The well correlated DIBs at λλ 6196, 6614 Å and their possible companions
aa r X i v : . [ a s t r o - ph . GA ] J un MNRAS , 1–11 (2020) Preprint 5 June 2020 Compiled using MNRAS L A TEX style file v3.0
The well correlated DIBs at λλ A. Bondar ⋆ International Center for Astronomical and Medico-Ecological Research, Zabolotnoho Str. 27, Kyiv, Ukraine
Last updated 2020 May 29; in original form 2019 Dec 26
ABSTRACT
Considered here is the interrelation between five diffuse interstellar bands (DIBs): λλ λ λ λ/δλ ≈ λλ µ m. Key words:
ISM: molecules, lines and bands
During nearly a century astrophysicists have been searchingfor a key to understanding of the origin of diffuse interstellarbands. A great deal of works dedicated to the DIBs prob-lem directly concerns the subject of interrelation betweenDIBs as well as of the nature of their carriers. So, Herbig(1975) considered 39 diffuse bands in spectra of 56 hot starsof diverse spectral types (from O6 till A5). 17 DIBs showedquite a good correlation between their strengths — evenbetter than either with color excess. He did not find anyregularities in the wavenumbers of the analysed DIBs thatmight be considered, e. g. as a series of vibrational transi-tions. As DIBs carrier the author suggested a constituent ofvery small interstellar grains ≈
300 ˚A in radius.Chlewicki et al. (1986) investigated several DIBs inspectra of stars from three associations. A special atten-tion was given to the shapes of profiles and correlationsbetween, so-called, “major” bands λλ ⋆ E-mail: [email protected]
DIBs, such as constancy of their central wavelengths, differ-ences in shapes of profiles amid strongly correlated bands,existence of narrow diffuse bands and correlations betweennarrow and broad DIBs, that suggests close relation of theircarriers. But as the hypothesis excludes appearance of thenarrow DIBs so solid particles should not also be responsiblefor existence of the broad ones.Kre lowski & Walker (1987) analysing intensities andcentral depths of some diffuse bands in spectra of reddenedearly B stars (E B − V = 0.16 – 0.42) detected three proba-ble families of DIBs, most evident in the spectrum of ζ Per.The authors supposed existence of three different carriersfor those DIBs.Concept of DIB families got its continuation in the pa-per by Kre lowski & Sneden (1995) and in quite recent works,for instance, by Friedman et al. (2011); Vos et al. (2011);Kre lowski, Galazutdinov & Bondar (2019), where spectra ofselected objects were classified accordingly to intensity ra-tios of two strong DIBs λλ σ ’ and‘ ζ ’ interstellar cloud types — called after sightlines towards σ Sco and ζ Oph.But Cami et al. (1997) besides ‘ σ ’ and ‘ ζ ’ cloud typesconsidered CS (circumstellar) and the Orion type clouds.Their classification of DIB families is based on differences inionization potentials (an ionization hypothesis) of DIB car-riers — as probably large carbon-bearing molecules. In theiropinion DIBs that correlate well with each other have carri-ers with similar ionization potentials and vise verse: poorly c (cid:13) A. Bondar correlated DIBs should have carriers with significant differ-ences in ionization and recombination properties. On thewhole the authors support molecular hypothesis of DIB car-riers as large carbonaceous molecules and explain existenceof ‘ ζ ’, ‘ σ ’ and Orion clouds as a sequence in intensity of localUV-field: from low in ‘ ζ ’ through strong in ‘ σ ’ to extremeUV radiation in Orion cloud types.Herbig (1988) described six weak features nearly uni-formly spaced at 35 cm − between 6779 – 6860 ˚A. He specu-lated about rotational or vibrational progressions with spac-ings at 35 cm − level, but did not find any plausible carrierfor this group of bands.Herbig & Leka (1991) explored the systematics in spac-ings of more than hundred of DIBs. Assuming that in inter-stellar medium only the lowest energy levels (no excitation)of possible complex molecules are involved into DIBs pro-duction, they sought for vibrational sequences of the type0 → v ′ . As a result, from 50 sequences of four or moreDIBs, found by their method, eight were considered as re-liable: thus some of the bands in these sequences could berelated to one common carrier. The authors had taken intoaccount quite a big number of known bands and concludedthat DIBs should be produced by a mixture of different car-riers.Moutou et al. (1999) tested 10 DIBs in spectra of 62stars and found good correlations between several of them:6614 ↔ ↔ ↔ ↔ ↔ ↔ ↔ λλ λλ λλ ≈ <
10 m˚A. Though Kre lowski et al. (2016) using spectra ofmuch higher resolution confirmed tight correlation betweenthese two diffuse bands, as it was presented in McCall et al.(2010).In this paper the author presents a new study of inter-relation between the well-correlated DIBs λλ λλ λ The selected objects (54 in total) cover wide range of spec-tral classes from O5 to B9.5 with several Wolf-Rayet stars.Many of them are members of well-known stellar associ-ations, such as Per OB2, Sgr OB1, Sco OB2, ρ Oph cloudand others, so they present regions of interstellar mediumwith variant physical conditions. These stars have redden-ing from moderate to strong: E B − V = 0.12 – 1.45 mag. Themain characteristics of the selected targets are presented inthe Table 1. Intrinsic values of (B − V) for calculating E B − V in the 5th column of the Table 1 were taken from work byPapaj, Krelowski &Wegner (1993).One of the selection criteria for the objects in the Ta-ble 1 is the absence of a splitting in interstellar atomic ormolecular lines visible in their spectra, i. e. a single cloud onthe sightline towards a star. Or at least, if such a splittingexists, a secondary component should have much lesser in-tensity in comparison with a main component, the strongestone. This criterion was not strictly kept as it is quite a diffi-cult task in case of high resolution spectra. The 1st columnof the Table 1 clearly demonstrates that 25% of sightlines(marked with ‘ • ’ sign) have more than one interstellar cloud.All the other objects in the Table 1 either do not have anysigns of splitting in the potassium K i line or can have slightasymmetries in this line. A special case is the star HD 169454as its spectrum reveals two evident interstellar clouds to-wards it.As a rule verification was done on the interstellar line ofpotassium K i at λ λ i at λ The used sample of spectral data was collected in differentyears with three spectrographs: mostly HARPS (2007 yr.,program ID 078.C-0403(A)) and UVES (2009 and 2015 yrs.,program IDs 082.C-0566(A), 092.C-0019(A)), see respectiveacronyms (‘H’, ‘U’) in the last column of the Table 1. Also,spectra obtained with HARPS and UVES spectrographsfrom ESO science archive have been used. They include theraw spectra as well as the reduced “Phase3” ones. Spec-tra of five objects were obtained with BOES spectrograph(Kim et al. 2007), letter ‘B’ in the last column of the Table 1.All the considered spectra are of a high resolution, λ/δλ ≈ ≈
260 to 2100. Minimum and maximumS/N (with noted spectrum of the star) and S/N mean valuesfor all the spectra in the data set are next: • DIB 5545: (HD 147889) – (HD 159176), < > ; • DIB 6113: (HD 144470) – (HD 142468), < > ; • DIB 6196: (15 Sgr) – ( σ Sco), < > ; • DIB 6445: (HD 185859) – (HD 142468), < > ; • DIB 6614: (HD 203532) – (HD 63804), < > .MNRAS , 1–11 (2020) orrelations between DIBs Table 1.
Main characteristics of the selected objects: visual magnitude (V), spectral type/luminosity (Sp/L), reddening (E B − V ), spec-trograph used (Inst.). Name a Type b V, mag Sp/L E B − V , mag Inst. c HD 14956 s ∗ b 7.20 B1.5Ia 0.89 BHD 23016 Be ∗ ∗∗ ∗ • ’ d ∗ ∗ • ’ ∗ ∗ b 9.19 B3Ia 0.99 UHD 63804 s ∗ b 7.78 B9.5Ia+(e) 1.05 UHD 73882 EB ∗ ∗ ∗ ∗ b 9.00 B2Ia+e 1.45 UHD 91824‘ • ’ SB ∗ • ’ SB ∗ • ’ SB ∗ ∗ ∗ ∗ ∗ ∗ ∗∗ ∗ ∗∗ ∗ ∗ ∗ • ’ s ∗ b 5.39 B1Iaeqp 0.52 Uzet Oph Be ∗ • ’ WR ∗ ∗ b 7.08 O9.7IabNwk 0.61 UHD 152233‘ • ’ SB ∗ • ’ s ∗ b 6.38 B0.5Ia 0.74 UHD 152249 s ∗ b 6.45 OC9Iab 0.42 UHD 152270 WR ∗ • ’ s ∗ b 4.87 B2Iab 0.45 HHD 154445 ∗ ∗ ∗ • ’ ∗ ∗ ∗ ∗ ∗ b 5.37 O9.7Iab 0.29 UHD 167771 SB ∗ • ’ s ∗ b 6.71 B1Ia 1.10 UHD 170740 SB ∗ ∗ ∗ • ’ ∗ ∗ • ’ WR ∗ ∗ ∗ b 8.67 B0.5Ia 1.34 BNotes. a HD number, Bayer or other designation. b Short type of an object (acronyms adopted from SIMBAD): ∗ — star,s ∗ b — blue supergiant star, ∗∗ — double star, SB ∗ — spectral binary, V ∗ —variable star, Or ∗ — Variable Star of Orion Type, bC ∗ — Variable Starof beta Cep type, Be ∗ — Be-type star, WR ∗ — Wolf-Rayet star, pr ∗ —pre-main sequence star, Em ∗ — emission-line star, EB ∗ — eclipsing binary. c ‘B’ — BOES, ‘H’ — HARPS, ‘U’ — UVES. d ’‘ • ” sightlines with strong asymmetry or signs of splitting in interstellaratomic K i line at λ , 1–11 (2020) A. Bondar
So, the spectral intervals utilized in this work have highmean S/N close to 700.
DIBs for the current study were selected on sugges-tion that they preserve their intensity ratios in thesame manner as DIBs 6196 and 6614 ˚A while chang-ing sightlines in the Galaxy:
EW(DIB) /EW(DIB) ≈ EW(6196) /EW(6196) ≈ EW(6614) /EW(6614) , where in-dices 1, 2 stand for two different sightlines. The intensityratios were calculated by division intensities of DIBs inspectrum of an arbitrary star by those in the spectrum ofHD 163800. The spectrum of this object was chosen as ref-erence for quite strong DIBs in it and because it is a simplestar: without signs of binarity, variability and so on, see theTable 1.At first EWs for ten stars were taken from work byBondar (2012). Later on the set of objects was widened to 54stars. Initially number of the selected DIBs reached 35, butafter a thorough verification most of them were rejected asimplausible and also due to big uncertainties in their EWs,their weakness or too noisy spectra. Thus only three bandswhich demonstrated variations in equivalent widths in agree-ment with those in DIBs λλ λλ Raw spectra were processed by the author eitherwith IRAF (Tody 1986) or ESO UVES pipeline: , “esorex”and “gasgano” programs. A set of reduced spectra fromHARPS (2007) and UVES (2015) spectrographs was kindlygiven by Prof. J. Kre lowski.
All the one-dimensional spectra were analysed with IRAFspectroscopic package tasks. A special procedure for mea-surement of equivalent widths with evaluation of uncertain-ties by method of Vollmann & Eversberg (2006) was createdon the basis of the IRAF “splot” task. Equivalent widthshave been measured by direct integration of DIB profiles.Typical profiles of the λλ σ Sco (the strongestand the weakest cases) or appeared structure in the profileof DIB 6196 ˚A in the spectrum of HD 147889. Profiles of theweak DIBs in spectra with lower S/N can also be signifi-cantly distorted. As to similarity of profiles for DIBs with common originthey do not match well. Though profiles of DIBs λλ α Vir) was used. But in many cases it has been realizedwith “Molecfit” package developed by Smette et al. (2015).It was quite effective, see e. g. mid picture at the right panelof the Fig. 1.Measurements of the equivalent widths with their1 σ uncertainties for the five diffuse bands are given inthe Table 2. The uncertainties were calculated for purephoton noise statistics by technique used in work byVollmann & Eversberg (2006). It remains indefinite howmuch such factors as continuum placement, peculiaritiesin profiles of diffuse bands themselves (wings, asymmetry);stellar peculiarities — binarity, variability (see the Table 1)which provoke blends with DIBs and also telluric lines rem-nants, in case of DIB 6445 ˚A, can contribute to the errorsin equivalent widths.Correct S/N determination can also present some diffi-culties. DIBs 6113, 6445 and especially 5545 ˚A have closelylocated neighbours and thus too short pieces of free con-tinuum to measure S/N. So, if those pieces of spectra wereinappropriate for measurements, S/N was to be measuredsomewhere in other place of the spectral order. Interrelations between the DIBs have been sought byweighted least squares fitting of measured equivalent widths.All the fits were done by method described in work by Orear(1982) — case when errors in two variables were allowed for.Parameters of the fits a, b as well as those for the invertedfits ( a ′ , b ′ ) are collected in the Table 3. The direct fits aredepicted at the Fig. 2. In all plots presented there equiva-lent widths of stronger DIB were adopted as arguments (X).Respectively, to every pair of the DIBs a weighted linear cor-relation coefficient, R, has been calculated. The values of R,the number degree of freedom of the fit (NDF) and reduced χ are shown at the upper left corners on the plots at theFig. 2.As it is seen from the figure, the highest value of R =0.988 was obtained for the pair: DIB 6196 vs DIB 6614 ˚A.The correlation coefficient R, parameters b = 0 . ± . b ′ = 3 . ± .
09 determined from the direct and inversefits between EW(6196) and EW(6614) (see the Table 3) —are in a good agreement with those from the earlier works(see McCall et al. 2010; Kre lowski et al. 2016). It concernsespecially the former one, where detailed statistical analysiswas done with data set twice larger than in the present work.Other pairs of the diffuse bands: 6445 vs 6614 ˚A, 6113vs 5545 ˚A and 6113 vs 6614 ˚A also show quite a high degreeof correlation R = 0.982 – 0.984, see the Fig. 2. Correlationsbetween other bands remain at nearly equal and still high
MNRAS , 1–11 (2020) orrelations between DIBs Wavelength, Å R e l a t i v e i n t en s i t y HD 147165HD 142468HD 185859HD 80077HD 63804
DIB 5545
Wavelength, Å R ea l a t i v e I n t en s i t y HD 147165HD 163800HD 14956HD 63804LS 719
DIB 6113
Wavelength, Å R e l a t i v e I n t en s i t y HD91824HD148688HD185418HD78344LS 719
DIB 6445
Wavelength, Å R e l a t i v e I n t en s i t y telluric absorption corrected spectrumcorrection function for telluric absorptionoriginal spectrum (HD 73882) Wavelength, Å R e l a t i v e I n t en s i t y HD 23016HD 147889HD185859HD78344LS 719
DIB 6196
Wavelength, Å R e l a t i v e I n t en s i t y HD 23016HD 151932HD 169454HD 78344LS 719
DIB 6614
Figure 1.
The selected DIBs of different intensities with characteristic profiles. The insets below right depict comparison between theweakest and strongest normalized profiles of the DIB. The mid picture at the right panel demonstrates telluric lines removing procedurefor DIB at 6445 ˚A. level R > < EW(6196)/EW(6614) ≈ ≈ ≈ .In vast majority of cases the most deviant points atthe Fig. 2 could be explained due to blends with stellarlines. So, in accordance with the synthetic spectra gener-ated with above mentioned programs, ATLAS9 and SYN-THE, DIB 5545 ˚A in the spectrum of B9.5Ia star HD 63804(see the plots on the Fig. 2) looks overestimated probablydue to the blend with the stellar Fe ii line. And further: DIB6113 ˚A in the spectra of HD 185418 (B0.5V), HD 152270 (WC7+O5-8), LS 719 (B3Ia) and HD 80077 (B2Ia) proba-bly has the blend with Ne ii or with Fe ii in the spectrumof HD 23016 (B9Vne). DIB 6445 ˚A can be blended withthe stellar Fe ii line in the spectra of V2052 Oph (B3III),HD 147889 (B2III/IV) or with N iii lines in the spectra ofHD 110432 (B0.5IVpe), HD 142468 (B1Ia/Iab), HD 185859(B0.5Iae), HD 228712 (B0.5Ia). As for DIB 6196 ˚A, e. g inthe spectra of HD 37367 (B2IV-V), HD 142468 (B1Ia/Iab)or other stars with the deviant points at the Fig. 2, it is dif-ficult to say what could increase the DIB’s equivalent widthas the stellar Fe iii , near λ MNRAS , 1–11 (2020)
A. Bondar
EW(5545), mA E W ( ) , m A R= 0.968NDF: 49reduced χ : DIB 6445 vs DIB 5545
HD 63804HD 228712
EW(6196), mA E W ( ) , m A DIB 5545 vs DIB 6196
R = 0.973NDF: 49reduced χ : HD 142468 HD 63804LS 719HD 37367HD 148688
EW(6196), mA E W ( ) , m A DIB 6113 vs DIB 6196
R = 0.974NDF: 50reduced χ : HD 142468 LS 419HD 161056HD 23016 HD 37367HD 193793
EW(6445), mA E W ( ) , m A DIB 6113 vs DIB 6445
R = 0.974NDF: 49reduced χ : HD 80077 LS 719HD 152270 HD 147889HD 142468V2052 Oph
EW(6614), mA E W ( ) , m A DIB 5545 vs DIB 6614
R = 0.978NDF: 50reduced χ : HD 185418HD 228712 HD 63804
EW(6196), mA E W ( ) , m A DIB 6445 vs DIB 6196
R = 0.978
NDF: 51reduced χ : LS 719HD 147889HD 185859HD 110432 HD 148688HD 152249
EW(6614), mA E W ( ) , m A DIB 6113 vs DIB 6614
R = 0.982NDF: 50reduced χ : HD 80077HD 152270 HD 170740
EW(5545), mA E W ( ) , m A DIB 6113 vs DIB 5545
HD 63804
R = 0.984NDF: 48reduced χ : HD 80077HD 152270HD 185418
EW(6614), mA E W ( ) , m A DIB 6445 vs DIB 6614
R = 0.984NDF: 51reduced χ : HD 110432 HD 152249
EW(6614), mÅ E W ( ) , m Å DIB 6196 vs DIB 6614
R= 0.988 NDF: 52reduced χ : Figure 2.
Weighted fits of the selected DIBs, ordered on increasing linear correlation coefficient R. NDF is short for “Number Degreeof Freedom”. In most cases the deviant points likely are caused by blends of the DIBs with stellar lines. Low values of χ < , 1–11 (2020) orrelations between DIBs Table 2.
Equivalent widths with their uncertainties (m˚A) of the two tightly correlated DIBs at λλ ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± · · · ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± · · · ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± · · · ± ± ± ± ± ± ± ± ± ± ± ± · · · ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± · · · ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± ± there the meaning of equivalent width is much lower of thatpredicted by fit.So, in some cases neither too big nor too small valuesof equivalent widths could be explained. If these effects havenot been caused by cosmic ray hits or by defects in spectrathen specific local conditions of ISM might be responsible. It is quite probable that such deviant points could be caused byvariations in the local physical conditions, especially in theintensity of radiation field and in the interstellar density,which can influence DIB equivalent widths and shapes oftheir profiles. This effect can be seen, for instance, in spectraof stars in Upper Scorpius (Vos et al. 2011) and in spectra MNRAS , 1–11 (2020)
A. Bondar
Table 3.
Parameters a, b, a ′ , b ′ obtained from direct and in-verted least squares linear fits of intensities of the selected DIBs.Direct fits are presented on the plots at the Fig. 2.Pair of DIBs a, b a ′ , b ′ − . ± .
92 7 . ± . . ± .
02 1 . ± . − . ± .
81 30 . ± . . ± .
01 5 . ± . . ± . − . ± . . ± .
01 1 . ± . − . ± .
55 6 . ± . . ± .
02 2 . ± . − . ± .
55 2 . ± . . ± .
03 1 . ± . − . ± .
45 28 . ± . . ± .
00 8 . ± . . ± . − . ± . . ± .
01 3 . ± . . ± . − . ± . . ± .
03 1 . ± . − . ± .
56 3 . ± . . ± .
02 1 . ± . . ± .
47 15 . ± . . ± .
00 7 . ± . of Local Bubble objects (Bailey et al. 2016; Farhang et al.2019) or in the spectrum of Her 36 where DIBs have ex-tended red wings (Dahlstrom et al. 2013; Oka et al. 2013).In many cases weakness of DIBs may be really explainedby diminishing of DIB carriers in interstellar medium be-cause of proximity of the hot stars, as it is in Orion Nebula.From the other hand DIBs can be more intense or of nor-mal strengths under nearly the same physical conditions (seeHerbig 1995). The DIBs explored here make a set of wavenumber dif-ferences or spacings, ∆˜ ν , which cover the range 218 –2913 cm − , see the 2nd column of the Table 4. These spac-ings pretty well match to the energies of vibrational transi-tions in some molecules including polyatomic ones, such asPAHs, which are quite abundant in ISM (see L´eger & Puget1984). Such transitions, often ascribed to vibrations inPAHs, are observable as IR emission bands, named ear-lier as Unidentified Infra-Red (UIR) emission features, see,e. g. review by Mathis (1990) and references therein. Sev-eral IR bands at λλ µ m arepresented in the 3rd column of the Table 4. The IR bandat 16.4 µ m is described in works by (Boulanger et al. 1998;Tielens et al. 1999; Moutou et al. 2000) about four otherssee, e. g. (L´eger & Puget 1984; Mathis 1990). In additionthere is one band, at 5.3 µ m, visible in calculated spectra ofPAHs (see L´eger et al. 1989).Of all spacings collected in the second column of theTable 4 six — with ∆˜ ν = 624.1, 842.5, 1237.4, 1675.8,1894.3 and 2913.2 cm − — demonstrate a relation withthe wavenumbers of IR bands. Thus spacings 842.5, 1237.4,1675.8 and 2913.2 cm − are lesser/greater than their coun-terparts from IR bands by nearly equal factors (1.04, 1.05),see the 4th column of the Table 4. These IR emissions is now Table 4.
Spacings between the DIBs in comparison withwavenumbers of some IR bands ascribed to vibrations in somePAHs.DIBs pair ∆˜ ν a , IR band b , Ratio c cm − ˜ ν , cm − ( λ , µ m)6196 ↔ · · · · · · ↔ · · · · · · ↔ ↔ ↔ · · · · · · ↔ ↔ ↔ d ↔ · · · · · · ↔ a — spacings between DIBs. b — wavenumbers and wavelengths of vibronic infraredbands ascribed to PAHs. c — the ratio spacing/wavenumber of IR band (∆˜ ν /˜ ν ). d — the band observed in calculated spectra for PAHs,see L´eger et al. (1989). regarded as indicatives of aromatic hydrocarbons due to CHout-of-plane bend vibration at λ µ m (˜ ν = 884.8 cm − ),C — C at λλ µ m (1612.9 cm − ), 7.7 µ m (1298.4 cm − )and CH at λ µ m (3029.7 cm − ) stretch vibrations, see,for instance, review by Mathis (1990). The spacing ∆˜ ν =624.1 cm − matches (it is greater by factor 1.02) to theband at λ µ m (609.8 cm − ) which a PAH producesdue to vibrations (see Moutou et al. 2000). This IR emissionhas been known from laboratory experiments (Moutou et al.1996) as 16.2 µ m mode in PAH and was first mentioned inBoulanger et al. (1998) as emission observable at λ µ min spectra of NGC 7023 and ρ Oph molecular cloud. Thespacing ∆˜ ν = 1894.3 cm − nearly exactly corresponds toIR band at 5.3 µ m (1886.8 cm − ) which was observed incalculated spectra of PAHs and recommended for search inastronomical objects by the authors (see L´eger et al. 1989).So, four pairs: ∆˜ ν vs ˜ ν from the Table 4 show the same ≈
5% and 4% differences; one pair differs by 2% and one by ≈ µ m can be produced in the wavelengthrange 16.21 – 16.52 µ m (˜ ν = 623 – 611 cm − ), thus vary-ing in its position ≈ by 2% (Moutou et al. 2000). The bandat 11.3 µ m, caused by out-of-plane C — H bending occursat λλ µ m (˜ ν = 862 – 800 cm − ) if there twoC — H bonds on the same aromatic ring, and at λλ µ m (807 – 752 cm − ) for three bonds (see Mathis (1990)and references therein). Tielens et al. (1999) demonstratedIR bands in the Red Rectangle with λλ µ m. According to the Table 4 the band at 11.9 µ m(840.3 cm − ) quite well corresponds to the spacing ∆˜ ν =842.5 cm − . MNRAS , 1–11 (2020) orrelations between DIBs Figure 3.
Quasi energy-level diagram. Energy of the upper level above the ground state in 10 cm − . Electronic transition X → A causesvibrational-rotational transitions from v ′′ = 0 to arbitrary upper levels, that produces the system of DIBs shown below. Rotational bandsare invisible at this scale and probably reveal themselves as substructures in the DIB profiles. Likely transitions from upper to lowervibrational states with ensuing IR emissions (IRE) due to vibrations in some PAHs are marked by numbers in µ m, see the Table 4.Radiationless transitions from upper to lower v ′ levels (vibrational relaxation) are also possible. Wavelengths of IR bands also depend on type of theircarriers: whether they are in aromatic or in aliphatic forms.In PAHs C — H stretching gives a band at λ µ m.Aliphatic hydrocarbons have a band of weaker intensity at λ µ m (2941.2 cm − ) which even better matches to thespacing ∆˜ ν = 2913.2 cm − , from the Table 4. According tothe recent studies (see e. g. Li & Draine 2012) the fractionof aliphatic constituent in IR bands carriers is <
15% —thus they are mostly aromatic. As to the possible carriers ofdiffuse bands, it is improbably that this group of five DIBswas due to both aliphatic and aromatic hydrocarbons.In any case a found proximity of spacings between theDIBs to wavenumbers of several IR bands may point to com-mon origin of their carriers: be they PAHs or any other com-plex molecules.If to suppose that the five diffuse bands analysed inthe current study present a fragment of spectrum of a com-plex molecule — e. g. an electronic transition with rotation-vibrational series — then DIB profiles could be consideredas packed, unresolved rotational structure and spacings be-tween the DIBs as differences in energies of vibrationalstates. Such a series perhaps might look as it is depictedat the Fig. 3. Electronic transition from the ground level toan upper state X → A (e. g. after absorption an energeticphoton) can give rise to a series of vibrational transitionsfrom the ground vibrational level v ′′ = 0 to arbitrary upperlevels v ′ = i, j, k, l, m etc. with a set of rotational transitions.Thus, the DIBs at corresponding wavelengths may appear. Later on due to different intramolecular processes the systemcan go to lower energy states in different ways. One of themvibrational relaxation — energy transfer from upper vibra-tional levels to different modes of vibrations (radiationlesstransition without changing electronic state), another — vi-brational de-excitation through IR emissions. Several likelytransitions between upper ( v ′ ) vibrational levels which couldlead to IR emissions are shown at the Fig. 3, see also the Ta-ble 4. Rotational P, Q, R branch structure may reveal itselfas substructure in DIB profiles, see the Fig. 1. Especially itis evident in DIB 6614 ˚A (see Sarre et al. 1995; Cami et al.2004). Sometimes a similar substructure (as well as the pro-file broadening) is developed in DIBs when their carriers arelocated in regions with higher temperature, determined e. g.on excitation temperature of C , as it was shown for DIB6196 ˚A in the study by Ka´zmierczak et al. (2009). On theFig. 1 the substructure is evident in the profile of DIB 6196 ˚Ain the spectrum of HD 147889. Conspicuous substructure isalso observable in the profile of the band at 5545 ˚A.If the five DIBs appear as a result of above mentionedfactors in such a case the intensity ratios of the DIBs,noted in subsection 3.3 — EW(6196)/EW(6614) ≈ ≈ ≈ MNRAS , 1–11 (2020) A. Bondar the found spacings. Moreover in complex molecules vibronicprogressions do not necessarily follow simple equation forenergies in function of vibrational quantum number ‘ v ’ forthe given fundamental frequency ‘ w e ’ and inharmonicity ‘ x e ’constants, which depend on electronic state of the molecule.So, finding a regularity in such a case may present a difficulttask.If to look into the whole DIBs’ spectrum there are manypairs of DIBs (not so strong) with values of spacings close tothe presented in the Table 4. Those with the best coincidence ∼ < cm − are next: 6600 ↔ cm − ); 6486 ↔ cm − ); 6768 ↔ cm − ); 6325 ↔ cm − ); 6467 ↔ cm − ). Many othersdo not match so well, but here, of course, the precision ofDIB wavelengths should be taken into account. CONCLUSIONS
High linear correlation coefficients R = 0.968 – 0.988 couldindicate close interrelation of the five considered bands. Cor-relation and intensities ratio found for two diffuse bands at λ λ > λ < χ could point to errors underestimation and alsoresult from big uncertainties in EW measurements of theweakest DIBs in the set: 6113 and 6445 ˚A.Do these DIBs have a common origin despite that factthe forms of their profiles are not alike? There is some sim-ilarity in the profiles of DIBs at λλ . Five spacings have values close to the wavenumbersof IR emission bands which were suggested to arise due tovibrations in PAHs: at λλ µ m;in addition there is one spacing (∆˜ ν = 1894.3 cm − ) thatcorresponds to the IR band at 5.3 µ m (1886.8 cm − ), visiblein calculated spectra of PAHs (see L´eger et al. 1989).There is not any evident regularity in the spacings amidthe considered DIBs. But a proximity of the spacings to thewavenumbers of the IR bands may be a sign of a close re-lation of the DIBs’ carriers to PAHs or even that the PAHsare these carriers themselves. It seems that DIBs may ariseas a result of excitation of PAHs. IR bands could appear dueto intramolecular processes such as de-excitation of vibra-tional levels in PAHs, see the Fig. 3. Perhaps, to shed lighton this issue it makes sense to search and explore the objectsin whose spectra both IR bands and DIBs are observable.Found ratios of equivalent widths have values:EW(6196)/EW(6614) ≈ ≈ ≈ ACKNOWLEDGEMENTS
Based on observations made with ESO Telescopes underprogrammes ID: 67.D-0439(A), 194.C-0833(A,B,D,F),382.D-0237, 266.D-5655(A), 67.C-0281(A), 097.C-0979(A),60.A-9036(A), 076.B-0055(A), 0102.C-0040(B), 080.D-0855(A), 083.C-0503(A), 078.C-0403(A), 187.D-0917(A).The author is very grateful to the referee for his relevantand useful comments and suggestions.
REFERENCES
Bailey M., van Loon J. T., Farhang A., Javadi A., KhosroshahiH. G., Sarre P. J., Smith K. T., 2016, A&A, 585, A12Bondar A., 2012, MNRAS, 423, 725Boulanger F., Boisssel P., Cesarsky D., Ryter C., 1998, A&A, 339,194Cami J., Sonnentrucker P., Ehrenfreund P., Foing B. H., 1997,A&A, 326, 822Cami J., Salama F., Jim´enez-Vicente J., Galazutdinov G. A.,Kre lowski J., 2004, ApJL, 611, L113Chlewicki G., van der Zwet G. P., Ijzendoorn L. J., Greenberg J.M., Alvarez P. P., 1986, ApJ, 305, 455Chlewicki G., de Groot M. S., van der Zwet G. P., Greenberg J.M., Alvarez P. P., Mampaso A., 1987, A&A, 173, 131Dahlstrom J., York, D. G., Welty, D. E., et al., 2013, ApJ, 773,41Farhang A., van Loon J. T., Khosroshahi H. G., Javadi A., BaileyM., 2019, NatAs, 3, 922Friedman S. D., et al., 2011, ApJ, 727, 33Herbig G. H., 1975, ApJ, 196, 129Herbig G. H., 1988, ApJ, 331, 999Herbig G. H., 1995, ARA&A, 33, 19Herbig G. H., Leka K. D., 1991, ApJ, 382, 193Ka´zmierczak M., Gnaci´nski P., Schmidt M. R., Galazutdinov G.,Bondar A., Kre lowski J., 2009, A&A, 498, 785Kim, K.-M., et al., 2007, PASP, 119, 1052Kre lowski J., Walker G. A. H., 1987, ApJ, 312, 860Kre lowski, J., Sneden, C., 1995, “Diffuse interstellar bands in in-dividual sightline”, edited by A. G. G. M. Tielens and T. P.Snow, Vol. 202, 13–24, “Kluwer Academic Publishers”, Dor-drechtKre lowski J., Galazutdinov G. A., Bondar A., Beletsky Y., 2016,MNRAS, 460, 2706Kre lowski, J., Galazutdinov, G., & Bondar, A., 2019, MNRAS,486, 3537Kurucz, R. L., 1970, SAOSR, 309.Kurucz, R.L., 1992, Proc. of the 149th Symposium of the IAU,225.L´eger A., Puget J. L., 1984, A&A, 500, 279L´eger A., D’Hendecourt L., 1985, A&A, 146, 81L´eger A., D’Hendecourt L. & D´efourneau D., 1989, A&A, 216,148Li A., Draine B. T., 2012, ApJL, 760, L35Mathis John S., 1990, ARA&A, 28, 37McCall B. J., et al., 2010, ApJ, 708, 1628MNRAS , 1–11 (2020) orrelations between DIBs Moutou C., L´eger A., D’Hendecourt L., 1996, A&A, 310, 297Moutou C., Kre lowski J., D’Hendecourt L., Jamroszczak J., 1999,A&A, 351, 680Moutou C.,Verstraete L., L´eger A., Sellgren K., Schmidt W.,2000, A&A, 354, 17Oka T., Welty D. E., Johnson S., York D. G., Dahlstrom J., HobbsL. M., 2013, ApJ, 773, 42J. Orear, ”Least Squares When Both Variables have Uncertain-ties”, Amer. J. Phys., 1982, 50, 912Papaj, J., Krelowski, J. & Wegner, W. 1993, A&A, 273, 575Sarre P. J., Miles J. R., Kerr T. H., Hibbins R. E., Fossey S. J.,Somerville W. B., 1995, MNRAS, 277, 41Sbordone, L., Bonifacio, P., Castelli, F., Kurucz, R. L., 2004,Mem. Soc. Astron. Italiana, 5, 93.Smette, A., Sana, H., Noll, S., Horst, H., Kausch, W.,Kimeswenger, S., Barden, M., Szyszka, C., Jones, A. M., Gal-lenne, A., Vinther, J., Ballester, P., Taylor, J., 2015, A&A,576, 77.Tielens A. G. G. M., Hony S., van Kerckhoven C., Peeters E.,1999, ESASP, 427, 579Tody, D., 1986, SPIE, 627, 733.van der Zwet G. P., Allamandola L. J., 1985, A&A, 146, 76Vollmann, K., Eversberg, T. 2006, Astron. Nachr., 327, 862.Vos D. A. I., Cox N. L. J., Kaper L., Spaans M., Ehrenfreund P.,2011, A&A, 533, A129This paper has been typeset from a TEX/L A TEX file prepared bythe author.MNRAS000