Rapid analyte recognition in a device based on optical sensors and the olfactory system

Area of Science:

• Biomimetic sensor design
• Artificial olfaction systems
• Optically based chemical sensing

Background:

• The vertebrate olfactory system inspires distributed sensor arrays for pattern recognition.
• Artificial neural networks (ANNs) offer powerful tools for analyzing complex sensor data.
• Fiber-optic sensors provide a versatile platform for chemical detection.

Purpose of the Study:

• To develop a novel vapor sensing device utilizing an array of optically based chemosensors.
• To integrate these sensors with an artificial neural network for pattern recognition.
• To evaluate the device's ability to identify single analytes, binary mixtures, and concentrations.

Main Methods:

• Constructed fiber-optic sensors by immobilizing Nile Red dye in various polymer matrices.
• Created differential sensor responses based on polymer properties (polarity, hydrophobicity, pore size, etc.).
• Trained ANNs using temporal fluorescent signal data from the fiber array for analyte identification.

Main Results:

• Individual fiber sensors exhibited broadly but differentially responsive characteristics.
• ANNs trained with temporal data from the entire fiber array demonstrated high accuracy in analyte identification.
• ANNs trained with integrated array data or single-fiber temporal data showed significant errors.

Conclusions:

• A fiber-optic sensor array coupled with ANNs can effectively identify chemical vapors.
• Temporal information from the sensor array is crucial for accurate analyte recognition.
• This biomimetic approach offers a promising direction for advanced vapor sensing technologies.