A Physical Interpretation of Imaginary Time Delay

Technical Summary: A Physical Interpretation of Imaginary Time Delay

Problem Statement
In linear scattering systems, the scattering matrix ($S$) relates incoming waves to outgoing excitations, encoding extensive information about the system's dynamics. While the Wigner-Smith time delay ($\tau_W$) and the transmission time delay ($\tau_T$) are well-established metrics for quantifying how long a wave lingers in a system, these quantities are generally complex for non-unitary (lossy or gain) systems. The real part of the time delay, representing a temporal shift, is well-studied and linked to phenomena such as negative group delay. However, the physical interpretation of the imaginary part of the transmission time delay, $\text{Im}[\tau_T]$, has remained largely unexamined experimentally. Although theoretical work by Asano et al. [71] proposed that $\text{Im}[\tau_T]$ corresponds to a shift in the carrier frequency of a scattered pulse rather than a time shift, this prediction lacked direct experimental verification. Furthermore, the behavior of $\tau_T$ in non-unitary systems over a broad range of conditions had not been systematically tested against theoretical models.

Methodology
The authors employed a dual approach combining frequency-domain characterization and time-domain pulse scattering experiments using a 2-port microwave ring graph as the scattering system. The ring graph, constructed from coaxial cables and T-junctions, supports a series of well-isolated absorptive modes, facilitating the analysis of sub-unitary scattering.

1. Frequency-Domain Characterization: Using a Keysight N5242A network analyzer, the authors measured the scattering parameters ($S$-matrix) across a broad frequency range (10 MHz to 18 GHz). From the transmission coefficient $S_{21}$, they calculated the complex transmission time delay $\tau_T$ using the definition $\tau_T = -i \frac{\partial}{\partial\omega} \ln[\det T(\omega + i\alpha)]$, where $\alpha$ quantifies uniform loss. This provided the theoretical prediction for the imaginary part, $\text{Im}[\tau_T]$.
2. Time-Domain Experiments: The authors generated Gaussian time-domain pulses with varying bandwidths (1 MHz and 5 MHz) using an arbitrary waveform generator and transmitted them through the ring graph. The output pulses were captured by a high-bandwidth real-time digital sampling oscilloscope.
3. Data Analysis: The study focused on the relationship between the pulse properties and the complex time delay. Specifically, they calculated the shift in transmission time ($D_t$) and the shift in center frequency ($D_\omega$) between input and output pulses. These experimental shifts were compared against theoretical predictions derived from Asano et al. [71] and Cao et al. [83], which posited:
   • $D_t = \text{Re}[\tau_T]$
   • $D_\omega = -\tilde{\Delta}^2 \text{Im}[\tau_T]$
     where $\tilde{\Delta}$ is related to the pulse's frequency bandwidth.

Key Results
The experimental data demonstrated a strong agreement between the measured pulse modifications and the predictions based on the complex transmission time delay:

• Time Shifts: The measured temporal shifts ($D_t$) aligned with the real part of the transmission time delay, $\text{Re}[\tau_T]$, including regions of negative time delay.
• Frequency Shifts: Crucially, the measured center-frequency shifts ($D_\omega$) of the transmitted Gaussian pulses matched the theoretical predictions derived from the imaginary part of the transmission time delay, $\text{Im}[\tau_T]$.
• Bandwidth Dependence: The results confirmed the predicted scaling behavior, where the frequency shift increases with the square of the pulse bandwidth ($\tilde{\Delta}^2$), while the time shift remains largely independent of the bandwidth within the tested limits.
• Systematic Deviations: Minor systematic deviations were observed between the time-domain results and frequency-domain predictions in the regions between Feshbach modes. The authors attribute these to standing waves on the coaxial cables used in the time-domain setup.

Significance and Claims
The primary contribution of this work is the experimental validation of the physical interpretation of the imaginary part of the transmission time delay. The authors establish that $\text{Im}[\tau_T]$ is directly responsible for the shift in the carrier frequency of a scattered pulse. This finding provides a concrete physical meaning to an abstract component of the complex time delay, linking it to the non-uniform distortion of the pulse's Fourier components as it traverses a dispersive medium.

The paper claims to generalize the theoretical framework of Asano et al. [71] by demonstrating that complex time delay derived from frequency-domain data can provide model-free predictions for pulse propagation properties in non-Hermitian systems. Unlike previous approaches that relied on specific resonant models, this work shows that the relationship holds even for frequencies far from resonant modes where analytical approximations might fail. The authors conclude that their results establish a detailed equivalence between complex scattering information derived from frequency-domain and time-domain approaches, offering a simplified and unified method for characterizing scattering systems across the electromagnetic spectrum.