0807.2398/ms.tex
1: %       Normal article style for journal.
2: %\documentstyle[12pt,aasms4]{article}
3: %       Preprint style.
4: %\documentstyle[12pt,aaspp4]{article}
5: %\documentstyle[emulateapj]{article}
6: %       Two-Column Preprint style.
7: %\documentstyle[10pt,aas2pp4,tighten]{article}
8: \documentstyle[preprint]{aastex}
9: \bibliographystyle{apj}
10: \received{}
11: \revised{}
12: \accepted{}
13: \journalid{}{}
14: \articleid{}{}
15: \paperid{}
16: \cpright{}{}
17: \ccc{}
18: 
19: \slugcomment{Submitted to the Astrophysical Journal}
20: 
21: \lefthead{Wood, Whitney, Robitaille, \& Draine}
22: \righthead{Emission from PAHs \& VSGs in Monte Carlo Radiation Transfer Codes}
23: %\righthead{Carcinogens in Gomez's Hamburger}
24: 
25: \begin{document}
26: 
27: \title{Emission from Very Small Grains and PAH Molecules in 
28: Monte Carlo Radiation Transfer Codes:\\ 
29: Application to the Edge-On Disk of Gomez's Hamburger}
30: 
31: %\title{Carcinogens in Gomez's Hamburger}
32: 
33: \author{Kenneth Wood\altaffilmark{1}, Barbara A. Whitney\altaffilmark{2}, Thomas Robitaille\altaffilmark{1}, 
34: \& Bruce T. Draine\altaffilmark{3} }
35: 
36: \altaffiltext{1}{School of Physics \& Astronomy, University of St Andrews, 
37: North Haugh, St Andrews, Fife, KY16 9AD, Scotland; kw25@st-andrews.ac.uk}
38: 
39: \altaffiltext{2}{Space Science Institute, 4750 Walnut St. Suite 205, Boulder, CO 80301; 
40: bwhitney@spacescience.org}
41: 
42: \altaffiltext{3}{Department of Astrophysical Sciences, Princeton University, Peyton Hall, 
43: Princeton, NJ 08544}
44: 
45: \authoremail{kw25@st-andrews.ac.uk}
46: 
47: 
48: \begin{abstract}
49: 
50: We have modeled optical to far infrared images, photometry, and 
51: spectroscopy of the object known as Gomez's Hamburger. We reproduce the 
52: images and spectrum with an edge-on disk of mass $0.3M_\odot$ and radius 
53: 1600~AU, surrounding an A0 III star at a distance of 280~pc. Our mass estimate is 
54: in excellent agreement with recent CO observations. However, our distance 
55: determination is more than an order of magnitude smaller than previous 
56: analyses which inaccurately interpreted the optical spectrum. To 
57: accurately model the infrared spectrum we have extended our Monte Carlo 
58: radiation transfer codes to include emission from polycyclic aromatic hydrocarbon (PAH) molecules 
59: and very small grains (VSG). We do this using pre-computed PAH/VSG emissivity files for 
60: a wide range of values of the mean intensity of the exciting radiation field. 
61: When Monte Carlo energy packets are absorbed by PAHs/VSGs we reprocess them to 
62: other wavelengths by sampling 
63: from the emissivity files, thus simulating the absorption and re-emission process without 
64: reproducing lengthy computations of statistical equilibrium, excitation and de-excitation 
65: in the complex many level molecules. 
66: Using emissivity lookup tables in our Monte Carlo codes gives the flexibility to use 
67: the latest grain physics calculations of PAH/VSG emissivity and opacity that are being 
68: continually updated in the light of higher resolution infrared spectra.
69: We find our approach gives a good representation of the 
70: observed PAH spectrum from the disk of Gomez's Hamburger. 
71: Our models also indicate the PAHs/VSGs in the disk have a larger 
72: scaleheight than larger radiative equilibrium grains, providing evidence for dust coagulation and settling 
73: to the midplane.
74: 
75: \end{abstract}
76: 
77: 
78: \keywords{accretion, accretion disks --- radiative transfer --- stars: circumstellar matter --- 
79: ISM: dust, extinction}
80: 
81: \section{Introduction}
82: 
83: The object known as Gomez's Hamburger \citep{Ruiz87} exhibits the spectral and imaging 
84: characteristics of an edge-on circumstellar disk. High resolution HST imagery shows 
85: a bipolar reflection nebula of size 5.5\arcsec by 2.5\arcsec bisected by an optically thick 
86: dust lane, reminiscent of 
87: models and images of disks surrounding low mass Classical T~Tauri stars 
88: \citep[e.g.,][]{WH92, Burrows96, Grosso2003}. The HST data support the 
89: circumstellar disk (or torus) interpretation first proposed by \citet{Ruiz87} based on lower 
90: resolution optical imaging 
91: polarimetry and the object's broad band spectral energy distribution (SED) which is 
92: faint in the optical and bright at mid- to far-infrared wavelengths. Such SEDs 
93: are characteristic of edge-on circumstellar disks \citep[e.g.,][]{Wood2002}. The disk hypothesis 
94: for Gomez's Hamburger has been strengthened by recent CO observations  \citep{Bujarr2008} 
95: that are reproduced with a model of a Keplerian disk of mass $0.9 M_\odot (d/500{\rm pc})^2$, 
96: where $d$ is the distance to the source.
97: 
98: Optical spectroscopy of Gomez's Hamburger indicates the central source is an A0 III star. 
99: Using parameters typical 
100: to this spectral type and the integrated broad band spectrum, \citet{Ruiz87}
101: estimated the object to be at a distance of 2.9~kpc with a circumstellar disk of radius 8000~AU. 
102: However, as we will show below the disk is smaller and closer than estimated by \citet{Ruiz87}. 
103: This is because they did not take into consideration that the integrated spectrum 
104: from an edge-on disk is strongly dependent on inclination \citep{Whitney2003a}. 
105: At optical wavelengths 
106: edge-on disks are only seen in scattered light and can be several orders of magnitude 
107: fainter than systems where the central star is directly seen. In addition to the error in distance determination, 
108: the spectral type of the central source may not be accurate because the optical spectrum is entirely 
109: scattered light. Therefore, the combination of high resolution optical spectroscopy 
110: and radiation transfer modeling of the scattered starlight from the disk is required for more accurate spectral 
111: typing. However, in the absence of higher resolution optical spectroscopy, in our modeling below 
112: we adopt the A0 III spectral type from \citet{Ruiz87}.
113: 
114: More recent Spitzer IRS observations of Gomez's Hamburger show strong mid infrared 
115: emission features typical of emission from polycyclic aromatic hydrocarbon (PAH) 
116: molecules and very small grains (VSGs). Such emission features are common in massive star 
117: forming regions \citep[e.g.,][]{Churchwell2006} as the emission mechanism is sensitive to 
118: the strength of the exciting UV 
119: radiation from O and B stars. However, charged PAH molecules exhibit significant opacity at 
120: optical and infrared wavelengths, so that the PAH excitation may arise from softer photons 
121: \citep{LD2002}, thus enabling PAH emission to be observed from the circumstellar environment of lower 
122: mass stars. Indeed, PAH emission has been detected from disks and envelopes around A stars 
123: and classical T~Tauri stars \citep[e.g.,][]{Habart2004, Pontop2007, Geers2006}.
124: 
125: In this paper we present models for the HST images, broad band spectral energy distribution, 
126: and high resolution Spitzer IRS spectrum of Gomez's Hamburger. In doing so, in \S2 we introduce 
127: extensions to our Monte Carlo radiation transfer codes that allow us to model 
128: emission from PAHs and VSGs. \S3 presents the radiation transfer modeling of Gomez's 
129: Hamburger, and in \S4 we summarize our findings and briefly discuss future applications of our new 
130: code developments.
131: 
132: \section{PAH/VSG Emission Processes}
133: 
134: Emission from PAH molecules and 
135: VSGs can dominate the mid-infrared spectrum 
136: from dusty regions associated with massive star formation.  New 
137: mid-infrared images from the {\it Spitzer Space Telescope} GLIMPSE survey 
138: \citep{Churchwell2004, Churchwell2006} dramatically show PAHs acting as a signpost for 
139: massive star formation.  
140: The characteristic infrared PAH emission spectrum has been studied in 
141: detail in many star-forming regions in the Milky Way and other galaxies using a variety of 
142: ground and space-based observatories 
143: \citep[e.g.,][]{Peeters2004, DraineSINGS2007}.  
144: A proper understanding of the 
145: infrared emission from star-forming regions and galaxies therefore requires 
146: the inclusion of PAHs and VSGs in radiation transfer codes.
147: 
148: Unlike the radiative equilibrium temperatures attained by large dust grains, 
149: the absorption and reprocessing of radiation by PAHs and VSGs does not 
150: yield an equilibrium temperature.  Absorption of a single photon can appreciably increase the energy 
151: content of the molecule, and the ensuing radiative de-excitation 
152: produces an infrared spectrum consisting of a continuum plus characteristic emission features.
153: Detailed modeling of the emission process therefore 
154: requires solving statistical equilibrium in complex multi-level molecules. 
155: 
156: Approximations exist for determining the reprocessing of photons and emission from PAHs and 
157: VSGs and have been incorporated into some radiation transfer codes 
158: \citep[e.g.,][]{Misselt2001, Pontop2007, Dullemon2007, Manske1998}. In this 
159: paper we introduce a new technique that is readily included into Monte Carlo 
160: radiation transfer codes and utilizes the detailed PAH and VSG opacity and 
161: emissivity calculations presented by \cite{DL2007}. 
162: 
163: \subsection{PAH/VSGs in Monte Carlo Radiation Transfer Codes}
164: 
165: Our current Monte Carlo radiation transfer codes \citep[see details in][]{Whitney2003a} 
166: treat multiple anisotropic scattering in three dimensions and calculate radiative 
167: equilibrium dust temperatures and spectra using the \citet{BW2001} technique 
168: or iteratively using methods described by \citet{Lucy1999}. 
169: The \citet{BW2001} technique works well for two and three dimensional systems \citep[e.g.,][]{Remy2006}, 
170: but to determine higher signal-to-noise temperatures for the same number of Monte Carlo 
171: energy packets, especially for three dimensional systems, 
172: the techniques described by \citet{Lucy1999} are more efficient \citep{CW2008}. 
173: The \citet{BW2001}
174: technique produces the same temperatures and spectra as Lucy's method, but requires more 
175: energy packets (or Monte Carlo photons) and hence longer run times for three dimensional 
176: simulations. 
177: 
178: \citet{Lucy1999} introduced efficient 
179: ways for determining mean intensities and hence temperatures and ionization fractions 
180: in Monte Carlo codes by summing the path lengths within grid cells of individual "energy 
181: packets" as they take their random walks in the radiation transfer simulation. In particular, 
182: the Monte Carlo estimator for the mean intensity in a cell is given by,
183: %
184: \begin{equation}
185: J_i = {\epsilon \over{4\pi\Delta t\Delta V_i}}\sum l
186: \end{equation}
187: %
188: where the $\epsilon = L \Delta t /N$ is the energy of each of the $N$ Monte Carlo energy packets, 
189: $L$ is the system luminosity, and $\Delta t$ is the time interval of the Monte Carlo simulation. 
190: The summation is over all pathlengths, $l$, of energy packets passing through 
191: cell $i$ of volume $V_i$. We will make use of this Monte Carlo estimator below. 
192: 
193: We have extended our Monte Carlo codes to allow for multiple dust species within any given 
194: cell in the simulation grid. When Monte Carlo energy packets interact with dust in a cell the 
195: dust type they interact with is chosen from the relative opacity of each dust species present,
196: %
197: \begin{equation}
198: P_i = {{\rho_i\kappa_i}\over{\sum_j \rho_j\kappa_j}}
199: \end{equation}
200: %
201: Where $P_i$ is the probability of the energy packet interacting with dust type $i$ which has 
202: density and opacity $\rho_i$ and $\kappa_i$.
203: 
204: If a Monte Carlo energy packet is absorbed by a dust type that attains a radiative equilibrium 
205: temperature, the energy packet is reprocessed and the incremental temperature 
206: change of the cell is calculated using the Bjorkman \& Wood technique. Therefore at the end of 
207: the simulation we have a temperature grid for each dust type that attains radiative equilibrium 
208: temperatures. 
209: 
210: If however, a Monte Carlo energy packet is absorbed by PAH/VSG opacity, the packet is 
211: reprocessed sampling a new frequency from the pre-computed emissivity files of \citet{DL2007}. 
212: Since the PAH/VSG emissivity has been computed for a wide range of values of the 
213: mean intensity of the exciting radiation field, we choose the relevant emissivity file based on 
214: the value of the mean intensity in the cell where the Monte Carlo energy packet was absorbed. 
215: This requires an iterative procedure whereby on the first iteration energy packets absorbed by 
216: PAHs/VSGs are reprocessed using the emissivity file for $J=1$, where $J$ is 
217: the mean intensity relative to the average value for the interstellar radiation field, 
218: $J_{ISRF} = 2.17\times 10^{-2} {\rm erg}\, {\rm cm}^{-2}\, {\rm s}^{-1}$ 
219: \citep{MathisISRF}. On subsequent iterations energy packets are reprocessed 
220: based on the mean intensity from the previous iteration, calculated using Eq.~1. 
221: Due to the small PAH/VSG fractions in our simulations (typically less than $4\%$) photons 
222: reprocessed by PAHs/VSGs do not contribute significantly to the mean intensity. Therefore 
223: we find that this procedure requires at most three iterations for convergence.
224: 
225: \subsection{Approximations for PAH/VSG Opacities and Emissivities}
226: 
227: The extensions to our Monte Carlo codes of allowing for multiple dust types mean that 
228: in principle we could employ opacity and emissivity files for grains of many different sizes. 
229: However, in this first paper we have taken a simpler approach of introducing a cutoff in the 
230: grain sizes at $a=200$~{\AA}, below which we assume grains are transiently heated and 
231: energy packets absorbed by these small grains are reprocessed as described above. This 
232: is a somewhat arbitrary cutoff, but it is a size above which grains generally attain radiative 
233: equilibrium temperatures \citep{DL2007}.
234: 
235: In addition to introducing a size cutoff between PAH/VSGs and grains in radiative equilibrium, 
236: we make the assumption that the PAH/VSG emissivity is a function only of the mean intensity 
237: of the radiation field and not the spectral shape of the exciting radiation field. The \citet{DL2007} 
238: emissivities that we use were computed for an exciting radiation field with spectral shape of the model 
239: interstellar radiation field of our Galaxy \citep{MathisISRF}. We make this 
240: approximation in the absence of emissivity calculations for a wide range of exciting spectra. This is 
241: an appropriate first approximation since the Monte Carlo code 
242: accurately treats the transfer and absorption of photons at all wavelengths by the PAH/VSG opacity. 
243: So in our technique the correct number of photons are reprocessed by PAHs/VSGs, and their 
244: wavelengths are chosen from the pre-computed mean intensity dependent emissivities. 
245: 
246: We use the opacity and emissivity files for PAHs/VSGs with sizes $a < 200$~{\AA} taken 
247: from the \citet{DL2007} silicate-graphite-PAH grain model.  Their model assumes 
248: a power law grain size distribution with the silicate and graphite abundances, PAH charging, and 
249: PAH mass fraction adjusted to reproduce the opacity and infrared 
250: spectrum of the Milky Way's diffuse interstellar medium. In particular they find a PAH mass fraction 
251: of 4.5\% is required to reproduce the observed infrared spectra in the Milky Way and 
252: other galaxies. 
253: 
254: Figures~1 and 2 show the wavelength dependent opacity and emissivities for PAHs/VSGs 
255: with grain sizes $a < 200${\AA}. This is the grain size below which we use the 
256: described approximation for computing the PAH/VSG reprocessing of absorbed energy 
257: packets. Notice that the PAH/VSG opacity is not negligible at long wavelengths, so 
258: even photons with just a few eV of energy can excite PAH emission, as described by \citet{LD2002}. 
259: The shape of the continuum emission in the $20-35\mu$m region is sensitive to the intensity of the illuminating 
260: radiation field, but the spectral shape and relative strengths 
261: of the PAH emission features are relatively 
262: insensitive to U for $U < 10^4$ \citep[see, e.g., Fig. 13 of][]{DL2007}, where $U$ is the energy density 
263: of the radiation field relative to that of the interstellar radiation field of \citet{MathisISRF}. 
264: Our reprocessing 
265: technique accounts for the fact that single photons may excite PAHs/VSGs in environments where 
266: the mean intensity is very low. Our technique also incorporates the effect of multiphoton excitation 
267: because the \citet{DL2007} emissivity calculations include this effect which becomes 
268: important in regions of high mean intensity. 
269: 
270: 
271: \section{Gomez's Hamburger Models}
272: 
273: Figures 3 and 4 show spectra and images of our model of Gomez's Hamburger that reproduces the 
274: observed spectrum and {\it HST} images. The broadband optical and near-infrared photometry, as 
275: well as the IRAS 60\,$\mu$m flux, were taken from \citet{Ruiz87}. The IRAC images and IRS spectrum 
276: were extracted from the Spitzer Space Telescope data archive (PI: Marwick-Kemper, program P01094). 
277: Aperture photometry was performed on the IRAC and MIPS 24\,$\mu$m data, with the appropriate point 
278: source aperture corrections applied. The source is slightly extended at IRAC wavelengths, but SSC 
279: recommends that for extended sources smaller than 8-9'' in size, the point source corrections should be 
280: used. The HST WFPC2 images were extracted form the MAST archive (PI: Noll, HST proposal 9315).
281: 
282: The model comprises a flared disk illuminated by a central star.  
283: The central star was spectroscopically identified as being A0 III by \citet{Ruiz87}, so we adopt  
284: $L_\star=160 L_\odot$ and $T_\star=10^4$~K. The input spectrum is a $10^4$~K NextGen model 
285: atmosphere \citep{NextGen}. 
286: As mentioned earlier, the {\it HST} images resemble those of edge-on protoplanetary disks, 
287: therefore we adopt a flared disk geometry for the circumstellar density structure parameterized by
288: %
289: \begin{equation}
290: %\rho (r) =  \rho_0
291: \rho (r) = \rho_0 \exp\left[{-\frac{1}{2}z^2/h(r)^2}\right]\left( {r\over{R_\star}}\right)^{-\alpha} 
292: \end{equation}
293: %
294: for $r> R_{\rm min}$ where $r$ and $z$ are the cylindrical coordinates and $R_{\rm min}$ is the 
295: inner radius of the disk. The disk scaleheight $h(r) = h_0 (r/R_\star)^\beta$ 
296: where $h_0$ and $\rho_0$ are the scaleheight and density at the stellar radius, $R_\star$. The surface 
297: density therefore has the form $\Sigma (r) \sim r^{\beta-\alpha}$. Models of disks heated by a central 
298: star typically yield surface densities $\Sigma \sim r^{-1}$ and flaring parameters in the range 
299: $1.1\la \beta\la 1.3$ \citep[see review by][]{DullPPV}. 
300: 
301: For the radiative equilibrium circumstellar dust, we adopt the opacity and scattering properties of the 
302: dust model described in \citet{Wood2002}. This dust model extends to larger sizes than ISM grains and 
303: produces a shallower wavelength dependent opacity. Larger grains are expected from theories of 
304: coagulation in disks and this dust model gives a good representation of 
305: the SEDs of disks in Taurus \citep[e.g.,][]{Rice2003}. We use this dust model in the absence of mm observations 
306: of Gomez's Hamburger which can help to determine the slope of the long wavelength dust opacity and 
307: hence information on the dust size distribution.
308: 
309: In addition to a radiative equilibrium dust component we introduce the PAH/VSG 
310: opacity and emissivity as described above. The PAH/VSG component of the dust has the same 
311: radial density structure but is allowed to have a larger scale height from the radiative equilibrium 
312: grains. The scaleheights ($h_0$) of the PAH/VSG and thermal dust components are parameters which are 
313: varied in a grid of models we ran to reproduce the observations (see below). 
314: This approximates dust settling in disks \citep[see also][]{Scholz2006} where the smaller 
315: grains (PAHs/VSGs in our model) have a larger scaleheight than the larger radiative equilibrium 
316: grains. Therefore the density structures of the PAHs/VSGs and radiative equilibrium grains are 
317: given by equation~3 with only the $\rho_0$ and $h_0$ terms differing to reflect the different 
318: mass fractions and scaleheights of radiative equilibrium grains and PAHs/VSGs. 
319: The mass fraction of PAHs/VSGs is one of the free parameters in our modeling.
320: 
321: In our modeling procedure we ran a small grid of models to determine the distance 
322: to Gomez's Hamburger in addition to the disk mass, inner hole size, scaleheights of the radiative 
323: equilibrium dust and PAH/VSG components, and the PAH/VSG fraction required to reproduce the 
324: features observed in the IRS spectrum. As the disk is observed to be close to edge-on we restricted 
325: the inclination to be $i > 85^\circ$ because lower inclination disks produce a larger contrast 
326: between the upper and lower parts of the scattered light disk. 
327: 
328: Assuming the illuminating source is an A0 III star and the circumstellar disk is viewed close to edge-on, 
329: allows us to determine the distance, interstellar extinction, and disk radius by fitting the optical flux 
330: levels and image size as shown in figures~3 and 4. From 
331: this we determine $d = 280$~pc, $R_d = 1600$~AU, and $A_V = 1.8$. 
332: The disk inclination, flaring parameter ($\beta$),  scaleheight, and mass of the radiative equilibrium dust component are determined from comparing models with the $HST$ 
333: images. We varied these parameters in the range $85^\circ \le i \le 90^\circ$ and $1.1\le \beta\le 1.3$.
334: 
335: The two SED models shown in figure~3 are our best fit models with and without PAHs/VSGs. 
336: We calculated goodness of fit $\chi^2$ values for the two models using broadband fluxes only and 
337: separately using only the IRS data. Assuming 15\% errors on each data point, we find that for the 
338: broadband fluxes (thirteen data points) the model without 
339: PAH/VSG effects has $\chi^2 = 381.2$, while the model including PAHs/VSGs has $\chi^2 = 50.8$. Using 
340: only the IRS data (273 data points) gives $\chi^2 = 19350$ (model without PAHs/VSGs) and $\chi^2 = 376$ 
341: (model including PAHs/VSGs). For the model including PAH/VSG effects and considering all the data, we 
342: find $\chi^2 = 1.49$ per data point, i.e., an average deviation of 1.22$\sigma$. 
343: 
344: Many degenerecies exist in modeling disk images and SEDs, notably in the disk flaring parameters 
345: $h_0$ and $\beta$. Within the context of the smooth, axisymmetric power law disk structure model in 
346: the parameter space we explored we found values that reproduce the infrared SED and images 
347: (Figures 3 \& 4) are $i = 87\pm 1^\circ$, $\beta = 1.2$, 
348: $M_{\rm dust} = 0.003 \pm 0.0005 M_\odot$ 
349: and $h_0 = 0.0077 \pm 0.0005$. The value of $\beta$ together with our assumption that the 
350: surface density scales as $\Sigma \sim r^{-1}$  therefore gives $\alpha = 2.2$. Assuming a gas to dust ratio 
351: of 100 results in a total disk mass of $0.3M_\odot$. This disk mass is in excellent agreement with that 
352: determined from the CO observations of \citet{Bujarr2008}, who derived a disk mass of 
353: $0.9 M_\odot (d/500{\rm pc})^2$.
354: 
355: We allow the scaleheight of the 
356: PAHs/VSGs to differ from that of the larger radiative equilibrium grains and find that a PAH/VSG mass fraction 
357: of $1.5\pm 0.25$\% and scaleheight $h_0^{PAH} = 0.009\pm 0.001$ reproduces the {\it Spitzer} IRS spectrum as shown in 
358: figure~3. The larger scaleheight for the PAH/VSG component is consistent with many other disk models that 
359: predict the coagulation and settling of large grains to the midplane. 
360: We also found that the IRS spectrum was best reproduced with an inner disk hole size of 
361: $R_{\rm min} = 500 \pm 25 R_\star$ which is larger than the dust destruction radius of radiative 
362: equilibrium grains of $40 R_\star$, assuming a typical dust sublimation temperature of 1600~K.
363: 
364: Figures~3 and 4 show that compared to a model without PAHs/VSGs our circumstellar disk model together 
365: with our new technique for incorporating PAH/VSG emission processes provides a very good match to 
366: the observed spectral energy distribution 
367: and high resolution optical images of Gomez's Hamburger. In particular the relative strengths of the 
368: PAH features in the {\it Spitzer} IRS spectrum are well reproduced, whereas they are absent in the model 
369: without PAHs/VSGs. We do not fit the IRAS 60$\mu$m data 
370: point, but this is not a great concern due to potential contamination problems with the IRAS data. In fact the 
371: IRAS 100$\mu$m data is not useable for this source.
372: 
373: The PAH/VSG emissivities we use in 
374: our code were developed by \citet{DL2007} specifically to model post-{\it Spitzer} data of the diffuse 
375: interstellar medium in the Milky Way and other galaxies. 
376: Further applications of our code to model mid-infrared spectra 
377: of other disks will enable us to test whether the \citet{DL2007} model can fit PAH features in a broad range of 
378: environments. Recent studies of PAH features in Herbig Ae stars \citep{Boersma2008} 
379: suggest there are multiple components that differ from the ISM-like model we have adopted.
380: 
381: In Figure~3 we also show our disk model SED {\it without} PAHs/VSGs which are seen to dominate in the 
382: wavelength range $5\mu{\rm m}\la\lambda\la 20\mu{\rm m}$ and also with a strong feature around 3$\mu$m. 
383: The excess emission from disks in this mid-infrared region is very sensitive to disk heating 
384: by viscous accretion. Therefore care must be taken when analyzing mid-infrared observations of disks 
385: because radiation transfer models 
386: that do not include PAH/VSG emission could lead to overestimating the accretion rate in order to ``fill in'' 
387: the mid-infrared excess. High resolution mid-infrared spectra that can identify PAH features will greatly 
388: assist in the analysis and interpretation of infrared spectra and accretion rates of protoplanetary disks.
389: 
390: In summary we find that if the central object is an A0 III star, then Gomez's Hamburger lies at a distance of 
391: 280~pc, much closer than the 2.9~kpc estimated by \citet{Ruiz87}. The nebulosity may be modeled with a 
392: highly inclined disk of mass $0.3M_\odot$ and radius $1600$~AU. 
393: We reproduce the PAH features present in the {\it Spitzer} IRS spectrum 
394: using the PAH/VSG model and associated opacities and emissivities computed by \citet{DL2007} for 
395: their model of the infrared spectrum of the diffuse ISM in our Galaxy.
396: 
397: Our modeling of the circumstellar structure suggests Gomez's Hamburger is an edge-on protoplanetary disk 
398: around an A0 star. However, if the spectral type of the source is inaccurate, we cannot rule out the possibility 
399: that Gomez's Hamburger is much closer and the disk has a smaller radius (for a lower luminosity source). Conversely, a higher luminosity central source would yield a larger disk radius and distance. 
400: As the optical spectrum is entirely scattered light, the combination of high resolution optical spectroscopy 
401: and radiation transfer modeling of the scattered starlight from the disk is required for more accurate spectral 
402: typing. The original suggestion from \citet{Ruiz87} is that Gomez's Hamburger is a protoplanetary nebula. 
403: While our model does not explicitly address this scenario, our derived circumstellar disk structure is consistent 
404: with dynamical models of planetary nebulae requiring a pre-existing circumstellar disk or torus which can 
405: act to focus and shape the stellar material ejected during the planetary nebula phase \citep[e.g.,][]{BF2002}. 
406: 
407: \section{Summary}
408: 
409: We have presented a straightforward method to include the absorption and re-emission from PAH molecules and 
410: VSGs into Monte Carlo radiative equilibrium codes. Our technique utilizes pre-computed intensity-dependent 
411: emissivities from which we sample the emission spectrum of Monte Carlo energy packets that have been 
412: absorbed by PAHs/VSGs. Our codes are readily adaptable to include a range of sizes for the 
413: PAHs/VSGs, but in this first paper we adopt a grain size cutoff of 200~{\AA}, above which we 
414: assume the dust grains attain radiative equilibrium temperatures. For grain sizes smaller than 200~{\AA}, 
415: absorbed energy packets are reprocessed using the pre-computed emissivity files as described in \S2. 
416: Incorporating the PAH/VSG emission in this manner allows us to compute model spectra and images 
417: very quickly without repeating lengthy calculations of statistical equilibrium and transient heating and 
418: cooling of the PAH/VSG component --- these calculations are in effect already incorporated in the 
419: pre-computed PAH/VSG opacities and emissivities presented in figures~1 and 2.
420: 
421: In a future paper we will extend our technique for sampling emissivities to be dependent not only on the 
422: intensity of the exciting radiation field, but also on the spectrum of the absorbed energy in each grid cell of 
423: our Monte Carlo simulation. Similar to the mean intensity calculation, determining the spectrum of the 
424: absorbed energy throughout our grid is straightforward to implement in Monte Carlo radiation transfer 
425: simulations. This more accurate treatment of the PAH/VSG reprocessing will be possible using planned 
426: calculations of PAH/VSG emissivity files for a range of exciting spectra comprising model stellar atmosphere 
427: spectra and black-bodies for a range of temperatures and reddening. 
428: 
429: As a first application of our new code we modeled the object known as Gomez's Hamburger. We find that 
430: the broad band SED, high resolution images, and {Spitzer} IRS spectrum can be reproduced with an edge-on 
431: circumstellar disk with a component of PAHs/VSGs comprising 1.5\% of the dust mass.  
432: Our Monte Carlo radiation transfer using the adopted dust opacities ensures that very few high 
433: energy photons penetrate into regions of large optical depth. Therefore there is little or no PAH/VSG 
434: emission from such regions and the PAH/VSG emission is largely confined to a surface layer exposed 
435: to the direct stellar radiation. This is what is observed in the limb brightened bubbles in the GLIMPSE 
436: survey \citep{Churchwell2006}. In future papers we will use our code to model the bubbles observed in the 
437: GLIMPSE survey to address the question of PAH/VSG survival around hot stars and in H~{\sc ii} regions 
438: in addition to modeling the dust content within the H~{\sc ii} regions and the effects this may have on 
439: determining the ionizing luminosity of the central sources. We will also investigate emission 
440: from PAHs/VSGs in protostars, thus extending the radiative equilibrium dust modeling presented by 
441: \citet{Whitney2003b}.
442: 
443: 
444: \begin{figure}
445:   \plotone{f1.eps}  \caption{Solid line shows the wavelength dependent opacity for the 
446:   PAH/VSG (dust plus gas) model of 
447:   \citet{DL2007} for grains with sizes $a < 200$~{\AA}. Dot-dashed line is for the radiative equilibrium 
448:   dust model of \citet{Wood2002}. For comparison the dashed line shows the opacity for an ISM dust 
449:   plus gas mixture.}
450: \end{figure}
451: 
452: \begin{figure}
453:   \plotone{f2.eps}
454:   \caption{Emissivity for PAHs/VSGs with size $a < 200${\AA} for different values of 
455:   $U$, the energy density of the exciting radiation field relative to the average interstellar value. The 
456:   emissivities shown are in the range $1 \le U\le 10^7$. The solid line shows the emissivity 
457:   for $U=1$ and the long dashed line is for $U=10^7$.}
458: \end{figure}
459: 
460: \begin{figure}
461:   \plotone{f3.eps}
462:   \caption{SED data (dots and red curve) and our best disk model (upper solid line) for 
463:   Gomez's Hamburger. The lower solid line at mid-IR wavelengths 
464:   shows the same disk model without the PAH/VSG components.}
465: \end{figure}
466: 
467: 
468: \begin{figure}
469:   \plottwo{f4a.eps}{f4b.eps}
470:   \caption{Three-color images at $B$ (blue), $V$ (green), and $R$ (red) bands showing HST data (left) and 
471:   model (right) images for Gomez's Hamburger}
472: \end{figure}
473: 
474: 
475: \acknowledgements
476: 
477: TR is supported by a Scottish Universities Physics Alliance Studentship, 
478: BAW is supported in part by the NASA Theory Program (NNG05GH35G), 
479: BTD was supported in part by NSF grant AST-0406883. 
480: 
481: %\bibliography{reference}
482: 
483: \bibliography{ms.bbl}
484: 
485: \end{document}
486: