Diagnostics for PopIII galaxies and Direct Collapse Black Holes in the early universe

TL;DR

The paper develops comprehensive photoionisation models to construct diagnostic diagrams that identify and distinguish PopIII galaxies and Direct Collapse Black Holes in the early, metal-poor universe. By exploring rest-frame optical and UV lines for three ionising sources (PopIII, PopII, DCBHs) across wide metallicity and density ranges with Cloudy, it reveals that He II lines, together with key metal lines, offer robust separation from metal-enriched populations and provide metallicity and density constraints. The study provides explicit selection criteria and metallicity diagnostics for both optical and UV regimes, highlighting the continued strength of [O III] 5007 and CIV 1549 even at ultra-low metallicities. These diagnostics will be instrumental for JWST and ELT campaigns targeting primitive sources and SMBH seed candidates in the early universe, while acknowledging the simplifications inherent in single-cloud models.

Abstract

Forthcoming observational facilities will make the exploration of the early universe routine, likely probing large populations of galaxies at very low metallicities. It will therefore be important to have diagnostics that can solidly identify and distinguish different classes of objects in such low metallicity regimes. We use new photoionisation models to develop diagnostic diagrams involving various nebular lines. We show that combinations of these diagrams allow the identification and discrimination of the following classes of objects in the early universe: PopIII and Direct Collapse Black Holes (DCBH) in pristine environments, PopIII and DCBH embedded in slightly enriched ISM (Z~10^{-5}-10^{-4}), (metal poor) PopII and AGN in enriched ISM. Diagnostics involving rest-frame optical lines (that will be accessible by JWST) have a better discriminatory power, but also rest-frame UV diagnostics can provide very useful information. Interestingly, we find that metal lines such as [OIII] 5007A and CIV 1549A can remain relatively strong (about a factor of 0.1-1 relative H-beta and HeII 1640A, respectively), even in extremely metal poor environments (Z~10^{-5}-10^{-4}), which could be embedding PopIII galaxies and DCBH.

Figures (12)

• Figure 1:
• Figure 2: Distribution of the Cloudy models on the EW(He ii$\lambda 4686$) versus He ii$\lambda4686$/H$\beta$ diagram, for the primitive sources (Left) and for the chemically-evolved systems (Right). As shown in the legend, different symbols present different kinds of modelled objects, different colors for different gas metallicities ($\rm Z=0$ (blue) to $0.028 = 2$ Z$_{\odot}$ (red)), and larger symbol sizes for higher ionisation parameters (Left: $\log U = -0.5$ to $-2.0$ for PopIII and PopII galaxies, $-0.5$ to $-3.0$ for DCBHs; Right: $\log U = -1.0$ to $-3.5$ for evolved galaxies, $-0.5$ to $-3.0$ for AGNs). A fiducial gas density of $10^3$ cm$^{-3}$ is assumed. The black dotted lines denote the demarcation (Eq. \ref{['eq:ewhe2_he2hb']}) between PopIII and the other populations.
• Figure 3: Distribution of the Cloudy models on the He ii$\lambda4686$$/$H$\beta$ versus He i$\lambda5876$$/$H$\beta$ diagram, only for PopIII and PopII galaxies (Z $<10^{-3}$) for three different gas density values, $n_{\rm e}=10^2$, $10^3$, and $10^4$ cm$^{-3}$, as indicated with the dotted curves (Eqs. \ref{['eq:he2hb_he1hb_5876_n2']}--\ref{['eq:he2hb_he1hb_5876_n4']}). Symbols as in Fig. \ref{['fig:ewhe2_he2hb']}.
• Figure 4: Distribution of the Cloudy models on the He ii$\lambda4686$$/$H$\beta$ versus He i$\lambda4471$$/$H$\beta$ diagram, for primitive sources (Left) and for chemically-evolved systems (Right). Symbols as in Fig. \ref{['fig:ewhe2_he2hb']}. The black dotted lines can be useful to disentangle some types of DCBHs (Eq. \ref{['eq:he2hb_he1hb_4471']}).
• Figure 5: Distribution of the Cloudy models on the optical metal-line diagnostics ([O iii]$\lambda5007$$/$H$\beta$ versus [N ii]$\lambda 6584$$/$H$\alpha$ (1st row), [O iii]$\lambda5007$$/$H$\beta$ versus [S ii]$\lambda\lambda 6717,6731$$/$H$\alpha$ (second row), and [O iii]$\lambda5007$$/$H$\beta$ versus [O ii]$\lambda 3727$$/$H$\beta$ (3rd row), for the primitive sources (Left) and for the chemically-evolved systems of galaxies (Middle) and AGNs (Right). Symbols as in Fig. \ref{['fig:ewhe2_he2hb']}. In the extremely low-metallicity regime (Z $<10^{-3}$; left panels), all models follow a sequence for a given gas-phase metallicity irrespective of the different ionising sources. The dotted curves show the average sequences for the models with Z $=10^{-5}$ (Eqs. \ref{['eq:o3hb_n2ha_zem5']}--\ref{['eq:o3hb_o2hb_zem5']}) and $10^{-4}$ (Eqs. \ref{['eq:o3hb_n2ha_zem4']}--\ref{['eq:o3hb_o2hb_zem4']}). The conventional demarcation curves between AGNs and galaxies on the BPT-diagrams are depicted as introduced by kewley2001 (dot-long dashed; "Kew01") and kauffmann2003 (short dash-long dashed; "Kau03"). The demarcation curve of lamareille2010 is plotted on the bottom diagram using [O ii] (dot-long dashed; "Lam10"). Clearly, these curves work for isolating AGNs if the metallicity is relatively high (Z $\gtrsim 0.5$ Z$_{\odot}$) and the spectrum is hard ($\alpha\gtrsim -1.6$).