Less is more: dimensionality reduction as a general strategy for more precise luminescence thermometry

Area of Science:

• Materials Science
• Nanotechnology
• Spectroscopy

Background:

• Thermal resolution is critical for luminescent thermometers, defining their minimum discernible temperature change.
• Existing methods focus on probe optimization and artifact correction, with limited exploitation of calibration datasets.
• Novel approaches are needed to enhance the performance of luminescence-based thermometry.

Purpose of the Study:

• To introduce dimensionality reduction techniques for defining luminescence-based thermometric parameters.
• To improve the thermal resolution of luminescent thermometers by better utilizing calibration data.
• To enable more precise monitoring of small temperature variations.

Main Methods:

• Application of linear Principal Component Analysis (PCA) to calibration datasets.
• Application of non-linear t-distributed Stochastic Neighbor Embedding (t-SNE) to calibration datasets.
• Comparison of dimensionality reduction methods with classical intensity-based and ratiometric approaches.

Main Results:

• Dimensionality reduction techniques significantly improved thermal resolution compared to classical methods.
• Rare-earth nanoparticles and semiconductor nanocrystals were used as calibration materials.
• Precise temperature monitoring of changes smaller than 0.1 °C was achieved.

Conclusions:

• Dimensionality reduction offers a superior approach for selecting thermometric parameters in luminescent thermometers.
• These methods push the performance of luminescent thermometry closer to experimentally achievable limits.
• The presented techniques provide a new perspective for optimizing luminescence-based temperature sensing.