Stephen St.Vincent - Swarthmore College
Research Home | Journal | Articles | To-Do | IDL code
Back to main images page
Same as below, but with an upper limit on the radius of 3 stellar radii. Most of the emission is still there, so most of this red shoulder is within 2 stellar radius of the star, but at least 0.1 stellar radii above it.
Here's the same image as below, but with a radius cutoff at 1.1 stellar radii. Clearly, the offending material (the redshifted peak) is not coming from right next to the star.
Here's the phase 0 line profile, with all of the h-alpha material (including that which would normally be occulted) being included in the red line. I was kind of hoping for more blueshifted material.
10 August 2006
I made one more Hα line profile, this time so that I could figure out which timesteps were in it. The redshifted peak was still there. It turns out that the order of the slices goes counter-clockwise from 9 o'clock if β=0 and i=0. The phi slices that contribute the most are 8, 9, 18, 19, and 23 (all have non-occulted points with EM>10^25.5 cm-6). These correspond to timesteps 20, 59, 47, 23, and 49, respectively.
Here's the XZ-plane color contour of phi-slice 23 (timestep 49), where I've changed the line-of-sight velocity color range to be +/- 200 km/s. The big blobs of white are off the color bar, so they have too high or too low a vlos.
It's a little hard to see, but I've plotted white asterisks at the highest emission points. They're all just barely not occulted and very close to the star. (Points plotted have EM > 10^26; they also would fall into the redshifted peak in the line profile.)
Here's another image, from slice 18 (timestep 47), but with the EM cutoff at 10^25.7. There looks to be more emission below the star, but it's all so close to the star that the stuff on the bottom half never gets seen. Not all of these points would fall into the reshifted peak in the line profile; only the ones in similar locations to the points above would do so.
09 August 2006
and wondered where the material in the rightmost peak was coming from. So today, I made this 3D plot:
where I've just plotted all of the points in black so that they're easier to see. I just only plotted points where 50 km/s ≤ vlos ≤ 200 km/s.
In addition, I've made the &phi=0.6 and φ=0 line profiles with a fresh simulation to see if this behavior is consistent. Below are my results, which appear to be consistent.
φ=0
φ=0.6
So, to test the continued presence of the righthand peak, I made another 3D plot:
Less points show up, which is good since the peak is lower. It just seems kind of arbitrary, in that if the observer looked at the grid from the opposite side, he would see a blushifted peak.
But then I made another 3D plot with a higher EM cutoff:
This makes me think that maybe this is inherent. A possible explanation is that this is material that's early in it's path along a magnetic field line close to the star and is rushing towards the equator; its blueshifted counterparts may be occulted by the star. Indeed, the 38 points with the highest levels of emission are all within 0.03 stellar radii of the stellar surface. 13 of those points are occulted by the star.
08 August 2006
Here's a line profile of a 3D simlation at phase 0.5. There's a lot more emission from redshifted material than blueshifted material, suggesting that the infalling material that isn't occulted by the star is the cause. This agrees with theories that had been previously put forth.
Here's a similar image as above, only for phase 0. I've got to figure out a way to tell where the material that's emitting in each of the peak is.
07 August 2006
Another composite line profile. Here, all zones have log(T)=3.69, or T≈5000 K.
Yet another composite line profile. Here, all zones have log(T)=7.
04 August 2006
Composite line profile; all temperature zones are log(T)=6. The peaks switched again; maybe that's because there was just more EM in the right peak to begin with? Probably doesn't make sense, though.
Composite line profile; all temperature zones are log(T)=0. Everything overlaps perfectly, as expected.
Composite line profile; all temperature zones are log(T)=4.66. It looks pretty much just like the real one.
Here's an overlapping image of two line profiles of the same timestep, one with thermal broadening (red) and one without (black).
02 August 2006
Here's the corresponding x-ray EM(phase) plot, which looks a lot more normal.
This is the EM(phase) histogram. Highly unusual. There are 40 divisions, as it is the same simulation file as before.
Here's the line profile of material that's H-alpha emitting and is within 0.1 stellar radii of the stellar surface and its corresponding EM(T) histogram. Again, the red line takes into account occultation of stellar wind material.
Here's the line profile of material that's H-alpha emitting and is above 0.1 stellar radii off the stellar surface and its corresponding EM(T) histogram.
Above is the same line profile as below, but with a particle mass of 16 mH. The lower image is the corresponding emission measure vs. temperature histogram. The line profile still has the same center, but is a little bit broader.
Here are the X-ray line profiles created using the correct particle mass (9.352e-23 g, or 56 times the mass of the hydrogen atom). The top image is with thermal broadening, while the bottom image is without. The line center is essentially unchanged, while the width changed by a few percent. In addition, the structure on the right side of the bottom plot appears to be smoothed out in the upper plot.
01 August 2006
Above are two X-ray line profiles. The top image employs the thermal broadening, while the bottom image does not. All EM is present and accounted for. And don't be fooled, the non-broadened image actually has a higher peak; the y-axis scale has changed by a factor of 10. The line centers are nearly identical, within a fraction of a percent, so I'm alright with chalking that up to computational boundary issues and moving on with my life.
So it turns out that I just wasn't removing points based on my thresholding for all of the variables. Which was stupid. Here's the latest image. Of course, I think that this lack of thresholding will affect the non-broadened image as well, so I'll remake that...
...and here it is! Looks narrower! Yay! The broadening is subtle, but I can assure you that the total emission measure in each plot is the same.
The image with the error function calculation fixed. It looks exactly the same, which means (as I half-expected) that I was doing everything correctly but in a convoluted way.
Here's the line profile from yesterday with thermal broadening corrected to include the volume of each point in the calculations. Still looks pretty bad.