A new publication from Opto-Electronic Sciences; DOI 10.29026/oes.2025.240004 , discusses tailoring temperature response for a multimode fiber.
A multimode fiber with a larger core diameter can transmit multiple modes of light waves. Due to its advantages of low optical loss, resistance to electromagnetic interference, compact size, and good stability, it is widely used in fields such as imaging, communication, spectroscopy, and high-power lasers.
However, modal dispersion, the inevitable random mixing of modes with different propagation constants, inherent defects of the fiber, and disorder caused by external disturbances all lead to the distortion of light signals transmitted through multimode fibers. When coherent light is coupled into a multimode fiber, it produces a seemingly chaotic pattern with bright and dark spots at the output end, known as a speckle pattern. Understanding and controlling this distortion in multimode fibers remains a significant challenge in applications such as fiber-optic communication, endoscopic imaging, and micro-manipulation.
Center for Complex Optical Fields and Meta-Optical Structures (COSMOS) at the University of Shanghai for Science and Technology has tackled the issue of multimode fiber response to environmental temperature fluctuations. They constructed a generalized Wigner-Smith operator based on the multimode fiber's multi-temperature transmission matrix. Using wavefront shaping technology, they experimentally generated a temperature principal mode that exhibits significant resilience to temperature-induced distortion. Additionally, this approach was employed to create a temperature anti-principal mode with an extremely narrow temperature bandwidth. To illustrate the practicality of the proposed special state, a learning-empowered fiber specklegram temperature sensor based on temperature anti-principal mode sensitization is proposed.
The researchers began by measuring the transmission matrix of the multimode fiber at various environmental temperatures. They then constructed a generalized Wigner-Smith operator, with its eigenstates representing the wavefront of the temperature principal mode. By designing a loss function, they were able to generate the temperature anti-principal mode based on the temperature principal mode. As illustrated in Fig. 1, compared to an unmodulated wavefront, the temperature principal mode can increase the temperature bandwidth by approximately 40%, while the temperature anti-principal mode can reduce the temperature bandwidth by about 30%.
By leveraging the temperature sensitivity of the temperature anti-principal mode, its application in optimizing the performance of learning empowered fiber specklegram temperature sensors is feasible, as illustrated in Fig. 2. The process begins with calibrating the temperature anti-principal mode and then collecting its output fields at various environmental temperatures to create a dataset. A regression-based deep learning model is employed to learn the mapping relationship between the speckle patterns and temperature fluctuations. Once trained, the model can directly predict the corresponding environmental temperature based solely on the speckle patterns collected at the fiber's output end.
Experiment validation of the effectiveness of the sensitivity enhancement scheme is presented in Fig. 3. The average measurement error of the sensor based on the unmodulated wavefront is 0.84°C, with a prediction error range of ±2°C. In contrast, the sensor based on the temperature anti-principal mode exhibits an average measurement error of 0.12°C, with a prediction error range of ±0.4°C, significantly improving sensor performance.
Keywords: multimode fiber / principal mode / wavefront shaping / optical fiber sensor / temperature response
