Draft version October 28, 2018

Preprint typeset using LATEX style emulateapj v. 04/03/99

FUSE OBSERVATIONS OF MOLECULAR HYDROGEN IN TRANSLUCENT INTERSTELLAR

CLOUDS: THE LINE OF SIGHT TOWARD HD 73882
T. P. Snow1 , B. L. Rachford1 , J. Tumlinson1 , J. M. Shull1 , D. E. Welty2 , W. P. Blair3 ,
R. Ferlet4 , S. D. Friedman3 , C. Gry5,6 , E. B. Jenkins7 , A. Lecavelier4 , M. Lemoine8 , D.
C. Morton9 , B. D. Savage10 , K. R. Sembach3 , A. Vidal-Madjar4 , D. G. York2 , B.-G.
Andersson3 , P. D. Feldman3 , and H. W. Moos3
Draft version October 28, 2018

ABSTRACT
arXiv:astro-ph/0005090v1 4 May 2000

We report the results of initial FUSE observations of molecular hydrogen (H2 ) in translucent clouds.
These clouds have greater optical depth than any of the diffuse clouds previously observed for far-UV H2
absorption, and provide new insights into the physics and chemistry of such regions. Our initial results
involve observations of HD 73882, a well-studied southern hemisphere star lying behind substantial
interstellar material (EB−V = 0.72; AV = 2.44). We find a total H2 column density, N(H2 ) = 1.2 × 1021
cm−2 , about three times larger than the values for diffuse clouds previously measured in the far-UV.
The gas kinetic temperature indicated by the ratio N(J=1)/N(J=0) is 58 ± 10 K. With the aid of
ground-based data to calculate an appropriate multi-component curve of growth, we have determined
column densities for all rotational levels up to J = 7. The J ≥ 2 states can be reasonably fitted with
a rotational excitation temperature of 307 ± 23 K. Both the kinetic and rotational temperatures are
similar to those found in previous investigations of diffuse clouds. The ratios of carbonaceous molecules
to hydrogen molecules are also similar to ratios in diffuse clouds, suggesting a similar chemistry for this
line of sight.
Subject headings: ISM: abundances — ISM: clouds — ISM: lines and bands — ISM: molecules — stars:
individual (HD 73882) — ultraviolet: ISM

1. INTRODUCTION difficulty in the IR, where the weak quadrupole transition

Molecular hydrogen is the most abundant molecular can be observed in absorption only for very high column
densities (Lacy et al. 1994).
species in the Galactic interstellar medium. While H2
is an important constituent of diffuse interstellar clouds Previous instruments capable of resolving the far-UV
with visual extinctions AV less than about 1 magnitude, absorption bands of H2 include the Copernicus mission
(Savage et al. 1977; Spitzer & Jenkins 1975), the IMAPS
it becomes the dominant form of hydrogen in the so-called
“translucent” clouds, which are characterized by AV in experiment (Jenkins & Peimbert 1997), and the ORFEUS
the range 1–5 mag. H2 dominates the dynamics, chem- spectrometer (Richter et al. 1998; de Boer et al. 1998). All
of those, however, were limited to relatively bright stars
istry, and physics of denser clouds, and is the basic raw
material for star formation. with AV < ∼ 1.
Due to the homonuclear structure of H2 and its lack The FUSE observatory (Moos et al. 2000) is well suited
of a dipole moment, ro-vibrational transitions within the for observations of cold H2 in the diffuse and translucent
electronic ground state are quadrupolar with low sponta- interstellar medium, owing to its high throughput in the
neous emission coefficients. They are therefore very diffi- wavelength region from 905 Å to 1187 Å, encompassing
cult to observe except where special excitation conditions the strong Werner and Lyman bands of H2 . A survey of
produce detectable ro-vibrational emission in the near-IR H2 in these regions was therefore identified by the FUSE
(e.g. Gautier et al. 1976) or from a space-borne observa- PI team as a project of high priority.
tory such as ISO, when a long enough line of sight (AV ≈ We have assembled a list of 35 lines of sight, which sam-
20) allows detection of the mid-IR pure rotational emis- ple a wide variety of interstellar environments. These lines
sion from warm diffuse gas (Falgarone et al. 2000). Cold of sight, with 0.3 mag <
∼ E(B−V ) < ∼ 1.1 mag, are also char-
H2 , the dominant form, can be widely observed through acterized by a range in extinction properties (determined
electronic transitions in the far UV (λ < 1150 Å), or with both from RV and from the shape of the far-UV extinc-
1 CASA, University of Colorado, Campus Box 389, Boulder, CO 80309
2 Astronomy and Astrophysics Center, University of Chicago, 5640 S. Ellis Ave., Chicago, IL 60637
3 Department of Physics and Astronomy, The Johns Hopkins University, 3400 N. Charles St., Baltimore, MD 21218
4 Institut d’Astrophysique de Paris, CNRS, 98bis, Blvd. Arago, Paris, F-75014, France
5 Laboratoire d’Astronomie Spatiale, B.P. 8, 13376 Marseille cedex 12, France
6 ISO Data Center, ESA Astrophysics Division, PO Box 28080 Madrid, Spain
7 Princeton University Observatory, Princeton, NJ 08544
8 DARC, UMR 8629 CNRS, Observatorie de Paris, F-92195 Meudon, France
9 Herzberg Institute of Astrophysics, National Research Council, 5071 West Saanich Road, Victoria, BC V8X 4M6, Canada
10 Department of Astronomy, University of Wisconsin, 475 N. Charter St., Madison, WI 53706

1
2

tion curve). Most of the lines of sight have information on (R ∼ 150,000–250,000) spectra of Na I, K I, and Ca II
molecular abundances from previous optical and/or mm- absorption toward HD 73882. A profile-fitting analysis of
wave data. the K I and Na I lines reveals no fewer than 20 velocity
In conjunction with the FUSE observing program, we components, spread over 55 km s−1 . However, the ab-
and several collaborators are obtaining additional data us- sorption is dominated by three narrow (b ∼ 0.7 km s−1 ),
ing various ground-based telescopes. We have obtained blended components at vLSR ≈ +2.7, +4.3, and +5.9 km
very high-resolution (R ∼ 150,000–250,000) spectra of in- s−1 , with significant contributions from two others at +7.7
terstellar K I, Na I, Ca II, and CH absorption along most of and +11.0 km s−1 . The blend centered at +4.3 km s−1
the lines of sight to understand the cloud velocity structure probably corresponds to the single (unresolved at R ∼
(Welty, Morton, & Snow 2000). We (led by DGY) are ob- 60,000—100,000) absorption component seen in CH, CN,
taining moderately high-resolution (R ∼ 50,000), very high CH+ , and C2 , and to the CO emission component at +5.9
S/N, nearly complete optical spectra to measure many of km s−1 . The two slightly weaker atomic components may
the diffuse interstellar bands, and the molecules CH, CH+ , correspond to the two other CO components at +8.9 and
CN, C2 , and C3 . T. Oka and B. J. McCall are pursuing +11.2 km s−1 . (Note that the CO column density quoted
near-IR measurements of H3 + , and we and F. Chaffee are in Table 1 is only for the +5.9 km s−1 component.) The
conducting near-IR observations to compare grain mantle numerous weaker features seen in the optical data may
features such as water ice, the 3.4-µm hydrocarbon band, arise in outlying less dense clouds.
and the silicate feature at 9.6 µm, with the UV data on
dust extinction and gas-phase depletions. The ultimate 3. FUSE OBSERVATIONS AND DATA ANALYSIS
goals of this program are to fully understand the masses Our FUSE spectrum of HD 73882 is derived from time-
of dense interstellar clouds, to probe the chemistry and tagged observations over the course of 8 orbits on 1999 Oct
physics of translucent clouds, and to probe the transition 30. Several “burst” events occurred during the observation
region between diffuse and dense interstellar clouds. (Sahnow et al. 2000). We excluded all photon events that
In this Letter, we describe the first observations of H2 occurred during the bursts, reducing effective on-target
in a translucent cloud line of sight, toward the star HD integration time from 16.8 ksec to 16.1 ksec. Strong in-
73882. The following sections describe the properties of terstellar extinction and lack of co-alignment of the SiC
the line of sight deduced from previous observations (§2), channels with the LiF channels prevented the collection of
the FUSE observations and our analysis of the data (§3), useful data shortward of 1010 Å.
and the insights gained from these new data (§4). We performed a simple “collapse” of the 2-dimensional
2. THE LINE OF SIGHT TOWARD HD 73882
spectral image to a 1-dimensional spectrum, summing 100
pixels (LiF 1A segment) or 120 pixels (LiF 2A) in each col-
The star HD 73882 is well known in the translucent umn, and we determined the background level in a similar
cloud community because it is of early spectral type way. This extraction compares favorably to a subsequent
(O8.5V), it is bright enough (V = 7.27) to allow high-S/N pipeline processed spectrum. In the regions of interest,
observations at optical wavelengths, and it has sufficient we obtained a maximum S/N of 20 per 15-pixel resolution
foreground dust (EB−V = 0.72, AV = 2.44) and gas to element in the LiF 1A segment, and 30 in LiF 2A. The
have large column densities of many atomic and molecular wavelength solution was derived from narrow lines in the
species. Fitzpatrick & Massa (1986, 1988, 1990) derived diffuse cloud line of sight to LMC star Sk−67 111 (Shull
the UV extinction curve and the column density of atomic et al. 2000). Some residual scatter remained in the wave-
hydrogen from low-resolution IUE spectra. The extinction length solution (∆vrms ≈ 6 km s−1 ), precluding a detailed
curve shows a steep far-UV rise with significant curvature, analysis of line velocities in this study. Figure 1 depicts a
generally similar to other “dense cloud” curves, as char- portion of our spectrum, along with our model of the H2
acterized by Cardelli, Clayton, & Mathis (1988) – a con- lines described in the following paragraph.
clusion reached previously for this star by Massa, Savage, The column densities for the J = 0 and J = 1 levels
& Fitzpatrick (1983). However, both the ratio of total to come from direct fits to the profiles of the corresponding
selective extinction (RV = 3.9) and the wavelength of max- heavily damped lines. The resolution of the FUSE spec-
imum polarization (Serkowski, Mathewson, & Ford 1975) tra (R ∼ 12,000) is not sufficient to separate the narrow,
suggest that small grains are not as important in this case closely spaced components contributing to the interstellar
as in most other molecular-cloud lines of sight. The molec- line profiles of the high-J lines. In view of the generally
ular column densities listed in Table 1 support the view good correlation between the column densities of Na I and
that this line of sight is dominated by one or more dense H2 (Federman 1981; Welty & Hobbs 2000) for N(Na I) > ∼
clouds, consistent with translucent cloud models (e.g. van 1012 cm−2 , we have used the five strongest Na I compo-
Dishoeck & Black 1986). nents described in §2 to model the H2 lines. The velocity
A series of papers have provided ground-based data on structure for Na I was preserved for H2 , and the observed
CO, CH, CN, CH+ , and C2 (van Dishoeck et al. 1991; Gre- b-values for the Na I components were scaled to H2 by the
del et al. 1993; see Table 1). The CO emission line pro- relation
files reveal three distinct components toward HD 73882,
r
2kT 2
at LSR velocities of +5.9, +8.9, and +11.2 km s−1 . The b= + vturb .
absorption from the other molecular species, however, is m
seen only in a single component near +4.5 km s−1 . Us- Given the relatively narrow range of column densities, we
ing the 3.6-m reflector and coudé echelle spectrometer at have assumed a one-to-one scaling between Na I and H2 .
ESO and the 0.9-m coudé feed telescope and coudé spec- The predicted H2 component structure is given in Table 2.
trograph at KPNO, we have obtained very high-resolution From this structure we calculated a curve of growth and
 3

determined the column densities for J ≥ 2 using 29 H2 J ≥ 2 in this line of sight are not in thermal equilibrium at
lines. This curve of growth (similar to a single component Tkin , but instead follow a different distribution character-
with b = 3 km s−1 ) is given in Figure 2, while the column ized by an excitation temperature, Tex = 307 ± 23 K. Non-
densities are given in Table 3. The data in Tables 2 and 3 thermal excitation of the high-J levels has been explained
were used to calculate the model of the H2 spectrum given in terms of UV pumping, in which the molecules cascade
in Figure 1. down through upper rotational and vibrational lines fol-
Lying on or near the saturation portion of the curve of lowing the absorption of a UV photon (Black & Delgarno
growth, the column densities for J ≥ 4 are highly sensi- 1973). In this case, the excited H2 may arise in the outer,
tive to the assumed component structure. If more than optically thin regions of the cloud, rather than in the self-
five Na I components trace significant amounts of H2 , the shielded interior. If so, the high-J lines may exhibit small
larger “effective” b-value yields column densities for J = velocity shifts or enhanced b-values. However, the limited
4 and J = 5 decreased by an order of magnitude or more. spectral resolution and uncertain wavelength scale of our
Similarly, if the H2 structure is less complex than we have FUSE data precludes a sensitive search for these effects.
assumed or if components 4 and 5 trace less H2 (as sug- It is noteworthy that even in a line of sight thought to
gested by the CH data of Gredel et al. 1993), the opposite be dominated by molecular cloud material within a single
situation occurs. The J = 5 lines would be shifted ∼0.5 cloud complex, the velocity structure is very complicated.
dex to the damping portion of the curve of growth, and This structure, along with the similarities between the cur-
the column densities for the J = 6 and J = 7 lines could rent results and previous results for diffuse clouds, suggests
be increased by ∼0.5 dex or more. the possibility that we looking through several Zeta Oph-
type clouds rather than a single very dense cloud.
4. DISCUSSION Our analysis of the excited rotational levels of H2 de-
The total H2 column density of 1.2 × 1021 cm−2 we pends heavily on high-resolution ground-based optical ab-
have measured toward HD 73882 is statistically identical sorption and mm-wave emission line data to which we had
to the largest previously observed via far-UV absorption access. It is thus imperative for future FUSE interstellar-
spectroscopy, 1.1 × 1021 cm−2 toward HD 24534 (Mason, line observations to obtain appropriate high-resolution
et al. 1976). The molecular fraction f = 2N(H2 )/[2N(H2 ) spectra in order to have any hope of unambiguous inter-
+ N(H I)] is f = 0.65, nearly identical to that found to- pretation of the lower resolution data.
ward Zeta Oph (f = 0.63; Bohlin, et al. 1978). The only
known larger value, f = 0.80, toward HD 24534 (Snow et We thank the referee, J. H. Black, for helpful comments.
al. 1998), will be redetermined via a planned FUSE obser- This work is based on data obtained for the Guaranteed
vation. Time Team by the NASA-CNES-CSA FUSE mission oper-
The abundances of CO, CN, CH, and C2 , relative to ated by the Johns Hopkins University. Financial support
the total hydrogen abundance, are more similar to those to U.S. participants has been provided by NASA contract
found in diffuse clouds (i.e. toward Zeta Oph) than in NAS5-32985.
dark clouds. Because these molecular abundances repre-
sent only a small fraction of the presumed total carbon
abundance, we conclude that the clouds toward HD 73882
have not reached the predicted transition point where car-
bon becomes primarily molecular (e.g. van Dishoeck &
Black 1986). The limited data presently available also sug-
gest that the depletions toward HD 73882 are similar to
those found in cold diffuse clouds, an issue we will explore
in more detail in a future paper (Rachford et al. 2000).
In principle, the CO abundance can be derived from our
FUSE spectrum. However, the best candidate lines, from
the C–X (0,0) band at 1088 Å, lie on the flat portion of
the curve of growth where the derived abundance is highly
sensitive to the chosen b-value.
Our derived H2 column density, combined with the line-
of-sight extinction properties cited above, show that the
ratio of hydrogen, both molecular and total, to dust ex-
tinction, is also similar to the value found for diffuse clouds
(Bohlin et al. 1978).
The ratio of molecules in J = 1 to J = 0 (ortho- to
para-hydrogen) is usually interpreted as a measure of the
gas kinetic temperature, because in these clouds the col-
lisional timescale for depopulating these levels is shorter
than the radiative timescale. Our N(1)/N(0) ratio yields
Tkin = 58 ± 10 K, consistent with the average value
Tkin = 77 ± 17 K found for diffuse clouds with Coper-
nicus data (Savage et al. 1977). Shull et al. (2000) find a
somewhat larger value in a small sample of FUSE data.
The excitation diagram in Figure 3 shows that the levels
4

REFERENCES

Bohlin, R. C., Savage, B. D., & Drake, J. F. 1978, ApJ, 224, 132 Massa, D., Savage, B. D., & Fitzpatrick, E. L. 1983, ApJ, 266, 662
Black, J. H. & Dalgarno, A. 1973, ApJ, L101 Mason, K. O., White, N. E., Sanford, P. W., Hawkins, F. J., Drake,
Cardelli, J. A., Clayton, G. C., & Mathis, J. S. 1988, ApJ, 329, L33 J. F., & York, D. G. 1976, MNRAS, 176, 193
de Boer, K. S., Richter, P., Bomans, D. J., Heithausen, A., Moos, H. W., et al. 2000, ApJ, this issue
Koornneef, J. 1998, A&A, 338, L5 Rachford, B. L., Snow, T. P., et al. 2000, in preparation
Falgarone, E., Verstraete, L., Pineau des Forêts, G., Flower, D., & Richter, P., et al. 1998, A&A, 338, L9
Puget, J.-L. 2000, in H2 in Space, Eds.: F. Combes, G. Pineau Savage, B. D., Drake, J. F., Budich, W., & Bohlin, R. C. 1977, ApJ,
des Forêts (Cambridge University Press, Astrophysics Series), in 216, 291
press Sahnow, D. et al. 2000, ApJ, this issue
Federman, S. R. 1981, A&A, 96, 198 Serkowski, K., Mathewson, D. L., & Ford, V. L. 1975, ApJ, 196, 261
Ferlet, R., et al. 2000, ApJ, this issue Shull, J. M. et al. 2000, ApJ, this issue
Fitzpatrick, E. L. & Massa, D. 1986, ApJ, 307, 286 Snow, T. P., Hanson, M. M., Black, J. H., van Dishoeck, E. F.,
Fitzpatrick, E. L. & Massa, D. 1988, ApJ, 328, 734 Crutcher, R. C., & Lutz, B. L. 1998, ApJ, 496, L113
Fitzpatrick, E. L. & Massa, D. 1990, ApJS, 72, 163 Spitzer, L. & Jenkins, E. B. 1975, ARA&A, 13, 133
Gautier, T. N., III, Fink, U., Larson, H. P., & Treffers, R. R., 1976, van Dishoeck, E. F., Phillips, T. G., Gredel, R., & Black, J. H. 1991,
ApJ,207, L129 ApJ, 366, 141
Gredel, R., van Dishoeck, E. F., & Black, J. H. 1993, A&A, 269, 477 van Dishoeck, E. F. & Black, J. H. 1986, ApJS, 62, 109
Jenkins, E. B. & Peimbert, A. 1997, ApJ, 477, 265 Welty, D. E., & Hobbs, L. M. 2000, in preparation
Lacy, J. H., Knacke, R., Geballe, T. R., & Tokunaga, A. T. 1994 Welty, D. E., Morton, D. C., & Snow, T. P. 2000, in preparation
ApJL, 428, L69
 5

Table 1
Ancillary Abundance Data for HD 73882

Species N Reference
(cm−2 )
HI 1.3 × 1021 Fitzpatrick & Massa 1990
KI 2.4 × 1011 Welty et al. 2000
Na I 2.6 × 1013 Welty et al. 2000
Ca II 4.2 × 1012 Welty et al. 2000
CH 3.7 × 1013 Gredel et al. 1993
CH+ 2.4 × 1013 Gredel et al. 1993
C2 6 × 1013 Gredel et al. 1993
CN 3.8 × 1013 Gredel et al. 1993
CO 3.2 × 1016 van Dishoeck et al. 1991

Table 2
Assumed H2 component structure

Comp. Rel. strength b vLSR

(km s−1 ) (km s−1 )
1 0.195 0.98 + 2.62
2 0.358 0.98 + 4.20
3 0.187 0.98 + 5.83
4 0.097 1.21 + 7.56
5 0.163 1.16 +10.85

Table 3
Column density in each rotational state

J log NJ Error
(cm−2 ) (cm−2 )
0 20.91 0.1
1 20.59 0.1
2 19.1 0.2
3 18.5 0.3
4 17.5 0.3a
5 17.0 0.3a
6 14.5 0.8a
7 14.4 0.5a
Total 21.08 0.1

a
The high-J lines are sub-
ject to shifts larger than the
quoted errors with changes in
the component structure; see
text.
6

Fig. 1.— Portion of LiF 1A spectrum with H2 model overlaid. The dotted line depicts the continuum derived from modeling the overlapping
wings of the lines from different bands. The HD line is not included in the model (see Ferlet, et al. 2000 for the HD analysis).
 7

Fig. 2.— Multi-component curve of growth for HD 73882. The two curves represent the range in damping constants for the measured lines.
The error bars lines only give the formal uncertainty in the line fits and may underestimate the total uncertainties for the stronger lines.

Fig. 3.— Excitation plot for HD 73882 for J = 0 through J = 7. The solid line corresponds to the assumed kinetic temperature derived
from J = 0 and J = 1, while the dashed line gives a weighted fit to the J ≥ 2 lines.