Information about http://fuse.pha.jhu.edu/aas207/AAS_207_175_03_Linsky.pdf
…
Tags: deuterium depletion, dust grains, fuse satellite, galactic disk, galactic halo, gas phase, imaps, intermediate regime, interstellar gas, jhu, jila, lacour, linsky, nasa gsfc, northwestern univ, nuclear processes, parsecs, princeton univ, sonneborn, spatial variations,Pages: 1
Language: english
Created: Fri Jan 27 15:53:20 2006
Display cached document
Page 1
Galactic Disk?
J.L. Linsky (JILA/U.Colo), B.T. Draine (Princeton Univ.), H.W. Moos (JHU), E.B. Jenkins (Princeton Univ.), B.E. Wood (JILA/U. Colo.), C. Oliveira (JHU),
W.P. Blair (JHU), S.D. Friedman (STScI), C. Gry (LAS/Marseille), D. Knauth (Northwestern Univ.), J.W. Kruk (JHU), S. Lacour (Meudon), N. Lehner (U.
Wisc.), S. Redfield (U. Texas), J.M Shull (CASA/U. Colo.), G. Sonneborn (NASA/GSFC), G.M. Williger (JHU)
Abstract
Analyses of spectra obtained with the FUSE satellite, together with spectra from the Copernicus and IMAPS instruments,
reveal a very wide range in the observed deuterium/hydrogen (D/H) ratios for interstellar gas in the Galactic disk beyond
the Local Bubble. For gas located beyond the Local Bubble but within several hundred parsecs, the observed D/H ratios
differ by a factor of 45, which is difficult to explain solely on the basis of either: (i) small-scale spatial variations in stellar
nuclear processes that convert deuterium to heavier elements; or (ii) the infall of deuterium-rich gas from the Galactic halo
and the IGM. We argue instead that spatial variations in the depletion of deuterium onto dust grains can explain these local
variations in the observed gas-phase D/H ratios. We present a deuterium depletion model that naturally explains the constant
measured values of D/H inside the Local Bubble, the wide range of gas-phase D/H ratios observed in the intermediate regime
(log N(H I) = 19.220.7), and the low gas-phase D/H ratios observed at larger hydrogen column densities. We test
the deuterium depletion hypothesis by: (i) correlations of gas-phase D/H ratios with depletions of the refractory metals iron
and silicon, and (ii) correlation with HD in heavily reddened lines of sight. Both of these tests are consistent with deuterium
depletion from the gas phase in cold, not recently shocked, regions of the ISM, and high gas-phase D/H ratios in gas that has
been shocked or otherwise heated recently. We argue that the total (gas plus dust) D/H ratio within 1 kpc of the Sun has a
much larger value than D/H in the gas phase in the Local Bubble, indicating that over the lifetime of the Galaxy there has been
a relative small decrease in the total D/H ratio from its primordial value.
1 Introduction ppm for the Car LOS (Copernicus) to · In thermodynamic equilibrium
21.8+2.2 ppm observed for the 2 Vel LOS (D/H)dust 5 × 104 for
Observations of the D/H ratio in the Galac- -1.9
(IMAPS). This wide range in (D/H)gas values Tgas 90 K. So grains will
tic disk ISM provide an important constraint
is seen in observations by the Copernicus, gradually deplete the ISM gas phase
on models of Galactic chemical evolution, as
IMAPS, and FUSE satellites and is unlikely an of D until a shock destroys the grains
deuterium is thought to be formed only in the
instrumental effect. and returns the D to the ISM. A
Big Bang and is converted to 3 He and 4 He strong UV radiation field may do the
in stars. Over time, supernovae and stellar same.
winds pollute the ISM with deuterium-depleted For the longest lines of sight in the Galaxy
but metal-rich gas. The disk is also enriched studied so far, there is a pattern of D/H being
a factor of 2 below the Local Bubble value. · The time scale for deuterium deple-
by the infall of deuterium-rich but metal-poor
For log N(H I) 20.5, the five FUSE data tion is a few million years in the cold
gas from the IGM and Galactic halo.
points are clumped at (D/H)gas = 8.6 ± 0.8 neutral medium.
Because of its simple origin and evolu- F IGURE 3: (D/H)gas vs. Fe depletion for sightlines with
tion, D plays a key role in testing models of ppm. The 3 lines of sight (1 IMAPS and 2 log N(H I) 19.0. Removal of the shorter lines of sight
Galactic chemical evolution. Measurements Copernicus) in the range log N(H I) = 20.2 · Very large D/H ratios have been minimizes possible corrections for hydrogen partial ion-
of D/H have been obtained from the analy- 20.5 are also very low. Five lines of sight measured in interplanetary dust ization. The dashed lines is the fit in Figure 2.
sis of high resolution spectra of Lyman- and indicate a trend. grains that were likely formed in the
higher Lyman lines observed by the Coperni- ISM. This is an important "proof of
cus, IUE, HST/GHRS, HST/STIS, IMAPS, and All Galactic disk lines of sight ob- concept" for deuterium depletion.
now FUSE satellites. The Lyman lines show served so far have (D/H)gas smaller than
strong absorption by interstellar H I and D I (D/H)prim = 27.5+2.4 ppm. Most Galactic
-1.9
(81 km s-1 from the H I line). lines of sight have (D/H)gas values smaller 4 Our Hypothesis
There is now agreement that the D/H ra- than 22 ± 7 ppm reported by Sembach et al.
tio for interstellar gas within the Local Bub- (ApJS, 150, 387, 2004) for the PG 1259+593 We assume that the total D/H (gas and
ble, extending to 100 pc from the Sun (Complex C) line of sight. Complex C is a dust) is relatively constant in the Local Disk
or to log N(H I) < 19.2, is essentially con- rapidly infalling cloud of gas with 0.10.4 Z (within 1 kpc), but (D/H)gas varies because of
stant, (D/H)gas = 15.6 ± 0.4 parts per mil- at 510 kpc distance. time- and spatially-dependent D depletion.
lion (ppm), where 0.4 ppm is the standard de- Regions that have not been shocked in a
viation of the mean (Wood et al. ApJ, 609, What does this mean for Galactic chem- long time will have low values of (D/H)gas . F IGURE 4: Gas phase D/H vs. the depletion of Fe us-
838, 2004). Recent Galactic chemical evolu- ical evolution? Previous studies have as- The (J=0,1) rotational excitation temperature ing data only from STIS, GHRS, and FUSE. This plot
of H2 toward JL 9 is 89 ± 6 K and toward deletes the more uncertain Copernicus data.
tion models (e.g., Romano et al. MNRAS, 346, sumed two possible explanations for the large
295, 2003) have used similar values of D/H range of (D/H)gas values, but both explana- LSS 1274 is 64 ± 5 K. These are two of the
for comparison with their calculations. They tions have problems and we argue for a differ- five lines of sight with large N(H I).
compute an astration factor 1.5 for the ra- ent explanation. Regions that have recently been shocked
tio of primordial to present day D/H in the so- Variable Astration: The observed factor of will have high values of (D/H)gas which should
lar neighborhood. This astration factor is con- 45 range in (D/H)gas beyond the Local Bub- be close to the total value for D/H in the Lo-
sistent with the WMAP data (Spergel et al. ble is difficult to explain given nuclear reaction cal Disk, i.e., (D/H)LD . Note that (D/H)gas is
ApJS, 148, 175, 2003) and the Cyburt et al. rates, stellar evolution, and interstellar mixing high for the Cru and 2 Vel lines of sight.
(Phys. Lett. B, 567, 227, 2003) nuclear reac- timescales. Why should 5 lines of sight at These stars are in young star forming regions
tion rates which place the primordial value of large N(H I) show the same low values of D/H that were likely shocked recently.
(D/H)prim = 27.5+2.4 ppm. Recent ob-
-1.9
over a wide range of Galactic longitude, while The Local Bubble was recently shocked (1
shorter lines of sight show large variations? 2 million years ago), so (D/H)LBgas is close
servations with FUSE call this agreement
Variable Infall of primordial gas: Why to but somewhat lower than (D/H)LB .
into question.
should there be large variations in (D/H)gas Long lines of sight pass through many F IGURE 5: Gas phase N(D I)/N(Fe II) vs. the H I col-
over short (< 100 pc) distance scales? Why different shocked and nonshocked regions. umn density. In this plot N(H I) is used in only one axis
would the infalling gas be precisely placed in Since cold (not recently shocked) regions to avoid spurious correlations. The dominant stages of
ionization of D and Fe are clearly correlated.
2 An emerging pattern for the local disk and then not well mixed at larger usually have much larger N(H I) than the
distances? Note that abundances in Complex warm, recently shocked gas, the average
(D/H)gas beyond the Local C vary between 0.1 and 0.4 Z over distance value of (D/H)gas for lines of sight with
Bubble scales of 100 pc (Collins, Shull, and Giroux large N(H I) should be systematically low. 6 Implications for Galac-
ApJ, 585, 336 (2003)).
tic Chemical Evolution
(GCE) Models
5 Evidence for Deuterium
3 Why Variable Depletion? Depletion (1) We propose that the most likely value of
(D/H)LD (total for gas and grains in the local
Draine (paper presented at the Carnegie If deuterium is depleted onto dust grains, disk) is 21.9 ± 2.7 ppm. This is the mean
Observatories symposium on "The Origin and then (D/H)gas should be correlated with the value of the four lines of sight with the highest
Evolution of the Elements") presented a de- depletion of metals that are depleted onto (D/H)gas . Since even these lines of sight may
tailed argument for D depletion on to grains: grains. We find good correlations using both have some deuterium in grains, (D/H)LD is a
iron and silicon, but here demonstrate the ef- lower limit.
· Consider dust grains containing 200 fect with iron using different data samples. (2) The primordial D/H estimated from anal-
ppm of C relative to H (typical for the ysis of the WMAP data is (D/H)prim =
ISM). Solid C would likely be hydro-
27.5+2.4 ppm.
-1.9 The corresponding astra-
genated as PAHs with H/C = 0.25.
F IGURE 1: Gas phase D/H vs. H I column density So 50 ppm of H is in the grains. tion factor is (27.5+2.4 )/(21.9 ± 2.7)
-1.9
1.25± 0.17, well below 1.5 predicted by GCE
Figure 1 summarizes the published D/H · If 20% of the hydrogen in grains were models.
measurements for lines of sight in our Galaxy replaced by D, then Dgas /Htotal = (3) We conclude that GCE models should
now extending to beyond 1 kpc and to log 10 ppm and (D/H)gas would be re- be modified to explain this very low astration
N(H I) = 21.2. We call attention to an emerg- duced by 10 ppm. Could (D/H) in factor and perhaps include higher rates of in-
ing pattern in the data: grains be 104 times higher than in flowing gas from the halo or IGM.
the gas phase? FUSE is a joint mission of NASA, the Cana-
For lines of sight (LOS) extending beyond dian Space Agency, the French space agency
the Local Bubble, D/H has a wide range as · The difference in binding energies of CNES, the Johns Hopkins University, the Uni-
F IGURE 2: Gas phase D/H vs. the depletion of Fe using
seen toward white dwarfs, subdwarf OB stars, C-D vs C-H is 0.083 eV. Therefore, it versity of Colorado at Boulder, and the Univer-
all 36 lines of sight. The solid line is a weighted least-
and main sequence OB stars. Beyond the is energetically favorable for D atoms squares fit to the data. sity of California at Berkeley. We thank NASA
Local Bubble (D/H)gas ranges from 5.0 ± 1.6 to displace H atoms in grains. and the JHU for their support.