[ Other OHP preprints ]


C. Moutou 1, L. Verstraete 2, K. Sellgren 3, A. Léger 2

1Observatoire de Haute Provence, St Michel, France
2Institut d'Astrophysique Spatiale, Orsay, France
3Ohio State University, Columbus, USA


The ISO-SWS spectra of two bright reflection nebulae, NGC 7023 and NGC 2023, are presented. We discuss the emission of molecular hydrogen from these photodissociated interfaces. Details of the aromatic infrared band profiles as well as the continuum emission are also analysed. The ISO-LWS spectrum of NGC 7023 is also presented, at two positions in the nebula. The dust temperature at the brightest far-infrared position of NGC 7023 is estimated to be 45 K.

Key words: ISO; reflection nebula; dust emission bands; molecular hydrogen emission.


ISO is providing us with high-resolution spectra of the Aromatic Infrared Bands between 3 and 13 Ám (hereafter AIBs) in a wide variety of galactic environments. At the spectral resolution of ISOCAM-CVF (= 40), the AIBs have been shown to be similar in interstellar regions with effective radiation fields ranging from 1 to 104 times the interstellar radiation field (Boulanger et al. 1998 and references therein). Using SWS data at higher spectral resolution, we show here that there are differences between the AIB spectra of two reflection nebulae, NGC 2023 and NGC 7023. The physical conditions of the gas, associated with the AIBs, are discussed through the analysis of the molecular hydrogen emission.



The SWS spectrum of NGC 7023 was obtained with the AOT01 observing mode in 1996 and is fully described in Moutou et al. (1998) and Sellgren et al. (1999, in preparation). Its mean spectral resolution is = 1000 (speed 4). A restricted part of this spectrum has been observed at higher spectral resolution (Moutou et al. 1999). The SWS spectrum of NGC 2023 was observed at a spectral resolution of = 500 (AOT01, speed 3) in 1998. Data reduction has been done within the SWS-Interactive Analysis environment. Our observations have higher spectral resolution than previously published mid-IR spectra of NGC 7023 and NGC 2023 from the Kuiper Airborne Obsevatory (Sellgren et al. 1985) and of NGC 7023 with ISOPHOT-S and ISOCAM (Laureijs et al. 1996; Cesarsky et al. 1996).

NGC 7023 and NGC 2023 are two bright reflection nebulae irradiated by hot stars. The SWS aperture for NGC 7023 was centered 27'' W 34'' N (position 1) of HD 200775 (Teff=17,000 K). For NGC 2023, the SWS aperture was 60'' S of HD 37903 (Teff=22,000 K).

Figure 1: SWS01 spectra of the reflection nebulae NGC 7023 and NGC 2023. An offset of W m-2 sr-1 is added to the spectrum of NGC 7023 (top) for clarity. The 3.5 to 5.3Ám region suffers of bad dark current measurements and is currently under processing.
\epsfig {file=global.ps,width=9.0cm,height=7.cm}

Figure 1 shows the mid-IR spectra of the two nebulae, with an offset added to NGC 7023 for clarity (see caption of Fig.1). The relative contribution of continuous emission and AIBs to the total emitted energy is comparable in both objects. More than 60% of the 3 - 20Ám energy is emitted in the AIBs. Reflection nebulae have a high feature-to-continuum ratio: this makes the study of faint details in the AIB profiles much easier than in more strongly irradiated sources where the AIBs are drowned by a strong mid-IR continuum (e.g., planetary nebula, H II region interfaces).

Molecular hydrogen emission

We detect many pure H2 rotational lines in both nebulae as well as some ro-vibrational (or fluorescent) lines in NGC 7023. This emission comes from photodissociated gas lumped into filaments (Lemaire et al. 1996, Field et al. 1998). We derive excitation diagrams (Fig.2) from the line fluxes. In the case of NGC 7023, our upper level column densities (from the S(0) line at 28.22Ám through to the S(4) line at 8.02Ám) are well-fitted[*] by a single rotational temperature of 411 K if one adopts an ortho-to-para ratio (hereafter Rop) of 1. These results, although derived from observations with a much larger beamsize, are consistent with the findings of Lemaire et al. (1996). Indeed, the warm gas emission we detect is probably dominated by the filaments shown in the Lemaire et al. map. For NGC 2023, using the S(1) at 17.03Ám through to the S(3) at 9.66Ám lines, we find a rotational temperature of 333 K also with Rop=1. While determining the rotational temperature, we have left aside higher levels (Tu > 5000 K) because their populations are strongly affected by the UV pumping (and subsequent radiative decay) of the H2 molecule. Densities of about 105 cm-3 have been inferred for both nebulae (Lemaire et al. 1996; Field et al. 1998; Martini et al. 1997). At these high densities, the low rotational transitions (from upper levels with Tu < 5000 K) we are discussing here are collisionally thermalized and the rotational temperature is equal to the gas temperature, Tgas.

Figure 2: The excitation diagrams of H2 in a) NGC 7023 and b) NGC 2023: Nu is the column density of the transition upper level, gu is the degeneracy of the upper level, gI the nuclear spin degeneracy and Tu is the upper level energy in Kelvin. The crosses with error bars indicate the values for Rop=1 (see text), while the asterisks show the case of Rop=3. The arrows show 3 upper limits. The rotational temperatures have been estimated with Rop=1.
\epsfig {file=h2_7023....
 \end{center} \end{figure}

Our low ortho-to-para ratio confirms earlier work (Chrysostomou et al. 1993 and references therein) on photodissociated interfaces. The H2 molecule forms on the surface of dust grains and is expected to be rejected in the gas with a high vibration-rotation energy content and a Rop value close to 3. After formation, Rop can be changed by gas-phase spin exchange reactions of H2 with atomic hydrogen and protons for Tgas>300 K or by H2-grain collisions for Tgas<300 K. A value of Rop=1 corresponds to an equilibrium temperature Teq 80 K (Burton et al. 1992). The fact that the dust temperature ( 40 K, see §3) and the gas temperature (300 - 400 K) are both very different from Teq points at the importance of out-of-equilibrium effects; this point is also highlighted by the high Rop values (between 2 and 3) predicted by recent stationary models (Draine & Bertoldi 1996). We note that H2 rotational populations in equilibrium at 40 K and 300 - 400 K correspond to Rop=0.15 and 3 respectively.

Two scenarios have been proposed by Chrysostomou et al. to explain the observed low Rop values. First, if the newly formed molecular hydrogen resides long enough on the surface of the dust grain, Rop will be between 3 and the equilibrium value of 0.15 set by the grain temperature Tdust. To get Rop=1 , the residence time of H2 on the dust grain should be approximately half the time required for the H2 rotational populations to reach equilibrium at Tdust after formation (see Fig.9 in Chrysostomou et al.).

Alternatively, in cold gas, Rop is fixed at a low value by gas-grain interactions (0.15 for Tdust=40 K). As the photodissociation front propagates into the molecular cloud at velocities of the order of 1 km/s, cold gas is advected through the hot (Tgas of a few 100 K) interface. In the hot gas, spin exchange reactions can drive the low Rop to the observed value.

The residence time of H2 on the grain goes as exp (450K/Tdust) (Tielens & Allamandola 1987) while the rate for spin exchange reactions goes as exp(-3200K/Tgas). From photodissociation region models, variations of Tgas are expected to be much larger than those of Tdust: Rop should thus vary much more rapidly in the second scenario (cold gas advection) than in the former (modified H2 formation). High spatial resolution observations yielding Rop profiles across photodissociated interfaces should help discriminate the 2 scenarios.

Band Profiles

We compare here qualitatively the AIB profiles of both nebulae. More quantitative results will be presented in a forthcoming paper (Sellgren et al. 1999). We remark that the AIB spectrum of NGC 2023 multiplied by 2.2 superimposes nicely to that of NGC 7023. As the AIB flux is proportional to the radiation field intensity (Sellgren et al. 1985), this suggests that the radiation field in NGC 7023 is twice as strong as that of NGC 2023 (at the positions given in §1). The two AIB spectra look very similar (width, position of the AIBs) as expected in view of the similar physical conditions (density, radiation field) derived for NGC 7023 (Lemaire et al. 1996) and NGC 2023 (Steinman-Cameron et al. 1997; Field et al. 1998). There are, however, significant spectral differences in the AIB profiles which we detail below. These differences reflect changes in the physico-chemical state of the AIB carriers.

Figure 3: The profiles of the AIBs are shown for NGC 7023 (continuous line) and NGC 2023 (dot-dashed line). The intensity of NGC 2023 is multiplied by ~ 2.2 for comparison to NGC 7023.
\epsfig {file=bande.ps,width=8.0cm,height=9.0cm}
 \end{center} \end{figure}

Figure 3a shows that the 3.3 and the 3.4Ám AIBs have identical profiles in both nebulae. The 3.4/3.3Ám band ratio remains thus the same while the radiation field intensity is multiplied by a factor 2.

Figure 3b shows that the 6.2 Ám AIB is asymmetrical and vary slightly between the two nebulae. Both profiles show a pronounced wing towards long wavelengths, which is interpreted as the consequence of anharmonic couplings during the cooling of the molecule (Barker et al. 1987; Joblin et al. 1995). In NGC 2023 the local continuum is more important, with respect to the 6.2Ám band, and has a steeper rise than in NGC 7023. This suggests that the underlying continuum has a different origin from the AIBs.

Figure 3c demonstrates that the 7 - 9Ám range is the most complex. The main AIBs are at 7.6, 7.8, and 8.6Ám. The intensity of the 7.6Ám-component is stronger in NGC 7023 than in NGC 2023. Furthermore, the blue shoulder at 7.45Ám observed in NGC 7023 (see also Moutou et al. 1998) is not seen in NGC 2023. Roelfsema et al. and Verstraete et al. (1996) have shown that the profile of the ``7.7''Ám AIB and the distribution of energy between the 7.6, 7.8 and 8.6Ám AIBs varies with the radiation field. In the PAH model, the ``7.7''Ám AIB falls in the spectral range where the effects of ionisation are the most dramatic (Pauzat et al. 1997; Langhoff 1996). Such a prominent 7.6Ám feature as well as the 7.45Ám blue wing are unusual and have only been seen towards compact H II regions (Roelfsema et al. 1996) and towards the post-AGB star HR 4049 (Molster et al. 1996). These latter authors attributed the 7.45 and 7.6Ám new features to small ionized PAHs. Ionized PAHs have a strong 7.7/11.3Ám band ratio (Langhoff 1996). Since the AIB spectrum in NGC 2023 is scalable to that of NGC 7023 (see Fig.3), the 7.7/11.3Ám band ratio is the same for both nebula implying that the fraction of ionized PAHs is the same in both cases. >From the absence of the 7.45Ám-band and the weaker 7.6Ám feature in NGC 2023, we conclude that the carriers of these bands are produced by other processes than photoionization (photochemical evolution, fragmentation...)[*].

Figure 3d shows that the 11.3Ám AIB has a profile similar to the 6.2Ám band. A red asymmetry is observed, which again suggests strong anharmonic effects. We note that the red wing of the 11.3Ám band is more pronounced in NGC 2023. The 11.3Ám band also shows some weak sub-features within its profile. In particular, the 11.0Ám band (previously detected by Witteborn et al. 1989 and Roche et al. 1991) is clearly seen in both nebulae. The 11.0Ám band also appears in many objects with a similar ratio to the 11.3Ám band [*] (Molster et al. 1996; Roelfsema et al., Verstraete et al. 1996) while the 7.7/11.3Ám band ratio presents large variations (compare for instance the AIB spectrum of the M17-SW interface and that of NGC 7023). This again rules out photoionization as the cause of the 11.0Ám-band in moderatly excited sources.


Finally, we show the complete LWS spectrum (AOT01 observing mode) of NGC 7023, at two nebular positions labeled 1 and 2. Unfortunately, we did not obtain any LWS spectrum of NGC 2023.

The LWS position 1 in NGC 7023 is the same as for SWS (see §1) and is close to the far-infrared peak of Whitcomb et al. (1981). Position 2 is located 100''N of the star. The LWS beam is 80'' in diameter, so the fields slightly overlap. Position 1 has been observed in the guaranteed time program of J.P. Baluteau and Position 2 was part of our open-time program on reflection nebulae. Both spectra are shown in Figure 4. The data reduction was done with the LWS-Interactive Analysis and ISAP softwares.

Figure 4: Complete LWS spectra of the reflection nebula NGC 7023, at two positions (see text). A modified blackbody fit is applied, which gives an effective temperature of 45K in Position 1 and 30K in Position 2. The narrow emission lines are due to atoms and ions of the gas (see text).
\epsfig {file=lws_7023.ps,width=9.0cm,height=7.0cm}

The LWS continuum emission is due to big dust grains in thermal equilibrium with the radiation field. We estimate the dust temperature by fitting a modified blackbody emission curve to the spectra with a dust emissivity law proportional to the frequency . The fits are shown on Figure 4. For Position 1, we also used the SWS spectrum of NGC 7023 (not shown here) to further constrain the fit. The effective temperature of dust grains at Position 1 is Tdust = 45 ▒ 2K, while at Position 2, the temperature drops to Tdust = 30 ▒ 2K. These temperatures are compatible with the temperature map obtained by Whitcomb et al. (1981). Poorer fits to the LWS spectra were obtained with a dust emissivity proportional to 2 (in this latter case, the above temperatures would drop by approximately 6 K). For a dust emissivity ~ , the radiation field goes as T5 and so the change in Tdust implies that the radiation field is stronger at Position 1 by a factor of 8 than at Position 2.

Atomic emission lines accounting for the cooling of the gas are visible on both spectra, namely CII (158Ám), OI (63Ám) and NI (145Ám). They will be discussed elsewhere.


We are very grateful to J.P. Baluteau for providing us the LWS spectra of NGC 7023. We acknowledge NATO Collaborative Research Grant 951347.


Barker J.R., Allamandola L.J., Tielens AGGM, 1987, ApJ 315, L61

Boulanger, F. et al., 1998, A&A 339, 194

Burton, M.G., Hollenbach, D.J., Tielens, A.G.G.M., 1992: ApJ 399, 563

Cesarsky D. et al., 1996, A&A 315, L305

Chrysostomou, A., Brand, P.W.J.L, Burton, M.G., Moorhouse, A., 1993: MNRAS 265, 329

Draine,B.T., Bertoldi,F., 1996: ApJ 468, 269

Field D., Lemaire J. L., Pineau des Forêts G., Gerin M., Leach S., Rostas F. & Rouan D., 1998, A&A 333, 280

Joblin, C., Salama, F., Allamandola, L., 1995, J. Chem Phys. 102 (24), 9743

Langhoff S.R, 1996, J. Phys. Chem. 100, 8, 2819

Laureijs R. et al., 1996, A&A 315, L313

Lemaire J.L., Field D., Gerin M., Leach S., Pineau des Forêts G., Rostas F. & Rouan D., 1996, A&A 308, 895

Martini P., Sellgren K. & Hora J.L., 1997, ApJ 484, 296

Molster F.J. et al. 1996, A&A 315, L373

Moutou C. et al., 1998, in Star formation with the Infrared Space Observatory, ASP Conf. Ser. 132, eds. Joao L. Yun and René Liseau, p. 47

Moutou C., Sellgren K., Verstraete L., Léger A., 1999, submitted to A&A

Pauzat F., Talbi, D. & Ellinger, Y., 1997, A&A 319, 318

Roche P.F., Aitken D.K. & Smith C.H., 1991, MNRAS 252, 282

Roefselma et al., 1996, A&A 315, L289

Sellgren K., Allamandola, L.J., Bregman, J.D., Werner, M.W. & Wooden, D.H., 1985, ApJ 299, 416

Steinman-Cameron,T.Y., Haas,M.R., Tielens,A.G.G.M., 1997: ApJ 478, 261

Tielens, AGGM, & Allamandola, LJ 1987, in Interstellar Processes, ed. D. J. Hollenbach & H. A. Thronson (Dordrecht: Reidel), p. 397

Verstraete L., Puget J. L., Falgarone E., et al., 1996, A&A 315, L337

Whitcomb S.E., Gatley I., Hildebrand R.H., Keene J., Sellgren K. & Werner M.W., 1981, ApJ 246, 416

Witteborn F.C., Sandford S.A., Bregman J.D., Allamandola L.J., Cohen M., Wooden D.H. & Graps A.L., 1989, ApJ 341, 270

About this document ...


This document was generated using the LaTeX2HTML translator Version 97.1 (release) (July 13th, 1997)

Copyright © 1993, 1994, 1995, 1996, 1997, Nikos Drakos, Computer Based Learning Unit, University of Leeds.

The command line arguments were:
latex2html -split +0 -local_icons pp133.tex.

The translation was initiated by on 1/15/1999


ISO is an ESA project with instruments funded by ESA Member States (especially the PI countries: France, Germany, the Netherlands and the United Kingdom) and with the participation of ISAS and NASA.

We did not include the S(5) line at 6.91Ám in the fit because its profile is much broader than the other pure H2 rotational lines suggesting contamination by another line.

We also note that the 7.7/11.3 Ám band ratio in HR 4049 is 4 times that of NGC 7023: such a variation points at very different fractions of ionized PAHs. The fact that the 7.45 and 7.6 Ám features are seen in both spectra reinforces our conclusion that ionization cannot be the major way of producing these bands.

This is not true for strongly irradiated regions where the AIB spectrum is deeply modified (strong 8.6 and 11.0Ám-bands).