Balmer Decrements and Nebular-Stellar Reddening in JADES Galaxies at $2.7<z<7$

Abstract

We aim to characterize nebular and stellar reddening in star-forming galaxies as a function of global galaxy properties (stellar mass, SFR, metallicity) at $2.7 < z< 7.0$. We also provide a prescription to convert SED-based $E(B-V)_{\mathrm{star}}$ to $E(B-V)_{\mathrm{gas}}$ when direct measurements of nebular reddening are unavailable. Our results are based on JWST/NIRSpec measurements of both individual spectra, with a sample of 283 galaxies, and composite spectra, including a larger sample of 327 galaxies. We estimate nebular reddening using the Balmer decrement (H$α$/H$β$) above $10^{8.5}$ $M_{\odot}$, where the sample is representative. Stellar reddening and SFRs are derived through Prospector SED fitting, while gas-phase metallicities are estimated from strong emission-line ratios. At fixed stellar mass, Balmer decrements remain consistent within uncertainties across our redshift range, indicating that stellar mass primarily determines the overall dust column even by $z \sim 7$. We find that differential reddening ($ΔE(B-V) \equiv E(B-V)_{\mathrm{gas}} - E(B-V)_{\mathrm{star}}$) scales linearly with mass and SFR at $z \sim 2.7 - 4.0$, but shows no mass or SFR dependence above $z \sim 4.0$. We find evidence for smaller $ΔE(B-V)$ above $z \sim 5.0$, suggesting that nebular emission and stellar continuum probe increasingly similar dust columns towards higher redshift. Finally, we find that nebular reddening correlates strongly with metallicity out to $z \sim 5$, whereas the correlation between stellar reddening and metallicity is weaker or absent. Together, these results suggest that both dust mass and geometry play a significant role in shaping the observed reddening of high-redshift galaxies.

Figures (8)

• Figure 1: Galaxy counts by redshift for our sample. Grey bars indicate counts in the full observed mass range, while colored bars show counts in the mass-representative range ($\log(\mathrm{M}_\ast/\mathrm{M}_\odot)$$\geq 8.5$). $N_{tot}$ denotes the total number of galaxies in each redshift bin in the mass-representative range.
• Figure 2: H$\alpha$$/$H$\beta$ vs. stellar mass in bins of redshift. Binned medians of Balmer decrement are plotted as red triangles; bin widths are displayed as red line segments with the $1\sigma$ interval on H$\alpha$$/$H$\beta$ shaded. The grey 2D histogram shows the $z \sim 0$ sample from SDSS; purple diamonds show $z \sim 2.3$ running medians in bins of M$_{\star}$ from the MOSDEF survey Shapley+22. Cyan squares indicate measurements from composite spectra. We show the combined $5.0 \leq z < 7.0$ composites in both the $5.0 < z \leq 6.0$ and $6.0 < z \leq 7.0$ redshift bins. The horizontal dashed line shows the assumed intrinsic Case B recombination value of 2.79. N denotes the number of galaxies in each redshift bin with $\log(\mathrm{M}_\ast/\mathrm{M}_\odot)$$>$ 8.5. The greyed-out region and points on the left indicate measurements that are not in the mass-representative range of our sample.
• Figure 3: Comparison of Balmer decrement vs. stellar mass across redshift bins. The grey 2D histogram shows the $z \sim 0$ sample from SDSS. Top panel: Binned medians from individual galaxies are plotted as triangles. $z \sim 2.3$ MOSDEF binned medians are shown as purple diamonds. Bottom panel: Balmer decrement measurements from composite spectra.
• Figure 4: $E(B-V)_{\rm gas}$$-$$E(B-V)_{\rm star}$ versus stellar mass in bins of redshift. Individual galaxies are plotted as blue circles; stacked spectra are shown as cyan squares. We plot measurements from the combined $5.0 \leq z < 7.0$ composite spectra in both the $5.0 < z \leq 6.0$ and $6.0 < z \leq 7.0$ redshift bins. Greyed-out points indicate measurements that are not in the mass-representative range of our sample. The horizontal dashed line shows the median vertical offset computed from the stacks, with the $1\sigma$ uncertainty shaded in grey. $\langle\Delta E(B\!-\!V)\rangle$ reports the mean offset and uncertainty. The grey dotted line indicates where $\Delta E(B-V)$$= 0$. In the $2.7 < z \leq 4.0$ bin, we plot the best fit to the stacked spectra as a solid red line.
• Figure 5: $E(B-V)_{\rm gas}$$-$$E(B-V)_{\rm star}$ versus $\mathrm{SFR}_{100 \rm Myr}$ in bins of redshift. Individual galaxies are plotted as blue circles; stacked spectra are shown as cyan squares. We plot measurements from the combined $5.0 \leq z < 7.0$ composite spectra in both the $5.0 < z \leq 6.0$ and $6.0 < z \leq 7.0$ redshift bins. The horizontal dashed line shows the median vertical offset computed from the stacks, with the $1\sigma$ uncertainty shaded in grey. $\langle\Delta E(B\!-\!V)\rangle$ reports the mean offset and uncertainty. The grey dotted line indicates where $\Delta E(B-V)$$= 0$. In the $2.7 < z \leq 4.0$ bin, we plot the best fit to the stacked spectra as a solid red line.