Brett Berger graduated in the Harvey Mudd College class of 2015 with a Physics degree. In his senior thesis, he presented the design, calibration, and use of an inverted lab 532 nm Raman microscope. He also devised algorithms to refine the Raman signals and separate them from fluorescence peaks.
The lab worked with a doubled Nd:YVO4 laser that lacked a built-in infrared filter, so they implemented a glass filter in order to minimize infrared heating of organic samples. Below are a diagram and picture of the final apparatus:
The device has both a signal-collection mode for taking spectra and an imaging mode for taking high-resolution photographs of the samples. The team used a power meter (Thorlabs part PM100USB) to measure the device’s output power and power delivered to sample to ensure that the laser does not damage the samples.
Spectra collected by this device can fall into one of three categories. The ideal target produces a strong Raman signal that can be easily seen even in the raw data. Other samples will exhibit exceptionally strong absorption at 532 nm emitting neither reflectance, fluorescence, nor Raman signals. These samples are simply ill-fit for a 532 nm device and need to be interrogated using other excitation wavelengths (blue, red, or near-IR). Finally, some samples generate spectra that are a mixture of fluorescence and Raman signals. These signals can be deconvolved to provide both a LIFE spectral profile and a pure Raman spectrum.
Because of its sharp Raman peaks and relatively lack of fluorescence, Teflon is used as the calibration sample. The peaks “at 289 cm−1 and 380 cm−1 arise from torsion and deformation of CF2. The strong peak at 731 cm−1 is from symmetric CF2 stretching, that at 1216 cm−1 from anti-symmetric CF2 stretching, and those at 1300 cm−1 and 1382 cm−1 from C-C stretching.”
Berger produced fluorescence and Raman spectra for carbon tetrachloride (CCl4) and compared both with the spectra obtained using Ruiz’s SERDS device as well as spectra published in literature. Berger’s and Ruiz’s spectra were consistent with the literature, down to peaks as low as 217 cm-1.
The lab analyzed a solid sample of sulfur to test the low-Raman-shift capabilities of the instrument. The raw data obtained contained no fluorescence signatures and displayed the three peaks documented in the RRUFF Raman database (link here). A Raman-active electric quadrupole vibration exists at 84 cm−1. Analysis of the transmission curve of the system’s dichroic and edge ﬁlters indicates that the majority of an 84 cm−1 peak is blocked from collection but enough of its blue-ward photons pass through to register as a modest peak with apparent center frequency shifted to the blue. From this it would seem the device is capable of reliably identifying Raman shifts as low as 100 cm−1.
The formation of calcium carbonate, CaCO3, requires the presence of water. As a result, the mineral can be used as a marker of ancient hydrological activity on other planets. Berger’s instrument was able to fully separate the fluorescence and Raman spectra in calcite, and detect the 1085 cm-1 band in the calcite variant, aragonite. The team also analyzed an unknown fluorescent mineral deposit from the Pisgah crater lava tubes collected during Ramon field trials in summer 2014. The sample exhibited strong Raman peaks at 996 cm-1 and 1064 cm-1 constant with a mixture of sulfates and phosphates such as diadochite (Fe2PO4SO4.6H20). However, the sample lacked a water signature near 3500 cm-1 suggesting that this is an anhyhdrous mineral.