Narrow [O iii] emission lines as a potential proxy for the evolutionary stage of quasars

t is a widely accepted fact that supermassive black holes (SMBHs) reside at the center of host galaxies. These SMBHs power active galactic nuclei (AGN) through the accretion of surrounding material. This process releases large amounts of energy that, in the form of feedback, influences the physical conditions and dynamics of gas and dust within the host galaxies. Consequently, it is postulated that SMBHs and their host galaxies co-evolve and mutually regulate each other’s development (e.g. Merritt & Ferrarese, 2001; Marconi & Hunt, 2003; Häring & Rix, 2004; Gu et al., 2009; Kormendy & Ho, 2013; Heckman & Best, 2014; Xue, 2017; Ricarte & Natarajan, 2018; Lin et al., 2022). In order to thoroughly investigate the evolution of SMBHs and/or host galaxies, we must address a fundamental question: how can we characterize the corresponding objects that exist at different stages of evolution?

Unlike galaxies and stars, determining the ages of quasars is a challenging task. Various methods used to estimate quasar ages often yield a large uncertainty spanning millions of years (e.g., Martini, 2004; Kirkman & Tytler, 2008; DiPompeo et al., 2014; Morganti, 2017). The properties of absorption lines serve as one common method used to characterize the evolutionary stages of quasars (e.g., Carswell et al., 1982; Bajtlik et al., 1988; Boroson, 1992; Giallongo et al., 1996; Scott et al., 2000; Cao Orjales et al., 2012; Shen & Ménard, 2012; Zheng et al., 2015; Wang et al., 2016; Khrykin et al., 2019; Chen et al., 2022; He et al., 2022; Peng et al., 2024). Low-ionization associated absorption lines (AALs) typically indicate quasars in early stages of evolution, while quasars without AALs or those with only high-ionization AALs are considered to be in later stages of evolution. The proximity effect, which describes a decrease in the number density of Ly $\alpha$ or He ii forest absorption lines observed in quasar sightlines (e.g., Carswell et al., 1982; Bajtlik et al., 1988; Khrykin et al., 2019; Zheng et al., 2019), is yet another method employed for aging quasars. According to the proximity effect, a small proximity zone suggests a younger quasar. However, using absorption lines to determine quasar ages leaves some questions unanswered. The detection and properties of AALs in quasars could be affected by viewing angle (e.g., Urry & Padovani, 1995; Hamann et al., 2012). Furthermore, the proximity effect has its limitations. Neither the Ly $\alpha$ nor the He ii forest absorption lines can be applied to UV-optical spectra of quasars with low redshifts (e.g., $z<1.5$).

The powerful UV radiation from quasars produces ionization zones within or around host galaxies, and is the primary mechanism behind the proximity effect observed in forest absorption lines. Additionally, it can alter the physical conditions within narrow-line regions (NLRs). The cumulative ionizing radiation field promotes the growth of NLRs. Therefore, changes in narrow emission lines (NELs) can provide clues to a quasar’s age. This is corroborated by Zheng (2020), who found that young quasars only host weak NELs, while older ones exhibit strong NELs.

According to the unified schemes of AGNs Urry & Padovani (1995), the NLR is located outside the dust torus and can extend from 100 pc to several kpc (e.g., Harrison & Ramos Almeida, 2024). The broad-line region (BLR) is located inside the dust torus and is often obscured by it. The position of the NLR means its emission lines are less affected by quasar viewing angles compared to broad emission lines (BELs) that originate in BLRs. What’s more, the NLR is compact enough to be illuminated by radiation from the central regions of the quasar. Thus, NELs with high ionizations, such as [O iii] $\lambda\lambda$4959,5007, are often observed in quasar spectra and can be detected by ground-based telescopes for objects at redshifts $z\lesssim 1$. Consequently, the NLR serves as an excellent location for studying quasar properties. In this paper, we use narrow [O iii]emission lines to investigate the growth of NLRs throughout the life cycle of radio quasars.

Throughout this work, we assume a flat $\Lambda$CDM cosmology with $\Omega_{m}$ = 0.3, $\Omega_{\Lambda}$ = 0.7, and $h_{0}$ = 0.7.

The sample we used in our study consists only of type-I quasars. The [O iii] emission lines can be excited by radiation from the central regions of the quasar (photoionization) as well as the shock from outflows/winds. This is the primary reason why researchers typically use two Gaussian functions to model each line of the [O iii] doublet - one Gaussian function attempts to represent the photoionization component (core component), while the other represents the outflow component (wing component). The core component has been found to be useful in investigations concerning the physical conditions and kinematics of NLRs or host galaxy (e.g., Greene & Ho, 2005; Komossa et al., 2008; Zhang et al., 2011; Mullaney et al., 2013; Kormendy & Ho, 2013; Woo et al., 2017; Sexton et al., 2019; Zhang, 2022; Le et al., 2023; Jin et al., 2023). Consequently, in this study, we use the core component of [O iii] $\lambda$5007 to trace the development of NLRs.

The spectral shape of the radio spectrum serves as an effective indicator of a quasar’s evolutionary stage. As shown in Figure 2(a), there is a notable correlation between the equivalent width of the [O iii] $\lambda$5007 emission line and the radio spectral shape, with a Spearman’s correlation coefficient of $r=-0.489$ and a probability of $P<10^{-15}$. BELs are known to exhibit a strong inverse correlation between their equivalent width and the continuum luminosity, a relationship known as the Baldwin effect (Baldwin, 1977). While this effect is less pronounced for NELs (e.g., Zhang et al., 2011, 2013), it may still significantly influence the relationship between the equivalent width of the [O iii] $\lambda$5007 emission line and $\alpha^{\rm 144}_{\rm 3000}$. To assess the influence of the Baldwin effect, we examine in Figure 2(b) the distribution of the continuum luminosity at 5100 Å as a function of $\alpha^{\rm 144}_{\rm 3000}$. The data suggests a negligible evolution in the continuum luminosity with respect to the radio spectral shape - an indication that the Baldwin effect likely contributes minimally to the strong correlation between the $W_{\rm core}^{\hbox{[O\,{\sc iii}]}\lambda 5007}$ and $\alpha^{\rm 144}_{\rm 3000}$. The observed tight correlation between the $W_{\rm core}^{\hbox{[O\,{\sc iii}]}\lambda 5007}$ and $\alpha^{\rm 144}_{\rm 3000}$ might be primarily attributed to the growth of NLRs. This correlation is further corroborated by the significant relationship between $L_{\rm core}^{\hbox{[O\,{\sc iii}]}\lambda 5007}$ and $\alpha^{\rm 144}_{\rm 3000}$, as shown in Figure 2(c). In essence, younger quasars inherently exhibit weaker [O iii] emission lines, while older ones host more potent [O iii] emission lines. These observations align with a study by Zheng (2020), which reported that narrow emission lines are significantly stronger in older quasars compared to younger ones.

Out of our parent sample, 484 quasars exhibit [O iii] emission lines that are too weak or unreliable to be considered ($W_{\rm core}^{\hbox{[O\,{\sc iii}]}\lambda 5007}<4\sigma_{\rm w}$), thus excluded from our final sample. Looking at the bigger picture, the $\alpha^{\rm 144}_{\rm 3000}$ values of these quasars, wherein [O iii] is undetected, are generally larger compared to those of quasars with detected [O iii] (refer to Figure 3). This observation implies that quasars with undetected [O iii] might be in an earlier stage of evolution. Consequently, it is reasonable to expect an increase in the correlation between $W_{\rm core}^{\hbox{[O\,{\sc iii}]}\lambda 5007}$ and $\alpha^{\rm 144}_{\rm 3000}$ once reliable measurements become available for the core [O iii] $\lambda$5007 of quasars without detected [O iii].

Using the observations from the SDSS, VLASS, and LoTSS, we compiled a sample of 2288 quasars with $z\lesssim 1$. In term of the observations from the VLASS and LoTSS, we computed the radio spectral index between 144 MHz and 3000 MHz ($\alpha^{\rm 144}_{\rm 3000}$), which is served as a reliable proxy for quasar ages. Based on the measurements from the SDSS spectra, 1804 quasars have $W_{\rm core}^{\hbox{[O\,{\sc iii}]}\lambda 5007}>4\sigma_{\rm w}$, where the $W_{\rm core}^{\hbox{[O\,{\sc iii}]}\lambda 5007}$ is the equivalent width of the core component of the [O iii] $\lambda$5007 emission line. We find that the $W_{\rm core}^{\hbox{[O\,{\sc iii}]}\lambda 5007}$ is obviously correlated with the $\alpha^{\rm 144}_{\rm 3000}$, which is independent of the Baldwin effect.

The strong correlation between the $W_{\rm core}^{\hbox{[O\,{\sc iii}]}\lambda 5007}$ and $\alpha^{\rm 144}_{\rm 3000}$ suggests that [O iii] $\lambda$5007 could serve as a suitable indicative proxy of the evolutionary stage of quasars. The intensity of narrow [O iii] emission lines tends to increase as the quasar ages. Meanwhile, [O iii] $\lambda$5007 is the most intense narrow emission line noticeable in the optical spectra of quasars and it can be detected using ground-based telescopes for sources with $z\lesssim 1$. As such, the [O iii] emission line proves to be a valuable addition to the proximity effect of forest absorption lines, a method typically employed in dating high-redshift quasars.

We deeply thank the anonymous referees for her/his helpful and careful comments. This work is supported by the Guangxi Natural Science Foundation (2024GXNSFDA010069), the National Natural Science Foundation of China (12073007), and the Scientific Research Project of Guangxi University for Nationalities (2018KJQD01).