Molybdenum Disulfide-Assisted Spontaneous Formation of Multistacked Gold Nanoparticles for Deep Learning-Integrated...

Area of Science:

• Nanotechnology and Materials Science
• Biomedical Engineering
• Analytical Chemistry

Background:

• Existing label-free detection methods face challenges in fabricating high-density, ordered nanoscale architectures via soluble processes.
• Sensitive and rapid screening for infectious diseases like COVID-19 requires advanced diagnostic tools.

Purpose of the Study:

• To develop an advanced biosensing platform integrating deep learning with surface-enhanced Raman scattering (SERS).
• To create large-area, highly ordered three-dimensional (3D) nanoscale architectures for enhanced detection sensitivity.
• To enable rapid, on-site screening of COVID-19 using human tears.

Main Methods:

• Fabrication of molybdenum disulfide (MoS2)-assisted gold nanoparticle (AuNP) 3D architectures using a spontaneous synthesis and electroless plating-like process.
• Integration of a convolutional neural network (CNN)-based deep learning model with SERS.
• Utilized human tears as the sample matrix for COVID-19 screening.

Main Results:

• Achieved outstanding SERS performance at subterascale levels with microlevel irradiation power and millisecond-level acquisition time.
• Demonstrated accurate assessment of COVID-19 susceptibility.
• The developed MoS2-AuNP architecture facilitated the formation of close-packed 3D multistacked AuNP structures.

Conclusions:

• The integrated deep learning and SERS platform offers a promising approach for label-free detection.
• The platform shows potential for rapid, low-damage, and high-throughput screening of analytes at exceedingly low concentrations.
• This technology could revolutionize on-site diagnostics for infectious diseases.