Automated Raman imaging for first-time users and specialists
Confocal Raman Imaging combines a confocal microscope with a highly sensitive Raman spectrometer. Image acquisition is therefore not confined to the surface, but can also include depth-profiles (optical sections) and 3D images, typically resulting in very good signal-to-noise ratios and a high image contrast. In confocal Raman microscopy a complete Raman spectrum is recorded from 256 lines each containing 256 pixels. From this “multi-spectrum file” appropriate software can produce an image of the distribution of different molecular species in the sample.
Automation of Raman Microscopy
In confocal Raman microscopy, instrument setup is critically important in achieving high performance. With conventional systems, manual adjustment or exchange of various optical components is generally required. A push-button system has the advantage of drastically reducing the time required to become familiar with the operation of the instrument. With apyron, generally only the loading of the sample and the selection of the lens are handled manually. All other functions such as the choice and change of excitation wavelength, adjustment of laser power, positioning and focusing of the sample, calibration and signal maximization can be simply set on the computer screen – the instrument will then automatically adjust all components without further help. All set parameters are recorded during the measurement and can be recalled later, ensuring reproducibility.
The microscope is equipped with the standard UHTS 300 spectrometer for the visible range, but also the UHTS 400 for the red and infrared range and the UHTS 600 with a longer focal length for extremely high spectral resolution are now available. Common to all is the lens-based optical path that enables an ideal line shape for Raman spectra, the highest transmission, three grating positions for maximum flexibility, and a variety of CCD cameras for detection. Special emphasis should be placed on the adjustment of the laser power: Since it must be selected so that Raman intensity is maximized, while avoiding the sample being modified or damaged in this system the laser power can be adjusted in 0.1 mW steps and the actual laser power values are continuously measured and recorded, ready to be matched precisely in repeated measurements.
3D Raman Imaging on an Emulsion with CCl4
A standard substance in Raman spectroscopy is carbon tetrachloride (CCl4). For the following study CCl4 was emulsified with an alkane, water and oil. It should serve to demonstrate the extent to which high resolution spectroscopy can be combined with high resolution imaging. First, a 3D Raman image of the emulsion was obtained (Fig.1). It shows the three-dimensional distribution of the emulsion components, with the CCl4 clearly dissolved in the oil phase. One oil drop was investigated for a second time with a high-resolution grating. In this image structures smaller than one micrometer can be recognized in detail. The achievable spectral resolution when imaging is so high that the three individual peaks in the region around 460 wavenumbers can be easily resolved. These CCl4 Raman bands are a measurement standard for high-resolution Raman spectroscopy.
Raman in the Sharpest Detail
The amount of data that such measurements generate is enormous. Therefore, a powerful piece of software for fast data and sophisticated image and data processing is essential. With the WITec FOUR suite a tool is available that allows both the recording and subsequent evaluation of several million spectra. Fig.2, for example, shows a dataset of 16.8 million spectra (pixels). It again depicts the water-alkane-oil-CCl4 emulsion which in this case was scanned over a total area of 1000 x 1000 μm2. The complete dataset concerning the distribution of the components was evaluated, on the left two very small structures were brought into view without further measurement. The picture on the right was produced by simply zooming in on the image data plane. The result: Raman in “HD” shows the smallest sample features with diffraction-limited lateral resolution.
Statistical Analysis -Dexpanthenol Ointments
When investigating the distribution of chemical components, statistical methods can also be employed to automate the process. Here for example is the technique known as cluster analysis, which can simply and quickly extract spectral characteristics. Moreover, image data can be very clearly evaluated by statistical histogram analysis. In the following two dexpanthenol-containing wound healing ointments were analyzed. Areas of application of the ointments were the skin and the eye. Both resulting images (Fig.3) already display differences in the concentration of the individual components.
When initially considering the spectra averaged over the entire measurement range, the two ointments appear to be very similar (data not shown). If a large excitation laser spot was used to illuminate the sample, little difference between the spectra could be ascertained. Only with the aforementioned lateral resolution is it possible to differentiate and statistically describe the individual features. In the histogram analysis, the intensity in counts was plotted against the number of pixels for each component. The relation of the incidence of a component to its intensity is immediately apparent. Fig.4 shows the resulting histogram analyses of the individual components. The mixing ratio of the individual components is shown to be quite different at each area of application. In this case the eye ointment contains less oil, and on average more water. Product developers obtain meaningful results through such analyses with which material properties can be improved.
Summary
In this paper we have shown that an automated Raman Imaging System (WITec apyron) can offer unparalleled ease of use without any compromise in performance. Imaging at high spatial resolution can be carried out simultaneously with the highest resolution spectroscopy, as evidenced by measurements on emulsions with CCl4. In the data analysis section it was demonstrated that with suitable software tools, large amounts of data can be quickly and easily statistically evaluated and depicted in the sharpest detail.