Intro

Along with capturing the above image of the region around Wolf-Rayet 134, which is the brightest star in the centre, I also took the star's spectrum: mousing over the image or selecting Revision "E" shows the spectrum overlaid on the image; that spectrum plot can be viewed separately by selecting Revision "F" (the difference between the plots in Revisions "F" and "G" is explained below). The image was taken with Ha and OIII filters (both 3-nm bandpass), along with RGB for the stars. The narrowband component of the image was processed using an "HOO" colour map, with hydrogen in red and oxygen in blue. Details of the data acquisition and an analysis of the spectrum are given further down.

Wolf-Rayet (WR) stars are highly evolved states of O-type progenitors, with stellar winds that are thick and fast, producing very strong emission lines, and shielding the star's photosphere such that few, if any, of the usual stellar absorption lines are visible (Wikipedia has a good overview of WR stars, and an authoritative review can be found here). The winds, with typical speeds of 1000s of km/sec, blow out the star's hydrogen-rich outer layer, and expose a region where the elements carbon, nitrogen, and oxygen have been dredged-up from the core. Correspondingly, there are three broad classes of WR stars: WC, WN, and WO, with strong emission lines coming mainly from the associated heavy element, along with helium, while hydrogen contributes weakly, if at all. WR winds eject on the order of 1 solar mass every 100,000 years, which makes this a very short phase in stellar evolution, of about 0.5-million years in duration; consequently, WR stars are rare, with only about five hundred discovered in the Milky Way so far. 

Some fun facts about the spectra of WR stars: many of them can be resolved with amateur equipment, and they can easily be used to estimate the wind speed and the WR-type (several "amateur" WR spectra are displayed in this spectral atlas). WR 134 was one of the three stars initially discovered by French astronomers Charles Wolf and Georges Rayet in 1867, by using a visual spectroscope!   

Image

My image of Wolf-Rayet 134 is the result of nearly 31 hours of integration, with about 12 hours in Ha and 16 hours in OIII, and about 2.5 hours in RGB for the stars. In addition, a further 21 hours of exposures were acquired but rejected based on seeing and sky background cuts. The image spans 30' on a side, at a plate scale of about 0.62"/pixel, and the data were gathered over the course of 17 nights, from the beginning of August through the end of September of 2022. 

Spectrum

The spectrum was acquired on August 1st 2022, using a Shelyak LISA spectrograph equipped with a monochrome CCD camera for greater sensitivity (native resolution about 3.6 Angstroms per pixel). The brightest emission line is so intense that an exposure time of only 5-seconds was used to avoid saturation, and 60 frames were stacked to reduce small-scale noise (just 5-minutes total to get a very clean spectrum!). Reduction and calibration of the raw data were done using Shelyak's Demetra software package, along with Christian Buil's ISIS package for some further processing (for an excellent guide to spectroscopy for amateurs, see this companion to the aforementioned spectral atlas).

To analyze the composition of a star, one usually divides out the contribution of its continuum radiation, which generates a so-called "rectified" spectrum, and is what I've shown in the image overlay and Revision "F". The continuum can be estimated from the spectrum itself by fitting, after cutting out the emission lines (ISIS has a handy tool to do both). The observed continuum of WR stars comes from the wind, since the photosphere is shielded. By way of comparison, the spectrum including the continuum is shown in Revision "G" (in arbitrary units), which exhibits a gradual rise towards the blue end. 

I've identified the primary atomic state that is responsible for each of the most prominent lines, bearing in mind that these are often a blend of transitions of similar wavelength from more than one ionized state, and in some cases also from different elements. The spectrum clearly shows that WR 134 is in the "WN" category, because of the multiple nitrogen lines (the C IV line at about 5808Å is present in most WN stars), and its subtype can also be pegged from the plot at "WN6", one of about ten numbered subtypes (several classification schemes are in use, and the number of subtypes varies). Only two absorption lines are seen in the plot, both due to our own atmosphere, and are labelled "telluric". A useful collection of rectified spectrum plots for many WN-type stars, including WR 134, can be found here, and I also found this older paper on the spectrum to be helpful.

Wind speed estimate

The spectral lines are strongly broadened, mostly due to Doppler shifts produced by motion in the fast wind, along with blending of more than one transition of similar wavelength in some lines. The width of most of the lines is very similar, which is indicative of Doppler broadening. The wind speed v_wind can be estimated from the FWHM Δλ of an emission line, relative to its peak wavelength λmax, according to v_wind=cΔλ/λmax, where c is the speed of light. As a typical example, I estimate from my plot that the He II line at λmax=5411Å has a width Δλ≈40Å, yielding v_wind≈2200 km/sec, which is within 10% of the accepted value of 2050 km/sec, as quoted in this exhaustive compendium of WR data!

He II spectral lines and hydrogen

A very interesting aspect of WR stars has to do with the fact that the spectrum produced by singly-ionized helium (He II) is closely related to the hydrogen spectrum, since they are both one-electron atoms. In particular, a subset of He II transitions known as the Pickering series alternates in wavelength between lines that are almost identical in position to the hydrogen Balmer series, and intermediate lines that have no analogue in hydrogen. This alternation leads to a clever and simple qualitative method to assess whether or not the WR wind contains an appreciable amount of hydrogen; to close-out this too-long post, I'll try to summarize this approach (drawing in part from this classic textbook on stellar classification, and this authoritative paper).

The correspondence between the Pickering and Balmer series arises because the energy of an atomic state is proportional to the square of the nuclear charge number, which introduces a factor of 4 in the energies of He II relative to hydrogen, and is also proportional to the inverse-square of the principal quantum number "n" of the state. The hydrogen Balmer emission lines are produced by transitions to a final state with n=2: the Hα line is a 3-2 transition, Hβ is 4-2, Hγ is 5-2, and so on. It follows that the Pickering series in He II corresponds to transitions to that atom's n=4 level, and of those, the transitions from states with *even* higher quantum numbers have wavelengths that are almost identical to the Balmer series (they are shifted to slightly smaller wavelengths — higher energies — because the electron "orbits" slightly closer to a heavier nucleus). So the He II transitions 6-4, 8-4, 10-4, and so on, correspond to the Balmer lines, and they alternate with transitions 7-4, 9-4, 11-4, etc. that have no analogue in hydrogen (the amateur guide to spectroscopy includes a concise overview of the quantum mechanics of stellar spectra). 

All of the labelled He II peaks in the plot are from the Pickering series, except for the most prominent one (centred at about 4686Å), which is a 4-3 transition, and the blend at about 6680Å, which gets a contribution from the 13-5 He II transition. The line centred at about 6561Å is analogous to the Balmer Hα line, while the line to the left (centred at about 5412Å) has no analogue in hydrogen, the one after that (at about 4859Å) is analogous to Balmer Hβ, and so on. The He II line on the far left, at about 4100Å, is analogous to Hδ, but turns out to be strongly blended with Ne III transitions.

If hydrogen is present, the spectrum of a WR star near the Balmer wavelengths will be blends of He II and hydrogen transitions, and the intensities observed across the Pickering series wavelengths will tend to oscillate in strength: this is because the intensity of He II Pickering transitions decreases steadily (almost linearly) with decreasing wavelength across the optical range, so that if hydrogen is present, it will "boost" every other line above the He II Pickering trend. There is no sign of such an oscillation in the observed spectrum of WR 134 (the Hδ-analogue line on the far left is not considered, due to its strong mixing with NIII), while the intensity decrement of about 1.3 is typical of the He II transitions (by contrast, the Balmer decrement from Hα to Hβ is a factor of about 3). A sophisticated model of WN atmospheres confirms that there is no detectable hydrogen in the wind of WR 134. 

Oy, was that TMI?!

For what it's worth, if mistakes are spotted in this rambling narrative,  please let me know so that I can correct them!