Source-linked AI summary
The host galaxies of core-collapse supernovae and gamma ray bursts
K. M. Svensson, A. J. Levan, N. R. Tanvir, A. S. Fruchter, L. -G. Strolger
TL;DR
The study asks how GRB and CCSN host environments differ and how those differences affect using GRBs as probes of star formation. Using multiwavelength host data, it finds that GRBs preferentially occur in smaller, more actively star-forming galaxies and in brighter, more concentrated local regions than CCSN.
Problem
The study compares GRB and CCSN environments to constrain their progenitors and assess biases in using GRBs as probes of star formation.
Method
The authors fit spectral energy distributions to multiwavelength photometry of GRB and CCSN hosts to estimate magnitudes, stellar masses, star formation rates, sizes, morphologies, and metallicities.
Results
GRB hosts are consistently smaller, concentrated in the low-mass high-SSFR regime, and more concentrated on host light and high absolute surface brightness than CCSN hosts.
Takeaways & Limitations
The differing galactic and local environments support models in which GRBs form only in certain environments and provide constraints on GRB progenitors and cosmological-probe biases.
Takeaways & Limitations
GRB host redshifts are often obtained from emission or absorption lines, and emission-line redshifts generally favor brighter hosts, introducing a host-magnitude selection effect.
Abstract
from arXiv · showhide
We present a comparative study of the galactic and small scale environments of gamma-ray bursts (GRB) and core collapse supernovae (CCSN). We use a sample of 34 GRB hosts at z<1.2, and a comparison sample of 58 supernova hosts located within the Great Observatories Origins Deep Survey footprint. We fit template spectra to the available photometric data, which span the range 0.45-24 micron, and extract absolute magnitudes, stellar masses and star formation rates from the resulting fits. Our results broadly corroborate previous findings, but offer significant enhancements in spectral coverage and a factor 2-3 increase in sample size. Specifically, we find that CCSN occur frequently in massive spirals (spiral fraction ~50%). In contrast GRBs occur in small, relatively low mass galaxies with high specific and surface star formation rates, and have a spiral fraction of only ~10%. A comparison of the rest frame absolute magnitudes of the GRB and CCSN sample is less conclusive than found in previous work, suggesting that while GRB hosts are typically both smaller and bluer than those of CCSN their total blue light luminosities are only slightly lower. We suggest this is likely due to rapid periods of intensified star formation activity, as indicated by the high specific star formation rates, which both create the GRB progenitors and briefly significantly enhance the host galaxy blue luminosity. Finally, our analysis of local environments of GRBs and CCSN shows that GRBs are highly concentrated on their host light, and further occur in regions of higher absolute surface luminosity than CCSN.
1 INTRODUCTION
CCSN and GRBs can trace massive-star formation, but their host environments may differ in ways that constrain GRB progenitors and observational biases. This study compares their global and local host properties using multiwavelength data at comparable redshifts.
- CCSN and long-duration GRBs both trace massive-star formation because each requires a single progenitor more massive than ∼8M⊙.
- Supernova searches beyond z ∼1 remain technologically limited, motivating CCSN studies as future calibrators of GRB environmental dependencies.
- Dust extinction can bias detections of both CCSN and GRB optical afterglows, although GRB γ-rays and X-rays are less affected.
- Previous work suggests GRBs favor lower-metallicity environments, but metallicity varies within hosts and global host evidence remains incomplete.
- The paper compares CCSN and GRB hosts at similar redshifts using optical-to-mid-infrared data to derive masses and star formation rates without requiring evolutionary corrections.
2 HOST GALAXY SAMPLES
The study assembles CCSN and GRB host samples at broadly comparable redshifts, alongside a GOODS-MUSIC field-galaxy comparison sample. Selection and redshift-measurement differences remain important caveats for interpreting host-property comparisons.
- 58 CCSN hosts come from blind GOODS/PANS searches, while GRB hosts are restricted to z < 1.2 to approximately match the CCSN redshift distribution.
- CCSN typing combines supernova and host photometric or spectroscopic data, with Gold, Silver, and Bronze medals reflecting classification confidence.
- Faint GRB hosts lacking reliable redshifts may be excluded, while emission-line redshifts can favor brighter hosts and bias the observed GRB population toward higher luminosity.
- The GOODS-MUSIC comparison contains ∼6900 non-stellar, non-AGN field galaxies at 0.1 < z < 1.2, selected primarily in ACS z-band with Ks-band supplementation.
- The field-galaxy sample is magnitude-limited and not selected for high star formation, unlike transient-selected CCSN and GRB hosts.
3 PHOTOMETRY
Host photometry combines HST, ground-based near-infrared, and Spitzer observations, extending coverage across optical-to-mid-infrared wavelengths. The dataset includes new GRB-host IRAC measurements but is limited by source blending.
- Photometry spans up to 12 bands from ACS B, V, I, Z through VLT J, H, K and Spitzer 3.6–24 µm measurements.
- ACS photometry is cross-checked against GOODS catalog values, while near-infrared measurements are checked using automated and manual aperture photometry.
- Host radii from the GOODS catalogue are converted into physical sizes using an assumed ΛCDM cosmology.
- Spitzer imaging provides measurements for 56 CCSN hosts and 26 new 3.6–8.0 µm GRB-host magnitudes or limits.
- IRAC blending prevents reliable photometry for some galaxies, leaving their catalogue entries blank.
4 SPECTRAL ENERGY DISTRIBUTION FITTING
The authors fit reddened template spectral energy distributions to broad multiwavelength photometry and use the best fits to estimate rest-frame magnitudes and host physical properties. The fitting incorporates redshift information and internal-extinction corrections.
- Photometry from 0.4 µm to 24 µm is fit with template spectral energy distributions to estimate host physical parameters.
- CCSN redshifts are spectroscopic in 41 cases and photometric in 17, supporting the SED fitting of the comparison sample.
- The best fit minimizes χ² with respect to template scaling b and reddening A_v while adopting a fixed reddening law.
- The optimized template is transformed to the rest frame and used to derive absolute magnitudes, stellar masses, star formation rates, and metallicities.
- Internal-extinction corrections from the fitting procedure are applied when estimating host properties.
5 DERIVING PHYSICAL PARAMETERS
The study derives host-galaxy physical properties from multi-wavelength photometry, using K-band luminosity for stellar mass and U-band luminosity for star formation, while noting IMF and young-population caveats.
- Stellar masses: K-band luminosity estimates stellar mass because it mainly samples older stars, although red supergiants can overestimate masses in young populations.The calibration is especially cautioned for galaxies dominated by young to intermediate-aged stars.
- Star formation rates: U-band luminosity traces hot, massive, newly formed stars and is used to estimate star formation rates.The conversion includes a factor of 8.8 to account for star formation above 5 M⊙.
- Assumptions: The inferred star formation rates assume a Salpeter initial mass function, so strong IMF deviations could bias both SFR and stellar-mass estimates.The authors note that low-mass, metal-poor GRB hosts might have a more top-heavy IMF.
- Star formation intensity: Specific and surface star formation rates quantify star formation per unit stellar mass and area, respectively, providing measures of its intensity.The study compares these quantities with host stellar masses and with distant galaxy populations.
- Metallicities: Mass-metallicity relations are used because stellar mass or luminosity can proxy metallicity when direct high-redshift measurements are difficult.K-band luminosity varies less than B- or V-band luminosity during short starbursts, supporting its use for mass estimates.
6 LOCATIONS
The location analysis compares transient positions within host light and measures the absolute surface brightness of their local regions, using accurately registered imaging and comparable physical scales.
- Relative locations: The study extends a pixel-ranking analysis to compare where GRBs and CCSN occur within their host-galaxy light distributions.Pixels are ranked by surface brightness, and each transient is assigned the fraction of host light in pixels no brighter than its location.
- Absolute environments: Absolute local surface brightness provides a direct comparison of the luminosities of regions hosting GRBs and CCSN.Because host galaxies lie at similar redshifts, the physical scales probed are comparable.
- Sample and measurements: The analysis uses all 58 CCSN and a z < 1.2 GRB subset with positional accuracy better than roughly 0.08 arcseconds.The GRB selection avoids additional image smoothing at the HST point-spread-function scale.
- Sample and measurements: The local measurements use pixels roughly 150–200 pc on a side and are complemented by host absolute magnitudes and 80% light radii.Figure 8 compares cumulative absolute V-band magnitudes and 80% light radii for GRB, CCSN, and field-galaxy samples.
7 RESULTS
GRB and CCSN hosts have broadly similar global luminosities and colours, but GRB hosts are typically smaller and have higher specific star formation rates, while GRBs occur in brighter local regions.
- Field comparison: The comparison with field galaxies weights cumulative distributions by stellar mass or star formation, rather than treating each field galaxy equally.This weighting tests whether transient occurrence follows host mass or star formation proportionally.
- Global luminosities and colours: GRB and CCSN hosts show no statistically significant differences in absolute B- or V-band magnitudes, despite CCSN hosts having median V luminosity twice as bright.The same-population hypothesis has PKS = 0.4 for both absolute-magnitude distributions.
- Global luminosities and colours: Their rest-frame B−V colours are also similar, with a same-population probability of PKS = 0.2.This supports a less conclusive luminosity contrast than reported in some previous work.
- Host properties: GRB hosts are typically smaller and have higher specific star formation rates than CCSN hosts, while their surface star formation rates are not significantly different.The radius comparison gives PKS = 0.003, the SSFR comparison gives PKS = 0.04, and the surface-SFR comparison gives PKS = 0.14.
- Local environments: GRBs occur in regions of substantially higher surface brightness than CCSN, with differences in both absolute and relative host-light measures.The relative-light comparison has PKS = 5 × 10^-3, while the absolute surface-brightness comparison differs by a factor of 4 with PKS = 0.01.
8 SELECTION EFFECTS
The GRB and CCSN host samples are subject to dust, stellar-population, redshift, and supernova-typing selection effects, but the authors conclude that their environmental differences remain real.
- 8.1 Dust obscuration: Dust may affect CCSN detections more strongly than GRB detections because CCSN can be deeply embedded and are much fainter than GRB afterglows.GOODS observations detected approximately 60% of SN hosts in MIPS, compared with approximately 20% of GRB hosts in one comparison, although differing depths complicate the interpretation.
- 8.2 Evolution of global properties: Very young GRB hosts may have overestimated stellar masses and underestimated star formation rates because standard K-band and U-band calibrations are not constant for young populations.The authors argue that correcting these effects would likely make the GRB and CCSN samples appear more different.
- 8.3 Redshift: The GRB sample may be biased toward intrinsically brighter hosts because emission-line redshifts generally require brighter hosts, whereas absorption redshifts are less dependent on host magnitude.This redshift-selection effect does not apply equally to the CCSN sample, for which broadband data enable photometric redshifts.
- 8.4 SN typing: Approximately half of the CCSN have low-confidence Bronze classifications, creating possible contamination by SN Ia hosts and potential effects on colours and star-formation indicators.Restricting the sample to securely typed Gold and Silver events changes some distributions, but the authors state that the overall conclusions are not changed.
- 8.4 SN typing: SN Ic contamination could reduce the separation between CCSN and GRB populations by preferentially placing events on brighter host regions, although the effect is considered small.The expected influence is limited mainly to the Flight and surface-luminosity distributions.
- 8 SELECTION EFFECTS: The authors conclude that GRB and CCSN environments are different despite selection effects and systematic uncertainties in the host samples.The conclusion follows after considering flux limits, dust, stellar-population assumptions, redshift selection, and supernova typing.
9 DISCUSSION
GRB hosts differ from CCSN hosts in morphology, size, stellar density, and local environments, while their luminosity and mass comparisons remain partly ambiguous. These patterns are consistent with GRBs arising preferentially from compact, intensely star-forming regions and possibly lower-metallicity progenitors.
- Local environments: GRBs occur in regions with different local surface brightness and light distributions from CCSN, with KS probabilities of 0.01 and 5 × 10−3 for the reported comparisons.The authors connect these local-environment differences to very young, massive GRB progenitors.
- Progenitor implications: The discussion links GRB-environment differences to progenitor models in which low metallicity weakens stellar winds and helps massive stars retain rotation, while binaries could broaden the metallicity range.These mechanisms are presented as theoretical context rather than a direct measurement from the host comparison.
- Global host properties: GRB and CCSN host absolute-magnitude distributions differ only modestly: GRB hosts are roughly a factor of 2 fainter in the median, without highly significant KS-test differences.The apparent discrepancy with earlier work is attributed to template-based magnitudes and a fainter, larger CCSN comparison sample.
- Global host properties: GRB hosts occupy the low-mass, high-SSFR region, although their global mass and star-formation-rate distributions may be affected by age and dust biases.Young starbursts can lead to overestimated masses and underestimated star formation rates, while dust obscuration can narrow mass distributions.
- Morphology and size: GRB hosts are smaller and more irregular than CCSN hosts, with only two recognised spirals compared with an expected 15 ± 4 if both samples shared a 45% spiral fraction.The probability of observing two or fewer spirals under that expectation is ∼4 × 10−5.
- Global host properties: The mass interpretation is ambiguous because similar K-band-based mass distributions coexist with significantly different size distributions and may imply either higher GRB-host stellar densities or lower true masses.The authors suggest size may be a more stable mass proxy when comparing populations with different ages.
10 SUMMARY
GRB hosts differ from CCSN hosts in size, morphology, star-formation intensity, and local environments. These results support GRB formation in specific conditions associated with young, massive stellar populations, while very young populations may bias inferred masses and star-formation rates.
- GRBs are highly concentrated on host light and occur in regions of high absolute surface brightness.Their positions differ from those of CCSN both relative to cumulative host light and in absolute surface brightness.
- GRB hosts are consistently smaller than CCSN hosts, with statistically significant differences in physical sizes and morphologies.The authors connect these differences to models in which GRBs form only under particular environmental conditions, likely involving low mass and metallicity.
- High surface brightness, surface star-formation rates, and host locations suggest GRBs originate in younger, more massive stellar populations.This interpretation is presented as part of the paper’s comparison of GRB and CCSN host environments.
- Very young dominant stellar populations in GRB hosts may overestimate stellar masses and underestimate star-formation rates.The authors identify these as potential systematic errors in their inferred host properties.
APPENDIX
The appendix provides photometric catalogs for CCSN and GRB host galaxies, including Spitzer IRAC measurements and stated uncertainty conventions.
- Table 1 catalogs GOODS-field CCSN host photometry with 1-sigma errors and 3-sigma limiting magnitudes.
- Table 2 continues the CCSN host photometric catalog for Spitzer IRAC bands.
- Table 3 lists GRB host photometry in Spitzer IRAC bands with 1-sigma errors and 3-sigma background limits.