Next Contents Previous

5.7. Goddard High Resolution Spectrograph (GHRS)

The GHRS *(8) was one of the two original spectrographs on board HST. It was replaced during the second servicing mission. The GHRS obtained high resolution spectra between 1150 Å and 1800 Å with resolving power between R = 2000 (for the "L") gratings and 25000 (for the "M") on one of its sides; while for the other the resolving powers ranged between 25000 and 80000 and the wavelength range between 1680 Å and 3200 Å. Each of these sides had its own detector (a Digicon) with 512 diodes each. The central 500 diodes were the ones used to obtain the spectra, while the first and last six (the so-called "special diodes") collect information about the observation itself.

The GHRS had two apertures: the Large Science Aperture (LSA) with a pre-COSTAR size of 2.0" (1.74" post-COSTAR) and the Small Science Aperture (SSA) with a pre-COSTAR size of 0.25" (0.22" post-COSTAR). Almost all the light from a point source was captured by the LSA. The LSA was used to obtain precise fluxes, while for the determination of radial velocities or line profiles it was the SSA.

The seven gratings, echelles and mirrors (used for the acquisitions) are mounted on a movable carrousel. Depending on the configuration, either an image of the aperture, a one or two-dimensional spectrum was obtained.

Due to the granularities of the photocathode (that limit the signal to noise) spectra were obtained divided in two or four parts, moving slightly the carrousel each time. This procedure is called FP-SPLIT. The final spectrum was obtained by combining them.

The size of the GHRS diodes is larger than the PSF. To re-gain the actual resolution it was necessary then to oversample the spectrum, ie. to allow the light to lay over several diodes. This was done by diverting the beam a quarter or half a diode and reading the spectrum after each movement. The number of pixels in the calibrated spectrum depended how these deviations were made. The raw spectrum always has 500 (no deviations made), 1000 (0.5 diode shift) or 2000 (0.25 diode shift) pixels.

5.7.1 Gratings and echelles

The first order GHRS gratings are listed in Table 21. The width listed corresponds to the number of Å that were obtained per observation.

Table 21. GHRS gratings

Name Range (Å) Å/diode Width (Å) Side

G140L 1100-1900 0.572-0.573 286-287 1
G140M 1100-1900 0.056-0.052 28-26 1
G160M 1150-2300 0.072-0.066 36-33 2
G200M 1600-2300 0.081-0.075 41-38 2
G270M 2000-3300 0.096-0.087 48-44 2

5.7.2 What constitutes a GHRS observation?

An "observation" with the GHRS is composed of several pairs of archives, whose name always start with Z. The header, as usual, is a text archive whose extension ends with h and the binary data's extension with d.

The different files, in GEIS format, that constitute an "observation" are briefly described in the following table. Please refer to the HST Data Handbook for more details.

Note that a calibrated "spectrum'" is composed of two pairs of files:

GHRS extensions

Extension Contents

Uncalibrated Data

d0h /d0d uncalibrated science data
q0h /q0d data quality file
x0h/x0d data from the special diodes
d1h/d1d acquisition data (return to brightness mode)

Calibrated Data

c0h /c0d wavelength-calibrated data
c1h/ c1d  flux-calibrated data
c2h/c2d statistical errors
c3h/c3d special diodes calibrated data

The different exposure types are described by the OBSMODE header parameter, which are shown in Table 22.

Table 22. GHRS OBSMODE types.

OBSMODE exposure type TARGNAME

DEFCAL centering of the source, no spectrum source name
SPYBAL calibration lamp spectrum WAVE
ACQ/TARG acquisition source name
IMAGE image of the photocathode source name
ACCUM spectrum count integration source name
RAPID rapid acquisition without accumulation source name

TARGNAME is the name with which the source is referred to in the header. If a special calibration was requested, the keywords are DATE-OBS = ACCUM and TARGNAME = WAVE.

Table 23 describes some of the header parameters of interest from a GHRS observation:

Table 23. GHRS keywords.

APERTURE aperture used (LSA, SSA)

RA_APER1, DECAPER1 aperture position
EXPOSURE exposure time per group
EXPSTART exposure time in MJD
EXPTIME total exposure time
FP_SPLIT FPSPLIT mode (2 or 4)
GCOUNT number of groups
NAXIS1 number of pixels in the spectrum
RPTOBS number of repeated observations
STEPPAT substepping mode
OBSMODE observing mode

5.7.3 Working with GHRS spectra

If the spectra were obtained with an FP-SPLIT before starting the analysis, it is necessary to combine them using the STSDAS tasks poffsets and specalign. The spectrum is still divided in two: the wavelength and flux calibrated parts. To combine them and proceed with the study, the mkmultispec task is used.

The wavelength calibration has an RMS error of 1.25 diodes. If additional data was obtained (like SPYBALS, for example) it is possible to reduce this number. Additionally the task wavecal can be used to obtain new dispersion coefficients and zero-point differences.

* Archival instrument.
8 A detailed description can be found in the GHRS Instrument Handbook, version 6, May 1995. Available at Back.

Next Contents Previous