Discovery of a Supernova Explosion at Half the Age of the Universe and its Cosmological Implications

Abstract

The ultimate fate of the universe, infinite expansion or a big crunch, can be determined by measuring the redshifts, apparent brightnesses, and intrinsic luminosities of very distant supernovae. Recent developments have provided tools that make such a program practicable: (1) Studies of relatively nearby Type Ia supernovae (SNe Ia) have shown that their intrinsic luminosities can be accurately determined; (2) New research techniques have made it possible to schedule the discovery and follow-up observations of distant supernovae, producing well over 50 very distant (z = 0.3 -- 0.7) SNe Ia to date. These distant supernovae provide a record of changes in the expansion rate over the past several billion years. By making precise measurements of supernovae at still greater distances, and thus extending this expansion history back far enough in time, we can distinguish the slowing caused by the gravitational attraction of the universe's mass density Omega_M from the effect of a possibly inflationary pressure caused by a cosmological constant Lambda. We report here the first such measurements, with our discovery of a Type Ia supernova (SN 1997ap) at z = 0.83. Measurements at the Keck II 10-m telescope make this the most distant spectroscopically confirmed supernova. Over two months of photometry of SN 1997ap with the Hubble Space Telescope and ground-based telescopes, when combined with previous measurements of nearer SNe Ia, suggests that we may live in a low mass-density universe. Further supernovae at comparable distances are currently scheduled for ground and space-based observations.

Figures (5)

• Figure 1: Spectrum of SN 1997ap, after binning by 12.5 Å, placed within a time series of spectra of "normal" SNe Ia $^{17,18,19,20,21}$ (the spectrum of SN 1993O was provided courtesy of the Calán/Tololo Supernova Survey), as they would appear redshifted to $z = 0.83$. The spectra show the evolution of spectral features between 7 restframe days before and 2 days after restframe $B$-band maximum light. SN 1997ap matches best at $2\pm 2$ days before maximum light. The symbol $\oplus$ indicates an atmospheric absorption line and * indicates a region affected by night sky line subtraction residuals. The redshift of $z = 0.83 \pm 0.005$ was determined from the supernova spectrum itself, since there are no host galaxy lines detected.
• Figure 2: Photometry points for SN 1997ap (a) as observed by the HST in the F814W filter; (b) as observed with ground-based telescopes in the Harris $I$ filter; and (c) as observed with the ground-based telescopes in the Harris $R$ filter (open circles) and the HST in the F675W filter (filled circle); with all magnitudes corrected to the Cousins $I$ or $R$ systems$^{13}$. The solid line shown in both (a) and (b) is the simultaneous best fit of the ground- and space-based data to the $K$-corrected, ($1+z$) time-dilated Leibundgut $B$-band SN Ia template light curve$^{22}$, and the dotted line in (c) is the best fit to a $K$-corrected, time-dilated $U$ band SN Ia template light curve. The ground-based data was reduced and calibrated following the techniques of ref 5, but with no host-galaxy light subtraction necessary. The HST data was calibrated and corrected for charge transfer inefficiency following the prescriptions of refs. 23 and 24. K-corrections were calculated as in ref 25, modified for the HST filter system. Correlated zeropoint errors are accounted for in the simultaneous fit of the lightcurve. The errors in the calibration, charge transfer inefficiency correction and K-corrections for the HST data are much smaller ($\sim$4% total) than the contributions from the photon noise. No corrections were applied to the HST data for a possible $\sim$4% error in the zeropoints (P. Stetson, private communication) or for non-linearities in the WFPC2 response$^{26}$, which might bring the faintest of the HST points into tighter correspondence with the best fit lightcurve in (a) and (c). Note that the individual fits to the data in (a) and (b) agree within their error bars, providing a first-order cross check of the HST calibration.
• Figure 3: SN 1997ap at $z=0.83$ plotted on the Hubble diagram from ref 5 with the five of the first seven high-redshift supernovae that could be width-luminosity corrected and the 18 of the lower-redshift supernovae from the Calán/Tololo Supernova Survey that were observed earlier then 5 days after maximum light. Magnitudes have been $K$-corrected and corrected for the width-luminosity relation. The inner error bar on the SN 1997ap point corresponds to the photometry error alone while the outer error bar includes the intrinsic dispersion of SNe Ia after stretch correction. The solid curves are theoretical $m_B$ for ($\Omega_{\rm M}$, $\Omega_\Lambda$) = (0, 0) on top, (1, 0) in middle, and (2, 0) on bottom. The dotted curves are for the flat universe case, with ($\Omega_{\rm M}$, $\Omega_\Lambda$) = (0, 1) on top, (0.5, 0.5), (1, 0), and (1.5, $-$0.5) on bottom.
• Figure 4: [ Color Version] Contour plot of the 68% (1$\sigma$) and 90%, confidence regions in the $\Omega_{\Lambda}$ versus $\Omega_{\rm M}$ plane, for (blue shading) the five supernovae at $z \sim 0.4$ (see ref 5); (yellow shading) SN 1997ap at $z=0.83$, and (red contours) all of these supernovae taken together. The two labeled corners of the plot are ruled out because they imply: (upper left corner) a "bouncing" universe with no big bang$^{27}$, or (lower right corner) a universe younger than the oldest heavy elements, $t_0 < 9.6$ Gyr$^{28}$, for any value of $H_0\ge 50$ km s$^{-1}$ Mpc$^{-1}$.
• Figure 4: [ Black and White Version] Contour plot of the 68% (1$\sigma$) and 90%, confidence regions in the $\Omega_{\Lambda}$ versus $\Omega_{\rm M}$ plane, for the five supernovae at $z \sim 0.4$ (see ref 5), SN 1997ap at $z=0.83$, and (shown as dark, unfilled contours) all of these supernovae taken together. The two labeled corners of the plot are ruled out because they imply: (upper left corner) a "bouncing" universe with no big bang$^{27}$, or (lower right corner) a universe younger than the oldest heavy elements, $t_0 < 9.6$ Gyr$^{28}$, for any value of $H_0\ge 50$ km s$^{-1}$ Mpc$^{-1}$.