An Instruction Dataset for Extracting Quantum Cascade Laser Properties from Scientific Text

Quantum Cascade Lasers (QCL) are semiconductor laser devices with a lasing frequency range in the mid to infra red region. The QCL device has both design (materials stacking properties, design types) and working properties (performance properties such as temperature, power etc). These properties are in most cases captured in scientific literature. A QCL device with particular design features has corresponding working properties which implies that the laser design influences the working properties. In order to study these relationships in detail, we propose an instruction dataset for training and evaluation of large language models for QCL property extraction from text . This is important as the LLMs are not able to identify some domain specific properties without adaptation. Extraction of QCL properties data from text will provide structured data for analysis in order to understand the relationships between the QCL properties and hence provide insights for fabrication of QCL devices with target properties.