Most of these are M-type stars, but some late K's have snuck in. With good weather on the Kast, we've been able to get down to Vmag~15 in <1hr of integration. Tonight we had ~1.5 arcsecond seeing (not bad, not great).
We would like our spectra to look something like this (grabbed from Song, Zuckerman, & Bessell 2012):
The Kast uses a beam splitter to divide the light at 5500 Angstroms (when you're using a dichrotic mirror - called d55) into a "red side" and "blue side". The grating can be tilted so that the wavelengths you're interested in are centered on the detector. In this case, we used a grating tilt of ~25,300 so that we could place 6300 Angstroms in the center of the detector. Since the Kast 1200/5000 grating has a wavelength range of ~1400 A, we were able to get the sodium doublet (Na I D lines) at ~5900 A, the Ha line at ~6562A, the Li feature at 6707A, and the TiO bands which start around 7100A.
For calibrations, I'm taking 2 arclamps each for the blue and red side detectors using HeHgCdANe lamps. I will use these lamps to do a wavelength calibration later on based on template spectra available online. I am taking 11 flat frame images (for each read and blue side), and 11 bias frames (shutter closed, 0s integration).
To do a quick reduction of Kast spectra, I've been using a series of IRAF packages. First, I combine the bias frames together and flat frames together using the task "combine". I then remove the bias_comb image from my science frames using the task "imarith." I then flatten my science frames using ccdproc. Finally, I extract my spectrum using apall. I found these tutorials to be very helpful.
Our first science target looks like this:
I still need to do the flux calibration (using a spectro-photometric standard star taken from the Lick website) and the wavelength solution (using the arclamps). But you can already see that there is Ha in emission (the large peak around pixel #600). The sodium absorption features are all the way toward the left hand side (near pixel#50). Since the street lights of San Jose are largely sodium emission lamps, they contaminate our sodium absorption lines to the point of being useless. Tomorrow night we will change our grating tilt angle so that we get more of the continuum past the TiO lines on the red side, thus sacrificing our measurements of the Na I doublet. The lithium doublet (~6707A) is separated from the Ha line by ~125 pixels. It is not obvious in this star. You can see the TiO features which start at pixel #1000 with a weaker TiO feature starting around pixel #700.
The wobbly continuum (compared with the spectrum from Song et al. 2012) is expected for an M type star. This is due to the fact that since M stars have a cooler atmosphere, they are home to molecules, which cannot form in the hotter environments of earlier type (F, G, and K) stars. Since molecules have more ro-vibrational transition lines available to them, the spectrum of a molecule-rich M star has MANY absorption features which, when packed closely together, cause a suppression of the continuum (which looks like one giant absorption feature). The most obvious of these is due to the TiO molecule.
Parallactic Angle - Making sure your star is ACTUALLY in the slit
Although it's not much of a problem at the wavelength's we're working with, it is important to note that the peak sensitivity of the guide camera is ~6000A which can differ from the wavelengths of your observations. In this case, although the star may be in the slit on the guider window, the emission at the wavelengths of interest may be centered outside of the slit. This is due to the fact that the atmospheric refraction is wavelength-dependent. To get around this problem you want to calculate a parallactic angle (angle between the zenith and a vector pointing North) and rotate the telescope by that angle. This will ensure that the slit lies vertical relative to the refraction axis (that is, vertical relative to the horizon). That is the direction along which the airmass of your observations increases during the length of your exposure. The folks at Lick have a handy script written which calculates the parallactic angle (=position angle) appropriately called "pa." This script just needs you to input the hour angle at the start of your observation, the exposure time for the star in question, and the declination of the star in question. Yay science!
BOGO - Measuring two spectra at once
Some of our stars turned out to be binaries (or at least, they were close together in the sky and had similar proper motions in the literature). In this case, we rotated the slit until both stars were in the slit, and exposed for long enough to get good S/N on the less bright star (but not too long, lest we reach saturation ~65,000 counts or non-linearity at ~40,000 counts). When we did this, our raw spectrum looked like this:
We reduced the file twice, extracting each spectrum separately. BOGO! In this case, it turned out that both components were M stars, and both had strong Ha in emission. This supports the case that they are a true binary (and not just a visual double).
**UPDATE**
Having (roughly) reduced the spectra from night 1, it looks like we found 2 stars (out of ~25) with strong Li detections. One of our stars has a double-peaked Ha feature:
This can mean one of two things: either there is an accretion disk around this star, or there are turbulent chromospheric motions causing a broadening of the Ha emission feature. So this star is either extremely chromospherically turbulent, or it's so young that it's still accreting matter onto the host star. Interesting either way!
No comments:
Post a Comment