16515

The CALSPEC archive contains the SEDs of the standard stars that will be used for the absolute flux calibrations of JWST comma Roman comma and ground-based telescopes. The CALSPEC SEDs are based mostly on STIS. While the provenance of the three primary standards (G191B2B comma GD153 comma and GD71 with V=11.8-13.3) and brighter stars is robust comma the precision of fainter stars depends on substantial corrections for CTE losses. These fainter stars will be the primary standards for COS comma JWST comma Roman comma and many modes of modern and large telescopes. However comma the precision of these flux distributions in the STIS G430L and G750L CCD modes is limited by the applicability of the old empirical CTE loss correction of Goudfrooij et al. (2006) comma where the time dependence is now extrapolated virgul15 years past the date of derivation. These longer STIS wavelengths are also most important in constraining the extrapolation into the mid-IR for faint JWST standard stars. In particular comma the STIS SED for SDSSJ151421 shows indications of a virgul2% discrepancy with respect to WFC3 photometry (Bohlin et al. 2020 comma AJ comma 160 comma 21 comma Fig. 14). The amount of the SDSSJ151421 CTE correction for G750L ranges from 7-20% over 5500-10000A. Thus comma even the estimated uncertainty of 20% of the correction exceeds 2% at wavelengths longer than 8000A comma where a 10% CTE loss is predicted. Therefore comma to achieve the difficult goal of 1% accuracy for faint standards comma the antique STIS CTE correction formula must be upgraded. The bright standard BD+75 325 provides the best target for measuring CTE losses at the present epoch. STIS SED measures with excellent S-N exist at the time of Epoch-0 in 1997 comma so errors in the current correction scheme are quantified by comparisons to new observations with signal levels relevant to the important Primary Faint Standards (PFS) WD1057+719 comma WD1657+343 comma WD0308-565 comma and SDSSJ151421 for which 24 STIS orbits have been allocated to upgrading the S-N of their SEDs in Cycles 27-28. Because of the different CTE loss correction at the longer G750L wavelengths where the halo from scattering in the CCD substrate fills some traps comma calibration observations in both G430L and G750L are required. G230LB should have the same correction formula for CTE losses as G430L. The range of signal for G430L and G750L for the observations of these four PFS stars can be efficiently bracketed by short exposures of BD+75 325 that have similar CTE corrections. This signal range is summarized at wavelengths of minimum and maximum signal among the PFS data. The signal is the actual observed net in electrons-s times the typical exposure time in sec for a single member of the CR-split set. The minimums are for observations of the faintest PFS comma WD1657+343 comma while maximums are for the brightest WD0308-565. Single frame exposure times are from cycle 28 ID=16437 comma i.e. for G430L and G750L comma respectively comma 208 comma 428s for WD1657 (Min) and 170 comma 500s for WD0308 (Max). TABLE 1 Signal from WD1657+343 (min) and WD0308-565 (max) WL (A) min (e) max (e) ------ ------- ------- 3100 700 3050 4200 1150 7000 5800 1400 14000 9000 180 2050