Selectable resolution

Resolution in an optical scanning system is always the result of a combination of several factors. Most important parameters are: pixel size, laser beam size, confocality and the scanning scheme. Only if we consider the combination of all parameters do we get the full picture (in terms of resolution). Let's start with a fundamental law of information theory that says that you always need to sample a signal with twice the bandwidth or frequency you actually want to see or hear. In a first order approach for an optical scanner system this means that the pixel size should be approximately half of the laser beam diameter. For a laser beam diameter of 12-16 |m the ideal pixel size therefore would be 6-8 |m. Most of the microarray features have a 100 |m diameter. On the other hand this also means that reducing the size of the laser beam without reducing the pixel size does not give the full benefit in terms of resolution. Therefore a pixel size of 5 |m and a laser beam size with the same dimensions might be slightly better, but is no way near twice as good as 5 |m pixel size with a laser beam of 10 |m diameter. However, there are two drawbacks of a very tight laser focus at the sample. First there is a higher risk of bleaching because the intensity at the small focus is much higher. The intensity providing higher sensitivity, and therefore the risk of bleaching, scale with the square of the beam diameter.

The second problem is under-sampling. If a laser beam diameter is rated as 10 |m according to standard definition this means that within a range of ± 5 |m the intensity of the beam is already down 1/e2. At the borders of such a beam only 14% of the maximum intensity is left. If a fluorescent molecule happens to sit away from the center of the beam it will be excited by a much lower intensity. In other words many of the labeled molecules will not be very efficiently excited and contribute much less to the overall signal than they could. Obviously this is a very sub-optimal (less sensitive!) way to make use of the rare sample molecules found in a spot at the end of the whole array process. The LS is designed to scan microarray spots from 60 to 200 |m in order to generate a minimum 100 pixels per feature on small spots and without the issue of under-sampling on larger spots.

0 0

Post a comment