No Evidence for Enhanced [O iii] 88 μm Emission in a z ∼ 6 Quasar Compared to Its Companion Starbursting Galaxy

The most distant quasars at z > 6 are unique probes of galaxy and structure formation in the first Gigayear of the universe. Their rapidly accreting supermassive black holes make these quasars the most luminous (non-transient) sources known at this cosmic epoch (e.g., Fan et al. 2006; Venemans et al. 2015; Wu et al. 2015; Bañados et al. 2016). The interstellar medium (ISM) of more than 30 host galaxies of z ∼ 6 quasars has now been detected through the [C ii] 158 μm line (hereafter: [C ii]) and far-infrared (FIR) continuum emission. This suggests that the host galaxies have intense star formation rates (SFRs) at many hundreds of M_⊙ yr⁻¹ occurring coevally with the growth of the accreting central black holes (e.g., Bertoldi et al. 2003; Walter et al. 2003, 2009; Wang et al. 2013; Willott et al. 2013, 2015; Decarli et al. 2018; Venemans et al. 2018).

Using Atacama Large Millimeter/submillimeter Array (ALMA) observations, Decarli et al. (2017) found dust-obscured companions to four quasars in a [C ii] survey of 27 z ∼ 6 quasars. These massive star-forming companion galaxies show no signs of powerful active galactic nucleus (AGN) activity and were detected serendipitously within 8–60 kpc at the same redshifts in four systems, suggesting that they are physically associated with overdensities or proto-clusters of galaxies at z > 6. These companion galaxies rival the FIR and [C ii] luminosities of the neighboring quasars, in some cases even surpassing their (already extreme) luminosity. These quasar–galaxy pairs thus provide a unique opportunity to efficiently study the ISM in distinctly different galactic environments in the first Gyr of the universe: pure starbursts, and those in the presence of a powerful AGN.

To date, the [C ii] line has been the workhorse line of ISM studies at z > 6 (e.g., Carilli & Walter 2013), but it is only one of several fine structure lines that help to characterize the physical properties of the ISM (e.g., Díaz-Santos et al. 2017). Of these, the [O iii] 88 μm line (³P₁ → ³P₀) (hereafter: [O iii]₈₈), emitted by doubly ionized oxygen (O⁺⁺), is a particularly important diagnostic of the star formation process (e.g., Ferkinhoff et al. 2010, 2011, 2015; Vishwas et al. 2018): Given the O⁺ ionization potential of 35.1 eV, O⁺⁺ is almost exclusively found in dense H ii regions around O-type stars or in AGN environments. Indeed, an enhancement of the optical 5008 Å [O iii] line is a key diagnostic of AGN activity (e.g., Baldwin et al. 1981). In contrast, neutral carbon has an ionization potential of only 11.3 eV, and [C ii] is therefore emitted both in the neutral and ionized phase. So far, [O iii]₈₈ detections have been reported in a number of z > 6 galaxies (e.g., Inoue et al. 2016; Carniani et al. 2017; Laporte et al. 2017; Hashimoto et al. 2018b, 2018a; Marrone et al. 2018; Tamura et al. 2018). However, in this redshift range, only two systems have been detected in [O iii]₈₈, [C ii], and the underlying dust continuum: the Lyman break galaxy B14–65666 at z = 7.15 (Hashimoto et al. 2018a) and the lensed galaxy SPT0311–58 at z = 6.90 (Marrone et al. 2018, the latter actually consists of two sources).

In this Letter we report [O iii]₈₈ and underlying dust continuum observations with ALMA of the z = 6.08 quasar CFHQS J2100–1715 (hereafter J2100–Q) and its dust-enshrouded companion starburst (J2100–SB). J2100–Q was discovered by Willott et al. (2010b), and a black hole mass of (9.4 ± 2.6) × 10⁸ M_⊙ was reported in Willott et al. (2010a). J2100–SB was discovered by Decarli et al. (2017) in the FIR using ALMA. Multi-wavelength follow-up observations of the companion galaxy using Spitzer (3.6 μm, 4.5 μm) and Hubble Space Telescope (WFC3/140W) as well as Multi Unit Spectroscopic Explorer (MUSE) and Large Binocular Telescope (LBT) spectroscopy did not detect the companion in the rest-frame ultraviolet (UV)/optical, with an implied ratio of obscured to un-obscured star formation of >99% (C. Mazzucchelli et al. 2018, in preparation). No metallicity measurement of our targets exists, but they must be significantly enriched with heavy elements as they harbor significant amounts of dust, and have high (∼100 M_⊙ yr⁻¹)SFRs.

In Section 2 we describe our ALMA band 8 observations, in Section 3 we present our results, followed by a discussion and a summary in Sections 4 and 5. Throughout this Letter we use cosmological parameters H₀ = 70 km s⁻¹ Mpc¹, Ω_M = 0.3, and Ω_Λ = 0.7, leading to a scale of 5.67 kpc per arcsec at the redshift of our sources.

We observed the [O iii]₈₈ line (ν_rest = 3393.0062 GHz) of J2100–Q and J2100–SB redshifted to ν_obs = 479.2 GHz at z = 6.08 with ALMA in band 8 and configuration C43–1 on 2018 May 20 (two executions) and 2018 May 23 (one execution) for a total of 3 × 5200 s or 4.3 hr, of which 2.5 hr were spent on source. The redshifts of the sources were known from the earlier detections of [C ii] line emission (z = 6.0806 ± 0.0011 for the quasar, 6.0796 ± 0.0008 for the companion). The two sources are very close in redshift (Δ v_LOS = 41 km s⁻¹, Decarli et al. 2017), meaning that the [O iii]₈₈ emission in the two objects could be covered with the same frequency setup. However, the projected separation on the sky (∼11'') required two pointings to cover the sources in band 8. The software package CASA (McMullin et al. 2007) was used for data reduction and imaging. For optimal sensitivity, we employed natural weighting when imaging the data, leading to a synthesized beam size of 073 × 056 (∼3.7 kpc at z = 6.08), and an rms of 0.45 mJy beam⁻¹ per 50 km s⁻¹ (80 MHz) channel. The corresponding sensitivity in the continuum (excluding the channels that contain line emission) is 54 μJy beam⁻¹ over an effective bandwidth of 5.625 GHz. For all analysis of the [O iii]₈₈ line data, the continuum has been subtracted in the uv-plane using line-free channels.

We also re--analyzed the [C ii] and underlying continuum ALMA data published by Decarli et al. (2017) and Venemans et al. (2018). The beam size of these [C ii] data (073 × 057) is almost identical to that presented here. Therefore, no additional beam matching was required. In order to optimize the signal-to-noise ratio in the measurements, and given the fact that any extent, if present, is marginal, we extract [C ii] line and underlying continuum fluxes at the peak position of both sources. We report the [C ii] values adopted in this study in Table 1.¹³

Table 1.  [O iii]₈₈ and Rest-frame 88 μm Continuum Measurements of the Quasar CFHQS J2100–Q and Its Companion J2100–SB

	Units	Quasar	Companion
${z}_{[{\rm{O}}{\rm{III}}]88}$		6.0816 ± 0.0009	6.0806 ± 0.0005
${S}_{\mathrm{cont},88\mu {\rm{m}}}$ a	mJy beam−1	1.12 ± 0.05	2.87 ± 0.05
${S}_{\mathrm{cont},158\mu {\rm{m}}}$ a,b	mJy beam−1	0.46 ± 0.07	1.37 ± 0.14
${S}_{[{\rm{O}}{\rm{III}}]88}$ c	mJy beam−1	0.93 ± 0.20	2.55 ± 0.21
FWHM${}_{[{\rm{O}}{\rm{III}}]88}$ c	km s−1	454 ± 117	614 ± 62
${F}_{[{\rm{O}}{\rm{III}}]88}$ d	Jy km s−1	0.39 ± 0.06	1.52 ± 0.07
${F}_{[{\rm{C}}{\rm{II}}]}$ d	Jy km s−1	1.09 ± 0.08	1.89 ± 0.21
${T}_{\mathrm{dust}}$	K	41 ± 1.2	37 ± 1.2
${L}_{\mathrm{IR}[8-1000\mu {\rm{m}}]}$	1011 ${L}_{\odot }$	8.3 ± 0.37	19.1 ± 0.33
${L}_{\mathrm{FIR}[42-122\mu {\rm{m}}]}$	1011 ${L}_{\odot }$	6.8 ± 0.30	15.5 ± 0.28
${L}_{[{\rm{O}}{\rm{III}}]88}$	109 ${L}_{\odot }$	0.77 ± 0.12	2.86 ± 0.13
${L}_{[{\rm{C}}{\rm{II}}]}$	109 ${L}_{\odot }$	1.05 ± 0.08	1.81 ± 0.9
${L}_{[{\rm{O}}{\rm{III}}]88}$/${L}_{[{\rm{C}}{\rm{II}}]}$		0.74 ± 0.15	1.58 ± 0.24
${L}_{[{\rm{O}}{\rm{III}}]88}$/${L}_{\mathrm{FIR}}$	10−3	1.12 ± 0.11	1.86 ± 0.14
${L}_{[{\rm{C}}{\rm{II}}]}$/${L}_{\mathrm{FIR}}$	10−3	1.55 ± 0.13	1.17 ± 0.13

Notes. Re-measured [C ii] and rest-frame 158 μm continuum values from Decarli et al. (2017). All measurements obtained at continuum emission peak.

^aObserved continuum flux density underlying the line emission. ^bThe error of the companion is higher as it is not in the phase center. ^cFrom Gaussian fitting to the line. ^dBased on integrated line maps (values consistent with spectra).

Download table as:  ASCIITypeset image

3.1. Continuum Emission

In Figure 1 we show the rest-frame 88 μm (observed: 475 GHz) continuum map of J2100–Q and J2100–SB. This map is based on those channels in the datacube that do not contain the [O iii]₈₈ line. The sources are detected at high significance and are unresolved at the resolution of our beam (073 × 057, flux densities are reported in Table 1). The coordinates (R.A./decl. in J2000.0) of the sources, 21:00:54.70, −17:15:21.9 (J2100–Q), and 21:00:55.45, −17:15:21.7 (J2100–SB) are in agreement with Decarli et al.(2017).

Zoom In Zoom Out Reset image size

Together with our earlier rest-frame 158 μm continuum measurement around the [C ii] line (Decarli et al. 2017, see Table 1 for slightly updated numbers) we can use the new rest-frame 88 μm observations to further constrain the properties of the FIR continuum emission. The infrared (IR) and FIR luminosities are obtained by fitting the 88 and 158 μm data points (assuming that they are optically thin) with a modified blackbody, taking the effect of the cosmic microwave background into account (da Cunha et al. 2013). These results are reported in Table 1. The resulting L_IR-based SFRs (Kennicutt & Evans 2012) are 125 ± 14 M_⊙ yr⁻¹ (J2100–Q) and 284 ± 30 M_⊙ yr⁻¹ (J2100–SB).

3.2. [O iii]₈₈ Line Emission

We show the continuum-subtracted [O iii]₈₈ spectra of both sources in Figure 2 as extracted from the peak positions of the continuum map. The Gaussian fit results are reported in Table 1. For reference, we overplot scaled, continuum-subtracted [C ii] spectra from Decarli et al. (2017) in grayscale. Both the redshifts and the linewidths of the [O iii]₈₈ and [C ii] detections agree within the uncertainties. The [O iii]₈₈ line width of the quasar is much smaller than that of emission lines seen in the broadline region of the quasar (FWHM: ∼3600 km s⁻¹, Willott et al. 2010a). In Figure 3 we show the integrated [O iii]₈₈ maps of both sources, after continuum subtraction. 2D-Gaussian fitting shows that the [O iii]₈₈ line emission is unresolved at our resolution like the continuum emission. In Table 1 we summarize the resulting [O iii]₈₈ luminosities. Using the [O iii]₈₈–SFR scaling relations of De Looze et al. (2014) for their "high-redshift sample" we derive [O iii]₈₈-based SFRs of approximately 100 M_⊙ yr⁻¹ and 360 M_⊙ yr⁻¹ for the quasar and the companion, respectively, consistent with the FIR-based estimates (Section 3.1).

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

3.3. Line/FIR Luminosity Ratio

We now combine the [O iii]₈₈ and rest-frame 88 μm continuum measurements with the earlier [C ii] line measurements to constrain the properties of the ISM in our targets. This is particularly interesting, as our sources are physically quite different. In Figure 4 (left panel) we plot the [O iii]₈₈/FIR luminosity ratios versus the continuum flux density ratios (S_{88 μm}/S_{158 μm}) of our sources, the latter being a proxy for the temperature of the dust (color-coded by FIR luminosity), and compare them to the recent compilation of local luminous infrared galaxies (LIRGs) by Díaz-Santos et al. (2017). The mid-infrared (MIR) emission in the majority of the local sample is not dominated by AGN emission (circles versus stars in Figure 4).

Zoom In Zoom Out Reset image size

Our two sources have luminosities at the high end of the LIRG sample studied in Díaz-Santos et al. (2017), but their [O iii]₈₈/FIR luminosity ratios lie only slightly above the average LIRG value of similar dust temperature, within the scatter found locally. Perhaps unexpectedly, despite the presence of an accreting 10⁹ M_⊙ black hole, the quasar does not show enhanced [O iii]₈₈ emission compared to either the companion or the local comparison sample.

It is instructive to also compare the [C ii]/FIR luminosity ratios for our sources to the local LIRGs (middle panel of Figure 4). The [C ii]/FIR luminosity ratios of our targets (∼1.6 × 10⁻³) are located within the distribution observed for the local LIRGS at the same temperature (around ∼1 × 10⁻³). The right panel in Figure 4 shows the [O iii]₈₈/[C ii] luminosity ratios as a function of dust temperature. Again, our two targets fall within the distribution seen in local LIRGS, and the [O iii]/[C ii] ratio is actually lower in the quasar compared to the companion.

4.1. Origin of the [C ii] and [O iii]₈₈ Emission

4.2. Minimum Ionized Gas Mass

4.3. Minimum Number of Ionizing Photons and Stars

We present ALMA [O iii]₈₈ observations of the z = 6.08 quasar J2100–Q and its dust-enshrouded starburst companion J2100–SB. This system is unique, as it offers the possibility to study the physical properties of the ISM of two starbursts that are physically associated with each other within 1 Gyr of the big bang. They display distinct galactic environments: one source shows the presence of a powerful AGN, while the other does not. Interestingly, we find that the [O iii]₈₈/FIR luminosity ratio of the starburst companion J2100–SB is higher than that of the quasar J2100–Q, even though the accreting supermassive black hole in the latter provides additional photons with energies high enough to produce O⁺⁺ (which typically leads to strong optical 5008 Å [O iii] emission from the broad and narrow line regions, measurable with JWST in the future). The fact that we do not see enhanced FIR [O iii]₈₈ emission in J2100–Q indicates that its [O iii]₈₈ emission is dominated by star formation, and not AGN activity. This is supported by the finding that the linewidths and positions of the [O iii]₈₈ and [C ii] line are the same in both sources and consistent with previous findings that the ISM properties in distant quasar host galaxies are predominantly powered by star formation (e.g., Leipski et al. 2014; Barnett et al. 2015; Venemans et al. 2017). One caveat is that J2100–SB could in principle also host an obscured accreting supermassive black hole. In that case, the [O iii]₈₈ emission in the companion may not be solely due to star formation. However, the fact that the [O iii]₈₈/FIR and [O iii]₈₈/[C ii] luminosity ratios of J2100–SB lie within the range of local LIRGs (including those that are not dominated by an AGN) argues against such a scenario. This latter finding, together with our analysis of the number of photons needed for the creation of O⁺⁺, implies that no extreme (top-heavy) initial stellar mass functions are needed to explain the [O iii]₈₈ luminosity in our sources.

We thank the referee for helpful comments that improved our manuscript and Gordon Stacey for useful suggestions. F.W., B.V., M.N., and Ma.N. acknowledge support from the ERC Advanced grant 740246 (Cosmic_Gas). D.R. acknowledges support from the National Science Foundation under grant No. AST-1614213. R.W. acknowledges supports from the National Science Foundation of China (11721303). This Letter makes use of the following ALMA data: ADS/JAO.ALMA#2017.1.00118.S, ADS/JAO.ALMA#2015.1.01115.S. ALMA is a partnership of ESO (representing its member states), NSF (USA) and NINS (Japan), together with NRC (Canada), NSC and ASIAA (Taiwan), and KASI (Republic of Korea), in cooperation with the Republic of Chile. The Joint ALMA Observatory is operated by ESO, AUI/NRAO and NAOJ. The National Radio Astronomy Observatory is a facility of the National Science Foundation operated under cooperative agreement by Associated Universities, Inc.

Facility: ALMA. -

Software: CASA (McMullin et al. 2007).