overview
The UVOT Clocked Grism mode is the best mode for the instrument. The reasons are that it covers the ultraviolet part of the spectrum from about 1700 A upward, and that due to the special 'clocked' position it enjoys both a lower background, as well as no zeroth orders in the upper left half of the detector.The accuracy of the first order wavelength calibration is discussed in this page .
The second order separates from the first order in the upper right hand and lower left corners, while it overlaps the first order to some degree in the middle parts of the detector. The dispersion of the second order was determined using the WR spectra to calibrate. Details of the second order calibration can be found here .
The flux calibration of first and second order is dependant on both. Both white dwarfs and F-type stars were used in order to calibrate the separate contributions to the flux. Find out how here .
The background in the grisms is high enough to experience coincidence loss. The correction factor to the flux is in the range 1.0-1.10 using the usual method (see also Breeveld et al. 2011 for a discussion of the applicability of the standard correction method for point sources to the background). For the flux calibration/effective area determination images with a background with a correction factor higher than 1.05 were not used.The method described in the main flux calibration page was used. The individual effective area curves at the centre of the detector is shown in the figure below. The following sources were used: WD1657+343, WD1057+719 below 2750A, G202-65 from 2500-3800A, and P177D, P041C above 3100A. The effect of overlapping second order emission is minimized by that choice, but above 4900A can not be avoided. A redder calibration source will be required for that.
The curves at the bottom have been corrected using the algorithm described on the main flux page. The weighted mean of the curves and the RMS error derived from the spread of the individual curves is shown more clearly in the figure below.
The curve is very similar to that found for the nominal UV grism. The peak at 2750 is more pronounced and may partly be due to the onset of second order emission in the WD spectra. The error is larger than in the nominal grism, which is due to a smaller total exposure time in the available calibration images after filtering out data that did not qualify. An additional error above 4900A has been added ad-hoc to show the expected second order contamination.
