Calibration of Raman bandwidth in large Raman images using a mercury–argon lamp

Abstract

Raman imaging has proven to be an important tool in the analysis of astromaterials. Raman imaging can describe the composition and mineralogy of astromaterials in a nondestructive manner preserving precious samples for additional study. Raman bandwidth is a common spectroscopic marker used to elucidate information such as composition, latent strain, and thermal processing history. The observed Raman bandwidth of a sample is a result of a convolution of the true Raman lineshape and the spectrometer's slit function. Thus, to compare Raman bandwidths across spectra and instruments, one needs to calibrate the Raman spectrum to account for effects of the slit function. Astromaterials are heterogeneous requiring the collection of large Raman images for adequate sampling. It is well known that the wavenumber calibration of an instrument drifts with time throughout Raman image collection. Here, we find that the spectrometer slit function also varies with time and laboratory temperature throughout Raman image collection. We utilize a mercury–argon (Hg–Ar) lamp integrated into our Raman microscope in a novel manner to calibrate the bandwidth in each Raman spectrum within our Raman image. We can do this because the narrow Hg–Ar lamp emission lines approximate the spectrometer slit function for each Raman spectrum allowing for the calculation of the true Raman bandwidth from each Raman spectrum within the Raman image. Our technique allows for the comparison of Raman bandwidth throughout a Raman image and between Raman images/spectra collected using different instruments, facilitating scientific analyses that are reproducible across multiple laboratories.

A new model for the calculation of homogeneously broadened Raman spectral bandwidths and bandshapes

“…The spectrometer SF is the observed band that results solely from the spectrometer and can be observed using a monochromatic light source. [3,4] The spectrometer contains a single slit at the entrance of the spectrometer. This slit results in a SF with a Gaussian shape as has been previously described.…”

“…This slit results in a SF with a Gaussian shape as has been previously described. [3,4,8] The SF width is the width of the image of the SF on the CCD detector. The SF width is determined by the linear dispersion of the instrument and the entrance slit width F I G U R E 1 Generalized optical diagram of a Raman spectrometer.…”

“…[5] The Raman band reported by the spectrometer and detector has a shape resulting from contributions of both the intrinsic Raman bandwidth and the spectrometer SF. [3,4,6,7] Thus, this observed Raman band (ORB) has a line shape resulting from both Gaussian and Lorentzian contributions. To calibrate the bandwidth and shape of a Raman band, one needs to separate the Gaussian SF and Lorentzian intrinsic Raman band (IRB) components for the ORB.…”

“…[1,2] Instrumental calibration is required to determine Raman spectral bandwidth and shape. [3,4] The line shape of a homogeneously broadened Raman band is Lorentzian. This band has a width dependent on the physical properties of the sample.…”

“…This is generally done by modeling the ORB as a convolution of Lorentzian and Gaussian functions, which is a Voigt function. [3,4,6,7] One can then set the Gaussian in the convolution to the SF measured for the spectrometer and determine the IRB width. While the Voigt function is an excellent model, it has many limitations.…”