As you may know there was a total solar eclipse recently (29.03.2006), which I was lucky enough to see from the ancient town of Side, Turkey (nowadays an amazing sea resort, not so far from Antalia). Since several other folks and I were not officially engaged with the scientific activities of the Institute (IA-BAS), nor were we even officially representing it, we decided to fully enjoy the spectacular view. Yet, some amateur-level scientific experiments were to be performed, and I was dealing with the polarimetry. We chose a spot on the arena of an ancient Greek theater, one of the many relatively well preserved in the region. We had to share the place with several dozen guys from NASA, who, equipped huge expensive telescopes and cameras were performing live broadcast of the event. I say "from NASA" not because they said so. They even denied they were from NASA, and the t-shirts they wore had a very meaningful label - "crew". Nevertheless everybody knew who they were. It took me some time to figure out why they were hiding themselves. Osama, however, didn’t get it until the end, apparently.

Anyway, the guys from NASA weren't very lucky to share the stage with us, but they had no choice since the mayor of Side gave us a special permission to be there. Rumors spread that he did so because he had been attracted to a girl from our group, who on the other hand was with someone she wasn't supposed to be, but this is a different story. Especially taking into account she was not the only one. However ask somebody else for details.

Now, back to polarimetry, since I already smell troubles. As I said we used only amateur equipment (deliberately). I had taken a small military grade Russian scope (Turist-5) to monitor the partial phases and not to miss the totality, since we did not have an accurate watch and furthermore weren't sure when exactly the total phase was. (Actually don't believe if somebody tells you that the 3 keys to success of the observation of a solar eclipse are preparation, preparation and again preparation - I would trade "preparation" for a good weather, for instance.) I also had my 3 Mpx digital camera with 10x optical zoom (Toshiba PDR M-700, Canon optics). Taking pictures trough a digital camera of a solar eclipse spells as two things - large optical zoom and proper mounting for the camera. I used a Russian military grade mount for my camera (much better than similar from Carl Zeis) and the camera remote control. Most of the camera owners are not lucky enough to have remote controls for their cameras, and if you are one of them then use self-timer to avoid trembling. And next time think before buying a camera.

I also had a polarizing filter, unfortunately not of extremely good quality. To measure properly the polarization, one is supposed to take 3 pictures rotating the polarizer at different angles in front of the lens. Less than 3 is not enough since there are 3 unknowns for each pixel - the intensity of the total light, the polarization rate and the angle of the polarization (for more details send "Stokes parameters" to google). Some say the polarizer has to be turned at a 45 degrees each time, but I don't see how it has to be 45, not 60, where the errors, I believe, should be minimized. So I did 60-degree turns, of course.

Once the pictures are taken the best way to proceed is to convert them into fits format and to use IRAF for the reductions, since there is a serious picture math involved (see below). After the alignment of the solar disk (there was no guiding) I converted the jpg frames into fits with IRIS. From these 3 frames one can make images for the unpolarized light, the polarization rate and the angle of the polarization after some math. Unless I made a mistake, the expressions are as follows:

Io=0.67*[I1+I2+I3 - 2*(I1^2 + I2^2 + I3^2 - I1*I2 - I2*I3 - I1*I3)^0.5]

Ip=0.33*[I2 + I3 +3*I1 - 2.5*Io + {(2*I2 - Io)*(2*I3 - Io)}^0.5]

mu=[(2*I1 - Io)/(2*Ip)]^0.5

where Io is the unpolarized light, Ip - the polarized one, mu=cos(theta) - the cosine of the polarization angle, and the polarization rate is simply p=1/(1+Io/Ip). I1, I2 and I3 are the 3 frames, taken at different angles (0, 60 and 120 degrees, respectively).

Once again - these formulas are supposed to work for 60 degree turns, and of course do not detect circularly polarized light (but only the linear component). After spending some time deriving these formulas, someone told me they were in the books and it was probably so, so feel free to check proper sources, which I, by the way, did not find. If you have used turns different from 60 degrees, good luck in deriving your own formulas.

Finally some results. There were 2 series of exposures taken through the polarizer - with 1/16sec and 4sec exposure times. The short exposures were to reveal the inner corona, the long ones - the outer corona (while the inner shows up saturated). See below firstly the direct images and after that - the reductions showing (from left to right) the unpolarized light image (Io), the polarization rate (p) and the angle (mu) for the two exposure times.

Pictures of the Sun through a very red filter (K8) 1/16 and 4 sec, no polarizer used. Using such a red filter makes sense since it reveals well the outer corona (in red light) and also the inner one (in white-yellow), where the blue colors might be saturated, but so are the red ones and the overall color is neutral. Note that no artificial colors are used in the right picture.

3 frames through a polarizing filter, oriented at 0, 60 and 120 degrees respectively from the vertical direction. 1/16sec exposure.

The same as above, 4sec exposure. The inner corona is saturated.

Polarization of the inner corona (1/16sec exposure). From left to right - the unpolarized image, the polarization rate (white colors indicate higher polarization), and the angle (white colors indicate vertical polarization, black - the horizontal one). Outside 1-2 solar radii spurious effects appear, due to the faintness of the corona.

The same as above, this time for the 4 sec exposure. The spurious effects in the inner ~1 solar radius are due to the saturation of the images.

Although one should not use the results above for any serious scientific purposes it is still seen that mostly streamers are polarized. The measured degree of polarization is about ~30-50% for the inner corona and ~15% for the outer (streamers). Slight cirrus clouds during the total phase may also affect the results.

Yet, the results are quite similar to the prediction http://shadow.adnc.net/corona/mar06eclipse/mar06eclipse.html especially taking into account a small (~10 degree) orientation difference between the images.

At the end – few pictures:

The theater. NASA people are on the stage, the ruins of the ancient Side are seen behind.

The theater again.

Our group on the left, NASA on the right.

Our group. My equipment is on the left (barely visible).

A better picture of the equipment and the crowd behind.

Projecting the partial phases through holes in a sheet, in front of a scripted ancient stone.

The partial phase through a Russian military equipment.

The temple of Apollo, Side, Turkey.

Shortly before attacking the enemy’s boat. (Antalia)

Already on board. (Manavgat river).

A captured enemy sailor, shortly before hanging him on the mast

The ancient city of Perge.

Sultan palace, Istanbul.