Comprehensive Connection among the Quasars with Different Types of Outflow Absorption Lines

It is commonly accepted that outflows from the central regions of quasars play a substantial role in regulating the global properties of the host galaxy. These outflows are typically detected through blueshifted absorption lines. However, the question remains whether outflows observed with different absorption line types indeed reflect the same environmental or evolutionary stage of the host galaxy. In this study, we use the Sloan Digital Sky Survey quasar catalog and employ the flux ratio of [O II] and [Ne V] emission lines as indicators to compare star formation rates (SFRs) within host galaxies of quasars exhibiting various outflow absorption line types: low-ionization broad absorption line (LoBAL), low-ionization Mini-BAL (LoMini-BAL), low-ionization narrow absorption line (LoNAL), high-ionization broad absorption line (HiBAL), high-ionization Mini-BAL (HiMini-BAL), and high-ionization narrow absorption line (HiNAL). Our findings indicate that the SFR of LoMini-BAL quasars is comparable to that of LoNAL quasars, somewhat less than that of LoBAL quasars, but markedly greater than that of HiBAL quasars. Furthermore, the SFR of HiMini-BAL quasars mirrors that of HiNAL or Non-abs (no associated absorption lines) quasars, but is significantly higher than that of HiBAL quasars. If we consider that differing absorption line types are indicative of the quasar evolution stage, our results propose an inclusive evolution sequence: LoBALs evolve into LoMini-BALs/LoNALs, then progress to HiBALs, and ultimately morph into HiMini-BALs/HiNALs/Non-abs. Concomitantly, the SFR within the host galaxies of quasars appears to decline noticeably nearing the LoNAL phase’s end and rejuvenates before the HiMini-BAL phase.


Introduction
The outflows from quasars, driven by accreting supermassive black holes (SMBHs), serve as vital links in understanding the coevolution of SMBHs and their host galaxies (e.g., Di Matteo et al. 2005;Kormendy & Ho 2013;Heckman & Best 2014;Weinberger et al. 2017;Valentini et al. 2021;Chen et al. 2022).These outflows are frequently identified through blueshifted absorption lines in the rest-frame ultraviolet (UV) spectra of quasars, also known as outflow absorption lines.These lines can form over a wide range of distances from the central SMBHs (e.g., Chen et al. 2013;Chamberlain et al. 2015;Xu et al. 2018Xu et al. , 2019;;He et al. 2022).Moreover, these outflows can purge the host galaxy of interstellar gas, thereby inhibiting star formation and preventing the inflow of gas into both the galaxy and its central SMBH (e.g., Germain et al. 2009;Hopkins & Elvis 2010;Zubovas & King 2012;Smith et al. 2020;Gao et al. 2021;Laha et al. 2021;Chen et al. 2022).Hence, it is commonly believed that the physical characteristics of host galaxies are significantly impacted by this outflow feedback.
Quasar outflows can produce a diverse range of absorption lines.Broad absorption lines (BALs) exhibit line widths exceeding 2000 km s −1 (e.g., Murray et al. 1995;Elvis 2000;Capellupo et al. 2011;He et al. 2017), narrow absorption lines (NALs) present widths less than a few hundred kilometers per second (e.g., Chartas et al. 2009;Hamann et al. 2011Hamann et al. , 2012;;Chen et al. 2018cChen et al. , 2020b)), and Mini-BALs exhibit intermediate widths between NALs and BALs (e.g., Misawa et al. 2007;Gibson et al. 2009a;Horiuchi et al. 2016;Xu et al. 2018;Chen et al. 2021).Optical spectroscopic surveys of quasars show that outflow NALs or Mini-BALs are more prevalent than outflow BALs.Approximately 30% of optically selected quasars contain outflow NALs or Mini-BALs (e.g., Nestor et al. 2008;Wild et al. 2008;Chen et al. 2016Chen et al. , 2018aChen et al. , 2021)), while only about 10%-20% exhibit outflow BALs (e.g., Gibson et al. 2009b;Allen et al. 2011;Guo & Martini 2019).Yet, it remains unclear how BALs, Mini-BALs, and NALs coalesce into a holistic perspective of quasar outflow phenomena.The two most invoked scenarios to synergize different absorption line types within a single quasar outflow phenomenon are orientation and evolution.In the orientation scheme (e.g., Hamann et al. 2012), BALs are observed when the quasar's line of sight aligns with specific directions close to the accretion disks, while Mini-BALs and NALs are observed at higher latitudes above the accretion disk.In contrast, in the evolution scheme (e.g., Boroson 1992;Wang et al. 2016;Chen et al. 2022), BALs can evolve into Mini-BALs and subsequently into NALs, and potentially in reverse order.
Under the orientation scenario, the properties of host galaxies with various types of outflow absorption lines should remain statistically consistent.On the other hand, the evolutionary scenario suggests that quasars with different types of outflow absorption lines are at different evolutionary stages and are therefore likely to demonstrate diverse properties in their host galaxies.For instance, Chen et al. (2022) observed Original content from this work may be used under the terms of the Creative Commons Attribution 4.0 licence.Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI.substantial differences in both color and star formation rate (SFR) between the host galaxies of low-ionization broad absorption lines (LoBALs) and high-ionization broad absorption lines (HiBALs).Moreover, Wang et al. (2016) found the cover factor of outflows could reach up to 20% for the Fe II scattering region.This is significantly larger than the fraction of FeLoBAL quasars (LoBAL quasars with Fe II absorption lines), which constitute less than 1% of all quasars (e.g., Trump et al. 2006;Dai et al. 2012).Both findings from Chen et al. (2022) and Wang et al. (2016) are unresolvable in the orientation scenario but can be neatly elucidated under the evolutionary scenario.While Chen et al. (2022) merely unveiled the linkage between different types of BAL quasars, this information is a valuable clue to broadly comprehend the connection among myriad types of outflow absorption lines.This connection can further be understood by examining the properties of their host galaxies.This forms the foundation and central goal of this work.

Data
The Sloan Digital Sky Survey (SDSS) employed spectroscopy to observe quasars in the wavelength range between 3800 and 9200 Å with a resolution of approximately R ≈ 2000 during its first and second stages (SDSS-I/II) (e.g., Abazajian et al. 2009).In its third and fourth stages (SDSS-III/IV), SDSS observed quasars in the wavelength range of 3600-10,500 Å, with a resolution varying from R ≈ 1300-2500 (e.g., Dawson et al. 2013;Smee et al. 2013;Dawson et al. 2016).From SDSS-I to SDSS-IV, the SDSS collected spectra for 750,414 quasars (data release 16, i.e., DR16Q; Lyke et al. 2020).This work focused on quasars having an emission redshift 1.4 < z em < 1.6 from SDSS DR16Q.This selection criterion ensures the imprinting of the [O II] λ3729 emission line and blueshifted C IV λλ1548,1551 absorption lines on the SDSS spectra.Additionally, we excluded quasars whose spectra had a median signal-to-noise ratio (S/N) < 3 pixel −1 to ensure the robustness of the detections and avoid spurious signals.
Following the standard convention (e.g., Shen et al. 2011;Chen et al. 2018b), we modeled a power-law continuum ( f λ = Aλ α ) plus an iron template for each spectrum in line-free regions (Vestergaard & Wilkes 2001;Véron-Cetty et al. 2004).The regions were carefully chosen to avoid obvious contamination from strong absorption or emission lines.The C IV λ1549 and Mg II λ2798 emission lines were then fitted using multiple Gaussian functions.The underlying continuum of each spectrum encompassed all of the aforementioned fitting components.Subsequently, we searched for C IV and Mg II outflow absorption lines in the spectra, normalized by the underlying continuum.Multiple Gaussian functions were invoked to model each absorption component (e.g., Chen & Qin 2015;Chen et al. 2018a).The equivalent widths of absorption features were obtained by integrating the fitting profiles of the absorption features.
LoBALs are typically narrower than HiBALs.We thus uniformly define the C IV and Mg II BALs that have a line width greater than 1500 km s −1 (e.g., Zhang et al. 2010).Here, line width refers to the sum of the FWHM of both the blue and red components of each doublet.Specifically, for C IV, the FWHM is the sum of the FWHMs at λ1548 and λ1551, and for Mg II, the FWHM is the sum of the FWHMs at λ2796 and λ2803.Based on the statistical results from previous research (e.g., Nestor et al. 2008;Chen et al. 2016Chen et al. , 2018aChen et al. , 2021)), we define Mini-BALs and NALs as absorption features with line widths of 800 < FWHM < 1500 km s −1 and FWHM < 800 km s −1 , respectively.
Subsamples of quasars with different types of absorption lines are defined as follows: LoBAL quasar sample.The LoBAL quasars should include Mg II BALs, which yields a sample of 179 LoBAL quasars.
LoMini-BAL quasar sample.The low-ionization Mini-BAL (LoMini-BAL) quasars should include at least one Mg II Mini-BALs 4 with −6000 < υ r < 0 km s −1 , might include Mg II NALs, C IV Mini-BALs, and C IV NALs, but not include Mg II, Al III, Fe II, and C IV BALs.We obtain 291 quasars with Mg II Mini-BALs, but without Mg II, Al III, Fe II, and C IV BALs.
HiBAL quasar sample.The HiBAL quasars should include C IV BALs, might include C IV Mini-BALs and C IV NALs; but not include Mg II BALs, Al III BALs, Fe II BALs, Mg II Mini-BALs with −6000 < υ r < 0 km s −1 , and Mg II NALs with −2000 < υ r < 0 km s −1 .We obtain 1724 quasars with C IV BALs.
Non-abs quasar sample.We require the Non-abs quasars that do not include Mg II BALs, Al III BALs, Fe II BALs, C IV BALs, Mg II Mini-BALs with −6000 < υ r < 0 km s −1 , C IV Mini-BALs with −6000 < υ r < 0 km s −1 , Mg II NALs with −2000 < υ r < 0 kms −1 , and C IV NALs with −4000 < υ r < 0 km s −1 .In addition, in order to reliably characterize the quasar continuum, we only consider the spectra with median S/N 5.This results in 8105 robust quasars without outflow BALs, Mini-BALs, and NALs.
The redshifts of the quasars involved in all the above samples were determined using narrow [O III], [O II], or [Ne V] emission lines, whenever available.If such data were not accessible, we utilized the "BEST REDSHIFTS" listed in DR16Q of the SDSS Lyke et al. (2020).We discovered that 4 3736 quasars had their redshifts determined using narrow or [Ne V] emission lines.The remaining 9123 quasars had their redshifts acquired from the "BEST RED-SHIFT."Further, for over 75% of quasars with narrow or [Ne V] emission lines, the redshifts from the "BEST REDSHIFT" align consistently within 500 km s −1 .

Composite Spectra
Mirroring the method utilized in previous studies (e.g., Chen et al. 2020aChen et al. , 2022)), we have produced median composite spectra for all the categories of quasars including LoBAL, LoMini-BALs, LoNALs, HiBAL, HiMini-BALs, HiNALs, and Non-abs quasars.The outcomes are depicted in Figure 1.The figure illustrates that the composite spectra for all categories of BAL, Mini-BAL, and NAL quasars are visibly redder than that of the Non-abs quasars.This discrepancy can be ascribed to dust extinction within the BAL, Mini-BAL, and NAL quasars.To describe this dust extinction, we draw on the extinction curve of the Small Magellanic Cloud, expressed as A λ = 1.39λ −1.2 E(B − V ) (e.g., Prevot et al. 1984;Pei 1992), where E(B − V ) indicates the color excess.The results considering this extinction curve are also represented by colored dotted-dashed lines in Figure 1.
In the case of each composite spectra, we apply a power-law function ( f λ = Aλ α ) alongside an iron template (Vestergaard & Wilkes 2001;Véron-Cetty et al. 2004) to establish the underlying continuum.The iron template is defined by four parameters: the velocity shift, velocity dispersion, and separate amplitudes for the wavelength regions 2200-3100 and 3250-3500 Å.The results of this process can be seen in Figures 2-8.The strengths of the [Ne V] λ3426 and [O II] λ3729 emission lines are calculated by integrating the emission line profiles in the composite spectra normalized by the underlying continuum (the result of the power-law plus iron continuum fitting).The outcomes of this are detailed in Table 1.
We note that there is a significant residual in the wavelength range of 2300-2600 Å for the cases with BALs or Mini-BALs when fitting the iron template to composite spectra.These discrepancies could potentially be ascribed to variations in dust content, gas density, and ionization level of gas across different evolutionary stages.We plan to explore these physical mechanisms in more detail in future work.However, it is reassuring to note that these imperfect fits do not impact the results derived from other spectral regions.

Results and Discussion
As can be clearly seen in Figure 1, all the quasars with BALs, Mini-BALs, and NALs are notably redder than the ones without any outflow absorption lines.This is especially true for the LoBAL, LoMini-BAL, and LoNAL categorized quasars.Figure 9(a) likewise displays a trend where extinctions decrease as quasars transition from LoBALs → LoMini-BALs → LoNALs → HiBALs → HiMini-BALs → HiNALs.This observation can be plausibly explained through the evolving life cycle of quasars.In the early stage of a quasar's lifetime (LoBAL, LoMini-BAL, and LoNAL phases), the outflow only expands to a small scale, and the central SMBH is still shrouded by thick gas and dust.At this point, the quasar exhibits significantly enhanced dust-reddening and the gas presenting in the outflow has lower ionization levels.As the quasar continues to evolve, the powerful outflow extends to larger scales, eventually breaking through and clearing away the surrounding gas and dust cocoon.This leads the subsequent stages of a quasar's life cycle (HiBAL, HiMini-BAL, and HiNAL phases) to experience reduced dust extinction.Furthermore, as the outflowing gas is exposed to this progressively cleared environment, its ionization state reaches higher levels.
As shown in Table 2, the low-ionization [O II] λ3729 and high-ionization [Ne V] λ3426 emission lines can be equally excited by all sources including shocks, extended emission line regions (EELRs), and narrow emission line regions (NLRs) (e.g., Maddox 2018, and references therein).Given that the ionization potential of [Ne V] is much higher than that of [O II], star formation activity within quasar host galaxies is more likely to produce [O II] than [Ne V].Because of this, the combination of [O II] and [Ne V] is commonly used as an indicator of star formation activity (e.g., Shen & Ménard 2012;Maddox 2018;Chen et al. 2020aChen et al. , 2022)), typically by comparing their relative strengths.An enhanced ratio of R = EW [O II] /EW [Ne V] corresponds to more intense star formation activity.We delineate the ratio R of quasars with different types of absorption lines in Figure 9(b), which clearly demonstrates a gradual decline in R from a high value (R = 2.257 ± 0.240) to a substantially lower one (R = 0.717 ± 0.023) as quasars evolve from LoBALs → LoMini-BALs → LoNALs → HiBALs, followed by a rebound and a plateau as quasars continue to evolve from HiBALs → HiMini-BALs/ HiNALs/Non-abs.The R of the LoBAL quasars is marginally higher than those of the LoMini-BAL (2σ) and LoNAL (2.3σ) quasars.While the LoMini-BAL and LoNAL quasars each have similar values (1σ) for R, they, together with the LoBAL quasars, all present an R that is notably (>6.3σ) higher than that of the HiBAL quasars.Furthermore, all the HiMini-BAL, HiNAL, and Non-abs quasars have similar R values within their categories (1σ), but are much higher (>4.7σ) than that of the HiBAL quasars.Within the evolution scenario of a quasar life cycle, consistent SFRs suggest that quasars are likely nestled in comparable environments or close evolutionary stages.In view of this, our results suggest that (1) the environment of the LoMini-BALs is likely similar to that of the LoNAL quasars and also shows some similarity to that of the LoBAL quasars; (2) the environments of all LoBAL, LoMini-BAL, and LoNAL quasar phases are noticeably different from that of HiBAL quasars; and (3) the environments of all HiMini-BAL, HiNAL, and Non-abs quasars share similarities yet are distinctly different from that of HiBAL quasar phase.Considering all samples of quasars with differing absorption lines do not demonstrate significant differences in luminosity and black hole mass (see Figure 10), we deduce that variations in R are not dominated by these two factors.
Our sample comprises 179 LoBAL, 623 LoMini-BAL/ LoNAL, 1724 HiBAL, and 2228 HiMini-BAL/HiNAL quasars.These represent about 1.4%, 4.8%, 13.4%, and 17.3% of the sample, respectively.Based on quasar evolution models, the LoBAL, LoMini-BAL, and LoNAL quasars make up a brief transition phase (around 6.2%) in the total lifetime of quasars and are thought to exist in the early stages of quasar evolution.During this stage, quasars possess enough cold gas to sustain high SFR.Assuming a quasar lifetime of 50 Myr (e.g., Hopkins et al. 2005), the LoBAL quasar phase lasts about 0.7 Myr, much shorter than the 2.4 Myr (0.0048 × 50 Myr) of the LoMini-BAL/LoNAL quasar phase.With an outflow velocity of 10,000 km s −1 , the outflow reaches 7 kpc from the galaxy's center during the LoBAL quasar phase.Assuming a slower shock velocity of 1000 km s −1 due to interactions with the interstellar medium, the impact can reach up to 3.1 kpc by the end of the LoMini-BAL/LoNAL quasar phase.Simulations (e.g., Scannapieco & Oh 2004;Di Matteo et al. 2005) have     6 The Astrophysical Journal, 963:3 (9pp), 2024 March 1 Peng et al.  shown that an outflow carrying only about 5% of the quasar's bolometric luminosity can produce significant feedback on its host galaxy, fully sweeping out interstellar gas and thus reducing star formation.Consequently, LoMini-BAL and LoNAL quasars, which exist close to the LoBAL quasars phase, may exhibit a slightly reduced but still significant SFR until they approach the blow-out phase (e.g., Sanders & Mirabel 1996;Hopkins et al. 2006).The longer duration (around 2.4 Myr) of the LoMini-BAL/LoNAL quasar phase allows the powerful outflow that has enough time to reach and impact the galactic scale (up to 3.1 kpc), and thus blow away large-scale gas.This could explain the significant decrease in star formation activity, as the LoMini-BAL/LoNAL quasar phase nears its end.The HiBAL quasar phase lasts around 6.7 Myr (0.134 × 50 Myr), during which the outflow has enough time to sweep out the gas.This could lead to a much lower SFR in HiBAL quasars compared to LoBAL, LoMini-BAL, or LoNAL quasars.As quasars evolve from the HiBAL phase to the HiMini-BAL phase, the outflow would have subsided and the suppression of star formation terminated, leading to a rekindling of star formation.The resulting SFR of the HiMini-BAL quasars is much higher (6.6σ) than that of the HiBAL quasars.Interestingly, the SFR in the HiMini-BAL quasar phase aligns with that in the HiNAL and Non-abs quasar phases, suggesting that the SFR rebound phenomenon probably starts before the HiMini-BAL quasar phase.Although the exact boundaries marking the points at which star formation declines and rebounds are not clearly identified, it appears that star formation likely begins to drop near the end of the LoBAL phase, significantly decreases by the end of the LoNAL phase, and begins to rebound prior to the HiMini-BAL phase.Furthermore, the effects of star formation suppression seem to be transient relative to the total quasar lifetime.
We use the flux ratio of the [O II] to [Ne V] emission lines as a proxy to compare the SFR within the host galaxies of quasars.Our findings suggest a slight suppression in the SFR as quasars evolve from the LoBAL to LoMini-BAL or LoNAL stages, followed by a significant quenching to the lowest value during the HiBAL stage.However, the SFR rebounds to a higher value as the quasars evolve from HiBAL to HiMini-BAL stages.The SFR of LoMini-BAL quasars is similar to that of the LoNAL quasars.In the same vein, HiMini-BAL, HiNAL, and Non-abs quasars share a consistent SFR.This similar SFR proposes that LoMini-BAL and LoNAL quasars may occupy a similar environment or evolutionary stage, and the same could be said for the HiMini-BAL, HiNAL, and Non-abs quasars.Moreover, we find that the SFR appears to decrease substantially toward the end of the LoNAL quasar phase, before beginning to rebound ahead of the HiMini-BAL quasar phase.This transitory suppression of star formation is relatively brief compared to the quasar's total lifespan.Acknowledgments   Note.The shocks only with high velocity (>600 km s −1 ) produce a strong [Ne V] emission line.See also Maddox (2018).
We are grateful to the anonymous referee for the careful comments that helped improve this manuscript.Z.-F.C. is supported by the National Natural Science Foundation of China (12073007), the Guangxi Natural Science Foundation (2023JJD110002), and the Scientific Research Project of Guangxi University for Nationalities (2018KJQD01).Zhicheng He is supported by the National Natural Science Foundation of China (Nos. 441 12222304, 12192220, and 12192221).

Figure 2 .
Figure 2. The median composite spectra of the LoBAL quasars.Green points mark the data used to model the power-law continuum (black dashed-dotted-dot lines).Black horizontal thick lines indicate the spectral regions used to model the iron template, and the orange solid lines are for the fitting results (power law plus iron template).Gray-shaded regions indicate the [O II] and [Ne V] lines.(a) Shows the global result from fitting the power-law continuum and iron template.(b) and (c) exhibit the composite spectra after subtracting the best-fitting power law and iron template.

Figure 3 .
Figure 3.The median composite spectra of the LoMini-BAL quasars.See Figure 2 for the meaning of symbols.

Figure 4 .
Figure 4.The median composite spectra of the LoNAL quasars.See Figure 2 for the meaning of symbols.

Figure 6 .
Figure6.The median composite spectra of the HiMini-BAL quasars.See Figure2for the meaning of symbols.

Figure 5 .
Figure 5.The median composite spectra of the HiBAL quasars.See Figure 2 for the meaning of symbols.

Figure 7 .
Figure 7.The median composite spectra of the HiNAL quasars.See Figure 2 for the meaning of symbols.

Figure 8 .
Figure8.The median composite spectra of the Non-abs quasars.See Figure2for the meaning of symbols.

Figure 9 .
Figure 9. (a) The distribution of the extinction of quasars with different types of absorption lines relative to the Non-abs quasars.(b) The line ratio R = EW [O II] /EW [Ne V] is a proxy of SFR.

Figure 10 .
Figure 10.The distributions of the 3000 Å luminosity and the black hole mass for the quasars with different types of absorption lines.The legend gives the median values and standard deviations of the distributions.

Table 1
Equivalent Widths of [Ne V] and [O II]

Table 2
Emission Lines of [Ne V] and [O II] of the Quasar Host