Conference Agenda

Raman spectroscopy was explored as a tool for the characterization of canine mast cell tumors based on their biochemical signatures. Representative spectra revealed distinct features in key regions associated with proteins, lipids, and nucleic acids. Variations in the relative intensity of these bands suggest differences in molecular composition between tumor samples. The 860 cm⁻¹ region, associated with tyrosine-related protein structures, showed noticeable changes that may reflect alterations in tumor biology. These findings demonstrate that Raman spectroscopy is sensitive to biochemical variations in canine mast cell tumors.

Fully fiber-optic coherent Differential Absorption Lidar (DIAL) operating in the 1.5 µm telecom band is a promising technique for range-resolved atmospheric greenhouse gas sensing. However, retrieval of weakly absorbing species at ambient concentration levels with sub-ppm sensitivity, such as CO₂, requires power-measurement accuracy below 10⁻⁴ dB, making even small instrumental nonlinearities critical. In practice, the long averaging times required to suppress random uncertainty mask such deterministic biases, complicating their identification under field conditions.

We present a methodology based on an external fiber-optic delay loop that emulates a range-resolved lidar echo. Parasitic polarization cross-coupling within the loop produces a controlled DAOD modulation, while the high signal level and absence of speckle enable observation of instrumental behavior over a wide dynamic range with high temporal resolution. Using this approach, ADC nonlinearity was identified as the dominant bias source in our system, producing DAOD errors up to ±0.012.

Two independent mitigation techniques were applied. We propose active equalization of the ON/OFF local oscillator power, a method generalizable to other instrumental nonlinearities in coherent DIAL, which reduced the effective DAOD nonlinearity by approximately one order of magnitude. Additionally, RF dithering, targeting ADC nonlinearity specifically, provided a further sixfold reduction.

This work proposes a controllable optical arrangement capable of simultaneously generating linear and nonlinear optical mappings required for training physical machine learning models. We demonstrate the advantage of nonlinear over linear optical encoding by training optical reservoir models for chaotic system forecasting, reducing the average model error by approximately 50% in short-term predictions. This method contributes toward simple-to-implement optical nonlinear functions, advancing the development of physical optical implementations of neural networks.

Hybrid-integrated lasers enable compact devices for sensing applica-

tions. A GaSb/SiN widely tunable integrated laser operating at around 2.3 µm

is demonstrated. Using the Vernier effect implemented with a double-ring res-

onator, the laser can be tuned by about 83 nm, with a maximum power of

∼5 mW, and a threshold of 189 mA.

Access to deep geothermal energy and subsurface mineral resources is increasingly constrained by the limitations of conventional mechanical drilling and the associated cost. Laser drilling offers a contactless, non-mechanical alternative that avoids bit wear entirely and has demonstrated penetration rates competitive with or exceeding conventional methods at laboratory scale.

Interaction between a high-power laser beam and rock produces three distinct observable phenomena: spallation, melting, and vaporization, governed primarily by the delivered power density and depending strongly on the lithology.

Thermal spallation is the most energy-efficient rock removal process, but operates within a narrow power density window of 300-1000W/cm^2 in granite that must be actively maintained to achieve sustained penetration over meters then kilometers. Precise control of the power density on the rock surface, real-time process monitoring, and robustness to mineralogical heterogeneity are non-trivial challenges when operating in a deep borehole environment.

We present laboratory demonstrations of high-power laser deep drilling using a surface-mounted laser source coupled to a downhole drilling head, achieving penetration rates exceeding 20m/h at large diameters (>5-10cm), and discuss the key optical and engineering challenges on the path toward industrialization.