Raman Microscope optical design.

Introduction.

               When  I look through the web and search  for "amateur spectroscopy", I usually find a lot of material about astronomical spectroscopy: how to build spectrographs and couple them to  telescopes, interesting results about star and nebulae spectra but relatively few about mineralogical spectroscopy  except perhaps fluorescence spectroscopy and of course professional or academic work. In particular I have not  found anything about amateur Raman spectroscopy but I could have missed something.   I have already shown in other web pages of this site the possibility for amateur microscopist to come into the field of minerals spectroscopy, in particular the transmission spectroscopy of rock sections in the visible and NIR . Although this kind of technique allows to make interesting studies of minerals, it has some drawbacks to help for the identification of  minerals in rock sections. First of all, it requires the preparation of a thin or thick section depending on the mineral to be studied or identified. The reason is that  the extinction coefficient will vary greatly from a mineral to the next. For instance we can consider the olivine series: the absorption spectrum  in the NIR around 1200 nm of such a material  is  due to the presence of iron in the crystal structure. The Iron concentration is low in forsterite and high in fayalite so such is the absorption spectrum: forsterite requires a rather thick section to get a good signal to noise ratio, fayalite a thin section to avoid the saturation of the spectrum. Another example is gypsum: to get my  spectra in the near infrared, I needed a 1 mm thick section. For some other minerals like high iron content orthopyroxenes, the standard 30µ thickness is enough. A second problem is of course the impossibility to measure a spectrum of opaque mineral although the reflection coefficient method could be used for an identification purpose. Finally, some minerals have no absorption at all in the wavelength range considered.

With those possibilities and limitations  in mind, I decided to consider the transformation of my microscope and the construction of a Raman instrument. The goal was to identify small crystal particles in a section of rock. By section, I mean a roughly polished section, one face,  in order to minimize sample preparation. I began my work with an old helium neon laser I had recovered from an old out of work granulometer instrument so all the work described in these pages was done with the 632.8 nm He-Ne line which is not the most common wavelength used in Raman spectroscopy. This project was a little bit challenging because it is well known that the Raman effect is very weak so efforts have been made to keep the sensitivity of the system has high as possible. Moreover, to reduce the cost, I used parts that were already available to me even if better choice could be made buying new ones. It is especially true for the spectrograph. Probably in the future I'll try to improve the system by redesigning some of the parts.

     The design drawing of the first version of my Raman microscope is the following:

Description:   the light coming out of the laser (1) pass through a narrow band pass filter (2) centered on the laser wavelength to isolate the 632.8 nm line and thus to avoid that weak emission lines from the laser at other wavelengths can reach the detector. Remember that the Raman effect is very weak so the acquisition time  will be sufficiently large to detect any parasitic emission lines. The laser light pass then through a modified spatial filter (3) made up of 2 short focal length objectives. In professional designs there is also a pinhole in the spatial filter because commercial equipments use  confocal microscopy. Due to mechanical difficulties and in order to get as much light as possible, I removed the pinhole from the system. The disadvantage of this removal is a worse spatial resolution in the image plane of the microscope in the z direction parallel to the optic axis of the microscope lens (5). The distance between both objectives (3) can be adjusted by a micrometric screw to finely change the width of the beam. This allows to fully cover the entrance pupil of the microscope objective (5) with the laser beam in order to get a fine laser spot on the sample (6). The light pass then through an important element of the Raman microscope: the beam splitter (7). The model used is a dichroic beam splitter. When this device is placed at 45° in a light beam, it reflects the light towards the microscope objective lens (5) at the laser wavelength while the light at higher wavelengths ( in fact a few nanometers above the laser line ) can pass through the filter. So the majority of the laser incident light can reach the sample while the Raman light moving upwards after interaction with the sample (6) can pass through the beam splitter without being deflected and  thus continue its way to the spectrograph. A polarizer (4) can been inserted into the light beam to reduce the intensity of the laser in order to examine safely the laser spot in the microscope field of view to help focusing the beam and to choose the region of interest on the sample. The helium neon laser used here is a polarized model. Above the beam splitter, an edge filter (8) removes the 632.8nm laser light. Only the Raman light  3 nm above the laser wavelength and over can pass through the filter. The absorbance at the laser wavelength is around 6. The Raman light is reflected by a mirror (10) and refocused (18) onto the entrance slit of the spectrograph (11). The flip mirror (10) can been removed from the path of the main beam so that the neon light, used as a reference to calibrate the wavelength range of the spectrograph, can be fed into the entrance slit. The spectrograph is a modified Jobin Yvon H20 monochromator equipped with a concave holographic grating (12) and two plane mirrors (13). The exit slit of this monochromator has been removed and replaced by a large window in order to give access to the whole spectrum.  The image of the spectrum is focused by a camera lens (15) onto the CCD detector of an astronomical camera (16). Two bellows are used (14), one to prevent spurious light to reach the detector and the second one to allow the movement of the objective lens and thus to change the magnification of the spectrum image or rather, in our case, the demagnification to increase the light intensity on the detector and adjust the spectral range to an acceptable value.