文档库

最新最全的文档下载
当前位置:文档库 > Tiny HI Clouds in the Local ISM

Tiny HI Clouds in the Local ISM

a r X i v :a s t r o -p h /0505055v 3 8 J u n 2005Astronomy &Astrophysics manuscript no.msc?n

February 2,2008

(DOI:will be inserted by hand later)Tiny H I Clouds in the Local ISM

Robert Braun 1and Nissim Kanekar 2

1

ASTRON,Postbox 2,7990AA Dwingeloo,The Netherlands e-mail:braun@astron.nl 2

National Radio Astronomy Observatory,Socorro,NM 87801,USA e-mail:nkanekar@http://www.wendangku.net/doc/d0a840bcc77da26925c5b077.html Received ;accepted Abstract.We report deep,high spectral resolution WSRT H I 21cm observations of four high latitude compact radio sources,that have revealed a new population of tiny,discrete clouds in the diffuse ISM,with peak optical depths τ~0.1?2%,H I column densities of 0.4–8×1018cm ?2and core temperatures of 20–80K.Imaging detections con?rm these low column densities and imply linear core dimensions of a few thousand AU,assuming a distance of 100pc.The physical origin of these tiny H I structures and their distribution in the ISM is at present unknown.Further observations will be required to determine whether they are a ubiquitous component of the ISM.Key words.ISM:clouds –ISM:bubbles –Stars:winds,out?ows –(Galaxy:)solar neighbourhood 1.Introduction Despite more than three decades of H I 21cm absorption &emission studies,we still do not have a good understanding of the physical characteristics of the atomic component,one of the most important phases of the interstellar medium (ISM).Crucial observational lacunae include the morphology of neu-tral hydrogen “clouds”,the time-scales of their formation and destruction (i.e.whether they are transient or relatively stable structures),the H I column density distribution function,the nature of the equilibrium between “cold”and “warm”phases,etc.In recent times,there have been several indications of a surprising degree of small-scale structure in the atomic ISM.

Perhaps the ?rst of these was the observation of spatially vari-able H I absorption towards compact radio sources (e.g.Dieter et al.1976;Davis et al.1996;Faison et al.1998).In the most extreme cases,for example towards 3C138,there is ev-idence for changes of as much as ?τ=0.1in the H I opacity on transverse scales of 20AU (Faison et al.1998).The sim-plest interpretation of these observations requires large vari-ations in the volume density,?n H i ~105cm ?3,assuming that all of the other relevant variables (speci?cally,the path-length and the temperature)are kept ?xed.However,as ar-gued by Deshpande (2000),realistic ISM structure functions can lead to large variations of τwith small angular offsets,simply from statistical ?uctuations in the effective pathlength with position.Further evidence for small-scale atomic structure stemmed from searches for time variability in the H I absorp-tion seen toward pulsars (e.g.Frail et al.1994).However,

the

Tiny HI Clouds in the Local ISM

2Robert Braun and Nissim Kanekar:Tiny H I Clouds in the Local ISM and B21325+32(1.4Jy,11hr),by observing in an in-band

frequency-switching mode,utilising a1MHz throw every5

minutes inside a total bandwidth of2.5MHz,with a chan-

nel width of0.5km/s.The in-band frequency switching al-

lowed exceptionally good band-pass calibration while provid-

ing100%of observing time on-source.We achieved RMS op-

tical depth sensitivities?τof1.8×10?4(for3C286),2.9×10?4

(for3C287)and4.0×10?4(for4C+32.44)at1km/s veloc-

ity resolution.These are the most sensitive H I21cm absorp-

tion measurements of which we are aware.A somewhat lower

sensitivity(?τ~0.0028per1km/s)was obtained towards

B21325+32,due to its lower?ux density.Finally,we also used

the simultaneously acquired total power spectra of the WSRT

single dishes to derive sensitive H I emission pro?les(with ef-

fective integration times of more than100hr)towards all four

sources.The?ux scale is based on the SEFD of the WSRT

dishes at1400MHz of300±10Jy/Beam.We emphasise that

only a constant term has been subtracted from each spectrum

(both absorption and emission)and not even a?rst order poly-

nomial,let alone a higher order spectral baseline.

3.Results

Since the primary motivation for the experiment was a search for broad,shallow absorption along“simple”high latitude lines of sight,we were surprised to?nd instead that multiple,narrow (FWHM~2–3km/s)absorption features were detected at dis-crete line-of-sight velocities,with peak opacities in the range 0.1–2%.A representative cool phase spin temperature of100K (corresponding to a thermal FWHM of2.2km/s)yields ex-tremely low H I column densities,0.4–8×1018cm?2,for these components.The detection of discrete absorption features was particularly surprising since the peak emission brightness seen in these directions(with the35′total power beam)is only be-tween2and5K.The absorption and emission spectra towards the four

Tiny HI Clouds in the Local ISM

Tiny HI Clouds in the Local ISM

Tiny HI Clouds in the Local ISM

Tiny HI Clouds in the Local ISM

Tiny HI Clouds in the Local ISM

Tiny HI Clouds in the Local ISM

Tiny HI Clouds in the Local ISM

Tiny HI Clouds in the Local ISM

sources are compared in the top and bottom panels of Figure1.It is striking that moderately nearby lines-of-sight (separation<105′between3C286and4C+32.44)show es-sentially uncorrelated absorption spectra,and only weakly cor-related emission features.Only in the case of4C+32.44and B21325+32,which have an angular separation of only15′,is there a hint of similar velocity components having been de-tected.In all cases,the emission spectra show wide wings ex-tending out to±50km/s,while the absorption is restricted to

the velocity range0to?30km/s.The very broadest emission in the spectra of Fig.1(b)is actually due to so-called stray radi-ation,H I emission attenuated by30to40dB,but still detected in the far side-lobes of the telescope response when the bright Galactic plane happens to be above the local horizon.

The same WSRT observations that provided the absorption spectra also allow an interferometric imaging search for emis-sion counterparts.Cubes of H I line emission were produced at a range of angular resolutions(15,30,60and120′′).Obtaining suf?cient brightness sensitivity to achieve detections in emis-sion typically required smoothing to60′′.At those velocities where total power emission exceeding about1K is seen in Fig.1,compact emission clumps of2–3K brightness are de-tected at apparently random locations in the?eld,superposed Fig.1.H I absorption(top)and total power emission(bottom) spectra after Hanning smoothing to1km/s velocity resolution. on the poorly-sampled(by the interferometer)diffuse back-ground emission.One such emission clump,immediately adja-cent to the3C286line-of-sight is shown in Fig.2(a).The intrin-sic angular size of these clumps appears to be about30′′,while their FWHM line-widths are1–2km/s,corresponding to tem-peratures of~20–80K.A representative peak column density for the clumps in the3C286?eld is N H i~5×1018cm?2.At an assumed distance of,say,100pc,the clumps would have a size of~3000AU,with a central volume density of~100cm?3.

An even more interesting emission structure is detected in the4C+32.44?eld.The line-of-sight toward this background

Robert Braun and Nissim Kanekar:Tiny H I Clouds in the Local ISM

3

Tiny HI Clouds in the Local ISM

Tiny HI Clouds in the Local ISM

Tiny HI Clouds in the Local ISM

Tiny HI Clouds in the Local ISM

Fig.2.(a)Top panel:H I emission clump adjacent to 3C286in map (left)and spectrum (right).(b)Bottom panel:H I emission shell toward 4C+32.44in map (left)and spectrum (right).

source appears to intersect a 15′diameter shell of H I emis-sion,as shown in Fig.2(b).Image quality is limited in Fig.2(b)by a non-ideal con?guration and no primary beam correction has been applied.Although it may be a chance superposi-tion,this apparent shell contains the G0III star HD 116856at (α,δ)2000=(13:25:55.835,+31:51:40.629).The measured par-allax of this star places it at 105±11pc,where the shell would have a diameter of ~0.45pc.The stellar proper mo-tion (?α,?δ)=(+14.73,?43.40)mas/yr is directed to the SE.Peak N H i ~1019cm ?2is seen in this structure,with FWHM line-widths of 2–3km/s.It seems possible that this structure is associated with the termination shock of a stellar wind.

4.Cloud models

Various approaches are used in the literature to constrain phys-ical conditions from an analysis of H I emission and absorption spectra (eg.Kanekar et al.2003,Heiles &Troland 2003).An exhaustive comparison of the various methods is beyond the scope of the current paper,but will be addressed in Braun &Kanekar (2005).We simply note from the outset that while it is always possible to model both H I absorption and emission pro?les as arising from the sum of multiple Gaussians,it is far from clear whether a plausible combination of physical con-ditions might exist along any real line-of-sight that would pro-duce such an arti?cially simplistic observable (ie.very long,yet iso-thermal,and non-turbulent path-lengths).In fact,the most sensitive observations of both emission and absorption lines (cf.Fig.1)show pro?les that appear to be semi-Lorentzian,with narrow line-cores that merge smoothly into broader wings.As more sensitive data become available,an increasing number

of broader Gaussians becomes necessary to ?t such pro?les.Motivated by the non-Gaussian line pro?les we have ex-plored some simple,spherically symmetric,isobaric two-phase cloud models of the form:n H (r )=n o exp[?(r /s )α1]for T <4000K and (1)n H (r )=n o exp[?(r /s )α2]for T >4000K (2)where we further relate volume density to temperature using a constant thermal pressure P /k B =n H T =1500cm ?3K.The gas temperature was allowed to vary between T min =20K and T max =15000K ,yielding an assumed thermal velocity disper-sion of σ2=0.0086T .The predicted H I absorption and total power emission spectra were calculated for a “cloud”placed at the central velocity of each observed feature in an attempt to simultaneously reproduce the observed spectra shown in Fig.1.The most important free parameters in this process were the cloud scale-length,s ,the cloud distance,d (which most strongly in?uences the predicted total power emission)and an impact parameter,b .This last parameter was used to allow for the likely circumstance that each spherical model cloud may not be penetrated exactly on-axis by the background absorber.The two power-law indices of the scaled radius,α1and α2de-termine the characteristic line shapes in the cold core and warm halo respectively.Although these were,in principle,also free parameters,it was found that only minor variations from “stan-dard”values of about α1=1/4and α2=1/8were needed.Our arbitrary choice of a transition temperature of 4000K to sepa-rate the cool and warm phases is also http://www.wendangku.net/doc/d0a840bcc77da26925c5b077.htmlparable ?ts are possible with other choices of the transition tempera-ture within a broad range.The broad-band (~90km/s FWHM)stray radiation contribution to the emission spectra is approxi-mated by a simple Gaussian,rather than a physical component.An illustration of the simultaneous emission and absorption line ?tting is shown in the top panel of Fig.3for the 3C286line-of-sight.A formal least squares ?t has not been carried out here,but merely a χ-by-eye to illustrate the possibilities of this approach.The bottom panel of the ?gure shows the entire set of density and temperature pro?les as well as the fractional distri-bution functions for the modelled clouds along all four lines-of-sight.This illustrates the basic similarities of the modelled fea-tures,although the apparent scale-lengths do vary by about an order of magnitude.The apparent distance used in reproducing the spectra was about 10pc,and the typical half-density radius was 100AU,although these were not very well-constrained given the very low angular resolution of our total power data (35arcmin).These modelling results can be easily scaled to other assumed thermal pressures by a linear scaling of n H to-gether with an inverse linear scaling of both s and d .For exam-ple,with a typical thermal pressure of only 150cm ?3K,com-parable ?ts would be obtained at apparent distances of 100pc and typical half-density radii of 1000AU.Variations from the idealised spherical cloud symmetry would have a similar im-pact on apparent distances and sizes.The preserved quantity in these scalings is the distribution of column density over tem-perature (or more precisely line-width).The fractional column density at a particular density and temperature with respect to the total line-of-sight column density is plotted in the lower

4Robert Braun and Nissim Kanekar:Tiny H I Clouds in the Local

Tiny HI Clouds in the Local ISM

Tiny HI Clouds in the Local ISM

Tiny HI Clouds in the Local ISM

Tiny HI Clouds in the Local ISM

Tiny HI Clouds in the Local ISM

Tiny HI Clouds in the Local ISM

Tiny HI Clouds in the Local ISM

Tiny HI Clouds in the Local ISM

Tiny HI Clouds in the Local ISM

ISM Fig.3.Overlay of observed,modelled and residual spectra for the 3C286l-o-s (top).(The broad stray radiation contribution to the emission pro?le is plotted separately.)Density and temper-ature pro?les (left)and fractional distribution functions (right)of all modelled “clouds”(bottom).

right panel of Fig.3.Column densities of the cool cores are typ-ically only a few percent of those in the warm envelopes.The total column densities of the modelled clouds vary between 5–35×1018cm ?2.While it should be emphasised that the current attempts are fairly simplistic,it is interesting that we obtain scale lengths quite similar to those of the structures seen di-rectly in emission in Fig 2.We also stress that our cloud model “temperatures”may not be wholly thermal in nature,but are likely to include a non-thermal,turbulent contribution.

5.Discussion

The narrow absorption features of Fig.1have the lowest column densities that have ever been detected in the CNM,two orders of magnitude lower than the mean CNM column density N H i =2.7×1020cm ?2of the “classical”McKee-Ostriker (1977)model (see also Stanimirovic &Heiles 2005).Our current results thus suggest that even the most diffuse re-gions of the Galaxy are populated by hitherto undiscovered,tiny distinct structures of very high density and temperature contrast.This implies substantial injection of ?uctuation power on very small scales,whose source is presently unclear.The detection of a narrow shell-like structure with small transverse separation from a G0III star in one of our ?elds suggests that the stellar winds of intermediate mass stars might play a role in the formation of these tiny clumps,whenever such stars ?nd themselves within a diffuse atomic structure.More dis-tributed sources of energy injection might also lead to ther-mal condensation in a turbulent ?ow as discussed by Audit &Hennebelle (2005).These authors also demonstrate that lo-cal pressure equilibrium is still approximately preserved in the vicinity of condensations,despite a wide range of associated gas temperatures,consistent with the isobaric two-phase cloud models we develop in §4.The density and temperature distri-bution functions of our model clouds,Fig.3bottom right,also compare very favourably with those of Audit &Hennebelle (their Fig.8)in the highly turbulent regime.We emphasise that the physical origin of these tiny H I structures and their general importance in the ISM is still un-known.Their very low column densities should cause them to evaporate rapidly (e.g.McKee &Cowie 1977).Perhaps the ?rst issue that needs to be clari?ed is whether they are truly ubiq-uitous in the ISM.We plan to address this with high sensitiv-ity WSRT absorption spectra towards a larger sample of bright high-latitude sources.Sensitive total power emission spectra with both higher angular resolution and a lower stray-radiation contribution using the Green Bank Telescope or Arecibo would also be invaluable.Clearly,more work needs to be done to fully characterise the nature of sub-structure in the diffuse ISM.Acknowledgements.The WSRT is operated by ASTRON with sup-port from the Netherlands Foundation for Scienti?c Research (NWO).References Audit,E.,Hennebelle,P.,2005,A&A,433,1Braun,R.,Kanekar,N.,2005,in preparation Davis,R.J.,Diamond,P.J.,Goss,W.M.,1996,MNRAS,283,1105Deshpande,A.A.,2000,MNRAS,543,227Dieter,N.H.,Welch,W.J.,Romney,J.D.,1976,ApJ,206,L113Faison,M.,Goss,W.M.,Diamond,P.J.,Taylor,G.B.,1998,AJ,116,2916Frail,D.A.,Weisberg,J.M.,Cordes,J.M.,Mathers,C.,1994,ApJ,436,144Heiles,C.,Troland,T.H.2003,ApJS,145,329Johnston,S.,Koribalski,B.,Wilson,W.,Walker,M.,2003,MNRAS,341,941Kanekar,N.,Subrahmanyan,R.,Chengalur,C.,Safouris,V .2003,MNRAS,346,L57McKee,C.F.,Cowie,L.L.1977,215,213McKee,C.F.,Ostriker,J.P.1977,218,148Points,S.D.,Lauroesch,J.T.,Meyer,D.M.2004,PASP,116,801Stanimirovic,S.,Weisberg,J.M.,Hedden,A.,Devine,K.E.,Green,

J.T.2003,ApJ,598,L23

Robert Braun and Nissim Kanekar:Tiny H I Clouds in the Local ISM5 Stanimirovic,S.,Heiles,C.2005,in preparation

Watson,J.K.,Meyer,D.M.,1996,ApJ,473,L127