Spectroscopic Atlas for Amateur Astronomers 1 Spectroscopic Atlas for Amateur Astronomers A Guide to the Stellar Spectral Classes Richard Walker Version 3.0 03/2012 Spectroscopic Atlas for Amateur Astronomers 2 Table of Contents 1 Introduction ....................................................................................................................... 7 2 Selection, Preparation and Presentation of the Spectra ........................................... 9 3 Terms, Definitions and Abbreviations........................................................................ 12 4 The Fraunhofer Lines .................................................................................................... 14 5 Overview and Characteristics of Stellar Spectral Classes ..................................... 15 6 Appearance of Elements and Molecules in the Spectra......................................... 20 7 Spectral Class O ............................................................................................................ 21 8 Wolf Rayet Stars ............................................................................................................ 28 9 Spectral Class B............................................................................................................. 32 10 LBV Stars......................................................................................................................... 39 11 Be Stars ........................................................................................................................... 43 12 Be Shell Stars ................................................................................................................. 46 13 Herbig Ae/Be and T Tauri Protostars ......................................................................... 48 14 Spectral Class A............................................................................................................. 53 15 Spectral Class F ............................................................................................................. 60 16 Spectral Class G ............................................................................................................ 65 17 Spectral Class K............................................................................................................. 70 18 Spectral Class M............................................................................................................ 78 19 Spectral Sequence on the AGB ................................................................................... 81 20 M(e) Stars on the AGB .................................................................................................. 83 21 Spectral Class S on the AGB ....................................................................................... 85 22 Carbon Stars on the AGB ............................................................................................. 91 23 Spectra of Extragalactic Objects ................................................................................ 97 24 Spectra of Emission Nebulae ....................................................................................100 25 Reflexion Spectra of Solar System Bodies..............................................................116 26 Telluric Molecular Absorption ...................................................................................121 27 The Night Sky Spectrum.............................................................................................123 28 Terrestrial Lightsources .............................................................................................125 29 Spectral Classes and (cid:1822)(cid:1819)(cid:1809)(cid:1814)(cid:1809) – Values of Important Stars ......................................135 30 Required Ionisation Energies for the Individual Elements ...................................138 31 Bright Planetary Nebulae sorted by Excitation Classes........................................139 32 Terminology of the Spectroscopic Wavelenght Domains ....................................140 33 Appendix .......................................................................................................................141 34 Bibliography and Internet ..........................................................................................147 Spectroscopic Atlas for Amateur Astronomers 3 Change log of the atlas versions Version 1.3: Corrections of some labeling-/typing errors and temperature of the O-class stars, some English language corrections. Version 1.4: Corrections of some labeling and typing errors. sect. 1: correction of citation error regarding Spectroweb. Version 1.5: Sect. 7.4, (8) WR133, expansion velocity of the stellar wind, EW value replaced by FWHM value. sect. 8.3 (9.3): chart of the theoretical B-class continuum, correction of the wrongly labeled Ca II line (thanks to Robin Leadbeater!). Version 2.0: New: Sect. 12: Be Shell stars and some adjustments to sect. 11. New: Sect. 18 – 21: Spectral sequence of the Mira Variables on the AGB with the classes M(e), S- and carbon stars C. General revision of the title structure, the table of contents and the sect. 31 “Bibliography and Internet”. Version 3.0: Sect. 3.5: Additional Abbreviations, Sect. 3.7: New: Formula for metallicity. Sect. 7: General revision and expansion to the early O Class. Sec. 10: P Cygni, higher resolved profile sections (900L grating) in the green and blue range of the spectrum. Sect. 13: New: Herbig Ae/Be and T Tauri Protostars Sect. 21: Correction of the wrongly labelled BD Camelopardalis: new HD 22649 Sect. 24: General revision of the section, new: determination of the excitation classes of emission nebulae and table 85: SNR M1/ NGC 1952. Sect. 25: Reflexion spectra: former Tables 85 and 86 ==> new Tables 90 and 91 New: Comet C/2009 P1 Garradd, Table 94 Sect. 27: New: Nigth Sky Spectrum Table 96 Sect. 28: Table 106, Calibration spectrum of the glow starter OSRAM ST 111 Sect. 31: List of bright Planetary Nebulae sorted by excitation classes. All Parameter tables of the individual spectral classes: Correction of the wrong (german) abbreviation “Mrd” by bn (thanks to Dave Dowhos!). Spectroscopic Atlas for Amateur Astronomers 4 Directory of Tables Wavelength Gra- Table Page Topic Objects domain ting 01 16 Overview on the spectral classes Spectral strips of 3950 – 6690 200L various stars 02 17 Overview on the spectral classes Intensity profiles of 3950 – 6690 200L various stars 1 25 Spectral features of the Alnitak ζ Ori 3920 – 6710 200L late O-class Mintaka δ Ori 2 26 Detailed spectrum of a late Alnitak ζ Ori 3950 – 4750 900L O-class star 5740 – 6700 3 27 Spectral features of the early to Θ1 Ori C 3800 – 6700 200L middle O-class 68 Cygni 5 31 Wolf-Rayet WR stars, final stage of WR 133 3850 – 7250 200L the O- and early B-Class WR 140 10 34 Development of spectral features Alnilam ε Ori 3900 – 6700 200L within the B-class Gienah Corvi γ CrV 11 36 Effect of the luminosity on spectra Regulus α Leo 3920 – 4750 900L of the late B-class Rigel β Ori φ Sagittarii 12 38 Detailed spectrum of an early Spica α Vir 3800 – 6750 200L B-class star 3900 – 4750 900L 4800 – 5100 5700 – 6050 6450 – 6600 13 41 LBV star, early B-class P Cygni, 34 Cyg 3900 – 6950 200L P Cygni profiles, 6000 – 6800 900L 13A 42 Detailed spectrum LBV star, early P Cygni 34 Cyg 3850 – 4650 900L B-class, P Cygni profiles 4700 – 6050 14 44 Be- star, early B-class Dschubba δ Sco 3650 – 7000 200L 4820 – 4940 900L 6500 – 6700 6670 – 6690 14A 45 Be-star, early B-Class Tsih γ Cassiopeiae 3970 – 6750 200L 15 47 Be Shell Star, comparison to an ζ Tauri 3800 – 6800 200L “ordinary” Be Star Dschubba δ Sco 17 51 Herbig Ae/Be Proto star R Monocerotis 3900 – 7200 200L NGC 2261 18 52 T Tauri Protostar T Tauri 3900 – 7000 200L 20 55 Development of spectral features Castor α Gem 3900 – 6800 200L within the A-class Altair α Aql 21 56 Detailed spectrum of an early Sirius A α CMa 3900 – 6700 200L A-class star 3900 – 4700 900L 4780 – 5400 22 58 Effects of the luminosity on spectra Vega α Lyr 3900 – 4700 900L of the early A-class Ruchbah δ Cas Deneb α Cyg Spectroscopic Atlas for Amateur Astronomers 5 Wavelength Gra- Table Page Topic Objects domain ting 23 59 Metallicity: Vega vs. Sirius Vega α Lyr 3920 – 4700 900L Sirius A α CMa 30 62 Development of spectral features Adhafera ζ Leo 3830 – 6700 200L within the F-class Procyon α CMi 31 64 Effects of the luminosity on spectra Porrima γ Vir 3920 – 4750 900L of the early F-class Caph β Cas Mirfak α Per 40 67 Development of spectral Muphrid η Boo 3800 – 6600 200L features within the G-class Vindemiatrix ε Vir 41 68 Detailed spectrum of an early G- Sun 3800 – 7200 200L class star 3900 – 4800 900L 42 69 Detailed spectrum of an early G- Sun 4700 – 5700 900L class star 5650 – 6700 50 73 Development of spectral features Arcturus α Boo 3900 – 6800 200L within the K-class Alterf λ Leo 51 74 Detailed spectrum of an early Pollux α Gem 3900 – 6800 200L K-class star 3800 – 4800 900L 52 76 Effects of the luminosity on spectra Alsciaukat α Lyncis 4000 – 4900 900L of the late K-class 61 Cygni B 53 77 Detailed spectrum of a later Aldebaran α Tau 5150 – 5900 900L K-class star 5850 – 6700 60 80 Development of spectral features Antares α Sco 3900 – 7200 200L within the M-class Ras Algethi α Her 63 84 Mira Variable M(e), comparison to Mira o Ceti 3900 – 7200 200L a late classified M-Star Ras Algethi α Her 65 87 Extreme S-Class star, comparison R Cygni 4100 – 7300 200L to a Mira Variable M(e) Mira o Ceti 66 89 Development of spectral features Omikron1 Orionis 4100 – 7300 200L within the S-Class Chi Cygni R Cygni 66A 90 Comparison of an „intrinsic“and BD Camelopardalis 4300 – 7200 200L „extrinsic“ S-Class star HR Pegasi 67 95 Comparison of differently WZ Cassiopeiae 4600 – 7300 200L classified carbon stars Z Piscium W Orionis 67A 96 Merril Sanford Bands, details at a W Orionis 4730 – 5400 900L higher resolved spectrum 70 99 Comparison of Spectra: Spiral Andromeda M31 3900 – 6700 200L galaxy vs. star of the late G-class Vindemiatrix ε Vir 80 108 Emission Nebula: H ll Region M42 3800 – 7300 200L 80A 109 Intensity profiles of Hβ and [OIII] M42 n.a. 200L (λ5007) in the central area of M42 81 110 Emission Nebula: Planetary Nebula IC418 Spirograph Neb 4100 – 7100 200L 82 111 Emission Nebula: Planetary Nebula NGC6210 Turtle Neb. 3850 – 6600 200L 83 112 Emission Nebula: Planetary Nebula NGC7009 Saturn Neb 3800 – 6700 200L Spectroscopic Atlas for Amateur Astronomers 6 Wavelength Gra- Table Page Topic Objects domain ting 84 113 Emission Nebula: Planetary Nebula M57 Ring Nebula 4600 – 6800 200L 84A 114 Intensity profiles of [O III] and [N II] M57 Ring Nebula n.a. 200L in the longitudinal axis of M57 85 115 Emission Nebula: SNR M1 / NGC 1952 4600 – 6800 200L 90 118 Reflexion spectra of Mars, Venus 4300 – 7800 200L solar system bodies 91 119 Reflexion spectra of solar Jupiter, Saturn 4400 – 7800 200L system bodies 94 120 Comet spectrum C/2009 P1 Garradd 3800 – 6400 200L 95 122 Spectral absorptions due to the Earth’s atmosphere 6800 – 7800 900L earth’s atmosphere 96 124 Night sky spectrum Light Pollution 4000 – 7400 200L Airglow 101 127 Gas discharge lamp Neon glow lamp 5800 – 8100 900L 102 128 Gas discharge lamp ESL Osram Sunlux 3900 – 6400 200L 103 129 Gas discharge lamp Xenon strobe tube 3900 – 8100 200L 900L 104 130 Gas discharge lamp High pressure 4700 – 7250 200L sodium vapor lamp 105 131 Gas discharge lamp High power xenon 4900 – 6900 200L lamp 106 132 Gas discharge lamp Glow Starter OSRAM 4000 – 7700 200L ST 111 3900 – 4800 900L 110 134 Swan Bands / Hydrocarbon gas Butane-gas-torch 3800 – 6400 200L flames: Comparison of superposed Comet Hyakutake spectra WZ Cassiopeiae Spectroscopic Atlas for Amateur Astronomers 7 1 Introduction Probably most amateur astronomers have a common experience at the beginning of their spectrographic “career”. Full of expectations they test the new device for the first time and look with great pride at the first self recorded spectrum. This first enthusiasm then quickly gives way to the perplexity, how these numerous lines shall now be identified and even analyzed. Fortunately, some recommendable introductory literature, e.g. [1], for amateurs to the interpretation of stellar spectra is available in English. However, a real “Spectral Atlas”, which systematically covers the classes O, B, A, F, G, K, M, by commenting most of the lines, visible in a medium resolved spectral profile was still missing. In the Internet we find many professional papers, mostly focused on small sections of a spectral profile. Nevertheless, they often contribute valuable puzzle pieces to the exciting “detective work” of line identification. Very well documented is the spectrum of the Sun (G2 V), where almost all of the lines are identified and commented. Here, at least two easily readable and freely downloadable atlases are available on the net [80], [81]. These sources can even help, with the necessary precautions, to identify spectral lines of adjacent spectral classes. Spectral atlases from professional sources, published on the internet, and covering all important spectral classes can rarely be found. Amazingly, even in such professional papers, usually only a few intense lines, mostly in very lowly resolved spectra, are commented. However, the focus of many practical applications, even in professional works, is mostly reduced to relatively few, highly intense and isolated lines, e.g. for determination of the rotation speed or the metal abundance. In such publications the ultraviolet, the blue- and red- to far infrared part of the spectra are preferably treated - the sections "green" and "yellow" however only rarely. In Appendix 33.3 some excerpts from several historical as well as state of the art spectral atlases are shown. Probably the best known, and highly important one, even for the history of science, is the out of print standard work "An atlas of stellar spectra, with an outline of spectral classification" by Morgan, Keenan and Kellman [50]. This seminal work from 1943 can now be downloaded from the Internet but is limited to the short-wavelength part of the visible spectrum, reflecting the state of technology in the 1940’s. It presents photographic spectral strips, commented with handwritten notes! Made in a similar style is the “Revised MK Spectral Atlas for Stars Earlier Than the Sun” from 1978 by Morgan, Abt, and Tapscott [51]. Some spectral atlases were available in the past but are out of print today. The most comprehensive “reference work” is surely “Bonner Spektralatlas”, the title of the English edition: “Atlas for Objective Prism Spectra” by Waltraut C. Seitter, 1975 [5]). This atlas, covering the spectral types from O5 to M2, is out of print today, but recently available as download [5]! Still based on photographic spectral strips it’s the only reviewed work, documenting most of the lines between about λ 3400 – 5000, which are visible with low to medium resolving spectrographs. In the green-red range this atlas partly shows quite large gaps. It’s striking, that despite this work undoubtedly being outdated, it has never been adapted to today's requirements, but is still referenced in various current papers. Written in a similar style, but much less detailed, are also the atlases [6], [7] and [8]. A stroke of luck that Martin Brunold [705] collected all these books and has kindly provided me with a review (see Appendix 33.3). Nowadays calibrated and intensity-normalized profiles, plotted against the wavelength are required to meet the current standards. Fairly recent (2000) is "A Digital Spectral Classification Atlas" by R.O. Gray [52]. However it’s limited to a lowly resolved, short-wave part of the visible spectrum. Furthermore several atlases exist, focusing on specific Spectroscopic Atlas for Amateur Astronomers 8 wavelength ranges. Some of them are listed in the bibliography. For some of the brighter stars monographs are available, with fully commented spectra of e.g. Sirius, Procyon and Aldebaran. Unfortunately, such papers are mostly available as “abstracts” only. The download prices for the full versions are usually pretty high. Very useful is Spectroweb [59], to find on the home page of Dr. Alex Lobel, Royal Observatory of Belgium. It’s an interactive internet platform providing highly resolved and almost completely commented profiles of some bright stars, belonging to the middle and late spectral classes F, G, K, M. Due to the enormous line density it’s, particularly for beginners, very difficult to make the link to the highly blended lines of their own, in most cases lowly resolved profiles. Here at last it becomes clear, why spectral atlases need to be created for specific degrees of resolution. Recently the pocket guide “A Spectroscopic Atlas of Bright Stars” [9] was published. It’s essentially a collection of non-normalized pseudo-continua of bright stars, showing a few, intense spectral lines. These have been recorded by Jack Martin under the London night sky on conventional (chemical) film with a slitless transmission grating spectrograph. This huge diversity of information explains why different sources have been used for the line identification. References for each and every line would not be feasible. However, for each spectral class, the mainly used sources are referenced. An alternative option would be to compare the spectra with synthetically generated profiles, based on models of stellar atmospheres. This allows for example the software "Spectrum" from R.O. Gray, which runs on Linux with a command-oriented interface. The installation for non Linux users, and also the operation are demanding. So this is probably a realistic option for very few amateurs only. This atlas is primarily intended to be used as a tool for the line identification – as a supplement to my " Spectroscopic Applications for Amateur Astronomers” [30] and the practical part," The processing and analysis of spectral profiles ..." [31]. In these documents also detailed information on the classification system of the spectra and the Hertzsprung- Russel diagram (HRD) are included (currently available in German only). Knowledge of these topics are presupposed here and therefore treated briefly only. From the beginning it was clear, that this atlas would not be reduced to an isolated collection of some labeled spectral tables. Therefore, each spectral class is presented with their main characteristics and typical features. The atlas will be continuously updated if new information or characteristic spectra are available. This intended future expansion was considered in the numbering system of the tables. After such updates, direct or "hot links" on the file will necessarily lead to the void. Therefore, I strongly recommend linking to the atlas only to: http://www.ursusmajor.ch/astrospektroskopie/richard-walkers-page/index.html Many thanks to Martin Huwiler, Urs Flükiger and Dr. Helen Wider for proofing the German edition, and Urs again for kindly providing his homepage for downloading! Richard Walker, CH 8911 Rifferswil © [email protected] About the author: Born in 1951, I began to show interest in astronomy when I was about 12 years old – after my grandparents took me to a public astronomical observatory in Zurich one fine night, which led to my first glance at Saturn. Later I started my own observations with one of the typical department store telescopes, followed by a self constructed 6 inch Newton reflector. For the last 10 years, my interest has increasingly focused on theories about astrophysics and stellar astronomy, somewhat later also on the indispensable key to these topics – the spectroscopy. My professional background is civil engineering. For a long time I worked for a Swiss engineering company, specialized in planning powerplants, dams and tunnels. For the last 12 years of my professional career I was involved in risk assessments and worldwide inspections for a Swiss Reinsurance Company – for so-called “large risks” like dams, powerplants, large construction sites, highrise buildings etc. For a short time now I have been enjoying my early retirement. My further interests are archeoastronomy, windsurfing, paragliding, diving and mountain climbing – and sometimes I am also busy with my granddaughters. Spectroscopic Atlas for Amateur Astronomers 9 2 Selection, Preparation and Presentation of the Spectra 2.1 Selection of spectra Main criteria for the selection of the spectra have been the documentation of the spectral characteristics of the classes O, B, A, F, G, K, M, and further the demonstration of certain effects e.g. due to the different luminosity classes. The consideration of bright “common knowledge stars” was of secondary importance. In principle, every spectral class, including some extraordinary star types, is presented at least with an “early” and a “late” representative spectrum to show the development of characteristic features in the profile. Commented in separate chapters are spectra of emission nebulae, composite spectra of extragalactic objects, reflection spectra of solar system bodies, absorption bands generated by the earth’s atmosphere and some profiles of terrestrial light sources. 2.2 Recording and Resolution of the Spectra The spectra have been recorded with the DADOS spectrograph [603], equipped with reflection gratings of 200- or 900 Lines/mm. Unless otherwise noted, the recording was made through the 8 inch Schmidt-Cassegrain Celestron C8, the 25 μm slit of the DADOS spectrograph, and finally with the monochrome camera Meade DSI III Pro. The spectra on the following Tables: 5 (WR133), 70 and 84 have been recorded together with Martin Huwiler through the CEDES 36 inch telescope of the Mirasteilas observatory in Falera (see Appendix 33.4). The processing of the spectra with Vspec indicates the following dispersions: 900L grating about 0.65 Å/pixel and 200L grating some 2.55 Å/pixel. Data of the Sony Chip ICX285AL: 1.4 Mega-pixel, 2/3" Monochrome CCD, Pixelsize 6.45μm x 6.45μm [606]. For longer exposure times a dark frame was subtracted. The processing of a "flatfield" was omitted. The R-value, defines the resolution of the spectrograph. According to the DADOS Manual it corresponds to (cid:1844) =647 at 6160 Å and (cid:1844) = 3,000 at 5610 Å. My own (cid:2870)(cid:2868)(cid:2868)(cid:3013) (cid:2877)(cid:2868)(cid:2868)(cid:3013) measurements in this wavelength domain, using several averaged FWHM values of Neon emissionlines, yielded R-values in the order of (cid:1844) =900, respectively (cid:1844) = 4,000. (cid:2870)(cid:2868)(cid:2868)(cid:3013) (cid:2877)(cid:2868)(cid:2868)(cid:3013) Generally these resolutions have proven to be ideal for the presentation of the stellar spectral sequence. Substantially higher R-values would be soon comparable to reading a newspaper with the microscope. Therefore it’s not surprising that Gray/Corbally [2] denote profiles with a resolution of ~ 3 Å as "classification resolution spectra”. Even the professional astronomy uses spectrographs with lower resolutions for certain tasks. 2.3 The Processing of the Spectra The monochrome fits images have been processed with the standard procedure of IRIS [550]. In most cases, about 5–7 spectral profiles have been stacked, to achieve noise reduction. The generating and analizing of the final profile was performed with Vspec [551]. The procedure is described in detail in the IRIS and Vspec manuals. For the software functions, applied to process the spectra in this atlas, a step by step description is available in [31] (German only). With one exception (M31) in all broadband spectra (200L) the pseudo-continuum was removed. The profiles have been rectified – divided by the course of their own continuum. Thus, the intensity of the spectrallines becomes visually comparable over the entire range and further it results a flat (rectified), space-saving and easily readable profile. The relative depth or height of a line in relation to the height of the according continuum level is crucial for the intensity comparison. The profile of a pseudo-continuum presents strong lines at the blue or red end of the spectrum optically as relatively too weak and vice versa, weak lines in the middle part, -as too high. Spectroscopic Atlas for Amateur Astronomers 10 But just this reasonable correction may confuse the beginners, if they try to find lines of their uncorrected pseudo-continuum in the flat, rectified atlas profile. The following graphic illustrates this effect with the superposition of the uncorrected pseudo-continuum (blue) and the atlas profile of the solar spectrum (red). In rare cases with an increased noise level (by faint objects), the profile was sometimes smoothed, using filters such as the Vspec MMSE filter (minimum mean squared error).The goal of this process was here exclusively to improve the readability of the documented lines. A reduction of the telluric H O/O absorption in the yellow/red range of the spectral 2 2 profile was omitted. Therefore the line documentation in this domain was restrained accordingly. 2.4 Calibration of the Wavelength Most of the spectra have been calibrated relatively, based on known lines and not absolutely with the calibration lamp. This prevents that the profile, as a result of possible high radial velocities (Doppler Effect), is shifted on the wavelength axis. The focus here is the presentation of the spectral class and not the documentation of the individual star. Only for spectra of late spectral classes and extraordinary stars the calibration was carried out with the lamp. The unit for the wavelength is here generally Angstrom [Å]. These values are shown according to convention [2] with the prefix [λ]. E.g. 5000 Å correspond to λ 5000. 2.5 Normalization of the Intensity An absolute flux calibration of the intensity profile would be very time consuming and is not necessary for the purpose of this atlas. But the continuum intensity of the rectified profiles was always normalized to unity, so that the medium continuum level yields about (cid:1835) =1. An (cid:3030) absorption line is usually saturated to the maximum if it reaches from the continuum level down to the wavelength axis (cid:1835) =0. At the applied low to (cid:3030) medium resolutions, in stellar spectra this can rarely, if at all, be seen. Hence, for space saving reasons, in most of the cases, not the entire range of saturation is presented. Instead the corresponding level on the wavelength axis is indicated with amounts mainly in a range of 0.3–0.6. In case of montages showing several spectra in one chart, the single profiles have been normalized to unity, based on the same continuum section, to enable a rough comparison of the line intensities. 2.6 Line Identification For the sometimes complex identification process of spectral lines nearly all referenced information sources have been used. Some intensive, but in all the listed sources not