate values with low error across realistic operating conditions, providing unbiased estimates of interface efficiencies. Experiments using periodically poled lithium niobate waveguides show that BNOT accurately estimates interface efficiencies, crucial for benchmarking nonlinear processes like squeezed light generation and second harmonic generation. Specifically, the team measured squeezing levels exceeding 4.

9 dB and achieved 3. 1 dB in wafer-scale integration, approaching the threshold required for advanced applications like quantum computing and gravitational wave detection. The BNOT method delivers narrower confidence intervals for efficiency estimates compared to conventional methods, enabling reproducible and coupling-resolved benchmarking for scalable optical systems.

Bidirectional Tomography Maps On-Chip Efficiency

The team developed bidirectional nonlinear optical tomography (BNOT), a new method for accurately evaluating nonlinear photonic integrated circuits. Conventional calibration techniques determine only the product of input and output coupling efficiencies, introducing systematic errors when assessing on-chip performance. BNOT overcomes this limitation by employing forward and backward probing with complementary nonlinear signals, allowing for the separate and precise estimation of each interface efficiency. This approach links off-chip measurements to on-chip quantities by explicitly accounting for fluctuations and noise, framing efficiency extraction as a constrained optimization problem.

Extensive Monte Carlo simulations demonstrate that BNOT converges to accurate efficiency values with minimal error across realistic operating conditions. Applying these refined efficiency estimates to reconstruct on-chip nonlinear figures of merit yields distributions centered on true values with reduced variance, a significant improvement over the biased results produced by conventional methods. BNOT is compatible with existing hardware and is platform-agnostic, offering unbiased characterization of coupling efficiencies for a range of nonlinear processes and enabling reproducible benchmarking for scalable optical systems. Future research directions include integrating BNOT into adaptive control systems and self-calibrating architectures, potentially automating coupling-loss compensation and dynamically tracking degradation in large photonic arrays, ultimately supporting the transition of nonlinear photonic integrated circuits from laboratory demonstrations to scalable technologies.