Noisy rainbow has a story to tell
This raw data set made up of what’s called Raman spectra can help scientists determine the identity and solid form of a compound
Every day, Bristol Myers Squibb scientists encounter images that show the wonder of science and illustrate the cutting-edge work of company research teams. This vibrant image was submitted by Stacey Crawford, a scientist in Chemical Process Development, within the company’s Global Product Development and Supply (GPS) organization.
Part of her job is to determine the best conditions for crystal growth and evaluate the growing crystals through a microscope. When strong crystalline structure is observed, Crawford’s team then uses laser technology called Raman spectroscopy to get a “second opinion” on the material’s crystalline potential.
A strong crystalline structure means a material can hold its form, which is important for making a potential drug compound, Crawford explained. “Good structure has good purity as well, and we want pure structure for potential hits for our drug discovery,” she said.
Raman spectroscopy is also non-destructive, which means the material won’t be diluted or destroyed during analysis.
“It’s so important to keep the material preserved, especially because it could be the only material of its kind,” she said. “Now if a chemist wants the material back, they can have it back.”
The above image is actually an accident, Crawford said. The Raman spectra readouts her team prefers to review are much cleaner and less “noisy” than the one seen above. The hundreds of squiggly lines represent the laser hitting or “exciting” the material; the peaks and dips seen in the image can help identify crystalline structure and also whether those structures are similar or different from one another.
“The Raman outputs two types of data; raw data like what you see here, and then data with all the extraneous background removed. There is a lot of noise in this image; it’s all over the place,” Crawford said. “When I opened this image, I realized I opened the wrong data set, but it looked so cool and pretty, and it reminded me of string art.”
Here is an example of a “quieter” Raman spectra readout:
