STScI Logo

Hubble Space Telescope
COS now offers resolutions of 8,000 - 12,000 between 900 and 1100

STScI is pleased to announce that the spectral resolution of the COS G130M 1055 and 1096 central wavelength settings has been dramatically improved. When combined with the short wavelength capability of the G130M 1222 setting, these changes now allow observations in the 900 - 1100 angstrom range with spectral resolutions between 8,000 and 12,000, a capability that has not been available since the FUSE satellite ceased operations in 2007.

The dramatic improvement in the resolution of the short wavelength settings of the G130M grating was made possible by changing the focus to values not used previously on orbit. The new focus values will apply by default to any 1055/1096 observations executing after 10 December 2012. Previous cycle observations that were kept "on-hold" pending a final decision on the focus values will be released for execution any time after 10 January 2013 unless the PI makes other arrangements with STScI. Please contact STScI by 14 December 2012 if you have any concerns about such pending observations. However, the full set of reference files needed to properly calibrate data taken using the 1055 and 1096 central wavelength settings in the OTFR pipeline may not be available until later in the first quarter of 2013. As a result, there may be delays in calibrating and archiving data taken using these settings.

Users are warned that the best resolution for these settings is available only over a narrow portion of the wavelength range for each grating (see figure 1), and so obtaining the highest possible resolving power over the full wavelength range may require the use of multiple overlapping CENWAVE settings.

Figure 1: The predicted spectral resolving power (R = λ / FWHM) as a function of wavelength for each segment of the COS FUV detector is shown for the G130M grating at the 1055, 1096, 1222, and 1291 central wavelength settings. The results for the B segment of the COS G140L grating are shown for comparison. These predictions are based on CodeV optical models that include the MFWFEs of the HST primary and secondary mirrors. The model predictions have been confirmed only for a limited set of wavelength regions.

These improvements in resolution were made possible by adjusting the focus mechanism positions for these settings to a value well outside those previously used on-orbit or well tested on the ground. Considerable analysis was necessary before it could be demonstrated that this did not present a risk to the instrument.

Because of the late decision to allow the focus values needed for this improvement, the version of the COS Exposure Time Calculator (ETC) released for phase 1 of cycle 21 (v 21.1) does not yet reflect these changes in spectral resolution. As a result it will be necessary to manually scale results calculated using ETC version 21.1. While the throughput and wavelength coverage of the 1055 and 1096 have not significantly changed, the size of the resolution element has decreased substantially. Because ETC calculations report the S/N per resolution element, the time computed to reach a given S/N will need to be adjusted by up to a factor of four to take into account the smaller resolution element. Another side effect of the increased resolution is the increase in the cross-dispersion height of the spectra, leading to an increase in detector background over the spectral extraction region. This can affect in particular ETC calculations for faint targets. Additional details on how to scale the ETC results to account for the effects described above are available at:

For observers who initially proposed to use these settings at lower spectral resolution, it should usually be possible to recover their planned S/N by binning or smoothing the spectra back to the originally planned resolution. Users should note however that for some background-limited observations the increased detector background under the more extended spectral height might prevent the original S/N from being reached.