Raman Labeled Nanoparticles: Characterization of Variability and Improved Method for Unmixing

Area of Science:

• Nanotechnology
• Spectroscopy
• Analytical Chemistry

Background:

• Raman spectroscopy offers high multiplexing potential for nanoparticle analysis due to unique spectral fingerprints.
• Quantitative unmixing of Raman spectra is hindered by peak overlaps and spectral variations (location, height, width).
• Inaccurate unmixing leads to significant errors in nanoparticle concentration determination.

Purpose of the Study:

• Develop and validate a mathematical model for quantitative unmixing of Raman spectra from nanoparticles.
• Enable simultaneous quantitation of multiple nanoparticle concentrations in complex samples.
• Improve accuracy in characterizing nanoparticle mixtures.

Main Methods:

• Developed the Narrow Peak Spectral Algorithm (NPSA) for quantitative spectral unmixing.
• Evaluated NPSA's performance in unmixing spectra from up to 7 Raman nanoparticles.
• Corrected for spectral variations including 30% intensity changes and peak shifts up to 10 cm⁻¹ (50% FWHM).
• Compared NPSA performance against conventional least squares (LS) analysis.

Main Results:

• NPSA successfully unmixed spectra from up to 7 Raman nanoparticles.
• NPSA demonstrated approximately 50% lower error than LS analysis.
• Error rates for NPSA were 10-16% (equal ratios) and 13-19% (unequal ratios) for 7 composite organic-inorganic nanoparticles (COINs).
• LS analysis showed significantly higher errors: 25-38% (equal) and 45-68% (unequal).
• Achieved quantitative unmixing of 7 nanoparticles with a maximum RMS % error below 20%.

Conclusions:

• NPSA provides a robust and accurate method for quantitative unmixing of complex Raman nanoparticle mixtures.
• The algorithm effectively handles spectral variations, improving unmixing accuracy.
• This represents a significant advancement for simultaneous multi-nanoparticle quantitation in various applications.