The Physical Origins of the Identified and Still Missing Components of the Warm–Hot Intergalactic Medium: Insights from Deep Surveys in the Field of Blazar 1ES1553+113

Cosmological simulations predict that gravitational shocks associated with structure formation will heat a large fraction of the cool (T ≈ 10⁴ K) intergalactic medium (IGM) that dominates the baryon budget in the early universe to form a warm–hot intergalactic medium (WHIM; T ≈ 10⁵–10⁷ K) at z ≲ 1 (e.g., Cen & Ostriker 1999). The predicted physical state and enrichment levels of the WHIM depend sensitively on stellar and black hole feedback that provides additional heating and chemical enrichment (e.g., Rahmati et al. 2016; Nelson et al. 2018; Wijers et al. 2019). Observations of the WHIM and its relationship to galaxies can, therefore, serve as a check of our cosmological paradigm and as a laboratory for studying galaxy evolution.

While observationally elusive, the WHIM can be detected via absorption spectroscopy through ionic transitions in the ultraviolet (UV) and X-ray as well as through metallicity-independent probes such as broad H i Lyα absorption (e.g., Danforth et al. 2010), the Sunyaev–Zel'dovich (SZ) effect (e.g., de Graaff et al. 2019), and the dispersion measure of fast radio bursts (FRBs; e.g., Bannister et al. 2019; Ravi et al. 2019). Surveys of the highly ionized phases of the circumgalactic/intergalactic medium (CGM/IGM) traced by O vi (e.g., Danforth et al. 2016), Ne viii (e.g., Pachat et al. 2017; Frank et al. 2018), and Mg x (Qu & Bregman 2016) with the Cosmic Origins Spectrograph (COS; Green et al. 2012) on the Hubble Space Telescope (HST) can account for a large fraction of the baryons expected in the WHIM; however, it leaves ∼30% missing (e.g., Shull et al. 2012) and potentially in a chemically pristine or more highly ionized phase.

Surveys of CGM/IGM around galaxies find that metal ion absorption is common in the CGM at projected distances (d) less than the estimated galaxy host halo virial radii (R_h) but comparatively rare at larger distances (e.g., Liang & Chen 2014; Turner et al. 2014; Johnson et al. 2015, 2017; Burchett et al. 2019). These observations suggest that feedback may be ineffective at enriching the IGM far beyond galaxy halos. Indeed, the statistical detection of a SZ signal from the filaments between massive galaxies (de Graaff et al. 2019) can potentially account for the remaining missing baryons, suggesting that a substantial portion of the WHIM exhibits low metallicities ($\lt \tfrac{1}{10}$ solar; Liang & Chen 2014; Johnson et al. 2015) or high temperatures (T > 6 × 10⁵ K) not traced in the UV.

New insights into chemical enrichment mechanisms and the physical state of the CGM/IGM require deep galaxy surveys in fields with UV and X-ray absorption spectra. Blazars are ideal for such studies because of their high UV/X-ray flux levels. Recently, Nicastro et al. (2018) obtained a 1.7 Ms XMM-Newton X-ray spectrum of the blazar 1ES 1553+113, reaching the signal-to-noise ratio (S/N) levels required to detect hot CGM/IGM absorbers individually over a large redshift pathlength for the first time. The X-ray spectrum revealed two candidate O vii absorption systems at z = 0.4339 and 0.3551, each with statistical significance of ≈3–4σ, though we note that systematic/non-Gaussian errors (e.g., Nevalainen et al. 2019) and contamination (e.g., Nicastro et al. 2016) have led to past controversies over X-ray absorbers. Nevertheless, taken together, the two O vii absorbers reported by Nicastro et al. (2018) suggest that the hot phase of the CGM/IGM is metal rich and accounts for 10%–70% of the baryon budget. However, the combination of a poorly constrained blazar redshift (due to a featureless spectrum) and limited complementary galaxy surveys complicates the interpretation of absorbers toward 1ES 1553+113.

Here, we present a deep and highly complete galaxy redshift survey in the field of 1ES 1553+113. When combined with UV absorption spectra, the survey enables a precise measurement of the redshift of 1ES 1553+113 and provides insights into the origins of intervening IGM/CGM systems. This Letter proceeds as follows. In Section 2, we describe the galaxy survey and UV spectroscopy. In Section 3, we combine these data sets to infer the blazar redshift. In Section 4, we characterize the galactic environments of the candidate WHIM absorbers and draw insights into their origins.

We adopt a flat Λ cosmology with Ω_m = 0.3, Ω_Λ = 0.7, and ${H}_{0}=70\,\mathrm{km}\,{{\rm{s}}}^{-1}\,{\mathrm{Mpc}}^{-1}$. All magnitudes are in the AB system. We define the knee in the galaxy luminosity function, L_*, as M_r = −21.5 (Loveday et al. 2012).

2.1. Galaxy Survey Data

To study the relationship between galaxies and the IGM, we conducted a deep and highly complete redshift survey targeting galaxies of m_r < 23.5 mag in the field of 1ES 1553+113 with multi-slit spectrographs on the Magellan Telescopes. We acquired deep g-, r-, and i-band images with MOSAIC on the Mayall telescope with 1800 s of exposure in each filter under 1'' seeing (PI: Johnson; PID: 2015A-0187) and an HST image with the ACS+F814W filter and 1200 s of exposure (PI: Mulchaey; PID: 13024). We processed the data as described in Chen & Mulchaey (2009) and Johnson et al. (2015). In total, we measured spectroscopic redshifts for 921 galaxies at angular distances of Δθ < 14' from the blazar sightline and also include 25 redshifts from Prochaska et al. (2011) and one from Keeney et al. (2018). Redshift histograms and completeness levels for galaxies of L ≳ 0.25L_* (>90% at projected distances of d < 500 kpc and z < 0.5) are shown in Figure 1.

Zoom In Zoom Out Reset image size

The survey results are summarized in Table 1, which reports coordinates, apparent magnitudes (${m}_{g},{m}_{r},{m}_{i}$), redshift quality ("g" for secure redshifts and "s" for single-line redshifts), redshift (z_gal), rest-frame ${M}_{g}-{M}_{r}$ color, absolute rest-frame r-band magnitude (M_r), stellar mass ($\mathrm{log}{M}_{* }/{M}_{\odot }$), and projected angular and physical separations from the blazar sightline (Δθ and d). The absolute magnitudes include k-corrections, and the stellar masses are estimated as in Johnson et al. (2015) assuming a Chabrier (2003) initial mass function (IMF). Typical uncertainties in the redshifts, magnitudes, and stellar masses are 60 km s⁻¹, 0.1 mags, and 0.2 dex, respectively. Table 1 is separated into sections by redshift within ±1000 km s⁻¹ of the two candidate O vii absorbers (z = 0.4291–0.4387; 0.3506–0.3596), those at higher redshift (z > 0.4387), and all other redshifts. Figure 2 displays an image of the field with galaxy redshifts labeled.

Zoom In Zoom Out Reset image size

Table 1.  Redshift Survey Summary with Galaxies Separated by Redshift

R.A.	Decl.	mg	mr	mi	Quality	zgal	${M}_{g}-{M}_{r}$	Mr	$\mathrm{log}{M}_{* }/{M}_{\odot }$	Δθ	d
(J2000)	(J2000)	(AB)	(AB)	(AB)		(AB)	(AB)			(arcsec)	(pkpc)
zgal > 0.4387
15:55:44.20	+11:11:04.6	23.1	22.1	21.4	g	0.5721	0.4	−21.3	10.3	26.1	170
15:55:43.91	+11:11:54.4	22.9	21.6	20.8	g	0.5963	0.5	−21.9	10.7	32.7	218
15:55:45.22	+11:11:20.7	23.3	21.9	21.1	g	0.6626	0.4	−22.2	10.6	32.3	226
15:55:45.38	+11:11:09.9	23.6	23.1	22.9	s	1.1130	0.2	−22.5	10.3	37.2	305
15:55:42.08	+11:10:32.1	23.1	22.2	22.0	g	0.5244	0.2	−20.2	9.4	54.1	338
15:55:44.62	+11:12:05.5	23.7	23.2	22.8	g	0.7753	0.2	−20.4	9.5	47.3	351
15:55:40.28	+11:10:55.3	23.9	23.1	22.3	g	0.8778	0.4	−21.9	10.5	49.9	386
15:55:47.24	+11:11:11.6	23.7	23.1	22.3	g	0.7270	0.2	−20.8	9.7	63.0	457
15:55:37.80	+11:11:43.4	23.0	21.4	20.9	g	0.4695	0.4	−21.2	10.3	79.5	469
15:55:38.87	+11:11:54.8	22.5	21.6	21.1	g	0.6935	0.2	−21.9	10.1	68.5	488
15:55:37.26	+11:11:23.9	23.2	21.5	20.9	g	0.4699	0.5	−21.3	10.5	85.2	503
⋯	⋯	⋯	⋯	⋯	⋯	⋯	⋯	⋯	⋯	⋯
zgal = 0.4291–0.4387
15:55:43.51	+11:11:13.0	23.9	22.7	22.5	g	0.4300	0.5	−20.8	10.2	13.2	74
15:55:43.19	+11:11:02.8	23.6	22.2	21.5	g	0.4347	0.6	−20.5	10.3	21.6	122
15:55:41.45	+11:11:29.1	25.4	23.3	22.9	s	0.4297	0.4	−19.1	9.5	23.9	134
15:55:42.76	+11:11:56.4	24.9	23.1	22.6	g	0.4344	0.7	−19.6	9.9	32.3	183
15:55:43.95	+11:11:56.7	21.4	19.6	18.9	g	0.4343	0.8	−23.2	11.4	35.0	198
15:55:42.94	+11:10:47.6	23.6	22.1	21.5	g	0.4329	0.6	−20.4	10.2	36.8	207
15:55:39.87	+11:11:46.8	24.3	23.3	23.4	g	0.4347	0.0	−18.5	8.7	51.9	293
15:55:46.15	+11:10:54.1	22.5	21.1	20.6	g	0.4328	0.5	−21.3	10.5	54.8	308
15:55:46.69	+11:11:07.7	23.1	22.3	22.2	g	0.4332	0.0	−19.5	9.1	56.1	316
15:55:47.25	+11:11:16.0	22.3	20.6	20.0	g	0.4328	0.7	−22.1	11.0	62.4	351
15:55:42.66	+11:12:29.8	23.4	22.1	21.9	g	0.4335	0.3	−20.0	9.6	65.7	370
15:55:42.53	+11:12:30.2	24.5	22.8	22.4	g	0.4339	0.5	−19.7	9.8	66.3	374
15:55:37.79	+11:10:51.4	24.3	23.0	22.8	g	0.4344	0.3	−19.2	9.3	84.0	474
15:55:44.56	+11:09:57.9	24.4	23.0	22.7	g	0.4327	0.4	−19.5	9.5	89.3	503
⋯	⋯	⋯	⋯	⋯	⋯	⋯	⋯	⋯	⋯	⋯	⋯
zgal = 0.3506–0.3596
15:55:44.49	+11:09:19.3	22.8	21.7	21.4	g	0.3531	0.4	−19.9	9.7	126.8	630
15:56:02.12	+11:12:44.1	22.7	21.6	21.3	g	0.3537	0.4	−20.0	9.8	291.8	1451
15:55:29.10	+11:05:37.7	24.3	23.3	23.2	g	0.3547	0.1	−18.3	8.6	402.8	2007
15:55:47.81	+11:04:41.8	22.8	21.9	21.6	g	0.3596	0.4	−19.8	9.6	408.6	2054
15:55:41.14	+11:03:51.8	21.2	19.5	18.8	g	0.3580	0.8	−22.7	11.2	453.4	2273
15:56:13.90	+11:09:06.0	24.1	23.3	23.4	g	0.3590	0.0	−18.1	8.5	474.8	2384
15:55:24.61	+11:18:04.5	22.2	20.7	20.3	g	0.3538	0.6	−21.1	10.4	483.4	2405
15:55:12.93	+11:17:48.8	21.9	20.9	20.6	g	0.3542	0.4	−20.7	10.0	586.6	2920
15:55:52.15	+11:21:42.4	20.6	19.4	18.9	g	0.3531	0.6	−22.5	11.0	632.4	3142
⋯	⋯	⋯	⋯	⋯	⋯	⋯	⋯	⋯	⋯	⋯	⋯
Other Redshifts
15:55:07.77	+11:01:42.3	21.1	19.8	19.3	g	0.0017	1.2	−9.6	5.8	780.0	28
15:55:46.11	+11:11:49.4	23.5	22.9	22.8	g	0.1022	0.3	−15.6	7.8	51.6	97
15:55:44.01	+11:11:09.1	23.1	22.6	22.3	g	0.3892	0.3	−19.2	9.3	20.8	110
⋯	⋯	⋯	⋯	⋯	⋯	⋯	⋯	⋯	⋯	⋯	⋯

Only a portion of this table is shown here to demonstrate its form and content. A machine-readable version of the full table is available.

Download table as:  DataTypeset image

2.2. UV Absorption Spectroscopy

Optical−X-ray spectra of 1ES 1553+113 exhibit no detected emission lines, preventing systemic redshift measurements (Landoni et al. 2014). The lack of a precise redshift measurement complicates the interpretation of absorption features in the spectrum of 1ES 1553+113 due to an inability to differentiate intervening IGM/CGM systems from associated absorption. Previous estimates of the redshift of 1ES 1553+113 based on the detection of intervening H i Lyα absorption (e.g., Danforth et al. 2010) and the shape of its γ-ray spectrum (e.g., Abramowski et al. 2015) imply 0.413 ≲ z_sys≲ 0.6.

Blazars are typically hosted by luminous elliptical galaxies (e.g., Urry et al. 2000) in massive groups (e.g., Wurtz et al. 1997). Moreover, 1ES 1553+113 is a high-energy peaked blazar which are thought to arise from beamed FR-I radio galaxies (e.g., Rector et al. 2000). 1ES 1553+113 exhibits a complex, one-sided radio-jet morphology (see Figure 2; Rector et al. 2003), indicating disturbance by a hot intragroup or intracluster medium. Identification of the blazar's host group therefore represents a precise means of inferring its redshift (e.g., Farina et al. 2016; Rovero et al. 2016).

To identify the host group of 1ES 1553+113, the top panel of Figure 3 displays the redshift histogram for galaxies of L > 0.25L_* from our survey at d < 500 and <1000 proper kpc (pkpc) from the blazar sightline. With high completeness levels of 100%, >90%, and >80% for galaxies of L > 1.0, 0.5, and 0.25L_* respectively, at d < 500 pkpc and z < 0.6, our redshift survey is sensitive to galaxy groups over the full range of possible systemic redshifts for 1ES 1553+113. The only significant overdensity with multiple luminous galaxies near the blazar sightline is at z ≈ 0.433, a strong indication that the blazar is a member of the z = 0.433 group.

Zoom In Zoom Out Reset image size

The blazar host group consists of seven (14) members of L > 0.25L_* at d < 500 (1000) pkpc from the blazar and exhibits a light-weighted redshift of z_group = 0.433. Not including the blazar host, the total stellar mass (luminosity) of the group is ≈8 × 10¹¹M_⊙ (18L_*) with ≈60% (50%) coming from three massive, quiescent galaxies of $\mathrm{log}{M}_{* }/{M}_{\odot }\gt 11.0$. The measured line-of-sight velocity dispersion of the group is σ_group ≈ 300 km s⁻¹, which corresponds to an estimated dynamical mass of ${M}_{\mathrm{dyn}}\sim 2-5\times {10}^{13}\ {M}_{\odot }$. Such a massive group is consistent with expectations for the environment of blazars like 1ES 1553+113. Assuming that the luminosity of the blazar host galaxy is L = 6L_* (Urry et al. 2000), the group light-weighted center is 13'' E. (70 pkpc) and 7'' N. (40 pkpc) of the blazar position.

The presence/absence of H i Lyα absorption in the blazar spectrum as a function of redshift can be used for an independent estimate of the blazar redshift. The archival COS spectrum of 1ES 1553+113 enables searches for H i Lyα absorption at λ < 1796.7 Å, which corresponds to a maximum redshift of ${z}_{\mathrm{Ly}\alpha }=0.478$. In this wavelength range, the S/N is sufficient to detect absorbers of W_r > 0.03 Å at 3σ significance. The spectrum reveals seven Lyα absorbers at z = 0.350–0.413, implying z_sys ≳ 0.413 but none over the similar interval of z = 0.413–0.478 (bottom-left panel of Figure 3).

To quantify the redshift constraint on 1ES 1553+113 from the Lyα forest with objects of similar luminosity, we identified 59 available quasi-stellar objects (QSOs) with measured systemic redshifts, archival COS spectra, and IGM absorption line identifications from Danforth et al. (2016). For each QSO, we computed the difference between the systemic redshift and that of the highest redshift H i Lyα line with W_r > 0.03 Å in the spectrum, ${\rm{\Delta }}z={z}_{\mathrm{sys}}-\max ({z}_{\mathrm{Ly}\alpha })$. The resulting Δz distribution is shown in the bottom-right panel of Figure 3. When combined with the highest redshift Lyα line in the spectrum of 1ES 1553+113 at z = 0.413 (50 pMpc from z = 0.433 where the UV background dominates), this distribution implies a 95% confidence interval for the redshift of 1ES 1553+113 of z_sys = 0.411–0.435, which is consistent with membership of the z = 0.433 galaxy group. This constraint assumes that blazars and QSOs reside in similar intergalactic environments and is subject to small number statistics in the wings of the distribution. It will be further tested with new HST near-UV (NUV) spectra (PI: Muzahid, PID: 15835) for improved Lyα searches.

4.1. The Origins of Candidate WHIM X-Ray Absorbers

Nicastro et al. (2018) identified two candidate WHIM O vii absorbers at z = 0.4339 and 0.3551 in the XMM spectrum of 1ES 1553+113, suggesting that the hot phase of the IGM is metal rich and potentially closing the missing baryon problem. Neither O vii candidate is detected in O vi or O vii, similar to recent non-detections in an X-ray emitting cosmic filament (Connor et al. 2019). Here, we discuss the origins of these WHIM candidates based on our redshift survey.

As discussed in Section 3, 1ES 1553+113 is most likely a member of a galaxy group at z = 0.433. The identified O vii system at z = 0.4339 is therefore associated with the blazar environment and cannot be used in cosmic baryon censuses.

The blazar host group is part of a larger-scale overdensity consisting of three additional groups at ≈1.5 pMpc southeast, ≈1.5 pMpc northwest, and ≈2.5 pMpc east-southeast from the blazar. The O vii candidate could be due to WHIM from this overdensity, but photoionization from the blazar and UV background would be important. To evaluate the feasibility of the WHIM interpretation under these circumstances, we ran a series of Cloudy (Ferland et al. 2017) models to calculate the equilibrium O vi, O vii, and O viii ion fractions as a function of distance from the blazar for gas with ${n}_{{\rm{H}}}={10}^{-5}\mbox{--}{10}^{-3}\ {\mathrm{cm}}^{-3}$ and temperatures of T = 10⁵–10⁷ K including photoionization from the blazar ($\lambda {L}_{\lambda }={10}^{46}\ \mathrm{erg}\,{{\rm{s}}}^{-1}$ at 1 Rydberg and UV spectral slope of α = −1.4 based the COS spectrum) and UV background (Khaire & Srianand 2019) as shown in Figure 4.

Zoom In Zoom Out Reset image size

Nicastro et al. (2018) demonstrated that the O vii detection and O vi/O viii non-detections at z = 0.4339 require a gas temperature of T ≈ 10⁶ K with little contribution from photoionization. This rules out WHIM gas with ${n}_{{\rm{H}}}\lt {10}^{-4}\ {\mathrm{cm}}^{-3}$ at any distance from the blazar because photoionization by the UV background is significant at such low densities (see Figure 4; Wijers et al. 2019). Denser hot gas of n_H = 10⁻⁴ (10⁻³) cm⁻³ can reproduce the absorber properties but only at >10 (1) pMpc from the blazar (see Figure 4). The z = 0.4339 O vii candidate is, therefore, unlikely to be due to low-density WHIM but may arise from hot CGM or intragroup medium (Mulchaey et al. 1996) in the blazar environment.

The z = 0.3551 O vii candidate resides in a comparatively isolated region with no galaxies at d < 500 pkpc from the sightline within Δv = ±1000 km s⁻¹ from the absorber redshift despite 100% completeness levels for galaxies of L > 0.1L_*. The nearest galaxy to the sightline is a star-forming galaxy of $\mathrm{log}{M}_{* }/{M}_{\odot }=9.7$ at z = 0.3531 and d = 630 pkpc or ≈5× its virial radius (estimated with the stellar-to-halo mass relation from Kravtsov et al. 2018 and virial radius definition from Bryan & Norman 1998). The nearest massive galaxy has a stellar mass of log M_*/M_⊙ = 11.2 and is at d = 2273 pkpc or ≈5 × its virial radius. The z = 0.3551 candidate could be due to an undetected dwarf in principle, but surveys of the CGM/IGM around dwarfs (Johnson et al. 2017) indicate that metal absorption systems are rare beyond the virial radius, and dwarfs are not expected to maintain a hot halo (e.g., Correa et al. 2018).

To determine whether strong O vii systems are expected from the WHIM in isolated environments, we calculated the fraction of predicted strong O vii absorbers as a function of environment using WHIM predictions (Wijers et al. 2019) from the EAGLE cosmological hydrodynamical simulations (Crain et al. 2015; Schaye et al. 2015; McAlpine et al. 2016). We calculated column densities within 2000 km s⁻¹ simulation slices and cross-correlated with galaxies as a function stellar mass and projected distance. In total, only 1%–3% of the predicted, comparably strong O vii ($\mathrm{log}N({\rm{O}}\,{\rm{vii}})/{\mathrm{cm}}^{-2}=15.6$) systems occur in similarly isolated environments (d > 630 pkpc to the nearest galaxy of $\mathrm{log}{M}_{* }/{M}_{\odot }\gt 9.7$). While the model predictions are subject to non-negligible uncertainties due to treatment of peculiar velocities and feedback, we nevertheless conclude that strong O vii WHIM systems are not expected to be common in isolated environments. Moreover, Bonamente (2018) estimated a 4% probability that the z = 0.3551 O vii candidate arises from noise fluctuations.

We conclude that neither of the two candidate O vii absorbers in the spectrum of 1ES 1553+113 are of confident and unbiased intergalactic origin. This implies a 95% upper limit on the number of WHIM O vii absorbers with W_r ≳ 6 mÅ per unit redshift of $\tfrac{{dN}}{{dz}}\lt 8$. The lack of strong WHIM X-ray absorption systems suggests that metal enrichment is primarily confined to galaxy halos and their immediate outskirts. This is consistent with the EAGLE simulations that predict that most strong O vii systems arise from metal-rich (> 0.5Z_⊙) gas at over-densities of δ ≳ 100 (see Wijers et al. 2019). Further exploration of the relationship between the WHIM and galaxies requires metallicity-independent probes.

4.2. The Origins of Broad H i Lyα Systems

4.3. Summary and Conclusions

We are grateful to J. Nevalainen, F. Nicastro, and M. Petropoulou for insightful comments. S.D.J. is supported by a NASA Hubble Fellowship (HST-HF2-51375.001-A). This research was partially supported by a NASA grant through HST-GO-13024. M.R.D. acknowledges support from the Dunlap Institute at the University of Toronto and the Canadian Institute for Advanced Research (CIFAR). J.C.C. acknowledges support by the National Science Foundation under Grant No. AST-1517816. Based on observations from the Magellan, the NOAO Mayall, and NASA/ESA Hubble Telescopes. The authors are honored to conduct research on Iolkam Duág (Kitt Peak), a mountain with particular significance to the Tohono Oódham. We made use of the NASA Astrophysics Data System.

Facilities: Magellan - , HST - , Mayall. -