Subject: Results of Spectroscopy Survey

From: Bernadette Rodgers

Submitted: Tue, 22 Apr 1997 17:05:36 -0700 (PDT)

Message number: 41 (previous: 40, next: 42 up: Index)

Thanks to those that responded to my spectroscopy query.
There were 5 responses.  Because of the small number, I've gone 
ahead and listed the respondees and their comments at the end of 
this message.  (I hope that none of you mind.) 

The general response was that noone has tried spectroscopy with GRIM 
much, and those that have gave up after initially dismal results.  
The reasons:

	Insufficient instrument characteristics:  Yup, too low-res.
	Telescope faults (ie, tracking too poor to stay on slit)_X__
		(Lack of an auto-guider was the killer.)
	Lack of Documentation X
	Other __X__ JHU has joined ARC only recently 

Compared to this, my experience has been relatively positive.
Therefore, I will bias this report by putting my own response first,
and commenting briefly on other responses.  

My general comment is that GRIM spectroscopy is do-able for the
right types of projects- basically low resolution ones.
The resolution is about 400 at f/10 and 800 at f/20 (3 pixel wide slit).  
I hope to put together a web page shortly offering info based on 
my experience, and eventually contribute to Chapter 6 of the manual.  

If anyone has further questions/comments/suggestions concerning
GRIM II Spectroscopy, please contact me.  I'd love to hear from you.


____________________SURVEY RESPONSES____________________
Respondees:  Bernadette Rodgers, Bernie Rauscher, James Rhoads, 
Michael Strauss, Zlatan Tsvetanov, Alan Watson

Project: monitoring variability of young stars (mostly Herbig Ae/Bes,
some T Tauris).  Bright (K=7) sources; using f/10, f/20, J, H and K

I have been able to obtain reasonable, high S/N spectra at f/10 in H 
and K.  I have also tried f/20, J filter, and slitless modes, with 
some success.  f/20 is feasible WITH decent guiding; J filter is
feasible at least for strong lines, ie Pa beta (the overlapping
orders is not a significant problem); slitless mode is useful for very
bright objects (max. exp <15 secs) and low resolution (R~250 with 1.2" seeing).
Wavelength calibration using skylines is doable at f/10 with 20s exposures
in H, 30s in K (easy with 60 sec exposures).   Sky saturation is not
an issue in spectral+slit mode (sky counts<50 in 120 sec exp.).

Besides weather, the biggest problems I've had is with calibration and
lack of documentation.  The lamps are not particularly helpful, mainly 
because long exposures are needed in order to see anything.  Spectral 
flats using the bright quartz lamp require 60 sec. exposures to get 
1200 counts above bias at f/10 (120sec only gets ~125DN at f/20!).  I gave up 
on lamps for wavelength cal, but find that sky lines are sufficient anyway.
I have run into a couple technical problems which I will report on
when I know more (banding at f/20; quadrant gain differences?).

R~1000 spectroscopy in Herbig-Haro Objects.f/20 spectra in J, H, and K
bands. This project required positioning the slit within 1 arcsec
and fixing the rotation within a few degrees. As the exposures were >
5 min. long, the lack of an appropriate autoguider rendered the
project undoable.  

COMMENT: f/20 is feasible now if there is a good guide star in the guider
field.  A wider field guide camera will soon be on the telescope.
Max. resolution at f/20 I believe is about 850.

 I once considered doing K band spectroscopy of bright knots in galaxies
to back up some photometric CO index measurements I was doing; I gave it
up on the advice of another observer who thought that the throughput
and resolution available did not make GRIM a competitive instruement
for spectroscopy.  This was in late 1994 / early 1995, and things
may have improved since then. 
I wanted to do moderate-resolution spectroscopy of ultraluminous IRAS
galaxies to look for faint broad wings to Pa alpha.  The resolution of
GRIM was completely inadequate for the task.  

COMMENT: Pa alpha @ 1.87um is outside of either the H or K bandpasses and 
is hard to get from the ground anywhere because of atmospheric absorption.
Pa beta is possible at f/20 in J band: full-width velocity resolution ~370 km/s.
Perhaps you could see wings for velocities > ~200 km/s.

Tsvetanov: [JHU- hasn't had a chance to try it yet]
low-to-mid resolution spectroscopy of nearby AGN; combination 
for my project - f/10, H and K  

Jon Holtzman and I tried to do CO absorption spectroscopy of stellar
clusters in a starburst galaxy. We used f/10 in K. We sent a report on
our lack of success to the GRIM mailing list about 18 months ago.
Basically, the show-stopper for us was the high read noise. 

COMMENT: Read noise hasn't changed, but throughput should have
increased with realuminization.  However, I find a peak signal of 
450 DN/sec/pix for a K=6.45 star at f/10 in photometric (~.7") seeing.  
This is compared to 500 DN/s/pix reported by Alan in Oct. 95 (msg#2).  
I was not guiding or worrying too greatly about lining up on the slit, 
but this is the _highest_ count I got for this standard star.  
Perhaps we are calculating this differently??

APO APO APO APO APO  Apache Point Observatory 3.5m  APO APO APO
APO  This is message 41 in the apo35-grim archive. You can find
APO  the archive on
APO  To join/leave the list, send mail to
APO  To post a message, mail it to