Poster Session 3, 1:15 PM - 2:00 PM: Concourse [B1]

Presenter: Diana Do

Faculty Sponsor: Joanna B. Dahl

School: UMass Boston

Research Area: Biology

ABSTRACT

Raman spectroscopy identifies biomolecules by their vibrational “fingerprints,” but thin samples are difficult to detect when substrate peaks and background signals dominate. This challenge is relevant to emerging diagnostics based on extracellular vesicles (EVs), which are membrane-bound particles in body fluids that carry lipids, proteins, and nucleic acids, which require analytical methods capable of distinguishing subtle, composition-driven differences between EV populations.

To model this signal-versus-background problem, this project tested Raman detection of glucose deposited as a thin dried film on calcium fluoride (CaF₂) and evaluated integration time. A 1.0 M D-(+)-glucose solution was prepared; 5 µL was pipetted onto CaF₂ and dried at 40°C. Spectra were collected using a 785 nm laser (100 mW), a 20× objective lens, and four coadditions, with integration times of 10 s and 80 s. Bare CaF₂, bulk glucose granules (~3 mm depth), and the dried glucose droplet on CaF₂ were analyzed using the same acquisition settings. Spectra were exported, baseline-corrected using an ALS routine in MATLAB, and averaged.

Increasing integration time raised intensity ~8.6× for CaF₂ and ~7.35× for the dried droplet, confirming improved signal collection. However, the dried-droplet spectra were dominated by a peak near 318–320 cm⁻¹ that overlaps the CaF₂ feature at 320–323 cm⁻¹, indicating substrate-driven measurements likely caused by a thin and/or non-uniform glucose layer. In contrast, bulk glucose showed bands consistent with carbohydrate vibrations (e.g., 841–915 cm⁻¹ and 1456 cm⁻¹), demonstrating detectability when sufficient thickness is present. Overall, improving film uniformity/thickness and reducing substrate contributions will be essential for reliable Raman identification on CaF₂, supporting measurement strategies needed for EV-based diagnostics.