Saturation Levels and Bias

To examine the saturation level of the CCD we took several identical images with increasing exposure times. As you can see in Figure 1, the mean value of the pixel array in ADU increases linearly at first and eventually starts to level off at around 2.6 seconds exposure time. However, this leveling off occurs at a much lower intensity than the expected 65,535 ADU. The reason for this becomes apparent if we examine the image itself (Figure 2). As you can see, the number of photons received by the CCD is not uniform over the surface of the image. The image has darker areas in which the pixel intensity values do not reach even close to the saturation limit. Since we are taking a mean of the image to determine the saturation limit of the CCD, these darker areas will effectively lower the overall mean of the image when compared to an image that is uniformly saturated.

Figure 1: This is a plot of the mean value of the pixel array in DN as a function of the exposure time. The plot begins linearly and eventually levels off at around 2.6 seconds at a value of around $3.7\times 10^{4}$ DN. Notice that the saturation limit does not reach close to the ceiling of 65,535 that we would expect.

Figure 2: Here is the image that we used for the saturation plot. This image is taken at an exposure time of 2.5 seconds which is very close to the point in which Figure 1 starts to level off. The important thing to note here is that the image is not uniformly illuminated.

To get around the problem of non-uniform saturation, we can crop the image so that we are only looking at an area on the image that is very nearly uniformly lit. The resultant plot of the saturation of the cropped image is shown in Figure 3. Here you can see that the mean illumination of the image does approach the expected value of 65,535 ADU. There is still a small but noticeable curve before the image becomes saturated, however. This is due to the fact that the image is still not perfectly uniformly lit. If this were the case, then we would expect the plot to increase linearly as a function of exposure time until it reaches the saturation limit, where it will stop increasing abruptly and level off at 65,535 ADU. This would be the case if we plotted the value of a single pixel with increasing exposure time.

Figure 3: Here is a plot of the same image as Figure 1 but with the image cropped only to an area on the image that becomes fully saturated. Here you can see that the plot starts leveling off at the same exposure time, but the mean value of the pixels actually reach the ceiling of 65,535 as expected.