Eclipse results - a quasi real time success

Science Nugget: August 13, 1999

1. Introduction

As the limited readership of these notes knows quite well, a satellite in an orbit like Yohkoh's gets many opportunities to observe partial eclipses of the Sun. In our case, of course, it is partial eclipses of the X-ray corona. From the point of view of spacecraft control, an eclipse is a tricky business because the Sun sensors don't work as advertised. Accordingly we have made many, many mistakes in the past, even though we have made some rather spectacular observations. This time we were brave enough to promise movies of the August 11, 1999 eclipse to an extremely prominent Japanese TV program. To succeed, we would have to get the movies made within a few minutes after reception of the data. So... this Science Nugget is to boast that we succeeded! The material provided for TV use may be found on the "quasi-real-time" eclipse pages, which include large movies (which may be slow to download) in .gif format. A still from this page appears on the right:

2. So what good is it to observe an eclipse?

This is a frequently asked question. The SXT chief observer often answers "because it's fun," but in fact there are potentially important applications of these data. The Moon is opaque even to our powerful X-rays, and the Moon's edge is almost smooth on the angular scales of our pixels. Thus it is a rather good mask or knife-edge; it is good for probing the level of scattered light, the solar astronomer's chief bugaboo. H. Hara has already used the latter property to estimate the scattered-light level in a coronal hole on the solar disk, a crucial measurement, from an earlier eclipse observation. We now comment below briefly on the knife-edge property of the images.

3. A lunar knife edge

The image on the right shows how SXT viewed the eclipse of an active region on the NE limb. This image looks odd -- it has three separate parts -- because of the way the SXT on-board software works: an image 192x192 pixels, as this one, requires three separate exposures, one for each EW strip.   The Moon's image is in three different  places at the times of the three different exposures, so the three parts of the image don't fit together very well.

To carry out a simple analysis using this exposure as a knife edge, we've taken the left middle part of the above image, and rotated it by 42 degrees to make the Moon's limb vertical.  This makes it easy, using simple IDL software, to make a plot of how well the telescope images the sharp drop of brightness that occurs at the edge of the moon's image.   To reduce the effect of counting statistics, we've summed the bottom 10 rows of pixels in this image, since we don't expect significant variation in the vertical dimension.

The plot on the right, which we've made in this manner, is pretty self-explanatory.  This shows how the brightness varies along this horizontal strip of 10 combined rows of pixels in the plot to the left.  It shows a precipitous drop in the brightness at the limb of the Moon; the part to the left is the solar active region, but the part to the right is in the shadow of the moon. Here we see a bit of scattered light. The distance between the 3/4 and 1/4 intensity levels, a crude measure of the width of the core of the point response function, is 3.6" (about 1.5 pixels), confirming that the 2.4" SXT pixels are too big to sample the point response function of the instrument properly (we really need 2 pixels to critically sample the image).  We knew this already!  But more detailed analysis of this data, plus others in the archive, could really teach us a lot about how the SXT optics work. Unfortunately we did not get enough exposures that were long enough, so this eclipse will be limited by counting statistics for applications like this. At this position the Moon's limb was moving rapidly, so a longer exposure might have caused appreciable blurring.

Now that Chandra, NASA's huge grazing-incidence X-ray telescope for astrophysics, has been launched, we'll expect to see some really, really good grazing-incidence images of distant bits of the Universe, but in the meanwhile SXT is making nice images of bits of the Universe at almost exactly one A.U. distance, and we need to understand these too.

H. Hudson (

August 13, 1999