GA-NIFS & JADES: Confirmation of pristine gas near GN-z11

Abstract

According to the leading cosmological model, a first generation of stars called Population III (PopIII), condensed almost entirely out of hydrogen and helium, must have initiated the creation of all heavier chemical elements. Here we report the detection of ionised hydrogen (H$γ_{4342}$) with $S/N$=5.9 in a region about 3 pkpc (projected) North-East from the z~10.6 galaxy GN-z11, where line emission compatible with doubly ionised helium (HeII$_{1640}$) had been found. Our new JWST/NIRSpec-IFS G395H data confirm the authenticity of the previous detection, at a redshift of $z_{\rm Hγ}$=$10.5862$$\pm$$0.0003$. H$δ$ is marginally detected ($S/N$$\sim$$2$). No metal lines are detected in our observations spanning $λ_{\rm rest}$=$0.25$-$0.45μ$m. We derive a $3σ$ upper limit on the gas phase metallicity of 12+log(O/H)$<$6.96 ($Z_{\rm gas}$$<$$0.019~Z_\odot$). Through comparison with NIRCam imaging, we constrain a lower limit on the equivalent width of EW$_0$(H$γ$)$>$350Å. We compare our emission line constraints to model predictions and find them compatible with photoionization by PopIII stars, possibly intermixed with next-generation (PopII) stars. We infer an upper limit on the dynamical mass of $M_{\rm dyn}$$\lesssim$$3$$\times$$10^8M_\odot$. Our data provide novel support for the presence of PopIII stars nearby GN-z11, 440 Myr after the Big Bang.

Figures (4)

• Figure 1: Left: NIRCam rgb image showing the location of the H$\gamma$ clump (pink contours) in reference to GN-z11, the 'haze' Tacchella23, and a $z$$\sim$$2$ galaxy. The white circle indicates the astrometric uncertainty ($r$=$0.05^{\prime\prime}$) in registering the NIRCam images and NIRSpec-IFS cubes, comparable to the NIRCam PSF at 2.8$\mu\mathrm{m}$ ($\sim$0.09). Right: H$\gamma$ linemap ($\lambda_{\rm obs}$=5.029-5.031$\mu\mathrm{m}$). The pink square shows the extraction aperture for Fig. \ref{['f:fit_hg']}. Red contours trace Heii$_{1640}$ by M26. The circle shows the NIRSpec PSF at 5.3$\mu\mathrm{m}$.
• Figure 2: Left: Integrated spectrum of the H$\gamma$ clump (pink aperture in Fig. \ref{['f:map_extraction']}). The error spectrum is shown in red. We indicate the theoretical locations of several emission lines at $z$$=$$10.586$. Right: Best fit to H$\gamma$, detected with $S/N$$\sim$$6$. In grey we mark the wavelength of H$\gamma$ in GN-z11.
• Figure 3: Diagnostic diagrams of [Neiii]$_{3870}$/H$\gamma$ ( left) and EW$_0$(H$\gamma$) ( right) vs. Heii$_{1640}$/H$\gamma$. Pink triangles shows our constraints for Hebe from this work and M26, using either the total Heii flux (filled) or only the Heii component matching our H$\gamma$ detection (open; 'C2', see M26 and Appendix \ref{['a:velmatch']}). Other symbols show model predictions by Nakajima22 and Rusta25 (small stars) for ionisation by PopIII (filled stars), mixed PopIII+PopII (open stars; fractions of total stellar mass), PopII (open circles) and AGN or DCBHs (crosses). For PopIII stars, the models shown have stellar masses $M_\star$=$1$-$500 M_\odot$, absolute gas phase metallicity $Z$=0-$10^{-4}$, neutral hydrogen density $n_H$=$10^{3-6}$cm$^{-3}$, ionisation parameter $\log(U)$=$[-0.5;-2]$ for Nakajima22, and $M_\star$=$0.8$-$1000 M_\odot$, $Z$=0-$2.7$$\times$$10^{-4}$, $n_H$=$10^3$cm$^{-3}$, $\log(U)$=$-1$ (and $n_H$=$10^6$cm$^{-3}$, $\log(U)$=$-2$ in the right panel) for Rusta25. Models with gas metallicities conflicting with our upper limit are greyed out. Our observational constraints are compatible with ionisation by (self-enriched) PopIII, or PopIII mixed with PopII.
• Figure 4: Overlay of the H$\gamma$ and Heii$_{1640}$ emission in velocity space (zero velocity is with reference to $z$=10.5862 for both lines). The Heii$_{1640}$ spectrum is taken from M26.