1: %\documentclass{aastex}

2: %\documentclass[12pt,preprint]{aastex}

3: %\documentstyle[12pt,aaspp4,psfig]{article}

4: %\documentclass[12pt,preprint]{aastex}

5: %\usepackage{emulateapj5}

6: %\usepackage[twocolumn,numberedappendix]{emulateapj5}

7: \documentclass[numberedappendix]{emulateapj}

8: 

9: 

10: %\documentstyle[emulateapj5]{article}   % emulate apj

11: %\documentclass[preprint]{aastex}

12: %\documentclass[manuscript]{aastex}

13: %%\usepackage{epsfig, bm}

14: %\usepackage{graphicx}

15: %\usepackage{rotating}

16: %%%%%%%%%%%%%%%%%%%%% macro definitions %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

17: \tightenlines

18: 

19: \font\slantsym=cmsy10

20: 

21: %------- delete following for submission to ApJ --------

22: %%\pagestyle{myheadings}

23: % today's date

24: %use number register 200 for "decade"

25: \countdef\decade=200

26: \decade=0

27: \advance\decade by \year

28: %\advance\decade by -2000       %to suppress two leading digits of yearb

29: \countdef\hours=201

30: \hours=0

31: \advance\hours by \time

32: \divide\hours by 60

33: \countdef\mins=202

34: \mins=0

35: \advance\mins by \hours

36: \multiply\mins by 60

37: \multiply\hours by 100

38: \countdef\miltime=203

39: \miltime=0

40: \advance\miltime by \hours

41: \advance\miltime by \time

42: \advance\miltime by -\mins

43: \newcommand     \todayd{\number\decade.\number\month.\number\day.\number\miltime}

44: %\markright{\todayd: DRAFT}

45: %---------------------------------------------------------------

46: 

47: %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

48: 

49: \newcommand{\ltsim}{\protect\raisebox{-0.5ex}{$\:\stackrel{\textstyle <}

50:         {\sim}\:$}}

51: \newcommand{\gtsim}{\protect\raisebox{-0.5ex}{$\:\stackrel{\textstyle >}

52:         {\sim}\:$}}

53: \newcommand{\ltsimscript}{\protect\raisebox{-0.5ex}{$\stackrel{\scriptstyle <}

54:         {\sim}$}}

55: \newcommand{\gtsimscript}{\protect\raisebox{-0.5ex}{$\stackrel{\scriptstyle >}

56:         {\sim}$}}

57: 

58: \newcommand{\msun}{M_{\odot}}

59: \newcommand{\lsun}{L_{\odot}}

60: \newcommand{\sfr}{\dot{M}_*}

61: \newcommand{\pgal}{P_{\rm gal}}

62: \newcommand{\phigal}{\phi_{\rm gal}}

63: \newcommand{\phicomp}{\phi_{\rm comp}}

64: \newcommand{\Sigmag}{\Sigma_{\rm g}}

65: \newcommand{\sigmag}{\sigma_{\rm g}}

66: \newcommand{\rhog}{\rho_{\rm g}}

67: \newcommand{\mcomp}{M_{\rm comp}}

68: \newcommand{\rcomp}{R_{\rm comp}}

69: \newcommand{\pcomp}{P_{\rm comp}}

70: \newcommand{\Sigmacomp}{\Sigma_{\rm comp}}

71: \newcommand{\Sigmashield}{\Sigma_{\rm sh}}

72: \newcommand{\rshield}{R_{\rm sh}}

73: \newcommand{\sigmacomp}{\sigma_{\rm comp}}

74: \newcommand{\avir}{\alpha_{\rm vir}}

75: \newcommand{\krho}{k_{\rho}}

76: \newcommand{\mmol}{M_{\rm mol}}

77: \newcommand{\Sigmamol}{\Sigma_{\rm mol}}

78: \newcommand{\Sigmacrit}{\Sigma_{\rm crit}}

79: \newcommand{\sigmad}{\sigma_{\rm d}}

80: \newcommand{\sigmahmol}{\sigma_{\rm H_2}}

81: \newcommand{\fmol}{f_{\rm H_2}}

82: \newcommand{\fatm}{f_{\rm HI}}

83: \newcommand{\ehat}{\hat{e}}

84: \newcommand{\flw}{\mathbf{F}_{\ellW}}

85: \newcommand{\flux}{F_{\ellW}}

86: \newcommand{\elw}{E_{\ellW}}

87: \newcommand{\calr}{\mathcal{R}}

88: \newcommand{\sigmadust}{\sigma_{\rm d,1000}}

89: \newcommand{\rmol}{r_{\rm H_2}}

90: \newcommand{\rdust}{r_{\rm d}}

91: \newcommand{\rd}{r_{\rm d}}

92: \newcommand{\Fd}{\mathcal{F}_{\rm d}}

93: \newcommand{\zd}{z_{\rm d}}

94: \newcommand{\zm}{z_{\rm H_2}}

95: \newcommand{\zmol}{z_{\rm H_2}}

96: \newcommand{\xd}{x_{\rm d}}

97: \newcommand{\xm}{x_{\rm H_2}}

98: \newcommand{\tauR}{\tau_{\rm R}}

99: \newcommand{\taudust}{\tau_{\rm d}}

100: \newcommand{\phicnm}{\phi_{\rm CNM}}

101: \newcommand{\edm}{E_{\rm dust}}

102: \newcommand{\fdissoc}{f_{\rm diss}}

103: \newcommand{\nsh}{N_{\rm sh}}

104: \newcommand{\mumol}{\mu_{\rm H_2}}

105: \newcommand{\taud}{\tau_{\rm d}}

106: \newcommand{\jh}{J^{\rm H}}

107: \newcommand{\jfuv}{J^{\rm FUV}}

108: \newcommand{\psifuv}{\Psi^{\rm FUV}}

109: \newcommand{\cale}{\mathcal{E}}

110: \newcommand{\calf}{\mathcal{F}}

111: \newcommand{\Fmol}{F_{\rm H_2}}

112: \newcommand{\ym}{y_{\rm H_2}}

113: \newcommand{\yd}{y_{\rm d}}

114: \newcommand{\Eone}{{\rm E}_1}

115: \newcommand{\xim}{\xi_{\rm H_2}}

116: \newcommand{\xid}{\xi_{\rm d}}

117: \newcommand{\Istar}{I^{*}}

118: \newcommand{\Jstar}{J^{*}}

119: \newcommand{\Estar}{E^{*}}

120: \newcommand{\Fstar}{F^{*}}

121: \newcommand{\bFstar}{\mathbf{F}^{*}}

122: \newcommand{\sigmae}{\sigma_{\rm e}}

123: \newcommand{\taueR}{\tau_{\rm eR}}

124: \newcommand{\taum}{\tau_{\rm HI}}

125: \newcommand{\nhi}{N_{\rm HI}}

126: 

127: \begin{document}

128: 

129: \title{The Atomic to Molecular Transition in Galaxies.\\I: An Analytic

130: Approximation for Photodissociation Fronts in Finite Clouds}

131: 

132: %\centerline{DRAFT: \today}

133: \slugcomment{Accepted to the Astrophysical Journal, August 8, 2008}

134: 

135: \author{Mark R. Krumholz\footnote{Hubble Fellow}}

136: \affil{Department of Astrophysical Sciences, Princeton University, Peyton Hall,

137: Princeton, NJ 08544 and Department of Astronomy \& Astrophysics, University of California, Santa Cruz, Integrative Sciences Building, Santa Cruz, CA 95060}

138: \email{krumholz@ucolick.org}

139: 

140: \author{Christopher F. McKee}

141: \affil{Departments of Physics and Astronomy, University of California, Berkeley, Campbell Hall,

142: Berkeley, CA 94720-7304}

143: \email{cmckee@astro.berkeley.edu}

144: 

145: \author{Jason Tumlinson}

146: \affil{Yale Center for Astronomy and Astrophysics, Yale University, PO Box 208121, New Haven, CT 06520 and Space Telescope Science Institute, 3700 San Martin Dr., Baltimore, MD 21218}

147: \email{tumlinson@stsci.edu}

148: 

149: \begin{abstract}

150: In this series of papers we study the structure of the atomic to molecular transition in

151: the giant atomic-molecular complexes that are the repositories of most molecular

152: gas in galaxies, with the ultimate goal of attaining a better understanding of what determines

153: galaxies' molecular content. Here we derive an approximate analytic solution for the

154: structure of a photodissociation region (PDR) in a cloud of finite

155: size that is bathed in an external dissociating radiation field. Our solution extends previous work, which

156: with few exceptions has been restricted to a one-dimensional treatment of the radiation field. We

157: show that our analytic results compare favorably to exact numerical

158: calculations in the one-dimensional limit. However, our more general geometry

159: provides a more realistic representation than a semi-infinite slab

160: for atomic-molecular complexes exposed to the interstellar radiation

161: field, particularly in environments such as low-metallicity dwarf

162: galaxies where the curvature and finite size of the atomic envelope

163: cannot be neglected. For clouds that are at least 20\% molecular we obtain analytic expressions for the molecular fraction in terms of properties of the gas and radiation field that are accurate to tens of percent, while for clouds of lower molecular content we obtain upper limits. As a side benefit, our analysis helps

164: clarify when self-shielding is the dominant process in H$_2$

165: formation, and under what circumstances shielding by dust makes a

166: significant contribution.

167: \end{abstract}

168: 

169: \keywords{ISM: clouds --- ISM: molecules --- molecular processes ---

170: radiative transfer}

171: 

172: \section{Introduction}

173: 

174: In galaxies such as the Milky Way, where atomic and molecular phases

175: of the interstellar medium coexist, molecular clouds represent the

176: inner parts of atomic-molecular complexes \citep{elmegreen87}. The

177: bulk of the volume of

178: the interstellar medium is filled with far-ultraviolet (FUV) photons

179: capable of dissociating hydrogen molecules, and this radiation field

180: keeps the majority of the gas atomic. Gas that is predominantly

181: molecular is found only in dense regions where a combination of

182: shielding by dust grains and self-shielding by hydrogen molecules

183: excludes the interstellar FUV field. These molecular regions are

184: bounded by a photodissociation region (PDR) in which the gas is

185: predominantly atomic \citep[and references therein]{hollenbach99}.

186: 

187: To date most work on the structure of PDRs has been limited to one-dimensional geometries, including unidirectional or bidirectional beams of radiation impinging on semi-infinite slabs or purely radial radiation fields striking the surfaces of spheres \citep[e.g.][]{federman79, vandishoeck86, 

188: black87, sternberg88, elmegreen93, draine96,

189: hollenbach99, browning03, allen04}. For these one-dimensional problems the

190: literature contains both detailed numerical solutions and analytic

191: approximations for the problem of radiative transfer and

192: H$_2$ formation-dissociation equilibrium. These approaches yield good

193: results when the PDR is thin compared to the cloud as a whole, or for PDRs that are in close proximity to hot stars whose radiation and winds have compressed the PDR into a slab-like geometry. Many nearby well-studied PDRs, such as the Orion bar, fall into this latter category. However, the one-dimensional approximation is much less appropriate for giant clouds being dissociated by the combined starlight of many distinct stars and star clusters, particularly when the atomic region constitutes a significant fraction of the total cloud volume. The problem is especially severe in galaxies with low

194: metallicities and interstellar pressures, where the predominantly

195: molecular parts of cloud complexes generally constitute a small part

196: of the total mass and volume \citep[e.g.][]{blitz06b}. In this case

197: one cannot neglect either the curvature of the PDR or the finite

198: size of the molecular region, and a higher-dimensional approach is preferable. Previous work in one dimension is therefore of limited use for the problem on which we focus: determining the atomic and molecular content of galaxies on large scales, under the combined effects of all the sources of dissociating radiation in that galaxy.

199: 

200: Considerably less work has focused on higher-dimensional

201: geometries, since these require a treatment of the angular dependence of the radiation

202: field and its variation with position inside a cloud. As a result all

203: treatments of two- or three-dimensional radiation fields to date are purely numerical.

204: \citet{neufeld96} consider spherical clouds and

205: \citet{spaans97} allow arbitrary geometries, but their method

206: applies only in translucent clouds, and involves an approximate

207: numerical integration of the transfer equation. Similarly,

208: \citet{liszt00} and \citet{liszt02} present models for PDRs in

209: spherical clouds involving angular integration over the radiation

210: field in radial bins, coupled with a relaxation method to determine the

211: H$_2$ abundance at each radius. Neither of these approaches yield a

212: simple analytic estimate of the size of the PDR or the molecular

213: region, nor do they provide any insight into the dimensionless numbers

214: that can be used to characterize the problem of PDR structure. Such estimates, and the accompanying physical insights, would allow modeling of clouds over a wide range of galactic environments without the need for a complex and numerically costly radiative transfer calculation to cover each case.

215: 

216: Our goal in this work is to revisit the problem of determining the

217: size of a PDR in a finite cloud embedded in a multi-dimensional radiation field, and to derive analytic

218: approximations for the structure of a PDR that will yield gross yet observable

219: quantities such as the total atomic hydrogen column around a molecular

220: region and the fraction of a cloud's volume in the atomic and

221: molecular phases. As part of this work we determine the important

222: dimensionless numbers that characterize the problem, and we provide a rough

223: classification of PDRs based on them. The model we develop is capable

224: of spanning the range from galaxies where the gas in atomic-molecular

225: complexes is predominantly molecular and a slab treatment is

226: appropriate to dwarf galaxies where only a tiny fraction of the ISM

227: resides in the molecular phase. In future work (Krumholz, McKee, \&

228: Tumlinson, in preparation), we provide a more detailed application of

229: the results derived here to the problem of determining the atomic to

230: molecular ratios in galaxies. Before moving on, we do note that our focus on an analytic solution with a multi-dimensional radiation field, characterized by a few dimensionless numbers, has a price: our approach to the chemical and thermal physics of PDRs is significantly simpler than much previous work. We do not account for factors such as the temperature-dependence of rate coefficients or H$_2$ dissociation by cosmic rays. Our work is therefore less suited to making detailed predictions of the structures of individual PDRs than it is to making predictions for galactic-scale trends in atomic and molecular content.

231: 

232: 

233: We approach the problem of finite clouds by idealizing to the case of

234: a spherical cloud embedded in an isotropic radiation field, since this

235: allows us to explore the effects of finite cloud size and curvature

236: while at the same time keeping the problem simple enough to admit an

237: approximate analytic solution. Our approach is as follows: in

238: \S~\ref{formal} we state the formal problem and introduce some

239: physical approximations that are independent of geometry. In

240: \S~\ref{slab} we derive an approximate analytic solution to the

241: one-dimensional semi-infinite slab case, which allows us to

242: demonstrate the underlying physical principles of our approach. In

243: this section we also compare to a grid of numerical solutions and show

244: that our approach produces good agreement. Then in \S~\ref{sphere} we

245: extend our approach to handle the case of a spherical cloud embedded

246: in an isotropic radiation field. Finally we summarize and draw

247: conclusions in \S~\ref{conclusion}.

248: 

249: 

250: \section{The Formal Problem}

251: \label{formal}

252: 

253: Consider a region of hydrogen gas where the number density of hydrogen nuclei is $n$, mixed with

254: dust which has a cross section $\sigma_{{\rm d},\nu}$ per H nucleus

255: to radiation of frequency $\nu$. The hydrogen is a mix of atoms and

256: molecules, with a fraction $\fatm$ of the nuclei in the form of H\textsc{i} and

257: a fraction $\fmol=1-\fatm$ in the form of H$_2$. We consider

258: frequencies $\nu$ that fall within the Lyman-Werner (LW) band from

259: $\nu_1=c/1120\mbox{ \AA}$ to $\nu_2=c/912\mbox{ \AA}$, such that photons

260: of that frequency can be resonantly absorbed by hydrogen

261: molecules.

262: 

263: The equation of radiative transfer for a beam of radiation

264: in direction $\ehat$ passing through this gas is

265: \begin{equation}

266: \label{transferfreq}

267: \ehat \cdot \nabla I_{\nu} = -n \left(\frac{1}{2} \fmol \sigma_{{\rm

268: H}_2,\nu} + \sigma_{{\rm d},\nu}\right) I_{\nu},

269: \end{equation}

270: where $I_{\nu}$ is the radiation intensity at frequency $\nu$ and

271: $\sigma_{{\rm H}_2,\nu}$ is the cross section for absorption of

272: radiation at frequency $\nu$ by a molecule of hydrogen. The value of

273: $\sigma_{{\rm H}_2,\nu}$ may change with position as the fraction of

274: H$_2$ molecules in different quantum states changes. The

275: total fraction of the gas in the molecular phase is determined by the

276: balance between the rate of H$_2$ formation and dissociation,

277: \begin{equation}

278: \label{dissocfreq}

279: \fatm n^2 \calr

280: =

281: \frac{\fmol}{2} n \int d\Omega\, \int_{\nu_1}^{\nu_2} d\nu\,

282: \frac{I_{\nu}}{h \nu} \sigma_{{\rm H}_2,\nu} f_{{\rm diss},\nu}.

283: \end{equation}

284: where $\calr$ is the rate coefficient for formation of H$_2$ molecules

285: on dust grain surfaces and $f_{{\rm

286: diss},\nu}$ is the fraction of absorptions at frequency $\nu$ that

287: yield dissociation of the H$_2$ molecule rather than decay back to a

288: bound state.

289: 

290: Note that we do not include a source term in the transfer

291: equation (\ref{transferfreq}), because although most FUV photons

292: absorbed by H$_2$ molecules do decay through a vibrational ladder via

293: photon emission, the photons released in this process do not fall into

294: the LW band. Thus, the transfer equation we have written is only valid

295: for frequencies in the LW band. We have also neglected scattering of

296: FUV photons by dust grains. Since scattering is highly forward-peaked

297: at FUV wavelengths \citep[e.g.][]{roberge81}, this approximation is

298: reasonable as long as we take $\sigma_{{\rm d},\nu}$ to be the

299: absorption cross section, not the total cross section. We have also omitted other H$_2$ dissociation mechanisms other than LW photons, such as cosmic ray collisions and chemical reactions. These are significant only in nearly fully molecular regions where there are no significant numbers of LW photons present.

300: 

301: Equations (\ref{transferfreq}) and (\ref{dissocfreq}), together with

302: the atomic and dust physics that specify $\sigma_{{\rm H}_2,\nu}$ and

303: $\sigma_{{\rm d},\nu}$ and a boundary condition that specifies

304: $I_{\nu}$ on all rays entering the surface of a cloud, fully

305: determine $I_{\nu}$ and $\fmol$ at all positions. We cannot

306: solve them exactly, but we can obtain an approximation that exposes

307: the basic physical outlines of the solution. We begin by making two

308: standard approximations, following \citet{draine96}, to simplify the

309: atomic physics.

310: 

311: First, it is convenient to simplify the transfer equation

312: (\ref{transferfreq}) by dividing

313: by $h\nu$ to transform from intensity to photon number, and then by

314: integrating over frequency in the LW band. In so doing we can exploit

315: the fact that for realistic dust $\sigma_{{\rm d},\nu}$ is nearly

316: independent of frequency in the LW band \citep{draine96} to replace

317: $\sigma_{{\rm d},\nu}$ with a constant value $\sigmad$. Doing so gives

318: \begin{equation}

319: \label{transfer}

320: \ehat \cdot \nabla \Istar =  - n \sigmad \Istar -\frac{1}{2} n \fmol

321: \int_{\nu_1}^{\nu_2} d\nu\, \sigma_{{\rm H}_2,\nu} \Istar_{\nu},

322: \end{equation}

323: where $\Istar_{\nu}=I_{\nu}/(h\nu)$ is the photon number intensity,

324: i.e.\ the number of photons per unit

325: time per unit area per unit solid angle per unit frequency that cross

326: a given surface, and $\Istar=\int_{\nu_1}^{\nu_2} d\nu\, \Istar_{\nu}$ is

327: the photon number intensity integrated over the LW band.

328: 

329: Second, we note that $f_{{\rm diss},\nu}$ varies only weakly when

330: integrated over frequency and over position within a

331: PDR. \citet{draine96} show that over the width of a PDR it stays roughly

332: within the range $0.1-0.2$. Its value in free space depends on the assumed radiation spectrum; \citet{draine96} find $\fdissoc=0.12$ in free space for their fiducial choice, while \citet{browning03} suggest

333: $\fdissoc=0.11$ as a typical value. For simplicity we adopt a constant

334: value $f_{{\rm diss},\nu} = \fdissoc=0.1$ and

335: take this constant out of the integral, reducing the dissociation

336: equation to

337: \begin{equation}

338: \label{dissoc}

339: \fatm n^2 \calr = 

340: \frac{\fmol \fdissoc}{2} n \int d\Omega\, \int_{\nu_1}^{\nu_2} d\nu\,

341: \Istar_{\nu} \sigma_{{\rm H}_2,\nu}.

342: \end{equation}

343: 

344: It is convenient at this point to produce a combined

345: transfer-dissocation equation from (\ref{transfer}) and

346: (\ref{dissoc}). If we integrate equation (\ref{transfer}) over solid

347: angle $d\Omega$, we obtain

348: \begin{equation}

349: \nabla \cdot \bFstar = -n \sigmad c \Estar -

350: \frac{1}{2} \fmol n \int

351: d\Omega\, \int_{\nu_1}^{\nu_2} d\nu\, \sigma_{{\rm H}_2,\nu}

352: \Istar_{\nu},

353: \end{equation}

354: where

355: \begin{eqnarray}

356: \Estar & \equiv & \frac{1}{c} \int d\Omega \, \Istar \\

357: \bFstar & \equiv & \int d\Omega \, \ehat \Istar

358: \end{eqnarray}

359: are the photon number density and photon number flux integrated over

360: the LW band, respectively. We can then use equation (\ref{dissoc}) to

361: substitute for the last term, yielding the combined

362: transfer-dissociation equation

363: \begin{equation}

364: \nabla \cdot \bFstar = -n \sigmad c \Estar

365: - \frac{\fatm n^2 \calr}{\fdissoc}.

366: \label{transdissoc}

367: \end{equation}

368: 

369: \section{Solution in One Dimension}

370: \label{slab}

371: 

372: \subsection{Analytic Solution}

373: \label{slabanalyt}

374: 

375: We start by giving an approximate analytic solution to this formal

376: problem for unidirectional radiation impinging on a one-dimensional

377: semi-infinite slab in order to illustrate the physical principles

378: behind our approach. Consider a region of gas of density $n$ filling

379: the half-space $z>0$, subjected to a dissociating radiation field of

380: photon number intensity $\Istar = 4\pi \Jstar_0

381: \delta(|\ehat-\hat{z}|)$ that fills the half-space $z<0$, where $\Jstar_0$ is the angle-averaged intensity in free space. The corresponding free-space photon number density is $\Estar_0=4\pi \Jstar_0/c$, and the magnitude of the free-space photon flux is $\Fstar_0=c\Estar_0$. For simplicity we neglect the (relatively weak) temperature-dependence of $\calr$.

382: 

383: Since the

384: radiation intensity everywhere at all $z$ remains proportional to

385: $\delta(|\ehat-\hat{z}|)$, it immediately follows that

386: \begin{equation}

387: \bFstar = \Fstar \hat{z} = c \Estar \hat{z},

388: \end{equation}

389: at all points, and the combined transfer-dissociation equation reduces to

390: \begin{equation}

391: \label{slabeqn}

392: \frac{d\Fstar}{dz} = -n \sigmad \Fstar - \frac{\fatm n^2

393: \calr}{\fdissoc},

394: \end{equation}

395: subject to the boundary condition that $\Fstar = \Fstar_0$ at

396: $z=0$. Since numerical calculations show that the transition from

397: predominantly atomic gas to predominantly molecular gas in a PDR generally occurs in thin band bounded by much larger regions where the gas is either predominantly atomic or predominantly molecular, we can obtain a good approximation to the exact

398: solution by treating $\fatm$ as having a constant value near unity

399: over the bulk of the PDR, and then dropping to zero as a step function once the fully molecular surface is reached. For constant $\fatm=1$, we can

400: non-dimensionalize equation (\ref{slabeqn}) to

401: \begin{equation}

402: \label{slabnondim}

403: \frac{d\calf}{d\tau} = -\calf - \frac{1}{\chi},

404: \end{equation}

405: where $\calf=\Fstar/(\Fstar_0)$ is the fraction of the free-space

406: flux remaining, $\tau=n\sigmad z$ is the dust optical depth from the

407: slab surface, and 

408: \begin{equation}

409: \chi = \frac{\fdissoc \sigmad c \Estar_0}{n \calr}.

410: \end{equation}

411: Equation (\ref{slabnondim}) has the exact solution

412: \begin{equation}

413: \calf(\tau) = \frac{1}{\chi} \left[e^{-(\tau-\taum)}-1\right],

414: \end{equation}

415: where

416: \begin{equation}

417: \label{slabexact}

418: \taum = \ln\left(1+\chi\right)

419: \end{equation}

420: is the depth at which the flux goes to zero, which we take to be the optical depth through the H\textsc{i} region.

421: Of course in reality the flux should never go to zero exactly. That is

422: does in our solution is an artifact of our choice to treat $\fatm$

423: as constant. Nonetheless, since the transition from $\fatm\approx 1$ to

424: $\fatm \approx 0$ is sharp,

425: $\zm=\taum/(n\sigmad)$ should be a good approximation of the depth at

426: which the gas becomes predominantly molecular.

427: 

428: The dimensionless parameter $\chi/\fatm$

429: is the ratio of the two terms on the right hand side of equation

430: (\ref{slabeqn}) with $\Fstar$ set equal to its value $4\pi\Istar_0$ at

431: the slab edge. This makes its physical meaning clear: $\chi/\fatm$

432: represents the ratio of the absorption rate of LW photons by dust

433: grains to the absorption rate by H$_2$ molecules for a parcel of gas

434: exposed to the unattenuated free space radiation field. If the gas at

435: the edge of free space is predominantly atomic, as is the case for

436: example at the edge of an atomic-molecular complex, then $\fatm\approx 1$

437: and this ratio is simply given by $\chi$. For $\chi>1$ absorptions by

438: dust grains dominate, while for $\chi<1$ absorptions by H$_2$

439: molecules dominate.

440: 

441: For a giant molecular cloud in the Milky Way and its outer

442: atomic envelope, typical values of the number density, dust cross

443: section, and H$_2$ formation rate coefficient are $n\sim 30$

444: cm$^{-3}$, $\sigmad\sim 10^{-21}$ cm$^2$ and $\calr\sim

445: 3\times 10^{-17}$ cm$^3$ s$^{-1}$ \citep{draine96}. Using the \citet{draine78} functional form for the local FUV radiation energy density as a function of wavelength, 

446: \begin{eqnarray}

447: \lambda E_{\lambda} & = & 6.84\times 10^{-14} \lambda_3^{-5}

448: \left(31.016\lambda_3^2-49.913\lambda_3+19.897\right)

449: \nonumber \\

450: & & \qquad \mbox{ erg cm}^{-3}

451: \end{eqnarray}

452: the free-space photon number density from $912-1120$ \AA\ is $\Estar_0=7.5\times 10^{-4}$

453: cm$^{-3}$. For an H$_2$ molecule in the ground state, this corresponds to a free-space dissociation rate of $3.24\times 10^{-11}$ s$^{-1}$. (In principle for a slab computation we should divide the observed value of $\Estar_0$ by two to account for the fact that one can only see half the sky at the surface of an opaque cloud, but we do not do so here because in \S~\ref{sphere} we will account for this effect self-consistently.) Thus, for Milky Way conditions not near a local strong source of FUV, $\chi$ of order a few might be typical. Thus, in the Milky Way dust shielding

454: is marginally significant in determining the structure of atomic-molecular complexes.

455: 

456: \subsection{The Two-Zone Approximation}

457: \label{slabapprox}

458: 

459: We can integrate the transfer-dissociation equation

460: (\ref{transdissoc}) directly in one dimension because, due to the

461: constant angular distribution of the radiation, there is a trivial

462: relationship between $\Estar$ and $\Fstar$. In multiple dimensions,

463: however, there is no simple relationship between the two, because the

464: angular distribution of the radiation intensity is not constant with

465: position inside a cloud. To overcome this problem, we adopt what we

466: call the \textit{two-zone approximation}. When the photon number density $\Estar$ is large, the

467: first term on the right hand side of equation (\ref{transdissoc}),

468: representing absorptions of photons by dust grains, is much larger

469: than the second term, representing absorptions by H$_2$

470: molecules. This makes intuitive sense: in regions where there many

471: dissociating photons present, the molecular fraction

472: will be very low, so there will be few H$_2$ molecules available to

473: absorb LW photons and most photons will be absorbed by dust. In regions

474: where $\Estar$ is small, the number of molecules will increase, and

475: for any appreciable number of molecules these will dominate the

476: absorption rate.

477: 

478: In the two-zone approximation, we divide the cloud

479: into a zone where dust absorption dominates and a zone where molecular

480: absorption dominates. In the dust-dominated region we drop the molecular absorption term in the radiative transfer or transfer-dissociation equations (equations \ref{transfer} or \ref{transdissoc}), and

481: approximate the opacity as having a constant value $n\sigmad$. In the zone where molecular absorptions dominate, we drop the dust absorption term in equation (\ref{transdissoc}) and approximate the molecular absorption term by $\phi n^2 \calr/\fdissoc$, where $\phi>1$ is a constant of order unity, whose precise value we determine below, that we include to account for the fact that some LW photons will be absorbed by dust grains even in the molecular-dominated region. We define a boundary between these two zones by the condition that the dust and molecular absorption  terms be equal, which is satisfied when

482: \begin{equation}

483: \label{dustcondition}

484: \frac{\Estar}{\Estar_0} = \frac{\phi}{\chi} \equiv \frac{1}{\psi},

485: \end{equation}

486: where we have set $\fatm=1$ because at the

487: point of equality the molecular fraction is $\ll 1$, and for convenience we have defined the modified dust to molecular absorption ratio $\psi = \chi/\phi$. With this approximation, the one-dimensional non-dimensionalized transfer-dissocation equation becomes

488: \begin{equation}

489: \label{slabtwozoneeq}

490: \frac{d\calf}{d\tau} = -\left\{

491: \begin{array}{lr}

492: \calf, & \calf > 1/\psi \\

493: 1/\psi, & \calf < 1/\psi

494: \end{array}

495: \right..

496: \end{equation}

497: 

498: We shall see in \S~\ref{sphere} how the two-zone approximation enables

499: us to solve the problem in the spherical case. First, though, we examine the solution in the one-dimensional case. If $\psi<1$, then $\calf < 1/\psi$ is satisfied everywhere and equation (\ref{slabtwozoneeq}) has the trivial solution

500: \begin{equation}

501: \calf=\frac{\psi-\tau}{\psi}.

502: \end{equation}

503: The flux goes to zero at a depth $\taum = \psi$. If $\psi > 1$, the solution is

504: \begin{equation}

505: \calf = \left\{

506: \begin{array}{lr}

507: e^{-\tau}, & \tau < \taud \\

508: \psi^{-1} (\taum-\tau)/(\taum-\taud), & \tau > \taud

509: \end{array}

510: \right.,

511: \end{equation}

512: with

513: \begin{eqnarray}

514: \taud & = & \ln \psi \\

515: \taum & = & 1 + \ln \psi.

516: \end{eqnarray}

517: Here $\taud$ represents the dust depth into the slab at which the absorption

518: begins to be dominated by H$_2$ molecules, while $\taum$ is the

519: optical depth where we expect a transition from mostly atomic to

520: mostly molecular gas. Combining the two cases, we have

521: \begin{equation}

522: \label{slabtwozone}

523: \taum = \left\{

524: \begin{array}{lr}

525: \psi, & \psi < 1 \\

526: 1+\ln\psi, & \psi > 1

527: \end{array}

528: \right..

529: \end{equation}

530: 

531: We now turn to the question of determining the constant $\phi$. Physically, we expect to have $\phi\rightarrow 1$ for $\chi \ll 1$, because in that case dust absorptions contribute negligibly throughout the cloud. For $\chi \gg 1$ we expect to have $\phi$ to asymptote to a value greater than unity, accounting for the contribution of dust to absorptions even in the molecular-dominated region. A comparison of the limiting behavior of the analytic solution (\ref{slabexact}) with the two-zone approximation (\ref{slabtwozone}) confirms this physical argument, and suggests that the appropriate limiting behavior is $\phi\rightarrow 1$ as $\chi\rightarrow 0$ and $\phi\rightarrow e$ as $\chi\rightarrow \infty$. We therefore adopt

532: \begin{equation}

533: \phi = \frac{2.5+\chi e}{2.5+\chi},

534: \end{equation}

535: which has the correct limiting behavior, and where the value $2.5$ is chosen to optimize agreement between the two-zone approximation and the analytic solution in the intermediate $\chi$ region.

536: 

537: \subsection{Comparison to Numerical Calculations}

538: \label{numeric}

539: 

540: Before using the two-zone approximation to compute the case of a

541: finite cloud with an isotropic radiation field, we check its accuracy

542: for the one-dimensional case by comparing with detailed numerical

543: calculations using the \citet{browning03} H$_2$ formation and

544: radiative transfer code. We refer readers to that paper for a full

545: description of the physics included in this calculation, but a brief

546: summary is that the code numerically integrates the

547: frequency-dependent equation of radiative transfer for a

548: unidirectional beam of radiation incident on an isothermal,

549: constant-density slab of gas mixed with dust. The transfer equation is

550: coupled to a statistical equilibrium calculation that determines the

551: populations of H\textsc{i} atoms and a large number of rotational and

552: vibrational levels of the H$_2$ molecule that are excited by LW band photons

553: in each computational

554: cell. The output of this calculation is the fraction of H nuclei in

555: molecules as a function of depth within the cloud. The code we use here differs from that described in \citet{browning03} only in that the earlier version accounted for absorptions of LW photons by dust grains by modifying the photodissociation rate using the method of \citet{vandishoeck86}, whereas the version we use here computes radiation attenuation by dust grains directly from the radiative transfer equation.

556: 

557: For the models we present here we use a density and temperature of

558: $n=5000$ cm$^{-3}$ and $T=90$ K. These values are chosen purely for

559: computational convenience, and have no significant impact on the

560: results. The incident radiation field is a unidirectional beam of photons uniformly distributed in frequency over the wavelength range $912-1120$ \AA. The frequency-dependent photon flux in this beam is $\Fstar_{\nu}$, so $\Estar_0=\Fstar_{\nu} (\nu_2-\nu_1)/c$ and $\Jstar_0=\Fstar_{\nu} (\nu_2-\nu_1)/(4\pi)$.

561: We adopt a dust extinction curve following

562: the functional form of \citet{cardelli89a}, scaled to give a dust cross

563: section per H nucleus at 1000 \AA\ of $\sigma=\sigma_{\rm d,MW} Z'$, where $Z'$

564: is the metallicity relative to solar and we take $\sigma_{\rm d,MW} = 6.0\times

565: 10^{-22}$ cm$^2$ or $2.0\times 10^{-21}$ cm$^2$ to be two fiducial dust opacities for the Milky Way. These two values of $\sigma_{\rm d,MW}$ correspond to the estimated attenuation cross sections at 1000 \AA\ estimated by \citet{draine96} for dense and diffuse clouds Milky Way, respectively. We adopt a

566: rate coefficient for H$_2$ formation on grain surfaces of $\calr=\calr_{\rm MW} Z'$ with $\calr_{\rm MW}=3\times 10^{-17} Z'$ cm$^{-3}$ s$^{-1}$ as our fiducial Milky Way value \citep{wolfire08}. We do our computations for a grid of models running from $\Fstar_{\nu}=10^{-7}-10^{-3}$ photons

567: cm$^{-2}$ s$^{-1}$ Hz$^{-1}$

568: in steps of nine steps of 0.5 dex, from $Z'=10^{-2}-10^{0.5}$ in six steps of

569: 0.5 dex, and for the two values of

570: $\sigma_{\rm d,MW}$ mentioned above. These values $\Fstar_{\nu}$ are significantly higher than are typical in the Milky Way, but are chosen so that, in conjunction with our choice of $n$, the ratio $\Estar_{\nu}/n$ that appears in $\chi$ is within the typical Milky Way range. With this parameterization

571: \begin{eqnarray}

572: \chi & = & 0.75 (\sigma_{\rm d,MW,-21}/\calr_{\rm MW}) (E_0^{*'} / n_2) \\

573: & = & 4.07 \sigma_{\rm d,MW,-21} \Fstar_{\nu,-5}

574: \end{eqnarray}

575: where $\sigma_{\rm d,MW,-21}=\sigma_{\rm d,MW}/10^{-21}$ cm$^2$ is the Milky Way 1000 \AA\ dust absorption opacity normalized to $10^{-21}$ cm$^2$, $E_0^{*'}=E_0^*/7.5\times 10^{-4}$ cm$^{-3}$ is the free-space dissociating photon number density normalized to the Milky Way value, $n_2 = n/100$ cm$^{-3}$ is the number density of hydrogen nuclei in units of 100 cm$^{-3}$, and 

576: $\Fstar_{\nu,-5}=\Fstar_{\nu}/10^{-5}$ photons cm$^{-2}$ s$^{-1}$ Hz$^{-1}$.

577: Thus the

578: calculation covers a broad range of parameters from strongly

579: dust-dominated to strongly molecular-dominated, thereby bracketing the real Milky Way value of $\chi\sim 1$. Note that $\chi$ is

580: independent of $Z'$ because for the parameterization we have chosen the $Z'$-dependences of $\sigmad$ and $\calr$ cancel. Since we predict that the dust optical depth through

581: the PDR, $\taum = n\sigmad \zm$, depends only on $\chi$, and $\chi$ in turn depends only on the ratio $\sigmad/\calr$, $\taum$ should be independent of $Z$. Since

582: we use a range of $10^{2.5}$ in $Z'$, our numerical calculations represent a

583: strong test of this prediction.

584: 

585: \begin{figure}

586: %\plotone{fh2prof.pdf}

587: \plotone{f1.eps}

588: \caption{

589: \label{fh2prof}

590: The plots show $\fmol$ versus dust optical depth $\tau=n\sigmad z$ for

591: our numerical radiative transfer calculations with $\log Z'=-1.5$

592: (\textit{dashed lines}), $\log Z'=-0.5$ (\textit{solid lines}), and

593: $\log Z'=0.5$ (\textit{dot-dashed lines}). The gray vertical lines

594: indicate the optical depth of the transition to fully molecular as

595: calculated with the two-zone approximation, equation

596: (\ref{slabtwozone}). Each cluster of three

597: curves plus a vertical line indicating a prediction

598: corresponds to a radiation flux $\log F_{\nu}^*=-7$, $-5$, or

599: $-3$, as indicated. The two panels are for the cases

600: $\sigma_{\rm d,MW,-21}=0.6$ and $2.0$, as indicated.

601: }

602: \end{figure}

603: 

604: \begin{figure}

605: %\plotone{hicolcomp.pdf}

606: \plotone{f2.eps}

607: \caption{

608: \label{hicolcomp}

609: The upper panel shows the dust opacity $\nhi\sigmad$

610: through the numerically-determined H\textsc{i} column (\textit{various

611: symbols}) as a function of $\chi$. The values of $Z'$ and

612: $\sigma_{\rm d,MW,{-21}}$ for each calculation are indicated by the plot

613: symbol. This is compared to the optical depth $\taum$ computed from

614: the two-zone approximation (equation \ref{slabtwozone}, \textit{thin solid

615: line}) and computed using the analytic solution (equation \ref{slabexact},

616: \textit{thick dashed line}). The lower panel shows the error in the two-zone

617: approximation, defined as $\mbox{Error} = \nhi\sigmad/\taum-1$; the dotted line indicates zero error.

618: }

619: \end{figure}

620: 

621: Figure \ref{fh2prof} shows $\fmol$ versus depth within a cloud as

622: computed numerically for a sample of our input parameters, overlaid

623: with the corresponding locations of the atomic to molecular transition

624: as calculated via the two-zone approximation. As the figure shows, the

625: two-zone approximation does a very good job reproducing the location

626: of this transition over an extremely broad range of parameters.

627: 

628: To quantify the quality of the approximation, we must define a fiducial measure for the depth of the H$_2$ region in the numerical calculations, since in these runs $\fatm$ approaches but never reaches unity. The most reasonable measure is

629: \begin{equation}

630: \label{nhiintegral}

631: \nhi = \int_0^{\infty} dz\, \fatm n,

632: \end{equation}

633: the total H\textsc{i} column integrated through the cloud. Since the radiation field is attenuated exponentially or faster, and  $\fatm$ is proportional to radiation intensity in the region where $\fmol\approx 1$, this integral is guaranteed to converge. In the limit where the transition from HI to H$_2$ is sharp it approaches the total gas column up to the transition point. In practice we cannot continue the numerical integration to $z=\infty$, so we truncate the integral at the value of $z$ where $\fatm=5\times 10^{-3}$; using $\fatm=10^{-2}$ instead changes the value by less than 8\% for all our runs, and by less than 2\% for all runs with $\chi>0.1$, so our evaluation of the integral should be accurate to this level.

634: 

635: We plot the dust opacity through this hydrogen column, $\nhi\sigmad$, and the corresponding value $\taum$ predicted by

636: the two-zone approximation, in Figure \ref{hicolcomp}. As the figure shows, the two-zone

637: approximation recovers the numerically-computed H\textsc{i} column to

638: better than 50\% accuracy over almost a five-decade range in

639: $\chi$. The error in the two-zone approximation is generally

640: comparable to or smaller than the spread between models with different

641: dust opacities but the same value of $\chi$. 

642: 

643: The error in our approximation is largest at small $\chi$, and examination of Figure \ref{fh2prof} suggests the reason why: by evaluating the equations with $\fatm=1$ inside the PDR, we have assumed that the transition from atomic to molecular is sharp. This is true for $\chi\sim 1$ or greater, but begins to fail for $\chi \ll 1$. In our runs with $\chi\ge 1$, typically 95\% of the gas is atomic in the region where $\fatm>0.5$; and even at the depth where $\fatm$ drops to $5\times 10^{-3}$ more than half the gas column above that point is atomic. This indicates a very sharp atomic-molecular transition, so our approximation that $\fatm=1$ until the gas is almost entirely molecular is a good one. For $\chi \approx 0.01$, on the other hand, roughly 80\% of the gas at $\fatm<0.5$ is atomic, and H\textsc{i} contributes only 10\% of the total gas column above $\fatm=5\times 10^{-3}$. The transition from atomic to molecular is therefore much more gradual, and our accuracy suffers as a result.

644:  

645: Nonetheless, we note that $\chi \ll 1$ does not appear to be physically realized in normal galactic environments. For Milky Way molecular clouds $\chi\sim 1$ or greater, and reducing $\chi$ to 0.01 would require some combination of reducing the ISRF and increasing the atomic gas density by a factor of 100. Such a combination of very high atomic ISM density and very low radiation field is generally not observed. We conclude that, for realistic physical parameters, and given that these parameters (such as

646: $\sigma_0$ and $\calr$) are themselves uncertain at the factor of a

647: few level \citep[e.g.][]{wolfire08}, the error in the two-zone

648: approximation is unlikely to be the dominant one.

649: 

650: 

651: \section{Solution for Spherical Clouds}

652: \label{sphere}

653: 

654: We now extend the two-zone approximation to a spherical cloud of

655: radius $R$ embedded in a uniform, isotropic radiation field of

656: angle-averaged intensity $\Jstar_0$. (Note that this radiation field has the same LW

657: photon number density as the unidirectional radiation field considered

658: in \S~\ref{slab}, so it gives the same dissociation rate in free

659: space.) Figure \ref{diagram} illustrates the

660: basic geometry of the problem and our approximation: we consider the

661: dust-dominated region to run from radius $r=\rd$ to $r=R$, and the

662: molecular self-shielding region to run from $r=\rmol$ to $r=\rd$. For

663: convenience we introduce the dimensionless position variables $x=r/R$

664: and $y=1-x$, and we define the dust optical depths from the surface to $\rd$ and to $\rmol$ as $\taud$ and $\taum$, respectively.

665: 

666: In \S~\ref{spheredust} and \S~\ref{spherenodust}, we develop the basic

667: equations that describe the two-zone approximation for clouds with and

668: without dust opacity-dominated envelopes. We then explore three

669: limiting cases of these equations. We consider the behavior at the

670: boundary between the presence and absence of a dust-dominated

671: zone in \S~\ref{dustboundary}, and we explore several interesting limits in \S~\ref{limitsection}.

672: We then give a

673: numerical solution and an analytic approximation to it in

674: \S~\ref{numericsol}. In \S~\ref{planarcomp} we compare our solution

675: for a finite cloud to the standard slab approximation, to determine

676: when the slab approximation is valid and when it fails. In \S~\ref{error},

677: we address the level of uncertainty introduced by the approximations we 

678: make in the spherical case. Finally,

679: in \S~\ref{example} we present some example calculations using our

680: analytic approximation.

681: 

682: 

683: \begin{figure}

684: \epsscale{0.7}

685: \plotone{f3.eps}

686: \epsscale{1.0}

687: \caption{

688: \label{diagram}

689: Illustration of the two-zone approximation in spherical geometry.\\\\

690: }

691: \end{figure}

692: 

693: 

694: \subsection{Clouds with Dust-Dominated Zones}

695: \label{spheredust}

696: 

697: First consider the case where $\psi$ is large enough so that there

698: is a dust-dominated zone in the outer part of the cloud where

699: molecular self-shielding is negligible, i.e.\ $\rd < R$. The transfer

700: equation in this region becomes

701: \begin{equation}

702: \label{transferdust}

703: \ehat \cdot \nabla \Istar =  - n \sigmad \Istar,

704: \end{equation}

705: which for rays originating at the cloud surface and staying entirely

706: within the dust-dominated region has the trivial solution

707: \begin{equation}

708: \Istar(x, \mu) = \exp(-\tauR\xi) \Jstar_0,

709: \end{equation}

710: where $\tauR=n\sigmad R$ is the center-to-edge dust optical depth of

711: the cloud,

712: \begin{equation}

713: \label{xidef}

714: \xi = \left(\sqrt{1 - x^2 + x^2\mu^2} - x\mu\right),

715: \end{equation}

716: is the distance, normalized to the cloud radius, from radius $r$ to

717: the cloud surface on a ray that makes an angle $\theta$ relative to

718: the radial vector (see Figure \ref{diagram}), and

719: $\mu=\cos\theta=-\ehat\cdot\hat{r}$. This solution applies for $\mu >

720: 0$. On the other hand, if $\mu<\mumol$, 

721: \begin{equation}

722: \label{mumoleqn}

723: \mumol \equiv -\sqrt{1-\left(\frac{\xm}{\xd}\right)^2},

724: \end{equation}

725: then the ray passes through a part of the cloud that is fully

726: molecular. Since the fully molecular region will be extremely opaque,

727: to good approximation along these rays $\Istar(x,\mu)=0$.

728: Finally, rays for which $0 > \mu > \mumol$ pass through the region

729: where the opacity is dominated by molecules rather than dust,

730: but where the gas is not yet fully molecular. Again, we use the

731: approximation that the transition from a low molecular

732: fraction to fully molecular is sharp, so that over most of this region

733: the molecular fraction is not vastly larger than it is at the region's

734: edge. This is consistent with numerical solutions of the problem (e.g.\ Figure \ref{fh2prof}),

735: which show that the atomic fraction $\fatm$ is nearly constant through

736: the bulk of the PDR, and rises from its free space value to

737: unity over a small region. This means that the opacity is not much

738: greater than its value

739: of $n\sigmad$ at the outer edge of the molecule-dominated region. We

740: therefore approximate that the optical depth along these rays is the

741: same as for those in the dust-dominated region, $\tauR\xi$. This approximation is perhaps the least certain part of our calculation, and we quantify the level of uncertainty that it produces in \S~\ref{error}.

742: Combining these three regions, we have an

743: approximate intensity

744: \begin{equation}

745: \label{istarray}

746: \Istar(x, \mu) = \left\{

747: \begin{array}{lr}

748: \exp(-\tauR\xi) \Jstar_0, & \mu > \mumol \\

749: 0, & \mu < \mumol

750: \end{array}

751: \right..

752: \end{equation}

753: 

754: The location of $\xd$, the crossover point from dust- to

755: molecule-dominated absorption, is defined by the condition that

756: $\Estar/\Estar_0=1/\psi$ (equation \ref{dustcondition}). For

757: convenience we define the function $\eta_0$ by

758: \begin{equation}

759: \label{f0def}

760: \frac{\Estar(\xd)}{\Estar_0} = \eta_0(\xd, \xm; \tauR) = 

761: \frac{1}{2} \int_{\mumol}^1 d\mu\, \exp\left(-\tauR \xid \right),

762: \end{equation}

763: where $\xid=\xi(\xd)$.

764: We show that $\eta_0$ can be evaluated in terms of exponential

765: integrals in Appendix \ref{fintegrate}. The first equation for the

766: two-zone approximation is therefore

767: \begin{equation}

768: \label{f0eqn}

769: \eta_0(\xd,\xm;\tauR) = \frac{1}{\psi}.

770: \end{equation}

771: 

772: Inside $\xd$, we drop the dust opacity term, so that the combined

773: transfer-dissociation equation (\ref{transdissoc}) becomes

774: \begin{equation}

775: \frac{1}{r^2}\frac{d}{dr}(r^2\Fstar) = \frac{\phi n^2 \calr}{\fdissoc},

776: \end{equation}

777: where we have again assumed that $\fatm\approx 1$ outside the fully

778: molecular region, and for convenience we have inverted the sign by

779: defining $\bFstar=-\Fstar \hat{r}$. The solution is

780: \begin{equation}

781: \label{fluxsol}

782: \Fstar = \frac{\tauR}{3\psi} x

783: \left[1-\left(\frac{\xm}{x}\right)^3\right] \Fstar_0,

784: \end{equation}

785: where we have chosen the constant of integration by requiring that

786: $\Fstar=0$ at $x=\xm$. To determine $\xm$ from the boundary

787: conditions, however, we must determine the flux at some other

788: location. Thus we evaluate the flux $\Fstar$ at $\xd$ by integrating

789: the intensity over solid angle using equation (\ref{istarray}). For

790: convenience we define the function $\eta_1$ by

791: \begin{equation}

792: \label{f1def}

793: \frac{\Fstar(\xd)}{\Fstar_0} = \eta_1(\xd, \xm; \tauR) 

794: =\frac{1}{2} \int_{\mumol}^1 d\mu\,

795: \mu \exp\left(-\tauR \xid \right).

796: \end{equation}

797: As with $\eta_0$, in Appendix \ref{fintegrate} we evaluate $\eta_1$ in terms

798: of exponential integrals. Combining (\ref{f1def}) with (\ref{fluxsol})

799: gives an implicit equation for $\xm$:

800: \begin{equation}

801: \eta_1(\xd,\xm; \tauR) = 

802: \frac{\tauR \xd}{3 \psi} \left[1-\left(\frac{\xm}{\xd}\right)^3\right].

803: \label{f1eqn}

804: \end{equation}

805: Together, equations (\ref{f0eqn}) and (\ref{f1eqn}) constitute two

806: equations in the two unknowns $\xd$ and $\xm$, and thus fully

807: determine the location of the transition from predominantly atomic to

808: predominantly molecular in the two-zone approximation.

809: 

810: \subsection{Clouds without Dust-Dominated Zones}

811: \label{spherenodust}

812: 

813: Now consider the case where $\psi$ is small enough so that

814: there is only one zone, because even gas at the edge of the cloud is

815: sufficiently molecular for absorptions by molecules to outnumber those

816: by dust grains. In this case equation (\ref{fluxsol}) applies

817: throughout the cloud, so we must fix $\xm$ directly from the boundary

818: conditions. To do so we need to know the flux

819: $\Fstar(1)$ at the cloud surface. This is not simply

820: $\Fstar_0=c\Estar_0$ as in the case of a unidirectional radiation field; in free space for an isotropic radiation field $\Fstar$ vanishes, and $\Fstar(1)$ is non-zero only because rays

821: passing through the cloud do not carry the same intensity as rays

822: that do not pass through it, preventing the integral over angle from

823: vanishing. Thus for a sufficiently transparent cloud, $\Fstar(1)$ approaches zero, its value in free space. Conversely, at the surface of a cloud that is opaque and extremely large, $\Fstar(1)=\Fstar_0/4$. The factor of $1/4$ relative to the unidirectional case arises because because half the solid angle is blocked by an opaque object (providing one factor of $1/2$), and because in the part of the sky that is not blocked the radiation is isotropic, and one must average over all the directions in which photons are traveling to find the fraction of that motion in the $-\hat{r}$ direction (providing another factor of $1/2$).

824: 

825: The problem of determining the intensity is exactly the same as in

826: \S~\ref{spheredust}. At the surface of the cloud rays at

827: angles $\mu>0$ do not pass through the cloud and therefore contribute

828: the unattenuated free space intensity $\Istar_0$. Those with $\mu <

829: \mumol = -(1-\xm^2)^{1/2}$ pass through the fully molecular region and

830: therefore contribute zero intensity. For rays at angles $\mumol < \mu

831: < 0$, we make the same approximation as in \S~\ref{spheredust}, that

832: the molecular absorption rate per unit distance that a photon travels is

833: roughly constant until one approaches the sharp transition from atomic

834: to molecular. Thus the molecular opacity-dominated part of the PDR has

835: a constant effective opacity, which we can determine by computing its

836: value at the cloud surface. For convenience we characterize this

837: opacity via an effective cross-section per H nucleus $\sigmae$. By

838: examining the transfer-dissociation equation (\ref{transdissoc}), it

839: is clear that for $\fatm\approx 1$ this opacity is

840: \begin{equation}

841: \label{sigmaedef}

842: \sigmae = \frac{\phi n \calr}{\fdissoc c \Estar(1)}.

843: \end{equation}

844: 

845: With this approximation the transfer equation through the region

846: outside where the gas becomes fully molecular is simply equation

847: (\ref{transferdust}) with $\sigmad$ replaced by $\sigmae$, and the

848: solution for the intensity along each ray is given by equation

849: (\ref{istarray}) with $\tauR$ replaced by $\taueR =

850: (\sigmae/\sigmad)\tauR$. The photon number density and flux at the

851: cloud surface are therefore given by

852: \begin{eqnarray}

853: \frac{\Estar(1)}{\Estar_0} & = & \eta_0(1, \xm; \taueR) \\

854: \frac{\Fstar(1)}{\Fstar_0} & = & \eta_1(1, \xm; \taueR).

855: \end{eqnarray}

856: With these arguments, evaluating equations (\ref{f0eval}) and

857: (\ref{f1eval}) shows that $\eta_0$ and $\eta_1$ reduce to

858: \begin{eqnarray}

859: \label{f0xd1}

860: \eta_0(1,\xm;\taueR) & = & \frac{1-e^{2\mumol\taueR}}{4\taueR}+\frac{1}{2} \\

861: \label{f1xd1}

862: \eta_1(1,\xm;\taueR) & = &

863: \frac{(1-2\mumol\taueR)e^{2\mumol\taueR}-1}{8\taueR^2} + \frac{1}{4}.

864: \end{eqnarray}

865: Using these values of $\Estar(1)$ and $\Fstar(1)$ in equations

866: (\ref{sigmaedef}) and (\ref{fluxsol}), we find

867: \begin{eqnarray}

868: \label{f0eqnodust}

869: \frac{1-e^{2\mumol\taueR}}{4\taueR}+\frac{1}{2} &

870: = & \frac{\tauR}{\psi \taueR} \\

871: \label{f1eqnodust}

872: \frac{(1-2\mumol\taueR)e^{2\mumol\taueR}-1}{8\taueR^2} + \frac{1}{4}

873: & = & \frac{\tauR}{3\psi} \left(1-\xm^3\right).

874: \end{eqnarray}

875: We therefore again have two equations in two unknowns, with the

876: unknowns in this case being $\taueR$ and $\xm$. Note that in this case

877: $\mumol$ can be isolated in the first equation, to give

878: \begin{equation}

879: \label{musol}

880: \mumol = \frac{1}{2\taueR}

881: \ln\left(1+2\taueR-\frac{4\tauR}{\psi}\right).

882: \end{equation}

883: Together with the relation between $\mumol$ and $\xm$ (equation \ref{mumoleqn}), this reduces the problem to a single non-linear equation, which is

884: convenient for numerical solution. However, the two-equation form is

885: more convenient for an analytic approach.

886: 

887: At this point it is worth making a few remarks about the behavior of equations (\ref{f0eqnodust}) and (\ref{f1eqnodust}). First, equation (\ref{f0eqnodust}) implies that $\tauR/(\psi\taueR)>1/2$, so the argument of the logarithm in equation (\ref{musol}) is always less than unity and $\mumol$ is negative. Second, in all of these equations $\tauR$ and $\psi$ appear only through the combination $\tauR/\psi$, so values of $\xm$ and $\taueR$ must be constant on lines of constant $\tauR/\psi$. Finally, note that in \S~\ref{numericsol} we give an approximate analytic solution to equations (\ref{f0eqnodust}) and (\ref{f1eqnodust}).

888: 

889: \subsection{The Dust-Dominated Zone Boundary}

890: \label{dustboundary}

891: 

892: \begin{figure}

893: %\plotone{chitauzone.pdf}

894: \plotone{f4.eps}

895: \caption{

896: \label{chitauzone}

897: The thin black curves show the boundaries in the $(\tauR,\psi)$-plane at

898: which $\xd=1$, $\taueR=\tauR$, and at which $\xm=0$, as indicated by

899: the text accompanying each curve. The thick dashed curve shows our

900: approximation to the $\xd=1$ curve, equation (\ref{weakapprox}).

901: The asterisks along the curve for $\xd=1$ mark the points at which

902: $\xm=0.1$, $0.3$, $0.5$, $0.7$, and $0.9$, as indicated.

903: }

904: \end{figure}

905: 

906: We first identify the boundary between the presence and absence

907: of a dust-dominated region. In the case of a perfectly beamed

908: radiation field impinging on a semi-infinite planar slab, which we treated

909: in \S~\ref{slab}, this is $\psi=1$. The result is more complex in the case of

910: an isotropic radiation field and a cloud of finite size. The boundary

911: between the two cases is defined by the condition that $\xd=1$ or

912: $\taueR=\tauR$, i.e.\ that the cross-over between dust-dominated and

913: molecular-dominated absorptions occur at the cloud surface, or

914: equivalently that the dust and molecular effective opacities at the

915: cloud surface are equal.

916: It is immediately obvious that equations (\ref{f0eqn})

917: and (\ref{f1eqn}) become identical to equations (\ref{f0eqnodust}) and

918: (\ref{f1eqnodust}) in this limit. If we set $\taueR=\tauR$, then

919: equations (\ref{musol}) and (\ref{f1eqnodust}) define a curve in the

920: $(\tauR,\psi)$-plane that

921: corresponds to the point where dust-dominated layer disappears. 

922: Above the curve the radiation field is

923: strong enough so that the outer part of the cloud is dust

924: opacity-dominated, while below it molecular opacity dominates

925: throughout. Along the bounding curve, 

926: \begin{equation}

927: \xm = \sqrt{1 - \frac{1}{4\tauR^2}

928: \left[\ln\left(1+2\tauR-\frac{4\tauR}{\psi}\right)\right]^2}.

929: \end{equation}

930: We show the curve at which the dust-dominated layer vanishes, and the

931: value of $\xm$ along the curve, in Figure \ref{chitauzone}.

932: 

933: That along this curve $\psi\rightarrow 1$ for $\tauR\ll 1$ and

934: $\psi\rightarrow 2$ for $\tauR\gg 1$ makes intuitive sense. If

935: $\tauR\ll 1$, the dust optical depth through the cloud is tiny and so

936: the radiation field has its unattenuated, free-space value regardless

937: of position in the cloud. Thus the condition that dust shielding and

938: molecular shielding contribute equally at the cloud surface

939: ($\xd= 1$) can only be fulfilled is they are equal or nearly

940: so in free space, which is simply a requirement that

941: $\psi=1$. Similarly, if $\tauR \gg 1$ then the cloud is

942: effectively a semi-infinite slab whose curvature is negligible. In

943: this case the radiation field at the cloud surface only contains

944: contributions from rays with $\mu>0$, i.e. those that never pass

945: through the cloud; rays with $\mu<0$ are infinitely attenuated. Thus,

946: the radiation field has exactly half its free-space value, the

947: molecular fraction has double its free-space value, and the

948: condition that dust shielding and molecular shielding are equal

949: reduces to the requirement that $\psi=2$.

950: 

951: These considerations suggest that a function that interpolates between

952: these two limiting behaviors is likely to produce a good

953: approximation. Numerical experimentation shows that the curve

954: \begin{equation}

955: \label{weakapprox}

956: \psi \approx \frac{1.4+2\tauR}{1.4+\tauR}

957: \end{equation}

958: reproduces the true value of $\psi$ along the curve $\xd=1$,

959: $\taueR=\tauR$ to better than 3\% for all $\tauR$. We show this

960: approximate solution with the thick dashed curve in Figure \ref{chitauzone}.

961: 

962: 

963: \subsection{Limiting Cases}

964: \label{limitsection}

965: 

966: We can better understand the behavior of PDRs in finite clouds by exploring several limiting cases of our equations, corresponding to clouds that are very large or very small, and to radiation fields that are very strong or very weak.

967: 

968: \textit{Case 1: Strong Radiation Fields.} The first limit we consider is one in which the radiation field is so strong that there is no fully molecular core, so $\xm=0$. It is easy to verify that if there is no dust-dominated region, so equations (\ref{f0eqnodust}) and (\ref{f1eqnodust}) apply, then there are no finite values of $\psi$ and $\tauR$ such that $\xm=0$. (However, see \S~\ref{error}, where we show that this behavior is probably not physical.) On the other hand, if there is a dust-dominated region, equations (\ref{f0eqn}) and (\ref{f1eqn}) apply and $\xm=0$ can be reached at finite $\psi$ and $\tauR$. This becomes clear if we note that $\xm=0$ implies $\mumol=-1$ (following equation \ref{mumoleqn}), and equation (\ref{f1eqn}) then admits the solution $\xd=\eta_1(\xd,0;\tauR)=0$. Since $\xd=0$, it immediately follows that $\xi_{\rm d}=1$ and $\eta_0(0,0;\tauR)=e^{-\tauR}$. Equation (\ref{f0eqn}) then gives $\psi=e^{\tauR}$. The physical meaning of this solution is that $\psi=e^{\tauR}$ is the critical curve along which $\xm=\xd=0$; at this value of $\psi$ or larger the radiation field is too intense for a fully molecular core to exist. We plot the critical curve in Figure \ref{chitauzone}. In Appendix \ref{strongappendix} we solve

969: equations (\ref{f0eqn}) and (\ref{f1eqn}) perturbatively in the vicinity of the critical curve, and show that

970: near the strong radiation boundary the

971: solution is

972: \begin{eqnarray}

973: \xd & = & \left[\frac{6}{\tauR (\tauR+2)} \left(\frac{e^{\tauR}}{\psi}-1\right)\right]^{1/2} \\

974: \xm & = & \left[\frac{1536}{25\tauR(\tauR+2)^3}\right]^{1/4} \left(\frac{e^{\tauR}}{\psi}-1\right)^{5/4}.

975: \end{eqnarray}

976: This solution obviously only applies for $e^{\tauR} \geq \psi$.

977: 

978: \textit{Case 2: Small Clouds.} Before analyzing this case, we warn that in \S~\ref{error} we show that our solution in this case should be regarded as giving an upper limit on the molecular fraction rather than a direct estimate. However it is still useful to consider this case, both in order to derive upper limits and to provide expressions that can be incorporated into approximations in parts of parameter space where our method does provide estimates rather than upper limits. We have shown that for $\psi>1$, there is a finite value of $\tauR$ at which the fully molecular core vanishes, and conversely that if $\psi<1$ there is no finite $\tauR$ for which $\xm=0$. However, one can easily verify that when there is no dust-dominated zone ($\psi<1$) and

979: equations (\ref{f0eqnodust}) and (\ref{f1eqnodust}) apply, 

980: $\xm\rightarrow 0$ as $\tauR\rightarrow 0$. Physically, this corresponds to the case of a cloud that is so small that its dust is optically thin to LW photons. In Appendix

981: \ref{largesmallappendix} we show that in this limit the solution may be

982: approximated by

983: \begin{eqnarray}

984: \label{smallcloudeqn1}

985: \taueR & = & \frac{\tauR}{\psi} + \frac{\tauR^2}{2\psi^2} +

986: \frac{\tauR^3}{4\psi^3} + \frac{7\tauR^4}{60\psi^4} \\

987: \xm & = & \frac{\tauR}{\sqrt{3}\psi} + \frac{2\sqrt{3}\tauR^2}{5\psi^2}.

988: \end{eqnarray}

989: Note that by definition $\taueR/\tauR = [\Estar_0/\Estar(1)]/\psi$, so equation (\ref{smallcloudeqn1}) is effectively a series expansion for the photon number density at the cloud surface: $\Estar(1)=\Estar_0 [1 - \tauR^2/(2\psi) + \cdots]$. Thus the leading-order approximation reduces to the statement that a small cloud blocks no radiation in any direction, so $\Estar(1)=\Estar_0$, the unattenuated value. The next-order correction accounts for the small fraction of photons that are blocked at the cloud surface.

990: 

991: \textit{Case 3: Large Clouds.} Our final limiting case is that of a cloud so large that

992: the transition from atomic to molecular gas occurs in a thin layer at

993: the cloud surface, so that the cloud's curvature is negligible. Before

994: proceeding we note that this case is \textit{not} the same as the

995: case of a one-dimensional slab subject to a unidirectional beam of

996: radiation that we analyzed in \S~\ref{slab}. The difference is that

997: here the radiation field is isotropic, so it has an angular dependence

998: that can vary with depth within the cloud. For this reason the large cloud limit with an isotropic radiation field is a

999: two-dimensional problem even if the cloud is a semi-infinite slab.

1000: To analyze this case we perform a series expansion around the limit

1001: $\tauR\rightarrow \infty$, but with $\ym\tauR$ finite, so that there

1002: is a finite optical depth to the molecular region. If $\psi>2$ then a

1003: dust-dominated zone exists, and we solve this problem by starting from

1004: equations (\ref{f0eval}) and (\ref{f1eval}) and series expanding $\eta_0$

1005: and $\eta_1$ to first order in $\tauR^{-1}$. Doing so gives

1006: \begin{eqnarray}

1007: \eta_0(\xd,\xm;\tauR) & = & \frac{{\rm E}_2(\taud)}{2} + \frac{e^{-\taud}+\taud {\rm

1008: E}_2(\taud)}{4\tauR} \\

1009: \eta_1(\xd,\xm;\tauR) & = & \frac{{\rm E}_3(\taud)}{2} + 

1010: \frac{\taud{\rm E}_3(\taud)}{2\tauR},

1011: \end{eqnarray}

1012: where $\taud=\yd\tauR$. Equation (\ref{f0eqn}) therefore becomes

1013: \begin{equation}

1014: \label{taupeqn}

1015: \frac{1}{\psi} = \frac{{\rm E}_2(\taud)}{2} + \frac{e^{-\taud}+\taud {\rm

1016: E}_2(\taud)}{4\tauR}

1017: \end{equation}

1018: to first order in $\tauR^{-1}$, which is straightforward to solve

1019: numerically to determine $\taud$ for a given $\psi$ and $\tauR$. Alternately, one may obtain a purely analytic expression by dropping the $1/\tauR$ correction term. In this case the equation becomes ${\rm E}_2(\taud) = 2/\psi$, which has the approximate solution

1020: \begin{equation}

1021: \taud \approx 0.83 \, \ln(0.2\psi + 0.6);

1022: \end{equation}

1023: this expression is accurate to better than 2\% for $2 \le \psi \le 100$. Since

1024: $\xd=1-\taud/\tauR$, this fixes $\xd$. Similarly, once

1025: $\taud$ is known it is straghtforward to solve equation

1026: (\ref{f1eqn}) to first order in $\tauR^{-1}$ to obtain

1027: \begin{equation}

1028: \xm = 1 - \left[\taud+\frac{\psi}{2} {\rm E}_3(\taud)\right]

1029: \tauR^{-1}.

1030: \end{equation}

1031: 

1032: For $\psi<2$, there is no dust-dominated zone, and we must instead

1033: solve equations (\ref{f0eqnodust}) and (\ref{f1eqnodust})

1034: in the limit $\tauR\rightarrow \infty$. We note that for a very large cloud $\Estar(1)=\Estar_0/2$ because the cloud blocks half the sky, and it therefore follows immediately from the definition of $\sigmae$ (equation \ref{sigmaedef}) that

1035: \begin{equation}

1036: \taueR = \frac{2 \tauR}{\psi},

1037: \end{equation}

1038: i.e.\ that the effective molecular opacity is a factor of two larger than its free-space value because the radiation intensity at the cloud surface has half its free-space value. Similarly, the flux is $\Fstar(1)=\Fstar_0/4$ because half the sky is blocked and the radiation direction is random over the other half. Using this boundary condition to integrate the one-dimensional transfer-dissocation equation (\ref{slabeqn}) with $\sigmad=0$ and the molecular absorption rate multiplied by $\phi$ then gives 

1039: \begin{equation}

1040: \xm = 1-\frac{\psi}{4\tauR}.

1041: \end{equation}

1042: We verify that these intuitive arguments in fact give the correct leading order terms in the series expansion in Appendix \ref{largesmallappendix}.

1043: Thus we have the limiting solution to first order in $\tauR^{-1}$ for

1044: both $\psi<2$ and $\psi>2$. We illustrate this solution for

1045: $\tauR\rightarrow \infty$ in Figure

1046: \ref{planeplot}.

1047: 

1048: \begin{figure}

1049: %\plotone{planeplot.pdf}

1050: \plotone{f5.eps}

1051: \caption{

1052: \label{planeplot}

1053: The plot shows the solution in the large cloud limit. The curves shown

1054: are the dust optical depth to the fully molecular

1055: region $\taum=\tauR(1-\xm)$ (\textit{solid curve}), the dust optical depth to the

1056: point of dust-molecular absorption equality $\taud=\tauR(1-\xd)$ (\textit{dashed

1057: curve to the right of $\psi=2$}), and the ratio of the effective molecular

1058: opacity to the dust opacity $\taueR/\tauR$ (\textit{dashed

1059: curve to the left of $\psi=2$}). The dotted vertical line at $\psi=2$

1060: indicates the boundary between the presence and absence of a

1061: dust-dominated zone in the weak radiation limit.

1062: }

1063: \end{figure}

1064: 

1065: \subsection{Numerical Solution and Analytic Approximation}

1066: \label{numericsol}

1067: 

1068: \begin{figure*}

1069: %\plotone{chitaucont.pdf}

1070: \plotone{f6.eps}

1071: \caption{

1072: \label{chitaucont}

1073: Contours showing the solution as a function of $\tauR$ and $\psi$ for

1074: the structure of the PDR in the two-zone approximation. The values

1075: shown are, clockwise from the upper left, $\xm$, $\xd$ or

1076: $\tauR/\taueR$, $\tau_{\rm HI}$, and $\xm^3$. The hatched region is

1077: the region in which there is no primarily molecular part of the cloud.

1078: The dotted line

1079: indicates the boundary between the presence and absence of a

1080: dust-dominated zone. In the panel labelled $\xd$, $\tauR/\taueR$, the

1081: contours above the dotted line indicate the value of $\xd$, those

1082: below it show the value of $\tauR/\taueR$, and on the dotted line both

1083: of these quantities are exactly $1.0$. We caution that the contours for $\xm = 0.1$ and $\xm=0.3$, and for $\xm^3 = 0.1$, should be regarded as giving upper limits on $\xm$, not precise estimates -- see the discussion in \S~\ref{error}.

1084: }

1085: \end{figure*}

1086: 

1087: We now proceed with a numerical treatment of the general case. We

1088: solve equations (\ref{f0eqn}) and (\ref{f1eqn}), or (\ref{f0eqnodust})

1089: and (\ref{f1eqnodust}), on a grid of points in the

1090: $(\tauR,\psi)$-plane, and plot the results in Figure \ref{chitaucont}. In

1091: addition to plotting $\xm$ and either $\xd$ or $\taueR$, we

1092: also show two derived quantities of interest. The first is the dust

1093: optical depth to the molecular transition along a radial trajectory, $\taum = nR(1-\xm) \sigmad$. We may think of this as the H\textsc{i} ``shielding column'' times the dust cross

1094: section. The second is $\xm^3$, which is the fraction of the cloud's

1095: volume that is within the predominantly molecular region.

1096: 

1097: The general behavior of these curves can be understood intuitively. 

1098: If one fixes the

1099: cloud density $n$ and dust opacity $\sigmad$, then as the cloud radius

1100: increases so does $\tauR$, and for fixed external radiation field

1101: $\psi$ the molecular transition moves outward, but the H\textsc{i} column to

1102: that transition approaches a constant value. Similarly, at fixed

1103: cloud size and hence $\tauR$, increasing the external illumination

1104: $\psi$ raises the amount of atomic hydrogen that is required to shield

1105: the molecules. Thus $\xm$ drops when $\psi$ increases at fixed $\tauR$.

1106: 

1107: \begin{figure}

1108: %\plotone{nodust.pdf}

1109: \plotone{f7.eps}

1110: \caption{

1111: \label{nodust}

1112: Contours showing the factor by which dust shielding changes the atomic or molecular volume at a given $(\tauR, \psi)$. The quantity plotted is $\mbox{max}[\xm/x_{\rm H_2,nd}, (1-x_{\rm H_2,nd})/(1-\xm)]$, where the subscript ``nd" indicates the value with no dust shielding in the limit, i.e. in the limit $\sigmad\rightarrow 0$. The quantity plotted is therefore the fractional amount by which dust shielding increases the radius of the molecular zone or decreases the radial path length through the atomic zone, whichever is larger.

1113: }

1114: \end{figure}

1115: 

1116: These curves also enable us to determine under what circumstances dust makes a significant contribution to shielding the gas, a subject that has been discussed considerably in the literature \citep[e.g.][]{vandishoeck86, draine96}. To evaluate the importance of dust, we can consider how the molecular and atomic volumes change as $\sigmad\rightarrow 0$. In terms of our parameters, this amounts to taking the limit as $\chi\rightarrow 0$ and $\tauR\rightarrow 0$, but the ratio $\chi/\tauR$ remains constant. Graphically, this is equivalent to sliding toward the lower right of Figure \ref{chitaucont}, along a trajectory that is close to a line of slope unity -- it is not precisely a line of slope unity because of the slight non-linearity of the relationship between $\psi$ and $\chi$. We plot the factor by which dust shielding changes the radius of the molecular zone or the radial path length through the atomic zone, whichever is larger, in Figure \ref{nodust}. As the figure shows, dust shielding changes the radius of the molecular zone by a factor of $\sim 2$ only when $\psi$ is of order unity or larger or when $\tauR$ is very large, which is about what one might expect. Dust shielding can affect the size of the molecular region even if there is no dust-dominated zone because we have allowed for dust absorptions even in the molecular shielding region. However, the effect is at most tens of percent. A larger change is possible only if the radiation field is intense enough to create a dust-dominated zone (i.e. $\psi \gtsim 1$) or the cloud is so large (i.e.\ $\tauR\gg 1$) that even weak dust shielding becomes significant because it attenuates the radiation exponentially rather than in a powerlaw fashion as do the molecules.

1117: 

1118: We can deduce approximate analytic fitting formulae for $\xm$ by

1119: interpolating between the solutions for the limiting cases. The

1120: following fitting formulae are reasonably accurate and can be evaluated with no numerical iteration:

1121: \begin{equation}

1122: \label{xmapprox}

1123: \xm \approx

1124: \left\{

1125: \begin{array}{ll}

1126: \frac{\psi^4 x_{\rm s} + (3\tauR)^4 x_{\ell}}

1127: {\psi^4+(3\tauR)^4}, \quad &

1128: \psi < 1 \\

1129: \max(x_{\rm s}, x_{\ell}), \quad &

1130: 1<\psi<\psi_{\rm b} \\

1131: \left(x_{\rm s}^{-3/2}+x_{\ell}^{-3/2}\right)^{-2/3}, \quad &

1132: \psi_{\rm b} < \psi < 2 \\

1133: \left(x_{\rm s}^{-3}+x_{\ell}^{-3}\right)^{-1/3}, \quad &

1134: 2 < \psi < e^{\tauR} \\

1135: 0, \quad & e^{\tauR} < \psi

1136: \end{array}

1137: \right.

1138: \end{equation}

1139: where $\psi_{\rm b} = (1.4+2\tauR)/(1.4+\tauR)$ is the approximate

1140: value of $\psi$ at the boundary between the existence and

1141: non-existence of a zone of dust-dominated opacity,

1142: \begin{equation}

1143: \label{xstapprox}

1144: x_{\rm s}^2 =  

1145: \left\{

1146: \begin{array}{ll}

1147: \left(\frac{\tauR}{\sqrt{3}\psi} + \frac{2\sqrt{3}\tauR^2}{5\psi^2}\right)^2, 

1148: \quad & \psi < 1 \\

1149: 1-\frac{1}{4\tau_{\rm

1150: b}^2}\left[\ln\left(1 + 2 \tau_{\rm b}-4\frac{\tau_{\rm

1151: b}}{\psi}\right)\right]^2,

1152: \quad & 1 < \psi < \psi_{\rm b} \\

1153: \left[\frac{1536}{25\tauR(\tauR+2)^3}\right]^{1/2}

1154: \left(\frac{e^{\tauR}}{\psi}-1\right)^{5/2}, 

1155: \quad & \psi_{\rm b} < \psi

1156: \end{array}

1157: \right.

1158: \end{equation}

1159: is the approximate value of $\xm$ in the strong radiation (for $\psi>1$) or small cloud (for $\psi<1$) limits (note that we only evaluate $x_{\rm s}$ when $\psi < e^{\tauR}$), and

1160: \begin{equation}

1161: \label{xweapprox}

1162: x_{\ell}^2 = 

1163: \left\{

1164: \begin{array}{ll}

1165: \left(\max\left[1-\frac{\psi}{4\tauR}, 0\right]\right)^2,

1166: \quad & \psi < \psi_{\rm b} \\

1167: 1-\frac{1}{4\tau_{\rm

1168: b}^2}\left[\ln\left(1 + 2 \tau_{\rm b}-4\frac{\tau_{\rm

1169: b}}{\psi}\right)\right]^2

1170: \quad & \psi_{\rm b} < \psi < 2 \\

1171: \left[1 - \left(\taud+\frac{\psi e^{-\taud}}{4+2\taud}\right)\frac{1}{\tauR}\right]^2,

1172: \quad & 2 < \psi

1173: \end{array}

1174: \right.

1175: \end{equation}

1176: is the approximate value of $\xm$ in the large cloud limit.

1177: Here

1178: \begin{equation}

1179: \taud={\rm E}_2^{-1}\left(\frac{2}{\psi}\right) \approx 0.83 \,\ln(0.2\psi+0.6)

1180: \end{equation}

1181: is the approximate optical depth

1182: to the dust-molecular opacity crossover in the large cloud limit

1183: when dust shielding is important,

1184: \begin{equation}

1185: \tau_{\rm b} \approx 1.4 \frac{\psi-1}{2-\psi}

1186: \end{equation}

1187: is the value of $\tauR$ at the dust-no dust

1188: boundary for a given $\psi$, and for convenience we have used the approximation ${\rm E}_3(x) \approx e^{-x}/(2+x)$. Note that these approximations can fail

1189: if one is very near the dust-no dust boundary because the

1190: approximation $\tau_{\rm b}

1191: \approx 1.4(\psi-1)/(2-\psi)$ is insufficiently accurate; in this case

1192: one may still use the approximate expressions by replacing $\tau_{\rm

1193: b}$ with a more accurate value of $\tauR$ on the dust-no dust boundary

1194: computed as described in \S~\ref{dustboundary}.

1195: 

1196: \begin{figure}

1197: %\plotone{chitaufit.pdf}

1198: \plotone{f8.eps}

1199: \caption{

1200: \label{chitaufit}

1201: Error in the approximate analytic fit given in \S~\ref{numericsol} as

1202: a function of $\tauR$ and $\psi$. The shaded regions indicate errors

1203: below $2.5\%$ (\textit{no shading}), $2.5-5\%$, $5-10\%$, $10-20\%$, and $>20\%$ (\textit{darkest shading}). The maximum error is 24\%. The dotted lines show the boundaries of our

1204: different approximation regions: $\psi<1$, $1<\psi<\psi_{\rm b}$,

1205: $\psi_{\rm b} < \psi < 2$, and $2<\psi$. The

1206: hatched region is $\psi>e^{\tauR}$, where there is no

1207: predominantly molecular core. Note that the error jumps at the $\chi=1$ and $\chi=2$ lines because the fitting formula is slightly discontinuous there.

1208: }

1209: \end{figure}

1210: 

1211: Figure \ref{chitaufit} shows the error in our analytic approximation

1212: as a function of $\tauR$ and $\psi$, where we define the error

1213: as $|\xm-x_{\rm H_2,approx}| / \max(\xm,1-\xm)$ and $\xm$ is the

1214: solution obtained by numerically solving the appropriate equations.

1215: As the plot shows the fitting formulae are generally good to the $\sim

1216: 10\%$ level, as good as the two-zone approximation itself. The

1217: maximum error over the range $0.1<\tauR<100$ and $0.1<\psi<100$ is

1218: 24\%, and occurs near $\tauR=1.8$, $\psi=1.9$.

1219: 

1220: We can also obtain an even simpler approximation formula if we specialize to the case where there is no or almost no dust-dominated zone and $\xm\gtsim 0.5$, which we show in Paper II is the most common case in nearby galaxies. Consider equation (\ref{f1eqnodust}), which describes the surface flux for the case of no dust. The two terms on the LHS represent the contributions to the flux from rays that do and do not pass through the cloud, respectively. If the cloud has a significant molecular core, $\xm \gtsim 0.5$, then only for a small range of angles do rays pass through the cloud but not strike the opaque molecular core, and thus the first term on the LHS is small in comparison to the second. For convenience we define

1221: \begin{equation}

1222: \delta = \frac{1-(1-2\mumol \taueR) e^{2\mumol\taueR}}{2\taueR^2},

1223: \end{equation}

1224: which enables us to rewrite equation (\ref{f1eqnodust}) as

1225: \begin{equation}

1226: \xm^3 = 1 - \frac{3\psi}{4\tauR}\left(1-\delta\right) \approx 1 - \frac{3\psi}{4\tauR}\left(\frac{1}{1+\delta}\right)

1227: \end{equation}

1228: where $\delta$ is a small, positive number. Now consider how $\delta$ varies with $\tauR$: we show in \S~\ref{limitsection} that for either small or large $\tauR$, to first order $\taueR\propto \tauR/\psi$. If we consider the series of expansion of $\delta$, this implies that $\delta$ approaches a constant at small $\tauR$ and varies as $\psi^2/\tauR^2$ for large $\tauR$. To generate our approximation we adopt an intermediate scaling

1229: \begin{equation}

1230: \delta \approx a \frac{\psi}{\tauR},

1231: \end{equation}

1232: where $a$ is a constant to be chosen to optimize the approximation. This gives

1233: \begin{equation}

1234: \label{chrisapprox}

1235: \xm^3 \approx 1 - \frac{3\psi}{4(\tauR+a\psi)}.

1236: \end{equation}

1237: For the choice $a=0.2$, equation (\ref{chrisapprox}) agrees with the numerical solution to equations (\ref{f0eqn}) and (\ref{f1eqn}) or (\ref{f0eqnodust}) and (\ref{f1eqnodust}) to better than 15\% whenever $\psi<3$ and equation (\ref{chrisapprox}) gives $\xm^3 > 0.15$. A corresponding approximate formula for the optical depth through the atomic layer is

1238: \begin{equation}

1239: \label{chrisapprox1}

1240: \taum \approx \frac{\tauR \psi}{4\tauR-a'\psi},

1241: \end{equation}

1242: with $a'=\frac{3}{2}-4a$. For $a=0.2$ this expression agrees with the numerical solution to better than 15\% for $\psi<3$ whenever equation (\ref{chrisapprox}) gives $\xm^3>0.1$.

1243: 

1244: \subsection{Comparison to the One-Dimensional Case}

1245: \label{planarcomp}

1246: 

1247: Now that we have solved the spherical case, we are in a position to

1248: compare to the case of a one-dimensional beam of radiation striking an infinite slab that is

1249: often treated in the literature. This will allow us to determine when this approximation 

1250: yields reasonably accurate results, and when it is it gives significantly different results. Figure \ref{slabcomp} shows a comparison between our

1251: solution with isotropic radiation and varying cloud sizes versus the most common approximation in the literature: an infinite cloud and a beam of radiation whose photon number density is half the free-space value. In the calculations for finite clouds and isotropic radiation we end each curve at the value of $\psi$ for which the fully molecular region vanishes. For the beamed radiation and infinite cloud case, we use the analytic solution described in \S~\ref{slabanalyt}.

1252: 

1253: As the plot shows, when $\tauR \ll \psi$, the one-dimensional slab approximation can

1254: produce significantly different estimates of the depth of the dust shielding layer than does our higher-dimensional approach. The difference

1255: becomes larger as we consider smaller

1256: clouds. For $\tauR > 1$ the slab approximation generally

1257: underestimates the depth of the atomic layer by tens of percent,

1258: primarily because it assumes neglects the photodissociation provided

1259: by non-radial rays. Even for a cloud that is infinitely large, 

1260: $\tauR=\infty$, this difference between an isotropic radiation field and a beamed one can be significant at moderate $\psi$ because even though there are no rays reaching a given position from the ``back side" of the cloud, $\mu<0$, when the radiation field is isotropic there are still non-radial rays that raise the photodissociation rate at a given position above what it would be in the purely beamed radiation field of smaller intensity.

1261: 

1262: For $\tauR < 1$ the sign of the error depends on $\psi$. When the

1263: radiation field is weak, the slab approximation also underestimates

1264: the depth of the atomic layer, for the same reason as when

1265: $\tauR>1$. When the radiation field becomes strong, though, the sign

1266: of the error reverses, although as we discus in \S~\ref{error} our fiducial model is of limited accuracy for small $\tauR$ and large $\psi$.

1267: 

1268: Physically, clouds of a wide range of sizes and densities are of

1269: course present in the ISM. For the atomic envelopes of GMCs in the

1270: Milky Way, a typical density is $n\sim 30$ cm$^{-3}$ and a typical

1271: dust cross-section is $10^{-21}$ cm$^2$, so that $\sim 10$ pc of path

1272: provides an optical depth of about 1. Since these envelopes are a few

1273: tens of pc in size, a typical one might have $\tauR$ of a few, in

1274: which case the slab treatment underestimates the true size of the

1275: envelope at the tens of percent level. In low-metallicity galaxies

1276: with low molecular fractions, however, the error is likely to be much

1277: worse because $\tauR$ will be significantly smaller.

1278: 

1279: \begin{figure}

1280: %\plotone{slabcomp.pdf}

1281: \plotone{f9.eps}

1282: \caption{

1283: \label{slabcomp}

1284: The upper panel shows the dust optical depth to the point where the

1285: gas becomes predominantly molecular $\tau_{\rm HI}$ for an isotropic

1286: radiation field of normalized intensity $\psi$ and various values of $\tauR$ (\textit{solid lines}) and for a unidirectional radiation field of normalized intensity $\psi/2$ (\textit{dashed line}). The lower panel shows the fractional difference between the results for finite $\tauR$ and isotropic radiation and for an infinite slab illuminated by unidirectional radiation,

1287: defined as $\mbox{Difference} = [\tau_{\rm HI,iso}-\tau_{\rm

1288: HI,beam}]/\tau_{\rm HI,iso}$. In

1289: all cases the curves for finite $\tauR$ end at the value of $\psi$ for

1290: which fully molecular region disappears ($\psi=e^{\tauR}$).

1291: }

1292: \end{figure}

1293: 

1294: \subsection{Uncertainties in Spherical Geometry}

1295: \label{error}

1296: 

1297: \begin{figure*}

1298: %\plotone{chitaucontho.pdf}

1299: \plotone{f10.eps}

1300: \caption{

1301: \label{chitaucontho}

1302: Same is Figure \ref{chitaucont}, but assuming infinite attenuation along rays that enter the  molecular absorption-dominated region (equations \ref{f0eqnho}--\ref{nodustho}).

1303: }

1304: \end{figure*}

1305: 

1306: We have shown that in the case of a one-dimensional beam of radiation impinging on a slab, the two-zone approximation is capable of determining the neutral hydrogen shielding column to better than $\sim 50\%$ accuracy. This characterizes the level of error imposed by most of our physical assumptions. However, in spherical geometry we have an additional uncertainty, imposed by the fact that we must assign an effective optical depth to rays that pass at arbitrary angles through the region where molecular shielding dominates, but the gas is not yet fully molecular. In particular, rays from the ``back side" of our cloud, those with $\mu<0$, contribute to the energy density and flux throughout the cloud. Such rays are absent in the case of an infinite planar cloud, because rays with $\mu<0$ pass through the fully  molecular region and are therefore infinitely attenuated.

1307: 

1308: In this section, we seek to determine how much additional uncertainty is introduced into our calculations in spherical geometry by this complication. To do so, we note that in our treatment above we assume that the opacity in the molecular shielding region will be roughly equal to that at its surface, i.e.\ that it does not rise sharply until one is very close to the transition to fully molecular gas. The represents a minimum attenuation along $\mu<0$ rays. To check the importance of that assumption, we consider an extreme assumption in the opposite direction: that all rays with $\mu<0$ are infinitely attenuated. This assumption is obviously unphysical, since if it were true then the transition to fully molecular would occur as soon as self-shielding began to dominate over dust shielding. However, it provides a worst case with which we can compare our fiducial model as a way of characterizing our uncertainty. Since, as we shall see, the value of $\xm$ that we obtain by making this assumption is always smaller than what we obtain for the fiducial case, we may regard the fiducial case as giving upper limit on $\xm$ and the case we calculate here as giving a lower limit.

1309: 

1310: If we take the intensity along rays with $\mu<0$ to be zero, this is equivalent to replacing $\eta_0(\xd,\xm;\tauR)$ with $\eta_0(\xd,\xd;\tauR)$ in the equations derived in \S~\ref{spheredust} and \S~\ref{spherenodust}, and similarly for $\eta_1$. Doing this and simplifying gives

1311: \begin{eqnarray}

1312: \label{f0eqnho}

1313: \eta_0(\xd,\xd;\tauR) & = & \frac{1}{\psi} \\

1314: \eta_1(\xd,\xd;\tauR) & = & \frac{\tauR\xd}{3\psi}\left[1-\left(\frac{\xm}{\xd}\right)^3\right]

1315: \end{eqnarray}

1316: for $\psi>2$ (dust shielding is significant) and

1317: \begin{equation}

1318: \label{nodustho}

1319: \xm=\left(1-\frac{3\psi}{4\tauR}\right)^{1/3}

1320: \end{equation}

1321: for $\psi<2$ (dust shielding is not significant). Note that in this case when dust shielding is negligible it is possible to solve the equations analytically, which we have done to obtain equation (\ref{nodustho}). It is immediately obvious from this equation that the fully molecular region vanishes in the region $4\tauR/3<\psi<2$. We plot the solutions to equations (\ref{f0eqnho})--(\ref{nodustho}) in the $(\tauR,\psi)$ plane in Figure \ref{chitaucontho}, and we show the difference between this solution and our fiducial one in Figure \ref{chitaucomp}.

1322: 

1323: \begin{figure*}

1324: %\plotone{chitaucomp.pdf}

1325: \plotone{f11.eps}

1326: \caption{

1327: \label{chitaucomp}

1328: The plots show the difference between $\taum$ and $\xm^3$ computed under the fiducial assumption that the opacity throughout the molecule-dominated region equal to that at its surface (\S~\ref{spheredust} and \S~\ref{spherenodust}) and under the assumption of infinite attenuation in this region (\S~\ref{error}).  The difference is defined as $|\tau_{\rm HI,fiducial}-\tau_{\rm HI,attenuated}|/\tau_{\rm HI,fiducial}$, and similarly for $\xm^3$. The hatched region is the region in which there is no predominantly molecular core under either assumption; the difference in this region is obviously zero.

1329: }

1330: \end{figure*}

1331: 

1332: As the plots show, the difference between the two models is negligibly small over most of parameter space;  there is a significant difference only in the region roughly bounded by the curves $\psi \la e^{\tauR}$, $\psi\ga 4\tauR/3$, and $\psi \la 2$. Alternately, we can phrase these constraints in terms of values of $\xm^3$. This is particularly useful for $\psi < 1$, where contours of both constant $\xm$ and constant uncertainty are both straight lines corresponding to fixed $\psi/\tauR$. The 10\%, 50\%, and 100\% uncertainty contours in $\xm^3$, shown in the right panel of Figure \ref{chitaucomp}, correspond to values of $\xm^3 = 0.47$, $0.28$, and $0.20$, respectively, as computed using our fiducial model and shown in Figure \ref{chitaucomp}. Since we have already established that the two-zone approximation is uncertain at the tens of percent level, the geometric uncertainty is probably only dominant when $\psi < 1$ and our predicted molecular volume fraction $\xm^3$ is less than about a quarter. If $\psi \ga 1$, the errors at a given value of $\xm^3$ are considerably smaller, so the geometric uncertainty is not important except for clouds with very small molecular fractions. At such molecular fractions, one should interpret our fiducial case as giving only an upper limit on the molecular content of a cloud.

1333: 

1334: The significant geometric uncertainty for such clouds is not surprising, since these clouds are near the limit of having no molecules at all. For them any change in our physical assumptions that increases or reduces the amount of shielding even a small amount produces a significant change in the results. Indeed, our fiducial calculation shows some unphysical behavior in this regime, in that we find that for $\psi<1$ there is no finite value of $\tauR$ for which the fully molecular region vanishes and the cloud remains atomic throughout. This seems unlikely, given that for a chosen value of $\psi$ and very large $\tauR$ the thickness of the atomic region approaches a finite value; one would expect that clouds much smaller than this should be atomic throughout regardless of their shape, and indeed under the assumption of infinite attenuation for backside rays that we make in this section, the molecular core always vanishes at some finite value of $\tauR$ for any finite $\psi$.

1335: 

1336: Fortunately, as we discuss in \S~\ref{example}, for realistic parameters describing giant atomic-molecular complexes, we generally have $\psi\ga 1$, $\tauR\ga 1$, and, as we show in \S~\ref{example}, $\xm^3 \ga 0.5$, and in this part of parameter space the uncertainty introduced by the $\mu<0$ rays in spherical geometry is $\la 10\%$.

1337: 

1338: 

1339: \subsection{Example Calculations}

1340: \label{example}

1341: 

1342: Here we provide examples that illustrate the use of our analytic approximations for PDR structure. Since these calculations are intended to be illustrative rather than to analyze real situations (which we will discuss in the next paper in this series), we choose parameters to yield examples that span the possible combinations of parameters without worrying how well they agree with observations. As a first case, consider a ``typical" Milky Way cloud, with $\sigmad=1.1\times 10^{-21}$ cm$^2$, $n=30$ cm$^{-3}$, $\calr=3\times 10^{-17}$ cm$^3$ s$^{-1}$, and $c \Estar_0=10^8$ cm$^{-2}$ s$^{-1}$. Note that we have used a value of $\Estar_0$ somewhat larger than the solar neighborhood value because most molecular clouds are closer to the galactic center, where the radiation field is more intense. This combination of parameters gives $\chi = \fdissoc \sigmad c \Estar_0/(n\calr) = 12.2$ and $\psi=\chi (2.5+\chi)/(2.5+\chi e)=6.3$. Now consider a giant molecular cloud complex with a radius $R=50$ pc, which gives $\tauR=n\sigmad R=4.64$. Since $\psi>2$, this cloud has a significant dust shielding zone, and since $\psi < e^{\tauR}$ it also has a fully molecular core, as we expect. Using the approximation equations (\ref{xmapprox})--(\ref{xweapprox}) for this case, we find $\taud=0.52$, $x_{\ell} = 0.74$, $x_{\rm s}=14.1$, and an approximate value of $\xm\approx 0.74$. (Note that although $\xm$ is strictly less than unity, it is possible for $x_{\rm s}$ to be larger than unity because we only retained a finite number of terms in the series expansion used to generate it. However, due to the way $x_{\rm s}$ and $x_{\ell}$ are combined, our approximate expression for $\xm$ is always less than unity.) Numerical solution for these parameters gives $\xm=0.70$. Such a cloud is 34\% molecular by volume, and is shielded by an atomic column that is 15 pc deep and has a column density of $\nhi=1.4\times 10^{21}$ cm$^{-2}$ from the edge the cloud to the edge of the molecular zone.

1343: 

1344: Now consider moving this cloud to a point farther out in the Galaxy where the ambient FUV radiation field is weaker, so that all cloud parameters remain the same but now $c \Estar_0=2\times 10^7$ cm$^{-2}$ s$^{-1}$, a factor of 5 below our previous value. In this case we have the same $\tauR$, but $\chi=2.44$ and $\psi=1.43$. For this $\tauR$ we have $\psi_{\rm b}=1.40$, slightly smaller than $\psi$, so this cloud just barely still has a dust-dominated zone. Evaluating equations (\ref{xmapprox})--(\ref{xweapprox}), we have $x_{\ell}=0.87$ and $x_{\rm s} = 95$, so get an approximate value $\xm\approx  0.87$; the numerical solution is $\xm=0.92$. Thus moving the cloud to this reduced-radiation environment raises the molecular volume fraction to 77\%, and reduces the H\textsc{i} shielding column to a layer 4 pc deep containing a column of $\nhi=3.7\times 10^{20}$ cm$^{-2}$ hydrogen atoms. If the cloud were slightly denser, $n=40$ cm$^{-3}$ instead of 30 cm$^{-3}$, then $\tauR$ would increase to 6.18 from 4.64, and $\psi$ would decrease to $1.11$ from 1.43. Since $\psi_{\rm b}=1.82$ in this case, the cloud would be dominated by molecular absorption throughout. Evaluating the approximation equations gives $x_{\ell} = 0.95$, $x_{\rm s} = 0.35$, and $\xm\approx 0.95$; the numerical solution is $\xm=0.95$. Thus, the increase in density would slightly increase the molecular volume and the column density through the shielding layer to 87\% and $\nhi=5.2\times 10^{20}$ cm$^{-2}$, respectively.

1345: 

1346: Finally consider a cloud in a low-pressure dwarf galaxy with a very low star formation rate, so the cloud has lower density and metallicity than a Milky Way cloud, $n=10$ cm$^{-3}$, $\sigmad=2.2\times 10^{-22}$, and $\calr=6\times 10^{-18}$, and is exposed to a lower level of radiation, $c \Estar_0=10^6$ cm$^{-3}$ s$^{-1}$. We keep the cloud radius unchanged. This cloud has $\tauR=0.31$ and $\psi=0.29$, which from our approximate formulae gives $x_{\ell}=0.77$, $x_{\rm s}=1.4$, and $\xm\approx 0.77$. The numerical solution is $\xm=0.73$. This cloud would be 38\% molecular by volume, and would have a shielding layer of $\nhi=4.2\times 10^{20}$ cm$^{-2}$, 14 pc deep.

1347: 

1348: 

1349: \section{Summary and Conclusion}

1350: \label{conclusion}

1351: 

1352: In this paper we develop an approximate analytic solution to the

1353: problem of determining the size of the PDR that

1354: bounds a cloud of gas embedded in a dissociating background radiation

1355: field. This is a reasonable approximation to the problem of finding

1356: the location of the transition between the atomic envelope and the

1357: molecular core in a giant atomic-molecular cloud complex, such as

1358: those which contain the bulk of the molecular gas in the Milky Way.

1359: 

1360: We show that the location of the transition is determined by two

1361: dimensionless parameters. These are $\tauR$, the dust optical depth

1362: through the cloud, and $\chi$, the ratio of the rate at which

1363: dissociating photons are absorbed by dust grains to the rate at which

1364: they are absorbed by H$_2$ molecules in the absence of any shielding. We may intuitively

1365: think of these parameters as characterizing the size of the cloud and

1366: the intensity of the radiation field to which it is subjected. Within

1367: this parameter space we identify two critical curves, which define the

1368: boundaries at which a fully molecular region in the cloud center

1369: appears, and at which dust shielding begins to contribute

1370: significantly to the shielding of H$_2$ molecules. We develop the

1371: equations that determine the sizes of the molecular and atomic

1372: regions in this parameter space, and we provide an approximate

1373: analytic solution for them (equations \ref{xmapprox} -- \ref{xweapprox} and equations \ref{chrisapprox} and \ref{chrisapprox1}). Our solutions are accurate to tens of percent for clouds that are $\gtsim 20\%$ molecular by volume, and provide upper limits on the molecular content at this accuracy for clouds with lower molecular content.

1374: Using this formalism we find that for typical giant atomic-molecular complexes in the Milky Way $\chi\sim 1$, which indicates that dust shielding and self-shielding each make order unity contributions to determining the location of the atomic-molecular transition. 

1375: 

1376: Our work shows that the procedure of determing the structure of PDRs

1377: by treating them as semi-infinite slabs illuminated by unidirectional beams of

1378: dissociating radiation is a reasonable approximation for extremely

1379: opaque clouds, but that it fails badly for small clouds or weak

1380: radiation fields, i.e.\ in cases where the transition from atomic to

1381: molecular is sufficiently far into the cloud that the cloud's

1382: curvature cannot reasonably be neglected. In such cases the slab

1383: approximation can either overestimate or underestimate the size of the

1384: atomic layer by factors of order unity, depending on the particular

1385: parameters of the cloud and the ambient radiation field.

1386: 

1387: The development of an analytic model for the structure of the atomic

1388: envelopes of finite molecular clouds opens up the possibility of

1389: developing a more general theory of the atomic to molecular ratio in

1390: galaxies. In a galaxy, the mean interstellar radiation field and the

1391: conditions in the atomic portion of the atomic interstellar medium are

1392: determined by the star formation rate, which determines the abundance

1393: of young, hot stars. In turn, the star formation rate depends on the

1394: fraction of the ISM of that galaxy in molecular form, and therefore

1395: available for star formation. At some level, therefore, star formation

1396: in galaxies much be a self-regulating process, with the formation and

1397: dissociation of molecular clouds representing one step in that

1398: regulation. Developing a simple model for how the molecular fraction

1399: in a cloud is determined by its properties and those of the ambient

1400: radiation field represents a step toward a complete theory of the star

1401: formation rate. In future work, we plan to develop this

1402: theory further by applying the model demonstrated here to molecular

1403: clouds in galaxies.

1404: 

1405: \acknowledgements We thank B.~Draine and J.~Goodman for helpful discussions, and the anonymous referee for useful comments. Support

1406: for this work was provided by NASA through Hubble Fellowship grant

1407: \#HSF-HF-01186 awarded by the Space Telescope Science Institute, which

1408: is operated by the Association of Universities for Research in

1409: Astronomy, Inc., for NASA, under contract NAS 5-26555 (MRK), and by

1410: the National Science Foundation through grants AST-0098365 (to CFM) and PHY05-51164 (to the Kavli Institute for Theoretical Physics, where MRK, CFM, and JT collaborated on this work). JT gratefully acknowledges the support of Gilbert and Jaylee Mead for their namesake fellowship in the Yale Center for Astronomy and Astrophysics.

1411: 

1412: \appendix

1413: 

1414: \section{Evaluation of $\eta_0$ and $\eta_1$}

1415: \label{fintegrate}

1416: 

1417: Here we evaluate the two functions

1418: \begin{eqnarray}

1419: \eta_0(\xd,\xm;\tauR) & = & \frac{1}{2} \int_{\mumol}^1 d\mu\, \exp(-\tauR

1420: \xi) \\

1421: \eta_1(\xd,\xm;\tauR) & = & \frac{1}{2} \int_{\mumol}^1 d\mu\, \mu \exp(-\tauR

1422: \xi).

1423: \end{eqnarray}

1424: To evaluate the integrals, we change the variable of integration from

1425: $\mu$ to $\xi$. Using the definition of $\xi$ (equation \ref{xidef}),

1426: we find that

1427: \begin{equation}

1428: \mu = \frac{(1-\xd)(1+\xd)-\xi^2}{2\xd\xi}

1429: \end{equation}

1430: and

1431: \begin{equation}

1432: \frac{d\mu}{d\xi} = -\frac{(1-\xd)(1+\xd)+\xi^2}{2\xd\xi^2}.

1433: \end{equation}

1434: Making the change of variable, we find that

1435: \begin{eqnarray}

1436: \eta_0(\xd,\xm;\tauR) & = &

1437: -\frac{1}{4\xd}\left[

1438: \int_{\xim}^{1-\xd} d\xi\, \exp(-\xi \tauR) + 

1439: (1-\xd)(1+\xd)\int_{\xim}^{1-\xd}

1440: d\xi\, \xi^{-2} \exp(-\xi \tauR)

1441: \right]\nonumber\\ \\

1442: \eta_1(\xd,\xm;\tauR) & = &

1443: \frac{1}{8\xd^2}\left[

1444: \int_{\xim}^{1-\xd} d\xi\, \xi \exp(-\xi \tauR) -

1445: (1-\xd)^2(1+\xd)^2

1446: \int_{\xim}^{1-\xd} d\xi\, \xi^{-3} \exp(-\xi \tauR)

1447: \right],\nonumber \\

1448: \end{eqnarray}

1449: where

1450: \begin{equation}

1451: \xim \equiv \xi(\xd,\mumol) = \sqrt{1-\xd^2 + \xd^2\mumol^2}-\xd\mumol.

1452: \end{equation}

1453: Of these integrals, the first one for $\eta_0$ may be evaluated directly,

1454: while the first one for $\eta_1$ may be evaluated by parts. The second

1455: integral on each line may be evaluated via the identity

1456: \begin{eqnarray}

1457: \int_{x_0}^{x_1} dx\, x^{-n} e^{-ax} & = &

1458: \int_{x_0}^{\infty} dx\, x^{-n} e^{-ax} - 

1459: \int_{x_1}^{\infty} dx\, x^{-n} e^{-ax} \\

1460: & = & x_0^{1-n} \int_{1}^{\infty} du\, u^{-n} e^{-ax_0 u}

1461: - x_1^{1-n} \int_{1}^{\infty} dv\, v^{-n} e^{-ax_1 v} \\

1462: & = & x_0^{1-n} {\rm E}_n (a x_0) - x_1^{1-n} {\rm E}_n (a x_1),

1463: \end{eqnarray}

1464: where in the second step we made the change of variables $u=x/x_0$ and

1465: $v=x/x_1$, and ${\rm E}_n$ is the exponential integral

1466: function of order $n$, defined by ${\rm E}_n(x)=\int_{1}^{\infty} t^{-n}

1467: e^{-x t} dt$. Using this identity gives

1468: {\small

1469: \begin{eqnarray}

1470: \label{f0eval}

1471: \eta_0(\xd,\xm;\tauR) & = &

1472: \frac{1}{4\xd} 

1473: \left\{

1474: \frac{e^{-\yd\tauR}-e^{-\xim\tauR}}{\tauR} +

1475: (1+\xd) \left[

1476: {\rm E}_2 (\yd\tauR) - \frac{\yd}{\xim} {\rm E}_2(\xim\tauR)

1477: \right]

1478: \right\}\nonumber \\ \\

1479: \eta_1(\xd,\xm;\tauR) & = &

1480: \frac{1}{8\xd^2} \left\{

1481: \frac{(1+\xim\tauR) e^{-\xim\tauR}-(1+\yd\tauR)e^{-\yd\tauR}}{\tauR^2}

1482: + (1+\xd)^2\left[

1483: {\rm E}_3(\yd\tauR)-\frac{\yd^2}{\xim^2} {\rm E}_3(\xim\tauR)

1484: \right]

1485: \right\},

1486: \label{f1eval}

1487: \end{eqnarray}

1488: }

1489: where $\yd\equiv 1-\xd$. Note that exponential integrals obey the

1490: recurrence relation $n{\rm E}_{n+1}(x) = e^{-x} - x {\rm E}_n(x)$, so

1491: we could alternately have written these in terms of ${\rm E}_1(x)$ or

1492: the classical exponential integral ${\rm Ei}(x) = - {\rm E}_1(-x)$.

1493: 

1494: \section{Solution by Series Expansion in the Strong Radiation Limit}

1495: \label{strongappendix}

1496: 

1497: Here we solve equations (\ref{f0eqn}) and (\ref{f1eqn}) in the strong radiation limit, i.e. $\xm\ll 1$, $\xd\ll 1$, and $\mumol+1 \ll 1$, by means of series expansion. Let $\beta=\xd \tauR$ and $\gamma = 1+\mumol$. Then we have

1498: \begin{equation}

1499: \label{betaexpansion}

1500: e^{-\tauR \xi} = e^{-\tauR} 

1501: \left[1 + \mu \beta + \left(\frac{1}{2\tauR}+\frac{\tauR+1}{2\tauR}\mu^2\right) \beta^2

1502: + \left(\frac{1}{2\tauR} \mu + \frac{\tauR-3}{6\tauR} \mu^3\right) \beta^3 + O(\beta^4)\right].

1503: \end{equation}

1504: Note that we have retained terms out to order $\beta^3$. We shall see below that this is required for a consistent solution. Using this expansion in the integrals

1505: \begin{eqnarray}

1506: \eta_0(\xd,\xm;\tauR) & = & \frac{1}{2} \int_{\gamma-1}^{1} d\mu\, e^{-\tauR\xi} \\

1507: \eta_1(\xd,\xm;\tauR) & = & \frac{1}{2} \int_{\gamma-1}^{1} d\mu\, \mu e^{\tauR\xi}

1508: \end{eqnarray}

1509: and expanding in powers of $\gamma$, we find

1510: \begin{eqnarray}

1511: \eta_0(\xd,\xm;\tauR) & = & \frac{e^{-\tauR}}{2} 

1512: \left[

1513: 2-\gamma+\gamma\beta+\frac{\tauR+2}{3\tauR}\beta^2 + O(\beta^4) + O(\gamma\beta^2) + O(\gamma^2)

1514: \right] \\

1515: \eta_1(\xd,\xm;\tauR) & = & \frac{e^{-\tauR}}{2}

1516: \left[

1517: \frac{2}{3}\beta + \gamma - \gamma\beta + \frac{\tauR+2}{15}\beta^3 + O(\beta^4) + O(\gamma\beta^2) + O(\gamma^2)

1518: \right].

1519: \end{eqnarray}

1520: If we similarly expand the right-hand sides of (\ref{f0eqn}) and (\ref{f1eqn}) in powers of $\beta$ and $\gamma$, we obtain the two equations

1521: \begin{eqnarray}

1522: \label{f0eqnstrong}

1523: \frac{e^{-\tauR}}{2} 

1524: \left[2-\gamma+\gamma\beta+\frac{\tauR+2}{3\tauR}\beta^2\right]

1525: & = & \frac{1}{\psi} + O(\beta^4) + O(\gamma\beta^2) + O(\gamma^2)

1526: \\

1527: \label{f1eqnstrong}

1528: \frac{e^{-\tauR}}{2} 

1529: \left[\frac{2}{3}\beta + \gamma - \gamma\beta + \frac{\tauR+2}{15}\beta^3\right]

1530: & = & \frac{\beta}{3\psi}  + O(\beta^4) + O(\gamma\beta^2) + O(\gamma^2).

1531: \end{eqnarray}

1532: If we combine these two equations by eliminating the common factor $e^{\tauR}/\psi$, we obtain an equation for the relationship between $\beta$ and $\gamma$:

1533: \begin{equation}

1534: -\frac{\gamma}{2}+\frac{\gamma\beta}{2} + \frac{\tauR+2}{6\tauR} \beta^2 =

1535: \frac{3}{2}\left(\frac{\gamma}{\beta}\right) - \frac{3}{2}\gamma + \frac{\tauR+2}{10\tauR} \beta^2

1536: + O(\beta^3) + O(\gamma\beta) + O\left(\frac{\gamma^2}{\beta}\right).

1537: \end{equation}

1538: The only way for this equation to have a consistent solution in which the orders on both sides balance is if $\gamma$ is of order $\beta^3$. In this case the leading order on both sides is $\beta^2$ (an order we retained only be performing the expansion in equation \ref{betaexpansion} to order $\beta^3$), and balancing the leading order terms gives

1539: \begin{equation}

1540: \label{gammasol}

1541: \gamma = \frac{2\tauR+4}{45\tauR} \beta^3.

1542: \end{equation}

1543: Since we now know the order of all terms, we can solve equation (\ref{f0eqnstrong}) to leading order to obtain

1544: \begin{equation}

1545: \label{xdsol}

1546: \xd = \frac{\beta}{\tauR} = \left[\frac{6}{\tauR(\tauR+2)}\left(\frac{e^{\tauR}}{\psi}-1\right)\right]^{1/2}.

1547: \end{equation}

1548: Similarly, we know that

1549: \begin{equation}

1550: \left(\frac{\xm}{\xd}\right)^2 = 1-\mumol^2 = 2\gamma + O(\gamma^2).

1551: \end{equation}

1552: Substituting (\ref{gammasol}) for $\gamma$ and (\ref{xdsol}) for $\xd$ and re-arranging, we obtain to leading order

1553: \begin{equation}

1554: \xm = \left[\frac{1536}{25\tauR(\tauR+2)^3}\right]^{1/4} \left(\frac{e^{\tauR}}{\psi}-1\right)^{5/4}.

1555: \end{equation}

1556: 

1557: \section{Solution by Series Expansion in the Large and Small Cloud Limits}

1558: \label{largesmallappendix}

1559: 

1560: Here we solve equations (\ref{f0eqnodust}) and (\ref{f1eqnodust}) by

1561: series expansion in the limits $\tauR\rightarrow 0$ and

1562: $\tauR\rightarrow \infty$. We approach this

1563: problem by defining $\beta = \taueR/\tauR$, so that with some

1564: rearrangement the equations are

1565: \begin{eqnarray}

1566: \label{f0ndlim}

1567: 1-e^{2\beta\mumol\tauR}+2\beta\tauR-\frac{4}{\psi}\tauR & = & 0\\

1568: \label{f1ndlim}

1569: (1-2\beta\mumol\tauR)e^{2\beta\mumol\tauR}-1+2\beta^2\tauR^2 -

1570: \frac{8}{3\psi} \beta^2\tauR^3 \left[1-(1-\mumol^2)^{3/2}\right] & = &

1571: 0.

1572: \end{eqnarray}

1573: For the case $\tauR\rightarrow 0$ we then let

1574: \begin{eqnarray}

1575: \beta & = & \beta_0 + \beta_1 \tauR + \beta_2\tauR^2 + \cdots \\

1576: \mumol & = & \mu_0 + \mu_1 \tauR + \mu_2\tauR^2 + \cdots.

1577: \end{eqnarray}

1578: Expanding equations (\ref{f0ndlim}) and (\ref{f1ndlim}) to

1579: leading order in $\tauR$ and re-arranging gives

1580: \begin{eqnarray}

1581: 2\beta_0(1-\mu_0) - \frac{4}{\psi} & = & 0 \\

1582: 2\beta_0^2(1 - \mu_0^2) & = & 0,

1583: \end{eqnarray}

1584: which has the solution $\mu_0=-1$, $\beta_0=1/\psi$. Using these

1585: values and continuing the expansion to the next order, we obtain

1586: \begin{eqnarray}

1587: -2\beta_1 + \frac{1+\mu_1\psi}{\psi^2} & = & 0 \\

1588: \frac{4\mu_1}{\psi^2} & = & 0,

1589: \end{eqnarray}

1590: so $\mu_1=0$ and $\beta_1=1/(2\psi^2)$. Continuing to one more order,

1591: we have

1592: \begin{eqnarray}

1593: 4\beta_2-\frac{2+6\beta_2\psi^2}{3\psi^3} & = & 0\\

1594: -1 + 6\mu_2 \psi^2 & = & 0,

1595: \end{eqnarray}

1596: so $\mu_2=1/(6\psi^2)$ and $\beta_2=1/(4\psi^3)$. It improves the

1597: accuracy of the approximation for $\xm$ significantly at small $\psi$

1598: to include one more order, so we do so:

1599: \begin{eqnarray}

1600: 4\beta_3+\frac{1}{3\psi^4}-\frac{2}{\psi}\mu_3 & = & 0\\

1601: \frac{4}{\psi^2}\mu_3-\frac{8}{5\psi^5} & = & 0,

1602: \end{eqnarray}

1603: so $\mu_3=2/(5\psi^3)$ and $\beta_3=7/(60\psi^4)$. Therefore to order

1604: $\tauR^4$ we have

1605: \begin{eqnarray}

1606: \frac{\taueR}{\tauR} & = & \frac{1}{\psi} + \frac{\tauR}{2\psi^2} +

1607: \frac{\tauR^2}{4\psi^3} + \frac{7\tauR^3}{60\psi^4}\\

1608: \mumol & = & -1 + \frac{\tauR^2}{6\psi^2}+\frac{2\tauR^3}{5\psi^3}\\

1609: \xm & = & \sqrt{1-\mumol^2} =

1610: \frac{\tauR}{\sqrt{3}\psi}+\frac{2\sqrt{3} \tauR^2}{5\psi^2}.

1611: \end{eqnarray}

1612: 

1613: For the case $\tauR\rightarrow\infty$, we let

1614: \begin{eqnarray}

1615: \beta & = & \beta_0 + \beta_{-1/2}\tauR^{-1/2} + \beta_{-1}\tauR^{-1}+\cdots

1616: \\

1617: \mumol & = & \mu_0 + \mu_{-1/2}\tauR^{-1/2} + \mu_{-1}\tauR^{-1}+\cdots,

1618: \end{eqnarray}

1619: and if we expand equations (\ref{f0ndlim}) and (\ref{f1ndlim}) in

1620: powers of $\tauR^{-1}$ then the leading order equations are

1621: \begin{eqnarray}

1622: 2\beta_0 - \frac{4}{\psi} & = & 0 \\

1623: -\frac{8}{3\psi} \left[1-(1-\mu_0^2)^{3/2}\right] & = & 0.

1624: \end{eqnarray}

1625: Therefore $\beta_0=2/\psi$ and $\mu_0=0$. Continuing the expansion to

1626: the next order,

1627: \begin{eqnarray}

1628: 2\beta_{-1/2} & = & 0\\

1629: -\frac{16}{\psi^3}\mu_{-1/2}^2+\frac{8}{\psi^2} & = & 0,

1630: \end{eqnarray}

1631: so $\beta_{-1/2}=0$ and $\mu_{-1/2} = -\sqrt{\psi/2}$. Continuing one

1632: more order,

1633: \begin{eqnarray}

1634: 1+2\beta_{-1} & = & 0\\

1635: \frac{16\sqrt{2}}{\psi^5/2}\mu_{-1} & = & 0,

1636: \end{eqnarray}

1637: so $\beta_{-1}=-1/2$ and $\mu_{-1}=0$

1638: Thus to order $\tauR^{-1}$ in the limit $\tauR\rightarrow\infty$ we have

1639: \begin{eqnarray}

1640: \frac{\taueR}{\tauR} & = & \frac{2}{\psi} \\

1641: \mumol & = & -\sqrt{\frac{\psi}{2\tauR}} \\

1642: \xm & = & 1 - \frac{\psi}{4\tauR}.

1643: \end{eqnarray}

1644: 

1645: 

1646: %\bibliographystyle{apj}

1647: %\bibliography{refs}

1648: \input{ms.bbl}

1649: 

1650: %

1651: %\clearpage

1652: 

1653: %\begin{figure}

1654: %%\plotone{fh2prof.pdf}

1655: %\plotone{f1.pdf}

1656: %\caption{

1657: %\label{fh2prof}

1658: %The plots show $\fmol$ versus dust optical depth $\tau=n\sigmad z$ for

1659: %our numerical radiative transfer calculations with $\log Z'=-1.5$

1660: %(\textit{dashed lines}), $\log Z'=-0.5$ (\textit{solid lines}), and

1661: %$\log Z'=0.5$ (\textit{dot-dashed lines}). The gray vertical lines

1662: %indicate the optical depth of the transition to fully molecular as

1663: %calculated with the two-zone approximation, equation

1664: %(\ref{slabtwozone}). Each cluster of three

1665: %curves plus a vertical line indicating a prediction

1666: %corresponds to a radiation flux $\log F_{\nu}^*=-7$, $-5$, or

1667: %$-3$, as indicated. The two panels are for the cases

1668: %$\sigma_{\rm d,MW,-21}=0.6$ and $2.0$, as indicated.

1669: %}

1670: %\end{figure}

1671: 

1672: %\clearpage

1673: %\thispagestyle{empty}

1674: %\begin{figure}

1675: %\vspace*{-10mm}

1676: %%\plotone{hicolcomp.pdf}

1677: %\plotone{f2.pdf}

1678: %\caption{

1679: %\label{hicolcomp}

1680: %The upper panel shows the dust opacity $\nhi\sigmad$

1681: %through the numerically-determined H\textsc{i} column (\textit{various

1682: %symbols}) as a function of $\chi$. The values of $Z'$ and

1683: %$\sigma_{\rm d,MW,{-21}}$ for each calculation are indicated by the plot

1684: %symbol. This is compared to the optical depth $\taum$ computed from

1685: %the two-zone approximation (equation \ref{slabtwozone}, \textit{thin solid

1686: %line}) and computed using the analytic solution (equation \ref{slabexact},

1687: %\textit{thick dashed line}). The lower panel shows the error in the two-zone

1688: %approximation, defined as $\mbox{Error} = \nhi\sigmad/\taum-1$; the dotted line indicates zero error.

1689: %}

1690: %\end{figure}

1691: 

1692: %\clearpage

1693: 

1694: %\begin{figure}

1695: %%\epsscale{0.7}

1696: %%\plotone{diagram.pdf}

1697: %%\epsscale{0.7}

1698: %\plotone{f3.pdf}

1699: %\caption{

1700: %\label{diagram}

1701: %Illustration of the two-zone approximation in spherical geometry.

1702: %}

1703: %\end{figure}

1704: 

1705: %\clearpage

1706: 

1707: %\begin{figure}

1708: %%\plotone{chitauzone.pdf}

1709: %\plotone{f4.pdf}

1710: %\caption{

1711: %\label{chitauzone}

1712: %The thin black curves show the boundaries in the $(\tauR,\psi)$-plane at

1713: %which $\xd=1$, $\taueR=\tauR$, and at which $\xm=0$, as indicated by

1714: %the text accompanying each curve. The thick dashed curve shows our

1715: %approximation to the $\xd=1$ curve, equation (\ref{weakapprox}).

1716: %The asterisks along the curve for $\xd=1$ mark the points at which

1717: %$\xm=0.1$, $0.3$, $0.5$, $0.7$, and $0.9$, as indicated.

1718: %}

1719: %\end{figure}

1720: 

1721: %\clearpage

1722: 

1723: %\begin{figure}

1724: %%\plotone{planeplot.pdf}

1725: %\plotone{f5.pdf}

1726: %\caption{

1727: %\label{planeplot}

1728: %The plot shows the solution in the large cloud limit. The curves shown

1729: %are the dust optical depth to the fully molecular

1730: %region $\taum=\tauR(1-\xm)$ (\textit{solid curve}), the dust optical depth to the

1731: %point of dust-molecular absorption equality $\taud=\tauR(1-\xd)$ (\textit{dashed

1732: %curve to the right of $\psi=2$}), and the ratio of the effective molecular

1733: %opacity to the dust opacity $\taueR/\tauR$ (\textit{dashed

1734: %curve to the left of $\psi=2$}). The dotted vertical line at $\psi=2$

1735: %indicates the boundary between the presence and absence of a

1736: %dust-dominated zone in the weak radiation limit.

1737: %}

1738: %\end{figure}

1739: 

1740: %\clearpage

1741: 

1742: %\begin{figure}

1743: %%\plotone{chitaucont.pdf}

1744: %\plotone{f6.pdf}

1745: %\caption{

1746: %\label{chitaucont}

1747: %Contours showing the solution as a function of $\tauR$ and $\psi$ for

1748: %the structure of the PDR in the two-zone approximation. The values

1749: %shown are, clockwise from the upper left, $\xm$, $\xd$ or

1750: %$\tauR/\taueR$, $\tau_{\rm HI}$, and $\xm^3$. The hatched region is

1751: %the region in which there is no primarily molecular part of the cloud.

1752: %The dotted line

1753: %indicates the boundary between the presence and absence of a

1754: %dust-dominated zone. In the panel labelled $\xd$, $\tauR/\taueR$, the

1755: %contours above the dotted line indicate the value of $\xd$, those

1756: %below it show the value of $\tauR/\taueR$, and on the dotted line both

1757: %of these quantities are exactly $1.0$. We caution that the contours for $\xm = 0.1$ and $\xm=0.3$, and for $\xm^3 = 0.1$, should be regarded as giving upper limits on $\xm$, not precise estimates -- see the discussion in \S~\ref{error}.

1758: %}

1759: %\end{figure}

1760: 

1761: %\clearpage

1762: 

1763: %\begin{figure}

1764: %%\plotone{nodust.pdf}

1765: %\plotone{f7.pdf}

1766: %\caption{

1767: %\label{nodust}

1768: %Contours showing the factor by which dust shielding changes the atomic or molecular volume at a given $(\tauR, \psi)$. The quantity plotted is $\mbox{max}[\xm/x_{\rm H_2,nd}, (1-x_{\rm H_2,nd})/(1-\xm)]$, where the subscript ``nd" indicates the value with no dust shielding in the limit, i.e. in the limit $\sigmad\rightarrow 0$. The quantity plotted is therefore the fractional amount by which dust shielding increases the radius of the molecular zone or decreases the radial path length through the atomic zone, whichever is larger.

1769: %}

1770: %\end{figure}

1771: 

1772: %\clearpage

1773: 

1774: %\begin{figure}

1775: %%\plotone{chitaufit.pdf}

1776: %\plotone{f8.pdf}

1777: %\caption{

1778: %\label{chitaufit}

1779: %Error in the approximate analytic fit given in \S~\ref{numericsol} as

1780: %a function of $\tauR$ and $\psi$. The shaded regions indicate errors

1781: %below $2.5\%$ (\textit{no shading}), $2.5-5\%$, $5-10\%$, $10-20\%$, and $>20\%$ (\textit{darkest shading}). The maximum error is 24\%. The dotted lines show the boundaries of our

1782: %different approximation regions: $\psi<1$, $1<\psi<\psi_{\rm b}$,

1783: %$\psi_{\rm b} < \psi < 2$, and $2<\psi$. The

1784: %hatched region is $\psi>e^{\tauR}$, where there is no

1785: %predominantly molecular core. Note that the error jumps at the $\chi=1$ and $\chi=2$ lines because the fitting formula is slightly discontinuous there.

1786: %}

1787: %\end{figure}

1788: 

1789: %\clearpage

1790: %\thispagestyle{empty}

1791: %\begin{figure}

1792: %\vspace*{-10mm}

1793: %%\plotone{slabcomp.pdf}

1794: %\plotone{f9.pdf}

1795: %\caption{

1796: %\label{slabcomp}

1797: %The upper panel shows the dust optical depth to the point where the

1798: %gas becomes predominantly molecular $\tau_{\rm HI}$ for an isotropic

1799: %radiation field of normalized intensity $\psi$ and various values of $\tauR$ (\textit{solid lines}) and for a unidirectional radiation field of normalized intensity $\psi/2$ (\textit{dashed line}). The lower panel shows the fractional difference between the results for finite $\tauR$ and isotropic radiation and for an infinite slab illuminated by unidirectional radiation,

1800: %defined as $\mbox{Difference} = [\tau_{\rm HI,iso}-\tau_{\rm

1801: %HI,beam}]/\tau_{\rm HI,iso}$. In

1802: %all cases the curves for finite $\tauR$ end at the value of $\psi$ for

1803: %which fully molecular region disappears ($\psi=e^{\tauR}$).

1804: %}

1805: %\end{figure}

1806: 

1807: %\clearpage

1808: 

1809: %\begin{figure}

1810: %%\plotone{chitaucontho.pdf}

1811: %\plotone{f10.pdf}

1812: %\caption{

1813: %\label{chitaucontho}

1814: %Same is Figure \ref{chitaucont}, but assuming infinite attenuation along rays that enter the  molecular absorption-dominated region (equations \ref{f0eqnho}--\ref{nodustho}).

1815: %}

1816: %\end{figure}

1817: 

1818: %\clearpage

1819: 

1820: %\begin{figure}

1821: %%\plotone{chitaucomp.pdf}

1822: %\plotone{f11.pdf}

1823: %\caption{

1824: %\label{chitaucomp}

1825: %The plots show the difference between $\taum$ and $\xm^3$ computed under the fiducial assumption that the opacity throughout the molecule-dominated region equal to that at its surface (\S~\ref{spheredust} and \S~\ref{spherenodust}) and under the assumption of infinite attenuation in this region (\S~\ref{error}).  The difference is defined as $|\tau_{\rm HI,fiducial}-\tau_{\rm HI,attenuated}|/\tau_{\rm HI,fiducial}$, and similarly for $\xm^3$. The hatched region is the region in which there is no predominantly molecular core under either assumption; the difference in this region is obviously zero.

1826: %}

1827: %\end{figure}

1828: 

1829: %--------------------------------------------------------------------------

1830: \end{document}

1831: 

1832: