1: \documentclass[preprint,11pt]{aastex}
2:
3: \slugcomment{Accepted for publication in The Astrophysical
4: Journal Supplement Series}
5: \shorttitle{{\it FUSE} Translucent H$_2$ Survey}
6: \shortauthors{Rachford et al.}
7:
8: \begin{document}
9:
10: \title{Molecular Hydrogen in the {\it FUSE} Translucent Lines of Sight: The Full Sample}
11:
12: \author{Brian L. Rachford\altaffilmark{1},
13: Theodore P. Snow\altaffilmark{2},
14: Joshua D. Destree\altaffilmark{2},
15: Teresa L. Ross\altaffilmark{2},
16: Roger Ferlet\altaffilmark{3},
17: Scott D. Friedman\altaffilmark{4},
18: Cecile Gry\altaffilmark{5},
19: Edward B. Jenkins\altaffilmark{6},
20: Donald C. Morton\altaffilmark{7},
21: Blair D. Savage\altaffilmark{8},
22: J. Michael Shull\altaffilmark{2},
23: Paule Sonnentrucker\altaffilmark{9},
24: Jason Tumlinson\altaffilmark{10},
25: Alfred Vidal-Madjar\altaffilmark{3},
26: Daniel E. Welty\altaffilmark{11},
27: Donald G. York\altaffilmark{11,12}
28: }
29:
30: \altaffiltext{1}{Department of Physics, Embry-Riddle Aeronautical University, 3700 Willow Creek Road, Prescott, AZ 86301; rachf7ac@erau.edu}
31: \altaffiltext{2}{Center for Astrophysics and Space Astronomy, Department of Astrophysical and Planetary Sciences, University of Colorado, Boulder, CO 80309-0389}
32: \altaffiltext{3}{Institut d'Astrophysique de Paris, UMR7095 CNRS, Universite Pierre \& Marie Curie, 98bis Boulevard Arago, 75014 Paris, France}
33: \altaffiltext{4}{Space Telescope Science Institute, 3700 San Martin Drive, Baltimore, MD 21218}
34: \altaffiltext{5}{Laboratoire d'Astrophysique de Marseille, CNRS, 38 av. Frederic Joliot-Curie, 13388 Marseille cedex 13, France}
35: \altaffiltext{6}{Princeton University Observatory, Peyton Hall, Princeton, NJ 08544}
36: \altaffiltext{7}{Herzberg Institute of Astrophysics, National Research Council, 5071 W. Saanich Road, Victoria, BC V9E 2E7, Canada}
37: \altaffiltext{8}{Department of Astronomy, University of Wisconsin, 475 North Charter Street, Madison, WI 53706}
38: \altaffiltext{9}{Department of Physics and Astronomy, Johns Hopkins University, 3400 North Charles Street, Baltimore, MD 21218}
39: \altaffiltext{10}{Yale Center for Astronomy and Astrophysics, Department of Physics, P.O. Box 208121, New Haven, CT 06520}
40: \altaffiltext{11}{Department of Astronomy and Astrophysics, University of Chicago, 5640 South Ellis Avenue, Chicago, IL 60637}
41: \altaffiltext{12}{Enrico Fermi Institute, University of Chicago, 5640 South Ellis Avenue, Chicago, IL 60637}
42:
43:
44: \begin{abstract}
45: We report total abundances and related parameters for the full
46: sample of the {\it FUSE} survey of molecular hydrogen in 38
47: translucent lines of sight. New results are presented for the
48: ``second half'' of the survey involving 15 lines of sight to
49: supplement data for the first 23 lines of sight already
50: published. We assess the correlations between molecular
51: hydrogen and various extinction parameters in the full sample,
52: which covers a broader range of conditions than the initial
53: sample. In particular, we are now able to confirm that many,
54: but not all, lines of sight with shallow far-UV extinction
55: curves and large values of the total-to-selective extinction
56: ratio, $R_V$ = $A_V$/$E(B-V)$ --- characteristic of larger
57: than average dust grains --- are associated with particularly low
58: hydrogen molecular fractions ($f_{\rm H2}$). In the lines of
59: sight with large $R_V$, there is in fact a wide range in
60: molecular fractions, despite the expectation that the larger
61: grains should lead to less H$_2$ formation. However, we see
62: specific evidence that the molecular fractions in this
63: sub-sample are inversely related to the estimated strength of
64: the UV radiation field and thus the latter factor is more important
65: in this regime. We have provided an update to previous values
66: of the gas-to-dust ratio, $N$(H$_{\rm tot}$)/$E(B-V)$, based on
67: direct measurements of $N$(H$_2$) and $N$(H I). Although our
68: value is nearly identical to that found with {\it Copernicus}
69: data, it extends the relationship by a factor of 2 in reddening.
70: Finally, as the new lines of sight generally show low to
71: moderate molecular fractions, we still find little evidence for
72: single monolithic ``translucent clouds'' with $f_{\rm H2}$
73: $\sim$ 1.
74:
75: \end{abstract}
76:
77: \keywords{ISM: abundances --- ISM: clouds --- ISM: lines and bands ---
78: ISM: molecules --- ultraviolet: ISM}
79:
80:
81: \section{Introduction}
82: Molecular hydrogen is the most abundant molecule in the
83: universe, and a detailed knowledge of H$_2$ is crucial for a
84: full understanding of the physics and chemistry of the
85: interstellar medium. A broad overview of interstellar H$_2$ is
86: provided by Shull \& Beckwith (1982), while a recent overview of
87: chemical processes in diffuse and translucent clouds is provided
88: by Snow \& McCall (2006).
89:
90: A major goal of the {\it Far Ultraviolet Spectroscopic Explorer}
91: ({\it FUSE}) was a survey of molecular hydrogen in 45 lines of
92: sight with an emphasis on interstellar clouds with as much
93: extinction as possible. The extinctions covered the range of
94: the so-called ``translucent clouds'' with $A_V$ in the range
95: 1--5 mag. Within these limits, lines of sight were chosen from
96: a variety of environments and dust characteristics. The first
97: results from the study were presented by Rachford et al.\ (2002
98: [Paper I]).
99:
100: Paper I contains the details regarding the previous history of
101: H$_2$ observations and the rationale for the {\it FUSE} survey.
102: Slightly more than one-half of the planned targets had been
103: observed and analyzed and those lines of sight were detailed. A
104: main finding was that in most cases a line of sight was likely
105: composed of multiple clouds, suggesting a change in terminology
106: to ``translucent lines of sight'' pending the clear
107: identification of a high-extinction line of sight made up of one
108: highly molecular cloud.
109:
110: The main purpose of the present paper is to provide the
111: community with the overall H$_2$ results for the full sample, as
112: well as refine and strengthen some of the conclusions of Paper
113: I. The additional lines of sight give us much better coverage
114: of the variety of dust characteristics than the original partial
115: sample and we emphasize those results. Also, we have taken
116: advantage of the publication of the 2MASS All-Sky Survey of
117: Point Sources (Skrustkie et al.\ 2006) to provide more reliable
118: extinction parameters for the full sample.
119:
120: The rest of the paper is organized as follows: in \S\ 2 we
121: describe the remaining target for the FUSE translucent H$_2$
122: survey, along with comments on the stars chosen; in \S\ 3 we
123: give the details of the observations, data reduction, and
124: analysis; in \S\ 4 we discuss the results and their
125: implications; and we summarize the paper in \S\ 5.
126:
127:
128: \section{Target Selection and Stellar Properties}
129:
130: \subsection{General Comments on the Sample}
131: Specific information on the selection criteria is given in Paper
132: I, which includes details on the {\it FUSE} programs containing
133: the observed stars. In this paper, we report new H$_2$ results
134: for observations towards 15 stars, whose basic parameters are
135: listed in Table 1. In a few cases, the data quality was poor
136: for the new targets, but we were able to obtain reasonable H$_2$
137: measurements.
138:
139: While the overall observing program was very successful, for 7
140: out of the original 45 targets {\it FUSE} was unable to
141: adequately observe the target, or not observe it at all, as
142: follows. HD 37021 and HD 37061 are both very bright targets in
143: the Orion molecular cloud and the lines of sight display unusual
144: extinction characteristic of larger than normal dust grains. We
145: had anticipated that the sensitivity of {\it FUSE} would
146: decrease to the point where the UV fluxes would be safe, but
147: that did not happen. HD 147889 is in the $\rho$ Oph complex,
148: which also shows unusual extinction in the same sense as the
149: Orion targets. However, a dearth of suitable guide stars in the
150: field prevented this observation from taking place.
151: Fortunately, {\it FUSE} was able to observe nearby HD 147888 as
152: noted below. Walker 67 (in open cluster NGC 2264) and HD 166734
153: are two targets on the high end of the extinction range we
154: wished to cover, but both targets were fainter in the far UV
155: than anticipated and {\it FUSE} did not collect enough counts in
156: either observation for an adequate analysis. HD 94414 was also
157: a faint target and had some data quality issues. HD 21483 was
158: also not observed.
159:
160: \subsection{Stellar and Extinction Properties}
161: We have generally applied the techniques of Paper I to determine
162: the relevant extinction parameters for each line of sight,
163: namely the color excess ($E(B-V)$), the total-to-selective
164: extinction ratio ($R_V$), and the total visual extinction
165: ($A_V$). However, in the interim between Paper I and the
166: current paper we have made some improvements. Many of the
167: authors of both papers have been involved in a major project to
168: understand the diffuse interstellar bands (DIBs). Part of this
169: project includes deriving a consistent set of extinction
170: measures for the {\it FUSE} translucent lines of sight plus a
171: much larger sample of reddened stars. This project has led to
172: some modifications in the tabulated $V$ magnitudes, spectral
173: types, and extinction values used for our {\it FUSE} sample.
174: These modifications result in the best consistency between the
175: translucent sample, our DIB project, and the papers on chemical
176: depletions in the translucent sample in which the two lead
177: authors of the present paper were involved (Jensen, et al.\
178: 2005, 2007). We have thus not only provided extinction
179: parameters for the new targets, but also the revised values for
180: the targets from Paper I, and these values are given in Table 2.
181:
182: The differences between the new procedure for determining
183: extinction parameters and the original procedures in \S\ 2.2 of
184: Paper I are as follows. First, the values of $E(B-V)$ were
185: homogenized as much as possible, which has led mostly to
186: cosmetic changes from our values in Paper I. Two exceptions are
187: HD 24534 and HD 110432, which are both emission-line stars where
188: some of the observed reddening is likely circumstellar, and the
189: $E(B-V)$ values we quote here are somewhat larger than those
190: given in Paper I. However, these changes do not result in any
191: significant changes in the overall results.
192:
193: Our primary technique for determining $R_V$ involves fitting a
194: functional form to color excess ratios at optical and IR
195: wavelengths. This is based on the method of Martin \& Whittet
196: (1990), which we discuss in detail in Paper I. Subsequent to
197: the completion of Paper I, the full version of the 2MASS All-Sky
198: Survey of Point Sources (Skrutskie et al.\ 2006) became
199: available which includes very high quality $JHK$ photometry for
200: all of our sources. Thus, we have used these values in our
201: derivation of $R_V$, which has significantly improved the
202: quality, consistency, and completeness of extinction values in
203: our sample. We also used correlations between $R_V$ and the
204: wavelength of maximum polarization ($R_V$ = 5.6$\lambda_{\rm
205: max}$, with $\lambda$ in $\mu$m; Whittet \& van Breda 1978) and
206: the far-UV rise in the extinction curve ($c_2$ = $-$0.824 +
207: 4.717$R_V^{-1}$; Fitzpatrick 1999) as ancillary proxies for the
208: photometric values, again as discussed in Paper I.
209:
210: The listed uncertainties in the extinction parameters in Table 2
211: are necessarily estimates, but are mostly based on the scatter
212: observed for that particular technique or correlation and the
213: notion that systematic effects are likely important. For the
214: $R_V$ values that did not come from an analysis of the IR
215: photometry, and for the latter values when the fit appeared
216: qualitatively reasonable, we assumed uncertainties of 0.3. In
217: several cases, the IR photometry deviated from the expected
218: relationship and we thus adopted larger error bars. However, in
219: those cases, if at least one of the other two techniques for
220: determining $R_V$ agreed with the IR method, we then adopted the
221: IR value and used smaller error bars. Finally, when we did not
222: have any $R_V$ information (only for HD 186994), we adopted the
223: Galactic average value of 3.1 (Draine 2003). As can be seen in
224: Table 2, in only 3 cases, HD 24534, HD 102065, and HD 186994 did
225: we not use the $R_V$ derived from the IR photometry. HD 24534
226: is a Be star (as discussed below) for which we could not confirm
227: the photometric value with other methods. The other two stars
228: have $E(B-V)$ $<$ 0.2 for which we were concerned about the large
229: relative effect that a small error in the color indices or
230: spectral type will have on the photometric $R_V$. Published
231: values of $R_V$ for HD 102065 are near 4.0 (Boulanger, et al.\
232: 1994; Paper I).
233:
234: It is worth noting that as we pointed out in Paper I, the
235: photometric method is the most direct and the other two methods
236: are basically correlations between another parameter and $R_V$
237: derived from IR photometry. Thus, it is not a surprise that
238: there is generally excellent agreement between the three
239: indicators. Of the 22 points with both photometric and
240: polarimetric measurements only 4 disagree by more than 2
241: $\sigma$, and only an additional 3 disagree by more than 1
242: $\sigma$. The statistics comparing the photometric and
243: extinction curve values for $R_V$ are nearly identical. One
244: limitation with the polarimetric method is that it is poorly
245: constrained for $R_V$ $\gtrsim$ 4 (Whittet \& van Breda 1978;
246: Figure 1). In fact, we do not derive $R_V$ $>$ 3.9 for any of
247: our targets with this method, even when the other two methods
248: give larger $R_V$. This is one reason we have not attempted to
249: average together values from different methods.
250:
251: There are two particular problems with the photometric method of
252: determining $R_V$ that we wish to address. First, 7 of the
253: stars in our sample are Be stars\footnote{In addition to the 5
254: Be stars in Table 1, HD 24534 and HD 110432 from Paper I are
255: also Be stars.} and thus have circumstellar material. Processes
256: in this material skew not only the optical photometry via
257: emission lines and continuous free-bound emission, but also
258: optical polarization and IR photometry. The other issue is
259: variability. Not only are Be stars generally variable, but
260: there are other types of variable stars within the O and early B
261: types.
262:
263: In a similar context to ours, Bowen et al. (2008) discussed
264: these issues as they related to determining extinction
265: parameters for the {\it FUSE} survey of O VI in the Galactic
266: disk. These targets generally had $E(B-V)$ $\lesssim$ 0.4. One
267: advantage we have in the present survey is that our targets
268: generally have $E(B-V)$ $\gtrsim$ 0.4. This is important
269: because the greater the color excess, the smaller the relative
270: error caused by variability, and variability is often small in
271: this stellar temperature range. The Be stars are somewhat more
272: problematic, although again we expect the fractional effects in
273: our extinction parameters will be lessened for the more highly
274: reddened stars, and most of the Be stars have small variability
275: amplitudes (e.g. Hubert \& Floquet 1998). Furthermore, issues
276: like these are one reason that we have typically adopted
277: relatively large errors for $R_V$. In fact, all 7 of the Be
278: stars in our sample were flagged as not fitting the Martin \&
279: Whittet (1990) relation very well and thus were given
280: particularly large errors, which were only reduced in the
281: adopted $R_V$ if either of the other methods agreed with the
282: photometric value. Still, because these values may be
283: unreliable for Be stars, in some cases we have eliminated these
284: stars when considering a particular correlation.
285:
286: While preparing our present manuscript, Fitzpatrick \& Massa
287: (2005, 2007) published an updated version of their seminal work
288: in parametrizing extinction curves. They slightly modified
289: their parametrization of the UV portion of the extinction
290: curves, and greatly expanded the sample of lines of sight,
291: including a few from our sample which have not previously been
292: analyzed. However, since other authors have published curves in
293: the original 6-parameter scheme (Fitzpatrick \& Massa 1986,
294: 1988, 1990), an exclusive use of the new scheme would result in
295: a significant decrease in the fraction of stars available in a
296: self-consistent system. Thus, we have continued to use values
297: based on the older parametrization which are given in Table 3
298: for our new targets. The key parameter for the present paper is
299: $c_2$, the linear coefficient of the far-UV rise in the curve.
300:
301: We note that Fitzpatrick \& Massa (2005, 2007) also used 2MASS
302: $JHK$ photometry for the IR portion of the extinction curves,
303: and derived $R_V$ for their sample using a technique similar to
304: ours (Paper I). Indeed, there is an excellent match between our
305: values and theirs, as expected.
306:
307:
308: \subsection{Special line-of-sight characteristics}
309: Many of the lines of sight are of special interest due to their
310: location and/or environment. As in Paper I, we give a brief
311: overview of each line of sight in the following sections,
312: including the mention of particularly relevant values from
313: Tables 1--3.
314:
315: \subsubsection{HD 37903}
316: This star lies within the Orion molecular cloud and illuminates
317: the reflection nebula NGC 2023. UV spectra indicate significant
318: quantities of vibrationally excited H$_2$ along the line of
319: sight (Meyer et al.\ 2001) and the {\it FUSE} spectrum confirms
320: this. Meyer et al. (2001) concluded that the excitation was due
321: to UV fluorescence in a dense area of gas within 1 pc of the
322: star. The line of sight $R_V$ is above average and may be even
323: larger within the dense material local to the star. While the
324: UV extinction curve does not have a particularly unusual shape,
325: it does show smaller than normal extinction at all wavelengths
326: below 2500 \AA\ (Fitzpatrick \& Massa 1990).
327:
328: \subsubsection{HD 38087}
329: This star illuminates the reflection nebula IC 435, with
330: estimates that about one-quarter or less of the line-of-sight
331: material is local to the star (Snow \& Witt 1989). The line of
332: sight shows far less than normal far-UV extinction as well as a
333: significant shift of the 2175 \AA\ bump to longer wavelengths
334: (Fitzpatrick \& Massa 1990). IR photometry indicates an
335: exceptionally large total-to-selective extinction ratio, $R_V$ =
336: 5.57, consistent with the small far-UV extinction. Also, Witt,
337: Bohlin, \& Stecher (1986) found evidence for scattering
338: associated with the 2175 \AA\ interstellar feature. These data
339: strongly suggest that larger than normal dust grains have
340: developed in this presumably quiescent environment. Enhanced
341: (but uncertain) depletions of manganese and zinc in the line of
342: sight also suggest grain mantle growth (Snow \& Witt 1989),
343: while the abundances of oxygen (Jensen, et al.\ 2005)
344: and nitrogen (Jensen, et al.\ 2007) are normal.
345:
346: \subsubsection{HD 40893}
347: The extinction curve shows a relatively narrow and somewhat weak
348: 2175 \AA\ bump and slightly enhanced far-UV extinction
349: (Jenniskens \& Greenberg 1993). IR photometry also indicates an
350: abnormally small total-to-selective extinction ratio, $R_V$ =
351: 2.46.
352:
353: \subsubsection{HD 41117, HD 42087, HD 43384}
354: These stars all lie within the Gem OB1 association and sample a
355: roughly 2.5 $\times$ 4.5 degree field in the foreground of this
356: association. Extinction curves are available for the first two
357: stars and are very similar and normal, and the other extinction
358: measures for all three stars are similar and near average. HD
359: 41117 and HD 42087 are Be stars and the uncertainties in the
360: photometric $R_V$ are large, but there is excellent agreement
361: with the other two techniques.
362:
363: \subsubsection{HD 46056, HD 46202}
364: These stars lie within a quarter-degree of each other in open
365: cluster NGC 2244 within the Mon OB2 association and the lines of
366: sight show very similar extinction parameters. In both cases,
367: $R_V$ is slightly smaller than average and the amount of far-UV
368: extinction is above average. Also, the 2175 \AA\ bumps are
369: relatively weak and narrow. As will be shown, the H$_2$
370: parameters are indeed nearly identical.
371:
372: \subsubsection{HD 53367}
373: The line of sight to this star shows a relatively large amount
374: of reddening. IR photometry suggests a smaller than normal
375: value of $R_V$, but this is a Be star and the value is very
376: uncertain. There is no published extinction curve or
377: polarization data to verify the value of $R_V$.
378:
379: \subsubsection{HD 147888}
380: This star is also known as $\rho$ Oph D and lies within the
381: $\rho$ Oph cloud complex, which is well known for having unusual
382: dust characteristics. All three methods of assessing $R_V$
383: result in significantly above average values ($\sim$4--5) and
384: the extinction curve (Fitzpatrick \& Massa 1990) shows a
385: correspondingly small far-UV extinction. This particular line
386: of sight has small reddening relative to other stars in this
387: complex, and when combined with the shallow far-UV extinction
388: curve, it was accessible to {\it FUSE} even though the star is
389: of relatively late spectral type and thus has a small UV flux.
390:
391: \subsubsection{HD 149404}
392: This star is relatively bright optically and there have been
393: many studies of the material along the line of sight. Despite
394: the significant number of optical and mm-wave observations, a UV
395: extinction curve has not been published for this star. The
396: photometric value of $R_V$ is uncertain as HD 149404 is a Be
397: star. However, polarization data also indicates a normal value
398: of $R_V$.
399:
400: \subsubsection{HD 152236}
401: This star, also known as $\zeta^1$ Sco, is part of the Sco OB1
402: association. The star is relatively bright given the amount of
403: extinction and thus provides one of several lines of sight with
404: $A_V$ $\approx$ 2 that were easy for FUSE to observe. The UV
405: extinction curve is normal. It is a Be star, but polarization
406: and extinction curve data confirm the apparently normal
407: photometric value of $R_V$.
408:
409: \subsubsection{HD 164740}
410: Also known as Herschel 36, this star excites a compact H II
411: region within the Lagoon Nebula (M8, NGC 6523) known as the
412: Hourglass (Thackeray 1950, Wolff 1961). The value of $R_V$ for
413: this line of sight, 5.36, is exceptionally large and
414: a correction for material in the foreground of M8 would give an
415: even larger $R_V$ (Hecht et al.\ 1982). The unparametrized
416: far-UV extinction curve published by Hecht et al.\ (1982) is
417: also exceptionally shallow, and the new parametrization by
418: Fitzpatrick \& Massa (2007) gives a similar result. That
419: characteristic allowed us to take the first UV spectrum of this
420: interesting line of sight at sufficient resolution and S/N to
421: investigate interstellar abundances despite the large total
422: extinction. This spectrum shows a number of unusual and
423: interesting features which will be covered in a separate paper
424: (B.\ L.\ Rachford, in preparation, 2008), but we will give the
425: overall H$_2$ results here.
426:
427: \subsubsection{HD 179406}
428: Also known as 20 Aql, this star shines through apparently
429: typical diffuse cloud material and lies toward the lower end of
430: the amounts of extinction covered in this study. However, given
431: the small extinction, the abundances of carbonaceous molecules
432: are relatively large indicating that this line of sight may
433: sample a core of denser material (Hanson, et al.\ 1992).
434: Fitzpatrick \& Massa (2007) included this target in their
435: updated UV extinction curve parametrization work, which shows a
436: stronger than normal 2175 \AA\ bump, but is otherwise normal.
437:
438: \subsubsection{HD 186994}
439: This star was primarily observed because it is bright and had
440: been previously observed with {\it Copernicus}. Despite its
441: relatively large distance ($\sim$2500 pc) the quantity of
442: interstellar material along the line of sight is small.
443: Extinction information is limited for this star simply because
444: the small amount of reddening makes it difficult to properly
445: analyze the UV extinction curve or to determine $R_V$. (In
446: fact, this star has been used as a lightly reddened comparison
447: star in extinction studies; e.g., Sasseen et al.\ 2002.) The
448: H$_2$ column density observed with {\it Copernicus} was quite
449: small, log N(H$_2$) = 19.59 (Savage et al.\ 1977) and we confirm
450: this result.
451:
452: \section{Observations and data analysis}
453: Table 4 gives information on our {\it FUSE} observations for the
454: new targets. For all targets with multiple integrations, we
455: performed a shift-and-coadd procedure to combine the data for
456: each detector segment, but did not combine data from different
457: segments. Note that in two cases (HD 179406 and HD 186994), we
458: obtained multiple observations of the target. In both cases,
459: this was required due to the brightness of the targets as a
460: precaution against saturation of the detector. A very short
461: preliminary observation was obtained from which the true flux
462: was determined before spending time on the full observation.
463: The preliminary observation represented a non-trivial fraction
464: of the total data, so we included both datasets in the final
465: co-added spectra.
466:
467: We originally planned to use spectra that were uniformly
468: processed with version 2.4 (or later) of the CALFUSE pipeline,
469: including a revision of the column densities in Paper I that
470: were measured from earlier reductions. However, in comparing
471: results from differing versions of CALFUSE, we have not seen
472: significant differences in the derived column densities,
473: presumably because the extremely broad profiles are not affected
474: by subtle changes in the algorithms. In only a few cases did
475: the differences approach the value of the 1-$\sigma$
476: uncertainty. Thus, we have not revised the older values, nor
477: used CALFUSE versions beyond 2.4 for the newer targets.
478:
479: We used the same measurement techniques described in Paper I and
480: we will only give a very brief overview of those techniques. We
481: fitted H$_2$ line profiles to the Lyman series (4,0), (2,0), and
482: (1,0) ro-vibrational bandheads, including the $J$ = 0, 1, and 2
483: lines. Weaker lines from higher rotational states and other
484: interstellar species were fitted and removed from the broad
485: bandhead profiles. Our profile fits included the effects of
486: overlapping wings from adjacent bandheads as the $J$ = 0 and 1
487: profiles are heavily damped and extremely broad. The $J$ = 2
488: lines from these bandheads are often strong enough to show
489: damping wings as well, and these lines had to be included due to
490: blending with the broad $J$ = 0 and 1 profiles.
491:
492: At most, we obtained 9 independent measurements of the $J$ =
493: 0--2 column densities from the combinations of three
494: ro-vibrational bands and one to four detector segments covering
495: each band. Poor data quality, severe stellar interference in a
496: particular H$_2$ band, and/or problems during the observation
497: itself occasionally led to fewer suitable combinations of bands
498: and detector segments. The differences from one band/segment
499: combination to another were generally considerably larger than
500: the formal fit uncertainties, so we averaged each individual
501: measurement and used the sample standard deviation as our formal
502: uncertainty.
503:
504:
505: \section{Results and Discussion}
506:
507: \subsection{Overall comments}
508: Table 5 gives our primary results for the lines of sight. The
509: H$_2$ column densities were measured directly from the spectra
510: as already described. We also include two derived quantities,
511: the hydrogen molecular fraction ($f_{\rm H2}$), and the
512: rotational temperature ($T_{01}$) for each line of sight.
513:
514: The hydrogen molecular faction is defined in terms of the
515: molecular and atomic hydrogen column densities as
516:
517: \begin{equation}
518: f_{\rm H2} = \frac{2N(H_2)}{2N(H_2) + N(H I)}.
519: \end{equation}
520:
521: The rotational temperature (in Kelvin) is determined by applying the
522: Boltzmann equation to the ratio of the column densities in the
523: first two rotational states, yielding
524:
525: \begin{equation}
526: \frac{N(1)}{N(0)} = 9e^{-171/T_{01}}.
527: \end{equation}
528:
529: Solving for the temperature and expressing the column densities
530: as base-10 logarithms gives
531:
532: \begin{equation}
533: T_{01} = \frac{74.0}{\log N(0) - \log N(1) + 0.954}.
534: \end{equation}
535:
536: As in Paper I, we generally interpret the ratio between $N$(1)
537: (ortho-H$_2$) and $N$(0) (para-H$_2$) as the kinetic temperature
538: on the assumption that collisions with H$^{+}$ (and H3$^{+}$
539: when enough is present) dominate over other processes in
540: controlling this ratio.
541:
542: There is evidence that under some circumstances ortho-H$_2$ can
543: be rapidly converted to para-H$_2$ on grains (Le Bourlot 2000).
544: In this case, the $N$(1)/$N$(0) ratio becomes lower, yielding a
545: lower rotational temperature. This process appears to be more
546: likely at the low temperatures in the core of a relatively dense
547: and opaque cloud (Shaw et al.\ 2005). The fact that we seem to
548: be seeing multiple diffuse clouds along the lines of sight
549: (Paper I) suggests that this process may not important for our
550: sample.
551:
552: Shaw et al.\ (2005) have modeled the physical conditions in one
553: of our lines of sight from Paper I, HD 185418. Their model
554: includes the Le Bourlot (2000) treatment of the ortho to para
555: conversion process. Their derived kinetic temperature for the
556: gas was about 25\% lower than than our derived $T_{01}$ = 101
557: $\pm$ 14 K. This line of sight has the largest $T_{01}$ of our
558: entire sample, and the gas density is relatively low ($n_H$ = 27
559: cm$^{-3}$, but Sonnentrucker et al.\ 2003 find an even lower
560: value of $n_H$ = 6.3 cm$^{-3}$).
561:
562: Lacking a complete knowledge of whether or not our lines of
563: sight may have non-thermal $N$(1)/$N$(0) ratios, we will
564: generally assume that the measured $T_{01}$ values correlate
565: with kinetic temperature in some manner. However, this
566: potential uncertainty should be kept in mind in the
567: interpretation of correlations of $T_{01}$ with other
568: parameters.
569:
570: Overall, the second half of the sample shows column densities,
571: molecular fractions, and temperatures similar to those in Paper
572: I. The primary difference is the presence of a sample of lines
573: of sight with large extinction, small N(H$_2$), and large $R_V$,
574: which we will discuss in detail in \S\ 4.3.
575:
576: \subsection{General correlations with reddening}
577: We have updated several correlations with reddening from Paper I
578: using the new data. In Figure 1 we present a plot of the H$_2$
579: column density versus color excess (unlike Figure 2 in Paper I,
580: we give $N(H_2)$ on a linear scale). While this plot shows the
581: expected increase in H$_2$ as we look through more material, it
582: also reflects that our targets probe a wide variety of
583: environments in that we usually see a range of column densities
584: at a given color excess.
585:
586: One particularly important relationship we can investigate with
587: the complete dataset is the gas-to-dust ratio. Using {\it
588: Copernicus} data, Bohlin, et al.\ (1978) found
589: $N$(H$_{\rm tot}$)/$E(B-V)$ = 5.8 $\times$ 10$^{21}$ atoms
590: cm$^{-2}$ mag$^{-1}$. In Figure 2, we present plots of
591: $N$(H$_{\rm tot}$) versus $E(B-V)$. In the upper panel we
592: include the Bohlin et al.\ {\it Copernicus} data and our {\it
593: FUSE} data. In the lower panel we only include our {\it FUSE}
594: data for lines of sight with direct determinations of $N$(H I)
595: and $N$(H$_2$), and also excluded Be stars as the color excesses
596: might be overstated due to the circumstellar emission. This
597: should give the most homogeneous sub-sample possible that covers
598: a broad range in color excess. When constrained to pass through
599: the origin, our error-weighted best-fit slope is (5.94 $\pm$
600: 0.37) $\times$ 10$^{21}$ atoms cm$^{-2}$ mag$^{-1}$, essentially
601: identical to the {\it Copernicus} value for less reddened lines
602: of sight. The solid line in both panels is the best fit, and it
603: is worth noting that many of the most discrepant {\it FUSE}
604: points in the upper panel are Be stars, which are removed in the
605: bottom panel.
606:
607: For reference, we have also included an unconstrained fit in
608: these panels which furthermore does not include the
609: low-reddening point at $E(B-V)$ = 0.17 (HD 186994). This
610: illustrates the significance of constraining the fit with points
611: at small reddening. Clearly this line is not a good fit to the
612: low-extinction {\it Copernicus} points. One possibility is that
613: the high-extinction sample has a different slope than the
614: low-extinction sample. This would represent a difference in the
615: gas-to-dust ratio that might indicate a change in dust
616: properties in the clouds with higher extinction. Such a
617: difference is not clearly seen in our data. The slope of the
618: dashed line is (4.2 $\pm$ 1.7) $\times$ 10$^{21}$ atoms cm$^{-2}$
619: mag$^{-1}$, less than the value for the constrained fit, but
620: with large enough uncertainty to be consistent with the
621: constrained fit.
622:
623: We performed the same analysis for the total visual extinction,
624: $A_V$, which is simply the product of $E(B-V)$ and $R_V$.
625: Figure 3 shows panels corresponding to the ones shown in Figure
626: 2. Again, we have fitted a line to our ``best'' sub-sample in
627: the lower panel, which yields $N$(H$_{\rm tot}$)/$A_V$ = (2.15
628: $\pm$ 0.14) $\times$ 10$^{21}$ atoms cm$^{-2}$ mag$^{-1}$. This
629: value is nearly identical to a simple division of 5.94 $\times$
630: 10$^{21}$ by the average $R_V$ = 2.93 for this sub-sample; the
631: latter value is slightly less than the galactic average of 3.1
632: (Draine 2003). Again, excluding HD 186994 and not constraining
633: the fit to pass through the origin gives a slightly smaller
634: slope of (1.50 $\pm$ 0.43) $\times$ 10$^{21}$ atoms cm$^{-2}$
635: mag$^{-1}$. Visually, there is a hint of a slope change in the
636: $A_V$ = 1.5--2.0 interval in the upper panel, but with so few
637: reliable points with $A_V$ $\geq$ 2, such a trend is not clear.
638:
639: As Figure 4 shows (an update of Figure 5 in Paper I), there
640: appears to be a weak inverse correlation between rotational
641: temperature and reddening. However, it is critical to note that
642: much of this trend disappears if we were to exclude the {\it
643: Copernicus} data points with $N$(H$_2$) $<$ 10$^{20}$ cm$^{-2}$.
644: In the region of overlap with {\it Copernicus} data, the {\it
645: FUSE} lines of sight show smaller average temperatures, while
646: the {\it FUSE} data for large reddening do not deviate from the
647: similar data for small reddening. The mean $T_{01}$ for our
648: entire {\it FUSE} sample is 67 K with a standard deviation of 14
649: K. This result is similar to other studies, including the
650: $<T_{01}>$ = 86 $\pm$ 20 K from the {\it FUSE} Galactic disk
651: survey (J.\ M.\ Shull et al., in preparation). However, the
652: $T_{01}$ distribution of our sample is not normal, having a
653: general rise in frequency up to around 75--80 K, and just a few
654: stars with temperatures above that; the latter fact can easily
655: be seen in Figure 4.
656:
657: In principle, the larger color excesses could be associated with
658: denser clouds, which in turn might be expected to show lower
659: temperatures. In Paper I we noted a correlation between the
660: rotational temperature and the fractional abundance of the CN
661: radical, the latter of which is a good density indicator
662: (Federman, et al.\ 1984). However, as we discussed in
663: Paper I, the lines of sight in the translucent sample seem to
664: mostly be made up of multiple diffuse clouds. Thus, while the
665: slight trend is in the right direction to suggest that we
666: are probing somewhat denser clouds in the {\it FUSE} sample,
667: this conclusion is somewhat speculative.
668:
669: Figure 5 (an update of Figure 7 in Paper I) shows a similar
670: pattern to Figure 1 in that once H$_2$ becomes self-shielded and
671: relatively abundant, the molecular fraction covers nearly the
672: entire possible range at any given reddening. Again, part of
673: this may be a selection effect as the lines of sight were chosen
674: to sample a variety of environments. However, it is notable
675: that even with the additional sample we have not found a line of
676: sight with $f_{\rm H2}$ $>$ 0.8. In fact, as seen in Table 5,
677: the new lines of sight preferentially have relatively small
678: $f_{\rm H2}$ given the large extinctions.
679:
680:
681: \subsection{The high $R_V$ sample}
682: \subsubsection{Significance}
683: An important aspect of the completed translucent sample is that
684: we have determined the molecular hydrogen column densities for
685: several lines of sight with particularly large values of $R_V$
686: combined with larger $A_V$ and total hydrogen column densities
687: than previous samples. In some cases, these lines of sight show
688: unusually small molecular fractions and are of particular
689: interest because of the relationship between dust parameters and
690: molecular hydrogen formation and destruction.
691:
692: Snow (1983) first investigated the possibility that H$_2$
693: abundances may be inversely correlated with grain size. Larger
694: grains, through the coagulation of small grains, provide less
695: surface area per unit dust mass. With less surface area available,
696: the H$_2$ formation rate should be diminished. Using the
697: compilation of {\it Copernicus} data of Bohlin et al.
698: (1978), Snow demonstrated that the mean molecular fraction in
699: the $\rho$ Oph cloud was a factor of 2.6 less than the
700: rest of the sample.
701:
702: Cardelli (1988) explored the related issue of the relationship
703: between H$_2$ abundances and $R_V$. The H$_2$ abundances again
704: came from Bohlin et al. (1978), while the values of
705: $R_V$ came from an analysis of IR photometry similar to that
706: which we have applied in the present work.
707:
708: Cardelli's sample displayed an inverse relationship between the
709: H$_2$ to $A_V$ ratio and $R_V$, with a weaker, but similar
710: dependence of the hydrogen molecular fraction and $R_V$. When
711: splitting the {\it Copernicus} lines of sight at $R_V$ = 3.5,
712: the high $R_V$ group had an average molecular fraction of
713: approximately a factor of 2.5 less than the low $R_V$ group. As
714: $R_V$ is thought to be positively correlated with grain size (or
715: grain size distribution) this result provides further support
716: for the role of grain size in regulating the H$_2$ abundance.
717: Furthermore, the large values of $R_V$ are associated with
718: smaller than normal far-UV extinction, which allows more
719: photodissociating radiation to penetrate the interstellar clouds
720: and provides further reduction in the hydrogen molecular
721: fraction. In fact, it appears that only the lines of sight with
722: the largest $R_V$ have extinction curves that deviate in a
723: consistent manner from average (Fitzpatrick \& Massa 2007).
724:
725: In Paper I, we had minimal coverage of high $R_V$ lines of sight
726: in our {\it FUSE} sample, but the newly added targets improve
727: the situation. In Figure 6 we show the relationship between
728: $f_{\rm H2}$ and $R_V$ for the {\it FUSE} and {\it Copernicus}
729: samples. This provides a major update to Figure 8 in Paper I as
730: we have also included the {\it Copernicus} sample, using 2MASS
731: photometry to derive the values of $R_V$ as with the {\it FUSE}
732: sample.
733:
734: Interestingly, using the Spearman rank correlation coefficient,
735: at best we only see an anti-correlation between molecular
736: fraction and $R_V$ at about the 1 $\sigma$ level whether or not
737: we restrict the sample to the best determinations of either
738: parameter. A more significant trend (3 $\sigma$) appears when
739: considering the H$_2$ to $A_V$ ratio versus $R_V$ similar to
740: Cardelli's result, but as he discussed, this division works best
741: when one can make the assumptions of similar average densities
742: in the various clouds and that most of the extinction occurs in
743: the regions where H$_2$ is significantly present. However, it
744: is not clear that this is an appropriate way to look at the {\it
745: FUSE} translucent sample. In particular, much of the trend
746: disappears if we exclude lines of sight with $A_V$ $>$ 2. As
747: discussed in Paper I, we believe that most of our lines of sight
748: are sampling multiple clouds that may be spread out in space,
749: and thus may have a variety of conditions.
750:
751: We wish to consider in more detail the lines of sight with
752: $R_V$ $>$ 4, which are labeled in Figure 6, split equally
753: between the {\it FUSE} and {\it Copernicus} samples. We will
754: ignore the {\it Copernicus} target HD 10516 ($\phi$ Per) due to
755: the very large uncertainty in $R_V$ and small extinction
756: ($E(B-V)$ $\approx$ 0.2 and $A_V$ $\sim$ 1). We will thus focus
757: our discussion on the five remaining lines of sight with high
758: $R_V$ and large extinction.
759:
760: Detailed modeling of an individual line of sight is complex and
761: requires a large amount of data to constrain the model, although
762: this has been done for a few favorable lines of sight from Paper
763: I (Rachford et al.\ 2001; Sonnentrucker et al.\ 2003). Barring
764: such an analysis, we can get a sense of the strength of the
765: local UV radiation field using the high $J$ lines of H$_2$, and
766: thus an estimate of the importance of photodissociation. A
767: preliminary analysis of lines from $J$ = 2 up to the highest
768: observable levels in a line of sight yields column densities
769: that can be used to estimate the amount of radiative pumping to
770: these levels. This has proven challenging in many of the
771: translucent lines of sight due primarily to data quality and the
772: resulting difficulty in measuring weak lines and properly
773: interpreting saturated lines. This is particularly acute in
774: these heavily reddened lines of sight as in many cases the S/N
775: rapidly decreases with decreasing wavelength. Also, the
776: strengths of other lines are typically larger than in less
777: reddened lines of sight which causes more interference with the
778: high-$J$ lines. However, we have enough information for most of
779: these five lines of sight to use $N$(6) and $N$(7) as potential
780: indicators of the strength of the radiation field. These lines
781: are typically weakly saturated and not as difficult to interpret
782: as the stronger lines from lower rotational states, when the
783: data quality permits their detection or the derivation of well
784: constrained upper limits. For reference, the column densities
785: for HD 110432 (Rachford et al.\ 2001) were log $N$(6) = 14.20
786: $\pm$ 0.20 and log $N$(7) = 13.25$^{+1.25}_{-1.00}$ for a
787: radiation field that was modeled as twice the strength of the
788: average curve of Draine (1978). Particle density also plays a
789: significant role in controlling H$_2$ excitation, so we should
790: look at the high $J$ column densities as an indicator of the
791: strength of the radiation field, but not as a definitive
792: measurement.
793:
794: \subsubsection{HD 38087}
795: HD 38087 has $f_{\rm H2}$ greater than half, the largest of the
796: group, but this value is uncertain because we do not have a
797: direct measurement of interstellar N(H I) due to the very late
798: spectral type. Fitzpatrick \& Massa (1990) report log N(H I) =
799: 21.48 from Lyman $\alpha$. However, at spectral type B5 V, the
800: interstellar line is strongly contaminated or even dominated by
801: the stellar line (Shull \& van Steenberg 1985; Diplas \& Savage
802: 1994).
803:
804: Atomic hydrogen column density is highly correlated with the
805: strength of certain diffuse interstellar bands, including
806: $\lambda$5780 (S.\ D.\ Friedman et al.\ 2008, in preparation; see
807: Herbig 1993 or Welty et al.\ 2006 for similar correlations), and
808: applying this correlation to HD 38087 gives log N(H I) = 21.08,
809: similar to our quoted value and which would only reduce the
810: derived molecular fraction from 0.52 to 0.42. Thus, it seems
811: likely that this line of sight is genuinely rich in molecules.
812:
813: We have already noted in \S\ 2.3.2 that this line of sight
814: samples a quiescent environment where grain growth might be
815: expected to occur (Snow \& Witt 1989). However, only about 25\%
816: of the line of sight material is thought to be in the reflection
817: nebula. The extinction curve indicates that the interstellar
818: material should be relatively transparent to photodissociating
819: far-UV radiation -- if such radiation exists at the cloud
820: location. HD 38087 itself is of rather late spectral type, so
821: it may not contribute significantly to the far-UV radiation
822: field at the location of the bulk of the line of sight material.
823: However, the {\it FUSE} spectrum clearly shows H$_2$ lines up to
824: the $J$ = 7 level and we derive a logarithmic column density of
825: about 15.0 in that level, a very large value that implies a
826: significant excitation mechanism.
827:
828: It is important to note that even a ``large'' quantity of
829: excited H$_2$ corresponds to a very small fraction of the total
830: H$_2$ column density. One simply needs a particularly high
831: excitation temperature to produce a relatively ``flat''
832: rotational distribution with a high percentage of the H$_2$ in
833: the excited states, while keeping the total column density of
834: this material several orders of magnitude below the 10$^{20-21}$
835: cm $^{-2}$ totals seen in the translucent lines of sight. It
836: thus may be the case that the excited H$_2$ is produced in the
837: reflection nebula, while the bulk of the low-excitation H$_2$ is
838: found farther from the star where photodissociation is not as
839: important.
840:
841: \subsubsection{HD 148184}
842: The other line of sight from the high $R_V$ sample with
843: significant molecular material is HD 148184 ($\chi$ Oph), with
844: $f_{\rm H2}$ = 0.33 (Bohlin et al.\ 1978).
845: This star is about 5 degrees from the $\rho$ Oph grouping and
846: samples material from the general Sco-Oph cloud complex. The
847: {\it Hipparcos} parallax is 6.21 $\pm$ 0.23 mas (van Leeuwen
848: 2007), corresponding to a distance of 161 $\pm$ 6 pc, which
849: places it near the distant edge of the interstellar material.
850: The star itself has spectral type B2 IVpe and may contribute
851: significant UV radiation to the material in the line of sight,
852: nearly all of which is likely associated with the cloud complex.
853: However, in the overall sense this portion of the complex is not
854: as highly populated by B-type stars as in the immediate $\rho$
855: Oph area.
856:
857: This star was observed with {\it Copernicus} at high resolution,
858: but with very limited wavelength coverage. Frisch (1980)
859: reported upper limits for two $J$ = 6 lines, yielding an upper
860: limit of log $N$(6) $<$ 14.33. Thus, there does not appear to
861: be an unusually large source of excitation.
862:
863: One final important consideration for this star is that it is
864: the only one of the five which is a Be star and thus the caveats
865: given in \S\ 2.2 apply. As the large error bar in Figure 6
866: implies, we found the color excesses to be a poor fit to the
867: Martin \& Whittet (1990) relation. There is no available
868: extinction curve for this star. The wavelength of maximum
869: polarization is 0.55 $\mu$m (Coyne et al. 1974), corresponding
870: to $R_V$ = 3.08. It is thus possible that this line of sight
871: should not be in the high $R_V$ group.
872:
873: \subsubsection{The $\rho$ Oph cloud: $\rho$ Oph A \& D}
874: The $\rho$ Oph cloud has long been known as a location which
875: exhibits large $R_V$ and larger than normal dust grains
876: (Carrasco, et al.\ 1973; Whittet 1974). The name $\rho$
877: Oph is applied to four stars within a few arcminutes of each
878: other labeled A through D. The A component (HD 147933) is the
879: brightest star in both the visible and UV and was observed by
880: {\it Copernicus}. The D component (HD 147888) is part of the
881: present {\it FUSE} sample. Improved {\it Hipparcos} distances
882: to these stars (van Leeuwen 2007) have been used as part of a
883: study of the distribution of interstellar material in the cloud
884: by Snow et al.\ (2008). These distances are 111$^{+12}_{-10}$ pc
885: for $\rho$ Oph A and 125$^{+14}_{-11}$ pc for $\rho$ Oph D. The
886: overall analysis indicates that both stars are in front of the
887: denser material that is sampled by the more distant and more
888: heavily obscured line of sight to HD 147889 that we had hoped to
889: study with {\it FUSE} as noted in \S\ 2.1.
890:
891: The uncertainties in the {\it Copernicus} measurements of
892: $N$(H$_2$) and $N$(H I) for $\rho$ Oph A are relatively large,
893: but still strongly indicate a small molecular fraction.
894: Unfortunately, the spectral type of $\rho$ Oph D (B5 V) is late
895: enough that the interstellar Ly$\alpha$ line will be severely
896: contaminated by the stellar line, as with HD 38087, thus $N$(H
897: I) is very uncertain. Cartledge et al.\ (2004) indirectly
898: estimated a value log $N$(H$_{\rm tot}$) = 21.73 $\pm$ 0.09
899: based on measurements towards nearby $\rho$ Oph A. For this
900: paper, we have used our preferred method for stars later than
901: spectral type B2, namely, applying the Bohlin et al. (1978)
902: $N$(H$_{\rm tot}$)/$E(B-V)$ = 5.8 $\times$ 10$^{21}$ atoms
903: cm$^{-2}$ mag$^{-1}$ value we discuss in \S\ 4.2. This gives
904: log $N$(H$_{\rm tot}$) = 21.44, much smaller than the value for
905: $\rho$ Oph A, but in excellent agreement with the correlation
906: between $N$(H I) and the equivalent width of the $\lambda$5780
907: DIB discussed in \S\ 4.3.2. It should be noted that $\rho$ Oph
908: A is well known as a line of sight with a larger than normal
909: gas-to-dust ratio (e.g.\ Bohlin et al.\ 1978). For that reason,
910: our derived value of log $N$(H$_{\rm tot}$) for $\rho$ Oph D may
911: also be too low. A larger value of $N$(H$_{\rm tot}$) would
912: produce an even smaller molecular fraction than the $f_{\rm H2}$
913: = 0.21 depicted in Figure 6. Thus, it appears that the
914: molecular fractions toward both stars are relatively small given
915: the amount of reddening and extinction.
916:
917: There are a number of B-type stars in the vicinity and the UV
918: radiation field is thought to be strong in this area despite the
919: lack of O-type stars (e.g., Kulesa et al. 2005). Our
920: preliminary analysis of the high $J$ lines in the {\it FUSE}
921: spectrum of $\rho$ Oph D reveals no conclusive detections beyond
922: $J$ = 6, from which we derive an uncertain log $N$(6) = 14.3. This
923: is comparable to that found toward HD 110432 (Rachford et al.
924: 2001) which was modeled with a radiation field twice the
925: interstellar average curve of Draine (1978).
926:
927: \subsubsection{Herschel 36}
928: As noted in \S\ 2.3.10, HD 164740, better known as Herschel 36,
929: lies within the Lagoon Nebula (M8) in a region with recent and
930: ongoing star formation. The molecular fraction (0.03) is the
931: smallest known for a line of sight with $E(B-V)$ $>$ 0.3. It is
932: believed that some of the intervening material lies close to the
933: star and is thus subject to the very strong far-UV radiation
934: field of the late O-type star. In fact, our {\it FUSE} spectrum
935: shows extreme H$_2$ excitation, including numerous lines from
936: vibrationally excited states (B.\ L.\ Rachford, in preparation)
937: demonstrating the likelihood of significant H$_2$ lying close to
938: the star and the resultant exposure to a large far-UV flux.
939: Furthermore, the exceptionally small far-UV extinction will
940: allow the radiation from the O-type stars in M8 to influence a
941: greater distance along the line of sight. Thus, while this line
942: of sight samples a complicated environment, it is plausible to
943: assume that the radiation field is contributing to the small
944: line of sight molecular fraction.
945:
946: \subsubsection{Overall properties of this sample}
947: The 5 lines of sight were chosen for having evidence for larger
948: than normal grains, thus providing smaller than normal H$_2$
949: formation rates per unit dust mass. Also, the far-UV extinction
950: is small which allows greater than normal penetration of the
951: photodissociating radiation. However, there remains a large
952: range in molecular fractions that spans most of the total range
953: we see in the overall translucent sample.
954:
955: In formation-destruction equilibrium, the hydrogen molecular
956: fraction will be proportional to the factors
957: controlling formation, which are density ($n_H$) and the
958: formation rate coefficient ($R$), and inversely proportional to
959: the far-UV radiation field which controls destruction. For the
960: moment, we will focus on the last point.
961:
962: There seems to be some evidence that the strength of the local
963: radiation field is responsible for this range, particularly when
964: considering the $\rho$ Oph cloud and Herschel 36. However, the
965: line of sight toward HD 38087 is an exception. As already
966: noted, the distribution of material along this line of sight
967: likely also plays a significant role.
968:
969: Interestingly, these lines of sight have small H$_2$ rotational
970: temperatures. In Figure 7 (an update of Figure 9 in Paper I) we
971: show the relationship between molecular fraction and rotational
972: temperature. In Paper I, we highlighted the high $f_{\rm H2}$,
973: low $T_{01}$ lines of sight. Much rarer are lines of sight with
974: low $f_{\rm H2}$ and low $T_{01}$. In fact, $\rho$ Oph A and D
975: and Herschel 36 are the only lines of sight in the {\it
976: Copernicus}/{\it FUSE} sample with $f_{\rm H2}$ $<$ 0.2,
977: $T_{01}$ $\leq$ 60 K, and $E(B-V)$ $>$ 0.2 (or log N(H) $>$
978: 21.1).
979:
980: In contrast to the high $R_V$ lines of sight with small
981: molecular fraction and low temperature, there are numerous lines
982: of sight with normal or low values of $R_V$ that have small
983: molecular fractions and higher than normal temperatures,
984: particularly within the {\it Copernicus} sample. Gas heating
985: due to grain electron photoemission (Draine 1978) may contribute
986: to this fact, although there is not a significant overall
987: relationship between $R_V$ and $T_{01}$. The 13 {\it
988: Copernicus} lines of sight with $R_V$ $<$ 4 and $f_{\rm H2}$ $<$
989: 0.2 all have $E(B-V)$ = 0.2--0.4, and all but one (HD 147165;
990: $T_{01}$ = 64 K) have rotational temperatures greater than the
991: average from our {\it FUSE} sample of 67 K. Of the 4 {\it FUSE}
992: lines of sight with $R_V$ $<$ 4 and $f_{\rm H2}$ $<$ 0.2, only
993: one has $T_{01}$ $<$ 67 K (HD 152236; $T_{01}$ = 62 K). Of
994: these 17 {\it Copernicus} and {\it FUSE} total lines of sight
995: none have $T_{01}$ as small as $\rho$ Oph A, $\rho$ Oph D, or
996: Herschel 36.
997:
998: It is clear from Figure 7 that these three lines of sight are
999: part of a small group that stand out from the rest of the
1000: sample, going against the generally inverse relationship between
1001: molecular fraction and temperature and lying in the bottom left
1002: of the figure. We would expect the cold lines of sight to
1003: contain denser material and represent the expected environment
1004: for the large grains. But, for at least Herschel 36 and $\rho$
1005: Oph D, the material is subject to considerable far-UV radiation,
1006: contributing to the small molecular fractions.
1007:
1008: In these cases, we may be seeing the effect of a broad
1009: distribution of material. Perhaps there is both ``cold''
1010: diffuse material which contains most of the H$_2$ and is still
1011: not dense enough to exhibit a level of self-shielding that would
1012: allow a large molecular fraction, yet there is also a relatively
1013: small amount of material closer to the hot stars that not only
1014: has few molecules, but also considerable H$_2$ excitation. In
1015: particular, for Herschel 36 there seems to be a significant
1016: velocity difference between the excited material and the
1017: ``cold'' material as indicated by the low-$J$ lines of H$_2$.
1018: Thus, the cold material may be significantly in the foreground
1019: of Herschel 36.
1020:
1021: As previously noted, the radiation field is one of three factors
1022: that influence the molecular fraction in the models of Browning
1023: et al.\ (2003), along with formation rate coefficient and
1024: density. The models show that for the column densities
1025: we are sampling in this paper, there can be degeneracy in the
1026: molecular fractions that result from different combinations of
1027: the three factors. In particular, a small formation rate
1028: coefficient can lower the molecular fraction in a manner similar
1029: to an increase in the UV flux.
1030:
1031: The total column densities in our sub-sample of 5 high $R_V$
1032: lines of sight cover the range log $N$(H$_{\rm tot}$) $\approx$
1033: 21.2-22.0 ($N$ in cm$^{-2}$). In this range, Browning et al.\
1034: (2003) find that a UV radiation field 50 times the Galactic mean
1035: reduces the molecular fraction by $\sim$5 orders of magnitude at
1036: the low end, to factors of $\sim$2--3 at the high end. A
1037: reduction in the formation rate coefficient by a factor of 10
1038: reduces the molecular fraction by $\sim$2 orders of magnitude at
1039: the low end and by a minimal amount at the high end. These two
1040: factors in combination were required to match H$_2$ data for the
1041: Small and Large Magellanic Clouds at levels of extinction
1042: generally below those in our present sample.
1043:
1044: The most likely variables that would change the formation rate
1045: coefficient are grain size and temperature. Since we have
1046: limited this sub-sample to high $R_V$, differences in grain size
1047: distribution have presumably been minimized. If the
1048: $N$(1)/$N$(0) rotational temperature is a meaningful indicator
1049: of kinetic temperatures, our sub-sample is more or less
1050: isothermal.
1051:
1052: Density, the third factor considered in the Browning et al.\
1053: (2003) models, is generally not as important within our range of
1054: column densities for typical Galactic values of radiation field
1055: and formation rate coefficient, particularly since the molecular
1056: fractions are usually relatively large already. When combined
1057: with large radiation field or small formation rate coefficient,
1058: density variations can cause a large spread in the resulting
1059: small molecular fractions.
1060:
1061: In conclusion, given the relatively large column densities in
1062: the regime we are probing with our sample, it appears more
1063: likely that large variations in radiation field are the dominant
1064: cause for the variations in molecular fraction. Formation rate
1065: coefficients far below the Galactic average may contribute to
1066: smaller molecular fractions in our regime, but it is more
1067: certain that some of the clouds we are studying lie very close
1068: to major UV sources that dramatically increase the local
1069: radiation field.
1070:
1071: \section{Summary}
1072: We have completed the primary molecular hydrogen analysis for
1073: the {\it FUSE} translucent lines of sight. Total H$_2$ column
1074: densities have been measured for a total of 38 lines of sight
1075: with $A_V$ $\gtrsim$ 1 via profile fitting of transitions from
1076: the lowest two vibrational levels which contain $\sim$99\% of
1077: the material. In addition, we have derived the H$_2$ molecular
1078: fractions and rotational temperatures for the lines of sight and
1079: considered various correlations between parameters. In
1080: particular, using these data we have found that the gas-to-dust
1081: ratio ($N$(H$_{\rm tot}$)/$E(B-V)$) remains identical to that
1082: found with {\it Copernicus} data out to $E(B-V)$ $\approx$ 1, a
1083: factor of 2 farther than the previous determination.
1084:
1085: These lines of sight were chosen to sample a wide variety of
1086: environments, including those with unusual dust characteristics,
1087: as dust grains are thought to provide the primary environment
1088: for H$_2$ formation in these clouds. An important consequence
1089: of the updated sample is that we have much better coverage of
1090: lines of sight with large total-to-selective extinction ratios
1091: ($R_V$) and smaller than normal far-UV extinction. These
1092: unusual characteristics are thought to indicate larger than
1093: normal dust grains for which the grain area per unit mass will
1094: be lower. In the lines of sight with large grains, we still see
1095: a large range in molecular fraction and can mostly attribute
1096: this to a range in the strength of the local interstellar far-UV
1097: radiation field, perhaps enhanced by material being widely
1098: distributed along the line of sight and/or by variations in the
1099: H$_2$ formation rate coefficient. Overall, we do not see a
1100: statistically significant trend of decreasing molecular fraction
1101: with increasing $R_V$.
1102:
1103: As our new lines of sight all have molecular fractions $f_{\rm
1104: H2}$ $\lesssim$ 0.5, our conclusions regarding the presence of
1105: truly ``translucent clouds'' with $f_{\rm H2}$ near unity are
1106: unchanged. Namely, the lines of sight in our survey are
1107: primarily sampling multiple clouds without a high-extinction core
1108: that is dominated by molecules.
1109:
1110: Work is still ongoing to fully utilize the {\it FUSE} data for
1111: the translucent sample, including a survey of the HD molecule
1112: (Snow et al.\ 2008), and a detailed analysis
1113: of the line of sight to Herschel 36 (B.\ L.\ Rachford 2008, in
1114: preparation).
1115:
1116: With the official end of the {\it FUSE} mission in 2007, direct
1117: far-UV measurements of another significant sample of reddened
1118: lines of sight will have to wait until another mission is
1119: launched. One of the goals of the planned Hubble
1120: Service Mission 4 in late 2008 will be to install the Cosmic
1121: Origins Spectrograph (COS) in the Hubble Space Telescope. This
1122: instrument is primarily designed to observe UV wavelengths
1123: longward of 1150 \AA\ at high sensitivity and similar resolution
1124: to that of {\it FUSE}. This will provide access to lines from
1125: numerous atomic and molecular species in more heavily reddened
1126: lines of sight than any previous mission. However, the
1127: combination of the {\it HST} optics and {\it COS} may have
1128: enough residual sensitivity at 1108 \AA\ to sample the longest
1129: wavelength (0,0) ro-vibrational bandhead of H$_2$. These
1130: potential observations of H$_2$ would occur at much lower
1131: resolution than {\it FUSE} ($R$ $\sim$ 3000) and will thus be
1132: more difficult to analyze. However, this might provide a
1133: significant constraint on molecular fractions for very heavily
1134: reddened lines of sight that were not accessible with {\it
1135: FUSE}, possibly revealing true ``translucent clouds'' with
1136: $f_{\rm H2}$ $\approx$ 1.
1137:
1138:
1139: \acknowledgments
1140: We thank the referee for useful comments. This work is based on
1141: data obtained for the Guaranteed Time Team by the NASA-CNES-CSA
1142: {\it FUSE} mission operated by the Johns Hopkins University.
1143: Financial support to U.S. participants has been provided by NASA
1144: contract NAS5-32985 and NASA grant NNX08AC14G. This research
1145: has made use of the SIMBAD database, operated at CDS,
1146: Strasbourg, France.
1147:
1148: {\it Facilities:} \facility{FUSE (LWRS,MDRS)}
1149:
1150:
1151: \begin{references}
1152:
1153: \reference{} Andersson, B.-G., \& Wannier, P. G. 2000, \apj, 491, L103
1154: \reference{} Anderson, C.\ M. et al. 1996, \aj, 112, 2726
1155: \reference{} Bohlin, R.\ C., Savage, B.\ D., \& Drake, J.\ F. 1978, \apj, 224, 132
1156: \reference{} Boulanger, F., Pr\'{e}vot, M. L., Gry, C. 1994, \aap, 284, 956
1157: \reference{} Bowen, D.\ V. et al. 2008, \apjs, 176, 59
1158: \reference{} Browning, M.\ K., Tumlinson, J., \& Shull, J. M. 2003, \apj, 582, 810
1159: \reference{} Cardelli, J.\ A. 1988, \apj, 335, 177
1160: \reference{} Carrasco, L., Strom, S.\ E. \& Strom, K.\ M. 1973, \apj, 182, 95
1161: \reference{} Cartledge, S.\ I.\ B., Lauroesch, J.\ T., Meyer, D.\ M., \& Sofia, U.\ J. 2004, \apj, 613, 1037
1162: \reference{} Coyne, G.\ V., Gehrels, T., \& Serkowski, K. 1974, \aj, 79, 581
1163: \reference{} Diplas, A., \& Savage, B.\ D. 1994, \apjs, 93, 211
1164: \reference{} Draine, B.\ T. 1978, \apjs, 36, 595
1165: \reference{} Draine, B.\ T. 2003, \araa, 41, 241
1166: \reference{} Federman, S.\ R., Danks, A.\ C., \& Lambert, D.\ L. 1984, \apj, 287, 219
1167: \reference{} Fitzpatrick, E.\ L. 1999, \pasp, 111, 63
1168: \reference{} Fitzpatrick, E.\ L. \& Massa, D. 1986, \apj, 307, 286
1169: \reference{} Fitzpatrick, E.\ L. \& Massa, D. 1988, \apj, 328, 734
1170: \reference{} Fitzpatrick, E.\ L. \& Massa, D. 1990, \apjs, 72, 163
1171: \reference{} Fitzpatrick, E.\ L. \& Massa, D. 2005, \aj, 130, 1127
1172: \reference{} Fitzpatrick, E.\ L. \& Massa, D. 2007, \apj, 663, 320
1173: \reference{} Frisch, P.\ C. 1980, \apj, 241, 697
1174: \reference{} Hanson, M.\ M., Snow, T.\ P., \& Black, J.\ H.\ 1992, \apj, 392, 571
1175: \reference{} Hecht, J., Helfer, H.\ L., Wolf, J., Donn, B., \& Pipher, J.\ L. 1982, \apjl, 263, 39
1176: \reference{} Herbig, G.\ H. 1993, \apj, 407, 142
1177: \reference{} Hubert, A.\ M. \& Floquet, M. 1998, \aap, 335, 565
1178: \reference{} Jenniskens, P., \& Greenberg, J.\ M. 1993, \aap, 274, 439
1179: \reference{} Jensen, A.\ G., Rachford, B.\ L., \& Snow, T.\ P.\ 2005, \apj, 619, 891
1180: \reference{} Jensen, A.\ G., Rachford, B.\ L., \& Snow, T.\ P.\ 2007, \apj, 654, 955
1181: \reference{} Kulesa, C.\ A., Hungerford, A.\ L., Walker, C.\ K., Zhang, X., \& Lane, A.\ P. 2005, \apj, 625, 194
1182: \reference{} Larson, K.\ A., Whittet, D.\ C.\ B., \& Hough, J.\ H. 1996, \apj, 472, 755
1183: \reference{} Le Bourlot, J. 2000, \aap, 360, 656
1184: \reference{} Martin, P.\ G., Clayton, G.\ C., \& Wolff, M.\ J. 1999, \apj, 510, 905
1185: \reference{} Martin, P.\ G., \& Whittet, D. G. B. 1990, \apj, 357, 113
1186: \reference{} McDavid, D. 2000, \aj, 119, 352
1187: \reference{} Meyer, D.\ M., Lauroesch, J.\ T., Sofia, U.\ J., Draine, B.\ T., \& Bertoldi, F.\ 2001, \apj, 553, L59
1188: \reference{} Orsatti, A.\ M., Vega, E.\ I., \& Marraco, H.\ G. 2000, \aaps, 144, 195
1189: \reference{} Rachford, B.\ L. et al. 2001, \apj, 555, 839
1190: \reference{} Rachford, B.\ L. et al. 2002, \apj, 577, 221 (Paper I)
1191: \reference{} Roche, P. et al. 1997, \aap, 322, 139
1192: \reference{} Sasseen, T.\ P., Hurwitz, M., Dixon, W.\ V., \& Airieau, S. 2002, \apj, 566, 267
1193: \reference{} Savage, B.\ D., Bohlin, R.\ C., Drake, J.\ F., \& Budich, W. 1977, \apj, 216, 291
1194: \reference{} Serkowski, K., Mathewson, D.\ L., \& Ford, V.\ L. 1975, \apj, 196, 261
1195: \reference{} Shaw, G., Ferland, G.\ J., Abel, N.\ P., Stancil, P.\ C., \& van Hoof, P.\ A.\ M. 2005, \apj, 624, 794
1196: \reference{} Shull, J. M., \& Beckwith, S. 1982, \araa, 20, 163
1197: \reference{} Shull, J.\ M. \& van Steenberg, M.\ E. 1985, \apj, 294, 599
1198: \reference{} Skrutskie, M.\ F.\ et al. 2006, \aj, 131, 1163
1199: \reference{} Snow, T.\ P. 1983, \apj, 269, L57
1200: \reference{} Snow, T.\ P., Destree, J.\ D., \& Welty, D.\ W. 2008, \apj, 679, 512
1201: \reference{} Snow, T.\ P. \& McCall, B.\ J. 2006, \araa, 44, 367
1202: \reference{} Snow, T.\ P., Ross, T.\ L., Destree, J. D., Drosback, M.\ M., Jensen, A.\ G., Rachford, B.\ L., Sonnentrucker, P., Ferlet, R. 2008, \apj, in press
1203: \reference{} Snow, T.\ P. \& Witt, A. 1989, \apj, 342, 295
1204: \reference{} Sonnentrucker, P., Friedman, S.\ D., Welty, D.\ E., York, D.\ G., \& Snow, T.\ P. 2003, \apj, 596, 350
1205: \reference{} Thackeray, A.\ D. 1950, \mnras, 110, 343
1206: \reference{} van Leeuwen, F. 2007, Hipparcos, the New Reduction of the Raw Data, Springer
1207: \reference{} Welty, D.\ W., Federman, S.\ R., Gredel, R., Thorburn, J.\ A., \& Lambert, D.\ L. 2006, \apjs, 165, 138
1208: \reference{} Whittet, D.\ C.\ B. 1974, \mnras, 168, 371
1209: \reference{} Whittet, D.\ C.\ B., Gerakines, P. A., Carknet, A. L., Hough, J. H., Martin, P. G., Prusti, T., Kilkenny, D. 1994, \mnras, 268, 1
1210: \reference{} Whittet, D.\ C.\ B., \& van Breda, I. G. 1978, \aap, 66, 57
1211: \reference{} Witt, A.\ N., Bohlin, R.\ C., \& Strecher, T.\ P. 1986, \apj, 305, L23
1212: \reference{} Woolf, N.\ J. 1961, \pasp, 73, 206
1213: \end{references}
1214:
1215:
1216:
1217: \clearpage
1218:
1219: \begin{deluxetable}{ccccclc}
1220: \tablecaption{Target list}
1221: \tablewidth{0pt}
1222: \tablehead{
1223: \colhead{Star} & \colhead{$\ell$} & \colhead{$b$} & \colhead{Assoc.}
1224: & \colhead{$V$} & \colhead{MK type}
1225: }
1226: \startdata
1227: HD \phn37903 &206.85 & $-$16.54 &Orion &\phn7.83 &B1.5 V \\
1228: HD \phn38087 &207.07 & $-$16.26 &Orion? &\phn8.30 &B5 V \\
1229: HD \phn40893 &180.09 &\phn$+$4.34 & &\phn8.90 &B0 IV \\
1230: HD \phn41117 &189.65 &\phn$-$0.86 &Gem OB1 &\phn4.63 &B2 Iae \\
1231: HD \phn42087 &187.79 &\phn$+$1.77 &Gem OB1 &\phn5.75 &B2.5Ibe \\
1232: HD \phn43384 &187.99 &\phn$+$3.53 &Gem OB1 &\phn6.25 &B3 Ib \\
1233: HD \phn46056 &206.34 &\phn$-$2.25 &Mon OB2 &\phn8.16 &O8 V \\
1234: HD \phn46202 &206.31 &\phn$-$2.00 &Mon OB2 &\phn8.19 &O9 V \\
1235: HD \phn53367 &223.71 &\phn$-$1.90 & &\phn6.96 &B0 IVe \\
1236: HD 147888 &353.65 & $+$17.71 &Sco-Oph &\phn6.74 &B5 V \\
1237: HD 149404 &340.54 &\phn$+$3.01 & &\phn5.47 &O9 Iae \\
1238: HD 152236 &343.03 &\phn$+$0.87 &Sco OB1 &\phn4.73 &B1 Ia+pe \\
1239: HD 164740 &\phn\phn5.97 &\phn$-$1.17 &M8 &10.30 &O7.5 V \\
1240: HD 179406 &\phn28.23 &\phn$-$8.31 & &\phn5.34 &B3 V \\
1241: HD 186994 &\phn78.62 & $+$10.86 & &\phn7.50 &B0 III \\
1242: \enddata
1243: \end{deluxetable}
1244:
1245:
1246: \begin{deluxetable}{cccccccccccc}
1247: \tablecaption{Extinction parameters for Paper I sample and
1248: present sample\tablenotemark{a}}
1249: \tabletypesize{\footnotesize}
1250: \tablewidth{0pt}
1251: \tablehead{
1252: & & \multicolumn{4}{c}{$R_V$} \\
1253: \cline{3-7} \\
1254: \colhead{Star} & \colhead{$E(B-V)$}
1255: & \colhead{Phot.\tablenotemark{b}}
1256: & \colhead{Polar.\tablenotemark{c}} & \colhead{Ref}
1257: & \colhead{E.C.\tablenotemark{d}}
1258: & \colhead{Adopted} & \colhead{$A_V$}
1259: }
1260: \startdata
1261: BD $+$31$^{\circ}$ 643 &0.85 &3.13$\pm$0.30 &3.75 &1 &3.54 &3.13$\pm$0.30 &2.66$\pm$0.26 \\
1262: HD \phn24534 &0.59 &1.70$\pm$1.00 &3.47 &2 & &3.47$\pm$0.30 &2.05$\pm$0.18\\
1263: HD \phn27778 &0.37 &2.72$\pm$0.30 & & & &2.72$\pm$0.30 &1.01$\pm$0.11\\
1264: HD \phn62542 &0.35 &2.83$\pm$0.30 &3.27 &3 &2.14 &2.83$\pm$0.30 &0.99$\pm$0.14\\
1265: HD \phn73882 &0.70 &3.37$\pm$0.30 &3.51 &3 &2.93 &3.37$\pm$0.30 &2.36$\pm$0.23\\
1266: HD \phn96675 &0.30 &3.85$\pm$0.30 &2.80 &4 &4.02 &3.85$\pm$0.30 &1.16$\pm$0.15\\
1267: HD 102065 &0.17 & & & &2.89 &2.89$\pm$0.30 &0.49$\pm$0.10\\
1268: HD 108927 &0.22 &3.14$\pm$0.30 & & &3.73 &3.14$\pm$0.30 &0.69$\pm$0.11\\
1269: HD 110432 &0.51 &3.95$\pm$0.60 &3.30 &5 & &3.95$\pm$0.60 &2.02$\pm$0.33\\
1270: HD 154368 &0.78 &3.00$\pm$0.30 & & &3.14 &3.00$\pm$0.30 &2.34$\pm$0.25\\
1271: HD 167971 &1.08 &3.17$\pm$1.00 & & & &3.17$\pm$1.00 &3.42$\pm$1.08\\
1272: HD 168076 &0.78 &3.55$\pm$0.30 &3.19 &6 &3.62 &3.55$\pm$0.30 &2.77$\pm$0.26\\
1273: HD 170740 &0.48 &2.71$\pm$0.30 &3.08 &5 & &2.71$\pm$0.30 &1.30$\pm$0.17\\
1274: HD 185418 &0.50 &2.32$\pm$0.30 & & &3.98 &2.32$\pm$0.30 &1.16$\pm$0.17\\
1275: HD 192639 &0.66 &2.84$\pm$1.00 & & & &2.84$\pm$1.00 &1.87$\pm$0.67\\
1276: HD 197512 &0.32 &2.35$\pm$0.30 & & &2.56 &2.35$\pm$0.30 &0.75$\pm$0.12\\
1277: HD 199579 &0.37 &2.95$\pm$1.00 & & &2.74 &2.95$\pm$0.30 &1.09$\pm$0.14\\
1278: HD 203938 &0.74 &2.91$\pm$0.30 & & &3.00 &2.91$\pm$0.30 &2.15$\pm$0.24\\
1279: HD 206267 &0.53 &2.67$\pm$0.30 & & & &2.67$\pm$0.30 &1.41$\pm$0.18\\
1280: HD 207198 &0.62 &2.42$\pm$1.00 &2.30 &7 &2.66 &2.42$\pm$0.30 &1.50$\pm$0.20\\
1281: HD 207538 &0.64 &2.25$\pm$0.30 &2.23 &3 & &2.25$\pm$0.30 &1.44$\pm$0.20\\
1282: HD 210121 &0.40 &2.08$\pm$0.30 &2.13 &8 &2.01 &2.08$\pm$0.30 &0.83$\pm$0.14\\
1283: HD 210839 &0.57 &2.78$\pm$0.30 &2.86 &9 & &2.78$\pm$0.30 &1.58$\pm$0.19\\
1284: \\
1285: HD \phn37903 &0.35 &3.67$\pm$0.30 &3.89 &3 &3.90 &3.67$\pm$0.30 &1.28$\pm$0.15\\
1286: HD \phn38087 &0.29 &5.57$\pm$0.30 &3.06 &3 &4.48 &5.57$\pm$0.30 &1.61$\pm$0.19\\
1287: HD \phn40893 &0.46 &2.46$\pm$0.30 & & &3.18 &2.46$\pm$0.30 &1.13$\pm$0.16\\
1288: HD \phn41117 &0.45 &2.74$\pm$1.00 &3.02 &5 &2.89 &2.74$\pm$0.30 &1.23$\pm$0.16\\
1289: HD \phn42087 &0.36 &3.06$\pm$1.00 &3.08 &5 &2.78 &3.08$\pm$0.30 &1.10$\pm$0.14\\
1290: HD \phn43384 &0.58 &3.06$\pm$0.30 &2.97 &5 & &3.06$\pm$0.30 &1.78$\pm$0.20\\
1291: HD \phn46056 &0.50 &2.60$\pm$0.30 & & &2.81 &2.60$\pm$0.30 &1.30$\pm$0.17\\
1292: HD \phn46202 &0.49 &2.83$\pm$0.30 & & &2.79 &2.83$\pm$0.30 &1.39$\pm$0.17\\
1293: HD \phn53367 &0.74 &2.38$\pm$1.00 & & & &2.38$\pm$1.00 &1.76$\pm$0.74\\
1294: HD 147888 &0.47 &4.06$\pm$0.30 &3.82 &3 &4.93 &4.06$\pm$0.30 &1.91$\pm$0.19\\
1295: HD 149404 &0.68 &3.28$\pm$0.60 &3.08 &5 & &3.28$\pm$0.60 &2.23$\pm$0.42\\
1296: HD 152236 &0.68 &3.29$\pm$1.00 &3.25 &5 &2.82 &3.29$\pm$0.30 &2.24$\pm$0.23\\
1297: HD 164740 &0.87 &5.36$\pm$0.30 &3.75 &5 & &5.36$\pm$0.30 &4.66$\pm$0.31\\
1298: HD 179406 &0.33 &2.86$\pm$0.30 &2.86 &5 & &2.86$\pm$0.30 &0.94$\pm$0.13\\
1299: HD 186994 &0.17 & & & & &3.10$\pm$0.30 &0.53$\pm$0.13\\
1300: \enddata
1301: \tablenotetext{a}{Revised values for Paper I targets are given
1302: first, followed by the values for the new targets}
1303: \tablenotetext{b}{Derived using 2MASS photometry}
1304: \tablenotetext{c}{Derived from the wavelength of maximum polarization}
1305: \tablenotetext{d}{Derived from the linear far-UV rise in the
1306: extinction curve based on the $c_2$ parameters given in Table 3.}
1307: \tablerefs{(1) Andersson \& Wannier 2000;
1308: (2) Roche et al. 1997; (3) Martin, Clayton, \& Wolff 1999;
1309: (4) Whittet et al. 1994; (5) Serkowski, Mathewson, Ford 1975;
1310: (6) Orsatti, Vega, \& Marraco 2000; (7) Anderson et al. 1996;
1311: (8) Larson, Whittet, \& Hough 1996; (9) McDavid 2000
1312: }
1313: \end{deluxetable}
1314:
1315:
1316: \begin{deluxetable}{cccccccc}
1317: \tablecaption{Extinction curve parameters\tablenotemark{a}}
1318: \tablewidth{0pt}
1319: %\tabletypesize{\footnotesize}
1320: \tablehead{
1321: \colhead{Target} & \colhead{$\lambda_0^{-1}$} & \colhead{$\gamma$}
1322: & \colhead{c$_1$} & \colhead{c$_2$} & \colhead{c$_3$} & \colhead{c$_4$}
1323: & \colhead{Ref.} \\
1324: & \colhead{($\mu$m$^{-1}$)} & \colhead{($\mu$m$^{-1}$)}
1325: }
1326: \startdata
1327: HD \phn37903 &4.615 &1.045 &\phm{$-$}0.965 &0.384 &3.300 &0.440 &1 \\
1328: HD \phn38087 &4.563 &1.026 &\phm{$-$}1.137 &0.230 &4.508 &0.311 &1 \\
1329: HD \phn40893 &4.591 &0.83\phn &\phm{$-$}0.26\phn &0.66\phn &3.13\phn &0.55\phn &2 \\
1330: HD \phn41117 &4.621 &0.97\phn &$-$0.38\phn &0.81\phn &3.64\phn &0.56\phn &2 \\
1331: HD \phn42087 &4.636 &1.05\phn &$-$1.28\phn &0.87\phn &4.36\phn &0.53\phn &2 \\
1332: HD \phn43384 & \\
1333: HD \phn46056 &4.611 &0.932 &$-$0.527 &0.857 &3.032 &0.541 &1 \\
1334: HD \phn46202 &4.599 &0.842 &$-$0.348 &0.864 &2.542 &0.515 &1 \\
1335: HD \phn53367 & \\
1336: HD 147888 &4.587 &1.022 &\phm{$-$}1.611 &0.133 &3.823 &0.339 &1 \\
1337: HD 149404 & \\
1338: HD 152236 &4.622 &1.06\phn &$-$0.51\phn &0.85\phn &3.71\phn &0.38\phn &2 \\
1339: HD 164740 & \\
1340: HD 179406 & \\
1341: HD 186994 & \\
1342:
1343: \enddata
1344: \tablenotetext{a}{In the parameterization scheme of Fitzpatrick \& Massa 1990}
1345: \tablerefs{(1) Fitzpatrick \& Massa 1990; (2) Jenniskens \& Greenberg 1993}
1346: \end{deluxetable}
1347:
1348: \begin{deluxetable}{cclccc}
1349: \tablecaption{{\it FUSE} observations}
1350: \tablewidth{0pt}
1351: %\tabletypesize{\footnotesize}
1352: \tablehead{
1353: \colhead{Target} & \colhead{{\it FUSE} data ID} & \colhead{Date} &
1354: \colhead{N$_{\rm int}$\tablenotemark{a}} &
1355: \colhead{t$_{\rm int}$\tablenotemark{b}} &
1356: \colhead{S/N\tablenotemark{c}} \\
1357: & & & \colhead{(ksec)} &
1358: }
1359: \startdata
1360: HD \phn37903 &P1160601 &2001 Oct 18 &\phn4 &\phn4.0 & 20.4 \\
1361: HD \phn38087 &P1160701 &2001 Oct 18 &\phn4 &\phn4.0 & 16.2 \\
1362: HD \phn40893 &P2160101 &2001 Oct 14 &\phn3 &\phn7.1 & 12.7 \\
1363: HD \phn41117 &P2160201 &2004 Mar 03 &\phn1 &\phn0.1 & \phn2.0 \\
1364: HD \phn42087 &P2160301 &2001 Oct 15 &\phn6 &\phn2.9 & 12.1 \\
1365: HD \phn43384 &P2160401 &2001 Oct 15 &\phn6 &\phn8.1 & \phn1.8 \\
1366: HD \phn46056 &P2160901 &2003 Jan 25 &\phn3 &\phn7.1 & \phn8.1 \\
1367: HD \phn46202 &P2161001 &2001 Oct 16 &\phn2 &\phn5.0 & 10.0 \\
1368: HD \phn53367 &P1161101 &2001 Oct 26 &\phn7 &11.4 & \phn9.9 \\
1369: HD 147888 &P1161501 &2003 Aug 21 &23 &12.2 & 12.8 \\
1370: HD 149404 &P1161701 &2001 Aug 07 &38 &17.9 & 28.9 \\
1371: HD 152236 &P1161801 &2001 Aug 08 &\phn4 &\phn4.6 & \phn6.1 \\
1372: HD 164740 &P1162001 &2000 Aug 30 &\phn3 &\phn5.9 & \phn9.0 \\
1373: HD 179406 &P2160701 &2001 Apr 27 &\phn1 &\phn0.1 & \phn3.7 \\
1374: HD 179406 &P2160702 &2002 Jun 11 &\phn3 &\phn1.0 & 14.4 \\
1375: HD 186994 &P2160801 &2001 Jul 02 &\phn1 &\phn0.1 & \phn6.3 \\
1376: HD 186994 &P2160802 &2001 Sep 07 &\phn1 &\phn0.4 & 17.2 \\
1377: \enddata
1378: \tablenotetext{a}{Number of integrations}
1379: \tablenotetext{b}{Total integration time}
1380: \tablenotetext{c}{Average per-pixel S/N for a 1 \AA\ region of the LiF 1A
1381: spectrum near 1070 \AA , between the Lyman (3,0) and (2,0) bandheads of H$_2$.
1382: One resolution element corresponds to about 9 pixels.}
1383: \end{deluxetable}
1384:
1385:
1386: \begin{deluxetable}{cccccccccc}
1387: \tablecaption{Molecular and atomic hydrogen parameters}
1388: \tabletypesize{\footnotesize}
1389: \tablewidth{0pt}
1390: \tablehead{
1391: \colhead{Target} & \colhead{Bands} & \colhead{log $N$(H$_2$)}
1392: & \colhead{log $N$(0)}
1393: & \colhead{log $N$(1)} & \colhead{T$_{01}$} & \colhead{log $N$(H I)}
1394: & \colhead{Ref} & \colhead{$f_{\rm H2}$} \\
1395: & & \colhead{($N$ in cm$^{-2}$)} & \colhead{($N$ in cm$^{-2}$)} & \colhead{($N$ in cm$^{-2}$)}
1396: & \colhead{(K)} & \colhead{($N$ in cm$^{-2}$)} & &
1397: }
1398: \startdata
1399: HD \phn37903 &\phn6 &20.92$\pm$0.06 &20.68$\pm$0.07 &20.54$\pm$0.05 &\phn68$\pm$\phn7 &21.17$\pm$0.10 &1 &0.53$\pm$0.09 \\
1400: HD \phn38087 &\phn7 &20.64$\pm$0.07 &20.39$\pm$0.08 &20.29$\pm$0.05 &\phn70$\pm$\phn8 &20.91$\pm$0.30 &2 &0.52$\pm$0.20 \\
1401: HD \phn40893 &\phn9 &20.58$\pm$0.05 &20.27$\pm$0.05 &20.28$\pm$0.05 &\phn78$\pm$\phn8 &21.50$\pm$0.10 &3 &0.19$\pm$0.06 \\
1402: HD \phn41117 &\phn2 &20.69$\pm$0.10 &20.51$\pm$0.10 &20.22$\pm$0.10 &\phn59$\pm$\phn8 &21.40$\pm$0.15 &4 &0.28$\pm$0.13 \\
1403: HD \phn42087 &\phn7 &20.52$\pm$0.12 &20.31$\pm$0.12 &20.11$\pm$0.12 &\phn64$\pm$11 &21.39$\pm$0.11 &1 &0.21$\pm$0.10 \\
1404: HD \phn43384 &\phn2 &20.87$\pm$0.14 &20.59$\pm$0.10 &20.54$\pm$0.18 &\phn74$\pm$16 &21.27$\pm$0.30 &5 &0.44$\pm$0.24 \\
1405: HD \phn46056 &\phn9 &20.68$\pm$0.06 &20.40$\pm$0.06 &20.35$\pm$0.06 &\phn74$\pm$\phn8 &21.38$\pm$0.14 &1 &0.29$\pm$0.09 \\
1406: HD \phn46202 &\phn9 &20.68$\pm$0.06 &20.38$\pm$0.07 &20.38$\pm$0.07 &\phn78$\pm$\phn9 &21.58$\pm$0.15 &1 &0.20$\pm$0.09 \\
1407: HD \phn53367 &\phn9 &21.04$\pm$0.05 &20.89$\pm$0.04 &20.52$\pm$0.07 &\phn56$\pm$\phn4 &21.32$\pm$0.30 &2 &0.51$\pm$0.19 \\
1408: HD 147888 &\phn7 &20.47$\pm$0.05 &20.39$\pm$0.04 &19.71$\pm$0.10 &\phn45$\pm$\phn4 &21.44$\pm$0.30 &5 &0.18$\pm$0.14 \\
1409: HD 149404 &\phn9 &20.79$\pm$0.04 &20.60$\pm$0.03 &20.34$\pm$0.05 &\phn61$\pm$\phn4 &21.40$\pm$0.15 &1 &0.33$\pm$0.09 \\
1410: HD 152236 &\phn1 &20.73$\pm$0.12 &20.53$\pm$0.12 &20.29$\pm$0.12 &\phn62$\pm$10 &21.77$\pm$0.15 &1 &0.15$\pm$0.08 \\
1411: HD 164740 &\phn1 &20.19$\pm$0.12 &19.92$\pm$0.12 &19.86$\pm$0.12 &\phn60$\pm$10 &21.95$\pm$0.15 &6 &0.03$\pm$0.02 \\
1412: HD 179406 &\phn4 &20.73$\pm$0.07 &20.55$\pm$0.07 &20.26$\pm$0.08 &\phn59$\pm$\phn6 &21.23$\pm$0.15 &7 &0.39$\pm$0.12 \\
1413: HD 186994 &\phn9 &19.59$\pm$0.04 &19.18$\pm$0.06 &19.37$\pm$0.03 &\phn97$\pm$10 &20.90$\pm$0.15 &8 &0.09$\pm$0.04 \\
1414:
1415: \enddata
1416: \tablerefs{(1) Diplas \& Savage 1994, Ly$\alpha$;
1417: (2) Jensen et al. 2007, $N$(H I) = 5.8$\times$10$^{21}$$E(B-V)$ -- 2$N$(H$_2)$;
1418: (3) Jensen et al. 2007, Ly$\alpha$;
1419: (4) Present work, Ly$\alpha$;
1420: (5) Present work, $N$(H I) = 5.8$\times$10$^{21}$$E(B-V)$ -- 2$N$(H$_2)$;
1421: (6) Fitzpatrick \& Massa 1990, Ly$\alpha$;
1422: (7) Hansen et al. 1992, Ly$\alpha$;
1423: (8) Bohlin et al. 1978, Ly$\alpha$}
1424: \end{deluxetable}
1425:
1426:
1427: \clearpage
1428: \begin{figure}
1429: \plotone{f1.eps}
1430: \caption{Molecular hydrogen column density versus color excess.
1431: Crosses: {\it FUSE}; diamonds: {\it Copernicus}. Typical H$_2$
1432: uncertainties for the {\it Copernicus} data points ($\sim$25\%)
1433: are slightly larger than for the {\it FUSE} data points.}
1434: \end{figure}
1435:
1436: \clearpage
1437: \begin{figure}
1438: \epsscale{.95}
1439: \plotone{f2.eps}
1440: \caption{Total hydrogen column density versus color excess.
1441: Symbols as in Figure 1. Top panel: All data points. Bottom
1442: panel: Only {\it FUSE} points with reliable H I and H$_2$
1443: measurements toward non-Be stars. Solid line in both panels is
1444: the best fit to the data in the lower panel, constrained to pass
1445: through the origin. The dashed line is an unconstrained fit
1446: that does not include the point at $E(B-V)$ = 0.17 (HD 186994).
1447: Typical $N(H_{\rm tot})$ errors for the {\it Copernicus}
1448: data points ($\sim$30\%) are slightly larger than for the {\it
1449: FUSE} data points.}
1450: \end{figure}
1451:
1452: \clearpage
1453: \begin{figure}
1454: \plotone{f3.eps}
1455: \caption{Total hydrogen column density versus total visual
1456: extinction. Symbols as in Figure 1. Top panel: All data
1457: points. Bottom panel: Only {\it FUSE} points with reliable H I
1458: and H$_2$ measurements toward non-Be stars. The solid line in
1459: both panels is the best fit to the data in the lower panel,
1460: constrained to pass through the origin. The dashed line is an
1461: unconstrained fit that does not include the point at $A_V$ =
1462: 0.53 (HD 186994). Typical $N(H_{\rm tot})$ errors for the {\it
1463: Copernicus} data points ($\sim$30\%) are slightly larger than
1464: for the {\it FUSE} data points.}
1465: \end{figure}
1466:
1467: \clearpage
1468: \begin{figure}
1469: \epsscale{1}
1470: \plotone{f4.eps}
1471: \caption{Rotational temperature versus color excess. Crosses:
1472: {\it FUSE}; diamonds: {\it Copernicus} points with $N$(H$_2$) $>$
1473: 10$^{20}$ cm$^{-2}$; squares: {\it Copernicus} points with
1474: $N$(H$_2$) $<$ 10$^{20}$ cm$^{-2}$.}
1475: \end{figure}
1476:
1477: \clearpage
1478: \begin{figure}
1479: \plotone{f5.eps}
1480: \caption{Molecular fraction versus color excess. Crosses: {\it
1481: FUSE}; diamonds: {\it Copernicus} points with
1482: $N$(H$_2$) $>$ 10$^{20}$ cm$^{-2}$; squares: {\it Copernicus}
1483: points with $N$(H$_2$) $<$ 10$^{20}$ cm$^{-2}$. Typical $f_{\rm
1484: H2}$ errors for {\it Copernicus} data points ($\sim$30\%) are
1485: slightly larger than for the {\it FUSE} data points. Error
1486: bars are not given for {\it FUSE} values derived without direct
1487: measurements of $N$(H I).}
1488: \end{figure}
1489:
1490: \clearpage
1491: \begin{figure}
1492: \plotone{f6.eps}
1493: \caption{Molecular fraction versus total-to-selective extinction
1494: ratio. Symbols and comments as in Figure 5. To make the plot
1495: easier to read, error bars for $R_V$ are only given for the high
1496: $R_V$ sample discussed in \S\ 4.3. We also give error bars for
1497: $f_{\rm H2}$ for the two {\it Copernicus} targets discussed in
1498: \S\ 4.3.}
1499: \end{figure}
1500:
1501: \clearpage
1502: \begin{figure}
1503: \plotone{f7.eps}
1504: \caption{Molecular fraction versus rotational temperature. Symbols
1505: and comments as in Figure 5}
1506: \end{figure}
1507:
1508:
1509: \end{document}
1510:
1511: