ebook img

SEQUENCE STRUCTURE EMISSION IN THE RED RECTANGLE BANDS PDF

13 Pages·2008·0.41 MB·English
by  
Save to my drive
Quick download
Download
Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.

Preview SEQUENCE STRUCTURE EMISSION IN THE RED RECTANGLE BANDS

Draftversion February5,2008 PreprinttypesetusingLATEXstyleemulateapjv.11/26/04 SEQUENCE STRUCTURE EMISSION IN THE RED RECTANGLE BANDS R.G. Sharp Anglo-AustralianObservatory,POBox296,Epping,NSW,1710, Australia and N.J. Reilly, S.H. Kable, T.W. Schmidt SchoolofChemistry,BuildingF11,UniversityofSydney,NSW,2006,Australia Draft versionFebruary 5, 2008 ABSTRACT 5 We reporthighresolution(R∼37,000)integralfieldspectroscopyofthe centralregion(r<14arcsec) 0 of the Red Rectangle nebula surrounding HD44179. The observations focus on the λ5800˚A emission 0 feature, the bluest of the yellow/red emission bands in the Red Rectangle. We propose that the 2 emissionfeature,widelybelievedtobeamolecularemissionband,isnotamolecularrotationcontour, t but a vibrational contour caused by overlapping sequence bands from a molecule with an extended c chromophore. We model the feature as arising in a Polycyclic Aromatic Hydrocarbon (PAH) with O 45-100 carbon atoms. 8 Subjectheadings: planetarynebulae: individual(RedRectangle),ISM:molecules,ISM:linesandbands 2 1 1. INTRODUCTION vised distance estimate of 710pc2) the nebula represents v The Diffuse Interstellar Bands (DIBs), first identified an excellent candidate for the study of the injection of 0 material into the interstellar medium by a dying star. asinterstellarabsorptionfeaturesbyMerrillandcowork- 9 ers (Merrill 1934, 1936; Merrill & Wilson 1938), repre- 7 2. THE RED RECTANGLE sent one of the longest outstanding mysteries in astro- 0 nomical spectroscopy. The bi-conical Red Rectangle (RR) nebula 1 The DIBs are widely believed to represent the origin (Cohen etal. 1975, 2004) surrounding the evolved, 5 0 bands of molecular transitions. Ultra-high resolution mass-losing star HD44179 is one of the strongest / spectroscopy of a selection of the stronger DIBs con- sources in the sky when observed in the mid-IR h firmstheirmolecularorigin(Sarre etal.1995a;Kerr etal. (IRAS06176-1036), where a set of emission fea- p 1998), while the lack of strong correlations between tures commonly attributed to polycyclic aromatic - o known bands precludes the identification of any tran- hydrocarbons (PAHs, Allamandola etal. (1985)) r sitions to higher vibronic states. dominate the spectrum. PAH molecules provide an st IdentificationofthecarriersoftheDIBswillultimately elegant solution to a number of astronomical issues a require accurate measurement of the gas phase spec- ( UV exctinction curve - (Donn, Hodge & Mentall : trum of a range of extraterrestrial molecules to allow 1968); DIBs carriers - (van der Zwet & Allamandola v i direct comparison with astronomical observations (for 1985; L´eger & Dhendecourt 1985); UIB X a review of current work see Schmidt & Sharp (2005); carriers(Geballe,Tielens, Allamandola etal. 1989)). r Fulara & Krelowski (2000); Herbig (1995)). However, With the coming of age of the new generation of a the parameter space occupied by the range of candidate mid IR instrumentation, PAH emission may prove to carriers is prohibitively large for full investigation with be a valuable tracer of extragalactic star formation the often complex techniques required to record the gas (Peeters etal. 2004). However, at this time, not a single phase spectra. PAHhasbeen unambiguouslyidentifiedinthe ISM(but In order to restrict the parameter space one can turn see Allamandola etal. (1999) for details of composite tostudiesofalternateenvironmentstothediffuse clouds spectraoftheunidentifiedInfra-redBands(UIBs)inthe within which most DIBs are observed. These alternate mid-IR). At the heart of the problem is the simple fact environments may represent the Rosetta stone for DIB that while the λλ3.3, 6.2, 7.7, 8.6 & 11.3µm emission spectroscopy1, offering the possibility of presenting ad- features (Peeters etal. (2004)) most likely represent ditional features of the molecular species, providing the excellent tracers of PAH emission, their origins in the much needed additional constraints to the properties of C-H and C-C stretches, abundant in PAH molecules, the carrier molecules. meantheypresentlittleornodiagnosticpower. Ifoneis In this paper we present a high spatial, and spectral, to use PAH emission as a key diagnostic on interstellar resolution study of one of the emission features seen in processes, within and outside of our Galaxy, it would the Red Rectangle nebula. At a distance of only 330pc seem prudent to understand the phenomenon. (but see Men’shchikov etal. (2002) for details of a re- Spatiallyextendedopticalemissionbandsnearλ5800˚A were first reported by Schmidt etal. (1980) but cur- Electronicaddress: [email protected] rentlyremainunidentified. Anumberofdetaileddescrip- Electronicaddress: [email protected] 1WethankA.Wittforsuggestingthiselegantmetaphor,private tions of the form of the Red Rectangle Bands (RRB) communication 2 Hipparchus satellite measurements give a value of 2.62±2.37mas/yrcorrespondingtotherange∼200-4000pc. 2 Sharp, Reilly, Kable and Schmidt are available in the literature. We refer the reader 0.3arcsec,yieldingfieldsofviewof11.44×7.28arcsecand to Van Winckel etal. (2002), rather than repeat a de- 6.6×4.2arcsec respectively. We used the 0.52arcsec lens tailed description. Following Van Winckel etal. (2002) scale for our observations due to the low surface bright- we adopt the standard molecular terminology Red De- ness of the nebula in the regions of primary interest. graded (RD) to refer to RRBs which show a sharp flamesisoperatedinaseriesofstandardusermodes. edge in the blue but an extended tail to the red. The We use the HR11 mode for our observations, with a λ5800˚AbandisatypicalRDRRB.Severalauthorshave central wavelength of 5728˚A and covering a wavelength noted the close correspondence of many of the RRBs rangeof5597-5840˚Aatadispersionof∼0.06˚A/pixeland with prominent DIBs (Scarrott etal. 1992; Sarre etal. a resolution element of 2.6pixels (0.156˚A, R∼37,000). 1995b; Van Winckel etal. 2002). While initial observa- The detector was a 2k×4k EEV CCD operated in the tions found strong support for convergence of the RRBs standard flames readout mode. to DIBs, the early works were hampered by low resolu- TableA1andFigureA1detailtheobservationsunder- tion and low signal-to-noise observations. Later works taken. Observationswereperformedduringtwonightsin (Glinski & Anderson 2002) indicate that the blue-ward classicallyscheduledtime,29th and30th December2004. shifts of the RRBs, with increasing nebular radius, do While classical mode was not ideal for the observations, not appear to asymptotically approach the DIB wave- due to the large range in airmass for the target during lengths. We concur that the λ5800˚A RRB, in emission thenight(zenithdistance60◦-10◦-60◦),classicalmodeis does not blue shift to the wavelength of the prominent a requirement for guaranteedtime observations. λ5797˚ADIB,atleastouttoanebularradiusinexcessof argus attached flat frames are observed, after stan- 14.0arcsec. Glinski & Nuth (1997) initially proposed C3 dard star spectra, twice per night, in order to provide asthecarrieroftheλ5800˚ARedRectanglebands. How- relativefibretransmissioninformationandfibreposition ever, more recent observations suggest this is not the traces. case (Glinski & Anderson 2002). Molecular band struc- Wavelength calibration is derived from standard vlt tures are reported towardsthe centre of the nebula with flames daytime calibrations. The giraffe spectro- Miyata etal. (2004) presenting molecular PAH emission graph offers the ability to observe a number of simul- in the inner regions of the Red Rectangle in the N band taneous calibration arcs from a Th-Ar lamp during the at ∼10µm. science observations. Unfortunately, a strong line is lo- The observations of Glinski & Anderson (2002), from cated close to the primary wavelength region of interest the Densepak IFU on the WIYN telescope, show ra- leadingtosignificantcontaminationofthelowlightlevel dial evolution of the molecular emission features. How- nebulaspectrumduringa600secexposuresandsosimul- ever, the low surface brightness3 of the Red Rect- taneous calibration was not used. angle in the outer regions of the nebula probed by After discussion with ESO support staff we used the Glinski & Anderson (2002) (10-14arcsec) has hampered argus fast acquisition method whereby acquisition re- previous attempts to identify the molecular spectra. lies on accurate coordinates for a guide star and will be Recent years have seen considerable progress in lab- accuratetotheorderof1-2arcsec. Amoreaccurate,and oratory determination of the spectra of complex as- repeatable acquisition would incur a considerable over- trophysicallyrelevantmolecules(Thaddeus & McCarthy head for repositioning of guide fibre bundles during the 2001;Linnartz etal.2004;Schmidt etal.2003;Ding etal. night (with each change in PA). The loss of positional 2003). There has also been progress in the astronomical accuracy has not degraded the observations presented. detection of molecular species in the absorption spec- Five base positions were observed within the neb- traofdiffuseclouds(Maier etal.2001;A´d´amkovics etal. ula. One centred on HD44179, and four aligned along 2003)andinmillimetre-waveemission. However,thelow the arms of the nebula, where limb brightening of the surface brightness and complex spatial structure of the bi-conical outflow gives the nebula its highest surface Red Rectangle have prevented the identification of the brightnessfor a givenradialdistance, andextending out carrier of its strong molecular emission. to ∼14arcsec. Figure A1 illustrates the IFU locations within the nebula. A three point telescope dither was 3. OBSERVATIONS AND DATA REDUCTION performed at each location to allow removal of IFU lens WeusedtheflamesinstrumentatoneoftheNasmyth artifacts and detector defects, of which there are few in fociofUnitTelescope2(UT2-Kueyen)oftheEuropean the argus/giraffe system. SouthernObservatory’sVeryLargeTelescope(ESOvlt) argus sky fibres were placed, as required by the ar- located upon Cerro Paranal in northern Chile. The ob- gus fast acquisition strategy, in a ring centred on the servations were undertaken in visitor mode during 29th IFU. The ring radius was set to 3arcmins after visual and 30th December 2004 as part of the Australian guar- inspection of the DSS image of the region showed that anteed time awarded for the construction of the OzPoz this would avoid the majority of brighter stars in the Fibre positioner by the Anglo-Australian Observatory field regardless of position angle. This avoids the need (Gillingham etal. 2003). The flames instrument com- torepositionthefibresbetweenobservationswhichwould prises a robotic fibre positioner which provides a range introduce a significant overhead. of fibre optic feeds to the bench mounted giraffe spec- SimultaneousobservationswiththeUVESfibresystem trograph. We used the argus Integral Field Unit (IFU) werenotpossibleduetolimitsontheproximityofUVES to record spectra of a 2D region of the nebulosity. The fibre placements to the argus IFU. argus IFU, a rectangular array of 22×14 micro lenses, 3.1. Data reduction can provide two spatial sampling scales, 0.52arcsec and During the observing period, pipeline data processing 3 HST PCF622Wm<17.6mag(Vega) arcsec2 was not available for the HR11 argus setting used. We Sequence Structure emission in The Red Rectangle 3 thereforeprocessedthedatausingasuiteofcustomwrit- tation to the IFU internal x/y coordinate system, re- ten idl routines and elements of iraf. In most cases, trieved from the flames header, to account for the five custom idl implementations of common iraf tasks are different position angles used during the observations. used to allow accurate propagationof error information. A number of techniques were subsequently investigated A 2D bias subtraction was not deemed necessary on for aligning individual data cubes. Cross-correlation of inspection of the data and hence the data were overscan pseudo-continuum images created from the data cubes correctedonly. Onpreliminarydatareductionitwasde- proveddifficult, due tosmallimageoverlaps. Ultimately cided a known dark current feature seen on the upper alignment was performed using an iterative χ2 proce- left corner of the giraffe CCD represented a signifi- dure, and the idl implementation of the Amoeba algo- cant contamination to the low light levels recorded in rithm (Press etal. 1993), as applied to a subsection of the outer reaches of the nebula and hence subtraction each dispersed data cube. of a dark frame was required. A master dark was con- The final analysis of the full data set is ongoing. Data structed from 6×1200sec dark exposures. This master from only a single arm of the nebula (pointing 2 in Fig- dark was scaled for the 600sec and 1800sec exposures as ure A1) is used to illustrate our model for the λ5800˚A no alternate dark frames were available. emission. Preliminaryinvestigationsshowthattheemis- Flatfields arecombinedusing varianceweightingwith sionintheremainingthreenebulaarmsarequalitatively outlier rejection. Where three or more science observa- similar. tions are available, at a constant dither position, these frames are also combined in a similar manner. 4. INTERPRETING THE λ5800˚A BAND. A fibre trace is generated using the iraf apall task, A number of origins have been proposed for the appliedtotheflatfieldframe. Thedatabasetraceoutput λ5800˚A emission. Most interpretations propose the fromthetask(aseriesof5thorderLegendrepolynomials) band profile represents a molecular rotation contour wasthen usedto providethe fibrecentresforanoptimal (Scarrott etal. 1992; Rouan etal. 1997; Glinski & Nuth extraction algorithm which accounts for fibre cross talk 1997). Glinski & Anderson (2002) demonstrate that the in the spectra. Crosstalk was perceived to be a problem λ5800˚A cannot be generated via Q, R or P branch ro- during the initial data reduction due to the high surface tational emission structures derived from the unknown brightness contrast between the inner and outer nebula. molecular carrier DIB at λ5797˚A. Rouan etal. (1997) The extracted fibre flat frames are applied to the sci- proposethattheRRBhasitsoriginsinasupra-thermally enceframesafterthespectraareextractedinordertore- rotating PAH molecule, and go on to argue that a PAH taintheprojectedrelativeintensityinformationbetween of around 40 carbon atoms is implied by the required the fibres for the extraction process. rotational constants. This model assumes that the DIB A wavelength solution was derived, using the iraf identifyandreidentifytasks(fittinga6th orderLegendre carrier at λ5797˚A is identical to the λ5800˚A RRB car- rier and that angular momentum is accumulated due to polynomial, with a RMS residual of the order of 0.01˚A, a “rocket effect” induced by photo-dissociation events. for each fibre) using ThAr arc exposures. The spectra This effect is required to attain an estimated rotational are then independently transformed to a common wave- temperatureof450Kfortheλ5800˚Acarriers. Rotational lengthsolution. Asingle2Dtransformation,asmightbe band profiles are always expected of molecular spectra used for long slit data, is not possible due to disconti- since there exist a large number of energetically acces- nuities in the wavelength solution as a function of fibre sible rotational states, even when the molecule in ques- number at the boundaries between the sub slits which tion is electronically and vibrationally cold. Thus a sin- make up the argus slit in the giraffe spectrograph. gle vibronic transition will be derived from a number of On rectification an estimate of the flat field lamp illu- individual transitions originating in different rotational mination spectrum is created from the rectified flat field states. Energeticallyaccessiblevibrationalstates,poten- andmultiplied backinto the spectrato removethe lamp tiallypumpedbyphotoexcitation,canproducehot-band signature. features in an electronic spectrum, where the vibronic An absolute flux calibration for the data is not re- transitionsoriginatefromoneoftheseexcitedvibrational quired. However, a good relative calibration internal to states. Where the potential energy surfaces (PES), gov- the spectrum is important for molecular modeling. A erning the forces on the nuclei in the electronic states, responsefunction is derivedvia repeatedobservationsof are parallel, the structure of the excitation or emission the standard star EG-21, a DA white dwarf with few spectrum will be dominated by sequence bands. This spectral features in the wavelengthrange of interest and will be the case for a molecule with an extended chro- favourablylocated onthe night sky during observations. mophore since the shape of the underlying molecule will Spectra are subsequently corrected to a heliocentric be largely unaltered upon electronic excitation. wavelength solution. 3.2. Mosaicing the data 4.1. Sequence structure Data from individual cosmic-ray splits (typically Basedonourstudy oftheradialevolutionoftheemis- 3×1200sec) are combined using variance weighting and sionstructure along the arms of the bi-cone (Figures A1 outlierrejection. Threeditherpositionswereobservedat & A2) we propose that the λ5800˚A feature arises from each location in the nebula. In order to align the dither sequence structure, either associated with the nearby data, and to facilitate the ultimate combination of all λ5797˚A DIB carrier, or an alternate molecule. The ab- five telescope pointings into a uniform mosaic, we re- sence of any perfect correlations between DIBs leads sample the data onto a common regular Cartesian grid one to conclude that if the DIB carriers are molecu- after applying the appropriate position angle (PA) ro- lar (Thorburn etal. 2003) then the Franck-Condon fac- 4 Sharp, Reilly, Kable and Schmidt tors for electronic excitation are dominated by the ori- sitions in the visible region (Brechignac & Pino 1999). gin band. For a molecule in its vibrational and elec- Ground state geometries and normal mode frequencies tronic ground state, electronic excitation will similarly of the naphthalene, pyrene, tetracene, perylene and leave the molecule with zero vibration in the electronic coronene cations were obtained with the B3LYP density excited state. The dominance of the originband in elec- functional and 6-31G basis set using the gaussian 03 tronic excitation is indicative of a transition between (Frisch etal. 2003) suite of programs. Time-dependent states with very similar potential energy surfaces. If the DensityFunctionalTheory(DFT)calculationswereper- equilibriumgeometryorthevibrationalfrequencieswere formedonthe optimisedgroundstatestructurestoyield to change appreciably upon excitation, vibrational pro- excited states. The electronic transition in the visible gressions would be observed. regionwiththelargestoscillatorstrengthwaschosen. In If the carrier of the λ5800˚A feature behaves similarly orderto replicatethis excitedstate,the groundstate or- to the DIBs, then we would expect any emission from a bitaloccupancywasalteredtoresemblethe leading con- vibrationally excited population of molecules to exhibit figuration of the selected excited state. Geometry opti- sequence structure. Sequence structure arises from elec- misations and normal mode analyses were performed on tronic transitions where the molecule remains in vibra- these excited states for naphthalene, pyrene, tetracene tional states of the same character. If the potential en- and perylene. ′′ ergy surfaces of two electronic states are parallel, then The vibrational frequencies in the ground {ν } and i ′ the sequenceemissionwill occuratthe same wavelength excited states {ν } were used to generate transition en- i as the origin band. ergies with the origin transition (arbitrarily) centred at Inmanylargemolecules,electronictransitionsbetween λ5800˚A, in order that we be able to compare the band states of increasing vibrational excitation result in tran- shape with the new RRB observations. At the time of sitions ofprogressivelylowerenergydue to the tendency writing, only indicative origin positions may be calcu- for vibrationalfrequencies to be slightly lower in the ex- lated by quantum theory for molecules of the size being citedelectronicstate. Inlargemoleculeswhereelectronic invokedhere. Bywayofbenchmarking,pentaceneisseen transitions from the ground state are dominated by the toabsorbatλ5362˚A(Heinecke1998)butiscalculatedby origintransition,populationofhighervibrationalexcita- TD-DFT to absorbat 6190˚A(Reilly & Schmidt, private tion states results in a red-wardbroadening of the emis- communication). Sequencebandprofileswerecalculated sion structure. in the manner laid out in appendix A. This sequence structure interpretation requires the For an example of an observed molecular sequence moleculeresponsiblefortheemissiontohaveanextended structure see Figure 2 of Linnartz etal. (2004). chromophore, in order that the electronic transitions do notchangethestructureofthemoleculeappreciably(i.e. 4.3. DFT Modeled emission spectra the PESs are almost parallel). In deference to the current popularity in astronomyof It must be emphasised that the modelling of emis- PAHs we, in the firstinstance, follow the trend and pro- sion spectra is, in essence, an exploration of the pa- pose an intermediate size PAH molecule as the emitter. rameter space available to the molecules responsible for Alternative molecular forms (C60 - Kroto etal. (1985), theλ5800˚Aemission. Assuch,byinvokingnaphthalene, carbon chains - Douglas (1977)), are not considered at pyrene, tetracene and perylene cations we are exploring this time. theexpectedvibrationalsequencestructureofamolecule Using the Ultra-High Resolution Facility (UHRF, of that size and shape, with an extended chromophore, R∼600,000) of the Anglo-Australian telescope (AAT), rather than the species itself. Sarre etal. (1995a) show that the λ5797˚A DIB carrier Notwithstanding the fact that naphthalene cation is fulfils the requirements to generate sequence structure. likely photodestroyed in harsh circumstellar environ- The slightly red-degraded shape of the UHRF spectrum ments, it serves as a starting point for the exploration shows that the molecule is slightly larger in the excited ofourparameterspace. As canbe seeninfigure A3, the state, indicating a weakening of bonds and thus a prob- modeled naphthalene spectra are far too structured and able slight decrease in the vibrational frequencies of the covertoo largea wavelengthregionto be responsible for molecule. It is thus suggested that if a vibrationally ex- the RRBs. What’s more, the density of states does not cited population of the carrier of the λ5797˚A DIB were increase fast enough with energy to bring about a shift to fluoresce (or phosphoresce) it would exhibit sequence in the maximum of the emission spectrum. structure as manifested by red-ward broadening of the On moving to pyrene, we see that the spectra become DIB band shape. However, as the λ5800˚A feature does less resolved due to an increased number of emissive not asymptotically approach the λ5797˚A DIB position, states. At first glance it would appear that pyrene be- the simplest assumption to be made is that they arise haves in a very similar manner to the new RR spectra. from a different carrier. For the bands to arise from the However,the shift in the emissionpeak for pyrene is too samecarrierwerequirethatthecoldestemissionbecom- large to be considered a serious candidate for the RR pletely self-absorbed by the nebula. emission carrier (cf. Figures A2 & A3). Tetracene is calculated to have small vibrational fre- 4.2. PAH Model quency shifts as compared to naphthalene and pyrene. In order to explore the sequence band hypothesis, we As such it yields narrow sequence structure features performednormalmodeanalysesonPAHcationsofvar- whicharebroadlyinkeepingwiththebehaviourobserved ious sizes in ground and electronically excited states. in the RR. Cations, rather than neutrals, were chosen for these Perylene behaves in a similar way to pyrene yet its quantumcalculationssothattheyexhibitedstrongtran- sequence structure is calculated to be somewhat domi- Sequence Structure emission in The Red Rectangle 5 nated by a small frequency mode with a relatively large Given the similarity of the sequence structure pre- shiftuponexcitation. ThegreaternumberofBoltzmann dicted by the MM model it is thus difficult to pin-down accessiblevibrationalstatesaccessedintheelectronicex- the exact size of the λ5800˚A carrier. cited state brings about a shift in the maximum of the Despite the molecules invoked here displaying similar emissionspectrumtolongerwavelengths. However,even red-shiftedemissionspectra,thered-shiftoftheemission for perylene the magnitude of the red-shift of the calcu- maximum was found to be very temperature sensitive. latedsequencestructureisfargreaterthanthatobserved As such the MM modeled temperature of the Red Rect- in the RRBs. In order that the RRBs be explained by angle due to sequence emission is relatively insensitive sequence structure, we require a larger molecule with a to carrier(!). We found that the broadest emission spec- much larger density of states. tra observedin the Red Rectangle could be explained in Thesizeofthemoleculerequiredissuchthatthemod- termsofmoleculesatvibrationaltemperaturesof∼90K. eling of its sequence structure explicitly using DFT be- ModeledMMspectraforourfiducialmolecule,dicircum- comes intractable. While it may be possible to obtain coronene, are superimposed on λ5800˚A RRB in Figure ground state electronic structures and vibrational fre- A8. An initial estimate of the radial vibrational tem- quencies of very large molecules, the excited state elec- perature of the λ5800˚A band carrier is given in Figure tronicstructuresbecomeincreasinglycomplicated,being A9. brought about by mixing several excitations out of the 5.2. Deficiencies of the MM model groundstate. Assuch,representationofanexcitedstate withanalteredorbitaloccupancyisunreliable. Optimiz- While the redward shifts in the maxima of emission ing excitedstates by usingTD-DFT is beyondthe scope arereproducedwithlittle trouble,ourmodelmayneces- ofcontemporarycomputercodesforthesizeofmolecules sarily over-estimate the size of the PAH required for the being employed here (vide infra). observed (putative) sequence structure. This is because For molecules the size of coronene and larger, excited weselecttherangeoffrequencyshiftsfromDFTcalcula- statefrequencieswerechosenguidedbythebehaviourof tionsnecessarilyperformedonsmallmolecules. Onemay the smaller systems at the DFT level of theory. Figure expect that larger molecules will exhibit smaller shifts A4demonstratesthatmostfrequencyshiftsuponexcita- upon excitation as a consequence of the electronic tran- tion fall within an envelope (shaded). Randomly select- sitionbeingmoredelocalized. Theeffectofthiswillbeto ing shifts from this envelope changes the nature of the pull the sequence band contour into the origin band po- sequence structure very little, as shown in figure A5 for sition, reducing the amount of redshifting. As such, the pyrene. Similarly, calculating the ground state frequen- observed structure in the RR may be due to a smaller cies using molecular mechanics (MM) produces rather molecule than dicircumcoronene, at a higher tempera- similar results. The correlation between MM and DFT ture. calculatedfrequenciesistight,asshowninFigureA6. To Additionally,ourmodelonlycontainsasingletemper- investigatesystemslargerthancoronenewehaveinvoked ature population at this time. The effect of a compound a purely molecular mechanicalmodel, andassumed that populationalongthelineofsight,witharestrictedrange these structuresbehave inthe samemanner observedby temperature, would be to smear out the emission peaks DFT for the smaller PAH cations. Coronene frequen- somewhat. cies were calculatedwith both DFT andMM, with both 5.3. Should sequence structure be observed? results presented. Sarre etal. (1995b) note that one might expect se- 5. MOLECULAR MECHANICAL MODEL quence structure to be a feature of molecular emission bands. If one accepts that large PAH molecules exist UsingparameterstakenfromtheCHARMMforce-field in appreciablequantities inthe Red Rectangle,and that (Brooks etal.1983),vibrationalfrequenciesforelectronic they will be undergoing collisions and photoexcitation, groundstatesoflargePAHsweregeneratedbydiagonal- then one concludes that they will exhibit a significant ization of the mass-weighted Hessian matrix. Excited vibrationaltemperature. While manyvibrationalmodes state vibrational frequencies were chosen from the enve- can radiate energy (indeed, the 3.3µm and 7.6µm UIB lopeofDFTshiftsasoutlinedabove. Sequencestructure emissionisproposedtooriginatefromPAHC-HandC-C wasthensimulatedinanidenticalmannertotheimplicit stretches) there will be many modes which cannot cou- DFT results. ple to electromagnetic radiation. As such, vibrational energy can be “stored” in the PAH for an appreciable 5.1. Molecular Mechanical model results time. This phenomenon and the high photon flux in Using a molecular mechanical (MM) model, we ex- the nebula will bring about a vibrational temperature. plorea rangeoflargerPAHmolecules, including pyrene, This temperature may exceed 100K, and electronically coronene,hexabenzocoronene,circumcoroneneanddicir- excited molecules will radiate to vibrational wavefunc- cumcoronene. tions with nearly identical character to that in the ex- Figure A5 compares the modeled emission spectra of cited state. We conclude that a large PAH molecule in pyreneat200KfromtheDFTmodelandtheMMmodel. an environment such as the RR will inevitably exhibit The effect of increasing the size of the molecule in the sequence structure in its emission. MM model is that the density of states increases very 5.4. De-hydrogenation rapidly with energy and thus the sequence structure is shifted to the red. Spectra for the coronene, hexaben- Kokkin&Schmidt (in prep.) demonstratethatthe ef- zocoronene, circumcoronene and dicircumcoronene at a fect of de-hydrogenationon the in-plane electronic tran- range of temperatures are shown in Figure A7. sitions in large PAH molecules is primarily to shift the 6 Sharp, Reilly, Kable and Schmidt band a small amount (of the order of a few ∼10˚A) al- arbitrary choices. However, there are other good rea- though bond-length alteration induced by dehydrogena- sons to suggest a highly symmetric carrier. The Born- tion may be important in systems as large as decacyl. Oppenheimer breakdown which prevents many large Since in-plane transitions are the strongest and only in- molecules from fluorescing is facilitated by a large back- ′′ volve molecular orbitals of π-symmetry (A ), dehydro- ground of vibronic states of the same symmetry. By in- genationhaslittleeffectonthepositionoftheexcitation. vokingamoresymmetriccandidate,thenumberofback- Out-of-planetransitionsofPAHradicalsareexpectedto ground states of the same symmetry naturally falls, as be significantly shifted from the strong in-plane transi- there are more irreducible representations of the molec- tions of the neutral close-shell parent, and such spectra ular point group. As such, it is more likely that a highly will also exhibit a degree of vibrational structure due to symmetric molecule will fluoresce than a less symmetric the localization of the lone-pair chromophore. one. If the λ5800˚A band should indeed be attributed to emission from a 45-100 carbon atom PAH (such as di- 6. CONCLUSIONS circumcoronene)then one may postulate a secondband, We have demonstratedthat the radialvariationof the atlongerwavelengths,derivedfromthede-hydrogenated profile of the λ5800˚A Red Rectangle band is well mod- molecule. Such bands would almost certainly be ob- eled as molecular sequence structure arising in a PAH served in vibrationally hotter states since one expects molecule with 45-100 carbon atoms. Our model has the these bands to be more prominent closer to the central attractive properties of : star, an energy source capable of removing the hydro- • Not requiringanexotic molecule, onceone accepts genatomfromthe PAH.They wouldnotbe observedas the presence of PAHs in the ISM. sharpbandssuchastheRDemissionfeature. Suchbands would better resemble the Van Winckel etal. (2002) • Not requiring high nebular temperatures. Symmetric (S) bands. The Red Rectangle spectrum is replete with such features. Thus we tentatively propose • Not requiring an exotic emission mechanism. thesecondfeature,atλ5826.5(Van Winckel etal.2002), While sequence structure does not appear promi- the extreme red limit of our argus spectra, as sequence nently in the astronomicalliterature, it is a simple emissionfromadehydrogenatedPAHresponsibleforthe molecular emission phenomenon. λ5800˚A band. If sequence structure is the origin of the λ5800˚A RRB 5.5. Other Red Degraded (RD) bands. then it is natural to extend the hypothesis to the other RedDegradedRRBs,invokingsequencestructureforad- Ifoneacceptsasequencestructureexplanationforthe λ5800˚A RRB, it is tempting to propose similar expla- ditional molecules. nations for the remaining RD bands in the Red Rectan- gle. A numberofRD bands,eachofwhichcanplausibly Acknowledgements : The authors wish to thank ESO be assigned an associated S band derived from the de- staff for invaluable discussion regarding observing strat- hydrogenated molecule, are listed by Van Winckel etal. egy prior to observation. We thank the reviewer for (2002). Such bands wouldnot arise in a commoncarrier insightful comments which led to a greatly improved but would instead arise in similar 45-100 carbon atom manuscript. We thank ATAC for awarding two nights PAH molecules. Where a PAH has more than one sym- of the Australian flames guaranteed time to this pro- metry unique dehydrogenation site, a single RD band gram. RGSthanksANSTOforfinancialsupporttocover could have associated with it more than one S band. travelexpenses. NJR acknowledgesthe awardof a Grit- ton Scholarship. We thank Prof. P. Thaddeus and Dr. 5.6. Symmetry and candidate selection A. Gray-Weale for helpful discussions. Preparations for When suggesting candidate carriers for the RRBs, the observations reported made extensive use of excel- one is naturally tempted to choose highly symmetric lent archival material from HST and the 2MASS and molecules in order to avoidthe plurality associatedwith UCAC-2 programs. APPENDIX DFT CALCULATIONS OF BAND PROFILES Assuming no change in vibrational state, the photon energy for sequence transition j, T , is simply j 3N−6 ′ ′′ Tj =T0+h X nij(νi−νi), (A1) i=1 where T is the energy of the origin transition, N is the number of atoms in the PAH cation and n is the number 0 i of quanta of vibrational energy in mode i. Simulated emission spectra were produced by convolving the predicted transitions energies with a Gaussian profile simulating the rotational structure of each sequence band. The vibrational population in the excited electronic state was assumed to have an exponential dependence on vibrational energy of the ground vibrational state j ′′ P =exp(−βE ), (A2) j j Sequence Structure emission in The Red Rectangle 7 where 3N−6 ′′ ′′ Ej =h X nijνi. (A3) i=1 The simulated emission spectrum I(λ) is thus a sum over sets of quanta (vibrational states) I(λ)=XPjexp(−α(λ−hc/Tj)2) (A4) j where α controls the FWHM of the gaussian (rotational) profile. In cases where the number of states needed to converge the shape of the calculated sequence emission profile was prohibitively large, a Monte Carlo sampling technique was employed. Here, states, j, were chosen as a Monte Carlo Markov Chain and added to the simulation one at a time (starting at the ground vibrational level). The probability of accepting a step, in the space of the number of quanta in each vibrational mode, was given by the Metropolis technique,wherebyallstepsdowninvibrationalenergywerechosenwithunitprobabilityandthoseupinenergywere ′′ acceptedwithprobabilityP =exp(−βE ). Since the probabilityofstatej being occupied is naturallytakeninto the j j selection algorithm, the emission spectrum was taken as a simple sum over the sampled states. I(λ)=Xexp(−α(λ−hc/Tj)2). (A5) j Typically, 10000 states were enough to converge the shape of the emission band profile. REFERENCES A´da´mkovics,M.,Blake,G.A.,McCall,B.J.2003,ApJ,595,235 Maier, J.P., Lakin, N.M., Walker, G.A.H., Bohlender, D.A. 2001, Allamandola,L.J.,Tielens,A.G.G.M.,Barker,J.R.1985,ApJ,290 ApJ,553,267 ,L25 Men’shchikov, A.B, Schertl, D. Tuthill, P.G., Weigelt, G., & Allamandola, L.J., Hudgins, D.M., & Sandford, S.A. 1999, ApJ, Yungelson,L.R.2002,A&A,393,867 511,L115 Merrill,P.W.1934,Publ.Astron.SocPac.,46,206 Brechignac,P.,Pino,T.1999, A&A,343,L49 Merrill,P.W.1936,Publ.Astron.SocPac.,48,179 Brooks,B.R.,Bruccoleri,R.E.,Olafson,B.D.,etal.1983,J.Comp. Merrill,P.W.,&WilsonO.C.1938,ApJ,87,9 Chem.,4,187 Miyata,T.,Kataza,H.,Okamoto,Y.K.etal.2004,A&A,415,179 Cohen,M.,Anderson,C.M.,Cowley,M.,etal.1975,ApJ,196,179 Peeters,E.,Spoon,H.W.W.,&Tielens,A.G.G.M.2004,ApJ,613, Cohen, M., Winckel, H.V., Bond, W.E., & Gull, T.R., 2004, AJ, 986 127,2362 Press, W.H., etal. 1993, Numerical Recipes in C : The Art of Ding,H.,Schmidt,T.W.,Pino,T.,Boguslavskiy,A.E.,Gathe,F., ScientificComputing(2nded.;CambridgeUniversityPress) &Maier,J.P.,2003,J.Chem.Phys.,119,814 Rouan,D.,L´eger,A.,&LeCoupanec, P.1997,A&A,324,661 Donn,B.,Hodge,R.C.&Mentall,J.E.1968,ApJ,154,135 Sarre,P.J.,Miles,J.R.,Kerr,T.H.,Hibbins,R.E.,Fossey, S.J.,& Douglas,A.E.,1977,Nature,269,130 Somerville,W.B.1995,MNRAS,277,L41 Fulara,J.,&Krelowski,J.,2000,NewAR,44,581 Sarre,P.J.,Miles,J.R.,Scarrott,S.M.1995,Sci,269,674 Frisch, M.J., Trucks, G.W., Schlegel, H.B., Scuseria, G.E., Robb, Scarrott,S.M.,Watkin,S.,Miles,J.R.&Sarre,P.J.1992,MNRAS, M.A.,Cheeseman,J.R.,Montgomery,J.A.Jr.,Vreven,T.,etal. 255,11 2003gaussian03,RevisionB.05Gaussian,Inc.,PittsburghPA Schmidt, T.W., Boguslavskiy, A.E., Ding, H., etal. 2003, Geballe, T.R.,Tielens, A.G.G.M.,Allamandola, L.J.,Moorhouse, International JournalofMassSpectrometry, 228,647 A.&Brand,P.W.J.L.,1989, ApJ,341,278 Schmidt,G.D.,Cohen,M.,&Margon,B.1980, ApJ,239,L133 Glinski,R.J.,&Anderson,C.M.2002,MNRAS,332,17 Schmidt, T.W., & Sharp, R.G. 2005, Aust. J. Chem., 58N2, 69 Glinski,R.J.,&Nuth,J.A.iii1997,Ap&SS,249,143 (arXiv:astro-ph/0501180) Gillingham, P.R., Popovic, D., Farrell, T.J., Waller, L.G. 2003, Thaddeus,P.,&McCarthy,M.C.2001, Spectrochimica,57,757 SPIE,4841,1170 Thorburn,J.A.,Hobbs,L.M.,McCall,B.J.,Oka,T.,Welty, D.E., Heinecke,E.,Hartmann,D.,Muller,R.,&Hese,A.1998,J.Chem. Friedman, S.D., Snow, T.P., Sonnentrucker, P., & York, D.G. Phys.,109,906 2003,ApJ,584,339 Herbig,G.H.1995,ARA&A,33,1 VanWinckel,H.,Cohen,M.,&Gull,T.R.2002,A&A,390,147 Kerr,T.H.,Hibbins,R.E.,Fossey,S.J.,Miles,J.R.,Sarre,P.J.1998, vanderZwet,G.P.,&Allamandola,L.J.,1985,A&A,146,76 ApJ,495,941 Kroto, H.W., Heath, J.R., O’Brien, S.Co, Curl, R.F., & Smalley, R.E.1985,Nature,318,14 L´eger,A.&Dhendecourt, L.1985, A&A,146,81L Linnartz, H., Araki, M., Ding, H., Boguslavskiy, A.E., Kolek, P., Schmidt, T.W., Motylewski, T., Cias, P. etal. 2004, ApJ, 616, 1301 8 Sharp, Reilly, Kable and Schmidt TABLE A1 Observation log,20041229-30,VLT-UT2 ARGUSHR11. Pointing RA/Dec(J2000) Pos. Ang. Exp. time Date P0 06:19:58.14-10:38:14.0 12 3×2×600sec 2005-12-29 P1 06:19:58.32-10:38:23.3 165 3×2×1800sec 2005-12-29 P2 06:19:57.68-10:38:20.7 230 2×3×1200sec 2005-12-29 ’’ ’’ 1×3×1200sec 2005-12-30 P3 06:19:57.95-10:38:04.5 340 3×3×1200sec 2005-12-30 P4 06:19:58.63-10:38:06.4 43 3×3×1200sec 2005-12-30 Sequence Structure emission in The Red Rectangle 9 Fig. A1.— Thelocations andorientations ofbasetelescope pointings andtheargusIFUareshownoverlayed ontheHST PCimage, atF622W, of Cohenetal. (2004). Thelargecirclesaredrawnat aradius of10.7 and14.3arcsec, whichcorrespond to thefiducial WIYN DensepakspectraofGlinski&Anderson(2002). 10 Sharp, Reilly, Kable and Schmidt Fig. A2.— Asanexampleofthestrongevolutionoftheλ5800˚Aband,asafunctionofradialdistancefromthecentralstar,asequence ofspectrafromthesouthwesternarm(P2inFigureA1)ofthenebulaisshown. TheradialdistancefromHD44179isindicatedtotheleft. The spectra arevertically baseline offset for clarity. Spectra in the left-hand Figure have been continuum subtracted, whilethe identical spectraintheright-handFigurehavebeennormalisedbythecontinuum. Inthisillustrationthecontinuumhasbeenestimatedasasimple constantvalueoverthespectralrangeλλ5780-5790˚A.Averticaldashedlineillustratesthelocationoftheλ5797˚ADIB.Tickmarksindicate thepeak oftheλ5800˚A bandwithineachspectrum. Fig. A3.— Calculating the sequence structure emissionprofileusingDensity Functional Theory(DFT), weconclude that smallPAHs altertheirstructureconsiderablyonexcitation. Consequently,thered-shiftinthebandpeakistoogreatandtoomuchstructureisresolved withinthebandprofileformoleculesofthissizetorepresenttheλ5800˚Abandevolution.

Description:
material into the interstellar medium by a dying star. 2. THE RED .. cies using molecular mechanics (MM) produces rather similar results Donn, B., Hodge, R.C. & Mentall ,J.E. 1968, ApJ, 154, 135. Douglas, A.E. Kerr, T.H., Hibbins, R.E., Fossey, S.J., Miles, J.R., Sarre, P.J. 1998,. ApJ, 495, 941.
See more

The list of books you might like

Most books are stored in the elastic cloud where traffic is expensive. For this reason, we have a limit on daily download.