The principle of confocal optical microscopy is shown in Figure 1(b). In this configuration, a confocal aperture is placed in a remote image plane to reduce the sampling depth of field. In this simple example, the aperture blocks the Raman light scattered from Z2, thus resulting in a spectrum exclusively from Z1. Only the in-focus and on-axis Raman light rays are recorded by the spectrograph system, because the confocal aperture blocks the out-of-focus and off-axis light rays. The result is that confocal optical and Raman microscopes restrict the sampling depth to a region that is smaller than that obtained using conventional optics.
In addition, confocal measurements can improve the rejection of stray light and reduce fluorescence interference. Not all confocal Raman microscopy designs are the same. Traditional confocal Raman microscopes, shown schematically in Figure 1(b), utilize a pinhole aperture placed in front of the spectrograph entrance slit. The Raman light is focused onto the pinhole and the diverging beam after the pinhole is then refocused onto the entrance slit of the spectrograph. Different pinhole apertures can be used to control the degree of confocality, while the entrance slit is used to control the spectral resolution of the spectrometer. While this true confocal configuration provides independent control of spatial and spectral resolutions, it is very difficult to align and to maintain optimum performance. This is because the beam is focused twice through two very small apertures. In practice, independent control of the two apertures offers little value. Typical slit widths must be less than 100 um to achieve acceptable spectral resolution, and above 25 um to avoid diffraction effects. The size of the pinhole must be greater than the diffraction limit, but too large a pinhole defeats its purpose. Furthermore, using slit widths larger than the pinhole diameter makes no sense, since it does not improve signal throughput (assuming perfect 1:1 imaging between the pinhole and the slit). Therefore, for the most part, the pinhole diameter and the slit width are kept at similar dimensions, making them redundant. As shown in figure 2(a), the spatial resolution can be controlled by a combination of the entrance slit in one direction, and the spatial resolution of the CCD detector in the orthogonal direction. This configuration depends on the imaging quality of the spectrometer; i.e. a point source at the entrance slit must be imaged to a very small spot. In reality, the non-ideal performance of the spectrograph optics makes this pseudo-confocal configuration inferior to the true confocal approach in terms of spatial resolution. Due to the reduced number of optics, the overall throughput is greater than that achievable with a true confocal design, but less than with a non-confocal or high throughput design.
