Observations of GRB 230307A by TESS

We present the TESS light curve of GRB 230307A. We find two distinct components: a bright, prompt optical component at the time of the Fermi observation that peaked at TESS magnitude 14.49 (averaged over 200 seconds), followed by a gradual rise and fall over 0.5 days, likely associated with the afterglow, that peaked at 17.65 mag. The prompt component is observed in a single 200s Full Frame Image and was undetectable in the next TESS image ($T_{\rm mag}>17.79$). Assuming that the onset of the optical transient was coincident with the gamma-ray emission, the prompt emission lasted less than 73.6 seconds, which implies the true peak was actually brighter than $T_{\rm mag} =$ 13.40. We also fit parametric models to the afterglow to characterize its shape. The TESS light curve can be retrieved at https://tess.mit.edu/public/tesstransients/light_curves/lc_grb230307A_cleaned.txt.

• March 12, 6:40 UTC-GRB data are made public on MAST, announced on the MAST holdings page, 2 and announced on MAST social media.
Once the data were public, we processed the images with the difference imaging analysis pipeline described by Fausnaugh et al. (2021).
We identified a transient point source in the TESS difference image overlapping with the Fermi trigger at BJD−2, 460, 000=11.15518 days, close to the coordinates reported by Levan et al. (2023). We measured the position of the source using flux-weighted centroids in a 5x5 square pixel aperture. We found a position of RA = 04:03:25.36, Dec = −75:22:41.31, which agrees with the position from ULTRACAM to 1.8 arcseconds. This difference is consistent with the typical 1σ precision of TICA World Coordinate Solutions, about 2 arcseconds.
We extracted light curves from the difference images using forced PSF photometry at the location of the afterglow reported by Levan et al. (2023). Figure 1 shows the resulting light curve. Based on the scatter of the light curve prior to the GRB discovery, we estimate a 5σ limiting magnitude of 18.63 in 1600 seconds. The TESS light curve is available at https://tess.mit.edu/public/tesstransients/light_curves/lc_grb230307A_ cleaned.txt. Information on the file format and processing steps is available on the same website. 3 The light curve shows two distinct components. First is a prompt rise from zero flux to TESS mag T mag = 14.49 ± 0.05 in a single FFI at BJD−2, 460, 000 = 11.15518 days (mid exposure). The quoted uncertainty includes a statistical uncertainty of 0.02 mag, but is dominated by uncertainty in the TESS instrument absolute flux calibration (Vanderspek et al. 2018). The source is not detected in the next FFI, with a 3σ upper limit of 17.79 mag. The second emission component starts just after the prompt emission, and consists of a gradual rise and fall in the TESS light curve over 0.5 days, likely associated with the afterglow.

ANALYSIS OF THE AFTERGLOW
We fit two parametric models to the afterglow light curve, a fast-rise exponential-decay (FRED) model and a double power law model. We fit the original data sampled at 200 seconds. We excluded the prompt emission at T mag = 14.49 from the fits to the afterglow emission. The reported uncertainties correspond to the 68% confidence intervals of the fits. For both models, we fit a residual background error as a nuisance parameter, which is consistent with zero at 1σ. From the FRED model, we find that • the peak magnitude of the afterglow is 17.65 ± 0.06, • the rise-time t rise = 0.057 ± 0.007 days, • the time containing 90% of the flux t 90 = 0.5 ± 0.1 days, • and the exponential decay timescale t exp = 0.23 ± 0.03 days.
For the double power law model, we used the parameterization of Li et al. (2012): where F 02 is the flux normalization at the break time t break , while α 1 , and α 2 are slope parameters. We fixed the smoothing parameter ω to a value of 5.0. The best-fit model yields • t break = 0.09 ± 0.02 days after the Fermi trigger, • α 1 = 0.6 ± 0.2, • and α 2 = −0.21 ± 0.07.
We also fit the TESS light curve with the deep r-band magnitudes reported by Levan et al. (2023) andO'Connor et al. (2023), and found consistent results.

DISCUSSION
The end time of the FFI with the prompt emission is BJD−2, 460, 000=11.15634 days, while the Fermi trigger is at BJD−2, 460, 000=11.15549 days. Assuming that the onset of the optical light curve is coincident with the onset of the gamma-ray emission observed by Fermi, the prompt emission observed by TESS lasted less than 73.6 seconds. A shorter duration implies a higher peak magnitude for the same fluence; we find T mag =13.40 mag for a 73.6 second burst, and the peak could be even brighter if the duration of the prompt emission is shorter. For a given luminosity distance D, the isotropic monochromatic luminosity at 7839 Å (the pivot wavelength of the TESS filter) would be greater than 9 × 10 44 (D/1 Gpc) 2 erg s −1 . This calculation is corrected for Galactic extinction at 7839 Å assuming a Cardelli, Clayton, & Mathis 1989 extinction law and E(B −V ) = 0.0758 mag from Schlafly & Finkbeiner 2011, but does not take into account the unknown spectral shape of the GRB in the broad TESS filter (6,000-10,000 Å).
These data demonstrate the utility of wide-field, continuous monitoring for studies of fast extragalactic transients. TESS data released approximately every 7 days as TICA High Level Science Products will facilitate these studies over the next several years. Flux (e s ¡1 ) Figure 1. TESS light curve of GRB 230307A. The large star shows the prompt emission coincident with the Fermi discovery at BJD−2, 460, 000=11.15549 days (vertical red line). The grey marks show the 200 second cadence light curve, and the black points show the light curve binned to 1600 seconds. The limiting magnitude in 1600 seconds is 18.63, based on the scatter of the light curve prior to the GRB discovery. All timestamps are given relative to 2,460,000 BJD. Fits for two models of the afterglow are shown: a double power law (PL) and a fast-rise exponentialdecay (FRED) model. For both fits, we excluded the prompt emission at the time of the Fermi trigger. Models were fit to the original light curve sampled at 200 seconds-data are binned here only for display purposes. The horizontal dashed line shows the residual background of the FRED model, which is a nuisance parameter in our fits. The vertical blue line shows the break time t break of the double power law. The TESS light curve is available at https://tess.mit.edu/public/tesstransients/light_curves/lc_grb230307A_cleaned.txt.