A Far Ultraviolet Spectroscopic Explorer Survey of Interstellar Molecular Hydrogen in the Galactic Disk
J. Michael Shull, Charles W. Danforth, Katherine L. Anderson
AA Far Ultraviolet Spectroscopic Explorer Survey ofInterstellar Molecular Hydrogen in the Galactic Disk
J. Michael Shull, Charles W. Danforth
Department of Astrophysical and Planetary Sciences and CASAUniversity of Colorado, 389-UCB, Boulder, CO 80309
Katherine L. Anderson
Green Mountain High School, Lakewood, CO 80228, USA [email protected], [email protected],[email protected]
ABSTRACT
We report results from a
Far Ultraviolet Spectrographic Explorer ( FUSE ) survey of interstellarmolecular hydrogen (H ) in the Galactic disk toward 139 O-type and early B-type stars at Galacticlatitudes | b | ≤ ◦ , with updated photometric and parallax distances. The H absorption ismeasured using the far-ultraviolet Lyman and Werner bands, including strong R(0), R(1), andP(1) lines from rotational levels J = 0 and J = 1 and excited states up to J = 5 (sometimes J = 6 and 7). For each sight line, we report column densities N H2 , N HI , N ( J ), N H = N HI +2 N H2 ,and molecular fraction, f H2 = 2 N H2 /N H . Our survey extends the 1977 Copernicus H surveyup to N H ≈ × cm − . The lowest rotational states have excitation temperatures and rmsdispersions, (cid:104) T (cid:105) = 88 ±
20 K and (cid:104) T (cid:105) = 77 ±
18 K, suggesting that J = 0 , , J states exhibit mean excitation temperatures, (cid:104) T (cid:105) = 237 ±
91 K and (cid:104) T (cid:105) = 304 ±
108 K, produced primarily by UV radiative pumping.Correlations of f H2 with E ( B − V ) and N H show a transition to f H2 ≥ . N H (cid:38) cm − and E ( B − V ) (cid:38) .
2, interpreted with an analytic model of H formation-dissociation equilibriumand attenuation of the far-UV radiation field by self-shielding and dust opacity. Results of thisdisk survey are compared to previous FUSE studies of H in translucent clouds, at high Galacticlatitudes, and in the Magellanic Clouds. Using updated distances to the target stars, we findaverage sightline values (cid:104) f H2 (cid:105) = 0 .
20 and (cid:104) N H /E ( B − V ) (cid:105) = 6 . × cm − mag − .
1. Introduction
Molecular hydrogen (H ) is the most abundant molecule in the universe, constituting the majority ofthe interstellar molecular clouds that eventually form stars. Even though H plays an important role inthe chemistry of the interstellar medium (ISM), many questions remain about its distribution, formation,and destruction in both diffuse and protostellar clouds. The Copernicus satellite in the 1970s provided thefirst large-scale survey of interstellar H (Spitzer et al. 1974; Savage et al. 1977). With the 1999 launch ofthe Far Ultraviolet Spectrographic Explorer ( FUSE ) satellite, astronomers once again gained access to thefar-ultraviolet (FUV) wavelengths needed to study H in its resonance absorption lines. This survey extendsthese studies to 139 OB-type stars, many of them fainter and more distant than observed by Copernicus . a r X i v : . [ a s t r o - ph . GA ] F e b FUSE mission, H studies were part of the science plan. The FUSE satellite,its mission, and its on-orbit performance were described in Moos et al. (2000) and Sahnow et al. (2000).Initial FUSE studies of H were reported in papers from the Early Release Observations (Snow et al. 2000;Shull et al. 2000). Later studies included a survey of 70 sight lines to OB stars probing H in the Largeand Small Magellanic Clouds (Tumlinson et al. 2002), surveys of 38 sight lines through translucent clouds(Rachford et al. 2002, 2009), studies of chemical relationships of H with other molecules (CO, CH, CH + ,CN) and atomic species (Burgh et al. 2007; Sheffer et al. 2008; Jensen et al. 2010), and observations of H in the low Galactic halo (Gillmon et al. 2006; Wakker 2006; Gillmon & Shull 2006). At high redshift, H Lyman/Werner lines have been detected in damped Ly α absorbers (Noterdaeme et al. 2008; Jorgenson et al.2014; Balashev et al. 2019).We present the results of a FUSE survey of interstellar H absorption in the Galactic disk, usingtransitions from the ground electronic state, X Σ + g , to excited electronic states, B Σ + u (Lyman bands) and C Π u (Werner bands). These lines are rovibrational transitions from lower states ( v l , J l ) to upper states( v u , J u ). In the cold, low-density ISM, essentially all the molecules are in the ground vibrational state( v l = 0). Absorption lines were observed up to J l = 5 − J l = 7. Because the wavefunctionfor the homonuclear H is anti-symmetric under interchange of the identical (fermionic) protons, the even-parity rotational states ( J = 0 , , ... ) have total nuclear spin S = 0 (spin anti-symmetric para-H ) whileodd-parity states ( J = 1 , , ... ) have S = 1 (spin-symmetric ortho-H ). The statistical weights of thesestates are g J = (2 S + 1)(2 J + 1), with odd- J states having a factor of three higher weight. Absorption bandsleading to the upper level are identified by changes in vibrational state ( v u − v l ) and rotational angularmomentum ( J u − J l ). The upper electronic state ( Σ + u ) of the Lyman bands has angular momentum Λ = 0along the internuclear axis. Dipole-allowed changes in rotational state (∆ J = ±
1) are denoted as R-branch( J u = J l + 1) and P-branch ( J u = J l − Π u ) of Werner bands has Λ = 1,allowing a Q-branch ( J u = J l ) in addition to the R and P branches.We observed multiple lines in the Lyman and Werner bands in the FUV (930–1126 ˚A) towards back-ground OB stars located near the Galactic disk plane at | b | ≤ ◦ . Several target stars were outside thislatitude range, but all show strong H absorption. Lines from J = 0 and J = 1 nearly always exhibitdamping wings. Analysis of absorption-line equivalent widths and damping wings yields column densities, N ( J ), in individual rotational states. Figure 1 shows a
FUSE spectrum of the sight line to HD 46150, anO5 Vf star at 1.5 kpc distance and E ( B − V ) = 0 .
45, with the Lyman and Werner bands labeled including aclose-up of the (4-0) Lyman band. The total H column density, N H2 , is found by summing over all observed J states. Typically, 98% to 99% of the molecules reside in the lowest two levels, J = 0 and J = 1.Observations of H column densities, the molecular fraction in diffuse clouds, and its rotational excitationprovide diagnostics of diffuse ISM (Shull & Beckwith 1982). The excitation temperature, T , of the lowestrotational states (para-H in J = 0 and ortho-H in J = 1) is an approximate measure of the gas kinetictemperature. In many cases, the three lowest levels ( J = 0 , ,
2) appear to be thermalized, with both T and T coupled to the kinetic temperature (Gry et al. 2002; Le Petit et al. 2006). In diffuse interstellar clouds,thermal equilibration requires sufficiently large gas densities for proton interchange collisions to produceortho-para conversion between J = 0 and J = 1 (Dalgarno et al. 1973; Gerlich 1990). Column densities N H2 ≥ cm − are needed to produce strong self-shielding in absorption lines from the dissociating FUVradiation. The current observations show that both T and T provide estimates of the heating and coolingprocesses in the diffuse ISM. The excitation temperature, T exc , of higher rotational states ( J = 3 −
6) isinfluenced by the FUV radiation field, which excites and photodissociates H (Jura 1974; Black & Dalgarno1976). From FUSE data on these excited states, we compute temperatures, T and T , based on population 3 –ratios of J = 2 to 4 and J = 3 to 5, respectively.Section 2 begins with a description of the OB-star sample and our methods of data acquisition andanalysis. Section 3 provides the survey results, including H and H I column densities, molecular fractions,rotational excitation temperatures, and gas-to-dust ratio, N H /E ( B − V ). We also present a simple analyticmodel of the transition from H I to H , including H formation rates, FUV dissociation, self-shielding, anddust opacity. The transition occurs at f H2 ≈ . τ d ≈ (cid:104) T (cid:105) = 88 ±
20 K and (cid:104) T (cid:105) = 77 ±
18 K. Higher rotational states ( J ≥
3) have larger excitation temperatures, T exc ≈ −
650 K, arising from fluorescent cascade following FUV radiative excitation in the Lyman andWerner bands. Section 4 summarizes our findings, with comparisons to the
Copernicus survey of H (Savageet al. 1977) and the N H /E ( B − V ) ratio (Bohlin et al. 1978). The FUSE mean value for the Galactic disk, (cid:104) N H /E ( B − V ) (cid:105) = (6 . ± . × cm − mag − , is lower than estimates from 21-cm/far-IR studiesat high Galactic latitudes (Liszt 2014a,b), suggesting that gas and dust have different spatial distributionsabove the disk.
2. Data Acquisition and Reduction2.1. FUSE Observations
The 139 targets in this survey were drawn primarily from
FUSE programs designed to study OB starsand interstellar gas in the Milky Way.
Figure 2 shows the target distribution in Galactic longitude andlatitude, with color coding for their distances. Most stars are O-type and early B-type (B0, B0.5) andgenerally more distant than those in the 1977
Copernicus survey. Many were part of
FUSE science teamprojects to study O VI from hot gas in the Galactic disk (Bowen et al. 2008), interstellar D/H (Moos et al.2002; Hoopes et al. 2003; H´ebrard et al. 2005), and hot stars and their winds (Massa et al. 2003). Severalstars were analyzed specifically for their interstellar H (Shull et al. 2000; Snow et al. 2000). Table 1 lists information on the 139 stellar targets and their observational characteristics. These starshave updated spectral types (SpT) and both photometric and
Gaia -DR2 parallax distances determinedby Shull & Danforth (2019) using new information from the Galactic O-star Spectroscopic Survey (Ma´ızApell´aniz et al. 2004). The GOS project generated a large sample of O stars within several kiloparsecsof the Sun with updated spectral classification (Sota et al. 2011, 2014), together with digital photometryand optical-NIR dust extinction (Ma´ız-Apell´aniz & Barb´a 2018). The first ten columns of Table 1 provideour internal target ID, star name, Galactic coordinates ( (cid:96) , b ), photometry ( B and V magnitudes), colorexcess E ( B − V ), SpT, and both photometric and parallax distances. The last two columns list the FUSE program ID and exposure time of the primary observation. In some cases, we used other
FUSE observationsto supplement or confirm our measurements. Footnotes explain the sources for photometry and distances.All column densities are expressed in units of cm − and often quoted in logarithmic format (log N ). Thetotal hydrogen column density, N H = N HI + 2 N H2 , is used to derive the molecular fraction, f H2 = 2 N H2 /N H ,and explore its correlations with E ( B − V ) and line-of-sight pathlength. Most H I column densities, N HI ,were taken from previous Ly α profile-fitting surveys (Shull & Van Steenberg 1985; Diplas & Savage 1994;Jenkins 2019) supplemented by a few individual Ly α measurements. Table 2 presents our adopted valuesof N HI and a comparison to previous measurements. We adopted N HI in priority order of: (J19) 57 starsfrom Jenkins (2019); (DS94) 51 stars from Diplas & Savage (1994); (FM90) two stars from Fitzpatrick &Massa (1990); and three stars from other sources. A comparison between J19 and DS94 found generally 4 –good agreement, within 0 . − .
08 in log N . Only seven sight lines differed by larger amounts (0 . − . α fits, we estimated N HI from the scaling relation (Bohlin et al.1978) of total hydrogen column density with color excess, N H = (5 . × cm − mag − ) E ( B − V ), where N HI = N H − N H2 . The data analysis for this survey used numerous absorption lines from H Lyman and Werner transitionsbetween 930 - 1126 ˚A. Our methodology is identical to that used in the
FUSE survey of H in the LMC andSMC (Tumlinson et al. 2002), and we refer readers to that paper for details. We search for absorption linesfrom J = 0 − J = 5. Most lines exhibit shifts in wavelength (0.03–0.13 ˚A)with offsets determined by measuring narrow interstellar metal lines such as Ar I (1048.218 ˚A, 1066.660 ˚A),Fe II (1055.269 ˚A, 1063.177 ˚A, 1144.938 ˚A), Si II (1020.699 ˚A), P II (1152.818 ˚A), and other lines of Fe II , O I ,N I . This procedure is similar to that employed by Tumlinson et al. (2002), Wakker (2006), and Gillmon et al.(2006). We also considered the R(0) lines of HD in its (3-0), (4-0), (5-0) bands at 1066.271 ˚A, 1054.286 ˚Aand 1042.847 ˚A, respectively. However, the HD lines were often offset in velocity from H , possibly becauseof its presence in only one of the components of the H line profile.The most accurate determinations of column densities N come from weak lines where equivalent width W λ ∝ N , or from strong lines with damping wings where W λ ∝ N / . In both limits, the column density canbe determined without knowledge of the Doppler parameter ( b ). Between these two regimes, from line-centeroptical depth τ ≈ τ damp ≈ − , one must employ curve-of-growth (CoG) methods that requirean estimation of b . The linear relation between W λ and N and the onset of line saturation are measured byoptical depth at line center, W λ = (cid:18) πe m e c (cid:19) N f λ c = (88 .
53 ˚A) N λ (cid:20) f λ
10 ˚A (cid:21) (1) τ = (cid:18) πe m e c (cid:19) N f λ √ π b = (1 . N b − (cid:20) f λ
10 ˚A (cid:21) . (2)Here, we scaled to column density N = (10 cm − ) N , Doppler parameter b = (10 km s − ) b , andwavelength λ = (1000 ˚A) λ . Line strengths are normalized to f λ = 10 ˚A, a typical value for many linesin the Lyman and Werner bands.Equivalent widths of the highly saturated lines from J = 2 and J = 3 are frequently quite large, withdimensionless values W λ /λ ≈ (1 − × − . On the flat portion of the CoG, these widths are primarilydetermined by the effective b -value, with asymptotic values W λ /λ ≈ (2 b/c )[ln( τ / ln 2)] / . As shown inthe CoG suite prepared by McCandliss (2003), strong lines on the equivalent width plateau often requireDoppler parameters, b = 10 −
20 km s − , much higher than the expected thermal values, b th = (2 kT /m ) / ≈ (0 . − ) T / , for H at temperatures T = (100 K) T . They likely result from velocity components The equivalent width ( W λ ) of a saturated absorption line rises slowly with increasing column density N and line-centeroptical depth τ . As shown by asymptotic analysis, W λ /λ ≈ (2 b/c )[ln( τ / ln 2)] / for large τ and a Doppler-broadenedGaussian velocity profile, φ ( v ) ∝ exp( − v /b ). Draine (2011) noted that this formula is accurate to 5% for 1 . < τ < τ damp ,up to the onset of damping wings. Inverting this relation, we find τ ≈ (0 . W λ /λ ) / (2 b/c ) ]. Clearly, τ and N areexponentially sensitive to measurements of W λ and b . When equivalent widths of the accessible lines lie on a flat CoG, errorscan exceed ± .
40 for log N = 16 −
17 and ± .
50 for log N = 17 − b -values increase with rotation level. Although we found severalcases where the higher- J lines required larger b -values, we saw no evidence of a uniform trend in the currentsurvey or in our study of H in the Magellanic Clouds (Tumlinson et al. 2002).For absorbers with N ( J ) ≤ cm − , we measured equivalent widths of the accessible H lines,producing a CoG that yields a b value and column densities, N ( J ), in rotational states J = 0 −
5, andoccasionally J = 6 and J = 7. Table 3 lists column densities in J = 0 − b -value whenmeasurable. Most column densities range from log N ( J ) ≈ . − . J = 0 − N ( J ) ≈ . − . J = 4 −
6. In 48 targets, we found detectable column densities in J = 6. For other sight lines, we quoteupper limits, which range from log N (6) < . − .
40 depending on the data quality. Six targets haddetectable column densities in J = 7, described in footnote (a) to Table 3.For absorbers with damping wings in R(0), R(1), and P(1) lines, we fitted line profiles to derive N (0)and N (1), a technique used in previous studies of H (Savage et al. 1977; Tumlinson et al. 2002; Jenkins2019). The errors on these column densities depend on data quality, both the signal-to-noise (S/N) and ourability to define the continuum on either side of the blended R (0), R (1), P (1) complex (see the bottom panelof Figure 1 for an example). Based on our experience with fitting damping wings, we define three levels ofS/N ratio and corresponding errors, σ log N , on log N (0) and log N (1): (1) S/N ≥
15 (0.03–0.05 errors); (2)5 ≤ S/N ≤
15 (0.05–0.10); and (3)
S/N ≤ absorption lines as possible in each sight line. We then use CoGs to derive column densities for the undampedlines in J ≥
2. The programs are written to consistently measure all available (unblocked) absorption lines inthe spectra. In many instances, lines from J = 2 and J = 3 were strongly saturated, with equivalent widthslying on the flat (Doppler-broadened) portion of the CoG where column densities are difficult to determine.In this case, W λ depends primarily on the Doppler parameter, which is often much larger than the thermalvalue because of multiple velocity components. The process of measuring the numerous lines that enter theCoG fitting cannot be automated completely. Our software requires the user to decide on a line-by-line basiswhich H lines will be fitted. We ignore H lines in spectral regions near strong interstellar absorption orbright geocoronal emission. For example, the H Lyman (6-0) band is obscured by strong damping wingsof the interstellar Ly β line (1025.722 ˚A), and the Lyman (5-0) band lies among the resonance absorptionlines of C II (1036.337 ˚A) and C II ∗ (1037.018 ˚A). We neglect these bands except when the higher- J linesare separated from the intervening absorption.For the higher- J lines, the software steps through the expected positions of the lines, band by band,shifted by the approximate velocity offset from the metal lines. Going through each complete band oflines consecutively, the routine displays the area where a line should appear. If the line is present andunblended with any other transitional or metal line, it is fitted to a Gaussian, and its equivalent width,errors, wavelength, full width half maximum, and velocity offset are entered into a table. The Lyman (6-0)band is typically omitted, and the Lyman (5-0) band is measured only partially, because of overlap with C II λ b -values. Asymmetric error bars are generated for b and column densities. 6 –As noted above, column densities from J = 2 and J = 3 are difficult to measure, because most of theirlines are highly saturated. To alleviate CoG uncertainties and determine an accurate b -value it is helpful tomeasure the weakest available lines, with strengths f λ ≤ J = 2: (0-0)P(2) at 1112.495 ˚A ( f λ = 0 .
740 ˚A); (0-0) R(2) at 1110.120 ˚A ( f λ = 1 .
199 ˚A); and (1-0) R(2) 1094.244 ˚A( f λ = 4 .
016 ˚A). These lines are blended with (0-0) P(1), (0-0) R(3), and (1-0) P(1), respectively. For J = 3,we were unable to measure the two weakest lines: (0-0) P(3) at 1115.896 ˚A ( f λ = 0 .
784 ˚A) blended with (0-0)R(4) at 1116.013 ˚A, and (0-0) R(3) at 1112.584 ˚A ( f λ = 1 .
135 ˚A) blended with (0-0) P(2) at 1112.495 ˚A.The weakest line that could be measured was (1-0) P(3) at 1099.788 ˚A ( f λ = 2 .
639 ˚A). For J = 4, theweakest lines, (0-0) P(4) at 1120.247 ˚A ( f λ = 0 .
808 ˚A) and (0-0) R(4) at 1116.013 ˚A ( f λ = 1 .
116 ˚A), areblended with (0-0) P(3) 1115.896 ˚A and (0-0) R(5) 1120.300 ˚A, respectively. In some cases we were ableto separate (0-0) P(4). In general the most accessible lines were the P(4) lines in (1-0), (2-0), (3-0), (4-0),(5-0), (7-0) bands and the R(4) lines in (3-0), (4-0), (5-0) bands. For J = 5, we found a sufficient number oflines with a range of strengths for a reliable CoG.A systematic uncertainty in our results comes from the possibility of multiple components in the ab-sorption lines caused by more than one cloud in the line of sight. Many spectra ( ∼
40 sight lines) havevelocity components that make the neighboring lines visually identifiable, but not separable without carefulprofile fitting using other information from higher resolution optical lines. Some absorbers in this surveyhave components that are not resolvable; those lines are treated as though they are a single component. Thespectra of 15 targets exhibited obvious multiple components separated by ∆ v ≥
20 km s − . We measuredthose individually using a double Voigt profile fit. The components have their values entered into separatedata tables, enabling us to generate two CoGs with individual b values. In these cases, we report the to-tal column densities measured for saturated lines or lines without visible structure and individual columndensities where components are measurable. Lines at J ≤
3. Results3.1. Molecular Abundances
The
FUSE survey finds H everywhere in the disk of the galaxy ( Figure 2 ) at typical Galactic latitudes | b | < ◦ . The H column density rises rapidly above N H2 ≥ . cm − for sight lines with color excess E ( B − V ) (cid:38) .
2, as illustrated in
Figure 3 . The molecular fraction, f H2 = 2 N H2 /N H , quantifies thenumber of hydrogen nuclei bound into H molecules, where N H = N HI + 2 N H2 is the total hydrogen columndensity. Later in this section, we will adopt N H2 = 10 . cm − and f H2 = 0 . N H2 = 17 . N H2 = 18 . N H2 = 18 . N H2 ≥ . Table 4 summarizes the column densities ( N HI , N H2 , N H ) together with the molecularfractions and rotational excitation temperatures inferred from populations in levels J = 0 −
5. Rotationalexcitation temperatures, T , T , T , and T , are discussed further in Sections 3.4 and 3.5.The atomic-to-molecular transition arises from both H self-shielding in optical thick lines and dustattenuation of the dissociating FUV radiation (Browning et al. 2003; Krumholz et al. 2008, 2009; Sternberget al. 2014). The Copernicus survey of 61 stellar targets (Savage et al. 1977) noted that f H2 rises rapidlyabove 1% at E ( B − V ) ≥ .
08 and N H ≥ × cm − . A transition to f H2 = 0 . N H ≈ .
4) in sight lines at high Galactic latitude (Gillmon et al. 2006). Inthe lower-metallilcity LMC and SMC (Tumlinson et al. 2002; Browning et al. 2003) the transition shifted tohigher column densities (log N H ≈ . − . formation rate. It also allows deeper penetration of FUV photons into the cloud, resultingin more H dissociation. These effects are discussed further in Section 3.3. Figure 4 presents the distributions of molecular fraction vs. E ( B − V ) and N H for 139 targets in the FUSE survey.
Figure 5 plots mean values of f H2 and hydrogen density n H = N H /D along the sight lines vs.target distance D . These averages were evaluated as (cid:104) n H (cid:105) = (cid:80) N H / (cid:80) D and (cid:104) f H2 (cid:105) = (cid:80) N H2 / (cid:80) [ N HI +2 N H2 ]. Both parallax and photometric distance are shown in the figure, with error bars reflecting formalparallax uncertainties from Gaia -DR2. With many sight lines to bright OB stars ( D ≤ Copernicus survey showed an increase of f H2 ≥ .
01 at E ( B − V ) ≥ .
08 (Savage et al. 1977). The
FUSE survey includesmore distant sight lines, with column densities up to log N H ≈ .
65. We found a transition at f H2 ≥ . E ( B − V ) (cid:38) . N H (cid:38) cm − and N H2 (cid:38) . cm − . The molecular fractions rise from f H ≈ N H ≈ .
40 to 21.65. Using new photometric distancestoward 129 stars at D phot ≤ (cid:104) n H (cid:105) = 0 .
50 cm − andmolecular fraction (cid:104) f H2 (cid:105) = 0 .
20. These mean values shift toward higher values (
Table 5 ) in sub-sampleswith D phot ≤ D ≤ (cid:104) n H (cid:105) = 0 .
81 cm − and molecular fraction (cid:104) f H2 (cid:105) = 0 . N H /E ( B − V ) Ratio
The optical extinction along Galactic sight lines is often taken to be proportional to the dust columndensity, and therefore to the gas column density. This “gas-to-dust ratio” assumes a homogeneous mixtureof interstellar hydrogen and grains, which may be a good assumption for most regions of the diffuse ISM.Deviations can be produced by changes in the grain size distribution and other physical properties that arisewithin dark clouds such as ρ Oph (Bohlin et al. 1978; Green et al. 1992) or in regions where shock waveshave sputtered or destroyed some of the grains (Seab & Shull 1983). The total hydrogen column density, N H = N HI +2 N H2 , is often compared to dust content through its ratio to color excess, N H /E ( B − V ), derivedfrom UV surveys of H I (Ly α ) and H toward early-type stars (Bohlin et al. 1978; Savage et al. 1977).Our FUSE survey of the Milky Way disk should be more robust, with more stellar sight lines, updatedO-star photometry and SpTs from the GOS survey, and newly derived values of E ( B − V ) and targetdistances (Shull & Danforth 2019). Figure 6 shows the distribution of N H /E ( B − V ) vs. E ( B − V ). For129 stars at D ≤ (cid:104) N H /E ( B − V ) (cid:105) = (6 . ± . × cm − mag − with (rms)variations shown as blue wash. A sub-sample of 56 stars at D ≤ (cid:104) N H /E ( B − V ) (cid:105) =6 . × cm − mag − . Both values are slightly above the values 5 . × cm − mag − in the Copernicus survey of 75 stars (Bohlin et al. 1978) and 5 . × cm − mag − in the FUSE survey of 38 translucentsight lines with A V ≈ . − . N HI is measured from 21-cm emission and E ( B − V ) is inferredfrom all-sky maps (Schlegel et al. 1998; Schlafly & Finkbeiner 2011) of far-infrared (FIR) dust emission from IRAS and
COBE /DIRBE. Liszt (2014a,b) found N HI /E ( B − V ) = 8 . × cm − mag − in high-latitudesight lines ( | b | > ◦ ) with low extinction, 0 . < E ( B − V ) < . . × cm − mag − . These studies used only H I , but as the authors comment, corrections 8 –for H are normally small for E ( B − V ) < . E ( B − V ) have different dust-to-gas ratios in the UVsurveys. For 25 stars with 0 . ≤ E ( B − V ) ≤ .
08 in the
Copernicus / IUE survey (Bohlin et al. 1978) themean ratio is 4 . × cm − mag − . For the 21 stars in our FUSE survey with E ( B − V ) ≤ .
25, we find amean ratio N H /E ( B − V ) = 5 . × cm − mag − . We omitted one outlier (HD 3827) with log N H = 20 . E ( B − V ) ≈ .
02, and a high ratio N H /E ( B − V ) ≈ × cm − mag − . Located atphotometric distance D phot ≈ .
88 kpc, HD 3827 lies 700 −
800 pc below the Galactic plane at b = − . ◦ .Its color excess, E ( B − V ) = 0 .
02, was based on magnitudes B = 7 .
76 and V = 8 .
01 (Deutschman et al.1976) and intrinsic color ( B − V ) = − .
27. Jenkins (2019) listed E ( B − V ) = 0 .
05 for this star, based on B = 7 .
76 and V = 7 .
95, which would reduce the ratio to 7 × cm − mag − .The difference between the two techniques (UV absorption and radio/FIR emission) appears to be aneffect only seen at high Galactic latitudes (Liszt 2014b; Hensley & Draine 2021). Elevated ratios fromUV data do not appear toward disk stars with low E ( B − V ). We suggest that the high ratios in 21-cm/FIR measurements result from different distributions of gas and dust above the disk. Dust grains areproduced by stars in the disk and grow in the ISM through accretion of refractory elements. Some grainsare transported above the disk plane by radiation pressure and supernova-driven outflows. Other grainsmay settle gravitationally into lower scale-height distributions, separating from the high-latitude H I . Dustat high latitudes may also come into contact with hot coronal gas at 10 − K and experience erosion bythermal sputtering and destruction by fast interstellar shock waves (Jones et al. 1996; Slavin et al. 2004).In hot, low-density halo gas, grain lifetimes from sputtering are t sp ≈ (1 Gyr)(10 − cm − /n e ). Thus, the21-cm and far-IR surveys at | b | > ◦ likely probe systematically lower dust-to-gas ratios. As we will describe, the molecular transition from H I to H is consistent with models involving H self-shielding and efficient H formation by atomic processes on grain surfaces (Hollenbach et al. 1971; Jura1975a,b; Shull & Beckwith 1982). In equilibrium, the abundance ratio, n H2 /n H , can be expressed as abalance between H formation and photo-dissociation. For number densities n HI , n H2 , and total hydrogen n H , molecule formation occurs at a rate per unit volume, Rn H n HI . The total hydrogen density serves asa proxy for dust grains, whose surfaces are catalysts for H formation. The coefficient R depends on thedust-to-gas ratio, metallicity, and atom-grain collisional rates, and it likely varies with gas temperature,grain temperature, and surface physics (Hollenbach & McKee 1979).Previous studies (Browning et al. 2003; Krumholz et al. 2008, 2009; Sternberg et al. 2014) have analyzedthe molecular transition with radiative transfer. Here, we present a simple analytic description of thetransition (at f H2 ≈ .
1) tied directly to parameters that control the H equilibrium between formation anddestruction and the attenuation of FUV flux by H self-shielding and dust opacity. In equilibrium, n HI n H R = ( f diss Gβ ) n H2 S H2 e − τ d , (3)where β is the average unshielded absorption rate of H in the Lyman and Werner bands. For the localISM, Jura (1974) estimated β = 5 × − s − with f diss ≈ .
11. With updated line-dissociation data andmodels of self-shielding, Draine & Bertoldi (1996) found a mean fraction (cid:104) f diss (cid:105) ≈ .
15 for all H absorptionsinside the cloud. The parameter G allows for local elevation of the FUV radiation relative to its averagevalue (Habing 1968), τ d is the dust optical depth at 930–1130 ˚A, and the factor S H2 accounts for H self-shielding as the absorption lines become optically thick. For clarity, we define the local molecular fraction 9 – f = 2 n H2 /n H , with n HI = (1 − f ) n H . This fraction varies with depth into the cloud, whereas the observed fraction, f H2 ≡ N H2 /N H , depends on the integrated column densities. This leads to an expression, f (1 − f ) = (cid:18) n H Rf diss β G (cid:19) S − e τ d . (4)Draine & Bertoldi (1996) provided a reliable approximation, S H2 = [ N H2 / cm − ] − . , for N H2 ≥ cm − . Combining S H2 = AN − . ( A = 3 . × cm / ) with the approximation, f ≈ n H2 /n H for f ≤ .
1, we can write the local molecular density as n H2 = (cid:18) Rn f diss β GA (cid:19) N . e τ d . (5)In a planar slab of constant density n H , with column density N H2 at a distance x into the absorber, wecan write n H2 = dN H2 /dx . The dust optical depth is τ d ( x ) ≈ n H x/N d , where N d ≈ . × cm − at λ ≈ − dN H2 dx = (cid:18) Rn f diss β GA (cid:19) N . e n H x/N d , (6)with the analytic solution N H2 = (cid:18) RN d n H f diss β GA (cid:19) (cid:104) e τ d ( x ) − (cid:105) . (7)The power-law approximation for S H2 ∝ N − . breaks down at low column densities, since S H2 = 1 at N H2 < cm − . However, this only occurs in a small surface layer ( f H2 < − ). Thus, the aboveexpressions are valid up to to f H2 = 0 . absorbers observed with FUSE . The observed atomic-to-molecular transition at N H2 ≈ . cm − occurs at dust optical depth τ d = N H /N d given by[ e τ d −
1] = (1 . (cid:18) N H2 . cm − (cid:19) / G n − (cid:18) × − cm s − R (cid:19) (cid:18) . × cm − N d (cid:19) (8)Here, we adopted f diss = 0 .
15 and
A/N d = 7 . × − cm / for metallicities and grain opacities inthe local ISM. We scaled the cloud density to n H = (30 cm − ) n , appropriate for thermal pressures, P/k ≈ − K, inferred from observations of C I fine-structure populations (Jenkins & Tripp 2011).Setting the parenthetical terms equal to unity, we find a dust optical depth τ d = 1 . G/R N d n H ), which is insensitive to metallicity if the product ( RN d ) remainsconstant. This would be expected if grain abundances decrease at lower metal abundances. In that case, N d would increase and R would decrease. Thus, the transition should be governed by the FUV/density ratio( G/n H ). In low-metallicity environments such as the LMC/SMC, the transition will occur at similar τ d ≈ N H = τ d N d .Previous studies of H abundances estimated that R ≈ × − cm s − (Jura 1975a). The observedtransition implies a rate coefficient, R = (cid:20) f diss β GAN d n H (cid:21) N / [ e τ d − − ≈ (3 . × − cm s − ) Gn − (cid:20) N H2 . (cid:21) / , (9)for τ d ≈
1, as found above. This result suggests that absorbers with densities n H >
30 cm − are associatedwith elevated radiation fields ( G >
1) from their proximity to hot stars. 10 –We see that dust can be an important factor, along with H self-shielding, in the onset of the moleculartransition. Our results show that the transition occurs when τ d ≈
1. Because the dust-to-gas ratio dependson abundances of C, Si, O, and other refractory heavy elements, changes in metallicity will have offsettingeffects on the H fractions through the formation coefficient ( R ) and radiative attenuation ( τ d ). This analysismay explain the observed hydrogen column densities of the molecular transition at f H2 = 0 .
1, which occursat log N H ≈ . N H ≈ . N H ≈ . N H ≈ . FUSE survey of the Milky Way disk exhibits no obvious shift with distance, but large changesin metallicity are not expected over the range of target star distances.More precise determinations of R require knowledge of n H , FUV radiation field, and dust properties,which may depend on metallicity and environment. Interstellar clouds are inhomogeneous, with internalvariations in temperature, hydrogen density, and metallicity. Irradiated cloud models have been constructed(Browning et al. 2003; Le Petit et al. 2006; Nehm´e et al. 2008; Klimenko & Balashev 2020) that followthe attenuation of UV radiation into the cloud and the resulting changes in gas temperature ( T g ), dusttemperature ( T d ) and molecular fraction. These models depend sensitively on n H , β , G , and R ( T g , T d , Z ).The best measures of the FUV radiation field and cloud density are the high- J excitation ratios, N (4) /N (2)and N (5) /N (3). These issues are discussed further in Section 3.5. J = 0 , , ) The excitation temperature T of the lowest two rotational states, J = 0 (para-H ) and J = 1 (ortho-H ) is frequently used as a measure of the kinetic temperature in diffuse clouds. This requires that the gasdensity and column density be sufficiently high for thermal proton collisions (Gerlich 1990) to couple the orthoand para forms of H and set the ratio N (1) /N (0). Recent experiments suggest that ortho-para conversionmight occur on silicate grain surfaces (Tsuge et al. 2021). The expectation is that ortho/para production ongrains would be in the 3:1 spin-statistical ratio. The observed low values of T ≈ −
90 K suggest thatthis formation channel is subdominant. Theoretical models of H formation and destruction in diffuse cloudsfind that the lowest three rotational states ( J = 0 , ,
2) are usually thermalized, with populations dependingprimarily on gas temperature. We assume that the observed populations obey Boltzmann ratios, with T and T determined from the expressions, T = ∆ E /k ln[( g /g ) N (0) /N (1)] , (10) T = ∆ E /k ln[( g /g ) N (0) /N (2)] . (11)Here, g /g = 9 and g /g = 5 are ratios of statistical weights of the rotational levels, and ∆ E /k = 170 .
48 Kand ∆ E /k = 509 .
86 K come from the rotational energies computed by Komasa et al. (2011).The initial
Copernicus study of H in 13 clouds with N (0) ≥ cm − found a mean temperature (cid:104) T (cid:105) = 81 ±
13 K (Spitzer & Cochran 1973). A more extensive
Copernicus survey (Savage et al. 1977)of 61 stars with log N H2 ≥
18 found (cid:104) T (cid:105) = 77 ±
17 K (rms). In the current
FUSE survey of 139 stars,we find (cid:104) T (cid:105) = 88 ±
20 K.
Figure 7 shows the distributions of T with E ( B − V ) and N H . Removingthe four labeled outliers reduces the mean slightly to (cid:104) T (cid:105) = 87 K. The distributions of T are shownin Figure 8 . For the reduced sample (four outliers removed) of 128 stars with measured J = 2 column 11 –densities, we find (cid:104) T (cid:105) = 77 ±
18 K. In their analysis of H in the LMC and SMC, Tumlinson et al. (2002)found (cid:104) T (cid:105) = 82 ±
21 K for 22 sight lines with N H2 ≥ . cm − . Kruczek et al. (2019) explored thecontributions of higher rotational lines to the damping wings of J = 1 lines. Their re-analysis of nine Copernicus sight lines and 13 from
FUSE altered N (0) and N (1), resulting in a reduced T = 68 ±
13 K(12% lower than their previous values). Our
FUSE survey extends to larger column densities and greaterstellar distances than
Copernicus and contains more than twice the number of stars. Given the dispersionsin T and T distributions, their agreement suggests that the diffuse ISM has similar heating and coolingrates over a wide range of cloud densities, metallicities, and FUV radiation fields.Even for the best determinations of log N (0) and log N (1), rotational temperatures T have errors of σ T ≈ T can be higher, when log N (2) is poorly determined. From errors on log N ,log N , and log N , and neglecting covariance, the propagated errors on T and T derived from equations(10) and (11) are: σ T T = 2 . (cid:20) T .
48 K (cid:21) (cid:2) σ N + σ N (cid:3) / (12) σ T T = 2 . (cid:20) T .
86 K (cid:21) (cid:2) σ N + σ N (cid:3) / . (13)For equal errors on J = 0 and J = 1 column densities, σ log N = σ log N , this expression simplifies to σ T ≈ [ T / . σ log N . For the best fits to the damping wings ( σ log N = 0 .
03) the temperature uncertaintyis σ T = 4 . T ≈
87 K. In poorer quality data, with σ log N ≈ σ log N = 0 .
07, the uncertainty ishigher, σ T ≈
10 K. Errors on J = 2 column densities are usually much larger than those for J = 0. Thus, σ T ≈ [ T / . σ log N . At T ≈
77 K, the uncertainty σ T ≈ σ log N = 0 . − . J ≥ ) The higher rotational levels ( J ≥
3) of H are generally believed to be populated by the fluorescentcascade following FUV radiative pumping in the Lyman and Werner bands (Black & Dalgarno 1976; Spitzer& Zweibel 1974; Jura 1974). The FUV radiation field includes the ambient Galactic radiation field (Jura1974; Habing 1968; Draine 2011) augmented by local flux from O stars. Measurements by Copernicus (Spitzer et al. 1974; Morton 1975) found that levels J ≥ T . The column densities in J = 3 , , T exc ≈ −
500 K, andsometimes as high as 1100 K near luminous early O-type stars such as Zeta Puppis (Morton & Dinerstein1976). Subsequent studies of H excitation using FUSE data (Browning et al. 2003; Sheffer et al. 2008;Nehm´e et al. 2008; Jensen et al. 2010) also found T exc ≈ −
500 K for J ≥
3. The mean rotationalexcitation temperature, fitted to J ≥ T exc = 326 ±
125 K with a typicalrange from 150–650 K. These temperatures are similar to those seen toward selected
Copernicus targets, andthey likely reflect fluorescent pumping of high- J states. Some observations suggest that J ≥ T >
500 K) heated by turbulent dissipation (Moseley et al. 2020). Observationally, CH + iscorrelated with rotationally-excited H (Jensen et al. 2010), leading to suggested production schemes forCH + involving hot H (Falgarone & Puget 1995; Myers et al. 2015).In our survey, we choose to focus on individual pairs of upper ( u ) and lower ( l ) rotational states ( J l , J u ),in particular (0,2), (2,4), and (3,5). These parameters capture the fact that H exists over a range ofcloud temperatures, with radiative pumping changing throughout the cloud because of H self-shielding 12 –and attenuation by FUV extinction. In addition, the rate of ortho-para conversion may change, dependingon cloud density. This will affect radiative pumping from J = 0, J = 1, and sometimes J = 2, creatingdepartures of rotational populations of J ≥ T , T , and T , each remaining within para (even- J ) andortho (odd- J ) forms of H . The temperatures T and T are not as useful diagnostics. Detailed modelsof the excitation processes and radiative transfer may help to distinguish the relative contributions of FUVpumping and collisional excitation and to estimate the FUV radiation field and gas density.Based on column densities N ( J ) in the FUSE survey,
Table 4 lists four excitation temperatures, T , T , T , and T , between upper and lower rotational states defined by T lu = ( E u − E l ) /k ln[( g u /g l )( N l /N u )] , (14)corresponding to Boltzmann population ratios, N (1) /N (0) = (9 /
1) exp[ − .
48 K /T ] (15) N (2) /N (0) = (5 /
1) exp[ − .
86 K /T ] (16) N (4) /N (2) = (9 /
5) exp[ − .
78 K /T ] (17) N (5) /N (3) = (11 /
7) exp[ − .
66 K /T ] . (18)These excitation temperatures were derived from the relativistic quantum calculations of Komasa et al.(2011), using the J -level dissociation energies in their Table 1. From the observed populations of higher- J states, we find mean excitation temperatures, (cid:104) T (cid:105) = 237 ±
91 K and (cid:104) T (cid:105) = 304 ±
108 K. Even with awide range of these temperatures, they generally exhibit a correlation between the two parameters.
Figure 9 shows the rotational distributions in four of the six sight lines with detections up to J = 7.These include Star N (7) = 15 . ± .
15; Star N (7) =14 . ± .
06; Star N (7) = 14 . ± .
11; and Star N (7) = 15 . ± .
15. The absorbers along these sight lines have different excitation temperatures. Twoexhibit higher excitation temperatures for J = 4 and J = 5, likely produced by elevated FUV radiationfields from absorber proximity to hot stars. Figure 10 displays distributions of the four excitation temperatures with stellar target distance. Wesee no obvious trend with increasing distance, although stars beyond 4–5 kpc are unlikely to be an unbiasedsample. The similar values of of T and T suggest that the J = 0 , , T and T show several outliers, well above their distribution means of 237 Kand 304 K, respectively. For this 10-15% population, higher rotational excitation is expected from exposureto local FUV radiation above the background. Figure 11 shows the relation of T and T , color-coded by the SpT of the target star. With theexception of a few labeled outliers, the para-H levels ( J = 2 and 4) and ortho-H levels ( J = 3 and 5) havecorrelated excitation temperatures above 300 K. Most sight lines have T > T , with data points abovethe dashed line of unit slope. This difference could arise from strong pumping of ortho-H populations outof J = 1. Alternatively, thermalization of the J = 2 population to the gas kinetic temperature may alterthe pumping out of J = 2. The expectation that high- J populations would be greater toward hotter O-typestars is not consistently reflected in this plot. The labeled sight lines with T >
500 K include two earlyO-type stars ( T >
450 K), while others have T <
400 K. 13 –This mixed distribution suggests that radiative pumping of J ≥ N (4) /N (2) and N (5) /N (3), as performed byBrowning et al. (2003) and Klimenko & Balashev (2020).
4. Discussion and Summary
This
FUSE survey of interstellar H in the Milky Way disk complements the pioneering survey by Copernicus , with several important extensions. Because
FUSE was a more sensitive spectrograph, thesurvey includes more OB-star targets at greater distances and larger H column densities. Many FUSE sightlines exhibit molecular fractions above the value, f H2 ≥ .
01 at log N H ≥ .
7, noted in the
Copernicus survey (Savage et al. 1977). We also measured H populations in higher rotational states ( J ≥
2) as well as J = 0 and J = 1. The FUSE sample includes the 139 OB-star targets with updated distances and E ( B − V )from Shull & Danforth (2019) derived from updated spectral types, digital photometry, and optical-NIR dustextinction in the Galactic O-star Spectroscopic Survey (Ma´ız Apell´aniz et al. 2004; Ma´ız-Apell´aniz & Barb´a2018). Our measurements of H populations in the lowest three rotational states ( J = 0 , ,
2) found similarexcitation temperatures, T ≈ T ≈ −
90 K, suggesting thermal coupling to the gas kinetic temperature.Populations of higher rotational states ( J ≥
3) could be use to distinguish between radiative pumping andcollisional excitation. The pairwise excitation temperatures, T and T , are correlated at the high end ofthe distribution, with T > T . After deriving total hydrogen column densities, N H , from those of H andH I , we compared them to updated values of selective extinction, E ( B − V ), to find a mean gas-to-dust ratioin the Galactic disk, (cid:104) N H /E ( B − V ) (cid:105) = (6 . ± . × cm − mag − .The FUSE survey finds that a typical atomic-to-molecular transition in the ISM of the Galactic diskoccurs at molecular fraction f H2 ≈ . N H2 ≈ . cm − . This transition canbe understood with a simple analytic model describing its dependence on H formation on dust-grain surfacesand photo-dissociation by FUV radiation, including self-shielding and dust attenuation. This formulationshows that the transition depends on the ratio of FUV flux to gas density, analogous to the photoionizationparameter used in the analysis of nebular lines. The transition occurs at dust optical depth τ d ≈ N H = τ d N d , where the dust-opacity parameter N d ≈ . × cm − at solarmetallicity ( Z ≈ Z (cid:12) ). The H I -to-H conversion is mediated by both dust opacity and H self-shielding.The optical depth τ d ≈ formation rate coefficient ( R ) and dust opacity ( N d ).With a mean fractional abundance (cid:104) f H2 (cid:105) ≈ . µ m emission from the J = 2 level of H (510 K excitation) augments the dominant cooling fromthe 157.74 µ m [C II ] fine-structure line (91.21 K excitation). At low densities, the cooling rate from H ◦ -H collisions is L H2 = (3 . × − erg cm s − ) n HI n H2 at T = 100 K (Forrey et al. 1997), while that from [C II ]is L CII ≈ (3 . × − erg cm s − ) n . We adopted an electron-impact collision strength Ω = 1 .
56, a solarcarbon abundance n C /n H = 2 . × − , and electron density n e ≈ . × − n H donated by trace metalions. Thus, L H2 / L CII ≈ . f H2 (1 − f H2 ) at 100 K. In higher density clouds, the H µ m emission willbe reduced by collisional de-excitation. The H cooling rises at T >
100 K, but drops off exponentially inlower temperature clouds at higher N H , owing to the 510 K excitation temperature of the J = 2 level. 14 –The following summarizes the results of the FUSE survey of interstellar H abundances and inferred physicalparameters in the Milky Way disk:1. The FUSE survey measured column densities of H in the Galactic disk toward 139 OB stars withrecently updated SpT, photometry, and distances (Shull & Danforth 2019). The survey extends the Copernicus H survey (Savage et al. 1977) up to total hydrogen column densities N H ≈ × cm − and complements FUSE surveys of H in the Magellanic Clouds (Tumlinson et al. 2002), translucentclouds (Rachford et al. 2002, 2009), and gas at high Galactic latitude (Gillmon et al. 2006; Wakker2006).2. For each sight line, we report column densities N H2 , N HI , N ( J ), N H = N HI + 2 N H2 , and f H2 =2 N H2 /N H , with mean values listed in Table 5. The mean gas-to-dust ratio, (cid:104) N H /E ( B − V ) (cid:105) = (6 . ± . × cm − mag − , is slightly above the value of 5 . × cm − mag − in the Copernicus survey(Bohlin et al. 1978). The larger ratios seen in 21-cm/far-IR surveys at high Galactic latitudes (Liszt2014a,b) suggest different distributions of gas and dust above the disk, produced by grain sedimentationto the disk plane or dust destruction when transported above the disk.3. Using an analytic model of H formation-destruction equilibrium with dust opacity and H self-shielding, we derive an expression for the atomic-to-molecular transition, which occurs at optical depth τ d ≈
1, molecular fraction f H2 ≈ . N H2 ≈ . cm − , and N H ≈ cm − . An H formation ratecoefficient R ≈ × − cm s − is consistent with the observed transition, with occasional elevatedFUV radiation fields in 10-15% of the sight lines. These parameters can be constrained with modelsof H rotational populations, supplemented by C I fine-structure abundances (Jenkins & Tripp 2011;Klimenko & Balashev 2020).4. The lowest three rotational states ( J = 0 , ,
2) appear to be thermally coupled by collisions to the gaskinetic temperature. The survey mean excitation temperatures are (cid:104) T (cid:105) = 88 ±
20 K and (cid:104) T (cid:105) =77 ±
18 K. For sight lines with E ( B − V ) > . N H > .
7, these temperatures decrease to50 −
70 K.5. Populations of higher- J states are produced primarily by radiative pumping from FUV radiation.From column-density ratios of rotational levels N (4) /N (2) and N (5) /N (3), we find mean excitationtemperatures, (cid:104) T (cid:105) = 237 ±
91 K and (cid:104) T (cid:105) = 304 ±
108 K (rms). In most cases, these two temperaturesare correlated, with T > T , but we find no consistent connection with SpT (from O3 to B1).Elevated radiative pumping is likely produced by close proximity to hot stars, possibly with somecollisional excitation from heating by turbulent dissipation. Acknowledgements.
This work was supported by the
FUSE mission, with financial support from NASAContract NAS5-32985 to Johns Hopkins University and a sub-contract to the University of Colorado atBoulder. We thank Jason Tumlinson for developing software for
FUSE studies of H and former CU under-graduates Teresa Ross and Kristen Gillmon for their assistance with data analysis during early stages of thisproject. We have benefitted from discussions on interstellar gas, molecules, and dust with Sergei Balashev,John Black, Bruce Draine, Kevin France, Ed Jenkins, Slava Klimenko, Harvey Liszt, Chris McKee, BlairSavage, and Don York. 15 – REFERENCES
Balashev, S. A., Klimenko, V. V., Noterdaeme, P., et al. 2019, MNRAS, 490, 2668Black, J. H., & Dalgarno, A. 1976, ApJ, 203, 132Bohlin, R. C., Savage, B. D., & Drake, J. F. 1978, ApJ, 224, 132Bowen, D. V., Jenkins, E. B., Tripp, T. M., et al. 2008, ApJS, 176, 59Browning, M. K., Tumlinson, J., & Shull J. M. 2003, ApJ, 582, 810Burgh, E. B., France, K., & McCandliss, S. R. 2007, ApJ, 658, 446Dalgarno, A., Black, J. H., & Weisheit, J. C. 1973, Astrophys. Letters, 14, 77Deutschman, W. A, Davis, R. T., & Schild, R. E. 1976, ApJS, 30, 97Diplas, A., & Savage, B. D. 1994, ApJS, 93, 211Draine, B. T. 2011, Physical Processes in the Interstellar and Intergalactic Medium(Princeton: Princeton Univ. Press)Draine, B. T., & Bertoldi, F. 1996, ApJ, 468, 269Falgarone, E., & Puget, J.-L. 1995, A&A, 293, 840Fitzpatrick, E. L., & Massa, D. 1990, ApJS, 72, 163Forrey, R. C., Balakrishnan, N., Dalgarno, A., & Lepp, S. 1997, ApJ, 489, 1000Gerlich, D. J. 1990, J. Chem. Phys., 92, 2377Gillmon, K., & Shull, J. M. 2006, ApJ, 636, 908Gillmon, K., Shull, J. M., Tumlinson, J., & Danforth, C. 2006, ApJ, 636, 891Green, J. C., Snow, T. P., Cook, T. A., Cash, W. C., & Poplawski, O. 1992, ApJ, 395, 289Gry, C., Boulanger, F., Nehm´e, C., et al. 2002, A&A, 391, 675Habing, H. J. 1968, B.A.N., 19, 421Hanson, M. M., Snow, T. P., & Black, J. H. 1996, ApJ, 392, 571H´ebrard, G., Tripp, T. M., Chayer, P., et al. 2005, ApJ, 635, 1136Hensley, B. S., & Draine, B. T. 2021, ApJ, 906, 73Hollenbach, D. J., & McKee, C. F. 1979, ApJS, 44, 555Hollenbach, D. J., Werner, M. W., & Salpeter, E. E. 1971, ApJ, 163, 165Hoopes, C. G., Sembach, K. R., H´ebrard, G., Moos, H. W., & Knauth, D. C. 2003, ApJ, 586, 1094Ingalls, J. G., Bania, T.M., Boulanger, F., et al. 2011, ApJ, 743, 174Jensen, A. G., Snow, T. P., Sonneborn, G., & Rachford, B. L. 2010, ApJ, 711, 1236Jenkins, E. B. 2019, ApJ, 872, 55Jenkins, E. B., & Peimbert, A. 1997, ApJ, 477, 265Jenkins, E. B., & Tripp, T. M. 2011, ApJ, 734, 65Jones, A. P., Tielens, A. G. G. M., & Hollenbach, D. J. 1996, ApJ, 469, 740Jorgenson, R. A., Murphy, M. T., Thompson, R., & Carswell, R. F. 2014, MNRAS, 443, 2783Jura, M. 1974, ApJ, 191, 375Jura, M. 1975a, ApJ, 197, 575Jura, M. 1975b, ApJ, 197, 581
16 –
Klimenko, V. V., & Balashev, S. A. 2020, MNRAS, 498, 1531Komasa, J., Piszczatowski, K., Lach, G., et al., 2011, J. Chem. Theory Comput., 7, 3105Kruczek, N., France, K., Hoadley, K., Fleming, B., & Nell, N. 2019, ApJ, 878, 77Krumholz, M. R., McKee, C. F., & Tumlinson, J. 2008, ApJ, 689, 865Krumholz, M. R., McKee, C. F., & Tumlinson, J. 2009, ApJ, 693, 216Lacour, S., Ziskin, V., H´ebrard, G., et al., 2005, ApJ, 627, 251Le Petit, F., Nehm´e, C., Le Bourlot, J., & Roueff, E. 2006, ApJS, 164, 506Lenz, D., Hensley, B. S., & Dor´e, O. 2017, ApJ, 846, 38Liszt, H. 2014a, ApJ, 780, 10Liszt, H. 2014b, ApJ, 783, 17Ma´ız Apell´aniz, J., & Barb´a, R. H. 2018, A&A, 613, A9Ma´ız Apell´aniz, J., Walborn, N. R., Galu´e, H. A., & Wei, L. H. 2004, ApJS, 151, 103Massa, D., Fullerton, A. J., Sonneborn, G., & Hutchings, J. B. 2003, ApJ, 586, 996McCandliss, S. R. 2003, PASP, 115, 651Moos, H. W., Cash, W. C., Cowie, L. L., et al. 2000, ApJ, 538, L1Moos, H. W., Sembach, K. R., Vidal-Madjar, A., et al. 2002, ApJS, 140, 3Morton, D. C. 1975, ApJS, 197, 85Morton, D. C., & Dinerstein, H. L. 1976, ApJ, 204, 1Moseley, E. R., Draine, B. T., Tomida, K., & Stone, J. M. 2021, MNRAS, 500, 3290Myers, A. F., McKee, C. F., & Li, P. S. 2015, MNRAS, 453, 2747Nehm´e, C., Le Bourlot, J., Boulanger, F., et al., 2008, A&A, 483, 485Noterdaeme, P., Ledoux C., Petitjean, P., & Srianand, R. 2008, A&A, 481, 327Rachford, B. L., Snow, T. P., Tumlinson, J., et al. 2002, ApJ, 577, 221Rachford, B. L., Snow, T. P., Destree, J. D., et al. 2009, ApJS, 180, 125Sahnow, D. J., Moos, H. W., Ake, T. B., et al. 2000, ApJ, 538, L7Savage, B. D., Bohlin, R. C., Drake, J. F., & Budich, W. 1977, ApJ, 216, 291Schlafly, E. F., & Finkbeiner, D. P. 2011, ApJ, 77, 103Schlegel, D. J., Finkbeiner, D. P., & Davis, M. 1998, ApJ, 500, 525Seab, C. G., & Shull, J. M. 1983, ApJ, 275, 652Sheffer, Y., Rogers, M., Federman, S. R., et al. 2008, ApJ, 687, 1075Shull, J. M., & Beckwith, S. 1982, ARA&A, 20, 163Shull, J. M., & Danforth, C. W. 2019, ApJ, 882, 180Shull, J. M., Tumlinson, J., Jenkins, E. B., et al. 2000, ApJ, 538, L73Shull, J. M., & Van Steenberg, M. E. 1985, ApJ, 294, 599Slavin, J. D., Dwek, E., & Jones, A .P. 2004, ApJ, 803, 7Snow, T. P., Black, J. H., van Dishoeck, E. F., et al. 1992, ApJ, 465, 245Snow, T. P., Rachford, B. L., Tumlinson, J., et al. 2000, ApJ, 538, L65Sota, A., Ma´ız Apell´aniz, J., Morrell, N. I., et al., 2014, ApJS, 211, 10
17 –
Sota, A., Ma´ız Apell´aniz, J., Walborn, N. R., et al., 2011, ApJS, 193, 24Spitzer, L., & Cochran, W. D. 1973, ApJ, 186, L23Spitzer, L., Cochran, W. D., & Hirshfeld, A. 1974, ApJS, 28, 373Spitzer, L., & Zweibel, E. G. 1974, ApJ, 191, L127Sternberg, A., Le Petit, F., Roueff, E., & Le Bourlot, J. 2014, ApJ, 790, 10Tsuge, M., Namiyoshi, T., Furuya, K., et al. 2021, ApJ, in press (arXiv:2101.02357)Tumlinson, J., Shull, J. M., Rachford, B. L., et al. 2002, ApJ, 566, 857Wakker, B. P. 2006, ApJS, 163, 282
This preprint was prepared with the AAS L A TEX macros v5.2.
18 –
920 940 960 980 1000 102005101520
Werner Lyman
HD 46150
R0 R1 P1 R2 P2 R3 P3 R4 P4L5-0P4 L5-0R5 L5-0P5L 4-0Ar IAr IAr IAr IAr IAr IAr IAr IAr I Fe IIFe IIFe IIFe IIFe IIFe IIFe IIFe IIFe IIHDHDHDHDHDHDHDHDHD
Fig. 1.—
FUSE spectrum of the sight line to HD 46150, an O5 Vf star at 1.5 kpc distance and E ( B − V ) =0 .
45, located inside the Rosette Nebula. Locations of the H Lyman and Werner bands are shown as purpleand green templates in the top two panels. Each template shows a different vibrational band ( v u − v (cid:96) ) witha comb of absorption lines from rotational levels J in the ground vibrational state ( v (cid:96) = 0) to the uppervibrational state ( v u ) in the excited electronic states, B Σ + u (Lyman bands) and C Π u (Werner bands).For each band, the template shows locations of R-branch and P-branch absorption lines from J = 0 , , , ... .The bottom panel shows a close-up of the (4-0) Lyman band, with nearby lines of Ar I (1048.220 ˚A), Fe II (1055.262 ˚A), HD (4-0) R(0) at 1054.294 ˚A, and P(4), R(5), P(5) lines from J = 4 and J = 5 levels of the(5-0) Lyman band. The R(0), R(1), P(1) lines have damping wings and are blended. In some bands, the P(1)and R(2) lines are separable. The P-branch and R-branch lines from higher- J states are sufficiently shiftedto be measured. We find log N H2 = 20 . ± .
04, with log N (0) = 20 . ± .
05, log N (1) = 20 . ± .
05, androtational temperatures T = 97 ± T = 55 ± (cid:96) and b ) with distancescoded by color. Several sight lines with multiple velocity components that are separably measurable (∆ v ≥
20 km s − ) are shown as asterisks. 20 – l o g N ( H ) HD3827 HD92554 HD201638
Fig. 3.— Distribution of H column density with color excess, labeling three sight lines with log N H2 < . FUSE survey (blue circles) and translucent sight lines (orangestars) also studied by
FUSE (Rachford et al. 2002, 2009). We re-analyzed 11 of these translucent sight lineswith our H software for stars with new GOS photometry and SpTs (ID numbers 32, 82, 97, 98, 105, 107,114, 116, 120, 122, 127). 21 – f H f H HD3827 HD91651 HD92554 HD201638
Fig. 4.— Molecular fraction f H2 compared to color excess (left panel) and total hydrogen column density, N H = N HI + 2 N H2 (right panel). Vertical dashed lines show the transition to f H2 > .
01 at E ( B − V ) (cid:38) . N H (cid:38) . Copernicus data (Savage et al. 1977). Most of the
FUSE targets are more distantand have f H2 between 3% and 75%. Symbols are color-coded as in Figure 3. The outlier (orange star)with f H2 = 0 .
034 at E ( B − V ) = 0 .
87 and log N H = 21 .
96 is the translucent sight line toward HD 164740(Rachford et al. 2009). f H PhotometricGaia 0 2 4 6 8 10D (kpc)0.00.51.01.52.02.53.03.54.0 < n H > ( c m ) PhotometricGaia
Fig. 5.— Molecular fraction f H2 (left panel) and total hydrogen density n H (right panel), averaged overphotometric distance to target stars. Updated values (Shull & Danforth 2019) are shown for both photometricand Gaia -DR2 parallax distances. For 129 stars with photometric distances D ≤ (cid:104) n HI (cid:105) = 0 .
50 cm − , (cid:104) f H2 (cid:105) = 0 .
20, and (cid:104) N H /E ( B − V ) (cid:105) = 6 . × cm − mag − . Table 5 lists these quantities for the full survey and for sub-samples ( D ≤ D ≤ N H / E ( B - V ) ( c m m a g ) Fig. 6.— Distribution of the “gas-to-dust” ratio, N H /E ( B − V ), for 138 sight lines, in units of 10 cm − mag − . One star (ID E ( B − V ). Blue pointsare the 112 sight lines with N HI determined from Ly α profile fits. Red points are the 26 stars lacking Ly α fits for H I (Tables 2 and 4), where N HI was scaled from E ( B − V ). The mean ratio for the 112 stars is (cid:104) N H /E ( B − V ) (cid:105) = (6 . ± . × cm − mag − . The mean and (rms) deviations are shown as horizontallines and blue wash. 23 – T ( K ) HD3827 HD15642 HD92554 HD201638 )]406080100120140160180 T ( K ) HD3827 HD15642 HD92554 HD201638
Fig. 7.— Rotational temperature T vs. color excess (left panel) and total hydrogen column density (rightpanel). The mean value (cid:104) T (cid:105) = 88 ±
20 K (horizontal dashed line with 1 σ dispersions) should trackthe gas kinetic temperature in high-density clouds. Lower temperatures appear in translucent clouds at E ( B − V ) (cid:38) . N H (cid:38) .
5. Symbols are color-coded as in Figure 3 with four outlier targets labeled. T ( K ) HD3827 HD66788 HD14434BD354258 )]406080100120140160180 T ( K ) HD3827 HD66788 HD14434BD354258
Fig. 8.— Rotational temperature T vs. color excess (left panel) and total hydrogen column density (rightpanel). Four outlier targets are labeled. The other 128 stars have a mean value (cid:104) T (cid:105) = 77 ±
18 K (horizontaldashed lines with mean and 1 σ dispersions). Within individual sight-line errors (Section 3.4) and spreads ofthe distributions, (cid:104) T (cid:105) is similar to (cid:104) T (cid:105) = 88 ±
20 K. The lowest three rotational levels ( J = 0 , ,
2) arelikely coupled to the gas kinetic temperature. 24 – l o g [ N ( J ) / g J ] HD 93250HD163892HD 199579HD 303308
Fig. 9.— Populations of H rotational states, log[ N ( J ) /g J ], vs. their excitation energies, E ( J ) /k , expressedas temperatures. The level statistical weights are g J = (2 S + 1)(2 J + 1), where S = 0 (even- J ) and S = 1(odd- J ). We show distributions ( J = 0 −
7) for four sight lines with different excitation temperatures oflow- J and high- J states. The lowest levels ( J = 0 , ,
2) appear thermally coupled. For HD 199579 (O6.5 V)we find T = 74 K and T = 75 K, but T = 225 K and T = 307 K. For HD 163892 (O9 IV) we find T = 62 K and T = 69 K, but T = 230 K and T = 319 K. Two sight lines toward hotter stars showeven higher excitation ( T = 582 K, T = 686 K) for HD 93250 (O4 III) and ( T = 569 K, T = 1358 K)for HD 303308 (O4.5 V). 25 – T ( K ) T T ( K ) T T ( K ) T T ( K ) T Fig. 10.— Excitation temperatures, T and T , of the lowest rotational levels and T and T for higherexcited levels, versus photometric distance to the target stars. Several sight lines with missing or uncertainvalues of T , T , T , have been omitted. Horizontal lines show means and 1 σ dispersions. A significantfraction (10-15%) in the lower two panels have high values of T and T , lying above the (rms) dispersions. 26 –
100 200 300 400 500 600 700T (K)100200300400500600700 T ( K )
4 122 32 59 2 48 126 109
O2-4O5-7O8-9.5ON,WNB0-1B2-3
Fig. 11.— Relation between pairwise rotational excitation temperatures, T and T , connecting ortho (odd- J ) and para (even- J ) states. These two temperatures are correlated above 300 K, usually with T > T .Most data points lie above the dashed line of slope unity. Stars of similar SpT are color-coded as follows:dark blue (O2–O4); cornflower blue (O5–O7); cyan (O8–O9.5); dark green (ON and WN); red (B0–B1); andorange (one B3 star). Using internal ID numbers, we label stars with high excitation temperatures, plusseveral that lie off the trend-line (see Section 3.5). Star T = 1358 K and T = 569 K. 27 –Table 1. Stellar Parameters a and Distances bID Target (cid:96) b B V E ( B − V ) SpT D photb D Gaiab
Program t exp (deg) (deg) (mag) (mag) (mag) (kpc) (kpc) (FUSE ID) (ksec)1 BD 35 ◦ ◦ ◦ ◦ ◦ ◦ ∗ z 5.75 6.00[4.87,7.80] P1221104 4.15429 HD 66695 245.01 +2.21 9.77 9.78 0.27 B0.5 IV ∗ +O3 IIIf ∗
28 –Table 1—Continued
ID Target (cid:96) b B V E ( B − V ) SpT D photb D Gaiab
Program t exp (deg) (deg) (mag) (mag) (mag) (kpc) (kpc) (FUSE ID) (ksec)55 HD 99857 294.78 -4.94 7.56 7.45 0.35 B0.5 Ib <
14 kpc P1016201 3.85695 HD 167659 12.20 -1.27 7.60 7.39 0.53 O7 II-IIIf 1.87 1.58[1.40,1.80] P1028001 5.81696 HD 167771 12.70 -1.13 6.66 6.54 0.44 O7 IIInf 1.40 1.82[1.67,2.00] P1028101 3.83697 HD 167971 18.25 +1.68 8.27 7.50 1.08 O8 Iafn 1.42 1.92[1.58,2.43] P1162101 9.45098 HD 168076 16.84 +0.84 8.61 8.18 0.75 O4 IIIf 1.97 data problem P1162201 6.60199 HD 168941 5.82 -6.31 9.41 9.34 0.37 O9.5 IVp 3.72 2.32[1.96,2.84] P1016502 3.983100 HD 172140 5.28 -10.61 9.90 9.96 0.22 B0.5 III < .
29 –Table 1—Continued
ID Target (cid:96) b B V E ( B − V ) SpT D photb D Gaiab
Program t exp (deg) (deg) (mag) (mag) (mag) (kpc) (kpc) (FUSE ID) (ksec)109 HD 190429A 72.59 +2.61 7.20 7.09 0.43 O4 If 2.38 2.04[1.91,2.20] P1028401 5.390110 HD 190918 72.65 +2.07 6.88 6.75 0.45 O9.5Iab+WN4 a Updated stellar parameters, photometry, spectral types, and distances are from a survey of 139 OB-type stars (Shull & Danforth 2019) withtheir internal target ID (column 1). These parameters are based on optical and near-IR digital photometry and extinction corrections from theGalactic O-Star Spectroscopic Survey (Ma´ız Apell´aniz et al. 2004) and new spectral types (Sota et al. 2011, 2014). Columns 3 and 4 list theGalactic longitude and latitude. Later columns list the stellar photometry [
B, V, E ( B − V )], spectral types (SpT), photometric distances ( D phot ),and parallax distances ( D Gaia ) listed in Tables 1 and 2 of Shull & Danforth (2019). The last two columns give the Program ID and exposuretime of the primary
FUSE spectroscopic observations. b Photometric distances ( D phot ) are based on GOS photometry, extinctions, and SpTs and a new set of absolute magnitudes (see Shull &Danforth 2019). When GOS data were not available, we based photometric distances (shown in boldface ) on photometry and SpT in theliterature. Parallax distances ( D Gaia ) and error ranges are based on parallaxes and errors from the
Gaia -DR2 archive, after adding a constantparallax offset of 0.03 mas.
30 –Table 2. H I Column Density Measurements aID Target E B − V b log N HIc log N HI log N HI log N HI log N HIc (Star) (mag) (Adopted) (SVS85) (DS94) (J19) (Other)1 BD 35 ◦ . +0 . − . ◦ . +0 . − . ◦ . ± .
10 21 . ± .
07 21 . ± .
15 (FM90)4 CPD -59 ◦ . ± .
15 21 . ± .
07 21 . +0 . − . ◦ . ± .
11 21 . +0 . − . ◦ . ± .
087 HD 3827 0.02 20.55 20 . ± .
09 20 . +0 . − . . ± .
15 21 . ± .
109 HD 12323 0.29 21.19 21 . +0 . − .
10 HD 13268 0.44 21.34 21 . ± .
18 21 . +0 . − .
11 HD 13745 0.46 21.34 21 . ± .
10 21 . +0 . − .
12 HD 14434 0.48 21.45 21 . ± . . ± .
16 21 . +0 . − .
14 HD 15558A 0.82 21.52 21 . ± . . ± .
07 21 . ± . . ± . . ± .
07 21 . ± .
08 21 . +0 . − .
19 HD 42088 0.39 21.15 21 . ± . . ± . . ± .
10 21 . ± . . ± . . ± .
05 21 . ± . . ± . . ± . . +0 . − .
28 HD 64568 0.37 21.16 21.16 (scaled)29 HD 66695 0.27 21.12 21.12 (scaled)30 HD 66788 0.22 21.23 21 . +0 . − .
31 HD 69106 0.19 21.07 21 . ± .
06 21 . +0 . − .
32 HD 73882 0.69 21.11 21 . ± .
15 (FM90)33 HD 74194 0.50 21.23 21.23 (scaled)34 HD 74920 0.35 21.15 21 . ± . . +0 . − .
36 HD 90087 0.28 21.19 21 . ± .
06 21 . +0 . − .
37 HD 91597 0.30 21.40 21 . ± .
10 21 . ± . . ± . . +0 . − . . ± .
15 (FM90)40 HD 92554 0.39 21.34 21 . ± .
10 21 . +0 . − .
41 HD 93028 0.24 20.95 20 . ± .
15 (FM90)42 HD 93129A 0.57 21.47 21 . +0 . − .
43 HD 93146A 0.35 21.18 21 . ± . . ± .
10 21 . ± . . ± .
05 21 . ± .
10 21 . +0 . − .
46 HD 93206 0.39 21.34 21 . ± . . +0 . − . . ± .
15 (FM90)48 HD 93250 0.49 21.39 21 . ± .
10 21 . ± . . ± .
05 21 . ± .
08 21 . +0 . − .
50 HD 96670 0.46 21.28 21.28 (scaled)51 HD 96715 0.42 21.20 21 . ± .
10 21 . ± . . ± .
31 –Table 2—Continued
ID Target E B − V b log N HIc log N HI log N HI log N HI log N HIc (Star) (mag) (Adopted) (SVS85) (DS94) (J19) (Other)55 HD 99857 0.35 21.27 21 . ± .
12 21 . +0 . − .
56 HD 99890 0.24 21.12 20 . ± .
13 21 . +0 . − .
57 HD 100199 0.30 21.18 21 . +0 . − .
58 HD 100213 0.34 21.18 21 . ± . . ± . . ± .
10 21 . ± .
11 21 . +0 . − .
62 HD 101205 0.38 21.20 21 . ± . . ± . . ± . . ± . . ± .
10 21 . +0 . − .
67 HD 104705 0.26 21.15 21 . ± .
07 21 . +0 . − .
68 HD 115071 0.51 21.39 21 . ± .
10 21 . +0 . − .
69 HD 116781 0.43 21.21 21 . +0 . − .
70 HD 116852 0.22 20.96 20 . ± .
10 20 . ± .
08 20 . +0 . − .
71 HD 118571 0.26 20.98 20.98 (scaled)72 HD 124314A 0.53 21.41 21 . ± .
10 21 . +0 . − .
73 HD 124979 0.41 21.27 21 . ± .
11 21 . +0 . − .
74 HD 148422 0.35 21.24 21 . ± .
12 21 . +0 . − .
75 HD 152218 0.48 21.34 21 . ± .
10 21 . ± . . ± .
10 21 . ± . . ± .
05 21 . ± . . ± . . ± . . ± .
05 (Sno96)83 HD 156292 0.56 21.29 21.29 (scaled)84 HD 157857 0.49 21.30 21 . ± . . ± .
10 21 . ± . . ± . . ± . . ± . . +0 . − .
92 HD 166546 0.34 21.19 21.19 (scaled)93 HD 166716 0.38 21.27 21.27 (scaled)94 HD 167402 0.23 21.13 20 . ± .
13 21 . +0 . − .
95 HD 167659 0.53 21.30 21 . ± . . ± .
12 21 . ± .
15 (FM90)97 HD 167971 1.08 21.60 21 . ± .
30 (Rat02)98 HD 168076 0.75 21.65 21 . ± . . ± .
09 21 . +0 . − .
100 HD 172140 0.22 21.11 21 . ± . . ± .
05 21 . ± . . ± . . ± .
09 20 . +0 . − .
104 HD 178487 0.40 21.22 21 . ± .
10 21 . +0 . − .
105 HD 179406 0.33 21.23 21 . ± .
30 (Han92)106 HD 179407 0.33 21.20 21 . ± .
11 21 . +0 . − .
107 HD 185418 0.50 21.19 21 . +0 . − . . ± .
15 (FM90)108 HD 187459 0.44 21.31 21.31 (scaled)
32 –Table 2—Continued
ID Target E B − V b log N HIc log N HI log N HI log N HI log N HIc (Star) (mag) (Adopted) (SVS85) (DS94) (J19) (Other)109 HD 190429A 0.43 21.33 21.33 (scaled)110 HD 190918 0.45 21.40 21 . ± . . ± .
10 21 . +0 . − .
113 HD 192035 0.34 21.20 21 . ± .
19 21 . +0 . − .
114 HD 192639 0.66 21.32 21 . ± . . ± .
07 20 . ± .
09 20 . +0 . − .
116 HD 199579 0.37 21.04 21 . ± .
10 21 . ± .
11 21 . ± .
15 (FM90)117 HD 201345 0.15 21.00 20 . ± .
10 21 . +0 . − .
118 HD 201638 0.11 20.80 20.80 (scaled)119 HD 203374A 0.53 21.20 21 . +0 . − .
120 HD 206267 0.53 21.22 21 . +0 . − . . ± .
15 (Rat02)121 HD 206773 0.51 21.09 21 . +0 . − .
122 HD 207198 0.60 21.28 21 . ± .
15 21 . ± .
17 21 . +0 . − .
123 HD 207308 0.53 21.20 21 . +0 . − .
124 HD 208440 0.28 21.24 21 . ± .
07 21 . +0 . − .
125 HD 209339 0.30 21.20 21 . +0 . − .
126 HD 210809 0.34 21.31 21 . ± .
07 21 . +0 . − .
127 HD 210839 0.57 21.24 21 . ± .
10 21 . ± .
12 21 . +0 . − .
128 HD 216044 0.38 21.28 21.28 (scaled)129 HD 216532 0.85 21.38 21.38 (scaled)130 HD 216898 0.84 21.44 21.44 (scaled)131 HD 217035 0.74 21.46 21 . ± . . ± . . ± .
10 21 . ± .
13 21 . +0 . − .
134 HD 224151 0.44 21.35 21 . ± .
10 21 . +0 . − .
135 HD 224257 0.24 21.08 21.08 (scaled)136 HD 224868 0.34 21.16 21.16 (scaled)137 HD 303308 0.46 21.41 21 . ± .
10 21 . ± .
09 21 . +0 . − .
138 HD 308813 0.34 21.20 21 . ± .
10 21 . +0 . − .
139 HD 332407 0.41 21.24 21 . ± . a The adopted column densities N HI (in cm − ) in column 4 (SD20) are compared to three previous Ly α -fittingsurveys in columns 5–7: SVS85 (Shull & Van Steenberg 1985); DS94 (Diplas & Savage 1994); and J19 (Jenkins2019). Column 8 lists individual measurements from: FM90 (Fitzpatrick & Massa 1990); Rat02 (Rachford et al.2002); Han92 (Hanson et al. 1992); Sno96 (Snow et al. 1996); and scaled estimates of N HI from E ( B − V ). Footnote(c) to Table 3 provides further details. b Updated values of color excess (Shull & Danforth 2019) were derived using digital photometry, visual extinction,and new SpTs from the Galactic O-star Spectroscopic Survey (Ma´ız Apell´aniz et al. 2004; Sota et al. 2011, 2014). c The adopted source for N HI was selected in priority order of: J19 (57 stars), DS94 (51 stars), FM90 (2 stars),and three other sources (3 stars). For 26 stars that lack fits to Ly α profiles, we estimated N HI by scaling with colorexcess E ( B − V ); see Section 2.1.
33 –Table 3. Column Densities a and Doppler Parameters aID Target log N(0) log N(1) log N(2) log N(3) log N(4) log N(5) log N(6) b (km s − )1 BD 35 ◦ +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . ◦ +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . < .
15 20 +1 . − . − ◦ +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . − ◦ +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . . +0 . − . +0 . − . − ◦ +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . < .
85 13.3 +0 . − . − ◦ +0 . − . +0 . − . +0 . − . +1 . − . +1 . − . < . < .
10 . . .7 HD 3827 16.84 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . < .
85 12.8 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +3 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . < .
00 7 . +1 . − .
10 HD 13268 b +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . < .
00 . . .(blue) . . . . . . . . . . . . 14.75 +0 . − . +0 . − . . . . 15.6 +4 . . (red) . . . . . . . . . . . . 15.45 +0 . − . +0 . − . . . . 13.2 +0 . − .
11 HD 13745 b +0 . − . +0 . − . +0 . − . +0 . − . +1 . − . +0 . − . < .
00 . . .(blue) . . . . . . . . . 17.27 +0 . − . +1 . − . +1 . − . . . . 3.9 +1 . − . (red) . . . . . . . . . 16.74 +0 . − . +0 . − . +0 . − . . . . 9.6 +0 . − .
12 HD 14434 b +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . < .
00 . . .(blue) . . . . . . . . . 16.93 +0 . − . +0 . − . +0 . − . . . . . . .(red) . . . . . . . . . 16.50 +0 . − . +0 . − . +0 . − . . . . . . .13 HD 15137 19.77 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . . . . 9.5 +0 . − .
14 HD 15558A 20.50 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . < .
60 12.6 +0 . − .
15 HD 15642 b +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . < .
00 . . .(blue) . . . . . . . . . . . . 14.46 +0 . − . +0 . − . . . . . . .(red) . . . . . . . . . . . . 15.58 +0 . − . +0 . − . . . . . . .16 HD 34656 19.80 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . < .
95 10.8 +0 . − .
17 HD 39680 19.16 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . < .
15 6 . +1 . − .
18 HD 41161 19.62 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . < .
15 7.7 +0 . − .
19 HD 42088 20.18 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . < .
70 7.5 +0 . − .
20 HD 45314 20.23 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . < .
08 6 . +1 . − .
21 HD 46150 20.23 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . . +0 . − . +0 . − .
22 HD 47360 20.02 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . < .
00 6 . +1 . − .
23 HD 47417 20.15 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . < .
08 9.0 +0 . − .
24 HD 60369 20.25 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . < .
85 5.3 +0 . − .
25 HD 61347 b +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . < .
20 . . .(blue) . . . . . . . . . 15.53 +1 . − . +0 . − . +1 . − . . . . . . .(red) . . . . . . . . . 17.41 +0 . − . +0 . − . +0 . − . . . . . . .26 HD 62866 19.80 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . < .
30 17.7 +2 . − .
27 HD 63005 19.68 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . < .
01 5.4 +0 . − .
28 HD 64568 20.35 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . < .
10 15.9 +2 . − .
29 HD 66695 19.82 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . < .
00 6 . +1 . − .
30 HD 66788 19.33 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . < .
18 6.4 +1 . − .
31 HD 69106 19.41 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . < .
11 9.6 +0 . − .
32 HD 73882 21.00 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . < .
40 18.6 +1 . − .
33 HD 74194 20.14 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . < .
00 5.7 +1 . − .
34 HD 74920 19.77 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − .
35 HD 89137 19.53 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . < .
85 9.6 +0 . − .
36 HD 90087 19.53 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . . +0 . − . +1 . − .
37 HD 91597 19.45 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . < .
00 10.9 +0 . − .
38 HD 91651 18.74 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − .
39 HD 91824 19.61 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . < .
00 7 . +1 . − .
40 HD 92554 18.88 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − .
41 HD 93028 19.08 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . < .
00 9.1 +0 . − .
42 HD 93129A 19.68 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − .
43 HD 93146A 19.38 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − .
44 HD 93204 19.37 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . < .
85 9.2 +0 . − .
45 HD 93205 b +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . < .
85 . . .
34 –Table 3—Continued
ID Target log N(0) log N(1) log N(2) log N(3) log N(4) log N(5) log N(6) b (km s − )(blue) . . . . . . 14.36 +0 . − . +1 . − . +0 . − . +0 . − . . . . . . .(red) . . . . . . 16.47 +0 . − . +0 . − . +0 . − . +0 . − . . . . . . .46 HD 93206 19.18 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . < .
85 8.2 +0 . − .
47 HD 93222 19.49 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . < .
15 12.9 +0 . − .
48 HD 93250 19.79 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − .
49 HD 93843 19.14 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . < .
78 14.6 +0 . − .
50 HD 96670 20.33 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . < .
00 9.1 +0 . − .
51 HD 96715 20.45 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . < .
10 10.1 +0 . − .
52 HD 96917 c +0 . − . +0 . − . . . . . . . 16.18 +0 . − . +0 . − . +0 . − . +3 . − .
53 HD 97471 19.67 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . < .
95 10.6 +0 . − .
54 HD 97913 19.76 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . . +0 . − . +0 . − .
55 HD 99857 19.90 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . < .
04 5.7 +1 . − .
56 HD 99890 19.09 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . < .
85 9.4 +0 . − .
57 HD 100199 19.76 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . < .
08 8.4 +0 . − .
58 HD 100213 20.11 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . < .
98 7.2 +1 . − .
59 HD 100276 19.49 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . < .
10 17.0 +2 . − .
60 HD 101131 19.92 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . < .
18 7.1 +1 . − .
61 HD 101190 20.21 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . < .
95 4.9 +0 . − .
62 HD 101205 19.94 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . < .
18 6.2 +1 . − .
63 HD 101298 20.23 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . < .
95 5.4 +0 . − .
64 HD 101413 20.30 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . < .
22 4 . +1 . − .
65 HD 101436 b +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . < .
90 . . .(blue) . . . . . . . . . 15.49 +1 . − . +0 . − . . . . . . . 5.3 +1 . − . (red) . . . . . . . . . 17.50 +0 . − . +0 . − . +0 . − . . . . 3.3 +0 . − .
66 HD 103779 19.33 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . < .
88 13 +0 . − .
67 HD 104705 b +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . < .
95 . . .(blue) . . . . . . 15.85 +0 . − . +0 . − . +0 . − . . . . . . . 7.1 +1 . − . (red) . . . . . . 17.87 +0 . − . +0 . − . +0 . − . +0 . − . . . . 3.4 +0 . − .
68 HD 115071 20.30 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . < .
90 10.2 +0 . − .
69 HD 116781 19.56 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . < .
00 8.5 +0 . − .
70 HD 116852 19.50 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . < .
85 7.1 +1 . − .
71 HD 118571 19.86 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . < .
04 7.5 +1 . − .
72 HD 124314A 20.17 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . < .
64 9.1 +0 . − .
73 HD 124979 20.02 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . < .
90 10.0 +0 . − .
74 HD 148422 c +0 . − . +0 . − . . . . . . . 15.04 +0 . − . +0 . − . < .
30 . . .75 HD 152218 20.32 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +1 . − .
76 HD 152233 20.01 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +1 . − .
77 HD 152248 20.01 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +1 . − . < .
90 4.9 +0 . − .
78 HD 152314 c +0 . − . +0 . − . . . . . . . 15.70 +0 . − . +0 . − . < .
90 7.0 +1 . − .
79 HD 152623 19.88 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . . +0 . − . +0 . − .
80 HD 152723 19.97 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . . +0 . − . +0 . − .
81 HD 153426 19.91 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . < .
81 6.5 +0 . − .
82 HD 154368 c +0 . − . +0 . − . . . . . . . 15.41 +0 . − . +0 . − . . . . . . .83 HD 156292 20.41 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . < .
78 9.4 +0 . − .
84 HD 157857 20.31 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . < .
11 7.1 +1 . − .
85 HD 158661 19.89 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . < .
10 8.2 +0 . − .
86 HD 161807 19.35 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . < .
90 6.4 +1 . − .
87 HD 163758 19.53 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − .
88 HD 163892 20.38 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − .
89 HD 164816 19.69 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . < .
30 2.8 +0 . − .
90 HD 165052 19.78 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . < .
00 7.7 +0 . − .
91 HD 165246 19.83 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . < .
08 8.7 +0 . − .
92 HD 166546 b +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . . +0 . − . . . .(blue) . . . . . . 18.01 17.54 15.41 14.83 13 . +0 . − . . . .(red) . . . . . . 15.32 15.34 15.16 14.90 13 . +0 . − . . . .
35 –Table 3—Continued
ID Target log N(0) log N(1) log N(2) log N(3) log N(4) log N(5) log N(6) b (km s − )93 HD 166716 c +0 . − . +0 . − . . . . . . . 15.65 +0 . − . +0 . − . < .
10 . . .94 HD 167402 19.90 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . < .
00 15.7 +2 . − .
95 HD 167659 b +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . . . .(blue) . . . . . . 15.73 +0 . − . +0 . − . +0 . − . +0 . − . . . . 8.4 +0 . − . (red) . . . . . . 17.58 +0 . − . +0 . − . +0 . − . +0 . − . . . . 7.6 +0 . − .
96 HD 167771 20.39 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . < .
96 6.1 +0 . − .
97 HD 167971 c +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . < .
31 . . . 15.0 +2 . − .
98 HD 168076 c +0 . − . +0 . − . . . . . . . 16.50 +0 . − . +0 . − . < .
40 . . .99 HD 168941 19.82 +0 . − . +0 . − . +0 . − . +0 . − . +1 . − . +0 . − . < .
08 4.0 +1 . − .
100 HD 172140 18.97 +0 . − . +0 . − . +1 . − . +1 . − . +0 . − . +0 . − . < .
90 5.1 +1 . − .
101 HD 175754 19.06 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . < .
95 8.4 +0 . − .
102 HD 175876 b +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . < .
06 . . .(blue) . . . . . . 16.03 16.15 14.73 14.38 < .
75 . . .(red) . . . . . . 16.71 16.16 14.69 14.16 < .
75 . . .103 HD 177989 19.93 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . < .
85 4.4 +0 . − .
104 HD 178487 20.18 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . < .
06 7.2 +1 . − .
105 HD 179406 20.41 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . < .
10 5.5 +1 . − .
106 HD 179407 19.88 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . < .
00 5.5 +1 . − .
107 HD 185418 20.23 +0 . − . +0 . − . +0 . − . +0 . − . +1 . − . +0 . − . < .
90 4.1 +0 . − .
108 HD 187459 20.12 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . < .
78 9.7 +0 . − .
109 HD 190429A 19.89 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − .
110 HD 190918 19.31 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +2 . − .
111 HD 191495 19.68 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − .
112 HD 191877 19.73 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . . +0 . − . +0 . − .
113 HD 192035 20.37 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . < .
08 6.8 +1 . − .
114 HD 192639 20.27 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +1 . − .
115 HD 195965 19.90 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . < .
90 6.4 +0 . − .
116 HD 199579 20.23 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − .
117 HD 201345 18.83 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . < .
60 7.4 +0 . − .
118 HD 201638 17.67 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . < . < .
30 6.5 +0 . − .
119 HD 203374A 20.32 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . < .
00 8.0 +0 . − .
120 HD 206267 20.65 +0 . − . +0 . − . +0 . − . +0 . − . +1 . − . +1 . − . +0 . − . +1 . − .
121 HD 206773 20.05 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − .
122 HD 207198 20.61 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +1 . − .
123 HD 207308 20.52 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . < .
30 5.2 +1 . − .
124 HD 208440 19.94 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . < .
30 8.1 +0 . − .
125 HD 209339 19.80 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . < .
10 6.8 +1 . − .
126 HD 210809 b +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +1 . − . (blue) . . . . . . . . . . . . 14.72 +0 . − . +0 . − . . . . . . .(red) . . . . . . . . . . . . 15.50 +0 . − . +0 . − . . . . . . .127 HD 210839 20.51 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . . +1 . − .
128 HD 216044 19.75 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . . +0 . − . +0 . − .
129 HD 216532 c +0 . − . +0 . − . . . . . . . 15.78 +0 . − . +0 . − . +0 . − . . . .130 HD 216898 20.66 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − .
131 HD 217305 20.65 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − .
132 HD 217312 20.43 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − .
133 HD 218915 19.78 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . . +0 . − . +0 . − .
134 HD 224151 b +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . . +0 . − . . . .(blue) . . . . . . 15.69 +0 . − . +0 . − . +0 . − . +0 . − . . +0 . − . . . .(red) . . . . . . 18.15 +0 . − . +0 . − . +0 . − . +0 . − . . +0 . − . . . .135 HD 224257 b +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . < .
20 . . .(blue) . . . . . . 16.12 +0 . − . +0 . − . +0 . − . +0 . − . < .
90 6.4 +1 . − . (red) . . . . . . 15.53 +0 . − . +0 . − . +0 . − . +0 . − . < .
90 19.5 +0 . − .
136 HD 224868 20.02 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . < .
95 10.0 +0 . − .
137 HD 303308 b +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . . . .
36 –Table 3—Continued
ID Target log N(0) log N(1) log N(2) log N(3) log N(4) log N(5) log N(6) b (km s − )(blue) . . . . . . . . . . . . 16.39 +0 . − . +0 . − . +0 . − . +0 . − . (red) . . . . . . . . . . . . 15.86 +0 . − . +0 . − . +0 . − . +1 . − .
138 HD 308813 19.97 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . < .
95 7.8 +0 . − .
139 HD 332407 20.03 +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . +0 . − . Column densities, N ( J ) (in cm − ), in rotational states J derived from curve-of-growth fitting with Doppler parameter ( b in lastcolumn) and damping wings for lines from J = 0 and J = 1. No value is listed when b was poorly determined. Error bars on low- J states ( J = 0 , J = 2 ,
3) depend on data quality (see footnote in Table 3). In damping-wing fits, errors on log N typically range from ± .
03 to ± .
10. Six targets had detectable column densities in J = 7: N (7) = 15 . ± .
15, N (7) = 14 . ± .
05, N (7) = 14 . ± .
06, N (7) = 14 . ± .
11, N (7) = 14 . ± . N (7) = 15 . ± .
15 (15.32 in blue component and 14.80 in red component). b In 15 sight lines, multiple velocity components were observable and measured in high- J states. We report individual columndensities, denoted as “blue and red” components, and sum them to find total column densities. c A few targets have no listed column densities N (2) or N (3). In several cases this was a result of poor data quality, so that linesfrom J ≥ J = 2 and J = 3 were highly saturated,with equivalent widths of 200-300 m˚A and large effective Doppler parameters ( b ≥
10 km s − ) likely produced by unresolved velocitycomponents. When equivalent widths of all the accessible lines lie on the “flat portion” of the curve of growth, errors can exceed ± .
40 for log N = 16-17 and ± .
50 for log N = 17-19.
37 –Table 4. Column Densities a and Rotational Temperatures bID Target log N H2 T T T T log N HIc log N Hc D photd (cid:104) n H (cid:105) d f H2d
Notes a (K) (K) (K) (K) (kpc) (cm − )1 BD 35 ◦ ◦ − ◦ − ◦ − ◦ − ◦
38 –Table 4—Continued
ID Target log N H2 T T T T log N HIc log N Hc D photd (cid:104) n H (cid:105) d f H2d
Notes a (K) (K) (K) (K) (kpc) (cm − )55 HD 99857 20.26 88 92 150 184 21.27 21.35 3.04 0.24 0.163 256 HD 99890 19.47 91 75 244 282 21.12 21.14 3.22 0.14 0.043 257 HD 100199 20.17 98 65 198 253 21.18 21.26 2.77 0.21 0.163 258 HD 100213 20.42 79 86 162 205 21.18 21.33 2.15 0.32 0.285 259 HD 100276 19.83 85 59 313 554 21.19 21.23 2.94 0.19 0.080 260 HD 101131 20.28 87 88 164 196 21.15* 21.25 1.78 0.32 0.212 161 HD 101190 20.42 64 86 149 181 21.15 21.35 2.09 0.35 0.232 162 HD 101205 20.26 79 103 155 181 21.24 21.29 1.64 0.39 0.187 363 HD 101298 20.45 65 56 259 180 21.26 21.28 2.58 0.24 0.237 264 HD 101413 20.51 63 98 143 200 21.23 21.37 2.13 0.36 0.276 265 HD 101436 20.54 61 88 155 197 21.23 21.28 1.85 0.33 0.290 166 HD 103779 19.82 117 59 300 507 21.17 21.21 3.99 0.13 0.082 267 HD 104705 19.99 98 92 175 214 21.15 21.21 4.18 0.13 0.122 168 HD 115071 20.64 84 82 254 304 21.39 21.42 1.87 0.46 0.262 269 HD 116781 20.06 119 94 172 206 21.21 21.27 1.79 0.34 0.124 270 HD 116852 19.78 74 53 317 322 20.96 21.01 4.88 0.068 0.117 171 HD 118571 20.44 94 86 172 193 20.98* 21.18 2.70 0.18 0.366 172 HD 124314A 20.47 77 78 185 265 21.41 21.50 1.25 0.82 0.187 173 HD 124979 20.41 93 88 172 230 21.27 21.30 3.09 0.21 0.258 274 HD 148422 20.13 63 . . . . . . . . . 21.24 21.30 8.26 0.078 0.134 375 HD 152218 20.57 69 82 185 253 21.34 21.47 1.42 0.67 0.254 276 HD 152233 20.29 73 92 311 222 21.29 21.37 1.52 0.50 0.167 277 HD 152248 20.29 73 95 285 217 21.27* 21.44 1.61 0.55 0.143 278 HD 152314 20.51 68 . . . . . . . . . 21.35* 21.46 1.44 0.65 0.224 279 HD 152623 20.21 82 55 325 381 21.28 21.35 1.04 0.70 0.145 280 HD 152723 20.29 79 87 176 203 21.43 21.49 1.94 0.52 0.127 281 HD 153426 20.30 93 77 195 264 21.34 21.41 1.83 0.46 0.154 282 HD 154368 21.17 55 . . . . . . . . . 21.00 21.60 1.08 1.20 0.747 383 HD 156292 20.81 94 72 180 345 21.29* 21.51 1.51 0.69 0.398 284 HD 157857 20.62 78 88 150 207 21.30 21.45 2.56 0.36 0.295 285 HD 158661 20.17 73 99 192 265 21.28* 21.34 4.08 0.17 0.134 286 HD 161807 19.86 121 95 171 213 21.08* 21.13 1.94 0.23 0.108 187 HD 163758 19.86 80 96 215 219 21.23 21.27 4.11 0.15 0.079 288 HD 163892 20.58 62 69 230 319 21.32 21.46 1.37 0.68 0.267 289 HD 164816 20.00 78 78 177 239 21.18 21.23 1.08 0.51 0.117 290 HD 165052 20.20 99 83 187 213 21.36 21.42 1.37 0.62 0.122 191 HD 165246 20.16 83 45 429 503 21.41 21.46 1.38 0.68 0.101 292 HD 166546 20.32 83 82 191 252 21.19* 21.29 1.46 0.43 0.212 293 HD 166716 20.23 74 . . . . . . . . . 21.27* 21.34 2.73 0.26 0.154 294 HD 167402 20.13 66 45 396 519 21.13 21.21 7.61 0.069 0.167 295 HD 167659 20.55 99 68 277 257 21.30 21.43 1.87 0.47 0.262 296 HD 167771 20.67 72 109 134 190 21.08 21.33 1.40 0.49 0.438 197 HD 167971 20.87 68 69 152 . . . 21.60 21.74 1.42 1.25 0.271 398 HD 168076 20.65 63 . . . . . . . . . 21.65 21.73 1.97 0.88 0.167 399 HD 168941 20.10 74 83 211 227 21.18 21.25 3.72 0.15 0.143 2100 HD 172140 19.25 74 57 360 370 21.11 21.12 6.54 0.065 0.027 2101 HD 175754 19.54 113 79 221 262 21.04 21.07 2.55 0.15 0.059 2102 HD 175876 19.47 117 77 250 311 21.04 21.06 2.57 0.14 0.051 1103 HD 177989 20.12 61 74 203 225 20.99 21.09 5.14 0.078 0.202 2104 HD 178487 20.47 76 81 169 197 21.22 21.35 5.22 0.14 0.262 2105 HD 179406 20.62 64 73 164 230 21.23 21.40 0.21 3.88 0.329 2106 HD 179407 20.22 83 95 234 193 21.20 21.28 7.72 0.080 0.173 2107 HD 185418 20.71 113 87 166 209 21.19 21.41 0.78 1.07 0.398 2108 HD 187459 20.39 73 68 212 250 21.31* 21.40 1.68 0.48 0.194 2
39 –Table 4—Continued
ID Target log N H2 T T T T log N HIc log N Hc D photd (cid:104) n H (cid:105) d f H2d
Notes a (K) (K) (K) (K) (kpc) (cm − )109 HD 190429A 20.22 83 53 473 364 21.33* 21.39 2.38 0.33 0.134 2110 HD 190918 19.80 117 64 357 379 21.40 21.42 2.39 0.36 0.048 2111 HD 191495 20.07 94 62 343 356 21.32* 21.37 1.69 0.45 0.101 2112 HD 191877 20.02 75 76 202 258 21.03 21.11 2.07 0.20 0.163 2113 HD 192035 20.66 76 84 156 292 21.20 21.40 2.29 0.36 0.366 2114 HD 192639 20.65 91 82 176 326 21.32 21.47 2.14 0.45 0.300 2115 HD 195965 20.37 110 88 175 219 20.92 21.11 1.03 0.41 0.360 1116 HD 199579 20.51 74 75 225 307 21.04 21.24 0.92 0.61 0.371 2117 HD 201345 19.24 97 84 237 306 21.00 21.01 2.50 0.13 0.034 2118 HD 201638 18.23 136 97 249 . . . 20.80* 20.80 8.98 0.023 0.0054 2119 HD 203374A 20.68 87 76 163 237 21.20 21.41 0.78 1.07 0.377 2120 HD 206267 20.89 68 70 204 268 21.22 21.51 0.73 1.44 0.483 2121 HD 206773 20.45 94 84 175 227 21.09 21.24 0.69 0.82 0.314 1122 HD 207198 20.83 65 77 251 564 21.28 21.51 1.04 1.01 0.415 2123 HD 207308 20.76 68 75 160 263 21.20 21.44 0.76 1.17 0.421 2124 HD 208440 20.27 83 76 195 254 21.24 21.32 1.04 0.65 0.176 1125 HD 209339 20.19 94 67 222 332 21.20 21.28 1.08 0.57 0.163 1126 HD 210809 19.95 98 57 482 413 21.31 21.35 3.88 0.18 0.080 1127 HD 210839 20.78 73 70 201 370 21.24 21.47 0.87 1.21 0.409 2128 HD 216044 20.20 105 79 216 269 21.28* 21.35 2.57 0.28 0.123 2129 HD 216532 21.10 70 . . . . . . . . . 21.38* 21.69 0.94 1.69 0.512 3130 HD 216898 21.03 89 86 170 273 21.44* 21.65 0.91 1.59 0.438 2131 HD 217035 20.94 74 92 170 269 21.46 21.67 0.72 2.10 0.377 2132 HD 217312 20.79 88 82 183 296 21.48 21.63 0.60 2.30 0.290 2133 HD 218915 20.17 93 88 191 229 21.20 21.27 3.62 0.17 0.157 2134 HD 224151 20.53 98 83 185 222 21.35 21.46 0.91 1.03 0.232 2135 HD 224257 19.96 127 57 332 473 21.08* 21.14 2.05 0.22 0.132 2136 HD 224868 20.41 93 61 228 304 21.16* 21.29 3.03 0.21 0.262 2137 HD 303308 20.23 131 68 569 1358 21.41 21.46 2.66 0.35 0.117 2138 HD 308813 20.29 80 85 157 206 21.20 21.30 3.45 0.19 0.197 2139 HD 332407 20.38 86 56 317 499 21.24 21.35 2.56 0.28 0.216 2 a All column densities, N H2 , N HI , and N H = N HI + 2 N H2 are in cm − , with errors on log N HI discussed in Table 2 andSection 3.2. Values of N (0) and N (1) come from fitting damping-wing Voigt profiles of R(0), R(1), and P(1) lines. Theirerrors depend on data quality. Notes in last column refer to three levels of S/N ratio: (1) S/N ≥
15 with errors of ± . − . N ; (2) 5 ≤ S/N ≤
15 with errors of ± . − .
10; (3)
S/N ≤
5, with errors of ± . − . b The rotational excitation temperature T is determined from the ratio N /N of column densities of the J = 1 and J = 0 rotational states. Similar excitation temperatures T , T and T follow from ratios N /N , N /N , and N /N ofpopulations in levels J = 0 , , ) and J = 3 , ) respectively, as described in Sections 3.4 and 3.5. c Neutral hydrogen column densities N HI are taken from previous surveys by Diplas & Savage (1994), Jenkins (2019),Fitzpatrick & Massa (1990), and several individual papers. Details are discussed in Section 2.1 and Table 2. For 26 sightlines with no Ly α fits, labeled with asterisks after log N HI , we use the relation, N H = (5 . × cm − mag − ) E ( B − V )from the Copernicus survey (Savage et al. 1977), with N HI = N H − N H2 . Updated values of color excess E ( B − V ) for all139 target stars are listed in Table 1, taken from Shull & Danforth (2019). d Mean total hydrogen density along each sight line, (cid:104) n H (cid:105) = N H /D phot , is estimated from stellar photometric distances D phot in Table 1. The molecular fraction is defined by f H2 = 2 N H2 / [ N HI + 2 N H2 ].
40 –Table 5. Observational Averages from
FUSE