VUV spectroscopy of carbon dust analogs: contribution to interstellar extinction

4.1 MB·

by  L. Gavilan

Astronomy&Astrophysicsmanuscriptno.27098˙ap (cid:13)c ESO2016
February1,2016
VUV spectroscopy of carbon dust analogs:
contribution to interstellar extinction
L.Gavilan1,I.Alata1,2,K.C.Le2,T.Pino2,A.Giuliani3,andE.Dartois1
1 Institutd'AstrophysiqueSpatiale(IAS),CNRS,Univ.ParisSud,Universite´Paris-Saclay,F-91405Orsay,France
2 InstitutdesSciencesMole´culairesd’Orsay(ISMO),CNRS,Univ.ParisSud,Universite´Paris-Saclay,F-91405Orsay,France
3 DISCObeamline,SOLEILsynchrotron,SaintAubin,France
e-mail:[email protected]
Preprintonlineversion:February1,2016
6 Abstract
1
0 Context.Afullspectralcharacterizationofcarbonaceousdustanalogsisnecessarytounderstandtheirpotentialascarriersofob-
2 servedastronomicalspectralsignaturessuchastheubiquitousUVbumpat217.5nmandthefar-ultraviolet(FUV)risecommonto
interstellarextinctioncurves.
n
Aims.OurgoalistostudythespectralpropertiesofcarbonaceousdustanalogsfromtheFUVtothemid-infrared(MIR)domain.We
a
J seekinparticulartounderstandthespectraofthesematerialsintheFUVrange,forwhichlaboratorystudiesarescarce.
Methods. We produced analogs to carbonaceous interstellar dust encountered in various phases of the interstellar medium: amor-
9 phoushydrogenatedcarbons(a-C:H),forcarbonaceousdustobservedinthediffuseinterstellarmedium,andsootparticles,forthe
2
polyaromaticcomponent.Analogstoa-C:Hdustwereproducedusingaradio-frequencyplasmareactoratlowpressures,andsoot

] nanoparticlesfilmswereproducedinanethylene(C2H4)flame.Wemeasuredtransmissionspectraofthesethinfilms(thickness<
M 100nm)inthefar-ultraviolet(190-250nm)andinthevacuum-ultraviolet(VUV;50-190nm)regionsusingtheAPEXchamberat
theDISCObeamlineoftheSOLEILsynchrotronradiationfacility.Thesewerealsocharacterizedthroughinfraredmicroscopyatthe
I SMISbeamline.
h. Results.Wesuccessfullymeasuredthetransmissionspectraoftheseanalogsfromλ=1µmto50nm.Fromthese,weextractedthe
p laboratoryopticalconstantsviaKramers-Kroniginversion.Weusedtheseconstantsforcomparisontoexistinginterstellarextinction
- curves.
o Conclusions.WeextendthespectralmeasurementsofthesetypesofcarbonaceousanalogsintotheVUVandlinkthespectralfeatures
r inthisrangetothe3.4µmband.Wesuggestthatthesetwomaterialsmightcontributetodifferentclassesofinterstellarextinction
t
s curves.
a
[ Keywords.ISM:dust,extinction-Infrared:ISM-Galaxies:ISM-Methods:laboratory:solidstate-Ultraviolet:ISM

1
v
1. Introduction terparts of the 3.4 µm band at 6.85 µm and 7.25 µm were later
7
3 observedbyspacetelescopes.Thecarriersofthesebandswere
0 1.1. Generaloverview namedHACbytheastrophysicalcommunity(e.g.,Pendleton&
8 Allamandola 2002; Furton et al. 1999; Scott et al. 1997; Duley
Carbon-based materials are abundant in circumstellar environ-
0 1994).TheHACdenominationforlaboratoryanalogshasbeen
ments and in the diffuse interstellar medium (DISM) (van
. used for a wide variety of hydrocarbon solids, even though the
1 Diedenhoven et al. 2004; Henning & Salama 1998; Greenberg
0 structureofthesematerialsvariesfromamorphoustodisordered
& Li 1999; Draine 2003; Ja¨ger et al. 2011). They are mainly
6 carbons.
produced in the envelopes of asymptotic giant branch (AGB)
1 Inthefollowing,werefertotheplasma-producedanalogsas
stars(Gailetal.2009;Molster&Waters2003)andareinjected
: a-C:H,atermusedinthephysicscommunity.Amorphousrefers
v into the DISM. Carbonaceous dust can then be processed (by
to the lack of order at any scale in these materials. The labo-
i photons, ions, cosmic rays) and can further evolve during their
X ratory infrared spectra of these materials match observed spec-
lifetimeintheseheterogeneousenvironments(Jonesetal.2014).
r tra,particularlyregardingtheshapeandrelativestrengthofthe
Carbonmaterialscanbeclassifiedaccordingtotheirsizes,which
a 3.4µmband(Dartoisetal.2007).Thestructureassociatedwith
range from (large) polycyclic hydrocarbon molecules to solid-
the hydrogen-rich a-C:H materials is characterized by few aro-
state carbon grains, their aliphatic vs aromatic content, and
maticunitslinkedinamainlyaliphaticbackbone(e.g.,Fig.5of
their spectra, among other properties (e.g., Henning & Salama
Dartoisetal.(2005)).
(1998)).Insolidcarbon,sp2,sp3,andmixedhybridizationsare
possible,givingrisetomaterialswithdifferentshort-,medium-, Some astrophysical carbon materials subjected to UV light
and long-range orders (Ja¨ger et al. 1998; Papoular et al. 1996; re-emitinthemid-infrared.Observationsoftheseemissionfea-
Tielens2008). tures, called AIB for aromatic infrared bands, appear consis-
Differentcarbondustcomponentsinspaceareidentifiedby tently between 3 to 20 µm and have been classified into three
their observed spectral signatures. In the infrared, observations mainclasses(Peetersetal.2002).Theintimatestructuresofthe
ofthe3.4µmabsorptionbandwereattributedtovibrationalCH carriersofthesedifferentclassesofAIBsarestillunknown,al-
2
andCH aliphaticmodesincarbonaceousmaterials.Thecoun- thoughtheyareprobablybuiltonapolyaromaticskeletonspan-
3
1
L.Gavilanetal.:VUVspectroscopyofcarbondustanalogs:contributiontointerstellarextinction
ning sizes from large molecules to small nanograins (Leger & tra from 300 nm to 40 nm for amorphous carbons. Colangeli
Puget 1984; Allamandola et al. 1985; Goto et al. 2007; Sloan et al. (1993) measured the spectra of carbon samples down to
et al. 2007; Pino et al. 2008; Tielens 2008; Joblin & Tielens ∼50 nm through transmission or reflection on a gold-coated
2011; Li & Draine 2012; Carpentier et al. 2012; Yang et al. quartzwindow.
2013). Thispaperisorganizedasfollows.InSect.2wedescribethe
Interstellar extinction curves, derived from observations of experimental setups used in sample preparation and characteri-
several lines of sight from the near-IR to the UV, provide an- zation of the carbonaceous materials. In Sect. 3 we present the
other clue to the type of dust present in the ISM. A persistent UVandMIR spectroscopyofbothmaterials. InSect.4we de-
spectralfeatureisthestrongabsorptionbumpat4.6µm−1(217.5 scribe the method used to derive the optical constants. In Sect.
nm), discovered by Stecher & Donn (1965), whose central po- 5 we discuss the astrophysical implications of these measure-
sition is surprisingly constant over several lines of sight, and ments.Finally,inSect.6wepresentourconclusions.
whoseoriginisstilldebated(Fitzpatrick&Massa2005,2007).
A steep far-ultraviolet (FUV) rise also characterizes many of
these curves, tracing extinction variations. The lack of correla- 2. Experiments
tionbetweenthespectralpropertiesoftheUVbumpandtherel-
ativestrengthsofthebumpandtheFUVrisepointstodifferent We studied two types of carbonaceous dust: a-C:Hs and soot
samples. The a-C:Hs were produced using an R.F. plasma re-
typesofdustcarriers(Greenberg&Chlewicki1983;Fitzpatrick
actor at low pressures. Their typical structural unit is shown in
& Massa 2007). Graphite particles were initially proposed as
Dartois et al. (2005). Soot materials have been produced in an
the carrier of the UV bump by Draine & Lee (1984). More re-
ethylene(C H )flameandprovidesamplesdominatedbyapol-
cently, hydrogenated carbons were proposed as potential carri- 2 4
yaromaticcarbonskeleton(e.g.,Carpentieretal.(2012)).While
ers(Colangelietal.1997;Schnaiteretal.1998;Mennellaetal.
our a-C:H materials have a high H/C, our soot materials have
1999; Ja¨ger et al. 2008), supporting theoretical work by Duley
lowH/C.
&Seahra(1998).
Twotypesofsubstrateswereused:MgF windowsandTEM
The central goal of this study is to obtain absolute spectral 2
properties of carbon dust analogs in both the VUV/Vis regions grids(TedPella,Inc.)coveredwithanm-thickmetallicorcarbon
membrane. Samples of MgF were used to ensure UV-visible
andtheinfraredranges,complementingearlierstudiesofhydro- 2
transmission measurements down to λ = 120 nm. Samples de-
carbonsolids,particularlyintheVUVrange,whereopticalmea-
positedonthe3mmTEMgridswereusedfortransmissionmea-
surements are scarce. A full spectral characterization of these
surementsfromλ=120nmto50nm.Thegridswere300mesh
materials will clarify their potential contributions as carriers of
AuorCusubstratescoveredwitha2-3nmlayerofgoldorplat-
observedspectralfeaturesintheinfraredandintheUV,suchas
inum, or 10-25 nm of carbon. Manipulation of the TEM grids
theFUVriseandalsothe217.5nmbumpofinterstellarextinc-
wasdonewithhigh-precisiontweezers,butmoreoftenthannot,
tioncurves.Inaddition,infraredspectraallowustocharacterize
the fragile grid supports were bent and/or the thin membrane
thequalityandsuitabilityofoursamplesasinterstellaranalogs.
shattered,eitherduringinstallationinthesampleholderordur-
These transmission spectra allow the derivation of optical con-
ingdepositionofthefilm.Weinspectedthefilmswithabinocu-
stantsfromthesecarbonaceousmaterials,necessarytosimulate
larmicroscopeattheInstitutd'AstrophysiqueSpatialeandwith
observedastronomicalextinctioncurves.
microscopicimagestakenattheSMISbeamlineattheSOLEIL
Carbon analogs have been produced in the laboratory us-
ing different methods and their spectra in transmission, reflec- synchrotron.
Asaresultoffourexperimentalcampaignsin2014and2015
tion,orelectronenergylossspectroscopy(EELS)(seeHenning
at the DISCO VUV beam line of the SOLEIL synchrotron, we
etal.(2004)forareview)hasbeenusedtoextractopticalcon-
foundthatwhileAuorPtmembranes(2-3nm)hadgoodtrans-
stants.Stecher&Donn(1965)firstderivedopticalconstantsof
mittanceintheVUV(T≤30%),theywereveryfragileduring
nanometer-sized graphitic particles. Draine & Lee (1984) de-
manipulation and during film deposition. Most VUV measure-
rived the optical constants from laboratory and observational
mentsonthesesubstratesshowedthatthemembranesweredam-
data for graphite down to 80 nm by the Kramers-Kronig ap-
aged either partially or completely. Instead, the carbon mem-
proachusingEELSdatafromTosatti&Bassani(1970).Zubko
branesonTEMgrids(T≤20%)provedtobemorerobust.We
et al. (1996) used the same method to obtain the optical con-
discuss the effect of these substrates on the film transmission
stants of several solid hydrocarbons from the millimeter to the
spectrainAppendixB.
FUV down to 40 nm. Schnaiter et al. (1996, 1998) derived the
optical constants through UV measurements of nano-sized car-
bongrainsusingthematrix-isolationtechniqueinthe200-400
2.1. Preparationofa-C:Hfilms
nm region. Furton et al. (1999) also obtained the optical con-
stants of a hydrogenated carbon film by trial and error fitting SICAL-ICP is anexperimental setup developed forthe purpose
in the range of 300 - 1000 nm. These values have a dispersion ofproducinghydrogenatedamorphouscarbonfilms,seeFig.1.
of between ∼10% and 20% at any given wavelength, and they Itconsistsofaninductivelycoupledradiofrequency(R.F.)gen-
comparecloselytootherpublisheddataforsimilara-C:Hmate- eratorworkingat13.56MHz(Eliteseries,MKS)withadynamic
rials,forexample,Smith(1984)between253nm-900nm,and range of 1 - 600 W. A background pressure of < 0.01 mbar is
for organic residues, for example, Jenniskens (1993) from 115 maintainedbyaprimarypump.Theprecursorhydrocarbongas
nm to 1.7 µm. Ja¨ger et al. (1998) derived the optical constants (CH ) supply is manually controlled with a fine valve and de-
4
of carbonized cellulose samples using the Lorentz-oscillator fit positsaremadeincontinuousflowkeptat∼0.1mbar,afterpas-
methodintherangeof200-500nmbasedonreflectancemea- sivationofthechamberwiththeprecursor.
surements. This was done later for gas-phase condensed soots TheR.F.dischargeisoperatedinaglasstube(length=1m,
from120nmtothemid-infrared(Ja¨geretal.2008).Fewlabora- φ =65mm),allowingfortheplasmatosymmetricallyenvelop
ext
torybaseddeterminationsofopticalconstantsofcarbonanalogs thesampleholder.TheR.F.powerisinductivelycoupledtothe
existforλ<120nm.Finketal.(1984)measuredtheEELSspec- plasma by means of an impedance matching network linked to
2
L.Gavilanetal.:VUVspectroscopyofcarbondustanalogs:contributiontointerstellarextinction
Figure1:SICAL-ICPisasetupallowingplasma-inducedchem-
icalvapordeposition.Itiscomprisedofaglasstubeinwhichan
inductively coupled plasma (ICP) on a continuum hydrocarbon
gasflowismaintainedbyanR.F.source.Theplasmaenvelopes
Figure2: Nanograins, an experimental setup for the prepara-
thesampleholderplacednearthematchingbox.
tionofsoots.Thisschemeshowstheburningregion,theextrac-
tionconeinquartz,andthegasflowregionshapedbyanozzle
wherethesootsamplesarecollectedbyimpactiononawindow
a copper coil around the glass cylinder. The tuning box allows (Carpentieretal.2012).
extracting a portion of the R.F. power coupled to the substrate
holder acting as an externally applied bias. For most samples,
the power was kept at 100 W, with 5 W reflected. The sam-
ple holder consists of a gold-coated copper block capable of
holding 20 mm windows. The holder can be internally cooled
by water, and the substrate can be installed by removing a me- The byproducts were then extracted by a water-cooled quartz
chanical cover. The estimated ion energies are on the order of cone (hole diameter = 1 mm) inserted into the flame at 30 mm
80 eV, thereby producing films with a hydrogen content up to fromtheburner.Theburnedgasandsootflowthroughanozzle
40% (Dworschak 1990). This R.F. source allows lower ion en- into a chamber at 5×10−2 mbar. Soot particles were deposited
ergydensitiesthanwiththe2.45GHzsourcepreviouslyusedby on MgF /KBr/NaCl windows or Au/C/Pt membranes, inserted
2
Dartois et al. (2005), giving a-C:H films with higher hydrogen directly into the molecular flow. The thicknesses of the soot
content due to the decreased preferential sputtering of the C-H nanoparticlefilmsvariedinarangeofafewtensofnmto1µm,
bondsvs.theC-Cbonds.Lowerpressuresleadingtohigherion depending both on the production rate and the exposure time,
energies also lead to sputtering of hydrogen. Samples are pre- tooptimizethespectralresponseinthedifferentwavelengthdo-
pared at room temperature, although the substrate temperature mainsfromtheIRtotheVUV.
hasastronginfluenceonthefilmproperties(Dworschak1990).
Tocontrolthedepositionrateforthefilms,weusedthecor-
relation between the optical depth of the 3.4 µm band and the 2.3. InstrumentsforUVandIRspectroscopy
thicknessofthefilms(d)forfilmswithd>100nm,usingtheir
Fabry-Perot interference fringes that are due to internal reflec- Ultravioletmeasurementsofourcarbonanalogswereperformed
tions within the film itself. We extrapolated this correlation for with three instruments. At the IAS we use a UV-Vis spectrom-
films with d < 100 nm (see Appendix A for details). As a first eter (λ = 210 nm - 710 nm) with a sub-nm resolution. It con-
estimateoftherefractionindexofthesefilmsintheoptical,we sists of a deuterium lamp, whose light is conducted by optical
usedtherelationin(Godard2011)betweentheopticalgap(E ) fibersandilluminatesthesample.Thetransmittedsignalisthen
04
and the refraction index measured for different a-C:H and a-C collected and carried by optical fibers toward a grating spec-
(amorphous carbon) films (e.g., Dischler et al. (1983); Lazar trometer (Maya PRO UV, Ocean Optics). For measurements in
(1998); Kassavetis et al. (2007)). For our a-C:H materials, we thevacuumultraviolet,weusetheAPEXbranchoftheDISCO
derived the Tauc band gap E = 2.5 eV, or E = 3.1 eV, and beamline at the SOLEIL synchrotron. The APEX chamber al-
g 04
usingtheaforementionedrelation,weestimaten ∼1.4-1.8. lowsmeasurementsatλ=50nm-260nminwindowlessmode
0
(Giulianietal.2009),withabeamsizeatthesampleof≤2mm.
Fromnowon,weinterchangebetweenthetermsVUVandFUV
2.2. Preparationofsootfilms to refer to this wavelength range. Finally, a spectrophotometer
(Specord, Analytikjena) allowed measurements on MgF win-
Soot samples were produced at the Institut des Sciences 2
dowsatλ=190-1000nm.
Mole´culairesd’Orsay(France)usingthesetupNanograins(Pino
etal.2008;Carpentieretal.2012).Inthesetup,whosescheme Infraredmeasurementsforthethicker(d>300nm)filmsde-
is shown in Fig. 2, a flat burner (McKenna) provides flames of positedonMgF windowsweremadeusinganevacuatedFTIR
2
premixedhydrocarbonandoxygengas.Suchalaminarflameis spectrometer (Bruker Vertex 80V) equipped with a KBr beam-
a one-dimensional chemical reactor offering a broad range of splitterandan(MCT)detectorworkinginthe7000to400cm−1
combustion conditions and sampling of byproducts. Here the (2.5 to 15 µm) spectral range at 1 cm−1 resolution. To mea-
pressure was maintained at 40 mbar by continuous pumping. sure the infrared spectra of ultra-thin films (d < 100 nm) we
Thefuelusedwasethylene(C H )premixedwithoxygenbefore used the infrared microscope available at the SMIS beam line
2 4
flowingthroughtheburneratacontrolledflowrate(4liters/min), attheSOLEILsynchrotron.Thismicroscopebenefitsfromhigh
and a C/O ratio set to 1.05. An N shield around the burn- synchrotron brilliance in the 2.5 - 100 µm spectral range and a
2
ing region was applied at a controlled flow rate of 3 liters/min. 10×10µmspotsizeonthesample.
3
L.Gavilanetal.:VUVspectroscopyofcarbondustanalogs:contributiontointerstellarextinction
(a)
Figure3:VUVmeasurementsofultra-thinfilms:a-C:Handsoot
optical depth spectra normalized to τ(∼10 µm−1). a-C:H film,
—: MgF (Specord), —: MgF (DISCO), —: TEM membrane
2 2
(DISCO). Soot film, —: MgF (Specord), —: MgF (DISCO),
2 2
—:TEMmembrane(DISCO).Dashedlinesshowmeasurement
dispersion.
3. Measurements
3.1. UVspectra:a-C:Handsoots
Transmissionspectrawereobtainedbycorrectingeachspectral
measurementbyanelectronicoffsetandthendividingthespec-
traofthedesiredcarbonfilmdepositedonthesubstrate(MgF or
2
amembrane)bythespectraofthesubstratealone.Weassumed
thattherearenoopticaleffectsduetothesubstrates,whichwe
further discuss in Appendix B. The resulting normalized opti-
cal depth spectra for an a-C:H and a soot nanoparticle film are (b)
showninFig.3.
Figure4: VUV optical depth in the VUV of (a) an a-C:H film
The UV-VUV spectra reveal the electronic structure of the
and(b)asootnanoparticlefilm(describedinthetext).Thede-
carbon analogs, arising from transitions between threefold (sp2
convolutedspectralfeaturesarelistedinTable1.
bonding)andfourfold(sp3)bondingcoordination,andanymix-
tureofthesetwobondings(spbondingisnotconsideredhereas
itisaminorcomponentinoursamples).Theopticalgapreaches
fromstatesinthevalenceσandπbandtostatesintheconduc- π∗ transitionsand/orn-σ∗ transitionsfromoxygencontamina-
tionσ∗ andπ∗ bands.Theπ-π∗ bandsduetosp2 sitestypically tionasdiscussedinGadallahetal.(2011).Thethirdabsorption
occur from 200 to 500 nm (2 - 5 µm−1) and dominate the op- (G3)peakat ∼9 µm−1 (115nm)in thefar-UVisassignedto σ
tical gap value. The σ - σ∗ transitions peak below 200 nm (5 - - σ∗ transitions. A fourth peak (G4) is found at ∼16 µm−1 (58
10 µm−1). Both of these bands produce an absorption from the nm).Atentativeidentificationofthisfourthpeakisgivenbelow.
UVtothevisibleandinfrareddomain.Theopticalgapinamor-
phousordisorderedcarboncanextendtotheIRinducedbyde- ThesootUVabsorptionspectrumwasfittedbyaLorentzian,
fectstates.Itcantakeavalueof0eVforgraphite,alongaplane, three Gaussians, and a linear function (slope ∼0.01 in energy
andupto5.5eVfordiamond. units).Thefirstpeak(L1)iswell-adjustedbyaLorentzianfunc-
An a-C:H sample of thickness ≥ 30 nm becomes optically tion peaking at 3.92 µm−1 and assigned to π - π∗ transitions.
thickintheVUV.However,thesefilmsareopticallythininthe The relatively strong intensity of this feature points to a highly
UV-Visrange,whichallowsustodiscernweakerabsorptionfea- poly-aromaticmaterial,closeinshapebutnotinpositiontothe
tures, such as the absorption bands at ∼4 - 5 µm−1. The recon- interstellarUVbumpat4.6µm−1.Aweakpeak(G2)at5.4µm−1
structed VUV spectra of an ultra-thin a-C:H film (d ∼20 nm ) isassignedtoπ-π∗ transitionsfromcarbonand/orn-σ∗ tran-
and soot film (d ∼50 nm) are shown in Figs. 4 (a) and (b), re- sitionsfromoxygencontamination.Athirdbroadpeak(G3)oc-
spectively.FittingparametersaregiveninTable1. curs at 10 µm−1 (100 nm) and a fourth peak (G4) is present at
Forthea-C:Hfilm,weidentifyfourabsorptionfeatures,fit- about 19 µm−1 (52 nm). The linear function helps fit the soot
tedbyGaussiandistributions.Twooverlappingabsorptionpeaks opacityfrom1µm−1(whilenolinearfunctionisrequiredforthe
(G1, G2) occur at ∼4 and at 5 µm−1. G1 is assigned to π - π∗ a-C:Hfit).
transitionsduetolocallyaromaticcarbonclusters(Robertson& Thepositionofpeaksinthesootfilmarefoundathigheren-
O’Reilly1987).G2isassignedtoasuperpositionofcarbonπ- ergiesthanforthea-C:H,with∆(L1-G1)=0.19eV,∆G2=0.45
4
L.Gavilanetal.:VUVspectroscopyofcarbondustanalogs:contributiontointerstellarextinction
Table1:Best-fitparametersforVUVspectra 3.2. Mid-infraredspectra:a-C:Hsandsoots
WemeasuredthecarbonanalogthinfilmsintheMIRwithtwo
Function Parameter[µm−1] soot a-C:H
objectives.Thefirstistocorrelatetheintegrated3.4µmbandto
area 3.33 the thickness of an a-C:H film to normalize the corresponding
Lorentzian(L1) center 3.92 laboratoryUVspectratointerstellarextinctioncurves.Thesec-
width 2.01 ondobjectiveistoassessthehomogeneityofthefilmsandcom-
height 0.06 parespectrafromthesamespatialregionwheretheUVspectra
Gaussian(G1) center 3.83 is measured. To this end, spectra with the SMIS FTIR micro-
width 5.64 scopeweretakenevery0.2mmacrosstheultra-thina-C:Hfilms
depositedonMgF andonTEMgridstocheckthereproducibil-
2
height 0.22 0.11
ityofdepositions.
Gaussian(G2) center 5.42 5.05
width 6.37 6.13
height 1.14 0.71
Gaussian(G3) center 9.71 8.67
width 1.81 1.83
height 1.58 1.21
Gaussian(G4) center 19.2 17.4
width 4.41 4.55
eV, ∆G3 = 1.62 eV, ∆G4 = 2.47 eV. Two π-electronic transi-
tionsarefoundbetween200-260nminbothmaterials.Forthe
hydrogen-richera-C:H,thelowersp2 carbonfractiondecreases
theπdensityofstates,whichaccountsfortheweakerπ-π∗band
strengths.BytakingtheopticaldepthratiooftheL1toG3fea-
ture in the soot spectra (Fig. 4a) and dividing it by the optical
depth ratio of the G1 to G3 feature in the a-C:H spectra (Fig.
4b), we find that the π - π∗ resonance is ∼10 times stronger in
sootthanina-C:H.
LaboratoryspectraofgasphasePAHshaveshowntheircon- Figure5:Rawtransmissionspectraofana-C:Hfilmandasoot
tributiontoaFUVrisefromabandwithanextrapolatedmaxi- nanoparticlefilm(shiftedby+65%forcomparison).
mumat∼17.7eV(70nm)(Joblinetal.1992).Lateron,Malloci
etal.(2004)computedtheabsolutephoto-absorptioncrosssec- Figure 5 shows the raw transmission mid-infrared spectra
tions of 20 PAHs in the VUV. They found a good agreement for an a-C:H and soot film. The continuum absorption is at-
both for the lower-lying bands of π - π∗ character and for the tributed to low-lying electronic transitions (Carpentier et al.
collective single broad absorption peak at ∼17 - 18 eV, com- 2012), while the superimposed bands are due to vibrational
posed of σ - σ∗, π - σ∗, σ - π∗ and Rydberg state transitions, modes. The CH stretching bond region is seen at 2800 - 3100
with possible contributions from collective effects (e.g., plas- cm−1. The bands between 1460 and 1380 cm−1 correspond to
mons). In graphite, in addition to the σ and π interband tran- the aliphatic CHn bending mode counterparts. The region be-
sitions, a band at 27 eV, measured in reflectance, was assigned tween1000-720cm−1correspondstoout-of-planearomaticCH
to the collective excitations of the π + σ plasmon (Djurisˇic´ & bendingmodesthatarenon-detectableduetothelowamountof
Li1999).Indiamond-likecarbonfilms(DLC),inadditiontothe sp2CH.Carbonylisdetectedat∼1718cm−1,mainlyduetooxy-
σandπinterbandtransitions,aFUVcontributiontoabsorption gencontamination.AweakOHbandispresentat∼3400-3500
wasdetectedusingsynchrotronellipsometry,peakingataround cm−1.
22 eV (56 nm, Franta et al. (2011), Fig. 5). The transitions in The intensity of the 3.4 µm band compared to the 6.2 and
thisrangeareattributedbytheauthorstovalenceσtoextended 8 µm bands is much lower for soots than for the a-C:H, where
states transitions σ - ξ∗. Figure 3 and Table 1 in Franta et al. aliphaticfeaturesdominate.Forthea-C:HfilmsweproduceH/C
(2011) depict the density of states for the corresponding band ∼1, while for soots, the H/C ratio is lower (∼0.01). The soot
model.Amarkeddouble-peakstructurewaspredictedbytheo- spectrumshowsa6.2µmband(at1584cm−1),duetotheC=C
reticalstudiesofthefrequency-dependentdielectricpermittivity sp2 stretchingmodes.The8.2µmband(at1225cm−1)isdueto
intheopticaltoVUVdomainofmodeledcarbonsootnanopar- C-C(sp3)stretchingmodesprobablyarisingfromdefects,with
ticles(Moulinetal.2008;Langletetal.2009).Thecalculations a contribution at 1160 cm−1 due to sp2 aromatic CH bending
predictedtwopeaksinthe∼20-30eVrange.Thedouble-peak modes. An alkyne CH stretching mode is present for soots at
structurewassuggestedtoarisefromthepolyaromaticunitswith 3300cm−1 andisnon-detectableforthea-C:H.Bandpositions
asizedistributionrangingfromafewangstromsuptoafewnm. arecoherentwithpreviouslaboratorysootinfraredspectra(Pino
Inlightofthesefindings,wetentativelyassigntheG4bandob- etal.2008;Carpentieretal.2012).
served in both analogs to the features observed in this energy Figure 6 (a) shows the deconvolution of the vibrational
rangeinothercarbonaceousmaterials.Thesetransitionsinvolve modesofthe3.4µmbandfora-C:HasinDartoisetal.(2004b).
excitationsfromπandσstatestowardextendedstatesabovethe The-CH and-CH groupsdominatethespectrumoftheband
2 3
σ∗ band.Thenatureoftheseexcitedstatesisstillunknownand at3.4µm,buttheshoulderat∼3050cm−1indicatesthepresence
workisinprogresstoexploreit. ofolefinicandaromaticC-Hbonds.
5
L.Gavilanetal.:VUVspectroscopyofcarbondustanalogs:contributiontointerstellarextinction
mode at 3.289 µm, A(CH) = 1.9 × 10−18 cm/group (Joblin
et al. 1994), we find N(CH) /N(CH ) = 1.1 ± 0.3,
arom 2,3 aliph
in agreement with Carpentier et al. (2012). Using the band
strengths from methyl-substituted PAHs (Yang et al. 2013), we
findN(CH) /N(CH ) =0.96±0.2.
arom 2,3 aliph
4. VUVopticalconstants
Optical constants were extracted from the transmission spectra
ofa-C:Hfilmsofvaryingthickness.Forthesootfilm,athickness
∼50nmwasestimateddirectlyviaAFM.Fromthetransmission
spectra, we obtained first-order values for the imaginary index
ofrefraction,k(ν¯),accordingtotheBeer-Lambertlaw,
τ(ν¯)
k∼ , (1)
4πν¯d
where τ is the optical depth, ν is in wave number units, and d
(a)
is the sample thickness. Before this, a baseline for the interfer-
ence fringes due to multiple internal reflections was removed.
Byfittingtheabsorptionink byaseriesofGaussianprofiles,k
was then extrapolated to 0 eV and to higher energies, to apply
the Kramers Kronig relations, closure relations linking the real
and imaginary parts of the complex dielectric index of refrac-
tion (Bohren & Huffman 1998). To derive the real part of the
complex index of refraction from our absorption data, we used
the subtractive Kramers-Kronig (SKK) method (Warren 1984).
In this method, we fixed n(ν ), known as the anchor point, a
0
knownvaluefortheindexofrefractionatareferencefrequency
farfromanystrongabsorptioninthematerialconsidered,gener-
allyintheNIR-visible.SKKisthepreferredmethodifkisonly
knownoveralimitedintervalandtheintegralsareevaluatednu-
merically.Fortherealindex,
2 (cid:90) ∞ ν(cid:48)k(ν(cid:48))
n(ν)=1+ P dν(cid:48), (2)
π ν(cid:48)2−ν2
0
where P denotes the Cauchy principal value of the integral.
(b)
Inversely,fortheimaginaryindex,
Figure6: Gaussian deconvolution of the 3.4 µm band of an (a) 2ν (cid:90) ∞ n(ν(cid:48))
a-C:H film and (b) soot film. The black lines correspond to the k(ν)= P dν(cid:48). (3)
-CH3 groups,thebluelinestothe-CH2 groups.Thefitresidual π 0 ν2−ν(cid:48)2
isingray.
ForimprovedaccuracyweusedtheformulationinTrotta&
Schmitt(1996),whichreducestheerrorbyusingafinitewave-
lengthintegrationrange,
Theobservedinterstellar3.4µmbandinthediffusemedium
andthemeasuredIRspectraoftheproduceda-C:Hfilmsagree
well (Dartois et al. 2005, 2007; Godard & Dartois 2010). We n(ν)=n(ν )+ 2P(cid:90) ν2(cid:20)k(ν(cid:48))ν(cid:48)−k(ν)ν − k(ν(cid:48))ν(cid:48)−k(ν0)ν0(cid:21)dν(cid:48).
usedtheoscillatorstrengthoftheobservedmodestodetermine 0 π ν(cid:48)2−ν2 ν(cid:48)2−ν2
the aliphatic CH /CH ratio of our a-C:H films (Dartois et al. ν1 0 (4)
2 3
2004a,2007).TheoscillatorstrengthfortheCH antisymmetric For a-C:H we took n = 1.5 as an anchor point at 1 µm.
3 0
mode,at3.38µm,A(a-CH )=12.5×10−18cm/group,whilefor Forsootwetookn =1.7(Schnaiteretal.1998).TheKramers-
3 0
CH antisymmetric + Fermi resonance at 3.377 µm, A(a-CH ) Kronignumericalintegrationcalculatesthedesirednvalueover
2 2
=8.4×10−18 cm/group(Dartoisetal.2004a,2007).Thecalcu- the input energy range. Subsequently, the complex index of re-
latedN(CH )/N(CH )ratioobtainedfromfittingthevibrational fraction values were used to numerically simulate a synthetic
2 3
modes of the 3.4 µm band is 2.4 ± 0.3, in agreement with ob- transmission value at each wavelength. For this, we used the
servationstowardCygOB2No.12,whichsamplesonlydiffuse modelofSwanepoel(1983),whoderivedtheexpressionforthe
ISMdust(Pendleton&Allamandola2002),wherearatioof2.5 transmissionofathinabsorbingfilmonatransparentsubstrate
± 0.4 is found, and to observations toward the infrared galaxy (MgF in our case). We constructed a real index template for a
2
IRAS08572+3915(Dartoisetal.2007). substratemadeofMgF from120-1000nm.Forλ=50-120
2
Figure 6 (b) shows the deconvolution of the vibrational nmaninversionwasmade,keepingaconstantnvalue,buteven-
modes of the 3.4 µm band for the soot nanoparticle film. From tually, a more sophisticated scattering model should take into
this, the calculated aliphatic N(CH )/N(CH ) ratio = 7 ± 3. accounttheeffectsofanon-transparentsubstrate.Thedifference
2 3
Using the oscillator strength for the aromatic CH vibrational betweenthesyntheticandmeasuredtransmissionwascalculated
6
L.Gavilanetal.:VUVspectroscopyofcarbondustanalogs:contributiontointerstellarextinction
5. Astrophysicalimplications
(a)
Figure8: Normalized extinction, A(λ)/A(V), of representative
lines of sight from Fitzpatrick & Massa (2007) (solid purple
and black). Laboratory spectra for an a-C:H re-normalized to
theopticaldepthoftheobserved3.4µmband(dottedgreen)and
renormalizedlaboratorysootspectratoitsVbandopticaldepth
(dotted blue). Calculated extinction curves from laboratory de-
rivedopticalconstantsforana-C:H(dashedgreencurve)anda
soot(dashedblue).
Fitzpatrick&Massa(2007)analyzed328Galacticinterstel-
lar extinction curves from the UV to the IR. Figure 8 shows
(b)
twoselectedextinctioncurves(HD21007andALS8107)which
may represent a-C:H and soot components. The laboratory a-
Figure7:Derivedopticalconstantsfor(a)ana-C:Hfilmand(b)
asootnanoparticlefilm,measureddowntoλ=50nm.Thedot- C:Hspectrumisscaledtoitslaboratory3.4µmopticaldepthand
to the expected observational ratio of 250 toward the Galactic
tedlinesrepresentthecompoundederrorbarsdominatedbyun-
center(Fig.8).Observationsoftheopticaldepthofthe3.4µm
certaintyinthethicknessandanchorpoints(n ).
0
bandhavebeencomparedtothevariationsinvisualextinction,
A ,towardGalacticsourcesanddiffuseISMlinesofsight,and
V
acorrelationbetweenthesetwohasbeenfound(Sandfordetal.
1995;Godard2011).AratioofA /τ ∼250hasbeenestimated
V 3.4
ateachpointuntilthedifferencebetweentheobservedandsyn- for lines of sight of the local ISM, while observations toward
theticspectrawasminimized,whichoccurredafterafewitera- theGalacticcentergivearatioofAV/τ3.4 ∼150(Sandfordetal.
tions. 1995; Rawlings et al. 2003), although the difference is proba-
blylower(Godardetal.2011).Thesootopticaldepthspectrum
The optical constants derived for the laboratory a-C:H and
(Fig.8)isscaledtoitsopticaldepthintheVband(551nm).
sootfilmsareshowninFig.7(a)and(b).Fora-C:Hthesystem-
We used the laboratory optical constants obtained in Sect.
atic uncertainty is mostly due to the thickness estimated from
4 to calculate the extinction, Q , that is due to a distribution
then*thicknessvs.3.4µmcorrelation(seeApp.A),whereun- ext
of grain sizes, computed using the Mie formulae (Bohren &
certaintiesinthethicknessoftheultra-thinfilmgoupto∼50%
Huffman 1998). Our primary objective is to illustrate the effect
fora-C:H.Theuncertaintyforthesootfilmthicknessisaslarge
ofthelaboratoryopticalconstantsfromthisstudyonextinction
as∼25%(estimatedfromAFMmeasurements).Theimaginary
curves.Asafirstattempt,wetookintoaccountasimplesizedis-
indexofrefractionreflectsthemajorabsorptivefeaturesseenin
tributionforVerySmallGrains(VSGs)fromDesertetal.(1990),
the laboratory spectra, that is, two FUV bumps for both mate-
thatis,
rials at ≥ 10 µm−1 and a Lorentzian peak at ∼4 µm−1 for the
soot. The uncertainty in the thickness measurement dominates n(a)∝a−p (5)
the systematic uncertainty, demonstrating the experimental dif-
where p = 2.6, and the grain sizes varied between a = 1 to
ficulty of measuring ultra-thin films in the VUV. As in Franta min
a =15nm.Thetotalextinctionisthencalculatedas
etal.(2011),whomeasuredtheopticalpropertiesofDLCfilms max
byellipsometryinthe5-30eVrange,thepresentstudyshows (cid:90) amax
thatdirectopticalmeasurements(e.g.,intransmission)arenec- Cext[cm−1]∝ πa2n(a)Qextda. (6)
essarytoremoveambiguitiesincalculatingopticalconstantsin amin
the VUV, especially concerning the contribution of the band at This analysis is limited to spherical, homogeneous, optically
∼22eV,whichareexpectedbythemodelsdescribedinSect.3.1. isotropic particles, and the resulting extinction is given by the
7
L.Gavilanetal.:VUVspectroscopyofcarbondustanalogs:contributiontointerstellarextinction
greenandbluedashedcurvesinFig.8.Theseopticalconstants sorption, which peaked in the 16 - 19 eV range. This band is
should be implemented in more sophisticated extinction mod- predictedbycalculationsandispresentatvaryinglevelsinother
els.Thecalculatedextinctionshowsthatbothmaterialspresent carbonaceous materials. It involves interband transitions to ex-
asteepFUVrise.However,thepositionoftheFUVriseispri- tendedstates.
marily dictated by the position of the maximum of the σ - σ∗ By combining UV measurements to IR measurements, we
band. For a-C:H band maximum positions occur at lower ener- scaled the UV contribution to the 3.4 µm band to compare the
giesthanforsoots,whichdominatethesteepFUVrise.TheUV observedUVextinctiontoourlaboratorya-C:Hmeasurements.
peakofthecalculatedextinctioncurvecalculatedforlaboratory We derived the optical constants for a laboratory a-C:H and
sootsiscloserinpositiontotheaveragegalacticextinctionUV sootfromseveralsetsoftransmissionmeasurementsatdifferent
bumpat4.6µm−1(Fitzpatrick&Massa2007). thicknesses. We used these optical constants to calculate
Draine&Lee(1984)suggestedthatcrystallinegraphitepar- extinction curves that show that a-C:Hs contribute to the steep
ticles could be a carrier of the UV bump, but due to the diffi- FUV rise, while soots are suitable carriers of the UV bump.
culty of condensing graphite in space and the need for specific Astronomical spectra of dust shortward of the Lyman limit are
size distributions for spectral agreement (Czyzak et al. 1981), currentlylacking,mostlyduetotheabsorptionlinesofhydrogen
othercarriersforthisbumpweresought.Mennellaetal.(1996, and helium, which have large FUV and extreme-UV (EUV)
1999) and later Gadallah et al. (2011) showed that an increase effective absorption cross sections (Hawkins & Wright 1991).
of the sp2 carbon fraction as a result of UV irradiation of their However, in dusty regions (where the opacity due to the gas
carbon materials could give rise to the UV bump. Considering component may be lower than the dust opacity), the laboratory
themid-infraredandUVfeatures,theirsamplesaremorecom- FUVspectralfeaturesofcarbondustmaybeobserved.
parable to our soot than to our a-C:H sample. Li et al. (2008)
used buckyonions to reproduce observed interstellar extinction Acknowledgments. We thank Christophe Sandt (SOLEIL) for his
curves. Ja¨ger et al. (1998) synthesized carbon analogs by py- helpwiththeinfraredmicroscopeattheSMISbeamline.VUVmea-
rolizingcellulosematerialsandlaterbylaserablationofgraphite surements were made at the DISCO beam line of the SOLEIL syn-
(Ja¨geretal.2008).However,theironion-likecarbonnanoparti- chrotron radiation facility (projects: 20141074 & 20130778). This
cles did not show a strong UV bump. Polycyclic aromatic hy- workhasbeensupportedbytheFrenchprogram“PhysiqueetChimie
drocarbons (PAH) were proposed to simultaneously contribute du Milieu Interstellaire”, the ANR COSMISME project (grant ANR-
to the 217.5 nm bump and the FUV rise (Malloci et al. 2004; 2010-BLAN-0502).L.G.acknowledgessupportfromtheCNESpost-
Joblin et al. 2009). Steglich et al. (2010, 2013) suggested that doctoralfellowshipprogram.
PAHswithsizesaround50-60carbonatomspermoleculecould
beresponsible fortheUVbump throughaheterogenous distri-
butionofsizesandnotthroughsingleisolatedspecies.Cecchi- References
Pestellini et al. (2008) showed that these two spectral features
Allamandola,L.J.,Tielens,A.G.G.M.,&Barker,J.R.1985,ApJ,290,L25
and their variations can be accounted for by mixtures of PAHs Bohren,C.F.&Huffman,D.R.1998,AbsorptionandScatteringofLightby
indifferentionizationstates.Morerecently,hydrogenatedamor- SmallParticles,WileyScienceSeries(Weinheim:Wiley-VCH)
phous carbon grains, believed to be more easily formed in car- Carpentier,Y.,Fe´raud,G.,Dartois,E.,etal.2012,A&A,548,A40
Cecchi-Pestellini,C.,Malloci,G.,Mulas,G.,Joblin,C.,&Williams,D.A.2008,
bonstaratmospheres,havebeenconsideredaspotentialcarriers
A&A,486,L25
of the UV bump, particularly when irradiated by UV photons
Colangeli,L.,Bussoletti,E.,Pestellini,C.C.,etal.1997,AdvancesinSpace
(Mennellaetal.1996,1997;Gadallahetal.2011). Research,20,1617
Li&Greenberg(1997)modeledinterstellarextinctionbyus- Colangeli,L.,Mennella,V.,Blanco,A.,etal.1993,ApJ,418,435
ingathree-componentdustmodel:silicates,amorphouscarbons, Czyzak,S.J.,Hirth,J.P.,&Tabak,R.G.1981,VistasinAstronomy,25,337
Dartois,E.,Geballe,T.R.,Pino,T.,etal.2007,A&A,463,635
and PAHs. Recent extinction models including only two com-
Dartois,E.,Marco,O.,Mun˜oz-Caro,G.M.,etal.2004a,A&A,423,549
ponents (silicates and a-C/a-C:H) (Jones 2012a,b) have found Dartois,E.,Mun˜ozCaro,G.M.,Deboffle,D.,&d’Hendecourt,L.2004b,A&A,
goodagreementwithpreviouswork.Moreimportantly,theyac- 423,L33
count for the FUV rise and the UV bump only by means of Dartois,E.,Mun˜ozCaro,G.M.,Deboffle,D.,Montagnac,G.,&D’Hendecourt,
L.2005,A&A,432,895
the size-dependent properties of the carbonaceous component.
Desert,F.-X.,Boulanger,F.,&Puget,J.L.1990,A&A,237,215
The calculated extinction curves (using a a size distribution of
Dischler,B.,Bubenzer,A.,&Koidl,P.1983,AppliedPhysicsLetters,42,636
carbondustgrainscharacterizedbyouropticalconstants)show Djurisˇic´,A.B.&Li,E.H.1999,JournalofAppliedPhysics,85,7404
that soot-like nanograins could be an important contributor to Draine,B.T.2003,ARA&A,41,241
theUVbump,whilea-C:Hscancontributemoresignificantlyto Draine,B.T.&Lee,H.M.1984,ApJ,285,89
Duley,W.W.1994,ApJ,430,L133
the steep FUV rise. However, other dust components (like sili-
Duley,W.W.&Seahra,S.1998,ApJ,507,874
cates)shouldbeconsideredinacompletemodeloftheobserved Dworschak,W.1990,ThinSolidFilms,189,257
extinction. Fink,J.,Mu¨ller-Heinzerling,T.,Pflu¨ger,J.,etal.1984,Phys.Rev.B,30,4713
Fitzpatrick,E.L.&Massa,D.2005,AJ,130,1127
Fitzpatrick,E.L.&Massa,D.2007,ApJ,663,320
6. Conclusions Franta,D.,Necˇas,D.,Zaj´ıcˇkova´,L.,Bursˇ´ıkova´,V.,&Cobet,C.2011,ThinSolid
Films,519,2694
Wehavedescribedthepreparationoflaboratorya-C:Handsoot Furton,D.G.,Laiho,J.W.,&Witt,A.N.1999,ApJ,526,752
nanoparticleanalogstodustfoundintheISM.Wecharacterized Gadallah,K.A.K.,Mutschke,H.,&Ja¨ger,C.2011,A&A,528,A56
Gail,H.-P.,Zhukovska,S.V.,Hoppe,P.,&Trieloff,M.2009,ApJ,698,1136
thinfilms(d<100nm)ofthesematerialstomeasuretheirtrans-
Giuliani, A., Jamme, F., Rouam, V., et al. 2009, Journal of Synchrotron
mission spectra in the λ = 1 µm to 50 nm range. These curves Radiation,16,835
canbedecomposedintoabsorptionfeaturesinvolvingtransitions Godard,M.2011,Theses,Universite´ParisSud-ParisXI
from states in the valence π band to states in the conduction σ Godard,M.&Dartois,E.2010,A&A,519,A39
Godard,M.,Fe´raud,G.,Chabot,M.,etal.2011,A&A,529,A146
bandaround5and10eV,thepositionandrelativeintensityvary-
Goto,M.,Kwok,S.,Takami,H.,etal.2007,ApJ,662,389
ing with the sp2 content of the material. The description of the
Greenberg,J.M.&Chlewicki,G.1983,ApJ,272,563
measurements required adding another band to explain the ab- Greenberg,J.M.&Li,A.1999,AdvancesinSpaceResearch,24,497
8
L.Gavilanetal.:VUVspectroscopyofcarbondustanalogs:contributiontointerstellarextinction
Hawkins,I.&Wright,E.L.1991,inExtremeUltravioletAstronomy,ed.R.F.
Malina&S.Bowyer,333
Henning,T.,Ja¨ger,C.,&Mutschke,H.2004,inAstronomicalSocietyofthe
PacificConferenceSeries,Vol.309,AstrophysicsofDust,ed.A.N.Witt,
G.C.Clayton,&B.T.Draine,603
Henning,T.&Salama,F.1998,Science,282,2204
Ja¨ger,C.,Mutschke,H.,&Henning,T.1998,A&A,332,291
Ja¨ger,C.,Mutschke,H.,Henning,T.,&Huisken,F.2008,ApJ,689,249
Ja¨ger,C.,Mutschke,H.,Henning,T.,&Huisken,F.2011,inEASPublications
Series,Vol.46,EASPublicationsSeries,ed.C.Joblin&A.G.G.M.Tielens,
293–304
Jenniskens,P.1993,A&A,274,653
Joblin,C.,Berne´,O.,Simon,A.,&Mulas,G.2009,inAstronomicalSociety
ofthePacificConferenceSeries,Vol.414,CosmicDust-NearandFar,ed.
T.Henning,E.Gru¨n,&J.Steinacker,383
Joblin,C.,D’Hendecourt,L.,Leger,A.,&Defourneau,D.1994,A&A,281,923
Joblin,C.,Leger,A.,&Martin,P.1992,ApJ,393,L79
Joblin,C.&Tielens,A.G.G.M.,eds.2011,EASPublicationsSeries,Vol.46,
PAHsandtheUniverse:ASymposiumtoCelebratethe25thAnniversaryof
thePAHHypothesis
Jones,A.P.2012a,A&A,540,A1 FigureA.1: Correlation between n*thickness of films (d > 300
Jones,A.P.2012b,A&A,540,A2 nm)vs.theintegrated3.4µmbandarea.Thisisusedtoestimate
Jones,A.P.,Ysard,N.,Koehler,M.,etal.2014,ArXive-prints
n*thicknessfortheultra-thinfilms(d<100nm)usingtheirinte-
Kassavetis,S.,Patsalas,P.,Logothetidis,S.,Robertson,J.,&Kennou,S.2007,
grated3.4µmvalues.Theinsetshowsthe3.4µmbandofa-C:H
DiamondandRelatedMaterials,16,1813
Langlet, R., Vanacharla, M., Picaud, S., & Devel, M. 2009, Journal of filmsproducedwithdifferentdepositiontimes.
Quantitative Spectroscopy and Radiative Transfer, 110, 1615 , {XI}
Conference on Electromagnetic and Light Scattering by Non-Spherical
Particles:2008
Lazar,G.1998,ActaPhysicaPolonicaA,100,67
Leger,A.&Puget,J.L.1984,A&A,137,L5
Li,A.,Chen,J.H.,Li,M.P.,Shi,Q.J.,&Wang,Y.J.2008,MNRAS,390,L39
whereαistheabsorptioncoefficient,anddisthefilmthickness.
Li,A.&Draine,B.T.2012,ApJ,760,L35
Li,A.&Greenberg,J.M.1997,A&A,323,566 The absorption coefficient α is related to the imaginary part of
Malloci,G.,Mulas,G.,&Joblin,C.2004,A&A,426,105 thecomplexindexofrefraction,k(ν)by
Mennella,V.,Baratta,G.A.,Colangeli,L.,etal.1997,ApJ,481,545
Mennella,V.,Brucato,J.R.,Colangeli,L.,&Palumbo,P.1999,ApJ,524,L71
Mennella,V.,Colangeli,L.,Palumbo,P.,etal.1996,ApJ,464,L191 4πκ(ν)
Molster,F.J.&Waters,L.B.F.M.2003,inLectureNotesinPhysics,Berlin α(ν)= , (A.2)
SpringerVerlag,Vol.609,Astromineralogy,ed.T.K.Henning,121–170 ν
Moulin,F.,Devel,M.,&Picaud,S.2008,J.Quant.Spec.Radiat.Transf.,109,
1791
which gives a first iteration of the κ(ν) index. From the Snell-
Papoular,R.,Conard,J.,Guillois,O.,etal.1996,A&A,315,222
Peeters,E.,Hony,S.,VanKerckhoven,C.,etal.2002,A&A,390,1089 Decarteslaw,weknowthattheopticalpathdifferenceδbetween
Pendleton,Y.J.&Allamandola,L.J.2002,ApJS,138,75 abeamtraversingthefilmandbeingreflectedovereachfaceof
Pino,T.,Dartois,E.,Cao,A.-T.,etal.2008,A&A,490,665 thefilmis
Rawlings,M.G.,Adamson,A.J.,&Whittet,D.C.B.2003,MNRAS,341,1121
Robertson,J.&O’Reilly,E.P.1987,Phys.Rev.B,35,2946 δ=2n dcos(θ ). (A.3)
Sandford,S.A.,Pendleton,Y.J.,&Allamandola,L.J.1995,ApJ,440,697 0 f
Schnaiter, M., Mutschke, H., Dorschner, J., Henning, T., & Salama, F. 1998,
ApJ,498,486 Theinterferencefringesofthefilmwillhaveamaximumwhen
Schnaiter,M.,Mutschke,H.,Henning,T.,etal.1996,ApJ,464,L187
Scott,A.D.,Duley,W.W.,&Jahani,H.R.1997,ApJ,490,L175
Sloan,G.C.,Jura,M.,Duley,W.W.,etal.2007,ApJ,664,1144 m+1
Smith,F.W.1984,JournalofAppliedPhysics,55,764 δ=mλ= , (A.4)
Stecher,T.P.&Donn,B.1965,ApJ,142,1681 ∆σ
Steglich,M.,Ja¨ger,C.,Huisken,F.,etal.2013,ApJS,208,26
Steglich,M.,Ja¨ger,C.,Rouille´,G.,etal.2010,ApJ,712,L16
where ∆σ is the interfringe in wave numbers. Combining the
Swanepoel,R.1983,JournalofPhysicsEScientificInstruments,16,1214
Tielens,A.G.G.M.2008,ARA&A,46,289 previoustwoequations,wefindthatthethicknessofthefilmis
Tosatti,E.&Bassani,F.1970,NuovoCimentoBSerie,65,161 givenby
Trotta,F.&Schmitt,B.1996,inNATOAdvancedStudyInstituteontheCosmic
DustConnection,3rdCourseoftheInternationalSchoolofSpaceChemistry, 1
p.179-184,179–184 d= 2n ∆σ, (A.5)
vanDiedenhoven,B.,Peeters,E.,VanKerckhoven,C.,etal.2004,ApJ,611,928 0
Warren,S.G.1984,Appl.Opt.,23,1206
Yang,X.J.,Glaser,R.,Li,A.,&Zhong,J.X.2013,ApJ,776,110 wheren isassumedtobewavelengthinvariant,whichisreason-
0
Zubko,V.G.,Mennella,V.,Colangeli,L.,&Bussoletti,E.1996,MNRAS,282, ableforvisibleandnear-infraredwavelengths.Measurementsin
1321 themid-infrared(from400to8000cm−1)givealimitforthede-
tectablehalfinterfringeof∼200nm.Thismeansthattheinfrared
interfringeofultra-thinfilmsisnotmeasurable.Weusedthecor-
relation between the integrated 3.4 µm band and n*thickness
AppendixA: Thicknessestimationofana-C:Hfilm
(from the interfringe measurements of the thicker films, from
Theopticaldepth,τ,obtainedfromtransmissionspectra,isequal Eq.A.5)toestimatethethicknessfortheultra-thinfilms.Figure
to A.1 shows this correlation fitted by a linear function y = m*x,
τ=αd, (A.1) wherem=0.059±0.01,alongwith3σerrorbars.
9
L.Gavilanetal.:VUVspectroscopyofcarbondustanalogs:contributiontointerstellarextinction
AppendixB: SubstratesintheVUV
FigureB.1:Transmissionspectraofdifferentcarbonmembranes
(10-25nmthick)usedforVUVmeasurements(dashedlines),
andcarbonmembranes+a-C:Horsootdeposits(solidlines).
Tostudytheinfluenceofthesubstrateonthetransmissionspec-
tra of a-C:H and soot films in the VUV, we studied the spec-
traldispersionofthesubstratesandthefilmsdepositedonsuch
substrates, seen in Fig. B.1. These spectra are divided by the
scaledspectralirradianceoftheVUVbeam.Spectraofsubstrate
C1 (thickness = 15 - 25nm) show very low dispersion(< 0.01
∆τ)forconsecutivemeasurements(C1a)andlowdispersion(<
0.1 ∆τ) between two series of measurements. Spectra of refer-
ence C2 (thickness ∼10 nm) show low dispersion (< 0.2 ∆τ)
betweenconsecutivemeasurements.Thefilms(a-C:Horsoots)
depositedonC1orC2showtheexpectedincreasedopticaldepth
with characteristic peak positions different from the substrate.
Divisionbytherespectivesubstrategivesusthespectraintrans-
missionforsuchfilms,asshowninFigs.3-4.
10