1: % ApJ paper about the electron fraction in Barnard 68
2:
3: %\documentclass[manuscript]{aastex}
4: \documentclass[12pt,preprint]{aastex}
5: %\documentclass[preprint2]{aastex}
6:
7: \slugcomment{To appear in the Astrophysical Journal}
8: \shorttitle{The Ionization Fraction of Barnard 68}
9: \shortauthors{Maret and Bergin}
10:
11: \newcommand{\eg}{\emph{e.g.}\ }
12: \newcommand{\ie}{\emph{i.e.}\ }
13:
14: \begin{document}
15:
16: \title{The Ionization Fraction of Barnard 68: Implications for Star
17: and Planet Formation}
18:
19: \author{S\'ebastien Maret\altaffilmark{1} and Edwin A. Bergin\altaffilmark{1}}
20: \affil{Department of Astronomy, University of Michigan, 500 Church
21: Street, Ann Arbor, MI 48109-1042, USA}
22:
23: \begin{abstract}
24: We present a detailed study of the ionization fraction of the
25: Barnard 68 pre-stellar core, using millimeter
26: $\mathrm{H^{13}CO^{+}}$ and $\mathrm{DCO^{+}}$ lines
27: observations. These observations are compared to the predictions of
28: a radiative transfer model coupled to a chemical network that
29: includes depletion on grains and gas phase deuterium
30: fractionation. Together with previous observations and modelling of
31: CO and isotopologues, our $\mathrm{H^{13}CO^{+}}$ and
32: $\mathrm{DCO^{+}}$ observations and modelling allow to place
33: constraints on the metal abundance and the cosmic ionization
34: rate. The $\mathrm{H^{13}CO^{+}}$ emission is well reproduced for
35: metals abundances lower than $3 \times 10^{-9}$ and a standard
36: cosmic ray ionization rate. However, the observations are also
37: consistent with a complete depletion of metals, \ie with cosmic rays
38: as the only source of ionization at visual extinctions greater than
39: a few $A_{v}$. The $\mathrm{DCO^{+}}$ emission is found to be
40: dependent of the ortho to para H$_{2}$ ratio, and indicates a ratio
41: of $\sim 10^{-2}$. The derived ionization fraction is about $5
42: \times 10^{-9}$ with respect to H nuclei, which is about an order of
43: magnitude lower than the one observed in the L1544 core. The
44: corresponding ambipolar diffusion timescale is found to be an order
45: of magnitude larger than the free fall timescale at the center of
46: the core. The inferred metal abundance suggests that magnetically
47: inactive regions (\emph{dead zones}) are present in protostellar
48: disks.
49: \end{abstract}
50:
51: \keywords{astrochemistry -- stars: formation --- ISM: abundances
52: --- ISM: molecules --- ISM: individual (Barnard 68)}
53:
54: \section{Introduction}
55:
56: The ionization fraction (or the electron abundance) plays an important
57: role in the chemistry and the dynamics of prestellar cores. Because of
58: the low temperature, the chemistry is dominated by ion neutral
59: reactions \citep{Herbst73}, and electronic recombination is one of the
60: major destruction pathways for molecular ions. Furthermore, the
61: ionization fraction sets the coupling of the gas with the magnetic
62: field \citep{Shu87}.
63:
64: Several attempts have been made to estimate the electron fraction in
65: dense clouds and prestellar cores
66: \citep{Guelin82,Wooten82,deBoisanger96,Williams98,Caselli98}. These
67: studies rely on measurements of the degree of deuterium fractionation
68: (though the $\mathrm{DCO^{+}}$ over $\mathrm{HCO^{+}}$ abundance ratio
69: for example), which has been found to be roughly inversely
70: proportional to the electron abundance \citep{Langer85}. However, this
71: simple approach has caveats \citep{Caselli02d} as it does not consider
72: line-of-sight variations of the electron fraction. Large density
73: gradients exist in prestellar cores, and therefore one may anticipate
74: similar variations in the electron abundance. In addition, the
75: freeze-out of molecules onto the grain surfaces
76: \citep[\eg][]{Tafalla02,Bergin02a} influence the degree of deuterium
77: fractionation independently of the electron fraction
78: \citep{Caselli98}. Finally, these studies usually consider simple
79: chemical networks that may neglect important ingredients for the
80: electron fraction.
81:
82: In this paper, we study the ionization fraction in the Barnard 68
83: core, using $\mathrm{H^{13}CO^{+}}$ and $\mathrm{DCO^{+}}$ line
84: observations. These observations are interpreted with a chemical
85: network including gas-grain interactions that is coupled to a
86: radiative transfer model. This technique allows us to infer the
87: electron abundance along the line-of-sight, and to place constraints
88: on the abundance of metals, the cosmic ray ionization rate, and the
89: ionization state of material that is provided by infall to the forming
90: proto-planetary disk.
91:
92: The paper is organized as follows: in Section
93: \ref{sec:observations-results}, we present the observations. The model
94: used to interpret these observations is detailed in Section
95: \ref{sec:analysis}. Implications of our findings are discussed in
96: Section \ref{sec:discussion}, and Section \ref{sec:conclusions}
97: concludes this paper.
98:
99: \section{Observations}
100: \label{sec:observations-results}
101:
102: The H$^{13}$CO$^{+}$ (1-0) ($\nu = 86.754288$ GHz), and the DCO$^{+}$
103: (2-1) ($\nu = 144.077319$ GHz) transitions were observed towards B68
104: ($\alpha = 17^\mathrm{h} 22^\mathrm{m} 38.2^\mathrm{s}$ and $\delta =
105: -23 \degr 49 \arcmin 34.0 \arcsec$; J2000) in April 2002 and September
106: 2002 using the IRAM-30m telescope. The core was mapped with a spatial
107: sampling of 12\arcsec. The half power beam size of the telescope is
108: $29\arcsec$ at 87 GHz and $17\arcsec$ at 144 GHz. System temperature
109: were typically $\sim 110-150$ K at 3 mm and $\sim 160-350$ K at 2
110: mm. Pointing was regularly checked using planets and was found to be
111: better than $\sim 2\arcsec$. The data were calibrated in antenna
112: temperature ($T_\mathrm{a}^{*}$) units using the chopper wheel method,
113: and were converted to the main beam temperature scale
114: ($T_\mathrm{mb}$), using the telescope efficiencies from the IRAM
115: website. All observations were carried out in frequency switching
116: mode. The H$^{13}$CO$^{+}$ (1-0) data were also presented in
117: \citet{Maret06}.
118:
119: Fig. \ref{fig1} shows a comparison between the integrated line
120: intensity maps of DCO$^{+}$ (2-1) and H$^{13}$CO$^{+}$ (1-0) with the
121: visual extinction map obtained by \citet{Alves01}. The C$^{18}$O (1-0)
122: map from \citet{Bergin02a} is also shown. On this figure, we see that
123: the peak of H$^{13}$CO$^{+}$ (1-0) line emission does not correspond
124: to the maximum visual extinction in the core\footnote{This is also
125: clearly seen on Fig. \ref{fig2}, which shows the H$^{13}$CO$^{+}$
126: (1-0) line emission as a function of the visual extinction. The line
127: emission increases as a function of the $A_{v}$ between 0 and 20,
128: but decreases at $A_{v} \sim 20$.}. The C$^{18}$O (1-0) line
129: emission shows a similar behavior: it peaks in a shell-like structure
130: with a radius of $\sim 50\arcsec$ around the maximum visual
131: extinction. The DCO$^{+}$ (2-1) line emission, on the other hand,
132: seems to correlate well with the visual extinction. These differences
133: are likely a consequence of chemical effects. Because of the
134: freeze-out on grain mantles, the abundance of CO and its isotopologues
135: decrease by about two orders of magnitude towards the center of the
136: core \citep{Bergin02a}. Since H$^{13}$CO$^{+}$ is mainly formed from
137: the reaction of $^{13}$CO with H$_{3}^{+}$, its abundance is also
138: expected to decrease towards the core center. DCO$^{+}$ should also be
139: affected by the depletion of CO. However, the deuterium fractionation
140: increases as CO is removed from the gas phase. Thus the disappearance
141: of CO might be compensated by the increased deuterium
142: fractionation. In the following, we interpret the emission of these
143: species using a chemical model coupled with Monte-Carlo radiative
144: transfer model, in order to derive precisely their abundance profiles.
145:
146: \section{Analysis}
147: \label{sec:analysis}
148:
149: We have used a technique that combines the predictions of a chemical
150: network with a Monte-Carlo radiative transfer
151: \citep{Bergin02a,Bergin06b,Maret06}. The outline of this technique is
152: the following. Chemical abundances are computed as a function of the
153: visual extinction in the core. Using these abundance profiles, the
154: line emission is computed with a Monte Carlo radiative transfer
155: code. The resulting map is convolved to the resolution of the
156: telescope, and is compared to the observations. Free parameters of the
157: chemical model (\eg cosmic ionization rate, metal abundances, etc.)
158: are adjusted until a good agreement is obtained between the model and
159: the observations. Thus, this technique allows for a direct comparison
160: between the predictions of the chemical network and the observations.
161:
162: We have used the chemical network of \citet{Bergin95}. This network is
163: contains about 150 species (including isotopologues, see below), and
164: focuses on the formation of simple molecules and ions (\eg CO and
165: HCO$^{+}$). The network includes the effect of depletion on grains,
166: and the desorption by thermal evaporation, UV photons, and cosmic rays
167: \citep{Hasegawa93a, Bringa04}. It also includes the effect of
168: fractionation of $^{13}$C and $^{18}$O, using the formalism described
169: by \citet{Langer93}. We have extended this network to include the
170: effect of deuterium fractionation, following the approach used by
171: \citet{Millar89}. Because of the importance of multiply deuterated
172: species in the deuterium fractionation process, these species were
173: also included in the network, following \citet{Roberts04}. It also
174: include neutralization reactions of ions on negatively charged
175: grains. The predictions of our network were checked against the UMIST
176: network \citep{Millar97} for consistency.
177:
178: We adopt the density profile determined by \citet{Alves01}, from
179: observations of near infrared extinction from background stars. This
180: profile is assumed to be constant as a function of time. The dust
181: temperature profile was computed using the analytical formulae from
182: \citet{Zucconi01}. For the gas temperature we have adopted the profile
183: determined by \citet{Bergin06b} from observations and modelling of CO
184: and its isotopologues. The gas temperature is relatively low (7-8 K),
185: and increases sightly (10-11 K) at the center of the core as indicated
186: by ammonia lines observations \citep{Lai03}. This increase in the
187: temperature is a result of grain coagulation at the center of the
188: core, which produces a thermal decoupling between the gas and the
189: cooler dust.
190:
191: The cloud is supposed to have the initial composition summarized in
192: Table \ref{table:initial-abundances}. In our model, we assume that the
193: density profile of the core does not evolve with time. Therefore, we
194: also assume that the chemistry has already evolved to a point where
195: hydrogen is fully molecular, and all the carbon is locked into CO.
196: Our treatment of the initial atomic oxygen pool deserves special
197: mention. \citet{Bergin02b} examined this question in the context of
198: the non-detection of water vapor emission in B68 by SWAS. They found
199: that if atomic oxygen were present in the gas phase in the dense core
200: center, then the well studied reaction chain that forms
201: $\mathrm{H_{2}O}$ (via $\mathrm{H_{3}O^{+}}$) would have yielded
202: detectable water vapor emission. The simplest way to stop this
203: reaction chain is to remove the fuel for the gas-phase chemistry:
204: atomic oxygen. This happens when oxygen is trapped on grain surfaces
205: in the form of water ice \citep[\eg][]{Bergin00}. Thus we have assumed
206: initial conditions in which all non-refractory oxygen is in the form
207: of water ice and CO gas with no atomic oxygen left. In this fashion
208: our initial abundances assume the core formed out of gas that reached
209: at $A_{v} \sim 2$ -- where $\mathrm{H_{2}}$ and CO have formed and
210: water ice mantles are observed. On the other hand, nitrogen is assumed
211: to be mostly in atomic form \citep{Maret06}.
212:
213: A grain size of 0.1 $\mu m$ is assumed. The cosmic ray ionization rate
214: and the abundance of low ionization potentials metals ($<$ 13.6 eV)
215: are free parameters of our study (see \S\ref{sec:metals-depletion} and
216: \S\ref{sec:constr-cosm-ioniz}). In our models we combine all metals
217: (\eg Fe$^+$, Mg$^+$, ...) into one species, labeled as M$^{+}$ with
218: the Fe$^+$ recombination rate of $\alpha(\mathrm{M}^+) = 3.7 \times
219: 10^{-12} (T/300\, \mathrm{K})^{-0.65}\, \mathrm{cm^{3}\, s^{-1}}$.
220: Due the low ionization potential these metals are assumed to be fully
221: ionized at the start of the calculation. The network also includes the
222: neutralization of ions of negatively charged grains with one electron
223: per grain.
224:
225: The core is assumed to be bathed in a UV field of 0.2 (in Habing
226: units; \citeyear{Habing68}), as determined by \citet{Bergin06b}. The
227: chemical abundances are computed as a function of time by solving the
228: rate equations using the DVODE algorithm \citep{Brown89}. This is done
229: until a time of $10^5 \ \mathrm{yr}$ is reached. This corresponds to
230: the ``best-fit'' model of \citet{Bergin06b}. However, as discussed by
231: \citeauthor{Bergin06b}, this time is a lower limit of the real age of
232: the cloud, since the CO is assumed to be pre-existing at $t = 0$ in
233: these models.
234:
235: Modeling the line emission requires the knowledge of velocity profile
236: in the core. As a first approach, we have neglected systematic motions
237: \citep[see][]{Lada03,Redman06}, and we have used the turbulent
238: velocity profile determined by \citet{Bergin06b} from
239: $\mathrm{C^{18}O}$ and $\mathrm{^{13}CO}$ lines. The turbulent
240: velocity is $\sim 0.3$ km/s at the edge of the cloud, and decreases
241: significantly ($\sim 0.15$ km/s) towards the center of the core.
242:
243: \section{Results}
244: \label{sec:results}
245:
246: \subsection{Metals depletion}
247: \label{sec:metals-depletion}
248:
249: Metals ions (\eg Fe$^{+}$ and Mg$^{+}$) play an important role in in
250: setting the electron abundance in pre-stellar cores, because they are
251: destroyed relatively slowly by radiative recombination. For example,
252: the recombination rate of H$_{3}^{+}$ is four order of magnitude
253: higher than the rate for Fe$^{+}$.
254:
255: \citet{Guelin82} measured the electron abundance in a sample of dense
256: molecular clouds using HCO$^{+}$ and DCO$^{+}$ line observations, and
257: obtained values comprised between 10$^{-8}$ and 10$^{-7}$. The authors
258: concluded that the metal abundance is lower than 10$^{-7}$ in these
259: clouds. \citet{Caselli98} determined the electron abundance in a
260: sample of twenty four low-mass isolated cores (with embedded stars and
261: starless -- similar in properties to B68) from CO, HCO$^{+}$ and
262: DCO$^{+}$ observations, and obtained values in the range
263: 10$^{-8}$-10$^{-6}$. \citeauthor{Caselli98} argued that the
264: differences between cores are due to changes in metal abundance and a
265: variable cosmic ionization rate ($\zeta$). The best fit between their
266: chemical model predictions and the observations indicates metal
267: abundances in the range $2 \times 10^{-9} - 3 \times 10^{-7}$.
268: \citet{Williams98} determined the electron abundance in a similar
269: sample of low mass cores using a slightly different approach, and
270: obtained metal abundances comprised between $5 \times 10^{-9}$ and $4
271: \times 10^{-8}$ (assuming a constant $\zeta$). All these studies
272: indicate low metal abundances with respect to their solar
273: values.. Indeed, observations of FUV FeII absorption lines, and other
274: metal lines, towards diffuse clouds find depletion factors of over two
275: order of magnitude with respect to solar values
276: \citep{Savage79,Jenkins86,Snow02}.
277:
278: Our $\mathrm{H^{13}CO^{+}} \ (1-0)$ observations can be used to set
279: limits on the metal ion abundance in B68. $\mathrm{H^{13}CO^{+}}$ is
280: sensitive to the electron abundance inside the core, because it is
281: mainly destroyed by electronic recombination. It is also sensitive to
282: the $\mathrm{H_{3}^{+}}$ and $\mathrm{^{13}CO}$ abundances, since it
283: is formed from the reaction between these two
284: species. $\mathrm{H_{3}^{+}}$ itself is mainly formed from
285: $\mathrm{H_{2}}$ ionization by cosmic rays. The remaining parameter
286: in determining the chemical abundance profile is the time dependence
287: of the chemistry. In this case, our analysis is simplified because
288: \citet{Bergin06b} used multiple transitions of $^{13}$CO and C$^{18}$O
289: and a similar modeling technique to derive the $^{13}$CO abundance and
290: constrain the ``chemical age\footnote{See discussion in
291: \citet{Bergin06b} on the meaning of this ``chemical age''.}'' of
292: Barnard 68 to $t = 10^5 \, \mathrm{yr}$. Thus, the only free
293: parameters for our modeling of the $\mathrm{H^{13}CO^{+}}$ emission
294: are the cosmic ionization rate $\zeta$ and the metal ion
295: abundance. These two parameters are difficult to constrain
296: simultaneously. In \citet{Maret06}, we found that the
297: $\mathrm{H^{13}CO^{+}} \ (1-0)$ line emission in B68 is well
298: reproduced by our chemical network if one assume a metal abundance of
299: $3 \times 10^{-9}$ with respect to H nuclei and a standard cosmic
300: ionization rate ($\zeta = 3 \times 10^{-17} \ \rm{s}^{-1}$, see next
301: section). In the following, we explore the parameter space into more
302: details to place constrains on the metal abundance in the core.
303:
304: On Fig. \ref{fig2}, we show the predicted intensity of the
305: $\mathrm{H^{13}CO^{+}}$ (1-0) line for different metal ion abundances
306: and cosmic ionization rates. In these models, metals are assumed to be
307: initially fully ionized. In Fig. \ref{fig2}, we see that for $\zeta =
308: 3 \times 10^{-17} \ \rm{s}^{-1}$, our model predicts the same
309: intensities for $x(\mathrm{M}^{+})$ = 0 and $x(\mathrm{M}^{+}) = 3
310: \times 10^{-10}$. The predicted emission is in fairly good agreement
311: with the observations. On the other hand, for a higher metal
312: abundance ($x(\mathrm{M}^{+}) = 3 \times 10^{-9}$) the model predicts
313: a intensity slightly lower than the observed, but is in better
314: agreement with the observations at the center of the core. A metal
315: abundance of $3 \times 10^{-8}$ is clearly ruled out by the model and
316: observation comparison. We conclude that $x(\mathrm{M}^{+}) \le 3
317: \times 10^{-9}$. This value is at the low end of the one obtained by
318: \citet{Caselli98} and \citet{Williams98}. Compared to the abundance of
319: metals in the solar photosphere \citep[$x(\mathrm{M}) \sim 8.5 \times
320: 10^{-5 }$;][]{Anders89}, this represent a depletion factor of more
321: than four orders of magnitude. Indeed, our observations are also fully
322: consistent with a complete depletion of metals in the core, \ie with
323: cosmic rays as the only source of ionization at $A_v$ greater than a
324: few magnitudes (see Fig. \ref{fig2}). It should be noted, however,
325: that result depends on the value of $\zeta$ adopted. For example, our
326: observations are fully consistent with a cosmic ionization rate of $3
327: \times 10^{-16} \ \rm{s}^{-1}$ and $x(\mathrm{M}^{+}) = 3 \times
328: 10^{-8}$. The effects of varying $\zeta$ are discussed in the next
329: section.
330:
331: \subsection{Cosmic Ray Ionization Rate}
332: \label{sec:constr-cosm-ioniz}
333:
334: Cosmic rays play a crucial role in the chemistry of pre-stellar cores,
335: because they set the abundance of the pivotal H$_{3}^{+}$ ion, and are
336: the only source of ionization at $A_{v}$ greater than a few
337: magnitudes. Despite of its importance, the cosmic ray ionization rate
338: is difficult to constrain (see \citealt{LePetit04},
339: \citealt{vanderTak06} and \citealt{Dalgarno06} for recent
340: reviews). Early estimates in diffuse clouds from HD and OD
341: observations indicate $\zeta = 7 \times 10^{-17} \, \mathrm{s}^{-1}$
342: \citep{vanDishoeck86}, a value in agreement with the lower limit of $3
343: \times 10^{-17} \, \mathrm{s}^{-1}$ measured by the Voyager and
344: Pioneer satellites \citep{Webber98}. $\mathrm{H_{3}^{+}}$ observations
345: towards $\zeta$ Persei cloud suggest a significantly higher rate
346: \citep[$\zeta = 1.2 \times 10^{-15} \, \mathrm{s}^{-1}$;
347: ][]{McCall03}. However, \citet{LePetit04} argued that a value of
348: $\zeta = 2.5 \times 10^{-16} \, \mathrm{s}^{-1}$ is more consistent
349: with both $\mathrm{H_{3}^{+}}$ and HD observations. In denser regions,
350: $\mathrm{HCO^{+}}$ observations indicates a lower ionization rate than
351: in diffuse clouds: \citet{vanderTak00} obtained $\zeta = (2.6 \pm 1.8)
352: \times 10^{-17} \, \mathrm{s}^{-1}$ from $\mathrm{HCO^{+}}$ line
353: observations towards massive protostars. In pre-stellar cores,
354: \citet{Caselli98} inferred a value comprised between $10^{-18}$ and
355: $10^{-16} \, \mathrm{s}^{-1}$. The difference in the cosmic ray
356: ionization rate between diffuse and dense clouds could be due to the
357: scattering of cosmic rays \citep{Padoan05}. In addition, large
358: variations are inferred as a function of the Galactic Center distance
359: \citep{Oka05,vanderTak06}.
360:
361: Cosmic rays are also heating agents of the gas. \citet{Bergin06b}
362: examined the value of $\zeta$ in B68 by comparing the predictions of a
363: chemical and thermal model to observations of CO and its
364: isotopologues. \citeauthor{Bergin06b} found that their model provide
365: reasonable fits to the data for $\zeta = 1 - 6 \times 10^{-17} \
366: \mathrm{s}^{-1}$. Their ``best fit'' model has $\zeta = 1.5 - 3 \times
367: 10^{-17} \ \mathrm{s}^{-1}$. Here we examine the constraints placed by
368: our $\mathrm{H^{13}CO^{+}}$ observations. On Fig. \ref{fig2}, we see
369: that our model produces a good fit to the data for $\zeta = 3 \times
370: 10^{-17} \ \mathrm{s}^{-1}$, except for $x(\mathrm{M}^{+}) = 3 \times
371: 10^{-8}$, where the model predictions underestimate the observation by
372: a factor two. Models with $\zeta = 3 \times 10^{-18} \
373: \mathrm{s}^{-1}$, consistently underestimate the
374: observations. Conversely models with $\zeta = 3 \times 10^{-16} \
375: \mathrm{s}^{-1}$ overestimate the model, except the one with
376: $x(\mathrm{M}^{+}) \le 3 \times 10^{-8}$. This in agreement with
377: \citet{Bergin06b}, who found that their observations are not
378: reproduced by models with $\zeta = 6 \times 10^{-16} \
379: \mathrm{s}^{-1}$.
380:
381: To summarize our conclusions regarding the metals abundances and the
382: cosmic ionization rate, models with $x(\mathrm{M}^{+}) \le 3 \times
383: 10^{-9}$ provide a good agreement with the data\footnote{For
384: simplicity, we have assumed that the initial $x(\mathrm{M^{+}})$
385: abundance is constant as a function of a radius. A better fit to the
386: observations might be obtained with a variation of
387: $x(\mathrm{M^{+}})$ with the radius.}, although the model with
388: $x(\mathrm{M}^{+}) = 3 \times 10^{-8}$ and $\zeta = 3 \times 10^{-16}
389: \ \mathrm{s}^{-1}$ is also consistent with our data. However values of
390: $\zeta$ greater that $6 \times 10^{-17} {s}^{-1}$ are ruled out by
391: \citet{Bergin06b} based on core thermal balance. On the other hand,
392: models with $\zeta = 3 \times 10^{-18} \ \mathrm{s}^{-1}$ always
393: underestimate our observations. We conclude that $\zeta = 1-6 \times
394: 10^{-17} \ \mathrm{s}^{-1}$, and $x(\mathrm{M}^{+}) \le 3 \times
395: 10^{-9}$ in B68. This implies that the abundance of ionized metals is
396: reduced in the center of B68. Charge transfer from molecular ions
397: (\eg H$_3^+$, HCO$^+$) to metals can be important and a reduction in
398: the abundance of ionized metals also requires lowering the neutral
399: metal abundance. In the case of Fe a potential reservoir is FeS
400: \citep{Keller02}, or organometallic molecules \citep{Serra92}. An
401: other possility is that Fe is incorporated into grain cores.
402:
403: \subsection{Ortho to para H$_{2}$ ratio}
404: \label{sec:ortho-para-h_2}
405:
406: The ortho to para H$_{2}$ ratio influences the degree of ion and
407: molecule deuteration in prestellar cores
408: \citep{PineaudesForets91,Flower06a}. In the gas phase, deuterium
409: fractionation is mainly due to the following reaction \cite[see ][and
410: references therein]{Roberts04}:
411: \begin{equation}
412: \mathrm{H_{3}^{+} + HD \rightleftharpoons H_{2}D^{+} + H_{2}}
413: \end{equation}
414: The reverse reaction has an activation barrier of $\sim 232 \
415: \mathrm{K}$ and therefore the reaction becomes essentially
416: irreversible at low temperature. \citet{Gerlich02} measured the
417: forward and reverse rates of the above reaction at 10 K, and found
418: them to be very different than commonly adopted values. The forward
419: reaction rate was found to be about five times higher than previous
420: estimates \citep{Sidhu92}, while the reverse reaction rate was found
421: to be five orders of magnitude larger than previously used
422: \citep[\eg][]{Caselli98}. In addition, \citet{Gerlich02} determined
423: via a laboratory measurement that the reverse reaction rate is very
424: sensitive to the ratio of ortho to para molecular hydrogen. This is
425: because o-H$_{2}$, in its ground rotational level ($J=1$) has an
426: higher energy ($\Delta E \sim 170.5 \ \mathrm{K}$) when compared to
427: the ground state of p-H$_{2}$ ($J=0$). Consequently, o-H$_{2}$ can
428: more easily cross the energy barrier than p-H$_{2}$, and the rate of
429: the reverse reaction increases with the ortho to para H$_{2}$ ratio.
430:
431: Our $\mathrm{DCO^{+}}$ observations can be used to estimate the
432: $\mathrm{H_{2}D^{+}}$ abundance, and thus the efficiency of the deuterium
433: fractionation process. $\mathrm{DCO^{+}}$ is mainly formed by the
434: following reaction:
435: \begin{equation}
436: \mathrm{H_{2}D^{+} + CO \rightarrow DCO^{+} + H_{2}}
437: \end{equation}
438: \noindent and is mainly destroyed by electronic recombination. Thus
439: the $\mathrm{DCO^{+}}$ emission depends on both the CO and
440: $\mathrm{H_{2}D^{+}}$ abundances, the electron fraction (induced by
441: cosmic-rays and by pre-existing metal ions), the ortho to para H$_{2}$
442: ratio, and on time. Here we benefit from our previous analysis of CO
443: which constrained the CO abundance and ``chemical age'' and our
444: analysis of $\mathrm{H^{13}CO^{+}}$ which limit the metal ion
445: abundance and cosmic ray ionization rate. Thus the primary free
446: parameter is the ortho to para $\mathrm{H_{2}}$ ratio ($o/p$) when we
447: adopt our best fit parameters of $x(M^{+}) = 3 \times 10^{-9}$ and
448: $\zeta = 3 \times 10^{-17} \, \mathrm{s^{-1}}$.
449:
450: On Fig. \ref{fig3}, we compare the observed $\mathrm{DCO^{+}}$ (1-0)
451: line emission as a function of $A_{v}$, with the predictions of our
452: model for different ortho to para H$_{2}$ ratio. Note that in these
453: models, no $o/p$ conversion is considered: the ortho to para H$_{2}$
454: ratio is assumed to be constant. The best agreement\footnote{Although
455: the model predicts the correct intensity at the core center, one can
456: note that emission at lower $A_\mathrm{v}$ is slightly
457: underestimated. This may suggest an H$_2$ $o/p$ variation with the
458: radius of the core: increasing the ratio at low $A_\mathrm{v}$ would
459: increase the $\mathrm{DCO^{+}}$ abundance emission in this region
460: and would probably produce a better fit.} between the observations
461: and the model is obtained for an $o/p$ ratio of $\sim 1.5 \times
462: 10^{-2}$, well above the Boltzmann equilibrium value at 10 K ($3.5
463: \times 10^{-7}$). Fig. \ref{fig3} also show the derived
464: $\mathrm{DCO^{+}}$ abundance inside the core. The abundance peaks at
465: an $A_{v}$ of $\sim 5$, and decreases slightly towards the core
466: center, as a consequence of CO depletion (see Section
467: \ref{sec:electron-abundance}).
468:
469: It is interesting to compare the $o/p$ H$_{2}$ ratio we obtain with
470: the predictions of other models. \citet{Walmsley04} modeled the $o/p$
471: H$_{2}$ ratio in prestellar cores, assuming a complete depletion of
472: heavy elements. In their model, an initial o/p H$_{2}$ ratio of $3.5
473: \times 10^{-7}$ is assumed. For a density of $10^6 \
474: \mathrm{cm}^{-3}$, steady-state is reached in $10^{5}$ yr, a time
475: comparable to the age of B68 inferred from CO depletion observation
476: and modeling \citep{Bergin06b}. At steady state, the o/p H$_{2}$ ratio
477: obtained is $6 \times 10^{-5}$, \ie about two orders of magnitude
478: lower than the value determined in this work. However, as noted by
479: \citet{Flower06b}, the ortho to para H$_{2}$ ratio conversion reactions
480: are very slow, and it is not clear if the steady state equilibrium is
481: reached in molecular clouds prior to the formation of dense
482: cores. Using a initial ortho to para ratio of 3 (a value appropriate
483: for H$_{2}$ formation on grains), \citet{Flower06a} obtain a steady
484: state ratio of $3 \times 10^{-3}$. This value, although still about a
485: factor 5 lower, is in better agreement with our estimate. We note that
486: for $o/p = 3 \times 10^{-3}$, our model predicts an DCO$^{+}$ (1-0)
487: emission about 2 times higher than the observations (see
488: Fig. \ref{fig3}).
489:
490: \section{Discussion}
491: \label{sec:discussion}
492:
493: \subsection{Electron abundance and main charge carriers}
494: \label{sec:electron-abundance}
495:
496: On Fig. \ref{fig4}, we show the derived electron and main ions
497: abundances inside the core. The electron abundance is $\sim 5 \times
498: 10^{-9}$ with respect to H nuclei throughout most part of the core. At
499: low $A_{v}$, the electron abundance increases as a result of
500: photo-dissociation of CO. In this region, the most abundant ion is
501: C$^{+}$. At higher $A_{v}$, the most abundant ion is H$_{3}^{+}$,
502: which caries about $\sim 20\%$ of the electric charge. The remainder
503: of the charge is shared between more complex ions. Deuterated ions do
504: not contribute significantly to the ionization fraction. In the
505: innermost region of the core, where the deuteration increases as a
506: result of CO depletion, the main deuterated ion, D$_{3}^{+}$, is about
507: ten times less abundant than H$_{3}^{+}$. H$_{2}$D$^{+}$ and
508: D$_{2}$H$^{+}$ have similar abundances ($2 \times 10^{-11}$ with
509: respect to H). This is in agreement with recent observations
510: \citep{Vastel04}.
511:
512: Recently, \citet{Hogerheijde06} reported a probable detection of
513: $\mathrm{o-H_{2}D^{+}}$ fundamental line towards B68 which can be
514: compared to our model predictions. The measured flux is however quite
515: uncertain, given the relatively low signal to noise ratio of this
516: observation (2.7$\sigma$ and 5.2$\sigma$ on the peak and integrated
517: intensity, respectively). Assuming a thermal excitation (10 K) and
518: optically thin conditions, \citeauthor{Hogerheijde06} derive a
519: $\mathrm{H_{2}D^{+}}$ column density of $1.5 \times 10^{12}$
520: cm$^{-2}$. Assuming a H$_{2}$ column density of $3.6 \times 10^{22}$
521: cm$^{-2}$ \citep{Alves01}, this corresponds to an
522: $\mathrm{H_{2}D^{+}}$ abundance of $2.1 \times 10^{-11}$ with respect
523: to H nuclei, averaged in the APEX beam (17\arcsec), with respect to H
524: nuclei. This is in excellent agreement with our model, which predicts
525: an H$_{2}$D$^{+}$ abundance of $2 \times 10^{-11}$, roughly constant
526: across the envelope. Of course, if the excitation is non-thermal, the
527: detection implies an higher abundance. Assuming a 5 K excitation
528: temperature, \citeauthor{Hogerheijde06} derive a beam averaged
529: abundance of $1.5 \times 10^{-10}$ with respect to H nuclei. This is
530: about an order of magnitude higher than our model predictions. Since
531: no collisional rates exist in the literature for
532: $\mathrm{H_{2}D^{+}}$, it is unclear whether or not the excitation of
533: this line is thermal. \citeauthor{Hogerheijde06} estimate a critical
534: density of $2 \times 10^{6}$ cm$^{-3}$, which exceeds the density at
535: the center of B68 ($3 \times 10^{5}$ cm$^{-3}$) by about an order of
536: magnitude. However, the collisional rate, and therefore the critical
537: density, is uncertain by an order of magnitude
538: \citep{vanderTak05,Hogerheijde06}. Our model predictions regarding the
539: deuterium chemistry could be also tested via observations of the
540: $\mathrm{D_{2}H^{+}} \ 1_{1,0}-1_{0,1}$ ($\nu$ = 691.66044
541: GHz). Assuming a excitation temperature of 10 K, we predict a line
542: intensity of 10 mK. Unfortunately, this is too weak to be detected
543: with current ground based telescopes.
544:
545: We would like to compare the electron abundance profile we obtained
546: with the one derived by \citet{Caselli02b} in L1544. In the
547: \citeauthor{Caselli02b} best fit model, the electron abundance at the
548: center of L1544 is $5 \times 10^{-10}$ (with respect to H), while we
549: obtain an electron abundance an order of magnitude higher at the
550: center of B68. These differences are probably a consequence of
551: different central densities: the L1544 central density is about an
552: order of magnitude higher than the one of B68, and the electron
553: fraction is expected to scale as $n(\mathrm{H}_{2})^{-1/2}$
554: \citep{McKee89}. Another important difference is the dominant ion:
555: \citet{Caselli02b} predicts that the most abundant ion is
556: $\mathrm{H_{3}O^{+}}$, while in our modeling main charge carrier is
557: $\mathrm{H_{3}^{+}}$. These differences are due to different
558: assumptions on the atomic oxygen abundance. \citet{Caselli02b} assumes
559: that oxygen is initially mostly atomic. As a consequence, the
560: $\mathrm{H_{3}O^{+}}$ abundance is relatively large, because atomic
561: oxygen reacts with $\mathrm{H_{3}^{+}}$ to form $\mathrm{H_{3}O^{+}}$
562: (after successive protonations by H$_{2}$ followed by
563: recombination). In our modeling, oxygen is assumed to be initially
564: locked in water ices and gas phase CO (see Table
565: \ref{table:initial-abundances}), and the atomic oxygen gas phase
566: abundance is relatively low.
567:
568: Finally, we would like to comment on the effect of grain size
569: evolution on the electron fraction in the core. \cite{Walmsley04}
570: computed the electron abundance and main charge carrier in a
571: prestellar core for different grain sizes. For a grain size of 0.02
572: $\mu m$, the main charge carrier in their model is H$_{3}^{+}$, while
573: for larger grains (0.1 $\mu m$), the most abundant ion becomes
574: H$^{+}$. In their models, H$^{+}$ recombines primarily on grains,
575: while H$_{3}^{+}$ recombines with free electrons. Since the
576: recombination timescale on grains depends on the grain size, the
577: H$^{+}$ over H$_{3}^{+}$ abundance ratio, and in turn the electron
578: abundance, depends on the grain size as well. However, these models
579: assume a complete depletion of heavy elements, which is not the case
580: for B68. In B68 we do find evidence for strong molecular, but not
581: complete, heavy element freeze-out, at the core center. The reaction
582: with H$^{+}$ with molecules containing these elements (\eg NH$_3$,
583: OH, ...) can transfer the charge to molecular ions with faster
584: recombination timescales. This would probably reduce the dependence of
585: the electron abundance on the grain size.
586:
587: \subsection{Core stability}
588: \label{sec:core-stability}
589:
590: The electron abundance in the core is also important for its dynamical
591: evolution, since its affects the efficiency of ambipolar diffusion. In
592: a weakly ionized sub-critical core, the ions are supported against
593: collapse by the magnetic field, but neutrals can slowly drift with
594: respect to the ions \citep[see ][for a review]{Shu87}. The timescale
595: for this phenomenon is given by \citet{Walmsley04}:
596: \begin{equation}
597: \tau_{\mathrm{ad}} = \frac{2}{\pi G m_n^{2}} \sum_{i}
598: \frac{n_i}{n_n} \frac{m_i m_n}{m_i + m_n} \langle \sigma v \rangle_{in}
599: \end{equation}
600: \noindent
601: where G is the gravitational constant, m$_{n}$ and m$_{i}$ are the
602: masses of the neutrals and the ions respectively, n$_{n}$ and n$_{i}$
603: are the number densities, $\langle \sigma v \rangle_{in}$ is the rate
604: coefficient for the momentum transfer, and the summation goes over all
605: ions. At low temperature, the rate coefficient for momentum transfer
606: is \citep{Flower00}:
607: \begin{equation}
608: \langle \sigma v \rangle_{in}= 2 \pi e \left( \alpha \frac{m_i +
609: m_n}{m_i m_n} \right)^{1/2}
610: \end{equation}
611: \noindent
612: where $\alpha$ is the polarizability of $\mathrm{H_{2}}$. Assuming
613: that $\mathrm{H_{3}^{+}}$ is the dominant ion, we obtain:
614: \begin{equation}
615: \tau_{\mathrm{ad}} \sim 2 \times 10^{14} \, x(e)\ \mathrm{yr}
616: \end{equation}
617: \noindent
618: where $x(e)$ is the electron abundance, with respect to H. Thus at the
619: center at the core, the ambipolar diffusion timescale is $10^{6}$
620: yr. It is interesting to compare this to the free fall time scale,
621: which is given by:
622: \begin{equation}
623: \tau_{\mathrm{ff}} = \left( \frac{3 \pi}{32 G \rho} \right)^{1/2}
624: \end{equation}
625: \noindent
626: where $\rho = n_\mathrm{H_2} m_\mathrm{H_2}$ is the mass
627: density. When expressed as a function of $n_\mathrm{H_2}$, this gives:
628: \begin{equation}
629: \tau_{\mathrm{ff}} = 3.6 \times 10^{7} \, n_\mathrm{H_2}^{-1/2}\ \mathrm{yr}
630: \end{equation}
631: At the center of B68 we obtain $\tau_{\mathrm{ff}} = 7 \times 10^{4} \
632: \mathrm{yr}$ which is about an order of magnitude faster than the
633: ambipolar diffusion timescale. Thus, if present, the magnetic field
634: may provide an important source of support.
635:
636: The strength of the magnetic field that is needed to support the
637: cloud can be obtained from the critical mass \citep{Mouschovias76}:
638: \begin{equation}
639: M \sim \frac {0.13}{G^{1/2}} \phi_{B}
640: \end{equation}
641: \noindent
642: where $\phi_{B} = \pi R^{2} B$ is the magnetic flux, $R$ is the core
643: radius, and $B$ is the magnetic field strength. The strength of the
644: magnetic field that is needed to support the cloud is therefore:
645: \begin{equation}
646: B \sim \frac {G^{1/2} M}{0.13 \ \pi R^{2}}
647: \end{equation}
648: where $M$ is the mass the core. Using $R = 12500\ \mathrm{AU}$ and $M
649: = 2.1 \ \mathrm{M}_{\sun}$ \citep{Alves01}, we obtain a critical
650: magnetic field of 76 $\mu$G for B68. No magnetic field measurements
651: for B68 exist in the literature, but we can compare this value to the
652: one measured in other cores from dust sub-millimeter
653: polarization. \citet{WardThompson00} and \cite{Crutcher04} measured
654: plane-of-the-sky magnetic field strengths of 80 $\mu$G in L183, 140
655: $\mu$G in L1544 and 160 $\mu$G in L43. \citet{Kirk06} measured lower
656: fields of 10 and 30 $\mu$G in the L1498 and L1517B cores. Therefore,
657: if the magnetic field strength in B68 is at the lower end of the
658: values measured in other cores, then it might be super-critical (\ie
659: the magnetic field is too weak to balance gravity). If it is higher,
660: then the core is probably sub-critical. One may argue B68 has nearly
661: round shape (albeit with an asymmetrical extension to the southeast),
662: which potentially is indicative of a weak magnetic field.
663:
664: \subsection{Implications of the metals depletion for accretion in
665: protostellar disks}
666:
667: One important conclusion of this study is the large metal depletion
668: inferred for B68. Here we examine the implication of this findings for
669: the mechanism of angular momentum transport in protostellar disks. The
670: most favored theory for angular momentum transport in disks predicts
671: that accretion occurs via magneto-rotational instability \cite[MRI;
672: ][]{Balbus91} which produces MHD turbulence. Since this is a magnetic
673: process, the ion-neutral coupling is therefore important. Typically,
674: the ionization fraction should be greater than 10$^{-12}$ for disks to
675: be able to sustain MHD turbulence \citep[see][and references
676: therein]{Ilgner06}. \citet{Gammie96} suggested a model wherein the
677: accretion is layered. The electron abundance is high at the surface of
678: the disk, because of the ionization of the gas by UV, X-rays and
679: cosmic rays, but it decreases towards the mid-plane. Thus disks may
680: have a magnetically active zones at high altitude, where the electron
681: fraction is sufficient to maintain MHD turbulence, and \emph{dead
682: zones}, closer the mid plane of the disk, where the electron
683: fraction is lower, and accretion cannot occur.
684:
685: Our results have some import on this process because the chemical
686: structure of the pre-stellar stage sets the initial chemical
687: conditions of the gas that feeds the forming proto-planetary disk.
688: Because of their influence on the ionization fraction, metal ions can
689: have dramatic effects on the size of the dead zone, assuming that they
690: are provided by infall to the disk \citep{Fromang02,Ilgner06}. The
691: latter authors computed the ionization fraction in a protostellar
692: disk, and found that for $x(\mathrm{M}^{+}) \le 3 \times 10^{-10}$,
693: the dead zone extend between 0.5 and 2 AU, while it disappears
694: completely for $x(\mathrm{M}^{+}) \ge 10^{-8}$. In B68, we obtain a
695: metal abundance of $x(\mathrm{M}^{+}) \le 3 \times 10^{-9}$, which is
696: below the threshold for a complete disappearance of the dead
697: zone. Thus, if B68 is representative of the initial conditions for the
698: formation of protostellar disks, and cosmic rays do not penetrate
699: deeply to the midplane, dead zones should exist in those disks.
700:
701: \section{Conclusions}
702: \label{sec:conclusions}
703:
704: We have presented a detailed analysis of the electron abundance in the
705: B68 prestellar core using $\mathrm{H^{13}CO^{+}} (1-0)$ and
706: $\mathrm{DCO^{+}} (2-1)$ line observations. These observations were
707: compared to the predictions of time dependent chemical model coupled
708: with a Monte-Carlo radiative transfer code. This technique allows for
709: a direct comparison between chemical model predictions and observed
710: line intensities as a function of radius (or the visual extinction) of
711: the core. Our main conclusions are:
712: \begin{enumerate}
713: \item The metal abundance is difficult to constrain independently from
714: the cosmic ionization rate. However, accounting for thermal balance
715: considerations and to reproduce $\mathrm{H^{13}CO^{+}} (1-0)$
716: emission we estimate that $x(\mathrm{M}^{+}) \le 3 \times 10^{-9}$
717: and $\zeta = 1-6 \times 10^{-17} \ \mathrm{s}^{-1}$.
718: \item The $\mathrm{DCO^{+}} (2-1)$ line emission is sensitive to the
719: ortho to para ratio. The emission is well reproduced by our model
720: for an ortho to para ratio of $1.5 \times 10^{-2}$, well below the
721: equilibrium value, and in reasonable agreement with previous work.
722: \item The inferred electron abundance is $5 \times 10^{-9}$ (with
723: respect to H), and is roughly constant in the core at $A_\mathrm{v}
724: > 5$. It increases at lower $A_\mathrm{v}$ because of the
725: photo-dissociation of CO and photo-ionization of C. In the dense
726: part of the core, the dominant ion is
727: H$_{3}^{+}$. $\mathrm{H_{2}D^{+}}$ and $\mathrm{D_{2}H^{+}}$ have
728: similar abundances and are about two of magnitude less abundant than
729: H$_{3}^{+}$. In the center of the core, our model predicts
730: $\mathrm{D_{3}^{+}}$ to be the most abundant deuterated ion.
731: \item The inferred electron abundance implies an ambipolar diffusion
732: timescale of $10^{6}$ yr at the center of the core, which is about
733: an order of magnitude higher than the free fall timescale
734: ($\tau_{\mathrm{ff}} = 7 \times 10^{4}$ yr).
735: \item The metal abundance we obtain is below the threshold for
736: protostellar disk to be fully active. Consequently, if the chemical
737: composition of B68 is reprentative of the initial conditions for the
738: formation of a disk and cosmic rays do not penetrate to the disk
739: mid-plane, then dead zones should exist in protostellar disks.
740: \end{enumerate}
741:
742: \acknowledgments Both authors are grateful to C. Lada for a fruitful
743: collaboration that led to this work and to T. Huard and E. Aguti for
744: obtaining a portion of these data. We are also grateful to the referee
745: and to the editor J. Black for useful and constructive
746: comments. S.~M. wishes to thanks H. Roberts for helping us testing the
747: predictions of our network for deuterated species, E. Herbst for
748: useful discussions about chemical reactions rates, E. Roueff and
749: M. Walmsley for discussions about the ortho to para ratio, L.
750: Hartmann and F. Heitch for discussions about the dynamics of B68, and
751: S. Fromang for discussions about MRI in protostellar disks. This work
752: is supported by the National Science Foundation under grant 0335207.
753:
754: {\it Facilities:} \facility{IRAM:30m}, \facility{CSO}, \facility{APEX}
755:
756: \bibliographystyle{apj}
757: \bibliography{bibliography}
758:
759: \clearpage
760:
761: \begin{figure}
762: \centering \includegraphics[scale=.6,angle=-90]{f1.eps}
763: \caption{Comparison between integrated intensity maps
764: (\emph{contours}) of C$^{18}$O (1-0) (\emph{left}, from
765: \citealt{Bergin02a}), H$^{13}$CO$^{+}$ (1-0) (\emph{center}) and
766: DCO$^{+}$ (2-1) (\emph{right}) superposed on the map of visual
767: extinction obtained by \citet{Alves01}. C$^{18}$O (1-0) contours
768: start at 0.2 K km s$^{-1}$ and step by 0.2 K km
769: s$^{-1}$. H$^{13}$O (1-0) contours start at 0.15 K km s$^{-1}$ and
770: step by 0.15 K km s$^{-1}$. DCO$^{+}$ (1-0) contours start at 0.1
771: K km s$^{-1}$ and step by 0.1 K km s$^{-1}$. The $A_{V}$ image
772: range from 0 to 27 mag. \label{fig1}}
773: \end{figure}
774:
775: \clearpage
776:
777: \begin{figure}
778: \epsscale{1}
779: \plotone{f2.eps}
780: \caption{Comparison between the observations and the model
781: predictions of the $\mathrm{H^{13}CO^{+}}$ (1-0) line for
782: different cosmic ionization rates ($\zeta$) and metal
783: abundances. Blue points with error bars (1 $\sigma$) represent the
784: observed integrated line intensity as a function of the visual
785: extinction in the core ($A_{v}$). Dashed, solid and dash-dotted
786: lines represent the model predictions for $\zeta = 3 \times
787: 10^{-18}$, $3 \times 10^{-17}$ and $3 \times 10^{-16}\,
788: \mathrm{s^{-1}}$ respectively. In panel \emph{(a)} a complete
789: depletion of metals is assumed. In panels \emph{(b)}, \emph{(c)},
790: and \emph{(d)}, a metals abundance of respectively $3 \times
791: 10^{-10}$, $3 \times 10^{-9}$ and $3 \times 10^{-8}$ is
792: assumed. \label{fig2}}
793: \end{figure}
794:
795: \clearpage
796:
797: \begin{figure}
798: \epsscale{0.75}
799: \plotone{f3.eps}
800: \caption{Comparison between the observations and the model
801: predictions of the $\mathrm{DCO^{+}}$ (2-1) line for different ortho
802: to para $\mathrm{H_{2}}$ ratios. Blue points with error bars (1
803: $\sigma$) represent the observed integrated line intensity as a
804: function of the visual extinction in the core ($A_{v}$). The dashed,
805: dotted and solid black lines show the predicted line intensity for
806: different ortho to para $\mathrm{H_{2}}$ ratios. The inset show the
807: corresponding abundances of $\mathrm{DCO^{+}}$.\label{fig3}}
808: \end{figure}
809:
810: \clearpage
811:
812: \begin{figure}
813: \epsscale{0.75}
814: \plotone{f4.eps}
815: \caption{Derived abundances inside B68 for the electrons and main
816: ions. Abundances are relative to H nuclei. \label{fig4}}
817: \end{figure}
818:
819: \clearpage
820:
821: \begin{deluxetable}{l c}
822: \tablewidth{0pt}
823: \tablecaption{Initial abundances.\label{table:initial-abundances}}
824: \tablehead{
825: \colhead{Species} & \colhead{Abundance\tablenotemark{a}}
826: }
827: \startdata
828: $\mathrm{H_{2}}$ & 0.5 \\
829: $\mathrm{He}$ & 0.14 \\
830: $\mathrm{H_{2}O_{ices}}$ & $2.2 \times 10^{-4}$ \\
831: $\mathrm{H_{2}^{18}O_{ices}}$ & $4.4\times 10^{-7}$ \\
832: $\mathrm{CO}$ & $8.5 \times 10^{-5}$ \\
833: $\mathrm{^{13}CO}$ & $9.5 \times 10^{-7}$ \\
834: $\mathrm{C^{18}O}$ & $1.7 \times 10^{-7}$ \\
835: $\mathrm{N}$ & $1.50 \times 10^{-5}$ \\
836: $\mathrm{N_{2}}$ & $2.5\times 10^{-6}$ \\
837: $\mathrm{HD}$ & $1.6 \times 10^{-5}$ \\
838: $\mathrm{Grains}$ & $10^{-12}$ \\
839: \enddata
840: \tablenotetext{a}{Relative to H nuclei.}
841: \end{deluxetable}
842:
843: \end{document}
844: